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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
d29bef02-6a24-47f2-9a5a-738941d7a19b | modelling-sars-cov-2-coevolution-with-genetic | 2102.12365 | null | https://arxiv.org/abs/2102.12365v1 | https://arxiv.org/pdf/2102.12365v1.pdf | Modelling SARS-CoV-2 coevolution with genetic algorithms | At the end of 2020, policy responses to the SARS-CoV-2 outbreak have been shaken by the emergence of virus variants, impacting public health and policy measures worldwide. The emergence of these strains suspected to be more contagious, more severe, or even resistant to antibodies and vaccines, seem to have taken by surprise health services and policymakers, struggling to adapt to the new variants constraints. Anticipating the emergence of these mutations to plan ahead adequate policies, and understanding how human behaviors may affect the evolution of viruses by coevolution, are key challenges. In this article, we propose coevolution with genetic algorithms (GAs) as a credible approach to model this relationship, highlighting its implications, potential and challenges. Because of their qualities of exploration of large spaces of possible solutions, capacity to generate novelty, and natural genetic focus, GAs are relevant for this issue. We present a dual GA model in which both viruses aiming for survival and policy measures aiming at minimising infection rates in the population, competitively evolve. This artificial coevolution system may offer us a laboratory to "debug" our current policy measures, identify the weaknesses of our current strategies, and anticipate the evolution of the virus to plan ahead relevant policies. It also constitutes a decisive opportunity to develop new genetic algorithms capable of simulating much more complex objects. We highlight some structural innovations for GAs for that virus evolution context that may carry promising developments in evolutionary computation, artificial life and AI. | ['Aymeric Vie'] | 2021-02-24 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 3.08959305e-01 -1.20147690e-01 3.15463156e-01 2.88834274e-01
3.57673138e-01 -6.19994819e-01 5.81602156e-01 1.81686759e-01
-4.99268711e-01 1.15176427e+00 -1.27383932e-01 -4.79455709e-01
-3.15975517e-01 -7.61709213e-01 -4.09361482e-01 -1.06524932e+00
-5.45058012e-01 8.14931750e-01 -2.92788208e-01 -7.97604740e-01
1.92520320e-01 7.20969677e-01 -1.42807293e+00 -1.88459873e-01
1.04500842e+00 2.02598929e-01 4.15336549e-01 6.20652795e-01
2.16240194e-02 1.04528747e-01 -7.75928617e-01 -1.81582808e-01
3.13690901e-01 -6.72336817e-01 -2.02819064e-01 -3.61350089e-01
-6.31249487e-01 1.62801936e-01 4.12439197e-01 1.01068628e+00
5.43034554e-01 -9.98627916e-02 5.89953184e-01 -9.49090481e-01
-5.53638160e-01 3.69933218e-01 -3.76165748e-01 3.03442359e-01
1.00568101e-01 7.78243482e-01 3.91869813e-01 -2.13036001e-01
9.16626155e-01 1.11243141e+00 6.80939674e-01 9.29229200e-01
-1.04864216e+00 -2.02068731e-01 -6.45602122e-02 1.58353508e-01
-1.10000086e+00 -1.41220223e-02 5.57169259e-01 -5.59437096e-01
1.08728969e+00 7.90098310e-01 1.26232743e+00 1.26881337e+00
6.84602976e-01 1.67584896e-01 9.84932184e-01 -2.61651814e-01
4.57035512e-01 1.10782519e-01 -6.52109861e-01 5.65552771e-01
8.24091256e-01 5.65230191e-01 1.06517404e-01 -7.23908484e-01
4.11572784e-01 1.21948801e-01 -6.26208782e-01 -1.68932721e-01
-8.64106536e-01 9.73172784e-01 2.10739806e-01 6.66359484e-01
-8.04990888e-01 -1.59748331e-01 2.14436486e-01 5.02278730e-02
1.87622517e-01 1.24801672e+00 -6.66221440e-01 -1.34659901e-01
-3.33372831e-01 3.70650232e-01 6.01416647e-01 2.34163627e-02
3.36131662e-01 2.42595702e-01 4.48711812e-02 3.97676110e-01
1.61645338e-01 7.01971710e-01 2.85840541e-01 -4.33366269e-01
-2.79111117e-01 6.74082100e-01 2.03156173e-01 -1.04463744e+00
-6.94248855e-01 -7.23764300e-01 -9.88761425e-01 1.39207587e-01
1.85198821e-02 -5.01283288e-01 -7.17303276e-01 1.73597431e+00
5.82129359e-01 1.26804397e-01 3.25944275e-02 7.35895991e-01
-2.85109222e-01 8.74337554e-01 -4.83265631e-02 -8.64754915e-01
1.34459162e+00 -3.91302258e-01 -6.37267172e-01 2.78375950e-02
4.32063460e-01 -4.77929264e-01 6.02108479e-01 1.61208734e-01
-8.26955974e-01 -1.93179384e-01 -8.84652674e-01 1.23864102e+00
-5.12821138e-01 -5.41662097e-01 5.56318402e-01 1.00769663e+00
-1.22988641e+00 4.93490100e-01 -7.42617786e-01 -7.53150761e-01
1.28239721e-01 3.02462697e-01 3.46265733e-01 3.66056114e-01
-1.37149775e+00 1.08776379e+00 3.32854360e-01 5.11292756e-01
-5.91954470e-01 -6.03630662e-01 -2.84427404e-01 2.36087572e-02
4.94788706e-01 -1.16138136e+00 2.92922676e-01 -9.21420753e-01
-1.25204945e+00 5.48551559e-01 1.93440184e-01 -6.82918727e-01
5.94455302e-01 3.22636187e-01 -5.75881481e-01 -3.52397203e-01
-5.06804824e-01 4.17046905e-01 6.01438642e-01 -1.18583572e+00
-3.68190229e-01 -4.75929052e-01 -3.41409475e-01 -1.15729833e-03
-4.14677523e-02 2.67549366e-01 3.30560684e-01 -7.98242152e-01
-7.75129974e-01 -1.23486090e+00 -8.35613966e-01 -4.33794588e-01
1.15345336e-01 -1.77185498e-02 6.04629278e-01 -2.66478181e-01
1.11625218e+00 -1.66952205e+00 4.17815655e-01 3.67836326e-01
6.48566559e-02 8.50363672e-01 -2.59760380e-01 7.51316607e-01
3.22403103e-01 3.77736568e-01 -4.87502635e-01 5.55218339e-01
-4.71782774e-01 2.62768328e-01 -7.33910352e-02 2.94966549e-01
4.57747698e-01 9.66091037e-01 -1.04545581e+00 2.22745627e-01
-2.32059970e-01 6.42030716e-01 -6.14491880e-01 3.30455959e-01
-6.33227885e-01 7.44824231e-01 -5.44737518e-01 5.46557784e-01
6.10538602e-01 -2.53525507e-02 7.27609515e-01 3.47245842e-01
-3.49972844e-01 -5.02193928e-01 -4.39050078e-01 7.19938874e-01
1.96186565e-02 2.99596101e-01 1.76130459e-01 -8.25016558e-01
9.94621873e-01 1.78031772e-01 6.13252938e-01 -7.81478763e-01
3.78958404e-01 2.35131115e-01 5.69219470e-01 -6.84418738e-01
1.70308605e-01 -8.03753436e-02 2.24342689e-01 4.34810311e-01
-5.99744737e-01 -1.75804377e-01 6.26252964e-02 -3.79036933e-01
8.93326938e-01 -1.02431059e-01 2.43693396e-01 -3.92358571e-01
6.96480215e-01 2.58800179e-01 6.86072409e-01 5.58315575e-01
-3.59221518e-01 -4.99858819e-02 3.70839536e-01 -8.14929187e-01
-1.22911704e+00 -6.01857960e-01 1.21223971e-01 4.41199541e-01
-1.81704625e-01 1.87704563e-02 -7.01379657e-01 -2.57461131e-01
-1.99951559e-01 6.85254693e-01 -8.82486343e-01 -5.10972559e-01
-9.59760785e-01 -1.35280371e+00 5.51153302e-01 -2.11138308e-01
8.30396041e-02 -1.15708876e+00 -1.27536237e+00 5.13836741e-01
1.13002479e-01 -5.08803546e-01 -1.18047439e-01 -7.16333166e-02
-7.43724406e-01 -1.26692629e+00 -8.10567677e-01 -3.49150419e-01
6.63091123e-01 -4.27278131e-02 8.18945050e-01 6.72114909e-01
-7.06819177e-01 1.48745656e-01 -3.38784635e-01 -7.78007746e-01
-1.06324804e+00 -3.42089385e-01 4.18877542e-01 -1.69440314e-01
9.00210217e-02 -2.75836915e-01 -7.15377986e-01 3.39293718e-01
-1.00862241e+00 -3.02235752e-01 3.34803730e-01 9.61492419e-01
3.00932914e-01 2.00302303e-01 6.51258588e-01 -6.40328884e-01
9.63756442e-01 -7.66180813e-01 -7.34550357e-01 6.99041247e-01
-7.95204759e-01 8.79092515e-02 6.52620316e-01 -4.91535515e-01
-8.76553059e-01 -4.43649381e-01 -1.67787179e-01 -6.36628037e-03
7.23076910e-02 4.05870348e-01 1.81269616e-01 3.19837891e-02
5.97610116e-01 2.98131734e-01 5.25270343e-01 -2.34790102e-01
9.73979309e-02 3.76006782e-01 -6.67312816e-02 -4.89531368e-01
7.33582735e-01 4.54412133e-01 1.19273581e-01 -8.99448514e-01
1.01443931e-01 -9.59942564e-02 6.74547628e-02 -3.88606519e-01
9.32798862e-01 -2.33110592e-01 -8.16753626e-01 5.35781741e-01
-1.15531421e+00 -1.24821827e-01 -2.15824515e-01 1.93990722e-01
-4.46863562e-01 5.23966961e-02 -3.71205956e-02 -1.13496768e+00
-5.02919614e-01 -1.21863997e+00 3.14561933e-01 4.50831413e-01
-1.95527121e-01 -1.07600522e+00 8.14634442e-01 1.35333925e-01
1.06711686e+00 5.63378036e-01 1.20976412e+00 -6.49614990e-01
-3.97933125e-01 2.59852290e-01 4.33130294e-01 8.02288111e-03
4.07844543e-01 5.15627980e-01 -2.43699625e-01 -5.64367831e-01
2.76203662e-01 2.62444109e-01 4.86912221e-01 3.96276683e-01
4.79857266e-01 -7.33179331e-01 -6.19437635e-01 6.10888481e-01
1.43041849e+00 1.17432725e+00 6.58182859e-01 4.62915182e-01
4.63254862e-02 9.66351986e-01 5.34632981e-01 6.38399720e-01
1.99890845e-02 7.62131691e-01 6.04186416e-01 -1.36751994e-01
4.13108677e-01 2.21847847e-01 1.80784464e-01 7.00684547e-01
-7.22608685e-01 -4.70959246e-01 -1.08322811e+00 1.13071494e-01
-1.57537305e+00 -1.22544765e+00 1.61115199e-01 2.13717580e+00
5.83679795e-01 -4.20343541e-02 4.02345210e-01 -5.08903623e-01
6.38720214e-01 -2.04667494e-01 -5.54376721e-01 -9.59137261e-01
-5.59299886e-01 1.87911391e-01 3.29393297e-01 2.09506691e-01
-5.15242934e-01 5.51410079e-01 6.41276503e+00 3.53262722e-01
-1.41602790e+00 -2.06346020e-01 8.17986846e-01 7.76108950e-02
-6.63170099e-01 -2.70922240e-02 -3.36877048e-01 6.27285838e-01
1.13127923e+00 -2.43526071e-01 7.47588575e-01 1.55937508e-01
5.90746224e-01 1.49427354e-01 -2.48504743e-01 3.36563677e-01
5.79778068e-02 -1.61506140e+00 7.53113478e-02 3.10463637e-01
7.89744496e-01 1.00987919e-01 5.96768260e-02 -6.44980557e-03
2.24971071e-01 -1.01940143e+00 2.70942688e-01 6.86998248e-01
2.51048744e-01 -1.03151751e+00 8.66608024e-01 3.80621731e-01
-7.28767633e-01 -3.19289982e-01 -2.33043224e-01 1.77602217e-01
4.67103451e-01 1.10686429e-01 -1.08065128e+00 3.58368218e-01
5.39512455e-01 -1.15934178e-01 -3.76423180e-01 1.16466141e+00
2.37004086e-01 2.43544802e-01 -8.59106928e-02 -6.45266354e-01
3.31360489e-01 -5.81086338e-01 9.44747388e-01 9.19361949e-01
5.14424682e-01 2.40330443e-01 -2.88213938e-01 9.48742509e-01
5.74134767e-01 6.37199134e-02 -7.02546835e-01 -5.11217177e-01
4.24244612e-01 6.59136415e-01 -8.46683443e-01 -1.68285161e-01
1.82913244e-01 6.81967735e-01 -5.16324677e-02 3.54774177e-01
-9.86901104e-01 -1.09256404e-02 1.12453485e+00 4.15470079e-02
7.86102787e-02 -1.95305526e-01 1.87777743e-01 -8.87486041e-01
-5.24631619e-01 -1.48638523e+00 2.29557946e-01 -2.74814457e-01
-8.26310575e-01 1.01149106e+00 -1.62392467e-01 -7.05473840e-01
-3.42311144e-01 -5.38110554e-01 -7.36924112e-01 6.22075915e-01
-1.23237085e+00 -6.44490600e-01 4.17675912e-01 1.80408642e-01
4.20680881e-01 -3.53725135e-01 6.73889279e-01 -1.53618708e-01
-6.58125401e-01 1.96109295e-01 5.06435156e-01 -8.17408562e-01
1.29140809e-01 -5.01062036e-01 5.11551440e-01 6.25412285e-01
-2.16800615e-01 8.61828864e-01 1.09519756e+00 -1.08678615e+00
-1.69582522e+00 -7.41704404e-01 5.84931016e-01 -1.78622440e-01
6.53952718e-01 -1.90233395e-01 -8.02900255e-01 1.06633164e-01
3.01184237e-01 -9.42467928e-01 5.32282948e-01 -2.23496631e-01
3.22971232e-02 2.56047159e-01 -1.36805081e+00 9.50665653e-01
9.22020137e-01 1.63460076e-01 -2.58349270e-01 3.62397641e-01
7.15939462e-01 1.81128711e-01 -2.56883472e-01 7.54778922e-01
6.20961726e-01 -1.03313231e+00 9.93723214e-01 -8.52207124e-01
-3.77151102e-01 -2.07613677e-01 1.87971652e-01 -1.50598097e+00
-4.21391845e-01 -1.03212476e+00 3.05266410e-01 6.55730784e-01
4.17680860e-01 -1.21543705e+00 4.53727365e-01 1.73678309e-01
2.08514288e-01 -1.12299287e+00 -9.27843809e-01 -9.25803959e-01
-2.89084539e-02 2.95393199e-01 1.05648911e+00 1.03184044e+00
-3.31963867e-01 -2.76792675e-01 -5.71269274e-01 -4.95644510e-02
3.07877362e-01 -1.42514184e-01 6.01142943e-01 -1.16738391e+00
-5.11985719e-01 -9.79858279e-01 -2.40935549e-01 1.80382758e-01
-4.04466182e-01 -3.67819965e-01 -1.90725699e-01 -1.04013789e+00
5.37359416e-02 -7.07697272e-01 -1.82419747e-01 3.63987242e-03
-2.48834759e-01 -2.84725875e-01 3.17125022e-01 2.11485639e-01
1.47890911e-01 4.34948832e-01 1.16870058e+00 -1.19253263e-01
-4.90594566e-01 3.39774191e-02 -6.27669632e-01 4.37326342e-01
9.90652204e-01 -4.85557079e-01 -3.86879414e-01 -9.76338536e-02
7.00252116e-01 5.66579308e-03 4.69929948e-02 -7.72500932e-01
-2.34287575e-01 -8.46472502e-01 -1.01498589e-01 -1.40935943e-01
2.32542410e-01 -9.69217718e-01 1.11265290e+00 1.58829153e+00
2.66721457e-01 5.74380994e-01 2.82129496e-01 5.12971580e-01
2.67405242e-01 -9.70519334e-02 6.63592398e-01 -1.88553557e-01
-2.65804619e-01 1.87695920e-01 -7.49925017e-01 -1.11050546e-01
1.51112807e+00 -4.13619071e-01 -4.73684520e-01 7.30657130e-02
-3.82187128e-01 3.27987820e-01 7.78535247e-01 6.67135596e-01
5.46552539e-01 -6.64526880e-01 -1.01210225e+00 2.91710198e-01
-3.20125282e-01 -7.57785797e-01 4.82346684e-01 7.90596664e-01
-9.97454822e-01 5.58948874e-01 -5.91120958e-01 -3.82478684e-01
-1.23216033e+00 1.04861498e+00 5.93524039e-01 -3.62178057e-01
-1.08953565e-01 8.79938006e-01 -4.47617508e-02 -2.81483531e-01
-2.85532176e-01 2.17058718e-01 -3.42697859e-01 9.26412418e-02
4.95607764e-01 4.17407095e-01 -1.80875510e-01 -4.32233840e-01
-6.70270026e-01 5.83052933e-01 1.91577598e-01 3.97982657e-01
1.53380740e+00 1.50925696e-01 -5.54746330e-01 -1.22205138e-01
4.63710666e-01 2.04208240e-01 -8.30605567e-01 5.11262000e-01
1.30785778e-01 -3.36934865e-01 -6.23102248e-01 -1.17861307e+00
-9.41518128e-01 3.07609588e-01 6.87102377e-01 4.17453080e-01
1.36149228e+00 -2.64319867e-01 6.51913464e-01 1.42813459e-01
5.52732587e-01 -7.42111266e-01 -1.93044111e-01 3.64112288e-01
1.02785039e+00 -7.64817417e-01 -2.19210669e-01 1.73403531e-01
-4.57156330e-01 9.68865633e-01 4.17103112e-01 3.08642894e-01
3.02581102e-01 3.08563799e-01 1.21349074e-01 -2.39991084e-01
-1.06075180e+00 -1.15877725e-01 -2.37486828e-02 9.00249064e-01
8.30744579e-02 3.19065183e-01 -9.99890983e-01 -1.18023902e-01
2.65116006e-01 -2.90217787e-01 4.91711617e-01 8.38293970e-01
-6.50002301e-01 -1.49988198e+00 -7.21989453e-01 2.05342308e-01
-4.57639098e-02 -3.78905796e-02 -6.61555290e-01 6.68806195e-01
4.69191045e-01 5.32773614e-01 -1.22013595e-02 -9.24848765e-02
1.73074260e-01 -1.79007217e-01 4.18277413e-01 -7.76854530e-02
-9.48731303e-01 -2.38848701e-02 -2.04028878e-02 -1.54564202e-01
-2.03869358e-01 -7.55544364e-01 -1.04955995e+00 -5.32329202e-01
-3.12567234e-01 5.75252593e-01 8.63736391e-01 6.40496314e-01
8.90254855e-01 3.36136699e-01 7.67042041e-01 -5.28121471e-01
-6.10698342e-01 -4.16116089e-01 -8.52845535e-02 -2.06427258e-02
7.78214931e-02 -4.47978318e-01 -2.32861891e-01 -4.64461148e-01] | [5.72324800491333, 4.224107265472412] |
dca0f5ea-a015-4700-9905-d04003ec0592 | label-aware-double-transfer-learning-for | 1804.09021 | null | http://arxiv.org/abs/1804.09021v2 | http://arxiv.org/pdf/1804.09021v2.pdf | Label-aware Double Transfer Learning for Cross-Specialty Medical Named Entity Recognition | We study the problem of named entity recognition (NER) from electronic
medical records, which is one of the most fundamental and critical problems for
medical text mining. Medical records which are written by clinicians from
different specialties usually contain quite different terminologies and writing
styles. The difference of specialties and the cost of human annotation makes it
particularly difficult to train a universal medical NER system. In this paper,
we propose a label-aware double transfer learning framework (La-DTL) for
cross-specialty NER, so that a medical NER system designed for one specialty
could be conveniently applied to another one with minimal annotation efforts.
The transferability is guaranteed by two components: (i) we propose label-aware
MMD for feature representation transfer, and (ii) we perform parameter transfer
with a theoretical upper bound which is also label aware. We conduct extensive
experiments on 12 cross-specialty NER tasks. The experimental results
demonstrate that La-DTL provides consistent accuracy improvement over strong
baselines. Besides, the promising experimental results on non-medical NER
scenarios indicate that La-DTL is potential to be seamlessly adapted to a wide
range of NER tasks. | ['Li-Heng Chen', 'Wei-Nan Zhang', 'Shaodian Zhang', 'Yong Yu', 'Jian Shen', 'Zhenghui Wang', 'Yimei Gao', 'Yanru Qu', 'Ken Chen', 'Gen Gu'] | 2018-04-24 | label-aware-double-transfer-learning-for-1 | https://aclanthology.org/N18-1001 | https://aclanthology.org/N18-1001.pdf | naacl-2018-6 | ['medical-named-entity-recognition'] | ['natural-language-processing'] | [-1.27814664e-02 1.07913181e-01 -2.85339564e-01 -4.58374739e-01
-1.07675505e+00 -4.21721548e-01 1.68282479e-01 3.08452785e-01
-8.26093793e-01 7.47650981e-01 3.07200193e-01 -2.94678658e-01
-1.56549111e-01 -5.07375658e-01 -3.54688764e-01 -6.02650821e-01
3.71333569e-01 5.25194585e-01 1.08219266e-01 -1.75506219e-01
-1.79398343e-01 4.24298912e-01 -7.91931272e-01 3.22039425e-01
1.13254106e+00 7.85509527e-01 2.08141640e-01 3.05556953e-01
-3.92760962e-01 8.15607905e-01 -4.93746400e-01 -6.73377275e-01
1.26274481e-01 -3.71565223e-01 -1.12800741e+00 -2.11272582e-01
1.51009738e-01 9.77044851e-02 -2.60752559e-01 1.11415613e+00
7.45645285e-01 1.51452664e-02 7.57630348e-01 -8.05878997e-01
-8.68868351e-01 7.47419655e-01 -5.41227043e-01 5.16133644e-02
8.15021768e-02 -2.66803592e-01 7.43439376e-01 -4.74570215e-01
5.94645977e-01 1.02309883e+00 9.19198036e-01 1.03556919e+00
-9.13466692e-01 -8.18842173e-01 -5.40663162e-03 -1.24839917e-01
-1.36044085e+00 -2.84819514e-01 4.60340053e-01 -4.14690226e-01
5.26673794e-01 1.29417703e-01 -7.95058813e-03 1.06274390e+00
3.53391945e-01 7.63194084e-01 1.06563461e+00 -3.25912356e-01
-7.03017553e-03 4.81116354e-01 2.00362638e-01 7.86719918e-01
3.55525464e-01 -3.11976612e-01 -1.31246760e-01 -3.35434049e-01
6.63345814e-01 1.59573540e-01 -4.83172923e-01 -1.16453297e-01
-1.57054269e+00 6.03989840e-01 4.71507847e-01 6.25699043e-01
-2.66070276e-01 -3.98029953e-01 8.00984859e-01 4.32999790e-01
3.06911379e-01 7.26805687e-01 -8.32188547e-01 2.65359897e-02
-7.51909673e-01 -3.71370137e-01 1.03586495e+00 1.28146029e+00
2.39943996e-01 -3.84464085e-01 -5.13736367e-01 1.09715414e+00
4.61171456e-02 2.97173977e-01 8.57274175e-01 -3.86263460e-01
5.74134529e-01 6.49835885e-01 -1.31522212e-02 -6.65980458e-01
-6.70250118e-01 -4.35752332e-01 -1.24167657e+00 -5.18577874e-01
1.41979799e-01 -5.09651721e-01 -9.08007443e-01 1.74391365e+00
2.62632161e-01 1.12019591e-01 2.44193897e-01 5.48424840e-01
1.17709708e+00 4.35687065e-01 5.47238588e-01 -2.84277052e-01
1.88549936e+00 -1.09208381e+00 -9.16841209e-01 9.82932895e-02
1.11561441e+00 -7.41265833e-01 8.96355569e-01 3.32866162e-02
-6.32176161e-01 -4.13446158e-01 -6.25047863e-01 -2.49675944e-01
-3.47338200e-01 4.03714091e-01 6.08779192e-01 5.87032259e-01
-7.74853468e-01 3.53369445e-01 -9.41841066e-01 -5.15081227e-01
3.93856525e-01 4.31763828e-01 -6.28710568e-01 9.90984496e-03
-1.28445399e+00 7.81451464e-01 5.82367897e-01 9.87959057e-02
-2.93301016e-01 -7.75472462e-01 -8.17146420e-01 4.43837978e-02
1.89059004e-01 -9.09992337e-01 1.29502058e+00 -5.97863197e-01
-1.22570777e+00 1.12808084e+00 4.31655087e-02 4.53598090e-02
4.68446136e-01 -9.84302685e-02 -9.22020078e-01 -7.05420505e-03
4.16722685e-01 4.10082668e-01 1.36342302e-01 -1.02697682e+00
-8.13926935e-01 -2.76075989e-01 -2.23823979e-01 2.33889565e-01
-8.46525908e-01 7.95899928e-02 -6.02689147e-01 -8.56599987e-01
-2.25078002e-01 -8.47453654e-01 -4.65704471e-01 -2.62972474e-01
-5.53981125e-01 -5.37383318e-01 2.04327837e-01 -5.63414216e-01
1.46430266e+00 -2.22171974e+00 -8.83363262e-02 5.39314933e-02
2.71705300e-01 3.28317612e-01 -1.30928963e-01 2.59965271e-01
-4.19324897e-02 2.62541115e-01 -2.21363187e-01 -2.30009407e-01
-1.92184120e-01 6.10316470e-02 -6.64563999e-02 1.03644624e-01
9.33665186e-02 9.34867978e-01 -1.05490518e+00 -8.58184576e-01
-3.86894464e-01 3.50948513e-01 -2.57335156e-01 3.81442308e-01
2.68965572e-01 6.38579309e-01 -9.31524515e-01 6.38876140e-01
4.68891025e-01 -6.37301207e-01 2.15371072e-01 -5.11301041e-01
1.08068407e-01 3.46484631e-01 -9.33571041e-01 1.85701501e+00
-5.26640415e-01 -5.27484864e-02 -2.20416993e-01 -7.57216930e-01
7.95243680e-01 6.54974401e-01 5.97557306e-01 -3.71263951e-01
3.07215840e-01 4.24599886e-01 -1.21961549e-01 -6.84032798e-01
3.47499490e-01 -3.81510556e-01 -4.90304440e-01 3.51529568e-01
1.88421205e-01 4.72490638e-01 1.25319153e-01 -4.71042693e-02
1.20789790e+00 -2.84523576e-01 7.00470328e-01 -4.39913541e-01
6.22749150e-01 -9.22713503e-02 1.16004646e+00 7.28806138e-01
-3.46322864e-01 3.27179760e-01 1.37065247e-01 -2.85869837e-01
-6.55941844e-01 -8.56512904e-01 -5.18407524e-01 1.04834378e+00
6.46843463e-02 -1.63011074e-01 -7.61142433e-01 -1.32204974e+00
-1.90660879e-01 3.62046152e-01 -6.13096356e-01 -2.55786777e-01
-5.23170352e-01 -8.41958880e-01 8.96911860e-01 7.31825948e-01
6.32207155e-01 -1.21473336e+00 -2.81861275e-01 2.54846364e-01
-3.44229132e-01 -1.23346949e+00 -1.10979497e+00 3.33060205e-01
-1.03468716e+00 -8.62657070e-01 -1.00088203e+00 -1.38953245e+00
8.58077943e-01 1.22168191e-01 1.01277781e+00 3.13113164e-03
-1.34846851e-01 3.50201756e-01 -4.85258758e-01 -3.74105871e-01
-4.39944446e-01 7.28404284e-01 7.48963878e-02 -1.02379255e-01
8.38969767e-01 -2.73709983e-01 -5.65683126e-01 3.21043134e-01
-9.07244146e-01 -3.29635739e-02 1.01714039e+00 1.12818623e+00
6.67209923e-01 -1.06164850e-01 7.65852869e-01 -1.62849975e+00
6.11072302e-01 -6.15582347e-01 -1.31997719e-01 7.97506750e-01
-6.71339154e-01 2.83986330e-01 7.41828382e-01 -4.81256515e-01
-1.36358809e+00 9.89453420e-02 -2.23974094e-01 -4.84328251e-03
-3.58490258e-01 5.97767472e-01 -5.07859707e-01 2.07778051e-01
6.87456310e-01 8.83493796e-02 -4.41099823e-01 -7.67293811e-01
1.61051229e-01 1.12379253e+00 6.51667297e-01 -6.42808795e-01
4.98272508e-01 1.75523862e-01 -3.00837934e-01 -4.88339782e-01
-1.16432512e+00 -7.02738464e-01 -6.73999548e-01 4.55498427e-01
1.13885236e+00 -1.04123545e+00 -5.39127886e-01 2.13477165e-01
-1.03487551e+00 -4.37091850e-02 -1.05707243e-01 6.76603317e-01
-2.23158762e-01 3.75262409e-01 -1.11070454e+00 -2.04270303e-01
-7.26886630e-01 -1.01962185e+00 9.74713802e-01 5.99137723e-01
-1.65434659e-01 -1.33917928e+00 2.56228536e-01 4.02513176e-01
1.27846479e-01 -6.77506104e-02 1.13946950e+00 -1.19813097e+00
8.47514421e-02 -1.77116111e-01 -3.48238617e-01 2.26406008e-01
6.65759444e-01 -5.46839774e-01 -8.62390399e-01 -3.72653812e-01
1.11166060e-01 -3.92393172e-01 7.35374510e-01 -6.01623394e-03
1.27850258e+00 -1.99324936e-01 -7.86413133e-01 6.76674306e-01
1.11965930e+00 2.76794374e-01 3.79765183e-01 4.21029150e-01
1.02667630e+00 4.60809231e-01 5.93325138e-01 1.59855440e-01
6.25460982e-01 4.71777111e-01 -3.70948166e-01 -5.71532845e-01
-4.84552085e-02 -2.34889701e-01 6.99344352e-02 1.43038130e+00
-4.54189070e-02 -1.70143828e-01 -1.02885234e+00 6.34929895e-01
-1.71957111e+00 -5.94155967e-01 1.00654960e-01 2.03939652e+00
1.46222067e+00 -1.35023668e-01 -7.86525831e-02 -4.05200124e-01
8.66645575e-01 -4.27170128e-01 -5.80028057e-01 -2.33662069e-01
2.03568593e-01 2.35090584e-01 5.77881038e-01 -5.19681983e-02
-1.36197352e+00 7.41771877e-01 5.85216618e+00 8.49937558e-01
-9.26165581e-01 2.88013726e-01 6.38489425e-01 4.65991735e-01
-7.17594102e-02 -4.67608631e-01 -1.11396074e+00 4.97393191e-01
9.70787406e-01 -2.22105712e-01 -2.81164348e-01 8.25979650e-01
-2.65572399e-01 6.27878547e-01 -1.27490950e+00 9.86998141e-01
-9.11455378e-02 -9.98137593e-01 3.01110864e-01 1.23818085e-01
6.94590271e-01 -3.95020731e-02 -1.55211926e-01 6.24868274e-01
3.01790863e-01 -8.36709023e-01 2.71437205e-02 5.32663882e-01
1.09073949e+00 -7.41505027e-01 1.11994946e+00 3.20875734e-01
-1.17308879e+00 1.07968487e-01 -4.82297838e-01 7.57978261e-01
2.12124348e-01 6.17510378e-01 -7.17266262e-01 1.00329959e+00
7.38999784e-01 7.65800178e-01 -3.46526504e-01 1.12667084e+00
-9.55053195e-02 4.95702207e-01 3.10395602e-02 8.68135616e-02
-3.51156518e-02 4.44885120e-02 6.49443120e-02 1.78991747e+00
3.31705660e-01 2.50702888e-01 3.89034063e-01 5.23401141e-01
-6.64077401e-01 6.16235077e-01 -3.94215435e-01 -2.18015686e-01
5.33195078e-01 1.56192994e+00 -7.06354856e-01 -2.41100520e-01
-5.81602216e-01 1.23358476e+00 4.03980464e-01 9.11038294e-02
-7.53846169e-01 -8.31999719e-01 4.53990310e-01 -2.83380508e-01
2.07398772e-01 3.26872528e-01 -1.17958613e-01 -1.44097018e+00
-1.27612963e-01 -8.70904326e-01 8.38037014e-01 -7.20066503e-02
-2.01809812e+00 1.13211572e+00 -3.61839473e-01 -1.57793200e+00
1.33744776e-01 -6.92823946e-01 -2.46433586e-01 7.18512356e-01
-1.66245782e+00 -1.21655858e+00 -1.21276535e-01 7.21328795e-01
3.04134429e-01 -3.52430820e-01 1.27704620e+00 7.61972189e-01
-6.46998167e-01 1.24829006e+00 1.79418623e-01 7.33100653e-01
1.22059739e+00 -1.33068609e+00 2.00115532e-01 4.10632849e-01
-5.78360781e-02 8.49510849e-01 7.51022547e-02 -6.78714335e-01
-9.49343622e-01 -1.38785124e+00 1.23728514e+00 -6.01881504e-01
4.75455612e-01 -1.77358910e-01 -1.10889554e+00 6.72396302e-01
1.03353590e-01 -5.05798124e-02 1.46990120e+00 3.36430967e-01
-4.86726731e-01 -1.26702756e-01 -1.21230721e+00 4.42383498e-01
1.00229514e+00 -4.83135194e-01 -8.62713695e-01 4.23715830e-01
8.05593491e-01 -4.06248361e-01 -1.52400291e+00 6.18794024e-01
3.27943027e-01 -2.84603328e-01 5.90133011e-01 -9.28891420e-01
2.49661222e-01 -2.91407675e-01 1.04106948e-01 -1.18712890e+00
-5.47926843e-01 -4.24043328e-01 1.40486807e-01 1.55368745e+00
5.32983661e-01 -5.53702950e-01 3.58384758e-01 7.27706075e-01
-1.38136506e-01 -6.91982806e-01 -7.64596283e-01 -8.51133525e-01
2.79221714e-01 1.15010828e-01 4.83612776e-01 1.53204191e+00
1.96329698e-01 6.24195695e-01 -3.71442437e-01 8.46252143e-02
2.98127949e-01 1.67029370e-02 3.74850512e-01 -1.47750521e+00
-2.64916688e-01 -2.96702921e-01 -3.06939334e-01 -1.11040604e+00
1.93635315e-01 -1.21190274e+00 3.04287285e-01 -1.53132689e+00
7.13607430e-01 -8.06779623e-01 -9.28912938e-01 9.26917374e-01
-5.40131748e-01 -5.95270433e-02 -1.98342070e-01 4.44574386e-01
-7.85767615e-01 3.06696296e-01 1.19865549e+00 1.66953180e-03
-2.50246912e-01 1.58139355e-02 -1.08754599e+00 6.51589394e-01
5.99578440e-01 -8.81146312e-01 -2.78040558e-01 -4.85783339e-01
4.23948616e-02 6.99295253e-02 -4.10544187e-01 -7.21656978e-01
4.22436625e-01 -9.02966112e-02 2.62149334e-01 -9.16356370e-02
-2.78073937e-01 -8.53629529e-01 -1.14476293e-01 3.92495483e-01
-4.77293283e-01 7.49712903e-03 2.09774822e-01 6.33480608e-01
-2.21182287e-01 -2.14031875e-01 7.01149404e-01 -8.68794993e-02
-5.86581528e-01 5.52935183e-01 -1.09788112e-01 3.85560483e-01
8.09591830e-01 1.89986259e-01 -2.73742914e-01 1.68572068e-01
-6.23348773e-01 4.51207608e-01 2.25721419e-01 5.11222780e-01
2.41882086e-01 -1.41340506e+00 -8.29958737e-01 -1.56306744e-01
5.74879706e-01 -5.03866188e-02 3.68763179e-01 9.44562495e-01
-2.62239456e-01 4.19124246e-01 -1.10745005e-01 -4.22961950e-01
-1.14625025e+00 6.65485263e-01 3.36014599e-01 -7.88607419e-01
-7.44192481e-01 7.66478300e-01 5.03292203e-01 -9.55687702e-01
1.76245064e-01 -1.12644091e-01 -4.09168571e-01 -1.33135632e-01
7.00869858e-01 1.53458267e-01 2.07280308e-01 -5.01290262e-01
-4.37502772e-01 5.99598825e-01 -6.70738995e-01 3.36320162e-01
1.34912086e+00 3.08763608e-03 -1.47408918e-01 4.70514178e-01
1.13254499e+00 9.27796736e-02 -6.03227377e-01 -5.73255718e-01
4.32811767e-01 1.15767211e-01 -2.72484440e-02 -1.00278306e+00
-1.14517117e+00 6.34867430e-01 6.66991651e-01 -1.82192087e-01
1.35323763e+00 6.93094730e-02 1.21261525e+00 7.02603996e-01
3.89999002e-01 -1.01585245e+00 -3.82160306e-01 3.53898495e-01
3.74053121e-01 -1.20998299e+00 -3.42414439e-01 -4.23980981e-01
-8.94164324e-01 1.07731378e+00 5.89627743e-01 3.72001141e-01
7.56580055e-01 2.95715094e-01 4.77575719e-01 -2.02326521e-01
-4.32336926e-01 -2.66817361e-02 6.27950549e-01 3.82372320e-01
9.67223585e-01 7.57249221e-02 -6.06821835e-01 1.18630946e+00
4.22161408e-02 2.33044907e-01 2.35763907e-01 9.35766041e-01
-1.41523078e-01 -1.48013580e+00 2.65144594e-02 5.59631884e-01
-8.32067668e-01 -1.71070307e-01 -1.41780049e-01 6.15386128e-01
2.12912396e-01 8.07034314e-01 -3.25629592e-01 -3.92331213e-01
5.98478377e-01 1.94544792e-01 2.62727827e-01 -9.80144441e-01
-8.31331849e-01 2.23521646e-02 -9.28222537e-02 -2.17132553e-01
-4.60528612e-01 -4.27181691e-01 -1.58163416e+00 6.86828196e-02
-3.63550991e-01 6.52657032e-01 2.35053763e-01 8.77520859e-01
6.73795938e-01 6.86884820e-01 5.92965961e-01 3.97261791e-02
-6.33231401e-01 -9.42768991e-01 -7.58151770e-01 5.71167111e-01
4.92934510e-02 -4.96127874e-01 -1.49026548e-03 1.99733287e-01] | [8.620674133300781, 8.884127616882324] |
21a82ebb-c5ee-41ae-873f-9508cd73c642 | open-source-frame-semantic-parsing | 2303.12788 | null | https://arxiv.org/abs/2303.12788v1 | https://arxiv.org/pdf/2303.12788v1.pdf | Open-source Frame Semantic Parsing | While the state-of-the-art for frame semantic parsing has progressed dramatically in recent years, it is still difficult for end-users to apply state-of-the-art models in practice. To address this, we present Frame Semantic Transformer, an open-source Python library which achieves near state-of-the-art performance on FrameNet 1.7, while focusing on ease-of-use. We use a T5 model fine-tuned on Propbank and FrameNet exemplars as a base, and improve performance by using FrameNet lexical units to provide hints to T5 at inference time. We enhance robustness to real-world data by using textual data augmentations during training. | ['David Chanin'] | 2023-03-22 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 6.97574439e-03 7.04928815e-01 -3.18703353e-01 -6.63549364e-01
-9.52887356e-01 -7.10121691e-01 6.42383814e-01 8.88675973e-02
-4.66514796e-01 7.65208542e-01 6.52643383e-01 -6.45109475e-01
5.51844358e-01 -7.89864898e-01 -8.42559934e-01 2.29087770e-01
2.97277391e-01 4.90081400e-01 6.89025819e-01 -4.98120487e-01
-1.51822686e-01 -1.08182564e-01 -1.43989813e+00 8.81072402e-01
3.98190796e-01 6.24286234e-01 3.59927505e-01 7.11559832e-01
-4.37342852e-01 1.05890751e+00 -7.06426442e-01 -7.60105491e-01
1.84401870e-02 -2.93715298e-01 -1.34846807e+00 -2.35111341e-01
1.86074391e-01 -4.13529545e-01 -4.04343754e-01 6.98286235e-01
5.03506124e-01 2.02043593e-01 -1.87633619e-01 -9.38150287e-01
-7.44039178e-01 1.17255020e+00 1.60920396e-01 4.34278250e-01
7.56940901e-01 1.55332619e-02 1.30911803e+00 -7.28481472e-01
1.04697835e+00 1.53521383e+00 9.40199196e-01 9.11321700e-01
-9.67936039e-01 -4.72549230e-01 3.21583450e-01 2.06977323e-01
-8.35692704e-01 -7.06616998e-01 6.25702918e-01 -5.43123931e-02
1.69662189e+00 -7.04938099e-02 6.10417724e-01 1.12759924e+00
-8.63225460e-02 1.13651383e+00 6.35621190e-01 -7.70761967e-01
-4.99101467e-02 -5.05625784e-01 2.24756435e-01 7.88204372e-01
-2.46720538e-01 -1.80416573e-02 -7.72168100e-01 2.56856885e-02
7.98181176e-01 -5.08371592e-01 -8.07102323e-02 1.04705155e-01
-1.12091982e+00 8.31790626e-01 1.11948803e-01 4.22472239e-01
-6.95231333e-02 3.83014292e-01 7.51522601e-01 1.20791882e-01
8.07609916e-01 2.49071449e-01 -9.41432059e-01 -7.81468689e-01
-9.76403534e-01 6.87812686e-01 9.24098969e-01 1.05446827e+00
4.48156297e-01 -2.35126577e-02 -1.66367024e-01 8.30891132e-01
4.48639601e-01 3.85672823e-02 5.47554791e-01 -1.12422359e+00
8.10533583e-01 4.66591150e-01 -2.55629010e-02 -2.04025403e-01
-3.17516446e-01 -1.67267844e-02 1.40514612e-01 -8.88537169e-02
6.62210822e-01 -1.41816780e-01 -1.10754979e+00 1.89778185e+00
2.60416687e-01 3.88844877e-01 2.67058223e-01 5.96156716e-01
1.13408697e+00 5.57674944e-01 5.05037189e-01 2.31736138e-01
1.46668208e+00 -1.11283028e+00 -7.93719053e-01 -6.28461063e-01
9.66901660e-01 -8.26205492e-01 1.28621280e+00 -6.46979585e-02
-1.48465705e+00 -6.15470886e-01 -8.75207245e-01 -6.55532837e-01
-3.24696213e-01 -2.42162853e-01 7.40956247e-01 6.57032669e-01
-9.39684987e-01 6.40763581e-01 -1.41838086e+00 -6.34627879e-01
2.99782008e-01 9.65790302e-02 -3.87312233e-01 -1.15677118e-01
-1.46732485e+00 1.14559615e+00 7.26150095e-01 -2.41121441e-01
-6.52520537e-01 -1.08061290e+00 -1.38906813e+00 -1.49424598e-01
5.02296329e-01 -8.52854609e-01 2.09116149e+00 -6.35861695e-01
-1.71806085e+00 8.62112105e-01 -4.62593347e-01 -7.82917023e-01
3.40943784e-01 -6.43599212e-01 -4.78002310e-01 1.60934299e-01
3.67063254e-01 9.67135966e-01 3.93535674e-01 -6.92596674e-01
-8.69878411e-01 6.66910484e-02 7.34823585e-01 1.58092409e-01
1.71633318e-01 5.62703252e-01 -5.31914890e-01 -7.56304264e-01
-1.65311426e-01 -6.70145154e-01 -3.68510634e-02 -1.69137955e-01
-6.46042004e-02 -4.43205863e-01 1.03966439e+00 -8.09980512e-01
1.33550811e+00 -1.84194708e+00 -1.32166237e-01 -5.09110391e-01
-1.98625252e-01 4.77002650e-01 -1.49129955e-02 2.43784517e-01
-6.27435297e-02 3.43209893e-01 -1.47125557e-01 -7.28343070e-01
-5.35365604e-02 7.37772107e-01 -2.89374918e-01 3.47094983e-02
6.10284209e-01 9.68976974e-01 -1.28243232e+00 -5.22835255e-01
5.23411870e-01 5.00555515e-01 -9.61857319e-01 1.64391845e-01
-7.75645077e-01 3.79071087e-01 -5.05523443e-01 6.52764082e-01
3.46304551e-02 -5.66300601e-02 3.10320437e-01 -1.05563745e-01
-1.74390674e-01 1.19367015e+00 -8.80066156e-01 2.39858007e+00
-4.32204902e-01 5.39990962e-01 -1.42371431e-01 -9.06303883e-01
4.20576364e-01 7.02715397e-01 1.50829121e-01 -4.38900828e-01
6.23146221e-02 2.68652905e-02 -9.50827152e-02 -3.62040699e-01
7.73289859e-01 -2.10137114e-01 -2.61379629e-01 3.12676638e-01
4.82230544e-01 -1.37899444e-01 4.46347505e-01 1.60519425e-02
1.17006814e+00 9.61208642e-01 4.93750304e-01 -3.34724337e-01
2.65448540e-01 2.69573361e-01 7.16006875e-01 5.38045466e-01
-2.65736375e-02 8.43801022e-01 3.43805492e-01 -7.11342037e-01
-1.09651256e+00 -7.46117473e-01 4.98742163e-02 1.40550601e+00
-2.09934339e-01 -1.05066144e+00 -1.09114814e+00 -1.03513598e+00
-3.19106251e-01 1.09566343e+00 -4.90867466e-01 3.96547556e-01
-1.06111431e+00 -3.43220204e-01 1.02062130e+00 8.51899981e-01
6.62379503e-01 -1.21384203e+00 -9.48612213e-01 7.46790648e-01
-4.59251940e-01 -1.61630094e+00 -1.00778952e-01 4.02995087e-02
-6.82259202e-01 -1.04340982e+00 -3.59560192e-01 -5.33879220e-01
1.89130723e-01 1.41338604e-02 1.79456592e+00 2.05286458e-01
2.78264284e-01 2.16288105e-01 -7.46536434e-01 -5.98636031e-01
-6.36331856e-01 3.95432025e-01 -4.93512660e-01 -7.45669723e-01
3.93096805e-01 -1.00280605e-01 -2.05814987e-01 1.31898224e-01
-8.97414088e-01 2.85641015e-01 -2.41056219e-01 8.96381080e-01
4.24443036e-01 -2.41123870e-01 1.78057030e-01 -1.18234253e+00
2.81302273e-01 -2.61215061e-01 -3.68111163e-01 1.35847867e-01
3.83488312e-02 4.21139076e-02 4.98010874e-01 6.46489188e-02
-1.33905220e+00 -9.64770690e-02 -7.82060981e-01 -9.33441669e-02
-2.39304766e-01 6.83669388e-01 -1.55343339e-01 3.44168663e-01
6.30008280e-01 -4.22092140e-01 -2.53312409e-01 -5.73678613e-01
7.56157815e-01 3.26239467e-01 7.52387226e-01 -1.03537750e+00
4.18479323e-01 2.47299269e-01 -4.26993966e-01 -7.44598329e-01
-1.46169233e+00 -1.72841713e-01 -5.38132191e-01 1.20211676e-01
1.04303432e+00 -1.03653193e+00 -1.86129451e-01 2.19392180e-01
-1.45655096e+00 -6.97537482e-01 -4.25294191e-01 4.22800109e-02
-5.35732806e-01 3.57190162e-01 -8.03196847e-01 -3.58408123e-01
-3.25247824e-01 -1.16906762e+00 1.12553024e+00 9.46414694e-02
-4.83367920e-01 -1.27897167e+00 -1.45594880e-01 5.49951494e-01
2.06980318e-01 1.41406476e-01 6.61136389e-01 -7.90920973e-01
-4.13015097e-01 8.31030011e-02 3.48316394e-02 2.24697620e-01
-7.37124532e-02 -4.07670774e-02 -1.13354874e+00 4.50795181e-02
-1.54463649e-01 -3.58917773e-01 5.92585087e-01 4.32823658e-01
9.62100804e-01 -7.27795735e-02 -3.32059443e-01 4.55319434e-01
1.02995861e+00 3.27471197e-02 8.16613197e-01 7.77660847e-01
6.47898972e-01 1.53836012e-01 8.20248485e-01 1.26938477e-01
9.53368485e-01 6.03134930e-01 1.57352835e-01 4.53371853e-02
-4.63134974e-01 -6.08592212e-01 1.77647233e-01 6.23894632e-01
-1.51602581e-01 -3.30017984e-01 -1.10667729e+00 6.72943532e-01
-2.07744551e+00 -1.01312304e+00 1.50879994e-01 1.66113770e+00
1.09983182e+00 7.64471412e-01 8.24837983e-02 4.78646457e-02
6.08945489e-01 4.09352988e-01 -9.88499522e-02 -2.98282057e-01
-8.48593786e-02 7.09027827e-01 3.50535423e-01 6.63173139e-01
-1.21293008e+00 1.75318885e+00 7.33994961e+00 6.68818712e-01
-1.06878710e+00 5.16288579e-01 5.45682192e-01 2.03776643e-01
-2.32408166e-01 4.58636105e-01 -1.17879653e+00 3.46241862e-01
1.14123178e+00 6.18969323e-03 2.82130480e-01 9.66871440e-01
2.11206004e-01 9.18096453e-02 -1.15525270e+00 5.74477255e-01
-1.43889144e-01 -1.82190263e+00 -8.63038450e-02 -5.70750415e-01
3.30873907e-01 4.21524167e-01 -4.55704898e-01 6.91543937e-01
6.72572076e-01 -7.36488104e-01 1.21445286e+00 -1.07371502e-01
7.27887213e-01 -3.78802419e-01 6.80803180e-01 7.99794346e-02
-1.42375410e+00 1.26843572e-01 -3.30644459e-01 -5.79893827e-01
8.06257665e-01 3.83945614e-01 -7.98364401e-01 6.82590365e-01
6.18027568e-01 8.07758689e-01 -4.57938761e-01 6.30408466e-01
-7.59325564e-01 8.30738604e-01 -5.46719134e-01 2.06879914e-01
3.54797453e-01 4.58023936e-01 3.87921631e-01 1.42163539e+00
3.55678611e-02 2.86018878e-01 4.56497818e-01 6.26087844e-01
-1.91114232e-01 -2.92724222e-01 -4.91344422e-01 6.54008538e-02
6.52021706e-01 9.20699000e-01 -7.33584762e-01 -6.08175278e-01
-8.80744517e-01 6.27895713e-01 3.91545802e-01 1.94703341e-01
-9.53117132e-01 -1.53148517e-01 7.79848516e-01 1.59836784e-01
4.05026287e-01 -4.08070713e-01 -2.85948694e-01 -1.33398783e+00
-1.40854523e-01 -9.65018094e-01 6.68873072e-01 -8.86505842e-01
-8.65239799e-01 6.32744730e-01 3.85379225e-01 -6.97761238e-01
-5.65627337e-01 -6.89943731e-01 -5.26924312e-01 6.40829980e-01
-1.49158061e+00 -1.54327929e+00 1.49434224e-01 3.57027441e-01
1.08138740e+00 1.21983133e-01 1.06781328e+00 1.96572691e-01
-3.25636297e-01 4.65407997e-01 -6.51006579e-01 4.57966089e-01
5.86560249e-01 -1.21683502e+00 1.47554708e+00 1.15862167e+00
1.43106416e-01 6.35105968e-01 8.82044971e-01 -7.68846869e-01
-1.24907577e+00 -9.94554639e-01 1.14192796e+00 -6.48662865e-01
1.08519864e+00 -3.73339832e-01 -7.65227020e-01 1.22614062e+00
1.98721111e-01 3.10060233e-01 3.90964061e-01 4.07752067e-01
-4.79922265e-01 5.28434336e-01 -1.10845721e+00 5.82763731e-01
1.36682785e+00 -8.60538423e-01 -1.31098545e+00 1.77731037e-01
1.13112319e+00 -1.21076584e+00 -9.30140972e-01 4.76364881e-01
4.60478991e-01 -8.10687125e-01 8.97371292e-01 -5.57640851e-01
5.17171443e-01 -2.97033668e-01 -3.64008904e-01 -1.18082905e+00
-2.11894009e-02 -8.56592655e-01 -1.94768518e-01 1.40084374e+00
5.64709544e-01 -3.97787094e-01 1.00814176e+00 8.39367449e-01
-7.82655120e-01 -4.69787389e-01 -1.06105030e+00 -5.39574206e-01
9.13501456e-02 -9.46908593e-01 6.28411353e-01 9.06446099e-01
3.82200569e-01 6.20829880e-01 2.92334054e-02 7.68029783e-03
1.45265728e-01 -3.02090794e-01 7.09346294e-01 -9.49871659e-01
-4.85815138e-01 -1.64760053e-01 -3.12855691e-01 -1.23516035e+00
6.31854415e-01 -7.50715017e-01 -2.73037199e-02 -1.81682432e+00
-3.25901061e-01 -3.65297735e-01 -5.81831224e-02 1.02724981e+00
-3.31792563e-01 2.78649509e-01 5.41826248e-01 -1.33911535e-01
-7.15163529e-01 2.46051446e-01 8.98131788e-01 1.67270288e-01
-1.56011105e-01 -3.66552174e-01 -7.37521589e-01 9.82221723e-01
7.55923986e-01 -3.37896705e-01 -3.89754266e-01 -1.13092601e+00
-1.40042473e-02 2.19055757e-01 9.19037312e-02 -1.11748004e+00
-7.19834343e-02 1.05931498e-01 1.13324009e-01 -5.74906886e-01
4.29022312e-01 -2.99969316e-01 -2.03590617e-02 7.53515959e-02
-2.07243726e-01 3.17130446e-01 6.60025954e-01 6.00336529e-02
-2.56376565e-01 -4.68905896e-01 5.42586267e-01 -5.70230663e-01
-1.00556171e+00 -1.04843117e-02 -3.24306577e-01 5.74243248e-01
5.35985053e-01 2.07640883e-02 -6.88297033e-01 -2.14553833e-01
-5.98703742e-01 2.55327951e-03 6.02431774e-01 4.92337406e-01
1.69522822e-01 -1.00765860e+00 -5.57885468e-01 9.15415362e-02
9.54786614e-02 3.24608654e-01 -1.49754927e-01 4.05857787e-02
-6.63989484e-01 3.62015486e-01 -8.07384774e-02 -4.92839992e-01
-1.16665983e+00 8.87091458e-02 1.84233189e-01 -4.64940429e-01
-8.69155467e-01 9.70015109e-01 -2.19604969e-01 -4.79830325e-01
1.77510083e-02 -7.36211479e-01 -9.02956054e-02 -3.45048010e-01
6.23930156e-01 1.15280248e-01 2.38388434e-01 -4.87605482e-01
-3.01328808e-01 2.19531611e-01 -9.33165252e-02 -4.91058409e-01
1.32780015e+00 6.19957317e-03 1.78857788e-01 3.24156851e-01
7.32544601e-01 -2.57241964e-01 -1.38468683e+00 -1.54222026e-01
1.37292236e-01 -3.24524283e-01 5.34692854e-02 -8.35472703e-01
-6.81429446e-01 6.11888051e-01 3.09879892e-02 1.76175058e-01
7.32073188e-01 1.57242388e-01 1.10991132e+00 4.02845711e-01
4.58487868e-01 -1.06428576e+00 -2.95883119e-01 1.20573056e+00
3.56513858e-01 -9.49638605e-01 -6.57888278e-02 -7.31686473e-01
-4.29760903e-01 1.06418490e+00 4.18923587e-01 -7.31438994e-02
6.63445234e-01 4.95909840e-01 2.37328753e-01 -3.01021226e-02
-8.28723490e-01 -2.49646947e-01 -1.79295372e-02 6.51243925e-01
1.02412057e+00 -1.04696259e-01 -2.70081848e-01 6.83954120e-01
-5.82165539e-01 3.57604116e-01 3.53611082e-01 1.54938793e+00
-4.13578063e-01 -1.63310111e+00 -1.47085607e-01 6.75355038e-03
-1.22117364e+00 -4.47360963e-01 8.38039964e-02 1.01074088e+00
2.51557007e-02 1.03779948e+00 7.63809308e-02 -7.08802044e-02
5.16724825e-01 4.12479669e-01 7.11446166e-01 -1.03266430e+00
-8.14908028e-01 3.63097899e-02 9.25228000e-01 -6.68950140e-01
-7.83280671e-01 -7.24278808e-01 -1.47836912e+00 -2.54345357e-01
-1.80214003e-01 -2.65657194e-02 6.05193019e-01 1.37302577e+00
3.47854406e-01 6.15671396e-01 -4.45686728e-01 -1.08627021e+00
-5.27977124e-02 -1.07106090e+00 3.01481783e-01 5.15103400e-01
-9.13525596e-02 -7.07460880e-01 1.16782099e-01 4.85409141e-01] | [10.33467960357666, 9.34820556640625] |
c1a37924-0c14-4845-804a-a90f263f4beb | icdar-2019-competition-on-large-scale-street | 1909.07741 | null | https://arxiv.org/abs/1909.07741v1 | https://arxiv.org/pdf/1909.07741v1.pdf | ICDAR 2019 Competition on Large-scale Street View Text with Partial Labeling -- RRC-LSVT | Robust text reading from street view images provides valuable information for various applications. Performance improvement of existing methods in such a challenging scenario heavily relies on the amount of fully annotated training data, which is costly and in-efficient to obtain. To scale up the amount of training data while keeping the labeling procedure cost-effective, this competition introduces a new challenge on Large-scale Street View Text with Partial Labeling (LSVT), providing 50, 000 and 400, 000 images in full and weak annotations, respectively. This competition aims to explore the abilities of state-of-the-art methods to detect and recognize text instances from large-scale street view images, closing the gap between research benchmarks and real applications. During the competition period, a total of 41 teams participated in the two proposed tasks with 132 valid submissions, i.e., text detection and end-to-end text spotting. This paper includes dataset descriptions, task definitions, evaluation protocols and results summaries of the ICDAR 2019-LSVT challenge. | ['Dimosthenis Karatzas', 'Errui Ding', 'Chun Chet Ng', 'Chee-Kheng Chng', 'Junyu Han', 'Yuliang Liu', 'Lianwen Jin', 'Jingtuo Liu', 'Chee Seng Chan', 'Zihan Ni', 'Yipeng Sun', 'Canjie Luo'] | 2019-09-17 | null | null | null | null | ['text-spotting'] | ['computer-vision'] | [ 5.00984788e-01 -2.98731863e-01 -1.90763772e-02 -3.92096370e-01
-1.10777199e+00 -6.76427186e-01 6.82124317e-01 1.17599577e-01
-4.97204900e-01 3.34949553e-01 1.71693444e-01 -9.29612815e-02
6.21223569e-01 -4.89581287e-01 -5.92371941e-01 -5.89932978e-01
6.34392679e-01 7.22680986e-01 3.15226734e-01 5.81753291e-02
5.61609387e-01 -7.34393075e-02 -1.63259101e+00 7.28457272e-01
5.74973702e-01 8.29272985e-01 6.32404506e-01 8.42703938e-01
-4.21048135e-01 5.38530827e-01 -5.98257422e-01 -4.82472897e-01
2.09133968e-01 -6.31573796e-02 -8.76024544e-01 4.86402333e-01
1.17712545e+00 -3.48682284e-01 -2.68281460e-01 7.67999232e-01
8.34021986e-01 -1.34885207e-01 6.46539629e-01 -9.93156314e-01
-4.28268790e-01 6.56543195e-01 -1.00612569e+00 3.30777764e-01
3.74648690e-01 7.01503009e-02 1.21192586e+00 -1.29722285e+00
9.23457861e-01 1.10385644e+00 5.37572861e-01 4.37315226e-01
-8.43218505e-01 -6.31800234e-01 4.48118448e-01 7.10691046e-03
-1.21560061e+00 -6.98941886e-01 4.38791603e-01 -6.00321472e-01
1.19579351e+00 2.22728878e-01 3.98424983e-01 1.26432371e+00
-1.77061096e-01 1.35426390e+00 9.99193490e-01 -5.20723403e-01
-2.05184504e-01 2.51809031e-01 2.35518828e-01 6.87449396e-01
3.06753337e-01 -4.97877449e-01 -7.06594348e-01 4.45035920e-02
2.12275013e-01 -1.94119409e-01 -2.31895596e-01 -3.12901825e-01
-1.51850200e+00 6.52474105e-01 -1.25443593e-01 2.61677742e-01
1.22374646e-01 -1.11617990e-01 1.02335405e+00 9.66591686e-02
8.59411716e-01 1.22357190e-01 -6.50850773e-01 1.42200710e-02
-1.28012753e+00 1.13789603e-01 4.60131973e-01 9.85748649e-01
3.71778429e-01 -1.03637569e-01 -5.26827335e-01 1.41333985e+00
1.96021706e-01 1.13319385e+00 5.83187282e-01 -2.03593239e-01
1.28949416e+00 7.51907349e-01 -1.55172199e-01 -7.88394690e-01
-5.40706575e-01 -3.75009060e-01 -8.68298411e-01 -2.10034460e-01
3.42794955e-01 6.06575273e-02 -1.31369281e+00 1.07118189e+00
2.38557532e-01 -2.04239145e-01 -1.22454837e-01 6.84242308e-01
1.18794763e+00 6.82013512e-01 -2.57879913e-01 1.02647185e-01
1.56882930e+00 -1.35285258e+00 -6.20075881e-01 -6.81821346e-01
9.04204428e-01 -1.15120423e+00 1.46235168e+00 4.21345979e-01
-8.98488045e-01 -3.56195122e-01 -8.75974178e-01 -3.12517345e-01
-6.85682952e-01 8.35933328e-01 1.06803700e-01 6.17038131e-01
-8.83360028e-01 -1.28965288e-01 -5.21205068e-01 -7.36609161e-01
6.32911026e-01 1.96809843e-01 -3.77396584e-01 -2.27491230e-01
-8.26954544e-01 4.37867612e-01 2.70043463e-01 3.28189181e-03
-8.61347079e-01 -4.67961788e-01 -5.67002892e-01 -1.96925372e-01
5.43217242e-01 -5.49850345e-01 1.04736876e+00 -4.70606983e-01
-1.03075981e+00 1.57487905e+00 -2.86051244e-01 -5.98046601e-01
8.50417197e-01 -3.99539679e-01 -2.67997652e-01 2.02052221e-01
7.14660108e-01 8.03247988e-01 1.16061568e+00 -1.05671060e+00
-1.06312585e+00 -5.99551797e-01 -4.47613329e-01 4.31119025e-01
-4.57658201e-01 1.03960428e-02 -8.47928584e-01 -8.27064931e-01
1.02588169e-01 -9.32879984e-01 1.56196058e-01 -3.16244096e-01
-7.65628695e-01 -3.84689450e-01 1.23870492e+00 -7.07102656e-01
9.59676921e-01 -1.97692251e+00 -2.88052738e-01 -2.42712185e-01
2.97248393e-01 1.57584891e-01 4.95459372e-03 4.97672409e-01
1.61507174e-01 1.17261797e-01 -1.19830362e-01 -6.15437388e-01
2.27708071e-02 -2.17165202e-01 -7.80020952e-01 4.84637141e-01
-3.52789938e-01 9.73291636e-01 -5.94868660e-01 -7.44907439e-01
5.17415226e-01 1.07899293e-01 -8.06843713e-02 -1.93869583e-02
-3.41272086e-01 2.28809744e-01 -4.92784142e-01 7.95034111e-01
6.93208933e-01 -6.40939534e-01 -2.42158800e-01 -1.66252136e-01
-1.67772956e-02 1.74847186e-01 -1.07850814e+00 1.83226681e+00
-4.55707997e-01 1.21020377e+00 1.08766049e-01 -9.04215813e-01
7.04946518e-01 2.05643892e-01 4.91083831e-01 -8.52476895e-01
8.24495256e-02 9.43013430e-02 -8.49237263e-01 -5.72932243e-01
8.34422708e-01 3.34126800e-01 -1.04458317e-01 7.31783092e-01
-2.20347062e-01 -1.39138654e-01 6.98120713e-01 3.93472642e-01
9.73791301e-01 -1.08953820e-04 -1.56716015e-02 -5.78417629e-02
7.24271476e-01 2.07200393e-01 -3.37942615e-02 9.39504087e-01
-6.08389117e-02 8.42297018e-01 2.08450615e-01 -6.16556168e-01
-1.22168005e+00 -7.02210069e-01 -3.47207904e-01 1.51998758e+00
5.92909865e-02 -5.23118854e-01 -8.65781307e-01 -1.00071836e+00
-2.77461141e-01 3.36533248e-01 -7.00591922e-01 4.07611072e-01
-5.69052339e-01 -1.02741158e+00 9.83863711e-01 3.84744227e-01
9.50013220e-01 -9.20220673e-01 -7.18545139e-01 -1.10823713e-01
-7.84742951e-01 -1.75329328e+00 -6.41312838e-01 1.71720520e-01
-8.43230784e-01 -1.02515161e+00 -1.10363650e+00 -7.73920178e-01
6.48002386e-01 7.29302168e-01 1.37189102e+00 -1.18573807e-01
-7.00776756e-01 4.05758917e-01 -4.31782275e-01 -4.90389079e-01
-5.88325001e-02 5.50856054e-01 -4.28990662e-01 -1.78030267e-01
4.77380246e-01 2.35422418e-01 -4.68870223e-01 5.96167684e-01
-8.01616490e-01 3.06746393e-01 6.73197448e-01 9.61848378e-01
6.32047772e-01 -2.19820693e-01 2.43740290e-01 -1.05810487e+00
4.52229470e-01 1.93381339e-01 -6.57311201e-01 4.40024614e-01
-6.37691617e-01 -7.55820274e-02 4.36993092e-01 -1.06147967e-01
-1.01006877e+00 3.99369478e-01 -1.46710828e-01 -9.93820578e-02
-2.27426350e-01 1.28921792e-01 1.31436303e-01 1.53823301e-01
7.65659213e-01 7.14794993e-01 -6.30710483e-01 -4.43332553e-01
3.61975372e-01 1.07053232e+00 3.59863549e-01 -3.15497398e-01
5.79037368e-01 9.11068082e-01 -4.50655103e-01 -1.20722461e+00
-1.31729150e+00 -1.08906329e+00 -8.44538271e-01 -6.58623204e-02
9.01438594e-01 -1.20864213e+00 -1.52205601e-01 7.00656176e-01
-1.01951468e+00 -2.67051876e-01 -6.16152957e-02 6.66165277e-02
-4.58302170e-01 5.42411089e-01 -2.34688520e-01 -5.09510159e-01
-8.07851553e-01 -1.08156717e+00 2.06415606e+00 -3.81568849e-01
2.30212197e-01 -7.48677433e-01 1.49088740e-01 9.11174417e-01
1.05811441e-02 -3.08454409e-02 5.70812285e-01 -5.79824746e-01
-3.41223568e-01 -3.91002923e-01 -5.88622630e-01 9.05637071e-02
-2.69978672e-01 -2.23851994e-01 -1.38669300e+00 -4.54365343e-01
-3.13800186e-01 -5.56459963e-01 1.45374155e+00 4.77271587e-01
1.03444529e+00 3.43536250e-02 -7.07405329e-01 4.51934487e-01
1.28051817e+00 -2.55229235e-01 4.16576803e-01 4.76358742e-01
1.05847335e+00 7.01766729e-01 6.90948904e-01 5.55080533e-01
4.10334319e-01 7.22791374e-01 2.42651820e-01 -2.00435117e-01
-2.50029385e-01 -2.06292123e-01 2.56940722e-01 6.73220754e-01
3.63521904e-01 -6.87219262e-01 -1.24680066e+00 7.41925359e-01
-1.80126595e+00 -7.98943102e-01 -5.16419470e-01 2.01908016e+00
5.04544735e-01 1.52817145e-01 1.21977843e-01 3.40069979e-01
1.06401145e+00 3.79234046e-01 -6.06549025e-01 1.03973210e-01
-3.68303984e-01 -3.27396512e-01 8.52623999e-01 8.71787965e-02
-1.61509836e+00 1.28748047e+00 6.65931845e+00 1.12822151e+00
-1.14700997e+00 1.59471080e-01 7.65384614e-01 -1.93554148e-01
4.07864988e-01 -4.24772084e-01 -1.42513561e+00 2.77705967e-01
7.11074293e-01 2.64900267e-01 4.19219173e-02 9.39507782e-01
1.79057479e-01 -1.88276291e-01 -8.60062540e-01 1.35575759e+00
5.77550411e-01 -1.39271140e+00 2.28189096e-01 -3.53913754e-02
9.58157301e-01 7.34368801e-01 2.45085150e-01 6.83742911e-02
1.03966117e-01 -1.02696908e+00 7.06190228e-01 -1.16546564e-01
1.08702338e+00 -3.44421983e-01 7.64208019e-01 3.81507277e-01
-1.36002147e+00 2.90137883e-02 -4.95944709e-01 4.35803860e-01
2.37392914e-02 7.33582079e-01 -1.06101298e+00 3.38189721e-01
9.50716615e-01 1.05976641e+00 -9.60187137e-01 7.42441475e-01
-1.08498469e-01 5.11366248e-01 -3.78177226e-01 -5.95993660e-02
4.47870821e-01 1.74732685e-01 4.12235290e-01 1.47471797e+00
-5.69968373e-02 -4.07813072e-01 3.19046170e-01 4.72692460e-01
-4.53085065e-01 4.68704969e-01 -6.16281331e-01 3.98181612e-03
1.54375717e-01 1.32591820e+00 -1.31951070e+00 -5.56089401e-01
-4.93551165e-01 1.08736777e+00 1.16600275e-01 1.36199966e-01
-5.66343844e-01 -3.36923182e-01 -1.62239298e-02 1.24782532e-01
5.04342139e-01 4.51557375e-02 -6.76302195e-01 -1.38759911e+00
3.62886041e-01 -9.57406044e-01 5.20321727e-01 -7.86143899e-01
-1.19013906e+00 6.17748737e-01 -2.74946928e-01 -1.15596569e+00
8.21614731e-03 -6.28622770e-01 -1.19203784e-01 5.89104354e-01
-1.55884171e+00 -1.66231787e+00 -7.36690581e-01 6.25327229e-01
1.33080220e+00 -2.91263074e-01 4.48728174e-01 4.04421955e-01
-6.18245482e-01 7.31271565e-01 5.62889516e-01 3.41592699e-01
1.15814769e+00 -1.18132210e+00 9.78672087e-01 6.35591805e-01
4.91763800e-01 -2.05060351e-03 4.94805485e-01 -6.35208428e-01
-1.14301157e+00 -1.24874103e+00 8.42328608e-01 -8.75391662e-01
4.63241071e-01 -8.19307625e-01 -4.96419668e-01 5.87549806e-01
1.33556873e-01 -6.70539811e-02 3.15958411e-01 7.66946143e-03
-4.07789737e-01 2.04122111e-01 -8.75341475e-01 5.30757546e-01
9.71629143e-01 -5.07039428e-01 -4.02333647e-01 8.64512146e-01
3.54976535e-01 -6.38665259e-01 -3.12895834e-01 1.78471312e-01
4.87545371e-01 -7.77339578e-01 1.03776753e+00 -1.27537280e-01
5.57204068e-01 -8.40264261e-02 -2.08526440e-02 -7.10241497e-01
3.28578681e-01 -1.59613118e-01 2.94104755e-01 1.19881976e+00
4.82807159e-01 -2.86784679e-01 1.04748762e+00 -7.96675608e-02
-8.59715417e-02 -5.47009885e-01 -8.35626721e-01 -5.61561584e-01
1.59367204e-01 -5.19636154e-01 1.38142318e-01 8.82372856e-01
-2.66955554e-01 8.47809851e-01 -5.46130300e-01 -2.08690673e-01
5.48804343e-01 2.59874403e-01 1.08754146e+00 -1.25918221e+00
1.05823256e-01 -4.25399125e-01 -1.95105433e-01 -1.39115441e+00
7.37882927e-02 -1.03280699e+00 1.14764854e-01 -1.95299733e+00
4.57421631e-01 -1.04484819e-01 3.76183212e-01 4.34621930e-01
-1.65778220e-01 4.67254430e-01 1.39633432e-01 5.32235026e-01
-1.10181391e+00 3.32072854e-01 1.05855393e+00 -5.79995394e-01
3.67097519e-02 -1.18174285e-01 -3.18330735e-01 8.55349481e-01
5.84342778e-01 -3.75455230e-01 -3.18446159e-01 -7.32960105e-01
4.38579738e-01 -3.46173286e-01 2.42604807e-01 -7.66864359e-01
3.01232904e-01 2.70861626e-01 4.35261250e-01 -1.45858526e+00
2.10181043e-01 -6.66138411e-01 -6.11005604e-01 2.01801270e-01
-6.63055837e-01 -6.55309558e-02 1.16420597e-01 6.77426755e-01
-1.39952853e-01 -2.99969673e-01 8.18401098e-01 -3.37506264e-01
-6.72136188e-01 1.75935209e-01 -2.98917919e-01 6.11790121e-01
9.12385583e-01 -2.97520787e-01 -6.31904364e-01 -1.43377140e-01
-4.19873714e-01 4.01360840e-01 3.93493086e-01 5.90743959e-01
5.86750746e-01 -8.10375631e-01 -1.01032400e+00 1.15119778e-02
6.14135385e-01 3.47606689e-02 2.30559632e-01 7.22660244e-01
-5.51534951e-01 9.43669856e-01 8.31434950e-02 -9.74244595e-01
-1.68028116e+00 4.69085634e-01 8.03203136e-02 -8.51991355e-01
-1.14574349e+00 7.87672102e-01 4.45487767e-01 -6.14996076e-01
3.88229668e-01 -1.91528469e-01 -3.30000490e-01 2.14364171e-01
6.79296613e-01 4.98447001e-01 4.22287494e-01 -5.75482905e-01
-3.04280519e-01 1.11279929e+00 -3.73607010e-01 -2.11758450e-01
1.06031406e+00 -5.45631230e-01 1.67972967e-01 3.70687902e-01
1.03061521e+00 -1.41619518e-01 -9.29838955e-01 -3.96366388e-01
1.22229546e-01 -5.51364541e-01 2.46729791e-01 -9.99403298e-01
-9.58095908e-01 1.33366740e+00 8.18293154e-01 6.86480775e-02
8.53705704e-01 -7.01556280e-02 1.15116894e+00 7.45611846e-01
2.65833475e-02 -1.68359351e+00 2.21341878e-01 7.08438516e-01
6.69737220e-01 -1.70898974e+00 1.22671969e-01 -5.70727825e-01
-6.88715518e-01 1.05451548e+00 4.60066527e-01 2.39828855e-01
2.85381734e-01 2.73117959e-01 1.60069048e-01 -4.32217926e-01
-7.47427166e-01 -2.76940048e-01 2.70141572e-01 4.46922600e-01
4.76642132e-01 -1.65819958e-01 2.49725264e-02 4.84025441e-02
-6.17874563e-02 -3.18262666e-01 3.66313010e-01 8.44270587e-01
-6.74769938e-01 -8.21096420e-01 -4.55045611e-01 8.45648289e-01
-7.48311400e-01 -4.77747589e-01 -8.56196582e-01 7.22534716e-01
-2.58678555e-01 9.95145559e-01 -1.10018983e-01 -1.30914852e-01
3.80365312e-01 5.98062016e-02 1.25900328e-01 -7.15394497e-01
-5.23854852e-01 3.30278784e-01 1.52711630e-01 -1.57892674e-01
-3.70044082e-01 -6.51638567e-01 -1.05616009e+00 -2.01187208e-01
-5.95389009e-01 -1.96156442e-01 8.59529912e-01 9.99196053e-01
3.43685091e-01 3.85307163e-01 5.56438506e-01 -6.32473409e-01
-4.19146478e-01 -1.06627524e+00 -3.57826382e-01 3.00659746e-01
1.67037874e-01 -4.28246915e-01 -2.66822934e-01 6.11060500e-01] | [11.954310417175293, 2.293041467666626] |
b8f36839-7be2-408a-905d-b61d66463c1e | periodic-load-rejection-for-floating-offshore | 2104.0425 | null | https://arxiv.org/abs/2104.04250v1 | https://arxiv.org/pdf/2104.04250v1.pdf | Periodic Load Rejection for Floating Offshore Wind Turbines via Constrained Subspace Predictive Repetitive Control | Individual Pitch Control (IPC) is an effective control strategy to mitigate the blade loads on large-scale wind turbines. Since IPC usually requires high pitch actuation, the safety constraints of the pitch actuator should be taken into account when designing the controller. This paper introduces a constrained Subspace Predictive Repetitive Control (SPRC) approach, which considers the limitation of blade pitch angle and pitch rate. To fulfill this goal, a model predictive control scheme is implemented in the fully data-driven SPRC approach to incorporate the physical limitations of the pitch actuator in the control problem formulation. An optimal control law subjected to constraints is then formulated so that future constraint violations are anticipated and prevented. Case studies show that the developed constrained SPRC reduces the pitch activities necessary to mitigate the blade loads when experiencing wind turbulence and abrupt wind gusts. More importantly, the approach allows the wind farm operator to design conservative bounds for the pitch actuator constraints that satisfies safety limitations, design specifications and physical restrictions. This will help to alleviate the cyclic fatigue loads on the actuators, increase the structural reliability and extend the lifespan of the pitch control system. | ['Jan-Willem van Wingerden', 'Riccardo M. G. Ferrari', 'Yichao Liu'] | 2021-04-09 | null | null | null | null | ['pitch-control'] | ['audio'] | [ 2.09638998e-02 3.59270066e-01 1.53679550e-01 6.08462930e-01
5.32575130e-01 -9.73829448e-01 9.57653448e-02 -3.30964215e-02
1.83528483e-01 6.96978867e-01 2.28248164e-01 -1.74240083e-01
-9.15837586e-01 -5.59597611e-01 -1.48321241e-01 -9.69535112e-01
-9.41459090e-02 -1.58137619e-01 1.03865430e-01 -4.55295175e-01
1.45969361e-01 8.85278642e-01 -1.64006650e+00 -5.39146662e-01
7.10984349e-01 6.13668382e-01 5.95107555e-01 6.53313160e-01
7.52583683e-01 2.34524772e-01 -4.70054954e-01 6.61457241e-01
3.47423881e-01 -1.71665132e-01 -5.22950292e-01 5.10008395e-01
-3.43197137e-01 -5.14155865e-01 6.37569666e-01 9.30272996e-01
8.19680989e-01 7.75913000e-01 5.23287177e-01 -9.14103031e-01
8.17195103e-02 3.05739820e-01 -4.79927033e-01 5.54338358e-02
-1.33268192e-01 1.38284847e-01 7.63768613e-01 -7.80655980e-01
2.96153873e-01 8.84760261e-01 5.51510334e-01 1.93661317e-01
-1.14296114e+00 -8.24359730e-02 -1.81950759e-02 -4.24992114e-01
-1.12212861e+00 -2.05380857e-01 6.25127554e-01 -8.57003152e-01
8.17196846e-01 4.84046072e-01 8.44948769e-01 2.33397856e-01
5.75394154e-01 -3.33553880e-01 6.84398174e-01 -3.93896222e-01
4.31659520e-01 -3.35497469e-01 -3.01667958e-01 1.58120617e-01
7.45732009e-01 3.98009598e-01 -1.81397591e-02 4.10582013e-02
8.13199103e-01 -3.20584148e-01 -6.31058633e-01 -3.17465693e-01
-6.68684661e-01 6.67155147e-01 7.62937590e-02 1.25425294e-01
-6.95632041e-01 -3.92134756e-01 6.46312714e-01 -6.72854260e-02
1.73439771e-01 1.01347458e+00 -6.97496533e-01 1.68258592e-01
-7.41180539e-01 4.34466451e-01 7.01585650e-01 4.40159082e-01
-1.85540006e-01 9.04373229e-01 2.66383201e-01 4.68305677e-01
1.68668255e-01 6.87408209e-01 1.10176347e-01 -1.08185458e+00
-2.45045330e-02 3.23371142e-01 5.66643417e-01 -9.39933181e-01
-5.45162678e-01 -5.94396710e-01 -5.57997584e-01 5.08198798e-01
-3.18558961e-01 -8.55585217e-01 -3.51309717e-01 1.29457438e+00
6.25210226e-01 -4.66462642e-01 2.30404451e-01 1.23957539e+00
-2.00381890e-01 8.64979148e-01 -4.04620357e-02 -1.02689648e+00
1.44762588e+00 -3.42986584e-01 -1.25884509e+00 -3.45759802e-02
6.60251796e-01 -8.72093558e-01 7.44578719e-01 4.90734994e-01
-8.85822058e-01 -5.98640382e-01 -1.30953598e+00 7.80425131e-01
-3.77714522e-02 7.81049728e-01 -2.48839259e-01 3.86872232e-01
-6.21628582e-01 6.31727934e-01 -8.00770998e-01 -2.23534390e-01
-6.74503982e-01 2.15772331e-01 9.92368255e-03 7.79871047e-01
-1.12071681e+00 1.34775925e+00 6.45838618e-01 5.76544940e-01
-3.84736568e-01 -9.51133251e-01 -5.73237598e-01 1.37845308e-01
4.32323724e-01 -6.17337584e-01 1.12818670e+00 -3.64131212e-01
-1.86951828e+00 -2.55731761e-01 3.34904641e-01 -4.07232195e-01
2.12639824e-01 -7.54354358e-01 -6.28068149e-02 3.25926214e-01
-1.00555092e-01 -2.72400975e-02 1.11400366e+00 -9.82261300e-01
-5.62947690e-01 1.44121557e-01 -3.00416946e-01 4.67035919e-01
-4.15766150e-01 -3.34891587e-01 7.46492147e-01 -7.15426147e-01
-3.26402903e-01 -1.18712461e+00 -4.11933661e-01 -3.60807449e-01
-2.87603498e-01 -2.87837088e-02 1.32024407e+00 -6.87464774e-01
1.33360958e+00 -2.16372752e+00 6.53894484e-01 8.20084512e-02
-6.38444960e-01 6.52596951e-01 4.93837655e-01 1.05036986e+00
-3.15926708e-02 -1.19462259e-01 1.17819207e-02 3.88449758e-01
-5.39944112e-01 5.91273010e-01 -5.84666789e-01 3.70190859e-01
3.14730495e-01 -2.89904833e-01 -4.19085920e-01 2.16884777e-01
4.75656927e-01 1.94934860e-01 -5.53651869e-01 4.33400869e-01
-1.97648272e-01 3.49026442e-01 -7.05633700e-01 1.60478100e-01
5.51039636e-01 5.92290938e-01 2.27296919e-01 -4.63117003e-01
-9.58738685e-01 -4.67354000e-01 -1.42062640e+00 6.57024920e-01
-6.02288842e-01 -1.72969803e-01 1.16178203e+00 -9.64296937e-01
1.15376377e+00 8.48327339e-01 4.81595963e-01 3.53027910e-01
1.74275458e-01 2.91812718e-01 5.45446388e-02 -6.73022270e-01
3.82994145e-01 -2.21531868e-01 3.76943767e-01 -3.40211421e-01
-3.30689967e-01 -7.62295961e-01 1.45525336e-01 -3.46006185e-01
3.35390985e-01 1.83844611e-01 6.63465202e-01 -1.03040409e+00
1.04227817e+00 2.17760310e-01 1.11575997e+00 -5.06313801e-01
2.26954937e-01 -2.46179283e-01 1.80023178e-01 -4.84692864e-02
-1.05048907e+00 -3.64423245e-01 -6.84924647e-02 7.94897616e-01
-1.01720922e-01 8.91778246e-02 -5.74034929e-01 2.58027852e-01
1.72007039e-01 7.43857324e-01 4.24666069e-02 -3.06234986e-01
-7.24338889e-01 -3.89096528e-01 -3.06446373e-01 6.52595997e-01
1.34737864e-01 -4.91257042e-01 -1.41807592e+00 5.09976625e-01
5.48685074e-01 -8.80868137e-01 -4.61656541e-01 3.51061881e-01
-9.43779469e-01 -1.21329129e+00 -1.10742941e-01 -5.59139848e-01
7.85084784e-01 1.32038042e-01 1.34242982e-01 2.59624183e-01
-2.58711666e-01 2.54996091e-01 -5.59093177e-01 -5.46475947e-01
-2.25529298e-01 -1.19467489e-01 7.46859074e-01 -4.51153576e-01
-9.05514598e-01 -3.56283188e-01 -3.50774020e-01 6.66566610e-01
-6.82968020e-01 1.41240912e-03 2.97348768e-01 1.03935194e+00
5.47845244e-01 8.62099469e-01 1.02860689e+00 -3.34896415e-01
7.20127165e-01 -1.32784441e-01 -1.24371123e+00 -1.72451764e-01
-6.97968483e-01 -2.67056674e-01 1.45720315e+00 -5.76534033e-01
-1.23329723e+00 4.42411900e-01 1.17675722e-01 -5.62063217e-01
1.70415069e-03 7.53339231e-01 -3.05610031e-01 9.41263661e-02
2.00317815e-01 -1.14650309e-01 5.43672979e-01 -6.79892480e-01
2.88091928e-01 3.05973500e-01 2.67915308e-01 -2.96973169e-01
1.14077699e+00 -4.75773141e-02 8.73695016e-01 -1.29120791e+00
-6.04207635e-01 -3.82463425e-01 -7.05742657e-01 -3.21736515e-01
7.08150208e-01 -8.40848029e-01 -6.82959080e-01 1.41783070e-03
-7.83294559e-01 -1.02277279e-01 -3.81122202e-01 6.98195636e-01
-6.41668856e-01 2.01492339e-01 -3.24979454e-01 -1.36320484e+00
-8.82066548e-01 -1.10301161e+00 6.02457106e-01 3.27172071e-01
-5.01874089e-01 -9.69889581e-01 -9.60906595e-02 -1.16971716e-01
4.85557526e-01 8.19111586e-01 8.34693372e-01 1.02814943e-01
1.20098822e-01 -6.50534704e-02 7.71972656e-01 5.90942502e-01
3.32756788e-01 4.39615279e-01 -2.73690820e-01 -9.51392412e-01
3.41527998e-01 2.74193324e-02 -1.25373960e-01 5.08479416e-01
3.97294253e-01 -7.60169327e-01 -2.43898928e-01 2.92209744e-01
1.77205122e+00 3.14967006e-01 -2.59074837e-01 5.73761724e-02
3.78208399e-01 1.09060371e+00 1.34348667e+00 9.30099010e-01
-5.46398878e-01 6.26987815e-01 8.21637571e-01 4.28835861e-02
3.56170028e-01 3.72343836e-03 5.08397520e-01 7.81914890e-01
-1.12418264e-01 -1.85841575e-01 -4.37563330e-01 7.85456061e-01
-1.56824160e+00 -7.51395285e-01 -5.10611117e-01 2.05658674e+00
4.96428967e-01 -1.87362462e-01 -1.91230431e-01 6.88937664e-01
9.36319053e-01 -1.33993113e-02 -2.33993322e-01 -1.21697664e+00
2.03000486e-01 -6.77695349e-02 8.56413066e-01 6.50204599e-01
-1.04145539e+00 2.74277806e-01 5.46293068e+00 4.49377090e-01
-1.24626589e+00 -5.38119435e-01 -2.79854059e-01 -6.27947301e-02
3.58774781e-01 1.38748363e-01 -8.62549305e-01 3.65122527e-01
8.75860095e-01 -7.40245223e-01 3.00584495e-01 1.15647745e+00
1.26802003e+00 -2.73091376e-01 -3.84764999e-01 -1.37243167e-01
-4.97261256e-01 -8.55954647e-01 -2.26475820e-01 1.82654917e-01
6.86536789e-01 -7.85190403e-01 -1.28638372e-01 -3.04420739e-01
-2.54235297e-01 -2.99423277e-01 7.57163107e-01 4.43579465e-01
5.13425469e-01 -1.05872095e+00 9.21563506e-01 8.39898705e-01
-1.52243054e+00 -5.71748734e-01 -2.78664023e-01 -6.04672670e-01
8.77363086e-01 6.86543345e-01 -9.29597497e-01 8.26331735e-01
3.50341737e-01 4.78734881e-01 9.77811739e-02 6.09439015e-01
-4.03075129e-01 6.06692433e-01 -4.54159111e-01 -5.32644391e-02
1.60624206e-01 -3.86262804e-01 1.08120322e+00 3.44697565e-01
5.20257771e-01 4.80329067e-01 6.85410202e-01 5.10001779e-01
9.58700955e-01 3.67427260e-01 -7.17198431e-01 -9.36807618e-02
8.27621818e-01 1.44843137e+00 -1.47884309e-01 2.78445929e-01
2.13406980e-01 1.84489757e-01 -4.80845869e-01 -2.09806655e-02
-5.84716082e-01 -4.27960753e-01 9.48944092e-01 4.99861568e-01
3.19607884e-01 -5.86257696e-01 -2.12209478e-01 4.98545691e-02
-1.01271914e-02 -3.56183052e-01 6.90932795e-02 -6.91562176e-01
-1.04009748e+00 5.09207090e-03 3.38781118e-01 -1.59749198e+00
-6.70227885e-01 -6.65262699e-01 -1.05368853e+00 1.20905232e+00
-1.07752466e+00 -7.32861102e-01 2.77045429e-01 -1.36536807e-01
7.31569409e-01 -1.03958966e-02 8.08046818e-01 -2.35293314e-01
-8.34470987e-01 -4.65367287e-01 2.16990888e-01 -5.80843270e-01
2.37479746e-01 -9.09183025e-01 -5.58352530e-01 1.20683753e+00
-1.35223627e+00 6.49726510e-01 1.37761152e+00 -9.74824071e-01
-1.83474159e+00 -1.34380913e+00 5.43919027e-01 6.23037815e-01
8.39708567e-01 2.75969297e-01 -8.74972463e-01 3.71193253e-02
2.86506057e-01 -7.59274438e-02 3.14008385e-01 -6.34693742e-01
8.75153005e-01 -3.84564586e-02 -1.03810608e+00 4.04741883e-01
2.10400209e-01 8.47339705e-02 -6.28203690e-01 2.37321973e-01
8.47314298e-01 -3.86089265e-01 -1.34864414e+00 8.13979447e-01
1.76916778e-01 -1.10345269e-02 6.68530703e-01 -3.47789615e-01
-3.30999941e-02 -7.66296923e-01 3.39101076e-01 -1.69819283e+00
-2.17382520e-01 -9.75394666e-01 -1.68927461e-01 1.37471461e+00
3.79266776e-02 -3.90022695e-01 2.00898185e-01 1.68734536e-01
-4.91948485e-01 -9.62676883e-01 -9.77749765e-01 -9.11287367e-01
-5.62387221e-02 6.58214390e-01 7.21032545e-02 6.94399834e-01
3.70621413e-01 3.38176280e-01 -5.86506546e-01 1.16790998e+00
2.18297288e-01 -1.08448073e-01 4.78230298e-01 -1.23353064e+00
-1.17380975e-03 7.06963465e-02 2.77252764e-01 -3.19642425e-02
3.09012942e-02 -1.77212078e-02 2.09128216e-01 -1.46609199e+00
-1.05636692e+00 1.78889379e-01 4.99800891e-01 3.66321921e-01
-2.79118985e-01 -5.75338721e-01 6.48192614e-02 1.45967689e-03
7.22658217e-01 7.64642119e-01 1.43998575e+00 3.79760087e-01
-4.74020451e-01 3.59973431e-01 1.55197948e-01 6.08826578e-01
1.02291620e+00 1.36759421e-02 -1.13264036e+00 -2.31091052e-01
-3.22468653e-02 6.24525487e-01 -2.31119283e-02 -1.20653355e+00
9.31522921e-02 -6.05657458e-01 -1.57335922e-01 -8.95919323e-01
3.63222440e-03 -1.24631751e+00 8.29146504e-01 1.19372392e+00
-5.52328005e-02 4.04344946e-01 3.03864986e-01 5.01277864e-01
-2.01542273e-01 -2.38771349e-01 8.94073069e-01 2.62712777e-01
-3.70853722e-01 -4.03334528e-01 -9.37338769e-01 -5.69439948e-01
1.40213609e+00 -1.34149879e-01 -1.94212750e-01 1.43624589e-01
-7.79117167e-01 7.19204843e-01 1.09099939e-01 3.59490424e-01
3.37082386e-01 -6.13150835e-01 -3.57762545e-01 1.04378566e-01
-5.80513716e-01 7.58446194e-03 5.71686745e-01 8.41488421e-01
-6.57383323e-01 7.08652318e-01 -5.19550920e-01 -1.98495001e-01
-1.42078805e+00 6.42082751e-01 4.29996759e-01 4.24275696e-02
-6.02190435e-01 3.69772464e-01 -2.71396160e-01 7.97741488e-02
-4.13457632e-01 -2.73333162e-01 -7.03736484e-01 2.46113524e-01
2.21256420e-01 7.02438235e-01 1.84084520e-01 -5.25426030e-01
-1.76654384e-01 8.23583722e-01 6.37653232e-01 1.06117809e-02
1.19010258e+00 -3.42718303e-01 1.72475222e-02 9.02262032e-02
3.79772723e-01 2.90944904e-01 -1.45722830e+00 6.60046816e-01
-2.49134779e-01 -2.05017284e-01 3.87924671e-01 -5.29786646e-01
-8.51836681e-01 3.24307263e-01 2.64025271e-01 3.18381310e-01
1.37643135e+00 -1.02646959e+00 2.89595693e-01 1.05959110e-01
2.47690663e-01 -1.31498420e+00 -3.00588191e-01 6.82396948e-01
1.17504978e+00 -1.58759758e-01 4.99319613e-01 -9.20554161e-01
-7.28813231e-01 1.23569667e+00 7.33104050e-01 -4.23907489e-01
6.22876525e-01 5.98428249e-01 -2.08918095e-01 2.47855306e-01
-8.71851802e-01 -6.14750758e-02 6.02635406e-02 3.35391134e-01
5.16185105e-01 2.20353201e-01 -1.25282049e+00 5.89134991e-01
-3.72400731e-02 -2.38008544e-01 7.06720769e-01 1.10066652e+00
-9.48124588e-01 -8.98984611e-01 -7.59524822e-01 5.79825789e-02
-4.26048875e-01 3.76192898e-01 1.45121306e-01 9.42622244e-01
5.47364354e-01 9.33601141e-01 -1.70679480e-01 -1.75882831e-01
8.83437991e-01 3.17668095e-02 -4.39874887e-01 -8.02779078e-01
-7.11129725e-01 3.74431938e-01 3.32932681e-01 -1.29598275e-01
-8.80247056e-02 -5.49434125e-01 -1.49556267e+00 1.66405842e-01
-9.71613944e-01 8.03201914e-01 4.77712840e-01 6.50405407e-01
4.37217295e-01 8.13026607e-01 9.68533993e-01 -8.22375715e-01
-1.23336613e+00 -9.15715456e-01 -8.83970678e-01 -3.53626341e-01
1.17877945e-01 -1.32290077e+00 -6.57903552e-01 2.48712420e-01] | [5.402334213256836, 2.484386444091797] |
f978fa04-22a5-43b9-ac7b-734dd93b422a | complex-word-identification-convolutional | null | null | https://aclanthology.org/W18-0538 | https://aclanthology.org/W18-0538.pdf | Complex Word Identification: Convolutional Neural Network vs. Feature Engineering | We describe the systems of NLP-CIC team that participated in the Complex Word Identification (CWI) 2018 shared task. The shared task aimed to benchmark approaches for identifying complex words in English and other languages from the perspective of non-native speakers. Our goal is to compare two approaches: feature engineering and a deep neural network. Both approaches achieved comparable performance on the English test set. We demonstrated the flexibility of the deep-learning approach by using the same deep neural network setup in the Spanish track. Our systems achieved competitive results: all our systems were within 0.01 of the system with the best macro-F1 score on the test sets except on Wikipedia test set, on which our best system is 0.04 below the best macro-F1 score. | ['er', 'ro', "Daniel Alej P{\\'e}rez Alvarez", 'Segun Taofeek Aroyehun', 'Jason Angel', 'Alex Gelbukh'] | 2018-06-01 | null | null | null | ws-2018-6 | ['complex-word-identification'] | ['natural-language-processing'] | [-3.37899476e-01 1.50196835e-01 1.25967085e-01 -1.71676591e-01
-1.12078679e+00 -8.88721883e-01 7.31271088e-01 9.13769454e-02
-1.09405029e+00 5.31818509e-01 2.98237741e-01 -4.56156760e-01
1.01198711e-01 -3.80172104e-01 -4.13190007e-01 -1.94682002e-01
6.05545491e-02 7.20852375e-01 -1.16578132e-01 -4.46237236e-01
5.25234547e-03 2.68327534e-01 -1.33950102e+00 4.05013263e-01
9.46720123e-01 6.83387220e-01 6.96950629e-02 8.41798842e-01
-2.96386361e-01 6.07100546e-01 -8.39179635e-01 -5.73239326e-01
2.23072007e-01 1.43292755e-01 -9.40401554e-01 -7.86846578e-01
7.40623057e-01 -4.34862748e-02 -2.16498390e-01 1.09194088e+00
9.50730264e-01 -1.68897081e-02 7.52374530e-01 -8.54580402e-01
-6.52810991e-01 1.17103291e+00 1.60659298e-01 1.24771707e-01
6.71474993e-01 6.50671124e-02 1.29417276e+00 -1.41460824e+00
5.61032116e-01 1.38175201e+00 9.42118168e-01 6.57067299e-01
-1.00936723e+00 -8.75453651e-01 2.25761712e-01 1.95611641e-01
-1.65031028e+00 -7.28013337e-01 6.49330541e-02 -5.23582399e-01
1.68840671e+00 1.23555986e-02 4.15155023e-01 1.31291020e+00
-1.53005138e-01 1.05953252e+00 9.34937716e-01 -7.52560794e-01
-1.51011184e-01 2.42719337e-01 7.66048074e-01 7.43956029e-01
4.27391827e-01 8.79745632e-02 -5.24087608e-01 3.89976194e-03
-6.82738191e-03 -7.85132229e-01 -3.08053613e-01 3.74246329e-01
-1.33403361e+00 9.99415815e-01 -1.13045603e-01 6.77691460e-01
9.41100530e-03 -5.19845411e-02 5.10008633e-01 5.19509971e-01
4.42953825e-01 9.03585315e-01 -1.17087972e+00 -2.41184264e-01
-6.37298286e-01 2.36417294e-01 1.40000832e+00 1.09738243e+00
1.83328807e-01 1.06813788e-01 -4.70335633e-01 1.04004073e+00
1.37050033e-01 5.97550035e-01 8.67762804e-01 -3.97463709e-01
6.28799856e-01 3.39907885e-01 -9.14766490e-02 -5.68266571e-01
-7.17663884e-01 -7.03953683e-01 -4.32164252e-01 8.23981240e-02
9.04681385e-01 -4.40655708e-01 -9.91771400e-01 1.72505951e+00
-3.34130466e-01 -4.12974685e-01 2.99423277e-01 2.48888686e-01
1.45067573e+00 6.06947839e-01 4.22997802e-01 2.69889206e-01
1.50758123e+00 -1.05995488e+00 -6.14505827e-01 -2.58918226e-01
1.15269780e+00 -9.97742951e-01 1.35211420e+00 4.48550403e-01
-1.06216025e+00 -6.05735362e-01 -1.13300610e+00 -1.02279976e-01
-8.96241903e-01 4.00623590e-01 1.97680727e-01 8.87848318e-01
-1.09243238e+00 4.47888672e-01 -3.07739705e-01 -4.02274609e-01
5.51697388e-02 2.79604763e-01 -6.73345923e-01 1.02618694e-01
-1.42458856e+00 1.24315321e+00 4.51244533e-01 -1.74223855e-01
-6.99208796e-01 -1.10975516e+00 -9.07873034e-01 2.78226584e-01
9.21484083e-02 -1.52308673e-01 1.52574456e+00 -8.27733576e-01
-1.38977706e+00 1.35819995e+00 -5.91860004e-02 -4.26624507e-01
5.36970198e-01 -6.51155353e-01 -7.31642663e-01 -3.72683406e-01
2.06146121e-01 6.12230361e-01 4.06025112e-01 -7.89535701e-01
-7.47207820e-01 -5.86452559e-02 -1.09328203e-01 9.10449401e-02
-4.13820326e-01 3.03598702e-01 -3.47169459e-01 -4.74049568e-01
-5.16515911e-01 -8.71945679e-01 8.63643438e-02 -5.87118626e-01
-7.07189620e-01 -8.83542180e-01 1.36049151e-01 -9.32869792e-01
1.21478593e+00 -2.14083052e+00 -3.30336802e-02 7.16404244e-02
2.38438398e-01 7.86660194e-01 -5.69185436e-01 3.66774827e-01
-2.87111700e-01 5.81329942e-01 1.16587535e-01 -5.28107345e-01
3.26275229e-01 -3.65239501e-01 -5.74092939e-02 1.74153849e-01
3.92348200e-01 9.50221896e-01 -8.75513196e-01 -1.88610613e-01
-2.86371678e-01 3.38962585e-01 -2.80606061e-01 1.67025164e-01
-4.78798002e-02 -6.08517304e-02 2.57173598e-01 5.56622088e-01
5.21615565e-01 2.44727552e-01 1.55129015e-01 -1.82269454e-01
-3.81869406e-01 6.23057127e-01 -1.16964591e+00 1.31796014e+00
-6.66297674e-01 1.00170660e+00 -5.36486842e-02 -7.12504745e-01
7.26046085e-01 5.05218327e-01 -9.79135260e-02 -7.91840851e-01
-8.21251869e-02 5.92505455e-01 5.86566925e-01 -2.12264076e-01
2.77559549e-01 3.60973865e-01 -3.15621823e-01 2.70906180e-01
4.74395365e-01 -3.32872644e-02 4.35033590e-01 4.47687469e-02
9.67003584e-01 -1.51217520e-01 5.34684896e-01 -9.64927256e-01
7.40822077e-01 -3.94937806e-02 2.93215930e-01 1.00273156e+00
-2.81082004e-01 5.42741358e-01 4.87925828e-01 -6.68585598e-01
-8.53516221e-01 -1.10888696e+00 -3.21291953e-01 1.61830068e+00
-4.82807666e-01 -7.80497372e-01 -8.37923825e-01 -6.84145033e-01
-1.09487832e-01 8.41785848e-01 -5.61824322e-01 -3.73434052e-02
-7.11410642e-01 -4.67088759e-01 1.13336980e+00 4.88148659e-01
2.57600248e-01 -1.29125297e+00 -9.30544659e-02 2.63652533e-01
-2.01735541e-01 -1.25860953e+00 -5.95852494e-01 3.19137037e-01
2.39282891e-01 -9.59062636e-01 -7.56104946e-01 -1.28785241e+00
-2.45134272e-02 -3.14658821e-01 1.73717272e+00 1.43796384e-01
-3.88163060e-01 3.27167392e-01 -2.42077321e-01 -8.35245252e-01
-5.55137396e-01 8.52228940e-01 1.19960032e-01 -5.43809652e-01
8.45582306e-01 1.33126974e-01 -1.71567634e-01 -1.26608893e-01
-2.73535550e-01 -3.03689033e-01 2.68503457e-01 8.68209720e-01
1.28900468e-01 -4.49997008e-01 5.28968990e-01 -8.11155677e-01
8.49157691e-01 -3.31364870e-01 -6.25465572e-01 4.86753881e-01
-5.39875686e-01 1.58916429e-01 5.17942667e-01 -5.57740688e-01
-6.05879962e-01 1.98561952e-01 -6.86842859e-01 5.29978275e-01
-2.18887106e-01 5.22535801e-01 -4.36586440e-01 6.72165006e-02
6.74720883e-01 -2.62544933e-03 -4.27741379e-01 -6.83587074e-01
4.50507581e-01 7.89041162e-01 4.77743924e-01 -5.13560355e-01
4.73142326e-01 -4.32737291e-01 -5.74107707e-01 -1.02074313e+00
-1.02034938e+00 -4.33129102e-01 -8.02053034e-01 2.48799920e-01
8.58613133e-01 -1.26742315e+00 -7.09448397e-01 8.19674075e-01
-1.49730754e+00 -4.34340686e-01 -1.43230958e-02 3.48955661e-01
-1.00358658e-01 1.12160444e-01 -5.18993378e-01 -4.83498961e-01
-8.01666856e-01 -1.18976986e+00 1.07749426e+00 -1.49100861e-02
-6.66325450e-01 -1.16105294e+00 1.70159698e-01 -1.05994217e-01
7.06247747e-01 4.40782793e-02 1.12111950e+00 -1.36553872e+00
2.44612560e-01 -5.74431159e-02 -1.88807338e-01 4.34148073e-01
-2.87771225e-01 7.10243732e-02 -1.26467991e+00 -1.63910061e-01
-5.43836176e-01 -3.62418085e-01 9.02261853e-01 1.36461034e-01
7.76314318e-01 -7.66327158e-02 -1.82038814e-01 7.18249977e-01
1.12989795e+00 -1.59794595e-02 3.62569213e-01 5.40967405e-01
7.96137333e-01 6.57110393e-01 6.43860847e-02 1.28937721e-01
5.15656948e-01 6.00874484e-01 -1.17949665e-01 -5.76690361e-02
-4.55804437e-01 -9.74436756e-04 5.75695872e-01 8.84391904e-01
2.88103223e-01 -3.75044912e-01 -1.59034956e+00 7.80795336e-01
-1.35970581e+00 -6.90769792e-01 -3.54067653e-01 1.90471113e+00
1.03102887e+00 2.60960698e-01 1.43305495e-01 1.04227372e-01
5.96753895e-01 -2.18197241e-01 -8.00834149e-02 -7.75274038e-01
-5.18375337e-01 6.90524042e-01 2.43534848e-01 8.84518206e-01
-1.34139204e+00 1.43556261e+00 7.07755661e+00 8.77996206e-01
-9.14533675e-01 2.23140150e-01 3.42400938e-01 2.48355102e-02
-2.30930254e-01 -5.65857410e-01 -1.53016233e+00 2.50737965e-01
1.45634317e+00 -5.40235579e-01 4.45157856e-01 7.65263200e-01
-5.13820529e-01 2.28061214e-01 -1.31395960e+00 9.16573405e-01
1.02273323e-01 -9.93434191e-01 -7.03009218e-02 -9.27178785e-02
6.07052207e-01 7.55455017e-01 -3.26276086e-02 1.00009525e+00
5.49455464e-01 -1.50344038e+00 7.58570790e-01 4.20887411e-01
1.08827257e+00 -8.89087200e-01 8.36687565e-01 2.45485589e-01
-1.12771463e+00 5.27477823e-02 -6.32861778e-02 3.70981311e-03
-2.94480026e-01 5.57529092e-01 -7.22036481e-01 -3.70483622e-02
7.78743744e-01 5.09900331e-01 -8.29423010e-01 8.28097999e-01
-3.48941386e-01 8.15740108e-01 -4.30790722e-01 -4.35556024e-01
2.87038803e-01 3.41170967e-01 5.45318007e-01 1.91881812e+00
1.06975101e-01 -3.44245404e-01 1.94714963e-01 4.41935867e-01
-2.03846321e-01 5.43536305e-01 -6.85354114e-01 -2.24262983e-01
5.23547471e-01 1.34172571e+00 -1.41755879e-01 -4.69138861e-01
-4.25891399e-01 7.61971474e-01 7.82971621e-01 1.64210364e-01
-3.44499469e-01 -9.18851793e-01 8.61187935e-01 -2.65809298e-01
1.75995737e-01 -3.51975411e-01 -4.01062071e-01 -1.21313500e+00
-2.00427756e-01 -1.20200920e+00 3.67210209e-01 -9.65473056e-02
-1.46464288e+00 1.18247664e+00 -2.14560851e-01 -7.26269424e-01
-6.14291310e-01 -1.32547879e+00 -5.50487101e-01 1.20756161e+00
-1.43795204e+00 -9.04962778e-01 -1.54668853e-01 4.77996230e-01
5.08436322e-01 -8.94271433e-01 1.45727456e+00 3.17940056e-01
-5.07540464e-01 1.27926004e+00 2.85540283e-01 5.36906004e-01
8.72931182e-01 -1.56215084e+00 9.86748457e-01 7.32326567e-01
3.59729618e-01 6.85019016e-01 3.53274226e-01 -3.13518345e-01
-1.05850470e+00 -8.73002589e-01 1.59158206e+00 -6.60808563e-01
8.85110795e-01 -7.68582046e-01 -7.20934927e-01 5.67211211e-01
5.28942883e-01 -1.82735801e-01 7.65468955e-01 5.86067200e-01
-6.62658274e-01 2.28579536e-01 -9.94858384e-01 6.31075680e-01
9.98432994e-01 -7.43853688e-01 -6.48525536e-01 5.70531249e-01
9.30135429e-01 -2.65263677e-01 -8.67576361e-01 3.26199293e-01
8.37527931e-01 -5.50788820e-01 8.78984213e-01 -8.99770141e-01
1.59246117e-01 1.04067661e-01 -2.90881276e-01 -1.86128676e+00
-5.65485597e-01 -4.56643969e-01 1.65835500e-01 1.28872681e+00
9.97914195e-01 -6.76594317e-01 1.62074536e-01 3.06039035e-01
-1.02247261e-01 -5.14671564e-01 -9.47369158e-01 -9.15595889e-01
7.31506586e-01 -6.52472734e-01 4.62956816e-01 8.05713117e-01
-1.84415817e-01 7.27793753e-01 2.12013632e-01 1.64198678e-03
1.90637648e-01 -5.11685252e-01 5.54344654e-01 -1.40791571e+00
-2.07176104e-01 -9.02827919e-01 -2.93594480e-01 -3.48219931e-01
7.59415030e-01 -1.31224084e+00 1.76584929e-01 -1.27338004e+00
-1.67251676e-02 -1.13641173e-01 -2.41714969e-01 6.36746287e-01
-4.09396023e-01 2.36281112e-01 4.60252821e-01 -2.21266031e-01
-3.32460701e-01 1.80858523e-01 6.08072281e-01 -4.59687859e-01
-1.19388765e-02 -1.69499889e-01 -8.85621965e-01 7.12023854e-01
9.63372588e-01 -4.49947685e-01 6.78168982e-02 -9.68200266e-01
2.76391655e-01 -8.94718289e-01 -3.13228399e-01 -1.15349460e+00
9.00967717e-02 4.60296869e-01 3.27710211e-01 -4.66262043e-01
-4.48779352e-02 -3.46617371e-01 -3.80527854e-01 7.26156592e-01
-3.70338142e-01 3.56350332e-01 5.91105998e-01 -2.41221428e-01
-1.33473024e-01 -5.32723486e-01 7.13165581e-01 -9.59803760e-02
-7.49400675e-01 -2.71155988e-03 -6.98836267e-01 5.89055538e-01
5.51412106e-01 2.58972794e-01 -3.87280762e-01 -1.76114634e-01
-6.73736274e-01 1.79527745e-01 -2.59755641e-01 7.04541683e-01
1.37638837e-01 -1.21492434e+00 -1.29667902e+00 3.81808251e-01
4.52337801e-01 -5.54060757e-01 -4.56589937e-01 2.87291408e-01
-6.78973615e-01 6.82623982e-01 3.93633991e-02 -4.07534659e-01
-1.16212261e+00 1.87415659e-01 6.87669516e-01 -6.12647116e-01
-2.74695605e-01 1.10943127e+00 9.39670876e-02 -1.11441195e+00
5.93808174e-01 -2.71816134e-01 -4.78557378e-01 2.37331420e-01
9.94115591e-01 3.81237298e-01 5.27493477e-01 -6.45702302e-01
-6.19307876e-01 5.98459542e-01 -3.93675566e-01 -2.78288752e-01
1.20620799e+00 2.59554833e-01 -3.46931517e-02 5.49195588e-01
1.58775282e+00 3.13138366e-01 -2.24151388e-01 -3.20341170e-01
4.51600492e-01 3.04184675e-01 7.92200416e-02 -1.24925923e+00
-6.60244107e-01 8.92033637e-01 8.32234561e-01 1.83337182e-01
5.80745459e-01 -1.30431816e-01 7.16772735e-01 1.00908351e+00
-8.59406665e-02 -1.37278008e+00 -3.12443137e-01 1.55491471e+00
1.04851115e+00 -1.48792064e+00 -3.95023733e-01 -2.83938885e-01
-4.57485467e-01 1.36454880e+00 6.40535176e-01 -1.28279775e-01
9.36409831e-01 4.95805323e-01 3.02227497e-01 -5.31994961e-02
-7.70301402e-01 -5.33220768e-01 7.98703015e-01 7.42559195e-01
9.89418089e-01 3.83772552e-01 -4.99706000e-01 1.09173942e+00
-6.87976182e-01 -5.03519058e-01 2.34126359e-01 4.17126268e-01
-2.89947867e-01 -9.58646715e-01 -6.58798739e-02 2.78180182e-01
-7.08008111e-01 -7.32492089e-01 -6.62396908e-01 1.13726747e+00
1.25389919e-01 1.03940654e+00 1.14738636e-01 -3.47560495e-01
6.78247452e-01 5.59351444e-01 3.21224332e-01 -8.77569556e-01
-1.05212379e+00 -3.22474778e-01 5.88547170e-01 -6.00782752e-01
2.18232229e-01 -6.68334961e-01 -1.07297289e+00 -9.01104063e-02
-1.44238904e-01 6.56094998e-02 8.92584920e-01 1.00010276e+00
1.06094286e-01 2.98386604e-01 2.98824221e-01 -5.74205935e-01
-6.47426307e-01 -1.50723815e+00 -3.69599462e-01 2.40601614e-01
3.55217159e-01 -2.47479662e-01 -4.21933025e-01 -1.91259742e-01] | [10.450173377990723, 10.439269065856934] |
14d97d37-c74b-49eb-885c-625e890d9aec | highly-accurate-quantum-chemical-property | 2303.16982 | null | https://arxiv.org/abs/2303.16982v2 | https://arxiv.org/pdf/2303.16982v2.pdf | Highly Accurate Quantum Chemical Property Prediction with Uni-Mol+ | Recent developments in deep learning have made remarkable progress in speeding up the prediction of quantum chemical (QC) properties by removing the need for expensive electronic structure calculations like density functional theory. However, previous methods learned from 1D SMILES sequences or 2D molecular graphs failed to achieve high accuracy as QC properties primarily depend on the 3D equilibrium conformations optimized by electronic structure methods, far different from the sequence-type and graph-type data. In this paper, we propose a novel approach called Uni-Mol+ to tackle this challenge. Uni-Mol+ first generates a raw 3D molecule conformation from inexpensive methods such as RDKit. Then, the raw conformation is iteratively updated to its target DFT equilibrium conformation using neural networks, and the learned conformation will be used to predict the QC properties. To effectively learn this update process towards the equilibrium conformation, we introduce a two-track Transformer model backbone and train it with the QC property prediction task. We also design a novel approach to guide the model's training process. Our extensive benchmarking results demonstrate that the proposed Uni-Mol+ significantly improves the accuracy of QC property prediction in various datasets. We have made the code and model publicly available at \url{https://github.com/dptech-corp/Uni-Mol}. | ['Guolin Ke', 'Linfeng Zhang', 'Di He', 'Zhifeng Gao', 'Shuqi Lu'] | 2023-03-16 | null | null | null | null | ['graph-regression'] | ['graphs'] | [ 2.46682063e-01 -3.07248086e-01 -4.82555807e-01 -4.19190735e-01
-9.42940295e-01 -6.85022056e-01 2.90393412e-01 4.47778553e-01
-1.02413937e-01 1.19387937e+00 2.74978988e-02 -7.09627867e-01
1.85892373e-01 -9.42118645e-01 -1.12996268e+00 -9.53684032e-01
-9.62359011e-02 5.93199492e-01 -4.18051817e-02 -1.00635238e-01
5.18925250e-01 6.62212133e-01 -9.30457413e-01 3.68618309e-01
1.13302326e+00 7.89810836e-01 9.02671963e-02 4.79561061e-01
4.91862819e-02 8.17126751e-01 -1.75778434e-01 -2.68760353e-01
1.79198846e-01 -6.69699669e-01 -9.57873762e-01 -6.35248899e-01
2.94837624e-01 -1.97937116e-01 -3.16994846e-01 9.23677862e-01
9.56467152e-01 2.01016083e-01 7.40235686e-01 -6.54212236e-01
-6.08034551e-01 4.21145618e-01 -3.21053773e-01 -5.85867800e-02
6.14980161e-01 5.57695985e-01 1.07172287e+00 -7.06161082e-01
8.47918153e-01 8.29566717e-01 7.52956390e-01 5.97936869e-01
-1.52977812e+00 -8.30830097e-01 -1.89023972e-01 4.38300401e-01
-1.29012334e+00 -2.98814744e-01 8.70382428e-01 -3.23790282e-01
1.40364468e+00 -7.11965486e-02 7.22798765e-01 1.19987285e+00
5.82341611e-01 4.60641980e-01 9.22193050e-01 -1.38868868e-01
4.43157345e-01 -5.20284653e-01 9.00367498e-02 8.98179889e-01
1.74594969e-01 2.32143790e-01 -4.79562044e-01 -4.93297189e-01
3.70908797e-01 3.91385518e-02 -3.64054114e-01 -5.57867050e-01
-1.09731090e+00 7.98723698e-01 7.66110599e-01 -8.77998322e-02
-5.58642030e-01 3.47155452e-01 4.77036923e-01 3.06511909e-01
3.47041547e-01 7.29117930e-01 -6.37851357e-01 -2.50104874e-01
-5.74856460e-01 3.06081355e-01 8.74253690e-01 6.91252947e-01
8.96164477e-01 -2.87589520e-01 2.62395859e-01 3.95480812e-01
1.43804818e-01 1.49745867e-01 1.74927488e-01 -6.00222528e-01
3.13005894e-01 6.13896251e-01 1.30721718e-01 -6.63865685e-01
-6.43734872e-01 -4.17006761e-01 -8.49499464e-01 -5.88432178e-02
2.09076539e-01 -1.07337721e-01 -9.98826325e-01 1.48886240e+00
5.36292553e-01 -1.09916448e-03 5.59684858e-02 6.63245440e-01
8.72631192e-01 1.02933288e+00 2.14690521e-01 -3.29664379e-01
6.96159720e-01 -8.71474683e-01 -3.54276568e-01 4.83111113e-01
1.11858296e+00 -5.12790859e-01 6.57700956e-01 5.73824525e-01
-9.81024384e-01 -5.13286233e-01 -1.21816647e+00 -1.89411938e-01
-3.69189203e-01 -6.90177083e-02 8.21006775e-01 4.01001662e-01
-7.43991077e-01 1.38533783e+00 -8.85118544e-01 1.30995482e-01
5.23990929e-01 7.73694396e-01 -4.65677649e-01 -1.44300148e-01
-1.32097435e+00 8.66944015e-01 8.31195116e-01 -8.45708400e-02
-1.10367537e+00 -9.66207981e-01 -7.73880124e-01 -2.87841968e-02
3.29967916e-01 -7.57768214e-01 1.18776298e+00 -7.91835070e-01
-1.76842475e+00 4.01829123e-01 -2.80712903e-01 -3.33973646e-01
1.98725238e-01 -2.70651672e-02 -1.76334128e-01 3.00811119e-02
-2.34073028e-01 5.65617800e-01 4.55810487e-01 -9.82524395e-01
9.90140470e-05 -3.48885030e-01 -5.93119999e-03 6.37758672e-02
2.15857312e-01 -3.76499087e-01 -2.79909279e-02 -1.05132431e-01
-1.40202343e-01 -7.74528742e-01 -4.71622854e-01 -3.44877392e-01
-5.43735206e-01 -3.25758725e-01 2.98642725e-01 -3.40465218e-01
1.10199153e+00 -1.57067049e+00 4.21320319e-01 4.42827880e-01
4.62448388e-01 4.89121974e-01 -2.11269125e-01 1.07849491e+00
-4.88753110e-01 6.55730814e-02 -2.15297922e-01 3.97114456e-02
-2.45041758e-01 -1.92838967e-01 -5.35777323e-02 5.38641810e-01
1.64720386e-01 1.03194094e+00 -1.10789561e+00 -1.68080419e-01
1.79886565e-01 6.08922720e-01 -7.41596937e-01 2.87835181e-01
-7.16105878e-01 7.16869533e-01 -5.13911843e-01 5.78921854e-01
8.14956605e-01 -6.50707960e-01 7.80654967e-01 -4.18536365e-01
-2.49136791e-01 8.10492039e-01 -5.21000504e-01 1.82579398e+00
-8.01695585e-02 -2.87139714e-02 -5.36175847e-01 -1.03712356e+00
9.69196081e-01 1.72480613e-01 6.38585567e-01 -8.23408544e-01
6.02596775e-02 3.93741369e-01 4.53575701e-01 -3.27170640e-01
5.73599786e-02 -3.43506634e-01 4.03900087e-01 3.04784030e-01
1.04637317e-01 -1.15310878e-01 1.93614468e-01 1.45941615e-01
9.39683795e-01 6.87402725e-01 5.87068379e-01 -1.04403466e-01
7.39454448e-01 3.67683977e-01 4.99727786e-01 3.89445752e-01
3.53673436e-02 2.63707489e-01 8.15648913e-01 -7.89761782e-01
-1.26426077e+00 -7.92610586e-01 -2.25316986e-01 7.99653769e-01
-8.68077110e-03 -8.81225944e-01 -6.55678093e-01 -7.18256474e-01
5.60290255e-02 5.48495352e-01 -5.97415984e-01 -4.79166418e-01
-4.75553721e-01 -9.44630146e-01 2.97833741e-01 2.57314533e-01
2.00481579e-01 -1.09084666e+00 1.44336611e-01 6.02726400e-01
1.97986230e-01 -4.58445489e-01 -4.83355671e-01 4.44557607e-01
-9.08023775e-01 -1.28740203e+00 -2.98524648e-01 -5.07426202e-01
5.31618893e-01 1.19269581e-03 9.65132117e-01 1.61068246e-01
-1.50699243e-01 -3.74340177e-01 -2.83684790e-01 -3.10247481e-01
-6.28042161e-01 4.46619838e-01 4.14470099e-02 -8.28205124e-02
4.64133471e-01 -8.68316531e-01 -1.05016685e+00 -1.22890778e-01
-6.31862104e-01 2.57585913e-01 5.55047512e-01 8.90136242e-01
8.75163138e-01 -2.46396288e-01 6.05872929e-01 -9.89279807e-01
5.75876296e-01 -4.80012208e-01 -6.71357751e-01 6.15545474e-02
-7.49992192e-01 5.14618814e-01 1.11457193e+00 -1.40120417e-01
-6.19498909e-01 4.87413824e-01 -6.46359682e-01 -2.83229500e-01
-5.07951267e-02 8.79946649e-01 -2.03058645e-01 -2.64795065e-01
6.30897939e-01 5.74333668e-01 -4.86847460e-02 -5.49403250e-01
8.13017637e-02 4.03601348e-01 1.23952173e-01 -8.68859470e-01
5.66828609e-01 1.11658331e-02 3.92081469e-01 -4.85523254e-01
-9.29869294e-01 -3.07516217e-01 -8.53076518e-01 1.11022376e-01
7.79652774e-01 -7.46755421e-01 -1.39109921e+00 4.25899714e-01
-1.15273428e+00 -5.45480072e-01 2.24333569e-01 4.25629795e-01
-5.99688768e-01 7.53854573e-01 -5.82258582e-01 -4.45913762e-01
-8.42522621e-01 -1.38228798e+00 1.03543270e+00 -3.25625986e-02
-7.21465424e-02 -1.02474058e+00 6.09168351e-01 2.78262556e-01
2.06486076e-01 6.04130983e-01 1.37927341e+00 -6.23391986e-01
-6.96956635e-01 -1.81299403e-01 -2.30075065e-02 1.54382616e-01
2.12004200e-01 2.10489500e-02 -7.08878934e-01 -4.60428685e-01
-5.64184189e-01 -6.25568926e-01 9.07855451e-01 2.85507143e-01
1.31681657e+00 -7.15074390e-02 -4.61184710e-01 8.95289838e-01
1.44560540e+00 4.48064446e-01 6.09307349e-01 1.29444957e-01
9.36684072e-01 3.04551609e-02 3.45197320e-01 2.77765542e-01
1.49641544e-01 5.84415972e-01 5.72991073e-01 4.51471694e-02
2.30690882e-01 -7.20824003e-01 5.27693570e-01 7.57847071e-01
-5.36369681e-01 -2.20971584e-01 -9.71988916e-01 -1.83791220e-01
-1.70427346e+00 -1.12926781e+00 -2.34580979e-01 2.24263477e+00
1.40303111e+00 7.40056410e-02 2.37884879e-01 -1.16689838e-01
3.66782069e-01 1.57880299e-02 -1.05132282e+00 -4.29420918e-01
2.90884078e-01 5.60468137e-01 3.99553210e-01 6.00982726e-01
-8.70501161e-01 1.06296563e+00 5.85772038e+00 8.85451376e-01
-1.50314498e+00 -1.90609589e-01 6.50638998e-01 8.55723768e-03
-2.61290461e-01 1.61899239e-01 -8.27956498e-01 5.26162207e-01
1.17176044e+00 -3.15404236e-02 4.35179532e-01 7.62709022e-01
4.06075209e-01 2.36697048e-01 -1.32292163e+00 8.26176643e-01
-4.58605915e-01 -2.03683066e+00 2.82455027e-01 2.50661969e-01
5.91319144e-01 2.39534050e-01 -2.07841113e-01 2.55092204e-01
2.16709912e-01 -1.25931895e+00 2.64023691e-01 5.69990337e-01
1.06583154e+00 -9.88780975e-01 5.85679173e-01 3.20563495e-01
-1.04643655e+00 4.59709466e-01 -5.49235225e-01 -5.55354953e-02
-7.41662309e-02 5.46918631e-01 -1.01720095e+00 8.06501687e-01
1.65952653e-01 1.14829874e+00 -2.77033240e-01 8.13915908e-01
-4.52541448e-02 6.56334817e-01 -3.45134176e-04 -2.99782187e-01
3.30758989e-01 -6.11727655e-01 1.61940798e-01 8.87531757e-01
1.80105679e-02 2.14531064e-01 1.76464424e-01 9.48536694e-01
-4.19844776e-01 2.38185823e-01 -5.76033354e-01 -4.31920677e-01
2.55501449e-01 9.74425554e-01 -3.17860097e-01 -2.79375196e-01
-2.52618581e-01 8.17255676e-01 6.12619340e-01 2.16056749e-01
-9.57041979e-01 -4.60199296e-01 7.67281234e-01 9.02444497e-02
2.36583576e-01 -3.90486300e-01 3.43926907e-01 -1.22600543e+00
-2.75809556e-01 -1.01962805e+00 1.07560299e-01 -6.16027117e-01
-1.22035968e+00 4.12385970e-01 -3.99198174e-01 -8.58851016e-01
8.19926932e-02 -1.01512516e+00 -6.79884493e-01 9.35676217e-01
-1.53615916e+00 -7.40206897e-01 8.46556723e-02 3.78920972e-01
7.37619922e-02 7.38009512e-02 9.13492262e-01 1.55068249e-01
-7.49156833e-01 5.69308996e-01 5.93406141e-01 -9.30972472e-02
7.50641942e-01 -1.42586088e+00 5.83230257e-01 1.24532588e-01
-2.92354703e-01 9.41031694e-01 5.36934316e-01 -8.86390746e-01
-1.91434526e+00 -1.09518600e+00 5.62052190e-01 -4.65805024e-01
5.97890317e-01 -5.27696609e-01 -9.34534192e-01 4.46680546e-01
1.46615859e-02 -1.00054525e-01 9.56247985e-01 -4.18405868e-02
-3.81384611e-01 1.01854004e-01 -9.41360891e-01 4.19522345e-01
1.18204069e+00 -4.55274612e-01 -2.04469502e-01 7.27586746e-01
8.29893410e-01 -4.20403779e-01 -1.24747324e+00 3.68013144e-01
4.40398514e-01 -8.64151120e-01 9.52750206e-01 -1.22356474e+00
5.26593566e-01 -2.71216989e-01 9.35369879e-02 -1.30241334e+00
-4.41671252e-01 -8.47497642e-01 -2.94247419e-01 4.10956562e-01
7.38754034e-01 -8.12738061e-01 9.13586974e-01 3.89449269e-01
-3.89514804e-01 -1.18046963e+00 -7.91239142e-01 -5.52666903e-01
3.94852579e-01 -1.02766275e-01 7.21014380e-01 9.31936979e-01
1.14140697e-01 7.00294375e-01 -4.38028514e-01 -5.75020723e-02
4.78689790e-01 4.29120690e-01 8.05344343e-01 -1.09212387e+00
-4.61833507e-01 -1.15114085e-01 -3.20224941e-01 -1.13181472e+00
1.49644107e-01 -1.43743277e+00 -4.44435686e-01 -1.40034258e+00
4.60911900e-01 -1.27889454e-01 -3.27874511e-01 6.24282300e-01
-9.25242752e-02 -3.92350070e-02 -1.44565195e-01 2.09822461e-01
-6.86296701e-01 9.44689393e-01 1.63039541e+00 -1.89006761e-01
-3.31276894e-01 -8.80675688e-02 -4.74197596e-01 1.19556837e-01
1.00943029e+00 -5.57359755e-01 -2.58463264e-01 2.09482253e-01
5.68035841e-01 1.36180982e-01 1.09606571e-01 -8.88230264e-01
-1.10766165e-01 -2.87200272e-01 7.05153763e-01 -7.40564704e-01
3.10819834e-01 -4.88835841e-01 1.94817349e-01 8.75820220e-01
-3.39172751e-01 -6.70192614e-02 1.62723720e-01 6.44051671e-01
1.13632545e-01 -7.89526254e-02 9.29753900e-01 -9.09070000e-02
-1.15636796e-01 8.75015914e-01 -1.66588157e-01 -3.45461667e-01
9.18650568e-01 -1.01556897e-01 -3.73987794e-01 -1.83667019e-01
-7.65304029e-01 1.51340276e-01 6.62113726e-01 -1.86588973e-01
5.51438868e-01 -1.13722289e+00 -3.95319045e-01 1.25264153e-01
3.94267589e-03 -6.27100319e-02 1.13845736e-01 9.95285213e-01
-9.32507157e-01 6.93780899e-01 -2.00287580e-01 -4.11630154e-01
-1.00035274e+00 8.87301862e-01 7.95922697e-01 -3.16510022e-01
-4.11960930e-01 4.27932680e-01 -5.66111784e-03 -6.82427108e-01
-1.97988331e-01 -2.28149146e-01 1.36156574e-01 -2.46853888e-01
3.19432974e-01 1.28887624e-01 2.80343711e-01 -3.02743316e-01
-2.49498218e-01 4.65819538e-01 -5.20317376e-01 6.67186737e-01
1.76539016e+00 5.19744754e-01 -2.84464836e-01 1.47547377e-02
1.57388675e+00 -1.75672621e-01 -1.32443714e+00 -6.42871931e-02
-5.79374060e-02 -1.03499524e-01 -1.02977522e-01 -8.63905191e-01
-7.74981797e-01 8.93388212e-01 3.71430665e-01 -2.29839683e-01
7.52417207e-01 -2.52166212e-01 1.04168046e+00 9.71436441e-01
4.12114471e-01 -7.40396202e-01 4.21580262e-02 6.21111274e-01
7.05245435e-01 -1.24591625e+00 2.06023797e-01 -1.91662669e-01
-3.09673816e-01 1.48091686e+00 4.93372321e-01 -5.56707568e-02
4.62748796e-01 -1.99254707e-01 -1.87818840e-01 -5.46499372e-01
-8.51598680e-01 5.54337502e-02 2.72621870e-01 4.65096951e-01
9.45656717e-01 -9.76653695e-02 -2.14243889e-01 2.49030679e-01
-8.90494362e-02 1.89500600e-01 3.00256312e-01 9.51195240e-01
-4.52206612e-01 -1.65407360e+00 3.64969894e-02 4.14013326e-01
-3.26387197e-01 -3.53785604e-01 -8.15282226e-01 5.94820201e-01
-8.89379829e-02 4.30874169e-01 -4.96959120e-01 -4.95752782e-01
9.73925367e-02 3.46940070e-01 7.29218245e-01 -6.76233649e-01
-5.95405579e-01 -1.85313523e-01 3.83214541e-02 -7.07505643e-01
-2.94149369e-01 -3.76925468e-01 -1.48675251e+00 -7.63532698e-01
-3.02520007e-01 6.12617254e-01 4.76107687e-01 6.93995476e-01
7.98204303e-01 2.43074670e-01 7.38590240e-01 -9.89978313e-01
-6.75501466e-01 -7.69233346e-01 -2.23672867e-01 1.55720323e-01
3.18129241e-01 -5.28558016e-01 -8.69344398e-02 -2.23577097e-01] | [5.092316150665283, 5.712825775146484] |
0357a41b-cb3c-4288-9bf5-60d38537680d | learning-emotional-representations-from | 2306.05709 | null | https://arxiv.org/abs/2306.05709v1 | https://arxiv.org/pdf/2306.05709v1.pdf | Learning Emotional Representations from Imbalanced Speech Data for Speech Emotion Recognition and Emotional Text-to-Speech | Effective speech emotional representations play a key role in Speech Emotion Recognition (SER) and Emotional Text-To-Speech (TTS) tasks. However, emotional speech samples are more difficult and expensive to acquire compared with Neutral style speech, which causes one issue that most related works unfortunately neglect: imbalanced datasets. Models might overfit to the majority Neutral class and fail to produce robust and effective emotional representations. In this paper, we propose an Emotion Extractor to address this issue. We use augmentation approaches to train the model and enable it to extract effective and generalizable emotional representations from imbalanced datasets. Our empirical results show that (1) for the SER task, the proposed Emotion Extractor surpasses the state-of-the-art baseline on three imbalanced datasets; (2) the produced representations from our Emotion Extractor benefit the TTS model, and enable it to synthesize more expressive speech. | ['Damian Borth', 'Jón Guðnason', 'Shijun Wang'] | 2023-06-09 | null | null | null | null | ['speech-emotion-recognition'] | ['speech'] | [ 7.13586956e-02 3.87350649e-01 -7.46558979e-02 -5.53257763e-01
-8.71049404e-01 -1.91527262e-01 3.08048069e-01 -3.06274205e-01
-1.91168785e-01 5.93764067e-01 4.90084946e-01 -1.00609407e-01
6.30851150e-01 -3.31811249e-01 -5.12712717e-01 -4.85474676e-01
1.64931834e-01 8.98392051e-02 -4.54146087e-01 -5.12690246e-01
-3.07339013e-01 2.48202398e-01 -1.70198500e+00 4.58766848e-01
9.73628461e-01 1.32849848e+00 -2.74828434e-01 6.46411836e-01
-4.43511456e-01 1.06462026e+00 -1.15412641e+00 -5.81342459e-01
-2.70313621e-01 -5.36056519e-01 -6.50492430e-01 5.53251542e-02
-1.30969837e-01 -1.76282793e-01 -2.16242090e-01 9.52957213e-01
1.01422787e+00 2.97924697e-01 7.36568749e-01 -1.58460295e+00
-7.86787331e-01 5.38492620e-01 -5.34371138e-01 1.80344544e-02
3.13022047e-01 -1.08221777e-01 9.41868544e-01 -1.05015326e+00
4.87182379e-01 1.41082501e+00 3.99700224e-01 1.13926876e+00
-7.62385547e-01 -1.07587183e+00 3.56019825e-01 1.24400206e-01
-9.31765676e-01 -1.01227546e+00 1.14820123e+00 3.17558348e-02
1.28877103e+00 5.09973645e-01 7.62116790e-01 1.92156053e+00
-1.64229348e-01 1.41313994e+00 1.12809718e+00 -2.84499168e-01
2.76291937e-01 2.86283016e-01 -9.83657241e-02 1.44003242e-01
-3.87302756e-01 1.95345301e-02 -7.95560598e-01 2.89218780e-02
-5.04288748e-02 -3.62659812e-01 -4.77593422e-01 3.05574238e-01
-8.37947667e-01 8.22285891e-01 1.37862578e-01 3.43291283e-01
-6.79902375e-01 -2.04746187e-01 8.58802974e-01 4.78346944e-01
1.03448403e+00 5.18390179e-01 -4.09456521e-01 -7.14493990e-01
-7.89552927e-01 2.52903458e-02 8.83721411e-01 7.20569074e-01
1.65733859e-01 6.69021726e-01 -1.03892155e-01 1.42217720e+00
8.08379352e-02 6.12136781e-01 6.99343383e-01 -5.34651399e-01
4.93907750e-01 3.73934209e-01 -3.17817032e-01 -8.49676669e-01
-3.02660793e-01 -5.01920819e-01 -1.02180135e+00 -1.52232815e-02
-3.72592777e-01 -3.82184416e-01 -7.81570792e-01 1.74216163e+00
1.15736350e-01 -2.83038653e-02 6.23796582e-01 7.96994388e-01
1.24513102e+00 1.15426791e+00 3.10459524e-01 -5.02650261e-01
1.28176808e+00 -1.15427256e+00 -1.44880581e+00 -7.74619997e-01
4.97761965e-01 -6.22838855e-01 1.37327421e+00 5.07869124e-01
-1.13067913e+00 -2.90681899e-01 -1.19214404e+00 3.54568437e-02
-3.54381710e-01 2.22150296e-01 5.77215910e-01 7.54984856e-01
-8.94493222e-01 1.17916673e-01 -5.29695928e-01 1.10904597e-01
5.38482726e-01 -2.63100564e-02 -4.12838548e-01 2.73130447e-01
-1.60308993e+00 9.64468122e-01 1.65744886e-01 1.87778413e-01
-4.38971132e-01 -5.23612797e-01 -1.19134259e+00 1.19516969e-01
1.49724364e-01 -2.60382265e-01 1.35960543e+00 -1.70577598e+00
-1.94686210e+00 7.43123472e-01 -2.95062900e-01 -3.10821831e-01
1.52718782e-01 -3.06626111e-01 -7.68226445e-01 2.51141310e-01
-4.25078958e-01 7.47250497e-01 1.01931107e+00 -1.26620698e+00
-8.16901028e-02 -2.71634728e-01 -4.33066458e-01 3.07427466e-01
-7.64007568e-01 4.26252991e-01 -1.85351238e-01 -9.77289021e-01
-1.90224543e-01 -6.81247175e-01 2.31691003e-02 -3.36160302e-01
-4.70546216e-01 -2.53706723e-01 9.13441420e-01 -9.82562184e-01
1.27251053e+00 -2.42422915e+00 9.31271017e-02 -1.81199208e-01
-4.26139832e-02 4.51150358e-01 -2.33559579e-01 2.40760788e-01
-5.35459340e-01 3.27769667e-01 -1.04093246e-01 -6.67228758e-01
3.51226568e-01 2.10396662e-01 -5.64688325e-01 -1.89813543e-02
6.89590275e-01 9.72088099e-01 -6.63775980e-01 -3.93782556e-01
6.31035119e-02 7.72371054e-01 -4.72876966e-01 6.00760400e-01
-4.84764799e-02 2.51563519e-01 -2.02819243e-01 6.42848432e-01
7.48301029e-01 2.68575609e-01 -2.27552420e-03 -1.99969769e-01
3.43385011e-01 5.89382768e-01 -8.11445057e-01 1.28915346e+00
-7.92308927e-01 7.31413603e-01 5.88387661e-02 -9.78766203e-01
1.27839780e+00 6.69935942e-01 2.23176420e-01 -9.89793301e-01
3.23951691e-01 1.62900165e-01 -1.40246108e-01 -6.96503103e-01
5.51286876e-01 -6.66445136e-01 -3.58311057e-01 2.82636106e-01
1.28784955e-01 -3.50317866e-01 -3.76303613e-01 6.95068715e-03
8.28483701e-01 -2.14483216e-01 2.57489443e-01 2.58854777e-01
4.31132793e-01 -6.45138323e-01 7.08857536e-01 8.15477222e-02
-4.63610888e-01 3.74099731e-01 8.11781764e-01 -1.34745106e-01
-6.77532315e-01 -6.38763845e-01 1.42198652e-01 1.13387537e+00
-2.82224715e-01 -3.66355181e-01 -9.70046639e-01 -8.37633669e-01
-4.89775449e-01 8.86666238e-01 -5.13533294e-01 -7.28052855e-01
-1.10801466e-01 -6.41935349e-01 8.75141740e-01 6.13097012e-01
3.17618310e-01 -1.37640381e+00 -3.28937083e-01 6.33506924e-02
-8.21276665e-01 -1.37922847e+00 -2.82878786e-01 3.06739390e-01
-4.66764241e-01 -3.73157889e-01 -6.66356504e-01 -4.57867324e-01
3.95501018e-01 -1.27145737e-01 1.15204155e+00 -1.44642368e-01
1.85154885e-01 -1.47308493e-02 -6.36688530e-01 -9.30353463e-01
-8.20485294e-01 7.58430287e-02 3.40238214e-02 1.18465453e-01
5.16201973e-01 -5.12291849e-01 -1.83988437e-01 2.83378121e-02
-9.33122456e-01 9.20799524e-02 4.35080677e-01 8.39581609e-01
2.33229533e-01 -1.08509734e-01 1.22190106e+00 -5.68949819e-01
1.05416906e+00 -3.60010087e-01 1.57427579e-01 7.44878734e-03
-1.58628598e-01 -6.98117763e-02 7.13204563e-01 -6.61089301e-01
-1.25352514e+00 -2.36771941e-01 -9.01306987e-01 -5.51244020e-01
-1.87949836e-01 4.12482828e-01 -5.90674162e-01 3.04741085e-01
3.95486981e-01 1.18199941e-02 -7.12013245e-02 -1.14706278e-01
2.34977037e-01 1.46428955e+00 3.08430731e-01 -5.24299264e-01
2.90890366e-01 4.98264022e-02 -7.00041234e-01 -9.03754056e-01
-9.50160265e-01 -3.15958038e-02 -4.40475643e-02 -3.78103465e-01
6.37447238e-01 -1.19121671e+00 -5.59934080e-01 4.42289263e-01
-1.27392375e+00 -1.60302341e-01 -3.62746269e-01 4.40474629e-01
-4.82740223e-01 1.72853798e-01 -7.67026663e-01 -1.40334094e+00
-7.36265898e-01 -1.03259110e+00 1.42060661e+00 2.21414611e-01
-5.75885355e-01 -4.99407709e-01 -1.08709753e-01 5.67103088e-01
6.12415195e-01 3.78306240e-01 8.21930587e-01 -8.58074725e-01
3.77018243e-01 -1.66773349e-01 3.00530996e-02 8.68776500e-01
-1.51692703e-01 2.47674912e-01 -1.54196429e+00 1.12558074e-01
2.74074048e-01 -9.36553121e-01 7.70248473e-01 -1.98829453e-02
1.32768738e+00 -4.02064145e-01 2.14059487e-01 4.87386018e-01
5.65624237e-01 3.40749115e-01 8.96062493e-01 -6.43788725e-02
1.93683833e-01 8.94045293e-01 5.74641407e-01 4.14229006e-01
4.27696347e-01 4.91680712e-01 2.59180039e-01 -5.28414965e-01
6.47855774e-02 -1.70997083e-01 8.56223524e-01 1.55244553e+00
2.65393645e-01 -3.89433473e-01 -6.35196149e-01 5.28124213e-01
-1.57048655e+00 -9.13769484e-01 2.45553479e-01 1.52734387e+00
1.17605209e+00 9.57923662e-03 4.75148931e-02 4.36177015e-01
4.51058239e-01 5.41100264e-01 -6.74442410e-01 -1.00117671e+00
-3.36543888e-01 4.13054079e-01 -4.02339250e-01 3.18182051e-01
-9.43442345e-01 1.10061157e+00 6.07049465e+00 9.54135597e-01
-1.42688751e+00 1.83269486e-01 9.79576826e-01 -3.89426231e-01
-4.87619877e-01 -6.82110906e-01 -2.87306815e-01 4.59590554e-01
1.38875926e+00 -8.53983164e-02 3.05406213e-01 1.06437004e+00
5.18124104e-02 4.07368064e-01 -8.34659159e-01 1.28916061e+00
3.82243663e-01 -6.64102852e-01 -5.58294095e-02 -1.50138915e-01
4.51492846e-01 -1.44307747e-01 1.39026910e-01 9.82881486e-01
-6.39142767e-02 -1.32184589e+00 7.96151578e-01 2.69413024e-01
7.83868492e-01 -1.11850166e+00 8.98194969e-01 2.78889418e-01
-7.21957862e-01 -1.36370342e-02 -2.45085090e-01 8.54543224e-03
2.54929721e-01 7.11856365e-01 -6.92851603e-01 4.22033310e-01
6.10658169e-01 5.54280102e-01 -1.65808260e-01 1.92038178e-01
-3.50440800e-01 8.01999032e-01 -7.43673369e-02 -2.98407465e-01
1.10313624e-01 4.51326780e-02 4.46874440e-01 1.41746211e+00
3.76266301e-01 3.70525494e-02 -1.85999870e-01 7.88675427e-01
-5.39072692e-01 4.38083351e-01 -6.07665956e-01 -6.67736888e-01
3.43191147e-01 1.32503688e+00 -2.44971022e-01 -3.20267886e-01
-2.11701542e-01 1.24846125e+00 4.14478064e-01 4.94064927e-01
-1.03760016e+00 -5.66351891e-01 1.02592170e+00 -6.14572465e-01
5.02177365e-02 2.75045812e-01 -1.85772836e-01 -1.36552227e+00
3.03584248e-01 -1.56796598e+00 1.74340099e-01 -1.03628778e+00
-1.34128189e+00 1.18472052e+00 -5.29344261e-01 -9.66380477e-01
-4.87495393e-01 -4.20547903e-01 -3.45539033e-01 6.31184757e-01
-1.77580106e+00 -9.82519269e-01 -2.09542900e-01 3.48233402e-01
7.94462025e-01 -1.67334348e-01 9.93620157e-01 3.60302120e-01
-7.07479358e-01 8.20102870e-01 -5.01044631e-01 2.23536506e-01
7.56568968e-01 -1.00990570e+00 4.39446658e-01 5.98478198e-01
1.21302091e-01 3.39297384e-01 6.72696114e-01 -2.78722048e-01
-1.29991972e+00 -9.36793327e-01 9.52061832e-01 -2.01482117e-01
3.51367235e-01 -8.10505509e-01 -1.01621413e+00 4.76052284e-01
5.00160515e-01 -1.12668676e-02 1.07401204e+00 1.69815570e-01
-5.64453900e-01 -4.85100821e-02 -1.12623000e+00 6.58224165e-01
8.54256570e-01 -8.56509209e-01 -6.23044014e-01 -5.79342768e-02
1.03469098e+00 -3.48858058e-01 -7.96513617e-01 4.34678257e-01
5.89268386e-01 -9.16785002e-01 8.00255775e-01 -8.75234246e-01
7.47514307e-01 2.12085620e-01 -2.69558161e-01 -1.84798741e+00
2.61629909e-01 -7.33728766e-01 -3.42542529e-01 1.56837952e+00
5.67283750e-01 -6.26060128e-01 2.98815131e-01 4.45082277e-01
-2.21839532e-01 -1.09848392e+00 -1.06907976e+00 -6.15629196e-01
8.11911449e-02 -7.63052344e-01 8.66802931e-01 1.18034685e+00
3.68809432e-01 6.29428208e-01 -6.03711843e-01 -1.03243113e-01
2.03820132e-03 -1.33820429e-01 6.16051197e-01 -9.40732539e-01
-2.99374554e-02 -5.38953960e-01 -1.89857632e-01 -7.58020043e-01
9.74633217e-01 -7.85409212e-01 1.76423326e-01 -1.28517628e+00
-9.17379856e-02 -6.97212294e-02 -2.03139380e-01 5.97586811e-01
-3.66066366e-01 9.55381542e-02 2.14295730e-01 -6.04635954e-01
-4.64344114e-01 1.26173663e+00 9.60126221e-01 -1.52478456e-01
-3.10201081e-03 -3.14690650e-01 -1.12675881e+00 6.19582236e-01
8.69061232e-01 -3.44866425e-01 -4.55169231e-01 -2.63883499e-03
3.94014448e-01 1.42722398e-01 5.04588522e-02 -6.71438336e-01
-2.49016106e-01 4.90086414e-02 3.69132042e-01 -2.28368253e-01
8.20312858e-01 -7.65574455e-01 -2.06924811e-01 -2.56216340e-02
-3.71662647e-01 -1.31648034e-01 4.75625157e-01 1.94429949e-01
-6.12761796e-01 -8.89108256e-02 7.97757328e-01 3.28429699e-01
-2.77203292e-01 1.45573795e-01 -4.74038810e-01 3.93492490e-01
8.00595880e-01 7.17877373e-02 -4.57939237e-01 -1.04062581e+00
-4.34023589e-01 1.33061916e-01 1.65850401e-01 9.25433993e-01
8.38548779e-01 -1.61041129e+00 -7.81289279e-01 3.96158367e-01
2.46895790e-01 -2.78774559e-01 4.24416125e-01 8.20042193e-01
1.72023952e-01 -2.71005947e-02 6.94708247e-03 -3.42165619e-01
-1.27130902e+00 3.83304030e-01 5.29804468e-01 -1.97519943e-01
-3.66166443e-01 8.62424910e-01 8.79602656e-02 -8.15709412e-01
4.11154568e-01 -3.25853348e-01 -1.81769580e-01 4.32587624e-01
7.74858713e-01 -6.88007101e-02 2.52698690e-01 -7.89473891e-01
-4.02641833e-01 6.40358627e-02 6.17737584e-02 -2.74006069e-01
1.55128551e+00 -4.39125635e-02 6.36797100e-02 5.82482696e-01
1.33957314e+00 6.57810345e-02 -9.51164961e-01 6.33757561e-02
-1.34020492e-01 -2.09113926e-01 2.15767473e-01 -9.64968085e-01
-1.17583966e+00 1.31617451e+00 3.13290745e-01 2.65472025e-01
1.42547941e+00 -7.62766078e-02 1.32663584e+00 2.06748798e-01
-9.35718119e-02 -1.52207017e+00 1.87930748e-01 5.82399189e-01
1.24653220e+00 -1.10559881e+00 -6.80484772e-01 -4.01170194e-01
-1.43060470e+00 1.01045024e+00 6.58001542e-01 4.08454865e-01
3.12470078e-01 7.28295982e-01 6.05658710e-01 -1.18454732e-01
-1.27335167e+00 -8.88110697e-02 2.74795324e-01 6.29116297e-01
6.88793421e-01 5.57836927e-02 1.14623658e-01 1.31964016e+00
-4.74973500e-01 -1.60998642e-01 2.84550369e-01 5.23437858e-01
-2.31637284e-01 -9.07392681e-01 -2.17247099e-01 1.17807187e-01
-6.09884560e-01 -1.25916243e-01 -8.70085597e-01 4.91274893e-01
-2.58553177e-01 1.23724186e+00 5.40101603e-02 -5.56971669e-01
6.62082374e-01 7.03390718e-01 9.85136926e-02 -1.65546164e-01
-7.41326332e-01 1.67634517e-01 6.59476340e-01 -6.67246044e-01
-2.44008303e-01 -3.55494440e-01 -1.23173201e+00 1.39075115e-01
-3.48983675e-01 2.45304093e-01 8.39118600e-01 8.84881496e-01
6.39624417e-01 9.61012065e-01 9.23185587e-01 -7.05796182e-01
-5.99103272e-01 -1.20464218e+00 -4.95080739e-01 5.73313057e-01
5.39584696e-01 -4.56603408e-01 -6.05541468e-01 -2.20987365e-01] | [13.658821105957031, 5.848977565765381] |
a92ce856-fa96-4e6b-87a0-e44a40882977 | transfer-learning-on-electromyography-emg | 2210.06295 | null | https://arxiv.org/abs/2210.06295v2 | https://arxiv.org/pdf/2210.06295v2.pdf | Transfer Learning on Electromyography (EMG) Tasks: Approaches and Beyond | Machine learning on electromyography (EMG) has recently achieved remarkable success on a variety of tasks, while such success relies heavily on the assumption that the training and future data must be of the same data distribution. However, this assumption may not hold in many real-world applications. Model calibration is required via data re-collection and label annotation, which is generally very expensive and time-consuming. To address this problem, transfer learning (TL), which aims to improve target learners' performance by transferring the knowledge from related source domains, is emerging as a new paradigm to reduce the amount of calibration effort. In this survey, we assess the eligibility of more than fifty published peer-reviewed representative transfer learning approaches for EMG applications. Unlike previous surveys on purely transfer learning or EMG-based machine learning, this survey aims to provide an insight into the biological foundations of existing transfer learning methods on EMG-related analysis. In specific, we first introduce the physiological structure of the muscles and the EMG generating mechanism, and the recording of EMG to provide biological insights behind existing transfer learning approaches. Further, we categorize existing research endeavors into data based, model based, training scheme based, and adversarial based. This survey systematically summarizes and categorizes existing transfer learning approaches for EMG related machine learning applications. In addition, we discuss possible drawbacks of existing works and point out the future direction of better EMG transfer learning algorithms to enhance practicality for real-world applications. | ['Mohamad Sawan', 'Jie Yang', 'Di wu'] | 2022-10-03 | null | null | null | null | ['electromyography-emg'] | ['medical'] | [ 5.07307887e-01 4.42745164e-02 -7.14100718e-01 -1.40344635e-01
-9.77857888e-01 -3.39358866e-01 1.15425726e-04 -5.17169833e-01
-4.93053436e-01 1.16444910e+00 -1.24139503e-01 -4.32334095e-02
-2.60734260e-01 -3.98090571e-01 -1.01748872e+00 -8.46481740e-01
-2.64666736e-01 3.04656953e-01 5.34514780e-04 -1.43663406e-01
2.84147590e-01 3.06014627e-01 -1.17323601e+00 3.49746108e-01
6.62633181e-01 8.90432060e-01 2.16460317e-01 3.35772425e-01
3.19357932e-01 4.33761001e-01 -9.30509090e-01 -2.24130899e-01
1.41741872e-01 -7.39521086e-01 -1.00042677e+00 -8.64505917e-02
1.02306478e-01 -1.13867141e-01 -4.02198821e-01 9.82678711e-01
9.49483216e-01 4.34570089e-02 9.57134247e-01 -1.38262916e+00
-6.25391960e-01 4.02075917e-01 -4.17827874e-01 2.45490551e-01
9.48636457e-02 1.51226968e-01 4.73456740e-01 -4.63775754e-01
7.91647136e-01 6.29268765e-01 9.08005118e-01 1.16382492e+00
-1.09003007e+00 -1.02554882e+00 -8.57625902e-02 6.08609259e-01
-9.25123692e-01 -4.95405570e-02 1.02151644e+00 -4.50305641e-01
5.48950732e-01 2.50365436e-02 7.65648544e-01 1.66437399e+00
5.20350635e-01 1.02153969e+00 1.40917706e+00 -3.74028832e-01
2.93793350e-01 5.75104589e-03 -3.50558490e-01 -7.51075819e-02
-7.74949715e-02 2.70804614e-01 -8.46855164e-01 1.52775436e-03
9.51655865e-01 -3.81186634e-01 -3.44721317e-01 -4.94266152e-01
-9.50563073e-01 6.50665402e-01 3.27324003e-01 4.23180163e-01
-4.37423825e-01 1.06160074e-01 7.54220665e-01 5.65465689e-01
4.52301085e-01 3.27642858e-01 -6.45774424e-01 -4.18733746e-01
-6.76179409e-01 2.10770234e-01 5.36538124e-01 9.23280299e-01
3.68029833e-01 2.31270328e-01 3.39080602e-01 1.02949917e+00
-1.22189432e-01 1.79540858e-01 9.34474230e-01 -1.09916329e+00
3.19503397e-01 1.23686880e-01 -4.55183834e-01 -6.27648294e-01
-1.88566640e-01 -4.57055271e-01 -6.76260710e-01 5.15593350e-01
3.68281007e-01 -3.65770936e-01 -6.33449733e-01 1.78137958e+00
1.09635681e-01 3.02527368e-01 -1.58451304e-01 8.45753133e-01
4.09676731e-01 1.06796511e-01 1.50935635e-01 -2.43399605e-01
7.67166913e-01 -7.42548168e-01 -8.06634724e-01 -1.74652576e-01
6.00675523e-01 -4.34767514e-01 9.91364360e-01 7.34753907e-01
-8.56848776e-01 -5.73570967e-01 -8.45404863e-01 3.40621829e-01
-1.81177005e-01 -1.60698835e-02 6.77385747e-01 5.46661437e-01
-2.93587923e-01 1.14425695e+00 -1.12147820e+00 -4.72708791e-01
6.30357802e-01 8.43117952e-01 -6.58256292e-01 2.41374373e-01
-1.37520659e+00 1.20272303e+00 2.58035868e-01 7.55514130e-02
-7.48941064e-01 -8.39985669e-01 -3.80360186e-01 -5.91827095e-01
1.24715485e-01 -5.27171135e-01 1.13804054e+00 -1.07810891e+00
-1.83539116e+00 1.11208141e+00 3.66407931e-01 -4.24147159e-01
4.08937007e-01 -4.68901277e-01 -4.01996195e-01 -4.36081737e-03
-2.21788645e-01 3.66638809e-01 7.99914658e-01 -1.13172913e+00
-5.95523298e-01 -4.50105757e-01 -2.24178642e-01 1.36319757e-01
-5.40608883e-01 -5.95704764e-02 -8.23768147e-04 -9.31662202e-01
-2.73181111e-01 -1.07620645e+00 1.99825540e-01 1.97691277e-01
6.48695678e-02 -3.41664881e-01 8.33287001e-01 -6.77253544e-01
1.01168203e+00 -2.03582501e+00 5.97670317e-01 -1.36907309e-01
-1.20348558e-01 5.77904165e-01 -7.42994994e-02 6.56202137e-01
-3.87087792e-01 -1.86582670e-01 -1.84700832e-01 2.58135021e-01
-3.87332976e-01 3.74763578e-01 -2.16171429e-01 6.14190042e-01
8.66012052e-02 9.78957117e-01 -8.46496165e-01 -4.19308901e-01
3.19398612e-01 2.96251923e-01 -3.66103917e-01 4.32151079e-01
2.47033715e-01 1.04214656e+00 -2.49887690e-01 6.03589296e-01
1.82932064e-01 1.74871176e-01 2.37626806e-01 -5.18207014e-01
5.31167328e-01 1.39848530e-01 -8.63006771e-01 1.96847796e+00
-4.67030585e-01 7.64601171e-01 1.51056275e-01 -1.67532182e+00
8.97909582e-01 5.12868583e-01 1.06378555e+00 -2.86664248e-01
3.27426851e-01 4.54887986e-01 3.79208654e-01 -8.00324142e-01
-4.44250375e-01 -5.99802852e-01 1.09700851e-01 3.56587887e-01
5.41649163e-01 -9.56370831e-02 -3.76713872e-01 -4.68225628e-01
9.46166754e-01 8.98128152e-01 2.18093902e-01 -5.25117069e-02
4.10989523e-01 7.51677062e-03 7.22036958e-01 1.07576221e-01
-3.96646887e-01 4.32950705e-01 5.47635332e-02 -3.26163024e-02
-6.01647019e-01 -1.12026274e+00 -2.15465292e-01 9.67392743e-01
3.23408842e-02 -6.38852865e-02 -1.05982053e+00 -6.81458294e-01
1.23540767e-01 2.71515369e-01 -8.70627701e-01 -8.30671489e-01
-7.84741402e-01 -6.26096845e-01 7.60956049e-01 9.53764915e-01
5.40176891e-02 -1.39776325e+00 -2.82603562e-01 2.04353631e-01
-4.00377423e-01 -6.77864730e-01 -2.36384392e-01 4.55718160e-01
-1.35864902e+00 -1.25186110e+00 -1.03047705e+00 -1.15253532e+00
4.39271450e-01 -3.38964343e-01 4.97467369e-01 -3.77928376e-01
-2.73070097e-01 4.33538795e-01 -4.41328019e-01 -8.45158994e-01
-3.40120703e-01 2.00024500e-01 3.88620019e-01 -3.40932220e-01
7.67939508e-01 -9.93374348e-01 -4.43628907e-01 4.78582650e-01
-6.84102833e-01 -3.22036058e-01 8.29398096e-01 1.25618696e+00
6.98399365e-01 -8.50131810e-02 1.19038653e+00 -8.24290931e-01
7.53785253e-01 -5.53048909e-01 2.37095833e-01 1.45172194e-01
-7.44908750e-01 -3.08950484e-01 4.32034105e-01 -1.00860786e+00
-7.83381104e-01 -4.22954969e-02 -1.04725897e-01 -5.96349359e-01
-1.72848344e-01 5.19906819e-01 -8.27236399e-02 -3.85105669e-01
8.15189838e-01 2.28636041e-01 6.48335218e-01 -6.34357750e-01
5.48683181e-02 1.09654939e+00 8.34927142e-01 -8.24456573e-01
7.12917030e-01 2.83929318e-01 -8.00736770e-02 -5.65525353e-01
-6.02195799e-01 -3.49251002e-01 -9.99779344e-01 -6.32248521e-01
5.14836729e-01 -4.91911054e-01 -4.65779632e-01 8.02592278e-01
-7.68645883e-01 -6.09052181e-01 -4.06393766e-01 9.38335836e-01
-1.42456448e+00 2.88957089e-01 -7.85399318e-01 -5.36431432e-01
-4.56487179e-01 -8.85755599e-01 7.25688934e-01 -1.58610061e-01
-5.80119431e-01 -1.15463412e+00 2.89737076e-01 5.63343942e-01
2.99052238e-01 3.73739660e-01 8.88853908e-01 -6.87610209e-01
1.95606396e-01 -3.85432810e-01 5.22133052e-01 9.37763631e-01
6.28868401e-01 -4.61843222e-01 -8.09384167e-01 -4.10613954e-01
1.83801651e-01 -6.94493711e-01 2.63920069e-01 3.90528172e-01
1.30973947e+00 2.33166352e-01 -5.10888278e-01 3.53596330e-01
9.14135158e-01 3.02478164e-01 7.21901357e-01 4.06623572e-01
3.69841844e-01 9.94848073e-01 9.62763309e-01 -1.21740535e-01
-2.92456925e-01 7.64935553e-01 2.87194043e-01 -1.80127904e-01
-4.10572499e-01 -3.23646665e-01 3.12915534e-01 1.15153933e+00
-7.51112163e-01 1.25354648e-01 -4.16493118e-01 3.43470573e-01
-1.72534847e+00 -7.57355809e-01 1.81958050e-01 2.22437072e+00
1.15055037e+00 2.23256624e-03 4.13341314e-01 4.36482638e-01
8.22989106e-01 -5.53841293e-01 -9.22203779e-01 -2.63717137e-02
8.76393542e-02 6.58611834e-01 3.32201868e-01 -1.31688295e-02
-8.96335661e-01 7.60205150e-01 7.44103050e+00 8.94051969e-01
-1.37558305e+00 2.36527085e-01 -1.81814861e-02 -1.23544112e-01
3.78694415e-01 -3.68347943e-01 -3.86749059e-01 4.67346877e-01
9.21616793e-01 -3.21592987e-01 4.58600402e-01 8.02608967e-01
4.36624028e-02 1.47283882e-01 -1.31561661e+00 9.52609539e-01
8.10331106e-02 -9.98911619e-01 -3.48928750e-01 2.72530943e-01
5.12945056e-01 1.02345541e-01 -2.39619464e-02 4.00088757e-01
-2.85932720e-01 -8.91281307e-01 3.12565446e-01 4.38368559e-01
8.97782922e-01 -5.51859200e-01 9.21750724e-01 3.59257907e-01
-7.21897006e-01 -1.47358216e-02 -3.02157581e-01 -7.03248456e-02
2.33443007e-01 -2.14426126e-03 -2.15309650e-01 5.98709226e-01
9.15450811e-01 9.86205518e-01 4.34026867e-02 8.69512379e-01
-2.29512945e-01 9.72007513e-01 -2.31188610e-02 6.98221698e-02
-3.28398824e-01 -1.42514169e-01 4.77838218e-01 9.56608415e-01
1.13615938e-01 -2.20180184e-01 -1.84858978e-01 7.16458082e-01
3.76856178e-02 5.44278398e-02 -5.76918185e-01 -1.62100270e-01
2.36625403e-01 8.99238288e-01 -9.01856422e-02 1.47484452e-01
-4.91431803e-01 1.02669561e+00 3.01095694e-01 3.16520125e-01
-7.67145514e-01 -5.16219079e-01 5.59947073e-01 2.28971496e-01
-2.54270006e-02 -1.10958509e-01 -2.19851181e-01 -9.28448141e-01
2.33221099e-01 -1.08163452e+00 3.46993536e-01 -6.39706552e-01
-1.66357863e+00 1.51691601e-01 1.69145346e-01 -1.68518126e+00
-4.99699444e-01 -1.03990257e+00 -4.64706063e-01 7.88508654e-01
-1.24977303e+00 -1.11665082e+00 -1.59173980e-02 6.41921222e-01
5.27833641e-01 -3.30244988e-01 1.14289749e+00 5.34259677e-01
-3.20677370e-01 6.56184852e-01 1.78729996e-01 2.18409017e-01
1.24037933e+00 -1.02001345e+00 -1.33372992e-01 2.32532043e-02
-6.73969686e-02 6.67971790e-01 5.74429274e-01 -6.63076699e-01
-1.37588012e+00 -8.43426645e-01 2.93396801e-01 -3.89807671e-01
7.91556299e-01 -2.07944542e-01 -1.13044620e+00 8.52542341e-01
-2.91323327e-02 -4.80485596e-02 1.10388410e+00 1.10925160e-01
1.24228708e-01 -3.34868759e-01 -9.54928160e-01 3.70142430e-01
1.11088598e+00 -4.73822206e-01 -9.70619142e-01 4.38766360e-01
-5.64054511e-02 -4.31508780e-01 -1.56090355e+00 6.60361230e-01
9.36623573e-01 -3.27212512e-01 7.98857093e-01 -1.09515083e+00
3.93988758e-01 1.86588485e-02 1.31566331e-01 -1.63180971e+00
-2.81524777e-01 -6.45154715e-01 -2.85624385e-01 1.08345640e+00
4.60660160e-02 -4.06252742e-01 1.07239103e+00 3.93455356e-01
-2.53523082e-01 -9.78488982e-01 -1.08909202e+00 -1.36668181e+00
7.77315617e-01 -4.16330993e-01 -4.66626473e-02 9.11269665e-01
6.01387441e-01 1.40121222e-01 -6.03479683e-01 -4.74174261e-01
5.28322101e-01 -1.88848600e-01 8.47468138e-01 -1.13860309e+00
-3.96015137e-01 -2.09902704e-01 -8.16051781e-01 -8.07839870e-01
4.57306743e-01 -1.10164237e+00 1.15782984e-01 -1.33154190e+00
-3.50046195e-02 -1.36508748e-01 -5.76387703e-01 4.78207827e-01
-2.91399341e-02 3.74305606e-01 -2.26051778e-01 4.67692018e-01
1.95213687e-03 4.58109438e-01 1.40161228e+00 -1.72101427e-02
-1.22664869e-01 5.19758523e-01 -3.77401829e-01 6.90907061e-01
1.00353980e+00 -6.47928655e-01 -4.14265215e-01 -1.78416193e-01
-2.96238422e-01 -1.75641045e-01 2.00978905e-01 -1.07007921e+00
7.89229274e-02 -3.07260156e-01 1.57957628e-01 -1.24085151e-01
3.49252403e-01 -9.37414765e-01 1.61274537e-01 7.23626792e-01
-3.50239873e-01 -5.22252858e-01 1.84058279e-01 6.94479287e-01
-3.51637810e-01 -3.23078752e-01 8.30783665e-01 -3.02629378e-02
-6.76414192e-01 2.75910348e-01 -6.25801623e-01 -9.60851926e-03
1.32640409e+00 -4.78803128e-01 1.95110887e-01 -4.00191009e-01
-1.05500758e+00 -1.81045732e-03 6.82909638e-02 7.14144051e-01
3.39524627e-01 -1.43174338e+00 -4.01634604e-01 1.43056903e-02
2.58976012e-01 -4.48002249e-01 1.89199269e-01 1.04132462e+00
3.18567865e-02 1.93217024e-01 -8.49732518e-01 -5.30858815e-01
-1.26393783e+00 5.11970401e-01 5.27112842e-01 8.12203959e-02
-4.90941316e-01 5.06156147e-01 -6.64936826e-02 -7.04372227e-01
3.60875666e-01 1.87000319e-01 -1.18955947e-01 -4.93511289e-01
1.91869199e-01 6.50553942e-01 2.79016435e-01 -3.44663858e-01
-1.36608273e-01 8.36228073e-01 1.36170089e-01 7.26807341e-02
1.45401323e+00 9.48698223e-02 1.28320858e-01 8.39217484e-01
1.04508734e+00 -4.35804933e-01 -1.21538174e+00 -1.82224110e-01
-8.84823129e-02 -1.08638883e-01 -1.45454139e-01 -8.42333198e-01
-1.13446152e+00 1.27618361e+00 7.43192732e-01 -4.94189382e-01
1.31103265e+00 4.06444781e-02 8.79656971e-01 1.11423530e-01
8.33487511e-01 -1.48975801e+00 1.96098194e-01 -1.25134200e-01
1.19253135e+00 -1.03970277e+00 1.33637292e-02 -6.11598969e-01
-5.45982838e-01 1.28502440e+00 8.91428649e-01 -3.63581032e-01
8.65259767e-01 3.76004100e-01 2.57863879e-01 1.11856379e-01
-2.81872034e-01 2.35702306e-01 4.76918966e-01 1.46171796e+00
5.52267432e-01 -6.66065663e-02 -7.48117685e-01 8.91836464e-01
-1.06683277e-01 6.55327499e-01 -2.18629967e-02 1.22327209e+00
-1.19168103e-01 -1.76967597e+00 -7.42600933e-02 6.05404973e-01
-6.41573429e-01 4.06412333e-01 -4.47677404e-01 1.10097611e+00
9.26159993e-02 6.67838156e-01 -1.84937015e-01 -8.21992278e-01
4.41578776e-01 9.17387307e-02 9.85408068e-01 -7.44123876e-01
-5.38460314e-01 2.14637280e-01 -2.05011517e-01 -3.18479210e-01
-9.40902531e-01 -5.90303123e-01 -1.29501736e+00 -3.53614874e-02
-5.65924048e-01 5.83999939e-02 6.26859248e-01 9.98600304e-01
2.24249482e-01 5.19951701e-01 3.29674602e-01 -8.46733093e-01
-8.42463315e-01 -1.37132919e+00 -8.92406583e-01 5.87051153e-01
-1.75852299e-01 -1.04454482e+00 -3.24828953e-01 4.63502288e-01] | [6.966719150543213, 0.2406071275472641] |
75f8eae1-b4c0-49e1-9d94-f04d99eec266 | generalized-separable-nonnegative-matrix | 1905.12995 | null | https://arxiv.org/abs/1905.12995v2 | https://arxiv.org/pdf/1905.12995v2.pdf | Generalized Separable Nonnegative Matrix Factorization | Nonnegative matrix factorization (NMF) is a linear dimensionality technique for nonnegative data with applications such as image analysis, text mining, audio source separation and hyperspectral unmixing. Given a data matrix $M$ and a factorization rank $r$, NMF looks for a nonnegative matrix $W$ with $r$ columns and a nonnegative matrix $H$ with $r$ rows such that $M \approx WH$. NMF is NP-hard to solve in general. However, it can be computed efficiently under the separability assumption which requires that the basis vectors appear as data points, that is, that there exists an index set $\mathcal{K}$ such that $W = M(:,\mathcal{K})$. In this paper, we generalize the separability assumption: We only require that for each rank-one factor $W(:,k)H(k,:)$ for $k=1,2,\dots,r$, either $W(:,k) = M(:,j)$ for some $j$ or $H(k,:) = M(i,:)$ for some $i$. We refer to the corresponding problem as generalized separable NMF (GS-NMF). We discuss some properties of GS-NMF and propose a convex optimization model which we solve using a fast gradient method. We also propose a heuristic algorithm inspired by the successive projection algorithm. To verify the effectiveness of our methods, we compare them with several state-of-the-art separable NMF algorithms on synthetic, document and image data sets. | ['Nicolas Gillis', 'Junjun Pan'] | 2019-05-30 | null | null | null | null | ['audio-source-separation', 'hyperspectral-unmixing'] | ['audio', 'computer-vision'] | [ 2.92640358e-01 -1.85564712e-01 -1.97409526e-01 -7.06852227e-02
-6.17657244e-01 -5.59413433e-01 -1.08260915e-01 -1.48439288e-01
-3.45224112e-01 5.66928506e-01 7.66085535e-02 -4.86812145e-01
-7.34836638e-01 -7.20590651e-01 -4.34165806e-01 -8.95995975e-01
-4.65088964e-01 2.93375760e-01 -6.78436995e-01 -2.21510485e-01
5.64013645e-02 3.23108137e-01 -1.58210409e+00 1.22060806e-01
7.70474553e-01 1.05989408e+00 2.54014045e-01 6.06880307e-01
-2.38965943e-01 5.67317963e-01 -4.30038869e-01 -2.29454651e-01
8.01800549e-01 -4.82972085e-01 -8.02262008e-01 3.94660830e-01
1.15817562e-01 1.63886756e-01 -2.29979813e-01 1.41300023e+00
9.52376202e-02 3.51725459e-01 7.02324748e-01 -1.47598767e+00
-7.76941657e-01 5.98393500e-01 -1.30998874e+00 -1.43007681e-01
1.43031582e-01 -2.80511230e-01 1.10577166e+00 -1.03688407e+00
6.18686676e-01 9.56277072e-01 3.51270586e-01 4.08496171e-01
-1.31596839e+00 -1.06701672e+00 2.50459135e-01 2.53857626e-03
-1.53952014e+00 -3.92225325e-01 8.00587177e-01 -4.96484280e-01
7.95766771e-01 6.72469795e-01 6.44173384e-01 3.40939984e-02
-3.77744257e-01 5.16534448e-01 1.19379997e+00 -6.35582626e-01
3.02801967e-01 -1.26157477e-01 2.45023802e-01 6.89317584e-01
2.50601083e-01 -2.13952512e-01 -6.67029202e-01 -2.41597801e-01
3.33039612e-01 1.54451773e-01 -4.71253157e-01 -4.28647637e-01
-1.11673141e+00 9.52062964e-01 1.14744872e-01 2.73816139e-01
-4.59245533e-01 1.37663305e-01 -5.44181839e-02 5.21082461e-01
1.57863930e-01 2.53637582e-01 -3.10574502e-01 1.69827268e-01
-1.06710410e+00 1.79993566e-02 4.13262695e-01 8.93146098e-01
1.15381181e+00 1.72681212e-01 1.63763404e-01 1.12344098e+00
2.35725239e-01 8.48307371e-01 4.49893624e-01 -1.16589427e+00
6.76494956e-01 6.93269789e-01 3.05337548e-01 -1.09670436e+00
-4.47542787e-01 -1.28542660e-02 -1.29097247e+00 -1.76335946e-02
1.15487806e-01 -1.48910983e-02 -8.71328115e-01 2.04366302e+00
1.50454015e-01 2.28237525e-01 3.47197026e-01 9.74129081e-01
6.63838685e-01 9.39893484e-01 -4.13236976e-01 -9.77141440e-01
1.10389197e+00 -7.72779346e-01 -5.25664210e-01 -2.03344032e-01
2.89824456e-01 -1.01954401e+00 1.06561255e+00 6.52862012e-01
-1.23833609e+00 -1.75054148e-01 -8.76462698e-01 3.35594475e-01
-2.35151485e-01 4.48629677e-01 8.61151636e-01 8.14990878e-01
-1.03004086e+00 1.56068459e-01 -4.77466017e-01 -8.47627968e-02
-1.27096072e-01 6.16042852e-01 -6.22913301e-01 -4.97223705e-01
-9.04135585e-01 2.06470862e-01 1.46029696e-01 2.79229581e-01
-6.59429550e-01 -3.67024601e-01 -6.94679081e-01 -1.20423935e-01
4.01116163e-01 -4.89711136e-01 5.77333868e-01 -8.78308296e-01
-1.11525416e+00 9.83143985e-01 -6.16439402e-01 -1.61348119e-01
-4.20432948e-02 2.25703388e-01 -5.76920033e-01 3.67009878e-01
2.90499270e-01 3.88570756e-01 1.20543551e+00 -1.38631868e+00
-5.32971919e-01 -6.44237578e-01 1.92991346e-01 1.65836245e-01
-6.60152256e-01 1.18523017e-01 -4.51189816e-01 -6.45129621e-01
9.17247832e-01 -1.14077556e+00 -5.05845249e-01 -4.21884686e-01
-6.82878792e-01 2.50938181e-02 7.24998295e-01 -5.86243272e-01
1.61924934e+00 -2.41419411e+00 4.77489978e-01 7.21998632e-01
5.26957035e-01 2.09377840e-01 -1.99829698e-01 5.63664734e-01
-6.30574167e-01 1.21854886e-01 -5.89062870e-01 -4.44805384e-01
-2.31319796e-02 8.96066874e-02 -2.65715927e-01 6.41005516e-01
-3.40555668e-01 3.98644865e-01 -7.30530500e-01 -4.66441475e-02
4.15518656e-02 1.40312418e-01 -5.90381503e-01 2.44881347e-04
5.18565066e-02 -9.36319232e-02 1.06826916e-01 7.07108557e-01
9.51702356e-01 -2.43491381e-01 4.19001281e-01 -3.29860449e-01
-3.95433605e-01 -3.84739131e-01 -1.94048309e+00 1.56370735e+00
-1.38326928e-01 2.37163961e-01 6.67519748e-01 -1.54754186e+00
7.92331934e-01 3.03106487e-01 1.01039374e+00 -5.69353700e-01
-1.37061536e-01 3.78875524e-01 2.86246669e-02 -2.95180857e-01
6.94487989e-01 -5.68028390e-01 -1.18233869e-02 9.35940444e-01
-1.94257200e-01 9.91899148e-02 6.32919908e-01 6.26569211e-01
1.05662322e+00 -7.07303345e-01 -1.57959148e-01 -3.46057504e-01
6.56584799e-01 -2.07537368e-01 6.69435143e-01 5.85581720e-01
-1.22018695e-01 4.18858975e-01 4.26144212e-01 2.21233768e-03
-7.26976335e-01 -9.50551569e-01 5.12033068e-02 8.89451921e-01
1.92606941e-01 -5.44617474e-01 -5.89115024e-01 -1.97176427e-01
-7.11272061e-02 2.09581733e-01 -5.43236315e-01 -6.62923381e-02
-3.10045779e-01 -1.15596497e+00 -6.89660311e-02 1.46267265e-01
3.60226065e-01 -7.48034239e-01 -1.89487949e-01 7.53298998e-02
-3.64489526e-01 -7.85988569e-01 -5.55120885e-01 5.67900002e-01
-7.56067038e-01 -1.13172007e+00 -5.85619807e-01 -7.30219424e-01
1.17748415e+00 9.85275745e-01 9.06253159e-01 -1.71033703e-02
-3.09750080e-01 4.23021257e-01 -3.61017793e-01 5.37176728e-02
1.47402421e-01 -3.19961905e-01 5.21907508e-01 2.56925434e-01
-2.48012152e-02 -4.97737259e-01 -7.91936576e-01 3.03190708e-01
-1.22183311e+00 -1.42560959e-01 3.43394995e-01 8.32456410e-01
9.48248863e-01 6.56247556e-01 2.78207153e-01 -7.52733052e-01
6.79419041e-01 -3.27511787e-01 -4.62661594e-01 3.98196429e-01
-6.18654191e-01 -4.83158231e-02 7.73709595e-01 -3.28271121e-01
-4.17689472e-01 1.76070288e-01 2.10023373e-01 -7.99939990e-01
4.77616578e-01 8.71036172e-01 -3.11021447e-01 1.26449674e-01
6.44380629e-01 3.69230360e-01 -8.90618637e-02 -3.67757380e-01
5.99842668e-01 6.07760668e-01 5.82934439e-01 -6.72044754e-01
9.63373303e-01 8.30709100e-01 1.11052133e-01 -1.01151538e+00
-5.87635219e-01 -8.11217010e-01 -2.02515736e-01 -3.98218110e-02
5.74485600e-01 -9.73786771e-01 -7.63910472e-01 2.58548468e-01
-5.96684039e-01 1.06115350e-02 -4.22661215e-01 6.07610404e-01
-4.54001188e-01 5.36844492e-01 -3.17309618e-01 -1.14483392e+00
-3.63324255e-01 -9.62628007e-01 5.72417855e-01 -1.82749361e-01
-4.30612639e-02 -5.44745445e-01 2.98353676e-02 6.26414120e-01
8.15024674e-02 -7.94680789e-03 1.17017782e+00 -3.25481556e-02
-2.83665895e-01 -2.11682290e-01 -3.93558949e-01 4.89752650e-01
1.87042072e-01 2.83405324e-03 -5.60487449e-01 -5.47421992e-01
1.66057963e-02 -1.45035282e-01 9.16690469e-01 4.69056398e-01
1.15698588e+00 -5.60777009e-01 -2.55751163e-01 6.79675817e-01
1.43852079e+00 4.45570707e-01 6.05990589e-01 -1.17274866e-01
5.73585153e-01 4.69118476e-01 4.73276377e-01 6.09836817e-01
1.67330325e-01 2.14300141e-01 4.49240476e-01 -2.33436421e-01
4.06224191e-01 2.14643121e-01 3.66987258e-01 1.01228511e+00
-2.70908386e-01 -2.43707851e-01 -6.58232391e-01 5.49915254e-01
-1.76453471e+00 -9.90240693e-01 -2.34358743e-01 2.56254768e+00
5.00442982e-01 -2.81067938e-01 1.66465789e-01 6.00476444e-01
8.64723802e-01 1.09560892e-01 -4.60588574e-01 -3.62866729e-01
-2.96350300e-01 7.28315890e-01 5.26624084e-01 4.99206394e-01
-8.45830739e-01 6.56792343e-01 5.67075205e+00 8.35908771e-01
-1.06801081e+00 -1.21035250e-02 3.91943425e-01 -5.29595792e-01
-7.63212383e-01 4.69167344e-02 -4.38538909e-01 3.96681070e-01
4.97032970e-01 -7.79633746e-02 9.53328490e-01 5.84636629e-01
2.47143567e-01 -1.35627061e-01 -7.32391834e-01 1.57116878e+00
2.15595827e-01 -1.20321453e+00 1.98179577e-02 1.62550211e-01
9.04531300e-01 -3.21840882e-01 3.10098439e-01 3.09346877e-02
4.87717539e-01 -9.89322007e-01 5.43666542e-01 -1.17926747e-02
7.96873271e-01 -1.00553954e+00 5.66872209e-02 4.82318372e-01
-1.46248889e+00 -2.39427984e-01 -4.99977618e-01 2.87782447e-03
1.80436045e-01 1.07065296e+00 -4.80874293e-02 7.38461852e-01
7.12691963e-01 4.76002574e-01 3.74072641e-02 4.21920508e-01
-8.89154002e-02 2.71789700e-01 -4.16270316e-01 2.61966437e-01
2.82812268e-01 -8.91601861e-01 3.23034674e-01 7.34536469e-01
8.29306841e-01 5.83887756e-01 2.85734028e-01 4.08785135e-01
-2.56690562e-01 5.72401345e-01 -4.97676551e-01 -5.89244127e-01
4.31318611e-01 1.26340270e+00 -8.79239202e-01 -2.22415686e-01
-3.80860776e-01 1.13116848e+00 2.10147083e-01 4.89978641e-01
-5.55381477e-01 -2.45829597e-01 9.37890470e-01 7.76961818e-02
1.91243380e-01 -4.32857037e-01 7.48459250e-02 -1.47014952e+00
1.23917416e-01 -9.50358152e-01 5.98656476e-01 -7.10352898e-01
-1.04282415e+00 4.45422113e-01 -3.85083288e-01 -1.34899724e+00
1.22363158e-01 -5.72422922e-01 -1.55841187e-01 9.89277899e-01
-1.25824118e+00 -7.20342875e-01 9.13709104e-02 1.20778430e+00
-4.92879935e-02 -4.02936161e-01 8.63296628e-01 3.36174667e-01
-8.58587861e-01 4.63721305e-01 4.87791777e-01 -9.20420438e-02
4.50704753e-01 -9.19555366e-01 -3.90583396e-01 1.07168972e+00
3.17158490e-01 1.13706219e+00 4.45978105e-01 -3.55880499e-01
-1.88246691e+00 -9.41542447e-01 7.41292357e-01 2.31662199e-01
6.23385906e-01 -7.50778168e-02 -3.70979756e-01 4.78025228e-01
1.13111414e-01 -3.74225751e-02 1.42930043e+00 2.66672522e-01
-4.72184718e-01 -2.73066431e-01 -1.30641782e+00 3.91331166e-01
9.26368713e-01 -6.33903861e-01 9.82883573e-02 5.00182569e-01
4.48769510e-01 -2.11213335e-01 -7.84781277e-01 5.00523210e-01
4.87141490e-01 -1.04475474e+00 1.08603311e+00 -7.12472141e-01
3.38921517e-01 -6.43601298e-01 -7.53656447e-01 -1.17634737e+00
-4.59935218e-01 -6.54895842e-01 -1.83896586e-01 9.00253594e-01
5.55984974e-01 -5.36234081e-01 9.43274915e-01 7.04787791e-01
-2.04465874e-02 -5.60035229e-01 -1.13834131e+00 -8.29753578e-01
-6.76750615e-02 -7.33476698e-01 6.20077491e-01 1.24934101e+00
2.13408425e-01 2.55033761e-01 -7.14793146e-01 3.28375578e-01
5.75769007e-01 7.81688213e-01 5.97232997e-01 -1.10650945e+00
-6.03282154e-01 -1.59475908e-01 -2.46635862e-02 -9.50637937e-01
1.59901634e-01 -9.41364229e-01 -3.60497713e-01 -1.48287833e+00
4.48121578e-01 -6.81263387e-01 -4.13437605e-01 4.06623900e-01
5.05238473e-02 4.77426350e-01 4.79176968e-01 3.51453722e-01
-4.34935212e-01 1.49790511e-01 1.17576516e+00 -4.07368004e-01
-3.68265033e-01 -3.77014279e-01 -1.06274891e+00 5.70340753e-01
4.76623803e-01 -1.73617035e-01 -5.35066366e-01 -4.88035291e-01
7.26443946e-01 5.06821930e-01 -1.72371805e-01 -6.75016701e-01
1.10786334e-02 -3.82820100e-01 3.91898640e-02 -5.88936925e-01
5.23232162e-01 -9.76757169e-01 5.48593104e-01 5.08863986e-01
5.05477600e-02 1.15150958e-01 -2.28198722e-01 3.19454104e-01
-3.22919488e-01 -3.25040638e-01 7.14454710e-01 -1.04458705e-01
-4.68747914e-01 5.96513093e-01 -3.30136478e-01 -1.63100630e-01
9.95337486e-01 -2.53343105e-01 -1.76676214e-01 -4.20702845e-01
-9.18303609e-01 2.80359387e-01 3.39909375e-01 3.97442989e-02
6.72178566e-01 -1.45780349e+00 -4.44389194e-01 4.00664657e-01
1.36181131e-01 1.46379337e-01 4.99675989e-01 9.05097187e-01
-2.55614460e-01 2.37304777e-01 2.51388788e-01 -3.11748236e-01
-1.17121828e+00 7.92072892e-01 -1.08026169e-01 -2.74614900e-01
1.23648010e-02 9.67082739e-01 1.70889944e-02 -2.83288449e-01
-9.81057584e-02 -1.92180037e-01 6.64903671e-02 4.22820419e-01
5.99685550e-01 4.59403753e-01 8.99330378e-02 -8.89655292e-01
-2.71353006e-01 7.07643092e-01 2.37792566e-01 -4.57980096e-01
1.23745751e+00 -8.07025284e-02 -7.77615249e-01 6.98282942e-02
1.40778530e+00 3.25037748e-01 -5.78970075e-01 -1.31011099e-01
-6.22035921e-01 -5.57970881e-01 -1.87048882e-01 -3.18941891e-01
-1.59052002e+00 7.15205729e-01 6.60158813e-01 4.19561595e-01
1.59868205e+00 -1.27485693e-01 6.81331038e-01 3.39701712e-01
5.23705184e-01 -1.24936485e+00 1.72765367e-02 3.78807485e-01
6.27987742e-01 -9.71316993e-01 -2.28261386e-04 -4.97339845e-01
-3.48033249e-01 9.23571110e-01 1.99000329e-01 1.03659451e-01
9.14705276e-01 1.36783704e-01 3.67005616e-02 -2.71525800e-01
-2.44644821e-01 -3.80671471e-01 1.63442753e-02 2.05946162e-01
3.50280523e-01 4.41850871e-01 -5.89384973e-01 6.51625574e-01
-4.48651493e-01 -1.05040550e-01 2.93536484e-01 9.16405857e-01
-5.02786338e-01 -1.44453919e+00 -7.47934699e-01 7.35035419e-01
-1.68078318e-01 -2.52303690e-01 -2.52368122e-01 2.68539697e-01
3.07090104e-01 1.15400243e+00 -2.21788406e-01 -4.80227232e-01
9.11007002e-02 5.35766371e-02 4.58163023e-01 -6.48291349e-01
-9.60431397e-02 3.19351524e-01 -3.88397157e-01 -2.93253541e-01
-7.45128572e-01 -8.67948532e-01 -1.58354640e+00 -4.76752579e-01
-3.02393168e-01 5.70633948e-01 5.39633572e-01 6.84366584e-01
1.82124600e-01 -2.45878031e-03 1.02638650e+00 -3.58021736e-01
9.38800164e-03 -6.01912022e-01 -1.29433501e+00 4.57903355e-01
1.41994193e-01 -5.22295654e-01 -5.04308105e-01 2.43284747e-01] | [7.31822395324707, 4.4997944831848145] |
a8721f59-fb50-4b9c-9faf-d16ece33eec1 | an-affective-robot-companion-for-assisting | 1807.09825 | null | http://arxiv.org/abs/1807.09825v1 | http://arxiv.org/pdf/1807.09825v1.pdf | An Affective Robot Companion for Assisting the Elderly in a Cognitive Game Scenario | Being able to recognize emotions in human users is considered a highly
desirable trait in Human-Robot Interaction (HRI) scenarios. However, most
contemporary approaches rarely attempt to apply recognized emotional features
in an active manner to modulate robot decision-making and dialogue for the
benefit of the user. In this position paper, we propose a method of
incorporating recognized emotions into a Reinforcement Learning (RL) based
dialogue management module that adapts its dialogue responses in order to
attempt to make cognitive training tasks, like the 2048 Puzzle Game, more
enjoyable for the users. | ['Barros Pablo', 'Sutherland Alexander', 'Churamani Nikhil'] | 2018-07-12 | null | null | null | null | ['dialogue-management', '2048'] | ['natural-language-processing', 'playing-games'] | [ 8.83621052e-02 8.85228992e-01 9.93454680e-02 -5.32370031e-01
9.64330956e-02 -1.95474163e-01 5.07230163e-01 1.01636752e-01
-5.33509552e-01 1.01062429e+00 1.11056790e-01 -4.53580543e-02
5.64098619e-02 -6.54304147e-01 -8.69166385e-03 -3.42968524e-01
-3.96590568e-02 5.92696607e-01 -2.44875088e-01 -9.61554348e-01
4.73251373e-01 6.20663941e-01 -1.38860285e+00 2.30941489e-01
7.22634315e-01 4.26912367e-01 3.41556877e-01 8.03721309e-01
8.71770382e-02 1.32445598e+00 -6.78644717e-01 -1.16321817e-01
-2.20664099e-01 -7.80649543e-01 -1.17924309e+00 1.65883288e-01
-8.33286583e-01 -1.16027713e-01 1.02126285e-01 7.30917931e-01
4.67408746e-01 6.51462376e-01 8.14079821e-01 -1.06916928e+00
-7.16743842e-02 6.14484370e-01 3.55075933e-02 -3.79215091e-01
1.02142358e+00 1.27787322e-01 6.03396952e-01 -3.68818760e-01
5.45638561e-01 1.42098963e+00 2.46218920e-01 1.04252493e+00
-8.99096966e-01 -3.04473609e-01 -2.16225296e-01 3.49474639e-01
-8.19073021e-01 -4.16678131e-01 9.94702339e-01 -6.64482489e-02
1.06319821e+00 1.98103026e-01 8.04750741e-01 1.17349708e+00
1.09241962e-01 6.22453392e-01 1.21186900e+00 -7.47159660e-01
4.18462008e-01 8.35320532e-01 -1.46783143e-01 3.63101423e-01
-2.89407432e-01 -3.88950296e-02 -5.69386363e-01 -6.64532185e-02
6.76874220e-01 -6.94210172e-01 -9.12130103e-02 -1.72064409e-01
-9.33693826e-01 1.07838702e+00 1.80453926e-01 7.93891788e-01
-1.06608880e+00 -1.97402582e-01 6.64526343e-01 6.28493965e-01
1.34360269e-01 1.16833723e+00 -5.13445079e-01 -7.84187734e-01
3.09179515e-01 -2.62563262e-04 1.26232851e+00 4.83103544e-01
6.92005455e-01 1.06111884e-01 -8.62736627e-02 1.17173326e+00
3.53370160e-01 9.22500938e-02 7.56828725e-01 -1.11819458e+00
-3.62784863e-01 7.78758645e-01 2.21357509e-01 -1.00899255e+00
-8.59588563e-01 3.26594293e-01 -3.88399601e-01 5.55603862e-01
-5.35193682e-02 -7.08494723e-01 5.95223485e-03 1.45515943e+00
5.98751903e-01 -5.94822049e-01 7.25769579e-01 7.62465477e-01
7.95457184e-01 7.49954998e-01 4.14647520e-01 -3.65119845e-01
1.39424324e+00 -7.31693149e-01 -8.37745726e-01 -2.87058443e-01
8.16225886e-01 -5.22152901e-01 1.06852221e+00 5.87416947e-01
-9.05394435e-01 -5.65025628e-01 -8.79643738e-01 3.10579032e-01
-3.29361737e-01 -2.80617713e-03 7.28825986e-01 7.33940661e-01
-8.98972511e-01 4.18414682e-01 -1.04403242e-01 -8.26578557e-01
-5.37785888e-01 7.42811322e-01 -2.77925909e-01 5.81433415e-01
-1.51293480e+00 1.58717942e+00 5.87759435e-01 2.73907036e-01
-3.23831499e-01 2.95610905e-01 -8.10497582e-01 -1.38055921e-01
3.13517034e-01 -2.45378867e-01 1.44097519e+00 -1.26157057e+00
-2.87143636e+00 9.11448061e-01 2.26884604e-01 -1.68459937e-01
9.89886373e-02 9.16260108e-02 -1.16872959e-01 2.14919358e-01
-4.95032519e-01 9.80021119e-01 6.91830933e-01 -1.40622330e+00
-5.99655509e-01 -2.68875271e-01 4.90751028e-01 9.87197340e-01
-1.65557951e-01 2.03311101e-01 4.57634814e-02 -4.34764847e-02
-4.37671185e-01 -1.09551656e+00 -4.91259634e-01 -6.61629796e-01
-1.35997043e-03 -6.50305748e-01 4.42392647e-01 -3.99379253e-01
5.42756975e-01 -2.00420260e+00 4.44526285e-01 2.40113124e-01
-2.80733347e-01 4.18111891e-01 -6.35269731e-02 5.69002748e-01
8.60283375e-02 -3.17907006e-01 2.58966357e-01 -2.23676730e-02
2.40952805e-01 4.86727655e-01 3.10173690e-01 2.37316750e-02
2.15226814e-01 5.63426316e-01 -7.75042415e-01 -3.50604743e-01
4.26284790e-01 2.91208506e-01 -4.75905895e-01 9.73049641e-01
-2.50658542e-01 8.91737103e-01 -6.12197042e-01 1.40357301e-01
-9.71472487e-02 4.67113316e-01 5.12384117e-01 3.36553872e-01
-5.19023426e-02 1.83657199e-01 -6.96250498e-01 1.11031079e+00
-9.54066813e-01 4.45018142e-01 2.28910819e-01 -8.34139824e-01
1.33433807e+00 7.27370918e-01 4.71497148e-01 -9.65795636e-01
4.48422849e-01 -2.08009854e-01 3.00938994e-01 -8.23625863e-01
7.44734466e-01 -2.64462888e-01 -4.08849657e-01 5.01317143e-01
-1.35220200e-01 -4.48985785e-01 -2.38288298e-01 -1.27723530e-01
7.87908971e-01 9.83866081e-02 7.12371826e-01 1.28395140e-01
8.71896327e-01 8.65287557e-02 1.81232974e-01 4.84859079e-01
-4.01335746e-01 -4.54865664e-01 5.79728782e-01 -2.23635927e-01
-7.19145834e-01 -1.94518939e-01 4.26705152e-01 1.42315507e+00
-9.64580253e-02 1.49953023e-01 -9.06103730e-01 -2.59358138e-01
-6.80034935e-01 1.17584479e+00 -4.92803335e-01 -5.86566925e-01
-3.12090605e-01 -2.72560537e-01 4.13023800e-01 -2.55103916e-01
6.23774171e-01 -2.22260714e+00 -1.20819199e+00 6.03887975e-01
-2.53992170e-01 -7.98619688e-01 2.65729874e-01 5.88558137e-01
-5.04109919e-01 -3.92295837e-01 -4.00798857e-01 -1.01174533e+00
4.04428154e-01 -2.51313150e-02 7.59068906e-01 2.03884810e-01
-4.78773341e-02 7.92795300e-01 -9.84589458e-01 -5.24897158e-01
-9.92900848e-01 1.99033231e-01 -4.85137627e-02 -2.51052678e-01
2.45049134e-01 -3.66226256e-01 -1.58373088e-01 3.53588283e-01
-6.31202579e-01 -3.61901745e-02 5.00803530e-01 7.28919029e-01
-2.48254806e-01 4.45361361e-02 8.83325696e-01 -7.78379321e-01
1.45740235e+00 -4.23697740e-01 1.99281856e-01 2.57863104e-01
-3.75744969e-01 -3.39264087e-02 4.71084177e-01 -5.50164998e-01
-1.46565628e+00 2.92677879e-01 -4.94971693e-01 5.34933746e-01
-4.93911415e-01 5.11897802e-01 -2.17115670e-01 -3.94427717e-01
7.17249453e-01 -4.46038134e-02 3.45060259e-01 2.10144266e-01
2.87945777e-01 1.07822788e+00 2.94620752e-01 -6.76846504e-01
1.70927316e-01 -4.31403577e-01 -3.38995665e-01 -1.17333949e+00
-1.25351951e-01 -4.88588780e-01 -6.06147766e-01 -8.24320257e-01
8.80321920e-01 -5.27925432e-01 -1.44389892e+00 2.58073568e-01
-1.17703831e+00 -8.41681778e-01 -4.15969379e-02 3.33049446e-01
-1.08781469e+00 1.65091529e-01 -5.13435185e-01 -1.44124031e+00
-3.87369007e-01 -9.80152965e-01 5.18407941e-01 7.53644288e-01
-9.60554123e-01 -8.19009125e-01 8.70547295e-02 3.22863638e-01
5.05877912e-01 -5.04053421e-02 8.57386053e-01 -9.10367548e-01
3.95187050e-01 -1.54065058e-01 2.37059146e-01 1.51352689e-01
1.28502518e-01 -2.01988056e-01 -8.47860217e-01 3.06613982e-01
4.16861057e-01 -1.03785837e+00 -3.60784680e-01 -1.59698576e-01
4.60428029e-01 -3.71994346e-01 2.09281236e-01 -4.48472559e-01
7.29635715e-01 1.12371230e+00 1.12867010e+00 4.25158530e-01
-1.36342525e-01 1.20204389e+00 1.31820261e+00 7.33180881e-01
4.82591152e-01 7.08855629e-01 2.25433692e-01 -2.21176162e-01
7.74550080e-01 1.73416153e-01 7.22608984e-01 4.36556786e-01
-1.74429581e-01 -4.89466861e-02 -5.03654361e-01 -2.29070871e-03
-1.87474430e+00 -9.03514445e-01 1.52872354e-01 1.76642501e+00
1.10781264e+00 2.70811282e-02 1.89400688e-01 -7.89163262e-02
5.31952083e-01 -1.52927667e-01 -3.42175901e-01 -1.44503748e+00
3.10923338e-01 1.89035177e-01 1.13277314e-02 6.96296036e-01
-6.84470534e-01 1.31356609e+00 6.03479147e+00 1.05684079e-01
-9.78330791e-01 -1.31590590e-01 7.36142576e-01 6.78425431e-01
1.85996547e-01 -2.36810178e-01 -1.88450083e-01 -1.51545316e-01
1.26803100e+00 -4.82870638e-03 8.04996192e-01 9.96246934e-01
5.57896197e-01 -5.98505616e-01 -7.52226472e-01 5.64135909e-01
-5.44982916e-03 -2.64483988e-01 -5.35211146e-01 -2.10561290e-01
-8.67095776e-03 -7.39181042e-01 1.87311154e-02 8.53897393e-01
2.14466467e-01 -1.02836287e+00 1.66612193e-01 6.72459722e-01
9.94159356e-02 -1.10216963e+00 8.57956767e-01 5.57705581e-01
-4.17371005e-01 -7.14446092e-03 -3.12333792e-01 -4.46350724e-01
3.33647057e-02 -4.41760719e-01 -1.59869373e+00 1.31131411e-01
3.95705402e-01 -6.02710843e-02 -1.40446648e-01 5.18173695e-01
-1.97974995e-01 2.83042222e-01 2.67434288e-02 -9.68546093e-01
1.85550705e-01 -3.64430875e-01 4.70831066e-01 1.03817344e+00
-9.75727439e-02 6.90395892e-01 5.63927628e-02 4.23689812e-01
4.13782686e-01 5.97593725e-01 -8.79008114e-01 -2.86791533e-01
2.06197366e-01 1.34924889e+00 -7.85941958e-01 -1.27146959e-01
4.04308503e-03 1.53373933e+00 3.30245972e-01 -6.71211537e-03
-6.98547304e-01 -4.88073140e-01 3.85966569e-01 -6.48222148e-01
-2.37738132e-01 -2.55396545e-01 -1.31386966e-01 -5.29202580e-01
-5.25381982e-01 -1.37723792e+00 -9.84405130e-02 -1.03924477e+00
-9.30876732e-01 7.88200676e-01 -1.42006785e-01 -7.50441194e-01
-8.47716331e-01 -5.82120895e-01 -5.44470251e-01 6.83702886e-01
-9.26002622e-01 -8.49245906e-01 -2.41238981e-01 3.65931064e-01
6.68323398e-01 -1.77227944e-01 1.32792532e+00 -3.16632062e-01
-4.89239991e-01 2.80241787e-01 -5.00321329e-01 -2.87002355e-01
8.33887160e-01 -1.08524609e+00 -3.42558712e-01 -1.23087935e-01
-3.95693153e-01 3.43935817e-01 1.18866026e+00 -4.44450617e-01
-1.39834321e+00 -3.88912469e-01 8.60485971e-01 1.73285171e-01
3.52510989e-01 -2.07585711e-02 -7.94952869e-01 2.92742938e-01
6.69721603e-01 -9.48503017e-01 7.62043774e-01 2.47735023e-01
5.40386260e-01 2.69598573e-01 -1.39120960e+00 9.21704710e-01
4.26250607e-01 -4.74627167e-01 -6.37023747e-01 1.33427158e-01
5.36474824e-01 -2.21945435e-01 -9.56495285e-01 1.80379033e-01
3.09386194e-01 -7.88954735e-01 6.65100396e-01 -4.62114751e-01
1.20596893e-01 1.73578858e-01 1.04574442e-01 -1.54035783e+00
-8.34526792e-02 -8.11156392e-01 4.79921013e-01 1.04982150e+00
5.63842177e-01 -7.76482046e-01 5.99554062e-01 1.26902580e+00
1.45957783e-01 -4.08175051e-01 -7.07664073e-01 -3.51349800e-03
-2.56987900e-01 -1.16639681e-01 1.13863520e-01 8.39870870e-01
1.18208587e+00 7.36150682e-01 -7.74168551e-01 -8.33227485e-02
-3.46354157e-01 -5.58225453e-01 1.00226986e+00 -1.09860718e+00
-3.59259367e-01 -4.26833630e-01 -2.60962993e-01 -6.38752639e-01
4.83571202e-01 -2.54891306e-01 8.18237722e-01 -1.26537788e+00
-2.73652166e-01 -3.40795934e-01 1.37526244e-01 3.71957630e-01
-1.06044643e-01 -2.48915508e-01 1.39979005e-01 -2.37817138e-01
-6.80129707e-01 8.56394470e-01 1.02752614e+00 2.23881721e-01
-7.30379105e-01 3.21312100e-01 -7.39550173e-01 8.07462335e-01
1.14044809e+00 -2.12083191e-01 -4.70971346e-01 6.58575058e-01
1.65379137e-01 7.31191695e-01 -1.73897907e-01 -9.82223034e-01
3.20313394e-01 -4.70136166e-01 2.20242396e-01 -8.80634785e-02
5.33716321e-01 -9.49096560e-01 2.50313189e-02 6.27368808e-01
-6.70731366e-01 7.52962679e-02 2.39588603e-01 2.29899541e-01
-1.20161787e-01 -7.34948695e-01 7.16117382e-01 -2.36737326e-01
-8.37284207e-01 -5.64894438e-01 -1.60757530e+00 -5.66694140e-01
1.30289650e+00 -1.33488834e-01 2.78877646e-01 -1.20053172e+00
-6.52936935e-01 3.18245888e-01 2.46011466e-01 4.35945302e-01
6.70017242e-01 -6.88971102e-01 -2.82985032e-01 -3.16746294e-01
-1.97745897e-02 -5.13141096e-01 1.74822852e-01 4.68388796e-01
-5.17016768e-01 4.77119714e-01 -7.68657148e-01 7.04492554e-02
-1.52679443e+00 2.72260845e-01 4.21974868e-01 -4.35095459e-01
-3.29549760e-01 5.18764555e-01 -3.15115690e-01 -7.52642453e-01
5.01000166e-01 2.52493620e-01 -1.12603724e+00 1.76736459e-01
2.99038798e-01 1.24945827e-01 -3.06885451e-01 -6.21265948e-01
1.64443582e-01 -8.85688886e-02 2.32458487e-01 -6.20852292e-01
1.21047628e+00 -3.52598339e-01 -1.70602083e-01 7.01717377e-01
6.35382235e-01 -2.50673473e-01 -7.97305644e-01 1.01929136e-01
5.04262745e-01 -2.92517561e-02 -4.40700710e-01 -1.10614622e+00
-3.79503250e-01 4.87266749e-01 4.58242357e-01 4.08360988e-01
1.12917256e+00 -2.99500316e-01 3.45208168e-01 1.10996461e+00
7.05387533e-01 -1.57541144e+00 4.47832197e-01 8.06077600e-01
1.17688775e+00 -1.18324924e+00 -3.18609715e-01 -2.59997696e-01
-1.73599994e+00 1.44904721e+00 1.04030597e+00 1.14650533e-01
3.61906469e-01 -1.53031006e-01 2.89933920e-01 -2.00759381e-01
-8.45513821e-01 -3.35338324e-01 -2.40287542e-01 8.00588310e-01
5.83172381e-01 1.90908581e-01 -6.41933024e-01 4.46954280e-01
-1.94033787e-01 -6.29456863e-02 8.25198591e-01 9.39309418e-01
-7.38202751e-01 -1.34076512e+00 -3.44832897e-01 5.20261228e-02
7.96540007e-02 4.70981658e-01 -1.02817535e+00 7.38763511e-01
-2.14320645e-01 1.30132365e+00 -2.23652914e-01 -5.30183673e-01
5.30190527e-01 3.10964793e-01 2.20118999e-01 -7.69994080e-01
-1.02799618e+00 -2.47663066e-01 7.82908022e-01 -3.55186671e-01
-6.68003500e-01 -6.24890447e-01 -1.65076005e+00 -5.01143411e-02
3.83502021e-02 3.46865743e-01 8.20376456e-01 9.68124807e-01
6.57230020e-02 3.11861962e-01 1.04802465e+00 -1.09022045e+00
-4.13073003e-01 -1.28900337e+00 -5.61747909e-01 1.79180831e-01
-3.85385454e-01 -4.63134944e-01 -8.68359283e-02 -3.74885768e-01] | [13.143417358398438, 7.684025287628174] |
6bf465e8-4571-468d-868f-001990635c5b | comparing-offline-and-online-testing-of-deep | 1912.00805 | null | https://arxiv.org/abs/1912.00805v1 | https://arxiv.org/pdf/1912.00805v1.pdf | Comparing Offline and Online Testing of Deep Neural Networks: An Autonomous Car Case Study | There is a growing body of research on developing testing techniques for Deep Neural Networks (DNN). We distinguish two general modes of testing for DNNs: Offline testing where DNNs are tested as individual units based on test datasets obtained independently from the DNNs under test, and online testing where DNNs are embedded into a specific application and tested in a close-loop mode in interaction with the application environment. In addition, we identify two sources for generating test datasets for DNNs: Datasets obtained from real-life and datasets generated by simulators. While offline testing can be used with datasets obtained from either sources, online testing is largely confined to using simulators since online testing within real-life applications can be time-consuming, expensive and dangerous. In this paper, we study the following two important questions aiming to compare test datasets and testing modes for DNNs: First, can we use simulator-generated data as a reliable substitute to real-world data for the purpose of DNN testing? Second, how do online and offline testing results differ and complement each other? Though these questions are generally relevant to all autonomous systems, we study them in the context of automated driving systems where, as study subjects, we use DNNs automating end-to-end control of cars' steering actuators. Our results show that simulator-generated datasets are able to yield DNN prediction errors that are similar to those obtained by testing DNNs with real-life datasets. Further, offline testing is more optimistic than online testing as many safety violations identified by online testing could not be identified by offline testing, while large prediction errors generated by offline testing always led to severe safety violations detectable by online testing. | ['Donghwan Shin', 'Fitash Ul Haq', 'Lionel Briand', 'Shiva Nejati'] | 2019-11-28 | null | null | null | null | ['dnn-testing'] | ['adversarial'] | [-9.43019614e-02 2.06168175e-01 1.12831198e-01 -5.78045309e-01
6.89062104e-02 -7.63588190e-01 2.73801178e-01 -2.49190733e-01
-5.05296826e-01 7.65408218e-01 -8.79685342e-01 -9.15869594e-01
-1.89288825e-01 -9.29508746e-01 -1.10671782e+00 -3.18863451e-01
-1.47886366e-01 5.23389459e-01 7.41220057e-01 -2.63442844e-01
-3.42785358e-01 6.84761286e-01 -2.21038246e+00 -8.28221906e-03
8.51058006e-01 9.97742951e-01 4.64386903e-02 6.17885947e-01
2.62669802e-01 4.84953016e-01 -9.99668956e-01 -2.02630937e-01
6.06180787e-01 -4.44662482e-01 -2.89525270e-01 -1.13791510e-01
2.63810575e-01 -7.03451395e-01 -2.53316253e-01 1.15719354e+00
3.48844349e-01 -2.28720233e-01 -1.06690489e-01 -2.09511566e+00
1.39448822e-01 5.36122024e-01 5.42950332e-01 -2.49659225e-01
1.63986795e-02 5.47250330e-01 4.82460588e-01 -3.82338613e-02
5.30777752e-01 8.90642405e-01 5.11186838e-01 8.85922194e-01
-1.26378870e+00 -8.48360419e-01 -1.60197720e-01 7.37866610e-02
-9.25184667e-01 -3.90457839e-01 4.67924207e-01 -5.05853236e-01
1.07977474e+00 7.08097918e-03 7.36150980e-01 1.63276255e+00
5.28596759e-01 4.61951882e-01 6.67203367e-01 -4.43556048e-02
8.75135720e-01 3.77379268e-01 2.08551213e-01 1.82504728e-01
8.42958391e-01 9.24646437e-01 6.89330176e-02 -1.60452898e-03
6.32871449e-01 -2.25224286e-01 -1.30818471e-01 -5.85666001e-01
-9.86975670e-01 6.79025114e-01 -2.15020522e-01 9.96450186e-02
-3.10744286e-01 3.32791567e-01 8.17549229e-01 7.75533438e-01
-2.53981739e-01 2.47562468e-01 -9.13562715e-01 -6.66331291e-01
-5.39086580e-01 4.07669932e-01 1.16348946e+00 1.24980164e+00
6.09084964e-01 6.58486784e-01 2.66077638e-01 3.51256460e-01
2.45619997e-01 4.62679565e-01 5.18845558e-01 -7.74305642e-01
3.32867503e-01 5.86128354e-01 -1.20604455e-01 -2.95913368e-01
-4.04984117e-01 -1.67261720e-01 -3.02831113e-01 7.77481139e-01
4.20667768e-01 -8.65757525e-01 -7.46733367e-01 1.88165259e+00
1.95772480e-02 3.31746399e-01 4.55833763e-01 5.79443395e-01
4.62887615e-01 1.84910908e-01 -2.26855442e-01 1.00628912e-01
6.90524220e-01 -4.48413551e-01 -4.96849805e-01 -2.33863100e-01
9.53026891e-01 4.19829488e-02 7.72866309e-01 7.25500584e-01
-8.54361892e-01 -6.86924338e-01 -1.47221696e+00 6.29287481e-01
-5.62913775e-01 -1.16012827e-01 1.33635104e-01 9.41810906e-01
-1.45815480e+00 8.11911047e-01 -9.81552899e-01 -2.35187680e-01
2.07504444e-02 7.65871525e-01 -2.91800112e-01 8.41974318e-02
-1.46179450e+00 9.97985125e-01 4.87828046e-01 1.52434811e-01
-1.53220725e+00 -6.11436546e-01 -8.91291857e-01 -2.61480808e-01
2.62473971e-01 -4.11667116e-02 1.47973883e+00 -8.52585733e-01
-1.50459862e+00 2.63497412e-01 3.97561133e-01 -8.39080989e-01
7.33838081e-01 1.75341234e-01 -9.80206192e-01 -2.16393635e-01
-1.35886297e-01 6.30048811e-01 2.54373759e-01 -1.05455363e+00
-4.96688217e-01 -3.15489620e-02 3.61304849e-01 -6.05655789e-01
1.13794943e-02 -2.08163992e-01 5.65379858e-02 8.92459974e-03
-4.04833198e-01 -1.02347553e+00 -2.05108915e-02 8.47117603e-03
-1.83016568e-01 4.39251140e-02 1.20932686e+00 -2.52555281e-01
6.90843105e-01 -2.17822480e+00 -4.36987340e-01 5.10895252e-01
-5.25481030e-02 5.54106355e-01 -3.08962673e-01 2.19806507e-01
-2.61201084e-01 8.56815949e-02 1.76386687e-03 3.88293654e-01
4.24521476e-01 7.48385906e-01 -3.58025849e-01 1.30251184e-01
3.01172107e-01 7.07821131e-01 -7.67005384e-01 2.31422633e-02
2.33585328e-01 3.03667694e-01 -4.76382941e-01 9.74386092e-03
-4.12426889e-01 5.80427945e-02 -7.82760158e-02 4.43782151e-01
4.09815550e-01 5.19225597e-01 1.25698745e-01 2.71029174e-01
-4.01291847e-02 1.32914051e-01 -1.10502672e+00 6.87133431e-01
-3.24576080e-01 1.04238904e+00 1.27757061e-02 -1.00518799e+00
1.02627969e+00 6.05607271e-01 6.75178394e-02 -9.39222038e-01
4.24030006e-01 3.48152757e-01 6.94519162e-01 -5.16747475e-01
8.82076174e-02 -6.18259758e-02 -1.10350050e-01 3.21191221e-01
3.68244916e-01 -9.07980427e-02 5.22429764e-01 -4.88303691e-01
1.67838132e+00 9.10016894e-03 -2.30396464e-01 -1.98165581e-01
2.60410130e-01 1.01297475e-01 7.60264397e-01 7.28212833e-01
-5.14081359e-01 1.25429139e-01 1.16872609e+00 -1.62659436e-01
-1.10880268e+00 -1.13731110e+00 -2.16479406e-01 3.92543793e-01
-7.89261013e-02 -3.65336277e-02 -9.68100131e-01 -8.78414333e-01
1.63115650e-01 1.21942925e+00 -4.79553491e-01 -4.73327547e-01
-2.02295989e-01 -5.60352057e-02 1.03434944e+00 8.50802541e-01
4.10723418e-01 -1.28251624e+00 -1.04520702e+00 3.81725788e-01
6.90614343e-01 -1.29522645e+00 1.49832740e-01 6.41152382e-01
-1.03967190e+00 -1.29915738e+00 -3.19299363e-02 -4.81796861e-01
5.56771696e-01 -1.75336555e-01 9.77301478e-01 1.43889517e-01
-7.78408200e-02 4.13272738e-01 -1.91831723e-01 -7.02436566e-01
-9.75448489e-01 -4.59950954e-01 4.47418571e-01 -4.00534749e-01
4.55277979e-01 -5.37472010e-01 -5.17083928e-02 8.90689492e-01
-1.22767258e+00 -5.51551759e-01 5.74146271e-01 6.34296417e-01
1.51262611e-01 4.80767280e-01 7.66718149e-01 -7.84493268e-01
7.82842398e-01 -3.87733012e-01 -1.40282130e+00 1.22737944e-01
-6.93478584e-01 4.16549966e-02 1.00074565e+00 -8.25147212e-01
-1.85827464e-01 -4.17877100e-02 -3.04164797e-01 -6.68952346e-01
-4.70198989e-01 3.92323405e-01 -5.46112895e-01 -1.34842947e-01
6.95423365e-01 7.22909272e-02 4.20238078e-01 4.81953137e-02
-3.99032682e-01 3.79540563e-01 2.54405200e-01 -5.90003729e-01
6.88572586e-01 -1.08973265e-01 -1.67591974e-01 -7.02538550e-01
1.86495736e-01 3.53753656e-01 -2.74057865e-01 -5.06530881e-01
5.26685774e-01 -6.34056449e-01 -1.09231901e+00 6.39415622e-01
-8.83832812e-01 -1.04922676e+00 -4.17002469e-01 5.55328906e-01
-4.50127363e-01 -1.53022766e-01 -2.97004163e-01 -7.51878440e-01
2.35148966e-01 -1.71107411e+00 6.61845922e-01 9.87719148e-02
-3.12597185e-01 -9.56297100e-01 -1.11825131e-01 -2.49630675e-01
4.05873775e-01 2.92448372e-01 7.84002662e-01 -1.09188712e+00
-2.42109016e-01 -7.08337247e-01 2.66501993e-01 1.00105548e+00
-2.58740634e-01 3.02625358e-01 -9.53100979e-01 -2.66610861e-01
1.46290377e-01 -1.87743321e-01 -1.16335094e-01 5.18424362e-02
1.00195837e+00 -7.64218718e-02 -1.03074990e-01 2.88195591e-02
1.46147084e+00 8.99389386e-01 7.98765898e-01 3.60520720e-01
1.90266564e-01 5.58639348e-01 5.75699747e-01 7.23860338e-02
-8.38352889e-02 4.10063416e-01 6.70965552e-01 3.61029625e-01
3.89594585e-01 -1.10034689e-01 9.84201372e-01 1.47254705e-01
5.03869176e-01 -4.05824304e-01 -1.18805254e+00 4.33019668e-01
-1.67318678e+00 -7.83733845e-01 -2.60943711e-01 2.44567680e+00
4.99145001e-01 8.13041568e-01 2.01617092e-01 4.85758036e-01
6.61504924e-01 -7.11326480e-01 -8.78611445e-01 -6.03057683e-01
1.43483907e-01 2.45662719e-01 5.77807546e-01 2.59691298e-01
-5.52680731e-01 2.14656219e-01 6.54910803e+00 2.14734912e-01
-1.27235472e+00 -1.60121337e-01 3.30580026e-01 -1.86152652e-01
-2.09712103e-01 1.56240851e-01 -9.24156368e-01 7.34129369e-01
1.67708027e+00 -9.06274840e-02 2.03279570e-01 1.23373079e+00
3.72115314e-01 -2.00552642e-01 -1.71950340e+00 5.63125372e-01
-3.90989125e-01 -8.89836371e-01 -3.47912043e-01 2.63370603e-01
3.78011823e-01 3.86631668e-01 -3.22079390e-01 7.45072067e-01
4.60346878e-01 -7.98253000e-01 9.76081789e-01 2.30423287e-01
5.45357108e-01 -9.33910787e-01 1.24411356e+00 5.95062017e-01
-5.26104808e-01 -1.69700593e-01 -1.56499937e-01 -8.61557946e-02
-8.70831832e-02 5.75840712e-01 -8.78390908e-01 1.34449512e-01
5.40287614e-01 1.53940409e-01 -5.32612205e-01 1.10664546e+00
-2.63872743e-01 7.16400564e-01 -3.72847110e-01 -4.61629391e-01
7.31834844e-02 -7.06986040e-02 2.42346123e-01 7.11738229e-01
3.90245110e-01 -4.68643695e-01 -3.77836615e-01 1.02795160e+00
3.52848113e-01 -7.99135506e-01 -1.00985253e+00 -5.61920106e-02
3.20444554e-01 9.56901968e-01 -4.25938219e-01 -1.80286750e-01
-3.58899504e-01 4.12786126e-01 -2.47115225e-01 4.49170113e-01
-1.32454693e+00 -7.03761399e-01 9.16830778e-01 2.09072232e-01
2.56870180e-01 -3.11711937e-01 -2.03067929e-01 -6.67343259e-01
2.02468351e-01 -9.03697431e-01 -3.31033319e-01 -8.02103281e-01
-9.50923085e-01 7.74524927e-01 3.94764453e-01 -1.53016508e+00
-6.46911919e-01 -1.10004282e+00 -7.58703530e-01 5.34109354e-01
-8.42399597e-01 -8.67178738e-02 -2.14055091e-01 2.83161998e-01
5.51477075e-02 -2.66589761e-01 7.07975030e-01 5.30622542e-01
-7.96745837e-01 7.28424728e-01 -7.04981387e-02 2.12254420e-01
2.89399832e-01 -8.83006334e-01 3.97104889e-01 9.92127657e-01
-3.09878260e-01 4.02556658e-01 8.69912744e-01 -6.66032791e-01
-1.46443880e+00 -1.22094560e+00 4.21198070e-01 -2.20995545e-01
7.56552279e-01 -7.89857507e-01 -7.92473197e-01 6.54066324e-01
-4.90735881e-02 1.88701123e-01 3.91706496e-01 -2.83664852e-01
3.32625620e-02 -3.71173263e-01 -1.44781113e+00 7.73976147e-01
9.34767783e-01 -3.57884943e-01 -1.26333177e-01 -2.90754139e-02
5.98530293e-01 -4.29461449e-01 -7.74368227e-01 5.30337036e-01
3.89442563e-01 -1.11831987e+00 2.66338557e-01 -3.64159644e-01
1.91693678e-01 -4.28233027e-01 -7.89287612e-02 -1.25742137e+00
5.37326276e-01 -3.41339827e-01 7.12072998e-02 1.08965969e+00
6.78489268e-01 -1.25141859e+00 7.66396403e-01 1.02197170e+00
-4.17596221e-01 -6.46811187e-01 -1.03445005e+00 -1.32788193e+00
-6.76028058e-02 -1.14019322e+00 6.61885023e-01 5.22274196e-01
-1.31051848e-02 -1.08047418e-01 4.77344394e-01 1.89862534e-01
3.03862691e-01 -7.49768972e-01 8.18062663e-01 -1.23765540e+00
-2.52420217e-01 -4.58903491e-01 -1.07740831e+00 -3.95922929e-01
4.73738521e-01 -5.08325219e-01 5.39823115e-01 -1.02830708e+00
-4.84160870e-01 -1.99475765e-01 -4.30337706e-04 7.42349505e-01
7.30810165e-01 -1.51959434e-01 -2.69092172e-01 -5.99677444e-01
-1.00666650e-01 2.14729384e-01 5.76944351e-01 -6.51062801e-02
9.46793035e-02 2.66283691e-01 -1.70091406e-01 4.45747405e-01
7.15877175e-01 -3.30122292e-01 -1.00333238e+00 -1.68606848e-01
2.32070416e-01 2.89205283e-01 8.09189320e-01 -1.58241582e+00
3.56829345e-01 -2.07531713e-02 9.29363295e-02 -5.85781336e-01
-6.76096752e-02 -1.27981818e+00 4.15364772e-01 6.27954364e-01
1.70820300e-02 2.78824508e-01 8.20421338e-01 4.31513369e-01
-3.35888386e-01 -4.85913455e-01 4.48441505e-01 4.46283221e-01
-8.28313053e-01 8.94084051e-02 -8.42333972e-01 -1.72138378e-01
1.45578754e+00 -6.68209732e-01 -3.07696164e-01 -2.28254899e-01
-6.45286560e-01 3.82965773e-01 4.04289126e-01 4.78331417e-01
7.66844451e-01 -1.19708800e+00 -1.96209073e-01 8.67602348e-01
-5.50498627e-02 -4.65722708e-03 1.51048088e-02 6.87418759e-01
-5.90513110e-01 4.41700220e-01 -3.28626454e-01 -6.74914122e-01
-1.04466701e+00 4.07305449e-01 6.57788515e-01 3.99288535e-01
-3.43520075e-01 3.40893149e-01 -3.54194306e-02 -7.74261594e-01
4.43659723e-01 -7.70993292e-01 3.20943236e-01 -3.77641618e-01
1.48278221e-01 1.65876970e-01 5.92266500e-01 9.01834369e-02
-5.45186818e-01 -1.51245803e-01 2.57653683e-01 -4.18909907e-01
1.17547786e+00 5.80677807e-01 1.64859474e-01 5.83996475e-01
1.03841496e+00 -6.09695494e-01 -1.34028649e+00 5.47164798e-01
1.05532743e-01 1.28305912e-01 -1.03955716e-01 -8.08459222e-01
-1.21097600e+00 6.14544332e-01 7.96882451e-01 4.44418788e-01
1.15442634e+00 -4.74063516e-01 5.02351463e-01 6.45955563e-01
7.56666064e-01 -1.23605812e+00 -3.10304582e-01 7.19054937e-01
6.77675009e-01 -9.40719545e-01 -6.73743963e-01 1.49404198e-01
-4.23004240e-01 1.23587394e+00 1.18014276e+00 -4.07308698e-01
8.39102745e-01 8.58156502e-01 8.88761505e-03 3.78352366e-02
-1.16043520e+00 2.63324022e-01 -2.56272197e-01 7.36874878e-01
2.75482088e-02 -1.35417715e-01 4.11774889e-02 7.78375328e-01
-2.77768731e-01 4.66798961e-01 9.23602760e-01 1.10306668e+00
-1.75252676e-01 -1.24287927e+00 -1.98854923e-01 6.43100023e-01
1.57125190e-01 3.65976781e-01 -3.58117253e-01 1.40770686e+00
2.79320776e-01 9.61604536e-01 3.06976259e-01 -1.07667971e+00
6.98422611e-01 1.61919668e-01 2.98721373e-01 -4.51208442e-01
-4.97445852e-01 -5.36025763e-01 4.93951768e-01 -6.38891459e-01
1.73804209e-01 -7.18821585e-01 -1.40574527e+00 -3.38178486e-01
-3.64631474e-01 1.14501387e-01 9.74821150e-01 1.11375201e+00
3.32057267e-01 8.36357653e-01 5.07919908e-01 -5.59421778e-01
-7.76695251e-01 -7.16014743e-01 -5.15291333e-01 -8.36544558e-02
2.98535645e-01 -6.63654447e-01 -5.39143026e-01 -3.96539122e-01] | [6.47033166885376, 7.623467922210693] |
ea802bb2-bd09-493e-8bcd-6e6de59f6f58 | an-adversarially-learned-turing-test-for | 2104.08231 | null | https://arxiv.org/abs/2104.08231v1 | https://arxiv.org/pdf/2104.08231v1.pdf | An Adversarially-Learned Turing Test for Dialog Generation Models | The design of better automated dialogue evaluation metrics offers the potential of accelerate evaluation research on conversational AI. However, existing trainable dialogue evaluation models are generally restricted to classifiers trained in a purely supervised manner, which suffer a significant risk from adversarial attacking (e.g., a nonsensical response that enjoys a high classification score). To alleviate this risk, we propose an adversarial training approach to learn a robust model, ATT (Adversarial Turing Test), that discriminates machine-generated responses from human-written replies. In contrast to previous perturbation-based methods, our discriminator is trained by iteratively generating unrestricted and diverse adversarial examples using reinforcement learning. The key benefit of this unrestricted adversarial training approach is allowing the discriminator to improve robustness in an iterative attack-defense game. Our discriminator shows high accuracy on strong attackers including DialoGPT and GPT-3. | ['Bill Dolan', 'Michel Galley', 'Yizhe Zhang', 'Xiang Gao'] | 2021-04-16 | null | null | null | null | ['dialogue-evaluation'] | ['natural-language-processing'] | [ 3.90374064e-01 6.14760458e-01 1.47332534e-01 -3.17143738e-01
-1.12322009e+00 -1.31427944e+00 1.00569117e+00 -2.57632911e-01
-3.31346720e-01 9.51446533e-01 3.33742052e-01 -5.91373026e-01
3.37476820e-01 -8.19666564e-01 -3.68027717e-01 -3.92891586e-01
-1.60745140e-02 8.31216037e-01 -4.10544090e-02 -9.19597268e-01
7.72414207e-02 1.55122653e-01 -7.26848602e-01 3.63295943e-01
9.55844402e-01 4.66601342e-01 -6.56810224e-01 1.26176059e+00
2.11935326e-01 1.14740336e+00 -1.45368183e+00 -9.30748165e-01
2.67611414e-01 -8.04469883e-01 -1.37283254e+00 -4.74113971e-01
1.67175129e-01 -9.35115159e-01 -4.82017636e-01 9.82007205e-01
7.81801999e-01 1.27125412e-01 5.89516580e-01 -1.50480914e+00
-4.86299604e-01 9.91401196e-01 1.71552375e-01 -1.56230882e-01
9.46282446e-01 7.98323929e-01 9.67543125e-01 -1.44827634e-01
4.93149221e-01 1.42917466e+00 4.75880653e-01 1.35165834e+00
-1.44689214e+00 -5.22527754e-01 -3.83432597e-01 -2.69996196e-01
-6.13509655e-01 -4.98162240e-01 7.95085132e-01 -2.41193354e-01
5.85383058e-01 5.80740035e-01 1.63982511e-01 1.98999524e+00
-1.38593167e-01 8.05904150e-01 1.27100062e+00 -2.63510168e-01
2.84078985e-01 2.27180585e-01 -1.26204640e-01 4.02191728e-01
-3.36283863e-01 4.13374692e-01 -2.37652421e-01 -7.86804557e-01
2.36391112e-01 -6.95626795e-01 -2.35092148e-01 -1.50812529e-02
-9.09225821e-01 1.19196415e+00 5.11007681e-02 -4.29882705e-02
-1.19175121e-01 -1.58210084e-01 8.59694302e-01 7.97405601e-01
1.60645396e-01 1.42176127e+00 -4.48335528e-01 -6.10679626e-01
-2.47791961e-01 6.21538520e-01 1.59919190e+00 5.14280200e-01
2.46857315e-01 3.67827535e-01 -5.00011802e-01 7.73320317e-01
-2.48502031e-01 5.08932889e-01 5.27529240e-01 -1.22514892e+00
5.14076054e-01 3.74489665e-01 2.55033642e-01 -6.08558178e-01
-2.16853306e-01 9.22741555e-03 -3.81302714e-01 4.95788068e-01
8.06411147e-01 -8.93446207e-01 -2.96866804e-01 1.99118972e+00
3.44208151e-01 -2.44111329e-01 6.66042924e-01 7.57226527e-01
7.64236629e-01 4.43640500e-01 2.80306544e-02 1.14601821e-01
9.20999527e-01 -8.56060386e-01 -4.30282176e-01 -7.40867183e-02
7.30550945e-01 -5.31210005e-01 1.49448848e+00 3.69373649e-01
-1.09595263e+00 5.17672673e-02 -1.11437535e+00 2.82890469e-01
-2.47180983e-01 -6.53198421e-01 6.51581466e-01 1.20544410e+00
-7.75729954e-01 5.13418138e-01 -3.90490860e-01 5.81089817e-02
9.17253047e-02 4.89599347e-01 -3.45045596e-01 2.21763447e-01
-1.69811046e+00 1.20088661e+00 1.91597730e-01 -3.00848007e-01
-1.46072423e+00 -3.85722071e-01 -8.93580079e-01 -5.58781400e-02
4.13231075e-01 -2.79455096e-01 1.88226914e+00 -1.11270583e+00
-2.24775815e+00 6.94116771e-01 5.21594167e-01 -8.17182124e-01
9.93044794e-01 -5.07467575e-02 -2.68732190e-01 1.91996455e-01
-1.83174923e-01 3.41587216e-01 7.55333722e-01 -1.21534574e+00
6.55280575e-02 1.27809137e-01 8.57766747e-01 3.81890684e-01
-4.27899390e-01 2.21948102e-01 3.73913646e-01 -4.22944933e-01
-6.93246067e-01 -9.64748979e-01 -1.94965214e-01 -5.23872793e-01
-7.28386104e-01 -1.73143685e-01 8.74101818e-01 -4.81566548e-01
7.97409892e-01 -1.70229673e+00 1.11818045e-01 5.64738549e-02
3.03123504e-01 7.42218018e-01 -3.48795652e-01 5.54384708e-01
1.06585048e-01 3.77885163e-01 -1.56147256e-01 -7.09253401e-02
3.63475502e-01 1.13899849e-01 -4.83091384e-01 1.29047155e-01
4.35610384e-01 9.23561275e-01 -1.22867417e+00 -2.57731795e-01
1.11391488e-03 6.64902329e-02 -7.67909646e-01 8.09615195e-01
-5.27467906e-01 7.74193704e-01 -5.37215054e-01 4.41521347e-01
2.68350691e-01 2.48300642e-01 3.08268040e-01 4.84914035e-01
3.70922655e-01 6.33201003e-01 -5.38713515e-01 1.17642272e+00
-4.44734484e-01 6.19484603e-01 2.31343478e-01 -8.26434255e-01
9.11617696e-01 4.87047404e-01 8.28487277e-02 -3.98192108e-01
3.69171262e-01 -4.90477197e-02 4.42508280e-01 -2.66845942e-01
5.84548473e-01 -4.32446450e-02 -7.70074904e-01 9.26412940e-01
4.80090119e-02 -5.98820508e-01 -1.45152211e-01 5.17976999e-01
1.53856945e+00 -1.83059305e-01 3.07144336e-02 1.03545420e-01
7.17327952e-01 8.57402906e-02 3.60039175e-01 1.11914814e+00
-6.31947756e-01 3.19740802e-01 1.03951192e+00 -3.37012798e-01
-1.08818531e+00 -9.75552440e-01 1.97933465e-01 1.26844668e+00
-2.37084717e-01 -3.60879898e-01 -1.20070362e+00 -1.26178646e+00
-7.18216673e-02 1.01366305e+00 -5.51119626e-01 -7.78522313e-01
-7.62538970e-01 -4.73993719e-01 1.65812457e+00 1.18419565e-01
6.04074359e-01 -1.28078556e+00 -2.90309250e-01 1.18566163e-01
-4.18003380e-01 -9.37508941e-01 -3.61753911e-01 9.44149718e-02
-9.46349502e-02 -1.17154396e+00 -3.01357031e-01 -4.03615832e-01
1.97690234e-01 -2.94201523e-01 1.17703128e+00 1.15125582e-01
1.20894574e-01 5.65221012e-01 -4.71656978e-01 -4.34581846e-01
-1.52462399e+00 1.85814813e-01 2.52235383e-01 -3.50352824e-01
3.32388014e-01 -3.38036835e-01 -3.76883298e-01 5.20631433e-01
-7.68787146e-01 -3.75106245e-01 3.10816653e-02 1.29122913e+00
-4.75572050e-01 -4.25098956e-01 8.97114277e-01 -1.17426217e+00
1.52574456e+00 -4.82334912e-01 -3.81125152e-01 2.28490829e-01
-4.05210525e-01 1.43560052e-01 1.35765851e+00 -9.03366268e-01
-9.95913148e-01 -3.81704330e-01 -3.54915172e-01 7.06853718e-02
-1.05996467e-01 1.66430458e-01 -2.90585905e-01 -4.52850103e-01
1.33912611e+00 1.05773263e-01 2.72897631e-01 2.26460341e-02
4.50771332e-01 7.16544628e-01 5.91207325e-01 -1.13449943e+00
1.05975437e+00 -2.23086104e-01 -3.48800987e-01 -6.38113856e-01
-4.47380811e-01 3.14807087e-01 -1.21749029e-01 -3.38285327e-01
6.15620136e-01 -4.67087626e-01 -1.24168694e+00 5.03026605e-01
-1.08181524e+00 -7.43693173e-01 -2.41769210e-01 2.21854255e-01
-7.17353821e-01 4.70614403e-01 -9.47823882e-01 -9.64942455e-01
-5.83103299e-01 -1.20381761e+00 4.45825666e-01 1.25858173e-01
-8.02192509e-01 -8.92922282e-01 3.97289038e-01 6.21596694e-01
6.00128174e-01 6.10704422e-01 7.59413004e-01 -1.57958221e+00
-7.21633434e-03 -5.80108166e-01 3.63913655e-01 6.73670232e-01
-1.12983175e-01 -2.90656239e-02 -9.52677846e-01 -2.70228565e-01
2.29334071e-01 -1.36807680e+00 8.75890478e-02 -4.31623459e-01
7.22044468e-01 -8.26984167e-01 2.73837417e-01 3.42314750e-01
5.03775656e-01 3.25718641e-01 5.21618485e-01 3.32990319e-01
3.10103029e-01 5.95827878e-01 4.94423479e-01 3.87913138e-01
1.82642147e-01 6.27959847e-01 4.50403303e-01 7.58532435e-02
3.38488460e-01 -4.72130209e-01 7.15117812e-01 2.35641018e-01
1.86950266e-01 -4.16842550e-01 -9.35081124e-01 7.57559612e-02
-1.45854616e+00 -1.12694645e+00 3.58520418e-01 2.10759664e+00
1.36394560e+00 4.84421849e-01 3.76384944e-01 3.55327018e-02
5.62232375e-01 2.55655468e-01 -6.09629750e-01 -1.15479064e+00
-2.45365396e-01 4.03488606e-01 2.40750730e-01 7.42074370e-01
-9.82278228e-01 1.14895797e+00 6.85686731e+00 4.08781976e-01
-8.74673367e-01 1.23804882e-02 5.23722947e-01 4.66020145e-02
-3.70247871e-01 3.34414802e-02 -3.16487104e-01 4.20647413e-01
1.08894682e+00 -3.68317634e-01 8.22476327e-01 9.66762900e-01
-4.29409631e-02 3.50014657e-01 -1.12533116e+00 2.99687237e-01
6.90031946e-02 -9.70549524e-01 -5.40864877e-02 -3.07813101e-02
5.23275852e-01 -3.05083007e-01 2.14082554e-01 8.96555364e-01
1.38516986e+00 -1.21808171e+00 1.90821394e-01 -1.25584558e-01
7.37744689e-01 -7.82983601e-01 6.82386816e-01 4.97913778e-01
-1.70122743e-01 -4.26503122e-02 -1.26248166e-01 -1.88811049e-01
-1.81381613e-01 -3.34187120e-01 -1.44858479e+00 9.20783952e-02
-1.89090148e-02 -2.07197517e-01 -1.98639542e-01 2.45963052e-01
-4.72424060e-01 9.99807358e-01 -1.16776697e-01 -4.06224459e-01
4.33353245e-01 -1.94461737e-02 8.71030569e-01 1.07006598e+00
-5.28299689e-01 2.90799290e-01 2.57873416e-01 7.26498544e-01
-3.20081204e-01 -9.92937163e-02 -9.11861062e-01 -3.24962199e-01
5.35638750e-01 1.15572095e+00 1.58865839e-01 -1.47329450e-01
5.69637306e-02 1.00634634e+00 4.87870097e-01 1.17879599e-01
-7.90672481e-01 -4.89060849e-01 6.25269651e-01 -3.93918544e-01
-3.84860665e-01 -1.03747845e-01 -1.39308915e-01 -1.12258482e+00
-2.79038340e-01 -1.90020323e+00 3.13016564e-01 -2.74361938e-01
-1.35207772e+00 8.34468186e-01 -1.36478543e-01 -9.05731618e-01
-1.05310917e+00 -4.03779477e-01 -9.97789860e-01 9.82653439e-01
-6.55133724e-01 -1.12792850e+00 1.54798832e-02 7.43111312e-01
3.71819109e-01 -6.13204360e-01 1.33833325e+00 -2.24816442e-01
-4.79478091e-01 1.42140746e+00 -1.63416475e-01 6.04532301e-01
9.02714849e-01 -1.46259320e+00 6.39518678e-01 5.81122458e-01
-3.47650230e-01 7.08471358e-01 1.03585529e+00 -4.24872071e-01
-1.28361940e+00 -6.73747420e-01 3.52019936e-01 -8.60139608e-01
8.51892889e-01 -4.71821249e-01 -7.72393763e-01 5.46234727e-01
2.73100555e-01 -5.53996503e-01 8.81707668e-01 2.36569658e-01
-7.47442842e-01 3.16410035e-01 -1.61200643e+00 1.00829983e+00
5.65823913e-01 -9.12430942e-01 -7.46440709e-01 5.53627074e-01
8.38370204e-01 -5.92292428e-01 -8.98409784e-01 2.26888433e-01
6.04729950e-01 -7.41244853e-01 8.69947851e-01 -1.34011996e+00
6.11549437e-01 2.83643693e-01 -4.59641702e-02 -1.52403212e+00
2.96483696e-01 -1.44978762e+00 -1.85450554e-01 1.25597858e+00
5.05020380e-01 -8.08199108e-01 9.16466594e-01 1.08446193e+00
1.11676194e-01 -3.82780164e-01 -7.30314195e-01 -7.52061069e-01
6.81084514e-01 -1.26469672e-01 6.18806541e-01 1.28629208e+00
6.79879665e-01 6.19509876e-01 -7.56737351e-01 -8.23393613e-02
4.53777015e-01 -4.76165950e-01 1.37857258e+00 -8.17831278e-01
-6.50506139e-01 -4.74069357e-01 -2.26845250e-01 -8.38023305e-01
6.16738379e-01 -6.31936431e-01 1.62501335e-01 -6.77691162e-01
-1.99654147e-01 -2.98112124e-01 2.41338924e-01 4.32691753e-01
-4.15232211e-01 -4.52271290e-02 1.19150080e-01 -3.66543755e-02
-4.07471538e-01 5.51840365e-01 1.17204869e+00 -2.85528094e-01
-2.54904121e-01 3.51106882e-01 -8.12169611e-01 5.84218383e-01
1.04072738e+00 -4.44525272e-01 -4.51886922e-01 -1.99783389e-02
-1.59228548e-01 3.58770996e-01 2.52776384e-01 -6.55916154e-01
1.27144277e-01 -1.90781683e-01 -8.55084956e-02 2.81801015e-01
4.26710695e-01 -1.82764724e-01 -5.19101441e-01 5.14543355e-01
-9.39354420e-01 -1.62086338e-02 3.04150581e-02 3.05421084e-01
-5.51740900e-02 -3.85562390e-01 7.49628067e-01 -1.85174748e-01
1.42962737e-02 3.77799608e-02 -6.01974845e-01 6.44487441e-01
8.61832440e-01 1.29215330e-01 -8.12895238e-01 -1.09622085e+00
-4.63238329e-01 3.59341621e-01 4.43536967e-01 3.11447322e-01
4.18539613e-01 -8.27622473e-01 -1.01055682e+00 -5.49998507e-02
2.99920309e-02 -3.55824888e-01 -6.51150718e-02 7.59322569e-02
-5.44999003e-01 4.52209674e-02 -3.03544760e-01 -4.71551679e-02
-1.30450594e+00 4.26626503e-01 6.63982213e-01 -5.63357711e-01
-3.04216295e-01 8.28208685e-01 -5.91958053e-02 -9.72316146e-01
3.02135974e-01 5.28271317e-01 -1.46948323e-01 -5.01925945e-01
5.07137239e-01 2.54611850e-01 -1.68260470e-01 -5.12912154e-01
-1.91389304e-02 -5.09870589e-01 -4.35185522e-01 -6.68696582e-01
7.79592454e-01 5.28278828e-01 1.41229436e-01 -2.49177199e-02
9.83191133e-01 2.61355221e-01 -1.08833158e+00 -1.99787572e-01
-1.80017903e-01 -3.46658260e-01 -6.59635067e-01 -1.16735840e+00
-2.85766661e-01 5.90460896e-01 1.02139473e-01 7.38479018e-01
6.19334757e-01 -4.52335775e-01 7.38166630e-01 9.60534453e-01
3.13629389e-01 -9.78045106e-01 4.77423161e-01 9.49920595e-01
9.42227304e-01 -1.33645272e+00 -3.02813143e-01 -3.26126292e-02
-1.20660710e+00 1.00746119e+00 1.07869053e+00 -1.18294902e-01
-8.05172697e-02 3.60148966e-01 4.16330725e-01 4.53739800e-02
-8.25069606e-01 3.37067425e-01 -5.69913015e-02 9.44034517e-01
2.68040448e-01 2.54381180e-01 -2.60211945e-01 5.99170923e-01
-7.50324845e-01 -6.80895448e-01 8.95051599e-01 8.81419480e-01
-6.47949651e-02 -1.43553996e+00 -2.91529447e-01 2.01717004e-01
-7.04170883e-01 -8.93917903e-02 -1.26008523e+00 7.40431905e-01
-6.72826827e-01 1.44850147e+00 -5.93235373e-01 -9.46715295e-01
3.99973005e-01 2.31522277e-01 1.47475958e-01 -5.07400036e-01
-1.43545520e+00 -6.85705125e-01 8.08841348e-01 -3.72175843e-01
1.39774337e-01 -4.49013740e-01 -7.92424917e-01 -7.30990529e-01
-3.33842158e-01 5.86321354e-01 4.38800722e-01 9.12895560e-01
-1.13404959e-01 2.06712306e-01 1.24168980e+00 -4.59360868e-01
-1.62521613e+00 -1.13041615e+00 -6.13640249e-02 7.13528752e-01
2.80592382e-01 -3.35677058e-01 -6.09240472e-01 -5.10950744e-01] | [12.649599075317383, 8.23375415802002] |
2c30cab9-d740-4be4-a63c-7ade8c067209 | sceneednet-a-deep-learning-approach-for-scene | 1807.03464 | null | http://arxiv.org/abs/1807.03464v1 | http://arxiv.org/pdf/1807.03464v1.pdf | SceneEDNet: A Deep Learning Approach for Scene Flow Estimation | Estimating scene flow in RGB-D videos is attracting much interest of the
computer vision researchers, due to its potential applications in robotics. The
state-of-the-art techniques for scene flow estimation, typically rely on the
knowledge of scene structure of the frame and the correspondence between
frames. However, with the increasing amount of RGB-D data captured from
sophisticated sensors like Microsoft Kinect, and the recent advances in the
area of sophisticated deep learning techniques, introduction of an efficient
deep learning technique for scene flow estimation, is becoming important. This
paper introduces a first effort to apply a deep learning method for direct
estimation of scene flow by presenting a fully convolutional neural network
with an encoder-decoder (ED) architecture. The proposed network SceneEDNet
involves estimation of three dimensional motion vectors of all the scene points
from sequence of stereo images. The training for direct estimation of scene
flow is done using consecutive pairs of stereo images and corresponding scene
flow ground truth. The proposed architecture is applied on a huge dataset and
provides meaningful results. | ['Snehasis Mukherjee', 'Ravi Kumar Thakur'] | 2018-07-10 | null | null | null | null | ['scene-flow-estimation'] | ['computer-vision'] | [ 1.32567465e-01 -4.19639736e-01 1.30198345e-01 -3.06164384e-01
-1.57059059e-01 -3.13348442e-01 6.49021387e-01 -9.42577701e-03
-7.72464812e-01 5.54207325e-01 1.16347499e-01 -5.91727234e-02
1.11287616e-01 -7.33308673e-01 -6.85526907e-01 -5.05396008e-01
-1.23774208e-01 8.00785199e-02 4.87698883e-01 -3.28785181e-02
5.84075391e-01 7.67500103e-01 -1.60734975e+00 2.76021343e-02
2.95244783e-01 1.10253954e+00 4.84499902e-01 1.10321867e+00
-3.12901318e-01 1.41437817e+00 -3.06991011e-01 -1.29355848e-01
4.68429416e-01 -5.01504838e-01 -8.37961972e-01 1.53123453e-01
5.86326599e-01 -8.75374973e-01 -8.38180721e-01 6.41420841e-01
4.57565308e-01 2.83706605e-01 3.85372430e-01 -1.03894329e+00
8.57141912e-02 -1.21485710e-01 -5.03959179e-01 6.08345807e-01
5.89483619e-01 3.50260556e-01 4.24423754e-01 -6.02358997e-01
7.49963641e-01 1.11178768e+00 4.00582939e-01 2.92660385e-01
-6.94549024e-01 -2.88564473e-01 -2.52862394e-01 6.17561162e-01
-1.06219912e+00 -1.82652935e-01 1.03599501e+00 -6.57023728e-01
1.00342178e+00 -1.84164315e-01 9.09525692e-01 7.37914741e-01
1.44206733e-01 7.08288133e-01 9.47737455e-01 -3.17716002e-01
4.39266056e-01 -1.20492190e-01 -1.96919009e-01 7.02573955e-01
1.11027911e-01 1.85525596e-01 -5.54001391e-01 5.10026813e-01
1.16294932e+00 2.38595963e-01 -3.26359332e-01 -6.91230178e-01
-1.26617610e+00 5.97946227e-01 6.88548505e-01 3.71032387e-01
-5.20921528e-01 3.48028749e-01 5.33336282e-01 -1.27049401e-01
3.56755495e-01 8.30918774e-02 -2.95096815e-01 -5.85491538e-01
-1.00697601e+00 1.84273556e-01 8.14025879e-01 7.18338490e-01
8.87467921e-01 1.61110669e-01 3.63598168e-01 1.88124612e-01
4.67982352e-01 4.26314831e-01 6.61320567e-01 -1.27403915e+00
5.27851403e-01 8.21836710e-01 1.30221713e-02 -1.39022434e+00
-3.77143115e-01 -9.79882665e-03 -8.74536097e-01 4.04316783e-01
6.11219347e-01 -4.95239720e-02 -5.35420954e-01 1.13321102e+00
4.29516554e-01 5.76955557e-01 2.39547696e-02 1.28382480e+00
6.99142814e-01 8.38210166e-01 -2.72176415e-01 -9.60862488e-02
8.85270715e-01 -6.73496187e-01 -5.80557644e-01 -3.04258585e-01
5.42015612e-01 -6.63140237e-01 7.35934138e-01 3.08367610e-01
-8.72586727e-01 -7.86938369e-01 -1.01109636e+00 -4.12736773e-01
-3.70271116e-01 -9.57394391e-03 7.06837177e-01 3.48417729e-01
-9.15923715e-01 5.18990874e-01 -1.14551914e+00 -5.19230485e-01
4.41443712e-01 4.02668774e-01 -7.78987110e-01 -2.57644922e-01
-8.61127019e-01 9.09756660e-01 5.25974751e-01 4.29683685e-01
-1.06380510e+00 -5.02862453e-01 -1.17637992e+00 1.10449670e-02
-4.83819917e-02 -5.65718949e-01 1.00852752e+00 -9.13854361e-01
-1.72218835e+00 8.44209731e-01 8.92725959e-03 -4.42944556e-01
6.22859120e-01 -4.31074262e-01 2.38343328e-01 4.44098830e-01
-1.82889700e-01 5.92059791e-01 6.21310949e-01 -8.16325545e-01
-7.73425281e-01 -3.83120418e-01 2.70501018e-01 2.87230253e-01
-2.23057285e-01 -2.77517855e-01 -2.40544185e-01 -1.13367075e-02
-1.22280847e-02 -6.17311060e-01 -3.71176839e-01 2.56474912e-01
-2.84848452e-01 1.33105323e-01 1.00798035e+00 -6.46723211e-01
7.54730225e-01 -1.86506057e+00 2.57276356e-01 -3.57624769e-01
1.24263270e-02 4.47123349e-01 2.56999642e-01 2.39785939e-01
3.14335413e-02 -5.92212617e-01 -2.95879930e-01 -2.88674265e-01
-5.02983272e-01 3.03752154e-01 -2.52172589e-01 8.42342198e-01
4.25056666e-02 4.40657437e-01 -1.10996854e+00 -6.01149559e-01
1.23954451e+00 6.55612588e-01 -6.64921761e-01 6.68203890e-01
3.08222007e-02 8.99274468e-01 -4.35498118e-01 5.00117429e-03
7.04255819e-01 -8.29855874e-02 -2.15570107e-01 -3.63188207e-01
-5.21038055e-01 1.99949071e-01 -1.18654859e+00 2.27918673e+00
-5.31467915e-01 1.24362397e+00 -1.92899406e-01 -1.20671380e+00
8.49913418e-01 1.65463284e-01 8.42833877e-01 -7.00164437e-01
4.36170995e-01 1.75975576e-01 -2.19192773e-01 -8.87203574e-01
4.62237328e-01 4.78978455e-02 2.16827735e-01 1.79277703e-01
2.06054494e-01 -4.54035670e-01 3.61023635e-01 -3.72075848e-02
1.11538482e+00 6.36803746e-01 4.78290707e-01 1.22664735e-01
1.06489527e+00 1.37645587e-01 2.03294456e-01 2.02734992e-01
-3.15560967e-01 6.30703926e-01 1.81459710e-01 -1.00512981e+00
-1.29941034e+00 -6.59734547e-01 4.34599482e-02 2.63258278e-01
5.06626487e-01 -7.55928904e-02 -6.53573275e-01 -2.81932086e-01
-2.50991911e-01 2.06521764e-01 -2.95854151e-01 3.75313163e-02
-1.00962043e+00 -2.34380409e-01 2.79644132e-01 2.47324824e-01
1.18428445e+00 -9.99677420e-01 -1.48931408e+00 2.18266547e-01
-2.16842055e-01 -1.55701280e+00 6.12038514e-03 -3.42879705e-02
-1.17030394e+00 -1.32874167e+00 -8.93391311e-01 -6.47962153e-01
4.08078253e-01 3.52836609e-01 8.30265582e-01 -2.60089070e-01
-5.65444946e-01 4.16711271e-01 -2.08531603e-01 -5.05060107e-02
-3.00519615e-01 1.50395468e-01 -2.03140631e-01 2.39341613e-02
3.83654505e-01 -6.44248009e-01 -9.04490888e-01 6.74202666e-02
-8.80335629e-01 2.96398044e-01 4.61364597e-01 2.80407280e-01
2.81237215e-01 -2.01726273e-01 -1.38457209e-01 -3.58348399e-01
3.33229033e-03 -3.16260755e-01 -8.70276630e-01 -2.80679762e-01
3.27005163e-02 1.71193510e-01 5.95740378e-01 1.27906144e-01
-1.14376640e+00 4.51669514e-01 -3.42340291e-01 -5.76998472e-01
-5.19581020e-01 1.95298105e-01 1.80139124e-01 -8.29790719e-03
5.65702260e-01 2.72879392e-01 -5.71287610e-03 -3.07477027e-01
4.10875916e-01 6.18735909e-01 6.60184205e-01 1.33338282e-02
6.22778594e-01 8.63577366e-01 2.75634766e-01 -1.05479121e+00
-6.13768458e-01 -8.27181637e-01 -1.09251893e+00 -6.55896485e-01
1.26396930e+00 -8.96531940e-01 -9.24261212e-01 7.80880034e-01
-1.65149093e+00 -3.36250186e-01 -1.53964326e-01 9.95676219e-01
-8.23926866e-01 5.22089779e-01 -5.54153383e-01 -5.98461509e-01
-1.15544133e-01 -1.39953804e+00 1.06687558e+00 3.66888970e-01
1.23949237e-01 -1.16502666e+00 4.81963217e-01 2.80140340e-01
1.97695076e-01 4.13559049e-01 2.27884144e-01 -2.78472733e-02
-1.08536839e+00 -2.38077581e-01 -3.10475677e-01 4.68318254e-01
2.92588532e-01 -1.14651680e-01 -1.09703827e+00 1.60808116e-01
2.81151235e-01 -1.38395458e-01 4.95093852e-01 6.75451100e-01
1.04592669e+00 -2.52927113e-02 -8.52625221e-02 8.94173861e-01
1.79299116e+00 1.21339321e-01 6.63978815e-01 5.16445339e-01
9.83613849e-01 5.87785184e-01 3.54097188e-01 6.18090391e-01
3.96958530e-01 6.90488935e-01 8.01230609e-01 1.81288924e-02
-2.54169583e-01 -1.58300191e-01 2.56941974e-01 9.15257752e-01
-3.64444733e-01 -4.44182605e-02 -9.41870451e-01 5.00524640e-01
-1.77482188e+00 -1.02923417e+00 -2.45056614e-01 2.29027104e+00
4.45730925e-01 -5.39961196e-02 -1.67617664e-01 4.40260619e-01
5.62370539e-01 2.73041368e-01 -6.23710454e-01 -3.96169484e-01
1.24418877e-01 -8.42494369e-02 7.12824523e-01 6.57192469e-01
-1.07159770e+00 8.91029954e-01 5.73956537e+00 2.64084548e-01
-1.44271255e+00 -1.38069823e-01 5.28642952e-01 4.17774096e-02
1.88913122e-01 -1.30755037e-01 -5.79188347e-01 3.59981537e-01
7.57731736e-01 6.43069297e-02 3.17520291e-01 7.97172844e-01
5.13029337e-01 -6.11994445e-01 -1.05451167e+00 1.43379498e+00
1.95958957e-01 -1.45497036e+00 -1.18542418e-01 -2.38829665e-02
5.99236727e-01 1.65648118e-01 -3.66925776e-01 -1.86385006e-01
2.95587797e-02 -6.60578430e-01 5.40074766e-01 4.89460289e-01
4.42478180e-01 -4.95180309e-01 8.22814047e-01 4.42199707e-01
-1.09041905e+00 -5.49517535e-02 -4.81615782e-01 -5.75439215e-01
4.82341379e-01 6.23278320e-01 -6.91273689e-01 3.20268005e-01
8.16640973e-01 1.30392730e+00 -3.82574260e-01 1.17255998e+00
-1.26926422e-01 2.61646241e-01 -3.39953184e-01 -6.14017397e-02
3.64874929e-01 -3.97777706e-01 4.39656675e-01 1.14953709e+00
3.07174444e-01 -6.97236657e-02 -1.67893499e-01 7.66726017e-01
2.57958680e-01 -1.58841178e-01 -1.04344535e+00 2.07564116e-01
-7.67506436e-02 1.32945085e+00 -7.82007992e-01 -2.98685640e-01
-5.04920900e-01 9.64396238e-01 1.63400993e-01 2.01443881e-01
-6.32381499e-01 -3.16875309e-01 5.34317374e-01 1.05593599e-01
1.68513954e-01 -8.35074842e-01 -1.54204071e-01 -1.17843258e+00
-1.76518083e-01 -2.69765377e-01 1.23477150e-02 -1.08276725e+00
-6.87473893e-01 4.68738645e-01 4.19393554e-02 -1.57924187e+00
-5.62258780e-01 -8.32401931e-01 -4.87670749e-01 7.86789417e-01
-1.79094028e+00 -6.69017255e-01 -8.31828177e-01 8.27376366e-01
7.50433922e-01 -6.10255264e-02 3.62325251e-01 4.79794443e-01
-2.57584244e-01 -3.02176297e-01 -1.08261630e-01 3.55322570e-01
2.79223263e-01 -9.99811590e-01 2.95412302e-01 1.05340290e+00
8.06159005e-02 1.41344577e-01 6.89851701e-01 -2.12348104e-01
-1.61506128e+00 -8.28423440e-01 7.73436010e-01 -2.77931690e-01
3.65350723e-01 -1.58776745e-01 -4.10105973e-01 5.73013783e-01
3.29367697e-01 3.35263789e-01 1.45503461e-01 -7.21687138e-01
3.05766970e-01 -4.05204535e-01 -9.18788910e-01 2.24695995e-01
7.32104301e-01 -5.92764080e-01 -3.74472976e-01 9.03454423e-02
3.56130481e-01 -6.43414974e-01 -6.03211939e-01 1.47438273e-01
4.99333262e-01 -1.45807827e+00 1.06739950e+00 -4.56125326e-02
8.43013823e-01 -4.22452956e-01 -6.88412860e-02 -1.07764065e+00
2.55147666e-01 -2.88111567e-01 -8.64577517e-02 6.70923889e-01
-3.62376392e-01 -7.89710432e-02 1.17274439e+00 5.13702929e-01
-1.01523191e-01 -3.20569426e-01 -9.67051566e-01 -2.27540225e-01
-4.33522701e-01 -5.40904760e-01 5.25522120e-02 6.01764858e-01
-1.07846089e-01 3.51985008e-01 -3.56612176e-01 -4.52144556e-02
6.73690319e-01 -4.17652912e-02 1.13020444e+00 -1.00868773e+00
-9.34513286e-02 -2.27438733e-01 -1.27572966e+00 -1.42431343e+00
5.13550006e-02 -5.33272088e-01 1.39272466e-01 -1.74992383e+00
-2.70598203e-01 1.94558538e-02 2.26843953e-01 -1.28823474e-01
9.78633165e-02 2.87180185e-01 2.73352385e-01 1.75157025e-01
-5.47531903e-01 6.32392049e-01 1.34794676e+00 4.08992879e-02
-1.04417652e-01 -4.32963781e-02 3.27904433e-01 8.69071722e-01
5.99392414e-01 -1.62615314e-01 -4.64352041e-01 -7.13046670e-01
1.92621589e-01 3.16031277e-01 4.72107559e-01 -1.48538172e+00
5.94626009e-01 8.26983526e-02 5.05764782e-01 -7.11369753e-01
5.00689685e-01 -1.11690950e+00 3.00894342e-02 7.32461572e-01
-3.08827460e-01 2.13221043e-01 1.53239697e-01 4.80921865e-01
-6.07492268e-01 -1.46647930e-01 7.14540839e-01 -4.89478737e-01
-1.30983686e+00 4.79074955e-01 -3.55494976e-01 -1.12665147e-01
1.08113849e+00 -6.09846234e-01 -5.18546393e-03 -3.17044497e-01
-2.37407878e-01 -3.20419014e-01 1.97665006e-01 4.03008342e-01
8.57713342e-01 -1.08888614e+00 -3.90679240e-01 3.69423449e-01
-1.51045218e-01 5.52854478e-01 1.76694304e-01 6.91437006e-01
-1.45667851e+00 6.11159205e-01 -6.45941615e-01 -9.54431653e-01
-8.28642309e-01 4.36192304e-01 5.84409535e-01 -1.81410294e-02
-7.40570247e-01 6.45422101e-01 2.35418051e-01 -3.21975142e-01
1.32687002e-01 -5.18675566e-01 -3.94198835e-01 -3.02540213e-01
4.92744058e-01 7.36773074e-01 1.82306133e-02 -8.94171178e-01
-3.55431199e-01 9.44961846e-01 4.78697389e-01 -1.86225399e-01
1.20793951e+00 -3.26442301e-01 -1.17976658e-01 5.90597034e-01
1.61691761e+00 -5.44879735e-01 -1.65197694e+00 2.59361975e-02
-5.74949309e-02 -7.22121060e-01 3.61762732e-01 -7.63404146e-02
-1.28817749e+00 1.48455381e+00 7.73823082e-01 -5.50414436e-02
1.00272310e+00 -4.16844398e-01 7.92606115e-01 3.54145914e-01
5.57461858e-01 -7.65590310e-01 8.80513936e-02 5.78858316e-01
3.73196781e-01 -1.41145086e+00 -7.91630056e-03 -1.53813481e-01
-2.61391133e-01 1.52479064e+00 5.27824461e-01 -4.70837623e-01
6.37601793e-01 2.01008126e-01 4.55364324e-02 -3.91120203e-02
-2.46566862e-01 -1.91131637e-01 -1.13665245e-01 5.89715719e-01
5.49031198e-01 -5.65091014e-01 -1.40973255e-02 -4.43182707e-01
-2.94242743e-02 5.89979947e-01 6.47611260e-01 8.19506347e-01
-4.66350973e-01 -7.67931998e-01 -1.86228558e-01 -5.20631252e-03
-2.84794539e-01 2.42978454e-01 1.00312665e-01 8.34283412e-01
7.05149621e-02 6.91945612e-01 4.32758331e-01 -9.94033646e-03
3.85451227e-01 -4.29161698e-01 7.04647660e-01 -1.84421957e-01
-2.94180095e-01 -3.98282617e-01 -3.21624011e-01 -9.02591050e-01
-1.14614832e+00 -5.32733262e-01 -1.20743287e+00 -3.63490671e-01
8.10796916e-02 -1.07673615e-01 1.04178393e+00 1.08985949e+00
1.27414182e-01 4.48622942e-01 7.00030684e-01 -1.44842529e+00
1.05865762e-01 -8.45508516e-01 -4.83676672e-01 5.64644933e-01
6.47743523e-01 -5.13690174e-01 -3.20581168e-01 5.84383428e-01] | [8.676101684570312, -2.0002858638763428] |
7b8a2f13-6ed9-4aca-bf26-5b544587001d | real-time-video-super-resolution-by-joint | 2105.02794 | null | https://arxiv.org/abs/2105.02794v1 | https://arxiv.org/pdf/2105.02794v1.pdf | Real-Time Video Super-Resolution by Joint Local Inference and Global Parameter Estimation | The state of the art in video super-resolution (SR) are techniques based on deep learning, but they perform poorly on real-world videos (see Figure 1). The reason is that training image-pairs are commonly created by downscaling a high-resolution image to produce a low-resolution counterpart. Deep models are therefore trained to undo downscaling and do not generalize to super-resolving real-world images. Several recent publications present techniques for improving the generalization of learning-based SR, but are all ill-suited for real-time application. We present a novel approach to synthesizing training data by simulating two digital-camera image-capture processes at different scales. Our method produces image-pairs in which both images have properties of natural images. Training an SR model using this data leads to far better generalization to real-world images and videos. In addition, deep video-SR models are characterized by a high operations-per-pixel count, which prohibits their application in real-time. We present an efficient CNN architecture, which enables real-time application of video SR on low-power edge-devices. We split the SR task into two sub-tasks: a control-flow which estimates global properties of the input video and adapts the weights and biases of a processing-CNN that performs the actual processing. Since the process-CNN is tailored to the statistics of the input, its capacity kept low, while retaining effectivity. Also, since video-statistics evolve slowly, the control-flow operates at a much lower rate than the video frame-rate. This reduces the overall computational load by as much as two orders of magnitude. This framework of decoupling the adaptivity of the algorithm from the pixel processing, can be applied in a large family of real-time video enhancement applications, e.g., video denoising, local tone-mapping, stabilization, etc. | ['Noam Levy', 'Shahar S. Yuval', 'Alex Itskovich', 'Noam Elron'] | 2021-05-06 | null | null | null | null | ['video-denoising', 'video-enhancement', 'tone-mapping'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.99115217e-01 -1.97556406e-01 -5.07484414e-02 -6.24222197e-02
-5.98295927e-01 -8.42043459e-02 3.07900846e-01 -3.64789158e-01
-5.14687598e-01 7.08610117e-01 -8.56736600e-02 -1.76271200e-02
1.64126620e-01 -9.38807607e-01 -9.24936175e-01 -9.51490283e-01
1.83277112e-02 -1.49990588e-01 6.45456016e-01 -4.96133983e-01
1.00064255e-01 5.82533896e-01 -1.97196329e+00 7.67215610e-01
5.36902845e-01 1.15103841e+00 4.54964370e-01 1.09155130e+00
8.70907232e-02 1.02718830e+00 -5.67885637e-01 -9.18905213e-02
4.35594857e-01 -7.25275159e-01 -4.01615471e-01 1.32370949e-01
8.25981259e-01 -4.46528286e-01 -5.90886712e-01 1.26700926e+00
5.14665306e-01 6.28199428e-02 1.34695664e-01 -5.81543982e-01
-4.72211391e-01 4.08773720e-01 -4.62974310e-01 5.20425200e-01
2.37199068e-01 1.64309934e-01 5.41989446e-01 -6.32749617e-01
6.74053073e-01 1.17379618e+00 8.92002225e-01 8.84180903e-01
-1.44536662e+00 -5.34572601e-01 -2.28763685e-01 2.34838456e-01
-1.19972110e+00 -7.24751592e-01 6.66106403e-01 -6.07188195e-02
8.10636818e-01 2.86020488e-01 6.15549684e-01 1.00718808e+00
3.93860012e-01 3.21076244e-01 1.29611719e+00 -3.57609361e-01
2.87374765e-01 -1.57879367e-02 -4.95337039e-01 4.39672351e-01
1.10765686e-02 3.06342930e-01 -7.26657271e-01 3.10135901e-01
1.56463635e+00 -1.37143195e-01 -6.45590246e-01 -9.40032080e-02
-1.06592405e+00 4.16529953e-01 3.63074511e-01 6.36380911e-01
-3.46164167e-01 3.26759070e-01 4.96868283e-01 7.33052671e-01
4.37137067e-01 2.22893417e-01 -4.02462184e-01 -2.39892881e-02
-1.30042851e+00 1.90093309e-01 4.43949342e-01 6.28638923e-01
7.32610881e-01 5.30785143e-01 5.44770733e-02 7.12650776e-01
-4.14816976e-01 2.25400120e-01 6.75439656e-01 -1.17050469e+00
2.88482666e-01 -2.59295665e-02 2.53881738e-02 -8.89660895e-01
-3.47984791e-01 -4.10881877e-01 -1.29554629e+00 6.89375699e-01
4.35911179e-01 4.91636433e-02 -8.73442650e-01 1.68681026e+00
3.66539471e-02 3.61606061e-01 7.75386468e-02 1.11665261e+00
5.20879030e-01 8.71755123e-01 -8.73512775e-02 -6.00207746e-01
1.27993166e+00 -5.97159386e-01 -8.61357987e-01 -1.34713963e-01
1.41993821e-01 -7.45053291e-01 1.03768694e+00 7.80222893e-01
-1.58042800e+00 -1.17464209e+00 -1.21341145e+00 -2.35062346e-01
-1.20382600e-01 -4.58181091e-02 1.46725714e-01 4.63270009e-01
-1.47093499e+00 1.14606643e+00 -7.05541253e-01 -9.84467641e-02
3.21821809e-01 5.51862478e-01 -3.08218032e-01 6.88760430e-02
-1.26927757e+00 7.80037284e-01 3.82869214e-01 6.71098307e-02
-7.32073486e-01 -8.01065385e-01 -7.47586846e-01 3.36124688e-01
2.87042022e-01 -6.41454756e-01 1.09132373e+00 -1.76958203e+00
-1.90561152e+00 9.07203019e-01 -7.13738427e-02 -5.97344756e-01
6.42410398e-01 -1.52771175e-02 -7.54941583e-01 3.29005390e-01
-3.71396899e-01 4.15060282e-01 1.57150269e+00 -1.16513610e+00
-7.62323499e-01 -1.41056433e-01 1.63077307e-03 4.78905998e-02
-3.00616026e-01 6.18753918e-02 -5.19374967e-01 -9.68445361e-01
-1.69983525e-02 -5.83966732e-01 -3.96360546e-01 1.83107987e-01
2.66205400e-01 4.97319639e-01 9.75518823e-01 -6.02156579e-01
1.19248366e+00 -2.33848786e+00 3.08585495e-01 -1.85772598e-01
1.16169654e-01 5.44038236e-01 -2.45526135e-01 -8.23089629e-02
-4.07370269e-01 -1.55670866e-01 -2.44376779e-01 -2.31215492e-01
-6.95134819e-01 1.78911433e-01 -2.46617362e-01 4.48436648e-01
4.39620048e-01 5.42288423e-01 -8.05473149e-01 -3.84366751e-01
3.56707841e-01 8.52756023e-01 -5.58170259e-01 1.82568371e-01
-7.49170706e-02 4.51137275e-01 -3.62227038e-02 1.81446970e-01
8.31357598e-01 -5.06898053e-02 2.82943755e-01 -5.78906715e-01
-2.93230057e-01 -6.12990279e-03 -1.42710423e+00 1.62948060e+00
-8.26378167e-01 9.66121018e-01 4.93685752e-01 -1.04798770e+00
8.61792624e-01 3.10151786e-01 4.53701079e-01 -9.87297177e-01
1.24569856e-01 3.06807309e-01 -1.33700937e-01 -4.82359976e-01
6.43116474e-01 -3.42321992e-01 3.74023348e-01 1.52003199e-01
1.56127647e-01 -1.88886836e-01 2.07009166e-01 -2.08595514e-01
9.37343240e-01 2.13682100e-01 2.78255641e-01 -3.87021035e-01
8.61347318e-01 -3.50192815e-01 5.76390088e-01 5.81573486e-01
5.87045364e-02 8.86569977e-01 3.23780358e-01 -7.77597487e-01
-1.43943965e+00 -9.08054352e-01 -1.08734988e-01 1.02770329e+00
8.87414142e-02 -2.16696247e-01 -9.96609986e-01 -1.88914537e-01
-5.25843263e-01 1.73541322e-01 -5.03176153e-01 -1.31363139e-01
-1.09134269e+00 -6.68223917e-01 2.63911068e-01 4.19123709e-01
6.88283563e-01 -1.10059953e+00 -9.88149762e-01 4.63260621e-01
-3.86268459e-02 -1.43402886e+00 -3.29556316e-01 1.65898681e-01
-1.18601334e+00 -8.35676253e-01 -8.17716420e-01 -8.93435419e-01
4.63870317e-01 1.67600259e-01 1.29199576e+00 1.94761306e-01
-3.32214773e-01 -1.17768152e-02 -9.29092318e-02 8.71693641e-02
-6.98457241e-01 -1.82277665e-01 1.77767456e-01 2.95494735e-01
2.48186421e-02 -7.77895451e-01 -6.48926497e-01 1.49074018e-01
-1.31242919e+00 8.14521015e-02 6.76796496e-01 9.75884378e-01
6.98981166e-01 4.62416172e-01 3.62562805e-01 -8.33424330e-01
2.43114665e-01 1.23686567e-02 -8.24513853e-01 -1.29321694e-01
-2.93527782e-01 1.69154983e-02 1.13192797e+00 -8.06949317e-01
-1.25450444e+00 1.09876685e-01 -2.75833935e-01 -5.89279413e-01
-1.19294904e-01 -4.71164845e-02 -4.54990417e-02 -3.12614292e-01
8.74852836e-01 2.90146619e-01 -2.90072802e-02 -3.60791862e-01
1.82015151e-01 4.69863623e-01 9.51079667e-01 -1.22504607e-01
9.46428716e-01 7.12976992e-01 1.93087518e-01 -1.16222525e+00
-5.21473527e-01 -1.55446559e-01 -5.48354566e-01 -2.85317898e-01
8.79800498e-01 -1.07090378e+00 -5.14238298e-01 5.89779377e-01
-1.03086638e+00 -6.42772138e-01 -5.37191033e-01 3.18565607e-01
-8.44380081e-01 3.22871625e-01 -9.80566204e-01 -5.31919062e-01
-2.37828359e-01 -1.09095216e+00 9.07708764e-01 3.70809019e-01
2.49855608e-01 -8.73825312e-01 -1.48090005e-01 -1.31498510e-02
7.98142135e-01 7.58733414e-03 6.89436972e-01 1.98103547e-01
-6.83219075e-01 2.75435567e-01 -3.36197168e-01 7.29732990e-01
-2.59447098e-02 -6.97479621e-02 -1.19665205e+00 -3.46706599e-01
4.93260056e-01 -8.69101658e-02 9.64780211e-01 6.68550074e-01
1.43076205e+00 -1.33705795e-01 1.27900511e-01 8.88941407e-01
1.85405135e+00 7.65183866e-02 1.07983649e+00 2.84934968e-01
4.93809551e-01 4.69025224e-01 4.59778458e-01 2.50259161e-01
-3.14534307e-01 9.63748813e-01 3.16758871e-01 -5.40454984e-01
-5.79331338e-01 1.18409857e-01 7.01243877e-01 6.88517153e-01
-5.89551687e-01 -1.51440594e-02 -3.15202713e-01 3.75302166e-01
-1.61238074e+00 -1.31079543e+00 -1.84713393e-01 2.34055185e+00
9.91246402e-01 3.55273575e-01 3.72417807e-03 3.91827703e-01
7.57929802e-01 2.71083057e-01 -3.74146134e-01 -5.58442175e-01
-4.00323570e-01 5.99466145e-01 7.37911046e-01 5.44925809e-01
-9.91153240e-01 8.40346277e-01 6.31974125e+00 1.01274896e+00
-1.55468178e+00 2.56707966e-01 7.79247999e-01 -1.24338120e-01
4.39613163e-02 -3.30093175e-01 -4.67973411e-01 3.77081215e-01
1.12396395e+00 1.36542812e-01 7.11186290e-01 6.53326750e-01
4.14437711e-01 -1.54967070e-01 -1.10149479e+00 1.31027186e+00
1.03298575e-01 -1.55052245e+00 3.65438610e-02 -1.83674604e-01
8.26319814e-01 -5.89557327e-02 1.55109271e-01 3.64210010e-02
-2.98943937e-01 -8.67542505e-01 7.29355633e-01 3.98192853e-01
1.27314234e+00 -7.17329741e-01 6.90650165e-01 6.08292036e-02
-1.17887044e+00 -2.82303900e-01 -5.86037576e-01 -1.28722265e-02
1.90479264e-01 7.62267768e-01 4.08308245e-02 3.23182076e-01
1.10543144e+00 6.44358933e-01 -3.22164625e-01 5.92345178e-01
1.12720117e-01 2.91882217e-01 -5.63480966e-02 5.32136738e-01
-1.01177357e-01 -2.48451605e-01 5.12416959e-01 1.45891249e+00
3.80371332e-01 1.96176872e-01 -2.43647546e-01 6.92503512e-01
-8.23116004e-02 -1.89469948e-01 -3.77852678e-01 2.58047014e-01
-9.76698101e-02 1.20016861e+00 -7.09449410e-01 -5.20779610e-01
-5.66431284e-01 1.18332016e+00 -3.85747589e-02 4.19649780e-01
-8.15598726e-01 -3.56623143e-01 6.75556898e-01 4.69316810e-01
5.25214195e-01 -2.45145351e-01 -2.41104484e-01 -1.24866593e+00
4.11224328e-02 -1.12671578e+00 1.37275308e-01 -8.53307426e-01
-8.81708086e-01 8.70982409e-01 -3.10394585e-01 -1.36140072e+00
-2.11625248e-01 -7.22241282e-01 -3.93986732e-01 7.52958119e-01
-1.88227916e+00 -6.69841886e-01 -5.11717856e-01 8.52305651e-01
8.76775503e-01 -6.79359259e-03 5.11145532e-01 5.75034916e-01
-2.68672734e-01 3.31991971e-01 8.22369158e-02 2.65708249e-02
6.21221423e-01 -9.73264277e-01 3.75712603e-01 1.08787930e+00
-4.93556969e-02 9.72319767e-02 1.01069391e+00 -2.76910305e-01
-1.36107540e+00 -9.50949609e-01 5.53704977e-01 4.91310693e-02
5.98095536e-01 -3.87922674e-01 -1.22791672e+00 2.88486898e-01
2.01806799e-01 3.22792053e-01 1.65397916e-02 -4.05496538e-01
-1.78781018e-01 -5.07177532e-01 -1.12916803e+00 6.57230079e-01
1.03614724e+00 -6.75973773e-01 -3.07614058e-01 6.42400607e-02
5.20626962e-01 -4.86284107e-01 -8.97766173e-01 1.89622208e-01
3.86552989e-01 -1.52009690e+00 1.13998830e+00 -2.47598261e-01
7.45677531e-01 -3.91720593e-01 2.92814318e-02 -1.08944833e+00
-4.61617529e-01 -8.13009381e-01 -2.59183824e-01 8.97048056e-01
7.04327822e-02 -3.65526140e-01 7.65381277e-01 1.75390273e-01
-2.07257029e-02 -3.98750752e-01 -9.94110405e-01 -6.99104369e-01
-3.04273784e-01 -1.71606496e-01 1.93552300e-01 7.91162431e-01
-3.74658436e-01 8.03363398e-02 -5.06119192e-01 1.67562097e-01
7.10687220e-01 -2.06463322e-01 4.72903430e-01 -9.42853987e-01
-5.13060570e-01 -5.40302277e-01 -5.33582032e-01 -9.79220808e-01
-5.11130989e-02 -1.92782834e-01 7.33917207e-02 -8.81305158e-01
-5.72729558e-02 -2.18458161e-01 -2.81532139e-01 7.58383656e-03
-2.62033325e-02 7.12859392e-01 2.26193488e-01 1.55599684e-01
-1.34519950e-01 1.61722958e-01 1.30852115e+00 1.68941781e-01
-3.88839751e-01 -9.28788707e-02 -1.25846267e-01 7.24430799e-01
7.82040656e-01 -3.24612737e-01 -2.50458896e-01 -4.82789844e-01
3.32987785e-01 2.37019926e-01 4.08036202e-01 -1.39047945e+00
1.83283508e-01 -8.84322301e-02 5.52942574e-01 -1.91029627e-02
4.57150072e-01 -8.82944822e-01 2.78111011e-01 5.93744338e-01
-3.18141609e-01 6.47030845e-02 1.66926697e-01 3.52607936e-01
-4.82091308e-01 -1.79410279e-01 1.53021693e+00 -3.95819753e-01
-8.55796933e-01 1.83821276e-01 -4.80992615e-01 -1.44838139e-01
7.07477033e-01 -5.17919183e-01 -1.21372022e-01 -4.70841378e-01
-6.78623557e-01 -5.85478008e-01 6.83104217e-01 1.68010011e-01
6.02363765e-01 -1.12717879e+00 -7.54905641e-01 4.95409489e-01
-5.64661980e-01 3.63982804e-02 6.54235721e-01 7.34778702e-01
-7.95418084e-01 -5.15629128e-02 -6.87890112e-01 -6.14229143e-01
-1.32904196e+00 9.00517941e-01 6.47660553e-01 -1.37939394e-01
-8.32370460e-01 6.39380872e-01 2.16764316e-01 3.68013948e-01
5.37156165e-02 -4.27313030e-01 -1.71585649e-01 -1.47274747e-01
1.00116467e+00 2.94072956e-01 1.87704638e-01 -6.26744688e-01
9.99179780e-02 9.81383324e-01 9.53357071e-02 1.90118924e-02
1.42207503e+00 -1.89380944e-01 -1.25621602e-01 1.91036448e-01
1.22436512e+00 -7.21247941e-02 -1.51436639e+00 -2.39335462e-01
-2.35690847e-01 -6.08637452e-01 4.95000154e-01 -2.41980731e-01
-1.52354705e+00 8.94333422e-01 8.25363219e-01 2.74769098e-01
1.90047050e+00 -4.02342945e-01 7.75229990e-01 1.56564526e-02
2.77108759e-01 -1.42145002e+00 1.55700892e-01 2.18014568e-01
7.49769747e-01 -1.09682727e+00 9.44480523e-02 -4.03486490e-01
-3.13233167e-01 1.50068212e+00 4.65408176e-01 -3.15246671e-01
3.65828991e-01 7.30051756e-01 1.09715030e-01 2.23377004e-01
-6.98637903e-01 -1.57497868e-01 1.04104262e-02 7.16878057e-01
2.39126638e-01 -3.10762227e-01 -1.00448824e-01 2.51720157e-02
1.45526320e-01 3.28335464e-01 8.47232521e-01 6.10107899e-01
-3.80924881e-01 -8.60238135e-01 -4.88199055e-01 2.24760659e-02
-7.34693348e-01 -1.26355991e-01 2.33052045e-01 6.36270344e-01
2.87386745e-01 6.40148461e-01 3.44535440e-01 -2.44784683e-01
2.47033849e-01 -3.42262834e-01 6.23910725e-01 -2.44072363e-01
-6.57200575e-01 2.13047206e-01 -8.66136923e-02 -9.27178979e-01
-8.05242419e-01 -4.04800445e-01 -9.53575253e-01 -4.14149225e-01
-5.86643629e-02 -2.62691468e-01 5.86092234e-01 5.84565938e-01
9.82355475e-02 6.78036273e-01 5.63323557e-01 -1.38033414e+00
-3.56818825e-01 -5.92442632e-01 -6.04780614e-01 5.42943597e-01
7.15287745e-01 -2.23728374e-01 -3.79514992e-01 5.37695944e-01] | [11.139166831970215, -1.9827536344528198] |
5a91e889-4368-4e86-8ba8-1859c5c06e81 | 1m-parameters-are-enough-a-lightweight-cnn | 2306.16103 | null | https://arxiv.org/abs/2306.16103v2 | https://arxiv.org/pdf/2306.16103v2.pdf | 1M parameters are enough? A lightweight CNN-based model for medical image segmentation | Convolutional neural networks (CNNs) and Transformer-based models are being widely applied in medical image segmentation thanks to their ability to extract high-level features and capture important aspects of the image. However, there is often a trade-off between the need for high accuracy and the desire for low computational cost. A model with higher parameters can theoretically achieve better performance but also result in more computational complexity and higher memory usage, and thus is not practical to implement. In this paper, we look for a lightweight U-Net-based model which can remain the same or even achieve better performance, namely U-Lite. We design U-Lite based on the principle of Depthwise Separable Convolution so that the model can both leverage the strength of CNNs and reduce a remarkable number of computing parameters. Specifically, we propose Axial Depthwise Convolutions with kernels 7x7 in both the encoder and decoder to enlarge the model receptive field. To further improve the performance, we use several Axial Dilated Depthwise Convolutions with filters 3x3 for the bottleneck as one of our branches. Overall, U-Lite contains only 878K parameters, 35 times less than the traditional U-Net, and much more times less than other modern Transformer-based models. The proposed model cuts down a large amount of computational complexity while attaining an impressive performance on medical segmentation tasks compared to other state-of-the-art architectures. The code will be available at: https://github.com/duong-db/U-Lite. | ['Van-Truong Pham', 'Thi-Thao Tran', 'Thanh-Thu Nguyen', 'Binh-Duong Dinh'] | 2023-06-28 | null | null | null | null | ['medical-image-segmentation'] | ['medical'] | [ 1.35787606e-01 7.36676455e-02 -1.75159216e-01 -3.57566118e-01
-5.73805809e-01 -2.14130625e-01 9.20909569e-02 1.43025592e-02
-6.21695280e-01 4.15740639e-01 -6.87742233e-02 -7.21377075e-01
8.53595957e-02 -1.04036891e+00 -5.71494222e-01 -5.08688390e-01
6.52287006e-02 -4.76767384e-02 6.32691085e-01 -5.53851537e-02
-3.16274352e-02 4.39187706e-01 -1.11492944e+00 3.37539345e-01
8.04755390e-01 1.34038687e+00 3.63668114e-01 5.48313975e-01
-8.20948482e-02 6.80344880e-01 -1.19831048e-01 -3.60871077e-01
2.88865447e-01 -2.54264653e-01 -9.49559212e-01 -1.46711901e-01
2.15629503e-01 -4.83935267e-01 -4.68425542e-01 9.74462986e-01
6.18806362e-01 -6.64994195e-02 2.17349112e-01 -7.87378669e-01
-4.57492232e-01 6.76097751e-01 -5.88257015e-01 3.77451122e-01
-3.01341653e-01 1.38309896e-01 8.17389011e-01 -5.30916274e-01
2.58512616e-01 9.17917311e-01 8.12301397e-01 7.03897655e-01
-1.00913191e+00 -8.37522864e-01 3.61575708e-02 4.09603678e-02
-1.20208371e+00 -4.01532203e-01 4.46157277e-01 -7.83452690e-02
9.42714810e-01 2.66694516e-01 7.14644969e-01 7.14321554e-01
3.36929858e-01 9.41896260e-01 9.18320179e-01 -1.16128981e-01
1.02513984e-01 -1.66560471e-01 3.80407684e-02 9.93226469e-01
1.83302104e-01 -1.38350487e-01 -1.55367792e-01 1.69189095e-01
1.17855275e+00 3.57820064e-01 -4.48185772e-01 -4.20636358e-03
-1.23248768e+00 8.19143414e-01 9.30519819e-01 4.96103764e-01
-3.29459071e-01 4.00492370e-01 5.67521095e-01 1.00947529e-01
3.27593833e-01 4.69552875e-01 -6.06372476e-01 -1.39365330e-01
-1.03937113e+00 5.92091195e-02 2.61062413e-01 7.91974545e-01
6.85285509e-01 -1.82086289e-01 -1.90766364e-01 7.55916178e-01
3.10880281e-02 -6.82961345e-02 5.99481165e-01 -7.07827628e-01
3.39871526e-01 7.82530963e-01 -4.58934814e-01 -5.59944212e-01
-5.26519477e-01 -6.69809818e-01 -1.13303792e+00 6.90437332e-02
4.55432832e-01 -2.16624945e-01 -1.28745306e+00 1.52424908e+00
1.66373298e-01 1.47610530e-01 -1.46271437e-01 7.84231842e-01
7.88820207e-01 6.28225029e-01 -8.85124058e-02 2.68161893e-01
1.57064736e+00 -1.17014945e+00 -5.22573411e-01 -3.81604046e-01
9.36179221e-01 -8.21086884e-01 9.95872378e-01 9.37206596e-02
-1.20333540e+00 -5.01346946e-01 -1.04997909e+00 -3.94017994e-01
-2.49037609e-01 2.30072185e-01 8.15759718e-01 6.75703585e-01
-1.12600553e+00 7.59914219e-01 -1.13029063e+00 -1.30417407e-01
7.99589097e-01 5.25902748e-01 -1.93276107e-01 -1.81595504e-01
-1.12847161e+00 6.33660316e-01 4.62355584e-01 2.76015341e-01
-4.42472756e-01 -8.39457929e-01 -8.78375590e-01 3.07524741e-01
3.42646331e-01 -6.90070748e-01 1.32943392e+00 -5.52462816e-01
-1.32854939e+00 5.67362487e-01 -6.97360467e-03 -5.68239808e-01
4.80361849e-01 -3.20247039e-02 2.42980532e-02 1.59094781e-01
-1.03948608e-01 9.94476914e-01 4.01908010e-01 -5.23575902e-01
-7.02459633e-01 -2.78224349e-01 1.92529321e-01 -4.69201691e-02
-6.03800356e-01 -1.83796108e-01 -8.51007760e-01 -7.30603278e-01
3.10149193e-01 -9.21214998e-01 -4.53975499e-01 2.90872037e-01
-2.40128458e-01 -4.87763584e-02 8.77809644e-01 -4.77699041e-01
1.39901149e+00 -2.20811677e+00 -3.24040174e-01 -3.26634588e-04
4.75154221e-01 7.64407575e-01 2.25113630e-02 4.17695455e-02
5.62566705e-03 6.29779175e-02 -3.90902609e-01 -1.97560564e-01
-4.21129346e-01 2.97912985e-01 1.65589496e-01 3.13532680e-01
2.82952994e-01 1.24533689e+00 -8.01028788e-01 -4.71509844e-01
3.16225171e-01 5.71639597e-01 -8.06847274e-01 -1.55537590e-01
1.22855529e-01 3.56439054e-01 -6.33227229e-01 5.84883749e-01
7.92005122e-01 -4.78186488e-01 9.67707038e-02 -3.37036967e-01
-1.66578457e-01 4.05554235e-01 -8.69051218e-01 1.70247436e+00
-5.68933189e-01 6.33559406e-01 1.48801744e-01 -1.23272526e+00
7.82030761e-01 3.93110245e-01 4.82161164e-01 -8.35394979e-01
5.46334088e-01 4.54180539e-01 3.25332254e-01 -2.62758911e-01
1.72550797e-01 -1.39988765e-01 1.38030872e-01 2.19079033e-01
-4.57010269e-02 9.68900248e-02 1.70490250e-01 3.17306407e-02
1.14462125e+00 -3.64308618e-02 1.97141916e-01 -4.60200906e-01
3.77522796e-01 -2.26634398e-01 6.89971626e-01 6.24076307e-01
-2.45622724e-01 5.09253204e-01 5.71603477e-01 -6.47761405e-01
-8.59640896e-01 -6.27595901e-01 -3.89946252e-01 6.49953544e-01
6.34239763e-02 -4.79921907e-01 -8.94551873e-01 -5.17203629e-01
-2.13494465e-01 2.86098599e-01 -6.43059492e-01 -1.38417900e-01
-7.34113276e-01 -8.40754509e-01 7.98883736e-01 9.61141229e-01
9.81305182e-01 -8.66473913e-01 -1.18636465e+00 3.42816740e-01
-7.43216425e-02 -1.13649654e+00 -6.18619204e-01 4.69707936e-01
-1.26652205e+00 -8.40410829e-01 -9.03806269e-01 -8.42407823e-01
8.03362668e-01 2.58546829e-01 8.19466770e-01 1.43696785e-01
-4.47872519e-01 -3.47782910e-01 -3.28377247e-01 -4.48787719e-01
-5.03126420e-02 3.72393966e-01 -4.02062297e-01 -2.58961946e-01
2.77656883e-01 -5.14182448e-01 -9.81824338e-01 3.15942138e-01
-1.12102818e+00 4.23071504e-01 8.27896416e-01 1.00288141e+00
4.65878904e-01 1.37993887e-01 3.97798240e-01 -9.23964381e-01
2.99798489e-01 -9.90654677e-02 -4.95976835e-01 -7.01305941e-02
-5.91987371e-01 9.15791094e-02 7.70805418e-01 -3.56477350e-01
-8.05340588e-01 7.74266049e-02 -5.09089887e-01 -4.47947115e-01
2.24326983e-01 4.70158279e-01 1.02397308e-01 -2.04344362e-01
4.19179142e-01 7.90565610e-02 1.13941997e-01 -5.13971388e-01
6.05511516e-02 6.38372004e-01 2.99615890e-01 -3.73258322e-01
3.70627135e-01 5.40145099e-01 -2.30606012e-02 -6.63382292e-01
-8.19778144e-01 -4.72563624e-01 -4.41311300e-01 8.06768388e-02
8.73908877e-01 -8.00886333e-01 -6.98418677e-01 6.06220007e-01
-1.02634215e+00 -5.65313160e-01 -1.40627205e-01 5.32413721e-01
-2.20167056e-01 1.13128506e-01 -1.04248655e+00 -3.25600088e-01
-6.14802003e-01 -1.52896607e+00 8.89910638e-01 4.12819207e-01
-2.14112159e-02 -8.90525877e-01 -5.90745866e-01 3.42824459e-01
6.75120831e-01 9.97362658e-02 9.24100935e-01 -2.13403985e-01
-5.65100312e-01 -1.23890579e-01 -6.08972371e-01 3.88409793e-01
9.68101025e-02 -3.90637159e-01 -9.95519757e-01 -3.44075620e-01
-6.73402920e-02 -2.16418520e-01 1.14408410e+00 6.72868192e-01
1.54083633e+00 -1.19107947e-01 -3.39212835e-01 9.58969712e-01
1.47763598e+00 3.86112779e-01 8.93829167e-01 2.97851115e-01
6.39713764e-01 2.91911960e-01 2.88755894e-01 2.39274025e-01
2.48659655e-01 5.06472766e-01 3.89770836e-01 -6.92742109e-01
-1.70229480e-01 -2.40046927e-03 -1.17869943e-01 7.85693228e-01
-3.82371731e-02 -9.00913700e-02 -9.55117404e-01 6.86261892e-01
-1.79715800e+00 -6.34780705e-01 -4.04372960e-02 2.14048672e+00
8.64071012e-01 3.81019503e-01 -4.75718826e-02 2.60149628e-01
4.85604197e-01 6.60281330e-02 -5.80827534e-01 -4.47711229e-01
4.07392859e-01 4.95185912e-01 8.35888803e-01 2.82584161e-01
-1.05541956e+00 8.04090977e-01 5.71263504e+00 1.02006972e+00
-1.51973569e+00 1.68181881e-01 1.03614986e+00 -2.17362955e-01
6.14159591e-02 -2.12872162e-01 -7.63211668e-01 3.56406182e-01
7.49857843e-01 2.41217285e-01 6.17937297e-02 7.73301363e-01
7.02679753e-02 -1.26748353e-01 -9.13161278e-01 9.87780154e-01
-3.17394465e-01 -1.63074160e+00 -1.91002771e-01 3.06844503e-01
4.90845799e-01 2.11424962e-01 -9.58645425e-04 1.97599918e-01
-1.10136978e-01 -9.89691198e-01 5.41224778e-01 -1.23761036e-01
9.77056026e-01 -7.04425216e-01 9.24039721e-01 3.37294072e-01
-1.36465108e+00 -1.02223232e-01 -4.58019316e-01 -9.49113220e-02
8.63773301e-02 7.32083142e-01 -4.80979592e-01 3.98818046e-01
1.03845000e+00 5.69058061e-01 -2.87472725e-01 1.00306022e+00
1.33471161e-01 5.48733950e-01 -3.67637455e-01 -6.22139163e-02
7.74873078e-01 -7.62037095e-03 4.08874974e-02 1.27979410e+00
3.71077508e-01 1.95594996e-01 4.29442897e-03 7.27265060e-01
-2.88348168e-01 -1.13889813e-01 -2.74053812e-01 1.57422617e-01
2.43507966e-01 1.29750907e+00 -1.09893799e+00 -4.41388756e-01
-6.04468286e-01 9.32725430e-01 2.09800333e-01 -1.04937097e-02
-9.62130308e-01 -6.63129449e-01 6.46120310e-01 3.29773426e-01
4.96757179e-01 -7.15427548e-02 -4.33617711e-01 -1.03193486e+00
1.40422985e-01 -7.56854713e-01 1.86042488e-01 -3.12616438e-01
-7.55895257e-01 8.02004576e-01 -2.47982919e-01 -1.06573212e+00
2.32013971e-01 -8.92640710e-01 -5.68210065e-01 8.72787356e-01
-1.81422937e+00 -1.04561555e+00 -3.49203259e-01 4.81098443e-01
5.57396054e-01 3.17477286e-01 6.42854512e-01 7.22288787e-01
-6.47574663e-01 6.39754355e-01 -2.14091525e-03 4.50043947e-01
5.02530992e-01 -1.03089488e+00 6.42886937e-01 8.27716112e-01
-1.81799009e-01 7.84502089e-01 1.98212937e-01 -2.80922920e-01
-1.14451218e+00 -1.18049574e+00 7.81574428e-01 4.13228646e-02
4.42770213e-01 -3.01211923e-01 -8.74280214e-01 5.30386150e-01
-5.38995229e-02 3.78275841e-01 5.84394991e-01 -8.85812044e-02
-1.37767121e-01 -1.74765691e-01 -1.03294015e+00 6.42851830e-01
8.94894719e-01 -2.64862508e-01 -9.54015926e-02 5.19548394e-02
6.09271228e-01 -6.58178508e-01 -9.30952311e-01 5.09148479e-01
7.29216516e-01 -1.11514163e+00 9.59806085e-01 -6.75656646e-02
6.82124734e-01 -1.56029955e-01 1.13135979e-01 -1.01956940e+00
-5.37300050e-01 -4.91285294e-01 1.08158939e-01 6.71684861e-01
5.31999111e-01 -8.87465000e-01 9.43157971e-01 4.03320909e-01
-3.68125767e-01 -1.61057794e+00 -1.02801919e+00 -6.88051701e-01
2.67773092e-01 -3.84036690e-01 5.66170573e-01 7.53306091e-01
-7.50921369e-02 3.30974936e-01 -1.17963292e-01 -1.13856785e-01
3.96466941e-01 -4.80247773e-02 3.14652324e-01 -9.34444547e-01
-2.60191083e-01 -6.80400193e-01 -3.97548139e-01 -1.49146473e+00
-3.96120727e-01 -8.69320154e-01 -1.49486244e-01 -1.63021767e+00
1.42764837e-01 -6.49262369e-01 -3.72501522e-01 8.99336874e-01
-2.13601053e-01 6.48729384e-01 1.90600857e-01 1.14232011e-01
-1.10104047e-01 3.46346557e-01 1.68769979e+00 -8.54032114e-02
-1.98420674e-01 -6.14040755e-02 -8.66523087e-01 7.73411810e-01
8.64307582e-01 -2.05173656e-01 -4.49202269e-01 -7.70720780e-01
-4.90597412e-02 1.33417338e-01 3.19285125e-01 -1.10017264e+00
2.41236761e-01 2.32960552e-01 4.51537460e-01 -5.26023209e-01
2.63484806e-01 -7.18020141e-01 -2.34306872e-01 8.99034917e-01
-2.12559134e-01 7.47911632e-02 3.48433554e-01 8.37823078e-02
-3.33054751e-01 -1.34067908e-01 1.11178041e+00 -3.38139772e-01
-5.60316563e-01 6.56295359e-01 -2.32961163e-01 -2.94758350e-01
9.77074027e-01 -4.21875983e-01 -2.51366049e-01 -2.03291830e-02
-4.28198516e-01 2.19297424e-01 2.03101069e-01 1.86394244e-01
6.24689162e-01 -1.03915012e+00 -4.82591093e-01 2.99887896e-01
-9.62173864e-02 3.58527154e-01 4.43870544e-01 1.13744390e+00
-9.09691870e-01 6.46779239e-01 -2.24149182e-01 -5.10692239e-01
-1.15254414e+00 3.37131232e-01 4.59994674e-01 -5.58583617e-01
-1.01919365e+00 1.03364635e+00 5.03455997e-01 -3.45036536e-01
1.29593629e-03 -5.29626548e-01 -4.70220633e-02 -1.71367481e-01
6.33071005e-01 2.66819566e-01 2.29073137e-01 -2.61845052e-01
-4.20617521e-01 4.12365615e-01 -3.57068390e-01 3.12400281e-01
1.32046175e+00 1.55481443e-01 -1.19898841e-01 -1.90512732e-01
1.38461602e+00 -4.14219856e-01 -1.35030198e+00 -3.37637544e-01
-3.19754392e-01 -3.67012262e-01 6.71598196e-01 -5.19886494e-01
-1.60617936e+00 1.08331311e+00 6.68444514e-01 -8.10831487e-02
1.42728996e+00 -7.74855316e-02 1.45683014e+00 1.17322579e-01
2.46352136e-01 -8.57764006e-01 -6.83756322e-02 3.51379246e-01
3.07048082e-01 -1.14154935e+00 -7.54296372e-04 -6.28629982e-01
-2.62269676e-01 1.15189528e+00 6.69836938e-01 -1.72198102e-01
7.27885365e-01 7.40882635e-01 1.16299018e-01 -2.22297400e-01
-5.09995162e-01 -2.30481803e-01 1.96963206e-01 2.16983378e-01
7.85994112e-01 2.72525959e-02 -3.71553451e-01 4.02817369e-01
-6.22619763e-02 2.33707279e-01 3.10572535e-01 1.01986456e+00
-3.00045401e-01 -1.12436926e+00 -1.65307581e-01 7.46991217e-01
-7.64822304e-01 -3.09578776e-01 3.14334512e-01 5.89641809e-01
2.50983298e-01 6.75603867e-01 1.65852934e-01 -3.42484772e-01
1.96182236e-01 -2.94922084e-01 4.59426194e-01 -4.70363736e-01
-6.98561072e-01 3.42592895e-01 -2.08950356e-01 -7.04973638e-01
-3.13342005e-01 -2.73610502e-01 -1.24777973e+00 -4.73833203e-01
-4.54313725e-01 -1.59369662e-01 5.18883586e-01 7.92119682e-01
3.36513340e-01 1.01910388e+00 2.40082905e-01 -7.16561675e-01
-3.72143477e-01 -8.39391649e-01 -4.42738086e-01 -4.50767428e-02
2.18661100e-01 -4.51924324e-01 8.94412994e-02 -1.11837618e-01] | [14.53321647644043, -2.6128926277160645] |
5597a497-0319-4b48-8c29-e28739ecd227 | a-causal-lens-for-peeking-into-black-box | 2008.00357 | null | https://arxiv.org/abs/2008.00357v1 | https://arxiv.org/pdf/2008.00357v1.pdf | A Causal Lens for Peeking into Black Box Predictive Models: Predictive Model Interpretation via Causal Attribution | With the increasing adoption of predictive models trained using machine learning across a wide range of high-stakes applications, e.g., health care, security, criminal justice, finance, and education, there is a growing need for effective techniques for explaining such models and their predictions. We aim to address this problem in settings where the predictive model is a black box; That is, we can only observe the response of the model to various inputs, but have no knowledge about the internal structure of the predictive model, its parameters, the objective function, and the algorithm used to optimize the model. We reduce the problem of interpreting a black box predictive model to that of estimating the causal effects of each of the model inputs on the model output, from observations of the model inputs and the corresponding outputs. We estimate the causal effects of model inputs on model output using variants of the Rubin Neyman potential outcomes framework for estimating causal effects from observational data. We show how the resulting causal attribution of responsibility for model output to the different model inputs can be used to interpret the predictive model and to explain its predictions. We present results of experiments that demonstrate the effectiveness of our approach to the interpretation of black box predictive models via causal attribution in the case of deep neural network models trained on one synthetic data set (where the input variables that impact the output variable are known by design) and two real-world data sets: Handwritten digit classification, and Parkinson's disease severity prediction. Because our approach does not require knowledge about the predictive model algorithm and is free of assumptions regarding the black box predictive model except that its input-output responses be observable, it can be applied, in principle, to any black box predictive model. | ['Aria Khademi', 'Vasant Honavar'] | 2020-08-01 | null | null | null | null | ['severity-prediction'] | ['computer-vision'] | [ 5.94517767e-01 6.82309389e-01 -6.97544217e-01 -5.26020825e-01
-2.79042393e-01 -4.12322640e-01 6.63438261e-01 2.35084578e-01
-3.10948074e-01 9.21352148e-01 4.33377743e-01 -8.47688317e-01
-6.30778193e-01 -8.91897619e-01 -1.22953248e+00 -4.64404464e-01
-1.15893632e-02 6.90230668e-01 -3.12812299e-01 3.00625175e-01
7.46387392e-02 2.48391792e-01 -1.02160609e+00 3.05172324e-01
8.80874932e-01 7.24003673e-01 -1.00311436e-01 7.25265801e-01
3.58621657e-01 1.14725542e+00 -4.73301470e-01 -4.02791589e-01
-9.84390825e-02 -5.42966604e-01 -7.81970620e-01 -2.84483284e-01
2.98819840e-02 -4.18053418e-01 -1.64141417e-01 9.54326987e-01
-3.62663046e-02 -1.09471276e-01 1.13466811e+00 -1.33185208e+00
-7.54400194e-01 8.64364505e-01 -2.31324136e-01 -4.73924913e-02
-2.54421942e-02 1.45525560e-01 1.19277048e+00 -1.41984224e-01
3.82475644e-01 1.45873141e+00 5.89679718e-01 4.79210407e-01
-1.75028551e+00 -7.02996314e-01 1.23507783e-01 1.41828237e-02
-6.44216120e-01 -2.04028174e-01 3.10096353e-01 -9.64273393e-01
5.58019042e-01 6.27483279e-02 4.26678956e-01 1.35057425e+00
5.13489783e-01 3.30407619e-01 1.09239209e+00 -4.45822358e-01
4.76404369e-01 1.61737770e-01 2.21497327e-01 4.88693565e-01
3.23081642e-01 7.68393934e-01 -4.08714861e-01 -4.73168164e-01
8.26046824e-01 2.13253886e-01 -4.65081111e-02 -1.74599349e-01
-8.87636006e-01 1.00336444e+00 4.32951272e-01 -1.19804397e-01
-5.93761861e-01 6.35141611e-01 2.83155236e-02 6.93924353e-02
4.85941857e-01 5.21499515e-01 -8.15997243e-01 2.80751705e-01
-5.89189112e-01 3.42445612e-01 7.64976680e-01 4.46266860e-01
2.74267614e-01 -2.79635757e-01 -1.56902611e-01 2.89467633e-01
5.18080473e-01 4.83041078e-01 2.87589014e-01 -1.15334785e+00
4.37692583e-01 5.05165994e-01 3.82432193e-01 -7.60131299e-01
-4.34067816e-01 -1.58464119e-01 -5.36904991e-01 2.57048517e-01
6.42282784e-01 -4.17653382e-01 -9.02103007e-01 2.00130272e+00
3.18701006e-02 4.94889244e-02 2.33911742e-02 6.26147509e-01
4.74339157e-01 6.18352950e-01 4.42382276e-01 -8.99344012e-02
1.05917561e+00 -1.88531071e-01 -5.39160430e-01 -3.28473181e-01
8.48495901e-01 -2.25857824e-01 9.32228982e-01 3.37830707e-02
-1.03629696e+00 -2.37177029e-01 -8.31394494e-01 1.00579150e-01
2.50817109e-02 2.58215740e-02 4.41701174e-01 4.40610677e-01
-6.43042982e-01 9.30954099e-01 -9.07689154e-01 -1.29852906e-01
7.83423841e-01 7.06301093e-01 -2.94117421e-01 3.22889350e-02
-1.18488765e+00 1.08147132e+00 4.32200909e-01 -1.74646765e-01
-1.02123773e+00 -9.59938288e-01 -4.47490335e-01 7.29878068e-01
3.64413559e-01 -9.50241745e-01 1.44744682e+00 -1.37155676e+00
-9.96227205e-01 4.72661704e-01 -1.28343612e-01 -4.58096266e-01
5.06394088e-01 -1.35243654e-01 -5.12831621e-02 -4.49198246e-01
9.23072174e-02 2.84314364e-01 5.54768384e-01 -9.06788051e-01
-6.27275229e-01 -3.50368857e-01 -4.17306833e-02 -1.21726431e-01
8.89879912e-02 -6.05309904e-02 3.21448505e-01 -4.79506671e-01
-2.08172545e-01 -9.30080831e-01 -3.48416358e-01 -1.11999691e-01
-6.15561724e-01 1.90383885e-02 3.54168296e-01 -5.61302125e-01
8.98239732e-01 -1.94308269e+00 5.99414892e-02 3.57289076e-01
3.03554922e-01 -8.38604942e-02 7.41821453e-02 2.72756368e-01
-4.65019822e-01 6.07248902e-01 -3.73076051e-01 -9.85883996e-02
-7.24306405e-02 4.02955800e-01 -4.72340524e-01 4.86087739e-01
2.94999927e-01 9.90977824e-01 -6.99325264e-01 -3.05849850e-01
1.13044135e-01 2.71780431e-01 -7.27450669e-01 2.42165849e-01
-3.73856306e-01 5.36867857e-01 -6.97703362e-01 -1.49707362e-01
9.85765085e-03 -4.68943387e-01 3.74337167e-01 5.01042008e-01
2.69188613e-01 5.34193993e-01 -9.98530507e-01 5.67075074e-01
-4.24882293e-01 5.87588429e-01 -4.30386990e-01 -1.21471000e+00
5.23194909e-01 4.43887085e-01 1.70727924e-01 -4.88209039e-01
1.76107273e-01 1.97869223e-02 4.27447379e-01 -6.12126291e-01
-8.28440115e-02 -7.73941278e-01 -5.04923388e-02 7.54887044e-01
-2.20546667e-02 3.20593923e-01 -4.91020083e-01 4.97756787e-02
1.08515787e+00 -1.28847376e-01 6.10609710e-01 -2.45509550e-01
-4.76881862e-02 1.67189538e-01 6.18721485e-01 1.04833138e+00
2.63123512e-01 4.11718905e-01 1.21245348e+00 -5.49619257e-01
-1.24553728e+00 -9.04001474e-01 -2.24473029e-01 8.15966904e-01
-3.88080865e-01 1.81166947e-01 -5.59607744e-01 -4.94710922e-01
2.21584201e-01 1.27385271e+00 -1.21316195e+00 -4.00132447e-01
-3.66797537e-01 -8.38027179e-01 3.94991666e-01 7.21151233e-01
-1.75385535e-01 -9.80163753e-01 -8.31475496e-01 1.30133420e-01
1.54155061e-01 -6.80583537e-01 9.47713852e-04 2.36196876e-01
-1.12738073e+00 -1.35895622e+00 -9.42382663e-02 -1.18664406e-01
7.54266262e-01 -6.49928629e-01 1.08021367e+00 5.34450784e-02
1.53570980e-01 1.01358093e-01 1.62219852e-01 -8.78001571e-01
-9.12684798e-01 -2.42572188e-01 -2.25534141e-02 -9.76067185e-02
1.74416661e-01 -4.04536903e-01 -4.17401880e-01 9.85694081e-02
-8.64673495e-01 3.11674863e-01 5.43214440e-01 1.03964436e+00
3.35001051e-01 -2.94833556e-02 6.99028075e-01 -1.53324199e+00
5.78753710e-01 -7.98824430e-01 -7.86039174e-01 4.01874453e-01
-8.69668722e-01 3.65875840e-01 7.90841401e-01 -5.41722834e-01
-1.07073188e+00 -9.13294777e-02 3.41592431e-01 -1.22105159e-01
-2.16915056e-01 8.10852230e-01 -1.66976422e-01 6.52166069e-01
7.68771648e-01 -3.63735974e-01 3.47372256e-02 -4.72693026e-01
1.45943806e-01 4.00492817e-01 4.47477371e-01 -6.08092785e-01
3.52196306e-01 2.14436367e-01 5.08494198e-01 -1.77559614e-01
-9.47634518e-01 3.54649782e-01 -3.79999459e-01 8.38679299e-02
8.46204042e-01 -7.09068418e-01 -1.16602254e+00 -6.27431124e-02
-1.17679000e+00 -5.95320284e-01 -4.65032101e-01 6.73788667e-01
-7.34589517e-01 -6.17810130e-01 -1.85298979e-01 -9.57562149e-01
2.11323246e-01 -1.20430958e+00 5.92612982e-01 9.37954113e-02
-6.45917594e-01 -1.44874907e+00 -6.13565482e-02 2.93352872e-01
1.91513777e-01 1.79168209e-01 1.89040947e+00 -1.02857184e+00
-5.81951380e-01 -4.39607948e-01 -9.05726105e-02 1.47632226e-01
1.03032300e-02 9.78348330e-02 -9.56903458e-01 2.93311596e-01
-3.88343558e-02 -4.67757806e-02 6.54907703e-01 1.05280066e+00
1.40029633e+00 -9.57739592e-01 -4.19817835e-01 2.56725222e-01
1.02984142e+00 3.81556809e-01 3.95349771e-01 1.28038421e-01
4.64173526e-01 8.24881196e-01 2.42221043e-01 6.63028955e-02
4.26565439e-01 6.21686101e-01 4.85164613e-01 -1.12334423e-01
4.97494429e-01 -7.05230176e-01 2.64492482e-01 -2.72756368e-01
-3.33714904e-03 -1.39462516e-01 -1.04397857e+00 4.31834012e-01
-2.14375877e+00 -9.82812762e-01 -1.99026823e-01 2.48362947e+00
8.20902824e-01 5.33669144e-02 8.86277035e-02 -2.21243933e-01
5.55442452e-01 -4.99790370e-01 -8.70389760e-01 -7.02840388e-01
1.72613025e-01 2.24528939e-01 7.57699728e-01 6.68137550e-01
-8.11645865e-01 3.93276006e-01 7.10464239e+00 3.46786618e-01
-9.35839713e-01 1.18203633e-01 1.30657494e+00 -2.19761282e-01
-4.31403726e-01 2.96767116e-01 -4.39777374e-01 4.86580104e-01
1.37102425e+00 -5.17260075e-01 4.17526484e-01 5.43179393e-01
7.29277313e-01 -2.17947081e-01 -1.84506214e+00 4.98702191e-02
-5.88316739e-01 -1.41933286e+00 -9.74627491e-03 3.71728450e-01
8.15626383e-01 -3.12413067e-01 2.61668026e-01 8.06298107e-03
1.00385988e+00 -1.77955532e+00 7.30878174e-01 7.46116042e-01
6.61009490e-01 -4.57379907e-01 8.68602335e-01 7.55469382e-01
9.81941260e-03 -4.35628593e-01 -7.93034658e-02 -5.67572653e-01
-2.71510273e-01 5.52977145e-01 -1.18533599e+00 -1.01344556e-01
3.47881258e-01 4.99123305e-01 -4.43402082e-02 6.34047151e-01
-5.03850520e-01 1.31081951e+00 -6.06041811e-02 5.57935424e-02
-4.50725481e-02 6.51430562e-02 1.73528269e-01 6.94388747e-01
1.18942790e-01 3.06059480e-01 -2.64064521e-01 1.43193126e+00
-2.54385144e-01 -1.65959671e-01 -6.66379333e-01 -1.82533577e-01
2.59629071e-01 4.63509351e-01 -2.16344386e-01 -3.28363419e-01
-2.25010619e-01 -1.36779314e-02 1.30012378e-01 6.25683069e-01
-7.41721570e-01 3.33297789e-01 6.66577756e-01 2.99302876e-01
-7.85346851e-02 3.36828142e-01 -9.85763788e-01 -7.34605312e-01
-2.51620561e-01 -7.72238791e-01 4.93904978e-01 -9.49670672e-01
-1.19872975e+00 -2.75797993e-01 3.79313409e-01 -5.28441012e-01
-5.95607996e-01 -5.48552036e-01 -7.54317224e-01 1.55248141e+00
-1.08274913e+00 -8.42536986e-01 4.46961403e-01 2.14699537e-01
7.56246448e-02 5.15527017e-02 8.70018184e-01 -3.47077042e-01
-4.72166568e-01 2.08770230e-01 2.72873282e-01 1.27429038e-01
3.44639540e-01 -1.15690839e+00 3.54079664e-01 5.63197434e-01
-1.08704954e-01 6.04124546e-01 8.98665369e-01 -7.98554480e-01
-8.91100764e-01 -1.15844810e+00 1.11001825e+00 -7.33908713e-01
8.59090745e-01 -7.26874843e-02 -8.45185518e-01 1.14886439e+00
-3.81452173e-01 -1.05401985e-01 7.15672553e-01 5.16634583e-01
-1.91459030e-01 6.37309626e-02 -1.17290580e+00 4.64276880e-01
5.72923660e-01 -2.56814331e-01 -5.83340168e-01 3.03577870e-01
7.08847523e-01 -2.20806763e-01 -6.87120914e-01 1.48784488e-01
7.28062987e-01 -6.64072037e-01 8.41245413e-01 -1.61160803e+00
1.25760305e+00 9.01307017e-02 -6.25230223e-02 -1.52563679e+00
-3.38557959e-01 -4.30534892e-02 4.51543974e-03 9.22384918e-01
8.23674619e-01 -8.24099481e-01 5.75856268e-01 1.39703643e+00
3.81689250e-01 -8.16091478e-01 -9.90345359e-01 -1.59650207e-01
4.64606583e-01 -4.55524266e-01 7.62739182e-01 8.79289448e-01
-1.39324531e-01 4.22046125e-01 -4.15526897e-01 4.12418485e-01
5.90259790e-01 1.73015714e-01 5.87877572e-01 -1.46130586e+00
-6.52388453e-01 -3.64904463e-01 -2.69586474e-01 -3.88120323e-01
3.14369053e-01 -7.21657217e-01 -1.26725346e-01 -1.37726796e+00
6.14472091e-01 -6.84771717e-01 -2.58710682e-01 7.61787117e-01
-4.87630665e-01 -5.02438605e-01 2.88579941e-01 3.02263469e-01
1.93394095e-01 2.93858737e-01 1.03651452e+00 6.89857751e-02
-1.09680377e-01 3.79671335e-01 -1.03596354e+00 9.69988942e-01
5.51920116e-01 -1.12524450e+00 -4.60894942e-01 -4.52965528e-01
4.74403977e-01 5.19159973e-01 1.00849581e+00 -3.71791601e-01
2.05723330e-01 -8.14213216e-01 6.17688239e-01 3.49949181e-01
-7.04496354e-02 -8.39125097e-01 4.97789800e-01 7.04490185e-01
-1.25450385e+00 -2.25531533e-02 7.80254453e-02 6.17915571e-01
2.11998373e-01 -3.28120947e-01 6.74813032e-01 6.97678402e-02
-3.45376730e-02 1.20719085e-02 -3.15260619e-01 1.68638527e-01
9.13109183e-01 1.60960004e-01 -4.70387936e-01 -6.73210442e-01
-7.96575904e-01 2.82832921e-01 3.35058630e-01 1.82899594e-01
3.11125457e-01 -8.73379946e-01 -6.40875518e-01 8.81825984e-02
-2.31958702e-01 -1.29488975e-01 -6.73433347e-03 6.69053912e-01
1.04236796e-01 6.07378542e-01 -4.04525436e-02 -2.61468083e-01
-1.08085346e+00 4.73740935e-01 6.65159166e-01 -2.98299670e-01
-4.13854331e-01 3.26158851e-01 9.16272998e-01 -3.05750191e-01
-2.09256075e-02 -4.61365998e-01 -9.07479301e-02 -3.79941821e-01
2.15396732e-01 2.68707812e-01 -3.31362605e-01 -3.78095478e-01
-2.16416549e-02 2.82685626e-02 2.37454697e-01 -1.76038414e-01
1.65209281e+00 2.52050996e-01 2.48967540e-02 8.37224960e-01
7.75667727e-01 -4.65005189e-01 -1.20352113e+00 -8.27563331e-02
-4.78591248e-02 -2.91493565e-01 1.89781398e-01 -1.21480656e+00
-7.73972392e-01 9.36112463e-01 3.12086523e-01 8.63108709e-02
8.59325767e-01 5.53410091e-02 -2.84407493e-02 2.35662237e-01
-2.17759069e-02 -7.21595109e-01 -4.22446519e-01 5.19407820e-03
6.80334747e-01 -1.26149929e+00 6.98156804e-02 -3.60000879e-04
-5.86426079e-01 8.45974803e-01 7.71669671e-02 -1.86865479e-01
6.09674692e-01 6.01935163e-02 -2.45722756e-01 -2.00242430e-01
-1.25458717e+00 3.44177693e-01 4.95268971e-01 3.18509132e-01
4.75366235e-01 5.52078724e-01 4.61155288e-02 8.81025136e-01
-2.11134210e-01 4.97277588e-01 7.79624045e-01 3.02204311e-01
-7.31770992e-02 -9.42695796e-01 -4.37381417e-01 1.04976463e+00
-7.52210081e-01 -1.26777992e-01 -4.13043112e-01 8.81990969e-01
4.88807857e-02 7.79644728e-01 1.90037876e-01 -9.61600170e-02
3.92311424e-01 2.79719591e-01 1.24224303e-02 -7.26673543e-01
-4.52477008e-01 -4.65835214e-01 3.42960060e-02 -5.32565653e-01
-3.81283790e-01 -8.48699868e-01 -1.09831560e+00 -5.42689145e-01
-1.21968053e-01 -3.79203185e-02 6.18194401e-01 1.48401284e+00
3.37030023e-01 5.70407152e-01 4.15811479e-01 -3.38794112e-01
-8.34876657e-01 -7.62296438e-01 -4.60605264e-01 4.80223924e-01
6.14973128e-01 -7.08424926e-01 -1.69901744e-01 1.52176335e-01] | [8.397570610046387, 5.528372287750244] |
974d4556-c005-42a9-9115-5b499d847c6d | speaker-change-aware-crf-for-dialogue-act | 2004.02913 | null | https://arxiv.org/abs/2004.02913v3 | https://arxiv.org/pdf/2004.02913v3.pdf | Speaker-change Aware CRF for Dialogue Act Classification | Recent work in Dialogue Act (DA) classification approaches the task as a sequence labeling problem, using neural network models coupled with a Conditional Random Field (CRF) as the last layer. CRF models the conditional probability of the target DA label sequence given the input utterance sequence. However, the task involves another important input sequence, that of speakers, which is ignored by previous work. To address this limitation, this paper proposes a simple modification of the CRF layer that takes speaker-change into account. Experiments on the SwDA corpus show that our modified CRF layer outperforms the original one, with very wide margins for some DA labels. Further, visualizations demonstrate that our CRF layer can learn meaningful, sophisticated transition patterns between DA label pairs conditioned on speaker-change in an end-to-end way. Code is publicly available. | ['Jean-Pierre Lorré', 'Antoine Jean-Pierre Tixier', 'Michalis Vazirgiannis', 'Guokan Shang'] | 2020-04-06 | null | https://aclanthology.org/2020.coling-main.40 | https://aclanthology.org/2020.coling-main.40.pdf | coling-2020-8 | ['dialogue-act-classification'] | ['natural-language-processing'] | [ 2.62424082e-01 4.80690092e-01 -1.77615613e-01 -1.03842449e+00
-3.61957282e-01 -6.24517977e-01 9.61072803e-01 -1.74306780e-01
-4.65087891e-01 8.38165998e-01 5.73818564e-01 -3.71021330e-01
6.80154085e-01 -3.37293774e-01 -1.59691185e-01 -5.58418453e-01
1.40364897e-02 6.07951283e-01 2.49661848e-01 -3.40413243e-01
2.10098818e-01 1.29651934e-01 -1.03381670e+00 3.38527501e-01
6.38172626e-01 5.72882950e-01 2.22207680e-01 6.98263884e-01
-4.56763387e-01 1.10105371e+00 -9.26841915e-01 -2.01836839e-01
-5.45339137e-02 -5.57374835e-01 -1.40604639e+00 2.43235871e-01
4.21226174e-01 -3.37845773e-01 -6.27581356e-03 8.59625459e-01
2.73532361e-01 4.10917789e-01 7.51789570e-01 -1.02381325e+00
-4.10868108e-01 9.71220076e-01 -2.55450487e-01 -6.30040839e-02
6.78296149e-01 1.19493157e-01 1.02717185e+00 -7.42727518e-01
5.17180383e-01 1.50214827e+00 5.61289907e-01 8.56504858e-01
-1.46836114e+00 -5.22933841e-01 2.90300429e-01 -5.70142455e-02
-1.11727715e+00 -3.67127120e-01 8.10828507e-01 -6.22129321e-01
1.22672939e+00 2.72071790e-02 2.17589036e-01 1.28171790e+00
9.46133630e-04 9.99300718e-01 1.15077698e+00 -8.46428216e-01
2.78829515e-01 -5.67795932e-02 6.71786726e-01 6.72058761e-01
-8.70863020e-01 -1.35568008e-01 -4.00064468e-01 -2.10089892e-01
4.06142086e-01 -3.02087724e-01 -1.84736419e-02 9.22585651e-03
-1.03625143e+00 1.04310548e+00 -4.65205945e-02 4.94459569e-01
-3.75330970e-02 -2.04447448e-01 4.59940940e-01 3.69761825e-01
6.20692253e-01 2.03737020e-01 -5.57362795e-01 -4.48805273e-01
-7.40628004e-01 2.74967194e-01 1.17856455e+00 9.21013057e-01
7.87321806e-01 5.04190400e-02 -3.22146058e-01 9.83019948e-01
5.59436679e-01 2.17240125e-01 4.31434572e-01 -8.63714635e-01
3.87192994e-01 4.37579274e-01 1.57276168e-02 -4.73860472e-01
-5.75058699e-01 8.80773440e-02 -6.18240833e-01 2.18435377e-01
8.01194727e-01 -4.88443226e-01 -9.18143034e-01 1.98001659e+00
3.17789674e-01 1.11257032e-01 6.99123889e-02 6.94302142e-01
7.59191573e-01 7.55246818e-01 5.33851683e-01 -3.19510341e-01
1.30693388e+00 -1.23411727e+00 -1.21613991e+00 -2.46093929e-01
8.47522855e-01 -7.52739668e-01 1.09745049e+00 2.94171035e-01
-9.50120687e-01 -5.63253939e-01 -8.83076727e-01 -1.64825842e-01
-2.64748096e-01 1.83355615e-01 7.69278228e-01 8.04201841e-01
-1.09755051e+00 3.58488023e-01 -7.90111661e-01 -3.55375797e-01
-1.31990135e-01 3.13022286e-01 -1.47458136e-01 2.79179364e-01
-1.45705950e+00 1.20897019e+00 3.76615882e-01 -7.47473165e-02
-6.19373322e-01 -2.97385514e-01 -1.10603809e+00 3.67080569e-02
4.07648981e-01 -1.89852700e-01 2.00592399e+00 -1.18748713e+00
-2.42381907e+00 8.50416780e-01 -3.94609779e-01 -4.43246245e-01
4.45559531e-01 -2.09802955e-01 -4.79249626e-01 -5.58981448e-02
-2.65499115e-01 7.49682128e-01 6.24189913e-01 -1.05205977e+00
-7.08386660e-01 -2.05660492e-01 1.67317137e-01 2.09615290e-01
1.08976133e-01 3.69938225e-01 1.91323459e-02 -4.98299807e-01
-2.28124797e-01 -9.80002344e-01 -3.56587410e-01 -6.58733249e-01
-5.01757741e-01 -9.13181603e-01 7.02270627e-01 -7.66267657e-01
1.33960867e+00 -2.03620625e+00 8.16319883e-02 -8.91655218e-03
5.43948337e-02 1.86534032e-01 1.27719641e-01 6.26776040e-01
-2.58542120e-01 3.09666782e-03 -5.34178317e-01 -6.59955680e-01
1.70127600e-01 1.81247547e-01 -2.65614301e-01 3.38803887e-01
2.96337396e-01 4.97311443e-01 -8.01594317e-01 -3.31948817e-01
2.19440863e-01 2.07104430e-01 -3.23483914e-01 5.55726051e-01
-4.36384767e-01 7.41499126e-01 -1.48224279e-01 1.23684183e-01
6.52181566e-01 1.11732297e-02 5.63969672e-01 2.21511349e-01
-2.72474319e-01 8.37992609e-01 -9.33442116e-01 1.85483086e+00
-4.53243583e-01 6.21633232e-01 2.04894483e-01 -8.00874591e-01
1.24351835e+00 6.03195667e-01 2.73871481e-01 -2.21567065e-01
1.02666460e-01 -2.12563217e-01 1.87709183e-01 -4.22892302e-01
7.03754842e-01 -3.25324237e-01 -4.31646526e-01 7.52027571e-01
4.12462711e-01 -1.52184116e-02 1.70824677e-01 3.27208161e-01
8.32902849e-01 2.43597731e-01 5.51733136e-01 -2.40751296e-01
6.94938064e-01 -1.35375276e-01 5.56345940e-01 6.17586315e-01
-4.08429027e-01 3.41163784e-01 5.22285342e-01 -3.74386102e-01
-7.24430442e-01 -8.16346347e-01 -1.71816736e-01 1.45364463e+00
-1.93487957e-01 -3.46257865e-01 -9.60441411e-01 -1.11887288e+00
-3.76516789e-01 1.15992296e+00 -4.97891784e-01 1.35138378e-01
-8.63661885e-01 -4.41123277e-01 7.62173176e-01 2.87607849e-01
4.67668205e-01 -1.36991191e+00 -3.60498935e-01 3.75827640e-01
-2.76956499e-01 -1.06925499e+00 -5.54371834e-01 3.59149873e-01
-4.78857309e-01 -7.01120675e-01 -4.29528922e-01 -1.09678257e+00
3.63068908e-01 -2.74870515e-01 1.06720459e+00 -2.18403161e-01
2.56076783e-01 2.05835164e-01 -3.58137548e-01 -1.40947387e-01
-1.05794919e+00 1.33154213e-01 -3.52698527e-02 4.17444259e-02
7.20400572e-01 -3.31002384e-01 7.97093362e-02 3.33549559e-01
-6.27351999e-01 3.40450048e-01 9.91094112e-03 8.81530046e-01
-2.28656814e-01 -3.28499109e-01 7.68117249e-01 -1.17165112e+00
8.03170621e-01 -4.49565649e-01 -4.57210779e-01 2.89320350e-01
-3.43464255e-01 2.04390347e-01 5.36747277e-01 -3.25335294e-01
-1.71052086e+00 5.38288951e-01 -6.06985986e-01 1.99393034e-01
-9.08687472e-01 5.70729792e-01 -2.65575826e-01 7.22570360e-01
5.23043096e-01 1.15840726e-01 1.70927078e-01 -7.50972807e-01
4.38024998e-01 9.93486166e-01 5.60385287e-01 -4.97922301e-01
2.55056828e-01 -1.31020978e-01 -4.99635011e-01 -7.36899018e-01
-7.85816908e-01 -5.90079427e-01 -1.20195818e+00 -2.49629885e-01
1.05607784e+00 -8.23802650e-01 -8.52745056e-01 6.25222862e-01
-1.41284406e+00 -6.91040158e-01 -3.88561958e-03 5.52106500e-01
-5.94740450e-01 5.32560408e-01 -9.60915744e-01 -9.70532656e-01
-1.79404430e-02 -1.12606537e+00 6.47378504e-01 2.70057797e-01
-5.13384283e-01 -1.19314373e+00 3.67493242e-01 2.87102222e-01
2.12017193e-01 -1.19923472e-01 9.26646054e-01 -1.19549298e+00
1.07723594e-01 3.94000448e-02 2.25128680e-01 3.28706056e-01
3.95433903e-01 1.46737397e-01 -1.28546286e+00 -5.18088266e-02
1.87685013e-01 -4.63956505e-01 7.31388032e-01 2.19701678e-01
7.12963641e-01 -3.72845054e-01 -1.45507187e-01 -1.74028546e-01
7.46255100e-01 5.31289816e-01 5.77051580e-01 -1.28479406e-01
5.46717465e-01 9.11911845e-01 6.13636017e-01 4.12773401e-01
6.63718760e-01 9.17277753e-01 5.40394783e-02 -1.69329688e-01
-3.32109109e-02 -1.68225214e-01 7.21875966e-01 9.42690790e-01
2.10659727e-01 -3.88963103e-01 -1.06800699e+00 2.46554375e-01
-1.73018372e+00 -9.97559130e-01 -3.19553018e-01 2.02658939e+00
1.30499101e+00 1.03212669e-01 4.64825243e-01 -9.74636748e-02
9.18122292e-01 3.08164567e-01 -1.10914648e-01 -7.88927853e-01
4.03395668e-02 5.79938293e-02 -6.06912225e-02 1.05659485e+00
-1.30233800e+00 1.26734197e+00 7.08495665e+00 5.67899108e-01
-9.50198472e-01 1.38570368e-01 5.03651679e-01 4.55677032e-01
-3.27483118e-02 6.42479956e-02 -1.05674839e+00 4.50307429e-01
1.32219517e+00 1.48949385e-01 2.49392539e-01 8.47904325e-01
2.75374413e-01 -1.91792667e-01 -1.28124702e+00 4.66370553e-01
1.53921116e-02 -1.07324326e+00 -4.10880476e-01 -2.41814852e-02
3.38321358e-01 -1.12417661e-01 -3.93555045e-01 6.86103821e-01
1.10911489e+00 -9.46281791e-01 5.54206192e-01 3.24116677e-01
5.54559827e-01 -6.94109917e-01 7.49524295e-01 5.30504584e-01
-8.91587913e-01 2.03112304e-01 -9.96317044e-02 -2.80400604e-01
3.96758109e-01 3.53519171e-01 -1.66254282e+00 3.16305429e-01
2.76953518e-01 7.10408032e-01 -4.75957811e-01 5.78229666e-01
-4.66722161e-01 1.21072948e+00 -1.28428012e-01 -4.02909853e-02
3.55201304e-01 -1.61090016e-01 5.09499967e-01 1.86586964e+00
-3.40689510e-01 8.03450644e-02 7.53198326e-01 6.78214371e-01
5.95496334e-02 1.36705667e-01 -4.33384776e-01 1.13714322e-01
3.91075462e-01 1.13585663e+00 -5.48024893e-01 -3.95086348e-01
-4.47313130e-01 9.29281533e-01 5.09492099e-01 2.45466933e-01
-5.86721301e-01 -8.34679455e-02 4.54676092e-01 -4.83899474e-01
1.44910097e-01 -3.55670273e-01 -2.79342234e-01 -1.00028098e+00
-3.98296267e-01 -8.95194888e-01 4.71464217e-01 -2.28040606e-01
-1.43886161e+00 7.18858123e-01 1.83343198e-02 -1.07357097e+00
-8.39986503e-01 -3.97831798e-01 -7.10994184e-01 9.79214907e-01
-1.14819360e+00 -1.07737243e+00 1.62946194e-01 5.79999447e-01
9.80625451e-01 -1.55583218e-01 1.26854479e+00 5.94425611e-02
-6.11655474e-01 5.61059833e-01 -1.66836157e-01 5.03858864e-01
8.92816663e-01 -1.70114470e+00 5.16708732e-01 6.49231791e-01
-3.73326777e-03 4.67757285e-01 7.71960616e-01 -7.57672071e-01
-4.29638922e-01 -6.17178619e-01 1.40940070e+00 -3.61443639e-01
5.21983802e-01 -6.66877151e-01 -1.16511106e+00 9.28071201e-01
8.23059797e-01 -4.16008055e-01 1.06192505e+00 5.36769986e-01
-3.41356844e-01 5.99115789e-01 -1.09782517e+00 5.27449369e-01
7.12705493e-01 -5.36685467e-01 -9.57767665e-01 2.87597865e-01
8.73178422e-01 -3.79168123e-01 -8.08696866e-01 5.08769900e-02
2.93419540e-01 -8.15166593e-01 5.17449379e-01 -7.25072980e-01
2.55312979e-01 -2.36872301e-01 3.63664213e-03 -1.43816543e+00
-2.67631471e-01 -7.04127312e-01 1.18221156e-01 1.48658788e+00
5.59941053e-01 -3.11513752e-01 4.56667364e-01 9.01530385e-01
-4.27639484e-01 -9.92006212e-02 -9.24959719e-01 -6.25348926e-01
2.50672698e-01 -4.56590980e-01 5.18395782e-01 1.27726781e+00
6.14037395e-01 1.01416445e+00 -7.16266632e-01 -1.70306891e-01
-1.01309726e-02 -1.08747125e-01 7.51121402e-01 -1.28411746e+00
-3.63344043e-01 -2.78298020e-01 -1.94725879e-02 -1.54302406e+00
5.43669641e-01 -8.49749148e-01 5.50988615e-01 -1.26658058e+00
-2.50717729e-01 -3.79437894e-01 -8.24908819e-03 7.68148303e-01
-3.50262672e-01 -2.87329853e-01 2.81410903e-01 1.58329591e-01
-5.90301156e-01 6.97368324e-01 8.80243003e-01 -1.97362438e-01
-5.07885456e-01 3.94918680e-01 -6.71468675e-02 8.73859406e-01
8.62362385e-01 -5.47610939e-01 -1.22124940e-01 1.32377688e-02
-5.31977832e-01 4.00824517e-01 -2.31377855e-01 -6.74620867e-01
1.28180996e-01 -3.31622005e-01 6.07177615e-02 -7.71713793e-01
3.28769177e-01 -7.07890272e-01 -2.42250398e-01 3.59496266e-01
-9.95286882e-01 -1.67168245e-01 -2.47928873e-02 3.80761772e-01
-4.05496657e-01 -5.13145208e-01 6.98362112e-01 -7.06623250e-04
-6.24854803e-01 -2.15050668e-01 -1.16433942e+00 -2.23098308e-01
8.00706923e-01 9.62074474e-02 -2.83038735e-01 -6.18236661e-01
-9.96118963e-01 2.58020997e-01 8.49143509e-03 4.95519400e-01
7.58188292e-02 -9.80368257e-01 -8.21612477e-01 1.15638234e-01
-2.90814266e-02 -7.24778995e-02 1.38386309e-01 5.57767987e-01
-2.18698248e-01 4.50297803e-01 -1.77267194e-01 -6.91013813e-01
-1.36523974e+00 3.35175365e-01 3.04048806e-01 -6.59108281e-01
-4.87740904e-01 7.00018346e-01 1.33651271e-01 -1.07692587e+00
5.30188859e-01 -2.55760044e-01 -7.52290666e-01 3.17261592e-02
6.98038518e-01 1.02850504e-01 -1.73227802e-01 -9.23716843e-01
-2.59306937e-01 -1.57480985e-01 -4.62824434e-01 -3.72883946e-01
1.08003163e+00 -4.88169283e-01 9.14944708e-02 9.05995786e-01
9.64017749e-01 1.58347681e-01 -1.55188429e+00 -6.51044548e-01
5.32305837e-01 -1.28659964e-01 -3.27817291e-01 -1.10220373e+00
-5.40610194e-01 8.78260374e-01 6.09560847e-01 4.05480087e-01
7.35851407e-01 -1.24084748e-01 4.86602783e-01 5.27298331e-01
1.30078763e-01 -1.16934216e+00 -7.04298355e-03 1.20999217e+00
8.16821575e-01 -1.30412626e+00 -2.92723298e-01 -5.24516284e-01
-1.17571068e+00 1.28065181e+00 8.19088280e-01 1.58211261e-01
7.34801054e-01 2.77849942e-01 6.09574199e-01 4.01445925e-02
-7.85959005e-01 -5.80321886e-02 1.63685181e-03 5.53727567e-01
7.59797275e-01 2.84548104e-01 -3.89051348e-01 4.95960802e-01
-3.44891638e-01 -2.41448238e-01 6.38644814e-01 9.22411263e-01
-4.43983018e-01 -1.52706778e+00 -8.82163271e-02 1.53138582e-02
-5.50654292e-01 -1.76558688e-01 -6.01239860e-01 5.56190372e-01
-1.55499324e-01 1.26532936e+00 3.54199439e-01 -4.57969099e-01
7.32658282e-02 8.22188795e-01 1.00502083e-02 -9.73316669e-01
-1.15665162e+00 1.85720205e-01 5.17350733e-01 -1.84273839e-01
-7.00022876e-01 -8.02724898e-01 -1.64786422e+00 -5.42845130e-02
-4.07181442e-01 2.17363477e-01 6.24506891e-01 1.37884748e+00
-1.28379976e-02 1.75306380e-01 8.69417727e-01 -6.21142387e-01
-5.63495517e-01 -1.55231798e+00 -5.23481250e-01 4.29767251e-01
4.05179292e-01 -4.92045671e-01 -2.12679118e-01 4.59381789e-01] | [12.789055824279785, 7.692222595214844] |
f058295b-7fa8-422c-b835-8fa3b0938f1a | xmi-icu-explainable-machine-learning-model | 2305.06109 | null | https://arxiv.org/abs/2305.06109v1 | https://arxiv.org/pdf/2305.06109v1.pdf | XMI-ICU: Explainable Machine Learning Model for Pseudo-Dynamic Prediction of Mortality in the ICU for Heart Attack Patients | Heart attack remain one of the greatest contributors to mortality in the United States and globally. Patients admitted to the intensive care unit (ICU) with diagnosed heart attack (myocardial infarction or MI) are at higher risk of death. In this study, we use two retrospective cohorts extracted from the eICU and MIMIC-IV databases, to develop a novel pseudo-dynamic machine learning framework for mortality prediction in the ICU with interpretability and clinical risk analysis. The method provides accurate prediction for ICU patients up to 24 hours before the event and provide time-resolved interpretability results. The performance of the framework relying on extreme gradient boosting was evaluated on a held-out test set from eICU, and externally validated on the MIMIC-IV cohort using the most important features identified by time-resolved Shapley values achieving AUCs of 91.0 (balanced accuracy of 82.3) for 6-hour prediction of mortality respectively. We show that our framework successfully leverages time-series physiological measurements by translating them into stacked static prediction problems to be robustly predictive through time in the ICU stay and can offer clinical insight from time-resolved interpretability | ['Tingting Zhu', 'Peter Watkinson', 'Munib Mesinovic'] | 2023-05-10 | null | null | null | null | ['mortality-prediction'] | ['medical'] | [ 1.18302651e-01 -2.64934838e-01 6.21301346e-02 -3.06227267e-01
-7.93523848e-01 -4.37366843e-01 -1.76940411e-01 6.40402198e-01
-2.77962983e-01 7.60374248e-01 2.73457617e-01 -7.93121040e-01
-8.41574550e-01 -3.37391615e-01 -7.58876093e-03 -5.63103855e-01
-8.45495582e-01 7.85829484e-01 -4.15503591e-01 3.91094834e-02
-3.07720918e-02 5.84481418e-01 -9.56604421e-01 5.99208534e-01
7.89835691e-01 1.24181664e+00 -4.76950228e-01 9.76236403e-01
5.78867376e-01 9.61208940e-01 -2.41311342e-01 -1.38226999e-02
4.97070283e-01 -5.79804301e-01 -5.68901420e-01 -4.53995854e-01
6.26152847e-03 -4.86379236e-01 -2.52393074e-02 -1.57326341e-01
8.78113151e-01 -2.25957274e-01 6.62613034e-01 -1.06906343e+00
3.01764399e-01 4.48634923e-01 -4.60148491e-02 6.13977909e-01
1.04578706e-02 3.52357119e-01 6.79818392e-01 -7.19165683e-01
2.88233664e-02 7.17398286e-01 1.34278142e+00 4.59793031e-01
-1.31851184e+00 -3.03241491e-01 -2.49354839e-01 2.36406147e-01
-1.01434159e+00 -2.94111997e-01 4.84230548e-01 -5.37618101e-01
9.76847529e-01 5.39322257e-01 9.05750334e-01 9.73536313e-01
7.35899568e-01 -1.60017032e-02 1.15678549e+00 -1.89117327e-01
3.23887438e-01 -3.42712134e-01 5.55411220e-01 4.90709901e-01
3.15913916e-01 3.58174086e-01 -4.75095987e-01 -7.72846162e-01
4.14972693e-01 7.80953646e-01 -4.21194553e-01 -8.49820003e-02
-1.56294429e+00 4.95567858e-01 3.45795825e-02 -2.50609010e-01
-7.79340446e-01 -1.46989942e-01 7.70018101e-01 4.20910209e-01
3.49580675e-01 4.24080551e-01 -1.35418999e+00 -5.08245468e-01
-7.30617046e-01 -1.80577189e-01 7.96575606e-01 1.74925461e-01
-1.11187445e-02 -1.36557743e-01 -2.71343470e-01 5.36471367e-01
1.49671406e-01 5.14299452e-01 8.76257241e-01 -1.00167847e+00
3.19838881e-01 6.86754227e-01 2.07066089e-01 -6.20283186e-01
-1.05889165e+00 -7.89229989e-01 -1.06039643e+00 1.29308239e-01
3.34250540e-01 -4.21819597e-01 -4.62001354e-01 1.61338639e+00
7.36540109e-02 4.32128847e-01 3.48156005e-01 6.59484744e-01
5.21565735e-01 -4.74415049e-02 4.29130614e-01 -7.22858548e-01
1.40528202e+00 -5.44392824e-01 -3.57996315e-01 1.99709069e-02
9.97116208e-01 6.49276003e-02 7.38773048e-01 6.01267695e-01
-9.82846797e-01 -2.33703762e-01 -7.34888256e-01 4.05084938e-01
1.92945987e-01 8.55358243e-02 3.30972433e-01 4.44581836e-01
-8.33735228e-01 1.01719081e+00 -1.17823088e+00 -3.97474855e-01
4.15361077e-01 3.64902854e-01 -3.19491684e-01 2.90044844e-01
-1.25744712e+00 9.72857833e-01 1.79723382e-01 1.12543426e-01
-6.74802959e-01 -1.28610933e+00 -4.39264268e-01 1.09263510e-01
-3.38874012e-01 -1.52889800e+00 7.63895512e-01 -8.26682448e-01
-1.08132994e+00 7.99079835e-01 -1.79277435e-01 -8.49932134e-01
7.79393375e-01 -5.61611235e-01 -1.72259361e-01 5.05183280e-01
-1.73376992e-01 -1.11494422e-01 4.49996740e-01 -7.61404753e-01
-3.73133928e-01 -8.04207444e-01 -4.40672755e-01 1.16882004e-01
-1.64699793e-01 -4.09147628e-02 6.92468703e-01 -5.63852012e-01
1.33060619e-01 -9.77888346e-01 -2.86147803e-01 -4.36255693e-01
-2.63942406e-02 4.37808007e-01 6.82697237e-01 -9.64154184e-01
1.60447061e+00 -1.89471984e+00 1.39110722e-02 -7.23922849e-02
6.86320305e-01 8.63459855e-02 4.25305277e-01 4.07734424e-01
-6.18391454e-01 2.63773918e-01 -3.53787541e-01 -1.63660675e-01
-6.27648234e-01 9.97304171e-02 -4.60552245e-01 4.57628965e-01
1.37120977e-01 1.08348060e+00 -8.18325281e-01 -3.06770861e-01
4.15916502e-01 2.35631213e-01 -6.31531179e-01 7.00782418e-01
5.62793553e-01 8.81796658e-01 -2.94848680e-01 3.91115308e-01
8.83795172e-02 -3.79364610e-01 4.19563502e-01 -1.22306108e-01
1.94722906e-01 -2.20424235e-01 -4.39511389e-01 1.12721670e+00
-2.64332145e-01 2.10159466e-01 -4.62492406e-01 -1.09799993e+00
8.88076127e-01 7.43035793e-01 1.05658305e+00 -2.25092486e-01
2.01530173e-01 6.13301210e-02 1.75933987e-01 -7.68476367e-01
-7.85037339e-01 -7.32484519e-01 -3.20420112e-03 3.62932682e-01
-3.28042179e-01 3.47642511e-01 -3.58396471e-01 -1.53573573e-01
1.50436378e+00 3.04391957e-04 8.23320687e-01 -6.17890120e-01
5.61927378e-01 -1.10099274e-04 8.25062990e-01 8.63666773e-01
-6.59915090e-01 9.02139664e-01 4.88636971e-01 -1.15406656e+00
-8.49984109e-01 -1.30215919e+00 -4.90755945e-01 5.44623911e-01
-4.92429554e-01 -1.90008417e-01 -3.59007955e-01 -5.77325583e-01
2.02039436e-01 6.15924239e-01 -7.72074461e-01 -6.20919943e-01
-6.74581230e-01 -1.36209619e+00 5.66492915e-01 7.43223369e-01
7.79726058e-02 -8.23657334e-01 -1.40315473e+00 6.79505587e-01
-4.06351656e-01 -6.96144283e-01 1.46902263e-01 4.35071141e-01
-1.57209790e+00 -1.42428589e+00 -5.91510117e-01 -2.17556670e-01
1.24041282e-01 -2.91833758e-01 1.44649994e+00 3.01103771e-01
-7.25894868e-01 4.86554176e-01 -9.13569927e-02 -6.79412544e-01
-3.70025724e-01 -3.11890423e-01 5.27909398e-01 1.28504455e-01
1.82112530e-01 -9.73655343e-01 -1.27580690e+00 3.00845742e-01
-4.67529237e-01 1.58787653e-01 3.54418933e-01 1.17245555e+00
5.16390383e-01 -7.15269387e-01 1.42139351e+00 -7.02541411e-01
2.88715869e-01 -9.73522902e-01 1.49110041e-03 6.25009462e-02
-1.37728810e+00 -3.38206142e-01 8.52611780e-01 -1.64919510e-01
-5.45243740e-01 -1.82599291e-01 3.35591346e-01 -4.88823384e-01
-1.92893848e-01 5.29467404e-01 2.95366883e-01 5.72957337e-01
7.21558154e-01 3.63068022e-02 1.72781751e-01 -3.66799951e-01
-4.47855920e-01 5.78046143e-01 4.06079859e-01 -5.20422697e-01
2.22941130e-01 4.49792236e-01 7.32639909e-01 -2.58136243e-01
-9.32379305e-01 -6.43095255e-01 -1.08751464e+00 -1.62644684e-01
7.56003380e-01 -1.15684652e+00 -8.72220874e-01 3.98676157e-01
-5.70050776e-01 -5.38789034e-01 -3.01066935e-01 6.15904689e-01
-7.53777862e-01 4.71119076e-01 -5.18556535e-01 -7.65862107e-01
-1.02477503e+00 -6.75244749e-01 8.23127925e-01 -2.82431334e-01
-7.03591526e-01 -1.40051031e+00 5.04071057e-01 2.03668416e-01
4.26716387e-01 1.14668822e+00 1.28511119e+00 -1.06499970e+00
2.08627045e-01 -3.94939721e-01 5.00962064e-02 2.97059655e-01
3.06130707e-01 -2.95659918e-02 -7.84444332e-01 -2.55736351e-01
2.60669172e-01 -5.66928759e-02 5.28016090e-01 6.57447934e-01
1.13236678e+00 -2.89678663e-01 -2.18856514e-01 8.61861706e-01
1.45219648e+00 2.78150648e-01 5.05149066e-01 4.11175460e-01
5.25612414e-01 2.98588723e-01 2.68090785e-01 9.58099067e-01
4.44028676e-01 4.13567811e-01 2.62608320e-01 -8.68040919e-02
2.43747428e-01 3.42990428e-01 1.74903467e-01 7.63498247e-01
-3.96276772e-01 5.65595739e-02 -1.44378424e+00 3.45595241e-01
-1.99732637e+00 -9.58581567e-01 -5.81980467e-01 2.40231252e+00
6.69953048e-01 -2.53156066e-01 6.21243864e-02 2.12087035e-01
3.46086562e-01 -5.15604854e-01 -6.04745686e-01 -5.82456768e-01
-1.66129433e-02 2.21698493e-01 6.21866547e-02 2.15102673e-01
-9.27240431e-01 1.74347926e-02 7.06079102e+00 -4.24612075e-01
-1.07771659e+00 1.02573209e-01 1.45528448e+00 -4.01719660e-01
5.49302638e-01 -7.90937990e-02 -7.08051678e-03 4.45115685e-01
1.89210248e+00 -2.88705647e-01 2.10083872e-01 7.01773584e-01
8.25866401e-01 2.95155585e-01 -1.40789723e+00 9.23852563e-01
-2.53740281e-01 -1.04649055e+00 -2.92538106e-01 -2.80681312e-01
4.16790664e-01 5.57218418e-02 -2.30728418e-01 2.81113654e-01
-2.32797846e-01 -1.11839676e+00 3.22840452e-01 9.71889973e-01
1.16633105e+00 -5.44260144e-01 1.03726435e+00 1.88378185e-01
-7.33879924e-01 -5.59404910e-01 3.10924619e-01 -4.08610195e-01
1.10249221e-01 7.66979933e-01 -9.93775785e-01 4.81359124e-01
8.11589718e-01 8.67660522e-01 -5.12363195e-01 8.52733493e-01
3.73937815e-01 1.06654036e+00 -2.85212904e-01 6.46536648e-01
-3.91941369e-01 2.67399624e-02 6.29477143e-01 9.92163002e-01
4.54965293e-01 6.38850331e-01 1.09493896e-01 3.50341141e-01
2.35335141e-01 1.94777817e-01 -4.77360964e-01 5.35228372e-01
-4.86933589e-02 1.14944160e+00 -3.33917260e-01 -5.57003140e-01
-1.81581870e-01 6.89378142e-01 1.06426433e-01 2.50490755e-01
-9.01104033e-01 -4.64146538e-03 5.75779736e-01 2.99301565e-01
-3.00340563e-01 1.92629755e-01 -1.01388943e+00 -1.45606172e+00
-3.06313097e-01 -8.78891349e-01 8.68303776e-01 -6.53141797e-01
-1.29918301e+00 6.34893656e-01 -2.21428834e-02 -1.46078098e+00
-4.73988086e-01 -6.49003565e-01 -8.72397661e-01 1.14584148e+00
-1.39974761e+00 -5.40402710e-01 -6.26603365e-01 4.09506530e-01
2.72405446e-01 -1.96477249e-02 1.40631831e+00 4.34051082e-02
-8.15407753e-01 2.53177106e-01 1.25643313e-01 -8.93352255e-02
7.18902051e-01 -1.26374745e+00 -2.08694994e-01 3.23224336e-01
-8.43568265e-01 7.19141066e-01 4.95767564e-01 -5.98444223e-01
-1.21022069e+00 -1.23245728e+00 6.28552318e-01 -1.05226028e+00
5.85132897e-01 3.86523873e-01 -8.66416276e-01 5.99586666e-01
-3.08675528e-01 3.51185113e-01 1.31402826e+00 -1.03133786e-02
4.95394468e-02 -1.58322170e-01 -1.21558022e+00 2.28332043e-01
8.66033733e-01 -1.48724899e-01 -9.33324993e-01 2.46302888e-01
1.39681712e-01 -2.12131128e-01 -1.59453702e+00 1.06725037e+00
9.39904630e-01 -1.20486319e+00 1.14972305e+00 -1.27594221e+00
5.33520699e-01 1.39960393e-01 1.61598921e-01 -1.09857643e+00
-4.58584309e-01 -7.16194510e-01 -3.96808028e-01 5.51331580e-01
2.14582473e-01 -1.01747119e+00 3.24393690e-01 9.03218448e-01
-1.47465795e-01 -1.25654125e+00 -1.00497413e+00 -3.72394383e-01
2.27234632e-01 -2.02089980e-01 3.51002425e-01 7.82709360e-01
3.15364808e-01 1.94630340e-01 -2.36571521e-01 7.54451230e-02
6.48879647e-01 2.63315797e-01 3.43252212e-01 -1.68194497e+00
-3.43866676e-01 -2.45858967e-01 -3.26131344e-01 2.71697223e-01
-1.62487507e-01 -8.84764135e-01 -3.93254787e-01 -1.45314229e+00
6.19101286e-01 -4.93703932e-01 -1.03993714e+00 5.26614785e-01
-7.09568918e-01 4.13791090e-02 -1.53753266e-01 5.71782231e-01
-1.51311457e-01 2.95001894e-01 5.50957739e-01 4.50555384e-01
-3.66245449e-01 7.63098821e-02 -4.85608250e-01 6.18993282e-01
9.78928149e-01 -4.43691760e-01 -3.63243014e-01 1.21695898e-01
-6.10108115e-02 8.24953139e-01 6.31610930e-01 -1.06377435e+00
-6.42308772e-01 -2.72970259e-01 8.70760739e-01 -1.80946857e-01
6.32217014e-03 -9.86275196e-01 5.38785398e-01 1.05553174e+00
-2.96181083e-01 6.13792241e-01 4.44951832e-01 4.93577838e-01
1.03136376e-01 3.91125232e-01 7.37181962e-01 -2.54907347e-02
-3.46879736e-02 4.78428304e-01 -3.05614650e-01 4.16285366e-01
1.14771080e+00 6.53475150e-03 -1.51700690e-01 -1.83602914e-01
-1.05309105e+00 2.40581445e-02 2.43693590e-01 1.53111175e-01
5.15175581e-01 -1.01742280e+00 -1.06987441e+00 1.58264533e-01
2.16166571e-01 -4.78475928e-01 3.43296230e-01 1.58494902e+00
-8.02608609e-01 3.83334309e-01 -1.51842341e-01 -6.94225729e-01
-1.21252406e+00 6.51851177e-01 7.39599228e-01 -4.95064467e-01
-1.07544625e+00 2.67786711e-01 3.07397127e-01 6.34953305e-02
-1.02320746e-01 -3.59599382e-01 -1.38115644e-01 -6.25723898e-02
5.88905334e-01 5.60795963e-01 2.78716773e-01 -3.05718809e-01
-6.39013529e-01 3.41629475e-01 4.67483819e-01 2.89013892e-01
1.42781234e+00 -2.43200988e-01 -2.40188733e-01 7.06657588e-01
1.24694729e+00 -4.05885935e-01 -9.97717738e-01 3.38648021e-01
-9.00446177e-02 5.88503806e-03 -1.05482027e-01 -1.24738276e+00
-7.32261062e-01 9.76334274e-01 1.17243350e+00 -1.86218452e-02
1.30384350e+00 -4.93966162e-01 7.80037284e-01 3.52063328e-01
1.43688709e-01 -4.25881803e-01 -1.96271911e-01 1.44311398e-01
7.08179116e-01 -1.20777202e+00 -4.49961014e-02 1.21515885e-01
-6.60050273e-01 1.47749007e+00 6.53778315e-02 -8.31182525e-02
9.02952731e-01 9.46796965e-03 5.35056949e-01 -4.70052799e-03
-1.19319773e+00 6.56833172e-01 8.21405426e-02 4.02581215e-01
4.36481208e-01 3.62101525e-01 -3.72152746e-01 1.07396770e+00
1.53029919e-01 3.25476259e-01 3.73359025e-01 7.70517528e-01
-1.69448316e-01 -7.01138258e-01 -2.18385741e-01 9.83105779e-01
-9.20267165e-01 -3.87767881e-01 -5.12611605e-02 2.63497114e-01
-1.87466919e-01 9.78568316e-01 -1.86279401e-01 -2.32315689e-01
2.66864806e-01 8.98581803e-01 1.75594762e-01 -2.78677285e-01
-9.53762114e-01 -3.73205811e-01 1.51819184e-01 -5.01065195e-01
-1.66440383e-01 -8.41986001e-01 -1.36928618e+00 8.75569731e-02
2.16044798e-01 1.65378645e-01 2.02817589e-01 8.01761866e-01
9.37811852e-01 8.62342179e-01 9.82594490e-01 -5.47300935e-01
-7.37785280e-01 -1.04209018e+00 -3.53146970e-01 6.98978782e-01
7.10476756e-01 -4.07060266e-01 -8.45237017e-01 4.16362256e-01] | [8.003293991088867, 6.129432201385498] |
f695b620-92cb-4b07-b5e1-689ecde8ab6d | distill-knowledge-from-nrsfm-for-weakly | 1908.06377 | null | https://arxiv.org/abs/1908.06377v1 | https://arxiv.org/pdf/1908.06377v1.pdf | Distill Knowledge from NRSfM for Weakly Supervised 3D Pose Learning | We propose to learn a 3D pose estimator by distilling knowledge from Non-Rigid Structure from Motion (NRSfM). Our method uses solely 2D landmark annotations. No 3D data, multi-view/temporal footage, or object specific prior is required. This alleviates the data bottleneck, which is one of the major concern for supervised methods. The challenge for using NRSfM as teacher is that they often make poor depth reconstruction when the 2D projections have strong ambiguity. Directly using those wrong depth as hard target would negatively impact the student. Instead, we propose a novel loss that ties depth prediction to the cost function used in NRSfM. This gives the student pose estimator freedom to reduce depth error by associating with image features. Validated on H3.6M dataset, our learned 3D pose estimation network achieves more accurate reconstruction compared to NRSfM methods. It also outperforms other weakly supervised methods, in spite of using significantly less supervision. | ['Chen Kong', 'Chaoyang Wang', 'Simon Lucey'] | 2019-08-18 | distill-knowledge-from-nrsfm-for-weakly-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Wang_Distill_Knowledge_From_NRSfM_for_Weakly_Supervised_3D_Pose_Learning_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Wang_Distill_Knowledge_From_NRSfM_for_Weakly_Supervised_3D_Pose_Learning_ICCV_2019_paper.pdf | iccv-2019-10 | ['weakly-supervised-3d-human-pose-estimation'] | ['computer-vision'] | [-1.14039212e-01 4.72996324e-01 -6.20916843e-01 -5.90848327e-01
-1.04448426e+00 -6.57268703e-01 2.72167414e-01 -1.71130687e-01
-5.02884209e-01 6.59887195e-01 2.45565116e-01 -3.95861864e-02
5.09159081e-03 -5.23497999e-01 -1.08081222e+00 -6.38050735e-01
3.62785310e-01 7.74925351e-01 3.24912220e-01 1.24024965e-01
2.40791142e-01 4.55713540e-01 -1.19600189e+00 -1.72488376e-01
6.89866364e-01 9.29348886e-01 3.81485492e-01 3.24815363e-01
9.03361291e-02 4.83431369e-01 -3.06656063e-01 -1.98068455e-01
4.06904548e-01 -5.63107021e-02 -7.36970365e-01 2.86707699e-01
1.15443492e+00 -7.78285921e-01 -8.62420723e-02 8.41178477e-01
6.02729976e-01 1.61259547e-01 8.22715104e-01 -1.01113546e+00
-4.27161604e-02 1.58428311e-01 -9.97565746e-01 -1.21209122e-01
4.09237742e-01 -2.45836228e-01 8.40716362e-01 -1.17383444e+00
8.53317201e-01 1.45877576e+00 9.55176473e-01 7.52827466e-01
-1.28582978e+00 -5.46358168e-01 4.09207344e-01 -2.14583069e-01
-1.21367693e+00 -3.58870834e-01 6.95872366e-01 -4.93314922e-01
6.27864122e-01 -7.92763904e-02 6.19702101e-01 1.31608367e+00
-3.99047546e-02 1.09192979e+00 9.60750699e-01 -1.73825994e-01
1.36796340e-01 6.64777756e-02 -6.81767985e-02 1.01455367e+00
2.62074083e-01 -2.05400195e-02 -8.47337723e-01 5.34771755e-02
1.19822395e+00 3.68337929e-02 -2.60909885e-01 -1.21241903e+00
-9.92420495e-01 6.53686881e-01 4.70130920e-01 -3.47562373e-01
1.83441669e-01 3.42613101e-01 4.13591303e-02 1.36276409e-01
5.27392566e-01 2.71913081e-01 -7.68026292e-01 -1.81890503e-01
-8.60128701e-01 1.87904403e-01 5.74764371e-01 1.09309232e+00
7.71983504e-01 -1.71103805e-01 4.32163835e-01 5.40588439e-01
5.77164471e-01 5.86266875e-01 2.51204759e-01 -1.34963715e+00
6.64403856e-01 5.47214568e-01 9.92609784e-02 -7.29321778e-01
-5.13223112e-01 -5.05798936e-01 -5.18792629e-01 4.46701288e-01
6.35221064e-01 9.75855142e-02 -1.09827852e+00 1.74714410e+00
6.74202144e-01 1.85211718e-01 -1.10465378e-01 1.16916990e+00
8.40785563e-01 2.48334929e-01 -5.21906435e-01 1.21305138e-01
9.57412660e-01 -1.13540614e+00 -4.01753157e-01 -4.35078204e-01
7.03048885e-01 -5.96196175e-01 1.02434528e+00 6.05852604e-01
-1.15459657e+00 -4.06324774e-01 -1.02259994e+00 -4.43883389e-01
1.61478937e-01 1.11038536e-01 7.04383194e-01 3.42324197e-01
-8.19969237e-01 6.87509000e-01 -1.18893123e+00 -2.57869542e-01
4.42599028e-01 5.06495178e-01 -6.19482100e-01 -9.41162333e-02
-6.71081305e-01 1.03548729e+00 2.50129886e-02 -1.07980661e-01
-1.03741062e+00 -9.07603443e-01 -1.22180080e+00 -3.57995600e-01
5.06748080e-01 -7.97534227e-01 1.40455997e+00 -5.46905041e-01
-1.59296274e+00 1.04234767e+00 -1.82673767e-01 -1.56880572e-01
9.03663754e-01 -7.01042891e-01 4.90939885e-01 2.31408015e-01
2.84382731e-01 1.11411989e+00 9.09695566e-01 -1.35527205e+00
-3.41474921e-01 -5.78476131e-01 -3.76536697e-02 6.58668578e-01
-1.87639967e-01 -7.01475441e-01 -6.75672412e-01 -4.16336983e-01
8.24219525e-01 -1.01416171e+00 -3.09365332e-01 4.65003341e-01
-3.41673255e-01 -7.34023377e-02 9.14347053e-01 -4.16972429e-01
6.17875695e-01 -2.02723145e+00 3.39477926e-01 2.05323502e-01
3.24679941e-01 -1.82903543e-01 -1.28178224e-02 1.59870256e-02
1.25536233e-01 -1.15410142e-01 -9.46282148e-02 -6.75152719e-01
-1.56306013e-01 4.35061574e-01 -1.53355047e-01 6.44442260e-01
7.61827603e-02 6.87502623e-01 -1.01775444e+00 -5.44511199e-01
2.16498628e-01 5.13375759e-01 -9.09958661e-01 2.14064389e-01
-1.89902797e-01 8.56952488e-01 -4.60437536e-01 6.11239731e-01
7.12613285e-01 -5.28239787e-01 9.71332714e-02 -2.35668659e-01
5.70202470e-02 5.00968635e-01 -1.26838601e+00 2.40294647e+00
-4.94566560e-01 3.65407646e-01 1.50481731e-01 -9.26607728e-01
9.09945905e-01 2.00749114e-01 5.53562760e-01 -4.12542284e-01
-1.47960231e-01 3.03122878e-01 -4.98199940e-01 -3.79265785e-01
8.53047073e-02 -6.66576624e-02 1.34467989e-01 2.14773655e-01
2.36703023e-01 -4.62991506e-01 -3.86614174e-01 1.12964861e-01
9.64602649e-01 8.41418087e-01 1.65611833e-01 -1.84595153e-01
2.06549570e-01 -1.83345705e-01 9.82244134e-01 4.35882866e-01
-1.14878252e-01 9.57934618e-01 3.63549590e-01 -4.78010118e-01
-8.78311932e-01 -1.31485462e+00 -2.97299832e-01 7.84239352e-01
3.04967463e-01 -3.66153479e-01 -4.40288842e-01 -1.09089029e+00
3.37716579e-01 6.17701337e-02 -4.45131809e-01 -4.35068831e-02
-7.42508173e-01 -3.37371200e-01 2.34828576e-01 8.27173591e-01
2.94750512e-01 -2.88613230e-01 -5.37675679e-01 -1.28641695e-01
-2.30618313e-01 -1.18768585e+00 -6.89256132e-01 5.59127212e-01
-1.40881443e+00 -9.60309207e-01 -7.86151111e-01 -7.94503212e-01
1.12866211e+00 3.39916706e-01 1.15096879e+00 -1.72293991e-01
1.89339910e-02 4.77603585e-01 4.22758497e-02 -2.26830259e-01
8.85079876e-02 2.89889932e-01 3.02627921e-01 -5.36834180e-01
9.60650742e-02 -8.48356426e-01 -7.14837492e-01 5.70425570e-01
-3.30576628e-01 1.75046965e-01 5.84215879e-01 9.36994135e-01
8.47651362e-01 -3.58599037e-01 3.46607357e-01 -9.24872994e-01
-2.42236435e-01 1.49720199e-02 -8.28871906e-01 -1.43068001e-01
-7.60694921e-01 2.90544122e-01 2.52627462e-01 -4.92355049e-01
-1.01610219e+00 6.47561133e-01 -8.59119371e-02 -6.76828623e-01
1.11680841e-02 1.65413722e-01 -2.57561922e-01 -2.56161600e-01
6.40868127e-01 -1.74214661e-01 2.17440769e-01 -8.51997077e-01
-9.79800448e-02 1.06812418e-01 5.15330017e-01 -8.28112245e-01
9.85675812e-01 6.15153134e-01 4.30935383e-01 -5.85229576e-01
-1.53909659e+00 -5.97804368e-01 -1.05990064e+00 1.52318310e-02
8.26789796e-01 -1.37749147e+00 -7.89001524e-01 1.70253903e-01
-1.07127380e+00 -4.24520701e-01 -2.08308339e-01 7.92619407e-01
-7.46996284e-01 4.87318367e-01 -6.34508729e-01 -6.58681929e-01
1.12428786e-02 -1.26682150e+00 1.38258111e+00 -4.52651381e-02
-3.65168840e-01 -9.44927752e-01 4.31445055e-03 7.23410368e-01
-5.57553433e-02 2.44211704e-01 4.67036456e-01 -2.22779438e-01
-8.96504998e-01 5.27396090e-02 -8.92540440e-03 2.31953174e-01
-1.84858423e-02 -4.24751818e-01 -9.87450778e-01 -5.38000345e-01
-2.08534598e-02 -8.59975874e-01 9.14740562e-01 4.27450240e-01
1.23965228e+00 -1.06913328e-01 -1.89757213e-01 1.06441832e+00
1.15561390e+00 -3.08053762e-01 1.76105708e-01 3.50863844e-01
1.13838875e+00 6.83921993e-01 8.12252164e-01 2.30871186e-01
6.61327839e-01 7.74299741e-01 7.30514169e-01 -7.89446533e-02
-4.09695245e-02 -6.46553993e-01 6.36833847e-01 8.15403044e-01
-7.46961683e-02 1.33679584e-01 -1.01523590e+00 2.24180430e-01
-1.81237209e+00 -4.38870341e-01 -3.62500995e-02 2.30858445e+00
9.87742782e-01 2.27808669e-01 4.14667428e-02 -1.43807586e-02
1.84981123e-01 6.42208755e-02 -8.20939600e-01 1.38443381e-01
2.17540413e-01 2.59138104e-02 6.94810033e-01 7.40121603e-01
-1.07352686e+00 8.48259330e-01 5.90598917e+00 4.47847247e-01
-9.14651692e-01 -6.08612746e-02 3.73892009e-01 -2.58305907e-01
-3.69547158e-01 8.48402232e-02 -1.26971591e+00 1.58698440e-01
3.86519104e-01 5.50505221e-01 1.36960238e-01 9.33392346e-01
1.91091582e-01 -2.55769759e-01 -1.53318238e+00 1.19926775e+00
1.82005793e-01 -1.09007382e+00 -1.32970884e-01 1.20073609e-01
8.25914204e-01 -8.57083499e-02 6.15705885e-02 1.18122175e-01
2.74278790e-01 -1.06815886e+00 7.33943880e-01 2.94072956e-01
6.87009633e-01 -7.25118160e-01 5.28470218e-01 7.25532651e-01
-1.01711965e+00 3.10881943e-01 -3.95369083e-01 4.10746112e-02
1.13490239e-01 7.99136102e-01 -8.63174975e-01 3.00914228e-01
8.60767663e-01 8.92471492e-01 -3.74576539e-01 8.01993370e-01
-4.80870873e-01 3.76036912e-01 -5.90002716e-01 4.33961481e-01
2.28469335e-02 -1.42409742e-01 5.56126952e-01 7.99252212e-01
3.15123975e-01 -4.13662978e-02 5.24260581e-01 5.65377355e-01
-1.61980212e-01 -2.82548934e-01 -6.36771023e-01 3.60510737e-01
4.39406276e-01 1.15534544e+00 -5.96482575e-01 -3.28518227e-02
-3.50741148e-01 1.06311297e+00 4.73456383e-01 1.07686602e-01
-5.81641257e-01 2.11896494e-01 5.60605586e-01 3.98896664e-01
2.29445368e-01 -3.86155486e-01 -2.73034036e-01 -1.51604855e+00
1.32716924e-01 -6.41137242e-01 2.86816329e-01 -7.12367654e-01
-1.28197360e+00 -5.16148955e-02 2.21839789e-02 -1.39005101e+00
-3.07762116e-01 -8.12862813e-01 -2.49221563e-01 5.73966622e-01
-1.58563793e+00 -9.35412824e-01 -4.51649338e-01 3.53617787e-01
4.96666491e-01 1.82081893e-01 5.74384391e-01 3.82732838e-01
-3.43026280e-01 7.22248733e-01 -3.96261603e-01 1.07617274e-01
1.24656296e+00 -1.59238851e+00 2.15333760e-01 4.06999856e-01
1.27916649e-01 5.17516017e-01 5.20085990e-01 -5.89593589e-01
-1.51991880e+00 -8.25480342e-01 7.52932310e-01 -9.21334207e-01
2.43104577e-01 -5.03382981e-01 -8.24548721e-01 7.83493996e-01
-4.48827416e-01 3.81810933e-01 4.98243123e-01 2.80847669e-01
-4.86713260e-01 -6.04527518e-02 -1.14627349e+00 3.69104117e-01
1.39597011e+00 -4.32351351e-01 -4.75187659e-01 2.03010410e-01
8.32753003e-01 -1.06630886e+00 -9.12562549e-01 5.46514869e-01
6.62860870e-01 -8.65366340e-01 1.26734567e+00 -4.15199250e-01
5.82793772e-01 -3.62413734e-01 -2.28227764e-01 -1.14786375e+00
1.73648521e-02 -3.39413673e-01 -6.16432190e-01 9.00688291e-01
3.44386220e-01 -2.59841144e-01 1.48012149e+00 4.15999740e-01
-1.57400474e-01 -9.66189384e-01 -7.61868715e-01 -9.63596582e-01
5.85851483e-02 -2.25542024e-01 1.74889863e-01 1.03398097e+00
-4.26073849e-01 5.65905809e-01 -4.61706758e-01 2.81919420e-01
1.03425741e+00 1.61964029e-01 1.07486498e+00 -1.46520460e+00
-4.93383139e-01 1.99117162e-03 -4.24236745e-01 -1.86407793e+00
1.40999153e-01 -9.80511665e-01 1.30830780e-01 -1.36252594e+00
1.92382812e-01 -3.80859137e-01 5.42209409e-02 5.77435553e-01
9.40295458e-02 3.11761796e-01 -6.58483431e-02 1.60836637e-01
-6.50674284e-01 6.41174316e-01 1.75153172e+00 2.05032781e-01
-1.07008606e-01 1.31598175e-01 -4.12202239e-01 1.25565565e+00
5.47434866e-01 -4.88671541e-01 -6.47185147e-01 -7.04018235e-01
3.98695499e-01 2.43214414e-01 5.18053770e-01 -7.40794837e-01
1.79775715e-01 -1.90323249e-01 9.19056833e-01 -1.05689132e+00
5.42253852e-01 -9.05019701e-01 -2.59284168e-01 3.15624684e-01
-2.00091213e-01 7.71532059e-02 -1.50080398e-01 6.44327939e-01
-9.28761065e-02 -2.45850444e-01 8.92551184e-01 -3.55407953e-01
-5.08595824e-01 5.10562420e-01 6.53920323e-02 2.74781525e-01
6.48389697e-01 -3.20857584e-01 1.42315581e-01 -5.20262837e-01
-8.29952240e-01 3.97006184e-01 8.20798695e-01 1.86510429e-01
6.90631866e-01 -1.40612543e+00 -2.19826400e-01 1.93328232e-01
3.13524790e-02 1.03574896e+00 -1.48188964e-01 9.05638099e-01
-4.10193205e-01 2.39991799e-01 7.61065632e-02 -1.06855047e+00
-1.11952126e+00 1.40112668e-01 2.76933014e-01 -2.21347675e-01
-9.10010338e-01 1.02022696e+00 4.05668080e-01 -1.06364310e+00
6.93686247e-01 -3.22276413e-01 8.09557587e-02 2.78566498e-02
1.34171292e-01 3.17971885e-01 1.07291691e-01 -2.04087749e-01
-3.67685348e-01 9.01679158e-01 -1.00924894e-01 -2.58596659e-01
1.56775689e+00 -2.45355904e-01 2.12910861e-01 5.71661830e-01
1.31440604e+00 1.69765100e-01 -1.88595474e+00 -2.83018351e-01
7.64357075e-02 -5.74536741e-01 2.05176070e-01 -5.60746789e-01
-1.07754588e+00 9.85875964e-01 4.76530254e-01 -6.94391727e-01
6.46754444e-01 1.52612835e-01 8.95720184e-01 7.41536379e-01
5.19196868e-01 -1.13522696e+00 6.06492698e-01 7.88562000e-01
7.02166855e-01 -1.69126678e+00 3.41066003e-01 -5.63089550e-01
-3.07504773e-01 1.08045042e+00 1.21187878e+00 -1.66048661e-01
5.51810324e-01 4.00962949e-01 1.61590829e-01 -1.98684692e-01
-6.34704769e-01 1.27161011e-01 5.11595547e-01 5.04316390e-01
5.10174930e-01 -3.51961821e-01 -4.60922271e-02 2.11294189e-01
-3.73540968e-01 -3.82298708e-01 3.51258636e-01 1.04659581e+00
-4.02328968e-01 -1.30052710e+00 -3.58764410e-01 2.19885677e-01
-3.15357864e-01 2.99448222e-01 -2.28193015e-01 8.14108610e-01
-2.70974301e-02 2.44572058e-01 5.11086471e-02 -1.66389212e-01
1.63697883e-01 -8.10546428e-02 9.46503997e-01 -8.38532865e-01
-1.18322670e-01 2.06498206e-01 -1.51987061e-01 -8.33931863e-01
-4.02126878e-01 -6.53544068e-01 -1.48046410e+00 -1.37976333e-01
-2.87705868e-01 -7.87808597e-02 5.35992682e-01 8.20179224e-01
1.58226684e-01 9.10891742e-02 6.06444776e-01 -1.10581017e+00
-8.74041319e-01 -5.25281131e-01 -4.03636247e-01 3.28319490e-01
7.09272027e-01 -1.04070020e+00 -4.51681972e-01 -4.70828228e-02] | [7.0955281257629395, -1.1691763401031494] |
207e1f6f-46e6-48a9-a474-1508b9fa3a5f | statistical-properties-of-color-matching | 2007.02197 | null | https://arxiv.org/abs/2007.02197v2 | https://arxiv.org/pdf/2007.02197v2.pdf | Statistical properties of color matching functions | In trichromats, color vision entails the projection of an infinite-dimensional space (the one containing all possible electromagnetic power spectra) onto the 3-dimensional space that modulates the activity of the three types of cones. This drastic reduction in dimensionality gives rise to metamerism, that is, the perceptual chromatic equivalence between two different light spectra. The classes of equivalence of metamerism are revealed by color-matching experiments, in which observers adjust the intensity of three monochromatic light beams of three pre-set wavelengths (the primaries) to produce a mixture that is perceptually equal to a given monochromatic target stimulus. Here we use the linear relation between the color matching functions and the absorption probabilities of each type of cone to find particularly useful triplets of primaries. As a second goal, we also derive an analytical description of the trial-to-trial variability and the correlations of color matching functions stemming from Poissonian noise in photon capture. We analyze how the statistical properties of the responses to color-matching experiments vary with the retinal composition and the wavelengths of peak absorption probability, and compare them with experimental data on subject-to-subject variability obtained previously. | ['Inés Samengo', 'María da Fonseca'] | 2020-07-04 | null | null | null | null | ['metamerism'] | ['computer-vision'] | [ 4.50029224e-01 -6.05629206e-01 3.58621597e-01 -6.64937720e-02
2.35345624e-02 -8.02175105e-01 4.86562580e-01 -1.89904779e-01
-7.78554022e-01 5.29771984e-01 -6.24160748e-03 -3.10052037e-01
-1.30449310e-01 -6.35611951e-01 -4.35223281e-01 -1.00305057e+00
3.26789916e-01 3.96711342e-02 2.99363047e-01 1.29983081e-02
5.35829127e-01 5.56941926e-01 -1.93550026e+00 6.46402985e-02
7.35946059e-01 9.43609715e-01 1.14290431e-01 1.04190481e+00
9.02532414e-03 4.11723107e-02 -5.10103285e-01 -2.31718838e-01
6.39696956e-01 -1.06438100e+00 -4.42390978e-01 -8.87811091e-03
6.13815188e-01 3.39474618e-01 -1.30735636e-01 1.41553795e+00
3.69876564e-01 1.13085099e-01 7.41977096e-01 -7.88569152e-01
-1.04838383e+00 6.53377995e-02 -6.30191863e-01 3.50710034e-01
4.25099015e-01 3.18359792e-01 6.07393146e-01 -3.41180205e-01
3.29410821e-01 1.00910473e+00 -4.36761081e-02 6.68048263e-01
-1.63082576e+00 -3.02552819e-01 -4.47016120e-01 4.06779796e-01
-1.26149070e+00 -6.50673926e-01 5.03251135e-01 -6.32071435e-01
4.48217750e-01 4.71449226e-01 9.32137609e-01 4.40171450e-01
6.67940438e-01 -5.29672980e-01 1.86562932e+00 -8.84053767e-01
3.97056580e-01 3.25838387e-01 1.54901501e-02 5.13505578e-01
5.54812729e-01 6.42609656e-01 -5.39373457e-01 -9.82886851e-02
6.14842713e-01 -5.73867023e-01 -5.65709710e-01 -1.61446229e-01
-9.62393403e-01 1.68141901e-01 2.94110358e-01 4.15401071e-01
-2.18076408e-01 -2.12683946e-01 -3.79150510e-01 2.45348349e-01
-1.11125864e-01 8.71836126e-01 3.89205068e-02 2.75879800e-01
-1.27082124e-01 -8.90746042e-02 6.26296043e-01 3.03987205e-01
6.92611814e-01 -2.98671842e-01 -3.00143287e-02 6.07678711e-01
1.18390158e-01 8.38077843e-01 3.97497386e-01 -1.01718688e+00
-1.13552690e-01 4.57507581e-01 3.58286828e-01 -3.33082169e-01
-4.48730856e-01 -1.39368087e-01 -4.20765907e-01 8.63771081e-01
9.57975984e-01 -1.80213764e-01 -9.11058486e-01 1.76500344e+00
5.34054786e-02 -2.15679318e-01 1.12486936e-01 1.20134223e+00
3.73135984e-01 3.51176858e-01 1.20629393e-01 -7.15866864e-01
1.30726802e+00 1.00435011e-01 -5.02364755e-01 1.81740355e-02
-5.68028316e-02 -1.05251706e+00 1.04953039e+00 1.85148850e-01
-1.22317672e+00 -5.22611260e-01 -8.99176776e-01 3.69995207e-01
-1.92962974e-01 -2.70931810e-01 1.90223992e-01 9.92567301e-01
-8.91604662e-01 3.70036542e-01 -3.34251463e-01 -2.18079165e-01
-1.61661372e-01 -4.51336056e-02 -2.62559384e-01 2.25270279e-02
-7.29657590e-01 1.07175338e+00 6.73037320e-02 1.46214247e-01
-1.12459973e-01 -6.46793425e-01 -2.18458012e-01 -1.89750239e-01
-1.53221145e-01 -7.86305249e-01 1.00043857e+00 -1.05695808e+00
-1.51261878e+00 1.22922242e+00 -4.00327146e-01 5.29440381e-02
4.42895852e-03 4.42177176e-01 -6.56860590e-01 2.84990847e-01
-5.88084273e-02 1.23632707e-01 7.02879190e-01 -1.20022893e+00
-2.88456470e-01 -7.36037433e-01 -8.60608295e-02 2.13437736e-01
2.67715663e-01 3.81286740e-01 5.28893322e-02 1.67386517e-01
4.64077711e-01 -9.82650340e-01 -2.38353200e-02 -2.25068539e-01
-4.13024455e-01 2.54982039e-02 -4.62084562e-02 -4.54822630e-02
4.83667403e-01 -2.44537139e+00 1.93229288e-01 1.73503548e-01
2.11562172e-01 6.54414147e-02 -3.94210041e-01 3.01187992e-01
-2.80841708e-01 -1.78125441e-01 -7.75242597e-02 4.59939808e-01
-1.74586311e-01 -2.98787683e-01 2.40825880e-02 7.74205804e-01
6.07204856e-03 4.72747028e-01 -7.79435396e-01 -1.37291983e-01
1.04546078e-01 2.40975410e-01 -1.84675246e-01 1.50085226e-01
-8.46193731e-02 4.67507184e-01 2.53133532e-02 3.17857206e-01
1.00422680e+00 1.38024375e-01 2.25504816e-01 -4.22621638e-01
-6.85985565e-01 1.82398893e-02 -6.07465446e-01 9.62331831e-01
-6.87757730e-02 9.07586098e-01 -3.17646623e-01 -4.13505346e-01
6.96726382e-01 9.90359411e-02 2.47770756e-01 -1.10169721e+00
1.75994873e-01 2.52737463e-01 8.45652819e-01 -5.56267202e-01
2.26630345e-01 -7.19421506e-01 2.05121398e-01 3.61826032e-01
-6.64394721e-02 -3.69379252e-01 3.96207869e-01 -1.97984159e-01
5.05853593e-01 -1.82607025e-01 2.79988080e-01 -1.98725462e-01
1.39781982e-01 -4.67036396e-01 3.43189508e-01 5.89082599e-01
-2.32383177e-01 3.54732096e-01 6.69312298e-01 -2.81017542e-01
-9.75427568e-01 -1.62045574e+00 -6.06893003e-01 6.01609051e-01
5.06947100e-01 1.87387168e-01 -5.99854708e-01 5.13046920e-01
-1.37408510e-01 8.19970906e-01 -5.88114619e-01 -4.13368136e-01
-9.38724726e-03 -1.12786901e+00 -4.02571373e-02 -1.54442519e-01
3.96203488e-01 -1.04347098e+00 -1.15696192e+00 -2.56096572e-01
1.14567690e-01 -1.01477921e+00 -2.97830347e-02 3.51055205e-04
-4.73072320e-01 -1.50868642e+00 -3.65950376e-01 -2.40123972e-01
6.09998345e-01 2.76761979e-01 9.75991845e-01 -3.50872874e-01
-9.52926874e-01 7.41279900e-01 -2.75957376e-01 -7.07594156e-01
-5.77743530e-01 -9.52818036e-01 3.22366625e-01 2.84572989e-01
5.79167902e-01 -5.55236161e-01 -7.58739412e-01 1.13959573e-01
-5.96604168e-01 -2.53900766e-01 3.99867803e-01 1.14187777e-01
5.47148108e-01 3.52668427e-02 -3.50626372e-02 -2.91112125e-01
4.93524343e-01 2.82316376e-02 -1.02495813e+00 2.66513735e-01
-3.91995311e-01 1.55159816e-01 5.55640221e-01 -4.15716916e-01
-1.03054249e+00 -2.04717383e-01 6.18144095e-01 2.02780753e-01
-4.11145806e-01 -4.25278284e-02 1.11239240e-01 -6.52774513e-01
1.17832065e+00 3.72183472e-01 -2.60457575e-01 -2.11513221e-01
4.36974406e-01 3.92856538e-01 7.53876328e-01 -2.86756068e-01
7.98310578e-01 4.75165933e-01 6.06193125e-01 -1.03886139e+00
-8.05141270e-01 -2.81005204e-01 -5.06833375e-01 -4.41459000e-01
9.94976342e-01 -3.65975410e-01 -1.11570239e+00 6.72948539e-01
-1.04932559e+00 -6.81576133e-02 -3.53372157e-01 9.48850751e-01
-5.93574345e-01 2.33834893e-01 -2.81714797e-01 -1.10281205e+00
3.05792987e-01 -8.43618512e-01 3.25305820e-01 8.05951118e-01
3.60797554e-01 -5.22574425e-01 5.54941714e-01 -6.41219914e-02
2.13783711e-01 7.20672831e-02 1.32351732e+00 9.87429395e-02
-7.29408562e-01 -1.44219324e-01 -3.27662706e-01 2.56579071e-01
1.83731496e-01 2.62350023e-01 -1.09974122e+00 -5.36224153e-03
3.37986648e-01 -1.12934873e-01 8.91541481e-01 8.66592467e-01
5.97813129e-01 1.56169727e-01 -2.40094215e-02 6.04931593e-01
1.68402565e+00 7.70286679e-01 7.46295214e-01 1.42888069e-01
1.90841615e-01 7.78489053e-01 1.91863924e-02 2.14184567e-01
-3.22260648e-01 5.78256488e-01 2.45141476e-01 -9.02242959e-02
-1.61524564e-01 1.87457904e-01 -2.78188922e-02 4.27258760e-01
-6.46697581e-01 -6.39824197e-02 -3.63804072e-01 9.47458819e-02
-9.35235620e-01 -1.18024015e+00 -3.98166269e-01 2.87295747e+00
8.04052770e-01 -1.34548679e-01 2.53181487e-01 -9.45833027e-02
7.59340644e-01 -2.54153669e-01 -3.75821978e-01 -2.21888781e-01
-3.97856176e-01 4.20589864e-01 5.64911485e-01 6.83606267e-01
-4.61634874e-01 3.70765805e-01 7.49363804e+00 3.07050552e-02
-1.04725766e+00 -2.88136303e-01 9.79742333e-02 -2.70549417e-01
-3.69930267e-01 1.66518137e-01 -3.43036711e-01 7.15002418e-01
1.05577064e+00 -3.66598099e-01 7.88003445e-01 -2.29298651e-01
2.20338300e-01 -8.37992966e-01 -1.08697796e+00 1.04751921e+00
-2.32665017e-01 -1.03110421e+00 -7.59152025e-02 1.70210183e-01
4.41256016e-01 -1.40520409e-01 3.38312089e-01 -5.59767246e-01
-1.09201252e-01 -6.60797775e-01 5.23786724e-01 1.09997559e+00
1.00105810e+00 -3.01472187e-01 1.73143551e-01 -6.96548074e-02
-8.51920426e-01 -1.81752756e-01 -8.02398145e-01 -2.15909511e-01
-9.94219445e-03 4.51587737e-01 -4.45547253e-01 6.47916794e-02
4.21962738e-01 -1.66402429e-01 -4.52458948e-01 1.54565787e+00
5.94146438e-02 2.32856557e-01 -9.56266001e-02 -2.03589812e-01
-3.31778973e-01 -8.23846638e-01 7.51096666e-01 6.39862716e-01
8.93197581e-02 5.14039516e-01 -6.34948969e-01 1.47225094e+00
2.18912229e-01 -1.22196324e-01 -4.58642215e-01 -1.51781648e-01
4.79542136e-01 1.16369712e+00 -7.33939469e-01 8.64588320e-02
-5.19177616e-01 7.56422937e-01 -3.37261073e-02 8.55423629e-01
-2.28178903e-01 -5.21180451e-01 7.39042401e-01 4.87984121e-02
-2.07530528e-01 -2.30200335e-01 -6.45783544e-02 -7.71951973e-01
2.73982938e-02 -2.43986785e-01 -1.03096366e-01 -1.06073964e+00
-1.30330086e+00 5.94208360e-01 -3.22129987e-02 -9.08264458e-01
7.74682462e-02 -9.04091537e-01 -5.38429141e-01 1.59097016e+00
-1.21556854e+00 -3.66357654e-01 -2.59141058e-01 7.52648473e-01
-4.81420249e-01 -4.62667271e-02 8.60838652e-01 -3.92787188e-01
-2.59339660e-01 1.83091924e-01 1.78392708e-01 -1.05232880e-01
6.61970794e-01 -1.31376803e+00 2.52684772e-01 8.72590780e-01
3.06053758e-01 5.06623685e-01 8.84718776e-01 -2.60287136e-01
-1.23091245e+00 -3.51541668e-01 6.75835550e-01 -5.07056952e-01
5.11415720e-01 -9.13949609e-02 -3.93715620e-01 1.58462778e-01
1.31070554e-01 -6.31505847e-02 1.05285907e+00 -2.21571848e-02
-6.08161628e-01 -2.19107643e-01 -8.24519873e-01 6.94882810e-01
7.92011440e-01 -8.58212590e-01 -4.15399462e-01 4.50961649e-01
5.29097676e-01 1.52397873e-02 -3.93004656e-01 3.26584838e-02
8.77122462e-01 -1.58578622e+00 6.50838375e-01 -8.84119213e-01
4.81729279e-04 -3.91416520e-01 -2.56404698e-01 -1.33529663e+00
-6.60260618e-01 -3.97721082e-01 8.60246599e-01 3.77752900e-01
2.28177026e-01 -1.02112305e+00 -1.03404634e-02 5.57899415e-01
1.15918882e-01 -1.03375353e-01 -9.09263790e-01 -7.91044056e-01
-1.68528348e-01 9.89816245e-03 3.83909717e-02 5.73637247e-01
1.56162664e-01 2.95732975e-01 2.00218901e-01 2.74553537e-01
1.00927746e+00 3.60578209e-01 8.96917954e-02 -1.44072938e+00
-3.41481566e-01 -4.25686866e-01 -6.71240211e-01 -3.56509566e-01
-2.20699862e-01 -7.04416871e-01 -8.23660940e-03 -1.06216729e+00
3.97481173e-01 2.73895431e-02 -4.15026635e-01 -2.36451581e-01
-3.06909084e-02 3.14434558e-01 3.70542467e-01 2.21105412e-01
3.61926518e-02 1.37491077e-01 1.35324383e+00 4.35962081e-01
-6.79495275e-01 1.25788510e-01 -7.15875924e-01 5.16415834e-01
5.09781122e-01 -2.42458403e-01 -2.51087666e-01 -1.04438454e-01
5.91823876e-01 1.85854346e-01 8.34310591e-01 -1.28328061e+00
9.53341797e-02 -3.86585951e-01 4.40653056e-01 1.81849226e-02
2.60398060e-01 -5.26735485e-01 2.27419123e-01 3.82784516e-01
-3.60515147e-01 -3.00691962e-01 1.68557972e-01 6.34118140e-01
2.80580610e-01 -4.39868480e-01 1.28165019e+00 -4.04788971e-01
-3.62737179e-01 5.21546192e-02 -4.96357262e-01 8.96109715e-02
9.63946104e-01 -2.79616982e-01 -8.00401747e-01 -2.51675304e-02
-5.93081713e-01 -7.50596106e-01 6.73139453e-01 -3.50451581e-02
3.35300952e-01 -9.46100652e-01 -5.45590281e-01 3.16300035e-01
1.62086878e-02 -9.25101459e-01 3.65491837e-01 7.29930937e-01
-3.61157238e-01 3.90414536e-01 -8.45290959e-01 -3.65747303e-01
-1.20428836e+00 7.01519668e-01 1.03224206e+00 7.82981753e-01
-1.79621410e-02 8.33013356e-01 4.64318573e-01 3.51028442e-01
-4.95218754e-01 -1.09979838e-01 -4.62864377e-02 -3.29352796e-01
7.19566464e-01 3.87407839e-01 -1.93228334e-01 -3.48031849e-01
-3.01853031e-01 9.89618838e-01 3.36276919e-01 -3.43604922e-01
5.44204891e-01 -1.59080744e-01 -5.76994359e-01 8.86523783e-01
8.01026940e-01 3.72715980e-01 -1.02269113e+00 -1.20360866e-01
-5.86861968e-01 -8.64350021e-01 -2.52110362e-01 -9.79666114e-01
-5.34841478e-01 7.56149828e-01 9.06959057e-01 6.93383276e-01
1.43192613e+00 3.89887691e-01 -1.87309638e-01 8.31105039e-02
2.84595102e-01 -8.39470565e-01 -1.11309879e-01 -1.29789740e-01
5.77641487e-01 -7.52327502e-01 -3.25402737e-01 -2.87560672e-01
-2.23632202e-01 1.03928339e+00 4.66552377e-01 -5.85593320e-02
4.36359346e-01 -5.09022065e-02 1.46267667e-01 -2.54030555e-01
-5.78844666e-01 -4.78082776e-01 6.15307510e-01 1.15436029e+00
4.48116899e-01 4.09837097e-01 -4.11528885e-01 -4.34305072e-02
-2.92349249e-01 -1.73071519e-01 8.62326622e-01 1.05960391e-01
-6.77759111e-01 -6.85771227e-01 -6.03666723e-01 3.74217659e-01
-1.58425704e-01 -1.54239550e-01 -5.95277309e-01 3.83300066e-01
3.50104898e-01 1.06667113e+00 4.59695458e-01 -2.05894768e-01
5.49274147e-01 8.28889310e-02 1.33280456e+00 -4.37091619e-01
2.87908256e-01 3.52524891e-02 -3.31370175e-01 -4.93367851e-01
-8.20834816e-01 -5.42859674e-01 -9.09049273e-01 -1.64547026e-01
-8.04010183e-02 7.38556636e-03 5.88207543e-01 8.42559516e-01
-3.44417356e-02 1.83783904e-01 5.52195787e-01 -5.15260756e-01
-3.86336237e-01 -7.44860888e-01 -1.25936460e+00 3.56436133e-01
5.49804986e-01 -4.93141770e-01 -7.74264753e-01 -1.79736409e-02] | [10.094829559326172, 2.0486621856689453] |
9f5dc517-b56b-4b11-92e2-4b7872c543a7 | human-pose-transfer-by-adaptive-hierarchical | 2012.0694 | null | https://arxiv.org/abs/2012.06940v1 | https://arxiv.org/pdf/2012.06940v1.pdf | Human Pose Transfer by Adaptive Hierarchical Deformation | Human pose transfer, as a misaligned image generation task, is very challenging. Existing methods cannot effectively utilize the input information, which often fail to preserve the style and shape of hair and clothes. In this paper, we propose an adaptive human pose transfer network with two hierarchical deformation levels. The first level generates human semantic parsing aligned with the target pose, and the second level generates the final textured person image in the target pose with the semantic guidance. To avoid the drawback of vanilla convolution that treats all the pixels as valid information, we use gated convolution in both two levels to dynamically select the important features and adaptively deform the image layer by layer. Our model has very few parameters and is fast to converge. Experimental results demonstrate that our model achieves better performance with more consistent hair, face and clothes with fewer parameters than state-of-the-art methods. Furthermore, our method can be applied to clothing texture transfer. | ['Kun Li', 'Xingzi Liu', 'Jinsong Zhang'] | 2020-12-13 | null | null | null | null | ['pose-transfer'] | ['computer-vision'] | [ 3.40837568e-01 1.11175008e-01 2.02159926e-01 -4.61697936e-01
-2.95065373e-01 -4.13130105e-01 1.62308306e-01 -5.46773374e-01
-3.05292666e-01 6.81177318e-01 1.53321594e-01 3.26090485e-01
3.58212054e-01 -1.03818607e+00 -9.27044868e-01 -6.95327759e-01
5.66853225e-01 5.10852575e-01 4.61625606e-01 -3.39555889e-01
-2.03991055e-01 2.21440494e-01 -1.33062041e+00 3.79801929e-01
8.61533463e-01 1.08914685e+00 2.36630708e-01 4.08194274e-01
4.34111431e-02 1.89476684e-01 -2.10408583e-01 -5.43813288e-01
4.14443225e-01 -5.74559689e-01 -7.74547517e-01 3.61218244e-01
8.22607160e-01 -6.48882329e-01 -2.65713841e-01 1.19457150e+00
6.69575751e-01 1.13214329e-02 3.95153284e-01 -1.09143662e+00
-8.15707266e-01 2.40238428e-01 -7.03472853e-01 -5.60425937e-01
3.19247395e-01 2.18351826e-01 4.07705098e-01 -9.11802113e-01
7.81172395e-01 1.58949494e+00 8.19625378e-01 9.38400745e-01
-1.23505616e+00 -8.28444600e-01 3.29068571e-01 -1.25393093e-01
-1.24523330e+00 -1.17022991e-01 8.73963058e-01 -3.02199781e-01
1.73385218e-01 1.55377313e-01 1.03042161e+00 1.33207893e+00
2.37440150e-02 7.96126366e-01 1.35739040e+00 -3.12226802e-01
-1.48124352e-01 4.76801395e-02 -3.04774493e-01 1.16220713e+00
1.38872370e-01 2.59070117e-02 -5.18368840e-01 7.83966631e-02
1.32317293e+00 -7.83254765e-03 -1.79058924e-01 -4.98757452e-01
-1.17374361e+00 6.83517933e-01 5.79602003e-01 -5.00602275e-02
-3.21499288e-01 4.20006603e-01 1.49929509e-01 7.32490718e-02
5.18600345e-01 -6.00118227e-02 -2.85416573e-01 3.17099482e-01
-7.61164963e-01 5.57389915e-01 6.27263844e-01 1.03919935e+00
7.71847963e-01 -1.74423516e-01 -4.23176855e-01 8.85950446e-01
3.01125109e-01 7.36609697e-01 -9.85508785e-02 -9.23950016e-01
3.85075867e-01 4.48949069e-01 1.89229965e-01 -9.73053694e-01
-2.33108237e-01 -1.21271096e-01 -8.53961289e-01 2.39329726e-01
4.74139392e-01 -2.83006132e-01 -1.43774509e+00 1.81437588e+00
6.62518442e-01 -4.76692654e-02 -3.48668009e-01 1.29508889e+00
9.24052954e-01 4.81377780e-01 2.15371415e-01 2.15295285e-01
1.47553706e+00 -1.19597304e+00 -6.70071483e-01 -3.86977822e-01
-2.32802793e-01 -9.58260953e-01 1.18862140e+00 1.86484233e-02
-1.27698541e+00 -7.67475545e-01 -8.46682906e-01 -3.50313157e-01
-3.12728472e-02 1.37513116e-01 7.63677180e-01 4.72684562e-01
-1.02474785e+00 7.76700795e-01 -8.69252980e-01 -6.40876651e-01
5.57767689e-01 4.11475390e-01 -2.42158845e-01 -9.67683196e-02
-1.24812937e+00 6.65192008e-01 2.28418648e-01 3.10115546e-01
-7.63340473e-01 -5.89866519e-01 -1.00023735e+00 -1.76896974e-01
2.38064051e-01 -1.18773150e+00 1.03843272e+00 -1.24898899e+00
-1.85414875e+00 1.02139926e+00 -5.78545257e-02 1.51156604e-01
1.01475489e+00 -4.15063262e-01 2.67911740e-02 1.04010195e-01
1.35230631e-01 1.16994870e+00 1.02649045e+00 -1.44651794e+00
-4.49076444e-01 -4.33294654e-01 -1.96478337e-01 5.09358108e-01
-4.04896319e-01 -6.80672526e-02 -1.04591691e+00 -9.08481836e-01
1.48305431e-01 -1.15422738e+00 -1.80731490e-01 6.08006835e-01
-3.74179512e-01 9.28019434e-02 7.56831229e-01 -9.34965074e-01
8.27675998e-01 -1.87244058e+00 4.80794132e-01 3.43955964e-01
-4.86907251e-02 1.98598001e-02 -1.02491632e-01 9.58442464e-02
2.90556192e-01 -1.38408482e-01 -3.40745449e-01 -3.76858234e-01
9.06240568e-02 2.35013723e-01 -9.23652649e-02 2.81212986e-01
1.31269693e-01 1.14881480e+00 -8.45815122e-01 -6.49372995e-01
2.50220656e-01 7.49322176e-01 -6.99196339e-01 3.84141475e-01
-2.37987339e-01 7.71292567e-01 -6.63525105e-01 5.84653676e-01
7.97782242e-01 -2.02143446e-01 3.13447028e-01 -6.49516106e-01
2.32499555e-01 -1.63395122e-01 -1.13122022e+00 1.95817959e+00
-2.36642942e-01 1.34140283e-01 2.66173780e-01 -4.54658180e-01
8.12450826e-01 1.37648791e-01 4.58859712e-01 -6.51302516e-01
2.33755395e-01 7.94846117e-02 -3.57227802e-01 -5.24191916e-01
2.85131305e-01 -2.45598644e-01 -1.67386964e-01 2.42320284e-01
3.15251946e-02 -2.00666025e-01 -1.72621876e-01 -2.24028260e-01
4.03574288e-01 7.19004273e-01 -2.23713920e-01 -2.16519341e-01
2.24401399e-01 -3.03226978e-01 6.24767303e-01 3.79016995e-01
3.56360450e-02 9.24006343e-01 1.93399221e-01 -6.94728851e-01
-1.27027941e+00 -1.27680206e+00 1.34274915e-01 1.20884466e+00
5.60147941e-01 -5.22189550e-02 -1.38492048e+00 -7.52273381e-01
8.72821510e-02 2.28685096e-01 -7.45217562e-01 -1.13577560e-01
-8.36158276e-01 -6.57478034e-01 3.57050955e-01 7.54087567e-01
1.06397784e+00 -1.33736277e+00 -4.66468006e-01 1.22695900e-01
-4.46660012e-01 -1.02051342e+00 -1.05575657e+00 -4.59042728e-01
-5.97884238e-01 -7.79066920e-01 -9.53184068e-01 -1.06035328e+00
1.08720696e+00 -1.60834581e-01 9.16583419e-01 1.64471686e-01
-4.42467898e-01 1.46455541e-01 -3.92825007e-02 -2.58617699e-01
-1.34700924e-01 -2.06583608e-02 -9.06027332e-02 3.03095698e-01
-2.21467726e-02 -3.70267510e-01 -1.03646410e+00 4.91343021e-01
-7.07159877e-01 4.30964202e-01 4.62951183e-01 9.19902563e-01
6.29478872e-01 -1.20885842e-01 2.81464964e-01 -9.28724110e-01
3.98721099e-01 9.86583382e-02 -2.87477970e-01 3.51652741e-01
-3.78141612e-01 -9.86567885e-02 3.62152189e-01 -5.02902329e-01
-1.35941279e+00 4.47036177e-01 -1.88852936e-01 -4.63578731e-01
-1.53471649e-01 -1.42518997e-01 -3.68344039e-01 -1.61330566e-01
3.58772248e-01 5.92939965e-02 2.10259736e-01 -5.24973333e-01
4.45885688e-01 2.54375130e-01 6.07687950e-01 -9.38885212e-01
9.03127611e-01 5.91895640e-01 -2.26591185e-01 -4.47261333e-01
-1.00260878e+00 5.34206517e-02 -8.65477860e-01 -4.29780841e-01
1.24871182e+00 -7.76087999e-01 -6.66492403e-01 8.38934362e-01
-1.09944510e+00 -5.45789599e-01 -1.96250111e-01 1.25810608e-01
-5.04909039e-01 3.34416479e-01 -9.80210304e-01 -5.14051378e-01
-5.96390605e-01 -1.00029409e+00 1.40266585e+00 3.48651141e-01
-2.36616656e-01 -7.94133961e-01 -2.72297829e-01 4.82889414e-01
4.13474262e-01 6.29396081e-01 7.18012035e-01 2.91585028e-01
-5.53993285e-01 1.35667652e-01 -3.11156690e-01 2.43254542e-01
1.87729493e-01 -2.97089994e-01 -8.86697173e-01 -5.88257492e-01
-3.67589027e-01 -5.75447261e-01 8.62539709e-01 2.54900783e-01
1.39159453e+00 -3.58665258e-01 -2.33017072e-01 8.72497082e-01
1.22745478e+00 -1.22757971e-01 6.95525885e-01 -2.84866989e-02
1.10198522e+00 8.05896759e-01 5.45830011e-01 7.78829604e-02
4.38274622e-01 7.04023540e-01 9.16987062e-02 -7.47622252e-01
-4.29527193e-01 -6.51394546e-01 3.03450137e-01 6.54785454e-01
-4.45964575e-01 -6.26432151e-03 -4.75396246e-01 2.66469508e-01
-1.94333017e+00 -8.37402582e-01 2.46830001e-01 2.10797191e+00
1.06235147e+00 9.87813398e-02 2.04009846e-01 -4.77440029e-01
8.53144228e-01 -4.18303302e-03 -5.87528229e-01 -1.45453466e-02
8.43130201e-02 5.01098692e-01 5.31512380e-01 4.70064282e-01
-1.20100641e+00 1.38049006e+00 6.23313761e+00 6.15229964e-01
-1.04775119e+00 1.06775433e-01 5.50571740e-01 -7.90633261e-02
-2.72297025e-01 -3.63400549e-01 -6.27240419e-01 4.57437813e-01
5.18094338e-02 3.28355759e-01 5.31910121e-01 6.42340899e-01
5.83944749e-03 3.14454049e-01 -1.00897634e+00 9.81173754e-01
1.04134031e-01 -9.20643508e-01 2.28145555e-01 -7.66154230e-02
8.92985165e-01 -5.99303961e-01 2.17495248e-01 8.37962106e-02
3.29246938e-01 -1.04484916e+00 1.09141552e+00 7.36070037e-01
1.15062749e+00 -7.41027594e-01 4.92984623e-01 -1.07733086e-01
-1.26682448e+00 2.03177229e-01 -3.78874183e-01 1.49338886e-01
3.46988529e-01 2.43260488e-01 -4.32673544e-01 4.61183846e-01
1.02649319e+00 4.80175227e-01 -4.66070533e-01 6.37704253e-01
-2.81774372e-01 3.35213363e-01 -2.52569735e-01 3.29949483e-02
6.70912489e-03 -2.96067029e-01 2.25284114e-01 1.10630631e+00
3.38664472e-01 7.67820254e-02 6.62980735e-01 1.03178084e+00
-7.34261125e-02 6.90414980e-02 -3.87875944e-01 3.34453881e-01
3.13111275e-01 1.25384259e+00 -7.97833562e-01 -2.95807779e-01
-1.97097987e-01 1.69537365e+00 3.02400619e-01 4.55888212e-01
-9.65115249e-01 -1.19475260e-01 5.16061187e-01 4.13710535e-01
2.98594207e-01 -4.56212945e-02 -2.00737879e-01 -1.05709493e+00
1.55730560e-01 -7.79987931e-01 2.25133941e-01 -7.76691437e-01
-1.54336274e+00 5.73814154e-01 -1.32942975e-01 -9.22325313e-01
8.06575865e-02 -6.49224162e-01 -3.60330164e-01 9.71172631e-01
-1.07040584e+00 -1.90878868e+00 -6.86315358e-01 7.15461552e-01
4.71823066e-01 2.50690758e-01 7.33018517e-01 4.47879791e-01
-3.27341616e-01 7.60668576e-01 -5.32332957e-01 3.06170315e-01
1.00121653e+00 -1.21961474e+00 4.89317954e-01 4.74999726e-01
-5.02560794e-01 6.16565406e-01 4.74543929e-01 -1.01513171e+00
-1.09194195e+00 -1.38337207e+00 5.33485472e-01 -3.49542916e-01
1.40754968e-01 -6.63933694e-01 -7.10879147e-01 5.97705245e-01
2.75102526e-01 8.20924062e-03 3.21023047e-01 -1.24860249e-01
-1.96552455e-01 -2.89352119e-01 -1.19676709e+00 8.06658208e-01
1.39368522e+00 -3.03942829e-01 -3.59081864e-01 3.16435844e-01
6.99343443e-01 -8.37636530e-01 -9.05419230e-01 6.11169875e-01
8.93051326e-01 -6.62407517e-01 1.09615874e+00 -4.72927660e-01
4.29001510e-01 -3.23911041e-01 4.96724732e-02 -1.20430064e+00
-6.14574432e-01 -5.33835471e-01 2.62641013e-01 1.17593038e+00
2.48634726e-01 -3.86390924e-01 8.65937889e-01 5.56779563e-01
5.70719093e-02 -7.75147974e-01 -5.39866686e-01 -4.44069117e-01
4.04789485e-02 2.07916632e-01 6.05148971e-01 8.72243702e-01
-5.85750282e-01 3.37802142e-01 -8.15300643e-01 -3.01598180e-02
9.66461658e-01 3.69227707e-01 8.36185157e-01 -1.08488786e+00
-2.94442177e-01 -2.07759961e-01 -5.73621094e-02 -9.62051570e-01
1.67890992e-02 -8.04909587e-01 2.75504380e-01 -1.52112174e+00
4.74436820e-01 -4.47717130e-01 -1.24301687e-01 8.10271621e-01
-4.25740153e-01 6.76993728e-01 1.98971316e-01 -7.33660758e-02
-4.11271602e-01 5.98031044e-01 1.94467914e+00 -1.05227686e-01
-1.00250222e-01 -2.12069497e-01 -5.46172440e-01 9.42413807e-01
6.18849933e-01 -6.69490397e-02 -4.32128429e-01 -4.76588666e-01
-1.25577420e-01 -2.25662157e-01 7.62626350e-01 -7.84559548e-01
1.17296295e-03 -2.07454175e-01 1.07839572e+00 -3.80924910e-01
3.88329864e-01 -7.57986665e-01 3.81101161e-01 5.33680797e-01
-2.67162681e-01 -4.80287336e-02 6.13458902e-02 4.17467326e-01
1.84810787e-01 3.52181703e-01 1.07768309e+00 -4.35779244e-01
-5.33728123e-01 7.61502683e-01 1.75542682e-01 -1.33229539e-01
9.54521477e-01 -1.31878465e-01 2.66019665e-02 -2.22013429e-01
-9.46017981e-01 3.10967594e-01 8.51408124e-01 7.40872264e-01
7.23707020e-01 -1.69736362e+00 -7.12531388e-01 3.10889721e-01
-1.22332834e-01 2.10528553e-01 4.21952307e-01 4.93676782e-01
-6.85328245e-01 -1.43910751e-01 -5.68529963e-01 -6.09043062e-01
-1.23881447e+00 5.28372228e-01 4.80424911e-01 -7.69437030e-02
-7.92227566e-01 7.96290636e-01 5.92991114e-01 -7.55750358e-01
1.82964727e-01 -3.34874308e-03 8.54794085e-02 -2.45899156e-01
2.87348896e-01 1.01006761e-01 -2.54764915e-01 -5.95478356e-01
-1.79514393e-01 1.07963443e+00 -6.44603297e-02 -1.26039401e-01
1.17220700e+00 -1.71046764e-01 -2.96418607e-01 1.34039596e-01
9.23999548e-01 -8.32928494e-02 -1.75080287e+00 -9.43496600e-02
-6.37591600e-01 -5.73540390e-01 -2.71513581e-01 -8.99061143e-01
-1.33609855e+00 7.71553218e-01 8.05396855e-01 -3.19960535e-01
1.18310881e+00 -1.80695683e-01 1.20081592e+00 -1.03775859e-01
6.71758652e-01 -1.37747192e+00 3.13274562e-01 2.15016916e-01
1.14491737e+00 -8.79889309e-01 -2.07342803e-01 -8.54808211e-01
-6.62345052e-01 9.09373820e-01 1.19793057e+00 -1.71627775e-01
3.69212687e-01 2.43254080e-01 2.09744602e-01 -1.67421013e-01
-2.83120304e-01 -1.50485426e-01 5.09080708e-01 5.82078636e-01
2.68230587e-01 4.22165692e-02 -3.62804174e-01 3.87185454e-01
-2.72680819e-01 2.30526980e-02 -1.63838401e-01 7.91881621e-01
-2.98526615e-01 -1.14562082e+00 -4.98215139e-01 2.17332184e-01
-3.91707033e-01 1.02809835e-02 -4.28181559e-01 7.72910178e-01
4.84258533e-01 4.21689659e-01 3.60802300e-02 -3.48020136e-01
6.07962728e-01 5.38167683e-03 9.05498981e-01 -6.02416575e-01
-5.55449724e-01 4.27474052e-01 -5.01071960e-02 -8.04860651e-01
-3.15584660e-01 -4.49360639e-01 -1.19101763e+00 -4.03794229e-01
3.08953202e-03 -3.26211900e-01 3.29016030e-01 7.09221780e-01
2.27925822e-01 5.98505378e-01 3.85793239e-01 -1.13236785e+00
-3.64681572e-01 -8.45461428e-01 -4.90899026e-01 9.21483457e-01
-3.61523987e-03 -8.16494465e-01 2.55104274e-01 4.33646262e-01] | [11.973681449890137, -0.8481130003929138] |
fc95d966-c016-4eb3-b866-8c6a39b9f330 | geometric-constraints-in-probabilistic | 2307.04493 | null | https://arxiv.org/abs/2307.04493v1 | https://arxiv.org/pdf/2307.04493v1.pdf | Geometric Constraints in Probabilistic Manifolds: A Bridge from Molecular Dynamics to Structured Diffusion Processes | Understanding the macroscopic characteristics of biological complexes demands precision and specificity in statistical ensemble modeling. One of the primary challenges in this domain lies in sampling from particular subsets of the state-space, driven either by existing structural knowledge or specific areas of interest within the state-space. We propose a method that enables sampling from distributions that rigorously adhere to arbitrary sets of geometric constraints in Euclidean spaces. This is achieved by integrating a constraint projection operator within the well-regarded architecture of Denoising Diffusion Probabilistic Models, a framework founded in generative modeling and probabilistic inference. The significance of this work becomes apparent, for instance, in the context of deep learning-based drug design, where it is imperative to maintain specific molecular profile interactions to realize the desired therapeutic outcomes and guarantee safety. | ['Markus Lill', 'Justin Diamond'] | 2023-07-10 | null | null | null | null | ['denoising', 'specificity'] | ['computer-vision', 'natural-language-processing'] | [ 5.26853740e-01 -9.98770073e-02 1.66776255e-01 -8.34973082e-02
-4.31468904e-01 -6.81146204e-01 7.10036337e-01 2.77330011e-01
-4.14777726e-01 1.15412140e+00 1.20809287e-01 -1.45445690e-01
-8.75505388e-01 -8.51089537e-01 -6.58389986e-01 -1.47132134e+00
-9.09521505e-02 5.96695483e-01 -1.53059945e-01 -2.21868649e-01
2.90503025e-01 9.83532846e-01 -1.06825221e+00 1.20022237e-01
1.05770147e+00 6.19892418e-01 3.45302634e-02 5.91355264e-01
6.85826614e-02 3.80138397e-01 -4.12261993e-01 -2.52189875e-01
-3.29650462e-01 -5.98999143e-01 -4.09978092e-01 -4.27029207e-02
-8.26515853e-02 3.25565189e-01 -1.57888204e-01 1.18351769e+00
5.45283675e-01 2.06123516e-01 1.32231450e+00 -6.82905793e-01
-6.76230371e-01 2.75224477e-01 -7.12710321e-02 1.16805054e-01
-1.06592894e-01 2.91213602e-01 6.68881714e-01 -6.89961076e-01
6.35495067e-01 9.28420007e-01 3.78342927e-01 6.31413460e-01
-1.88188863e+00 -2.43506625e-01 1.13647103e-01 5.71415313e-02
-1.54951656e+00 -5.18046141e-01 7.64348626e-01 -8.22395205e-01
6.18272841e-01 2.47723386e-01 4.87251580e-01 1.49813712e+00
9.28804100e-01 1.45426035e-01 1.21649683e+00 -7.30526000e-02
9.98116076e-01 -4.22856398e-02 6.93563446e-02 2.40073219e-01
4.48175520e-01 1.75679862e-01 -6.05088949e-01 -4.52620506e-01
6.91815972e-01 1.01482183e-01 -4.48570460e-01 -8.15429151e-01
-1.04066563e+00 7.38168776e-01 -7.76944086e-02 5.43242216e-01
-8.28184843e-01 1.04384892e-01 7.73066059e-02 -1.83759391e-01
5.13172746e-01 2.84323305e-01 -4.52504486e-01 -3.49772312e-02
-1.02222371e+00 2.51831532e-01 8.35629284e-01 4.49713022e-01
4.85503018e-01 -2.32702699e-02 -2.05939189e-02 1.46607697e-01
5.96101820e-01 5.69695830e-01 -1.80438355e-01 -8.88378620e-01
-1.03340469e-01 2.36122638e-01 3.07025343e-01 -7.30777681e-01
-2.55872369e-01 -9.97659624e-01 -1.15589941e+00 4.19162869e-01
5.44826567e-01 -2.65641272e-01 -5.90808392e-01 1.89143348e+00
5.06744802e-01 1.12107083e-01 1.31738409e-01 5.88306963e-01
1.35655269e-01 4.93593544e-01 3.80954623e-01 -5.18666446e-01
1.02648282e+00 1.68144271e-01 -6.11182630e-01 3.37702811e-01
2.31191888e-01 -2.22048923e-01 5.87985218e-01 5.87177932e-01
-8.61091912e-01 -1.08360618e-01 -1.00334334e+00 3.34036559e-01
-4.74506021e-01 -5.09517677e-02 3.92588288e-01 7.55897164e-01
-9.43203151e-01 8.71495783e-01 -1.13800848e+00 -2.55065650e-01
5.90057373e-01 5.77500761e-01 -2.36114934e-01 1.64810315e-01
-1.05560672e+00 8.49537373e-01 1.33841738e-01 5.71800947e-01
-1.13810647e+00 -7.48529196e-01 -4.89069819e-01 1.15394786e-01
-1.03307448e-01 -1.00056970e+00 6.12193406e-01 -4.61028397e-01
-1.75507081e+00 4.02642727e-01 -2.83060104e-01 -3.34554404e-01
5.96143246e-01 7.29884803e-02 -7.61181787e-02 -2.71340758e-02
-4.03186709e-01 3.58194038e-02 9.57589388e-01 -1.21199727e+00
1.23913676e-01 -8.58036518e-01 -3.16912264e-01 -9.29741561e-02
-8.55886713e-02 -2.19102278e-01 3.52662712e-01 -3.18146020e-01
-8.89177248e-03 -8.28059673e-01 -6.64343596e-01 -2.67841905e-01
-6.07203901e-01 -1.26865983e-01 6.00901127e-01 -4.00342494e-01
9.44223225e-01 -1.75739598e+00 9.12611723e-01 6.34055257e-01
5.03786504e-01 5.59733398e-02 1.25296205e-01 9.19637680e-01
-7.93883875e-02 -1.48903450e-03 -4.49561983e-01 -1.82706993e-02
7.97710270e-02 1.01444833e-01 -5.12033939e-01 8.27809930e-01
2.24258035e-01 6.27043128e-01 -8.25469851e-01 -8.64766911e-02
3.23303878e-01 1.04141033e+00 -5.21341264e-01 -8.93628672e-02
-5.39567947e-01 1.16401899e+00 -1.00933897e+00 2.56998539e-01
6.06277823e-01 -2.68152833e-01 5.26168764e-01 -2.93493211e-01
-1.31125033e-01 1.81107651e-02 -1.10203755e+00 1.68417251e+00
-7.85048753e-02 2.12605795e-04 5.26580587e-02 -1.15270650e+00
7.06684947e-01 3.72577012e-01 7.63251364e-01 1.82897858e-02
2.78169394e-01 1.23520002e-01 1.74288943e-01 -1.80590063e-01
-7.24826455e-02 -4.23933506e-01 1.20160639e-01 4.47154701e-01
1.21127978e-01 -3.29994527e-03 -4.68645170e-02 8.56719017e-02
1.06285858e+00 2.52145618e-01 2.29227930e-01 -7.53813326e-01
6.38203323e-01 -1.41126633e-01 3.68412554e-01 7.06859052e-01
-3.06581229e-01 2.71079004e-01 6.17216885e-01 -2.31915236e-01
-1.04126918e+00 -1.34675646e+00 -6.21541977e-01 4.91647899e-01
-1.81485564e-01 -1.56369671e-01 -9.34922993e-01 -5.69561362e-01
-9.36277509e-02 6.02932513e-01 -8.86374891e-01 -4.32199925e-01
-2.89742053e-01 -1.24976182e+00 2.74313956e-01 6.99533224e-02
-1.22242585e-01 -8.00076425e-01 -3.60006034e-01 4.68965858e-01
3.69746357e-01 -6.33745849e-01 -1.09227166e-01 3.77681702e-01
-8.68828118e-01 -1.05115199e+00 -7.08860695e-01 -1.46414801e-01
6.10088825e-01 -4.18877661e-01 8.48344564e-01 -6.03759229e-01
-2.68447578e-01 5.17553806e-01 1.82683423e-01 -2.81993359e-01
-5.96396923e-01 3.00942045e-02 2.45139509e-01 2.55367070e-01
4.18447763e-01 -9.10884619e-01 -7.97139108e-01 1.09877087e-01
-1.00073099e+00 -3.63494009e-01 3.66632074e-01 6.71319604e-01
7.83951700e-01 2.49727145e-01 7.55973876e-01 -6.19235694e-01
7.38106608e-01 -5.33684194e-01 -5.63789904e-01 3.26253951e-01
-3.37713718e-01 2.40923062e-01 4.41700250e-01 -3.54918927e-01
-1.01005781e+00 2.03807816e-01 -2.37312675e-01 4.50661918e-03
-2.97959149e-01 5.26152015e-01 -6.44717813e-01 -2.38355454e-02
6.57257080e-01 5.45628369e-01 2.91455656e-01 -2.66903996e-01
5.45084655e-01 4.30672288e-01 1.19287640e-01 -9.27483916e-01
4.37536120e-01 8.29735756e-01 3.91164303e-01 -1.08053899e+00
-5.82319558e-01 -1.54508399e-02 -7.83105135e-01 -1.79791316e-01
1.02491796e+00 -4.99435335e-01 -1.03613687e+00 5.09366393e-01
-1.08649981e+00 -2.28747815e-01 -3.92457843e-01 4.84796137e-01
-8.39698672e-01 3.72764885e-01 -5.25772989e-01 -9.24822807e-01
-2.15777233e-01 -1.34992027e+00 9.88779902e-01 1.84880868e-01
-3.58514607e-01 -1.28376663e+00 5.03316998e-01 -8.77113119e-02
5.85694790e-01 4.54648435e-01 9.64548051e-01 -5.97037673e-01
-6.75759375e-01 -1.26712397e-01 3.46876234e-01 3.32686007e-01
2.59986490e-01 2.13491037e-01 -8.13247859e-01 -2.08994120e-01
3.65627944e-01 1.90580010e-01 8.26621294e-01 8.68141890e-01
9.09203470e-01 4.96729389e-02 -4.82716113e-01 2.04362392e-01
1.36562824e+00 3.27469170e-01 5.19669116e-01 -2.00010061e-01
4.27788585e-01 6.83311641e-01 -1.72291305e-02 6.09424531e-01
6.23343959e-02 6.46455526e-01 1.86805621e-01 2.67769933e-01
2.23882288e-01 -7.50981867e-02 1.81528568e-01 5.08912563e-01
-8.31090286e-02 -2.76261866e-01 -7.47551501e-01 2.63962150e-01
-1.89513958e+00 -1.17487776e+00 -1.03334144e-01 2.24773860e+00
1.09070504e+00 -7.37216920e-02 -2.59560999e-02 -7.58000910e-02
7.01869190e-01 -2.37471983e-03 -7.48211980e-01 -1.27885461e-01
-1.95484698e-01 4.14477319e-01 3.02903384e-01 6.45536602e-01
-8.57115507e-01 5.44439316e-01 6.96528482e+00 9.18816447e-01
-1.09066045e+00 4.99519743e-02 7.51162171e-01 -1.31766160e-03
-4.50591773e-01 -1.06779687e-01 -9.71738100e-01 5.53728759e-01
1.09008372e+00 -1.47687510e-01 3.02767426e-01 2.41150618e-01
7.50488877e-01 7.15700808e-05 -1.06989264e+00 4.60092664e-01
-3.32365125e-01 -1.49986947e+00 -5.10444604e-02 5.75131416e-01
7.51316845e-01 -1.06564179e-01 3.03958327e-01 -3.14910591e-01
4.81005549e-01 -1.13135874e+00 4.25895035e-01 1.16715682e+00
4.99580950e-01 -7.14851499e-01 3.39050800e-01 4.52075541e-01
-5.82249463e-01 2.62299836e-01 -3.59825641e-01 2.35429332e-01
2.96866715e-01 1.32033992e+00 -4.78110552e-01 2.67464459e-01
1.88435897e-01 7.21236110e-01 2.06377357e-01 8.71907353e-01
-5.51978126e-02 5.08735120e-01 -3.79398197e-01 -3.65824439e-02
-1.35096125e-02 -8.26737583e-01 6.96379185e-01 1.03917265e+00
2.41165131e-01 -2.65329033e-02 -2.79180944e-01 1.34377015e+00
2.56930441e-01 -1.52585611e-01 -7.89314687e-01 -3.44185919e-01
2.88947374e-01 1.17224491e+00 -6.62728846e-01 -6.16120286e-02
1.26656860e-01 6.19976223e-01 1.27353728e-01 7.06981659e-01
-7.83849180e-01 -2.20837593e-02 9.47642386e-01 9.68059823e-02
3.88665318e-01 -6.39394939e-01 -4.14673805e-01 -1.03398812e+00
-2.54679024e-01 -6.94905341e-01 5.11088036e-02 -2.13911235e-01
-1.41545773e+00 4.09493476e-01 -1.26063943e-01 -7.77657986e-01
1.00685604e-01 -7.32221127e-01 -4.95549709e-01 1.11588395e+00
-1.16988671e+00 -8.88880908e-01 3.23306620e-01 4.14367139e-01
1.52921682e-04 6.44180402e-02 8.55528355e-01 9.15054977e-02
-7.08393455e-01 -1.99112613e-02 7.41963387e-01 -5.52813411e-01
2.90299445e-01 -1.28224564e+00 -5.18234260e-02 5.70551932e-01
6.28222898e-02 1.06175756e+00 1.10232472e+00 -6.80453837e-01
-1.45430088e+00 -9.58187759e-01 7.71058798e-01 -8.20780277e-01
7.82743275e-01 -4.25887674e-01 -9.12012219e-01 1.81109563e-01
6.84945136e-02 -8.24757069e-02 9.83152866e-01 4.94523384e-02
-1.22519642e-01 2.26864312e-02 -1.18867731e+00 5.77497005e-01
8.97357583e-01 -5.97215950e-01 -2.09739849e-01 4.58191216e-01
1.32427171e-01 -1.20195977e-01 -1.31373191e+00 3.31568807e-01
5.60373306e-01 -7.98841119e-01 1.06078517e+00 -1.06678426e+00
1.65460765e-01 -5.47913790e-01 -2.45412752e-01 -1.41241884e+00
-2.81139791e-01 -7.26297319e-01 -1.64637253e-01 8.63183618e-01
4.57327127e-01 -5.82989812e-01 9.09850955e-01 7.64602661e-01
1.20673418e-01 -8.02609861e-01 -1.17161083e+00 -5.04500747e-01
5.21202207e-01 -2.81921059e-01 3.26820225e-01 5.63956499e-01
5.16095087e-02 2.53721714e-01 -1.17814079e-01 1.50866553e-01
9.68626797e-01 -1.17398806e-01 3.82921815e-01 -1.23268366e+00
-3.03080291e-01 -4.45575982e-01 -3.61335069e-01 -6.76102698e-01
1.82764024e-01 -6.12082541e-01 -1.05443828e-01 -1.42575657e+00
1.80594906e-01 -3.63591194e-01 -5.44698894e-01 -2.13667303e-01
1.66921139e-01 -2.80123837e-02 -3.05090547e-01 1.90897301e-01
-5.63335776e-01 9.74582493e-01 1.17859578e+00 -2.83656307e-02
-8.77648667e-02 2.58201599e-01 -7.81703770e-01 4.55272764e-01
8.11899781e-01 -4.56284374e-01 -5.30668259e-01 6.78863153e-02
3.16369981e-01 6.75714463e-02 4.66287643e-01 -8.15997064e-01
2.52601057e-01 -3.75737756e-01 3.40117455e-01 -4.44701016e-01
4.08550680e-01 -8.38624835e-01 5.88251531e-01 5.29702961e-01
-3.81915599e-01 -3.39856327e-01 -7.59689212e-02 1.02090740e+00
1.11070469e-01 5.74234836e-02 9.30410028e-01 -3.14056408e-03
-9.73548591e-02 3.54165554e-01 -6.88649952e-01 -1.38018250e-01
1.14597917e+00 -2.02283770e-01 -1.40314579e-01 -4.51475605e-02
-1.43934512e+00 -3.97865415e-01 4.27168787e-01 -1.09689064e-01
4.52055156e-01 -1.06391418e+00 -6.70781076e-01 1.12022720e-01
-2.17931792e-01 -2.41935551e-01 5.74702442e-01 1.12920880e+00
-2.05693468e-01 5.60324073e-01 -4.11076434e-02 -8.25895369e-01
-7.44110644e-01 5.01078665e-01 7.91005492e-01 -2.06226230e-01
-2.72550166e-01 5.53399503e-01 5.12527645e-01 -1.59157842e-01
-9.52020064e-02 -1.55214012e-01 -9.39345658e-02 -1.72197804e-01
4.29983109e-01 2.61606365e-01 2.03795105e-01 -5.55844724e-01
-4.95391756e-01 4.32399392e-01 -8.00014064e-02 -4.21035886e-02
1.45617533e+00 -8.38207528e-02 -2.53124148e-01 2.28396103e-01
8.64410996e-01 5.09424582e-02 -1.66567278e+00 -6.45279810e-02
1.30076617e-01 4.62662615e-02 2.19506502e-01 -7.74621367e-01
-6.63732290e-01 7.76302457e-01 5.40928245e-01 1.33777052e-01
6.77739382e-01 6.61803335e-02 1.71788886e-01 1.44715279e-01
4.45439041e-01 -8.19077253e-01 -2.42808536e-01 2.81889856e-01
8.24231088e-01 -6.67432249e-01 -8.80710110e-02 -2.27128625e-01
-2.35286668e-01 1.00160396e+00 3.82479350e-03 -3.13156664e-01
1.02620101e+00 3.22760582e-01 -2.63860226e-01 -4.64018047e-01
-8.23432386e-01 -2.61235237e-02 3.10957342e-01 8.43968630e-01
5.89747071e-01 5.81176430e-02 -4.60582078e-01 6.33308589e-01
4.36503351e-01 -2.45285090e-02 2.58541107e-01 8.38595331e-01
-5.05434990e-01 -1.37795532e+00 -2.06180215e-01 1.95326060e-01
-3.95573616e-01 -2.97212545e-02 -2.88312018e-01 2.68617868e-01
1.67973544e-02 7.98358023e-01 -3.53555888e-01 2.41899658e-02
1.07940979e-01 1.06299274e-01 7.65187979e-01 -4.78771478e-01
-2.16646984e-01 2.22300872e-01 -1.98114529e-01 -3.41083020e-01
-4.08216625e-01 -9.34408963e-01 -9.97163296e-01 -2.44338274e-01
-2.56685793e-01 4.34347838e-01 8.25446665e-01 1.08004177e+00
7.76351869e-01 5.44747233e-01 3.84853065e-01 -8.23055744e-01
-8.08272898e-01 -6.92687988e-01 -8.46456885e-01 -5.75045124e-02
2.59704143e-01 -6.96878076e-01 -2.37582073e-01 1.26687646e-01] | [5.315295696258545, 5.085062503814697] |
b9d7124b-ad5e-4086-8a87-d572b226434e | linearly-scalable-learning-of-smooth-low | 2306.10287 | null | https://arxiv.org/abs/2306.10287v1 | https://arxiv.org/pdf/2306.10287v1.pdf | Linearly-scalable learning of smooth low-dimensional patterns with permutation-aided entropic dimension reduction | In many data science applications, the objective is to extract appropriately-ordered smooth low-dimensional data patterns from high-dimensional data sets. This is challenging since common sorting algorithms are primarily aiming at finding monotonic orderings in low-dimensional data, whereas typical dimension reduction and feature extraction algorithms are not primarily designed for extracting smooth low-dimensional data patterns. We show that when selecting the Euclidean smoothness as a pattern quality criterium, both of these problems (finding the optimal 'crisp' data permutation and extracting the sparse set of permuted low-dimensional smooth patterns) can be efficiently solved numerically as one unsupervised entropy-regularized iterative optimization problem. We formulate and prove the conditions for monotonicity and convergence of this linearly-scalable (in dimension) numerical procedure, with the iteration cost scaling of $\mathcal{O}(DT^2)$, where $T$ is the size of the data statistics and $D$ is a feature space dimension. The efficacy of the proposed method is demonstrated through the examination of synthetic examples as well as a real-world application involving the identification of smooth bankruptcy risk minimizing transition patterns from high-dimensional economical data. The results showcase that the statistical properties of the overall time complexity of the method exhibit linear scaling in the dimensionality $D$ within the specified confidence intervals. | ['Lukas Pospisil', 'Illia Horenko'] | 2023-06-17 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [ 2.77639866e-01 -8.11529234e-02 -1.31219149e-01 -3.21808726e-01
-6.03385389e-01 -4.22560394e-01 1.15806304e-01 3.29767793e-01
-4.51296240e-01 5.08514881e-01 -8.68053585e-02 -3.39278251e-01
-1.07925558e+00 -7.25598752e-01 -4.36911017e-01 -9.64279294e-01
-6.94340885e-01 6.42479956e-01 -2.73915589e-01 1.88146666e-01
8.15384388e-01 7.01452255e-01 -2.01430154e+00 -1.14136048e-01
1.06941104e+00 1.43061852e+00 -1.12985685e-01 4.18651432e-01
-1.92739535e-02 9.34432074e-02 -1.33412123e-01 -3.69846411e-02
6.77989781e-01 -2.38655373e-01 -6.81840599e-01 6.52949035e-01
5.29437466e-03 1.08212031e-01 1.79993957e-01 1.12502015e+00
3.70048940e-01 3.68419647e-01 7.78411090e-01 -1.02268171e+00
-1.46846488e-01 9.50776041e-03 -5.42110085e-01 2.60460258e-01
1.77209437e-01 -9.05675888e-02 9.00143564e-01 -1.19242918e+00
5.50065637e-01 8.16594839e-01 4.59732622e-01 -7.77844191e-02
-1.42261755e+00 -2.86440879e-01 -1.80492535e-01 -5.96399046e-02
-1.43390906e+00 -3.20269436e-01 8.26871514e-01 -6.03556633e-01
8.99668455e-01 6.11113012e-01 5.07802784e-01 2.19425231e-01
2.46028811e-01 3.45834881e-01 1.01154900e+00 -4.27427739e-01
6.59681976e-01 4.78185788e-02 5.47082961e-01 5.66101551e-01
8.01594615e-01 1.93542883e-01 -4.36996281e-01 -2.83024758e-01
3.44599009e-01 2.30019737e-06 -3.81255858e-02 -6.82113767e-01
-1.08572555e+00 8.82165015e-01 -1.98232189e-01 3.61887455e-01
-6.05328321e-01 -5.27163684e-01 3.13084155e-01 4.91463691e-01
5.18118083e-01 6.89468324e-01 -5.02996266e-01 -3.66670489e-01
-8.21413338e-01 2.79794872e-01 8.59188676e-01 9.41599667e-01
5.59136868e-01 -4.00627591e-02 2.18956575e-01 6.77679837e-01
-6.02039099e-02 4.25858021e-01 5.37659407e-01 -7.50453055e-01
5.62859595e-01 1.00083458e+00 2.98144579e-01 -1.41212857e+00
-6.93508089e-01 -2.42524266e-01 -1.15874946e+00 2.23549698e-02
6.12964630e-01 -1.55992776e-01 -5.88794827e-01 1.28864241e+00
5.00798643e-01 -3.57891649e-01 5.99793084e-02 7.50881076e-01
-6.69657392e-03 5.14030278e-01 -2.89315820e-01 -9.47238564e-01
1.24785256e+00 7.25714937e-02 -6.18545353e-01 -1.09155932e-02
6.80878937e-01 -5.29119372e-01 1.24067509e+00 5.68716764e-01
-1.18686283e+00 -2.40204901e-01 -1.00479293e+00 3.28419298e-01
-2.06063583e-01 1.96169615e-01 6.15160108e-01 6.43462300e-01
-3.64504158e-01 5.87567031e-01 -5.64939559e-01 5.21276295e-02
2.92845607e-01 6.59461319e-01 -4.35746402e-01 8.50845799e-02
-8.24046552e-01 4.21251923e-01 4.15658265e-01 4.17270958e-01
4.39578295e-02 -7.48810589e-01 -6.18445158e-01 1.33620068e-01
2.53588825e-01 -1.37672782e-01 4.38788861e-01 -5.32745481e-01
-1.07565117e+00 8.73371005e-01 -1.77601963e-01 -2.99258083e-01
2.37480611e-01 5.71667738e-02 -5.13310611e-01 1.02814317e-01
6.83195964e-02 -1.59653232e-01 8.89645100e-01 -8.57787251e-01
-7.15271115e-01 -6.82359517e-01 -7.53166556e-01 8.15153345e-02
-4.64870274e-01 -2.35290125e-01 -1.00020811e-01 -7.41766036e-01
6.60253048e-01 -8.63011122e-01 -4.29686397e-01 -3.21813881e-01
-2.79484838e-01 -5.25052696e-02 5.70667922e-01 -5.37130952e-01
1.68401098e+00 -2.45024347e+00 1.71917990e-01 1.07458937e+00
-6.08711094e-02 -5.03682494e-02 3.03988248e-01 2.83845812e-01
-3.02755713e-01 3.75091024e-02 -6.91210151e-01 1.66442811e-01
1.47605881e-01 5.83088249e-02 -2.25052193e-01 9.02840257e-01
2.05155894e-01 2.09853604e-01 -5.77512205e-01 -2.85491586e-01
-3.94853577e-03 -1.07601039e-01 -7.07040250e-01 1.04806624e-01
1.73438728e-01 -3.00684702e-02 -4.84624714e-01 6.03153348e-01
7.18306780e-01 -2.68515259e-01 3.42880309e-01 -9.68463272e-02
-2.84315586e-01 -4.19765748e-02 -1.89770877e+00 1.25778842e+00
-9.61298123e-02 2.51784384e-01 -4.20731045e-02 -1.35960484e+00
1.21828973e+00 -2.10133288e-03 9.49401617e-01 -8.24670851e-01
1.64262161e-01 4.68450963e-01 -8.81682523e-03 -5.61371565e-01
5.71938932e-01 -4.39601839e-01 -5.09090066e-01 4.80298489e-01
-4.14889961e-01 9.62238386e-03 5.30222893e-01 -3.68958026e-01
1.03561532e+00 -5.66178560e-01 4.09025609e-01 -7.68021941e-01
4.64330733e-01 2.81461626e-01 7.52385139e-01 2.93447733e-01
2.37852231e-01 3.55329514e-01 9.19480026e-01 -4.55588818e-01
-1.22843921e+00 -8.05244684e-01 -5.68804860e-01 3.69379282e-01
1.86696768e-01 -9.64045972e-02 -5.59024453e-01 -3.70022714e-01
2.73687392e-01 7.49908090e-01 -5.43662608e-01 -4.91054654e-02
-6.59293294e-01 -1.24617326e+00 -1.95109949e-01 1.35461926e-01
1.38552681e-01 -8.15262437e-01 -8.65939140e-01 1.82307005e-01
3.04404914e-01 -7.74330974e-01 -3.91704679e-01 5.69443762e-01
-1.06814802e+00 -1.33169186e+00 -2.47988269e-01 -8.68380904e-01
1.24196351e+00 -1.21849641e-01 8.32313538e-01 -1.81348443e-01
-5.20286083e-01 1.56952038e-01 -1.83534682e-01 -9.27003995e-02
-6.59190193e-02 -2.30229441e-02 4.01071817e-01 1.76387608e-01
6.07228816e-01 -4.61906344e-01 -4.97289717e-01 5.25859058e-01
-1.01650560e+00 -3.19576263e-01 6.08753562e-01 9.92255569e-01
9.68843579e-01 6.86365604e-01 6.96331859e-01 -5.83517134e-01
9.78011250e-01 -4.38572973e-01 -9.41043377e-01 6.76726177e-02
-9.71928000e-01 4.30463344e-01 7.96984017e-01 -3.75919968e-01
-5.51670432e-01 3.19170743e-01 3.63496840e-01 -1.14511997e-01
-8.75876471e-02 5.56344926e-01 -1.95388585e-01 1.77678064e-01
3.57685834e-01 4.56897169e-01 3.00398737e-01 -5.03895938e-01
1.18791219e-02 6.49319291e-01 3.73296618e-01 -6.11370862e-01
6.56135678e-01 2.70591497e-01 5.17208695e-01 -9.56699252e-01
-2.81254560e-01 -5.82277238e-01 -6.90189540e-01 8.32964182e-02
5.30122757e-01 -2.46819258e-01 -1.15632546e+00 1.42684653e-01
-3.74325186e-01 2.28323936e-01 -7.78608978e-01 4.99139547e-01
-7.82489181e-01 5.08518338e-01 -2.00698152e-02 -9.43547547e-01
-2.18950137e-01 -9.46202338e-01 6.61316037e-01 -1.94917738e-01
-3.97158712e-01 -7.56053507e-01 1.14686780e-01 -2.14419469e-01
-7.43578449e-02 5.13446748e-01 1.37101889e+00 -7.29853630e-01
-1.98599935e-01 -6.79330409e-01 7.02512711e-02 2.23855913e-01
3.91359299e-01 -2.31574073e-01 -1.48736373e-01 -3.26684505e-01
4.72240865e-01 7.52481371e-02 2.92779416e-01 5.55091739e-01
1.12414777e+00 -6.97303057e-01 -9.96385515e-02 5.14468789e-01
1.30148232e+00 4.45758522e-01 3.19046080e-01 2.43270501e-01
1.84912652e-01 8.87369454e-01 1.06194139e+00 1.06167948e+00
-1.67402059e-01 4.99662220e-01 1.29206240e-01 -2.11068541e-02
6.19686484e-01 2.50720352e-01 3.96912098e-02 8.23231697e-01
1.15801424e-01 1.28943875e-01 -9.85333979e-01 5.80569088e-01
-1.68645906e+00 -1.02594554e+00 -1.01941630e-01 2.61324239e+00
7.77487636e-01 2.27843687e-01 4.04225886e-01 7.84687281e-01
7.14927793e-01 -2.35955745e-01 -6.55796647e-01 -7.26792395e-01
-1.55421406e-01 2.43414044e-01 6.29642189e-01 5.02531350e-01
-1.14205492e+00 1.13756284e-01 5.94221258e+00 6.33116663e-01
-8.34048092e-01 -8.26767683e-01 7.80909359e-01 -3.50837439e-01
-2.22362965e-01 -2.44452387e-01 -7.19391644e-01 7.48272300e-01
9.72305298e-01 -4.03056622e-01 3.45475227e-01 8.58478010e-01
4.62034464e-01 -2.89762080e-01 -1.10170102e+00 1.12811816e+00
-3.09404671e-01 -1.23424053e+00 -1.29925638e-01 4.31820929e-01
6.08578622e-01 -7.95779228e-01 2.64307201e-01 -2.22963333e-01
-1.91068694e-01 -9.44492936e-01 4.12834764e-01 5.23687184e-01
7.02176988e-01 -1.23091352e+00 4.15403068e-01 4.00879413e-01
-1.05014896e+00 -5.44963777e-01 -2.48339340e-01 3.37205008e-02
1.55052304e-01 1.13592529e+00 -6.72305703e-01 4.81675714e-01
6.42485619e-01 6.02974296e-01 -1.48666784e-01 9.66393888e-01
6.95233226e-01 1.13019660e-01 -7.14265943e-01 -1.12824358e-01
1.51018441e-01 -6.65800631e-01 4.61416006e-01 9.49988186e-01
6.90000236e-01 4.57117409e-01 -1.10988319e-01 4.69801962e-01
2.62668550e-01 3.50710541e-01 -5.51859438e-01 -1.03814252e-01
3.83332461e-01 6.11986160e-01 -8.18973541e-01 3.61030246e-03
-2.00257033e-01 4.39945966e-01 -1.34924829e-01 2.12162342e-02
-2.88873166e-01 -7.32624710e-01 7.55452693e-01 3.83226037e-01
5.11981189e-01 -4.17557657e-01 -9.84944642e-01 -8.24524641e-01
5.16948819e-01 -9.52332616e-01 6.05730057e-01 1.96510404e-01
-1.26469433e+00 3.90703946e-01 -2.76139658e-02 -1.78013837e+00
-4.62473273e-01 -6.80550098e-01 -8.75111222e-02 7.05434978e-01
-8.56428206e-01 -6.13835230e-02 1.82358801e-01 6.01468980e-01
3.14018369e-01 -3.65288049e-01 7.84079552e-01 3.40821445e-01
-7.36378968e-01 4.20608938e-01 6.41973376e-01 -2.35772833e-01
4.33393382e-02 -1.12021601e+00 1.25527298e-02 1.00163794e+00
-3.96354049e-01 6.03099585e-01 9.31600451e-01 -6.24785364e-01
-1.71403253e+00 -6.66397929e-01 8.95565867e-01 8.87390897e-02
5.92608333e-01 -3.62528533e-01 -9.39284563e-01 4.23345789e-02
-5.16127050e-01 6.82448149e-02 7.87587762e-01 -1.80213585e-01
3.22970957e-01 -2.75586039e-01 -1.51931202e+00 4.57501858e-01
8.37138653e-01 -3.22466367e-03 -4.50745702e-01 2.48692140e-01
1.04581363e-01 -1.35923639e-01 -1.40182674e+00 4.99398768e-01
3.88937682e-01 -8.65214586e-01 9.52460885e-01 -7.79736578e-01
1.55924305e-01 -2.02475354e-01 -2.12446764e-01 -8.32204878e-01
-1.48218364e-01 -7.21252799e-01 -9.85814705e-02 7.97386050e-01
3.52429271e-01 -5.52976310e-01 9.97497916e-01 1.03824151e+00
1.57647073e-01 -1.01758599e+00 -1.27205110e+00 -8.11951697e-01
-2.30591238e-01 -3.38448733e-01 5.35139322e-01 8.44643116e-01
3.61756623e-01 -8.61363187e-02 -2.06863001e-01 2.31419414e-01
8.20229828e-01 5.12702703e-01 5.80716193e-01 -1.42399621e+00
-1.15086928e-01 -4.76404190e-01 -4.09876525e-01 -7.78923333e-01
-1.32478863e-01 -6.68515563e-01 5.11298236e-03 -9.18084443e-01
-2.68893093e-01 -7.76569426e-01 -1.78912118e-01 1.84967220e-02
1.23528883e-01 -5.31583011e-01 -1.94824100e-01 2.00925216e-01
-1.01399273e-01 5.61876953e-01 9.06573296e-01 1.23174176e-01
-6.36036277e-01 2.37510115e-01 -5.95999599e-01 5.61009586e-01
5.41198432e-01 -5.43189526e-01 -5.85638344e-01 1.95702538e-01
3.31604570e-01 2.53836781e-01 -1.17644727e-01 -7.23573029e-01
-3.86558706e-03 -4.80863661e-01 2.87779689e-01 -7.57768512e-01
1.47028431e-01 -1.16155827e+00 2.76870966e-01 5.57883024e-01
-2.93037713e-01 2.27491662e-01 1.22880124e-01 5.28967381e-01
-2.48782337e-01 -1.37528047e-01 8.01971078e-01 3.60806048e-01
-6.13611519e-01 1.54404178e-01 -2.14911029e-01 -4.41279858e-02
1.29129255e+00 -5.54080129e-01 9.56088006e-02 1.35853767e-01
-8.38403523e-01 8.54076594e-02 2.92589873e-01 9.08676088e-02
7.48968542e-01 -1.31781840e+00 -4.17407602e-01 7.11967945e-01
-1.84210569e-01 2.35874847e-01 1.96036652e-01 9.53562379e-01
-4.24197108e-01 5.87131739e-01 -1.19197726e-01 -6.72321141e-01
-1.04928303e+00 7.51952887e-01 6.99009746e-02 -3.67384136e-01
-7.17977405e-01 4.87909019e-01 -2.81162739e-01 4.27077105e-03
1.93387359e-01 -6.68487370e-01 4.00524959e-02 3.69154662e-01
3.23152393e-01 7.01559424e-01 3.67098540e-01 -3.13889235e-01
-4.52253699e-01 6.71807706e-01 1.37180328e-01 5.71767762e-02
1.66884959e+00 -1.27968296e-01 -1.87589154e-01 2.25248307e-01
1.42281365e+00 -3.02173138e-01 -1.12085366e+00 3.98498960e-02
5.82325101e-01 -6.41499579e-01 -2.10560873e-01 -1.77247897e-01
-8.99596214e-01 4.12998527e-01 6.26357079e-01 6.52837455e-01
1.42209804e+00 -3.05312812e-01 4.62936074e-01 4.68276173e-01
2.98876673e-01 -1.46104991e+00 -1.23204723e-01 2.72786289e-01
8.79950881e-01 -8.05119336e-01 1.66525632e-01 -3.83524746e-01
-6.30259097e-01 1.02034461e+00 6.55971691e-02 -4.43378866e-01
9.96930897e-01 1.49463147e-01 -6.14886761e-01 -2.61804074e-01
-5.52088678e-01 1.65527716e-01 3.76150608e-01 8.01842436e-02
4.45913374e-02 5.21473140e-02 -8.09909582e-01 7.50204921e-01
-4.31636274e-01 -1.57167390e-01 3.69387507e-01 1.08407974e+00
-4.98325348e-01 -9.03833210e-01 -5.18808007e-01 9.21142459e-01
-3.70229483e-01 1.93264708e-01 2.34536938e-02 8.24827313e-01
-1.60724655e-01 7.07991064e-01 4.27252382e-01 -1.45953789e-01
6.73844337e-01 1.39431581e-01 9.39013883e-02 -2.41765618e-01
-2.13187844e-01 1.36262160e-02 5.03496788e-02 -5.94946444e-01
-1.82157665e-01 -1.10183716e+00 -1.33694375e+00 -3.78763407e-01
-3.24863717e-02 3.78305644e-01 7.10453808e-01 8.17345500e-01
6.04622722e-01 3.67450044e-02 1.16896796e+00 -5.61860085e-01
-6.38134181e-01 -3.87946844e-01 -1.09782159e+00 4.67969656e-01
3.37650239e-01 -6.75766170e-01 -6.47650421e-01 3.28921862e-02] | [7.584362506866455, 4.297757625579834] |
04bc18cf-10c3-4f7b-aa19-389b4be5e772 | image-reconstruction-with-predictive-filter | 1811.11482 | null | http://arxiv.org/abs/1811.11482v1 | http://arxiv.org/pdf/1811.11482v1.pdf | Image Reconstruction with Predictive Filter Flow | We propose a simple, interpretable framework for solving a wide range of
image reconstruction problems such as denoising and deconvolution. Given a
corrupted input image, the model synthesizes a spatially varying linear filter
which, when applied to the input image, reconstructs the desired output. The
model parameters are learned using supervised or self-supervised training. We
test this model on three tasks: non-uniform motion blur removal,
lossy-compression artifact reduction and single image super resolution. We
demonstrate that our model substantially outperforms state-of-the-art methods
on all these tasks and is significantly faster than optimization-based
approaches to deconvolution. Unlike models that directly predict output pixel
values, the predicted filter flow is controllable and interpretable, which we
demonstrate by visualizing the space of predicted filters for different tasks. | ['Shu Kong', 'Charless Fowlkes'] | 2018-11-28 | null | null | null | null | ['lossy-compression-artifact-reduction'] | ['computer-vision'] | [ 7.44181693e-01 2.62553804e-02 1.05384260e-01 -2.20258534e-01
-6.92169785e-01 -6.46454930e-01 4.34997380e-01 -7.48525321e-01
-1.22112922e-01 8.18229914e-01 6.80283964e-01 -2.01014563e-01
-6.59091398e-02 -2.02885777e-01 -8.17562938e-01 -6.67801142e-01
1.30030438e-01 -7.86637142e-02 -5.54364882e-02 1.57020420e-01
4.19249058e-01 3.52475256e-01 -1.31070054e+00 6.46656692e-01
9.17183638e-01 7.50895083e-01 6.46729648e-01 1.06205642e+00
3.08850020e-01 1.18284905e+00 -3.26679647e-01 1.17729619e-01
3.72125924e-01 -8.79078627e-01 -7.78133452e-01 4.41241235e-01
6.55702293e-01 -7.26251125e-01 -5.43361723e-01 1.20132291e+00
2.59791851e-01 2.56427735e-01 6.12828851e-01 -6.82098925e-01
-1.08521640e+00 2.07543492e-01 -3.80197823e-01 3.91112983e-01
3.38077664e-01 4.73431557e-01 8.30219030e-01 -1.04201901e+00
8.75851154e-01 1.32320905e+00 6.41611278e-01 8.16688001e-01
-2.04683089e+00 -1.88428149e-01 -4.33942117e-02 4.86404672e-02
-7.42719769e-01 -8.52490127e-01 3.96804094e-01 -5.85201919e-01
6.67634189e-01 3.49408418e-01 3.30261439e-01 8.39575589e-01
1.90778166e-01 5.97570658e-01 1.41631603e+00 -2.28361771e-01
1.84041739e-01 -2.17595994e-01 -1.81965396e-01 6.33805931e-01
-5.18041104e-02 3.35308671e-01 -7.70449877e-01 -7.14267269e-02
1.39987314e+00 -1.09765224e-01 -1.01005197e+00 -2.10985485e-02
-1.54970026e+00 3.23838770e-01 2.48693749e-01 -2.67886072e-02
-5.11197150e-01 6.36821032e-01 -1.05214223e-01 4.67639655e-01
7.34653115e-01 7.77850091e-01 -5.11465549e-01 7.05875224e-03
-1.13834250e+00 1.85809538e-01 6.75693750e-01 5.60540915e-01
6.54365122e-01 4.06452596e-01 -2.78711289e-01 6.34348810e-01
7.32444078e-02 3.67849827e-01 4.34827298e-01 -1.96758902e+00
6.87817112e-02 -1.61485702e-01 6.37133539e-01 -6.71981931e-01
-7.52968192e-02 -4.92772818e-01 -8.42074871e-01 9.53759909e-01
5.08280993e-01 -1.15458816e-01 -1.09689009e+00 1.63822472e+00
-1.79506212e-01 7.76451051e-01 6.02360964e-02 1.30937910e+00
6.45536721e-01 8.01741958e-01 -3.10234457e-01 -4.59963620e-01
9.89991844e-01 -1.22349668e+00 -1.08994102e+00 -4.88123477e-01
-1.02002628e-01 -8.71299326e-01 1.03767550e+00 5.43854415e-01
-1.71328366e+00 -5.18799663e-01 -8.89773190e-01 -3.55840802e-01
2.97610700e-01 1.40705526e-01 2.91629732e-01 1.03240393e-01
-1.40453386e+00 1.15383065e+00 -8.88506532e-01 -9.71944723e-03
5.90846300e-01 6.33027181e-02 -3.54208291e-01 -1.83874771e-01
-5.39311528e-01 8.80975127e-01 -2.20625877e-01 8.84756297e-02
-1.29548717e+00 -1.06978333e+00 -7.15523899e-01 1.64077148e-01
-6.62857620e-03 -1.17421818e+00 1.37648499e+00 -1.21630216e+00
-1.62228370e+00 9.23238993e-01 -7.68883169e-01 -5.43994188e-01
6.76127195e-01 -4.32290286e-01 -2.05083117e-01 3.77519757e-01
5.78320026e-02 5.06401658e-01 1.53017056e+00 -1.67182493e+00
-3.76071900e-01 1.58034470e-02 -2.43209839e-01 2.84424543e-01
7.55853578e-02 2.99913269e-02 -2.08253339e-01 -9.49492812e-01
3.22464794e-01 -5.86385727e-01 -4.14737046e-01 5.74794471e-01
-4.09875751e-01 7.76601672e-01 8.83081496e-01 -1.06245482e+00
1.01234722e+00 -2.05157971e+00 6.93852365e-01 -3.34986717e-01
4.84731525e-01 6.66908100e-02 -3.07434410e-01 -3.59814265e-03
-2.11772606e-01 8.36641937e-02 -6.24298573e-01 -6.69812918e-01
-3.12916875e-01 1.37099862e-01 -6.92250609e-01 6.32708430e-01
2.42999196e-01 9.03337777e-01 -9.73967850e-01 -5.48403710e-03
3.69807363e-01 6.25002563e-01 -6.21676922e-01 5.42114019e-01
-2.71182209e-01 1.17265904e+00 -3.77854370e-02 3.46025109e-01
6.46084666e-01 -5.10627866e-01 5.59642613e-02 -4.60944265e-01
-1.44812673e-01 8.13485160e-02 -1.09492445e+00 1.71146882e+00
-5.08799732e-01 1.19807446e+00 5.24143219e-01 -8.11761796e-01
4.99287158e-01 4.46069002e-01 2.10649282e-01 -4.26453322e-01
-2.89068431e-01 1.35524020e-01 -3.02748948e-01 -8.55970681e-01
2.58438975e-01 -2.33992785e-01 6.58649981e-01 6.67568505e-01
6.09340593e-02 -2.59947091e-01 -1.11161910e-01 1.14608720e-01
1.27548492e+00 4.63127673e-01 8.74092504e-02 -3.56472224e-01
4.17716205e-01 -4.96896654e-02 3.73009533e-01 9.56521451e-01
-4.81538922e-02 1.36273932e+00 3.24933261e-01 -4.43567008e-01
-1.23639321e+00 -1.28369582e+00 5.95874302e-02 8.06615591e-01
2.41119340e-01 -2.00622708e-01 -8.74076307e-01 -1.19917817e-01
-3.20181474e-02 6.22831225e-01 -5.57547152e-01 1.48259416e-01
-6.61304116e-01 -4.84076649e-01 1.35201871e-01 3.34202081e-01
5.89213848e-01 -1.00161123e+00 -6.38663054e-01 1.86874419e-01
-4.59553272e-01 -1.23307204e+00 -8.13881934e-01 -7.28390440e-02
-1.28608632e+00 -9.51437712e-01 -7.83432662e-01 -8.44672620e-01
9.65948701e-01 6.41674757e-01 1.27516484e+00 1.75885633e-01
-3.46910685e-01 5.26710868e-01 2.73491770e-01 2.49550700e-01
-5.65827549e-01 -7.20471919e-01 -2.35702708e-01 2.00714171e-01
-5.13210654e-01 -7.30315864e-01 -9.56375778e-01 1.95717290e-01
-1.01247346e+00 4.18428242e-01 2.83562005e-01 1.12891531e+00
8.19967210e-01 2.48768833e-02 1.91756889e-01 -9.55741048e-01
8.69123340e-01 -2.09561408e-01 -4.29350168e-01 7.67154917e-02
-5.16805351e-01 3.74580920e-01 7.93532491e-01 -6.18342280e-01
-1.50520718e+00 7.12591708e-02 3.82089198e-01 -6.51056826e-01
-1.11279085e-01 2.73110062e-01 2.71329880e-01 -1.50200948e-01
9.87848461e-01 3.74567479e-01 2.57990658e-01 -6.98715985e-01
5.38011253e-01 1.87269032e-01 1.33219051e+00 -2.33938709e-01
9.60271299e-01 9.47215140e-01 3.56355868e-02 -6.96693003e-01
-9.01000857e-01 -1.47376299e-01 -5.01109779e-01 -1.99919328e-01
9.99114215e-01 -1.07225502e+00 -5.24822712e-01 5.22228003e-01
-1.46854043e+00 -8.09920490e-01 -5.05051315e-01 3.61338317e-01
-8.82745922e-01 2.26666212e-01 -8.32409620e-01 -6.83254421e-01
-1.05543651e-01 -1.06793857e+00 1.00562346e+00 1.53560475e-01
-1.33450076e-01 -1.33114159e+00 -2.02189147e-01 3.20889920e-01
6.80280805e-01 2.18342528e-01 6.49773777e-01 3.92382175e-01
-1.05005002e+00 4.35325950e-01 -3.94369543e-01 7.20265925e-01
2.56682634e-01 -3.08859348e-01 -1.10255265e+00 -2.64580578e-01
4.60829020e-01 -1.10639252e-01 1.28640342e+00 9.25867975e-01
1.37859786e+00 -5.32167614e-01 -4.08341102e-02 1.14411306e+00
1.47339630e+00 -2.01474652e-01 8.95361245e-01 2.76517391e-01
5.52967906e-01 4.11515146e-01 1.52032450e-01 2.27905676e-01
-2.65407078e-02 3.28391314e-01 4.61848766e-01 -4.60275024e-01
-7.35399485e-01 -1.31836578e-01 3.58071923e-01 2.43832856e-01
-3.45806956e-01 -1.43143572e-02 -4.13912833e-01 5.62848151e-01
-1.95592690e+00 -1.16006267e+00 -1.93273395e-01 1.98862171e+00
9.77027833e-01 -1.54947624e-01 -3.95356894e-01 -2.60369271e-01
6.68248296e-01 3.14470500e-01 -8.47156465e-01 -8.25040340e-02
-3.66668105e-01 3.10202897e-01 5.60932755e-01 1.08657670e+00
-9.12103832e-01 8.45471680e-01 8.06751442e+00 1.84542090e-01
-1.00707364e+00 1.61609560e-01 7.76724696e-01 -1.42818809e-01
-4.92049128e-01 1.19357154e-01 5.63820591e-03 3.76606256e-01
5.96027553e-01 -3.17212522e-01 1.34371531e+00 2.66466886e-01
9.04781580e-01 -4.85647023e-02 -1.26277363e+00 1.11188948e+00
-2.90795565e-02 -1.81488800e+00 -1.34864956e-01 -2.41013840e-01
1.10910857e+00 -1.78064689e-01 3.48719984e-01 -6.25692308e-01
3.89710397e-01 -1.43721914e+00 7.90334821e-01 9.88259733e-01
9.33115661e-01 1.13657787e-02 1.62179962e-01 2.52668798e-01
-6.27741396e-01 -1.76564589e-01 -1.94939539e-01 -2.87432224e-01
4.41846073e-01 5.81055045e-01 -2.02421039e-01 -1.14827678e-02
8.02320957e-01 1.13319123e+00 -1.74481198e-01 1.00894928e+00
-5.09052455e-01 6.92652464e-01 2.02865317e-01 8.79610002e-01
-1.59691125e-01 -5.15793204e-01 7.86311805e-01 1.18756425e+00
3.57767224e-01 4.43619728e-01 -2.88402557e-01 1.34241974e+00
-1.93309233e-01 -5.59361994e-01 -4.31704551e-01 3.19523722e-01
2.02805713e-01 1.02639902e+00 -2.87747711e-01 -4.38002199e-01
-3.87277007e-01 1.51918650e+00 5.36161810e-02 1.13686776e+00
-5.94723642e-01 7.69869536e-02 1.05252218e+00 1.36703759e-01
3.98267448e-01 -3.09599817e-01 -6.63962126e-01 -1.52530527e+00
-1.62453055e-01 -8.64947259e-01 -5.73975481e-02 -1.50571966e+00
-1.19606602e+00 4.14719582e-01 -2.67879099e-01 -1.21741533e+00
-1.55179277e-01 -6.32163763e-01 -7.39533722e-01 1.15217066e+00
-1.79372096e+00 -6.08714223e-01 -5.64937234e-01 6.20659411e-01
6.80788457e-01 -2.21219864e-02 6.68579459e-01 3.27236243e-02
-2.41140693e-01 -1.77058220e-01 3.92390788e-01 -1.99305654e-01
8.10512304e-01 -1.29363525e+00 5.18822372e-01 1.16621006e+00
-9.16227177e-02 5.50882757e-01 1.19637883e+00 -3.75053972e-01
-1.41614842e+00 -9.63102698e-01 4.66592222e-01 -2.88600951e-01
5.79082370e-01 3.34361978e-02 -1.00213778e+00 9.34433460e-01
4.26227689e-01 5.04622161e-01 1.26449302e-01 -6.12763226e-01
-3.17095309e-01 1.21176410e-02 -1.20352972e+00 6.39357269e-01
1.22893918e+00 -6.88479185e-01 -4.74608332e-01 4.53065515e-01
6.94911778e-01 -6.50327861e-01 -5.60952544e-01 -1.22940138e-01
4.89472985e-01 -1.19630647e+00 1.30352497e+00 -8.40761006e-01
1.00978196e+00 -4.78931248e-01 4.08592895e-02 -1.67643321e+00
-6.59932554e-01 -1.05271697e+00 -4.32799041e-01 5.76182246e-01
3.02522063e-01 -5.83421171e-01 5.81157029e-01 6.75989211e-01
-1.59422129e-01 -2.40436763e-01 -5.98582208e-01 -4.74588633e-01
-2.32638761e-01 -9.06240866e-02 1.40262678e-01 7.94677913e-01
-3.63876730e-01 9.04684812e-02 -7.50667810e-01 2.69843638e-01
1.11622882e+00 2.26948380e-01 5.33315241e-01 -8.44096065e-01
-6.20625079e-01 -2.73992419e-01 -5.65447882e-02 -1.65168786e+00
1.12140030e-01 -6.26268864e-01 2.34898314e-01 -1.67056298e+00
1.27499700e-01 1.11760400e-01 3.71936746e-02 2.66075075e-01
-1.89538628e-01 4.91363466e-01 1.72621801e-01 5.95590711e-01
-2.64994711e-01 1.91641852e-01 1.78181100e+00 -3.97461765e-02
-2.55133301e-01 -1.49717361e-01 -8.13113689e-01 7.23847091e-01
4.99289691e-01 -2.09637329e-01 -3.84185910e-01 -1.00622511e+00
-1.69794187e-01 4.48447943e-01 7.35869229e-01 -6.16695344e-01
2.33411252e-01 -3.72648746e-01 5.86468756e-01 1.41138390e-01
3.40230495e-01 -5.48351824e-01 3.31009477e-01 3.42384756e-01
-6.59060895e-01 -1.92420676e-01 2.81189010e-02 5.49152076e-01
-2.13186786e-01 -3.59091051e-02 1.05444145e+00 -4.13532823e-01
-7.22566903e-01 4.67264839e-02 -3.91005009e-01 1.37075946e-01
4.63001370e-01 -2.84526110e-01 -6.30864382e-01 -8.79771888e-01
-1.11079741e+00 -1.43974781e-01 6.01865768e-01 -9.84724960e-04
9.51202989e-01 -1.07972825e+00 -9.41549599e-01 1.85404629e-01
-5.57057440e-01 -6.10305741e-02 1.37492388e-01 7.05343366e-01
-6.73448741e-01 -1.11337386e-01 -2.10701585e-01 -5.84950268e-01
-1.12297440e+00 3.12824816e-01 7.30328381e-01 8.79603848e-02
-9.11700189e-01 7.19848335e-01 4.55188483e-01 7.64764771e-02
3.91490608e-02 -2.87947416e-01 1.76397413e-01 -6.70363069e-01
8.69150817e-01 5.22839844e-01 -3.85766268e-01 -3.70992780e-01
1.48034140e-01 6.27551377e-01 3.13559085e-01 -4.47257847e-01
1.59954977e+00 -5.85633457e-01 -4.62694287e-01 1.80279016e-01
1.01329017e+00 -1.42685205e-01 -2.22250891e+00 -2.78709978e-01
-3.50156009e-01 -9.27121460e-01 3.99421304e-01 -9.61208522e-01
-1.24141490e+00 6.51813388e-01 5.44262588e-01 1.22312956e-01
1.51387072e+00 -7.47276917e-02 6.47136211e-01 8.62706080e-02
-1.36273205e-01 -7.25980282e-01 3.46832812e-01 2.53317147e-01
1.12300527e+00 -1.19158494e+00 1.95509732e-01 -4.77041483e-01
-5.61308265e-01 1.19913888e+00 1.51425421e-01 -2.94231534e-01
6.07951820e-01 5.33004284e-01 1.86698213e-01 1.08466260e-01
-9.49241996e-01 -3.97669636e-02 3.74282897e-01 7.15897441e-01
3.76620322e-01 -2.92872548e-01 2.01952700e-02 3.16515826e-02
1.37433976e-01 2.48688027e-01 8.47651660e-01 6.64706469e-01
-5.24857223e-01 -8.06711614e-01 -6.29886091e-01 3.09797943e-01
-3.95610869e-01 -3.41485351e-01 7.92724546e-04 1.05861090e-01
-1.24080367e-01 1.00150001e+00 1.31377978e-02 5.25522754e-02
1.10810548e-01 -1.95464358e-01 7.93675363e-01 -5.45779943e-01
-2.92397469e-01 2.61121362e-01 -9.85021219e-02 -7.99572349e-01
-6.25110209e-01 -7.14798629e-01 -1.12971914e+00 -2.75463581e-01
1.33998603e-01 -4.34034467e-01 3.94461423e-01 7.99034178e-01
4.04737025e-01 6.13342643e-01 4.99325365e-01 -1.29924190e+00
-4.61241812e-01 -6.47491276e-01 -4.82319862e-01 8.40826750e-01
1.09547007e+00 -1.80602953e-01 -8.13247025e-01 9.91989434e-01] | [11.582866668701172, -2.3903050422668457] |
84b64dae-814a-4020-9f1e-4f23b8e3dd09 | anglicized-words-and-misspelled-cognates-in | null | null | https://aclanthology.org/W19-4429 | https://aclanthology.org/W19-4429.pdf | Anglicized Words and Misspelled Cognates in Native Language Identification | In this paper, we present experiments that estimate the impact of specific lexical choices of people writing in a second language (L2). In particular, we look at misspelled words that indicate lexical uncertainty on the part of the author, and separate them into three categories: misspelled cognates, {``}L2-ed{''} (in our case, anglicized) words, and all other spelling errors. We test the assumption that such errors contain clues about the native language of an essay{'}s author through the task of native language identification. The results of the experiments show that the information brought by each of these categories is complementary. We also note that while the distribution of such features changes with the proficiency level of the writer, their contribution towards native language identification remains significant at all levels. | ['Carlo Strapparava', 'Vivi Nastase', 'Ilia Markov'] | 2019-08-01 | null | null | null | ws-2019-8 | ['native-language-identification'] | ['natural-language-processing'] | [-2.90697336e-01 2.74290126e-02 -3.36182803e-01 -2.78212070e-01
-8.06697786e-01 -1.15274346e+00 8.54053438e-01 3.17717433e-01
-6.30084634e-01 6.01452231e-01 5.08682013e-01 -7.38343358e-01
7.65112564e-02 -2.66550183e-01 -2.88612843e-01 -2.76867390e-01
7.40896702e-01 4.33721364e-01 -3.54980826e-02 -2.15773851e-01
1.01835823e+00 5.20926893e-01 -9.59495008e-01 6.69322014e-02
7.90957808e-01 3.80605310e-01 2.02352554e-02 4.32303190e-01
-4.07228023e-01 1.00818121e+00 -8.57840419e-01 -8.05022597e-01
1.54261366e-01 -3.35190386e-01 -8.82294595e-01 5.22915227e-03
6.57203317e-01 -2.25178912e-01 -3.35687667e-01 1.45287704e+00
3.14269096e-01 -9.02326927e-02 8.43922436e-01 -6.24172747e-01
-6.47824287e-01 1.04336774e+00 -3.98282349e-01 5.48157871e-01
6.62461460e-01 3.70405793e-01 1.04518664e+00 -1.18314242e+00
7.44977713e-01 1.54909730e+00 7.90455997e-01 4.94921029e-01
-1.23132038e+00 -7.00476587e-01 4.23511595e-01 -2.05211639e-01
-1.52083254e+00 -9.28558767e-01 7.28216827e-01 -1.05260849e+00
3.33770066e-01 1.51142597e-01 1.31484881e-01 1.45966196e+00
2.45002478e-01 4.81239051e-01 1.77278495e+00 -5.85125744e-01
1.08740322e-01 7.52299130e-01 8.60517263e-01 6.16177738e-01
5.70602655e-01 -2.91046593e-02 -8.42618108e-01 -2.82031566e-01
1.99111208e-01 -6.44676089e-01 -4.34069395e-01 3.90396595e-01
-1.15818226e+00 7.89614022e-01 -2.50821680e-01 6.99808776e-01
9.03642830e-03 -2.31200725e-01 2.82547265e-01 4.14374322e-01
3.68492782e-01 8.31120074e-01 -4.27107066e-01 -3.64842355e-01
-1.01547527e+00 1.59480676e-01 1.12087941e+00 5.94769418e-01
3.67300123e-01 -1.63168907e-02 -3.21065009e-01 8.44533086e-01
3.02936435e-01 4.76065248e-01 8.14214468e-01 -6.61030769e-01
5.42875886e-01 2.22341388e-01 4.48049426e-01 -1.04789376e+00
-2.56828785e-01 -4.77605790e-01 -1.16999447e-01 1.53721139e-01
9.76499856e-01 -1.58858895e-01 -5.11814773e-01 1.91142988e+00
-3.53051424e-01 -7.37584293e-01 -8.00756961e-02 6.47293866e-01
5.70880890e-01 4.50250655e-01 2.54911035e-01 -5.61581910e-01
1.59615147e+00 -6.49669111e-01 -7.21310616e-01 -5.59468150e-01
5.84970951e-01 -9.82830584e-01 1.40005898e+00 3.20003122e-01
-1.02152193e+00 -5.34792721e-01 -1.09789813e+00 1.71603542e-02
-2.74613112e-01 8.08703974e-02 2.49245048e-01 1.17636025e+00
-8.50250483e-01 5.55451393e-01 -2.46417284e-01 -3.42523783e-01
-1.28437981e-01 -9.57702845e-03 -1.66053757e-01 1.43899143e-01
-1.32946980e+00 1.23441219e+00 1.29191101e-01 -2.16501370e-01
-4.47631747e-01 -4.52850074e-01 -5.75219333e-01 -2.58856207e-01
9.90903974e-02 1.44225463e-01 1.19168448e+00 -1.36688197e+00
-1.27821052e+00 1.55181348e+00 -4.23802137e-01 2.49494195e-01
8.99013519e-01 -1.32835194e-01 -5.92985570e-01 -2.20006406e-01
4.97147858e-01 8.59273016e-04 9.34165239e-01 -1.07115161e+00
-4.10814196e-01 -5.38160563e-01 -1.08851880e-01 1.85759053e-01
-2.00219646e-01 5.95392942e-01 2.01177686e-01 -9.79650080e-01
-2.37589572e-02 -8.84009421e-01 4.06474531e-01 -1.98795140e-01
-6.32924557e-01 -6.75154984e-01 1.83985054e-01 -1.01164138e+00
1.45025325e+00 -2.37083006e+00 -1.05976060e-01 4.58052874e-01
3.35280389e-01 1.79646865e-01 1.94469213e-01 4.95224297e-01
-3.89373042e-02 7.34071195e-01 1.16529472e-01 -1.91750810e-01
2.76070982e-01 -2.66893655e-01 -2.29034871e-01 4.69603598e-01
-8.67976248e-02 8.57890725e-01 -7.30663955e-01 -2.22802922e-01
-4.62262273e-01 5.55609316e-02 6.64383247e-02 -1.20680563e-01
1.81312054e-01 3.64082843e-01 -1.63720280e-01 4.61056203e-01
3.78930211e-01 -2.87456866e-02 2.47924402e-01 2.70393640e-01
-5.88384628e-01 9.89614248e-01 -9.55368221e-01 9.55670536e-01
-2.40047053e-01 8.42900217e-01 1.51339620e-01 -1.98374987e-01
8.44912529e-01 1.35882840e-01 -6.32132530e-01 -3.11376482e-01
9.46996361e-02 7.23992109e-01 4.81758922e-01 -3.17705154e-01
4.27838713e-01 -3.80112797e-01 -3.05276752e-01 7.86947846e-01
-3.94057333e-01 7.76529685e-02 2.46946394e-01 1.47834733e-01
7.12189436e-01 -4.49782342e-01 6.30700111e-01 -1.02313888e+00
6.00214124e-01 -2.76064008e-01 4.68730211e-01 1.20676613e+00
-3.87924373e-01 1.62772268e-01 7.14137673e-01 -1.21579215e-01
-9.59876478e-01 -1.06353402e+00 -2.53055811e-01 1.27633965e+00
-4.96901684e-02 -3.60244900e-01 -6.80383742e-01 -5.55852115e-01
4.79727909e-02 1.29753995e+00 -5.22459030e-01 -2.44625360e-01
-5.35374582e-01 -4.63416606e-01 7.09882855e-01 5.23990989e-01
1.55499861e-01 -7.76994050e-01 -3.73628646e-01 -1.05963573e-01
-2.34147623e-01 -9.34942901e-01 -7.76438177e-01 2.70371940e-02
-3.54800969e-01 -8.84339631e-01 -4.12344724e-01 -7.09257245e-01
5.24472654e-01 -1.92703873e-01 1.09614027e+00 3.10352117e-01
1.89695254e-01 4.32989150e-01 -2.75268942e-01 -3.87346148e-01
-8.94260347e-01 1.59698114e-01 4.27725226e-01 -2.34727681e-01
8.33691776e-01 -1.47852704e-01 4.36600037e-02 -6.41702786e-02
-2.77746677e-01 -3.46942723e-01 2.84117818e-01 5.07647276e-01
-1.09176882e-01 7.45518133e-02 1.92903236e-01 -1.07931435e+00
1.01548016e+00 -2.58735001e-01 -3.76250505e-01 3.31462413e-01
-6.99965298e-01 1.42596662e-01 5.85516989e-01 -6.79233551e-01
-8.24039638e-01 -4.04376090e-01 -7.11234882e-02 2.59951025e-01
-2.68402755e-01 4.45906341e-01 -2.15446189e-01 -2.23133683e-01
6.93787575e-01 1.78602189e-01 -1.44334242e-01 -5.30116618e-01
-1.53797001e-01 9.44806039e-01 6.32219076e-01 -8.77572536e-01
9.03038025e-01 -2.52285063e-01 -6.04613841e-01 -9.61952209e-01
-9.99658108e-01 -1.22218601e-01 -6.49494350e-01 -4.35076877e-02
5.86225450e-01 -8.52164507e-01 -5.92226326e-01 5.83830833e-01
-1.20734739e+00 -2.38278568e-01 -8.72436762e-02 4.83542144e-01
-2.76993811e-01 4.87791985e-01 -8.06477368e-01 -9.61818278e-01
6.60566473e-03 -1.30681610e+00 4.43312585e-01 1.73315868e-01
-1.08760250e+00 -1.13834035e+00 -9.92981195e-02 2.86775440e-01
9.84285027e-02 -3.09983402e-01 1.46994865e+00 -1.47739387e+00
-1.42900825e-01 -1.72818601e-01 -4.73133810e-02 1.67236194e-01
-2.72125974e-02 1.20946117e-01 -8.86334240e-01 -4.22418453e-02
3.37987423e-01 -4.31236446e-01 6.28986597e-01 -1.00421421e-01
4.67794627e-01 -5.03827035e-01 -6.23235554e-02 3.23550612e-01
1.09292173e+00 9.86656845e-02 1.76642403e-01 3.36704224e-01
5.28747559e-01 5.88901699e-01 1.46650314e-01 2.62975127e-01
2.88147360e-01 4.88770366e-01 -4.11932290e-01 6.44330859e-01
-4.28733677e-02 -4.84111249e-01 8.64359319e-01 7.74036646e-01
2.06198707e-01 -3.01734447e-01 -1.19724798e+00 5.14906347e-01
-1.30357039e+00 -9.01103914e-01 -2.65272319e-01 2.29244065e+00
1.09610701e+00 3.60240847e-01 6.11377768e-02 3.10187042e-01
9.96205509e-01 3.17981541e-01 -2.87112236e-01 -7.09242642e-01
-4.25514758e-01 6.88956305e-02 4.23362762e-01 1.20745564e+00
-7.78589666e-01 1.20707703e+00 7.22496843e+00 8.06815684e-01
-1.08144939e+00 -1.48584306e-01 7.28933275e-01 3.17389339e-01
-6.74519837e-01 -1.88677251e-01 -1.01555741e+00 6.98565602e-01
7.48764932e-01 -4.61082816e-01 4.38035399e-01 5.41457891e-01
-1.28212031e-02 -2.74266630e-01 -1.22345197e+00 7.13644445e-01
2.74791628e-01 -5.59842229e-01 -2.76710149e-02 -1.23796284e-01
4.77235079e-01 -4.09509093e-01 2.86894053e-01 1.91734672e-01
3.99478436e-01 -1.06645632e+00 1.45412517e+00 2.68096000e-01
8.74969482e-01 -4.83072102e-01 5.68686306e-01 6.47056162e-01
-4.11384761e-01 -8.72330144e-02 -1.44218028e-01 -5.50309896e-01
-2.54087001e-01 4.43486243e-01 -5.11131525e-01 -3.26873213e-01
2.02267930e-01 8.37983936e-02 -1.04377270e+00 3.03065777e-01
-7.09058762e-01 9.32481349e-01 -7.46886358e-02 -3.52403343e-01
1.16551347e-01 1.16107665e-01 7.74661779e-01 1.21677041e+00
1.46153733e-01 2.42461473e-01 1.88214660e-01 9.79545236e-01
-1.25306189e-01 4.53322113e-01 -4.58397448e-01 -5.67904115e-01
7.25858629e-01 9.73940372e-01 -7.42431343e-01 -3.84605497e-01
-4.07620132e-01 1.13184011e+00 5.78276336e-01 2.62706608e-01
-1.07556731e-01 -3.20252448e-01 6.03523254e-01 2.36963719e-01
-1.82045713e-01 -2.84967303e-01 -7.25610852e-01 -1.28023410e+00
1.18243918e-01 -9.75608408e-01 1.34366855e-01 -5.44090271e-01
-1.69015789e+00 3.11092526e-01 -3.93880010e-01 -2.41527021e-01
-2.32301205e-01 -1.16666889e+00 -4.37238574e-01 1.28553116e+00
-9.04098690e-01 -6.12090051e-01 1.37680262e-01 1.64944634e-01
5.75458884e-01 -4.46184725e-01 6.30051196e-01 -2.53215522e-01
-5.38968921e-01 8.78926754e-01 -2.89931800e-03 4.38128173e-01
9.89946783e-01 -1.25418639e+00 5.09168327e-01 9.90417480e-01
3.00962150e-01 1.09429812e+00 8.39895189e-01 -8.62117589e-01
-6.47752345e-01 -3.88037115e-01 1.78428400e+00 -9.57859218e-01
1.09769869e+00 -4.15436238e-01 -7.26895213e-01 7.99575269e-01
1.48674816e-01 -4.72773522e-01 7.28005290e-01 3.31597149e-01
-7.91760206e-01 4.87097710e-01 -9.46169019e-01 7.53794193e-01
9.95369375e-01 -1.14143741e+00 -9.69532073e-01 3.32848132e-01
3.61162663e-01 -5.68076968e-02 -6.29275322e-01 -3.27005535e-02
6.38922036e-01 -1.18727994e+00 5.50802231e-01 -7.99166143e-01
5.15285730e-01 1.58849657e-01 -2.13117488e-02 -1.38121164e+00
-7.71829367e-01 -4.85455006e-01 3.79608333e-01 1.38396871e+00
5.10516405e-01 -5.99531353e-01 3.52466166e-01 1.01482785e+00
2.87076950e-01 -1.90648019e-01 -7.15079784e-01 -6.94396913e-01
6.68167710e-01 -3.05934608e-01 1.05538920e-01 1.31327701e+00
4.42681074e-01 5.66148460e-01 -5.09348139e-02 -8.79330412e-02
2.98473030e-01 -9.04309824e-02 2.15859175e-01 -1.51314497e+00
-2.56325781e-01 -8.68435144e-01 -1.78890422e-01 -1.00655901e+00
7.11097121e-01 -9.88380253e-01 1.52651705e-02 -6.79969370e-01
3.11597139e-01 -3.90570253e-01 -9.86201912e-02 1.39246225e-01
-6.51607931e-01 1.16168313e-01 4.66555655e-01 4.04206038e-01
-8.65762755e-02 -5.38568012e-02 6.11001492e-01 5.33993132e-02
-1.09385625e-01 1.44805923e-01 -1.11536551e+00 1.08732975e+00
8.23495924e-01 -3.90277416e-01 -1.70991868e-02 -4.16300833e-01
4.50404316e-01 -1.02531411e-01 3.59915435e-01 -6.32281303e-01
2.15798885e-01 -2.86945969e-01 4.08460498e-01 -2.18801089e-02
-2.38343999e-02 -4.78410423e-01 -3.29887331e-01 5.06597579e-01
-7.40455806e-01 4.83832061e-01 3.41875218e-02 -2.94220503e-02
5.46354540e-02 -8.01647246e-01 9.89081621e-01 -3.69902462e-01
-4.36434358e-01 -2.27166757e-01 -9.37995374e-01 5.58838487e-01
7.03814924e-01 -1.16537847e-01 -2.71871507e-01 -4.43957925e-01
-4.46794510e-01 -3.80540818e-01 8.64328265e-01 4.15041059e-01
7.54295290e-02 -1.15833282e+00 -7.17495739e-01 1.00989722e-01
2.00291529e-01 -1.04995584e+00 -4.91546154e-01 6.90228403e-01
-4.35548216e-01 5.71275711e-01 -1.11358285e-01 2.14911014e-01
-1.17743778e+00 4.85572159e-01 2.90270656e-01 -1.84745770e-02
-2.21918467e-02 1.09261620e+00 2.66898274e-01 -3.00886124e-01
7.48215169e-02 4.16710556e-01 -2.79239919e-02 3.56167853e-01
5.81676066e-01 5.50403118e-01 -2.29954347e-01 -1.10012126e+00
-5.79313695e-01 3.59857619e-01 -3.88786763e-01 -5.28159976e-01
4.51010853e-01 -3.47369313e-01 -3.41944039e-01 1.37399864e+00
7.57301509e-01 1.08786201e+00 -5.93443871e-01 -2.76489228e-01
4.06890243e-01 -5.03192127e-01 -2.68451750e-01 -1.07940066e+00
-3.11694384e-01 7.33456433e-01 -5.94006702e-02 -4.50515486e-02
2.49746099e-01 -1.47353932e-01 2.59196311e-01 4.11007732e-01
2.72241592e-01 -1.22868812e+00 -2.94328302e-01 8.13070238e-01
6.53134823e-01 -1.06571531e+00 7.56997615e-02 -4.03846473e-01
-5.86196125e-01 1.22915196e+00 6.64617240e-01 2.05830842e-01
4.45730776e-01 4.52729948e-02 4.38393265e-01 1.73596255e-02
-5.25387704e-01 4.44814451e-02 2.26573363e-01 3.67694288e-01
9.23397005e-01 1.03618197e-01 -8.93273294e-01 8.87367010e-01
-7.06972241e-01 -6.55461490e-01 6.91416621e-01 5.76053321e-01
-5.61214447e-01 -9.30579066e-01 -6.68098509e-01 3.98520231e-01
-7.23157644e-01 -5.53112507e-01 -1.09517825e+00 7.05195963e-01
1.00995377e-01 8.67438674e-01 2.66882349e-02 -3.37708443e-01
-3.33357155e-02 7.13846028e-01 4.08431828e-01 -6.42354727e-01
-7.92389095e-01 -3.11373353e-01 2.92718858e-01 7.75596418e-04
1.59472525e-01 -1.00004566e+00 -7.67241180e-01 -5.04169047e-01
-1.01070248e-01 3.80847007e-01 2.31543720e-01 1.22407067e+00
-2.88008869e-01 -2.35985383e-01 2.24722043e-01 -2.69143760e-01
-1.05930972e+00 -9.88095760e-01 -9.03434932e-01 4.79859114e-01
2.63081580e-01 -4.15728092e-01 -8.24718237e-01 -1.54475316e-01] | [10.408025741577148, 10.388558387756348] |
efd3e24f-e651-4807-9f68-924c6f99a5f3 | quantifying-generalization-in-reinforcement | 1812.02341 | null | https://arxiv.org/abs/1812.02341v3 | https://arxiv.org/pdf/1812.02341v3.pdf | Quantifying Generalization in Reinforcement Learning | In this paper, we investigate the problem of overfitting in deep reinforcement learning. Among the most common benchmarks in RL, it is customary to use the same environments for both training and testing. This practice offers relatively little insight into an agent's ability to generalize. We address this issue by using procedurally generated environments to construct distinct training and test sets. Most notably, we introduce a new environment called CoinRun, designed as a benchmark for generalization in RL. Using CoinRun, we find that agents overfit to surprisingly large training sets. We then show that deeper convolutional architectures improve generalization, as do methods traditionally found in supervised learning, including L2 regularization, dropout, data augmentation and batch normalization. | ['Tae-hoon Kim', 'John Schulman', 'Karl Cobbe', 'Oleg Klimov', 'Chris Hesse'] | 2018-12-06 | null | null | null | null | ['l2-regularization'] | ['methodology'] | [-1.10886544e-01 6.38209358e-02 -2.47853827e-02 -3.27713966e-01
-5.68981469e-01 -8.39105129e-01 6.31571293e-01 1.15961647e-02
-9.52194512e-01 1.39921522e+00 -4.19292599e-02 -3.35441470e-01
5.34882024e-02 -8.84207547e-01 -8.26284766e-01 -6.04268551e-01
-1.85942650e-01 4.85337257e-01 -8.47246125e-03 -3.47894996e-01
8.57205614e-02 6.55431688e-01 -1.41550744e+00 6.97652921e-02
8.83503199e-01 6.28933728e-01 -9.70286950e-02 7.71128535e-01
3.19658667e-01 1.06284022e+00 -1.10371304e+00 -4.57637668e-01
3.51598054e-01 -7.20890760e-01 -8.87004733e-01 1.19725727e-01
2.73690343e-01 -4.34860766e-01 -3.51081640e-01 7.50403106e-01
5.17780006e-01 4.25609440e-01 5.40417135e-01 -1.25645661e+00
-5.65838158e-01 7.27232099e-01 -1.51614055e-01 1.20344706e-01
1.80425078e-01 5.72510719e-01 7.20041573e-01 -2.93311447e-01
6.73366308e-01 1.17312062e+00 8.58694196e-01 9.35797215e-01
-1.61395884e+00 -6.29374444e-01 1.37150452e-01 -4.37474430e-01
-9.36043799e-01 -4.00125533e-01 4.38280612e-01 -2.77577072e-01
8.39588106e-01 1.52011439e-01 6.58843577e-01 1.66431117e+00
7.76551813e-02 9.35367167e-01 1.29823589e+00 -3.71329874e-01
6.17730439e-01 1.14785053e-01 -2.23691136e-01 5.27012050e-01
2.68030375e-01 6.66273296e-01 -2.22628295e-01 -8.49929750e-02
1.19114053e+00 -1.55969813e-01 -2.03419179e-01 -4.06895995e-01
-8.33616376e-01 9.81266260e-01 5.29712677e-01 1.23964764e-01
-2.36646205e-01 5.60787618e-01 5.26370585e-01 6.71498239e-01
2.18550548e-01 1.03048408e+00 -5.99877656e-01 -2.40629137e-01
-5.61925948e-01 7.06508279e-01 8.78210187e-01 7.95592308e-01
7.55087912e-01 4.10222232e-01 -2.21855734e-02 7.65473068e-01
-6.86666965e-02 6.53896928e-02 8.03540349e-01 -1.44736373e+00
2.72065639e-01 2.38531560e-01 3.61443579e-01 -4.76812512e-01
-6.06723905e-01 -9.28648412e-01 -5.11438727e-01 8.33885431e-01
5.60566127e-01 -7.22050488e-01 -8.04761946e-01 2.25832224e+00
-4.33212742e-02 6.45806566e-02 5.35478830e-01 5.32438397e-01
5.21017790e-01 1.50546387e-01 2.63852417e-01 5.23439571e-02
3.94587696e-01 -1.04502404e+00 -3.17968875e-01 -5.18056810e-01
9.30720925e-01 -2.46822283e-01 1.36558867e+00 3.86400700e-01
-1.26963770e+00 -4.12890255e-01 -9.96133685e-01 2.97583550e-01
-4.77433890e-01 -1.79808334e-01 1.24264395e+00 6.46306932e-01
-1.00742304e+00 1.07102752e+00 -8.33621383e-01 -3.71363133e-01
6.31281614e-01 4.50006753e-01 -4.12938803e-01 9.69360992e-02
-9.48656201e-01 1.03393924e+00 4.35158193e-01 -1.47505492e-01
-1.37142408e+00 -4.38264549e-01 -9.33358788e-01 -1.36200532e-01
3.12387586e-01 -6.51108980e-01 1.72499657e+00 -1.40011585e+00
-1.69816732e+00 7.48367548e-01 3.43945920e-01 -6.53593063e-01
6.78976953e-01 -1.78215161e-01 6.71096100e-03 -4.49811697e-01
-4.65232879e-02 7.91679025e-01 4.48958188e-01 -1.45688105e+00
-4.03676450e-01 -2.13674232e-01 4.39691812e-01 2.96275318e-01
4.89880741e-02 -2.14468062e-01 4.26276661e-02 -6.64848089e-01
-3.32677305e-01 -8.19165468e-01 -4.72871900e-01 -5.45391083e-01
-1.72742039e-01 -1.01112418e-01 3.31768751e-01 -1.36466220e-01
8.82676125e-01 -1.92488658e+00 1.79968894e-01 1.68649435e-01
1.03254475e-01 1.55536029e-02 -3.36038232e-01 2.43094698e-01
-2.95327995e-02 1.31973460e-01 -3.08300138e-01 -4.31427360e-01
1.33933619e-01 7.65379488e-01 -2.61770338e-01 3.53796840e-01
2.06417993e-01 1.05962384e+00 -9.43293869e-01 -5.76147251e-02
-1.87877312e-01 1.29890949e-01 -1.03113711e+00 1.79398164e-01
-5.03260672e-01 8.31619620e-01 -3.43496591e-01 3.25614005e-01
2.76146084e-01 -9.94198397e-02 1.16812989e-01 5.20231426e-01
3.85418050e-02 3.44942629e-01 -8.68841350e-01 1.79328763e+00
-5.22626579e-01 6.42713964e-01 -9.38467979e-02 -8.76215339e-01
7.97584951e-01 4.40112166e-02 1.54387042e-01 -7.35292494e-01
2.43661761e-01 3.00952166e-01 2.96824962e-01 -4.15601403e-01
4.31941360e-01 -3.18215370e-01 -1.34678766e-01 4.84171629e-01
2.76135087e-01 -3.38862240e-01 3.10230464e-01 -1.57930344e-01
1.31406426e+00 6.23980105e-01 3.65542099e-02 -1.33040130e-01
4.73360568e-02 2.50959009e-01 6.95144653e-01 1.12625754e+00
-9.88373458e-02 4.08256531e-01 6.79506302e-01 -5.00250340e-01
-1.17007399e+00 -1.00125766e+00 -3.00809257e-02 1.41379058e+00
-3.90176266e-01 -1.95781395e-01 -1.00466239e+00 -8.39623928e-01
1.34638458e-01 8.21685314e-01 -9.80082035e-01 -3.69923145e-01
-6.45183802e-01 -8.24907541e-01 9.91242707e-01 7.39399374e-01
5.26882231e-01 -1.54281235e+00 -9.04083908e-01 1.65774211e-01
4.64610696e-01 -6.10675812e-01 2.02919096e-02 9.11125004e-01
-1.01516855e+00 -9.62770045e-01 -6.39227688e-01 -7.71266580e-01
6.47682846e-01 -3.46935779e-01 1.53030300e+00 3.43712956e-01
-1.11052670e-01 4.36537087e-01 -3.25962514e-01 -2.25770667e-01
-6.14654124e-01 2.66957223e-01 1.14485733e-01 -6.65902734e-01
1.31781816e-01 -7.81742990e-01 -3.29550683e-01 2.23257378e-01
-1.01696324e+00 -3.30032229e-01 5.90133429e-01 1.03050840e+00
1.90131798e-01 -8.06997716e-02 7.90520966e-01 -1.18310773e+00
1.18273294e+00 -4.43581283e-01 -7.10708737e-01 2.41266824e-02
-3.42427731e-01 2.16188177e-01 8.58501434e-01 -5.13264596e-01
-7.30037987e-01 -2.17219651e-01 -2.06218407e-01 -2.12430224e-01
-5.81701100e-01 5.67414284e-01 1.66048985e-02 -3.19941640e-01
1.11836314e+00 9.73869786e-02 1.60225146e-02 -3.56928378e-01
2.43041351e-01 2.81955749e-01 4.44751203e-01 -9.59124684e-01
5.86480975e-01 6.78704157e-02 -1.32088929e-01 -5.67839265e-01
-8.26154411e-01 3.42783779e-01 -4.18816835e-01 5.15775476e-03
5.33946335e-01 -7.06202149e-01 -9.41228032e-01 2.16070712e-01
-8.64079535e-01 -1.36483705e+00 -8.20244491e-01 5.18035769e-01
-9.99829054e-01 -2.45351344e-01 -6.74383819e-01 -5.00155866e-01
2.41530195e-01 -1.41790640e+00 5.31034052e-01 3.92287135e-01
-2.73529649e-01 -1.34585798e+00 5.55289865e-01 -8.79212916e-02
5.56731343e-01 3.57063919e-01 7.60848761e-01 -1.09134519e+00
-2.27500752e-01 1.86822191e-01 3.09178412e-01 5.13471603e-01
5.45185469e-02 -2.29697302e-01 -1.03590345e+00 -5.14148533e-01
-5.27436808e-02 -1.15736639e+00 8.01402390e-01 2.05400378e-01
1.27110243e+00 -2.78898358e-01 2.95934882e-02 9.05034661e-01
1.17121613e+00 4.07771498e-01 8.19119751e-01 9.50278223e-01
3.50065589e-01 4.79150742e-01 2.60140866e-01 3.57913285e-01
1.23146206e-01 3.26080263e-01 4.36345935e-01 -7.80061930e-02
2.12474450e-01 -3.97795171e-01 3.18323195e-01 2.65514404e-01
-5.41172326e-02 -1.66631013e-01 -8.20037484e-01 2.12138042e-01
-1.77969444e+00 -9.83896136e-01 6.20552480e-01 2.18865871e+00
1.06027436e+00 1.83619604e-01 2.43812621e-01 -1.10028133e-01
1.96957245e-01 -5.88258468e-02 -7.54768014e-01 -5.08852482e-01
-3.48938465e-01 6.08096898e-01 5.24909794e-01 6.02157593e-01
-1.01051176e+00 1.20455277e+00 7.57190275e+00 5.38297355e-01
-1.04321826e+00 -9.43137482e-02 8.42275023e-01 -1.78628460e-01
-3.53884369e-01 -1.20903365e-01 -6.35584533e-01 1.56561807e-01
7.23837376e-01 2.07563385e-01 8.47680986e-01 8.26845467e-01
4.09214720e-02 -1.10266462e-01 -1.43896401e+00 6.17635965e-01
-4.85617369e-02 -1.18245816e+00 -2.08808601e-01 -6.71908483e-02
1.00706637e+00 8.88519660e-02 3.22083384e-01 9.31331158e-01
1.11728418e+00 -1.67340136e+00 5.53147316e-01 2.30650142e-01
4.16692704e-01 -8.99383128e-01 5.23728192e-01 3.19386572e-01
-2.83776760e-01 -2.59291261e-01 -5.93672156e-01 -2.47677892e-01
-4.75856483e-01 -1.87467858e-01 -7.87741601e-01 1.84885934e-02
4.56774682e-01 5.76293528e-01 -7.06620336e-01 1.27176845e+00
-4.06260550e-01 4.93889064e-01 -2.71426678e-01 -1.86274759e-02
6.60363078e-01 -2.06502765e-01 2.18745351e-01 9.49147701e-01
2.46225223e-02 -2.46223450e-01 1.94288537e-01 9.49780107e-01
-4.46514934e-01 -1.92309245e-01 -9.42656517e-01 1.03762329e-01
2.42087841e-01 8.63142133e-01 -4.91875827e-01 -3.38014290e-02
-2.25448638e-01 8.15517962e-01 8.77372503e-01 8.12630892e-01
-8.15771461e-01 -1.85818627e-01 6.42747164e-01 -2.48275116e-01
2.34947905e-01 -3.78801078e-01 -1.59131557e-01 -1.05780709e+00
-1.78268805e-01 -1.27270973e+00 2.12133765e-01 -6.65573895e-01
-1.13553452e+00 6.36631131e-01 -1.59504384e-01 -8.46462607e-01
-5.67719638e-01 -6.70646250e-01 -7.48547912e-01 7.64089406e-01
-1.41092217e+00 -6.24414027e-01 -2.08066314e-01 5.67862689e-01
3.71639073e-01 -4.64400083e-01 9.56389546e-01 -1.64195162e-03
-7.18160868e-01 6.84801400e-01 4.05841678e-01 2.76713371e-01
5.72855413e-01 -1.45682335e+00 3.10117990e-01 3.77687693e-01
2.50439674e-01 7.69449472e-01 7.84365654e-01 -4.92698938e-01
-1.01596332e+00 -1.00682437e+00 5.24673387e-02 -3.36730152e-01
5.82542181e-01 -2.23682061e-01 -8.00185740e-01 1.27419853e+00
2.86880702e-01 -6.00851364e-02 6.44405961e-01 3.06645483e-01
-2.82164633e-01 1.90713536e-02 -1.11420417e+00 8.79787207e-01
1.00112736e+00 -2.75960028e-01 -5.70169389e-01 3.35176528e-01
6.88072324e-01 -5.56571186e-01 -7.48234868e-01 2.45431378e-01
3.47886980e-01 -1.20590985e+00 7.43712544e-01 -1.43598235e+00
4.35062617e-01 1.03356227e-01 -1.55222699e-01 -1.85552979e+00
-8.15149471e-02 -9.04296994e-01 3.75103951e-01 1.01756823e+00
7.02124119e-01 -8.03620100e-01 1.27666581e+00 5.75884938e-01
-1.29962698e-01 -7.89397299e-01 -6.70193970e-01 -1.02246284e+00
8.02202106e-01 -2.45893255e-01 5.06902516e-01 9.40050006e-01
-7.07252175e-02 2.04009399e-01 -6.07707771e-03 -2.28013933e-01
4.39362735e-01 -8.84721726e-02 1.04118526e+00 -9.83659446e-01
-6.58311546e-01 -6.63949490e-01 -1.15934573e-01 -1.09924781e+00
5.33456504e-01 -9.03336346e-01 1.06188461e-01 -1.21390843e+00
7.30731264e-02 -9.22575831e-01 -1.42379224e-01 5.46246529e-01
1.71047002e-01 9.39758718e-02 6.22748546e-02 7.49752373e-02
-6.89389408e-01 7.35676527e-01 1.43453801e+00 1.19850650e-01
-4.40566391e-01 1.29528508e-01 -6.53472126e-01 9.10444796e-01
1.39622796e+00 -4.69223022e-01 -4.26260889e-01 -6.92738593e-01
3.40603113e-01 -1.99261531e-01 5.18284619e-01 -1.11046493e+00
9.89598632e-02 -4.04977053e-01 6.74159050e-01 1.77079454e-01
2.59637624e-01 -6.66399240e-01 -1.09990463e-01 3.64610672e-01
-8.00053537e-01 3.69284451e-01 5.99709809e-01 2.37436950e-01
-6.61204308e-02 -6.61320508e-01 7.70996928e-01 -4.72157925e-01
-4.41845119e-01 1.47323802e-01 -3.71453673e-01 5.13815582e-01
9.61503685e-01 -8.30010101e-02 -4.92735475e-01 -4.54982936e-01
-9.26084936e-01 3.44931960e-01 7.59388983e-01 -6.78084092e-04
5.23994386e-01 -1.18695557e+00 -6.23971164e-01 2.61963725e-01
-1.24207303e-01 1.65878430e-01 -2.66649693e-01 5.91842592e-01
-6.47354841e-01 1.06329091e-01 -4.86898690e-01 -2.90116936e-01
-7.78848946e-01 3.52027208e-01 8.80108297e-01 -2.50382960e-01
-3.43531042e-01 9.53092754e-01 2.46882990e-01 -8.49027991e-01
5.53413510e-01 -3.69411767e-01 -7.80033544e-02 -2.89383084e-01
2.20655560e-01 1.82044104e-01 -2.40793139e-01 -6.39687628e-02
9.43329185e-02 1.64037004e-01 -1.05589606e-01 -3.28441858e-01
1.55758643e+00 3.64322364e-01 1.13369964e-01 4.89069790e-01
8.97798121e-01 -6.08902574e-02 -1.80184174e+00 1.51838258e-01
7.20559135e-02 -2.70159632e-01 -2.87175298e-01 -9.70929563e-01
-9.93379414e-01 7.04699814e-01 2.43552729e-01 3.99987817e-01
7.60745466e-01 -1.87759101e-01 1.54608294e-01 9.17876780e-01
4.38169986e-01 -1.16683304e+00 1.55974984e-01 7.99418569e-01
1.01968026e+00 -1.19057727e+00 -9.75172222e-02 4.31388587e-01
-7.38171101e-01 9.61317480e-01 8.79453957e-01 -5.68788946e-01
1.16884045e-01 4.60892260e-01 3.47514637e-02 1.43044814e-02
-8.65171015e-01 -1.65951446e-01 -4.37266439e-01 8.99883807e-01
4.86398339e-01 -8.36357772e-02 1.94626153e-01 3.93028110e-01
-5.27111709e-01 -2.00054608e-02 6.08459711e-01 1.12206089e+00
-4.35231835e-01 -1.27686918e+00 -1.95651248e-01 2.29477674e-01
-5.73652864e-01 2.83940621e-02 -4.43061113e-01 1.24004161e+00
5.96169680e-02 6.50326014e-01 2.11423263e-03 -2.62064606e-01
3.19960654e-01 -5.62625080e-02 9.34859931e-01 -6.75120115e-01
-1.08531058e+00 -2.98020035e-01 1.71754956e-01 -5.67723215e-01
-3.06201220e-01 -4.96799231e-01 -1.18445873e+00 -2.87226647e-01
2.12762207e-02 3.27876836e-01 4.25798386e-01 8.59414458e-01
-5.41843250e-02 7.99363971e-01 4.41482425e-01 -8.86414349e-01
-9.41586554e-01 -1.00107944e+00 -5.42161703e-01 5.06558299e-01
4.79675293e-01 -7.94017315e-01 -6.10327423e-01 -3.28456461e-01] | [4.065098285675049, 1.7225666046142578] |
80b98daf-29a1-4cfb-8956-1e6101fedf5d | heuristics-based-mosaic-of-social-sensor | 2009.11663 | null | https://arxiv.org/abs/2009.11663v1 | https://arxiv.org/pdf/2009.11663v1.pdf | Heuristics based Mosaic of Social-Sensor Services for Scene Reconstruction | We propose a heuristics-based social-sensor cloud service selection and composition model to reconstruct mosaic scenes. The proposed approach leverages crowdsourced social media images to create an image mosaic to reconstruct a scene at a designated location and an interval of time. The novel approach relies on the set of features defined on the bases of the image metadata to determine the relevance and composability of services. Novel heuristics are developed to filter out non-relevant services. Multiple machine learning strategies are employed to produce smooth service composition resulting in a mosaic of relevant images indexed by geolocation and time. The preliminary analytical results prove the feasibility of the proposed composition model. | ['Tooba Aamir', 'Hai Dong', 'Athman Bouguettaya'] | 2020-09-21 | null | null | null | null | ['service-composition'] | ['miscellaneous'] | [ 2.87256002e-01 -2.21321434e-01 2.68125385e-02 -4.99634027e-01
-7.95308948e-01 -5.26733816e-01 8.29406381e-01 2.09255457e-01
-1.75766498e-01 3.07351887e-01 2.45872006e-01 1.53837293e-01
-4.49513465e-01 -7.75392413e-01 -4.46487188e-01 -7.53540039e-01
9.89072304e-03 2.90192008e-01 3.05565536e-01 -2.10858732e-01
4.76017207e-01 6.51218414e-01 -2.11951923e+00 4.63046342e-01
6.83518648e-01 1.29671621e+00 6.47577941e-01 5.59537172e-01
-4.06767011e-01 7.80893981e-01 -3.86175990e-01 -2.54788786e-01
5.84758222e-01 -1.75721169e-01 -3.81165147e-01 8.51692975e-01
-4.30501960e-02 -9.51258168e-02 1.75531834e-01 1.04551780e+00
2.52734691e-01 7.01465830e-02 2.48627394e-01 -1.65573645e+00
-3.58959585e-01 4.95690197e-01 -3.64518046e-01 -6.31920099e-02
7.83295989e-01 8.53874609e-02 7.53444612e-01 -9.05864358e-01
1.12544930e+00 6.93755746e-01 4.37175274e-01 -4.68191653e-02
-5.90751588e-01 -1.43060803e-01 2.55979180e-01 3.86244059e-01
-1.70384920e+00 -7.18988299e-01 8.14453602e-01 -1.21613197e-01
6.38891637e-01 7.47056484e-01 9.73263800e-01 4.18904364e-01
3.88479084e-02 5.29805362e-01 1.26620579e+00 -7.18499839e-01
6.31480992e-01 8.01443905e-02 -4.76386517e-01 5.20930886e-01
-2.98850145e-02 -3.13083082e-01 -8.50574791e-01 -5.12663126e-01
1.01442434e-01 5.52411795e-01 2.16366366e-01 -3.44349414e-01
-1.21150768e+00 3.65009725e-01 6.68050051e-02 1.32594779e-01
-1.09668863e+00 -2.58671790e-01 6.77867234e-02 9.49576944e-02
5.54248333e-01 -7.85434172e-02 -1.61880389e-01 1.16510212e-01
-8.79040182e-01 3.05731386e-01 6.43783569e-01 1.19152558e+00
9.89873111e-01 -1.80861548e-01 -1.53504908e-01 5.44602036e-01
4.26707655e-01 7.50411391e-01 -8.88631679e-03 -1.35009742e+00
1.48313090e-01 9.26178098e-01 4.37634259e-01 -1.12758183e+00
-4.21843715e-02 2.22947702e-01 -1.88879043e-01 1.44310510e-02
-1.87840253e-01 1.33529427e-02 -7.20882297e-01 8.24274004e-01
9.03509557e-01 2.93489844e-01 1.17270842e-01 1.14543402e+00
5.04140437e-01 4.94423330e-01 8.39625522e-02 -2.82011807e-01
1.15769076e+00 -5.13300180e-01 -5.79206645e-01 3.36151600e-01
-3.53570044e-01 -7.38207579e-01 6.76467896e-01 2.01885149e-01
-9.52846110e-01 3.13246511e-02 -7.15834379e-01 4.99273151e-01
-4.56866771e-01 -1.56722933e-01 6.73146367e-01 4.23496187e-01
-1.14023042e+00 2.65731841e-01 -1.44134969e-01 -6.68996990e-01
2.53919244e-01 2.25247070e-01 -1.28193945e-01 -1.60069808e-01
-5.84471464e-01 4.78863388e-01 -1.93883032e-02 2.66205482e-02
-1.11789763e+00 -1.76353991e-01 -3.54247123e-01 -1.60817131e-01
4.48397666e-01 -6.11546338e-01 8.24269712e-01 -1.46883488e+00
-1.12529647e+00 9.21784163e-01 -2.46343568e-01 -5.34954190e-01
5.20396471e-01 5.47916830e-01 -8.31959069e-01 5.50429583e-01
5.60209394e-01 2.34025329e-01 1.07118070e+00 -1.59189010e+00
-1.66038942e+00 -3.75472844e-01 1.85870528e-01 6.26986504e-01
-2.09339678e-01 4.33637083e-01 -9.18600857e-01 -1.48443580e-01
4.50331330e-01 -8.81948769e-01 -4.84761089e-01 -4.39162314e-01
-1.83549330e-01 2.18874678e-01 4.27500933e-01 -5.54748297e-01
7.95391142e-01 -1.83643997e+00 -2.82612890e-01 6.38751149e-01
-2.07787566e-02 -3.61029208e-01 -1.46234855e-01 7.80100524e-01
6.49990261e-01 -2.38030359e-01 -2.30680615e-01 -1.83719337e-01
-1.01420969e-01 1.70400485e-01 -2.85932481e-01 5.74130714e-01
-1.86000973e-01 4.40694362e-01 -1.06574225e+00 -6.77889585e-01
3.44729573e-01 4.45739746e-01 3.20006050e-02 2.59880900e-01
-4.57516879e-01 3.71339977e-01 -5.92236400e-01 1.34662485e+00
5.46206594e-01 -5.08808158e-03 4.75521356e-01 -5.21903299e-02
-2.71095634e-01 -1.79439753e-01 -1.24988556e+00 1.73305416e+00
-1.55518234e-01 2.42428362e-01 1.46086469e-01 -6.13671124e-01
7.52338588e-01 3.24575692e-01 1.36745203e+00 -8.08781683e-01
1.04562506e-01 1.04034983e-01 -7.80433774e-01 -9.30774152e-01
8.03394973e-01 3.86736125e-01 -3.29197556e-01 6.20016992e-01
-4.00117666e-01 1.16392739e-01 8.57643485e-02 1.93273887e-01
1.13539827e+00 4.07845259e-01 2.20737055e-01 -6.98947459e-02
5.12701154e-01 6.79873884e-01 5.64800322e-01 8.07698190e-01
-3.23878020e-01 5.65261245e-01 -9.47595537e-02 -8.42522204e-01
-9.42386031e-01 -7.68661737e-01 4.19756472e-01 1.19051743e+00
7.49728203e-01 1.74223178e-03 -8.61970544e-01 -5.75298488e-01
2.08440319e-01 3.45393747e-01 -4.18390036e-01 4.78668809e-01
-1.35844722e-01 -6.16169870e-01 -6.23303950e-02 -4.26753849e-01
2.00561240e-01 -1.22969759e+00 -9.84186709e-01 1.71922334e-02
-2.29859918e-01 -1.31427383e+00 -1.76538602e-01 -5.10906100e-01
-3.07023585e-01 -1.23874474e+00 -2.02968657e-01 -4.25174981e-01
1.02469862e+00 9.37297821e-01 8.58800054e-01 1.28534511e-01
-1.50804952e-01 8.76892209e-01 -6.70617163e-01 -2.87233859e-01
-1.77497536e-01 -2.16863945e-01 2.97824573e-02 9.15425241e-01
4.63559985e-01 -6.12129152e-01 -6.39572561e-01 2.40719870e-01
-1.02225220e+00 1.17450185e-01 4.06189233e-01 -7.85025656e-02
9.18226719e-01 9.86847803e-02 4.19954151e-01 -6.92784011e-01
6.63658500e-01 -1.00968301e+00 -7.96001315e-01 5.56034029e-01
-9.44525957e-01 -3.34933341e-01 2.21560881e-01 -2.18392357e-01
-1.11745930e+00 5.06720364e-01 6.33101225e-01 -3.44270915e-01
-2.46300131e-01 4.37160254e-01 -1.52171746e-01 -2.98560381e-01
3.35072786e-01 2.81163186e-01 -2.00723723e-01 -2.92294949e-01
4.78217125e-01 7.91137218e-01 5.06969571e-01 -3.62996429e-01
6.71685278e-01 1.18862319e+00 -2.59889722e-01 -5.22754431e-01
-3.57678503e-01 -8.54365408e-01 -3.02734554e-01 -8.76751781e-01
5.06018043e-01 -1.18682122e+00 -6.93829119e-01 1.05034180e-01
-8.32805276e-01 2.68301755e-01 -2.91111231e-01 6.55364245e-02
-5.02190530e-01 -1.31536722e-02 7.42866918e-02 -1.24259138e+00
-4.31515545e-01 -9.76572335e-01 1.14089000e+00 3.60693336e-01
2.39033908e-01 -2.44099498e-01 -2.74095982e-01 4.97714430e-01
3.65365833e-01 4.74436909e-01 -4.54367623e-02 -4.00566339e-01
-1.17800558e+00 -5.24552584e-01 -2.77431518e-01 -4.02967960e-01
1.50453806e-01 1.75512388e-01 -9.50084329e-01 2.44919762e-01
-1.65447503e-01 4.08921957e-01 1.89966351e-01 3.63648295e-01
6.70663893e-01 -4.21936512e-01 -2.60591745e-01 4.89794195e-01
1.89625871e+00 5.11449277e-01 4.66822416e-01 6.60908997e-01
6.27355814e-01 5.84676504e-01 1.05631959e+00 1.06863415e+00
1.00757110e+00 5.81625819e-01 6.89635634e-01 1.37866482e-01
2.02308059e-01 -1.93409577e-01 1.74457490e-01 4.84580934e-01
-4.73153330e-02 -8.25573802e-02 -8.83346260e-01 1.10906982e+00
-2.16841769e+00 -9.59148645e-01 8.41924399e-02 2.04539156e+00
1.02886744e-01 -2.91819096e-01 2.32549801e-01 -5.59259392e-02
8.66038024e-01 2.23672673e-01 -4.99486893e-01 -2.11829185e-01
-4.67206001e-01 -3.81514728e-01 9.93012071e-01 3.50009501e-01
-7.38954246e-01 8.50292206e-01 6.22828960e+00 2.99207956e-01
-7.99037278e-01 5.71884811e-01 3.02451611e-01 -2.75267214e-01
-7.53495097e-01 4.72735435e-01 -4.54756916e-01 4.82383728e-01
7.29828715e-01 -3.73945534e-01 1.08185506e+00 7.54735172e-01
5.72917759e-01 -6.03935540e-01 -2.05447331e-01 9.40410554e-01
4.15252209e-01 -1.68605149e+00 -1.61485299e-01 -6.84776381e-02
1.10060883e+00 3.59827757e-01 -2.58759141e-01 -5.55869460e-01
4.94791269e-01 -2.10257605e-01 1.28038681e+00 1.08234417e+00
4.93662387e-01 -6.81330502e-01 3.06981623e-01 1.92387804e-01
-1.13986123e+00 -4.82928872e-01 -1.67446166e-01 9.39125381e-03
2.90126413e-01 6.36073411e-01 -1.00085282e+00 6.59028113e-01
7.74923086e-01 2.44989142e-01 -5.51830411e-01 1.17680895e+00
1.56520262e-01 1.53543040e-01 -1.57269120e-01 -1.20056361e-01
-5.02705202e-02 -5.03373742e-01 7.56534636e-01 8.56126904e-01
4.26196247e-01 6.48298338e-02 3.24376673e-01 4.50011164e-01
7.27351010e-02 3.73430818e-01 -6.88671052e-01 1.96464807e-01
7.64179051e-01 1.43818545e+00 -9.12904680e-01 -2.79374242e-01
-4.14413810e-01 9.72290933e-01 -9.31699425e-02 5.83524764e-01
-5.99882483e-01 1.30766377e-01 5.33460796e-01 2.20721006e-01
3.01120132e-01 -1.18309155e-01 -3.74478251e-01 -1.08056593e+00
3.03665757e-01 -7.56036222e-01 3.71698856e-01 -9.74853396e-01
-9.74205196e-01 6.78941965e-01 -7.36467987e-02 -1.36244857e+00
-1.73060045e-01 3.81417632e-01 -2.08004922e-01 5.20582795e-01
-1.62635458e+00 -1.40397465e+00 -5.44613481e-01 7.31542647e-01
4.59548652e-01 -5.88372171e-01 7.24285960e-01 1.74581140e-01
-1.75339460e-01 -3.21426451e-01 1.76906168e-01 -4.00967151e-01
1.70856938e-01 -7.21602857e-01 2.51949012e-01 1.12596917e+00
2.53179669e-01 1.39872968e-01 7.84170270e-01 -8.03179741e-01
-1.80847168e+00 -9.29111004e-01 1.11264753e+00 -2.05119222e-01
3.84672672e-01 -1.10732749e-01 -5.02077676e-02 2.35402524e-01
1.85747892e-01 2.98837185e-01 5.27091682e-01 -4.56183195e-01
-6.54471144e-02 -7.99711585e-01 -1.58633208e+00 5.71457684e-01
1.07550001e+00 -6.60914063e-01 2.14063656e-03 5.54074705e-01
5.57677388e-01 -3.60958762e-02 -6.60795212e-01 1.50264069e-01
7.26136804e-01 -1.13399553e+00 6.65787458e-01 -2.18552217e-01
8.05616677e-02 -6.97296560e-01 -6.10077798e-01 -6.52794600e-01
1.08768612e-01 -7.67614663e-01 1.32993653e-01 1.18579912e+00
2.98710972e-01 -4.50780869e-01 6.74872816e-01 1.05713975e+00
1.42323181e-01 -1.42313883e-01 -9.25592840e-01 -3.12609702e-01
-1.32652366e+00 -5.27546525e-01 1.27505422e+00 8.17247808e-01
-2.34938100e-01 -6.54513478e-01 -2.50225633e-01 4.79601443e-01
9.40234184e-01 6.09885037e-01 7.99158931e-01 -9.93854702e-01
2.72296876e-01 3.01116314e-02 -3.86154294e-01 2.21819356e-01
3.10768485e-02 -5.57526231e-01 2.43890032e-01 -1.69194424e+00
9.36479717e-02 -8.72211754e-01 -5.94718695e-01 2.26932257e-01
1.15205646e-01 2.66321719e-01 2.21268430e-01 5.62975645e-01
-1.19608867e+00 3.89103927e-02 4.54239517e-01 -9.83804744e-03
-2.54350185e-01 3.65217552e-02 -7.10720062e-01 3.87283504e-01
8.08659375e-01 -7.45012581e-01 -3.42159301e-01 -4.64136004e-01
3.89337003e-01 1.76753327e-01 1.89952582e-01 -9.94317234e-01
5.29169381e-01 -9.95295584e-01 -1.68017522e-02 -4.50165510e-01
1.49710923e-01 -1.23159218e+00 9.96117175e-01 -1.65498540e-01
-3.81978840e-01 2.58402944e-01 -4.63381559e-01 6.25953436e-01
-3.18028443e-02 -9.35174972e-02 1.24584004e-01 -3.60315412e-01
-1.21058595e+00 4.20620799e-01 -2.97425896e-01 -6.53681815e-01
1.31983554e+00 -3.16972971e-01 2.14220658e-01 -4.08318102e-01
-4.44321752e-01 -4.47562821e-02 1.08301914e+00 6.90915525e-01
7.26800084e-01 -1.28292990e+00 -7.17836916e-01 1.35761231e-01
5.13332367e-01 -5.17494023e-01 2.86237836e-01 4.56053495e-01
-5.93863070e-01 -1.25889987e-01 -2.64568895e-01 -5.01538277e-01
-1.10400534e+00 5.20436108e-01 2.05474701e-02 4.53652352e-01
-5.43989122e-01 2.79213548e-01 -6.74883246e-01 -9.42183807e-02
7.55637288e-02 2.81629413e-01 -1.37078866e-01 1.81627408e-01
6.03510916e-01 2.52132356e-01 1.85159490e-01 -1.22722495e+00
-5.23464680e-01 2.23753080e-01 8.02068353e-01 -5.98422170e-01
1.59022093e+00 -8.18809390e-01 -3.40348125e-01 1.36487350e-01
7.11769640e-01 2.60259598e-01 -1.35255396e+00 -4.12128121e-01
3.98227632e-01 -9.07446921e-01 1.10463966e-02 -8.05154383e-01
-1.13886118e+00 -3.88894022e-01 8.82366240e-01 3.13028127e-01
1.56003773e+00 -2.67098606e-01 5.45430779e-01 -1.98984280e-01
8.65638852e-01 -1.50472403e+00 -7.09119916e-01 -3.37852448e-01
5.87560117e-01 -1.24845541e+00 2.04979762e-01 -2.35002413e-01
-1.02751386e+00 6.76728487e-01 -1.48668036e-01 -9.35960934e-02
5.47083437e-01 2.86277294e-01 2.09940150e-01 -3.98851573e-01
-7.19659746e-01 -8.07392597e-01 1.02835717e-02 8.51598859e-01
-5.75257421e-01 5.18153310e-01 -3.95802200e-01 3.34540635e-01
3.79070044e-01 2.86602855e-01 4.54573721e-01 1.21559954e+00
-5.48221350e-01 -8.60283911e-01 -6.87116385e-01 2.81487525e-01
-2.83912599e-01 2.81724095e-01 -3.07435095e-01 -5.25303856e-02
4.68109727e-01 1.44458067e+00 -1.43157288e-01 -5.20573795e-01
2.64138401e-01 -8.80127177e-02 3.73196676e-02 -2.44210392e-01
-7.70611823e-01 2.01317891e-02 1.98784441e-01 -7.72665262e-01
-9.81173635e-01 -1.05583298e+00 -1.04295409e+00 -8.82074833e-02
5.78252785e-02 1.17619550e-02 1.38896942e+00 1.03748333e+00
9.60867226e-01 -1.00172378e-01 1.19362819e+00 -7.93243289e-01
1.85673058e-01 -3.46306324e-01 -4.61310863e-01 7.06040144e-01
1.87784925e-01 -2.76658773e-01 -8.77960026e-02 5.55461466e-01] | [7.9733734130859375, -1.5317692756652832] |
4cac64c6-3dd8-49f4-9b92-254fbc50a06e | granger-causality-based-hierarchical-time | 2104.04206 | null | https://arxiv.org/abs/2104.04206v1 | https://arxiv.org/pdf/2104.04206v1.pdf | Granger Causality Based Hierarchical Time Series Clustering for State Estimation | Clustering is an unsupervised learning technique that is useful when working with a large volume of unlabeled data. Complex dynamical systems in real life often entail data streaming from a large number of sources. Although it is desirable to use all source variables to form accurate state estimates, it is often impractical due to large computational power requirements, and sufficiently robust algorithms to handle these cases are not common. We propose a hierarchical time series clustering technique based on symbolic dynamic filtering and Granger causality, which serves as a dimensionality reduction and noise-rejection tool. Our process forms a hierarchy of variables in the multivariate time series with clustering of relevant variables at each level, thus separating out noise and less relevant variables. A new distance metric based on Granger causality is proposed and used for the time series clustering, as well as validated on empirical data sets. Experimental results from occupancy detection and building temperature estimation tasks show fidelity to the empirical data sets while maintaining state-prediction accuracy with substantially reduced data dimensionality. | ['Soumik Sarkar', 'Gregor P. Henze', 'Margarite Jacoby', 'Homagni Saha', 'Sin Yong Tan'] | 2021-04-09 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [ 1.52719840e-01 -5.28874636e-01 -1.74012005e-01 -2.34169289e-01
-3.31623495e-01 -5.66398740e-01 3.81271690e-01 5.47076464e-01
-3.19734007e-01 7.41962314e-01 6.30917773e-02 -2.57503033e-01
-6.93652987e-01 -9.22138274e-01 -6.99647516e-02 -9.16823924e-01
-7.14277804e-01 6.13328636e-01 3.44048381e-01 -2.01902054e-02
1.85729358e-02 6.31112158e-01 -2.01968503e+00 -3.77689987e-01
9.73335147e-01 9.09688830e-01 2.69432757e-02 5.52972138e-01
-8.35790560e-02 7.00195491e-01 -6.58811212e-01 4.70572352e-01
2.30018899e-01 -5.59736133e-01 -2.91833997e-01 9.62270424e-02
-3.75506818e-01 1.40592664e-01 -1.34626284e-01 8.65070701e-01
2.71591723e-01 5.76764822e-01 6.05959237e-01 -1.67228520e+00
-2.73518804e-02 4.92425323e-01 -4.78962511e-01 2.24813357e-01
9.46825445e-02 -1.08297832e-01 6.89813197e-01 -3.39859098e-01
2.39957586e-01 1.10391700e+00 6.45129263e-01 2.31984869e-01
-1.69648802e+00 -8.14740539e-01 1.32413164e-01 1.80247888e-01
-1.60534084e+00 -2.33011290e-01 9.47937191e-01 -5.90940237e-01
8.01324844e-01 4.69452709e-01 7.81557858e-01 6.96642280e-01
1.90233260e-01 3.79831463e-01 1.07676947e+00 -2.06797510e-01
8.78953457e-01 -1.81584924e-01 2.56962717e-01 2.67980874e-01
1.38423309e-01 1.70020312e-01 -3.25552106e-01 -6.38106763e-01
4.61330265e-01 4.84173626e-01 -4.65993620e-02 -4.30758685e-01
-1.06772113e+00 7.90750742e-01 4.95697819e-02 2.03575209e-01
-4.36195582e-01 9.41208676e-02 4.69114274e-01 5.17197132e-01
5.09395838e-01 1.70496240e-01 -4.92150724e-01 -3.83250892e-01
-1.02110183e+00 1.13223940e-01 9.14002895e-01 7.50008881e-01
6.28588557e-01 4.09591436e-01 3.99168938e-01 7.08424151e-01
2.29761750e-01 8.55487943e-01 4.91178095e-01 -9.19025481e-01
1.15128281e-02 8.95109177e-01 7.49845132e-02 -1.27387643e+00
-6.84565067e-01 -7.03636464e-03 -1.26045287e+00 2.44069308e-01
2.86754549e-01 -1.29879102e-01 -8.53177249e-01 1.76956558e+00
4.92991626e-01 5.55670917e-01 -5.14680229e-04 7.39531577e-01
-7.51203811e-03 7.65388370e-01 -7.73972720e-02 -9.74672139e-01
8.33201587e-01 -1.10724121e-01 -9.46065724e-01 1.08220533e-01
4.92839068e-01 -2.27617472e-01 6.26811922e-01 5.63332498e-01
-5.37246883e-01 -5.15204787e-01 -8.56805205e-01 6.13274515e-01
-4.36876923e-01 -3.24223995e-01 8.17942619e-01 6.95681214e-01
-8.48334908e-01 7.02265739e-01 -1.40749383e+00 -4.59085464e-01
-5.78977838e-02 5.28897643e-01 -1.93121831e-03 4.16616321e-01
-1.28497791e+00 6.34709537e-01 4.34167147e-01 1.02644868e-01
-7.84258664e-01 -3.16809744e-01 -7.43526220e-01 -7.14285970e-02
2.62849957e-01 -2.22854719e-01 7.23322332e-01 -6.20922863e-01
-1.17954779e+00 5.98544665e-02 -2.46125668e-01 -6.03799939e-01
2.67538548e-01 8.63398891e-03 -9.39693451e-01 6.30914196e-02
2.87687648e-02 -8.18499625e-02 1.08555710e+00 -1.00695562e+00
-7.19828546e-01 -5.41313469e-01 -5.50175667e-01 -6.57446533e-02
-5.71334004e-01 -2.14936405e-01 1.11673512e-02 -4.33538079e-01
5.73420525e-01 -9.98779833e-01 -6.91367507e-01 -5.17883956e-01
-1.26353875e-01 -1.47756115e-01 1.29002976e+00 -4.43830132e-01
1.78325844e+00 -1.98790443e+00 2.80984193e-01 7.68056929e-01
2.85109282e-01 -1.35706663e-01 3.36508840e-01 7.90916622e-01
-2.58719206e-01 -5.76439686e-02 -5.30042231e-01 4.29048277e-02
-1.76859170e-01 4.52025115e-01 -4.97029245e-01 6.75924838e-01
1.53078824e-01 1.48994923e-01 -1.15387905e+00 -3.70255977e-01
6.10808611e-01 2.08812416e-01 -2.82207757e-01 1.16725788e-02
-7.40370005e-02 4.63221312e-01 -5.81073344e-01 4.80559736e-01
3.18891406e-01 -1.50133818e-01 2.56156236e-01 4.60168719e-02
-1.61173120e-01 -6.71752542e-02 -1.65538347e+00 1.27915049e+00
-1.91286191e-01 4.42884266e-01 -1.34310767e-01 -1.30088043e+00
1.11746502e+00 2.91991621e-01 1.23103905e+00 -3.65718335e-01
2.13605747e-01 -1.33359641e-01 7.77982250e-02 -5.77627301e-01
4.57470626e-01 -8.84605199e-02 -5.29942751e-01 5.66808999e-01
-3.14009875e-01 -2.05681801e-01 4.28454220e-01 2.83341825e-01
1.39461231e+00 -1.47015601e-01 3.76211643e-01 -4.88909513e-01
4.74090725e-01 3.17585140e-01 1.00298786e+00 3.81446272e-01
-2.64994383e-01 7.87066072e-02 4.14429307e-01 -3.96288395e-01
-9.98067915e-01 -1.06950307e+00 -4.52685468e-02 6.82064235e-01
1.63063437e-01 -6.63268566e-01 -3.36986840e-01 -1.94735527e-01
5.14134876e-02 5.64894974e-01 -5.88301420e-01 -2.85859734e-01
-2.71431893e-01 -9.98941362e-01 3.87952387e-01 4.75457788e-01
1.10647500e-01 -6.78658903e-01 -7.72874415e-01 5.16767919e-01
-1.14102013e-01 -6.86539352e-01 1.28207564e-01 6.74703181e-01
-1.03996432e+00 -1.00405872e+00 1.37743652e-01 -5.50974250e-01
6.92093432e-01 1.31822675e-01 7.39445448e-01 2.87874998e-03
-4.20901716e-01 3.16281289e-01 -4.38438386e-01 -2.28479519e-01
-4.06374335e-01 -2.05637679e-01 6.97789311e-01 6.23925356e-03
6.55767024e-01 -8.75552654e-01 -3.05221260e-01 3.67143333e-01
-9.24601793e-01 -3.90537560e-01 -1.68804284e-02 7.33415842e-01
5.53523004e-01 9.67398226e-01 6.72866881e-01 -4.62564349e-01
5.98384857e-01 -5.96542358e-01 -8.03229153e-01 -1.60012498e-01
-6.14059091e-01 1.29397796e-03 1.02019119e+00 -6.38844967e-01
-6.27498269e-01 3.61229151e-01 3.93200457e-01 -5.99975049e-01
-3.35399568e-01 5.88336051e-01 3.57796252e-02 4.51189995e-01
7.02454150e-01 3.48308504e-01 8.06726888e-02 -2.43864506e-01
2.24451020e-01 7.06872940e-01 4.67432141e-01 -4.02100682e-01
1.05752754e+00 7.76263058e-01 2.18038335e-01 -1.22861814e+00
-2.20949367e-01 -7.04295278e-01 -9.05147076e-01 -3.80782217e-01
5.83335936e-01 -7.39886642e-01 -8.48697066e-01 2.88185477e-01
-4.79541391e-01 -1.07733399e-01 -3.41612965e-01 6.43033743e-01
-4.70881671e-01 3.27097684e-01 -4.57147419e-01 -1.32584238e+00
7.87577331e-02 -7.45620251e-01 7.53807485e-01 7.63846049e-03
-5.56967080e-01 -1.05551767e+00 3.11831504e-01 -2.17812568e-01
7.48410746e-02 7.58831441e-01 7.70092368e-01 -6.99393213e-01
-1.46951482e-01 -3.48595262e-01 4.63739872e-01 4.58750241e-02
4.90522981e-01 4.34395403e-01 -7.47323811e-01 -4.37999159e-01
3.26300979e-01 -3.74471620e-02 5.74365616e-01 3.69542778e-01
9.83494580e-01 -1.94699410e-02 -5.65274477e-01 8.18574950e-02
1.32755005e+00 5.64772785e-01 2.65953034e-01 2.45432602e-03
7.19002783e-01 7.09427595e-01 7.85899162e-01 7.91245401e-01
1.94259942e-01 4.42963913e-02 2.14348570e-01 1.76620364e-01
6.99190259e-01 -1.04872726e-01 3.47026438e-01 1.42570484e+00
1.14064775e-01 2.07363181e-02 -1.22069669e+00 6.93491638e-01
-2.09377909e+00 -1.28522885e+00 -5.84728956e-01 2.56450057e+00
6.35280430e-01 1.32074401e-01 3.67342949e-01 9.04599428e-01
7.22455978e-01 9.82599519e-03 -7.44244039e-01 -1.27495527e-01
1.33372098e-01 -2.72563901e-02 5.23051977e-01 2.86716491e-01
-1.20166659e+00 5.27337968e-01 6.31143379e+00 5.55736721e-01
-1.00913644e+00 -1.17335185e-01 2.99769789e-01 -1.55497134e-01
9.32558849e-02 4.35993597e-02 -4.80448663e-01 7.62639642e-01
1.42219627e+00 -3.28167498e-01 4.52924252e-01 8.82627547e-01
9.07084405e-01 -3.08916748e-01 -9.60780501e-01 1.08280253e+00
-3.12434077e-01 -7.71154821e-01 -5.26915431e-01 3.73331122e-02
7.56972551e-01 -9.43465754e-02 -3.15250546e-01 2.13504732e-01
8.12822342e-01 -7.62771547e-01 3.09953988e-01 4.01419044e-01
2.89676875e-01 -1.02020741e+00 2.11727276e-01 6.77151859e-01
-1.68676555e+00 -4.41328377e-01 -1.36092797e-01 -5.28790355e-01
2.22275212e-01 8.60111237e-01 -4.59799647e-01 2.96115577e-01
8.35974991e-01 9.26666200e-01 -2.65286446e-01 8.70194376e-01
1.03703365e-01 8.20952415e-01 -9.26575184e-01 -7.38837197e-02
-9.36047882e-02 -4.31629717e-01 4.83531922e-01 7.21102774e-01
2.19623342e-01 2.01372728e-01 5.44833362e-01 4.89512831e-01
5.86173654e-01 -1.75737783e-01 -6.66959763e-01 -1.68587685e-01
9.51815248e-01 1.25289917e+00 -1.31051230e+00 -4.26080167e-01
-1.72842950e-01 6.84417725e-01 -1.54441237e-01 3.58581185e-01
-9.39927757e-01 -3.71624172e-01 6.61075950e-01 9.61629581e-03
5.10422513e-02 -7.16233552e-01 -1.62740886e-01 -1.12832952e+00
-2.17704147e-01 -7.78740108e-01 5.91690421e-01 -2.01976970e-01
-1.33838701e+00 3.42197984e-01 1.61274299e-01 -1.82329631e+00
-6.67493582e-01 -3.28009129e-02 -5.42860150e-01 5.23984313e-01
-9.03209627e-01 -5.10584354e-01 -1.62166715e-01 9.56676722e-01
4.18488890e-01 4.03267592e-02 9.10324395e-01 2.93183923e-01
-6.67136073e-01 -1.04198217e-01 6.93501711e-01 -6.85477480e-02
3.33929747e-01 -1.30477345e+00 2.00801134e-01 1.03463364e+00
-1.92580625e-01 7.84005821e-01 8.87729287e-01 -9.92241979e-01
-1.71910131e+00 -1.11467063e+00 4.55781043e-01 -3.39998156e-01
1.20810068e+00 -4.89489973e-01 -1.02908003e+00 3.61782640e-01
-4.20781195e-01 -2.04455286e-01 8.80814970e-01 1.55865669e-01
4.26570922e-02 -1.01984255e-01 -1.13847971e+00 5.00480711e-01
8.29168379e-01 -5.47074020e-01 -6.26547933e-01 1.43487707e-01
4.05624717e-01 1.48478717e-01 -1.24195123e+00 2.28945643e-01
2.37302125e-01 -6.41504705e-01 6.82295382e-01 -3.57903242e-01
-3.31748500e-02 -8.42624664e-01 -1.32135719e-01 -1.30645251e+00
-4.57553208e-01 -7.67367184e-01 -3.00206602e-01 1.32353175e+00
5.74204437e-02 -5.97919524e-01 7.51094520e-01 7.61767626e-01
4.23377842e-01 -4.59665135e-02 -9.93979871e-01 -1.03878677e+00
-3.25167537e-01 -7.93274105e-01 4.67656732e-01 1.34961987e+00
3.85661870e-01 4.26095515e-01 -3.56788367e-01 4.46775287e-01
1.01512086e+00 4.23361391e-01 7.71577179e-01 -1.71273756e+00
-1.20807126e-01 -2.58428782e-01 -6.47391915e-01 -4.45501387e-01
-3.14856358e-02 -5.18703878e-01 2.02195257e-01 -1.20869291e+00
-2.22259402e-01 -5.05837262e-01 -2.90636778e-01 2.96713054e-01
2.55877175e-03 -7.30730295e-02 -1.44297779e-01 4.89430279e-01
-4.19228107e-01 6.89311922e-01 4.03652847e-01 -4.15691622e-02
-7.29554832e-01 6.87442943e-02 6.18040152e-02 8.02365720e-01
8.74953032e-01 -4.92879331e-01 -7.94718981e-01 1.69265091e-01
-8.63088667e-03 2.75933176e-01 1.85892925e-01 -1.18698645e+00
2.72838175e-01 -3.57274979e-01 4.77169245e-01 -1.04951191e+00
3.31324726e-01 -1.40074337e+00 6.27404988e-01 6.06060684e-01
-8.98617581e-02 1.42900214e-01 1.59548670e-01 8.93460691e-01
-1.36356413e-01 3.25913787e-01 6.36985838e-01 1.42929927e-01
-9.18580711e-01 2.55750269e-01 -7.26020753e-01 -1.47942573e-01
1.26653802e+00 -3.09268206e-01 9.23467577e-02 -3.42427164e-01
-9.98342156e-01 6.23923838e-01 4.27570015e-01 5.63400447e-01
4.93972301e-01 -1.42233527e+00 -2.94648081e-01 4.05990273e-01
-1.14677409e-02 -2.17985600e-01 1.23541176e-01 8.57445896e-01
-1.37986913e-01 3.20488624e-02 -6.71438202e-02 -9.67221379e-01
-1.13482261e+00 9.02678549e-01 -1.23630226e-01 4.28132974e-02
-6.50924981e-01 1.62215263e-01 -5.31177342e-01 -2.71837294e-01
2.44184390e-01 -5.46474814e-01 -1.58321634e-01 5.49925148e-01
6.47608399e-01 6.53459013e-01 -1.38322547e-01 -6.97876394e-01
-4.18537438e-01 4.32321906e-01 3.77939254e-01 -1.13012247e-01
1.51289451e+00 -4.96489793e-01 -2.22570330e-01 9.97293591e-01
1.04031849e+00 -4.65760648e-01 -1.19745123e+00 -2.44962856e-01
4.70523894e-01 -3.75510663e-01 4.52625379e-02 -7.67025501e-02
-8.59363437e-01 6.60725594e-01 8.83827806e-01 9.77813542e-01
1.44776571e+00 -3.56281489e-01 5.49060345e-01 4.90506768e-01
6.06269419e-01 -1.49215746e+00 -3.57154340e-01 4.16653335e-01
1.96874902e-01 -1.06024539e+00 -9.98249352e-02 -6.02403060e-02
-3.16073567e-01 8.30025971e-01 3.39164525e-01 -3.75422031e-01
9.89913464e-01 6.72904909e-01 -1.57367796e-01 -1.40426174e-01
-8.79831731e-01 -3.15921605e-01 -3.50152165e-01 7.35809803e-01
-1.31476521e-02 1.79962024e-01 -2.66690910e-01 2.98908830e-01
-1.19142719e-01 -1.36096582e-01 4.06500161e-01 1.28068578e+00
-5.55300593e-01 -8.33499193e-01 -8.11558723e-01 7.52337813e-01
-1.75771907e-01 2.23512322e-01 -2.22538680e-01 6.64579570e-01
-1.32586047e-01 1.30638492e+00 2.56593227e-01 -6.36852920e-01
2.96040326e-01 1.50143400e-01 -1.93132997e-01 -4.45440114e-01
-4.55482453e-01 4.38914239e-01 -2.58219987e-01 -6.57874286e-01
-5.77394307e-01 -7.58005559e-01 -1.62283969e+00 -4.66977477e-01
-3.56806785e-01 3.86584550e-01 5.08682191e-01 8.17876458e-01
2.14706734e-01 6.17158830e-01 1.11449695e+00 -8.16966891e-01
-2.91057289e-01 -7.42920876e-01 -1.01653850e+00 5.82470119e-01
3.97929847e-01 -8.36579800e-01 -6.45646513e-01 3.69643539e-01] | [7.19122838973999, 3.3270130157470703] |
0c4eb3d3-3b1f-45d8-9218-d4c48e82555c | sample-size-in-arabic-authorship-verification | null | null | https://aclanthology.org/W19-7412 | https://aclanthology.org/W19-7412.pdf | Sample Size in Arabic Authorship Verification | null | ['Hossam Ahmed'] | 2019-09-01 | null | null | null | ws-2019-9 | ['authorship-verification'] | ['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] | [-7.344784736633301, 3.6997811794281006] |
23b40996-6a52-4ae9-a71a-86d7ea233298 | beyond-spectral-gap-the-role-of-the-topology | 2206.03093 | null | https://arxiv.org/abs/2206.03093v2 | https://arxiv.org/pdf/2206.03093v2.pdf | Beyond spectral gap: The role of the topology in decentralized learning | In data-parallel optimization of machine learning models, workers collaborate to improve their estimates of the model: more accurate gradients allow them to use larger learning rates and optimize faster. We consider the setting in which all workers sample from the same dataset, and communicate over a sparse graph (decentralized). In this setting, current theory fails to capture important aspects of real-world behavior. First, the 'spectral gap' of the communication graph is not predictive of its empirical performance in (deep) learning. Second, current theory does not explain that collaboration enables larger learning rates than training alone. In fact, it prescribes smaller learning rates, which further decrease as graphs become larger, failing to explain convergence in infinite graphs. This paper aims to paint an accurate picture of sparsely-connected distributed optimization when workers share the same data distribution. We quantify how the graph topology influences convergence in a quadratic toy problem and provide theoretical results for general smooth and (strongly) convex objectives. Our theory matches empirical observations in deep learning, and accurately describes the relative merits of different graph topologies. | ['Martin Jaggi', 'Hadrien Hendrikx', 'Thijs Vogels'] | 2022-06-07 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-4.28085744e-01 3.87306660e-01 -1.49881050e-01 -2.64751643e-01
-4.43615079e-01 -4.38656986e-01 3.25719267e-01 5.41393340e-01
-5.85549176e-01 8.06197107e-01 1.08836465e-01 -2.79527962e-01
-4.92822558e-01 -6.55991733e-01 -1.02415395e+00 -7.32400715e-01
-5.96915185e-01 9.48121369e-01 -3.00976008e-01 3.80514562e-02
4.25945327e-04 3.58306557e-01 -7.98810244e-01 -1.57401279e-01
6.00976110e-01 5.32666385e-01 1.67424291e-01 1.14911652e+00
1.05269924e-02 9.76071537e-01 -7.33450413e-01 -4.47960943e-01
5.85271597e-01 -4.97859001e-01 -9.88734782e-01 1.96316987e-01
5.65843761e-01 -3.96121815e-02 -6.37666047e-01 1.14281690e+00
4.02080089e-01 5.47803147e-03 5.84068596e-01 -1.38991702e+00
-7.48566687e-01 1.19632602e+00 -8.35583329e-01 3.75469267e-01
-3.86380792e-01 1.53398842e-01 1.27860391e+00 -2.37733245e-01
5.40173650e-01 1.27810669e+00 9.48737502e-01 3.72301370e-01
-1.42837107e+00 -3.15421671e-01 3.36949825e-01 -3.65784407e-01
-1.21691298e+00 -3.18441212e-01 5.20492613e-01 -3.87855738e-01
8.32222998e-01 4.31056619e-02 8.25474441e-01 6.45602465e-01
3.97485644e-01 3.93803507e-01 9.11694169e-01 -1.70962706e-01
1.75383866e-01 9.55726653e-02 1.83570981e-01 1.04799163e+00
8.75126481e-01 -9.70394388e-02 -1.04900587e+00 -2.83161402e-01
8.67138982e-01 1.54279858e-01 -4.12618160e-01 -5.30375004e-01
-1.09564495e+00 9.14092660e-01 7.84766853e-01 3.32943648e-01
-4.72310156e-01 9.75332618e-01 1.74223110e-01 1.04055333e+00
7.49245226e-01 5.15271783e-01 -5.63458025e-01 -1.56997100e-01
-6.07544839e-01 1.58521220e-01 1.24179029e+00 9.19894457e-01
1.14959097e+00 -1.46400258e-01 2.93837965e-01 3.39056611e-01
3.48211169e-01 3.56727660e-01 1.60382181e-01 -1.30560219e+00
8.42493951e-01 3.52678478e-01 3.49460505e-02 -1.22610629e+00
-8.42031479e-01 -9.14665461e-01 -1.07910705e+00 -2.66383141e-02
9.99446988e-01 -7.94309974e-01 -2.14744121e-01 1.81855142e+00
1.39171243e-01 -1.37169376e-01 -8.31900686e-02 1.11447275e+00
8.90804306e-02 4.49399948e-01 -3.72417063e-01 -1.24580488e-01
5.86870432e-01 -1.00232768e+00 -3.31298321e-01 -4.72171307e-01
1.18001199e+00 -3.49294275e-01 8.69834602e-01 3.32234293e-01
-1.12843168e+00 -9.35088024e-02 -6.90911591e-01 -8.45650658e-02
1.83343217e-02 -3.45689863e-01 1.00701439e+00 6.24050379e-01
-1.72918427e+00 1.08805954e+00 -1.00910139e+00 -2.49723494e-01
7.02334940e-01 6.80386424e-01 -2.12678090e-01 -9.25271660e-02
-5.84570110e-01 4.75709498e-01 3.61280367e-02 5.91499433e-02
-1.01466084e+00 -7.93257475e-01 -4.21900511e-01 2.14971572e-01
2.51869082e-01 -1.04105031e+00 1.17763174e+00 -1.40959001e+00
-1.16616821e+00 7.28623807e-01 1.82627365e-02 -6.38418317e-01
7.25704193e-01 -1.45169765e-01 4.51579094e-01 -1.02485515e-01
-2.35147372e-01 2.53163725e-01 7.43710816e-01 -1.10336602e+00
-3.23876321e-01 -7.53310800e-01 2.60483980e-01 4.60002512e-01
-6.03353500e-01 -4.56832439e-01 -1.19591225e-02 -1.36729613e-01
-2.61808217e-01 -1.00307941e+00 -5.57483375e-01 3.80326547e-02
-2.84300506e-01 -1.81829661e-01 3.53929788e-01 -1.53504878e-01
8.33706379e-01 -1.89546311e+00 3.69440764e-01 5.58889568e-01
1.19451594e+00 -4.80850369e-01 -2.92532533e-01 7.25735843e-01
3.95802706e-01 2.40648106e-01 4.96103726e-02 -7.70603120e-01
8.43555704e-02 2.72543758e-01 2.12716997e-01 1.18785906e+00
-3.22802305e-01 1.04769242e+00 -8.95872295e-01 -4.23406392e-01
-3.68584007e-01 1.78807050e-01 -8.33595037e-01 -1.50505513e-01
-2.57506091e-02 3.18620890e-01 -7.92661428e-01 -1.06956758e-01
4.55019385e-01 -1.13424623e+00 5.94111800e-01 4.32195961e-01
3.03400487e-01 3.26979272e-02 -1.09497011e+00 1.65997744e+00
-4.28762972e-01 1.03773975e+00 8.91630292e-01 -1.44375312e+00
5.81228554e-01 -5.75883165e-02 5.18023849e-01 -2.38916546e-01
1.05367012e-01 1.52878240e-01 2.42711544e-01 -3.33274275e-01
8.68585780e-02 -6.31076889e-03 3.68683368e-01 9.12337899e-01
4.32485566e-02 -1.04695283e-01 -9.16430280e-02 6.38925791e-01
1.43040550e+00 -9.09552276e-01 -1.41491428e-01 -8.98979008e-01
-2.03509569e-01 -1.23393536e-01 1.36678264e-01 1.25891912e+00
-4.11647093e-03 3.08969676e-01 1.01301253e+00 -4.72018510e-01
-1.16867495e+00 -7.29487360e-01 4.09198791e-01 1.37906790e+00
2.96126634e-01 -4.01850224e-01 -9.49577272e-01 -6.74332142e-01
5.02291381e-01 -1.17203221e-01 -7.37660587e-01 1.13600321e-01
-4.44664955e-01 -9.52910900e-01 1.04957104e-01 5.25735468e-02
6.19245209e-02 -4.76211429e-01 -1.88131824e-01 2.55886149e-02
2.10102484e-01 -7.22947061e-01 -8.74252796e-01 2.05711856e-01
-1.19570374e+00 -1.28578448e+00 -8.10740888e-01 -6.94099605e-01
9.77757931e-01 5.21276474e-01 1.37884152e+00 6.55091882e-01
-1.44554198e-01 9.11596656e-01 1.15978569e-01 -4.28040981e-01
-3.64731401e-01 6.59909844e-01 3.57427411e-02 8.42950717e-02
4.92392341e-03 -6.52779937e-01 -7.96722293e-01 1.13997068e-02
-6.81245327e-01 -1.40563414e-01 6.08706295e-01 6.91113949e-01
7.42252767e-02 -1.90217253e-02 5.66343248e-01 -1.16480982e+00
1.31603205e+00 -7.14411557e-01 -6.37781382e-01 3.11874032e-01
-9.12790954e-01 4.06401664e-01 9.64937449e-01 -4.32609379e-01
-5.42578459e-01 -1.80888563e-01 7.12681770e-01 -2.80652523e-01
3.38730216e-01 6.39600754e-01 6.71750367e-01 -4.28423941e-01
9.21899140e-01 -1.56882778e-01 5.40335417e-01 -4.48652059e-01
3.75770450e-01 4.19132024e-01 -7.68225417e-02 -8.10857594e-01
6.16226256e-01 5.57349145e-01 3.30505729e-01 -8.06997120e-01
-6.24571204e-01 -1.35641620e-01 -3.24836910e-01 -1.93104878e-01
2.89595157e-01 -9.59195435e-01 -1.21620750e+00 3.39254409e-01
-1.12120032e+00 -1.00382543e+00 -5.06003559e-01 6.12370968e-01
-4.27149653e-01 2.75293678e-01 -8.77730966e-01 -8.97486627e-01
-2.46634543e-01 -8.55163991e-01 8.63001287e-01 2.18413711e-01
2.05172062e-01 -1.62030506e+00 1.89734578e-01 2.04964340e-01
7.28223085e-01 -8.75394121e-02 5.82140982e-01 -5.40044546e-01
-7.53166974e-01 1.18977897e-01 -2.11188719e-01 8.81007165e-02
-1.96908176e-01 -2.26516336e-01 -5.93051851e-01 -7.27176130e-01
-6.66832377e-04 -5.03288865e-01 8.99401188e-01 6.95626915e-01
1.25058854e+00 -7.76096225e-01 -3.66108537e-01 7.30099142e-01
1.36221111e+00 -6.52914405e-01 -2.11656824e-01 -1.76690906e-01
8.53233159e-01 7.59058118e-01 -2.12745979e-01 6.37602746e-01
6.16884053e-01 1.26060605e-01 5.95978022e-01 -1.00473002e-01
-2.08335817e-02 -2.13605806e-01 2.84640700e-01 9.25825417e-01
-2.41409019e-01 -4.39337015e-01 -1.10050333e+00 3.79935056e-01
-1.98222685e+00 -5.16252935e-01 -3.02112043e-01 2.12059474e+00
8.59194458e-01 3.62186693e-02 4.08769339e-01 -5.06469786e-01
7.81898558e-01 1.62976533e-01 -8.09408188e-01 -1.30721465e-01
-9.98204947e-02 -1.21788204e-01 1.18829203e+00 9.64963198e-01
-4.77366358e-01 5.70004165e-01 6.89287710e+00 6.14961982e-01
-1.05399740e+00 4.86091197e-01 1.17997849e+00 -6.23168707e-01
-4.49431032e-01 -1.21061966e-01 -4.40180212e-01 3.08314890e-01
9.64112163e-01 -4.70060408e-01 1.26589572e+00 7.15772748e-01
9.23106372e-02 8.92770439e-02 -1.42761743e+00 1.13515365e+00
-1.50163919e-01 -1.67434859e+00 -4.06212598e-01 6.64981008e-01
1.17140901e+00 7.34697640e-01 1.88505948e-01 -1.49971858e-01
9.54186141e-01 -1.24393511e+00 4.53130245e-01 5.14287591e-01
3.88590604e-01 -7.03858674e-01 4.80766833e-01 8.15821290e-01
-6.29348099e-01 -3.14674795e-01 -6.18832827e-01 -4.30463165e-01
-1.73296317e-01 8.94004405e-01 -6.40193522e-01 8.36592764e-02
5.05831242e-01 7.47526646e-01 -6.10627234e-01 7.67623127e-01
3.86844099e-01 5.35509706e-01 -6.52767718e-01 -4.56536114e-01
2.93255776e-01 -4.59834456e-01 4.08623695e-01 9.53447640e-01
8.82963836e-02 -8.89745280e-02 2.21473619e-01 8.40229928e-01
-7.20583737e-01 1.29632086e-01 -7.59384274e-01 -4.44282651e-01
2.46053562e-01 1.02612722e+00 -7.43151009e-01 -9.15280655e-02
-4.71447021e-01 7.03191638e-01 1.09963226e+00 6.82617486e-01
-3.34199131e-01 -2.20738128e-02 6.32676244e-01 2.19336748e-01
-4.43615839e-02 -5.96652746e-01 -3.63912880e-01 -1.25846255e+00
8.21171403e-02 -6.53694689e-01 2.92685330e-01 -3.09359524e-02
-1.68815351e+00 4.73922715e-02 -3.69059175e-01 -2.41616234e-01
1.77551042e-02 -4.30539370e-01 -5.45501113e-01 6.22182608e-01
-1.26446116e+00 -6.54565573e-01 -7.85392225e-02 5.60340822e-01
1.41484708e-01 -2.01502725e-01 2.43403196e-01 7.25450218e-02
-5.09607077e-01 5.64364254e-01 5.74217737e-01 2.94788945e-02
3.95939291e-01 -1.52989733e+00 3.20634037e-01 5.51338792e-01
5.80349505e-01 6.05454266e-01 7.17990279e-01 -4.98260796e-01
-2.04015875e+00 -8.86094749e-01 5.35557687e-01 -5.87369204e-01
9.53322113e-01 -4.10520643e-01 -5.94967127e-01 8.13115180e-01
1.46081239e-01 3.59665692e-01 3.07490140e-01 5.74016809e-01
-1.39871091e-01 -3.40924442e-01 -8.58542979e-01 2.69089967e-01
1.34096229e+00 -4.72047150e-01 4.62500185e-01 9.21485960e-01
6.42081916e-01 -3.08882296e-01 -8.77205729e-01 -4.11133885e-01
1.93296403e-01 -9.76532698e-01 6.07227266e-01 -8.16038847e-01
2.86050498e-01 4.61930186e-01 1.76723927e-01 -1.59131086e+00
-4.58489098e-02 -1.16164064e+00 -4.64154691e-01 5.32905042e-01
5.78763247e-01 -1.21401453e+00 1.20585823e+00 6.24943316e-01
2.26947308e-01 -9.34265137e-01 -8.94345045e-01 -7.29789555e-01
5.08656800e-01 1.11886472e-01 3.54219049e-01 1.10362744e+00
1.03907252e-03 3.51330608e-01 -1.57950759e-01 1.25494480e-01
8.99652481e-01 -5.91523647e-02 9.93152022e-01 -1.27343404e+00
-7.78706849e-01 -8.36355090e-01 -1.92240328e-01 -1.47562933e+00
2.23828301e-01 -1.06191623e+00 -3.11837435e-01 -1.27940166e+00
4.16088402e-01 -8.40221465e-01 -9.18164626e-02 1.19241543e-01
1.49247916e-02 -4.62549888e-02 3.02805603e-01 5.01470506e-01
-9.02837873e-01 2.92391270e-01 1.40570271e+00 -1.40771851e-01
-1.49834603e-01 1.58195943e-01 -1.00228214e+00 4.20466155e-01
8.46915007e-01 -6.46939278e-01 -4.75576937e-01 -1.10316956e+00
9.16570902e-01 8.49347115e-02 3.15888435e-01 -5.87403357e-01
7.25968421e-01 -2.53925622e-01 2.23534837e-01 4.15931404e-01
-5.38992509e-02 -7.32807934e-01 -4.26281914e-02 6.75996125e-01
-8.33075583e-01 3.56366575e-01 -4.55237597e-01 1.02891839e+00
3.21233273e-01 -7.55444393e-02 7.62024343e-01 -1.65555164e-01
2.37022772e-01 6.02173567e-01 -4.52015025e-04 6.31278396e-01
8.99027884e-01 2.68113047e-01 -4.70124722e-01 -9.58916366e-01
-6.33902073e-01 5.16769886e-01 6.78649068e-01 -9.34958383e-02
2.36748010e-01 -8.62488925e-01 -9.09906268e-01 -3.48527580e-02
-4.59719002e-01 2.87418872e-01 9.06204954e-02 1.15526927e+00
-8.39349210e-01 1.40631506e-02 4.13303047e-01 -6.09076917e-01
-8.73185098e-01 4.05332118e-01 8.02888095e-01 -2.94511020e-01
-5.04098535e-01 1.17094088e+00 2.42314190e-01 -3.11170995e-01
4.50904250e-01 -2.38817513e-01 4.48676258e-01 -2.96112359e-01
1.03788748e-01 5.03627837e-01 -1.84319392e-01 -1.22705167e-02
-3.42430100e-02 1.85292110e-01 -9.26666409e-02 -6.48988336e-02
1.45675159e+00 -3.16119105e-01 -3.21561635e-01 4.97629464e-01
1.47762454e+00 -4.80999425e-02 -1.47224438e+00 -4.49139029e-01
-2.35818639e-01 -5.42349875e-01 2.03989416e-01 -1.55808985e-01
-1.58513951e+00 8.54687810e-01 1.39642373e-01 9.24536228e-01
4.32947546e-01 3.87207896e-01 5.22862196e-01 7.92257965e-01
3.76514733e-01 -1.08323169e+00 2.26356789e-01 3.75218630e-01
6.88638926e-01 -1.23889804e+00 1.17675029e-01 1.19821504e-02
-4.40019250e-01 1.12320244e+00 4.02083993e-01 -3.88216227e-01
8.97064507e-01 2.29420081e-01 -3.32209259e-01 -6.32851422e-01
-1.03159916e+00 2.11436823e-01 -1.26291215e-01 5.07911742e-01
3.14668387e-01 1.36585966e-01 -7.07047712e-03 -1.92034170e-02
-4.48688537e-01 -4.64209944e-01 5.67960083e-01 3.45929980e-01
-5.31992018e-01 -6.96912766e-01 2.13307957e-03 9.68340218e-01
-6.11109614e-01 8.64299312e-02 -7.01614618e-01 6.02360845e-01
-4.00115967e-01 8.41260731e-01 2.76586086e-01 1.65794380e-02
-2.94695407e-01 -5.56799114e-01 6.72563791e-01 -6.63216710e-01
-6.86539888e-01 -2.64724195e-01 -1.17644481e-01 -6.49460196e-01
-3.04162800e-01 -4.43946123e-01 -8.83868933e-01 -1.16819036e+00
-2.60989815e-01 4.79224354e-01 6.75506592e-01 8.27780783e-01
6.30361021e-01 2.02342585e-01 7.90640891e-01 -8.06084752e-01
-1.12725079e+00 -6.66218817e-01 -9.30399597e-01 2.33308211e-01
8.00167620e-01 9.68365185e-03 -7.58511007e-01 -3.11417758e-01] | [6.400151252746582, 5.206604480743408] |
f90ea398-2b38-403b-a4bc-f240a9507e3d | opt-open-pre-trained-transformer-language | 2205.01068 | null | https://arxiv.org/abs/2205.01068v4 | https://arxiv.org/pdf/2205.01068v4.pdf | OPT: Open Pre-trained Transformer Language Models | Large language models, which are often trained for hundreds of thousands of compute days, have shown remarkable capabilities for zero- and few-shot learning. Given their computational cost, these models are difficult to replicate without significant capital. For the few that are available through APIs, no access is granted to the full model weights, making them difficult to study. We present Open Pre-trained Transformers (OPT), a suite of decoder-only pre-trained transformers ranging from 125M to 175B parameters, which we aim to fully and responsibly share with interested researchers. We show that OPT-175B is comparable to GPT-3, while requiring only 1/7th the carbon footprint to develop. We are also releasing our logbook detailing the infrastructure challenges we faced, along with code for experimenting with all of the released models. | ['Luke Zettlemoyer', 'Tianlu Wang', 'Anjali Sridhar', 'Punit Singh Koura', 'Daniel Simig', 'Kurt Shuster', 'Sam Shleifer', 'Myle Ott', 'Todor Mihaylov', 'Xi Victoria Lin', 'Xian Li', 'Mona Diab', 'Christopher Dewan', 'Shuohui Chen', 'Moya Chen', 'Mikel Artetxe', 'Naman Goyal', 'Stephen Roller', 'Susan Zhang'] | 2022-05-02 | null | null | null | null | ['stereotypical-bias-analysis'] | ['natural-language-processing'] | [ 1.22392094e-02 1.22380987e-01 -2.22157866e-01 9.29794312e-02
-9.90349531e-01 -6.74740076e-01 7.62540400e-01 -2.39872724e-01
-3.00217837e-01 8.19722772e-01 1.64485327e-03 -8.75969112e-01
9.47590992e-02 -7.05010474e-01 -8.57592702e-01 -4.03906882e-01
-1.84459001e-01 6.65043294e-01 2.56826341e-01 -1.84383169e-01
1.37135535e-01 -4.48352322e-02 -1.33210373e+00 -5.46934791e-02
9.47066724e-01 6.99951589e-01 5.65817416e-01 6.88618541e-01
-8.79469588e-02 6.60907626e-01 -1.99440494e-01 -5.66143155e-01
2.63991088e-01 -5.88204674e-02 -7.26289034e-01 -4.78235543e-01
1.80031300e-01 -4.16396201e-01 -4.36367393e-01 8.72705102e-01
7.55502701e-01 -1.34867072e-01 5.19693553e-01 -1.07749879e+00
-5.76157570e-01 8.59413981e-01 -2.19091445e-01 2.94029683e-01
1.20229018e-03 4.46567863e-01 1.10438550e+00 -9.22032177e-01
4.07003194e-01 8.89187098e-01 7.95970142e-01 5.84040403e-01
-1.07554853e+00 -7.95555949e-01 -2.25610420e-01 8.72370303e-02
-1.45133996e+00 -1.08795130e+00 1.42026410e-01 -4.34181124e-01
1.89535773e+00 2.23733578e-02 7.73704410e-01 1.44067585e+00
2.55450338e-01 2.81470537e-01 9.08624113e-01 -4.22367334e-01
2.55759060e-01 1.31480679e-01 -9.95443687e-02 9.57544684e-01
4.69543993e-01 5.20507582e-02 -5.36430180e-01 -2.78651685e-01
3.69010657e-01 -3.33636016e-01 2.10336894e-01 -2.67862767e-01
-1.06843090e+00 6.73601210e-01 3.48042734e-02 6.89802840e-02
-8.04346651e-02 5.66080093e-01 2.81330615e-01 9.38746482e-02
5.54608941e-01 5.51213145e-01 -8.25177372e-01 -8.09764564e-01
-9.29133236e-01 6.58700839e-02 1.05824065e+00 1.43645060e+00
6.69977427e-01 2.49000475e-01 9.46280509e-02 7.66997457e-01
1.60873592e-01 4.45125252e-01 3.03046227e-01 -9.98778403e-01
3.69549394e-01 1.86758891e-01 1.94551468e-01 -2.79432595e-01
-2.97274858e-01 -2.40945205e-01 -5.54256737e-01 -3.73818547e-01
2.64725626e-01 -3.46831322e-01 -8.80175233e-01 1.37406850e+00
-1.68528542e-01 2.85563231e-01 -1.04769692e-01 2.19822049e-01
6.27838016e-01 8.92039835e-01 1.84937701e-01 -7.43823424e-02
1.31518090e+00 -1.25199962e+00 -2.18763009e-01 -5.12291789e-01
6.62244618e-01 -7.98680544e-01 1.34188449e+00 5.19390762e-01
-1.31395888e+00 2.35343594e-02 -1.02857590e+00 -7.84317479e-02
-5.87488353e-01 -3.15975785e-01 9.88663256e-01 9.46069717e-01
-1.15472734e+00 7.77078986e-01 -1.11381781e+00 -6.65137708e-01
5.42168319e-01 1.56897396e-01 9.19321701e-02 -1.07865795e-01
-1.10684335e+00 1.24639058e+00 9.38952491e-02 -4.67774659e-01
-1.44831085e+00 -1.06634092e+00 -6.21352971e-01 2.91838288e-01
3.76319975e-01 -6.68052375e-01 1.50759459e+00 -9.55741182e-02
-1.53903162e+00 7.15480983e-01 -1.09994061e-01 -4.82610762e-01
4.44606632e-01 -1.58061087e-02 -4.01130736e-01 -2.56889492e-01
-1.52833313e-01 5.78696370e-01 4.53552425e-01 -6.61927283e-01
-3.81266117e-01 1.21710047e-01 1.32535055e-01 3.70975845e-02
-9.16422129e-01 3.08222294e-01 -7.63131618e-01 -3.49385619e-01
-3.35579067e-01 -9.32260215e-01 -4.07891989e-01 -9.52771008e-02
-5.20925879e-01 -1.70643814e-02 4.59531903e-01 -7.97790349e-01
1.29633391e+00 -1.57398176e+00 -5.62879220e-02 -3.20805490e-01
-2.43520625e-02 2.36169741e-01 -4.02867138e-01 6.24073148e-01
2.62073457e-01 6.52643025e-01 -2.78607011e-01 -3.76885235e-01
4.15782422e-01 1.40353143e-01 -8.32139626e-02 2.55777419e-01
8.13528597e-02 9.83155131e-01 -9.28125024e-01 -4.18902487e-01
3.10937706e-02 4.65424001e-01 -5.90852916e-01 -3.84845305e-03
-6.25198126e-01 1.22404747e-01 -3.26451302e-01 8.31854224e-01
5.30643582e-01 -5.23683310e-01 3.79492223e-01 1.41902342e-01
-2.37140343e-01 8.09640944e-01 -6.16395831e-01 2.02210879e+00
-6.74583554e-01 3.97393495e-01 4.18250822e-02 -7.55125880e-01
4.46599752e-01 2.82213867e-01 4.43030685e-01 -7.82592237e-01
-5.94762079e-02 3.15252990e-01 -2.09826380e-02 -2.48915687e-01
4.35792536e-01 -2.16802265e-02 -2.33934104e-01 8.06989968e-01
1.78803071e-01 -4.22838539e-01 5.95604658e-01 3.04660708e-01
1.46355164e+00 1.16519772e-01 3.02877370e-02 -3.92721593e-01
-1.83785245e-01 -6.14575818e-02 6.65950537e-01 6.31260037e-01
-1.43864661e-01 3.49748194e-01 5.95486522e-01 -3.79945457e-01
-1.62100077e+00 -9.50623095e-01 -8.48528519e-02 1.23088074e+00
-3.77910703e-01 -8.48603904e-01 -7.02832103e-01 -2.50453532e-01
-2.14780927e-01 8.74333799e-01 2.86172926e-02 -4.14078981e-02
-1.99279815e-01 -9.69668925e-01 8.41827929e-01 3.44330430e-01
3.18429112e-01 -4.93728727e-01 -4.27012771e-01 2.44256228e-01
9.60563421e-02 -8.13833416e-01 -5.51444173e-01 5.20477831e-01
-7.74204373e-01 -7.22745717e-01 -6.00198030e-01 -5.73533654e-01
2.09993929e-01 1.14632405e-01 1.42556000e+00 2.64875054e-01
-3.82637352e-01 5.96511401e-02 -1.28755748e-01 -5.58508992e-01
-5.18388212e-01 5.69252431e-01 1.72751039e-01 -8.58646870e-01
1.91778228e-01 -8.25650632e-01 -3.37770253e-01 -1.04384227e-02
-4.12166536e-01 4.07158047e-01 4.95180994e-01 6.36188328e-01
2.87319273e-01 2.57376395e-02 2.80808896e-01 -8.89767528e-01
5.65422237e-01 -5.83274543e-01 -8.43831658e-01 5.45966804e-01
-1.10366619e+00 5.50169090e-04 3.59573632e-01 -2.02997610e-01
-9.05786693e-01 -9.16973874e-02 -2.11546734e-01 -1.69813544e-01
2.96356171e-01 4.68758255e-01 2.31057797e-02 -1.18593901e-01
5.08875310e-01 1.54337943e-01 -2.60352850e-01 -6.98441207e-01
4.54189539e-01 6.00192785e-01 7.64129832e-02 -7.81252444e-01
8.76216054e-01 -1.28283158e-01 -3.12942833e-01 -6.77479804e-01
-5.79057574e-01 -1.13166440e-02 -3.44659388e-01 1.95464745e-01
6.63435161e-01 -1.31414819e+00 -5.26902735e-01 5.16051233e-01
-1.01226377e+00 -8.96731675e-01 -7.84761757e-02 6.54367507e-02
-4.29276735e-01 1.18339762e-01 -9.07298386e-01 -7.25501597e-01
-6.80930257e-01 -1.12622762e+00 8.17576766e-01 1.05495580e-01
-4.87717465e-02 -1.05141759e+00 2.04859838e-01 4.74188596e-01
9.80885386e-01 -3.50216120e-01 1.41924751e+00 -4.82255101e-01
-6.78744614e-01 7.90638104e-03 -2.08913326e-01 1.65925726e-01
-1.79744139e-01 3.28503460e-01 -1.08166754e+00 -4.81977791e-01
-5.25751829e-01 -5.20227134e-01 7.30419517e-01 2.57271886e-01
1.19910204e+00 -2.57498533e-01 -5.47363818e-01 6.49331570e-01
1.33455968e+00 -8.11652765e-02 5.68988740e-01 1.69746801e-01
5.95729351e-01 -3.42386812e-02 1.23928539e-01 6.37124836e-01
5.79507351e-01 4.78752673e-01 2.82160550e-01 2.09242478e-02
-9.51787457e-03 -4.10399675e-01 5.84583580e-01 1.45092428e+00
-1.01752587e-01 -3.14954817e-01 -1.35460925e+00 3.94505858e-01
-1.35642827e+00 -7.18276560e-01 4.60585244e-02 2.20360756e+00
1.11353219e+00 4.62597400e-01 2.02909508e-03 -3.04962158e-01
4.31422144e-01 1.96983695e-01 -8.11232448e-01 -5.44480681e-01
1.32749349e-01 4.87562835e-01 7.60057271e-01 4.58678126e-01
-7.45331407e-01 1.21514189e+00 8.39178658e+00 1.03200722e+00
-9.13643241e-01 6.02017224e-01 6.03716791e-01 -6.34288013e-01
-6.69894755e-01 4.93735820e-01 -6.76492155e-01 5.78955889e-01
1.90468574e+00 -6.56130672e-01 1.09186816e+00 8.37514997e-01
3.40090185e-01 -1.30206361e-01 -1.05813682e+00 5.42824686e-01
-2.17776060e-01 -1.60077047e+00 -7.61404634e-02 2.40435481e-01
8.60340238e-01 8.67875576e-01 -3.96172181e-02 8.16885829e-01
8.29527199e-01 -1.19220638e+00 7.85881937e-01 5.01648307e-01
1.06736636e+00 -6.67525411e-01 -7.27899522e-02 4.76647675e-01
-1.08623743e+00 -1.79431215e-01 -8.47531378e-01 -3.98379773e-01
1.73167303e-01 6.77533090e-01 -8.49043369e-01 3.51990998e-01
7.41749942e-01 6.93438351e-01 -6.50723875e-01 1.01195502e+00
-1.00310475e-01 9.25746262e-01 -4.56063360e-01 -1.03260979e-01
1.35991471e-02 -1.80648133e-01 8.72883666e-03 1.05417299e+00
6.80005074e-01 -2.52839386e-01 -5.13650067e-02 9.41422224e-01
-5.00675082e-01 -2.18629748e-01 -6.32412374e-01 -5.84526539e-01
8.90048563e-01 1.27730799e+00 -4.10594255e-01 -5.04293263e-01
-7.08707213e-01 7.64718711e-01 4.41395074e-01 8.90466943e-02
-1.12231326e+00 -3.03148776e-01 8.01299930e-01 -1.04863338e-01
1.93253979e-01 -3.58684361e-01 -2.11900324e-01 -1.20566285e+00
-3.36883485e-01 -9.20611203e-01 5.22523224e-02 -1.03236294e+00
-1.16092300e+00 2.87870318e-01 -2.28666425e-01 -6.24990165e-01
7.54606724e-03 -5.59083402e-01 -7.61654139e-01 1.06740224e+00
-1.38891459e+00 -1.10397935e+00 4.81878221e-02 9.68219638e-02
4.81041551e-01 -2.42966101e-01 1.16807687e+00 5.22425413e-01
-1.00033760e+00 5.20623207e-01 4.56424564e-01 -3.94924343e-01
5.63366473e-01 -8.95007312e-01 1.05280650e+00 6.93031192e-01
-4.21179682e-02 7.21573353e-01 7.99152970e-01 -4.94324356e-01
-1.95368361e+00 -1.00509036e+00 1.03717887e+00 -7.08239019e-01
1.01528227e+00 -8.38176608e-01 -4.88275230e-01 8.61357033e-01
3.94891083e-01 -1.30960688e-01 6.08507156e-01 2.87315309e-01
-3.27634126e-01 -5.10434993e-02 -8.64567459e-01 4.90710527e-01
1.21431565e+00 -6.23207271e-01 -1.35947660e-01 7.72722960e-01
1.03308392e+00 -2.64122099e-01 -1.08826137e+00 1.01000099e-02
5.50728917e-01 -5.94661772e-01 8.90053272e-01 -6.25806928e-01
4.00496304e-01 1.44051909e-01 -1.55078098e-01 -1.23572731e+00
-3.93841028e-01 -8.37180555e-01 -2.82206833e-01 1.33414733e+00
9.17370796e-01 -5.56871653e-01 7.23662198e-01 8.13538134e-01
-1.61618710e-01 -7.69910932e-01 -8.58782232e-01 -9.98495996e-01
7.33720660e-01 -5.97076178e-01 7.66362607e-01 9.13149476e-01
1.60660818e-01 3.94554675e-01 -5.70114791e-01 -1.92004755e-01
4.33388889e-01 -2.47492850e-01 5.30278385e-01 -1.07419884e+00
-5.35968363e-01 -3.19816321e-01 3.32411915e-01 -6.71624064e-01
-2.15987265e-02 -1.23751450e+00 -5.14328629e-02 -1.62926137e+00
7.22202778e-01 -7.43397534e-01 -2.30482534e-01 8.39073479e-01
1.52710393e-01 1.71531022e-01 -6.17346875e-02 1.37247011e-01
-6.38351798e-01 5.84586978e-01 8.80248845e-01 -3.12509745e-01
2.34123498e-01 -3.53455901e-01 -9.96958494e-01 5.61156929e-01
6.74139798e-01 -5.68621755e-01 -5.30566394e-01 -8.90918136e-01
3.33973169e-01 -2.21622333e-01 -6.58927858e-03 -1.16390932e+00
2.19856739e-01 -2.06758186e-01 7.23933876e-02 -3.33844423e-01
4.79596078e-01 -3.02058309e-01 4.41316903e-01 4.86921519e-01
-1.30548298e-01 1.91744328e-01 3.13807130e-01 2.20460638e-01
5.41590989e-01 -3.86742234e-01 6.07900560e-01 -4.61927086e-01
-5.65029919e-01 5.02237380e-01 -7.55441189e-01 1.78899109e-01
8.97489667e-01 3.19816232e-01 -4.90866274e-01 -2.80747890e-01
-1.95509121e-01 2.61769503e-01 7.08618283e-01 2.27397829e-01
-1.82965014e-03 -1.00168312e+00 -4.99836653e-01 -1.75597653e-01
5.36122322e-02 -2.13815764e-01 1.22844204e-01 8.14456165e-01
-5.39925456e-01 6.36470318e-01 1.15724407e-01 -1.47915170e-01
-7.93946207e-01 7.07428992e-01 3.63197267e-01 -4.61268038e-01
-5.45349836e-01 7.58329391e-01 -4.71151292e-01 -5.56552112e-01
1.63259476e-01 -2.43538305e-01 4.88362074e-01 -1.36056513e-01
3.16340089e-01 4.23321366e-01 2.46927008e-01 3.63082848e-02
-3.25185150e-01 1.24654301e-01 3.09775081e-02 -1.37573808e-01
1.67056668e+00 1.17869228e-01 -2.54356533e-01 3.45222682e-01
8.00459921e-01 -2.41015404e-01 -1.34187877e+00 -1.12014666e-01
3.20915617e-02 -7.95101002e-02 3.17616582e-01 -1.00236964e+00
-8.61775756e-01 1.14330137e+00 5.08761227e-01 -1.32887317e-02
7.27271855e-01 -1.57478139e-01 9.04376209e-01 7.25986779e-01
8.47290039e-01 -1.27590394e+00 -1.11275941e-01 8.01304698e-01
2.71730870e-01 -9.19847012e-01 1.53139025e-01 1.06197380e-01
-1.22693166e-01 6.78204358e-01 4.69318718e-01 4.67473745e-01
6.26796603e-01 7.30955958e-01 -3.30844730e-01 9.46910121e-03
-1.58919394e+00 -2.13791197e-03 -4.58582133e-01 6.48250341e-01
6.31135404e-01 1.08199269e-01 -1.13054767e-01 7.02789307e-01
-2.20326364e-01 3.17010693e-02 6.34324133e-01 9.63113785e-01
-6.60736263e-01 -1.37039542e+00 -3.43041234e-02 7.85822451e-01
-3.43656659e-01 -8.08975935e-01 -5.88867031e-02 4.01826024e-01
-1.50002733e-01 9.86573875e-01 -5.32265827e-02 -5.01529276e-01
1.36627436e-01 4.11967993e-01 5.56940317e-01 -7.49338865e-01
-2.50579268e-01 -7.97974989e-02 5.87290704e-01 -3.76225024e-01
2.79991180e-01 -4.88101423e-01 -9.62662220e-01 -1.04099762e+00
-2.28692889e-01 -5.19514531e-02 8.52081418e-01 7.43103445e-01
5.74618638e-01 4.66904581e-01 2.69479334e-01 -1.12306249e+00
-6.71607077e-01 -1.08018482e+00 -5.11046529e-01 -3.99505526e-01
-2.93999434e-01 -5.12731016e-01 -1.66529357e-01 -6.64113984e-02] | [10.541096687316895, 8.18695068359375] |
e46f0a44-d437-4992-8538-e1f878a75672 | transcription-is-all-you-need-learning-to | 2010.11904 | null | https://arxiv.org/abs/2010.11904v1 | https://arxiv.org/pdf/2010.11904v1.pdf | Transcription Is All You Need: Learning to Separate Musical Mixtures with Score as Supervision | Most music source separation systems require large collections of isolated sources for training, which can be difficult to obtain. In this work, we use musical scores, which are comparatively easy to obtain, as a weak label for training a source separation system. In contrast with previous score-informed separation approaches, our system does not require isolated sources, and score is used only as a training target, not required for inference. Our model consists of a separator that outputs a time-frequency mask for each instrument, and a transcriptor that acts as a critic, providing both temporal and frequency supervision to guide the learning of the separator. A harmonic mask constraint is introduced as another way of leveraging score information during training, and we propose two novel adversarial losses for additional fine-tuning of both the transcriptor and the separator. Results demonstrate that using score information outperforms temporal weak-labels, and adversarial structures lead to further improvements in both separation and transcription performance. | ['Jonathan Le Roux', 'Gordon Wichern', 'Yun-Ning Hung'] | 2020-10-22 | null | null | null | null | ['music-source-separation'] | ['music'] | [ 3.95245492e-01 -9.15841237e-02 -1.55406699e-01 -2.80868173e-01
-1.22411156e+00 -1.11419308e+00 4.28638875e-01 -5.57342879e-02
-1.58669248e-01 6.04084551e-01 1.10223942e-01 6.46487484e-03
-1.26503155e-01 -2.77708471e-01 -6.91068172e-01 -1.00758100e+00
-1.15399472e-01 2.30466381e-01 1.68103516e-01 -3.19740996e-02
-1.59836560e-01 2.31723562e-01 -1.20979893e+00 1.19978122e-01
1.06658852e+00 9.79920864e-01 4.96420600e-02 8.50411594e-01
3.22406530e-01 8.98557365e-01 -9.56125855e-01 -2.78222948e-01
4.29319441e-01 -1.03395796e+00 -3.27392727e-01 -1.01749294e-01
3.29425275e-01 -5.09245172e-02 -2.11927757e-01 9.66726661e-01
7.34178901e-01 3.64828557e-01 6.43252015e-01 -1.02272415e+00
-3.25797528e-01 1.05157876e+00 -5.81826746e-01 -3.87656428e-02
1.88948780e-01 1.20619044e-03 1.30195630e+00 -5.48969626e-01
1.50288492e-01 8.10444117e-01 7.58737028e-01 4.79126602e-01
-1.58931661e+00 -1.03042185e+00 6.79781437e-02 -7.31431618e-02
-1.22161484e+00 -8.45832884e-01 1.25203943e+00 -3.27419370e-01
4.84182894e-01 3.18407446e-01 6.07145846e-01 1.12572646e+00
-3.96936119e-01 8.79227400e-01 6.77946746e-01 -6.04349852e-01
2.55837113e-01 9.55344886e-02 -1.91501435e-02 4.33863044e-01
-3.50044489e-01 1.69414386e-01 -8.95560920e-01 -3.12581360e-01
9.74685252e-01 -3.67212415e-01 -5.17436445e-01 -4.01147544e-01
-8.84344280e-01 6.39770865e-01 4.06405658e-01 1.77812561e-01
-2.21828640e-01 3.01054448e-01 3.04641634e-01 5.04903674e-01
4.31633443e-01 6.21119499e-01 -3.18774283e-01 -1.70796916e-01
-1.44366586e+00 1.23264216e-01 8.08118761e-01 6.19105160e-01
4.60235149e-01 4.30037022e-01 -4.42912400e-01 1.16455531e+00
-4.60854657e-02 7.32815146e-01 5.80924749e-01 -1.20102358e+00
4.65082884e-01 7.00411946e-02 1.61094517e-01 -6.56050265e-01
-6.60332963e-02 -9.92775321e-01 -5.92340112e-01 2.76406318e-01
6.78701937e-01 -2.86492646e-01 -9.83622193e-01 2.19224882e+00
1.58326581e-01 7.28475869e-01 -9.55170020e-02 1.05138350e+00
2.50015706e-01 5.93682826e-01 -4.82214749e-01 -3.90164793e-01
9.33048725e-01 -1.22349477e+00 -7.12440073e-01 -2.56745130e-01
3.11791390e-01 -8.10838938e-01 1.03434491e+00 7.76145816e-01
-1.38779867e+00 -5.17754912e-01 -1.21661973e+00 2.14234330e-02
1.78050593e-01 5.07056177e-01 3.69484425e-01 3.41157794e-01
-6.36806846e-01 8.48245084e-01 -9.99565303e-01 4.16833043e-01
1.94767594e-01 4.84156042e-01 1.48712266e-02 4.35823977e-01
-1.09706032e+00 4.43294913e-01 4.01670150e-02 2.56836992e-02
-1.07987046e+00 -5.42122364e-01 -7.60869801e-01 2.76516289e-01
4.38047230e-01 -4.16745901e-01 1.36218607e+00 -1.37468529e+00
-1.92719281e+00 6.41808927e-01 -3.88795212e-02 -5.38465023e-01
4.20781553e-01 -3.46647114e-01 -4.06370372e-01 3.23749632e-01
-9.76566412e-03 2.74264127e-01 1.37242448e+00 -1.28218973e+00
-3.70596528e-01 4.26232442e-02 -3.16448510e-02 2.55540699e-01
-3.68436188e-01 -5.32442592e-02 -6.71490848e-01 -1.12444055e+00
3.95173468e-02 -1.04969847e+00 -3.68146272e-03 -2.76772350e-01
-5.78773141e-01 1.26085997e-01 3.82339239e-01 -8.79328609e-01
1.41267467e+00 -2.48447704e+00 5.85749626e-01 4.06282902e-01
-1.46670625e-01 2.87349075e-01 -2.56066471e-01 5.03576398e-02
-6.33077249e-02 -2.52364933e-01 -5.64736724e-01 -8.27066779e-01
4.33647409e-02 2.82419194e-02 -5.57085752e-01 3.10059339e-01
2.31460035e-01 5.31399131e-01 -9.92455542e-01 -1.66345209e-01
-1.43348888e-01 4.66038048e-01 -7.21733510e-01 3.10289949e-01
-2.42826119e-01 7.99915910e-01 -1.16786897e-01 2.24020138e-01
3.35537225e-01 -7.69983456e-02 2.09511995e-01 9.59965289e-02
2.06438288e-01 8.28239322e-01 -1.40697074e+00 2.01736784e+00
-5.76461136e-01 4.93249059e-01 4.83023137e-01 -9.52522159e-01
8.69574666e-01 6.40622377e-01 3.68778527e-01 -2.75571942e-01
1.28360286e-01 3.29580337e-01 2.38800362e-01 -5.37752323e-02
6.54193088e-02 -2.97895253e-01 -2.26897836e-01 5.62180221e-01
2.53749639e-01 -3.81795943e-01 2.81472206e-01 1.41892880e-01
1.07042015e+00 2.53828079e-01 -2.43024483e-01 9.33762714e-02
5.15075147e-01 -3.65889192e-01 1.01437700e+00 5.04553020e-01
2.14622453e-01 9.42309260e-01 5.82172632e-01 3.31271440e-01
-7.93446660e-01 -1.18428361e+00 7.99292773e-02 1.12217653e+00
7.02943355e-02 -4.71525431e-01 -7.23700166e-01 -6.84481025e-01
8.22184086e-02 6.37808502e-01 -3.53030026e-01 -3.46873313e-01
-8.55264127e-01 -3.24727029e-01 8.44085038e-01 7.18922496e-01
-1.19052425e-01 -6.68198764e-01 -1.90889671e-01 3.03254187e-01
-2.29596645e-01 -7.72587121e-01 -1.13775826e+00 7.63833463e-01
-8.15763652e-01 -8.43744993e-01 -8.38691056e-01 -6.33753061e-01
4.29460287e-01 1.00951148e-02 8.73383343e-01 -1.70486748e-01
7.78746754e-02 -9.89333615e-02 -1.77868068e-01 -5.44864655e-01
-3.53791505e-01 2.35519245e-01 1.31232947e-01 4.01657462e-01
-1.58178046e-01 -1.20800209e+00 -4.55500185e-01 3.21384221e-01
-7.00889647e-01 -2.46939167e-01 3.46218050e-01 1.00977635e+00
4.45664555e-01 1.12119913e-01 7.63286829e-01 -7.54308760e-01
4.18750703e-01 -2.24779174e-01 -6.21390641e-01 -1.53685110e-02
-3.06275517e-01 1.62570879e-01 1.10673511e+00 -8.72141838e-01
-9.42137718e-01 8.78066793e-02 1.03190288e-01 -8.15574586e-01
2.35874355e-01 3.08178812e-01 -2.46872604e-01 2.32982025e-01
8.04720402e-01 2.34479774e-02 -4.24121208e-02 -8.51386964e-01
2.49386519e-01 5.32125175e-01 8.27938974e-01 -6.37558937e-01
1.18854988e+00 3.03438306e-01 -2.70993412e-01 -4.17352080e-01
-1.11869013e+00 -5.39285481e-01 -6.94069505e-01 1.93756178e-01
3.44779640e-01 -1.04154825e+00 -4.82993096e-01 4.64353502e-01
-9.08534050e-01 -6.75766468e-01 -5.54763258e-01 6.85436487e-01
-6.00362539e-01 -2.66197957e-02 -6.77021503e-01 -8.88125896e-01
-2.25395828e-01 -7.77431965e-01 1.00917733e+00 1.18892483e-01
-3.33665520e-01 -9.39966679e-01 1.85393155e-01 4.10523146e-01
2.55284101e-01 2.17542306e-01 5.52395284e-01 -5.33345938e-01
-4.67179835e-01 -1.99797839e-01 4.14676130e-01 8.14683437e-01
1.90402210e-01 -1.08900182e-01 -1.31373227e+00 -3.87370765e-01
3.42625409e-01 -5.74922860e-01 1.03817809e+00 2.74247319e-01
8.83933723e-01 -3.00641179e-01 -2.35218361e-01 9.45098221e-01
9.34045792e-01 2.96176285e-01 2.26300821e-01 -1.38619840e-01
7.76224792e-01 2.96445459e-01 3.31641227e-01 3.98906678e-01
-1.08177461e-01 8.98278058e-01 -3.81050967e-02 -2.33443677e-01
-3.88046503e-01 -2.68804044e-01 7.09129989e-01 1.17664480e+00
-1.69911906e-01 -7.01535642e-02 -4.96019840e-01 5.74818194e-01
-1.78939331e+00 -9.89889979e-01 2.78511375e-01 2.48402071e+00
1.48503876e+00 1.32837668e-01 4.38216358e-01 6.48003280e-01
4.45578098e-01 1.76724523e-01 -8.41421962e-01 1.09708481e-01
-1.28694937e-01 4.92908150e-01 2.51617432e-01 6.68805301e-01
-1.12478113e+00 6.25128329e-01 6.12206125e+00 8.27967048e-01
-1.40467274e+00 1.39403939e-01 -9.60514173e-02 -7.46352255e-01
-2.50587076e-01 1.43487528e-01 -3.37516278e-01 6.04117990e-01
8.32190335e-01 2.12127231e-02 8.12723756e-01 4.85272139e-01
1.62752658e-01 1.80229262e-01 -1.41250277e+00 7.62980223e-01
4.09977585e-02 -7.11370766e-01 -3.95395756e-01 -2.04182357e-01
6.79439962e-01 -2.70038694e-01 1.71602562e-01 5.53527512e-02
4.57243532e-01 -8.77388299e-01 1.23779798e+00 4.78153288e-01
6.86876297e-01 -9.94979858e-01 3.34342122e-01 4.09283459e-01
-1.19406700e+00 -1.76152185e-01 -2.10682862e-02 3.08104474e-02
1.70282051e-01 7.16242194e-01 -5.52617788e-01 5.88925719e-01
2.39297345e-01 6.45558715e-01 -9.11633521e-02 1.00176084e+00
-9.13715601e-01 1.10965919e+00 -3.78436774e-01 6.81963325e-01
-1.07852474e-01 -8.67314413e-02 8.40327501e-01 1.12373865e+00
1.74489081e-01 -1.10960968e-01 4.63027030e-01 8.42408895e-01
-1.71802178e-01 -1.96723655e-01 -1.83073059e-01 -3.15171070e-02
8.09707224e-01 1.12554312e+00 -5.88397026e-01 -2.42119402e-01
-9.17027518e-02 1.12176144e+00 3.68805170e-01 4.70573455e-01
-1.02957189e+00 -7.37235248e-01 5.74670136e-01 -9.20756534e-02
5.21391094e-01 -3.27098906e-01 -3.40406060e-01 -1.40793192e+00
9.92518589e-02 -1.07895780e+00 3.05789798e-01 -4.79636997e-01
-1.03453803e+00 3.34869057e-01 -3.05466890e-01 -1.39197695e+00
-6.03878081e-01 -1.75618753e-01 -7.66910613e-01 1.20664728e+00
-1.46931136e+00 -9.38988566e-01 1.71073481e-01 6.77363098e-01
1.71128392e-01 -8.20774883e-02 7.83588886e-01 3.48515332e-01
-6.84589267e-01 8.74942780e-01 1.95213616e-01 3.48521650e-01
1.19193482e+00 -1.47609901e+00 -1.42398346e-02 9.09512043e-01
7.17820585e-01 6.97708428e-01 7.02568889e-01 -3.12200069e-01
-1.12192774e+00 -7.68938601e-01 4.26411420e-01 -3.89963269e-01
5.87693214e-01 -7.77802646e-01 -9.99670863e-01 5.75550437e-01
8.75853673e-02 -1.88936219e-01 9.03153002e-01 2.67781317e-01
-5.19357681e-01 -4.36187953e-01 -6.93594992e-01 3.27018648e-01
8.69304061e-01 -8.31383407e-01 -6.26772285e-01 6.81050792e-02
6.41131639e-01 -5.25455475e-01 -6.82152152e-01 1.67035341e-01
5.17707288e-01 -7.06567347e-01 9.75925922e-01 -1.74098849e-01
1.50595590e-01 -4.67332542e-01 1.32799685e-01 -1.65854084e+00
-3.53242934e-01 -9.71187949e-01 -4.69622284e-01 1.60925591e+00
5.63954890e-01 -3.32554787e-01 6.53641820e-01 1.68803364e-01
-2.54570127e-01 -2.93589056e-01 -6.80123210e-01 -1.27531052e+00
-3.86795984e-03 -3.61361265e-01 4.78199691e-01 1.01455200e+00
9.85952094e-02 6.77221358e-01 -4.85941589e-01 2.72138447e-01
6.39521718e-01 3.39747429e-01 6.73383951e-01 -1.21359217e+00
-1.15619290e+00 -6.90988660e-01 2.12630555e-01 -1.09218884e+00
1.48856521e-01 -1.01892126e+00 2.69667447e-01 -9.72425401e-01
-1.44388154e-01 -5.68953335e-01 -6.33743584e-01 7.02628732e-01
-2.98022062e-01 4.32318330e-01 3.96349490e-01 4.82396036e-01
-2.54730552e-01 6.37862802e-01 1.02199888e+00 -3.10234636e-01
-6.57328844e-01 3.85933489e-01 -6.77717447e-01 9.38384235e-01
6.03277266e-01 -6.72489643e-01 -7.54000485e-01 -4.32570100e-01
2.58787274e-02 3.05011004e-01 1.21020004e-01 -1.18193853e+00
3.40826750e-01 -1.63620412e-02 2.44865552e-01 -1.57518074e-01
7.21703887e-01 -6.41976655e-01 2.78439838e-02 8.23441744e-02
-5.26070178e-01 -6.57049537e-01 1.12575844e-01 4.08317536e-01
-4.67227519e-01 -3.37323368e-01 8.32611322e-01 2.29326561e-01
3.08271974e-01 8.13081339e-02 2.10333318e-01 1.73373684e-01
5.79091012e-01 1.26459867e-01 1.63473636e-01 -5.42281270e-01
-8.81028891e-01 3.24160933e-01 3.72897089e-01 1.92908257e-01
9.88401920e-02 -1.53598356e+00 -6.60195470e-01 3.26685458e-01
-3.57219547e-01 2.19910458e-01 -1.19163776e-02 8.64431560e-01
1.70647576e-01 -3.46613042e-02 2.31073067e-01 -3.98088247e-01
-1.31965721e+00 4.25347120e-01 3.25370401e-01 -3.90474588e-01
-4.58339244e-01 1.13665080e+00 2.42126003e-01 -3.31445456e-01
7.62994170e-01 -2.20895559e-01 3.12856771e-02 3.82894605e-01
5.48384905e-01 2.61568606e-01 -2.94685247e-03 -3.01172167e-01
-2.70876020e-01 5.67677319e-01 1.85083419e-01 -6.56290948e-01
1.36309290e+00 -3.91300470e-02 8.91246274e-02 9.88545001e-01
8.96263480e-01 5.98143995e-01 -1.59766996e+00 -5.22845626e-01
-6.27896041e-02 -2.26109236e-01 -1.18460722e-01 -1.05226433e+00
-1.08238137e+00 9.84450758e-01 3.02699655e-02 3.20498616e-01
1.50492966e+00 -1.99040622e-01 9.36868906e-01 6.73813522e-02
1.70776263e-01 -1.10764050e+00 2.55768836e-01 6.14996016e-01
9.73366201e-01 -7.57938206e-01 -3.52202266e-01 -3.00392121e-01
-4.74166006e-01 8.21874022e-01 2.83015400e-01 -2.69443095e-01
4.80302244e-01 5.93475282e-01 2.10516185e-01 1.61587134e-01
-7.81328976e-01 -1.20968498e-01 7.08501935e-01 3.13340098e-01
5.22052705e-01 -1.82071347e-02 6.45348728e-02 1.09826505e+00
-5.46532750e-01 -2.34701425e-01 2.28519216e-01 7.93530881e-01
-2.94223756e-01 -1.42184389e+00 -5.96575499e-01 1.87123671e-01
-7.43306220e-01 -1.21440135e-01 -5.05444467e-01 1.47291288e-01
3.25067312e-01 9.43652689e-01 -6.29770979e-02 -3.93453568e-01
3.01989734e-01 4.54062819e-01 4.75277126e-01 -7.50772953e-01
-8.04475069e-01 5.59091389e-01 1.18580103e-01 -3.23395699e-01
-3.05963218e-01 -6.71602964e-01 -1.41324568e+00 1.51533812e-01
-7.69544840e-01 5.74424207e-01 4.79145110e-01 8.06489944e-01
1.86431706e-01 8.63272369e-01 9.88290370e-01 -9.10987973e-01
-7.64805436e-01 -9.81132567e-01 -9.10009444e-01 5.15922487e-01
8.14799249e-01 -4.82582092e-01 -5.05429626e-01 4.44494009e-01] | [15.458955764770508, 5.557227611541748] |
a614b6c4-a0f0-4d42-bc52-ab4d31a34e0e | augmentednet-a-roman-numeral-analysis-network | null | null | https://doi.org/10.5281/zenodo.5624533 | https://archives.ismir.net/ismir2021/paper/000050.pdf | AugmentedNet: A Roman Numeral Analysis Network with Synthetic Training Examples and Additional Tonal Tasks | AugmentedNet is a new convolutional recurrent neural network for predicting Roman numeral labels. The network architecture is characterized by a separate convolutional block for bass and chromagram inputs. This layout is further enhanced by using synthetic training examples for data augmentation, and a greater number of tonal tasks to solve simultaneously via multitask learning. This paper reports the improved performance achieved by combining these ideas. The additional tonal tasks strengthen the shared representation learned through multitask learning. The synthetic examples, in turn, complement key transposition, which is often the only technique used for data augmentation in similar problems related to tonal music. The name "AugmentedNet" speaks to the increased number of both training examples and tonal tasks. We report on tests across six relevant and publicly available datasets: ABC, BPS, HaydnSun, TAVERN, When-in-Rome, and WTC. In our tests, our model outperforms recent methods of functional harmony, such as other convolutional neural networks and Transformer-based models. Finally, we show a new method for reconstructing the full Roman numeral label, based on common Roman numeral classes, which leads to better results compared to previous methods. | ['Ichiro Fujinaga', 'Mark Gotham', 'Nestor Napoles Lopez'] | 2021-11-07 | null | null | null | ismir-2021-11 | ['chord-recognition'] | ['audio'] | [ 1.79334760e-01 5.47673777e-02 -1.16889909e-01 -9.10324305e-02
-8.68893981e-01 -3.60755950e-01 6.08000100e-01 -5.47104001e-01
-3.37042481e-01 8.02914381e-01 4.53048736e-01 -7.99494088e-02
-1.95999473e-01 -7.06952572e-01 -6.80253565e-01 -6.12074196e-01
4.98905079e-03 6.05237305e-01 -3.25874388e-01 -9.13169026e-01
4.13294174e-02 3.18107046e-02 -1.57712734e+00 7.73311257e-01
3.32122713e-01 9.94141459e-01 -1.57361582e-01 4.92474705e-01
-1.70327544e-01 1.02977049e+00 -6.27061605e-01 -3.25528830e-01
3.66695255e-01 -4.14307147e-01 -1.12042511e+00 -3.35889906e-01
4.39896137e-01 1.63179249e-01 -4.34540734e-02 5.65758049e-01
6.84683263e-01 1.71170980e-01 3.12351227e-01 -1.17523646e+00
-6.01172209e-01 1.36638296e+00 -3.76953036e-01 -1.53627574e-01
1.77261353e-01 -2.26578116e-01 1.56046259e+00 -8.78021121e-01
3.52080077e-01 1.08380103e+00 1.24983025e+00 3.71520966e-01
-1.24080563e+00 -1.07377625e+00 -2.08326161e-01 3.35109621e-01
-1.15177870e+00 -3.69425982e-01 1.07339370e+00 -2.43613049e-01
9.05124545e-01 5.05204916e-01 7.97352910e-01 1.39241683e+00
-4.71916974e-01 8.48986685e-01 1.44820642e+00 -6.26201987e-01
-3.43215466e-01 -1.85726471e-02 -3.67373258e-01 2.23335072e-01
-4.63874876e-01 2.12023079e-01 -8.94371092e-01 7.39513338e-02
7.90179312e-01 -4.15215552e-01 -1.40367940e-01 -5.68374880e-02
-1.60746670e+00 8.63523722e-01 3.33166987e-01 5.15591025e-01
-1.35941789e-01 4.70382035e-01 8.91279697e-01 6.39128625e-01
6.57863855e-01 1.17550099e+00 -8.01525474e-01 -3.15593183e-01
-1.00303113e+00 4.00856197e-01 8.79930496e-01 5.10177791e-01
5.64921379e-01 8.29421341e-01 -1.35819703e-01 1.27465129e+00
-3.05614442e-01 2.43171245e-01 6.26525164e-01 -8.35508823e-01
2.45507494e-01 2.62849033e-01 -6.79238364e-02 -5.85971713e-01
-6.95306718e-01 -8.05469811e-01 -1.02483547e+00 2.46080354e-01
4.98833627e-01 -3.38355042e-02 -8.46206188e-01 1.87376356e+00
-1.79198191e-01 3.90901506e-01 1.00938477e-01 8.35527420e-01
1.08564556e+00 5.39342046e-01 -2.38245130e-01 7.18937740e-02
1.33529770e+00 -1.30136371e+00 -8.30963492e-01 2.19811946e-01
4.43685800e-01 -1.44304085e+00 1.53327894e+00 9.36948776e-01
-8.62959027e-01 -9.47071135e-01 -9.68343437e-01 -1.52131200e-01
-6.37513757e-01 2.01271117e-01 8.73899937e-01 6.05240226e-01
-8.23836088e-01 8.92338455e-01 -1.20195426e-01 1.11824259e-01
-1.69383939e-02 5.19601882e-01 -9.64893401e-02 5.52112997e-01
-1.72907722e+00 1.01095903e+00 4.62351024e-01 6.77843094e-02
-6.89086795e-01 -8.91471565e-01 -7.69536853e-01 -3.13830972e-02
1.82309777e-01 -3.22393328e-01 1.54343450e+00 -9.39170003e-01
-1.85468543e+00 8.55488420e-01 4.67241853e-01 -6.91541135e-01
3.46434027e-01 -3.15108538e-01 -6.85017705e-01 -2.64328539e-01
-5.79995178e-02 8.30988646e-01 7.39172697e-01 -9.16829705e-01
-3.49153310e-01 3.33947927e-01 4.58882423e-03 1.97685480e-01
-3.61467451e-01 -1.11181336e-02 5.39631918e-02 -1.29643583e+00
-9.50143337e-02 -1.01321483e+00 -1.46967158e-01 -8.98832381e-01
-6.66443706e-01 -2.73959160e-01 7.40648031e-01 -5.41261435e-01
1.04681671e+00 -2.11633110e+00 2.31703058e-01 1.10203594e-01
-8.81202221e-02 -2.55830307e-02 -2.20602646e-01 3.54325801e-01
-6.85229301e-01 1.07570447e-01 4.68178615e-02 -5.45118749e-01
1.96829602e-01 1.90770060e-01 -6.96769238e-01 4.75728959e-02
1.34920001e-01 1.02950597e+00 -5.42154253e-01 6.83210343e-02
3.38885844e-01 4.72068459e-01 -4.48078841e-01 -1.80585906e-01
-1.82798266e-01 7.75518239e-01 2.65257239e-01 5.48985064e-01
1.65555179e-01 -1.16877198e-01 -5.62264100e-02 -4.01999980e-01
-2.62232989e-01 8.00816536e-01 -1.41422379e+00 1.93881178e+00
-7.78080344e-01 5.78489840e-01 -4.49188173e-01 -1.09715331e+00
1.25938296e+00 7.64151216e-01 6.48723245e-01 -7.53262997e-01
-6.80904952e-04 4.65070575e-01 2.10582376e-01 -1.00531131e-02
8.74031007e-01 -3.31491619e-01 -5.91292202e-01 5.55736661e-01
2.03890517e-01 -5.58126926e-01 1.70986086e-01 -9.48245525e-02
4.56835449e-01 4.57405508e-01 1.92276284e-01 -2.67542899e-01
9.93523002e-02 4.72314423e-03 5.41272998e-01 5.63314497e-01
3.07681680e-01 6.90645635e-01 4.09116060e-01 -1.00043118e+00
-1.44615221e+00 -6.06230259e-01 -2.94968933e-01 1.37728059e+00
-5.40923357e-01 -6.88582838e-01 -2.15074480e-01 -7.79482797e-02
-1.88303456e-01 4.97523904e-01 -1.00915742e+00 -1.12972595e-01
-8.42853487e-01 -8.23930144e-01 1.17481196e+00 3.59681726e-01
6.55789554e-01 -1.61366963e+00 -2.63790250e-01 3.90040994e-01
-5.25137126e-01 -9.22929645e-01 -3.28095227e-01 7.48276114e-01
-5.46937048e-01 -9.50848043e-01 -7.13651001e-01 -8.37334037e-01
-4.81975943e-01 -2.25144580e-01 1.28338325e+00 -4.32892621e-01
-2.59976208e-01 -2.71783769e-01 -5.57621658e-01 -6.25386238e-01
-3.45514268e-01 5.16824365e-01 1.20769948e-01 4.35627513e-02
1.84513792e-01 -7.76127458e-01 -2.52121806e-01 2.48161957e-01
-5.48166454e-01 4.94317055e-01 4.33946073e-01 1.05706048e+00
6.15092456e-01 -2.35186011e-01 9.39578414e-01 -1.08910358e+00
5.30467570e-01 -1.77955955e-01 -9.78141949e-02 -1.90557927e-01
-3.65836531e-01 2.89857890e-02 7.48043418e-01 -6.23244107e-01
-7.07316756e-01 7.22302422e-02 -2.99620569e-01 -5.12318432e-01
-8.41303263e-03 4.76035744e-01 3.34357947e-01 8.67174380e-03
8.51976752e-01 -8.23727250e-03 -2.02726796e-01 -8.56584370e-01
5.07845640e-01 4.66205657e-01 8.74119222e-01 -6.32205069e-01
7.80096650e-01 1.61012441e-01 2.38123350e-03 -4.93349165e-01
-1.26777232e+00 -3.58813137e-01 -4.23182070e-01 -4.46232371e-02
6.92133009e-01 -1.01915359e+00 -7.20137000e-01 7.19048381e-01
-8.71636033e-01 -5.86227775e-01 -1.02421200e+00 6.60557330e-01
-7.60125339e-01 -1.50638148e-01 -9.34940100e-01 -4.87103254e-01
-4.51176941e-01 -1.03369033e+00 9.81193900e-01 -1.58619925e-01
-4.72886175e-01 -8.04373562e-01 2.65307695e-01 2.64881492e-01
6.92123950e-01 4.32578534e-01 9.45642769e-01 -7.09011197e-01
-3.04487757e-02 1.56283900e-01 8.04435685e-02 4.80969399e-01
1.71789125e-01 -1.65292695e-01 -1.40729237e+00 2.23988127e-02
-2.54308760e-01 -9.17509437e-01 1.16541350e+00 3.37513655e-01
1.31109822e+00 -8.94860774e-02 4.74672019e-01 8.42846155e-01
9.27246392e-01 -4.37652282e-02 7.04915702e-01 8.53130698e-01
7.69551516e-01 2.76579946e-01 4.23425943e-01 4.46198612e-01
2.62600988e-01 1.01168036e+00 4.62718487e-01 -6.25932992e-01
-5.97239137e-01 -1.29818723e-01 7.44755194e-02 1.16833830e+00
-5.25021613e-01 3.76162201e-01 -7.42075861e-01 4.92846429e-01
-1.55881715e+00 -9.81058121e-01 -3.59942734e-01 1.94125271e+00
1.24725735e+00 5.14728725e-02 4.82712686e-01 9.71742928e-01
4.70558435e-01 1.37307748e-01 -2.85754293e-01 -6.86258376e-01
-5.57791770e-01 1.09504223e+00 2.31297716e-01 2.15354949e-01
-1.42520618e+00 1.19232798e+00 6.98787928e+00 1.01763940e+00
-1.23035443e+00 2.05391780e-01 5.03784835e-01 -2.79952615e-01
-1.66092932e-01 -1.79170281e-01 -5.93623936e-01 7.03901052e-02
9.51472580e-01 7.03606904e-02 7.00783312e-01 7.72738039e-01
-1.18371123e-03 4.72317189e-01 -9.32166338e-01 1.02724421e+00
-2.93266121e-02 -1.56522942e+00 2.48854589e-02 -1.15141749e-01
9.34845209e-01 1.27923176e-01 4.79288310e-01 8.11410606e-01
4.49086159e-01 -1.19995165e+00 8.38931024e-01 2.39447236e-01
1.22041595e+00 -9.70521808e-01 7.07782626e-01 -3.34532082e-01
-1.26705956e+00 6.36444464e-02 -3.67659628e-02 -3.40712816e-01
3.40176709e-02 3.90427291e-01 -8.74501228e-01 6.40518248e-01
9.06751871e-01 7.01371074e-01 -3.17803890e-01 7.96345830e-01
-3.33238155e-01 7.61147141e-01 -2.78863549e-01 3.21576774e-01
4.52692568e-01 6.74967915e-02 3.34814429e-01 1.23587656e+00
2.54214555e-01 -5.06407619e-01 2.11684063e-01 5.86096048e-01
-2.21556053e-01 3.66320252e-01 -5.18944561e-01 5.44986725e-02
1.07895605e-01 1.54665303e+00 -4.91288155e-01 -2.42687106e-01
-2.30720937e-01 7.05604076e-01 1.04577467e-01 2.45117694e-01
-8.61646116e-01 -5.75288832e-01 3.13902617e-01 -1.94218382e-01
1.52223021e-01 -9.90985408e-02 -4.81980830e-01 -8.93462002e-01
-3.57884824e-01 -1.10882962e+00 4.40538287e-01 -8.40228438e-01
-1.09606135e+00 8.62176180e-01 -2.60583431e-01 -1.34920061e+00
-4.34333086e-01 -5.35133660e-01 -5.48091829e-01 9.46289420e-01
-1.68538976e+00 -1.52149045e+00 -1.28867969e-01 8.56875777e-01
6.79940879e-01 -5.10580182e-01 1.45858562e+00 6.37738347e-01
-3.42407882e-01 7.00018644e-01 -1.37271702e-01 2.86759079e-01
1.13934648e+00 -1.55898201e+00 5.76810062e-01 1.85073704e-01
6.79025352e-01 4.19906005e-02 5.83721519e-01 -4.91121784e-02
-7.22895563e-01 -1.05605197e+00 8.27982247e-01 -1.56620309e-01
8.45067620e-01 -4.53086317e-01 -7.55066633e-01 7.41936386e-01
3.95817667e-01 -3.09585065e-01 9.85964179e-01 7.85243809e-01
-7.57275283e-01 -1.93587139e-01 -7.12745786e-01 3.76704216e-01
7.13261127e-01 -4.89086449e-01 -6.88095987e-01 4.98737246e-01
9.83200848e-01 -5.28934360e-01 -9.38056946e-01 6.71820700e-01
5.86882234e-01 -8.52136791e-01 8.58678997e-01 -7.49824762e-01
6.10569596e-01 -1.43758371e-01 -3.32726061e-01 -1.69375205e+00
-3.13637346e-01 -8.75550985e-01 5.86717278e-02 9.74706531e-01
6.68384850e-01 -2.59144843e-01 7.66893804e-01 -4.87140238e-01
-6.31993413e-01 -6.18201017e-01 -1.03803313e+00 -6.50257826e-01
8.95736888e-02 -7.81120777e-01 6.60951495e-01 1.43554807e+00
6.42695744e-03 5.76068461e-01 -1.17123938e+00 -5.77672124e-01
1.62492737e-01 4.03283000e-01 7.64072061e-01 -1.39334571e+00
-5.74797451e-01 -5.17776012e-01 -1.37499988e-01 -3.90699357e-01
8.19458663e-02 -1.16309845e+00 -3.37827414e-01 -1.20291698e+00
-1.42749786e-01 -7.61051476e-01 -6.69124305e-01 1.15394735e+00
3.62656742e-01 1.19351673e+00 2.54661053e-01 2.29888052e-01
-2.54197747e-01 5.61467409e-01 1.46825814e+00 -2.89781634e-02
-4.12727088e-01 -9.19855572e-03 -7.25393414e-01 9.18392003e-01
1.20711339e+00 -2.39296645e-01 -8.02053697e-03 -2.53901750e-01
6.69481456e-01 -2.72104591e-01 2.35111237e-01 -1.10168099e+00
-1.10948458e-01 2.18391225e-01 3.93696755e-01 -4.98069078e-01
9.18841600e-01 -3.93521607e-01 2.43143201e-01 4.99393314e-01
-4.37425464e-01 5.37520833e-02 4.52338308e-01 -5.35714701e-02
-5.31230330e-01 2.02884659e-01 8.82844329e-01 -3.49947453e-01
-6.88088417e-01 1.88328419e-02 -2.70859569e-01 6.67548738e-03
4.92798954e-01 -7.21264631e-02 -1.39203653e-01 -4.22673732e-01
-1.30649114e+00 -1.27571285e-01 -3.43976647e-01 7.37224519e-01
3.44676018e-01 -1.71203911e+00 -1.07712328e+00 2.50093311e-01
-3.38587724e-02 -2.62104392e-01 1.84149459e-01 1.06419444e+00
-2.82323807e-01 3.35035712e-01 -5.44412315e-01 -3.13444585e-01
-1.11794806e+00 1.97701439e-01 3.54835778e-01 -5.58514655e-01
-7.55149126e-01 7.26125777e-01 -1.11432143e-01 -9.73797858e-01
2.92174876e-01 -3.37964296e-01 -4.17318076e-01 4.30106759e-01
9.23606753e-02 1.61857843e-01 3.62666667e-01 -4.36629146e-01
1.39233142e-01 5.18495679e-01 -6.84876963e-02 -2.42934823e-01
1.57659602e+00 4.06194091e-01 -1.38111696e-01 1.00047266e+00
6.80423260e-01 1.34197697e-01 -6.10478640e-01 -4.09131289e-01
1.46341268e-02 3.25747393e-02 -8.44904557e-02 -1.13595915e+00
-1.13452649e+00 8.19080770e-01 4.04071718e-01 3.37323189e-01
1.21094978e+00 -1.18888900e-01 7.23224521e-01 5.11332333e-01
6.40944466e-02 -1.19398379e+00 4.15043861e-01 8.18209887e-01
1.11863256e+00 -9.37531292e-01 -3.62046063e-01 -6.97134435e-02
-7.51132071e-01 1.16626310e+00 4.73617226e-01 -2.65546650e-01
4.20073241e-01 4.25597548e-01 2.37699091e-01 -1.74343497e-01
-6.84033692e-01 -2.71614850e-01 3.49730164e-01 1.05900295e-01
7.70736456e-01 1.56426862e-01 -3.92306075e-02 6.50862753e-01
-1.12985516e+00 -2.38312647e-01 5.78800499e-01 2.80528605e-01
-1.78850934e-01 -1.35037768e+00 -3.10913265e-01 4.56028342e-01
-8.54765534e-01 -5.89906394e-01 -4.25253987e-01 1.18493068e+00
4.60668951e-01 6.61625683e-01 2.04020627e-02 -6.51603937e-01
4.22103763e-01 3.29257011e-01 3.55300784e-01 -5.90655625e-01
-1.10216868e+00 5.38891792e-01 4.95913804e-01 -1.84651092e-01
-6.61948085e-01 -2.75879413e-01 -8.83246303e-01 -2.65827894e-01
-2.19305918e-01 4.80949134e-01 6.39767706e-01 6.49549484e-01
-9.85863656e-02 1.07327962e+00 6.68403149e-01 -9.83499885e-01
-3.87670875e-01 -1.42946970e+00 -8.42191160e-01 4.98422742e-01
2.03440920e-01 -6.32052898e-01 -2.95955569e-01 -5.76322749e-02] | [15.829815864562988, 5.2507429122924805] |
d2394bdc-0828-42ac-a0ff-daeb0ffce7cc | high-perceptual-quality-jpeg-decoding-via | 2211.11827 | null | https://arxiv.org/abs/2211.11827v1 | https://arxiv.org/pdf/2211.11827v1.pdf | High-Perceptual Quality JPEG Decoding via Posterior Sampling | JPEG is arguably the most popular image coding format, achieving high compression ratios via lossy quantization that may create visual artifacts degradation. Numerous attempts to remove these artifacts were conceived over the years, and common to most of these is the use of deterministic post-processing algorithms that optimize some distortion measure (e.g., PSNR, SSIM). In this paper we propose a different paradigm for JPEG artifact correction: Our method is stochastic, and the objective we target is high perceptual quality -- striving to obtain sharp, detailed and visually pleasing reconstructed images, while being consistent with the compressed input. These goals are achieved by training a stochastic conditional generator (conditioned on the compressed input), accompanied by a theoretically well-founded loss term, resulting in a sampler from the posterior distribution. Our solution offers a diverse set of plausible and fast reconstructions for a given input with perfect consistency. We demonstrate our scheme's unique properties and its superiority to a variety of alternative methods on the FFHQ and ImageNet datasets. | ['Michael Elad', 'Theo Adrai', 'Guy Ohayon', 'Sean Man'] | 2022-11-21 | null | null | null | null | ['jpeg-artifact-correction'] | ['computer-vision'] | [ 7.14825928e-01 1.38397776e-02 1.14518449e-01 -2.87094831e-01
-8.77863884e-01 -2.47920245e-01 4.75540429e-01 -1.39802266e-02
-3.26901108e-01 7.82630920e-01 3.64902765e-01 8.10248032e-02
-1.43104374e-01 -6.88661337e-01 -9.02292430e-01 -7.67554998e-01
-2.95644328e-02 9.05314907e-02 1.88828602e-01 -7.04860911e-02
4.97502834e-01 3.93830389e-01 -1.66610205e+00 4.62198853e-02
9.92173195e-01 1.06324828e+00 5.31877458e-01 8.61473680e-01
2.62744665e-01 9.65510964e-01 -5.08307278e-01 -8.43091190e-01
3.38845640e-01 -6.98973417e-01 -7.00310946e-01 3.11437845e-01
3.10713083e-01 -4.28135902e-01 -3.81108969e-01 1.48329794e+00
4.67993855e-01 -5.63367549e-03 7.28542566e-01 -8.70174706e-01
-6.66553199e-01 2.52458602e-01 -4.87004817e-01 -1.02501828e-02
4.48454797e-01 3.20608705e-01 9.42533135e-01 -6.76135957e-01
6.91133678e-01 1.24082768e+00 7.45770633e-01 3.89158666e-01
-1.63954401e+00 -2.36124173e-01 -4.90107149e-01 3.06859851e-01
-1.60509181e+00 -7.47164488e-01 7.64457881e-01 -2.99313039e-01
5.30929029e-01 3.24569464e-01 5.07439792e-01 1.03760743e+00
5.49795270e-01 5.24264216e-01 1.06196630e+00 -5.27479231e-01
4.23445463e-01 1.18624218e-01 -6.33914411e-01 4.63163227e-01
1.44084141e-01 1.97128162e-01 -6.50602460e-01 -1.36664892e-02
9.14491475e-01 -1.33185491e-01 -7.49819934e-01 -3.93896937e-01
-9.97118771e-01 6.32786453e-01 3.27540278e-01 1.30093783e-01
-6.76089585e-01 4.13273931e-01 1.53111979e-01 2.65159220e-01
3.38302702e-01 2.50790358e-01 7.28264153e-02 -9.07293037e-02
-1.37705505e+00 3.59396517e-01 5.68037987e-01 8.58703673e-01
5.80010653e-01 8.66116360e-02 -1.83186218e-01 8.10236096e-01
4.75206494e-01 5.27865529e-01 4.68876332e-01 -1.37309444e+00
3.36784959e-01 -2.15660587e-01 2.33074799e-01 -1.36118042e+00
2.93778509e-01 -5.70629776e-01 -1.13855088e+00 5.69563687e-01
2.13888437e-01 4.20336187e-01 -7.26099133e-01 1.89306569e+00
-1.02092519e-01 8.98070931e-02 1.15593947e-01 8.48302066e-01
1.92967549e-01 8.21838856e-01 6.37157187e-02 -3.96847218e-01
8.40075016e-01 -5.11109054e-01 -9.22269404e-01 -4.08935919e-02
-1.64347365e-01 -1.03135180e+00 1.08934951e+00 8.03628385e-01
-1.64882052e+00 -6.39642298e-01 -1.27883899e+00 -1.55280426e-01
1.79608062e-01 -1.87250584e-01 5.37042320e-02 6.64931655e-01
-1.25713110e+00 1.03742945e+00 -6.62854373e-01 -4.94076274e-02
3.83520246e-01 5.06903455e-02 -1.82507202e-01 -2.04704016e-01
-8.92404616e-01 9.54894781e-01 2.69440174e-01 -9.82834697e-02
-1.09269285e+00 -4.39034402e-01 -6.53731167e-01 1.85300872e-01
1.69247873e-02 -9.55272079e-01 1.18399489e+00 -1.16638041e+00
-1.46370578e+00 8.61787975e-01 -1.16965979e-01 -7.63948321e-01
9.78843510e-01 -1.11471072e-01 -2.21456543e-01 4.46253002e-01
6.17872365e-02 6.73928499e-01 1.36968243e+00 -1.60501218e+00
-2.72569209e-01 -1.97943926e-01 -2.67793506e-01 2.27981567e-01
-2.91071624e-01 -1.28903523e-01 -5.56266308e-01 -8.57971787e-01
2.20888346e-01 -5.49248755e-01 -2.86892563e-01 3.95156026e-01
-4.17473525e-01 4.53708380e-01 3.72347146e-01 -9.44821179e-01
1.20799625e+00 -2.23965788e+00 2.83816338e-01 1.65744632e-01
2.42528543e-01 3.53243649e-02 -4.89212722e-02 4.46552366e-01
6.80090636e-02 4.20071110e-02 -7.32209086e-01 -5.70270717e-01
-1.03628702e-01 3.08629751e-01 -3.55560571e-01 7.41588056e-01
1.28264010e-01 5.62183201e-01 -9.38388228e-01 -6.70710683e-01
4.17344272e-01 7.87283957e-01 -7.18585134e-01 4.08533126e-01
-1.73540786e-01 4.20890808e-01 -2.18914077e-02 3.07967663e-01
8.97463858e-01 -2.43066594e-01 3.04785073e-02 -3.63956541e-01
4.28408720e-02 1.29989833e-02 -1.20049214e+00 1.80034232e+00
-4.46906179e-01 6.28342867e-01 1.84645817e-01 -7.30945826e-01
7.24966526e-01 3.07549000e-01 4.45478767e-01 -6.80932939e-01
1.09214649e-01 2.78395683e-01 -3.67709190e-01 -5.46809137e-01
6.55261755e-01 -3.94845694e-01 3.67795169e-01 7.32258111e-02
1.65032297e-01 -4.07230943e-01 6.32232130e-02 2.51713991e-01
9.48463738e-01 1.09322071e-01 4.35131311e-01 -1.93747804e-01
4.32761759e-01 -3.69188994e-01 4.33474213e-01 8.59240651e-01
-1.53127685e-01 1.23176157e+00 2.03989834e-01 -1.16904318e-01
-1.53308046e+00 -1.41282821e+00 -1.36249900e-01 2.27491871e-01
3.24702471e-01 -3.02387744e-01 -7.21058190e-01 -7.06613138e-02
-2.51849651e-01 6.99342251e-01 -2.85536528e-01 -2.13502809e-01
-3.90886635e-01 -4.35519844e-01 4.16188866e-01 5.93748838e-02
7.30327368e-01 -8.71191144e-01 -7.43756235e-01 4.96017486e-01
-5.41739821e-01 -1.07811940e+00 -4.21688855e-01 4.02190536e-02
-1.08491385e+00 -8.25049162e-01 -9.37089384e-01 -5.40381610e-01
6.48098230e-01 1.75635636e-01 1.36794353e+00 6.98044449e-02
-2.43010610e-01 2.28586242e-01 -1.71400800e-01 1.52128469e-02
-7.50215590e-01 -5.92941999e-01 -1.24149866e-01 2.47103900e-01
-3.94772261e-01 -9.17530417e-01 -7.68643558e-01 -4.86278310e-02
-1.36876953e+00 2.01533124e-01 6.24831796e-01 8.45784664e-01
8.56587887e-01 4.19452488e-01 2.06092671e-02 -6.66777372e-01
5.78179479e-01 -2.45875016e-01 -4.30966169e-01 1.01076839e-02
-8.35432291e-01 5.98669164e-02 9.93679881e-01 -1.43473342e-01
-1.06094909e+00 2.53404528e-02 -5.12716770e-01 -5.13248265e-01
-8.03850144e-02 2.20990926e-01 -1.83900952e-01 -1.45476505e-01
7.64307499e-01 7.38497853e-01 2.85855345e-02 -4.72978115e-01
4.05734181e-01 4.23048645e-01 1.06759691e+00 -4.61634845e-01
8.44267428e-01 5.44617712e-01 1.31492302e-01 -9.27233875e-01
-4.22433883e-01 -1.59084514e-01 -2.32497677e-01 -4.14867997e-01
7.26256132e-01 -8.25989664e-01 -4.19237554e-01 5.09934664e-01
-1.23566127e+00 -1.58270076e-02 -4.70258683e-01 3.41677040e-01
-1.04927158e+00 7.30760694e-01 -5.96087217e-01 -9.02797163e-01
-2.53619164e-01 -1.17552280e+00 8.33441019e-01 9.39428955e-02
-4.07206640e-02 -7.91218042e-01 7.17524290e-02 4.60250936e-02
5.85151911e-01 3.61120194e-01 7.41321325e-01 3.15074891e-01
-7.02496111e-01 1.03709660e-01 -3.25175524e-01 8.59393895e-01
3.77986319e-02 3.32478546e-02 -9.73989546e-01 -3.85222971e-01
4.38255072e-01 -2.68027395e-01 8.97167563e-01 6.18800044e-01
1.35534143e+00 -6.58075392e-01 1.00114524e-01 9.50393856e-01
2.08484197e+00 1.07880935e-01 1.12024498e+00 2.10954472e-01
2.33601883e-01 3.11885357e-01 2.52404869e-01 5.90742946e-01
7.72154704e-03 6.75573766e-01 7.09168255e-01 -1.87839419e-02
-4.56008464e-01 -3.99598062e-01 1.88653216e-01 7.41813958e-01
-1.19360179e-01 -4.46410954e-01 -4.73380148e-01 5.94143748e-01
-1.60034716e+00 -1.16671526e+00 6.63482547e-02 2.44376993e+00
9.64878261e-01 2.22873241e-01 -2.34087527e-01 5.06202817e-01
5.57187676e-01 2.49105573e-01 -2.87957013e-01 -1.36760667e-01
-2.64093727e-01 2.35636562e-01 6.08860612e-01 5.78076065e-01
-8.40661585e-01 3.43981862e-01 6.93067360e+00 8.42413902e-01
-9.11013603e-01 -4.15963074e-03 7.75743067e-01 1.64104685e-01
-4.85517949e-01 -1.18020326e-01 -1.54352099e-01 7.20068157e-01
9.74133193e-01 -1.60007209e-01 5.52665472e-01 5.47902942e-01
4.69801039e-01 -2.00862572e-01 -8.29506457e-01 1.18958724e+00
1.39062926e-01 -1.38270700e+00 2.43417278e-01 -1.44089498e-02
7.29189396e-01 -2.22331583e-01 2.66896009e-01 -4.24283653e-01
1.23475507e-01 -9.83944416e-01 1.21209347e+00 6.65332556e-01
1.00756133e+00 -7.74820387e-01 4.57019985e-01 2.47171357e-01
-7.43760943e-01 8.65463912e-02 -4.53510284e-01 2.65555829e-01
4.14049715e-01 9.06841338e-01 -2.90693879e-01 5.36837697e-01
6.93056345e-01 7.41757095e-01 -3.27274919e-01 1.40852869e+00
-2.64876544e-01 5.62265813e-01 -1.69706136e-01 4.14760977e-01
1.25820972e-02 -2.74516046e-01 7.33284473e-01 1.30012453e+00
5.20772994e-01 -5.59710115e-02 -1.32336900e-01 8.96481395e-01
-1.64420068e-01 4.41830903e-02 -5.93445182e-01 2.54993349e-01
2.31073499e-01 7.45619476e-01 -5.24403453e-01 -3.07821214e-01
-1.85356930e-01 1.43294775e+00 2.70989742e-02 3.01044494e-01
-8.55238974e-01 -3.58068645e-01 6.05626345e-01 1.73981085e-01
3.03132623e-01 -1.20638773e-01 -4.82145727e-01 -1.18151557e+00
1.59080893e-01 -9.37353551e-01 1.99046526e-02 -9.36521351e-01
-1.21978819e+00 6.87825799e-01 -2.32007682e-01 -1.54313612e+00
-2.09488630e-01 -1.85393021e-01 -2.38952339e-01 9.52706158e-01
-1.83462226e+00 -6.11230373e-01 -3.09626788e-01 7.15577841e-01
5.82067013e-01 1.77683547e-01 7.31334925e-01 5.06483793e-01
2.02658419e-02 4.45002764e-01 3.48829299e-01 -4.25730020e-01
6.21136129e-01 -1.19131100e+00 1.18709534e-01 1.18857527e+00
4.63816263e-02 3.12607080e-01 1.43027830e+00 -3.83384168e-01
-1.31493425e+00 -1.00363266e+00 9.49245811e-01 4.10052687e-02
2.80847758e-01 6.47322834e-02 -8.77414167e-01 3.94131154e-01
2.96942055e-01 -1.35095149e-01 2.05830380e-01 -7.31512845e-01
-3.14503700e-01 -1.42870650e-01 -1.48488784e+00 5.44263661e-01
8.34293544e-01 -5.23091018e-01 -4.34956282e-01 1.92122161e-01
5.04079103e-01 -3.29699934e-01 -6.74664974e-01 1.23106211e-01
4.85914320e-01 -1.45154202e+00 1.22162700e+00 -3.18671018e-02
1.03282118e+00 -3.78963023e-01 -4.51224506e-01 -1.28599250e+00
-4.26481277e-01 -7.70276248e-01 -1.25726745e-01 1.04538238e+00
2.13297661e-02 -3.20990503e-01 6.12725019e-01 2.00763687e-01
-1.03669465e-01 -5.00157177e-01 -9.20423925e-01 -8.98467481e-01
-3.73480767e-01 -5.14870584e-01 3.52371961e-01 4.92541730e-01
-3.56851846e-01 -8.28048289e-02 -8.71826649e-01 1.65180877e-01
1.31185436e+00 -1.44631848e-01 5.30326486e-01 -8.03568840e-01
-6.35151386e-01 -4.09328461e-01 -4.82016534e-01 -1.37652469e+00
-4.55404103e-01 -5.83630025e-01 4.16612029e-01 -1.37795138e+00
2.12775469e-01 -1.96870595e-01 -1.37982294e-01 -1.68716148e-01
3.59409191e-02 5.71648479e-01 2.06368357e-01 4.07651126e-01
-3.74262542e-01 7.13358879e-01 1.39250219e+00 -1.15341835e-01
1.15386046e-01 -8.12702179e-02 -5.61302543e-01 6.11967266e-01
4.86616462e-01 -6.50420368e-01 -4.74759370e-01 -6.22018754e-01
2.18157992e-01 3.92649680e-01 5.15091419e-01 -1.33931088e+00
1.41657338e-01 -4.90947850e-02 5.35896063e-01 -3.19065571e-01
5.22511482e-01 -8.74605894e-01 5.76056778e-01 5.28407693e-01
-5.34613788e-01 -1.93715263e-02 -2.36759529e-01 7.65774727e-01
-5.30657053e-01 -4.06218886e-01 1.38257539e+00 -3.00950468e-01
-5.13856888e-01 1.67331263e-01 -2.61064261e-01 -3.87299955e-02
6.50652409e-01 -3.02211940e-01 7.54992515e-02 -6.96491718e-01
-5.95090330e-01 -4.00197059e-01 8.23554099e-01 1.77441386e-03
1.02982867e+00 -1.21538568e+00 -8.92104208e-01 2.28186488e-01
-1.18964680e-01 -2.85632640e-01 2.28315577e-01 4.75085616e-01
-8.65359187e-01 6.63922280e-02 -4.03914869e-01 -5.64389884e-01
-1.01309812e+00 4.12570804e-01 2.87049860e-01 -2.56504148e-01
-8.75313997e-01 7.90335953e-01 -3.84306200e-02 3.10598284e-01
3.66230726e-01 -9.65499133e-02 1.12585105e-01 -4.92058992e-01
6.06229246e-01 2.80901611e-01 3.54800336e-02 -6.25352681e-01
-2.48151347e-02 3.92076075e-01 2.84687728e-01 -3.99338514e-01
1.43870580e+00 -5.22578955e-01 2.18667071e-02 4.04035822e-02
1.28365839e+00 3.00984224e-03 -1.57522523e+00 -1.27856314e-01
-2.91758418e-01 -1.06330371e+00 2.05030739e-01 -6.34003639e-01
-1.14200580e+00 6.62896574e-01 8.08054924e-01 3.90608966e-01
1.56430244e+00 -3.17096472e-01 9.16149199e-01 -2.38434449e-01
5.78829348e-01 -9.12210822e-01 7.80854002e-02 2.39572152e-02
1.06032765e+00 -1.05576086e+00 1.21038042e-01 -4.09508497e-01
-3.33536476e-01 1.01580691e+00 -1.29775349e-02 -3.65191370e-01
3.96327227e-01 1.75967082e-01 -8.64296407e-02 9.04494449e-02
-6.23794734e-01 1.29801512e-01 1.39722405e-02 7.79582143e-01
2.25220472e-01 -9.15273130e-02 -5.47195077e-01 4.87488098e-02
-2.42552295e-01 1.27058581e-01 6.77492976e-01 7.78356075e-01
-4.06950414e-01 -9.26736295e-01 -5.75524330e-01 3.63027424e-01
-6.82620168e-01 -1.43163338e-01 2.38585085e-01 2.98468709e-01
1.71875373e-01 9.78560209e-01 -2.05245912e-01 -2.94525266e-01
2.39441410e-01 -3.65330368e-01 7.04934835e-01 -1.98157147e-01
-2.68387705e-01 1.75913945e-01 -2.81465918e-01 -8.25746715e-01
-4.65096205e-01 -6.95969820e-01 -7.94312596e-01 -5.02769947e-01
-3.44290324e-02 -2.79642437e-02 7.87651360e-01 5.72324336e-01
6.58226088e-02 4.16661680e-01 7.62442350e-01 -9.54892755e-01
-8.16649139e-01 -5.31399250e-01 -7.19051600e-01 8.12069595e-01
5.34167290e-01 -3.02918017e-01 -6.12028480e-01 5.69595039e-01] | [11.4775972366333, -1.893999457359314] |
9fa2ffd6-3fed-4f6a-b362-04f0f04a70a5 | ai-playground-unreal-engine-based-data | 2007.06153 | null | https://arxiv.org/abs/2007.06153v1 | https://arxiv.org/pdf/2007.06153v1.pdf | AI Playground: Unreal Engine-based Data Ablation Tool for Deep Learning | Machine learning requires data, but acquiring and labeling real-world data is challenging, expensive, and time-consuming. More importantly, it is nearly impossible to alter real data post-acquisition (e.g., change the illumination of a room), making it very difficult to measure how specific properties of the data affect performance. In this paper, we present AI Playground (AIP), an open-source, Unreal Engine-based tool for generating and labeling virtual image data. With AIP, it is trivial to capture the same image under different conditions (e.g., fidelity, lighting, etc.) and with different ground truths (e.g., depth or surface normal values). AIP is easily extendable and can be used with or without code. To validate our proposed tool, we generated eight datasets of otherwise identical but varying lighting and fidelity conditions. We then trained deep neural networks to predict (1) depth values, (2) surface normals, or (3) object labels and assessed each network's intra- and cross-dataset performance. Among other insights, we verified that sensitivity to different settings is problem-dependent. We confirmed the findings of other studies that segmentation models are very sensitive to fidelity, but we also found that they are just as sensitive to lighting. In contrast, depth and normal estimation models seem to be less sensitive to fidelity or lighting and more sensitive to the structure of the image. Finally, we tested our trained depth-estimation networks on two real-world datasets and obtained results comparable to training on real data alone, confirming that our virtual environments are realistic enough for real-world tasks. | ['Aashis Khanal', 'Mehdi Mousavi', 'Rolando Estrada'] | 2020-07-13 | null | null | null | null | ['data-ablation', 'indoor-monocular-depth-estimation', 'surface-normals-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.76541999e-01 -1.58816919e-01 2.62677938e-01 -5.39020181e-01
-4.95073736e-01 -9.32075679e-01 5.18628716e-01 1.40681997e-01
-5.34324765e-01 5.67419767e-01 -1.74104437e-01 -2.46166557e-01
1.07102461e-01 -1.00637555e+00 -1.04238105e+00 -5.36635935e-01
1.67254493e-01 3.95838559e-01 5.13688564e-01 1.11170702e-01
4.25827116e-01 8.17371428e-01 -2.13622832e+00 2.32629821e-01
5.18360674e-01 1.09290338e+00 2.21942797e-01 7.69955993e-01
5.55492453e-02 4.24291015e-01 -7.32320905e-01 -5.15071526e-02
5.37661672e-01 -1.81029186e-01 -6.44723713e-01 2.03702107e-01
8.83213162e-01 -4.73056376e-01 -2.50837535e-01 7.65704989e-01
4.68870372e-01 4.91890945e-02 5.42567730e-01 -1.19499552e+00
-3.53622526e-01 -5.10766990e-02 -2.78449655e-01 -5.65893576e-03
4.68332916e-01 5.23368657e-01 4.47318405e-01 -4.03869361e-01
6.94977105e-01 9.53985512e-01 6.97251976e-01 5.21215498e-01
-1.28884482e+00 -5.30363679e-01 -3.07413917e-02 -3.26551162e-02
-1.30786037e+00 -4.44101185e-01 6.26229405e-01 -6.37594402e-01
6.79988980e-01 2.51976758e-01 8.19007218e-01 1.38768518e+00
2.81873584e-01 2.32326284e-01 1.43066263e+00 -4.31887180e-01
4.94846135e-01 2.78786659e-01 -2.48139635e-01 6.05403125e-01
1.85069844e-01 3.04731429e-01 -2.22618282e-01 1.97291404e-01
1.02156281e+00 -2.91295677e-01 -4.10205543e-01 -5.51322818e-01
-1.29832160e+00 6.55150235e-01 3.87956947e-01 2.42921710e-01
-8.25507492e-02 2.26176679e-01 2.72410393e-01 2.26224333e-01
1.48294047e-01 6.84321880e-01 -5.54093838e-01 -3.52824271e-01
-7.39188135e-01 1.26367018e-01 7.12592006e-01 6.65627301e-01
8.67418468e-01 -1.03221051e-02 -2.85644438e-02 6.41081095e-01
8.06933939e-02 3.53042692e-01 5.47052383e-01 -1.29020905e+00
2.68435448e-01 1.61432415e-01 2.30895624e-01 -1.13309133e+00
-7.44937837e-01 -1.46839581e-02 -3.65632951e-01 7.16958463e-01
9.74208891e-01 -2.48029262e-01 -1.09880388e+00 1.68172050e+00
2.06980795e-01 6.15479238e-02 -1.60357013e-01 1.13465524e+00
8.30809355e-01 3.45598608e-01 -4.34960648e-02 1.69969603e-01
1.11256444e+00 -6.23709440e-01 -3.91167343e-01 -4.70067471e-01
5.27125299e-01 -8.27068985e-01 1.56653106e+00 6.00987136e-01
-1.06204677e+00 -6.79946423e-01 -1.08072078e+00 -9.40101035e-03
-4.96986181e-01 -1.59624591e-01 7.51463294e-01 8.77004802e-01
-1.16595697e+00 9.22231495e-01 -8.67654800e-01 -4.92226899e-01
2.87588984e-01 2.18256310e-01 -4.06104386e-01 -1.16228051e-01
-9.25676346e-01 8.22533846e-01 3.38446140e-01 -2.52013765e-02
-1.00199699e+00 -6.70945346e-01 -8.22328508e-01 -2.88593203e-01
1.57846540e-01 -4.35592473e-01 1.22967446e+00 -1.04460239e+00
-1.64018953e+00 1.02847111e+00 1.47966728e-01 3.52392113e-03
6.20204926e-01 1.29870567e-02 -3.36230993e-01 -1.85915790e-02
-1.80002451e-02 9.53405619e-01 5.32521129e-01 -1.64031851e+00
-3.30660433e-01 -3.91198695e-01 3.52949440e-01 6.46469072e-02
-7.99916759e-02 -2.40177467e-01 -3.76523256e-01 -1.10129118e-01
1.48286894e-01 -8.89973044e-01 -2.60944087e-02 4.55518752e-01
-3.41983885e-01 4.58965778e-01 6.44321740e-01 -5.44044495e-01
5.84476173e-01 -2.22208428e+00 -3.11241388e-01 2.22077474e-01
-2.24295389e-02 1.31623909e-01 -1.31081685e-01 1.57920331e-01
-1.72676742e-02 2.37016872e-01 -1.96284577e-01 -1.59779534e-01
-1.24574415e-01 3.32712024e-01 1.49271205e-01 5.91793418e-01
-5.70500121e-02 5.44608355e-01 -7.97876298e-01 -5.50233841e-01
7.70840466e-01 7.51010418e-01 -4.33145165e-01 6.10442981e-02
-1.91865161e-01 6.59278870e-01 -1.79083254e-02 5.02894580e-01
6.86002374e-01 -2.44532842e-02 -8.83441940e-02 -3.49253327e-01
-1.36381015e-01 1.59203842e-01 -1.31559932e+00 1.63372695e+00
-9.53983903e-01 1.08451056e+00 -7.73450285e-02 -5.94782948e-01
9.64673638e-01 1.91831395e-01 3.56510043e-01 -1.11651206e+00
4.39649343e-01 7.47847334e-02 -5.75402416e-02 -6.72089696e-01
5.69166958e-01 -2.27018967e-02 1.65916219e-01 3.93918186e-01
-2.77723700e-01 -6.46104991e-01 3.96533199e-02 -2.87405640e-01
1.07434022e+00 2.58929461e-01 3.20946909e-02 -2.23190673e-02
7.58633157e-03 2.24713627e-02 2.84491360e-01 4.44018483e-01
-3.00521940e-01 1.18780041e+00 5.19131303e-01 -4.75804061e-01
-1.26931357e+00 -1.19367838e+00 -3.99665177e-01 6.58921242e-01
3.81774902e-01 -5.69115654e-02 -8.90087485e-01 -4.46257353e-01
-2.14915320e-01 9.27560687e-01 -7.57780790e-01 -1.02378592e-01
-2.56982088e-01 -4.16308522e-01 4.99086857e-01 5.52438378e-01
4.18321580e-01 -1.14628398e+00 -1.05961597e+00 -2.65609529e-02
-5.77637628e-02 -1.39044130e+00 -1.54917225e-01 2.04313353e-01
-7.86277413e-01 -1.04822075e+00 -2.73147315e-01 -4.45482880e-01
7.64275432e-01 2.00422451e-01 1.38392961e+00 1.78984895e-01
-2.61323422e-01 4.35054511e-01 -3.23500037e-01 -4.02548760e-01
-3.47719699e-01 -7.53925368e-02 -1.80063441e-01 -2.44492218e-01
-1.34277478e-01 -5.68673253e-01 -7.33838677e-01 6.59882665e-01
-1.16044366e+00 7.54464194e-02 1.89233154e-01 3.66287977e-01
7.97730029e-01 1.11024700e-01 1.42077135e-03 -7.12842703e-01
4.10345227e-01 -1.18140109e-01 -7.74858534e-01 -1.23037072e-02
-5.14361799e-01 -1.15991451e-01 5.90956867e-01 -4.62091774e-01
-8.90157521e-01 1.85326353e-01 -1.18140467e-01 -3.35095584e-01
-6.19197786e-01 5.59567399e-02 -2.41831556e-01 -1.56613216e-01
9.04947162e-01 -2.25012541e-01 -1.68511599e-01 -8.43321159e-02
1.15525037e-01 5.96985817e-01 5.17774820e-01 -5.92753410e-01
5.91823339e-01 7.84565210e-01 -1.61845043e-01 -7.63248324e-01
-7.58802235e-01 3.41728181e-02 -7.43366182e-01 -4.48651612e-01
9.50676084e-01 -6.25751019e-01 -5.40752530e-01 6.10000432e-01
-1.07622480e+00 -9.08216476e-01 -3.35845113e-01 5.16128540e-01
-5.63471317e-01 3.16228829e-02 -2.60528386e-01 -4.62288886e-01
1.71548814e-01 -1.24998760e+00 1.10033357e+00 3.50598186e-01
-3.16864550e-01 -1.13424611e+00 1.25688286e-02 5.02559125e-01
3.73873085e-01 6.42957687e-01 5.09019792e-01 -4.38397340e-02
-4.49687868e-01 -1.14691906e-01 -4.33591694e-01 3.81682247e-01
2.23034605e-01 5.08375823e-01 -1.33159626e+00 -1.27411038e-01
-1.83043003e-01 -4.65550363e-01 2.93179572e-01 3.77077490e-01
1.65925264e+00 5.98840974e-02 -5.16160280e-02 7.14229703e-01
1.60549664e+00 2.30268523e-01 1.08667505e+00 7.66277134e-01
8.09419096e-01 7.87545621e-01 6.60048664e-01 2.94180840e-01
1.99003667e-01 9.36752200e-01 6.72141612e-01 -3.40844452e-01
-1.87634349e-01 -1.50916889e-01 1.85775340e-01 1.82889298e-01
-9.97546762e-02 -3.12905192e-01 -9.62122560e-01 3.96701425e-01
-1.24773324e+00 -7.54930377e-01 -3.14785361e-01 2.41437149e+00
6.51248038e-01 2.18848616e-01 -4.36861478e-02 2.25030571e-01
4.81503427e-01 -4.46509495e-02 -4.84186351e-01 -8.09689879e-01
3.05508543e-02 2.30938435e-01 6.79674923e-01 3.57568681e-01
-8.06140184e-01 7.67922223e-01 6.65026760e+00 3.64053756e-01
-1.60266995e+00 -6.66257814e-02 6.68582320e-01 -2.95897990e-01
-3.73610675e-01 -1.72348350e-01 -2.58140326e-01 5.68875313e-01
9.06534076e-01 2.54428715e-01 5.65761924e-01 9.04761136e-01
3.65996182e-01 -4.80104268e-01 -1.28911340e+00 9.21256244e-01
2.63174884e-02 -1.11866057e+00 -4.19715643e-01 3.58423521e-03
7.80468643e-01 2.00040460e-01 -1.76865458e-01 -2.37682089e-02
2.77229756e-01 -1.28638709e+00 8.83035123e-01 4.49882746e-01
9.64524209e-01 -5.99334300e-01 7.58628666e-01 1.80505678e-01
-7.52490520e-01 2.24419579e-01 -2.24380717e-01 -1.34952426e-01
1.13044210e-01 5.94360054e-01 -7.37420321e-01 1.85837254e-01
1.01154912e+00 1.40198812e-01 -7.55096912e-01 9.73300517e-01
-2.56165802e-01 3.74970675e-01 -4.44673151e-01 2.09568739e-01
5.99465109e-02 -6.02115951e-02 -9.81700569e-02 1.02046347e+00
2.72402644e-01 -2.49777600e-01 -1.09460711e-01 8.32146525e-01
1.74754962e-01 -8.40004086e-02 -7.02658057e-01 2.34270796e-01
4.60725427e-01 1.27166545e+00 -1.15083909e+00 -1.76798180e-02
-4.06671613e-01 9.13224220e-01 1.54503673e-01 3.39731485e-01
-1.08042991e+00 -2.01986358e-01 6.26717329e-01 4.38284278e-01
1.73633754e-01 -2.87226737e-01 -5.47968388e-01 -8.10108602e-01
1.13075942e-01 -7.28368282e-01 -5.43961227e-02 -1.26029921e+00
-1.01866293e+00 6.07111454e-01 2.66625397e-02 -1.33491790e+00
-1.04065411e-01 -6.84301019e-01 -5.13889909e-01 6.24961078e-01
-1.40692878e+00 -7.46253192e-01 -1.04129362e+00 5.26478469e-01
1.76655039e-01 4.22410160e-01 7.52338648e-01 3.96424979e-01
-5.09072721e-01 4.29849714e-01 -7.83655941e-02 1.38477251e-01
6.81783736e-01 -1.02783239e+00 2.88721412e-01 5.21491528e-01
2.35772297e-01 3.06589067e-01 8.66825163e-01 -2.40327284e-01
-1.11855793e+00 -9.15095687e-01 2.12992147e-01 -5.90263546e-01
4.01052564e-01 -4.41411942e-01 -9.28785086e-01 6.99120700e-01
-1.21572569e-01 2.46655941e-01 3.73640239e-01 -1.28886640e-01
-2.20988005e-01 -1.42370030e-01 -1.46523833e+00 5.28398395e-01
1.12394977e+00 -3.60169262e-01 -1.22702889e-01 2.45811671e-01
5.97528338e-01 -8.83294821e-01 -8.25806022e-01 3.63512516e-01
7.55016685e-01 -1.63043475e+00 9.45922673e-01 -7.33014047e-02
7.30548263e-01 -3.35219413e-01 -3.21012557e-01 -1.44992399e+00
-1.00535592e-02 3.86242606e-02 3.75556111e-01 1.17816699e+00
5.18864453e-01 -6.86003268e-01 9.45978403e-01 9.58677351e-01
-2.69780338e-01 -7.15321779e-01 -7.77765214e-01 -7.55859613e-01
7.05374777e-02 -8.85336399e-01 7.28835464e-01 1.04074585e+00
-5.18632650e-01 -8.11237097e-03 6.82222098e-02 2.34200165e-01
1.87619537e-01 -8.56056809e-02 1.02325940e+00 -1.14587605e+00
-3.41981411e-01 -3.74940693e-01 -5.60600460e-01 -4.91138458e-01
5.40356003e-02 -5.36545098e-01 1.12551637e-01 -1.76171458e+00
-2.61405110e-01 -8.52943778e-01 1.29620805e-01 4.23735648e-01
1.38117790e-01 5.51672280e-01 -7.42677879e-03 2.96376068e-02
-2.44938806e-01 2.40038350e-01 1.45085585e+00 1.53765753e-01
-2.55343735e-01 -1.52071252e-01 -5.12243211e-01 7.59961009e-01
1.01001382e+00 -2.74560571e-01 -5.22745430e-01 -7.54061937e-01
2.21387073e-01 -1.65433139e-01 5.60037494e-01 -1.40281796e+00
-1.79076433e-01 -3.08232665e-01 6.48710549e-01 -2.49008730e-01
4.69219029e-01 -1.02654624e+00 2.74907291e-01 2.45587751e-01
-6.21124320e-02 -1.02691121e-01 5.50969422e-01 2.92223394e-02
8.57632980e-02 -3.60354215e-01 7.82684982e-01 -1.34065449e-01
-7.58596301e-01 1.11475050e-01 -1.72386393e-01 1.90320853e-02
1.18822658e+00 -8.56354773e-01 -2.84658492e-01 -3.90366793e-01
-3.50210369e-01 -8.26200992e-02 1.10024130e+00 4.28634137e-01
5.46400785e-01 -1.09012437e+00 -3.91934842e-01 2.96390563e-01
3.13735992e-01 4.07330900e-01 2.59493411e-01 4.38033283e-01
-9.42334950e-01 5.80343306e-02 -4.40979093e-01 -9.23010170e-01
-1.13502753e+00 4.48680073e-01 5.90406597e-01 1.89491749e-01
-4.51607466e-01 6.01543009e-01 7.25321099e-02 -6.94643676e-01
1.31162941e-01 -5.78623891e-01 6.35399818e-02 -1.52541786e-01
3.48939240e-01 1.80508062e-01 3.71238053e-01 -3.49552453e-01
-2.69451857e-01 8.68544757e-01 3.02086264e-01 -1.18357778e-01
1.05194855e+00 5.07521257e-03 2.95660883e-01 6.84522033e-01
1.04539812e+00 -6.40354678e-02 -1.65896797e+00 2.93573052e-01
-4.99103993e-01 -7.55715787e-01 2.63642818e-01 -7.10576117e-01
-1.41830528e+00 7.64514029e-01 8.61458004e-01 2.07050815e-01
1.01723909e+00 -1.12331405e-01 6.09994590e-01 1.20138697e-01
6.29014313e-01 -1.22753227e+00 1.98036447e-01 1.30064324e-01
7.33251333e-01 -1.36148453e+00 1.45303421e-02 -4.34740454e-01
-5.52538216e-01 1.00999534e+00 9.04814005e-01 1.92363858e-01
3.99899215e-01 5.15234888e-01 5.11645436e-01 -3.16647887e-01
-3.68378431e-01 1.50981024e-01 -4.82851192e-02 1.00833154e+00
5.47281802e-01 7.17445612e-02 1.08042821e-01 -2.42029369e-01
-7.67790377e-01 9.84907970e-02 8.10396969e-01 9.15688992e-01
-1.51564315e-01 -8.59073281e-01 -7.41058469e-01 4.23453301e-01
-1.20801955e-01 2.75041223e-01 -3.09329003e-01 1.11396182e+00
4.21983778e-01 9.95070815e-01 4.95820433e-01 -5.44119477e-01
4.65465188e-01 -2.55298406e-01 7.43831575e-01 -6.44781172e-01
-3.42580378e-01 -5.65558076e-01 1.02973960e-01 -8.33503067e-01
-3.64250839e-01 -7.67071068e-01 -1.46467721e+00 -5.41631818e-01
-1.72256500e-01 -3.33450109e-01 1.12333941e+00 7.91532338e-01
2.48969018e-01 7.16735363e-01 5.69089592e-01 -1.06739616e+00
1.35648623e-01 -8.07343721e-01 -4.35458899e-01 6.68529272e-01
9.54044238e-02 -8.37168276e-01 -6.10583425e-01 7.85749257e-02] | [9.169790267944336, -2.381455183029175] |
88b8c4a1-2f53-4792-9fdc-cafd0047ce56 | active-neural-localization | 1801.08214 | null | http://arxiv.org/abs/1801.08214v1 | http://arxiv.org/pdf/1801.08214v1.pdf | Active Neural Localization | Localization is the problem of estimating the location of an autonomous agent
from an observation and a map of the environment. Traditional methods of
localization, which filter the belief based on the observations, are
sub-optimal in the number of steps required, as they do not decide the actions
taken by the agent. We propose "Active Neural Localizer", a fully
differentiable neural network that learns to localize accurately and
efficiently. The proposed model incorporates ideas of traditional
filtering-based localization methods, by using a structured belief of the state
with multiplicative interactions to propagate belief, and combines it with a
policy model to localize accurately while minimizing the number of steps
required for localization. Active Neural Localizer is trained end-to-end with
reinforcement learning. We use a variety of simulation environments for our
experiments which include random 2D mazes, random mazes in the Doom game engine
and a photo-realistic environment in the Unreal game engine. The results on the
2D environments show the effectiveness of the learned policy in an idealistic
setting while results on the 3D environments demonstrate the model's capability
of learning the policy and perceptual model jointly from raw-pixel based RGB
observations. We also show that a model trained on random textures in the Doom
environment generalizes well to a photo-realistic office space environment in
the Unreal engine. | ['Emilio Parisotto', 'Ruslan Salakhutdinov', 'Devendra Singh Chaplot'] | 2018-01-24 | active-neural-localization-1 | https://openreview.net/forum?id=ry6-G_66b | https://openreview.net/pdf?id=ry6-G_66b | iclr-2018-1 | ['game-of-doom', 'fps-games'] | ['playing-games', 'playing-games'] | [-2.56068438e-01 2.23707259e-02 2.18997523e-01 -3.10806662e-01
-7.72342682e-01 -6.09921873e-01 4.58040595e-01 -1.07650291e-02
-1.00172520e+00 7.76600480e-01 -4.78930920e-02 -1.31109357e-01
-1.12599336e-01 -8.28521192e-01 -1.01645017e+00 -7.93478906e-01
-6.27862573e-01 3.58749628e-01 4.88499045e-01 -1.36220440e-01
2.83450603e-01 6.54994011e-01 -1.20270574e+00 -2.72122890e-01
4.25915450e-01 1.14070642e+00 7.68176615e-01 1.00787234e+00
2.65768617e-01 1.27539146e+00 -6.20501757e-01 5.60640156e-01
5.77773094e-01 -4.90085222e-02 -3.77397776e-01 4.71866615e-02
1.81943685e-01 -2.83021569e-01 -5.02001882e-01 1.07440996e+00
5.34589589e-01 6.76853418e-01 3.09736818e-01 -9.68465209e-01
-5.90482414e-01 1.23700835e-01 -3.36294740e-01 1.71547249e-01
3.22167516e-01 3.68756384e-01 4.52406734e-01 -4.58105683e-01
4.74774837e-01 1.24184251e+00 8.07080746e-01 4.21288341e-01
-1.05548370e+00 -3.06699127e-01 3.82509112e-01 9.83097553e-02
-1.45302212e+00 -4.88824725e-01 5.26103258e-01 -2.36294255e-01
1.13806129e+00 -1.72646224e-01 6.85142696e-01 9.58751082e-01
7.25155473e-01 3.76985222e-01 1.39995778e+00 -4.90836620e-01
1.05706191e+00 -1.11012369e-01 -3.65839481e-01 1.28674757e+00
-8.69040042e-02 6.71549916e-01 -7.83911884e-01 -2.44307205e-01
1.23776793e+00 -9.73301847e-03 -2.66355306e-01 -8.21483552e-01
-1.15259457e+00 7.98462868e-01 7.69358099e-01 -2.08530053e-01
-7.01346576e-01 8.61842334e-01 -2.76751786e-01 1.46079838e-01
5.20809352e-01 6.06923461e-01 -4.04050916e-01 -8.44925642e-02
-7.76432872e-01 2.26489067e-01 8.36926401e-01 8.21009040e-01
9.21508372e-01 1.51789695e-01 1.30055144e-01 4.10535216e-01
7.81927884e-01 6.32090569e-01 4.98757094e-01 -1.43804240e+00
5.25754094e-02 2.70062745e-01 6.96936190e-01 -9.13285792e-01
-3.49259853e-01 -4.40566838e-01 -2.20256612e-01 1.15414548e+00
2.09886044e-01 -5.76256394e-01 -1.32532430e+00 1.78659201e+00
2.98009455e-01 5.45283318e-01 2.61617541e-01 9.78512287e-01
3.77831340e-01 6.56192005e-01 -1.29858762e-01 2.20822021e-01
6.97852135e-01 -1.18242252e+00 -4.71233517e-01 -7.80612469e-01
2.43071929e-01 -3.88559699e-01 7.65974820e-01 3.05794567e-01
-9.17848706e-01 -6.69227779e-01 -1.18304491e+00 2.68076479e-01
-6.38130128e-01 1.77413508e-01 7.33879685e-01 2.85277814e-01
-1.55091488e+00 5.04169285e-01 -1.52825975e+00 -5.07925332e-01
1.36089966e-01 6.59014285e-01 -4.06398654e-01 2.13080779e-01
-7.35838830e-01 1.15796268e+00 4.96608615e-01 2.46894225e-01
-1.62965488e+00 -1.35624141e-01 -9.88218725e-01 -6.35775253e-02
4.73840117e-01 -4.45164829e-01 1.41366839e+00 -7.95671582e-01
-1.86406696e+00 2.01583892e-01 -8.43807831e-02 -7.76837647e-01
3.67791772e-01 -2.44049251e-01 -2.11241059e-02 9.85348895e-02
9.04929712e-02 9.73365784e-01 5.65595210e-01 -1.40660131e+00
-9.24733877e-01 -2.07256466e-01 5.75432241e-01 5.31630754e-01
3.46559316e-01 -6.59469843e-01 -2.68753648e-01 -8.02733377e-02
4.33003336e-01 -9.55169797e-01 -8.52124989e-01 2.23033458e-01
6.61613280e-03 2.47430399e-01 7.71999240e-01 -3.49270076e-01
2.70011574e-01 -1.97902060e+00 -5.03033586e-02 2.02930659e-01
9.18715671e-02 -4.46900696e-01 -2.65981555e-01 2.93420285e-01
3.99134308e-01 -3.37998003e-01 1.13684826e-01 -6.14339054e-01
3.89933400e-02 4.33110774e-01 -3.75175327e-01 7.81044185e-01
-1.95100948e-01 8.09780121e-01 -1.21822381e+00 -1.33286774e-01
2.96681136e-01 4.25911248e-01 -3.08910400e-01 3.07219148e-01
-3.66189241e-01 6.07301354e-01 -4.11621958e-01 4.35146332e-01
2.95174748e-01 -1.66566044e-01 -1.90283298e-01 4.00924772e-01
-5.24292707e-01 1.13575660e-01 -1.21579695e+00 2.10926867e+00
-6.09369576e-01 6.52603626e-01 3.69734377e-01 -6.41200304e-01
9.87606645e-01 1.00556530e-01 2.87211895e-01 -7.67290294e-01
-6.22231029e-02 1.50750965e-01 -3.02720249e-01 -4.11868989e-01
5.81372440e-01 1.90005481e-01 -1.42184973e-01 3.12869728e-01
2.40786791e-01 -2.32405454e-01 -8.60696584e-02 -1.33015951e-02
1.66623271e+00 5.78660905e-01 1.05985254e-01 -7.09436536e-02
-5.68315089e-02 3.84265393e-01 4.05435115e-01 1.46817434e+00
-2.92694777e-01 3.25826347e-01 -7.49095976e-02 -4.93000954e-01
-7.45626271e-01 -1.24916041e+00 5.01228034e-01 1.00436080e+00
5.92831135e-01 -7.05053611e-03 -4.94493425e-01 -7.37522125e-01
-1.44121975e-01 6.44111574e-01 -7.94125080e-01 -1.18086338e-01
-4.28789824e-01 -5.16608298e-01 2.28505298e-01 4.37516928e-01
8.52359593e-01 -1.10906780e+00 -1.16928566e+00 2.54277200e-01
1.23680465e-01 -8.46928596e-01 2.62274337e-03 7.92832434e-01
-5.79839766e-01 -7.96019554e-01 -4.34249401e-01 -8.02261829e-01
9.73247290e-01 2.68030971e-01 8.40203047e-01 -3.20384115e-01
3.04102469e-02 8.70855749e-01 -1.90838009e-01 -4.88488466e-01
-8.50007012e-02 -2.01934770e-01 1.52956843e-01 -3.17660332e-01
4.64960895e-02 -4.59668487e-01 -5.41838229e-01 1.93164825e-01
-4.87758636e-01 -3.52997929e-02 6.86551988e-01 8.82936835e-01
8.07227969e-01 3.17556381e-01 4.27160114e-02 -3.41942847e-01
5.90178430e-01 -4.25088078e-01 -1.27473819e+00 1.47652879e-01
-1.45283878e-01 1.14815585e-01 3.60103488e-01 -5.93760312e-01
-1.02919149e+00 6.03602409e-01 7.09584430e-02 -3.65215361e-01
-3.06041092e-01 5.07605374e-01 2.01543689e-01 -7.45102644e-01
7.93256164e-01 2.48327509e-01 -2.30090648e-01 -2.68948704e-01
3.05916071e-01 3.08720559e-01 7.05453455e-01 -5.29898584e-01
7.72902071e-01 6.64269865e-01 -3.32519747e-02 -5.06421208e-01
-5.58627248e-01 -3.95395577e-01 -5.49602926e-01 -2.36643627e-01
6.39184415e-01 -1.00719595e+00 -9.84819353e-01 2.99642146e-01
-1.14528358e+00 -1.08395827e+00 -5.62825382e-01 8.44620705e-01
-1.04317510e+00 -1.21953242e-01 -3.87309879e-01 -1.02232027e+00
2.56909341e-01 -1.18270564e+00 9.14256573e-01 5.73164582e-01
8.31044167e-02 -1.15843797e+00 5.52918375e-01 -3.23862195e-01
5.69668531e-01 2.53474623e-01 3.38157535e-01 -2.63037086e-01
-1.12411499e+00 -1.75445244e-01 1.37768954e-01 -1.97194200e-02
3.31163555e-02 -5.83830297e-01 -9.52983260e-01 -4.26602781e-01
1.34428039e-01 -4.44209605e-01 8.26781631e-01 5.83423853e-01
7.99064755e-01 -1.03526346e-01 -4.27005023e-01 5.85205734e-01
1.76833153e+00 6.70452595e-01 3.62499297e-01 5.44537306e-01
1.68879658e-01 1.36737868e-01 6.09022021e-01 2.13009298e-01
2.31917098e-01 4.75246787e-01 1.04318237e+00 -3.13352048e-01
1.84010193e-02 -4.03415740e-01 4.92029667e-01 3.56532693e-01
-3.24509363e-03 -3.35170805e-01 -8.27924311e-01 4.79808897e-01
-2.24173760e+00 -7.37303555e-01 7.18216300e-01 2.26484346e+00
4.46370333e-01 1.93143487e-01 -3.25737685e-01 -5.08018494e-01
4.45568323e-01 1.19231299e-01 -6.65662348e-01 -7.15570748e-02
8.89580399e-02 8.34266543e-02 1.03626907e+00 9.50120509e-01
-1.15906894e+00 1.12509835e+00 6.95205498e+00 4.59523916e-01
-1.33999908e+00 3.74779046e-01 2.84116268e-01 -8.20977762e-02
2.35197008e-01 1.50646120e-01 -6.96055949e-01 2.06367180e-01
9.06064451e-01 3.71801674e-01 7.96287715e-01 1.14319587e+00
4.99617726e-01 -7.79639840e-01 -9.52974319e-01 7.99826503e-01
-1.63366094e-01 -1.35746694e+00 -7.89834917e-01 -3.16891796e-03
6.83255196e-01 5.38251936e-01 1.71938807e-01 4.61922795e-01
1.12811255e+00 -9.35004234e-01 1.25302947e+00 7.94330120e-01
1.71762258e-01 -5.71782649e-01 7.52505898e-01 7.75407791e-01
-8.47838938e-01 -2.29522362e-01 -7.02751637e-01 -3.32792342e-01
2.00179145e-01 -6.91183060e-02 -1.19718266e+00 -2.03695834e-01
7.86561430e-01 2.96672612e-01 -5.39465666e-01 1.55422819e+00
-5.54041564e-01 3.62549186e-01 -5.74159205e-01 -4.23239112e-01
6.11912012e-01 -3.08808312e-03 5.52163899e-01 8.03105116e-01
3.69258791e-01 -1.25945628e-01 4.85507965e-01 9.71581876e-01
2.77510583e-01 -5.16396821e-01 -6.94085181e-01 1.77547961e-01
4.85011250e-01 1.03367043e+00 -1.08895254e+00 -8.90159756e-02
-3.77373397e-02 1.07660794e+00 6.39386237e-01 8.17858815e-01
-7.81692088e-01 -4.32155669e-01 2.37266913e-01 -1.83552459e-01
4.99193370e-01 -7.66711950e-01 9.96487066e-02 -7.69397378e-01
-3.13412845e-01 -3.11429054e-01 -2.78833777e-01 -1.21728349e+00
-8.19987893e-01 6.36345744e-01 -2.20408421e-02 -8.54087949e-01
-3.19153398e-01 -7.26728857e-01 -5.00263333e-01 9.93658602e-01
-1.28998065e+00 -8.78960907e-01 -3.97942156e-01 4.65024948e-01
4.58615035e-01 -2.85346597e-01 1.00970304e+00 -3.02487999e-01
-1.30999312e-01 -2.48032883e-02 3.80394220e-01 1.86204582e-01
4.00250882e-01 -1.48586905e+00 4.34874982e-01 9.03529108e-01
3.36097687e-01 7.11099803e-01 7.57887423e-01 -7.39482760e-01
-1.35674810e+00 -9.64301884e-01 2.67381132e-01 -3.50363523e-01
5.13753831e-01 -6.31773174e-01 -3.51077616e-01 9.03746367e-01
3.24217498e-01 3.60502660e-01 1.10034071e-01 -1.13881260e-01
9.85411108e-02 -7.12024570e-02 -1.42512405e+00 6.99820518e-01
8.80373359e-01 -4.15663898e-01 -3.88478607e-01 4.27549422e-01
7.50515819e-01 -7.92884886e-01 -3.47334832e-01 6.87180758e-02
3.89223099e-01 -9.58972335e-01 8.36847305e-01 -2.94986039e-01
-2.08796203e-01 -6.51687980e-01 -4.03015286e-01 -1.71726465e+00
-4.67137188e-01 -5.52292705e-01 -1.98831096e-01 5.37070632e-01
3.87918830e-01 -7.31154859e-01 1.13466930e+00 2.61703283e-01
-1.30545303e-01 -6.95564449e-01 -1.10599768e+00 -6.52558088e-01
-5.37809908e-01 -4.04549807e-01 3.23864281e-01 3.50536644e-01
-3.38416249e-01 6.06090501e-02 -1.22608766e-01 7.99163997e-01
7.35033453e-01 -2.95388281e-01 6.70357585e-01 -7.45912015e-01
-6.98078096e-01 1.44127563e-01 -6.13588214e-01 -1.38848412e+00
1.76551297e-01 -1.78764135e-01 8.21978748e-01 -1.76150823e+00
-2.52278358e-01 -7.06562042e-01 -3.15712541e-01 6.87103450e-01
4.05651510e-01 1.25895843e-01 -1.62182730e-02 9.37669501e-02
-1.02509356e+00 3.52157116e-01 9.61820245e-01 -2.48757735e-01
-3.78865629e-01 4.62344848e-02 -2.18592495e-01 9.89231229e-01
6.97811127e-01 -6.02554023e-01 -6.20561540e-01 -6.82418823e-01
2.15435699e-01 1.35418266e-01 6.41863525e-01 -1.44682884e+00
7.31410623e-01 -2.30311975e-01 7.51211464e-01 -4.94709074e-01
9.47335124e-01 -9.93375003e-01 -7.99347311e-02 5.75683057e-01
-5.04648983e-01 2.54509896e-01 2.46258035e-01 8.75164032e-01
9.64804888e-02 -3.63927156e-01 5.95698357e-01 -5.86932898e-01
-1.38035417e+00 1.93681479e-01 -5.97590566e-01 -3.19577366e-01
1.01739359e+00 -3.74421775e-01 -2.64755130e-01 -6.86046720e-01
-1.06402135e+00 3.15248221e-02 6.04244709e-01 2.10261166e-01
8.12061250e-01 -1.17581451e+00 -1.95667893e-01 3.34937125e-01
-3.19466352e-01 1.86358452e-01 -1.56945109e-01 5.65672100e-01
-9.86668587e-01 2.98039794e-01 -2.84539431e-01 -5.98702073e-01
-7.89947033e-01 2.90712416e-01 8.20942402e-01 -1.03549965e-01
-3.10732394e-01 1.02785790e+00 4.41791862e-02 -6.58312261e-01
6.94568813e-01 -4.37963814e-01 1.08184293e-01 -6.99403942e-01
4.83655304e-01 2.09359393e-01 -8.88993219e-02 -3.59030217e-01
-1.73418060e-01 2.31768981e-01 2.45725915e-01 -7.07367122e-01
1.36836588e+00 -2.71996886e-01 1.55312791e-01 5.17307162e-01
7.36876726e-01 9.99557748e-02 -2.00928092e+00 -1.11173920e-01
-1.76820815e-01 -3.59719664e-01 5.00966966e-01 -1.11588967e+00
-7.90268600e-01 6.13790154e-01 1.17018557e+00 -5.71314506e-02
7.85784543e-01 -2.05202937e-01 1.47329524e-01 8.66789877e-01
1.11429453e+00 -1.06201410e+00 2.74635255e-01 7.89769053e-01
7.16732502e-01 -1.07476211e+00 -2.05425426e-01 1.79933146e-01
-2.66358346e-01 9.05153453e-01 7.13545620e-01 -5.10594249e-01
6.39225960e-01 6.92664146e-01 4.43079144e-01 -3.48358527e-02
-6.05653584e-01 -2.91400224e-01 -2.74734795e-01 9.18551624e-01
-2.91452229e-01 -1.08350106e-01 5.43709040e-01 1.04923338e-01
-1.92208365e-02 1.36706978e-02 3.76968622e-01 1.33507323e+00
-9.39926744e-01 -5.41159868e-01 -7.13225245e-01 -2.32005656e-01
-4.75548282e-02 2.14481905e-01 -2.73912668e-01 8.26975465e-01
4.82264668e-01 9.61456060e-01 2.14414805e-01 -2.92792797e-01
4.50920276e-02 -3.64307463e-01 6.86366796e-01 -7.25791216e-01
-2.67195225e-01 1.08084626e-01 -2.05482394e-01 -8.04687619e-01
-2.91712880e-01 -3.11200589e-01 -1.41789758e+00 -3.85779552e-02
-2.47553974e-01 3.39953482e-01 9.72479165e-01 8.96177471e-01
3.21191221e-01 5.57156384e-01 2.98345119e-01 -1.25881839e+00
-5.47024608e-01 -8.15060437e-01 -6.34478152e-01 -3.33663553e-01
5.46296120e-01 -7.79299736e-01 -2.44774476e-01 -2.09923923e-01] | [4.6164231300354, 0.7386215925216675] |
a369a9b0-22e4-4b59-8dd9-3b69527225f8 | comprehensive-evaluation-of-deep-and-graph | 2306.05257 | null | https://arxiv.org/abs/2306.05257v1 | https://arxiv.org/pdf/2306.05257v1.pdf | Comprehensive evaluation of deep and graph learning on drug-drug interactions prediction | Recent advances and achievements of artificial intelligence (AI) as well as deep and graph learning models have established their usefulness in biomedical applications, especially in drug-drug interactions (DDIs). DDIs refer to a change in the effect of one drug to the presence of another drug in the human body, which plays an essential role in drug discovery and clinical research. DDIs prediction through traditional clinical trials and experiments is an expensive and time-consuming process. To correctly apply the advanced AI and deep learning, the developer and user meet various challenges such as the availability and encoding of data resources, and the design of computational methods. This review summarizes chemical structure based, network based, NLP based and hybrid methods, providing an updated and accessible guide to the broad researchers and development community with different domain knowledge. We introduce widely-used molecular representation and describe the theoretical frameworks of graph neural network models for representing molecular structures. We present the advantages and disadvantages of deep and graph learning methods by performing comparative experiments. We discuss the potential technical challenges and highlight future directions of deep and graph learning models for accelerating DDIs prediction. | ['Xiangxiang Zeng', 'Philip S. Yu', 'Bosheng Song', 'Haowen Chen', 'Li Zeng', 'Dong-Sheng Cao', 'Jian-Yu Shi', 'Wen Zhang', 'Zu-Guo Yu', 'Yafang Zhou', 'Lichang Dai', 'Xuan Lin'] | 2023-06-08 | null | null | null | null | ['drug-discovery'] | ['medical'] | [ 2.41264239e-01 -1.94733769e-01 -6.02495790e-01 -6.16092607e-02
5.13392352e-02 -4.89375204e-01 3.21651191e-01 8.36543322e-01
1.73085555e-02 1.04703927e+00 -9.67375860e-02 -7.18682826e-01
-3.99286330e-01 -8.12479198e-01 -5.66746652e-01 -9.36908185e-01
-6.68846667e-01 8.94814134e-01 -2.18255281e-01 -2.96741784e-01
1.82722837e-01 1.08950305e+00 -7.73611605e-01 3.34914982e-01
9.73605871e-01 7.37295091e-01 -3.54953296e-02 4.45446193e-01
-2.69406646e-01 7.90978789e-01 -7.09653974e-01 -3.79467666e-01
-2.05531806e-01 -5.96964240e-01 -7.37894297e-01 -5.16764283e-01
1.87131524e-01 8.88716057e-02 -6.10901177e-01 9.19660270e-01
9.55424249e-01 5.00674471e-02 8.69443893e-01 -1.00851738e+00
-1.02192569e+00 1.93228200e-01 -3.29415709e-01 3.25259089e-01
6.39187157e-01 1.12899803e-01 6.63490653e-01 -7.05585361e-01
8.28968227e-01 1.05956745e+00 7.45174766e-01 6.69733286e-01
-8.40744078e-01 -5.64632595e-01 -8.14733282e-02 6.10187709e-01
-1.38686681e+00 -7.06153139e-02 5.82414210e-01 -5.74464798e-01
1.53711569e+00 9.44929346e-02 7.32820094e-01 9.57580566e-01
8.84562433e-01 4.17254418e-01 3.98606658e-01 -2.58102864e-01
2.91584939e-01 -2.74433225e-01 1.38759464e-01 7.37132311e-01
5.14403343e-01 1.69130519e-01 -3.55404615e-01 -5.12627184e-01
5.64497054e-01 3.21281701e-01 -2.21123159e-01 -3.14026624e-01
-8.51718605e-01 1.00991237e+00 7.64654279e-01 4.56149161e-01
-6.27236187e-01 2.58073676e-02 3.94534767e-01 1.38655156e-01
3.29418629e-01 5.75266957e-01 -4.17453408e-01 3.00764292e-01
-5.70972741e-01 2.52676427e-01 7.53750503e-01 6.13628328e-01
1.25908956e-01 2.31342107e-01 4.40625027e-02 6.56346083e-01
3.36466134e-01 2.06261039e-01 2.39187926e-01 -1.21636614e-01
1.42306024e-02 7.12749302e-01 -3.27348173e-01 -1.23237181e+00
-7.95031846e-01 -5.95268786e-01 -1.26810229e+00 1.51306745e-02
4.39230092e-02 -4.76463921e-02 -1.07887304e+00 1.17728186e+00
4.57300276e-01 2.73311347e-01 2.28868145e-02 5.51652014e-01
1.43044543e+00 7.30303764e-01 5.85675061e-01 -4.18363750e-01
1.13140142e+00 -6.52274907e-01 -1.02993858e+00 1.98110774e-01
8.87959301e-01 -5.85766912e-01 9.33980197e-02 3.52137893e-01
-9.76879895e-01 -2.58624673e-01 -1.15069115e+00 -1.01016998e-01
-9.24594522e-01 -2.51079530e-01 1.03148890e+00 4.66348588e-01
-7.36687481e-01 1.12542641e+00 -7.17774808e-01 -2.89289862e-01
8.04793537e-01 1.03677285e+00 -4.80982155e-01 -2.08348185e-01
-1.53026295e+00 1.19899178e+00 2.88922518e-01 9.54191983e-02
-7.46758938e-01 -9.25497115e-01 -9.15070415e-01 -1.30079180e-01
-8.78525302e-02 -8.87411416e-01 5.58765769e-01 -4.19549823e-01
-1.27883542e+00 6.26137257e-01 1.11144677e-01 -6.10338926e-01
9.76833999e-02 2.12807223e-01 -3.78193766e-01 6.43827254e-03
-3.28394681e-01 1.97250813e-01 5.98899983e-02 -5.09160876e-01
-4.73259762e-02 -7.87853777e-01 -2.41342977e-01 1.22354373e-01
1.42462343e-01 -1.76397368e-01 -7.95708597e-02 -4.06063467e-01
-2.49341920e-01 -8.18912685e-01 -4.68625873e-01 4.14588191e-02
-3.92068058e-01 -3.57602000e-01 8.70331645e-01 -6.31260216e-01
1.19235551e+00 -1.71470797e+00 4.14002955e-01 2.40467757e-01
8.75675321e-01 9.47018087e-01 -2.53474623e-01 8.92024457e-01
-5.64727664e-01 3.14707845e-01 3.64073776e-02 5.85905612e-01
-5.62081337e-01 -6.28063362e-03 2.19492838e-01 6.58312440e-01
-2.57912674e-03 1.21329427e+00 -1.07503366e+00 -2.02663288e-01
3.80680323e-01 8.67641032e-01 -3.53833050e-01 2.13754132e-01
-3.56397718e-01 7.33971536e-01 -6.65882289e-01 7.66527474e-01
7.09657669e-01 -6.10473096e-01 6.59754038e-01 -3.65300089e-01
2.13103145e-01 3.31257463e-01 -4.96300995e-01 1.37360966e+00
8.64117146e-02 3.78823429e-01 -4.70231712e-01 -1.42068696e+00
7.38160551e-01 5.52319765e-01 6.95045292e-01 -8.43786716e-01
3.18428650e-02 2.89774239e-01 4.80042607e-01 -5.67593694e-01
-4.26006049e-01 -1.25851914e-01 5.32695591e-01 1.12404808e-01
8.58256891e-02 7.11404979e-02 2.06983134e-01 8.75573680e-02
1.04668534e+00 -1.69652160e-02 7.59588301e-01 -5.31141348e-02
5.77603579e-01 -1.39306942e-02 2.96732813e-01 3.46531510e-01
-1.91893816e-01 -1.47003263e-01 5.87406218e-01 -1.03223133e+00
-9.26168621e-01 -5.03607035e-01 -2.74169832e-01 6.93250895e-01
-8.19062442e-02 -3.80957693e-01 -6.27003968e-01 -5.78873694e-01
1.79617077e-01 1.90463468e-01 -6.87968016e-01 -4.11725283e-01
-4.36072320e-01 -1.19457674e+00 4.36418563e-01 1.88864991e-01
1.80951834e-01 -9.47256744e-01 2.90433466e-01 5.25993407e-01
3.25020313e-01 -6.83464944e-01 -6.19692840e-02 2.84193307e-01
-8.77029121e-01 -1.49310601e+00 -7.65159786e-01 -9.79580700e-01
3.84919375e-01 -3.35490517e-02 9.51261163e-01 4.02818292e-01
-8.28008413e-01 -3.16896051e-01 -8.22598413e-02 -6.90268278e-01
-5.78142762e-01 -7.21152648e-02 5.64473905e-02 -5.97847581e-01
5.23825228e-01 -6.02414191e-01 -9.13723588e-01 -8.62344950e-02
-7.03284025e-01 -1.56380802e-01 4.12928253e-01 9.26151335e-01
6.28065646e-01 -3.28279845e-02 6.50837243e-01 -1.16865253e+00
8.57298374e-01 -6.40932560e-01 -4.53033596e-01 4.02142972e-01
-8.05930316e-01 2.10961282e-01 4.99036014e-01 -1.89714655e-01
-4.07117635e-01 6.65501878e-02 -4.95063543e-01 -2.30084024e-02
-1.13044064e-02 1.06034541e+00 -1.56168997e-01 -6.07849896e-01
8.06686997e-01 7.18052536e-02 1.43117830e-01 -3.28944176e-01
1.82002053e-01 5.86613119e-01 -1.85097083e-01 -1.60582870e-01
8.31902325e-02 8.38017017e-02 5.78998148e-01 -1.05170965e+00
-4.85879809e-01 -1.84850648e-01 -5.59955657e-01 1.32096142e-01
9.23309743e-01 -4.60323751e-01 -1.09812498e+00 4.36348170e-01
-1.36880326e+00 -6.01412095e-02 3.66960615e-01 4.25903767e-01
-1.96381748e-01 6.56404674e-01 -8.08148146e-01 -1.97638556e-01
-8.51194620e-01 -1.45163202e+00 6.40172303e-01 2.36489028e-02
-2.47985438e-01 -1.56573844e+00 4.12020236e-01 2.48025402e-01
2.86546469e-01 7.08345175e-01 1.67049289e+00 -1.03664100e+00
-1.99029148e-01 -4.00046408e-01 -1.48741439e-01 -6.50131851e-02
3.99010599e-01 1.10734820e-01 -5.77479601e-01 -2.03159735e-01
-5.05940259e-01 -6.66293204e-02 5.15008986e-01 7.11557448e-01
1.11262357e+00 -3.14613551e-01 -8.90825450e-01 6.04097545e-01
1.47829235e+00 1.05332077e+00 7.31566787e-01 -6.24555126e-02
1.00491238e+00 3.72570127e-01 4.46575284e-02 1.86035946e-01
-9.90197584e-02 6.99058175e-01 3.73115897e-01 -4.76595312e-01
-1.81596294e-01 -1.01892150e-03 -2.56701797e-01 5.51535666e-01
-6.00884974e-01 -5.95947564e-01 -1.11507750e+00 -5.73954917e-02
-1.73702180e+00 -8.46082330e-01 -2.19054133e-01 2.14461350e+00
9.77051854e-01 -3.27920794e-01 8.17376375e-02 -1.95314288e-01
6.82459950e-01 -2.32545182e-01 -9.26401615e-01 -5.79099357e-01
-6.83327094e-02 7.42307365e-01 4.13070172e-01 4.78969812e-01
-9.90785599e-01 9.62680995e-01 7.20918798e+00 8.33957016e-01
-1.38740575e+00 -2.78190017e-01 8.11416268e-01 2.51798898e-01
-8.69959742e-02 -5.62755167e-01 -4.55968976e-01 3.58223438e-01
1.36043215e+00 -3.40467066e-01 3.51328462e-01 5.85263371e-01
3.27502012e-01 5.17549038e-01 -1.35795522e+00 1.15685344e+00
-2.13150680e-01 -2.20012474e+00 5.04809380e-01 3.78463209e-01
5.18556654e-01 1.59720287e-01 1.77648645e-02 -2.30215743e-01
9.74176079e-02 -1.53040326e+00 -3.12912583e-01 2.94033587e-01
9.19358194e-01 -7.89763212e-01 8.84379923e-01 -1.43346936e-01
-1.10327613e+00 3.21302295e-01 -4.32328254e-01 -4.32126597e-02
-2.96230078e-01 4.33028966e-01 -8.82823110e-01 6.44929945e-01
1.28722340e-01 1.25237608e+00 -1.07107207e-01 1.14826941e+00
-6.75284788e-02 1.31672278e-01 1.39322206e-01 -3.30486923e-01
2.18667433e-01 -3.86756688e-01 1.26974314e-01 1.20184577e+00
-5.82326502e-02 3.21973085e-01 2.18127757e-01 7.87816226e-01
-1.59501374e-01 2.15749830e-01 -8.62814128e-01 -7.69667625e-01
3.55865717e-01 8.30544889e-01 -4.64785546e-01 -3.15361530e-01
-3.92809004e-01 8.34529281e-01 1.85129911e-01 4.29286897e-01
-6.81132138e-01 -6.02901042e-01 8.71071279e-01 1.11312337e-01
-1.39499024e-01 -5.86903989e-02 8.43683705e-02 -7.52599657e-01
-6.14107728e-01 -1.08763826e+00 5.80869377e-01 -3.95094991e-01
-1.23762751e+00 6.61648571e-01 -2.10939810e-01 -9.46869791e-01
-1.42841801e-01 -1.13386500e+00 -4.08452451e-01 8.32235277e-01
-1.37699640e+00 -9.37944174e-01 9.79323238e-02 4.37452912e-01
2.26648524e-01 -3.50007504e-01 1.20450211e+00 5.79753816e-01
-7.34900594e-01 4.76917177e-01 4.39899355e-01 -2.33125947e-02
3.88674617e-01 -9.07822728e-01 4.86010998e-01 -2.66600475e-02
-1.13990404e-01 8.58327568e-01 7.03010857e-01 -8.95767331e-01
-1.60192370e+00 -1.05870044e+00 8.90969932e-01 -1.55564323e-01
5.69426000e-01 -7.15441704e-02 -9.65352952e-01 4.12201285e-01
1.25542566e-01 -3.38826366e-02 1.25837338e+00 -6.91848621e-02
-5.49768656e-02 2.40107477e-01 -1.20077169e+00 4.89832550e-01
9.48135555e-01 -3.60364318e-01 -1.79259956e-01 1.15511787e+00
4.37182516e-01 -4.76313233e-01 -1.09517598e+00 4.13342476e-01
5.06436467e-01 -5.77129602e-01 1.28565633e+00 -1.32996225e+00
2.31974557e-01 -1.33741304e-01 4.39476013e-01 -1.22008491e+00
-6.69020355e-01 -5.00668228e-01 -2.92125314e-01 6.71132430e-02
5.30674100e-01 -6.93601727e-01 8.10525835e-01 5.14796853e-01
-1.43006369e-01 -1.20543551e+00 -7.46125340e-01 -5.65081835e-01
3.47952992e-01 1.28713697e-01 5.69656074e-01 1.04519260e+00
4.61487979e-01 6.33202612e-01 -1.91765249e-01 2.18246169e-02
4.24001545e-01 -1.72949240e-01 2.77321398e-01 -1.44636440e+00
-7.50047117e-02 -4.99551594e-01 -1.03636634e+00 -4.66415167e-01
5.90665825e-03 -1.06586969e+00 -7.44101644e-01 -1.92431402e+00
3.82603496e-01 -7.27178296e-03 -4.35195088e-01 5.30644894e-01
4.09038104e-02 5.30937500e-02 -2.99911559e-01 2.55958945e-01
-4.21502978e-01 2.54131556e-01 1.35413766e+00 -5.80092788e-01
-2.80197173e-01 -2.31522992e-01 -4.98105735e-01 3.40842515e-01
7.19989955e-01 -3.45187157e-01 -3.77120376e-01 -2.39039734e-02
5.54399431e-01 8.05391967e-02 -1.09050766e-01 -5.43540359e-01
6.47648945e-02 -2.60462165e-01 7.44770110e-01 -3.14444751e-01
4.72590737e-02 -6.58949316e-01 3.22722346e-01 1.16688251e+00
-1.57459676e-01 1.60889216e-02 4.52382773e-01 6.07300341e-01
-2.03786522e-01 9.67291892e-02 9.29161489e-01 -3.37857336e-01
-4.57397610e-01 1.03515553e+00 -4.87385064e-01 -4.15892243e-01
1.17593110e+00 -3.15360427e-01 -3.44857097e-01 -2.93892980e-01
-1.00968790e+00 6.77683353e-02 -7.57099763e-02 1.66774929e-01
8.95508528e-01 -1.20252085e+00 -5.61532438e-01 1.69303030e-01
1.67901978e-01 -3.79108161e-01 1.92471758e-01 7.36824095e-01
-1.17265666e+00 9.03104305e-01 -3.45645934e-01 -2.06640676e-01
-1.42653346e+00 9.37412024e-01 7.65535414e-01 -3.08806509e-01
-2.87766308e-01 7.59877920e-01 3.88746232e-01 -1.50876641e-01
2.71416128e-01 1.87123604e-02 -5.83448589e-01 -2.19120488e-01
5.47578454e-01 3.58226597e-01 5.06517172e-01 -3.38364005e-01
-7.14754939e-01 5.58231711e-01 -4.85203683e-01 9.34110045e-01
1.52026260e+00 4.58337039e-01 -4.25950676e-01 -1.89092711e-01
1.36675525e+00 -6.45412803e-01 -3.41151714e-01 5.72362170e-02
-9.02122483e-02 1.62917197e-01 2.54109502e-01 -1.08216524e+00
-1.01881647e+00 1.11709142e+00 8.91609371e-01 -6.11424493e-03
6.04971349e-01 6.41315505e-02 8.82766426e-01 6.23909593e-01
1.52016401e-01 -7.78608501e-01 -3.06853782e-02 2.66608745e-01
1.03093791e+00 -1.08053088e+00 3.93344283e-01 -5.53387344e-01
-2.69311488e-01 1.37182152e+00 1.24301381e-01 6.20855652e-02
8.63236725e-01 -1.12258427e-01 -9.38059017e-03 -9.14680898e-01
-4.72953200e-01 1.46506235e-01 5.16002238e-01 8.83127689e-01
1.23759043e+00 4.91015576e-02 -6.05541110e-01 9.29737762e-02
4.44882154e-01 1.06720492e-01 -7.03431889e-02 9.95602012e-01
-1.55091047e-01 -1.55433440e+00 8.21904019e-02 3.94995511e-01
-5.65191031e-01 -4.02268201e-01 -9.94910598e-01 7.25342631e-01
5.68779185e-02 7.76343286e-01 -3.21008831e-01 -2.53233671e-01
1.61162302e-01 -3.73679139e-02 7.17741668e-01 -7.00712264e-01
-3.74466449e-01 6.58390298e-02 -2.00307040e-04 -4.30470377e-01
-3.85189444e-01 3.59978061e-03 -1.58771956e+00 -5.17301798e-01
-4.19679314e-01 3.38299751e-01 8.20287645e-01 8.96629632e-01
9.25774395e-01 7.11792350e-01 3.30670387e-01 -5.56577802e-01
2.55585667e-02 -7.15814114e-01 -5.87352812e-01 1.20299570e-01
3.86217207e-01 -6.94568217e-01 1.02442969e-02 -1.03509665e-01] | [5.2273077964782715, 5.848116397857666] |
c211e84b-d4cd-4bf4-b1c2-98e097382346 | pic-xai-post-hoc-image-captioning-explanation | null | null | https://ieeexplore.ieee.org/abstract/document/10158563 | https://ieeexplore.ieee.org/abstract/document/10158563 | PIC-XAI: Post-hoc Image Captioning Explanation using Segmentation | The rapid advancement in Deep Learning (DL) proposes viable solutions to various real-world problems. However, deploying DL-based models in some applications is hindered by their black-box nature and the inability to explain them. This has pushed Explainable Artificial Intelligence (XAI) research toward DL-based models, aiming to increase the trust by reducing their opacity. Although many XAI algorithms were proposed earlier, they lack the ability to explain certain tasks, i.e. image captioning (IC). This is caused by the IC task nature, e.g. the presence of multiple objects from the same category in the captioned image. In this paper we propose and investigate an XAI approach for this particular task. Additionally, we provide a method to evaluate XAI algorithms performance in the domain. | ['Gábor Szűcs', 'Modafar Al-Shouha'] | 2023-05-23 | null | null | null | ieee-17th-international-symposium-on-applied | ['explainable-artificial-intelligence', 'image-captioning'] | ['computer-vision', 'computer-vision'] | [ 2.20702752e-01 6.43855929e-01 -2.28879988e-01 -5.41363895e-01
-7.26298094e-02 -3.19950461e-01 8.15896213e-01 -1.25480726e-01
5.21174341e-04 9.94333804e-01 1.80640996e-01 -3.73227149e-01
-3.17414582e-01 -4.00895327e-01 -1.00958669e+00 -4.71198112e-01
8.86545852e-02 7.21256435e-01 -1.59355640e-01 1.24959797e-01
2.54270881e-01 4.74550784e-01 -1.57506883e+00 7.28753209e-01
8.21015954e-01 7.23793447e-01 2.82567531e-01 2.37130731e-01
-3.54405791e-01 1.06051397e+00 -7.71752238e-01 -4.44131464e-01
-1.24112144e-01 -4.72406417e-01 -8.25984538e-01 -8.38101748e-03
5.58239818e-01 -2.41868958e-01 -9.76380557e-02 9.94880795e-01
-1.28301948e-01 -2.15122059e-01 6.50268674e-01 -2.01475716e+00
-1.01723361e+00 7.98331797e-01 -5.00211179e-01 -4.82415296e-02
2.18993589e-01 -2.30199634e-03 1.07831407e+00 -7.01738000e-01
6.51647091e-01 1.55991602e+00 4.16948617e-01 9.58916962e-01
-1.17277884e+00 -5.55818677e-01 3.91163409e-01 5.95771253e-01
-1.01632977e+00 7.93748349e-02 9.15672541e-01 -4.77127954e-02
8.33304107e-01 4.62069601e-01 4.07129645e-01 1.38939536e+00
5.70264198e-02 1.21825039e+00 1.45550478e+00 -6.67151034e-01
3.31648797e-01 5.37277997e-01 4.04218763e-01 3.94350141e-01
6.34515643e-01 5.59882298e-02 -5.66515088e-01 2.24386781e-01
4.52311397e-01 -1.78371936e-01 -1.13860376e-01 -3.48970115e-01
-1.27455223e+00 9.33654249e-01 7.55243421e-01 4.40323204e-01
-3.95090699e-01 2.65534848e-01 1.06018953e-01 1.11984022e-01
2.98692375e-01 9.14045036e-01 -5.98626614e-01 8.58846828e-02
-8.59191895e-01 3.00218582e-01 6.71856701e-01 7.73868322e-01
8.09022486e-01 1.42993927e-01 -1.93520188e-02 2.43388802e-01
4.67721194e-01 1.43108457e-01 3.04213554e-01 -9.42009568e-01
3.32810163e-01 7.54282236e-01 -1.60024129e-02 -8.92891645e-01
-5.36945224e-01 -6.58213735e-01 -8.10095966e-01 7.24246144e-01
4.61778224e-01 1.94072977e-01 -1.16308081e+00 1.87752318e+00
7.17157274e-02 2.18274012e-01 1.40522793e-01 1.15808666e+00
7.72315919e-01 9.47925806e-01 2.62663245e-01 3.37517634e-02
1.24947405e+00 -1.26555347e+00 -9.15603638e-01 -7.11166561e-01
6.21687829e-01 -4.41152811e-01 1.33763349e+00 5.21663368e-01
-9.06408906e-01 -6.67253554e-01 -1.07560849e+00 -1.63204074e-01
-4.62302089e-01 1.32312104e-01 7.97422290e-01 6.11160994e-01
-1.00995564e+00 2.54256338e-01 -3.93072873e-01 -2.54369050e-01
6.13463402e-01 4.70485806e-01 -3.79678309e-01 -1.57042757e-01
-1.02629554e+00 1.18673193e+00 4.61413860e-01 8.78252238e-02
-8.92100751e-01 -5.46321690e-01 -6.03053153e-01 4.30966884e-01
3.76002073e-01 -7.35106528e-01 1.03893113e+00 -1.55872655e+00
-8.71832013e-01 8.59869778e-01 -9.54201892e-02 -7.53042221e-01
5.81675053e-01 -5.51755309e-01 -1.30869016e-01 -4.97980043e-02
3.05033773e-02 1.20778489e+00 9.22199070e-01 -2.09144258e+00
-4.38691288e-01 -3.94481033e-01 6.28015518e-01 7.98204616e-02
-2.35135630e-01 -1.59726664e-01 -5.97047396e-02 -4.55532849e-01
1.37549147e-01 -1.10841513e+00 8.32215697e-02 1.54581577e-01
-5.58314681e-01 -2.38378838e-01 1.33260405e+00 -4.87806737e-01
8.95447910e-01 -2.20243096e+00 6.28759861e-02 -6.50338382e-02
4.93927002e-01 5.41447282e-01 -3.19347262e-01 1.81756839e-01
-3.73522848e-01 5.27899206e-01 -3.25371474e-02 -6.05117023e-01
1.40444458e-01 7.29613662e-01 -3.60776395e-01 5.96185699e-02
3.64851236e-01 9.96691704e-01 -8.25565696e-01 -6.13041341e-01
3.18163961e-01 6.20767295e-01 -4.48883206e-01 2.70755619e-01
-7.31228948e-01 4.39264476e-01 -2.71690518e-01 4.00418818e-01
6.67663038e-01 -3.07994872e-01 -9.36235860e-02 -2.19245166e-01
2.55360734e-04 3.71095799e-02 -7.98282862e-01 1.55180502e+00
-3.85806024e-01 1.02550983e+00 -2.11551711e-01 -8.77584577e-01
8.26005757e-01 3.45218390e-01 1.67790711e-01 -7.15930343e-01
-5.03487252e-02 1.49033546e-01 3.32073748e-01 -4.11648840e-01
3.61174881e-01 -4.54534926e-02 3.85877192e-01 5.08927941e-01
-2.61756867e-01 5.25524616e-02 2.59139892e-02 3.60947788e-01
6.50192797e-01 3.29301149e-01 2.03287438e-01 -9.04830024e-02
4.60666716e-01 2.80195653e-01 3.28255475e-01 1.01346636e+00
-2.12448910e-01 8.65584135e-01 6.28403664e-01 -9.39866066e-01
-1.20586765e+00 -7.32675195e-01 9.77474451e-02 7.54428208e-01
3.55399221e-01 -1.74343213e-01 -9.10356104e-01 -7.73846269e-01
-2.51546562e-01 1.29394770e+00 -1.03015363e+00 -3.49224478e-01
-4.56939548e-01 -4.91674066e-01 1.57277927e-01 3.41548204e-01
5.72003663e-01 -1.41690624e+00 -9.38512802e-01 7.65789077e-02
-4.09434825e-01 -9.01554644e-01 2.81026632e-01 2.36187056e-01
-8.57362151e-01 -7.97079206e-01 -4.04103190e-01 -4.87895668e-01
7.11919725e-01 3.68711859e-01 1.38471138e+00 4.10166502e-01
-1.05472475e-01 2.86778748e-01 -5.67795813e-01 -9.19723630e-01
-6.80357695e-01 1.85534403e-01 -1.34577736e-01 -4.00447659e-02
5.26329160e-01 -1.57589972e-01 -4.80767876e-01 1.39870539e-01
-1.22620285e+00 5.84295332e-01 7.60783792e-01 7.31618524e-01
2.15178683e-01 2.07325574e-02 4.02263105e-01 -9.25683379e-01
7.05311239e-01 -2.84338713e-01 -1.60771951e-01 4.46052283e-01
-8.27158034e-01 3.64339054e-01 3.64603877e-01 -4.65658575e-01
-1.18406785e+00 -1.46417052e-01 1.91867530e-01 -3.23134601e-01
-4.23344791e-01 4.50579375e-01 -2.83467680e-01 1.74814880e-01
6.11874223e-01 -1.52404249e-01 -9.82678123e-03 -2.12742150e-01
3.91693383e-01 3.22798699e-01 2.50227213e-01 -3.44711483e-01
8.03483069e-01 3.41129661e-01 5.85429333e-02 -4.10403401e-01
-1.31514752e+00 2.09856421e-01 -4.13025081e-01 -4.68744546e-01
1.02234983e+00 -5.14011443e-01 -6.07448161e-01 -6.57868981e-02
-1.64642847e+00 -6.78704232e-02 -2.06641063e-01 4.09702033e-01
-5.27514875e-01 1.37087941e-01 -1.58086970e-01 -8.56537580e-01
-7.74305090e-02 -1.24894679e+00 8.46989274e-01 3.15000474e-01
-4.73076284e-01 -8.72531235e-01 -1.90378055e-01 6.71963751e-01
4.65217263e-01 2.73266912e-01 1.37827301e+00 -6.49212837e-01
-8.04066241e-01 -3.84566709e-02 -6.42078638e-01 4.20392573e-01
-3.34724002e-02 -1.39442906e-02 -1.21081030e+00 -9.83011425e-02
1.47391960e-01 -3.05725664e-01 5.60471892e-01 4.68882471e-01
1.18659008e+00 -4.40378189e-01 -3.10255140e-01 3.15466672e-01
1.39621043e+00 2.05099583e-01 8.97575796e-01 9.31740582e-01
7.68908799e-01 7.77091682e-01 7.95408487e-01 1.80960655e-01
6.31029904e-02 7.03261197e-01 1.26231778e+00 -6.69514954e-01
-5.14441371e-01 -1.73226252e-01 7.29595423e-02 4.44026813e-02
-1.58018529e-01 -5.25805175e-01 -1.11764467e+00 2.98735887e-01
-2.10595417e+00 -8.94189835e-01 -5.12227595e-01 1.68498755e+00
4.65438247e-01 3.11724424e-01 -3.89683664e-01 2.91211367e-01
5.30538857e-01 9.82644595e-03 -4.68596995e-01 -6.26061022e-01
-4.20700252e-01 -1.89668208e-01 1.50947683e-02 5.20603359e-01
-8.96302402e-01 8.79487574e-01 6.28780413e+00 3.73589158e-01
-9.98218358e-01 1.20437600e-01 9.39860880e-01 2.50078738e-01
-5.23347616e-01 6.76754862e-02 -4.71029907e-01 3.90362561e-01
6.42546058e-01 -7.54701421e-02 2.72706270e-01 1.12093949e+00
2.82290846e-01 -8.99811089e-02 -1.35136688e+00 9.55539525e-01
2.73183167e-01 -1.35984218e+00 5.19002259e-01 6.99051172e-02
6.34168863e-01 -3.92120779e-01 4.26767051e-01 1.08648427e-01
-2.08885193e-01 -1.26980972e+00 9.28216457e-01 3.35359246e-01
4.32983160e-01 -5.72503448e-01 9.50119138e-01 3.04014623e-01
-3.28212857e-01 -2.12076873e-01 -4.03056651e-01 -4.15424973e-01
-1.50090531e-01 3.90433609e-01 -1.05918896e+00 2.43295312e-01
6.98947847e-01 2.53116757e-01 -8.95281613e-01 9.94763553e-01
-5.17709136e-01 7.50492513e-01 4.27572522e-03 -5.85966744e-02
7.00623691e-01 -2.07986981e-01 3.69076341e-01 9.15076494e-01
1.93582624e-01 -1.50146872e-01 -2.91991591e-01 1.22235286e+00
1.13871060e-01 -3.57360542e-01 -7.15186179e-01 1.58910125e-01
1.64944574e-01 1.04484665e+00 -6.34112716e-01 -2.98136175e-01
-5.56129038e-01 1.04317307e+00 2.68571973e-01 4.51686084e-01
-1.22999096e+00 1.81319937e-01 5.46577513e-01 2.04575837e-01
-1.82076558e-01 -1.27547249e-01 -7.47286737e-01 -9.44912076e-01
-8.04204792e-02 -1.19967830e+00 3.64462957e-02 -1.50320148e+00
-1.01409936e+00 8.95470083e-01 1.61288977e-01 -1.02557814e+00
-2.18686000e-01 -8.36979806e-01 -3.48434597e-01 5.49554825e-01
-1.78375483e+00 -1.56860793e+00 -6.24514759e-01 4.21706975e-01
6.16237700e-01 -1.18961498e-01 7.36423612e-01 1.12859227e-01
-3.28253180e-01 1.67922273e-01 -1.05168752e-01 -2.98378766e-01
5.88046074e-01 -1.29497695e+00 2.64808387e-01 8.12115490e-01
5.96560597e-01 4.47675198e-01 1.32174993e+00 -4.26641881e-01
-1.08935809e+00 -7.49598384e-01 1.11169982e+00 -6.70108080e-01
3.57984275e-01 -3.27117950e-01 -1.19670808e+00 8.37266982e-01
4.95111972e-01 -3.43600452e-01 4.29571062e-01 8.89453217e-02
-4.30008262e-01 -2.17907578e-02 -1.06111288e+00 6.88036859e-01
8.28814805e-01 -2.06185728e-01 -8.14437389e-01 5.31376123e-01
7.13064969e-01 -4.09568064e-02 -4.71610017e-03 2.96049953e-01
2.22691387e-01 -1.20368898e+00 9.40990448e-01 -7.34048963e-01
7.95028389e-01 -3.78861785e-01 9.85877961e-02 -1.21437824e+00
-4.71246898e-01 -2.53227293e-01 -1.83931530e-01 1.09237444e+00
6.39054894e-01 -3.64746600e-01 1.05049574e+00 1.10327744e+00
-2.13088021e-01 -3.46624792e-01 -9.01110828e-01 -7.30865836e-01
-2.00534150e-01 -4.33162570e-01 5.91273904e-01 9.50367510e-01
-3.18206638e-01 3.02760661e-01 -6.49574697e-01 4.25736725e-01
6.68548584e-01 4.98845875e-02 7.06328988e-01 -1.34846056e+00
3.38656968e-03 -1.53762937e-01 -4.27000016e-01 -4.74891603e-01
3.35420817e-01 -5.12951016e-01 -1.59449384e-01 -1.88321567e+00
3.67108107e-01 -3.84740233e-01 -2.60299116e-01 6.83235288e-01
-2.19240874e-01 2.81520009e-01 5.80230713e-01 4.16814625e-01
-4.42457348e-01 3.41509819e-01 1.05595541e+00 -3.95421714e-01
1.24823786e-01 -3.11104238e-01 -6.07369184e-01 7.29547322e-01
1.11832929e+00 -7.15241432e-01 -6.83874786e-01 -8.69165480e-01
4.10533279e-01 -3.73036683e-01 5.64040482e-01 -1.38595152e+00
-5.62227145e-02 -1.62863195e-01 3.40107977e-01 -3.71688366e-01
5.19475877e-01 -1.19767070e+00 3.28777343e-01 5.93422532e-01
-5.48330426e-01 1.28047794e-01 2.86888152e-01 3.17531645e-01
-4.78958070e-01 -5.94847858e-01 5.97810507e-01 -1.84550360e-01
-8.29336286e-01 -2.10728757e-02 -1.29871249e-01 -4.39618617e-01
1.24622715e+00 -4.68215853e-01 -5.70174754e-01 -5.33233702e-01
-7.27509379e-01 -9.71425399e-02 4.92394179e-01 6.30270898e-01
7.38721907e-01 -1.03604269e+00 -7.23764122e-01 -1.03769027e-01
3.06712598e-01 -1.81670547e-01 3.86142284e-02 3.64216745e-01
-6.78344727e-01 5.68026483e-01 -6.63331091e-01 -4.36596900e-01
-1.47303259e+00 8.72438550e-01 2.73112446e-01 -2.29482174e-01
-5.44983387e-01 6.95784986e-01 6.99342847e-01 2.33562533e-02
4.66529936e-01 -2.58035392e-01 -2.70904809e-01 -3.63832086e-01
4.51076806e-01 -1.57904420e-02 -2.71160603e-01 -3.07296157e-01
-2.78295696e-01 2.11564735e-01 -2.19745800e-01 3.78564298e-02
1.31110716e+00 -1.11248314e-01 -1.78999364e-01 4.04750377e-01
8.39775860e-01 -5.93681991e-01 -1.30650961e+00 1.41893923e-01
2.23534122e-01 -6.18574917e-01 1.06776036e-01 -1.23312962e+00
-8.65394831e-01 1.06295955e+00 6.29042447e-01 3.61304462e-01
9.39656198e-01 2.27616876e-01 2.75380462e-01 3.96260262e-01
2.00125486e-01 -9.96533930e-01 2.11651087e-01 1.46340638e-01
1.07468259e+00 -1.48269439e+00 9.23600420e-02 -3.44228119e-01
-8.51117849e-01 1.07536376e+00 8.53635967e-01 3.59749287e-01
1.18261904e-01 1.03347294e-01 3.85030538e-01 -3.75173271e-01
-8.35831046e-01 -4.06647734e-02 1.38187751e-01 8.97458434e-01
4.94856298e-01 -2.63308957e-02 -2.52455503e-01 1.44185349e-01
3.04900091e-02 7.16162995e-02 8.45281243e-01 5.67143381e-01
-4.12445009e-01 -8.74461591e-01 -5.89589536e-01 1.79494679e-01
-1.48662493e-01 8.27295408e-02 -7.72778034e-01 1.12918353e+00
3.36674929e-01 1.00731814e+00 -3.96306701e-02 -1.30632043e-01
7.22515490e-03 9.62083694e-03 1.68653220e-01 -3.58181655e-01
-4.91777927e-01 -5.36307931e-01 -5.88873997e-02 -4.20700163e-01
-6.65680051e-01 -3.80983323e-01 -1.22613621e+00 -2.35687122e-01
-1.85355127e-01 1.69648379e-01 1.11905777e+00 9.61001098e-01
2.16375291e-01 5.10414839e-01 1.85306713e-01 -5.01480877e-01
-2.48245195e-01 -7.05295563e-01 -3.75050932e-01 8.92625868e-01
2.79693514e-01 -6.10424817e-01 -5.33792078e-01 6.48456365e-02] | [8.99783992767334, 5.479629039764404] |
66db4836-ccdf-49ea-8fd6-86f53a4ec117 | cltr-an-end-to-end-transformer-based-system-1 | null | null | https://aclanthology.org/2021.acl-demo.24 | https://aclanthology.org/2021.acl-demo.24.pdf | CLTR: An End-to-End, Transformer-Based System for Cell-Level Table Retrieval and Table Question Answering | We present the first end-to-end, transformer-based table question answering (QA) system that takes natural language questions and massive table corpora as inputs to retrieve the most relevant tables and locate the correct table cells to answer the question. Our system, CLTR, extends the current state-of-the-art QA over tables model to build an end-to-end table QA architecture. This system has successfully tackled many real-world table QA problems with a simple, unified pipeline. Our proposed system can also generate a heatmap of candidate columns and rows over complex tables and allow users to quickly identify the correct cells to answer questions. In addition, we introduce two new open domain benchmarks, E2E{\_}WTQ and E2E{\_}GNQ, consisting of 2,005 natural language questions over 76,242 tables. The benchmarks are designed to validate CLTR as well as accommodate future table retrieval and end-to-end table QA research and experiments. Our experiments demonstrate that our system is the current state-of-the-art model on the table retrieval task and produces promising results for end-to-end table QA. | ['Peter Fox', 'Alfio Gliozzo', 'Michael Glass', 'Mustafa Canim', 'Feifei Pan'] | 2021-08-01 | null | null | null | acl-2021-5 | ['table-retrieval'] | ['natural-language-processing'] | [-1.76408350e-01 3.60275596e-01 6.67122705e-03 -4.50863481e-01
-2.29813123e+00 -1.16112721e+00 3.80162090e-01 6.50710821e-01
7.59094860e-03 8.59209776e-01 5.80022395e-01 -5.85447431e-01
-1.75277978e-01 -1.23425317e+00 -8.79523456e-01 1.94866419e-01
3.56654316e-01 1.73003018e+00 6.21445119e-01 -1.02932918e+00
1.31308004e-01 -1.51049301e-01 -1.26852560e+00 1.17369843e+00
1.34413886e+00 1.40567911e+00 -2.18102887e-01 6.45884454e-01
-5.94047904e-01 1.30745006e+00 -7.76863873e-01 -1.24760568e+00
2.24124759e-01 -2.34931961e-01 -1.45592356e+00 -6.49458230e-01
9.05200481e-01 -1.57387078e-01 -2.80199915e-01 6.94756091e-01
5.85531294e-01 -1.33722192e-02 4.06080902e-01 -8.39906931e-01
-8.89307797e-01 8.78723383e-01 7.26945996e-02 2.45333567e-01
1.01751733e+00 1.55182078e-01 1.66362059e+00 -1.00706673e+00
8.62509727e-01 1.54967260e+00 4.32705611e-01 2.36318931e-01
-9.17051315e-01 -3.19438457e-01 -2.43936524e-01 4.13425326e-01
-1.27189720e+00 -5.23215353e-01 2.85066843e-01 1.51819706e-01
1.40491128e+00 6.42193556e-01 7.53525868e-02 4.19058889e-01
3.52686137e-01 1.03849983e+00 7.66553998e-01 -3.19117278e-01
1.06042661e-01 4.76086028e-02 1.52435482e-01 8.38270247e-01
2.44263545e-01 -9.13283706e-01 -6.52326405e-01 -3.81988078e-01
1.44149050e-01 -7.62731791e-01 -4.00503390e-02 -3.42540503e-01
-1.48403919e+00 8.60063851e-01 6.28228307e-01 -1.50658965e-01
-3.79951268e-01 -1.81593463e-01 5.99412262e-01 4.67356980e-01
-4.31759581e-02 1.06983209e+00 -7.84338176e-01 -1.78056940e-01
-3.26971024e-01 8.37460876e-01 1.25301933e+00 1.23151433e+00
6.45869315e-01 -7.62854993e-01 -1.02237785e+00 7.98645318e-01
5.53253070e-02 8.59047174e-01 -1.10872410e-01 -1.10048890e+00
1.52703273e+00 1.17473829e+00 2.87285715e-01 -7.69251645e-01
-1.71452150e-01 -1.82855755e-01 -4.28748965e-01 -6.45874023e-01
7.69539595e-01 3.67583662e-01 -7.86864281e-01 1.22930074e+00
3.39755982e-01 -1.15468490e+00 5.29527068e-01 6.03443325e-01
1.29519296e+00 9.40709293e-01 -3.92676480e-02 2.23924279e-01
2.18035388e+00 -1.07392240e+00 -1.00206077e+00 -5.26731253e-01
7.95556545e-01 -8.76063704e-01 1.72997570e+00 2.98302203e-01
-1.37599635e+00 -4.12908852e-01 -5.76672316e-01 -9.18209374e-01
-8.27703476e-01 6.61081895e-02 5.34403920e-01 6.25913203e-01
-9.33719873e-01 -1.81210563e-01 -3.27286780e-01 -2.27624342e-01
2.64384538e-01 2.22610891e-01 -2.22438544e-01 -5.66535234e-01
-1.74919915e+00 1.17262399e+00 2.53228098e-01 8.80040526e-02
-6.73466861e-01 -1.04682910e+00 -1.08055961e+00 3.28735173e-01
9.88701642e-01 -1.09567940e+00 1.64526033e+00 4.48220462e-01
-1.02828205e+00 8.49704742e-01 -5.87136567e-01 -5.07629812e-01
8.10631886e-02 -3.26761246e-01 -5.20738721e-01 2.48123869e-01
7.55831122e-01 6.40916646e-01 -1.95188284e-01 -9.73647654e-01
-5.53169847e-01 -3.70226026e-01 3.53384674e-01 4.98723447e-01
2.35292956e-01 -2.41060462e-02 -8.32694471e-01 -1.73398688e-01
1.65757462e-01 -3.28425199e-01 7.19099417e-02 -1.73453808e-01
-6.50780797e-01 -6.30145967e-01 1.38845518e-01 -1.10050106e+00
1.47593832e+00 -1.41346920e+00 -1.59645543e-01 1.43962279e-01
2.24470943e-01 -1.86334059e-01 -1.10149281e-02 9.21029389e-01
4.98889983e-01 -4.66077439e-02 -3.84128302e-01 7.27677122e-02
5.03758013e-01 7.51683116e-02 -7.48454511e-01 -3.96398008e-01
3.39961588e-01 1.50093234e+00 -8.65472555e-01 -7.81271935e-01
-1.18261993e-01 -1.23902105e-01 -5.67678988e-01 1.87207550e-01
-8.86201859e-01 -2.22388208e-02 -6.17416084e-01 1.06556165e+00
4.81925786e-01 -5.03571332e-01 2.86799818e-02 -2.62057900e-01
5.22185445e-01 1.31160343e+00 -9.41143095e-01 1.81273687e+00
-3.84005278e-01 1.93984389e-01 -2.54528731e-01 -3.32747288e-02
1.00452185e+00 -5.52463308e-02 8.30135718e-02 -1.37959754e+00
-5.58083914e-02 6.56168044e-01 -1.50673494e-01 -3.36990207e-01
9.24178898e-01 4.46363598e-01 -9.38317418e-01 2.50394076e-01
-2.14039180e-02 -6.80355132e-01 7.63133943e-01 6.80353463e-01
1.32138705e+00 -2.18052700e-01 2.27954879e-01 -4.75487232e-01
1.02520883e+00 8.02885592e-01 4.66020283e-04 8.89122963e-01
1.54923171e-01 4.38197970e-01 4.97322619e-01 -4.73559678e-01
-7.70735800e-01 -1.34647512e+00 8.50638468e-03 1.25114930e+00
1.13629669e-01 -8.62554729e-01 -8.12851131e-01 -8.18662286e-01
1.31455854e-01 9.50525880e-01 -6.29521012e-01 -9.35535654e-02
-8.73043060e-01 -2.80228555e-01 7.57005632e-01 5.33577085e-01
8.21425438e-01 -1.26080883e+00 -2.12533578e-01 3.98311377e-01
-1.16225064e+00 -1.15683103e+00 -5.88104010e-01 9.86714661e-02
-5.25577426e-01 -1.20036495e+00 -1.35709077e-01 -8.28628361e-01
1.56238660e-01 -1.90261781e-01 2.38483047e+00 -1.94417536e-01
1.02392025e-01 2.94948369e-01 -2.37755224e-01 -4.15235400e-01
-6.22228980e-01 5.51976204e-01 -5.25565207e-01 -5.42927742e-01
6.83705807e-01 2.90047854e-01 -5.09791791e-01 5.48550308e-01
-7.16924369e-01 -3.07717621e-01 4.34922874e-01 5.96140385e-01
1.00128949e+00 -2.71402121e-01 5.67044199e-01 -1.11182940e+00
1.23601186e+00 -2.94847906e-01 -8.28941762e-01 1.15297079e+00
-6.54377341e-01 6.29503310e-01 9.62316334e-01 4.21498001e-01
-9.94877577e-01 -2.85550058e-01 -6.24327600e-01 3.28682423e-01
2.21866369e-01 7.02550650e-01 -6.30623460e-01 4.45057213e-01
9.76243377e-01 3.14841509e-01 -3.42597932e-01 -4.89335179e-01
7.89076149e-01 5.09125650e-01 8.77373099e-01 -8.50476146e-01
8.73594403e-01 2.20471337e-01 -2.32067704e-01 2.58578956e-01
-1.15938222e+00 -5.17027617e-01 -5.70277750e-01 2.14072734e-01
8.30028296e-01 -1.05533838e+00 -1.09994686e+00 -4.16327305e-02
-9.18814361e-01 -1.73219800e-01 -4.60274845e-01 -4.04716253e-01
-6.21113539e-01 9.79346782e-02 -6.28298938e-01 -4.92332667e-01
-1.03855765e+00 -1.13248289e+00 1.58259058e+00 3.62729132e-02
-2.61965543e-01 -6.84655488e-01 1.84974134e-01 1.05539310e+00
4.77753818e-01 -2.11809754e-01 1.38989449e+00 -8.32522631e-01
-1.05664718e+00 -1.96094170e-01 -1.77063718e-01 -2.98782170e-01
-9.58194770e-03 -5.20769835e-01 -6.11569464e-01 3.90869975e-02
-3.84128124e-01 -9.42868233e-01 6.99738920e-01 -1.46638989e-01
9.48491395e-01 -4.40881848e-01 -1.05317451e-01 1.01879202e-01
1.22704268e+00 3.42545927e-01 1.19892120e+00 6.89754486e-01
4.95070189e-01 5.81257463e-01 1.16208065e+00 -7.30706230e-02
1.37363935e+00 6.61711216e-01 3.71145159e-01 1.42422497e-01
-1.43016040e-01 -6.94211900e-01 7.62865990e-02 8.12027693e-01
8.02118063e-01 -5.55889845e-01 -1.20723331e+00 6.96942449e-01
-1.46810532e+00 -5.67195535e-01 -2.03467280e-01 2.05321527e+00
1.08746338e+00 4.06394780e-01 8.17097165e-03 1.50761873e-01
1.88321799e-01 -6.15388714e-02 -4.82750863e-01 -5.20702899e-01
-2.28803039e-01 7.13758051e-01 2.25945786e-01 4.44885701e-01
-8.71350765e-01 1.14705932e+00 6.06668186e+00 9.58304703e-01
-2.43962452e-01 -2.25884199e-01 7.81995356e-01 3.88155043e-01
-7.46549666e-01 -2.77675819e-02 -1.14234257e+00 -5.29116467e-02
1.12372363e+00 -2.97223806e-01 5.76460660e-01 6.44585729e-01
-7.29952812e-01 -1.89029366e-01 -1.17640913e+00 7.62204587e-01
1.01255722e-01 -1.66909170e+00 6.42526269e-01 -4.19914603e-01
2.52655506e-01 -2.88648278e-01 2.72428077e-02 9.54651237e-01
4.62675303e-01 -1.17313313e+00 6.68602526e-01 3.53092819e-01
1.03812945e+00 -6.83214486e-01 9.39598858e-01 4.47197109e-02
-1.50922501e+00 -1.22216739e-01 -4.14612174e-01 2.62908906e-01
1.09763958e-01 7.58602247e-02 -1.19981086e+00 1.00463533e+00
8.94810379e-01 3.25905606e-02 -1.23919821e+00 7.15431809e-01
-2.04360932e-01 3.17179769e-01 -4.15950418e-01 -4.15641308e-01
3.03982228e-01 7.10165426e-02 9.41814762e-03 8.97945285e-01
2.98759174e-02 2.55479634e-01 -1.88162968e-01 8.16029668e-01
-5.57110310e-01 2.36695498e-01 -2.73618162e-01 1.26655400e-01
7.65430152e-01 1.01993668e+00 -5.39146960e-01 -6.80055201e-01
-1.22582495e-01 5.92440903e-01 4.62819517e-01 -1.32031575e-01
-5.69350123e-01 -7.59984612e-01 4.20619816e-01 -9.35999751e-02
3.18980098e-01 3.31717759e-01 -6.54577971e-01 -1.14852977e+00
6.23763323e-01 -1.74266493e+00 1.32274306e+00 -1.01310980e+00
-1.27825630e+00 9.12639081e-01 -4.13751155e-02 -1.03864789e+00
-6.50295019e-01 -4.67009962e-01 4.45388742e-02 1.05092525e+00
-1.44709873e+00 -8.84442449e-01 -3.60164970e-01 8.84060204e-01
6.37560904e-01 -4.99488078e-02 1.09024131e+00 4.89844054e-01
2.01622979e-03 9.52818513e-01 8.52075666e-02 4.63900983e-01
8.33541572e-01 -1.71847248e+00 1.10274780e+00 7.33243823e-01
2.44002894e-01 8.15466106e-01 5.07550180e-01 -6.22540832e-01
-2.02834368e+00 -8.99808645e-01 1.45605206e+00 -1.49749684e+00
5.48515677e-01 -9.83451545e-01 -1.07754314e+00 6.41988754e-01
4.04397964e-01 -1.75772920e-01 2.88655192e-01 1.53695062e-01
-6.66346371e-01 -7.10796416e-01 -1.40912044e+00 5.82496583e-01
8.30880582e-01 -7.60778427e-01 -9.41202700e-01 5.83663225e-01
1.08888185e+00 -1.05860949e+00 -1.17445493e+00 5.68296611e-01
4.29527819e-01 -5.39048612e-01 1.26238942e+00 -4.82285291e-01
5.61984718e-01 -4.65456337e-01 -3.72211665e-01 -1.01297474e+00
-1.62756249e-01 -4.68920350e-01 -3.30907613e-01 1.15023899e+00
1.13885689e+00 -4.84623373e-01 8.66962910e-01 8.36838663e-01
-2.58158237e-01 -9.95704591e-01 -1.06727111e+00 -3.36956620e-01
2.94997573e-01 -8.86212885e-02 1.10553813e+00 4.36249733e-01
-2.06889451e-01 7.94590890e-01 1.58501387e-01 1.63301244e-01
3.75991821e-01 4.79482234e-01 1.00814641e+00 -9.55938876e-01
-3.95829789e-02 -1.02505885e-01 1.95640117e-01 -1.42239141e+00
-4.17810768e-01 -9.48485136e-01 1.62499100e-01 -2.33115625e+00
2.82292869e-02 -4.49741215e-01 7.88172483e-02 4.30730611e-01
-4.88307178e-01 5.35379685e-02 2.20114678e-01 3.70904505e-02
-1.11744475e+00 5.07262349e-01 1.52064383e+00 -4.82459873e-01
1.15276776e-01 -2.23601609e-01 -1.29174674e+00 -2.95179963e-01
4.26247656e-01 -4.64493364e-01 -6.00428462e-01 -7.25933909e-01
7.11664498e-01 6.26229644e-01 -2.33837962e-01 -9.70065951e-01
4.68875229e-01 1.82029471e-01 3.66329402e-01 -1.29137337e+00
4.77566481e-01 -5.94577014e-01 -3.66394877e-01 3.05813611e-01
-5.48884630e-01 8.89563084e-01 4.39368159e-01 1.11641146e-01
-5.33186555e-01 1.32200509e-01 1.17215514e-01 -3.28702837e-01
-4.41392750e-01 1.01745436e-02 -2.78390199e-01 1.00966930e+00
3.82594436e-01 2.90847480e-01 -1.05482197e+00 -6.70599997e-01
-5.75678423e-02 1.01820040e+00 -3.31006236e-02 3.90090555e-01
6.13427937e-01 -1.12114370e+00 -9.61111069e-01 -9.56924558e-02
6.53735638e-01 3.35236996e-01 1.88155398e-02 1.91005051e-01
-8.73810709e-01 1.22834110e+00 -1.41647249e-01 -4.51706827e-01
-8.75985682e-01 5.71394265e-01 3.46522629e-01 -1.25256956e+00
-2.14830592e-01 9.90853488e-01 -1.93071514e-01 -1.16718423e+00
1.97022587e-01 -7.00726211e-01 -1.93624064e-01 -9.99785513e-02
5.58163345e-01 -4.24578087e-03 8.67222548e-01 -9.04606730e-02
-5.99570751e-01 1.33331299e-01 -3.37564111e-01 -5.28325588e-02
7.80972660e-01 -4.96405736e-02 -3.61630231e-01 1.65775508e-01
7.94086277e-01 2.28217229e-01 -1.83585778e-01 -3.67340088e-01
2.14790806e-01 -3.66652369e-01 -7.06226766e-01 -1.76937068e+00
-5.07169843e-01 6.76003098e-01 2.05689162e-01 8.17244425e-02
1.24158680e+00 1.72541335e-01 1.26953983e+00 1.22756267e+00
7.04045236e-01 -6.76971555e-01 9.87386033e-02 8.70629132e-01
1.13573730e+00 -1.20476806e+00 -1.14493534e-01 -7.07192063e-01
-8.11090887e-01 9.13967133e-01 9.56350923e-01 3.85314524e-01
-1.46237602e-02 1.50502408e-02 4.06948864e-01 -5.74390709e-01
-1.21785367e+00 -3.62603217e-01 6.47871852e-01 4.14210916e-01
3.34719151e-01 -1.84982479e-01 -8.89144689e-02 6.04991019e-01
-8.28469932e-01 -3.37858707e-01 2.42007554e-01 9.82701898e-01
-2.65259653e-01 -1.14161289e+00 -5.36118388e-01 7.73135245e-01
-4.58688945e-01 -5.25912166e-01 -7.69273698e-01 8.91141713e-01
-5.10718405e-01 1.18566489e+00 -1.39480397e-01 -2.74243742e-01
9.68103588e-01 4.19226587e-01 4.48988467e-01 -5.08025527e-01
-1.18751585e+00 -7.52263248e-01 7.01888144e-01 -5.75965703e-01
4.30990696e-01 -3.24580342e-01 -1.26968575e+00 -3.75928909e-01
-1.50368717e-02 7.51881778e-01 1.09471917e-01 5.40246487e-01
7.00434625e-01 4.15876091e-01 1.79932386e-01 5.19949734e-01
-7.35024989e-01 -1.13148868e+00 7.53597841e-02 3.95104855e-01
7.06311539e-02 -4.04722482e-01 2.12106347e-01 -3.20512503e-01] | [10.31632137298584, 7.836194038391113] |
8c662934-1f82-45e9-8f77-e71224c9d5a5 | long-term-leap-attention-short-term-periodic | 2207.05526 | null | https://arxiv.org/abs/2207.05526v2 | https://arxiv.org/pdf/2207.05526v2.pdf | Long-term Leap Attention, Short-term Periodic Shift for Video Classification | Video transformer naturally incurs a heavier computation burden than a static vision transformer, as the former processes $T$ times longer sequence than the latter under the current attention of quadratic complexity $(T^2N^2)$. The existing works treat the temporal axis as a simple extension of spatial axes, focusing on shortening the spatio-temporal sequence by either generic pooling or local windowing without utilizing temporal redundancy. However, videos naturally contain redundant information between neighboring frames; thereby, we could potentially suppress attention on visually similar frames in a dilated manner. Based on this hypothesis, we propose the LAPS, a long-term ``\textbf{\textit{Leap Attention}}'' (LA), short-term ``\textbf{\textit{Periodic Shift}}'' (\textit{P}-Shift) module for video transformers, with $(2TN^2)$ complexity. Specifically, the ``LA'' groups long-term frames into pairs, then refactors each discrete pair via attention. The ``\textit{P}-Shift'' exchanges features between temporal neighbors to confront the loss of short-term dynamics. By replacing a vanilla 2D attention with the LAPS, we could adapt a static transformer into a video one, with zero extra parameters and neglectable computation overhead ($\sim$2.6\%). Experiments on the standard Kinetics-400 benchmark demonstrate that our LAPS transformer could achieve competitive performances in terms of accuracy, FLOPs, and Params among CNN and transformer SOTAs. We open-source our project in \sloppy \href{https://github.com/VideoNetworks/LAPS-transformer}{\textit{\color{magenta}{https://github.com/VideoNetworks/LAPS-transformer}}} . | ['Chong-Wah Ngo', 'Yanbin Hao', 'Lechao Cheng', 'Hao Zhang'] | 2022-07-12 | null | null | null | null | ['video-classification'] | ['computer-vision'] | [ 1.72024518e-01 -1.13825329e-01 8.53252336e-02 -5.19203916e-02
-6.69364691e-01 -6.11953795e-01 1.32582262e-01 -3.72762412e-01
-7.32395768e-01 5.13029933e-01 -1.49362162e-01 -4.75485682e-01
-3.95002142e-02 -7.32506990e-01 -1.05386901e+00 -7.68373609e-01
-2.15960041e-01 -2.66377032e-01 6.52832806e-01 -1.25210479e-01
9.19996426e-02 9.35841799e-02 -1.52151453e+00 2.96582460e-01
6.36639953e-01 1.37467611e+00 5.55468380e-01 6.40183628e-01
2.44808290e-02 8.98151338e-01 -3.47248852e-01 -6.07351184e-01
5.41971862e-01 -3.28830451e-01 -7.09292650e-01 -1.97980896e-01
5.30978203e-01 -3.49011451e-01 -8.89947712e-01 1.18023384e+00
5.99014938e-01 3.21373910e-01 6.67269081e-02 -1.23064029e+00
-7.04378128e-01 5.04366398e-01 -9.07227159e-01 7.92958558e-01
1.94093823e-01 5.28587997e-01 9.35611904e-01 -1.02714825e+00
4.83784169e-01 1.03056669e+00 7.97460616e-01 4.20803189e-01
-1.02493954e+00 -9.70766604e-01 6.95469201e-01 3.75414610e-01
-1.52841067e+00 -3.59257102e-01 5.32106400e-01 -2.31826380e-01
9.88292038e-01 4.77629155e-01 7.56669879e-01 9.85578179e-01
1.41968817e-01 7.97253191e-01 9.55518901e-01 4.77825291e-02
-8.11580122e-02 -3.92725587e-01 1.43104598e-01 9.23398733e-01
-2.12628290e-01 -1.90153401e-02 -6.18552446e-01 2.07386091e-01
1.07720578e+00 3.58734250e-01 -6.61188304e-01 2.33898923e-01
-1.42136908e+00 4.85725492e-01 4.84699219e-01 1.81990296e-01
-2.69078523e-01 5.99747121e-01 4.73761111e-01 5.15415668e-01
4.70597208e-01 5.00014573e-02 -5.62042654e-01 -2.77763456e-01
-9.13933873e-01 5.81172258e-02 3.24052602e-01 1.29424918e+00
8.92334700e-01 -4.42730375e-02 -4.12569582e-01 6.90203011e-01
-3.06191649e-02 6.21357501e-01 4.39911902e-01 -1.32081544e+00
5.08475065e-01 2.37154245e-01 1.70966405e-02 -9.96416867e-01
-1.91153631e-01 -2.69707263e-01 -1.13004589e+00 1.22810110e-01
3.56001645e-01 -2.18532309e-01 -8.73407423e-01 1.83507526e+00
-1.48531154e-03 4.10601586e-01 -5.20020485e-01 9.20839310e-01
7.04940438e-01 7.60974765e-01 -1.27755120e-01 -5.53133547e-01
1.59534311e+00 -1.18393552e+00 -4.90049064e-01 -1.70088168e-02
2.29509935e-01 -9.30459321e-01 1.34307384e+00 3.39770079e-01
-1.51129103e+00 -7.22578704e-01 -7.07250357e-01 -2.47426137e-01
-2.57571727e-01 -6.94010332e-02 4.79290545e-01 1.61830783e-01
-1.38206720e+00 6.97750866e-01 -9.02847946e-01 -1.02652811e-01
5.08462846e-01 4.80732083e-01 -3.44181299e-01 -6.33726865e-02
-1.23847842e+00 4.32313174e-01 -2.74364911e-02 3.51545304e-01
-8.07290196e-01 -6.77441120e-01 -7.91421413e-01 1.47689313e-01
6.93901420e-01 -7.32336879e-01 1.14368367e+00 -1.03276718e+00
-1.26255286e+00 7.02210069e-01 -3.98685515e-01 -4.29634511e-01
7.00172126e-01 -2.93226540e-01 -1.41116142e-01 2.32248828e-01
1.65012956e-01 8.54386628e-01 1.06664419e+00 -7.05289483e-01
-7.28075385e-01 -2.26856902e-01 4.43812609e-01 1.69642046e-01
-3.46797973e-01 1.54675022e-01 -1.13248634e+00 -1.04583549e+00
4.14335839e-02 -1.12492442e+00 1.56490877e-02 1.81906432e-01
-2.48879611e-01 -2.16501325e-01 9.16552663e-01 -6.25702918e-01
1.54906321e+00 -2.29421878e+00 2.26341248e-01 -1.89730793e-01
4.91985381e-01 4.05955613e-01 -7.46431872e-02 1.41933814e-01
-2.52857536e-01 2.74844617e-01 -1.58129096e-01 -4.26784188e-01
-1.59541070e-01 1.54830376e-02 -2.09346697e-01 3.92682195e-01
-1.41888857e-01 8.78741086e-01 -8.05078566e-01 -3.33066255e-01
1.49604261e-01 5.67259610e-01 -6.38504803e-01 -5.09815849e-02
-9.09481943e-02 2.88106561e-01 -4.23210353e-01 5.97532630e-01
7.25666046e-01 -4.59445566e-01 -4.35513034e-02 -4.33477461e-01
-4.29695100e-01 7.09810480e-02 -9.53922927e-01 1.77712667e+00
-1.82532176e-01 7.03217387e-01 1.89154968e-01 -9.62187767e-01
4.84853715e-01 2.08707765e-01 6.61099494e-01 -7.17441201e-01
1.93428189e-01 8.49424452e-02 -2.67441541e-01 -5.27727365e-01
2.60868073e-01 1.59474254e-01 9.26155075e-02 3.21906269e-01
-7.44914040e-02 1.82995960e-01 2.09135473e-01 5.52023202e-02
1.31201649e+00 3.03218096e-01 -2.45460719e-01 -1.80726945e-01
4.78103876e-01 -3.79127532e-01 6.89774394e-01 6.52400136e-01
-4.33681101e-01 6.83764338e-01 6.34376943e-01 -7.08891571e-01
-8.77500296e-01 -9.96872842e-01 2.07951263e-01 1.37019694e+00
4.28412259e-01 -5.04767597e-01 -8.37283611e-01 -7.14111924e-01
-3.17490876e-01 1.94966108e-01 -6.41991079e-01 1.51846511e-02
-9.30753171e-01 -6.98838174e-01 6.05383396e-01 6.94398344e-01
9.02128994e-01 -1.22375500e+00 -8.91198575e-01 6.77344799e-02
-4.29151833e-01 -9.82757628e-01 -1.14957285e+00 1.49997696e-01
-5.69468439e-01 -8.68330896e-01 -1.09399879e+00 -7.97477365e-01
6.14286244e-01 6.40127659e-01 9.98850048e-01 1.56649381e-01
-1.86326042e-01 2.28153765e-01 -2.98767954e-01 -1.92849696e-01
5.18295467e-01 -4.23381962e-02 -4.35444787e-02 -1.08690366e-01
2.97713995e-01 -9.04576182e-01 -1.27991617e+00 6.17135763e-01
-9.64463115e-01 3.07767212e-01 4.03317362e-01 9.63699520e-01
7.78544307e-01 -9.99623071e-03 1.28382012e-01 -4.00303751e-01
3.80641818e-01 -1.95026100e-01 -6.15037441e-01 8.18981677e-02
-2.19275177e-01 -2.58244485e-01 6.60176158e-01 -5.76377332e-01
-6.85614347e-01 -2.30907843e-01 -1.00363992e-01 -1.12461662e+00
1.95664883e-01 3.36048305e-01 1.22181468e-01 5.24412617e-02
3.05368304e-01 5.31170607e-01 4.35735472e-02 -3.18403363e-01
1.07554585e-01 1.70756891e-01 5.28794229e-01 -4.96650934e-01
6.58570886e-01 6.43884838e-01 -3.59908581e-01 -5.10163188e-01
-6.02859378e-01 -1.92281753e-01 -2.91121811e-01 -2.46376827e-01
9.88258004e-01 -9.95637894e-01 -1.27086508e+00 5.29173911e-01
-1.09694171e+00 -4.89140093e-01 -3.31244051e-01 4.88663673e-01
-5.74559093e-01 4.83103156e-01 -7.76111901e-01 -4.82329577e-01
-4.33027983e-01 -1.37881851e+00 8.90603960e-01 2.40945160e-01
-3.13973129e-02 -4.64836210e-01 -4.23971295e-01 4.52861398e-01
4.21601951e-01 -5.04664071e-02 5.71636260e-01 -1.37906209e-01
-9.52854753e-01 3.67032774e-02 -5.66953540e-01 5.62117636e-01
-2.45649703e-02 -1.80973545e-01 -9.41288948e-01 -6.34000003e-01
3.34470533e-02 -6.34764647e-03 1.02391720e+00 5.65715492e-01
1.59559810e+00 -3.83640766e-01 -1.62699685e-01 9.55581546e-01
1.20500791e+00 4.92317349e-01 8.54370356e-01 1.84265777e-01
5.95475256e-01 2.16129512e-01 4.23544586e-01 4.52066004e-01
2.99778104e-01 6.35020256e-01 4.97308105e-01 -1.74995437e-01
-1.71495542e-01 5.93632460e-02 6.70461774e-01 8.07344139e-01
-7.26297498e-01 -3.30904037e-01 -6.54454648e-01 5.81508636e-01
-1.91675365e+00 -1.17798913e+00 2.11939827e-01 2.19931269e+00
6.73356712e-01 2.59201735e-01 1.92309059e-02 3.14551778e-02
8.58588159e-01 3.52137148e-01 -5.14123142e-01 -1.02636449e-01
-7.83712193e-02 3.60177040e-01 6.57905817e-01 3.81846875e-01
-1.10308278e+00 9.40273702e-01 4.16608667e+00 1.13757849e+00
-1.37293804e+00 2.00489089e-01 8.78591239e-01 -6.46267891e-01
-1.09269537e-01 -8.05717111e-02 -7.03908563e-01 8.82359743e-01
5.96598864e-01 -1.26605913e-01 6.17303908e-01 4.52496827e-01
4.36863065e-01 -1.08510636e-01 -9.04234469e-01 1.10846114e+00
-8.38305876e-02 -1.50847626e+00 -3.85949835e-02 -4.43259068e-02
3.24789464e-01 1.18952662e-01 3.07754815e-01 3.11834395e-01
-7.55109068e-04 -9.27207947e-01 9.88415062e-01 3.99857074e-01
1.26588833e+00 -5.13415277e-01 5.41481555e-01 -4.37736278e-03
-1.82521582e+00 -1.42095909e-01 -1.83275476e-01 -5.67054860e-02
1.62394717e-01 3.95237416e-01 1.49590392e-02 5.59763789e-01
1.57035124e+00 7.41393685e-01 -3.04152817e-01 7.14398980e-01
1.19372152e-01 2.64052927e-01 -4.26024616e-01 2.99812555e-01
3.50767195e-01 -1.97882637e-01 6.21570110e-01 1.23332906e+00
5.18628955e-01 5.03137052e-01 1.28003418e-01 4.94734108e-01
-1.74613863e-01 -1.01632774e-01 -4.44820911e-01 2.64247566e-01
2.94742525e-01 8.93434763e-01 -9.16815400e-01 -4.09605324e-01
-5.55063546e-01 1.19861722e+00 8.97441059e-02 5.72530389e-01
-1.45520079e+00 -5.62719107e-01 7.50016391e-01 3.53769690e-01
7.74842799e-01 -1.11619376e-01 -4.46599461e-02 -1.34350801e+00
4.71503913e-01 -8.44613791e-01 4.75032926e-01 -8.97162080e-01
-9.88600671e-01 8.46062005e-01 -1.33981302e-01 -1.39493132e+00
4.29956973e-01 -3.88212889e-01 -4.14345592e-01 8.39795291e-01
-1.37287939e+00 -1.01922202e+00 -5.34366727e-01 1.11453855e+00
8.79378021e-01 1.89350858e-01 4.61652070e-01 6.23128295e-01
-6.10359967e-01 8.60942841e-01 -8.00981671e-02 1.65486991e-01
7.38537133e-01 -8.70661557e-01 3.26390624e-01 7.20839739e-01
-2.68133342e-01 6.36220515e-01 5.37333727e-01 -4.07993406e-01
-1.16225219e+00 -1.15122998e+00 7.75706291e-01 -7.84081593e-02
7.33120561e-01 -2.58981973e-01 -8.32617700e-01 8.68357241e-01
4.08315450e-01 3.82692516e-01 3.73975068e-01 -4.57611442e-01
-4.13970679e-01 -3.35675448e-01 -8.77672434e-01 8.45467865e-01
1.45296502e+00 -5.83441138e-01 -2.82908618e-01 2.13136449e-01
1.01041353e+00 -3.27126503e-01 -8.63191783e-01 3.68264139e-01
6.52900159e-01 -1.30114424e+00 1.07963657e+00 -1.72301363e-02
4.14045721e-01 -4.74216074e-01 -1.33626625e-01 -6.21425450e-01
-3.90133619e-01 -1.17486417e+00 2.16030404e-02 8.58919144e-01
2.85225868e-01 -6.81948245e-01 6.75900102e-01 4.22462970e-01
-3.11232775e-01 -1.03201282e+00 -1.15013850e+00 -6.86187565e-01
-3.09454668e-02 -4.58862782e-01 3.33234042e-01 7.34296501e-01
-1.36420503e-01 -9.85235907e-03 -4.96954590e-01 5.04886284e-02
2.97450751e-01 -9.57860276e-02 3.64217997e-01 -6.24571264e-01
-5.15053749e-01 -5.69898367e-01 -2.46723294e-01 -1.50416338e+00
-3.14137727e-01 -4.71021980e-01 -1.06261753e-01 -9.84259427e-01
2.75218487e-01 -3.39364827e-01 -5.08403122e-01 7.75597513e-01
-1.31028071e-01 5.62679410e-01 5.35548985e-01 3.04773390e-01
-6.79622114e-01 4.07459497e-01 1.34560955e+00 -1.33471042e-01
-8.37667733e-02 -1.26243711e-01 -4.64722544e-01 8.13963294e-01
5.39986670e-01 -2.51321167e-01 -5.52705944e-01 -8.09648514e-01
2.05674127e-01 3.16077828e-01 5.00181854e-01 -9.94784415e-01
4.24483329e-01 -4.82699741e-03 1.95946455e-01 -4.99779880e-01
4.77268875e-01 -5.22023737e-01 1.54371127e-01 3.48815143e-01
-7.01363310e-02 6.42484546e-01 4.07587916e-01 5.40593803e-01
-3.00986886e-01 2.75358111e-01 8.32847893e-01 -4.74994153e-01
-6.12636268e-01 7.03894973e-01 -3.84865522e-01 -3.71360593e-02
1.15079474e+00 -3.96049201e-01 -4.17362154e-01 -2.96692401e-01
-1.01985574e+00 2.04620451e-01 4.40732926e-01 3.18112642e-01
5.67005455e-01 -1.02997911e+00 -3.66970658e-01 2.04314917e-01
-3.08944762e-01 3.44782025e-01 8.10227573e-01 1.25054133e+00
-5.57055593e-01 2.33517498e-01 -2.07025722e-01 -6.58723235e-01
-1.29631853e+00 6.96510315e-01 4.37807441e-01 -2.94528514e-01
-8.04297090e-01 1.35280418e+00 6.23578191e-01 1.87957615e-01
3.15423787e-01 -4.04760003e-01 9.71961915e-02 1.92319632e-01
5.41719258e-01 3.74210805e-01 -9.09758136e-02 -4.55878586e-01
-3.17161381e-01 7.85959601e-01 -1.90591738e-01 -2.16021910e-02
1.24590802e+00 -4.03400123e-01 -2.03888983e-01 2.09773600e-01
1.22759783e+00 -2.46968597e-01 -1.58742309e+00 -1.57478243e-01
-5.49084604e-01 -4.49056417e-01 -1.38727605e-01 -4.49978441e-01
-1.50641441e+00 8.10911298e-01 6.81604803e-01 2.74624258e-01
1.56427038e+00 -1.48595259e-01 9.78183866e-01 4.88107838e-02
3.44765037e-01 -9.01595414e-01 1.64391547e-01 5.79748929e-01
9.66521502e-01 -1.01471198e+00 -1.33768797e-01 -3.89633536e-01
-4.74628776e-01 9.36407030e-01 6.94764256e-01 -1.63842544e-01
8.81208718e-01 1.98725536e-01 -1.76500186e-01 -1.55069351e-01
-8.36855710e-01 -2.86067533e-03 -2.36629155e-02 2.04107612e-01
3.57347816e-01 -2.19503343e-01 -2.95835465e-01 4.89604712e-01
-9.77042988e-02 2.05838442e-01 2.61319101e-01 1.00191879e+00
-1.22263849e-01 -7.54898906e-01 -1.74248725e-01 2.84501791e-01
-6.18804753e-01 -4.91091609e-01 2.64720052e-01 6.51415408e-01
6.45508766e-01 6.89366341e-01 3.34428787e-01 -4.08708990e-01
3.22112858e-01 -6.37814552e-02 2.40358457e-01 -1.62261039e-01
-7.71151066e-01 5.43377697e-01 -2.66251415e-01 -9.64519858e-01
-3.94799352e-01 -6.34418011e-01 -1.09367895e+00 -7.91708827e-01
1.49039896e-02 -1.85175538e-02 -4.24043089e-02 5.82688212e-01
5.08054137e-01 7.06847966e-01 4.23563004e-01 -1.07777011e+00
-1.15401335e-01 -7.92298079e-01 -3.22448850e-01 3.62698138e-01
2.72301078e-01 -5.34823000e-01 -4.19342816e-01 4.21058446e-01] | [9.232989311218262, 0.02330116555094719] |
de3035c2-ec35-4e08-a171-066c80cac9d1 | multi-pooled-inception-features-for-no | 2011.05139 | null | https://arxiv.org/abs/2011.05139v1 | https://arxiv.org/pdf/2011.05139v1.pdf | Multi-pooled Inception features for no-reference image quality assessment | Image quality assessment (IQA) is an important element of a broad spectrum of applications ranging from automatic video streaming to display technology. Furthermore, the measurement of image quality requires a balanced investigation of image content and features. Our proposed approach extracts visual features by attaching global average pooling (GAP) layers to multiple Inception modules of on an ImageNet database pretrained convolutional neural network (CNN). In contrast to previous methods, we do not take patches from the input image. Instead, the input image is treated as a whole and is run through a pretrained CNN body to extract resolution-independent, multi-level deep features. As a consequence, our method can be easily generalized to any input image size and pretrained CNNs. Thus, we present a detailed parameter study with respect to the CNN base architectures and the effectiveness of different deep features. We demonstrate that our best proposal - called MultiGAP-NRIQA - is able to provide state-of-the-art results on three benchmark IQA databases. Furthermore, these results were also confirmed in a cross database test using the LIVE In the Wild Image Quality Challenge database. | ['Domonkos Varga'] | 2020-11-10 | null | null | null | null | ['no-reference-image-quality-assessment'] | ['computer-vision'] | [ 3.18910033e-01 -3.78286690e-01 1.69777498e-01 -4.06407326e-01
-7.68228173e-01 -3.63890052e-01 5.00131428e-01 2.34614432e-01
-7.67033160e-01 3.35165411e-01 -2.57924616e-01 -1.64422989e-01
-2.11229369e-01 -1.15903533e+00 -8.60332370e-01 -5.55522442e-01
-2.40472049e-01 -1.98237330e-01 4.44732785e-01 -2.93650180e-01
1.95067659e-01 8.74662876e-01 -1.70783377e+00 5.98713338e-01
6.81706548e-01 1.64330304e+00 -5.67451231e-02 7.75081992e-01
1.43495649e-01 6.34943187e-01 -8.58188510e-01 -6.59450591e-01
5.44458091e-01 -2.58136034e-01 -7.33521879e-01 2.13986263e-01
7.99800217e-01 -4.87127513e-01 -4.02596921e-01 1.03378952e+00
5.89321613e-01 -6.50438964e-02 3.84160906e-01 -1.21109986e+00
-3.89724612e-01 1.84447184e-01 -4.10576344e-01 4.76784885e-01
1.51272967e-01 4.24426496e-01 9.80191290e-01 -9.18761551e-01
5.01640856e-01 1.04524124e+00 4.80036497e-01 5.75693548e-02
-1.15918159e+00 -5.75036407e-01 -1.81818292e-01 4.89640474e-01
-1.21341658e+00 -4.01256442e-01 5.90818465e-01 -3.82297575e-01
1.04168749e+00 4.60213535e-02 6.70629323e-01 6.88443720e-01
3.79915923e-01 3.50485772e-01 1.12365377e+00 -3.07306498e-01
3.08859795e-01 -3.77823040e-02 4.26764078e-02 6.22240424e-01
2.96167523e-01 -3.83320525e-02 -5.53103089e-01 1.63970217e-01
6.19219482e-01 -1.63731977e-01 -2.46574238e-01 -3.97399932e-01
-1.15839064e+00 5.86757243e-01 7.14587688e-01 3.92762214e-01
-4.39239502e-01 2.82681018e-01 5.07265806e-01 5.70981205e-01
2.12954447e-01 2.99163014e-01 -2.06162244e-01 5.34611903e-02
-1.19062257e+00 2.87980378e-01 5.52924871e-01 4.79535073e-01
8.68495762e-01 2.95401271e-02 -3.81152928e-01 6.03048682e-01
-8.99446234e-02 3.28274041e-01 2.67616749e-01 -9.80004668e-01
4.81286645e-01 6.67294681e-01 -1.66018587e-02 -1.22110534e+00
-3.97467494e-01 -5.85738599e-01 -1.06123996e+00 9.92992878e-01
5.73311448e-01 2.27160737e-01 -7.53024280e-01 1.38521528e+00
-1.27201239e-02 -8.19930211e-02 6.41100630e-02 9.59986567e-01
8.60167444e-01 6.73195362e-01 4.95447516e-02 -1.49459532e-03
1.52939034e+00 -7.55362332e-01 -4.29181814e-01 3.07343379e-02
1.45870373e-01 -7.06031740e-01 1.15524101e+00 8.45683694e-01
-1.34440160e+00 -1.10503900e+00 -1.47632039e+00 -1.96555689e-01
-5.48669338e-01 2.90799081e-01 1.31261498e-01 7.42957592e-01
-1.30615366e+00 8.89073789e-01 -5.22047997e-01 -2.18784794e-01
7.41189837e-01 3.96476418e-01 -7.70506084e-01 -2.83022612e-01
-1.01422572e+00 6.91419661e-01 4.27036673e-01 1.10651314e-01
-9.61569667e-01 -5.42536378e-01 -6.88132405e-01 4.07217264e-01
1.07803293e-01 -7.15167165e-01 8.35928202e-01 -1.24686837e+00
-1.50822771e+00 8.98385763e-01 1.50101736e-01 -7.51867115e-01
5.52624047e-01 -1.78586408e-01 -3.55501801e-01 4.73533839e-01
-3.03458452e-01 7.04697490e-01 1.04028344e+00 -1.00066650e+00
-6.18577123e-01 -3.42099458e-01 4.09503758e-01 -7.72036016e-02
-3.01763088e-01 1.25902936e-01 -6.53800845e-01 -5.76346934e-01
-7.93838948e-02 -3.96264553e-01 7.54196495e-02 3.67692024e-01
-1.11979090e-01 9.65297371e-02 5.97728491e-01 -5.79346538e-01
1.13893640e+00 -2.32958150e+00 1.46584466e-01 2.33762354e-01
3.13168466e-01 4.96809393e-01 -4.42820668e-01 1.74700364e-01
-1.59447208e-01 1.26046062e-01 -2.18587160e-01 -3.18583339e-01
-8.19412097e-02 -2.02696234e-01 -2.03745253e-02 6.18939102e-01
6.14460766e-01 8.32112074e-01 -5.17753243e-01 -4.24718380e-01
4.50695246e-01 5.68696260e-01 -6.18441284e-01 2.98183143e-01
5.63121624e-02 1.29974931e-01 4.37003858e-02 5.22743404e-01
9.86265302e-01 -2.21315965e-01 -2.50486672e-01 -5.97892046e-01
-1.44757509e-01 1.04583250e-02 -1.20795691e+00 1.87704825e+00
-5.23836792e-01 8.28825653e-01 -5.74950352e-02 -1.04198790e+00
8.63250971e-01 1.01877302e-01 3.54387969e-01 -1.22832823e+00
2.67415166e-01 2.26170421e-01 1.73147455e-01 -4.32769656e-01
5.57433426e-01 2.51141399e-01 3.21750268e-02 1.87396213e-01
5.39157391e-01 8.76021534e-02 3.10207069e-01 9.12858397e-02
1.15169775e+00 -1.95716590e-01 3.67873281e-01 -2.55478919e-01
7.53384650e-01 -3.19177568e-01 2.42126599e-01 6.91836655e-01
-3.08567941e-01 1.03350937e+00 6.68732762e-01 -5.65647185e-01
-1.32583463e+00 -1.08158815e+00 -1.89458266e-01 7.88174808e-01
9.61649492e-02 -4.27999765e-01 -8.80237937e-01 -4.46442783e-01
-2.28013352e-01 5.94169609e-02 -7.25714505e-01 5.72463206e-04
-4.93982643e-01 -6.19483113e-01 5.18618941e-01 2.91512787e-01
8.86337280e-01 -1.36331952e+00 -9.03511822e-01 2.73284376e-01
1.21644311e-01 -1.29574203e+00 -8.56249705e-02 -5.73426597e-02
-6.10004842e-01 -1.17684245e+00 -7.58722723e-01 -3.99461180e-01
1.72244698e-01 1.65423676e-01 1.37630761e+00 1.88332900e-01
-4.71556097e-01 1.12641864e-01 -2.60457665e-01 -1.42040357e-01
-3.82733941e-01 1.17164075e-01 -4.27664310e-01 3.46838385e-01
-2.90322229e-02 -6.55829549e-01 -1.06107306e+00 1.76529124e-01
-1.20009100e+00 -1.98617831e-01 8.68787706e-01 5.55219114e-01
6.15993738e-01 2.09775984e-01 4.83393252e-01 -5.97070158e-01
5.71841717e-01 -1.79622799e-01 -8.05123508e-01 2.20676452e-01
-3.43732744e-01 -7.81617612e-02 5.83102942e-01 -1.06480725e-01
-6.35527432e-01 -9.95993689e-02 -4.43014085e-01 -3.94556612e-01
-3.44650030e-01 2.08592504e-01 -5.73628545e-01 -2.44081795e-01
7.21745074e-01 1.26753420e-01 -1.16452992e-01 -2.79174775e-01
3.41789871e-01 5.25878966e-01 7.10131645e-01 -2.88068384e-01
6.28228426e-01 6.30781770e-01 1.60905048e-01 -6.95923388e-01
-3.49638969e-01 -3.33990842e-01 -6.35462224e-01 -3.50466400e-01
1.00359714e+00 -8.55127752e-01 -7.94487000e-01 8.51041198e-01
-1.21892142e+00 -1.89652711e-01 -2.79328585e-01 2.71228909e-01
-5.11942804e-01 3.19133639e-01 -5.21692455e-01 -3.54169905e-01
-5.60461044e-01 -1.54349196e+00 8.60585749e-01 2.74505347e-01
3.26608092e-01 -5.74974120e-01 -8.77674222e-02 1.28439993e-01
7.74070263e-01 3.08891356e-01 8.12582374e-01 -3.38599354e-01
-7.53018379e-01 -1.31141052e-01 -7.20614135e-01 8.74228776e-01
-7.68629834e-02 1.39110804e-01 -1.07739687e+00 -4.08591032e-01
-2.41566613e-01 -4.35178876e-01 1.02728403e+00 3.05045813e-01
1.31464446e+00 -3.06780301e-02 3.96241993e-01 8.40602458e-01
1.78545070e+00 -8.25919509e-02 1.03703225e+00 5.38551569e-01
3.01607013e-01 3.97624671e-01 3.57906461e-01 3.87582153e-01
-1.05130067e-02 8.43930721e-01 7.96498299e-01 -4.74725783e-01
-3.59034866e-01 2.52034783e-01 3.38410497e-01 2.73677379e-01
-7.31404945e-02 -3.35982680e-01 -6.42490089e-01 5.09538412e-01
-1.29349387e+00 -8.90027761e-01 5.27879782e-02 2.18328857e+00
3.67825389e-01 2.88112611e-01 2.23240569e-01 5.24319708e-01
4.81859922e-01 2.10265443e-01 -4.51969922e-01 -4.65627640e-01
-2.62732089e-01 6.63838983e-01 5.35070240e-01 1.80186689e-01
-1.10214412e+00 4.78136688e-01 5.74380827e+00 8.99582267e-01
-1.36287487e+00 2.16357604e-01 8.32380950e-01 -6.46776855e-02
1.34275392e-01 -4.30855304e-01 -4.34631467e-01 2.97994256e-01
1.02102101e+00 -4.22918908e-02 2.91251063e-01 7.09938467e-01
1.04582280e-01 -2.35736251e-01 -1.02715862e+00 1.12573981e+00
8.36493000e-02 -1.37387824e+00 2.80671358e-01 6.70338199e-02
5.72068572e-01 6.54458180e-02 3.04074436e-01 4.33008038e-02
-2.89384216e-01 -1.27473938e+00 8.91787708e-01 4.92208451e-01
1.00773168e+00 -8.96006703e-01 9.45377290e-01 -1.12546355e-01
-1.09200811e+00 -1.58335447e-01 -5.48241913e-01 1.48033231e-01
-7.73238167e-02 7.20608532e-01 -1.99583441e-01 8.41943085e-01
1.12243187e+00 4.92331177e-01 -9.50872064e-01 1.35727537e+00
2.34330714e-01 3.29455495e-01 -4.07818146e-02 4.82378423e-01
2.15797991e-01 1.67940453e-01 2.10514203e-01 1.38550973e+00
4.98774260e-01 -1.45440862e-01 -2.98843503e-01 8.19522083e-01
-3.74086261e-01 3.53348672e-01 -3.38802248e-01 1.80412441e-01
1.75440274e-02 1.31493104e+00 -7.34263539e-01 -4.52217698e-01
-5.91245651e-01 1.02905869e+00 2.71391004e-01 2.01666147e-01
-6.40779197e-01 -6.52684331e-01 8.15353572e-01 2.35485613e-01
5.32425880e-01 -3.17842104e-02 1.00272978e-02 -1.12039232e+00
2.34497383e-01 -1.10934067e+00 1.84845328e-01 -7.65069842e-01
-1.01173866e+00 1.03308702e+00 -1.13417931e-01 -1.43693972e+00
-3.96349877e-02 -7.76826680e-01 -6.19815469e-01 1.00748229e+00
-1.81071341e+00 -8.82683456e-01 -7.21289992e-01 8.41142297e-01
3.90888035e-01 -4.39365059e-02 5.62646806e-01 5.67167044e-01
-3.89265865e-01 6.29565418e-01 -6.50524199e-02 3.21162403e-01
5.91974914e-01 -1.04846632e+00 5.20416200e-01 1.06924546e+00
6.67928755e-02 2.09838957e-01 4.34700072e-01 3.47107276e-02
-1.15505791e+00 -1.04911029e+00 4.46147680e-01 2.09830850e-02
3.89978021e-01 -3.55653286e-01 -1.14269686e+00 6.82234094e-02
5.50089061e-01 5.70438743e-01 3.58279973e-01 -3.59212160e-01
-4.88856584e-01 -5.66285729e-01 -1.16670454e+00 6.14299923e-02
7.02200532e-01 -4.76761132e-01 -2.34804839e-01 -2.08824724e-01
5.27938724e-01 -2.57785767e-01 -1.05095899e+00 5.60619473e-01
6.70128703e-01 -1.65611255e+00 1.06504905e+00 -2.48207629e-01
5.96077025e-01 -4.07155842e-01 -1.56269237e-01 -1.09967911e+00
-1.24993794e-01 -1.54494211e-01 1.54712081e-01 1.06565368e+00
2.28599489e-01 -4.25453365e-01 5.36814570e-01 1.21738516e-01
-4.53662723e-02 -7.05972075e-01 -1.01201177e+00 -6.57403886e-01
1.05546571e-01 -5.39709687e-01 6.84374332e-01 4.34931338e-01
-5.27691603e-01 3.71001959e-02 -2.39172921e-01 2.56138444e-01
5.98325133e-01 -6.51095062e-02 7.80748725e-01 -1.05604315e+00
-3.47397834e-01 -7.19900489e-01 -9.22533095e-01 -5.10314107e-01
-2.53561527e-01 -5.58715940e-01 -2.17908889e-01 -1.40013540e+00
1.03330813e-01 -1.42603621e-01 -6.40339434e-01 9.24035460e-02
3.69386263e-02 8.14382255e-01 5.06208420e-01 -2.99247950e-02
-6.76248848e-01 3.25050950e-01 1.25764906e+00 -3.13368469e-01
9.04032737e-02 -8.27193037e-02 -2.82116503e-01 3.34518969e-01
8.26348364e-01 -3.02805156e-01 -2.80904680e-01 -4.32029605e-01
3.87486994e-01 -5.05962584e-04 7.24249244e-01 -1.63665414e+00
1.62014633e-01 2.47871459e-01 4.83890474e-01 -4.59386796e-01
1.99752867e-01 -7.43778527e-01 1.06406488e-01 4.46362048e-01
-1.98530048e-01 2.99326569e-01 4.03234512e-01 2.41971880e-01
-5.05026639e-01 -1.06662437e-01 1.02027810e+00 -7.55424351e-02
-8.02792251e-01 3.42605531e-01 -6.85927570e-02 -1.89060912e-01
7.44987011e-01 -2.14378193e-01 -3.93111318e-01 -1.22020282e-01
-5.59974074e-01 -3.50167006e-01 6.50666416e-01 2.62441456e-01
7.48571515e-01 -1.20971096e+00 -8.91511083e-01 4.31957215e-01
2.62649924e-01 -1.60553619e-01 6.59408629e-01 5.37741363e-01
-8.53424251e-01 2.34236181e-01 -8.23704302e-01 -7.29138553e-01
-1.19420207e+00 6.45126879e-01 6.27128720e-01 -3.30266088e-01
-5.89816213e-01 4.22118276e-01 2.22861320e-01 6.41687587e-02
3.08845818e-01 -5.10678172e-01 -2.88813025e-01 -1.15016671e-02
8.39589238e-01 1.37186974e-01 6.53549373e-01 -7.55183041e-01
-2.06872344e-01 5.81343651e-01 1.16103835e-01 -9.26635042e-02
1.26872957e+00 -1.73146576e-02 -7.53892809e-02 1.32311061e-01
1.51142430e+00 -1.88499853e-01 -1.21733427e+00 -1.77203283e-01
-2.36677155e-01 -5.79572320e-01 1.14920594e-01 -7.49852538e-01
-1.50552797e+00 1.20028377e+00 1.14603531e+00 1.23632848e-01
1.58853602e+00 -1.98464721e-01 4.10571575e-01 3.04304451e-01
1.46930695e-01 -6.83620632e-01 3.41731071e-01 -2.24713292e-02
1.05907571e+00 -1.25395894e+00 -8.08548331e-02 -1.79651927e-03
-2.48242453e-01 1.26389086e+00 4.60868657e-01 -3.98922056e-01
6.45197392e-01 1.82930470e-01 1.13665804e-01 -1.62982062e-01
-5.91156006e-01 -3.21717799e-01 4.75701720e-01 4.06601012e-01
2.94127882e-01 -2.57377803e-01 -2.09168643e-01 3.91374022e-01
-5.93059026e-02 2.23119155e-01 5.62967837e-01 5.71101725e-01
-3.19885045e-01 -9.80990410e-01 -2.46005714e-01 2.30386212e-01
-7.22365260e-01 -1.35813698e-01 5.83117902e-02 9.14510965e-01
3.82621646e-01 8.34626675e-01 1.79996699e-01 -3.99581283e-01
6.19654953e-01 -2.22799420e-01 4.26784188e-01 -2.01758638e-01
-9.09438252e-01 -2.85327405e-01 -3.43938142e-01 -1.05301762e+00
-6.31360173e-01 -3.31453890e-01 -6.67741835e-01 -4.30627108e-01
-4.23778109e-02 -2.06929043e-01 9.37523425e-01 6.79993093e-01
3.07650357e-01 7.74947643e-01 5.01263916e-01 -9.36007321e-01
-1.62982047e-01 -1.00427735e+00 -4.03767616e-01 5.60482085e-01
4.76357222e-01 -3.80571783e-01 -1.97401986e-01 6.13508336e-02] | [11.748741149902344, -1.7810512781143188] |
85628063-f9c8-4fd2-a222-cb0afb23d54f | iterative-residual-refinement-for-joint | 1904.0529 | null | http://arxiv.org/abs/1904.05290v1 | http://arxiv.org/pdf/1904.05290v1.pdf | Iterative Residual Refinement for Joint Optical Flow and Occlusion Estimation | Deep learning approaches to optical flow estimation have seen rapid progress
over the recent years. One common trait of many networks is that they refine an
initial flow estimate either through multiple stages or across the levels of a
coarse-to-fine representation. While leading to more accurate results, the
downside of this is an increased number of parameters. Taking inspiration from
both classical energy minimization approaches as well as residual networks, we
propose an iterative residual refinement (IRR) scheme based on weight sharing
that can be combined with several backbone networks. It reduces the number of
parameters, improves the accuracy, or even achieves both. Moreover, we show
that integrating occlusion prediction and bi-directional flow estimation into
our IRR scheme can further boost the accuracy. Our full network achieves
state-of-the-art results for both optical flow and occlusion estimation across
several standard datasets. | ['Junhwa Hur', 'Stefan Roth'] | 2019-04-10 | iterative-residual-refinement-for-joint-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Hur_Iterative_Residual_Refinement_for_Joint_Optical_Flow_and_Occlusion_Estimation_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Hur_Iterative_Residual_Refinement_for_Joint_Optical_Flow_and_Occlusion_Estimation_CVPR_2019_paper.pdf | cvpr-2019-6 | ['occlusion-estimation'] | ['computer-vision'] | [-1.01986438e-01 -2.91852683e-01 -2.09965810e-01 -3.32077742e-01
-3.48328650e-01 -4.13717508e-01 4.17676538e-01 -1.58047870e-01
-3.19503099e-01 8.67564023e-01 5.94284296e-01 6.35023266e-02
-3.39283608e-03 -7.39278197e-01 -4.08765793e-01 -3.98068935e-01
-1.49265109e-02 1.31584227e-01 5.38447261e-01 -1.41671821e-01
2.67209977e-01 6.86149895e-01 -1.36344051e+00 -3.93892415e-02
9.60184693e-01 1.02695203e+00 -2.16998667e-01 6.46048903e-01
-1.37514323e-01 1.09341669e+00 -2.91856676e-01 -5.38794100e-01
5.10982156e-01 -2.83796489e-01 -1.01289809e+00 -7.01731890e-02
8.52306545e-01 -5.51248848e-01 -6.49003208e-01 7.27303922e-01
4.81423557e-01 3.92284065e-01 2.61492044e-01 -1.16525555e+00
-2.67508775e-01 1.66336730e-01 -6.50118113e-01 2.51182288e-01
7.23388270e-02 2.50018030e-01 1.08281755e+00 -7.67315567e-01
8.48145783e-01 1.32039118e+00 8.24801385e-01 3.97996306e-01
-1.54377520e+00 -4.89296824e-01 2.89315104e-01 2.65945673e-01
-1.21093428e+00 -6.56335950e-01 7.63231695e-01 -4.58073229e-01
9.64345455e-01 -1.53949663e-01 8.76638353e-01 6.43716395e-01
8.28013346e-02 7.73873866e-01 7.46547341e-01 -6.90522045e-02
-1.08893542e-02 -2.09461123e-01 -2.35686675e-02 9.16468501e-01
2.52992898e-01 1.80998698e-01 -4.44270939e-01 5.56140766e-02
1.07403171e+00 -1.23870976e-01 -6.77615821e-01 -5.71971357e-01
-8.54299307e-01 9.18929517e-01 7.33470857e-01 1.44112542e-01
-1.82396978e-01 3.93768400e-01 4.48310792e-01 1.98398139e-02
6.21172369e-01 4.84230310e-01 -3.86934906e-01 -2.51091301e-01
-1.19091034e+00 4.53266740e-01 8.10720205e-01 4.50273961e-01
1.25605619e+00 5.97576723e-02 -2.22728983e-01 6.52946949e-01
2.88747877e-01 -2.27900892e-02 7.68101364e-02 -1.60584033e+00
4.22912836e-01 5.56423962e-01 1.92590222e-01 -1.33039820e+00
-5.64195693e-01 -6.75152123e-01 -7.31125534e-01 5.09371877e-01
6.88803852e-01 -3.33205104e-01 -8.24162424e-01 1.96744227e+00
2.93450803e-01 6.37039542e-01 -3.77365261e-01 1.02302158e+00
5.11281013e-01 4.63048011e-01 -8.94422978e-02 9.30025056e-03
8.06563854e-01 -1.44156933e+00 -6.16671920e-01 -2.62419879e-01
6.05164826e-01 -7.91131556e-01 5.96135855e-01 3.16873521e-01
-1.29465556e+00 -6.75916970e-01 -8.64572942e-01 -5.11895895e-01
-2.99347355e-03 -2.21540421e-01 8.42921078e-01 6.06815159e-01
-1.44805872e+00 9.36702549e-01 -8.57795835e-01 -8.91179070e-02
6.61522329e-01 4.73204404e-01 -2.91339457e-01 -1.12581916e-01
-8.46443594e-01 7.16752350e-01 -7.05074891e-02 2.24299863e-01
-4.46923614e-01 -1.20886469e+00 -1.01246095e+00 2.32950538e-01
3.53142470e-01 -1.23075330e+00 8.39938581e-01 -7.78998852e-01
-1.70330393e+00 3.65251809e-01 -4.96755928e-01 -4.65936005e-01
9.23877239e-01 -5.42439461e-01 -5.55914536e-04 2.69382238e-01
-1.29785210e-01 1.11702931e+00 8.04338157e-01 -1.02662802e+00
-6.46661103e-01 -9.87611488e-02 1.66635305e-01 -7.73295294e-03
-1.71477139e-01 -1.90678567e-01 -5.60529888e-01 -6.59122229e-01
-1.22498527e-01 -9.39470470e-01 -5.01076579e-01 4.27985013e-01
-2.29606003e-01 -1.76350817e-01 8.58285785e-01 -4.45839465e-01
1.18873918e+00 -1.91612792e+00 4.08604920e-01 7.66680911e-02
6.36951566e-01 4.04850334e-01 -2.00261354e-01 5.95873371e-02
7.20880702e-02 1.32750690e-01 -2.10026562e-01 -7.95560598e-01
-1.19390376e-01 1.35398895e-01 -1.22753039e-01 4.73107070e-01
3.89517576e-01 1.05301034e+00 -9.26506817e-01 -5.00803828e-01
6.16580367e-01 9.61474717e-01 -1.02200747e+00 1.21631935e-01
7.40693808e-02 6.71339333e-01 -1.01648413e-01 1.14233434e-01
7.64794528e-01 -5.83585858e-01 5.03523909e-02 -4.79331374e-01
-1.72629133e-01 5.37211776e-01 -1.46296394e+00 1.94746971e+00
-4.55492735e-01 8.63056540e-01 1.68969467e-01 -8.98046792e-01
5.40917933e-01 2.44041800e-01 8.43641937e-01 -3.74547273e-01
5.73902242e-02 8.42913538e-02 -9.12571326e-02 -2.19248161e-01
6.64486110e-01 5.17808013e-02 4.82837975e-01 3.66661757e-01
1.52307153e-01 9.73048657e-02 4.90945548e-01 1.94280088e-01
1.03858960e+00 5.54576516e-01 -3.01166661e-02 -1.92093849e-01
7.35714972e-01 -2.40320101e-01 9.83428597e-01 5.73557198e-01
-4.95632291e-01 7.86042869e-01 5.33042669e-01 -7.96796024e-01
-8.29170704e-01 -8.55824232e-01 -1.54646128e-01 7.47056305e-01
2.15396762e-01 -6.62374616e-01 -5.39403439e-01 -6.36160016e-01
1.05437897e-01 -2.88929157e-02 -6.13684595e-01 1.91935539e-01
-1.02630591e+00 -5.51682055e-01 3.00999999e-01 6.53071105e-01
6.26672924e-01 -9.10567224e-01 -5.81298470e-01 4.35270876e-01
-3.47946048e-01 -1.44374061e+00 -5.48576176e-01 -2.71967173e-01
-8.73344302e-01 -9.70766127e-01 -6.41715527e-01 -4.89783823e-01
4.72398669e-01 2.43606776e-01 1.44258416e+00 3.20164859e-01
-3.43017071e-01 7.39910305e-02 1.83927882e-02 1.80223957e-01
2.33879853e-02 3.22186768e-01 -2.31940970e-01 1.80158079e-01
-1.25808597e-01 -7.73605645e-01 -1.07556248e+00 1.73603222e-01
-6.09840810e-01 -7.08701462e-02 1.71637371e-01 5.42422712e-01
1.90440819e-01 -2.43886605e-01 4.01866406e-01 -7.02511013e-01
2.36962721e-01 -5.89352436e-02 -6.33722067e-01 -1.51367322e-01
-6.24263167e-01 3.00386161e-01 5.93185782e-01 -1.56026900e-01
-1.02124858e+00 -7.97253028e-02 -2.29640469e-01 -4.98456419e-01
-4.20160182e-02 9.19455662e-02 3.27824920e-01 -5.97299039e-01
3.91566753e-01 -2.96824992e-01 3.62174469e-03 -3.65183800e-01
4.25922364e-01 -8.76037925e-02 3.77128363e-01 -3.89092356e-01
8.92446101e-01 7.16419458e-01 4.19720352e-01 -5.80533326e-01
-1.00916851e+00 -3.18049461e-01 -7.31936395e-01 -2.72706598e-01
8.16751599e-01 -7.81183302e-01 -1.00651872e+00 5.11043727e-01
-1.22384512e+00 -4.81421351e-01 -2.21488595e-01 4.62347716e-01
-3.27114135e-01 6.16768897e-01 -9.58483815e-01 -4.72750813e-01
-2.96842188e-01 -1.43532550e+00 8.82820249e-01 3.04697663e-01
-3.26382995e-01 -1.30373144e+00 2.16280788e-01 3.18401843e-01
8.27281535e-01 2.91139126e-01 5.97724497e-01 2.19433784e-01
-9.51645195e-01 1.41500592e-01 -5.72276354e-01 3.09491158e-01
9.88325328e-02 1.93462864e-01 -8.77474546e-01 -3.53607386e-01
-2.83415407e-01 -3.30295861e-01 1.17434657e+00 7.33390510e-01
9.74296629e-01 -6.26499355e-02 -2.44295537e-01 1.05597317e+00
1.55403960e+00 -2.69527256e-01 8.15339029e-01 1.48418829e-01
1.02325904e+00 6.53991222e-01 7.66025186e-02 3.19463968e-01
5.22792101e-01 8.49833488e-01 4.38131243e-01 -2.41037712e-01
-5.79294562e-01 -2.31291703e-03 7.41770789e-02 4.52968657e-01
-4.12372828e-01 -1.01420388e-01 -5.80979109e-01 5.76227605e-01
-1.97534668e+00 -1.00008106e+00 -1.50607884e-01 2.04284334e+00
5.93333244e-01 6.52052984e-02 3.23976159e-01 1.28075093e-01
4.97430921e-01 5.53165019e-01 -4.15095717e-01 -2.85148472e-01
7.13895336e-02 3.90164852e-01 3.60398591e-01 9.74462926e-01
-9.22186017e-01 1.02130735e+00 6.90697241e+00 4.15534079e-01
-1.32049918e+00 -6.50874600e-02 6.43714905e-01 -2.55313188e-01
-1.67086139e-01 8.96407664e-02 -7.84493625e-01 3.53449047e-01
4.25076693e-01 1.18408270e-01 5.95000148e-01 3.29224169e-01
2.96798706e-01 -8.66908655e-02 -9.28073525e-01 8.68355453e-01
-1.36331558e-01 -1.76249468e+00 8.02364722e-02 2.06876859e-01
9.01715517e-01 2.77638644e-01 -1.95343018e-01 -2.93462463e-02
3.73151749e-01 -9.65848386e-01 4.47874039e-01 4.25884485e-01
5.23969054e-01 -7.56267309e-01 6.62384570e-01 -9.89592075e-02
-1.54358530e+00 1.14702024e-02 -1.52802512e-01 -3.09427023e-01
6.02469265e-01 6.36525333e-01 -9.30485427e-02 5.01707256e-01
7.03401566e-01 1.15251279e+00 -4.44265872e-01 1.29838622e+00
-2.58956283e-01 3.56542349e-01 -2.60974318e-01 4.34793472e-01
3.18201363e-01 -3.87979865e-01 5.43658435e-01 1.19978988e+00
-1.61992591e-02 -2.42355466e-01 3.31126511e-01 9.89893436e-01
-2.33685195e-01 8.90383311e-03 -2.34459132e-01 5.41704416e-01
3.61708760e-01 1.46584725e+00 -5.48103929e-01 -3.15860927e-01
-5.25562763e-01 7.27618635e-01 7.27746844e-01 4.89407510e-01
-7.04110026e-01 -3.72157037e-01 1.16397119e+00 2.01695815e-01
4.91635561e-01 -3.63877982e-01 -2.59849101e-01 -1.24111438e+00
1.06127448e-01 -3.09586734e-01 1.26843452e-01 -5.03399074e-01
-1.01436841e+00 5.35475671e-01 -4.11641151e-01 -1.00785804e+00
-2.20208615e-01 -3.76652449e-01 -6.53270900e-01 8.70502174e-01
-2.12180591e+00 -8.81054461e-01 -5.93460083e-01 5.57280064e-01
3.88313055e-01 3.60861719e-01 4.95531648e-01 7.20528424e-01
-7.80780852e-01 4.42035019e-01 -4.91433233e-01 3.86471450e-01
7.86487460e-01 -1.14459658e+00 4.99845326e-01 1.15510690e+00
-1.10340612e-02 5.06810367e-01 5.17443955e-01 -2.87269831e-01
-8.88538420e-01 -9.88000453e-01 9.86110330e-01 -3.68583143e-01
6.18925810e-01 -1.54563427e-01 -9.44083869e-01 6.69314981e-01
2.07093030e-01 6.02582693e-01 1.57831803e-01 1.60552204e-01
-3.47363800e-01 -2.19660044e-01 -9.10417438e-01 6.05596781e-01
1.19799411e+00 -4.15697008e-01 -1.72193255e-02 -9.91931111e-02
6.31887496e-01 -5.30457735e-01 -9.11867678e-01 4.58881527e-01
6.65810704e-01 -1.47567809e+00 1.09368694e+00 -4.02821094e-01
5.70375502e-01 -4.18745518e-01 2.82131225e-01 -1.19963574e+00
-5.95148146e-01 -1.13391018e+00 -3.70958447e-01 1.04330790e+00
2.46217713e-01 -5.99515855e-01 1.08489704e+00 6.68521404e-01
-6.14159815e-02 -7.87746727e-01 -7.62720048e-01 -5.45654297e-01
2.30594277e-02 -4.22171324e-01 4.02457386e-01 7.64160275e-01
-2.96342403e-01 4.65474606e-01 -6.20319545e-01 -9.87437516e-02
7.51523256e-01 6.49125576e-02 9.33763862e-01 -1.42423570e+00
-2.21675843e-01 -7.42439508e-01 -5.50013185e-01 -1.52306151e+00
3.13758910e-01 -7.26510465e-01 -1.35994047e-01 -1.62811244e+00
-1.42249942e-01 -5.39802670e-01 -2.75805354e-01 3.27272058e-01
-4.35346246e-01 6.51335239e-01 4.32628572e-01 8.85762796e-02
-6.19044542e-01 4.67662483e-01 1.51113224e+00 2.79235780e-01
-5.22465408e-01 -1.63548425e-01 -4.95577216e-01 6.86030447e-01
6.19834840e-01 -3.16425979e-01 -3.40886891e-01 -7.22451866e-01
2.91237414e-01 4.96097766e-02 5.19340992e-01 -1.10412025e+00
2.34919220e-01 2.12388888e-01 3.11694294e-01 -2.80852377e-01
3.88827085e-01 -6.28959954e-01 -9.05557871e-02 4.57107306e-01
-1.69733241e-01 1.08128376e-01 1.46618322e-01 5.11577070e-01
-1.95215642e-01 1.47437990e-01 1.04642856e+00 9.03193094e-03
-6.51802301e-01 5.31126916e-01 -3.56425606e-02 3.35325897e-01
5.20774007e-01 -3.23839456e-01 -5.56019723e-01 -4.80414212e-01
-5.48751354e-01 1.41418457e-01 4.50848520e-01 4.47322518e-01
2.67702967e-01 -1.17686689e+00 -6.69240177e-01 1.80519179e-01
-3.80231112e-01 4.25569825e-02 8.69557709e-02 1.18809330e+00
-6.83896840e-01 3.75368506e-01 -2.23354459e-01 -7.03928649e-01
-8.94627512e-01 2.03116491e-01 6.19784832e-01 -5.93089700e-01
-9.16382372e-01 7.65872955e-01 1.44851014e-01 -1.82774395e-01
2.36387506e-01 -1.91891506e-01 -2.54036307e-01 -5.85645530e-03
5.46721339e-01 7.65365064e-01 -1.11147910e-01 -7.40896523e-01
-3.55283886e-01 1.08729887e+00 -7.12522818e-03 9.27275568e-02
1.40769124e+00 -3.14154208e-01 -2.13871911e-01 -1.52237630e-02
1.33410716e+00 9.47636589e-02 -1.86714184e+00 -2.43808776e-01
-2.08563447e-01 -7.11579561e-01 3.70318562e-01 -5.55532455e-01
-1.76931894e+00 8.38715851e-01 3.96110594e-01 6.97742328e-02
9.29469824e-01 -3.50884527e-01 1.16402566e+00 -3.63040180e-03
1.40909284e-01 -8.19254518e-01 2.89400145e-02 5.10901809e-01
3.63043964e-01 -1.30369377e+00 6.87573552e-02 -7.05696762e-01
-1.95773199e-01 1.17634153e+00 5.54254234e-01 -4.30180877e-01
7.31687129e-01 3.44798982e-01 1.36148468e-01 9.45833884e-03
-6.81007266e-01 -2.97726214e-01 3.68375808e-01 3.93734723e-01
6.72203839e-01 -4.22786832e-01 -1.24833055e-01 -3.23360950e-01
-1.91989522e-02 2.95153290e-01 5.16155183e-01 6.30147338e-01
-2.28919357e-01 -1.24917781e+00 8.17245767e-02 2.23049194e-01
-6.31423831e-01 -3.44625451e-02 9.39905867e-02 7.38839567e-01
8.51242617e-03 1.06409562e+00 2.16271862e-01 -7.87567347e-02
2.01874450e-01 -2.47215569e-01 5.94917417e-01 -1.58391282e-01
-6.49513125e-01 -7.12341592e-02 9.08878539e-03 -1.11984038e+00
-7.61501253e-01 -5.07432997e-01 -1.10466695e+00 -6.97036684e-01
-1.80690944e-01 -7.85752535e-02 2.37520933e-01 1.03891194e+00
6.87456667e-01 5.92045605e-01 4.94205087e-01 -1.22899616e+00
1.11090437e-01 -7.61374474e-01 -1.22257620e-01 4.16095167e-01
7.66418934e-01 -6.62475049e-01 -5.35024941e-01 -8.22702274e-02] | [8.81418514251709, -1.7429338693618774] |
72d166c2-c8fb-46f2-a8d6-96aea1485add | initial-explorations-of-ccg-supertagging-for | null | null | https://aclanthology.org/K17-3023 | https://aclanthology.org/K17-3023.pdf | Initial Explorations of CCG Supertagging for Universal Dependency Parsing | In this paper we describe the system by METU team for universal dependency parsing of multilingual text. We use a neural network-based dependency parser that has a greedy transition approach to dependency parsing. CCG supertags contain rich structural information that proves useful in certain NLP tasks. We experiment with CCG supertags as additional features in our experiments. The neural network parser is trained together with dependencies and simplified CCG tags as well as other features provided. | ['Ruket Cakici', 'Heval Azizoglu', 'Burak Kerim Akkus'] | 2017-08-01 | null | null | null | conll-2017-8 | ['ccg-supertagging'] | ['natural-language-processing'] | [-6.95381880e-01 4.42746937e-01 -3.11816394e-01 -9.83142436e-01
-7.05910265e-01 -6.98068500e-01 3.95067990e-01 3.77423465e-01
-6.99692488e-01 1.13615811e+00 8.32329214e-01 -8.59146535e-01
3.64921033e-01 -6.45152450e-01 -3.77179474e-01 -3.70111525e-01
-7.28041291e-01 6.61557257e-01 2.87943929e-01 -7.72987902e-01
-1.85672373e-01 3.61248940e-01 -6.33649230e-01 2.47725591e-01
4.47751492e-01 2.11573437e-01 4.64901626e-01 1.01307809e+00
-3.85358393e-01 9.13325667e-01 -6.29959166e-01 -4.51107860e-01
3.86966392e-02 -3.89525853e-02 -1.27557385e+00 -8.79742742e-01
4.75875251e-02 3.07767969e-02 -1.26505941e-01 7.70192981e-01
4.36976373e-01 -6.84440732e-02 2.31795236e-01 -6.08503342e-01
-3.92953217e-01 1.59708226e+00 -1.48231223e-01 6.41902864e-01
4.77091551e-01 -6.39030457e-01 1.39972746e+00 -5.29726863e-01
9.48003352e-01 1.46592164e+00 7.60660827e-01 5.04528701e-01
-6.84080720e-01 -4.84698176e-01 2.44891837e-01 -1.95579752e-01
-7.03096569e-01 -1.29762605e-01 5.65676987e-01 -9.24328417e-02
1.90129459e+00 -1.71497479e-01 2.30723366e-01 9.18165565e-01
6.75291061e-01 6.36321723e-01 1.04579222e+00 -1.02007496e+00
-3.28070998e-01 -2.77934134e-01 9.67993140e-01 8.27681601e-01
-6.08315021e-02 -6.35274425e-02 -2.73607433e-01 1.12165652e-01
6.56150877e-01 -6.29378140e-01 2.47547075e-01 5.36510766e-01
-9.52995241e-01 1.33735228e+00 4.46704209e-01 8.34102273e-01
-1.70758516e-01 2.46852309e-01 5.79820096e-01 6.35234237e-01
7.28937745e-01 1.99866772e-01 -1.41718996e+00 4.06521000e-02
-2.07500190e-01 -1.49368703e-01 1.34742975e+00 1.08607018e+00
8.51609468e-01 -9.01771635e-02 2.35411599e-01 1.03674924e+00
2.56333530e-01 1.48502380e-01 3.30254614e-01 -7.33067811e-01
8.37681413e-01 3.14235032e-01 -2.18371257e-01 -3.38414878e-01
-1.08220947e+00 -9.76899266e-02 -4.24470484e-01 -2.54323155e-01
3.72655034e-01 -8.37599277e-01 -1.12670612e+00 1.75787735e+00
1.79098144e-01 -6.39289260e-01 4.81613040e-01 4.20559138e-01
1.22393584e+00 5.32756209e-01 5.63330233e-01 -5.19985594e-02
1.38073766e+00 -8.99800241e-01 -8.39999855e-01 -3.55583221e-01
1.52321470e+00 -1.02412891e+00 6.08495891e-01 -6.21093065e-02
-8.00074220e-01 -3.01128060e-01 -8.07474613e-01 -2.62504339e-01
-7.49250948e-01 -1.01741210e-01 1.23044538e+00 5.49342513e-01
-1.48451245e+00 7.86211669e-01 -9.32196677e-01 -7.23629773e-01
-3.48314226e-01 8.17566097e-01 -8.85615051e-01 -7.67099932e-02
-1.69350970e+00 1.21640515e+00 1.05854559e+00 6.84484616e-02
-1.55464575e-01 2.10731328e-01 -1.32950294e+00 -3.09185773e-01
4.35802527e-02 -1.49142966e-01 1.58031845e+00 -6.17839277e-01
-1.34303331e+00 9.63504493e-01 1.18399382e-01 -4.92875904e-01
-1.77717790e-01 -3.68753225e-01 -5.99655211e-01 -1.33331493e-01
2.86407828e-01 6.69253290e-01 6.49706423e-02 -6.10924900e-01
-1.07498014e+00 -3.25797200e-01 4.15558249e-01 1.38323769e-01
1.51992112e-01 9.33377802e-01 -4.50809836e-01 -4.94942546e-01
3.34807158e-01 -9.76746082e-01 -5.89357853e-01 -1.48520458e+00
-7.08303273e-01 -8.28712046e-01 6.36958122e-01 -1.07381368e+00
1.12905359e+00 -1.54666078e+00 -2.92686559e-02 2.84200963e-02
-2.12125778e-01 -4.54895310e-02 -4.13039595e-01 7.96893775e-01
-4.76414561e-01 3.63062322e-01 -9.29349102e-03 -4.26049083e-01
-1.38147920e-01 1.03944111e+00 3.48305345e-01 8.10701549e-02
2.75017619e-01 8.65761340e-01 -9.35102105e-01 -6.71854734e-01
-6.46979511e-02 3.29162985e-01 -2.14972913e-01 3.85402888e-01
-1.90313160e-01 5.38528681e-01 -5.71025789e-01 8.24881196e-01
5.20827353e-01 3.69436353e-01 7.89140224e-01 -5.49939321e-03
-4.38698947e-01 9.33006823e-01 -4.21928436e-01 1.92961204e+00
-7.79029787e-01 5.63018739e-01 1.24537237e-01 -8.45767200e-01
8.78694892e-01 5.38000464e-01 -1.17944434e-01 -4.43494856e-01
4.80332315e-01 3.08968842e-01 2.68629909e-01 -5.75871408e-01
6.28838658e-01 -5.78882731e-02 -8.43080282e-01 6.17569864e-01
6.92388892e-01 4.10002880e-02 6.87744319e-01 5.00367343e-01
1.19092143e+00 4.25736725e-01 5.20628512e-01 -6.90167964e-01
2.35689864e-01 1.90168843e-01 6.15298629e-01 5.48532903e-01
2.53360607e-02 2.67885536e-01 7.06826210e-01 -7.39053845e-01
-8.55691671e-01 -4.87309903e-01 -8.47685933e-02 1.77242625e+00
-6.51817620e-01 -5.45974791e-01 -8.12609196e-01 -1.28049886e+00
-4.95821744e-01 5.22928357e-01 -8.49010408e-01 6.60429239e-01
-1.36692500e+00 -9.73220646e-01 6.00292027e-01 9.57818449e-01
1.43605798e-01 -1.47478962e+00 -1.31694168e-01 6.91556573e-01
-1.79264829e-01 -1.63625836e+00 -3.54289003e-02 1.33697295e+00
-9.29312885e-01 -1.04037786e+00 -8.51036757e-02 -1.40668559e+00
3.78246874e-01 -2.28974193e-01 1.71689034e+00 3.10318649e-01
1.15192696e-01 -2.81156451e-01 -9.80679154e-01 -4.17967737e-01
-7.77284801e-01 7.76195347e-01 -1.39316514e-01 -1.14338958e+00
5.41275203e-01 -6.67873681e-01 1.00739442e-01 -2.04828948e-01
-5.66405952e-01 -2.29700014e-01 5.49088955e-01 6.77865922e-01
3.40247191e-02 -5.86811006e-01 2.04719961e-01 -1.50790620e+00
7.82663107e-01 -4.51582670e-01 -4.15671170e-01 -2.23443415e-02
3.09602134e-02 3.93314004e-01 5.23779392e-01 2.09708989e-01
-1.13283420e+00 1.88503504e-01 -9.42232251e-01 5.27130067e-01
-4.01022494e-01 1.09522569e+00 -2.42342502e-01 1.72547191e-01
4.51968312e-01 -6.23340786e-01 -6.68986261e-01 -9.83434975e-01
5.08072376e-01 4.76289034e-01 5.61208367e-01 -8.50268543e-01
3.44937772e-01 -2.75670201e-01 -2.00521145e-02 -6.64959192e-01
-1.03999758e+00 -3.65185678e-01 -1.40013433e+00 3.91088545e-01
1.08227170e+00 -9.16191995e-01 -2.27665991e-01 3.09270829e-01
-1.48687637e+00 -5.52192032e-01 3.70685190e-01 6.80993974e-01
-6.05125464e-02 2.26477236e-01 -1.47938073e+00 -3.79226357e-01
-4.23943520e-01 -7.08748519e-01 5.97785413e-01 2.26866424e-01
-1.47833809e-01 -1.54331100e+00 5.78675508e-01 -3.16925168e-01
1.33780256e-01 3.91464561e-01 9.36849952e-01 -1.25264406e+00
-5.49029792e-03 -1.66772187e-01 -7.88823143e-02 4.23263639e-01
-5.20744286e-02 4.98755313e-02 -7.09050059e-01 -1.37302905e-01
-4.89831835e-01 -4.37531292e-01 8.96058440e-01 5.69135427e-01
1.12216741e-01 3.61757837e-02 -3.02298486e-01 5.64138055e-01
1.45726168e+00 1.84945002e-01 1.38170779e-01 6.21828735e-01
8.76411378e-01 7.90398359e-01 5.74450672e-01 -1.87130757e-02
7.17928946e-01 2.89985798e-02 4.26261663e-01 -1.42585918e-01
1.94996580e-01 7.88559988e-02 2.60464609e-01 1.40587187e+00
-3.74766916e-01 -3.71749669e-01 -1.24713814e+00 4.63161916e-01
-1.72587538e+00 -8.35027099e-02 -7.02279329e-01 1.31550920e+00
9.17356312e-01 3.44863087e-01 -1.68168098e-01 -2.06775069e-01
7.76179194e-01 1.45707652e-01 4.77297157e-01 -1.19266820e+00
-2.91828543e-01 7.56847799e-01 7.08874166e-01 8.51423860e-01
-1.45795643e+00 1.77575159e+00 7.24902153e+00 1.02638528e-01
-7.90862560e-01 4.89769995e-01 1.65379822e-01 5.03423452e-01
2.89481208e-02 2.74916649e-01 -1.20243692e+00 -1.14383332e-01
1.41397381e+00 5.59565485e-01 -1.33877948e-01 6.91356778e-01
-3.41877997e-01 -3.18379521e-01 -9.33777630e-01 8.67709294e-02
-3.27086031e-01 -9.86900508e-01 -4.25492465e-01 -2.25804776e-01
3.02278727e-01 1.03613830e+00 -6.51556134e-01 4.55844045e-01
1.46321213e+00 -7.93245912e-01 2.66292214e-01 -3.56586218e-01
7.20292568e-01 -1.15964210e+00 1.33297098e+00 1.06250480e-01
-1.11554086e+00 1.98586956e-01 -5.65024257e-01 -4.46164012e-01
2.74713337e-01 4.72576916e-01 -8.84581089e-01 8.36396217e-01
8.30009282e-01 6.35150194e-01 -5.63088894e-01 2.90627301e-01
-1.11509275e+00 8.17508578e-01 -5.34427285e-01 -5.06158210e-02
8.06039512e-01 -1.15381934e-01 2.67053992e-01 2.04127145e+00
-8.68893042e-02 8.17903504e-02 2.70412713e-01 -2.91072190e-01
-1.07247978e-01 4.46763575e-01 -5.95061541e-01 3.85054983e-02
4.27403420e-01 1.47587800e+00 -1.04019856e+00 -3.50829989e-01
-6.40022695e-01 7.05974400e-01 8.87546122e-01 1.66015968e-01
-3.73293698e-01 -5.72849452e-01 3.85755628e-01 -5.20188034e-01
2.50914931e-01 -5.33349872e-01 3.37447077e-01 -1.17734587e+00
-4.13271815e-01 -3.46460760e-01 9.77717102e-01 -5.41099370e-01
-1.06064844e+00 1.38517261e+00 1.52931541e-01 -4.55967426e-01
-7.67516613e-01 -9.79901135e-01 -6.79015160e-01 1.13784945e+00
-1.64486754e+00 -1.60594583e+00 3.87982249e-01 6.56539142e-01
4.11867797e-01 -1.48704454e-01 1.39916432e+00 6.22215606e-02
-7.40624011e-01 2.46110097e-01 -1.83250591e-01 9.44808424e-01
8.42375040e-01 -1.64005637e+00 1.15956223e+00 9.27418768e-01
1.73495129e-01 8.15755427e-01 7.66003728e-01 -7.12890327e-01
-8.32907438e-01 -9.52966809e-01 1.82285833e+00 -5.47463834e-01
1.04363453e+00 -4.32361245e-01 -6.98804855e-01 1.37857437e+00
9.13099170e-01 9.95774120e-02 7.80592740e-01 8.08258533e-01
-4.84529585e-01 4.77041185e-01 -7.91965365e-01 -3.66712287e-02
9.21627104e-01 -4.63609636e-01 -9.13079739e-01 4.44638461e-01
1.00212288e+00 -6.59009218e-01 -9.94089842e-01 2.15490222e-01
3.75615925e-01 -7.94746578e-01 4.86773044e-01 -8.53064060e-01
4.21363831e-01 2.28123412e-01 -2.96061665e-01 -1.64153862e+00
-3.39061469e-01 -5.06875634e-01 4.17510033e-01 1.21570551e+00
1.02542889e+00 -7.54256904e-01 5.61058819e-01 4.25468564e-01
-6.09994173e-01 -2.04683408e-01 -8.75422657e-01 -5.72631776e-01
5.20072162e-01 -5.52627265e-01 4.04121399e-01 1.19409144e+00
3.99270952e-01 8.87763083e-01 -1.67272836e-01 5.86260445e-02
2.77224451e-01 -1.12994820e-01 3.65445107e-01 -1.47933972e+00
-1.96744695e-01 6.06781617e-02 4.81687076e-02 -4.74211395e-01
5.69032907e-01 -8.56932163e-01 2.31419951e-01 -1.80096531e+00
-1.23279251e-01 -4.80997413e-01 -3.02414656e-01 8.58170033e-01
-1.36725038e-01 1.28902733e-01 1.21739037e-01 3.82478605e-03
-3.55457306e-01 -4.58314389e-01 7.77836382e-01 2.52719969e-01
-2.72006392e-01 -2.40804449e-01 -4.71321553e-01 9.60163653e-01
1.11732996e+00 -1.06236017e+00 1.80147901e-01 -7.54688442e-01
4.83216614e-01 3.22720230e-01 -5.25334179e-01 -5.76726019e-01
-4.45492081e-02 1.44392550e-01 2.09744930e-01 -8.55300486e-01
-8.72320309e-02 -4.67548490e-01 -3.48334819e-01 4.13480431e-01
5.22605553e-02 5.97509861e-01 1.56230435e-01 -1.87911049e-01
-3.94985557e-01 -5.49388826e-01 3.95183891e-01 -6.75103903e-01
-8.31908762e-01 6.91791400e-02 -5.54924607e-01 -9.22029391e-02
1.68899521e-01 3.57156932e-01 -1.89277649e-01 -1.00950174e-01
-1.20350707e+00 3.46811384e-01 1.46636635e-01 5.16152620e-01
-6.07536100e-02 -8.36832643e-01 -8.66798162e-01 2.48137508e-02
-1.16684422e-01 -1.36915103e-01 -3.51084203e-01 4.83285367e-01
-8.12539101e-01 6.74621761e-01 -3.97791058e-01 -1.47454113e-01
-1.37238657e+00 2.59926587e-01 2.14215238e-02 -8.27789605e-01
-5.31290233e-01 1.18366742e+00 -2.31968582e-01 -9.56943870e-01
-4.63356636e-02 -7.29181886e-01 -8.17620456e-01 2.41530687e-01
3.06516111e-01 -3.64440709e-01 3.51740777e-01 -8.39873493e-01
-5.02851844e-01 1.84782311e-01 -3.72045726e-01 -2.09501266e-01
1.62466526e+00 -6.52280301e-02 -6.41412377e-01 2.63926685e-01
1.29124689e+00 7.09186643e-02 -5.59284389e-01 -1.68891490e-01
6.64828956e-01 2.86276251e-01 -1.09523751e-01 -9.48229909e-01
-9.97330010e-01 7.68143594e-01 6.77649230e-02 3.96363705e-01
6.94788218e-01 5.16620219e-01 7.43888676e-01 7.95510232e-01
5.16253114e-01 -1.23810291e+00 -6.15685940e-01 1.57424641e+00
5.18929124e-01 -1.18648660e+00 -1.82178780e-01 -4.55575615e-01
-3.93374354e-01 1.40801692e+00 5.30140340e-01 -4.08621430e-01
8.61852765e-01 8.34995329e-01 8.73652756e-01 -2.65736789e-01
-6.58348799e-01 -6.93197727e-01 -3.75102997e-01 7.77850747e-01
1.19046569e+00 4.58197653e-01 -9.44608033e-01 6.09270215e-01
-4.89411443e-01 -5.06383061e-01 6.11915112e-01 1.34707904e+00
-4.72206205e-01 -1.87286901e+00 -2.13444874e-01 4.41409759e-02
-1.15252709e+00 -7.57637322e-01 -3.87434155e-01 1.43347108e+00
-7.36184716e-02 9.29848611e-01 3.71768586e-02 -3.08603138e-01
2.17398554e-01 1.64462164e-01 5.35492897e-01 -1.02180934e+00
-1.10036504e+00 3.13465655e-01 1.18327320e+00 -5.16908109e-01
-7.61328459e-01 -5.90442657e-01 -1.63938439e+00 3.54491621e-02
-3.44373316e-01 6.52970731e-01 1.01643324e+00 1.19000065e+00
-2.36929446e-01 2.90087849e-01 4.02057618e-01 -9.06545520e-01
1.97555870e-01 -1.60402179e+00 -5.52888155e-01 3.93803194e-02
1.23816237e-01 -2.94745266e-01 -3.61247100e-02 -2.66403444e-02] | [10.355940818786621, 9.841753959655762] |
30637dcd-36b2-4d34-b0e3-c4087236aab6 | lit-former-linking-in-plane-and-through-plane | 2302.1063 | null | https://arxiv.org/abs/2302.10630v1 | https://arxiv.org/pdf/2302.10630v1.pdf | LIT-Former: Linking In-plane and Through-plane Transformers for Simultaneous CT Image Denoising and Deblurring | This paper studies 3D low-dose computed tomography (CT) imaging. Although various deep learning methods were developed in this context, typically they perform denoising due to low-dose and deblurring for super-resolution separately. Up to date, little work was done for simultaneous in-plane denoising and through-plane deblurring, which is important to improve clinical CT images. For this task, a straightforward method is to directly train an end-to-end 3D network. However, it demands much more training data and expensive computational costs. Here, we propose to link in-plane and through-plane transformers for simultaneous in-plane denoising and through-plane deblurring, termed as LIT-Former, which can efficiently synergize in-plane and through-plane sub-tasks for 3D CT imaging and enjoy the advantages of both convolution and transformer networks. LIT-Former has two novel designs: efficient multi-head self-attention modules (eMSM) and efficient convolutional feed-forward networks (eCFN). First, eMSM integrates in-plane 2D self-attention and through-plane 1D self-attention to efficiently capture global interactions of 3D self-attention, the core unit of transformer networks. Second, eCFN integrates 2D convolution and 1D convolution to extract local information of 3D convolution in the same fashion. As a result, the proposed LIT-Former synergizes these two sub-tasks, significantly reducing the computational complexity as compared to 3D counterparts and enabling rapid convergence. Extensive experimental results on simulated and clinical datasets demonstrate superior performance over state-of-the-art models. | ['Hongming Shan', 'Ge Wang', 'Chuang Niu', 'Zhihao Chen'] | 2023-02-21 | null | null | null | null | ['deblurring'] | ['computer-vision'] | [-1.72067985e-01 2.08419621e-01 2.36996114e-01 -3.84899080e-01
-1.05823052e+00 -8.43877643e-02 2.31761932e-01 -2.87486672e-01
-3.61244053e-01 3.73366356e-01 5.81529021e-01 -1.81493908e-01
-2.26375714e-01 -7.22906351e-01 -6.51474357e-01 -1.01287639e+00
-5.03865331e-02 3.78406674e-01 3.69171590e-01 -2.78746709e-02
-3.59339833e-01 6.13335252e-01 -7.81998217e-01 4.09126371e-01
8.37094486e-01 8.82951856e-01 2.95961291e-01 5.53459525e-01
6.59418702e-02 8.25284302e-01 -8.27155858e-02 -7.76567236e-02
-6.38589710e-02 -3.13844115e-01 -7.62978613e-01 -1.55742466e-01
1.84597746e-01 -7.14388072e-01 -7.17080891e-01 1.03364503e+00
1.11638427e+00 -5.74866496e-02 5.82050800e-01 -4.27985400e-01
-5.03157079e-01 3.72332275e-01 -6.00513577e-01 6.58459246e-01
1.74222682e-02 2.88528532e-01 2.12388054e-01 -8.96762431e-01
1.92170411e-01 9.56507146e-01 9.21946526e-01 5.43912590e-01
-9.86334741e-01 -6.28599048e-01 -4.79044616e-02 8.99766684e-02
-1.01579452e+00 -1.84512421e-01 8.97169113e-01 -2.91896045e-01
1.07769728e+00 9.12737250e-02 6.29985332e-01 8.48529100e-01
5.33779740e-01 7.81263888e-01 1.14954960e+00 -1.70940373e-04
-3.96799482e-02 -3.33379596e-01 2.57501490e-02 5.95820904e-01
-3.86934169e-02 2.81007767e-01 -2.11033091e-01 5.24562076e-02
1.41531444e+00 2.19552010e-01 -7.60616422e-01 -4.22219597e-02
-9.72263753e-01 5.88431835e-01 9.55236435e-01 5.68357289e-01
-8.55409563e-01 2.23089337e-01 5.12845039e-01 -4.62474935e-02
8.36005151e-01 1.28589675e-01 -1.76650763e-01 1.32841587e-01
-9.51836944e-01 -1.35652861e-02 2.95743525e-01 6.13722682e-01
2.11763039e-01 2.85301954e-02 -3.72576594e-01 7.43784785e-01
1.77032143e-01 2.89981991e-01 7.77139425e-01 -4.20121253e-01
2.91027516e-01 2.48497531e-01 -2.09126443e-01 -6.34770513e-01
-8.86975229e-01 -9.30937707e-01 -1.45010304e+00 -6.22328334e-02
1.57200787e-02 -1.18430980e-01 -1.25778711e+00 1.74042368e+00
5.63402116e-01 4.62821364e-01 -2.20349193e-01 1.31709826e+00
1.23503292e+00 4.26635832e-01 1.00131899e-01 -3.18911940e-01
1.59766948e+00 -1.03423131e+00 -9.10597205e-01 -2.49813959e-01
4.68546480e-01 -7.30356872e-01 6.99542940e-01 7.73475915e-02
-1.57197809e+00 -4.08058941e-01 -9.52903926e-01 -2.11603716e-01
1.39108658e-01 -2.33536690e-01 4.76178378e-01 3.98572534e-01
-1.09264088e+00 7.03683555e-01 -1.23856258e+00 1.33001208e-01
7.57884622e-01 5.98305881e-01 -2.45249212e-01 -2.76456028e-01
-1.28551280e+00 9.58094776e-01 -1.39292389e-01 3.36241782e-01
-1.05282128e+00 -1.33163798e+00 -7.08865643e-01 1.52595460e-01
1.91815704e-01 -1.11374772e+00 1.47923338e+00 -4.35328782e-01
-1.56672359e+00 7.39335597e-01 -8.29881579e-02 -2.56989270e-01
5.68111718e-01 -3.29670846e-01 -2.88262218e-01 3.02570105e-01
1.58576846e-01 2.11499661e-01 8.48737240e-01 -9.45872128e-01
-1.42540745e-02 -7.58787930e-01 -3.02646339e-01 3.50625455e-01
-5.17942831e-02 -2.22096089e-02 -5.80482185e-01 -9.05786514e-01
6.40505850e-01 -4.65234905e-01 -3.66057783e-01 -4.70533147e-02
-3.22560668e-01 1.32684827e-01 8.09090376e-01 -8.38303328e-01
1.08784604e+00 -1.91696477e+00 6.58909827e-02 -5.34283184e-02
7.84413636e-01 2.62436956e-01 6.93735331e-02 -1.34324446e-01
-6.26022696e-01 -1.24895558e-01 -2.61022836e-01 -3.90701860e-01
-5.16762972e-01 1.27489790e-01 1.67003155e-01 7.72061110e-01
1.70633420e-01 1.15569973e+00 -8.39803696e-01 -3.76061141e-01
4.49490637e-01 8.92243326e-01 -6.79461360e-01 4.24001306e-01
2.86267608e-01 8.93168628e-01 -7.18654215e-01 3.95826966e-01
1.15231264e+00 -4.46739614e-01 -1.33638918e-01 -6.95855737e-01
-2.45995507e-01 3.16496223e-01 -7.12149024e-01 1.80112779e+00
-7.77404428e-01 9.31490809e-02 4.83025759e-01 -1.13989353e+00
3.64850640e-01 7.47020125e-01 8.80979240e-01 -1.09896159e+00
4.73052204e-01 3.21793407e-01 -1.45118296e-01 -7.89540946e-01
-1.75445471e-02 -8.25229168e-01 1.80287972e-01 5.23663700e-01
1.22841962e-01 -2.51483142e-01 -5.41935742e-01 -2.70662941e-02
1.12035644e+00 -1.87855452e-01 1.47057086e-01 -2.98195928e-01
5.84405601e-01 -3.66311133e-01 3.78061354e-01 6.20597661e-01
-1.64836168e-01 8.68941128e-01 2.25235745e-01 -5.14887154e-01
-9.91333902e-01 -9.39692497e-01 -2.37956479e-01 4.95591879e-01
1.58782929e-01 1.07899457e-01 -9.54788208e-01 -5.65664113e-01
-3.45850110e-01 3.75431716e-01 -7.44815111e-01 -1.57363892e-01
-9.67177629e-01 -1.00840998e+00 3.84614021e-01 7.31548727e-01
7.03103006e-01 -7.65158117e-01 -6.36783659e-01 5.16697884e-01
-4.46361095e-01 -1.11787689e+00 -7.44236827e-01 4.31068361e-01
-1.18401051e+00 -9.77787435e-01 -1.31304932e+00 -6.43726647e-01
5.18009126e-01 4.31331873e-01 1.14734304e+00 -1.07037507e-01
-2.77827889e-01 7.40033155e-03 -8.31019133e-02 -4.75488119e-02
-2.08187699e-01 9.73236747e-03 -1.20990716e-01 -2.08606765e-01
-6.16797246e-02 -1.03853536e+00 -9.86027241e-01 4.24776614e-01
-1.03178740e+00 2.65495807e-01 8.42680216e-01 1.22410882e+00
6.58219516e-01 1.49357677e-01 4.17149603e-01 -7.07145512e-01
5.32566011e-01 -4.32470024e-01 -2.99258262e-01 1.31743163e-01
-1.69970438e-01 8.90064836e-02 5.54405510e-01 -2.29108930e-01
-1.18312001e+00 -2.01768249e-01 -7.60380507e-01 -6.36232316e-01
-2.33101491e-02 4.24261302e-01 -1.43233640e-02 -3.03388804e-01
4.59346592e-01 5.07683635e-01 9.31137875e-02 -5.89772522e-01
-1.30869634e-02 5.17607510e-01 5.23499668e-01 -3.30030560e-01
5.84743261e-01 6.66746557e-01 1.41622387e-02 -6.27708912e-01
-1.10838521e+00 -3.59821528e-01 -6.55904889e-01 -2.22163260e-01
1.18350506e+00 -1.00345957e+00 -8.12908232e-01 8.82162154e-01
-1.08214581e+00 -2.35535681e-01 -2.48753875e-01 6.93657756e-01
-5.56607068e-01 4.37698632e-01 -8.34406316e-01 -2.98069865e-01
-9.41934228e-01 -1.70355225e+00 1.17412043e+00 2.88940787e-01
1.20053232e-01 -9.02991772e-01 -4.71215695e-02 2.91619718e-01
9.14016008e-01 1.23098262e-01 1.10182810e+00 -2.42412359e-01
-6.40666068e-01 -1.39073163e-01 -5.25182426e-01 2.53445685e-01
1.01000130e-01 -1.01532805e+00 -1.04064620e+00 -3.38424414e-01
6.31618261e-01 -2.37982139e-01 6.88520014e-01 1.12484324e+00
1.38771558e+00 4.15698923e-02 -2.48373166e-01 1.18386972e+00
1.30287230e+00 3.90403345e-02 7.39149809e-01 7.59997293e-02
7.52786040e-01 9.58160385e-02 4.90957983e-02 2.51187891e-01
2.94527501e-01 7.36223876e-01 5.20018339e-01 -6.37211978e-01
-4.88113493e-01 2.68352628e-02 -1.66720316e-01 1.08271849e+00
-1.02556519e-01 1.17063135e-01 -8.38045716e-01 3.77789795e-01
-1.50170958e+00 -6.92740023e-01 -1.57276973e-01 2.00630641e+00
8.00147355e-01 -1.14845231e-01 -1.79844499e-01 -7.02457279e-02
4.50050265e-01 1.42647430e-01 -7.37527251e-01 8.15250352e-02
1.67045519e-01 5.75787187e-01 4.81855750e-01 5.42479455e-01
-1.19113684e+00 3.52830142e-01 5.32114697e+00 1.05158579e+00
-1.26818132e+00 6.87784851e-01 8.34271371e-01 -1.43323928e-01
-2.21209049e-01 -5.48948586e-01 -4.11471635e-01 3.17115933e-01
4.78616089e-01 1.44334540e-01 7.62464404e-02 5.42379797e-01
4.23816234e-01 2.05857605e-02 -9.68680859e-01 1.03326070e+00
-2.15473846e-01 -1.27544498e+00 -1.99070916e-01 -1.33816570e-01
6.20135128e-01 2.94969887e-01 -6.78646332e-03 2.87336320e-01
-2.56571975e-02 -1.10349405e+00 4.40362364e-01 3.97338033e-01
1.06246066e+00 -8.11715662e-01 9.67866659e-01 4.58285034e-01
-1.08063734e+00 9.82946604e-02 -1.34203687e-01 4.16528642e-01
5.43826580e-01 9.25055623e-01 -3.68698537e-01 9.73090768e-01
8.97999585e-01 6.45596445e-01 5.04115820e-02 1.09976399e+00
-1.70124531e-01 3.26355159e-01 -3.55030894e-01 5.03656209e-01
5.26483536e-01 6.25974238e-02 5.97669542e-01 1.12627280e+00
4.33200419e-01 7.57944286e-01 -1.60563946e-01 8.80340755e-01
-4.87196818e-02 -3.53310615e-01 -1.38071522e-01 6.37854040e-01
2.26098597e-02 1.11653364e+00 -4.87220585e-01 -3.96050513e-01
-4.64302987e-01 9.87905562e-01 9.35968459e-02 4.07990068e-01
-1.02503383e+00 -6.72850385e-02 6.28227651e-01 5.58656216e-01
2.17954010e-01 2.13856585e-02 -4.47009534e-01 -1.10312331e+00
-4.29811664e-02 -6.45212471e-01 3.95290881e-01 -8.58544707e-01
-1.44411254e+00 9.36060965e-01 -1.50813803e-01 -1.09932256e+00
1.13433756e-01 -3.21101636e-01 -7.13377297e-01 1.28815842e+00
-1.89017034e+00 -1.13478065e+00 -7.26213872e-01 8.32301199e-01
4.04425710e-01 5.76563537e-01 6.07605100e-01 6.46101117e-01
-3.89335573e-01 5.78225374e-01 6.51477352e-02 1.25112161e-01
5.17567873e-01 -1.00965738e+00 3.51239949e-01 6.19947553e-01
-7.06514359e-01 3.36810708e-01 2.09877387e-01 -7.06298351e-01
-1.36141193e+00 -1.09936333e+00 4.36897278e-01 3.33023481e-02
4.22128975e-01 5.40824905e-02 -1.06035566e+00 3.89197141e-01
8.43807235e-02 4.30227220e-01 3.52379352e-01 -3.37951809e-01
-1.35674821e-02 2.43443903e-02 -1.31360567e+00 2.90254653e-01
8.59928548e-01 -4.25465494e-01 -4.55096006e-01 4.44852978e-01
8.03150117e-01 -1.03965604e+00 -1.11379933e+00 6.74199641e-01
2.67137468e-01 -1.05148625e+00 1.36932397e+00 -2.20985413e-01
7.62990177e-01 -1.18420176e-01 1.44354150e-01 -1.39855230e+00
-7.23135531e-01 -5.40562332e-01 -2.05461830e-02 6.00105464e-01
1.60610564e-02 -7.05568254e-01 8.52382600e-01 4.73172486e-01
-8.85296583e-01 -1.08802688e+00 -1.35777760e+00 -3.99453878e-01
3.78868669e-01 -4.32745039e-01 4.63204265e-01 9.38335538e-01
-3.99609476e-01 1.71658292e-01 -2.93252796e-01 4.55070734e-01
9.27833676e-01 -1.48960412e-01 6.21895492e-02 -7.06231296e-01
-4.24344569e-01 -5.50449848e-01 -1.30887508e-01 -1.49439692e+00
-3.14715654e-01 -8.89102042e-01 -6.50353581e-02 -1.57484674e+00
4.43750620e-01 -3.83418769e-01 -2.07753956e-01 2.75169998e-01
-2.29479343e-01 2.16405734e-01 -2.44925171e-01 8.35295171e-02
-8.15707967e-02 7.95956552e-01 2.06092739e+00 5.76788895e-02
-1.12646356e-01 1.83482822e-02 -4.99873161e-01 7.16211081e-01
4.11735982e-01 -4.74699467e-01 -2.44301096e-01 -7.27220535e-01
-3.90480876e-01 6.65453136e-01 6.62719190e-01 -8.77607942e-01
4.31244463e-01 3.45001042e-01 5.99940360e-01 -9.12195981e-01
3.59779537e-01 -8.68311524e-01 5.71388230e-02 4.76239502e-01
7.03191310e-02 -1.67982832e-01 3.10648173e-01 3.54112387e-01
-4.54385787e-01 1.36669293e-01 1.28646147e+00 -4.45139885e-01
-1.51592687e-01 7.11105883e-01 -2.47552946e-01 8.25008750e-02
7.37374902e-01 -1.66868880e-01 -1.35285826e-02 -2.76191354e-01
-1.09021616e+00 4.91327308e-02 1.95396580e-02 -6.13697767e-02
8.12340081e-01 -1.24342120e+00 -7.47635484e-01 4.80336249e-01
-4.92472678e-01 5.68567097e-01 1.24690449e+00 1.60257912e+00
-5.20667195e-01 4.45992619e-01 1.07332859e-02 -8.40826929e-01
-9.43219125e-01 3.56471956e-01 9.34632003e-01 -8.44469190e-01
-1.06668866e+00 1.16080201e+00 7.25413263e-01 -3.33309323e-01
1.81784835e-02 -4.41372752e-01 1.01232298e-01 -2.86143810e-01
6.58359647e-01 -6.73693269e-02 5.55176973e-01 -3.68680447e-01
-4.63660270e-01 8.88342619e-01 -4.16147977e-01 2.35827893e-01
1.74637282e+00 -8.74284059e-02 -1.43253326e-01 -2.92572260e-01
1.36287439e+00 -5.82815170e-01 -1.29370379e+00 -5.31974494e-01
-5.66433072e-01 -3.70678902e-01 8.26590836e-01 -8.10941696e-01
-1.51946795e+00 1.08997214e+00 7.38766789e-01 -1.77216098e-01
1.60198975e+00 1.17707290e-02 1.20049167e+00 -2.96448618e-01
1.04699604e-01 -5.23162067e-01 2.72011220e-01 4.67660367e-01
8.97070944e-01 -1.05508363e+00 1.61092319e-02 -6.01477504e-01
-3.76473188e-01 9.48865771e-01 5.68347573e-01 -1.61230415e-01
1.04775858e+00 7.10594654e-01 -9.47115645e-02 -5.84225357e-01
-4.11268473e-01 1.04348518e-01 2.99837887e-01 4.33571130e-01
6.33177996e-01 -8.94191489e-02 2.98659280e-02 9.59198892e-01
1.53494090e-01 3.04581285e-01 2.48313025e-01 8.05739880e-01
-1.52483061e-01 -6.33470118e-01 -4.56088334e-01 3.13516915e-01
-6.84912443e-01 -3.90973687e-01 4.42079008e-01 7.14575827e-01
-1.88838281e-02 3.96152645e-01 -1.45310879e-01 -2.22275540e-01
5.89566886e-01 -3.74901861e-01 6.83865070e-01 -4.02595699e-01
-9.79569018e-01 5.28309166e-01 -3.10273379e-01 -7.04166293e-01
-2.18979433e-01 -3.09774190e-01 -1.08421230e+00 -3.66919011e-01
-2.77119845e-01 -9.93857458e-02 5.32578826e-01 8.57530057e-01
2.69698679e-01 1.19434810e+00 5.80975354e-01 -1.19503880e+00
-6.35732293e-01 -1.03975666e+00 -4.72751051e-01 2.18053743e-01
4.67782915e-01 -6.59889817e-01 -1.95825458e-01 -4.29134935e-01] | [13.550154685974121, -2.485015869140625] |
1196a48f-73a9-4d78-92a8-2b0f74e7b568 | lordbert-embedding-long-text-by-segment | null | null | https://openreview.net/forum?id=b-064TCPoyB | https://openreview.net/pdf?id=b-064TCPoyB | LordBERT: Embedding Long Text by Segment Ordering with BERT | Although BERT has achieved significant improvements on many downstream NLP tasks, it has difficulty handling long text because of its quadratic computation complexity. A typical approach to this issue is splitting the input into shorter segments and utilizing order-independent attention mechanism to conduct inter-segment interaction, but the approach ignores the segment order information, which is greatly beneficial for capturing implicit relations across different segments. To address this problem, we propose a novel multi-task learning framework, named LordBERT, which fully exploits both intra- and inter-segment information in long text by segment ordering with BERT. LordBERT learns segment-level representations from segments through BERT and a reasoner, and utilizes an auxiliary segment ordering module to reorder disordered segments. With this module, the model implicitly encodes inter-segment relations and global information of long text into segment representations. The downstream task and the ordering task are jointly optimized during training, while for inferencing we mainly conduct the downstream task. Experimental results show that LordBERT outperforms the state-of-the-art models by up to 0.94% in accuracy for text classification tasks on long text. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['implicit-relations'] | ['natural-language-processing'] | [ 1.56460375e-01 3.78528982e-01 -7.08016694e-01 -6.65234923e-01
-9.72909510e-01 -5.68008542e-01 3.64333421e-01 4.78280872e-01
-3.69966567e-01 5.68353534e-01 4.06932056e-01 -5.10734677e-01
-5.41994534e-02 -7.85861671e-01 -8.22118759e-01 -4.98152554e-01
3.36106360e-01 1.03578246e+00 3.67747784e-01 -1.84521779e-01
2.83115000e-01 9.42455530e-02 -1.01503873e+00 7.11831033e-01
1.22866738e+00 7.78355062e-01 3.60044897e-01 6.55265212e-01
-7.08826900e-01 8.42999101e-01 -4.10994738e-01 -3.22338611e-01
-1.13078713e-01 -3.42776299e-01 -1.37686098e+00 -9.49418545e-02
-2.12200001e-01 -3.72219145e-01 -2.64621049e-01 6.65539861e-01
2.61466444e-01 2.30470255e-01 8.22974503e-01 -8.71372223e-01
-6.23314798e-01 1.29208457e+00 -6.97607577e-01 2.00206727e-01
1.89809158e-01 -2.85682023e-01 1.68725204e+00 -5.97559452e-01
3.54015559e-01 1.18946576e+00 5.09498179e-01 1.90911978e-01
-1.23139465e+00 -5.18168807e-01 6.88562512e-01 2.60739833e-01
-1.19010031e+00 -1.10439524e-01 7.14775503e-01 -3.76743734e-01
1.30431831e+00 8.31084475e-02 2.83715189e-01 7.24260390e-01
1.82499856e-01 1.34045494e+00 5.02305090e-01 -4.27678496e-01
-2.81404316e-01 -1.26637369e-01 9.33140457e-01 5.38647890e-01
-1.61428928e-01 -2.38096356e-01 -2.52498597e-01 9.83385183e-03
3.95551175e-01 1.72854792e-02 -1.37333184e-01 9.82497558e-02
-1.13280392e+00 9.00695264e-01 6.32928431e-01 3.95227045e-01
-1.72767803e-01 -5.56203723e-02 5.92045963e-01 3.26939344e-01
7.03692615e-01 2.48287275e-01 -8.61960113e-01 -1.50418952e-01
-9.21783984e-01 1.19872913e-01 1.02475560e+00 1.26699853e+00
8.49054635e-01 -6.22099876e-01 -2.80172586e-01 1.19125307e+00
4.73734826e-01 -2.41833534e-02 6.97486460e-01 -3.65478843e-01
1.25226498e+00 8.91429782e-01 -2.19468623e-01 -5.53297877e-01
-5.50123990e-01 -4.96029854e-01 -8.60063493e-01 -5.87811947e-01
2.83385903e-01 -1.87424988e-01 -1.00186217e+00 1.44181848e+00
1.81396112e-01 -2.05193117e-01 -4.50850055e-02 6.65207744e-01
9.02717948e-01 1.05266762e+00 1.83966130e-01 -9.28951725e-02
1.36677516e+00 -1.61999381e+00 -6.99966967e-01 -4.35400754e-01
1.16476238e+00 -6.74717784e-01 9.12859619e-01 2.23459199e-01
-9.13284838e-01 -5.19107342e-01 -8.58820796e-01 -8.02968860e-01
-3.79848331e-01 4.14966375e-01 4.15276378e-01 -3.53478417e-02
-5.80637813e-01 4.40113068e-01 -7.67962635e-01 -1.09183662e-01
3.31064612e-01 4.68603581e-01 -8.93874317e-02 -2.57914402e-02
-1.41944063e+00 7.62146115e-01 7.80609965e-01 3.12771857e-01
-4.98365223e-01 -6.03133082e-01 -9.32273746e-01 3.98597896e-01
3.80544335e-01 -5.04248679e-01 1.41725457e+00 -6.87227607e-01
-1.52209604e+00 7.19015419e-01 -7.29223728e-01 -3.63583386e-01
5.44366539e-01 -5.83935618e-01 -1.50304228e-01 -6.95279986e-02
2.64804810e-01 4.48177487e-01 5.12683094e-01 -1.10364246e+00
-5.45804620e-01 -3.34461242e-01 7.17575327e-02 5.10616541e-01
-1.75475199e-02 -8.71679857e-02 -7.68860281e-01 -6.23945236e-01
2.44246140e-01 -8.38246644e-01 -2.83245891e-01 -4.57971334e-01
-8.85105252e-01 -8.38125765e-01 7.77970314e-01 -7.02941775e-01
1.41770148e+00 -1.94730651e+00 3.02700669e-01 1.18729807e-01
2.28366613e-01 2.65600026e-01 -2.19774991e-01 6.84487522e-01
7.61134177e-02 1.54480532e-01 -3.85527790e-01 -6.05298102e-01
-1.08562492e-01 4.58511084e-01 -4.86053228e-01 1.40678987e-01
2.24560976e-01 1.13575649e+00 -7.71190107e-01 -6.87878132e-01
-1.05095282e-01 7.20693916e-02 -4.96638447e-01 3.59278530e-01
-6.32665932e-01 3.90603870e-01 -8.12467575e-01 3.01451743e-01
6.21193945e-01 -5.52331328e-01 2.10702509e-01 6.34235376e-03
-5.19524477e-02 9.10421908e-01 -5.87105334e-01 1.69345224e+00
-5.46327114e-01 4.79067087e-01 -1.23434171e-01 -1.42517745e+00
9.60545361e-01 3.73525292e-01 2.73610443e-01 -4.34936672e-01
9.31621939e-02 4.23550941e-02 3.89198624e-02 -8.25583875e-01
3.86725008e-01 -1.12140812e-01 -2.27509305e-01 7.62727737e-01
-1.72321916e-01 -1.83690265e-01 3.06756884e-01 3.63637716e-01
9.43381727e-01 2.57877052e-01 2.66581953e-01 -7.36005232e-02
5.85323036e-01 2.31821630e-02 5.76551437e-01 6.82335436e-01
1.44832497e-02 5.04607260e-01 9.01315272e-01 -3.45308721e-01
-8.72474313e-01 -7.68895984e-01 -8.17503110e-02 1.67631400e+00
3.62399578e-01 -5.60151041e-01 -5.88955522e-01 -1.00181746e+00
1.15005903e-01 7.47676551e-01 -7.29281962e-01 -4.19622473e-02
-7.61741221e-01 -7.63665140e-01 5.99821448e-01 7.35416830e-01
6.10126793e-01 -1.09198928e+00 2.06580713e-01 5.02460539e-01
-6.64255261e-01 -1.02307868e+00 -5.76503158e-01 3.79348069e-01
-9.57598031e-01 -8.79903018e-01 -6.26887619e-01 -1.07565284e+00
5.96075177e-01 1.41300291e-01 1.08024764e+00 6.45928830e-02
1.99943557e-01 -4.88494903e-01 -6.55731797e-01 -2.67189056e-01
-2.96786040e-01 8.80466044e-01 -7.17072308e-01 -2.30119944e-01
3.60361636e-01 -3.26011449e-01 -4.46113646e-01 3.91888678e-01
-7.37740040e-01 3.75919849e-01 8.68174255e-01 1.05240583e+00
4.84891504e-01 -1.33978844e-01 7.50824988e-01 -1.51509678e+00
5.62066734e-01 -6.50625408e-01 -1.77121148e-01 3.42913717e-01
-5.78552663e-01 2.43525013e-01 9.11879897e-01 -2.56666839e-01
-1.24237883e+00 -6.70914054e-02 -5.17176330e-01 1.75934806e-01
-6.25196099e-02 9.94407535e-01 -3.08815569e-01 6.03353083e-01
2.80999720e-01 2.24126890e-01 -3.50583345e-01 -6.79414093e-01
4.46003973e-01 1.00232351e+00 1.18576370e-01 -5.66847980e-01
3.09762985e-01 1.71845123e-01 -3.92515004e-01 -6.79514468e-01
-1.56423450e+00 -9.84930456e-01 -9.85993207e-01 4.05749112e-01
8.54682505e-01 -8.51580203e-01 -5.36369503e-01 5.41268229e-01
-1.37948740e+00 -6.30410850e-01 2.35976335e-02 2.38534108e-01
-3.63516271e-01 4.92149919e-01 -1.16561854e+00 -4.72065210e-01
-4.17212486e-01 -1.13889182e+00 1.27108705e+00 -1.10599771e-01
-2.87104160e-01 -1.12307632e+00 -7.18894377e-02 7.38846898e-01
4.31367867e-02 -3.69646490e-01 1.38197517e+00 -9.79667664e-01
-4.85062838e-01 -1.62532210e-01 -6.57826424e-01 1.56153783e-01
-3.95588344e-03 -2.47613832e-01 -8.21274638e-01 -1.48054017e-02
-2.53356338e-01 -3.87187660e-01 1.29714274e+00 2.98340291e-01
1.47704315e+00 -3.75760555e-01 -7.05142140e-01 5.16889274e-01
1.00777543e+00 1.24832772e-01 4.52182144e-01 2.57253408e-01
1.10799456e+00 8.60334098e-01 6.30255997e-01 1.92195028e-01
7.69903839e-01 5.62720299e-01 1.22082897e-01 -2.33797282e-02
1.12667538e-01 -2.88837761e-01 1.49830714e-01 1.04603815e+00
1.72285914e-01 -6.13170624e-01 -9.64576840e-01 4.35195982e-01
-2.18899345e+00 -6.67884171e-01 -3.78118128e-01 1.60367036e+00
1.11672854e+00 2.95977265e-01 -2.94576157e-02 1.44556746e-01
7.82559395e-01 2.75816113e-01 -6.97160423e-01 -5.73442400e-01
1.69956446e-01 -2.09916726e-01 3.52362484e-01 6.48337901e-01
-1.07174492e+00 1.32825291e+00 5.90127325e+00 8.99830937e-01
-8.13279510e-01 -6.46686777e-02 7.23103344e-01 2.87185103e-01
-3.50775778e-01 1.41257092e-01 -1.10389149e+00 4.29168224e-01
7.03898728e-01 -1.25225231e-01 1.17729269e-01 7.03239083e-01
2.37471551e-01 4.04649191e-02 -1.21257496e+00 5.20431519e-01
-5.66772148e-02 -1.01613295e+00 1.14873901e-01 -5.40625416e-02
7.19032764e-01 2.73561567e-01 -3.67005765e-01 6.83541894e-01
3.84348601e-01 -9.54887271e-01 5.24334610e-01 1.68571010e-01
6.22023284e-01 -7.51614571e-01 9.86701131e-01 7.88872182e-01
-1.35681784e+00 -8.57903287e-02 -3.65754932e-01 -1.77706584e-01
2.49432161e-01 9.05618668e-01 -8.54383945e-01 6.52013183e-01
2.67719686e-01 1.05788147e+00 -2.75221854e-01 8.09292138e-01
-5.28685033e-01 9.12988842e-01 -2.50684053e-01 -9.51241776e-02
4.62468535e-01 -3.08143198e-01 2.39376888e-01 1.49547827e+00
-1.15046315e-01 1.60568282e-01 4.10925567e-01 7.03902245e-01
-4.14099455e-01 1.70230553e-01 -2.73286194e-01 -3.31597552e-02
3.20624352e-01 9.56455767e-01 -6.67149723e-01 -6.79753602e-01
-4.32888746e-01 1.16411340e+00 7.74290085e-01 5.47848642e-01
-7.71039844e-01 -7.04671919e-01 1.41670242e-01 -1.27528131e-01
4.21804279e-01 -1.27301976e-01 -6.52012467e-01 -1.20433319e+00
4.12571467e-02 -5.18274963e-01 6.07444465e-01 -3.55277389e-01
-1.21845520e+00 5.76493800e-01 -1.41036212e-01 -7.65722096e-01
-2.08912179e-01 -3.48227233e-01 -6.83795094e-01 9.36026216e-01
-1.71362364e+00 -1.41009486e+00 9.11446288e-02 2.50091076e-01
1.02060056e+00 3.29480708e-01 6.20504737e-01 8.10698941e-02
-8.50755394e-01 6.37437820e-01 2.93909609e-01 4.58589047e-01
6.88907087e-01 -1.34656131e+00 6.15697980e-01 4.13854122e-01
-4.29198286e-03 6.68561876e-01 2.07127810e-01 -7.53977656e-01
-8.99307787e-01 -1.16950119e+00 1.47744584e+00 -1.68169528e-01
6.43750131e-01 -4.39806998e-01 -1.29104352e+00 1.17694294e+00
7.01439232e-02 -1.20057046e-01 8.53042185e-01 6.33233488e-01
-5.06420434e-01 1.61257237e-01 -5.75756609e-01 4.19722289e-01
1.08642423e+00 -6.51448369e-01 -1.02762198e+00 3.82283896e-01
1.06414437e+00 -4.50587541e-01 -7.03022957e-01 2.98591793e-01
4.13479745e-01 -6.34905815e-01 6.28686726e-01 -4.33203846e-01
7.05559611e-01 3.21808159e-02 3.81739467e-01 -1.31158388e+00
-4.72353548e-01 -4.42843258e-01 -1.94324464e-01 1.39859641e+00
9.12293315e-01 -6.61230147e-01 8.05961728e-01 5.07622540e-01
-4.89285916e-01 -1.06269681e+00 -6.87029302e-01 -5.36867082e-01
3.53077263e-01 -3.94054830e-01 5.96627116e-01 8.09746206e-01
3.80966008e-01 9.77310061e-01 -1.09148420e-01 4.56377603e-02
1.33789226e-01 7.36687064e-01 5.31524777e-01 -1.13053215e+00
-4.12373066e-01 -5.75798094e-01 1.97223023e-01 -1.98179471e+00
5.87993979e-01 -1.20164466e+00 4.53310341e-01 -1.99020290e+00
3.23303968e-01 -6.25158429e-01 -2.09977791e-01 5.24134159e-01
-5.56447089e-01 -2.54077464e-01 -2.94244140e-02 5.64614058e-01
-8.55423391e-01 7.28310585e-01 1.41545653e+00 -2.48351648e-01
-4.33629572e-01 2.29288802e-01 -8.01388443e-01 6.41714454e-01
8.38214576e-01 -5.34967840e-01 -3.83010864e-01 -8.23928714e-01
2.12548375e-01 1.62681371e-01 -2.38163024e-01 -4.34504837e-01
2.94855118e-01 2.22096704e-02 2.58179963e-01 -9.32586491e-01
1.61663458e-01 -6.24106586e-01 -5.64345062e-01 2.62154788e-01
-9.53300655e-01 -1.77509680e-01 7.61253713e-03 6.11607015e-01
-3.57625216e-01 -4.11443025e-01 4.73181725e-01 2.30155885e-02
-4.43021238e-01 3.54730755e-01 -4.32050675e-01 1.55252233e-01
8.13037694e-01 2.53987521e-01 -3.04645389e-01 -1.85577556e-01
-7.36981809e-01 8.01128685e-01 2.05052756e-02 3.60425562e-01
3.08209002e-01 -8.49040985e-01 -7.20068514e-01 3.10558856e-01
1.13407575e-01 7.80371547e-01 1.39079630e-01 7.31849313e-01
-4.30863529e-01 6.28078580e-01 4.92470413e-01 -5.77982247e-01
-1.02652538e+00 4.62086469e-01 2.17813943e-02 -9.89050329e-01
-6.60285175e-01 1.10773301e+00 5.00805855e-01 -7.44711041e-01
1.08353458e-01 -6.25947833e-01 -2.84239352e-01 2.61504084e-01
3.15273672e-01 2.16989234e-01 4.84732576e-02 -5.12881756e-01
-1.58660859e-01 6.97617054e-01 -7.27050126e-01 3.95334698e-02
1.19789422e+00 -4.67019022e-01 -2.08983317e-01 6.49210989e-01
1.44545531e+00 -2.08169699e-01 -1.18917596e+00 -3.91602725e-01
2.92694122e-01 -6.58737198e-02 -1.01186037e-01 -7.35969663e-01
-7.07540989e-01 1.02426982e+00 -5.37837148e-01 3.46150815e-01
7.88852692e-01 2.62702137e-01 1.33446813e+00 5.58104575e-01
-3.91970277e-02 -9.90457058e-01 -1.90770756e-02 1.17346478e+00
8.07432055e-01 -1.34707069e+00 -1.34704784e-01 -8.08641136e-01
-7.13255346e-01 9.94064927e-01 5.31851172e-01 -9.90397669e-03
7.54490197e-01 1.06192723e-01 -9.11964942e-03 -4.95369919e-02
-8.95236850e-01 -2.02794641e-01 2.87948281e-01 1.85313419e-01
8.81873012e-01 -3.65245156e-02 -4.30234045e-01 5.86827874e-01
-2.15580791e-01 -9.12682936e-02 -2.41850317e-02 7.37229884e-01
-4.68151093e-01 -1.33990693e+00 -2.11030077e-02 7.60726035e-01
-6.01885259e-01 -3.02628011e-01 -3.46306056e-01 5.60551763e-01
-2.58593392e-02 1.11186445e+00 1.15845069e-01 -1.52987480e-01
1.26502678e-01 9.96616855e-02 9.16682482e-02 -9.65395987e-01
-6.98626816e-01 1.75384864e-01 2.86777347e-01 -2.73165643e-01
-1.10909171e-01 -5.82572997e-01 -1.67683029e+00 -6.45810738e-02
-6.32280886e-01 6.04441226e-01 4.28877920e-01 1.34208667e+00
3.98479044e-01 7.87124634e-01 5.43818772e-01 -6.96723163e-01
-5.51335812e-01 -1.22539103e+00 -4.73668844e-01 4.36470270e-01
4.67586309e-01 -3.82915348e-01 -3.11825901e-01 8.85276720e-02] | [10.94691276550293, 8.547266960144043] |
bf3d8a1c-5584-4cbc-a6eb-f9ccfd55e5ae | denoising-based-image-reconstruction-from | 2205.11202 | null | https://arxiv.org/abs/2205.11202v1 | https://arxiv.org/pdf/2205.11202v1.pdf | Denoising-based image reconstruction from pixels located at non-integer positions | Digital images are commonly represented as regular 2D arrays, so pixels are organized in form of a matrix addressed by integers. However, there are many image processing operations, such as rotation or motion compensation, that produce pixels at non-integer positions. Typically, image reconstruction techniques cannot handle samples at non-integer positions. In this paper, we propose to use triangulation-based reconstruction as initial estimate that is later refined by a novel adaptive denoising framework. Simulations reveal that improvements of up to more than 1.8 dB (in terms of PSNR) are achieved with respect to the initial estimate. | ['André Kaup', 'Jürgen Seiler', 'Ján Koloda'] | 2022-05-23 | null | null | null | null | ['motion-compensation'] | ['computer-vision'] | [ 6.62406683e-01 -2.21881300e-01 1.89009383e-01 -1.77580535e-01
-4.87260848e-01 -2.91632801e-01 3.49702597e-01 2.03874782e-01
-5.22145212e-01 7.68601477e-01 1.33729994e-01 -7.73772225e-02
1.89268798e-01 -7.85666943e-01 -6.43574178e-01 -8.16020191e-01
-1.94598921e-02 -2.65625954e-01 1.78917959e-01 -6.56741187e-02
6.19340837e-01 6.39429271e-01 -1.35865963e+00 2.42854521e-01
8.70538652e-01 1.19904244e+00 1.58473834e-01 6.46347165e-01
-1.25202388e-01 3.33708256e-01 -6.62666261e-01 -2.55519181e-01
5.90801418e-01 -3.52342516e-01 -5.70450760e-02 4.90110248e-01
5.08092523e-01 -6.25698030e-01 -2.03494027e-01 1.44687760e+00
3.48532170e-01 -6.08158112e-02 3.62508833e-01 -5.57329297e-01
-4.40494090e-01 2.64383554e-01 -1.08187222e+00 -4.70122024e-02
4.74954873e-01 -3.09697930e-02 3.27518135e-01 -1.12509096e+00
5.37883341e-01 1.09338892e+00 7.69411147e-01 1.32412478e-01
-1.31618440e+00 -3.59878570e-01 -8.38603005e-02 2.85696089e-01
-1.55770421e+00 -4.72369283e-01 9.78204489e-01 -1.79887339e-02
4.12528068e-01 5.35554290e-01 6.46797478e-01 3.40714216e-01
4.93780315e-01 4.29650694e-01 1.26927745e+00 -6.43550754e-01
1.67375222e-01 -2.78034270e-01 -6.22045696e-02 2.83470929e-01
4.45305824e-01 -2.59295613e-01 -5.03884137e-01 1.38081727e-03
1.35116136e+00 2.23363210e-02 -4.97506201e-01 -2.03544438e-01
-1.35470641e+00 5.18554568e-01 4.93641466e-01 2.97661722e-01
-7.87011027e-01 3.53564411e-01 2.65021175e-01 4.16865438e-01
3.36920321e-01 8.34110007e-02 1.30040303e-01 1.70858294e-01
-1.15383005e+00 -3.99950966e-02 3.96177858e-01 8.59910488e-01
8.22446942e-01 1.78343832e-01 1.27723917e-01 6.46585941e-01
2.35987768e-01 4.44062322e-01 4.65401083e-01 -1.39103758e+00
3.89266640e-01 2.18539611e-01 2.37947822e-01 -1.48436630e+00
-1.54983550e-01 -5.36566079e-01 -1.48414779e+00 5.48279166e-01
2.74558574e-01 7.44447932e-02 -8.14648986e-01 1.17719662e+00
3.36066037e-01 2.84927070e-01 1.18935056e-01 9.11431134e-01
4.99645144e-01 1.01789975e+00 -3.50896597e-01 -7.45204151e-01
1.13744807e+00 -8.36210966e-01 -1.10733032e+00 -7.23539758e-03
-1.80763975e-01 -1.09208536e+00 5.54676116e-01 1.01142228e+00
-1.57103693e+00 -7.63876915e-01 -1.17482316e+00 -1.08430199e-01
1.55290425e-01 4.63629179e-02 1.07547343e-01 6.22470379e-01
-1.15425098e+00 4.98194933e-01 -5.34762859e-01 1.41943678e-01
4.35940363e-03 2.28119507e-01 -3.12023014e-01 -1.82788789e-01
-7.30755448e-01 5.21209061e-01 1.12922251e-01 6.40271127e-01
-3.43121618e-01 -3.55600387e-01 -5.58520615e-01 -2.44273663e-01
3.60090099e-02 -5.59554517e-01 7.24616528e-01 -1.01153719e+00
-1.33170676e+00 6.88814163e-01 -4.52086985e-01 -6.13896489e-01
6.52020752e-01 -5.26232086e-02 -4.11453158e-01 4.70660865e-01
-1.03833504e-01 4.07334268e-01 1.40854216e+00 -1.44006145e+00
-5.02265155e-01 -3.94137293e-01 -1.76004618e-01 2.15819761e-01
-2.54212558e-01 -1.99600026e-01 -5.84667265e-01 -9.11846817e-01
1.00429404e+00 -6.79327726e-01 -7.20549047e-01 4.95146871e-01
-2.79775560e-01 4.74362075e-01 5.69355190e-01 -8.83095086e-01
1.29484463e+00 -2.29019022e+00 2.49861032e-02 5.29936194e-01
2.77919710e-01 5.39994426e-02 -2.41828188e-02 3.37252647e-01
-6.84622005e-02 -1.03694662e-01 -3.85980874e-01 -2.64972836e-01
-4.46926385e-01 1.11129344e-01 -1.25905469e-01 6.99198425e-01
-5.19160889e-02 1.48349062e-01 -6.31795883e-01 -3.30313981e-01
4.61488038e-01 5.85967898e-01 -3.19844484e-01 -1.31378233e-01
3.14464986e-01 5.68333626e-01 -3.98277998e-01 6.23389721e-01
1.36528027e+00 1.77227318e-01 2.20074475e-01 -7.29229450e-01
-6.17129564e-01 -1.86683953e-01 -1.63248241e+00 1.45179570e+00
-3.65048379e-01 6.47740543e-01 5.04368067e-01 -9.49195802e-01
1.13930976e+00 3.01435262e-01 6.29958093e-01 -5.64274013e-01
-2.87478343e-02 3.76474291e-01 -2.27866992e-01 -1.21891052e-01
8.17763627e-01 1.21316619e-01 2.46305034e-01 -3.18439052e-05
-7.43658125e-01 -2.18737915e-01 2.53887743e-01 -1.95343137e-01
1.09323871e+00 -3.84692639e-01 4.06657577e-01 -1.86584562e-01
8.21842492e-01 8.14822316e-02 6.58374965e-01 6.01331592e-01
-8.40689912e-02 9.90095198e-01 2.97625601e-01 -4.60309595e-01
-1.25680900e+00 -9.30601597e-01 -5.06002486e-01 8.24529752e-02
6.31300032e-01 -1.91737905e-01 -6.46210611e-01 1.85055003e-01
-1.04166053e-01 3.87143880e-01 -1.51960745e-01 2.96489477e-01
-8.91825020e-01 -6.96357250e-01 1.51391581e-01 2.84800947e-01
1.04361439e+00 -5.90322316e-01 -7.56941199e-01 5.77316105e-01
7.57977590e-02 -1.16267860e+00 -4.41850185e-01 -2.66075730e-02
-1.43977964e+00 -6.21571124e-01 -1.00350404e+00 -9.15466666e-01
1.07506108e+00 5.91824472e-01 8.82784665e-01 2.91007102e-01
-1.36918306e-01 2.80045837e-01 -3.35554302e-01 2.13030577e-01
-2.30227530e-01 -6.14563525e-01 -8.80546868e-02 2.43464723e-01
-1.64767325e-01 -6.41244411e-01 -1.00145531e+00 3.18443567e-01
-1.13175154e+00 6.14402145e-02 8.00370157e-01 8.01861525e-01
9.99343336e-01 4.14994001e-01 2.65094310e-01 -6.77195966e-01
6.08969033e-01 -1.47170156e-01 -6.59309328e-01 -1.88608710e-02
-4.78692055e-01 -1.48269624e-01 5.70061684e-01 -4.15269881e-01
-9.70351875e-01 4.32544678e-01 -1.38109565e-01 -3.70690286e-01
1.06038459e-01 4.87356424e-01 6.21537864e-02 -4.38307613e-01
3.56610656e-01 3.22374374e-01 9.12444443e-02 -4.98711824e-01
2.38131225e-01 5.66433072e-01 8.19280148e-01 -2.00420991e-01
9.80445087e-01 5.92073083e-01 4.56161976e-01 -1.03781903e+00
-7.66644627e-02 -2.78420627e-01 -5.29192805e-01 -3.40423077e-01
5.96912086e-01 -1.01326931e+00 -4.72815156e-01 7.27749109e-01
-1.12248266e+00 8.79671425e-02 -1.56582445e-01 4.14313972e-01
-3.50000173e-01 8.16171467e-01 -7.53650308e-01 -7.26779699e-01
-3.83627892e-01 -1.19032300e+00 6.82383180e-01 2.41648123e-01
-6.20882735e-02 -7.53268063e-01 -4.14694577e-01 1.24624908e-01
5.10147870e-01 3.29757303e-01 6.03589475e-01 3.50376368e-01
-8.75125766e-01 -2.53172696e-01 -2.24738076e-01 3.74865174e-01
4.43452783e-02 7.45623782e-02 -5.14594495e-01 -3.55693847e-01
4.98572737e-01 2.16061920e-01 6.51260495e-01 5.40050626e-01
1.26471019e+00 -1.57973468e-01 -1.40632778e-01 6.96138084e-01
1.69464445e+00 3.44458431e-01 1.19101238e+00 4.36223179e-01
3.03605288e-01 3.09990197e-01 7.46262729e-01 6.62293911e-01
-4.99688387e-02 6.66556597e-01 5.68805635e-01 -1.05146274e-01
-2.15786815e-01 1.51913136e-01 7.61712417e-02 9.53119814e-01
-2.13422343e-01 -2.01713815e-01 -4.89724755e-01 4.52453285e-01
-1.54695737e+00 -8.22831750e-01 -7.76072800e-01 2.29593945e+00
8.97900403e-01 1.32712722e-01 -4.73927975e-01 5.98401725e-01
9.34537053e-01 3.16190958e-01 -4.35635030e-01 -5.25822163e-01
-3.26457739e-01 1.85684681e-01 1.03214800e+00 4.90660369e-01
-8.14079344e-01 2.76830018e-01 7.09660387e+00 8.76054764e-01
-1.23112655e+00 -1.93743050e-01 5.85251153e-01 3.85727316e-01
-4.38182026e-01 -8.17372277e-03 -4.14315432e-01 4.90940481e-01
3.70554417e-01 -1.07415207e-02 3.02341580e-01 4.29821402e-01
4.58897054e-01 -5.18134832e-01 -4.60307032e-01 1.36449194e+00
-1.72583107e-02 -1.03549564e+00 1.38188437e-01 9.30411369e-02
1.02597260e+00 -8.58260930e-01 3.42509389e-01 -4.17805701e-01
-3.87102067e-01 -7.52212226e-01 7.47303963e-01 5.41942537e-01
6.36436522e-01 -7.31466472e-01 5.60748339e-01 2.94110328e-01
-1.11064565e+00 1.72936320e-01 -7.15023935e-01 3.53009291e-02
3.98487657e-01 1.17655051e+00 -4.02776062e-01 5.33498526e-01
5.43798923e-01 6.06252253e-01 -7.54810870e-02 1.33183599e+00
-1.85050562e-01 4.52043921e-01 -3.56642395e-01 2.78210372e-01
9.64993238e-02 -8.08199644e-01 5.70071876e-01 7.31818080e-01
8.74802411e-01 4.03573871e-01 -7.67177641e-02 3.95699382e-01
-6.94436058e-02 1.26307577e-01 -4.66693729e-01 5.04742622e-01
4.53158557e-01 1.07277989e+00 -6.93020940e-01 -3.88348073e-01
-3.98537666e-01 1.40883374e+00 -4.15162683e-01 3.30277294e-01
-5.88933885e-01 -3.39369535e-01 4.67116416e-01 2.36376181e-01
3.62754643e-01 -7.19489098e-01 -6.19296670e-01 -7.17697144e-01
2.03761190e-01 -8.99322867e-01 -1.94594011e-01 -7.69114673e-01
-9.12607849e-01 4.36342627e-01 -3.36482108e-01 -1.68561113e+00
-2.63537377e-01 -4.73569930e-01 -4.00606751e-01 8.97844613e-01
-1.48683333e+00 -5.43363035e-01 -5.61720014e-01 4.71110851e-01
6.53895020e-01 2.15271279e-01 4.93972033e-01 4.72612143e-01
-3.69369507e-01 3.03057164e-01 6.49209917e-01 -1.66515768e-01
4.43195939e-01 -8.42840672e-01 2.06602857e-01 1.05165684e+00
-1.24355517e-01 5.32586217e-01 9.72463846e-01 -6.96493566e-01
-1.68651342e+00 -6.53070509e-01 6.28280878e-01 4.51408446e-01
2.66843736e-01 2.96369880e-01 -9.05032039e-01 3.88824791e-01
4.37744021e-01 2.34651387e-01 1.29790008e-01 -7.76912212e-01
1.24080993e-01 -3.41966331e-01 -1.61338568e+00 6.55606568e-01
8.16301227e-01 -1.33144915e-01 -1.55421078e-01 6.12740852e-02
1.16035879e-01 -6.42538607e-01 -1.01267874e+00 4.48030740e-01
4.33304489e-01 -1.14486241e+00 1.23220956e+00 3.59793514e-01
3.28541696e-01 -7.29559660e-01 -3.02064657e-01 -1.07554209e+00
-2.58896738e-01 -6.08655512e-01 1.25990912e-01 1.00074387e+00
6.34668767e-02 -5.06823719e-01 7.97802925e-01 2.79199272e-01
4.38839719e-02 -3.50969464e-01 -1.17906070e+00 -5.04308462e-01
-5.80074787e-01 -2.56134570e-01 2.42184132e-01 6.95131779e-01
-5.37850797e-01 -1.78423852e-01 -4.96665776e-01 3.58300388e-01
1.14919269e+00 3.54253575e-02 5.41022658e-01 -7.67755866e-01
9.15239379e-02 -1.82143807e-01 -7.81692863e-01 -1.61384594e+00
-4.32586491e-01 -1.54805422e-01 7.18245432e-02 -1.65849078e+00
-1.92174762e-01 -4.81001794e-01 -1.98473468e-01 -1.67191699e-01
1.70015410e-01 9.29742575e-01 -1.27393939e-02 3.19070905e-01
-7.63790235e-02 3.07948977e-01 1.45396543e+00 -5.29505238e-02
-2.57488098e-02 -2.19932012e-02 -1.70533985e-01 8.62068534e-01
7.08525836e-01 -1.03589356e-01 -1.49510726e-01 -8.01371872e-01
6.45952448e-02 4.00678098e-01 2.36272097e-01 -1.29031837e+00
2.80528367e-01 5.77132516e-02 6.92161679e-01 -9.92481709e-01
7.52712607e-01 -1.14122903e+00 7.76353240e-01 5.49903870e-01
-9.80702788e-02 2.17335597e-01 -3.19546275e-02 5.48464298e-01
-5.55533528e-01 -7.97908187e-01 7.14203775e-01 -5.44529185e-02
-8.15942526e-01 -2.23426238e-01 -6.37225986e-01 -6.63321018e-01
1.05205047e+00 -7.80658901e-01 2.35483274e-02 -6.46143377e-01
-6.98684573e-01 -2.86710441e-01 7.55323768e-01 -4.09822673e-01
9.32658911e-01 -1.41099286e+00 -6.97502077e-01 3.08693498e-01
-5.30087948e-01 6.50102869e-02 4.01566535e-01 9.21873629e-01
-1.23846543e+00 1.04145641e-02 -3.93567592e-01 -6.81440771e-01
-1.48852742e+00 2.88885772e-01 7.35101029e-02 -5.02718464e-02
-6.52123272e-01 5.45918047e-01 -3.17396134e-01 3.84690523e-01
3.82634811e-02 -3.39772820e-01 -4.65527689e-03 1.32931367e-01
6.74983263e-01 6.43401563e-01 3.70642245e-02 -6.53615654e-01
-6.48039579e-02 1.15830302e+00 1.74592972e-01 -2.62899518e-01
1.14735913e+00 -4.11274821e-01 -5.15382051e-01 -1.84443919e-03
1.13771379e+00 4.73754883e-01 -1.23419189e+00 -1.19258396e-01
-2.16820002e-01 -1.14393961e+00 2.43356258e-01 -2.27893353e-01
-1.24980736e+00 5.21535456e-01 8.45134199e-01 3.77975792e-01
1.75181770e+00 -6.67006433e-01 8.30752850e-01 2.32951999e-01
5.99878252e-01 -8.95382285e-01 -3.62478457e-02 1.71654254e-01
9.99749839e-01 -8.94875884e-01 3.53751570e-01 -7.96413839e-01
-8.51161927e-02 1.40261769e+00 1.28125295e-01 -4.84723091e-01
4.03928310e-01 4.41420615e-01 1.52720764e-01 2.67305225e-01
-3.06670964e-01 6.61114156e-02 -7.12734908e-02 3.75314862e-01
4.35198575e-01 -2.03555137e-01 -8.92735898e-01 -1.46495059e-01
1.73525840e-01 -8.98765698e-02 6.79405153e-01 8.86189699e-01
-7.06960618e-01 -1.15093267e+00 -1.27685702e+00 2.47258261e-01
-4.31101084e-01 -1.52578121e-02 3.76173347e-01 4.58081722e-01
-1.43546816e-02 9.87347603e-01 2.71789670e-01 -1.42791867e-01
3.13310266e-01 -6.06084347e-01 5.96778810e-01 -6.58141673e-02
-3.76031697e-01 3.76148671e-01 -5.63608967e-02 -4.98101115e-01
-4.61777180e-01 -7.03795254e-01 -1.28210938e+00 -3.78271699e-01
8.95972401e-02 -9.08611044e-02 9.31142688e-01 2.21905425e-01
-5.79621866e-02 4.54272866e-01 9.20104325e-01 -9.03713346e-01
-5.79384506e-01 -7.04739869e-01 -5.80857694e-01 2.65601784e-01
3.58821958e-01 -1.19214922e-01 -5.00836849e-01 2.50459731e-01] | [11.208309173583984, -2.359672784805298] |
6fdddc82-5199-4f22-9b42-030ec7cfa8b3 | towards-understanding-pixel-vulnerability | 2010.06131 | null | https://arxiv.org/abs/2010.06131v2 | https://arxiv.org/pdf/2010.06131v2.pdf | Learning to Attack with Fewer Pixels: A Probabilistic Post-hoc Framework for Refining Arbitrary Dense Adversarial Attacks | Deep neural network image classifiers are reported to be susceptible to adversarial evasion attacks, which use carefully crafted images created to mislead a classifier. Many adversarial attacks belong to the category of dense attacks, which generate adversarial examples by perturbing all the pixels of a natural image. To generate sparse perturbations, sparse attacks have been recently developed, which are usually independent attacks derived by modifying a dense attack's algorithm with sparsity regularisations, resulting in reduced attack efficiency. In this paper, we aim to tackle this task from a different perspective. We select the most effective perturbations from the ones generated from a dense attack, based on the fact we find that a considerable amount of the perturbations on an image generated by dense attacks may contribute little to attacking a classifier. Accordingly, we propose a probabilistic post-hoc framework that refines given dense attacks by significantly reducing the number of perturbed pixels but keeping their attack power, trained with mutual information maximisation. Given an arbitrary dense attack, the proposed model enjoys appealing compatibility for making its adversarial images more realistic and less detectable with fewer perturbations. Moreover, our framework performs adversarial attacks much faster than existing sparse attacks. | ['Dinh Phung', 'Tamas Abraham', 'Olivier De Vel', 'Paul Montague', 'Thanh Nguyen', 'Trung Le', 'He Zhao'] | 2020-10-13 | null | null | null | null | ['adversarial-attack-detection', 'adversarial-attack-detection'] | ['computer-vision', 'knowledge-base'] | [ 7.09339917e-01 4.75191087e-01 2.85715431e-01 -4.98472117e-02
-5.63825727e-01 -9.11296189e-01 1.00447309e+00 -9.33327079e-02
-4.08834010e-01 7.24925637e-01 1.99285656e-01 -1.30135454e-02
1.16137397e-02 -1.03312552e+00 -1.07115936e+00 -1.07243121e+00
6.82848180e-03 1.05137371e-01 8.95939246e-02 -4.17494118e-01
1.46520436e-01 6.46036029e-01 -1.19762278e+00 2.69355297e-01
6.90644741e-01 6.54623151e-01 -2.87339956e-01 5.02404034e-01
2.35206217e-01 9.62769389e-01 -1.12810421e+00 -8.50846827e-01
6.13806963e-01 -3.53595912e-01 -4.95408654e-01 -7.58652166e-02
4.49608415e-01 -2.48937532e-01 -7.98062742e-01 1.67520237e+00
4.61279690e-01 3.19713354e-02 7.15822756e-01 -1.21098542e+00
-6.38386428e-01 9.26156878e-01 -3.80105436e-01 1.35076419e-01
2.91736126e-01 4.35605466e-01 6.98781192e-01 -3.61358404e-01
3.36653799e-01 1.42669725e+00 4.72856849e-01 8.59522760e-01
-1.19302332e+00 -9.12295222e-01 2.08791777e-01 1.04010619e-01
-1.34591353e+00 -4.05617684e-01 1.05549634e+00 -1.85227826e-01
2.13309854e-01 7.38674223e-01 3.00872147e-01 1.74384928e+00
2.25959837e-01 5.86444318e-01 1.19776642e+00 -2.02801213e-01
4.11600292e-01 3.06970119e-01 -5.03486693e-01 3.21324438e-01
3.30158949e-01 5.24125934e-01 -4.51424569e-02 -4.66493696e-01
3.99057925e-01 7.70109147e-02 -6.04954600e-01 -2.50951886e-01
-9.16232765e-01 1.09682429e+00 9.07102227e-01 2.46286541e-01
-3.11474025e-01 6.28622845e-02 3.11567605e-01 4.29511726e-01
2.00006098e-01 7.88545966e-01 -5.30942604e-02 3.99857730e-01
-5.24741948e-01 3.77296746e-01 7.84256637e-01 5.25406241e-01
6.19201124e-01 3.51978660e-01 -2.94564277e-01 4.64389890e-01
1.34499833e-01 5.75991869e-01 4.06706512e-01 -4.53547686e-01
3.38990688e-01 3.76755983e-01 -1.89644530e-01 -1.49575579e+00
-5.58055080e-02 -7.24032760e-01 -1.23180950e+00 4.82087016e-01
4.84986126e-01 -3.11555952e-01 -8.28640342e-01 1.99343669e+00
3.47288489e-01 3.22500885e-01 3.73773336e-01 6.71636462e-01
4.38651890e-01 4.65390414e-01 7.82236978e-02 -9.00826305e-02
9.82423306e-01 -4.98619884e-01 -5.17237604e-01 -3.72835159e-01
3.46063733e-01 -6.50843203e-01 8.18287909e-01 4.57296312e-01
-8.23568463e-01 -4.07611609e-01 -1.20602691e+00 6.74574316e-01
-4.24990594e-01 -4.87918913e-01 3.67041796e-01 1.06653047e+00
-5.73286176e-01 5.45348823e-01 -4.82988387e-01 2.18423679e-01
8.18807602e-01 3.80939335e-01 -5.38368821e-01 -1.76864237e-01
-1.52009487e+00 8.53415191e-01 4.92233306e-01 1.18665077e-01
-1.28798759e+00 -6.37894630e-01 -8.49211752e-01 -8.37677121e-02
3.09518576e-01 -4.95632261e-01 6.46182120e-01 -1.24990392e+00
-1.42594302e+00 6.74273312e-01 4.86323178e-01 -9.58256066e-01
8.51498783e-01 -5.52884527e-02 -2.61168003e-01 2.17179567e-01
-1.87222674e-01 5.10911047e-01 1.60069656e+00 -1.44881415e+00
-7.50114173e-02 -3.53334457e-01 4.81849700e-01 7.35213012e-02
-6.96937919e-01 -1.77583277e-01 5.81040196e-02 -1.16385269e+00
-4.25350368e-01 -1.07935882e+00 -7.79550016e-01 -2.35089749e-01
-6.68284893e-01 3.18279207e-01 1.07344639e+00 -8.85610729e-02
8.92427802e-01 -2.25689507e+00 2.96269029e-01 4.12381113e-01
5.46913743e-01 6.86396241e-01 -2.97885537e-01 3.54268223e-01
-2.79579341e-01 2.37652361e-01 -5.18932164e-01 -1.74863427e-03
-4.25875522e-02 2.77802289e-01 -7.07133889e-01 7.70983160e-01
2.20159307e-01 8.18501592e-01 -8.98356378e-01 -4.51889336e-02
2.05139354e-01 5.22974730e-01 -6.10593021e-01 4.18969512e-01
-2.00455859e-01 6.24330819e-01 -5.83261490e-01 2.43771583e-01
9.20619249e-01 2.33962998e-01 -1.30724847e-01 -2.17882633e-01
5.98539054e-01 -3.76312137e-01 -1.01247227e+00 1.07096505e+00
-2.96413630e-01 3.62127513e-01 8.61713141e-02 -1.25990856e+00
1.00318754e+00 2.42873743e-01 1.95346296e-01 -2.44736835e-01
5.36183417e-01 -4.95320745e-02 2.62854815e-01 -1.97458982e-01
-6.00185469e-02 -2.59409994e-01 -3.12117368e-01 2.56477982e-01
-1.29364356e-01 -2.24860549e-01 -1.88990727e-01 3.68540436e-01
1.53824544e+00 -5.53529382e-01 3.09118837e-01 -2.20825057e-02
6.95525885e-01 -2.94537425e-01 4.65933532e-01 1.08275867e+00
-2.11607143e-01 5.64708591e-01 4.76686120e-01 -4.50792849e-01
-8.85509253e-01 -1.07936919e+00 -7.92225972e-02 4.83555913e-01
2.97101617e-01 -2.23334089e-01 -1.06843984e+00 -1.18674719e+00
1.88425276e-02 4.75693524e-01 -9.77373183e-01 -7.27093279e-01
-4.78440225e-01 -9.58866000e-01 1.05330002e+00 6.94004446e-02
7.74654448e-01 -1.08304250e+00 -3.01866740e-01 7.68349599e-03
2.02511102e-01 -1.11390996e+00 -3.22265893e-01 2.07728609e-01
-4.28217560e-01 -1.07568502e+00 -7.61502206e-01 -4.93657112e-01
8.95008504e-01 -5.27554601e-02 7.54540563e-01 2.12079622e-02
-1.85883239e-01 8.80296007e-02 -6.50141954e-01 -6.10327542e-01
-9.07897592e-01 -9.88401696e-02 2.64688879e-01 4.72063571e-01
1.17973626e-01 -8.79972816e-01 -3.19295943e-01 2.40763858e-01
-1.49797738e+00 -4.09448773e-01 7.66584337e-01 9.21965718e-01
2.33182147e-01 5.42282462e-01 4.22598451e-01 -1.26799083e+00
6.66140914e-01 -5.81490159e-01 -2.82002121e-01 -8.07088315e-02
-2.26385191e-01 1.65045694e-01 1.29896772e+00 -9.19526458e-01
-7.79893816e-01 1.11536987e-01 -4.14367735e-01 -9.28078711e-01
-4.70497668e-01 8.68987143e-02 -4.92680460e-01 -5.99908292e-01
1.26714337e+00 3.95801902e-01 1.52506186e-02 -1.16599359e-01
5.55977583e-01 2.79571116e-01 5.57740152e-01 -3.83245111e-01
1.78182685e+00 6.03223622e-01 2.02946246e-01 -7.57013559e-01
-8.37112784e-01 1.06714055e-01 -2.17865437e-01 -8.78676325e-02
5.23160398e-01 -6.20984316e-01 -4.93317723e-01 7.02687681e-01
-9.24935520e-01 -6.54312894e-02 -4.40768331e-01 1.94803983e-01
-3.29489470e-01 5.57908773e-01 -3.11028004e-01 -4.76226628e-01
-1.63598135e-01 -1.25111604e+00 6.81296408e-01 1.44462258e-01
-8.89456049e-02 -8.30328345e-01 6.69090673e-02 1.13953874e-01
5.43976426e-01 8.50611567e-01 6.70153379e-01 -1.06174207e+00
-3.54749471e-01 -6.30141258e-01 1.14659578e-01 6.44440055e-01
2.01737061e-01 -2.66201735e-01 -1.05493581e+00 -5.73317349e-01
4.16748047e-01 -3.91663522e-01 8.50481033e-01 7.89576843e-02
1.27704763e+00 -9.72095788e-01 -1.90241963e-01 8.91150951e-01
1.31148636e+00 8.42883512e-02 8.98014009e-01 3.72652680e-01
8.75886142e-01 5.41480660e-01 1.62301987e-01 3.75076234e-01
-4.73563701e-01 5.20213783e-01 1.09095061e+00 -1.63320169e-01
1.85395747e-01 -1.36681005e-01 4.11448926e-01 1.28763363e-01
1.99347824e-01 -4.43376511e-01 -4.80012715e-01 2.93815553e-01
-1.54800940e+00 -1.22975016e+00 2.19436392e-01 2.02586174e+00
8.07837963e-01 5.02073467e-01 -8.19724128e-02 4.37945783e-01
7.41684139e-01 6.59862459e-01 -5.69055378e-01 -2.94046789e-01
-3.52876842e-01 4.69524235e-01 6.59828663e-01 4.04347688e-01
-1.28062296e+00 8.79264116e-01 5.81743956e+00 1.08015466e+00
-1.00877118e+00 -1.56137079e-01 6.52330458e-01 -1.57062262e-01
-4.84147578e-01 -1.37360901e-01 -5.51573932e-01 6.46709740e-01
6.99359357e-01 -2.59182274e-01 3.59212905e-01 9.52607214e-01
-9.39067230e-02 4.66636807e-01 -8.89023006e-01 7.41370320e-01
2.18052655e-01 -1.31044841e+00 5.55606365e-01 2.35961989e-01
8.62437189e-01 -2.78925985e-01 4.72122163e-01 1.81727692e-01
6.41211927e-01 -1.28855968e+00 5.28364837e-01 2.44112521e-01
4.05068636e-01 -1.02581239e+00 7.53949821e-01 4.20365512e-01
-7.32198596e-01 -1.69810668e-01 -4.55718428e-01 4.86447737e-02
-1.51628256e-01 6.65688634e-01 -6.37071192e-01 3.16885293e-01
4.28176343e-01 4.71136481e-01 -5.73652565e-01 5.79645097e-01
-4.28943992e-01 6.43405199e-01 -1.39287412e-01 6.52592704e-02
4.43655193e-01 -6.22902252e-03 9.92083848e-01 9.18559670e-01
3.48771620e-03 1.70109086e-02 9.66105834e-02 8.62074494e-01
-3.36898118e-01 -6.04441576e-02 -1.19985354e+00 2.15903074e-01
4.83527333e-01 1.10182631e+00 -4.36992854e-01 -1.50881752e-01
-8.98759738e-02 1.14044607e+00 1.18611507e-01 2.48928070e-01
-9.45538640e-01 -4.60719556e-01 8.39965463e-01 -1.02624901e-01
2.13723898e-01 2.76928037e-01 -8.20132419e-02 -1.06719387e+00
-6.05250858e-02 -1.50617635e+00 2.61527181e-01 -2.47295499e-01
-1.61168480e+00 9.50663626e-01 -1.26388416e-01 -1.38013506e+00
-2.00669825e-01 -4.22174841e-01 -8.04501295e-01 7.65314758e-01
-1.17165506e+00 -9.47107971e-01 -2.27661297e-01 1.06365752e+00
3.73390496e-01 -4.49737161e-01 8.09735060e-01 -4.07179706e-02
-5.78673005e-01 9.94562209e-01 -8.70297104e-02 3.50866586e-01
4.83304828e-01 -1.17416906e+00 4.53103483e-01 1.16774595e+00
2.67032295e-01 5.12654781e-01 9.71436322e-01 -4.78806227e-01
-1.18228972e+00 -1.46591234e+00 5.02904654e-02 -5.36336303e-01
6.73832417e-01 -3.43097419e-01 -9.04225111e-01 4.85598922e-01
1.66558340e-01 2.80203819e-01 5.18245161e-01 -3.79116774e-01
-6.20006561e-01 -4.16772366e-02 -1.51057124e+00 8.79916012e-01
9.03261125e-01 -5.27064562e-01 -5.69571495e-01 3.06325406e-01
6.54328763e-01 -2.73963094e-01 -6.39863849e-01 4.50101286e-01
1.43256649e-01 -9.93158698e-01 1.19074750e+00 -6.03757143e-01
4.48500216e-01 -2.80372083e-01 -2.00198129e-01 -1.59970081e+00
-2.96237618e-01 -8.64243031e-01 -2.03134552e-01 9.53054368e-01
1.38948202e-01 -7.51160324e-01 8.42337728e-01 4.50413749e-02
1.70424283e-01 -5.08211017e-01 -9.02299762e-01 -8.06994796e-01
2.44824857e-01 -2.80662358e-01 6.05219841e-01 1.00257039e+00
-3.44526142e-01 2.56846063e-02 -5.72091520e-01 6.20418429e-01
9.36878979e-01 -4.93831456e-01 9.35531139e-01 -9.81909692e-01
-4.32009697e-01 -3.68063688e-01 -8.97966087e-01 -6.11644804e-01
3.79805744e-01 -7.24516571e-01 1.02675175e-02 -5.20129263e-01
-8.71014148e-02 -3.50205749e-01 -2.01724246e-01 4.24361765e-01
-4.12889808e-01 7.44433463e-01 1.67096719e-01 2.56048650e-01
-1.86066683e-02 6.41448319e-01 1.17563856e+00 -5.62934160e-01
1.25674129e-01 2.33541653e-01 -9.98331428e-01 8.80009890e-01
7.07200885e-01 -7.17536628e-01 -5.19349635e-01 -1.24162681e-01
5.01677468e-02 -3.52945000e-01 3.94210011e-01 -1.19027984e+00
-1.27312735e-01 -4.27014828e-02 4.48527664e-01 1.46539032e-01
2.09149376e-01 -9.07149136e-01 1.73903793e-01 7.07190394e-01
-5.03202856e-01 -3.03447932e-01 1.59927458e-01 8.27639163e-01
-2.62690455e-01 -3.01387250e-01 1.39500070e+00 -2.98353374e-01
-2.48191118e-01 4.91628319e-01 -3.45853269e-01 5.30916862e-02
1.41065514e+00 1.02925692e-02 -2.11190730e-01 -4.72497344e-01
-8.03133667e-01 -3.46205443e-01 3.67962450e-01 3.74564201e-01
5.88083923e-01 -1.33278668e+00 -8.31292272e-01 3.93365711e-01
-4.78444770e-02 -1.19908698e-01 2.79803216e-01 3.44388515e-01
-3.35127801e-01 -3.71112414e-02 -3.23334455e-01 -3.75621557e-01
-1.15677202e+00 9.93456900e-01 3.37548405e-01 -3.54434788e-01
-6.44708395e-01 1.05668902e+00 5.41436970e-01 -1.77555770e-01
3.08681309e-01 2.65409917e-01 -3.48701596e-01 -9.26487148e-02
7.76127994e-01 4.36391383e-02 3.28038931e-02 -7.09932327e-01
-2.12353602e-01 3.68602008e-01 -3.81293118e-01 2.17599124e-01
1.08017409e+00 2.84004271e-01 1.36336640e-01 -3.18315774e-01
1.32864213e+00 3.64346594e-01 -1.25348127e+00 -3.17405343e-01
-5.44597507e-01 -7.03437150e-01 2.96452567e-02 -4.69286233e-01
-1.35362864e+00 5.21884084e-01 5.32409966e-01 6.07847154e-01
1.28640890e+00 -2.07831606e-01 7.05885887e-01 3.51005137e-01
2.20343933e-01 -4.12360281e-01 1.58484206e-01 1.77514717e-01
9.10974443e-01 -1.11245787e+00 -2.09720969e-01 -3.64119381e-01
-5.19365013e-01 7.96374202e-01 5.81528306e-01 -5.48682332e-01
5.48328757e-01 4.92285967e-01 5.31373322e-02 -8.56071860e-02
-3.77110064e-01 2.69040503e-02 1.34614825e-01 8.87414873e-01
-3.50415230e-01 -1.91094708e-02 -1.03195660e-01 2.60462075e-01
-4.19457197e-01 -6.39842272e-01 5.67005873e-01 5.46796560e-01
-1.68744013e-01 -1.14001656e+00 -7.93496072e-01 3.68136913e-01
-6.28636897e-01 -8.81165415e-02 -6.52866006e-01 5.94373345e-01
1.13117263e-01 8.88357997e-01 -2.95902222e-01 -6.97872698e-01
3.07328969e-01 -3.64765346e-01 3.48523349e-01 -5.24151146e-01
-7.63332605e-01 -4.02413785e-01 -2.96942979e-01 -5.65025806e-01
-1.81683049e-01 -4.64863956e-01 -5.66396415e-01 -4.47989345e-01
-2.50471801e-01 2.51818895e-01 2.77773559e-01 1.03425694e+00
7.02550188e-02 5.46887994e-01 1.35490453e+00 -9.43517268e-01
-1.05338919e+00 -9.25510824e-01 -5.25552571e-01 7.88152158e-01
3.44381154e-01 -5.06382823e-01 -9.44654882e-01 -1.13625512e-01] | [5.58610725402832, 7.912164688110352] |
3c0f54e1-3038-4942-88eb-1d1b05df0f4c | montage-based-3d-medical-image-retrieval-from | 1812.04118 | null | http://arxiv.org/abs/1812.04118v1 | http://arxiv.org/pdf/1812.04118v1.pdf | Montage based 3D Medical Image Retrieval from Traumatic Brain Injury Cohort using Deep Convolutional Neural Network | Brain imaging analysis on clinically acquired computed tomography (CT) is
essential for the diagnosis, risk prediction of progression, and treatment of
the structural phenotypes of traumatic brain injury (TBI). However, in real
clinical imaging scenarios, entire body CT images (e.g., neck, abdomen, chest,
pelvis) are typically captured along with whole brain CT scans. For instance,
in a typical sample of clinical TBI imaging cohort, only ~15% of CT scans
actually contain whole brain CT images suitable for volumetric brain analyses;
the remaining are partial brain or non-brain images. Therefore, a manual image
retrieval process is typically required to isolate the whole brain CT scans
from the entire cohort. However, the manual image retrieval is time and
resource consuming and even more difficult for the larger cohorts. To alleviate
the manual efforts, in this paper we propose an automated 3D medical image
retrieval pipeline, called deep montage-based image retrieval (dMIR), which
performs classification on 2D montage images via a deep convolutional neural
network. The novelty of the proposed method for image processing is to
characterize the medical image retrieval task based on the montage images. In a
cohort of 2000 clinically acquired TBI scans, 794 scans were used as training
data, 206 scans were used as validation data, and the remaining 1000 scans were
used as testing data. The proposed achieved accuracy=1.0, recall=1.0,
precision=1.0, f1=1.0 for validation data, while achieved accuracy=0.988,
recall=0.962, precision=0.962, f1=0.962 for testing data. Thus, the proposed
dMIR is able to perform accurate CT whole brain image retrieval from
large-scale clinical cohorts. | ['Cailey I. Kerley', 'Yuankai Huo', 'Shikha Chaganti', 'Mayur B. Patel', 'Shunxing Bao', 'Bennett A. Landman'] | 2018-12-10 | null | null | null | null | ['medical-image-retrieval', 'medical-image-retrieval'] | ['computer-vision', 'medical'] | [ 1.85406983e-01 -2.67102718e-01 4.73087355e-02 -2.25017205e-01
-1.03505862e+00 -3.65814596e-01 2.71603286e-01 5.26784539e-01
-8.59201610e-01 5.39521396e-01 -4.97728074e-03 -1.74166277e-01
-2.97521263e-01 -8.41303110e-01 -4.16167200e-01 -6.83872283e-01
-1.49289727e-01 9.47091043e-01 1.91354692e-01 2.68113345e-01
4.43198644e-02 8.03139091e-01 -1.40393686e+00 1.75808042e-01
4.90327209e-01 1.33350849e+00 3.30354452e-01 4.50825274e-01
1.41808549e-02 6.53710306e-01 -6.01261079e-01 8.07745606e-02
2.86631227e-01 -2.94828601e-02 -6.34573221e-01 -4.48134877e-02
1.94362059e-01 -8.45248103e-01 -5.11064649e-01 8.36811423e-01
7.91142881e-01 -4.32472443e-03 8.93488884e-01 -1.10509670e+00
-1.30052105e-01 1.38740689e-01 -6.73003137e-01 5.20961165e-01
1.74428839e-02 1.08378470e-01 4.40515101e-01 -1.01759517e+00
5.55201292e-01 6.28935933e-01 2.95591950e-01 3.10916275e-01
-5.71086168e-01 -1.08882213e+00 -4.40452278e-01 3.21767837e-01
-1.53585136e+00 -1.01014249e-01 3.37597370e-01 -6.34427011e-01
5.28372586e-01 2.20738068e-01 1.07109189e+00 5.45278430e-01
6.74678504e-01 5.30167639e-01 9.43799078e-01 -3.61405760e-02
1.74563974e-01 -4.63100314e-01 3.95100266e-01 5.37786543e-01
4.53398377e-01 -6.51236027e-02 -1.28845543e-01 -2.15701953e-01
7.94801176e-01 4.84019697e-01 -3.82066041e-01 -5.50656095e-02
-1.04482150e+00 6.32257283e-01 5.15792906e-01 3.62023622e-01
-6.91407859e-01 -7.16380849e-02 7.31400490e-01 1.33074194e-01
4.24089700e-01 -2.52495438e-01 -4.73417155e-02 9.86534655e-02
-1.04902959e+00 2.55368680e-01 1.02276936e-01 7.03795552e-01
3.44919533e-01 -2.84300476e-01 -2.54692584e-01 1.18234503e+00
1.62937772e-02 7.10130513e-01 1.16539252e+00 -3.76503021e-01
3.64952713e-01 7.00527072e-01 -2.43215591e-01 -7.27989674e-01
-8.06892276e-01 -3.73104155e-01 -1.08417392e+00 -1.58383492e-02
2.38626495e-01 8.67888182e-02 -1.27612019e+00 1.41212559e+00
3.08546871e-01 -2.70536870e-01 -2.27455884e-01 1.27477145e+00
9.82595265e-01 2.01956481e-01 1.92520902e-01 -3.10340792e-01
1.65143132e+00 -6.72645569e-01 -3.98574412e-01 -1.01047412e-01
5.25029182e-01 -6.33671284e-01 9.00445580e-01 3.95566404e-01
-1.01479745e+00 -8.66784602e-02 -8.97512674e-01 1.67878374e-01
-1.86556637e-01 3.90632719e-01 3.73240978e-01 3.49668831e-01
-8.31764102e-01 1.02895185e-01 -7.97295690e-01 -3.05704176e-01
6.76929832e-01 4.49755460e-01 -7.87629008e-01 -4.84226406e-01
-9.76023734e-01 8.05754662e-01 4.24748182e-01 -6.70442730e-02
-1.12462223e+00 -8.54758680e-01 -4.22966421e-01 3.51343527e-02
8.27617645e-02 -6.24444604e-01 1.15775073e+00 -4.34258789e-01
-8.48168373e-01 1.13010538e+00 7.89339021e-02 -3.63957882e-01
6.02768540e-01 -1.12724535e-01 -1.99524090e-01 7.17760980e-01
2.10175589e-01 5.49913943e-01 6.50920689e-01 -1.01217842e+00
-4.57073629e-01 -8.03488791e-01 -3.14456403e-01 6.83718249e-02
-2.33599916e-01 1.79884240e-01 -4.13401544e-01 -6.59751713e-01
4.03352499e-01 -9.27285731e-01 -1.02503342e-03 -5.87462224e-02
-1.28201604e-01 6.02348633e-02 6.45387292e-01 -9.29322898e-01
9.45494652e-01 -2.02389908e+00 -2.49599054e-01 3.82203907e-01
6.69163287e-01 1.77665696e-01 3.99589576e-02 7.00404868e-02
-6.27798259e-01 4.46798168e-02 -4.62256163e-01 -6.73617348e-02
-4.80331510e-01 -2.42432147e-01 1.57446176e-01 6.39519215e-01
-1.16084091e-01 8.35600019e-01 -5.24104595e-01 -6.56783640e-01
1.94625095e-01 3.63974363e-01 -2.43123680e-01 3.39301318e-01
5.54246366e-01 6.08614683e-01 -4.94073331e-01 7.91418254e-01
6.73591256e-01 -1.32884324e-01 -1.84472278e-01 -1.99837759e-01
1.19747862e-01 -2.88953334e-01 -7.71337748e-01 1.55928290e+00
-2.84573317e-01 3.86044621e-01 1.51104525e-01 -1.12962520e+00
6.38205469e-01 4.17909384e-01 1.19885123e+00 -8.89380932e-01
5.66508591e-01 2.96440333e-01 6.39338493e-02 -6.66866481e-01
-1.64504889e-02 -4.51953858e-01 1.38245905e-02 8.64225864e-01
-3.12863052e-01 -1.97448671e-01 8.16592276e-02 1.72042355e-01
1.35145700e+00 -7.46825159e-01 1.85172573e-01 5.69788106e-02
4.34815824e-01 2.91310430e-01 2.59880781e-01 5.02100229e-01
-4.38207448e-01 1.05892980e+00 1.38023883e-01 -4.83454704e-01
-1.03469682e+00 -1.10736752e+00 -4.50144500e-01 5.09570122e-01
-1.04223669e-01 -1.11508265e-01 -9.19493079e-01 -2.63537407e-01
-2.65155286e-01 1.85570404e-01 -3.94983858e-01 -4.20828581e-01
-6.04048133e-01 -1.01484871e+00 6.64459646e-01 4.51042175e-01
6.04181468e-01 -9.25722957e-01 -9.80685711e-01 6.06117770e-02
-3.63275796e-01 -8.94988000e-01 -5.02678216e-01 -9.69749317e-02
-1.02315652e+00 -1.47529435e+00 -1.14737093e+00 -5.51045954e-01
7.68345773e-01 3.78501594e-01 6.83532536e-01 5.11360884e-01
-8.34304035e-01 4.77948308e-01 -3.27920288e-01 -3.79360855e-01
-3.46520059e-02 -1.39061928e-01 1.51044458e-01 -1.89599916e-01
-6.92023244e-03 -4.88700658e-01 -1.03077269e+00 3.95335406e-02
-1.39329875e+00 -1.28639981e-01 8.83498549e-01 1.00011504e+00
8.07335019e-01 1.66858435e-01 3.99498880e-01 -5.81950963e-01
7.22767949e-01 -6.55796349e-01 -2.66949058e-01 2.78152078e-01
-5.28957248e-01 -5.21060050e-01 4.46845710e-01 -4.89554048e-01
-6.04940891e-01 -1.43951401e-01 -9.96520817e-02 -6.63085639e-01
-2.71726519e-01 7.57806599e-01 1.99366137e-01 1.23075262e-01
4.15350765e-01 3.67421210e-01 2.74471462e-01 -4.36201215e-01
-3.59039187e-01 9.01300609e-01 5.35954833e-01 -5.50809383e-01
2.90033072e-01 4.43495929e-01 9.91986096e-02 -6.52553082e-01
-3.63004029e-01 -6.68309748e-01 -6.52156532e-01 -4.29844171e-01
8.21941018e-01 -7.81330228e-01 -5.53491950e-01 5.83480418e-01
-7.70700276e-01 -1.62519604e-01 7.40718618e-02 8.46280932e-01
-3.50596786e-01 4.53540146e-01 -5.99166691e-01 -5.53551972e-01
-1.15341556e+00 -1.55231500e+00 1.13069463e+00 -1.21562816e-01
-1.85306191e-01 -4.31243569e-01 -7.58962333e-02 6.85255945e-01
4.75804985e-01 1.82762101e-01 1.35141695e+00 -7.58968055e-01
-1.94932476e-01 -7.51075268e-01 -5.35447896e-01 2.48599172e-01
1.94172189e-01 -3.04867655e-01 -5.97113311e-01 -4.21006829e-01
1.57040104e-01 -4.06640708e-01 6.62943065e-01 4.61137742e-01
1.36119616e+00 2.02826768e-01 -1.89617470e-01 2.87060201e-01
1.34703684e+00 5.86525857e-01 6.10360682e-01 3.59313369e-01
5.03970504e-01 5.27585387e-01 6.29800618e-01 4.27547693e-01
2.69309372e-01 3.86541963e-01 3.87267143e-01 5.74156269e-02
-1.82304844e-01 2.88869709e-01 -3.32468711e-02 8.11752737e-01
-4.00352031e-02 -5.01934066e-02 -1.46264529e+00 5.89156866e-01
-1.32089365e+00 -5.32165408e-01 -1.76968619e-01 2.42530251e+00
4.84027982e-01 -1.32294685e-01 3.28625925e-02 3.02549273e-01
9.19303000e-01 -3.98508936e-01 -6.70135558e-01 2.99980193e-01
2.42818311e-01 6.75501406e-01 4.29809839e-01 -9.48879216e-03
-9.21269357e-01 4.82427031e-01 5.96437788e+00 8.86679173e-01
-1.43405080e+00 3.47523868e-01 7.16110647e-01 -4.47238654e-01
2.39788547e-01 -4.38435614e-01 -1.16354130e-01 4.60794717e-01
8.06635499e-01 -1.21691100e-01 1.97856560e-01 5.43220103e-01
3.33245784e-01 -4.94579762e-01 -6.84948862e-01 1.26899052e+00
4.88318801e-02 -7.82320440e-01 2.51242727e-01 2.29161251e-02
1.34957388e-01 2.69208163e-01 -6.12957701e-02 1.89968899e-01
-1.80945888e-01 -1.18462825e+00 7.02805221e-01 4.53895181e-01
1.33301485e+00 -7.39614248e-01 9.04989421e-01 4.29660976e-01
-9.95670259e-01 -6.30949065e-02 -2.60287344e-01 1.89401865e-01
5.35076894e-02 7.66761839e-01 -8.75470698e-01 6.22958541e-01
9.23049867e-01 2.18158916e-01 -4.52622354e-01 1.20707834e+00
8.33102465e-02 4.54729557e-01 -3.64066511e-01 3.52204531e-01
7.74066672e-02 -6.34231791e-02 3.44720274e-01 8.22218597e-01
5.30479193e-01 6.28068686e-01 1.64130941e-01 3.68534714e-01
-1.34813070e-01 4.29151654e-01 -3.90396595e-01 2.02435374e-01
3.34818006e-01 1.21440732e+00 -9.94540691e-01 -4.90044266e-01
-1.34602919e-01 4.40085471e-01 2.01741815e-01 9.17213708e-02
-7.98730433e-01 -1.32646263e-01 2.76280612e-01 2.23381847e-01
-3.24985772e-01 -5.07348888e-02 -2.85328865e-01 -1.05694866e+00
3.00760437e-02 -7.93064058e-01 6.83323085e-01 -9.80127931e-01
-1.25070930e+00 5.68570673e-01 2.86354065e-01 -1.26937354e+00
7.56776184e-02 -5.00167310e-01 -5.36455393e-01 7.94964254e-01
-1.01803994e+00 -9.10229921e-01 -7.33275533e-01 7.40093529e-01
4.14266407e-01 -1.23949751e-01 6.19949758e-01 6.49725974e-01
-7.14408875e-01 3.17229360e-01 6.54643252e-02 4.11554813e-01
7.43829489e-01 -6.85022116e-01 -4.05233800e-01 4.32596743e-01
-5.14362812e-01 6.52728677e-01 1.01077937e-01 -7.30374634e-01
-1.23496473e+00 -1.11801255e+00 4.51065987e-01 6.01907708e-02
4.48605031e-01 2.58143067e-01 -8.23738754e-01 4.54951614e-01
-3.08061391e-01 1.13917693e-01 8.23807776e-01 -6.68766916e-01
-1.02776147e-01 -6.37884587e-02 -1.47991312e+00 2.76951849e-01
6.33948326e-01 -3.31275940e-01 -5.77429354e-01 4.76109505e-01
2.20933244e-01 -5.53138673e-01 -1.36840558e+00 6.27726436e-01
8.74637187e-01 -4.97406602e-01 1.08321857e+00 -4.59351808e-01
4.61228877e-01 -7.86618888e-02 -1.23589478e-01 -9.36515570e-01
-1.26566067e-01 6.07988596e-01 4.48393017e-01 6.48358881e-01
-1.27025619e-02 -7.27108896e-01 5.90094090e-01 8.49166632e-01
-3.20938945e-01 -8.40805650e-01 -1.22715890e+00 -4.75808650e-01
3.96277994e-01 -7.08359957e-01 6.09246433e-01 6.86039865e-01
-2.71953374e-01 -1.99479640e-01 2.28888527e-01 4.46985429e-03
5.98285615e-01 8.51246268e-02 4.04562831e-01 -1.24511909e+00
1.45803362e-01 -5.27805269e-01 -5.86811662e-01 -3.46713424e-01
-7.29239210e-02 -1.08611739e+00 -2.65583694e-01 -1.64860630e+00
7.31779277e-01 -6.15879416e-01 -6.10540986e-01 5.11887252e-01
8.41047391e-02 5.38782775e-01 3.39124762e-02 6.85216904e-01
-5.71078621e-02 3.33783478e-01 1.34083223e+00 -4.63812888e-01
2.39822239e-01 -2.09515229e-01 -2.03192383e-01 4.30964768e-01
9.72616196e-01 -6.44312322e-01 -4.81046587e-01 -4.83342171e-01
-2.77660310e-01 3.76145929e-01 5.04498124e-01 -1.09443748e+00
2.90001214e-01 -3.02970018e-02 5.99431157e-01 -8.39925945e-01
3.44970107e-01 -8.59088242e-01 4.18708920e-01 7.28950024e-01
4.04282585e-02 1.29370242e-01 1.49165943e-01 2.30643123e-01
-5.11078715e-01 -3.13005477e-01 7.74097145e-01 -3.58640611e-01
-3.09246689e-01 8.27593565e-01 -6.64083064e-01 -8.96931253e-03
1.13806641e+00 -2.02049121e-01 -1.74489394e-01 -1.44225672e-01
-8.08292091e-01 -1.39940586e-02 2.81890064e-01 1.29260719e-01
9.69311357e-01 -1.28234220e+00 -6.19744420e-01 7.02054352e-02
2.50852317e-01 2.75079727e-01 5.52536905e-01 1.33447170e+00
-8.88760865e-01 4.77108151e-01 -4.15757418e-01 -6.86135411e-01
-1.39989448e+00 2.38992080e-01 2.60067940e-01 -3.17898244e-01
-8.96754444e-01 3.89650702e-01 3.36313218e-01 -8.96099657e-02
-2.62452904e-02 -3.56484264e-01 -2.64108211e-01 2.23175213e-01
6.66364610e-01 2.56219029e-01 6.23089671e-01 -8.25557530e-01
-4.50555712e-01 6.88573301e-01 -1.94568917e-01 -2.07818702e-01
1.47013795e+00 1.20964602e-01 -4.20982063e-01 1.78745896e-01
1.38431597e+00 -4.33452845e-01 -4.12551194e-01 -5.61797665e-03
-3.73028547e-01 -3.94358844e-01 3.16378683e-01 -6.41673625e-01
-1.53470075e+00 1.00633097e+00 9.71674383e-01 -2.08225235e-01
1.40107393e+00 -5.87730622e-03 1.05447781e+00 2.77895868e-01
6.61637008e-01 -6.93587244e-01 -1.62071720e-01 3.76358956e-01
9.89160478e-01 -1.08887243e+00 7.28290603e-02 -1.79189757e-01
-4.19262707e-01 1.26443970e+00 3.98641050e-01 -3.68986093e-02
7.08485782e-01 8.02139640e-02 -2.64441315e-02 -6.28443897e-01
-4.09038991e-01 1.18861906e-01 1.16366677e-01 2.12705448e-01
3.67788017e-01 1.58998802e-01 -5.93080759e-01 8.73712301e-01
-2.61079669e-01 6.38614073e-02 2.48337895e-01 1.23778927e+00
-3.78664851e-01 -7.88007617e-01 -7.23899126e-01 1.03815150e+00
-6.45752013e-01 -5.58094010e-02 -3.05169821e-01 8.54857802e-01
-2.75973696e-02 7.62242436e-01 2.80724429e-02 -3.29937845e-01
3.38384300e-01 2.83799529e-01 3.88261288e-01 -5.38088083e-01
-6.23658001e-01 -1.13320395e-01 -3.20100665e-01 -3.17446470e-01
-9.73016322e-02 -4.79500592e-01 -1.59403694e+00 -1.13201633e-01
-1.85702235e-01 -1.15915146e-02 8.05873096e-01 9.55097020e-01
1.60109028e-01 5.19963503e-01 3.72426629e-01 -6.67885065e-01
-9.55675617e-02 -1.16493213e+00 -6.36955202e-01 5.50787449e-01
2.00249806e-01 -1.03124213e+00 -2.09791943e-01 8.11792910e-03] | [14.521520614624023, -2.0650343894958496] |
87945a34-aeae-4cd6-9bfb-3648a0f70b80 | apollo-a-simple-approach-for-adaptive | 2212.09282 | null | https://arxiv.org/abs/2212.09282v2 | https://arxiv.org/pdf/2212.09282v2.pdf | APOLLO: A Simple Approach for Adaptive Pretraining of Language Models for Logical Reasoning | Logical reasoning of text is an important ability that requires understanding the information present in the text, their interconnections, and then reasoning through them to infer new conclusions. Prior works on improving the logical reasoning ability of language models require complex processing of training data (e.g., aligning symbolic knowledge to text), yielding task-specific data augmentation solutions that restrict the learning of general logical reasoning skills. In this work, we propose APOLLO, an adaptively pretrained language model that has improved logical reasoning abilities. We select a subset of Wikipedia, based on a set of logical inference keywords, for continued pretraining of a language model. We use two self-supervised loss functions: a modified masked language modeling loss where only specific parts-of-speech words, that would likely require more reasoning than basic language understanding, are masked, and a sentence-level classification loss that teaches the model to distinguish between entailment and contradiction types of sentences. The proposed training paradigm is both simple and independent of task formats. We demonstrate the effectiveness of APOLLO by comparing it with prior baselines on two logical reasoning datasets. APOLLO performs comparably on ReClor and outperforms baselines on LogiQA. The code base has been made publicly available. | ['Xiang Ren', 'Chenguang Zhu', 'Wenhao Yu', 'Reid Pryzant', 'ZiYi Yang', 'Shuohang Wang', 'Yichong Xu', 'Soumya Sanyal'] | 2022-12-19 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [ 0.15934381 0.58999836 -0.1497784 -0.6457775 -0.4813157 -0.6592972
0.68285567 0.5053576 -0.62176114 0.56649154 0.25411007 -0.8752408
-0.12570207 -1.1020671 -1.2409528 0.016948 0.02220781 0.53842455
0.16830592 -0.43841046 0.23630533 0.24763513 -1.2229255 0.9581489
1.1689507 0.80378616 0.17862377 0.59799594 -0.6004256 1.8530538
-0.48044148 -0.75744474 -0.06700571 -0.36769447 -1.2220783 -0.3711774
0.67692477 -0.29910696 -0.26623198 0.9136997 -0.09829564 0.06760163
0.69204307 -1.168346 -0.9101191 1.4548252 -0.07390325 0.32297468
0.4943799 0.15017535 1.2883847 -0.91480684 0.5465345 1.5860977
0.8541736 0.47079703 -1.2403753 -0.54737705 0.49195 0.53831667
-1.2398344 -0.41656372 0.69920164 -0.42048615 1.4668331 0.00656102
0.224514 0.9940698 0.27608132 0.91083777 1.085622 -0.77156883
0.19521335 0.33814204 0.5994972 1.0989484 0.26766214 -0.46747577
-0.81720513 0.11586573 0.09878157 -0.41870093 -0.06512071 -0.24948156
-1.0066808 0.62357044 0.17852522 -0.02383806 -0.17814733 0.20561145
0.55788237 0.626777 0.0434798 0.5874765 -0.7108038 0.14644673
-0.80777 0.37806505 1.0956476 0.9680302 0.3583051 -0.23051812
-0.307308 0.67404413 0.4661644 0.54908586 0.2899279 -1.0529202
0.9321452 0.7981337 -0.24316809 -0.77578974 -0.3518245 -0.24748605
-0.4330364 0.01148002 0.5269648 -0.06718049 -0.39377135 2.0956283
-0.05873837 -0.1738623 0.5351414 0.37809408 0.99251586 0.40630394
0.213459 0.1857252 1.5565387 -0.94248 -0.7733218 -0.558201
0.90020347 -0.38498574 1.5588437 0.5647488 -1.4414017 -0.38356796
-1.2221489 -0.7312563 -0.6538229 0.03212276 0.78939086 0.3596661
-0.88088316 0.3670429 -0.68509805 -0.18829736 0.5108762 -0.01133745
-0.14981431 -0.22103618 -1.6530349 1.0804768 0.70470446 0.05557576
-0.48156625 -0.8534454 -1.2435452 0.36373642 0.7806883 -0.7193647
1.3737857 -0.5801633 -1.3329633 0.9403957 -0.14295094 -0.83151275
0.62916654 -0.42748308 -0.16877769 0.11631358 0.17378096 0.50387233
0.52236336 -0.9918103 -0.33861837 0.00731793 0.6078351 0.15833303
-0.21302651 -0.1658284 -0.24031402 -0.53893286 0.06835009 -0.55212235
0.35264558 0.12071611 -0.5859899 -0.6497244 0.34568074 -0.73484766
1.2263625 -1.942726 0.06021742 0.06895074 0.18486267 -0.03325741
-0.18377788 0.36886594 -0.09104869 0.33660942 -0.21775985 -0.24711835
0.3421831 0.1999639 -0.7318352 -0.06494798 0.49934158 1.0336221
-0.8039421 -0.72635895 -0.11064982 -0.14803752 -0.8747801 0.14578018
-0.835229 -0.04631886 -0.12743908 0.4179402 0.5862095 -0.37754348
0.4756137 -0.21759886 0.19704862 0.9735661 -1.025694 1.6792003
-0.71413565 0.53446764 -0.31315756 -1.1767342 0.5476507 0.0088154
-0.27787268 -0.6463371 0.03346964 -0.03057814 0.30107588 -0.7640238
0.07953808 -0.22450848 -0.2610128 0.6114409 0.13173576 -0.31062356
0.6553458 0.67299306 1.1157391 0.3036492 0.3053689 -0.22144753
0.88471097 0.09131449 0.2887501 1.1344724 0.19044667 -0.1130865
1.003039 -0.1593596 -0.76162606 -1.242344 -0.1057224 1.1620035
-0.08514768 -0.64976215 -0.62949055 -0.72072357 0.2756207 1.3846499
-0.42065826 -0.3293811 -0.81159884 -0.30691925 0.92749226 0.715964
0.6946469 -1.195004 -0.34886217 -0.03376846 -0.2986986 -1.574296
-0.02803176 0.37034914 -0.7572309 -1.2789832 0.3712639 -0.7988418
0.73968506 -0.16683598 1.4167928 0.38391536 0.10029982 0.43086645
-0.20216708 -0.5270701 -0.71872395 -0.0562214 -0.02097636 -0.4075017
0.57289636 -0.35631546 0.14230773 -0.4382322 -0.7192627 0.38894653
0.6388295 0.8014675 0.2104832 0.45314598 0.27295694 -1.2732337
0.80392903 -0.53922415 -0.39332968 0.61344624 -0.5289166 0.61672145
1.0420692 -0.15763526 -1.3545709 -0.37196422 0.03700637 0.04358706
0.13218427 0.96853423 -0.1524562 0.35789314 0.54506135 0.29040864
0.06003286 -0.2354113 0.53293926 0.3403116 0.7307658 -1.2990052
0.98678404 0.15536773 -0.05771914 -0.70473015 -1.2652565 0.14175835
-0.6535145 0.32746336 0.61159223 -1.05318 -1.0674424 0.36465046
-1.3562154 -0.76344514 -0.17878224 0.23807405 -0.47695947 0.43901893
-0.7812401 -0.70901316 -0.22864653 -0.78447133 0.5407102 -0.16640726
-0.57952106 -1.2526321 -0.46111965 0.725494 0.1726842 -0.1497797
1.7571701 -0.9686879 -0.66833603 0.01542237 -0.27286667 0.62147963
-0.20987827 -0.06640512 -0.77157664 0.1297179 0.13179994 -0.91421676
0.9586508 -0.05365436 1.4678736 -0.5594914 -0.05131881 0.3855421
1.026157 -0.10353937 0.4757249 0.3757298 0.51812565 0.6522709
0.30991402 0.09420678 1.0636505 0.24214523 -0.05697811 0.1969731
-0.22501804 -0.5191882 0.52179533 0.63266134 0.3493527 -0.17656702
-1.2699121 0.2827673 -1.6935481 -1.0939126 0.01925384 1.7408594
1.5622851 0.6637946 -0.42256966 0.10540053 0.08703849 0.00790025
-0.59598786 -0.5875657 -0.23876923 0.38802564 0.15720075 1.053355
-0.97953886 1.1748029 5.818572 0.47558603 -0.8548287 -0.11510418
0.19287269 -0.11803424 -0.49376136 0.02951858 -0.843867 0.11600627
0.7922199 -0.2385753 0.5505023 0.5363521 0.16193224 -0.36495665
-1.7026507 0.48502842 0.27667892 -1.3368764 0.35728866 -0.5731633
0.21157394 -0.30684543 -0.06782854 0.89647233 0.51287866 -1.2149465
1.2157633 0.6144191 0.40071294 -0.3665287 0.7491081 0.5830163
-0.8993125 -0.23050846 -0.02049144 -0.4984407 -0.15988548 0.34540942
-0.81853163 0.32647258 0.51952773 0.6640601 -1.0162491 0.15576228
-0.9070597 0.56594807 -0.20563868 -0.24518558 0.0583126 0.11579358
0.2035707 1.3278699 -0.31060243 0.21811785 0.20292537 1.3318487
-0.35888153 -0.10527667 -0.44223437 -0.29640555 0.68065566 0.7839487
-0.37731338 -0.622678 -0.71234083 0.79691255 0.8460472 0.28697458
-0.87572145 -0.47607258 0.3397749 0.07728815 0.14074156 -0.23252808
-0.68817234 -1.3653098 0.39178807 -0.9966155 0.56866324 -0.9623607
-1.382782 0.22717573 0.38311917 -0.47471595 -0.1822795 -0.9847509
-0.5881202 1.0360383 -1.680392 -1.077003 -0.1631686 0.33626074
0.6184784 -0.23159012 0.6260127 -0.10999367 -0.47183385 0.85002494
-0.49602968 0.5758355 0.5453945 -1.4421737 0.3632376 0.9188706
-0.06900248 1.3235426 0.7391957 -0.6170032 -1.6224469 -0.8889045
1.3608489 -0.7941777 1.0795187 -0.69620067 -1.1431036 1.1409191
0.29150265 -0.26314396 0.67666346 0.38197598 -0.8778314 -0.1053438
-0.9729077 0.9357778 1.0617669 -0.90004325 -1.4018838 0.54454625
0.7695594 -0.5489662 -0.60752726 0.15675922 0.53565925 -0.6114563
0.95626706 -1.0367062 0.83328694 -0.2119076 -0.17574878 -0.86677814
-0.08146263 -0.21607238 -0.31840333 1.0458653 0.65702546 -0.6205657
0.49112025 0.8032386 -0.05988827 -0.7990655 -0.4979626 -0.84148604
0.56765145 -0.6777955 0.26653385 0.8984441 0.5788921 0.7894786
0.3160424 0.3488607 0.6513862 0.05604899 0.69106036 -1.1395766
-0.31672773 -0.6204583 0.12946166 -0.9941229 0.978016 -1.5396729
-0.03734071 -1.7523167 0.21832785 -0.28890112 -0.04158181 1.0093048
-0.11503696 -0.47617024 0.17488456 -0.1612907 -0.5955605 0.36428124
0.9932895 -0.3662203 0.04680458 -0.43098405 -0.8546357 1.0684329
0.64102334 -0.25974283 -0.6907216 -0.63211435 0.62460077 -0.16264883
0.54844785 -0.8976877 0.4294652 -0.2655735 0.19209486 -0.4953066
-0.04553789 -0.62677824 -0.5534878 0.5787847 -0.9578193 0.05081723
0.53610504 0.2871065 -0.09861487 -0.46045876 0.60578275 -0.23214848
-0.7578371 -0.37950864 -0.38587987 0.62402016 0.6671232 0.06227168
-0.5921194 -0.12977818 -0.5301078 0.54776406 0.12022362 0.3910373
0.52691346 -0.9961095 -0.65211517 0.12055655 0.16435 0.1590917
0.03208528 0.9075397 -0.795019 0.67568946 0.08553676 -0.17758305
-1.0600173 0.64356685 0.46605754 -0.41544932 -0.69670993 0.9719864
0.07050593 -0.84564924 0.48547414 -0.9737755 -0.16536568 -0.14821592
0.48340777 -0.04466131 0.12406407 0.02388419 -0.63245887 0.20112807
-0.28894737 0.05637962 1.1188321 0.03158901 -0.49937648 0.68656623
0.78913224 0.19637467 -0.79866326 -0.6105183 0.32936567 0.12356198
-0.1580537 -1.2939986 -0.34848142 0.9823381 -0.29467478 -0.01306913
0.87006253 0.13891542 0.5806083 1.1679933 0.15156564 -1.036872
0.25116453 1.1974585 0.975985 -1.339409 0.1744129 -0.59538406
-0.5094822 1.2265474 0.8374093 0.10019749 0.49446556 0.58217335
-0.05688331 -0.27231637 -1.1839715 0.09056448 0.21272019 0.15380423
0.80704266 -0.18177171 -0.31882116 1.0561533 -0.59300065 0.11651769
0.48601654 0.9921349 -0.18181752 -0.74714303 -0.25901735 0.43289337
-0.16964342 -0.6492661 -0.4464045 0.8619307 0.10180348 0.83184105
0.04033233 0.05938531 0.4824305 0.60515 0.7815959 -0.8153079
-0.3585032 -0.8655981 0.37692413 -0.31893438 -0.08607589 -0.6161354
-1.7925782 -0.5258882 0.1252251 -0.0368279 0.20837668 1.3974417
0.08114468 0.7570105 -0.08753642 -0.01110577 -0.9997599 -0.8428463
-0.1483651 0.48012498 0.29414836 -0.55471563 -0.32342124 0.26932392] | [9.658500671386719, 7.485533237457275] |
48311cfc-c19a-4dd8-b7dc-03870781258f | membership-inference-attacks-on-lottery | 2108.03506 | null | https://arxiv.org/abs/2108.03506v1 | https://arxiv.org/pdf/2108.03506v1.pdf | Membership Inference Attacks on Lottery Ticket Networks | The vulnerability of the Lottery Ticket Hypothesis has not been studied from the purview of Membership Inference Attacks. Through this work, we are the first to empirically show that the lottery ticket networks are equally vulnerable to membership inference attacks. A Membership Inference Attack (MIA) is the process of determining whether a data sample belongs to a training set of a trained model or not. Membership Inference Attacks could leak critical information about the training data that can be used for targeted attacks. Recent deep learning models often have very large memory footprints and a high computational cost associated with training and drawing inferences. Lottery Ticket Hypothesis is used to prune the networks to find smaller sub-networks that at least match the performance of the original model in terms of test accuracy in a similar number of iterations. We used CIFAR-10, CIFAR-100, and ImageNet datasets to perform image classification tasks and observe that the attack accuracies are similar. We also see that the attack accuracy varies directly according to the number of classes in the dataset and the sparsity of the network. We demonstrate that these attacks are transferable across models with high accuracy. | ['Amol Deshpande', 'Shruti Bidwalka', 'Shishira R Maiya', 'Aadesh Bagmar'] | 2021-08-07 | null | https://openreview.net/forum?id=4lyXal2ZWB3 | https://openreview.net/pdf?id=4lyXal2ZWB3 | icml-workshop-aml-2021-7 | ['membership-inference-attack'] | ['computer-vision'] | [ 2.03057796e-01 2.20839426e-01 -1.87143892e-01 -3.99157196e-01
-2.55773932e-01 -7.38927186e-01 5.97526908e-01 1.33038789e-01
-4.86931801e-01 7.67776251e-01 -4.63761598e-01 -6.73957229e-01
-1.18618555e-01 -1.14728606e+00 -1.12745702e+00 -6.39036655e-01
-3.98926705e-01 6.93360329e-01 3.83230001e-01 1.14544094e-01
4.38711345e-01 6.52940452e-01 -1.21737206e+00 4.95751411e-01
2.45002612e-01 1.15798438e+00 -7.52359271e-01 4.14902061e-01
5.05068481e-01 8.20018888e-01 -9.89494562e-01 -5.67211628e-01
5.39400399e-01 -3.39670777e-02 -8.79890919e-01 -3.19092304e-01
9.80965793e-01 -3.72003704e-01 -5.34487367e-01 1.48067653e+00
1.31240815e-01 -7.66306669e-02 5.66595078e-01 -1.73729670e+00
-9.61982235e-02 1.12801623e+00 -4.02290672e-01 2.55901664e-01
-2.79701889e-01 1.34201899e-01 8.31418693e-01 -5.73737919e-01
2.63190091e-01 1.33214772e+00 5.90442240e-01 5.69101572e-01
-1.43407834e+00 -1.48597944e+00 -3.94771923e-04 2.75995225e-01
-1.57509649e+00 -4.09903675e-01 3.63192290e-01 -3.25159937e-01
7.06036985e-01 3.78207922e-01 1.20740600e-01 1.31598949e+00
2.71745473e-01 1.39560133e-01 1.04455531e+00 -1.49281472e-01
6.57244146e-01 2.95698136e-01 5.11831880e-01 5.86151481e-01
6.62269354e-01 4.46199238e-01 -2.86979616e-01 -7.27270246e-01
5.95052779e-01 -1.80539846e-01 -2.53046542e-01 -2.18276933e-01
-6.43306553e-01 1.37812710e+00 3.87638986e-01 2.38432270e-02
-9.19738114e-02 4.92984831e-01 6.64842963e-01 5.65251112e-01
9.76222288e-03 6.26924455e-01 -4.52296972e-01 4.16150600e-01
-8.51856768e-01 1.91921532e-01 1.24460030e+00 3.82823259e-01
7.12704480e-01 1.82421312e-01 2.29077175e-01 2.07970455e-01
-3.70913781e-02 4.64588851e-01 2.41526932e-01 -9.29654002e-01
3.87505054e-01 1.22816503e-01 -1.71328202e-01 -1.18462920e+00
-8.27495568e-03 -3.85766596e-01 -9.88719106e-01 6.17839158e-01
6.99363172e-01 -4.25413311e-01 -6.69003308e-01 2.00006413e+00
7.05797970e-02 6.96297944e-01 -3.44263576e-02 4.57498103e-01
3.98037076e-01 4.85048741e-01 1.26384616e-01 2.48065948e-01
1.41493738e+00 -1.88552916e-01 -2.70063542e-02 -2.43002504e-01
5.92308640e-01 -3.96270931e-01 7.51821637e-01 6.57591820e-01
-8.19616497e-01 -3.19419533e-01 -1.46034837e+00 5.72805464e-01
-4.36653793e-01 -2.49432504e-01 5.82020342e-01 1.19386339e+00
-4.82541591e-01 8.02386582e-01 -8.00793111e-01 1.78404063e-01
9.58468080e-01 6.53270662e-01 -3.06914806e-01 1.62276886e-02
-1.52736604e+00 9.21701729e-01 7.50400424e-01 -2.94933498e-01
-1.46358037e+00 -7.87322879e-01 -5.97652733e-01 4.18402702e-01
3.31172436e-01 -3.22210759e-01 7.55363524e-01 -8.66598427e-01
-9.34775293e-01 6.59338355e-01 4.21771377e-01 -1.31569481e+00
3.61462265e-01 5.55710355e-03 -2.47834399e-01 2.21307695e-01
-1.99586719e-01 5.73602021e-01 1.15862405e+00 -9.21874821e-01
-5.48949480e-01 -3.94591510e-01 2.53837705e-01 -4.59459752e-01
-4.33978617e-01 -3.46290283e-02 3.11715573e-01 -5.34739912e-01
-1.88999280e-01 -9.97865498e-01 -6.53855130e-02 -9.31366161e-02
-8.58407021e-01 -1.24918394e-01 1.06473410e+00 -5.00120148e-02
1.06628501e+00 -2.12367606e+00 -5.47431707e-01 7.74616003e-01
3.30357760e-01 6.01048768e-01 -3.98572423e-02 6.26658946e-02
-4.57431167e-01 4.60212916e-01 -1.47242904e-01 3.33834663e-02
5.46503551e-02 3.04082245e-01 -1.20252132e+00 7.27938056e-01
-5.68488380e-04 4.83404845e-01 -3.00465912e-01 -8.86676982e-02
-2.41972189e-02 1.96397424e-01 -6.71989441e-01 5.11183543e-03
-2.45272130e-01 -1.00837559e-01 -2.01255068e-01 7.21317902e-02
8.76533508e-01 -2.47937813e-01 2.63691247e-01 -2.96101660e-01
5.30850947e-01 5.45217097e-01 -1.10292125e+00 5.00172973e-01
-1.88488588e-01 6.32149100e-01 -1.39512926e-01 -1.39749277e+00
8.08534145e-01 3.46177697e-01 -9.42345858e-02 -2.64213860e-01
2.94518381e-01 1.30611554e-01 4.68267918e-01 -3.19584459e-02
-3.05365678e-02 -2.24543303e-01 -1.18943378e-01 8.44426572e-01
-1.13593526e-01 2.24320069e-01 8.56209546e-02 1.85445607e-01
1.04660726e+00 -7.81553864e-01 -3.05069350e-02 -4.42196280e-01
3.34596664e-01 -6.40416667e-02 5.51725090e-01 1.34028137e+00
-8.63130689e-02 1.49283865e-02 7.94620812e-01 -7.07924366e-01
-1.03431559e+00 -1.36175907e+00 -5.35710931e-01 7.92239785e-01
-1.01924479e-01 -2.14027837e-01 -7.57931054e-01 -1.01812410e+00
1.31575704e-01 9.04620349e-01 -8.55561733e-01 -5.77276766e-01
-3.98521125e-01 -8.28500748e-01 1.34322977e+00 3.37454945e-01
7.22160697e-01 -1.01783216e+00 -5.79345822e-01 -8.46300349e-02
1.16830215e-01 -1.07850933e+00 -9.51354131e-02 1.29165456e-01
-8.22934508e-01 -1.43770003e+00 1.46934777e-01 -6.94389224e-01
7.95430005e-01 -2.61372387e-01 1.15097237e+00 1.96701318e-01
-2.70722568e-01 -1.76526010e-01 2.10699260e-01 -4.60731447e-01
-7.28463769e-01 7.06096292e-02 3.15869212e-01 -1.27468901e-02
6.86035812e-01 -8.11431289e-01 -3.94152403e-01 5.20138562e-01
-9.77780700e-01 -5.90211153e-01 2.53569633e-01 6.52203619e-01
1.46439284e-01 6.47118986e-01 8.03264976e-01 -1.25194013e+00
5.38285136e-01 -5.75185239e-01 -8.00406456e-01 1.39977887e-01
-2.98327982e-01 -1.19037397e-01 8.05737197e-01 -8.93527627e-01
-3.34366471e-01 -1.85735226e-01 -6.76605329e-02 -5.67352533e-01
-3.30740899e-01 4.04150128e-01 -1.57647267e-01 -9.41208899e-02
9.85065818e-01 -6.95955306e-02 -1.47593170e-02 -2.24368751e-01
1.43453315e-01 3.44248742e-01 6.82986736e-01 -6.27706647e-01
8.37006271e-01 7.17984796e-01 2.33854637e-01 -7.88780928e-01
-8.52169394e-01 2.07831845e-01 -1.66459844e-01 2.26980835e-01
5.22731423e-01 -5.62277377e-01 -1.13095236e+00 5.18556297e-01
-9.40961599e-01 -3.50725710e-01 -1.91084027e-01 3.73214424e-01
-1.79500863e-01 4.19640750e-01 -5.45948446e-01 -4.67147231e-01
-3.73657018e-01 -1.06024754e+00 5.66241965e-02 -7.23794550e-02
-3.99012119e-01 -9.12315845e-01 -3.67472559e-01 7.72551596e-02
2.17469394e-01 1.78270593e-01 1.28790474e+00 -1.44108164e+00
-5.01880884e-01 -5.62486172e-01 -3.03524703e-01 5.75287342e-01
-1.63071558e-01 -6.23444542e-02 -1.12930739e+00 -6.44448102e-01
3.71775180e-01 -7.05107272e-01 1.05972326e+00 3.55792582e-01
1.48818517e+00 -9.04471099e-01 -1.76008478e-01 4.04235214e-01
1.37486339e+00 6.00388721e-02 1.04153836e+00 2.48301670e-01
3.74645919e-01 4.71239477e-01 2.80336022e-01 2.59511083e-01
-4.92952317e-01 3.13591957e-01 7.46796906e-01 1.64395213e-01
4.20510739e-01 -1.86882377e-01 3.06892693e-01 -3.94950390e-01
5.92172801e-01 -1.61742136e-01 -7.68152237e-01 3.68525535e-01
-1.53351963e+00 -1.30852342e+00 2.38072738e-01 2.61177206e+00
1.00046563e+00 8.76152694e-01 5.23937307e-02 4.05354708e-01
8.34469199e-01 -1.07522003e-01 -5.32338917e-01 -5.86648166e-01
5.99486716e-02 5.10038972e-01 7.74864554e-01 4.86539751e-01
-1.30430508e+00 8.15781415e-01 6.38529110e+00 1.15401757e+00
-1.23375928e+00 -1.70318946e-01 1.10778487e+00 -1.47590280e-01
1.47549033e-01 1.28757894e-01 -1.00782979e+00 4.85998333e-01
1.08626616e+00 -9.48546007e-02 2.90196151e-01 9.98937786e-01
-2.86314696e-01 1.33048266e-01 -1.43893182e+00 6.28578782e-01
-3.54503244e-02 -1.68363035e+00 1.07580051e-01 2.65111446e-01
5.37278056e-01 2.82647938e-01 3.16436052e-01 3.12652856e-01
8.28300536e-01 -1.40662801e+00 4.67567325e-01 -8.69223699e-02
6.78467989e-01 -1.19440687e+00 6.29638195e-01 3.77037078e-01
-6.05717659e-01 -2.74051636e-01 -6.76219106e-01 1.04795061e-02
-3.04989994e-01 6.10532701e-01 -1.04054070e+00 -2.61263996e-01
6.31576955e-01 8.25502947e-02 -4.93307948e-01 8.59935343e-01
-2.52837151e-01 1.18495429e+00 -8.40552270e-01 2.11721659e-01
3.38704079e-01 1.24563456e-01 4.87317294e-01 8.18424523e-01
-2.11000249e-01 -3.38523299e-01 3.33902150e-01 1.17121458e+00
-3.89499307e-01 -6.29046500e-01 -6.25276089e-01 -5.89232780e-02
9.14334893e-01 8.17787230e-01 -5.70682406e-01 -5.95473230e-01
4.03716043e-02 2.36358047e-01 2.20156923e-01 1.67235032e-01
-7.92238176e-01 -5.73346853e-01 9.00298297e-01 1.95066497e-01
4.94316727e-01 -1.96601301e-02 -3.12979937e-01 -8.79434586e-01
-3.85885745e-01 -1.15467930e+00 6.41711831e-01 -1.70550764e-01
-1.61566818e+00 4.55407351e-01 -6.48912275e-03 -8.79458904e-01
-2.07006767e-01 -5.36240280e-01 -9.83451903e-01 7.65762329e-01
-1.02514076e+00 -7.19562769e-01 3.14857543e-01 7.33326614e-01
-1.42291784e-01 -4.55948442e-01 9.37153697e-01 1.43915921e-01
-7.54946887e-01 1.14349103e+00 -8.85009617e-02 9.38192368e-01
4.76471424e-01 -6.75858200e-01 5.72274089e-01 9.41503763e-01
3.55674207e-01 8.84179711e-01 8.28723729e-01 -5.73751092e-01
-7.47056663e-01 -1.18224776e+00 6.25009060e-01 -3.84451002e-01
8.18085790e-01 -3.08019698e-01 -1.07387829e+00 8.83246839e-01
-2.75674939e-01 1.04912527e-01 7.09829092e-01 1.07326195e-01
-1.07574272e+00 -2.81989593e-02 -1.38904655e+00 5.06800950e-01
2.82621920e-01 -6.89635813e-01 -6.08046412e-01 4.44961712e-02
5.85529625e-01 5.24236858e-02 -7.08107054e-01 3.52657825e-01
5.27628899e-01 -8.87801945e-01 1.27260876e+00 -1.02668977e+00
2.84789056e-01 5.55542149e-02 -2.50308543e-01 -8.02931368e-01
-1.30603999e-01 -2.69725591e-01 -1.04726508e-01 9.39564168e-01
4.27587390e-01 -1.05157864e+00 1.13764763e+00 6.21321797e-01
6.23168588e-01 -4.43425387e-01 -1.14810038e+00 -9.05352175e-01
5.28756320e-01 -2.82384068e-01 7.46561706e-01 1.15624177e+00
-6.00120611e-02 2.88813293e-01 -2.47071177e-01 5.63787818e-01
1.06352103e+00 1.01211518e-01 7.76368141e-01 -1.35740435e+00
-5.36279202e-01 -5.00207901e-01 -5.57122946e-01 -5.04555225e-01
3.04875791e-01 -7.85869539e-01 -4.09239262e-01 -3.93773943e-01
3.32332682e-04 -8.23291957e-01 -2.26287633e-01 7.55979598e-01
2.18323588e-01 6.89003766e-01 1.69418663e-01 2.95724541e-01
-1.15679637e-01 -7.64924148e-03 4.96082187e-01 -2.19772726e-01
8.44125003e-02 3.93430620e-01 -6.67562842e-01 8.61801684e-01
1.19722354e+00 -9.30564106e-01 -5.93157232e-01 3.79138328e-02
3.46233279e-01 -2.35238478e-01 7.85431027e-01 -1.08427155e+00
1.81658939e-01 2.31144950e-02 3.07768732e-01 -4.22239274e-01
1.94612756e-01 -8.38752151e-01 -1.22805014e-02 9.72567201e-01
-6.80669546e-01 -2.96099454e-01 3.45790505e-01 6.01216733e-01
1.74413338e-01 -4.95095819e-01 1.22595239e+00 5.68530476e-03
-1.69551224e-01 5.64920843e-01 -4.38827038e-01 3.62476379e-01
9.46948588e-01 -1.13907576e-01 -5.83163440e-01 -2.69486874e-01
-8.01023424e-01 2.22574007e-02 2.21967891e-01 1.15475506e-01
5.49353123e-01 -1.07332838e+00 -7.71770656e-01 5.15354693e-01
-2.35067114e-01 -1.94723114e-01 -1.00624420e-01 4.07826364e-01
-4.08398837e-01 1.85594723e-01 -4.65196639e-01 -5.08525491e-01
-1.33253384e+00 6.21602476e-01 6.03722334e-01 -3.33723128e-01
-5.29366136e-01 7.94056356e-01 4.40350920e-01 -1.34278312e-01
5.04362047e-01 1.06112406e-01 6.32251054e-02 -1.40088394e-01
8.40880513e-01 2.70164549e-01 1.00148506e-01 -1.91864759e-01
-2.88029402e-01 -1.11577012e-01 -4.73910928e-01 -3.17825675e-02
1.04472601e+00 4.58747387e-01 -1.04345746e-01 -2.86630299e-02
1.36191833e+00 -3.48502070e-01 -9.86359954e-01 -2.92728513e-01
-2.06801221e-01 -5.03004849e-01 8.69550481e-02 -6.71812057e-01
-1.34642625e+00 1.01019239e+00 4.46282774e-01 3.96210402e-01
6.79636300e-01 -2.30798036e-01 6.84137642e-01 6.33555830e-01
3.53824854e-01 -8.14178586e-01 4.98476140e-02 4.02962923e-01
5.43635368e-01 -8.31478238e-01 4.84379940e-02 -3.51104468e-01
-2.68250942e-01 8.50419641e-01 6.19682372e-01 -7.01730907e-01
1.03427315e+00 4.90766704e-01 -1.91445559e-01 -2.46705890e-01
-8.13958287e-01 6.09406531e-01 -1.70891210e-01 6.57091916e-01
-3.02510858e-01 -3.59354843e-03 1.98850095e-01 4.68969673e-01
-3.67427289e-01 -1.67175427e-01 6.38685167e-01 5.75380921e-01
-6.89603448e-01 -7.98251033e-01 -4.65899140e-01 6.44327223e-01
-7.96399117e-01 -1.11458473e-01 -3.44941765e-01 5.66089511e-01
-8.84410739e-02 8.81542981e-01 3.60488534e-01 -4.46311951e-01
-1.76878646e-01 7.29758069e-02 3.86192530e-01 -5.95088243e-01
-6.77392960e-01 -5.90576053e-01 1.48271948e-01 -4.90212142e-01
1.53434500e-01 -1.61089882e-01 -1.09934425e+00 -9.08353567e-01
-4.97548431e-01 2.82283306e-01 3.54662567e-01 8.99106383e-01
1.63170889e-01 -4.25301753e-02 7.82276690e-01 -4.34259623e-01
-1.29411519e+00 -7.59162545e-01 -6.92824781e-01 4.66976702e-01
1.27618626e-01 -4.13383812e-01 -8.59872580e-01 -3.11699539e-01] | [5.863270282745361, 7.41719913482666] |
c3c3732a-fcff-4dd6-bbdb-795414bae2c9 | rethinking-image-super-resolution-from-long | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Gou_Rethinking_Image_Super_Resolution_From_Long-Tailed_Distribution_Learning_Perspective_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Gou_Rethinking_Image_Super_Resolution_From_Long-Tailed_Distribution_Learning_Perspective_CVPR_2023_paper.pdf | Rethinking Image Super Resolution From Long-Tailed Distribution Learning Perspective | Existing studies have empirically observed that the resolution of the low-frequency region is easier to enhance than that of the high-frequency one. Although plentiful works have been devoted to alleviating this problem, little understanding is given to explain it. In this paper, we try to give a feasible answer from a machine learning perspective, i.e., the twin fitting problem caused by the long-tailed pixel distribution in natural images. With this explanation, we reformulate image super resolution (SR) as a long-tailed distribution learning problem and solve it by bridging the gaps of the problem between in low- and high-level vision tasks. As a result, we design a long-tailed distribution learning solution, that rebalances the gradients from the pixels in the low- and high-frequency region, by introducing a static and a learnable structure prior. The learned SR model achieves better balance on the fitting of the low- and high-frequency region so that the overall performance is improved. In the experiments, we evaluate the solution on four CNN- and one Transformer-based SR models w.r.t. six datasets and three tasks, and experimental results demonstrate its superiority. | ['Xi Peng', 'Hongyuan Zhu', 'Jiancheng Lv', 'Peng Hu', 'Yuanbiao Gou'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['image-super-resolution'] | ['computer-vision'] | [ 2.90921450e-01 -1.04433171e-01 -2.81394452e-01 -4.08109725e-01
-6.96896911e-01 6.33618683e-02 3.87341291e-01 -3.81346047e-01
-1.93742767e-01 7.26080239e-01 3.59779656e-01 1.78604275e-01
-3.44887406e-01 -7.39880800e-01 -6.39098108e-01 -1.09457147e+00
3.27075154e-01 -8.38888809e-02 5.95978379e-01 -1.77892298e-01
2.93105900e-01 6.76547214e-02 -1.62538266e+00 4.35263664e-01
1.23612273e+00 1.12163055e+00 6.59748256e-01 6.77344725e-02
-2.05039918e-01 9.68783915e-01 -2.18835756e-01 -2.20840308e-03
2.16972440e-01 -5.48833728e-01 -4.73770857e-01 8.90909806e-02
6.85964108e-01 -1.37386695e-01 -2.50096858e-01 1.39485788e+00
5.89794874e-01 -5.55423573e-02 4.50817794e-01 -8.92417431e-01
-9.16114032e-01 5.63271105e-01 -1.21593511e+00 6.08021617e-01
-2.61185598e-03 -1.32239666e-02 9.05262947e-01 -7.39333272e-01
3.46927881e-01 1.27817309e+00 6.09831393e-01 4.26429331e-01
-1.21963477e+00 -6.96696877e-01 1.84503987e-01 3.61263335e-01
-1.26547015e+00 -7.07671866e-02 1.04279232e+00 -4.19603825e-01
3.93436909e-01 -1.07473232e-01 3.59777868e-01 9.91575539e-01
4.64998692e-01 6.98127389e-01 1.59739578e+00 -3.06190938e-01
-4.43820879e-02 2.80857980e-01 5.93519658e-02 3.29021752e-01
1.63406283e-01 2.20904648e-01 -7.08383918e-01 1.63775131e-01
1.18950915e+00 -1.78327143e-01 -5.88438213e-01 -4.22012389e-01
-8.31639528e-01 6.88435197e-01 5.92206061e-01 5.72120547e-01
-2.86360443e-01 -2.27553040e-01 6.80247396e-02 -7.59936310e-03
6.64983153e-01 1.34857580e-01 -4.49955851e-01 2.91470706e-01
-1.01406038e+00 1.45320982e-01 7.97505304e-02 6.02625787e-01
7.54439831e-01 1.52245075e-01 -3.21443528e-01 1.11021078e+00
2.09599122e-01 2.79473633e-01 5.41668713e-01 -6.14765704e-01
3.02564085e-01 2.64420748e-01 1.83445036e-01 -1.15031779e+00
-2.55336136e-01 -9.51843321e-01 -1.11179066e+00 2.51095802e-01
4.91331100e-01 2.75706891e-02 -7.53277302e-01 1.94168139e+00
2.36723766e-01 5.10855556e-01 -1.83190212e-01 1.38088834e+00
7.80964375e-01 6.19308114e-01 -2.82220617e-02 -3.63525003e-01
1.44846988e+00 -1.03013110e+00 -9.63313639e-01 -2.38240451e-01
-1.85307652e-01 -8.15171599e-01 1.33243394e+00 4.48579252e-01
-1.12188256e+00 -1.07169425e+00 -1.05776739e+00 -2.52603680e-01
1.28121883e-01 2.02096537e-01 3.52802664e-01 2.84322143e-01
-9.13144946e-01 7.05356061e-01 -6.56365871e-01 -8.09774175e-02
3.24938834e-01 -9.95159447e-02 -8.38437304e-02 1.36913238e-02
-1.28319991e+00 8.50908339e-01 3.71318161e-01 1.39818117e-01
-4.00424212e-01 -8.22385252e-01 -5.03849745e-01 1.60261001e-02
2.54967958e-01 -5.74564695e-01 9.06562567e-01 -1.21715903e+00
-1.13215685e+00 9.14770663e-01 -7.83220008e-02 -3.35999399e-01
3.50964397e-01 -2.04737112e-01 -5.23119390e-01 1.34874851e-01
2.13334963e-01 2.86035001e-01 1.24234104e+00 -1.57056642e+00
-8.40587199e-01 -6.44692600e-01 -1.57393306e-01 1.83287814e-01
-3.14161032e-01 -7.86128715e-02 -1.12616047e-01 -8.64060462e-01
2.87024766e-01 -5.29528320e-01 1.48306876e-01 -1.20898247e-01
2.57594343e-02 -6.33769035e-02 7.39003062e-01 -7.61774242e-01
1.34992254e+00 -2.32366610e+00 2.84084886e-01 -2.54949301e-01
4.37079847e-01 4.00182307e-01 -3.57998274e-02 6.86667413e-02
-3.17185462e-01 -1.06242776e-01 -1.96957812e-01 -1.18982412e-01
-4.65598583e-01 6.91495314e-02 -4.87836331e-01 4.50943351e-01
2.07262605e-01 4.96510655e-01 -9.32325006e-01 -5.58221877e-01
6.88732490e-02 8.20631862e-01 -3.17615569e-01 2.68249899e-01
5.59179112e-02 8.36534679e-01 -5.52047849e-01 3.17442358e-01
1.17296827e+00 -3.41823876e-01 -9.72817317e-02 -6.80135787e-01
-4.14379507e-01 2.86112842e-03 -1.15589428e+00 1.50699675e+00
-2.52090842e-01 4.02767420e-01 2.08319843e-01 -9.20781255e-01
1.10787034e+00 -5.93872145e-02 4.68200028e-01 -1.13821185e+00
-7.91085213e-02 3.27436298e-01 -9.50751733e-03 -6.61801934e-01
3.21924627e-01 -6.10523641e-01 5.33700347e-01 2.68464908e-02
-2.95305680e-02 -1.53568923e-04 -1.95406079e-01 -1.11671478e-01
4.86056834e-01 2.84076303e-01 1.53784409e-01 -4.51977551e-01
6.77695334e-01 -4.87161845e-01 6.54615462e-01 4.94084656e-01
-1.81810439e-01 8.59517992e-01 2.99681604e-01 -2.68234462e-01
-1.00617278e+00 -1.09278882e+00 -4.47693557e-01 9.34261024e-01
4.67406720e-01 -1.77947089e-01 -8.60922515e-01 -1.96881086e-01
-2.15523496e-01 4.43864703e-01 -6.43849492e-01 -2.00035587e-01
-6.99782729e-01 -8.22365701e-01 3.62543315e-02 4.72123057e-01
1.02836192e+00 -7.83624351e-01 -6.16918981e-01 1.70500558e-02
-4.47155803e-01 -1.15450585e+00 -6.03175998e-01 2.46827174e-02
-8.62436891e-01 -8.75062048e-01 -8.83467555e-01 -8.60722184e-01
3.40114564e-01 6.97279334e-01 1.15575182e+00 -1.92263275e-02
-2.23289236e-01 -3.83781381e-02 -3.23823959e-01 -2.10836366e-01
2.35830456e-01 -1.89121932e-01 -2.83844054e-01 3.83106738e-01
3.04277658e-01 -8.45883846e-01 -7.58721650e-01 4.29405451e-01
-9.55890596e-01 1.35112420e-01 7.57086933e-01 9.99542832e-01
8.04517031e-01 5.54425240e-01 5.98464489e-01 -6.16308451e-01
4.74922001e-01 -4.49909657e-01 -5.71623802e-01 1.09250350e-02
-6.51061475e-01 1.36206701e-01 5.83801031e-01 -5.60852826e-01
-1.44978130e+00 -7.32553676e-02 -6.44508302e-02 -4.57676560e-01
2.88133286e-02 3.96259189e-01 -2.40280807e-01 -1.65137649e-01
5.31240642e-01 5.77985168e-01 -7.17315003e-02 -6.28647208e-01
2.92821735e-01 5.87502778e-01 6.71242297e-01 -5.72664618e-01
8.55220318e-01 6.71844602e-01 -1.47124559e-01 -9.61562991e-01
-1.39663303e+00 -4.35653627e-01 -3.24351043e-01 -2.53172606e-01
1.05059016e+00 -9.24615204e-01 -2.98952579e-01 6.55701816e-01
-8.81865203e-01 -1.74234882e-01 -3.76929194e-01 4.64345783e-01
-6.93474054e-01 4.07498598e-01 -5.77728271e-01 -7.54826427e-01
-1.53464243e-01 -1.08151126e+00 1.13989198e+00 6.12463415e-01
3.15721989e-01 -9.04242635e-01 1.33234300e-02 3.64688247e-01
6.87571943e-01 7.19122440e-02 8.67311060e-01 -2.02401318e-02
-7.09902287e-01 3.66489053e-01 -7.81991899e-01 4.69101965e-01
6.59707338e-02 -3.62951279e-01 -1.12820017e+00 -3.34691167e-01
7.96242595e-01 -1.92494184e-01 1.12003648e+00 7.78019786e-01
1.37149775e+00 -1.04205236e-01 -2.77509540e-02 8.24963808e-01
1.66984403e+00 -3.49548012e-02 1.05342126e+00 4.33979541e-01
4.63914514e-01 7.47228444e-01 8.15043628e-01 3.07512939e-01
2.52577782e-01 8.65541160e-01 4.70841944e-01 -2.92304248e-01
-6.01966083e-01 -3.71845990e-01 -4.15538857e-03 5.24070680e-01
-2.33668223e-01 1.04874574e-01 -4.67447251e-01 3.62519652e-01
-1.78971624e+00 -9.43143606e-01 -3.10224444e-01 2.07422209e+00
1.03532135e+00 7.50436559e-02 1.08108990e-01 5.18634953e-02
8.10388625e-01 4.16503251e-01 -4.47660387e-01 2.94519216e-01
-4.42239225e-01 6.52530789e-02 2.49107167e-01 4.98798251e-01
-9.98646855e-01 7.57717550e-01 5.81538486e+00 1.30605352e+00
-1.44204366e+00 6.32486045e-02 7.97012508e-01 1.15826808e-01
-4.10217971e-01 -1.28822193e-01 -8.88860106e-01 6.90318167e-01
4.18303668e-01 -6.68208674e-02 4.09895599e-01 5.38709939e-01
3.31650168e-01 -1.26709267e-01 -6.53907299e-01 1.28549290e+00
1.21090196e-01 -1.00312877e+00 -1.44805729e-01 -9.84468982e-02
6.51303411e-01 -1.62021264e-01 4.03301865e-01 9.75449756e-02
-3.60352606e-01 -1.05828798e+00 9.58527505e-01 6.44311965e-01
7.18536854e-01 -5.16865611e-01 6.90662503e-01 4.76078272e-01
-1.37824738e+00 -2.10514143e-01 -4.77271348e-01 2.90647652e-02
-5.60197979e-02 8.86726379e-01 -1.90326367e-02 6.20625556e-01
9.88958120e-01 8.25498700e-01 -4.01496947e-01 1.03842175e+00
-3.70215207e-01 3.61184239e-01 1.41626284e-01 5.66684544e-01
4.73140776e-02 -5.23270726e-01 5.52997947e-01 9.47401226e-01
2.79814243e-01 1.58544943e-01 1.19645074e-01 1.16348088e+00
2.94222772e-01 -2.05001771e-01 -2.84280092e-01 3.59625906e-01
2.79327303e-01 1.32333744e+00 -6.03216946e-01 1.40125128e-02
-4.36883956e-01 7.65334308e-01 3.17422390e-01 5.04299581e-01
-9.44902539e-01 -1.51565984e-01 3.04475337e-01 5.32350659e-01
4.14343983e-01 -1.02825537e-01 -4.60750222e-01 -1.25033844e+00
2.67526388e-01 -9.33566272e-01 3.29969823e-01 -9.49095309e-01
-1.49265397e+00 7.19178379e-01 -6.07007258e-02 -1.25668859e+00
6.75858781e-02 -4.46349531e-01 -3.91433418e-01 1.09544039e+00
-2.27188945e+00 -1.18711102e+00 -5.55921197e-01 6.64798975e-01
5.12392342e-01 1.30673468e-01 3.01462084e-01 5.76911092e-01
-3.89159352e-01 3.41683924e-01 -4.63967696e-02 -2.14879394e-01
7.95740247e-01 -1.18056154e+00 -2.66643792e-01 8.37010264e-01
-2.71950543e-01 4.78155851e-01 9.31467295e-01 -5.15690565e-01
-1.12382972e+00 -8.53743076e-01 7.12392330e-01 -1.52326703e-01
6.45525157e-01 -7.95786083e-02 -1.25738919e+00 3.94019157e-01
2.30882615e-01 9.38775241e-02 1.89721882e-01 -8.51923749e-02
-5.17326951e-01 -2.89973617e-01 -1.18548644e+00 2.13194028e-01
9.11559820e-01 -5.00930369e-01 -6.03230774e-01 4.09540571e-02
5.69177449e-01 -2.85767108e-01 -8.47261131e-01 6.16704822e-01
3.26273352e-01 -1.38678050e+00 1.21222746e+00 -1.85857311e-01
7.46111870e-01 -5.20630121e-01 -1.99214384e-01 -1.24062788e+00
-6.09124720e-01 -2.43667066e-01 -1.48319036e-01 1.16107380e+00
-1.72251218e-03 -6.07603669e-01 5.36060512e-01 3.26714702e-02
-1.38092384e-01 -8.93368840e-01 -7.84069002e-01 -6.83821797e-01
5.89216650e-02 1.04756542e-01 3.92136931e-01 9.78510797e-01
-4.72647905e-01 5.41230023e-01 -6.08657956e-01 2.78128892e-01
9.73025739e-01 2.72300422e-01 4.64191854e-01 -1.22206271e+00
-4.04882580e-01 -4.75607336e-01 1.03201501e-01 -1.22708845e+00
-9.90101546e-02 -4.22533870e-01 8.99079740e-02 -1.47084439e+00
5.92032611e-01 -3.96079212e-01 -3.52390349e-01 7.85114244e-02
-3.84043574e-01 1.62973002e-01 2.81821610e-03 4.84412581e-01
-2.65146464e-01 6.97284997e-01 1.67776012e+00 7.83789456e-02
4.06347476e-02 -1.65856749e-01 -9.11469817e-01 9.51961398e-01
7.37084925e-01 -1.67410791e-01 -6.07725024e-01 -5.68354249e-01
1.05060279e-01 1.16388261e-01 4.23848331e-01 -8.25642884e-01
1.65303662e-01 -2.28853136e-01 3.91289830e-01 -7.00933993e-01
2.37374321e-01 -7.70917058e-01 -1.12365887e-01 -1.96718518e-02
-3.57469201e-01 -4.35941935e-01 2.66323518e-03 5.45802832e-01
-4.76173788e-01 5.67702986e-02 1.49028695e+00 -2.21412092e-01
-7.24054635e-01 3.54728401e-01 1.68573037e-01 1.99262902e-01
7.37947345e-01 -3.56955737e-01 -4.53956187e-01 -2.36471549e-01
-5.57848811e-01 5.86156286e-02 3.10528725e-01 4.35479552e-01
7.50947177e-01 -1.17098951e+00 -8.47415090e-01 3.52553338e-01
-1.44644096e-01 -2.41008773e-02 8.02752733e-01 1.00952983e+00
-1.38493301e-03 1.29671648e-01 -4.23186183e-01 -6.04264498e-01
-1.03146172e+00 5.72290003e-01 6.04062915e-01 -3.66599411e-01
-8.05565655e-01 8.72179329e-01 5.83326459e-01 -1.32141814e-01
3.70401919e-01 -1.12859182e-01 -6.08835995e-01 9.77181643e-02
9.45684314e-01 3.72090191e-01 -1.95534125e-01 -6.35004282e-01
-1.04436032e-01 1.11781275e+00 -9.08006206e-02 2.74776667e-01
1.49341023e+00 -4.55295950e-01 -3.24101686e-01 4.99797255e-01
9.81142819e-01 2.27044262e-02 -1.48063588e+00 -4.26910192e-01
-1.61463007e-01 -8.56918693e-01 4.31938529e-01 -6.22874498e-01
-1.31764019e+00 9.85830188e-01 8.22624028e-01 4.13015455e-01
1.59964538e+00 -1.15904287e-02 8.87986839e-01 -5.27805865e-01
3.95923555e-01 -1.04207003e+00 4.43172246e-01 4.03429061e-01
9.91696119e-01 -1.15724504e+00 -6.63663521e-02 -7.67149746e-01
-5.52725613e-01 9.63766515e-01 7.00673938e-01 -3.71665329e-01
6.44826770e-01 2.46315986e-01 -1.47269234e-01 -1.11007214e-01
-4.78771776e-01 -2.85438508e-01 2.89850593e-01 7.04897761e-01
5.46608746e-01 -1.99725077e-01 -3.96718234e-01 6.58823371e-01
-3.68286073e-02 1.71142727e-01 3.54580551e-01 5.22021830e-01
-7.55613923e-01 -8.98700416e-01 -4.07542586e-01 1.88248068e-01
-5.94791412e-01 -1.81580722e-01 6.30228445e-02 5.38400948e-01
4.29966778e-01 9.04812157e-01 -9.20319930e-02 -2.72892177e-01
3.57956797e-01 -4.43229288e-01 6.59976244e-01 -3.10644269e-01
-1.13630064e-01 4.87789840e-01 -2.82527924e-01 -6.21552408e-01
-6.69575274e-01 -4.09045279e-01 -1.04125392e+00 -1.98035575e-02
-3.13677132e-01 -3.73758487e-02 3.32229018e-01 6.96399510e-01
1.66920066e-01 7.36642897e-01 6.18925750e-01 -7.97442853e-01
-8.68114531e-01 -9.53942955e-01 -1.06294549e+00 4.28273678e-01
4.76127177e-01 -8.34095955e-01 -6.12538457e-01 -1.36649683e-01] | [10.993551254272461, -2.1462206840515137] |
a0fc541d-80f6-4b1a-a930-824eedd21a63 | dialogue-response-selection-with-hierarchical | 2012.14756 | null | https://arxiv.org/abs/2012.14756v3 | https://arxiv.org/pdf/2012.14756v3.pdf | Dialogue Response Selection with Hierarchical Curriculum Learning | We study the learning of a matching model for dialogue response selection. Motivated by the recent finding that models trained with random negative samples are not ideal in real-world scenarios, we propose a hierarchical curriculum learning framework that trains the matching model in an "easy-to-difficult" scheme. Our learning framework consists of two complementary curricula: (1) corpus-level curriculum (CC); and (2) instance-level curriculum (IC). In CC, the model gradually increases its ability in finding the matching clues between the dialogue context and a response candidate. As for IC, it progressively strengthens the model's ability in identifying the mismatching information between the dialogue context and a response candidate. Empirical studies on three benchmark datasets with three state-of-the-art matching models demonstrate that the proposed learning framework significantly improves the model performance across various evaluation metrics. | ['Yan Wang', 'Nigel Collier', 'Shuming Shi', 'Yunbo Cao', 'Simon Baker', 'Zibo Lin', 'Qingyu Zhou', 'Deng Cai', 'Yixuan Su'] | 2020-12-29 | null | https://aclanthology.org/2021.acl-long.137 | https://aclanthology.org/2021.acl-long.137.pdf | acl-2021-5 | ['conversational-response-selection'] | ['natural-language-processing'] | [ 4.74661946e-01 1.37869492e-01 -2.66368747e-01 -5.71523249e-01
-1.20368576e+00 -4.82415974e-01 7.88395643e-01 5.17384410e-01
-4.58488822e-01 5.10719240e-01 2.68648446e-01 -4.52961564e-01
1.71599671e-01 -7.42404938e-01 -2.56939769e-01 -3.30221564e-01
3.98099214e-01 8.18162024e-01 6.60183549e-01 -8.26855421e-01
6.41622365e-01 -1.77562490e-01 -1.29786181e+00 8.78987730e-01
1.10653567e+00 8.67086232e-01 3.31998646e-01 7.48928428e-01
-3.81510228e-01 9.95959342e-01 -7.16680706e-01 -4.93283629e-01
7.73721933e-03 -6.34864271e-01 -1.37344396e+00 -1.16601445e-01
2.88484007e-01 -2.73453057e-01 -2.87368238e-01 8.87867987e-01
4.87126350e-01 1.99063525e-01 3.69321823e-01 -1.01938248e+00
-3.16165417e-01 5.84111214e-01 -3.06898624e-01 1.18364811e-01
9.66193974e-01 5.50077818e-02 1.15377653e+00 -9.33141887e-01
6.62895799e-01 1.28756678e+00 3.42738658e-01 9.71068799e-01
-1.17718172e+00 -5.25205076e-01 4.03595567e-01 3.16131294e-01
-7.36984491e-01 -2.47515634e-01 7.68380225e-01 -3.50917071e-01
9.14878845e-01 3.94653976e-01 3.75686854e-01 1.18358123e+00
-1.41239360e-01 1.25770497e+00 1.22560310e+00 -7.67507374e-01
1.40086114e-01 4.79540229e-01 4.63431090e-01 3.44597608e-01
-6.04840338e-01 6.88966587e-02 -7.19590783e-01 -5.15637934e-01
2.60337949e-01 -3.77885997e-01 -3.41250688e-01 -2.37093836e-01
-8.17781389e-01 1.10753345e+00 1.04572959e-01 1.74213663e-01
-5.92375249e-02 -4.38029796e-01 6.15139425e-01 8.03467929e-01
2.99327254e-01 7.76042759e-01 -6.08124793e-01 -1.82469428e-01
-5.82172930e-01 6.07629418e-01 1.11241961e+00 9.37932730e-01
5.08398890e-01 -4.25496548e-01 -4.10200775e-01 1.29666257e+00
6.10190071e-02 -1.18745372e-01 5.18808603e-01 -6.62481248e-01
8.53443086e-01 9.68743086e-01 1.47326857e-01 -6.40195668e-01
-2.09505469e-01 -1.81790099e-01 -5.21707833e-01 -1.24585219e-01
6.35843217e-01 -2.07957968e-01 -3.49612504e-01 1.76873326e+00
4.48400676e-01 1.77020773e-01 2.60565639e-01 8.05142045e-01
1.24403584e+00 4.23277408e-01 3.19909781e-01 -1.78159982e-01
1.24333489e+00 -1.14032745e+00 -5.25473833e-01 -3.73853415e-01
7.91107595e-01 -9.02641654e-01 1.25989330e+00 3.31426650e-01
-1.10382068e+00 -8.24768960e-01 -9.00501251e-01 1.86692625e-01
-1.04732893e-01 -1.76869631e-02 4.68531996e-01 3.49186212e-01
-8.42876256e-01 2.77306378e-01 1.05738252e-01 -2.98649907e-01
-3.07186693e-01 2.86977768e-01 -5.80658801e-02 -2.27427468e-01
-1.65370286e+00 1.01278114e+00 1.77885666e-01 -2.64756113e-01
-8.51656497e-01 -6.37392581e-01 -8.00155461e-01 6.83146417e-02
5.28911233e-01 -5.10794222e-01 1.71756375e+00 -1.10633636e+00
-1.64852762e+00 9.71599042e-01 -1.25021026e-01 -1.86411038e-01
8.13127518e-01 -1.21327035e-01 -2.69406676e-01 1.51555449e-01
-1.71682164e-01 8.24390173e-01 6.56264544e-01 -1.40075922e+00
-1.00427520e+00 -5.52019514e-02 5.05656600e-01 3.83347213e-01
-3.32429856e-01 1.41400874e-01 -3.28086376e-01 -3.25001568e-01
8.16729888e-02 -8.52345228e-01 -2.53891110e-01 -5.89337528e-01
-6.11522458e-02 -8.79841745e-01 6.81576312e-01 -3.22364777e-01
1.49630475e+00 -1.70046091e+00 1.04731563e-02 1.94711685e-01
1.51358917e-01 3.75765443e-01 -3.28330934e-01 7.90339708e-01
-1.00256898e-01 -1.22409202e-01 1.27566636e-01 -1.98334470e-01
1.29434327e-02 -8.54797512e-02 -2.68680364e-01 -2.05027044e-01
3.18866163e-01 4.95119780e-01 -1.39770079e+00 -4.33187872e-01
-1.18430354e-01 -1.60718068e-01 -7.23480463e-01 9.14942384e-01
-3.69939297e-01 5.63061893e-01 -5.71337342e-01 4.28161979e-01
4.44201499e-01 -2.74791062e-01 6.23400271e-01 1.03514843e-01
1.86208114e-01 7.58550107e-01 -1.12397659e+00 1.47719622e+00
-4.38408256e-01 3.00263137e-01 4.53079976e-02 -8.60594511e-01
1.18752503e+00 4.35040981e-01 1.22478202e-01 -9.94804919e-01
-1.75421774e-01 1.33254066e-01 3.73650640e-01 -6.73973083e-01
9.63073969e-01 1.28468171e-01 -3.85767639e-01 6.37265682e-01
2.06082195e-01 3.66592146e-02 1.73520103e-01 4.65029269e-01
1.14039123e+00 6.31366111e-03 1.83128729e-01 -2.55685419e-01
9.79042053e-01 4.01048809e-02 6.78102434e-01 8.76387358e-01
-4.35940713e-01 2.63461351e-01 6.74672663e-01 -3.81861717e-01
-6.33809507e-01 -6.61003530e-01 1.52798399e-01 1.70115995e+00
2.50268966e-01 -4.69487637e-01 -8.28488588e-01 -1.03785884e+00
-1.05052449e-01 6.34663820e-01 -6.58923090e-01 -3.10981542e-01
-8.36173773e-01 -4.40281838e-01 2.78435409e-01 4.35135633e-01
4.64324504e-01 -1.12242246e+00 -3.94940197e-01 2.34892935e-01
-6.49657845e-01 -9.47810113e-01 -5.39730310e-01 2.60656089e-01
-5.99860430e-01 -1.26682925e+00 -1.87373206e-01 -1.07720518e+00
7.39332914e-01 4.59954381e-01 1.54148209e+00 6.44874811e-01
1.60813630e-01 4.83758330e-01 -5.73778152e-01 -2.28592962e-01
-8.46293807e-01 1.01518035e-01 -2.54112750e-01 -1.81342036e-01
6.86331213e-01 1.02607630e-01 -6.54031396e-01 6.80751383e-01
-5.84759533e-01 8.51987824e-02 1.99750617e-01 1.25925291e+00
2.13020712e-01 -3.12516302e-01 1.10631371e+00 -1.18220520e+00
1.28740489e+00 -5.96573889e-01 -2.79245973e-01 4.45327401e-01
-7.63046801e-01 -1.40690759e-01 5.81273615e-01 -7.58691728e-01
-1.22929871e+00 -7.63678029e-02 -2.51818627e-01 1.33826301e-01
-1.78059593e-01 5.30784070e-01 3.99732031e-02 -1.24394529e-01
6.87997758e-01 1.71499193e-01 -2.99139708e-01 -3.53753209e-01
-8.70020781e-03 7.77478814e-01 4.68688965e-01 -1.25621450e+00
4.30769086e-01 -2.32446015e-01 -7.14963853e-01 -4.83343214e-01
-7.30810344e-01 -8.66592944e-01 -5.49176157e-01 -4.75095749e-01
5.48406422e-01 -6.78307176e-01 -9.28560615e-01 2.31589094e-01
-1.15113056e+00 -5.91838360e-01 9.42993090e-02 2.14334860e-01
-4.03970271e-01 3.69280905e-01 -8.14693451e-01 -7.81018436e-01
-3.21644425e-01 -1.36581576e+00 5.71273327e-01 5.94389975e-01
-6.47448301e-01 -1.06491506e+00 2.74273217e-01 5.78477621e-01
3.40950131e-01 -2.59821147e-01 1.31852448e+00 -1.25634587e+00
-2.26716056e-01 -3.02977830e-01 5.96778877e-02 1.07026815e-01
-3.15016001e-01 -1.61338553e-01 -1.18671536e+00 -3.39085281e-01
9.89723876e-02 -9.67280805e-01 5.17898977e-01 -2.70061612e-01
9.77955222e-01 -1.10621326e-01 -2.54509181e-01 -1.81885466e-01
1.00304711e+00 5.07018566e-01 6.37035310e-01 5.11811197e-01
1.38362512e-01 1.06253803e+00 1.08612490e+00 2.79490978e-01
6.36009872e-01 7.99944937e-01 2.57658958e-01 -2.08900362e-01
2.30450436e-01 -4.56250638e-01 1.69367373e-01 8.64098787e-01
5.19622624e-01 -1.93989575e-01 -9.21746612e-01 7.06487775e-01
-2.05530572e+00 -8.76475573e-01 -1.40480518e-01 2.37131810e+00
1.38434553e+00 3.31555635e-01 2.48327062e-01 -9.88363773e-02
6.24418974e-01 -7.18204007e-02 -2.77323991e-01 -8.09536159e-01
2.04620615e-01 2.10359752e-01 -3.59548956e-01 6.80813491e-01
-1.05521190e+00 9.45921302e-01 6.21557045e+00 6.81737483e-01
-8.55619967e-01 -3.79123501e-02 7.68927336e-01 4.39798743e-01
-4.80183601e-01 1.27453089e-01 -1.02715755e+00 2.64459252e-01
8.05860519e-01 -5.85592873e-02 -8.89710933e-02 7.66789973e-01
-1.88791137e-02 -1.82550803e-01 -1.39238238e+00 4.09316182e-01
-2.89634545e-03 -1.16246140e+00 7.20058903e-02 -3.51847082e-01
6.19607389e-01 -4.67342883e-01 -1.75884247e-01 9.21020865e-01
6.28397763e-01 -8.44884455e-01 3.28311712e-01 2.51197785e-01
3.33194435e-01 -7.87500560e-01 9.98504877e-01 6.63659871e-01
-9.82738495e-01 -1.23216242e-01 -4.13458645e-01 4.88002338e-02
-3.11147898e-01 -1.97602138e-01 -1.25368381e+00 4.77311730e-01
4.21913236e-01 -1.56823937e-02 -5.68400919e-01 9.33032274e-01
-4.32645917e-01 6.04586244e-01 3.11644167e-01 -2.34860465e-01
4.57936645e-01 -1.59609199e-01 2.87840277e-01 1.54326105e+00
-4.42093462e-01 3.05866897e-01 7.23820865e-01 5.56745589e-01
-1.01566814e-01 3.23501498e-01 -2.38278925e-01 3.58858109e-01
8.93175483e-01 1.31791985e+00 -3.73392969e-01 -4.09754932e-01
-4.67822373e-01 6.02867365e-01 6.10196054e-01 1.95430279e-01
-3.72743964e-01 -2.19725385e-01 3.86419117e-01 -4.92992885e-02
8.10177624e-03 3.72201413e-01 -2.06998155e-01 -7.57926404e-01
-1.32490054e-01 -1.54575551e+00 7.30317891e-01 -3.75891775e-01
-1.31794024e+00 5.91665745e-01 1.10980514e-02 -1.29063094e+00
-5.11550963e-01 -3.38965237e-01 -1.00590527e+00 1.04664886e+00
-1.54702532e+00 -9.67904806e-01 -3.21127415e-01 5.98987460e-01
9.66670632e-01 -2.56565064e-01 1.29206800e+00 1.52864203e-01
-3.89684498e-01 1.00877380e+00 -3.09265137e-01 3.26776892e-01
1.06226778e+00 -1.46257520e+00 2.45399863e-01 4.51232314e-01
-2.78314948e-01 8.68161440e-01 6.51809990e-01 -4.80480641e-01
-1.15761864e+00 -5.33181548e-01 1.24689150e+00 -2.88018018e-01
5.95520496e-01 -3.78382087e-01 -1.26423526e+00 1.94776192e-01
4.19021487e-01 -6.06169462e-01 1.07441854e+00 3.99069190e-01
-5.03724873e-01 9.88520682e-02 -1.17092550e+00 6.00408971e-01
3.74658704e-01 -6.48238957e-01 -8.35576057e-01 5.65781057e-01
6.74834430e-01 -7.88351774e-01 -9.05947268e-01 3.30555916e-01
5.71791470e-01 -7.69046068e-01 8.69153380e-01 -1.09223628e+00
6.28777742e-01 2.15577111e-01 -3.42470519e-02 -1.30492544e+00
-1.03953354e-01 -7.98964262e-01 -2.90062372e-02 1.19642627e+00
4.33142066e-01 -2.00010777e-01 7.27234304e-01 8.21261942e-01
-1.83289528e-01 -1.07770717e+00 -6.99250460e-01 -3.27901393e-01
3.10927123e-01 -1.69626161e-01 4.88008052e-01 1.37085772e+00
5.41178942e-01 7.23823309e-01 -2.50280082e-01 -2.00792566e-01
1.62906870e-01 2.98409969e-01 8.61444652e-01 -1.33610809e+00
-4.52996731e-01 -5.49494386e-01 2.25980222e-01 -1.57238162e+00
3.16059381e-01 -5.56321144e-01 1.86619133e-01 -1.13244283e+00
4.21578228e-01 -6.98729753e-01 -2.47170493e-01 2.02405155e-01
-7.79792070e-01 -4.15518403e-01 1.40201673e-01 5.65399379e-02
-8.36480200e-01 2.50211984e-01 1.14490104e+00 -1.86759576e-01
-3.08663756e-01 3.91534775e-01 -7.13932693e-01 5.96132517e-01
6.45858586e-01 -4.78190869e-01 -5.93671978e-01 -1.16525508e-01
1.16113879e-01 5.71590185e-01 -1.28265470e-01 -4.42008197e-01
5.37205637e-01 -2.73266584e-01 1.10959940e-01 -4.74623561e-01
6.42423704e-02 -4.17719752e-01 -7.09830046e-01 3.61538619e-01
-1.07694197e+00 2.58773744e-01 1.03310823e-01 5.10400891e-01
-3.15066516e-01 -7.17736006e-01 7.69841254e-01 -2.35382929e-01
-7.82710135e-01 -1.59723654e-01 -3.24326128e-01 3.80203158e-01
5.60901105e-01 -1.42024219e-01 -7.12420523e-01 -6.89983547e-01
-5.18549740e-01 7.22777605e-01 1.87188134e-01 6.89445317e-01
8.61439884e-01 -1.10965490e+00 -8.10774863e-01 6.21654652e-02
5.12434542e-01 -1.98301852e-01 1.68596670e-01 6.92481637e-01
3.19841653e-02 4.26524818e-01 7.59982169e-02 -6.26475215e-01
-1.65519512e+00 2.67267734e-01 4.32739913e-01 -8.76387358e-01
-1.38032183e-01 1.10367227e+00 2.32919514e-01 -1.05758762e+00
6.23004973e-01 8.32082406e-02 -6.76921844e-01 1.04833148e-01
7.42132127e-01 1.34613141e-01 1.42335713e-01 -3.61275315e-01
-2.45911051e-02 -1.56783145e-02 -6.21244431e-01 2.23668516e-02
8.57244432e-01 -1.17983095e-01 1.54791117e-01 3.00670445e-01
8.59142482e-01 -2.51556605e-01 -9.69184935e-01 -7.49357402e-01
4.75773156e-01 -3.53460282e-01 -2.50840664e-01 -1.13920820e+00
-4.65829492e-01 8.96063626e-01 4.59640026e-01 3.82517606e-01
9.57427084e-01 -3.50585848e-01 7.36453414e-01 5.39825797e-01
3.20284903e-01 -1.55859900e+00 6.25607312e-01 9.83351529e-01
8.06509733e-01 -1.53147626e+00 -1.98618561e-01 -4.01402116e-01
-9.49603021e-01 1.23836529e+00 1.29477465e+00 1.48971424e-01
1.63663357e-01 7.92237967e-02 3.89718801e-01 -1.40115976e-01
-1.35146451e+00 -7.76420161e-02 3.49305958e-01 5.10562003e-01
8.42635155e-01 -6.85619786e-02 -5.33592284e-01 5.97967684e-01
5.36582917e-02 -4.73741025e-01 4.56261069e-01 1.00857687e+00
-5.27455449e-01 -1.52709687e+00 -9.22888592e-02 1.77764282e-01
-2.68591195e-01 -1.10587724e-01 -1.01709974e+00 8.51458967e-01
-3.00323129e-01 1.48987508e+00 -9.73989964e-02 -7.97079086e-01
6.24476194e-01 3.21866155e-01 2.77063400e-01 -7.92643309e-01
-1.50794041e+00 1.87373593e-01 5.00735819e-01 -4.64500517e-01
-2.82939881e-01 -4.19542581e-01 -1.05887341e+00 -2.27951482e-01
-4.20542538e-01 6.14283919e-01 2.89218932e-01 1.09750891e+00
-1.21202968e-01 4.68404591e-01 9.39835787e-01 -2.31763408e-01
-1.18546534e+00 -1.08574593e+00 1.22644752e-01 6.27534270e-01
1.41742071e-02 -5.46774864e-01 -4.95639928e-02 -3.29283595e-01] | [12.511350631713867, 7.900093078613281] |
8969b938-736a-429e-ba63-09748097fd45 | understanding-uncertainty-sampling | 2307.02719 | null | https://arxiv.org/abs/2307.02719v1 | https://arxiv.org/pdf/2307.02719v1.pdf | Understanding Uncertainty Sampling | Uncertainty sampling is a prevalent active learning algorithm that queries sequentially the annotations of data samples which the current prediction model is uncertain about. However, the usage of uncertainty sampling has been largely heuristic: (i) There is no consensus on the proper definition of "uncertainty" for a specific task under a specific loss; (ii) There is no theoretical guarantee that prescribes a standard protocol to implement the algorithm, for example, how to handle the sequentially arrived annotated data under the framework of optimization algorithms such as stochastic gradient descent. In this work, we systematically examine uncertainty sampling algorithms under both stream-based and pool-based active learning. We propose a notion of equivalent loss which depends on the used uncertainty measure and the original loss function and establish that an uncertainty sampling algorithm essentially optimizes against such an equivalent loss. The perspective verifies the properness of existing uncertainty measures from two aspects: surrogate property and loss convexity. Furthermore, we propose a new notion for designing uncertainty measures called \textit{loss as uncertainty}. The idea is to use the conditional expected loss given the features as the uncertainty measure. Such an uncertainty measure has nice analytical properties and generality to cover both classification and regression problems, which enable us to provide the first generalization bound for uncertainty sampling algorithms under both stream-based and pool-based settings, in the full generality of the underlying model and problem. Lastly, we establish connections between certain variants of the uncertainty sampling algorithms with risk-sensitive objectives and distributional robustness, which can partly explain the advantage of uncertainty sampling algorithms when the sample size is small. | ['Xiaocheng Li', 'Shang Liu'] | 2023-07-06 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 2.17435688e-01 3.69689047e-01 -3.05166066e-01 -6.06647670e-01
-1.06605649e+00 -4.43739325e-01 4.73201424e-01 6.34953856e-01
-5.88039160e-01 1.13926792e+00 -1.84354812e-01 -1.02263153e-01
-6.22293532e-01 -1.13630867e+00 -8.40826988e-01 -6.80977523e-01
-1.47988662e-01 4.56119150e-01 7.61832371e-02 8.02773535e-02
2.25040823e-01 4.19535577e-01 -1.71170557e+00 -1.73564985e-01
1.07003844e+00 1.45520222e+00 -2.13308111e-01 3.92362535e-01
-2.81584293e-01 5.21505237e-01 -6.18374467e-01 -5.55147588e-01
3.95695329e-01 -2.85662919e-01 -7.44514585e-01 -8.40518475e-02
-1.35569349e-01 -2.07565159e-01 2.24981651e-01 1.05378842e+00
4.57788259e-01 4.02994335e-01 7.81329572e-01 -1.46609807e+00
1.61368418e-02 1.04256451e+00 -1.21016562e-01 3.38410474e-02
1.52117684e-01 -1.35753706e-01 1.11362481e+00 -5.94980955e-01
3.72450501e-01 8.54000688e-01 5.10973930e-01 4.86905813e-01
-1.08020389e+00 -3.12740088e-01 2.91009218e-01 8.81017074e-02
-1.35056186e+00 -3.27785164e-01 5.06035686e-01 -4.17276055e-01
2.01613680e-01 4.89142299e-01 3.93732399e-01 1.06411028e+00
1.05089806e-02 1.10400987e+00 9.16222572e-01 -4.65809882e-01
1.06349432e+00 5.58111250e-01 2.54042834e-01 2.46546701e-01
5.78457534e-01 2.41156846e-01 -5.60486913e-01 -4.70006198e-01
1.79344848e-01 -3.16953175e-02 -5.27870357e-01 -7.14617133e-01
-6.96036160e-01 1.01780033e+00 -1.07860968e-01 -3.82910334e-02
-1.84317276e-01 2.33951909e-03 5.61019957e-01 5.02059519e-01
9.25546587e-01 4.58752960e-01 -4.55752045e-01 -1.47590041e-01
-9.92621005e-01 3.20213586e-01 1.18982506e+00 1.03723931e+00
7.11591840e-01 -6.15657531e-02 -2.60442466e-01 4.91490781e-01
6.07728839e-01 1.87142268e-01 -2.55778208e-02 -7.03066885e-01
4.99952704e-01 3.90734375e-01 4.14491028e-01 -5.75700104e-01
-4.96089691e-03 -4.41862732e-01 -4.82445002e-01 4.68274742e-01
5.41203201e-01 -3.06894362e-01 -1.79829836e-01 1.88250458e+00
4.26637948e-01 6.62834197e-02 -8.48518237e-02 5.70133865e-01
1.29536480e-01 3.18808079e-01 -1.29536971e-01 -7.93540537e-01
6.00581646e-01 -2.21586570e-01 -6.33471191e-01 1.86623201e-01
4.17503268e-01 -2.00374305e-01 1.04245853e+00 6.57875955e-01
-1.00748539e+00 -3.81249972e-02 -1.34964502e+00 5.52969635e-01
-4.03836578e-01 -4.88456637e-01 4.19822603e-01 1.12795603e+00
-6.05124950e-01 9.56065059e-01 -7.09246397e-01 -1.06779635e-01
4.78698045e-01 8.02714750e-02 8.35281760e-02 1.62529141e-01
-1.30921400e+00 6.46941066e-01 5.84565401e-01 9.35628638e-02
-9.92456377e-01 -9.81046617e-01 -6.93986654e-01 8.35722983e-02
6.03177667e-01 -6.26481593e-01 1.17096841e+00 -1.02322984e+00
-1.53004158e+00 5.07020831e-01 2.31072396e-01 -7.78513134e-01
1.26125085e+00 -2.92563796e-01 -1.16568297e-01 -1.16002277e-01
-1.85453221e-01 -1.15843602e-01 9.78812397e-01 -1.16198635e+00
-7.13798106e-01 -4.44000006e-01 1.81885615e-01 1.77419320e-01
-1.51418149e-01 -1.89121902e-01 5.20559028e-02 -5.94586194e-01
-6.35443851e-02 -3.71926337e-01 -3.20322901e-01 3.41535777e-01
-4.17251974e-01 -2.03059658e-01 4.96775895e-01 8.50952938e-02
1.48185146e+00 -2.08172059e+00 2.49981321e-02 4.59695280e-01
-7.05873445e-02 -1.60423324e-01 3.67952168e-01 5.47575057e-01
1.02995478e-01 2.83748150e-01 -8.41692984e-01 -4.41815257e-01
3.51151526e-01 1.21210404e-01 -7.63098598e-01 7.71236122e-01
1.64299577e-01 4.67513889e-01 -1.06132686e+00 -3.75040591e-01
1.31767258e-01 1.82134137e-01 -2.12381855e-01 3.12663972e-01
-6.75014913e-01 2.74950624e-01 -5.16539037e-01 4.68283832e-01
6.18484139e-01 1.05343364e-01 -6.17614053e-02 1.77803114e-01
-1.59532323e-01 -4.29311283e-02 -1.62663448e+00 1.55482340e+00
-4.43349302e-01 1.69029742e-01 1.79280311e-01 -1.11836278e+00
9.67642069e-01 3.52843225e-01 6.82974577e-01 1.71732768e-01
9.09356698e-02 4.77903455e-01 -6.00035727e-01 -2.33531713e-01
2.61124998e-01 -2.05944940e-01 -9.18805078e-02 7.51674294e-01
3.42208445e-02 -6.64502159e-02 1.61638215e-01 8.26591998e-03
9.64395761e-01 2.85388798e-01 4.69484925e-01 -5.11265695e-01
4.09710824e-01 -3.55058134e-01 5.06983638e-01 1.03705108e+00
-2.21913069e-01 5.85800111e-01 5.18740535e-01 -1.68467358e-01
-7.16544569e-01 -1.33209801e+00 -7.32387364e-01 9.51655269e-01
-2.68039294e-03 -2.75653243e-01 -5.16419828e-01 -9.74439979e-01
2.41547734e-01 9.60786939e-01 -6.87766433e-01 -9.77345333e-02
7.25006163e-02 -9.56901729e-01 4.20104235e-01 9.54087451e-02
2.38562748e-01 -6.05102122e-01 -7.85401285e-01 1.79400355e-01
1.41115785e-01 -3.72533381e-01 -1.44868389e-01 4.82957631e-01
-1.03444481e+00 -1.23757684e+00 -4.67116624e-01 2.06929326e-01
3.80603313e-01 -4.90442753e-01 1.23838401e+00 -3.48972231e-01
6.43890053e-02 8.30140114e-01 -5.04158437e-01 -9.18050647e-01
-2.20606044e-01 -1.68729201e-01 8.64011273e-02 3.64318460e-01
1.61334842e-01 -5.84746718e-01 -3.29162925e-01 1.81407213e-01
-1.20159149e+00 -6.80739164e-01 1.54297322e-01 6.52261198e-01
6.90050423e-01 2.88417071e-01 7.28441715e-01 -1.03821683e+00
7.50818133e-01 -8.57057810e-01 -7.59542704e-01 5.29834807e-01
-1.00107706e+00 4.15015191e-01 3.87134612e-01 -7.18324631e-02
-1.08951390e+00 6.29349500e-02 1.90495960e-02 -2.61476308e-01
1.44793853e-01 3.90982449e-01 -5.31180799e-01 1.37025386e-01
6.90185249e-01 5.39177880e-02 -4.35086749e-02 -3.00771713e-01
4.17613596e-01 6.30635262e-01 1.08887941e-01 -1.13986015e+00
4.47994322e-01 5.98674476e-01 6.95002824e-02 -7.58493364e-01
-8.44606638e-01 -3.31475586e-01 -4.15618509e-01 -2.47337878e-01
2.98231453e-01 -4.90236044e-01 -8.29044819e-01 1.69149548e-01
-9.11135137e-01 -6.51200488e-02 -1.15046716e+00 5.51864147e-01
-1.12239373e+00 4.05227602e-01 3.73887713e-03 -1.65625000e+00
-4.31458116e-01 -1.02380168e+00 8.66469741e-01 -7.21370131e-02
-6.91098943e-02 -1.08545804e+00 2.48352095e-01 -1.35287896e-01
3.44216734e-01 4.30696398e-01 6.81620538e-01 -9.12463546e-01
-5.44534564e-01 -1.63164675e-01 1.84026554e-01 6.12921238e-01
-1.45991698e-01 1.82216942e-01 -1.12040901e+00 -1.52515724e-01
3.83027822e-01 -3.59441727e-01 9.05158222e-01 5.13091803e-01
1.32084572e+00 -3.19536865e-01 -6.44415095e-02 5.51281631e-01
1.66878939e+00 -3.68859358e-02 4.65059221e-01 1.09176785e-01
-7.09359944e-02 9.18622732e-01 7.03466296e-01 9.74309087e-01
-7.89197385e-02 3.59763056e-01 6.30469918e-01 6.80373371e-01
7.35337734e-01 -2.27713659e-01 3.85628432e-01 3.38332385e-01
2.56122440e-01 -2.67791212e-01 -5.78240991e-01 3.02061766e-01
-2.06431937e+00 -9.98496234e-01 3.77824306e-01 3.07378769e+00
1.09629822e+00 2.82031387e-01 2.21904263e-01 4.31000829e-01
7.51598835e-01 -4.98458603e-03 -7.43698657e-01 -3.25510800e-01
-5.77153563e-02 2.27209136e-01 5.33268034e-01 5.92220187e-01
-1.18432319e+00 -5.42568928e-03 6.08712435e+00 1.07427895e+00
-7.86959231e-01 1.22760281e-01 7.25349367e-01 -2.53197432e-01
-6.31137669e-01 -5.50403297e-02 -8.10515940e-01 7.39335656e-01
8.73930931e-01 -5.44739544e-01 3.53283845e-02 1.06043088e+00
1.22204229e-01 -1.90379679e-01 -1.58428860e+00 7.82709241e-01
-2.60524541e-01 -1.07610536e+00 -2.69299448e-01 -4.47730385e-02
4.76243347e-01 -2.20624968e-01 -1.07618004e-01 1.78209081e-01
2.94680119e-01 -9.01817918e-01 9.75158691e-01 9.84519958e-01
6.76286280e-01 -9.52005029e-01 5.82835019e-01 5.89865565e-01
-8.10628712e-01 -9.29916799e-02 -3.20245445e-01 1.51622653e-01
2.21091166e-01 1.17368877e+00 -6.76779449e-01 7.56352365e-01
3.90068531e-01 3.67246568e-01 -3.12416423e-02 1.30378044e+00
-5.51177049e-03 7.22673833e-01 -8.58519256e-01 -1.81607351e-01
-1.33325338e-01 -3.83571863e-01 9.19425964e-01 1.00391448e+00
3.50488335e-01 -3.49549770e-01 -3.75669748e-02 1.11588800e+00
4.73566167e-02 1.87126473e-01 -6.61552668e-01 1.27720475e-01
6.59856915e-01 8.71593356e-01 -5.43840468e-01 7.15904012e-02
-1.58947229e-01 5.82907081e-01 1.11665212e-01 2.21013516e-01
-6.09547615e-01 -4.28009480e-01 6.16865516e-01 4.78424095e-02
9.26152095e-02 1.35251939e-01 -4.95395988e-01 -1.13257384e+00
3.07804734e-01 -4.81770039e-01 6.77317321e-01 -1.66883603e-01
-1.67405856e+00 3.15938532e-01 3.64225388e-01 -1.23623395e+00
-2.53490835e-01 -5.43321073e-01 -6.88788712e-01 7.95451581e-01
-1.17163336e+00 -4.57214803e-01 7.60231838e-02 5.32631338e-01
2.94156283e-01 -9.59294364e-02 8.44028175e-01 3.39340568e-02
-5.44580519e-01 6.37368202e-01 4.99186844e-01 -2.20281556e-01
4.94130641e-01 -1.62908638e+00 -2.17506111e-01 8.61922026e-01
8.16744417e-02 5.50628006e-01 9.22982454e-01 -5.25446177e-01
-1.19906318e+00 -1.08946955e+00 4.56181645e-01 -6.78427100e-01
6.66618645e-01 -3.05557787e-01 -6.69562638e-01 4.07054126e-01
-2.58685589e-01 1.70162335e-01 7.31928706e-01 1.75018504e-01
-3.38395506e-01 -4.08056408e-01 -1.67739654e+00 1.78334087e-01
8.14654887e-01 -3.87928545e-01 -4.14073855e-01 2.92902052e-01
8.52140546e-01 -3.96355726e-02 -1.07576776e+00 5.91127634e-01
5.56351960e-01 -1.18092322e+00 6.60341382e-01 -8.21787715e-01
1.39380738e-01 -1.73274383e-01 -5.69932818e-01 -1.19630408e+00
3.34519565e-01 -7.93780208e-01 -6.13271952e-01 1.29197633e+00
5.93991399e-01 -8.28178823e-01 7.25358307e-01 8.37445378e-01
8.60114917e-02 -1.19394815e+00 -1.15353525e+00 -8.39167058e-01
1.63112074e-01 -1.06909335e+00 6.99983180e-01 6.39872313e-01
9.04201716e-02 -1.48696363e-01 -2.85983812e-02 -6.90926760e-02
1.08002067e+00 -1.25933364e-01 2.31064633e-01 -1.46548462e+00
-5.27625144e-01 -3.95046324e-01 -3.13865483e-01 -5.86530566e-01
1.56583443e-01 -4.83052522e-01 3.05728463e-04 -9.46634650e-01
-1.86055735e-01 -9.15644944e-01 -4.33144093e-01 -8.76642838e-02
1.82104073e-02 -5.39473116e-01 1.70909967e-02 9.31611210e-02
-6.10702753e-01 7.25740850e-01 6.61178291e-01 -9.48745012e-02
-2.64779419e-01 8.99936020e-01 -6.03876650e-01 5.11256576e-01
6.78238690e-01 -3.25914204e-01 -8.81620944e-01 1.18203312e-01
8.17273855e-01 4.43696119e-02 1.07121691e-01 -7.51856863e-01
2.24376187e-01 -2.01566726e-01 1.40960496e-02 -3.73093784e-01
5.28926775e-02 -1.00065231e+00 2.05292970e-01 1.69262096e-01
-7.90759981e-01 -4.80041116e-01 -2.21008688e-01 1.13946044e+00
-1.68195531e-01 -8.71321380e-01 7.56247222e-01 -1.88529462e-01
-3.82166624e-01 6.00291193e-01 -2.71716639e-02 3.67942750e-01
1.18256247e+00 -1.02114908e-01 8.97050053e-02 -4.06190008e-01
-7.28639662e-01 1.96785823e-01 3.15076917e-01 1.80116091e-02
4.52661186e-01 -1.14029229e+00 -8.74169886e-01 1.58219095e-02
2.28291318e-01 2.84791321e-01 -7.92716220e-02 7.24762559e-01
-1.85932353e-01 1.77020073e-01 3.71386945e-01 -5.94152808e-01
-5.63807130e-01 5.13433516e-01 4.55796748e-01 -1.52293518e-01
-2.52217978e-01 8.18047822e-01 -2.21106410e-01 -4.43427980e-01
6.96360111e-01 -2.90180951e-01 1.20594919e-01 3.32564384e-01
6.12567186e-01 7.90938556e-01 3.66706640e-01 5.10406233e-02
-2.15434983e-01 -4.26982269e-02 3.38096082e-01 -3.58726233e-01
1.02956688e+00 -2.56688297e-01 4.73235697e-02 9.40333843e-01
1.01129258e+00 -9.72124636e-02 -1.35971320e+00 -3.71860832e-01
2.98733771e-01 -3.70603859e-01 -1.25454217e-01 -6.52429163e-01
-7.09398985e-01 7.86783874e-01 6.24705970e-01 5.29758632e-01
1.10424650e+00 -2.04182506e-01 9.12601426e-02 3.96754146e-01
8.38277757e-01 -1.22217715e+00 -5.67436814e-01 2.54453331e-01
9.14805472e-01 -1.21638882e+00 1.44446522e-01 -1.58660457e-01
-4.79880810e-01 1.14877558e+00 1.16149388e-01 2.04788242e-02
1.05966711e+00 3.83839279e-01 -3.95144671e-01 2.35606268e-01
-8.43266189e-01 -1.89538583e-01 1.42073527e-01 6.29391551e-01
2.63581008e-01 2.05548391e-01 -5.69509566e-01 9.15325344e-01
5.70483021e-02 7.98733681e-02 2.80144274e-01 8.86725426e-01
-6.43855095e-01 -1.06922936e+00 -3.55857551e-01 5.61689019e-01
-4.46093440e-01 2.25605249e-01 -2.15846717e-01 5.07850528e-01
4.87910658e-02 9.45220292e-01 -8.67245570e-02 -5.99950030e-02
2.10328028e-01 3.51038843e-01 4.89730448e-01 -6.78196549e-01
-3.03587884e-01 -4.67048049e-01 9.73136723e-02 -4.76060480e-01
-3.69965881e-01 -6.98694885e-01 -8.64175677e-01 -6.94061369e-02
-5.03801823e-01 5.55280745e-01 7.38944292e-01 1.02394617e+00
1.06448978e-01 -2.88978238e-02 8.28642845e-01 -3.98255438e-01
-1.49219811e+00 -7.82104313e-01 -1.09634459e+00 1.16095528e-01
1.85450017e-01 -7.02918231e-01 -8.05427849e-01 -3.42388481e-01] | [7.146929740905762, 4.062348365783691] |
418d0b32-9cee-46da-9940-ef99831d7cef | online-adaptation-through-meta-learning-for | 1904.08462 | null | http://arxiv.org/abs/1904.08462v1 | http://arxiv.org/pdf/1904.08462v1.pdf | Online Adaptation through Meta-Learning for Stereo Depth Estimation | In this work, we tackle the problem of online adaptation for stereo depth
estimation, that consists in continuously adapting a deep network to a target
video recordedin an environment different from that of the source training set.
To address this problem, we propose a novel Online Meta-Learning model with
Adaption (OMLA). Our proposal is based on two main contributions. First, to
reducethe domain-shift between source and target feature distributions we
introduce an online feature alignment procedurederived from Batch
Normalization. Second, we devise a meta-learning approach that exploits feature
alignment forfaster convergence in an online learning setting. Additionally, we
propose a meta-pre-training algorithm in order toobtain initial network weights
on the source dataset whichfacilitate adaptation on future data streams.
Experimentally, we show that both OMLA and meta-pre-training helpthe model to
adapt faster to a new environment. Our proposal is evaluated on the
wellestablished KITTI dataset,where we show that our online method is
competitive withstate of the art algorithms trained in a batch setting. | ['Zhen-Yu Zhang', 'Stéphane Lathuilière', 'Jian Yang', 'Andrea Pilzer', 'Elisa Ricci', 'Nicu Sebe'] | 2019-04-17 | null | null | null | null | ['stereo-depth-estimation'] | ['computer-vision'] | [ 2.48565570e-01 1.76839251e-03 -6.95056021e-02 -4.89246935e-01
-6.11009598e-01 -3.28719199e-01 7.64018178e-01 -3.30765657e-02
-8.72643948e-01 5.67474484e-01 8.63923877e-02 2.64074355e-01
4.54365686e-02 -6.24895751e-01 -1.03286183e+00 -7.04455495e-01
1.72093213e-01 4.33854222e-01 4.94945735e-01 -1.15091801e-01
2.61279732e-01 6.90011322e-01 -1.89387202e+00 3.94214630e-01
7.01287448e-01 9.81957018e-01 3.87211263e-01 8.56705427e-01
-1.16047397e-01 7.74922073e-01 -3.52401674e-01 -1.38431057e-01
3.70609641e-01 -6.60670877e-01 -8.47302854e-01 4.47352678e-01
8.39049399e-01 -3.62689704e-01 -2.88596839e-01 8.69657218e-01
5.81668437e-01 4.42672849e-01 5.52093267e-01 -8.83376300e-01
8.47841203e-02 3.62941563e-01 -2.32055560e-01 4.65596944e-01
-6.60163313e-02 -4.64471616e-02 5.02230167e-01 -1.03333926e+00
8.26959670e-01 8.03446651e-01 6.50963366e-01 9.36708391e-01
-1.17499435e+00 -4.09751952e-01 4.59985942e-01 6.07895195e-01
-1.05155897e+00 -8.97764683e-01 7.63607264e-01 -3.39789271e-01
9.89232659e-01 -1.57651260e-01 7.55213559e-01 1.06819665e+00
1.47680029e-01 8.03999007e-01 9.84326303e-01 -6.12177312e-01
5.94047010e-01 1.58936054e-01 -3.28391612e-01 4.53742415e-01
-8.03609788e-02 2.74056256e-01 -8.62388730e-01 2.37911284e-01
8.66065860e-01 -2.03189000e-01 -2.40845412e-01 -8.26636255e-01
-9.28758323e-01 5.12613058e-01 2.79874951e-01 4.57770169e-01
-5.22220910e-01 1.63795888e-01 4.51421738e-01 5.37313044e-01
5.95777392e-01 4.28753085e-02 -4.98899817e-01 -1.55187413e-01
-9.61254656e-01 -1.86711885e-02 7.34570444e-01 8.74517322e-01
1.11743546e+00 9.48119983e-02 1.48208678e-01 8.79867554e-01
1.34402484e-01 4.41530854e-01 1.00828314e+00 -1.03240716e+00
5.13118207e-01 1.72264397e-01 -3.98114286e-02 -5.99223316e-01
-4.95569825e-01 -5.21615565e-01 -6.37149811e-01 4.63400722e-01
4.55412924e-01 -3.22909176e-01 -8.45699430e-01 1.82922792e+00
5.57371557e-01 5.34138978e-01 9.39108208e-02 6.45187438e-01
3.26069385e-01 3.92293096e-01 -2.61273116e-01 -2.36753970e-01
5.51581979e-01 -1.09046555e+00 -4.42797720e-01 -3.74261856e-01
6.38760686e-01 -4.36516076e-01 7.31390059e-01 5.89857757e-01
-1.15480602e+00 -6.59055829e-01 -1.03580987e+00 2.07929894e-01
-9.79883224e-02 -1.63477048e-01 3.88923079e-01 3.00960302e-01
-1.35397542e+00 9.74155426e-01 -1.01771593e+00 -6.43810213e-01
2.52286553e-01 4.88480806e-01 -5.92953980e-01 1.61461920e-01
-6.70842648e-01 7.14476824e-01 7.54390836e-01 -2.38943920e-02
-8.00612509e-01 -5.89043677e-01 -7.89063334e-01 -1.95022985e-01
3.66141379e-01 -9.36178923e-01 1.52384126e+00 -1.94171572e+00
-2.18390679e+00 8.54781866e-01 -2.16810197e-01 -6.55371845e-01
5.46529055e-01 -1.91517860e-01 -1.47858247e-01 2.47760043e-01
-2.47911900e-01 5.56886613e-01 1.23878336e+00 -1.07896662e+00
-9.76477444e-01 -4.98531848e-01 -6.28267601e-02 5.10674238e-01
-6.12123132e-01 -4.26938236e-01 -4.87525880e-01 -6.18980825e-01
1.82252377e-01 -7.83363998e-01 -3.64702791e-01 -5.17884642e-02
1.84916511e-01 2.04846069e-01 4.31930989e-01 -2.05856010e-01
1.01630020e+00 -2.23343778e+00 6.52004361e-01 2.89658993e-01
8.04774687e-02 3.35943073e-01 -1.43537745e-01 3.08415651e-01
-2.55436867e-01 -6.07257962e-01 -1.63617432e-01 -8.08290184e-01
-3.23476940e-01 3.49648356e-01 -2.04929098e-01 4.65766758e-01
-8.97673145e-02 5.19747615e-01 -9.35937345e-01 -4.11318421e-01
2.27390826e-01 1.69917524e-01 -7.61812150e-01 4.55498517e-01
-1.42795667e-01 5.85877717e-01 -1.67284235e-02 1.35911927e-01
5.91041684e-01 3.21433730e-02 7.67706782e-02 8.07529390e-02
-3.61962169e-01 -1.28160203e-02 -1.34399056e+00 2.14659643e+00
-7.91623235e-01 6.07955039e-01 7.54471719e-02 -1.01025712e+00
6.93318307e-01 3.16145509e-01 5.88771939e-01 -5.56235969e-01
3.83181781e-01 3.80282044e-01 -2.03135625e-01 -5.31588435e-01
4.20773685e-01 -1.73425987e-01 4.05996650e-01 3.90229791e-01
6.07056797e-01 -2.87766550e-02 1.04797363e-01 -2.12203816e-01
8.83844852e-01 4.10473645e-01 3.60750765e-01 -1.22747324e-01
8.02566826e-01 -3.63478601e-01 5.01809657e-01 8.62641931e-01
-1.94675729e-01 6.86327875e-01 -9.30001494e-03 -6.65575266e-01
-1.05397367e+00 -7.29523063e-01 1.20755807e-01 1.16342676e+00
-1.02805346e-01 -1.52017593e-01 -9.61509049e-01 -8.39373112e-01
-1.82647169e-01 2.93432087e-01 -7.16090620e-01 -2.01144710e-01
-6.82175398e-01 -4.86109108e-01 1.09583981e-01 7.26507068e-01
6.12823367e-01 -8.85912418e-01 -9.74632382e-01 4.66906786e-01
2.49547753e-02 -1.09809780e+00 -5.30966699e-01 2.54865527e-01
-1.24306798e+00 -8.33880186e-01 -8.27759922e-01 -8.29818845e-01
6.13122821e-01 1.80526093e-01 9.03359711e-01 -3.19227934e-01
1.15065806e-01 7.86568105e-01 -3.60995620e-01 -3.06994200e-01
-4.74406272e-01 4.17579114e-01 5.50881438e-02 5.01050472e-01
4.13704276e-01 -8.31568122e-01 -4.53318238e-01 8.11566263e-02
-9.92294073e-01 1.39185593e-01 5.23532629e-01 8.27486515e-01
6.91706538e-01 -2.33036205e-01 2.40403533e-01 -9.25493121e-01
6.22842573e-02 -3.21054667e-01 -8.91443968e-01 8.54839906e-02
-5.98524749e-01 4.22492743e-01 7.83493936e-01 -5.78462660e-01
-1.20656359e+00 4.58851576e-01 -4.05676216e-01 -4.82536376e-01
-3.32798511e-01 2.54810482e-01 -1.84853628e-01 -4.91168827e-01
8.62725973e-01 2.28395462e-01 -4.25753146e-02 -5.35065591e-01
2.83852935e-01 5.50914526e-01 6.85911655e-01 -3.29762280e-01
9.41302419e-01 7.67264187e-01 4.34538089e-02 -6.52634919e-01
-7.97081709e-01 -5.97416222e-01 -1.09486151e+00 -2.63502866e-01
5.71989179e-01 -9.08226252e-01 -3.23409826e-01 1.05495846e+00
-1.11621690e+00 -8.07419360e-01 -5.40794015e-01 6.05803430e-01
-1.08984888e+00 3.43663931e-01 -2.75200427e-01 -5.53677857e-01
-1.95231721e-01 -5.44093668e-01 7.47113049e-01 3.67755443e-01
1.86482325e-01 -1.45530272e+00 6.17609799e-01 -2.06148162e-01
5.05932152e-01 1.09616369e-01 2.53149718e-01 -6.60288513e-01
-5.30352473e-01 -6.88785091e-02 1.74989909e-01 4.71831650e-01
4.52717133e-02 -2.71610945e-01 -1.16364443e+00 -4.75590527e-01
3.35405976e-01 -2.67458498e-01 9.62334216e-01 1.38581753e-01
1.01516926e+00 -1.08858667e-01 -7.51145855e-02 1.13254547e+00
1.61666059e+00 -2.38747708e-02 5.47742486e-01 9.01141226e-01
5.96533597e-01 5.34469306e-01 5.37463725e-01 7.28451073e-01
5.38788736e-01 7.50817358e-01 4.50526386e-01 9.85314548e-02
-2.39116982e-01 -1.85331792e-01 4.65088516e-01 7.67988265e-01
-2.40655512e-01 -2.54462436e-02 -7.22752452e-01 7.12076485e-01
-1.97437000e+00 -9.29670095e-01 3.78486037e-01 2.62682819e+00
6.82694614e-01 1.19962625e-01 2.53383011e-01 -4.61710915e-02
4.89984304e-01 1.71957947e-02 -6.87337220e-01 -2.50826150e-01
-1.54958442e-01 2.36309737e-01 5.99113822e-01 7.60532498e-01
-1.19810259e+00 1.00510192e+00 5.85618210e+00 3.95084321e-01
-1.48736119e+00 3.68592292e-01 3.09944332e-01 -1.41794041e-01
3.21923420e-02 -1.25297740e-01 -8.71149182e-01 2.84355968e-01
1.19625330e+00 -2.12379202e-01 6.71754897e-01 8.50644827e-01
-1.13393841e-02 -7.14975521e-02 -1.31581366e+00 1.22898197e+00
5.92410147e-01 -1.14091134e+00 -9.66129899e-02 -1.33485913e-01
7.97135949e-01 4.14870590e-01 1.97677612e-02 1.58509329e-01
-1.26196206e-01 -3.85389060e-01 8.19141805e-01 5.11561036e-01
5.03053665e-01 -7.85808265e-01 7.94735789e-01 4.26845998e-01
-1.00990069e+00 -2.94229627e-01 -3.21313620e-01 4.32302803e-02
1.29329786e-01 4.01427239e-01 -7.70873725e-01 6.61126733e-01
6.75604224e-01 1.01592743e+00 -5.05472839e-01 1.44585729e+00
-2.17115089e-01 2.16817096e-01 -2.00246170e-01 3.91500145e-01
1.33965621e-02 1.32247671e-01 6.26404822e-01 1.13306737e+00
3.63851488e-01 -1.74149826e-01 6.29527122e-02 2.29252279e-01
3.67761180e-02 2.44689167e-01 -5.76991022e-01 3.20948780e-01
1.71276659e-01 9.90561962e-01 -4.52982724e-01 -4.11966145e-01
-3.19199532e-01 1.42571008e+00 6.26695991e-01 3.51791233e-01
-5.78136742e-01 -3.96692693e-01 3.34254235e-01 1.14182811e-02
6.02460086e-01 -2.58884698e-01 -2.08554789e-02 -1.55105686e+00
3.09257507e-02 -7.24332035e-01 4.66278255e-01 -4.71161693e-01
-9.23757076e-01 8.43694627e-01 1.34988666e-01 -1.37703919e+00
-7.35590041e-01 -5.50143540e-01 -4.45654869e-01 5.24576485e-01
-2.01581287e+00 -8.50414157e-01 -5.45386195e-01 6.90320492e-01
7.71510363e-01 -2.44172573e-01 6.97275937e-01 4.26820546e-01
-3.95711541e-01 7.47623444e-01 3.39947343e-01 -3.03992242e-01
9.47951734e-01 -1.27629304e+00 3.75576049e-01 1.02679515e+00
1.65162638e-01 1.56984165e-01 7.25421190e-01 -3.16145211e-01
-1.05044615e+00 -1.04777348e+00 9.11347866e-01 -2.68526345e-01
5.63964784e-01 -3.59887809e-01 -9.93098438e-01 8.20536435e-01
8.60218629e-02 1.30854309e-01 5.28407693e-01 -2.97274217e-02
-2.18167901e-01 -4.15950716e-01 -1.01639962e+00 3.47015113e-01
1.13600922e+00 -4.22161251e-01 -4.10230845e-01 2.39355668e-01
2.31445044e-01 -8.68118405e-01 -6.29388154e-01 2.19449520e-01
5.42865753e-01 -1.30505526e+00 6.47703111e-01 -4.60576296e-01
5.25161438e-02 9.76178795e-03 -6.74942508e-02 -1.63459027e+00
-3.19806069e-01 -8.04507017e-01 -4.02578950e-01 9.22954023e-01
9.64796618e-02 -9.28227186e-01 1.11199760e+00 3.20761651e-01
-2.63111562e-01 -4.57691461e-01 -1.19425809e+00 -1.00261760e+00
9.59675387e-02 -3.15920502e-01 5.44462323e-01 5.50348639e-01
-1.58569202e-01 1.00969404e-01 -3.35112542e-01 1.24242932e-01
6.65830851e-01 -8.82479846e-02 1.01138759e+00 -1.19256234e+00
-8.48268569e-01 -1.87435791e-01 -6.73517585e-01 -1.28088272e+00
2.46017709e-01 -5.45041203e-01 3.07279319e-01 -1.00274789e+00
-8.06713535e-04 -1.74637198e-01 -4.75336403e-01 4.41244483e-01
-1.49195641e-03 1.31931633e-01 1.49846718e-01 1.90579280e-01
-6.71873868e-01 7.44953275e-01 8.33943784e-01 1.38266385e-01
-4.83484596e-01 2.29980052e-01 -1.10693581e-01 6.85964167e-01
7.91352034e-01 -5.75550199e-01 -5.66302896e-01 -8.58782291e-01
1.27576441e-01 -2.80401647e-01 2.37667963e-01 -1.36381805e+00
4.67499137e-01 -7.03451484e-02 2.14587301e-01 -3.19093645e-01
3.28378558e-01 -1.00130475e+00 -2.77675781e-02 3.48476917e-01
-6.83911815e-02 5.89778349e-02 4.10999566e-01 6.00742757e-01
-1.35812864e-01 -5.59764028e-01 9.29533541e-01 -9.60527062e-02
-1.00384724e+00 3.52186739e-01 -4.35762674e-01 6.02681413e-02
8.56823802e-01 -4.08798635e-01 5.30802310e-02 -3.89313728e-01
-8.66326988e-01 -1.46324085e-02 8.07565808e-01 2.96865702e-01
5.90037107e-01 -1.23147345e+00 -5.72191179e-01 4.56272662e-01
1.82551250e-01 -2.15932708e-02 4.20343548e-01 8.13806295e-01
-5.68031669e-01 1.88722357e-01 -3.14954519e-01 -6.57951772e-01
-1.15961826e+00 4.75129515e-01 6.58107877e-01 -5.70833161e-02
-6.27131045e-01 1.02210224e+00 8.49700049e-02 -3.62260669e-01
4.00871694e-01 -2.93560475e-01 -3.13230932e-01 -8.44510645e-02
7.47499347e-01 2.35136732e-01 3.03851873e-01 -6.24316633e-01
-1.62802279e-01 6.58162594e-01 -2.32670516e-01 -3.81199509e-01
1.45346606e+00 -3.60992581e-01 1.72509417e-01 7.61823714e-01
1.10929275e+00 -9.91244242e-02 -1.80985403e+00 -5.12655020e-01
2.79630087e-02 -7.65975893e-01 -2.27483604e-02 -2.96907008e-01
-1.04018235e+00 7.37735450e-01 1.02125680e+00 -2.78765947e-01
1.41675615e+00 -2.68834233e-01 6.99177325e-01 6.29961789e-01
5.26930511e-01 -1.51083624e+00 2.90161073e-01 7.77031183e-01
6.92584932e-01 -1.35705507e+00 -2.90855914e-01 -1.41773215e-02
-2.90491074e-01 1.48685122e+00 6.78943217e-01 -9.95354801e-02
5.59867620e-01 -1.24739036e-01 1.87308922e-01 1.22555971e-01
-8.12698364e-01 -3.03233176e-01 1.11967154e-01 6.31401300e-01
7.38773718e-02 -4.98046517e-01 1.31601259e-01 1.81805477e-01
-6.44273162e-02 2.48104379e-01 6.13336444e-01 1.16559839e+00
-6.20329320e-01 -1.39217055e+00 -1.38528064e-01 -1.29140347e-01
-1.02971658e-01 3.26735526e-02 -2.01131180e-01 5.53144336e-01
1.68162540e-01 5.57459772e-01 1.25455379e-01 -3.41215968e-01
4.66562420e-01 4.73644994e-02 8.50234151e-01 -8.63505304e-01
-4.44838077e-01 1.03873715e-01 -2.31375903e-01 -6.70175076e-01
-6.20799959e-01 -8.41010511e-01 -1.05316401e+00 1.20273061e-01
-3.25609982e-01 4.28707264e-02 8.84414136e-01 1.21889496e+00
4.09101337e-01 3.63431364e-01 8.56057227e-01 -1.25271606e+00
-4.96307939e-01 -6.78330362e-01 -3.22606117e-01 3.19374025e-01
5.43218195e-01 -5.22580862e-01 -3.77944797e-01 2.00668365e-01] | [8.672350883483887, -2.3266947269439697] |
3782594c-16f7-4020-ae37-60a0ffaabc55 | cross-document-coreference-resolution-over | 2106.0121 | null | https://arxiv.org/abs/2106.01210v1 | https://arxiv.org/pdf/2106.01210v1.pdf | Cross-document Coreference Resolution over Predicted Mentions | Coreference resolution has been mostly investigated within a single document scope, showing impressive progress in recent years based on end-to-end models. However, the more challenging task of cross-document (CD) coreference resolution remained relatively under-explored, with the few recent models applied only to gold mentions. Here, we introduce the first end-to-end model for CD coreference resolution from raw text, which extends the prominent model for within-document coreference to the CD setting. Our model achieves competitive results for event and entity coreference resolution on gold mentions. More importantly, we set first baseline results, on the standard ECB+ dataset, for CD coreference resolution over predicted mentions. Further, our model is simpler and more efficient than recent CD coreference resolution systems, while not using any external resources. | ['Ido Dagan', 'Mandar Joshi', 'Gabriel Stanovsky', 'Alon Eirew', 'Arie Cattan'] | 2021-06-02 | null | https://aclanthology.org/2021.findings-acl.453 | https://aclanthology.org/2021.findings-acl.453.pdf | findings-acl-2021-8 | ['cross-document-coreference-resolution'] | ['natural-language-processing'] | [ 1.70214057e-01 6.84615493e-01 -5.76696694e-01 -2.92912126e-01
-1.64539695e+00 -1.01581120e+00 9.64911282e-01 2.26121083e-01
-5.71419537e-01 9.38984871e-01 1.05791914e+00 -3.73056643e-02
-4.01888013e-01 -2.50624806e-01 -4.59269106e-01 -2.34214827e-01
1.03216186e-01 1.43750012e+00 3.33466619e-01 -4.44605500e-01
4.37675565e-02 3.75016659e-01 -1.06698549e+00 6.24079585e-01
7.50482976e-01 2.30389312e-02 7.34278858e-02 5.60917139e-01
-1.69408292e-01 4.62773472e-01 -4.81836319e-01 -1.05386329e+00
-3.59467179e-01 2.05503739e-02 -1.87026870e+00 -1.02686608e+00
8.71956646e-01 2.33843565e-01 -4.60879534e-01 9.67391133e-01
8.07497978e-01 2.33367398e-01 3.75391781e-01 -9.78809357e-01
-3.43852967e-01 1.52407658e+00 -4.33231145e-01 4.79934692e-01
8.63540173e-01 -6.18242800e-01 1.60636568e+00 -7.07476437e-01
1.35258603e+00 1.67447209e+00 8.08463633e-01 9.05709684e-01
-1.37856555e+00 -9.77747262e-01 1.11524299e-01 5.30271947e-01
-1.21178997e+00 -6.91002667e-01 3.09126645e-01 2.14748476e-02
1.70324743e+00 4.92504686e-01 -2.04801708e-01 1.42197824e+00
-4.47215796e-01 9.85573769e-01 3.29077184e-01 -5.54972231e-01
-5.50437093e-01 -3.76635522e-01 5.75089514e-01 -8.39363113e-02
3.30065072e-01 2.75032341e-01 -3.97033781e-01 -1.97016791e-01
-1.99487396e-02 -6.10274971e-01 -7.18627036e-01 -2.06808344e-01
-9.97417033e-01 7.50474572e-01 3.58737886e-01 6.93744123e-01
-1.28719375e-01 -3.86681914e-01 6.83382988e-01 1.01327680e-01
1.14312708e-01 5.65499365e-01 -7.37392962e-01 -3.21767896e-01
-1.04661441e+00 6.47001565e-01 1.18037772e+00 1.26699042e+00
1.30501762e-01 -9.14163172e-01 -3.86978328e-01 1.00328755e+00
1.54364750e-01 5.54043233e-01 -1.29818143e-02 -1.19713545e+00
1.05874109e+00 2.21790180e-01 2.65022248e-01 -7.48817861e-01
-8.48994970e-01 -3.14282715e-01 -5.95428824e-01 -4.59801793e-01
5.16535759e-01 -7.98177347e-02 -8.28401893e-02 2.06946445e+00
2.71650404e-01 1.38481840e-01 4.24077779e-01 7.15193272e-01
1.36886883e+00 3.30323815e-01 4.41419840e-01 -6.10287070e-01
1.68746102e+00 -8.93227279e-01 -1.04241002e+00 -3.11219990e-01
9.14476812e-01 -8.92991662e-01 2.90728450e-01 1.41415834e-01
-1.41269386e+00 -1.14618734e-01 -6.52286232e-01 -4.33626831e-01
-1.33557335e-01 -3.60395432e-01 3.80544186e-01 2.41326377e-01
-8.34874988e-01 4.85751301e-01 -6.38578355e-01 -8.09244394e-01
-7.46964812e-02 3.78823727e-01 -6.72425508e-01 -4.76312749e-02
-1.90935254e+00 1.44827199e+00 7.79611647e-01 -2.42991537e-01
-2.53371149e-01 -1.17189348e+00 -8.25019598e-01 1.78804114e-01
4.41303015e-01 -7.17446685e-01 1.76609254e+00 -9.68564600e-02
-9.74443436e-01 1.24135160e+00 -4.08484817e-01 -7.19325125e-01
3.45861912e-01 -6.86110616e-01 -9.82972264e-01 6.83123246e-02
2.41352648e-01 7.30760634e-01 -1.62800655e-01 -1.32154787e+00
-1.05144358e+00 -3.07491064e-01 -1.00974135e-01 2.60211468e-01
3.09100598e-01 8.08280647e-01 -7.07403123e-01 -3.63675326e-01
-1.88364923e-01 -9.21561539e-01 1.29174232e-01 -8.38694394e-01
-5.84108233e-01 -7.53147840e-01 6.96262419e-01 -5.58747828e-01
1.56568611e+00 -1.85911393e+00 3.05693984e-01 -3.63504171e-01
-2.59472787e-01 6.38644218e-01 -2.64808953e-01 6.37324512e-01
-6.36473775e-01 3.16846222e-02 -1.63129166e-01 -5.24761856e-01
3.31690013e-01 9.69802141e-02 -6.29579782e-01 1.58248857e-01
-2.84972414e-02 9.26104963e-01 -1.15721953e+00 -8.31637383e-01
-2.04098195e-01 4.99498397e-01 -7.04257488e-01 1.87780753e-01
-1.35063678e-01 1.78338662e-01 -2.15623483e-01 1.58335447e-01
8.45235527e-01 1.00733720e-01 7.27231383e-01 -5.36347806e-01
-3.17650825e-01 1.05206013e+00 -1.25588191e+00 2.09602857e+00
-1.87000141e-01 3.88571978e-01 4.05544072e-01 -6.90986574e-01
6.83791339e-01 7.30997682e-01 2.50165194e-01 -6.83470905e-01
-1.32295623e-01 1.50645599e-01 -1.49987601e-02 -3.19063455e-01
8.57141435e-01 -9.80239585e-02 -5.84887922e-01 4.57703352e-01
3.38557333e-01 9.59911942e-02 3.85722607e-01 7.34907329e-01
9.97519255e-01 1.18029058e-01 3.35644662e-01 -1.35745600e-01
5.70562780e-01 3.43607754e-01 7.90706277e-01 7.08337367e-01
-1.23211652e-01 6.13568246e-01 1.29992425e-01 1.76299661e-01
-4.51729178e-01 -1.01944029e+00 -5.14603376e-01 1.21431172e+00
1.79704651e-01 -7.92578697e-01 -7.82431364e-01 -9.42075551e-01
-1.08447023e-01 1.17621469e+00 -4.29694086e-01 8.63752663e-02
-1.31426048e+00 -6.49100602e-01 1.34396434e+00 5.17252982e-01
2.40132660e-01 -1.28361392e+00 -9.88589376e-02 5.56305051e-01
-1.08637834e+00 -1.33394480e+00 -8.38914871e-01 1.32308111e-01
-6.28018439e-01 -1.30342352e+00 -2.64936537e-01 -9.44348872e-01
-3.33069116e-01 1.99896973e-02 1.78011596e+00 -1.15735374e-01
-1.22548595e-01 3.62115294e-01 -2.59143502e-01 -1.84944466e-01
-3.38219017e-01 5.53016961e-01 -5.62598333e-02 -8.55616808e-01
1.01145089e+00 -3.96460503e-01 -3.10013443e-01 5.69612868e-02
-3.47907335e-01 -4.27288979e-01 2.87282646e-01 7.16454387e-01
3.20434242e-01 -6.17145538e-01 4.13907379e-01 -1.22094834e+00
6.33816779e-01 -4.97867823e-01 -1.94229819e-02 4.12610024e-01
-5.99825501e-01 2.19400436e-01 4.89195772e-02 -6.80235177e-02
-1.80727589e+00 -4.09546524e-01 -4.19616193e-01 2.11980902e-02
-4.15252149e-01 3.89659077e-01 -4.87638831e-01 6.75222993e-01
6.06187105e-01 -3.67295891e-01 -6.64950967e-01 -1.01381910e+00
7.08211482e-01 6.07759833e-01 1.57578981e+00 -1.02659142e+00
4.03446496e-01 1.22811742e-01 -4.21902031e-01 -1.69753656e-01
-1.16234052e+00 -8.67415369e-01 -1.18748987e+00 4.61028069e-01
6.78152323e-01 -8.93453360e-01 -9.47766304e-01 -1.45487770e-01
-1.72967803e+00 2.05296472e-01 9.24775600e-02 4.64551568e-01
-3.71444792e-01 5.79316556e-01 -8.55088711e-01 -3.15351635e-01
-9.87616360e-01 -6.72042012e-01 1.04620433e+00 1.83340892e-01
-1.01296115e+00 -1.05084610e+00 8.85021329e-01 5.61377347e-01
-2.03677163e-01 -3.93154807e-02 9.09660399e-01 -1.08693647e+00
-1.84836775e-01 4.24841791e-03 -2.83527613e-01 -6.77708387e-01
-2.88278699e-01 -6.28502294e-02 -7.12827086e-01 -3.85319740e-01
-6.08156919e-01 -1.94015071e-01 9.62892890e-01 1.82083830e-01
1.14231586e-01 -1.64976984e-01 -1.18251061e+00 6.42062783e-01
1.15572095e+00 -8.71201220e-04 6.07153594e-01 8.32309783e-01
3.47826362e-01 5.56887090e-01 9.93807435e-01 1.67010315e-02
7.56800771e-01 1.12972188e+00 1.82623163e-01 2.34703064e-01
-5.11778772e-01 -3.09901118e-01 3.54131944e-02 6.40084684e-01
-8.27012062e-02 -3.87722760e-01 -1.04375935e+00 9.74674225e-01
-2.00413513e+00 -1.58717263e+00 -5.57982862e-01 1.68198431e+00
1.34464979e+00 2.10722163e-02 -2.82120872e-02 -1.76398784e-01
1.08184433e+00 5.48417419e-02 -1.13011651e-01 -4.95955795e-01
-5.28658628e-01 4.64068949e-01 1.15701325e-01 8.48509133e-01
-1.32158089e+00 1.15170169e+00 6.77879286e+00 3.75419378e-01
-3.91145885e-01 1.69520766e-01 -4.44831073e-01 -2.28906870e-01
-2.52125680e-01 2.46124178e-01 -1.58088911e+00 1.17962144e-01
1.08756185e+00 -1.68889448e-01 -5.72616085e-02 5.07298470e-01
-2.52122581e-01 3.19671214e-01 -1.38536978e+00 7.32614219e-01
-9.73364338e-02 -1.39544821e+00 2.13187817e-03 -4.11277086e-01
4.23590481e-01 4.17896599e-01 -4.82366920e-01 6.40877366e-01
1.08650815e+00 -8.83778155e-01 4.48448598e-01 2.61135191e-01
8.21476519e-01 -1.02877450e+00 9.04029012e-01 2.80109704e-01
-1.26053786e+00 1.85903937e-01 -1.94783702e-01 4.84043926e-01
7.68900037e-01 4.90758084e-02 -5.45502603e-01 1.02828646e+00
8.63792539e-01 6.61796629e-01 -4.98669967e-02 8.66404176e-01
-3.47778261e-01 3.23406249e-01 -9.21822861e-02 5.62342346e-01
1.65299401e-01 4.00121659e-01 8.31647515e-01 2.23637915e+00
8.39062780e-02 6.13166034e-01 -5.62618226e-02 6.18725955e-01
-3.08851749e-01 -3.90476659e-02 -6.24734201e-02 4.93617773e-01
1.35081875e+00 1.26569295e+00 2.30624184e-01 -8.46829265e-02
-5.31238556e-01 7.73247063e-01 9.39855993e-01 2.01337622e-03
-6.82497740e-01 -4.75039244e-01 9.55301344e-01 -4.47213262e-01
3.87309343e-01 4.27050799e-01 1.11478493e-01 -1.01634192e+00
-5.87603986e-01 -1.05218244e+00 1.48968744e+00 -4.59126353e-01
-1.24500918e+00 5.56673348e-01 2.47897267e-01 -6.74683094e-01
-6.25572681e-01 -1.51069820e-01 -6.08242035e-01 9.90578532e-01
-1.53420138e+00 -1.19749272e+00 1.18303761e-01 7.53882825e-01
2.21203566e-01 2.45608628e-01 1.29543114e+00 5.43360054e-01
-5.69454551e-01 1.05323076e+00 1.60088912e-02 6.01750731e-01
1.46593821e+00 -1.37641108e+00 5.58131218e-01 8.31356585e-01
1.82838619e-01 1.03466272e+00 1.16163468e+00 -6.20129824e-01
-9.72657263e-01 -8.52121830e-01 1.86864674e+00 -1.01985216e+00
6.04732096e-01 -1.24707762e-02 -1.22910798e+00 1.14346278e+00
6.73311770e-01 -2.96341449e-01 6.04389906e-01 8.81492674e-01
-8.11697721e-01 3.59793663e-01 -1.18482542e+00 3.58895779e-01
1.52699518e+00 -5.02035201e-01 -1.71623492e+00 -7.86391571e-02
7.86425889e-01 -7.72145450e-01 -1.39801240e+00 7.93036103e-01
2.86483675e-01 -5.24945676e-01 1.36768198e+00 -9.81679142e-01
6.53993338e-02 -8.78076106e-02 -5.77690303e-02 -1.32307196e+00
-8.57940495e-01 -6.81937754e-01 -4.08367842e-01 1.97402894e+00
6.54612541e-01 4.88121659e-02 3.89997751e-01 6.37698114e-01
-3.69593561e-01 3.07198167e-01 -9.94068563e-01 -5.16307831e-01
6.24516785e-01 -1.81024119e-01 6.44553006e-01 1.30274558e+00
8.52187276e-01 7.95849562e-01 -1.29032522e-01 3.11294943e-01
8.14036191e-01 5.93065798e-01 4.99866545e-01 -1.74330604e+00
-8.29552561e-02 -4.80050713e-01 3.91056091e-01 -9.80893373e-01
8.35390508e-01 -1.27985907e+00 2.66971827e-01 -1.41789782e+00
7.43962646e-01 -2.81162411e-01 -1.12334639e-01 4.61105704e-01
-5.06258547e-01 -4.70942333e-02 2.74617136e-01 3.50876361e-01
-8.19959462e-01 1.38882294e-01 7.59479940e-01 -1.85699433e-01
3.33646946e-02 -2.77252376e-01 -9.22672272e-01 5.32452464e-01
2.59378701e-01 -6.32193565e-01 9.68650281e-02 -5.19150853e-01
-4.34919856e-02 3.77040327e-01 -1.80799320e-01 -3.71680945e-01
6.81605637e-01 5.71413673e-02 -1.11064382e-01 -9.77062345e-01
2.21952554e-02 -4.54354376e-01 2.31441975e-01 1.70374721e-01
-5.72409630e-01 -1.82714779e-02 4.52667683e-01 3.79032791e-01
-3.80336910e-01 -3.15920413e-01 7.53419042e-01 -3.72759365e-02
-9.13604438e-01 6.79433271e-02 1.29669532e-01 1.00910318e+00
4.45480168e-01 3.35642129e-01 -8.65596414e-01 3.43905762e-02
-1.00483680e+00 4.90158617e-01 1.61003038e-01 8.74557495e-01
-7.52144977e-02 -1.24226868e+00 -1.14543974e+00 -6.23258710e-01
2.32430339e-01 -7.85992667e-02 2.15039581e-01 8.11188996e-01
2.54352897e-01 1.30596256e+00 8.68823677e-02 -4.21563148e-01
-1.86354864e+00 8.43384862e-01 3.30354184e-01 -8.10328484e-01
-8.75199914e-01 8.11257064e-01 -5.77478856e-02 -8.74713898e-01
4.55438018e-01 1.55895352e-01 -6.89554214e-01 4.34650570e-01
1.04851449e+00 4.25178111e-01 2.83577561e-01 -9.55175996e-01
-9.86164868e-01 5.27942121e-01 -5.87582111e-01 -5.86128235e-02
1.33792043e+00 -3.79521757e-01 -1.11431107e-01 -1.96451023e-01
9.70812380e-01 9.97720286e-02 -6.95676088e-01 -6.68532848e-01
8.60897422e-01 9.09110606e-02 -1.28778905e-01 -1.22520268e+00
-7.71473706e-01 5.43855727e-01 4.17684466e-02 -2.60645330e-01
6.86423838e-01 5.23031652e-01 7.95976281e-01 5.12825608e-01
3.76392663e-01 -9.45953667e-01 -7.45793045e-01 1.05899847e+00
9.91048217e-01 -9.32777703e-01 -2.28327826e-01 -4.72382218e-01
-5.14852285e-01 8.83252323e-01 6.55985713e-01 3.01219553e-01
4.34292495e-01 7.94015110e-01 3.27630676e-02 -2.20898643e-01
-1.07495570e+00 -2.23630145e-01 2.74757057e-01 7.21527278e-01
9.94088173e-01 2.83780191e-02 -5.07961392e-01 1.14119577e+00
-3.02349776e-01 -1.11520156e-01 1.43793121e-01 3.25743914e-01
-2.62849778e-01 -1.45249605e+00 -2.79813647e-01 -3.12997073e-01
-9.36570823e-01 -5.33418715e-01 -6.71213567e-01 1.12507796e+00
-9.96190161e-02 1.07223272e+00 3.00930828e-01 1.26396075e-01
8.86782110e-01 2.28739470e-01 6.35221243e-01 -4.92688239e-01
-8.57210219e-01 -3.29168469e-01 9.26795900e-01 -6.73599720e-01
-7.75589108e-01 -1.22618783e+00 -1.71456778e+00 -6.48240745e-01
-4.01562303e-01 7.20321834e-01 -2.18073979e-01 1.06197679e+00
4.07748967e-01 1.48546115e-01 -4.94425409e-02 -5.36075413e-01
-2.74282068e-01 -1.25778139e+00 -2.88203508e-01 8.82015049e-01
1.26785412e-01 -6.85782135e-01 -2.21289635e-01 -2.32739553e-01] | [9.306904792785645, 9.555166244506836] |
977f36f3-b301-4813-b823-d46617d6b38d | volumetric-lung-nodule-segmentation-using | 1912.13335 | null | https://arxiv.org/abs/1912.13335v2 | https://arxiv.org/pdf/1912.13335v2.pdf | Volumetric Lung Nodule Segmentation using Adaptive ROI with Multi-View Residual Learning | Accurate quantification of pulmonary nodules can greatly assist the early diagnosis of lung cancer, which can enhance patient survival possibilities. A number of nodule segmentation techniques have been proposed, however, all of the existing techniques rely on radiologist 3-D volume of interest (VOI) input or use the constant region of interest (ROI) and only investigate the presence of nodule voxels within the given VOI. Such approaches restrain the solutions to investigate the nodule presence outside the given VOI and also include the redundant structures into VOI, which may lead to inaccurate nodule segmentation. In this work, a novel semi-automated approach for 3-D segmentation of nodule in volumetric computerized tomography (CT) lung scans has been proposed. The proposed technique can be segregated into two stages, at the first stage, it takes a 2-D ROI containing the nodule as input and it performs patch-wise investigation along the axial axis with a novel adaptive ROI strategy. The adaptive ROI algorithm enables the solution to dynamically select the ROI for the surrounding slices to investigate the presence of nodule using deep residual U-Net architecture. The first stage provides the initial estimation of nodule which is further utilized to extract the VOI. At the second stage, the extracted VOI is further investigated along the coronal and sagittal axis with two different networks and finally, all the estimated masks are fed into the consensus module to produce the final volumetric segmentation of nodule. The proposed approach has been rigorously evaluated on the LIDC dataset, which is the largest publicly available dataset. The result suggests that the approach is significantly robust and accurate as compared to the previous state of the art techniques. | ['Byung-ilLee', 'Sung Hyun Kim', 'Byoung-Dai Lee', 'Muhammad Usman', 'Shi Sub Byon'] | 2019-12-31 | null | null | null | null | ['lung-nodule-segmentation'] | ['medical'] | [ 2.73633510e-01 2.79191136e-01 -9.05742049e-02 -3.55145074e-02
-4.21814978e-01 -2.82028019e-01 3.01768273e-01 -1.11299209e-01
-4.04727757e-01 4.48128104e-01 -1.71679556e-01 -3.89853716e-01
-2.93800890e-01 -7.84121096e-01 -2.10895360e-01 -7.48296797e-01
2.31919974e-01 7.37209201e-01 9.26508248e-01 2.23089144e-01
9.17505398e-02 9.33737278e-01 -1.13800144e+00 -4.18016501e-02
7.29818523e-01 1.05580497e+00 3.97462606e-01 4.86983925e-01
-3.98462802e-01 5.17596781e-01 -2.22570583e-01 2.88970649e-01
4.59787250e-01 -5.61149836e-01 -8.22544277e-01 5.95766902e-01
5.40782930e-04 -2.27032468e-01 1.62021443e-01 8.39025736e-01
4.12400931e-01 -1.21391684e-01 8.15860868e-01 -7.91488588e-01
1.81136891e-01 3.74353826e-01 -7.05031753e-01 3.93306524e-01
-3.90218914e-01 2.82091975e-01 6.54431999e-01 -1.01246011e+00
5.53865194e-01 6.69459701e-01 5.62310815e-01 3.07454914e-01
-7.02728152e-01 -4.18283820e-01 -2.69056082e-01 -8.14190879e-02
-1.61480045e+00 8.30409899e-02 7.56829143e-01 -5.56763887e-01
6.02094889e-01 4.99220401e-01 9.81577992e-01 1.28675714e-01
2.48382375e-01 4.95965719e-01 1.02409923e+00 -3.26326281e-01
7.86281005e-02 2.73485690e-01 1.08104832e-02 9.26550448e-01
2.07045466e-01 -3.93583551e-02 4.96622950e-01 -7.96603560e-02
1.19059169e+00 1.71246365e-01 -3.94598991e-01 -5.68994403e-01
-1.21924543e+00 7.49643326e-01 8.12180102e-01 9.51423764e-01
-6.66908503e-01 -6.98654577e-02 2.34971672e-01 -3.28379959e-01
3.10018748e-01 2.67355442e-01 -7.90665150e-02 5.00110984e-01
-1.43816471e+00 -5.93908690e-02 6.47399127e-01 3.74642044e-01
6.50570750e-01 -1.03184655e-02 -4.03660387e-01 6.22479856e-01
5.49023509e-01 8.14494025e-03 7.73596525e-01 -4.95457232e-01
2.51846045e-01 9.89205182e-01 -1.50664330e-01 -7.31476665e-01
-6.00572348e-01 -6.86007261e-01 -9.47811007e-01 2.35198557e-01
3.53036344e-01 -1.97594509e-01 -1.28530133e+00 9.42041218e-01
8.24288368e-01 2.26697221e-01 -2.90396869e-01 1.04771197e+00
1.06624901e+00 2.09926471e-01 -1.56951576e-01 -2.87262231e-01
1.47792566e+00 -1.00383604e+00 -4.91268426e-01 5.38057908e-02
4.90382671e-01 -7.03853428e-01 6.89443111e-01 -7.09565505e-02
-1.02261829e+00 -5.02683520e-01 -1.11579490e+00 3.70065242e-01
-1.00214489e-01 4.64469224e-01 1.43924788e-01 6.45895958e-01
-1.06315291e+00 5.90480924e-01 -9.76587951e-01 -2.88962781e-01
5.89122057e-01 6.10004485e-01 -1.08430751e-01 2.28402048e-01
-9.91985083e-01 8.35232079e-01 7.16147065e-01 3.64842206e-01
-8.33010733e-01 -5.94847023e-01 -4.81017798e-01 7.62007236e-02
6.32593453e-01 -6.58058167e-01 9.71321702e-01 -8.96324873e-01
-1.32235956e+00 6.57940865e-01 -3.36668454e-02 -3.09558898e-01
8.59375536e-01 5.49597919e-01 7.20892921e-02 2.50906616e-01
2.35404801e-02 5.34096837e-01 7.84553528e-01 -1.10958397e+00
-6.00951195e-01 -3.72057378e-01 -2.88265586e-01 3.82786542e-01
-8.66286755e-02 -1.98970780e-01 -6.67758286e-01 -4.46733385e-01
5.20409942e-01 -8.47882450e-01 -5.07396162e-01 6.72264919e-02
-5.99440098e-01 -2.06593975e-01 1.17857492e+00 -7.11011589e-01
1.23730707e+00 -1.84290373e+00 -7.87535030e-03 6.89845681e-01
4.29777175e-01 2.20394567e-01 4.43495154e-01 -2.93358564e-01
-1.34683564e-01 2.49131277e-01 -5.65446794e-01 2.21765682e-01
-5.42124689e-01 -2.43743267e-02 6.10258341e-01 4.54877049e-01
2.89885342e-01 9.82900083e-01 -4.42730218e-01 -1.17618001e+00
4.43388760e-01 4.18573469e-01 -4.14953798e-01 2.20467493e-01
-1.77225098e-01 5.97631276e-01 -7.57483602e-01 6.13127947e-01
7.33965039e-01 -3.61064196e-01 1.29419759e-01 -3.31317544e-01
-3.61475080e-01 -1.84383258e-01 -1.37250125e+00 1.15260541e+00
-1.78286254e-01 1.87074155e-01 3.27628702e-01 -8.56772780e-01
8.48477125e-01 7.18871534e-01 1.05746853e+00 -4.17233892e-02
5.67772329e-01 4.23892707e-01 5.21620750e-01 -5.28509915e-01
-2.08909977e-02 -2.11541504e-01 3.53732795e-01 5.68877816e-01
-2.89585173e-01 -3.86473805e-01 1.60875365e-01 -1.16484664e-01
1.02020860e+00 6.38212115e-02 6.57098353e-01 -2.57782996e-01
1.05456793e+00 2.73700088e-01 3.17253798e-01 3.22201848e-01
-3.92512202e-01 6.30558133e-01 3.29917222e-01 -1.48199677e-01
-9.43999231e-01 -8.59521866e-01 -3.34674001e-01 1.15614220e-01
2.37654764e-02 2.69157797e-01 -8.90182197e-01 -1.04196179e+00
-1.73501566e-01 3.01986486e-01 -6.58374310e-01 4.23534483e-01
-6.75785422e-01 -7.45174587e-01 1.86167285e-01 4.40063566e-01
8.57620001e-01 -1.27760744e+00 -8.97563577e-01 1.81726664e-01
-8.25985968e-02 -7.21195221e-01 -3.29805344e-01 3.31975430e-01
-1.19087744e+00 -1.26285267e+00 -1.12665355e+00 -6.97369814e-01
9.87527907e-01 1.83383435e-01 7.63113856e-01 3.86414081e-01
-6.14724457e-01 8.42723474e-02 2.93062720e-02 -2.23046169e-01
-4.50162172e-01 1.83571711e-01 -6.22289419e-01 -6.37734681e-02
-1.99011583e-02 -2.41836384e-01 -6.94338083e-01 5.47186196e-01
-8.57639909e-01 -2.15224028e-02 1.07157350e+00 6.51773274e-01
1.04895186e+00 3.92702162e-01 4.25905645e-01 -1.04957283e+00
5.19851089e-01 -6.84620798e-01 -4.91723895e-01 1.73482478e-01
-3.34337085e-01 -7.12233111e-02 5.11649966e-01 -3.33830982e-01
-1.11528575e+00 4.49557066e-01 -1.80861756e-01 -5.41784286e-01
-9.63171795e-02 5.44634879e-01 8.52233842e-02 -2.70719439e-01
5.24754047e-01 2.17903167e-01 1.85765252e-01 -6.80786893e-02
-6.89333901e-02 7.46815443e-01 1.13398045e-01 1.43682212e-01
9.08698022e-01 5.39052665e-01 2.91024715e-01 -7.52317309e-01
-4.47121799e-01 -8.45599353e-01 -1.17631698e+00 -5.84137142e-01
1.06260920e+00 -2.54027367e-01 -2.42379606e-01 5.84870670e-03
-8.29168200e-01 -6.36409596e-02 -3.47371519e-01 6.23363793e-01
-3.66662532e-01 4.79997307e-01 -3.23404908e-01 -6.05063140e-01
-6.93777859e-01 -1.45147943e+00 6.14773214e-01 4.01736140e-01
-2.12838501e-01 -7.74271071e-01 -4.90841046e-02 2.85567462e-01
5.14528513e-01 3.17883193e-01 9.59650695e-01 -7.85394967e-01
-7.80597448e-01 -2.53738075e-01 -3.75293761e-01 2.70109624e-01
4.26555812e-01 1.26958057e-01 -5.94182789e-01 -2.82697473e-02
3.35669249e-01 -9.42820590e-03 6.51945829e-01 7.10943460e-01
1.11746800e+00 8.79936665e-02 -6.11838877e-01 4.77675378e-01
1.62733889e+00 3.48302543e-01 3.42980802e-01 1.83481440e-01
7.83574402e-01 4.63155866e-01 5.82243741e-01 1.61303237e-01
-1.68258220e-01 3.03859979e-01 7.42798686e-01 -3.78464252e-01
-2.28383750e-01 3.14618915e-01 -2.98857510e-01 7.02155828e-01
-3.61257046e-01 1.60727929e-02 -9.42754984e-01 6.81268573e-01
-1.39779401e+00 -6.29449725e-01 -2.97759145e-01 2.12704277e+00
4.88674581e-01 1.75351605e-01 8.96904767e-02 3.25949311e-01
8.49198341e-01 7.30747804e-02 -6.54807925e-01 -1.01087637e-01
5.80945194e-01 3.81435275e-01 5.66672623e-01 2.60424465e-01
-1.14045870e+00 5.06031573e-01 5.65417862e+00 7.78302729e-01
-1.25078249e+00 2.39299536e-01 6.10281169e-01 -1.51410438e-02
-8.06169286e-02 -2.03545079e-01 -5.64446867e-01 1.66725293e-01
5.15455663e-01 1.34569015e-02 1.85492495e-03 7.61553586e-01
3.64063442e-01 -4.06869829e-01 -8.22319746e-01 4.03826982e-01
-4.82447743e-02 -1.09531045e+00 -8.34833756e-02 1.43888205e-01
6.50093913e-01 5.12038544e-02 -3.04495335e-01 1.36476144e-01
-2.00826868e-01 -1.11632597e+00 1.53014019e-01 4.71067041e-01
6.96530402e-01 -7.06019759e-01 1.06812584e+00 6.44199789e-01
-1.47133040e+00 1.55669168e-01 -3.66547942e-01 6.45065904e-01
-1.18062096e-02 4.81053591e-01 -1.65627170e+00 7.59475648e-01
4.59760368e-01 2.17615992e-01 -6.97529852e-01 1.36982739e+00
1.53409362e-01 6.23834610e-01 -5.08015573e-01 -1.67740032e-01
4.58023578e-01 -2.83122420e-01 6.08575583e-01 8.58033538e-01
5.94286799e-01 3.64486910e-02 2.00170815e-01 1.16387451e+00
1.91005487e-02 7.06461668e-01 -3.08958292e-01 2.38610268e-01
1.55960813e-01 1.57180941e+00 -1.44191504e+00 -3.41780066e-01
-3.58641922e-01 6.05332792e-01 9.84144285e-02 -5.07887732e-03
-8.35522175e-01 -3.24864648e-02 -5.06190658e-01 6.43987894e-01
4.85321790e-01 2.50797778e-01 -2.60906875e-01 -5.20136535e-01
-7.47672915e-02 -4.97589201e-01 4.54447329e-01 -5.09837806e-01
-6.80107355e-01 8.08766782e-01 -5.13457134e-02 -1.33104956e+00
-3.17457706e-01 -4.25648540e-01 -9.80910540e-01 9.51364994e-01
-1.28948355e+00 -8.87518942e-01 -7.17668056e-01 3.94060642e-01
5.92927873e-01 1.31948069e-01 4.24716413e-01 1.41385332e-01
-5.09286702e-01 4.47193123e-02 -9.20110866e-02 1.71440452e-01
3.21221948e-01 -1.19662035e+00 -3.04104418e-01 7.61633396e-01
-2.44517565e-01 1.14083409e-01 1.61111981e-01 -1.01720667e+00
-7.47613966e-01 -1.30939388e+00 4.29088473e-01 -1.11824408e-01
4.82241631e-01 2.39526749e-01 -9.71629024e-01 5.68406343e-01
1.55562265e-02 4.67785984e-01 3.34211260e-01 -7.84682155e-01
6.71806335e-01 2.17031166e-01 -1.50868654e+00 4.27278370e-01
3.43215883e-01 1.22343734e-01 -3.42668921e-01 1.33916572e-01
3.02442044e-01 -4.88625616e-01 -1.13258839e+00 6.59859836e-01
4.49942589e-01 -9.94586825e-01 9.04639363e-01 6.88977614e-02
3.38183135e-01 -5.13965011e-01 4.01652008e-01 -9.55857575e-01
-4.17976260e-01 1.49471045e-01 1.70709267e-01 8.80905688e-01
5.43751478e-01 -3.99603426e-01 1.17307556e+00 3.41278732e-01
-1.74314335e-01 -1.27957726e+00 -9.71974790e-01 -3.16150457e-01
-1.32252082e-01 -3.01942259e-01 3.83174896e-01 5.71208358e-01
-6.25046313e-01 -1.67220473e-01 2.95714766e-01 -3.71783003e-02
5.38192809e-01 -1.47346186e-03 3.63474727e-01 -1.39862216e+00
-2.15338841e-01 -6.34233236e-01 -2.57931143e-01 -6.15925431e-01
-4.46063876e-01 -1.20133376e+00 1.17462307e-01 -1.92125309e+00
2.80166149e-01 -5.04939616e-01 -3.06387722e-01 1.86227694e-01
-2.16142893e-01 2.88129508e-01 -7.38659874e-02 5.71822345e-01
-3.59004363e-02 1.08883344e-01 1.92371750e+00 -9.35691148e-02
-3.60281438e-01 6.25699580e-01 -3.00558448e-01 8.84105146e-01
7.99958646e-01 -3.99487823e-01 -4.58804965e-01 2.75458843e-01
-5.69852233e-01 3.08393061e-01 4.31215346e-01 -1.14912534e+00
2.07080871e-01 -5.04657365e-02 7.72430599e-01 -1.19271195e+00
1.01178989e-01 -1.29422665e+00 1.49911553e-01 8.43369246e-01
3.17075439e-02 -3.42507929e-01 6.03516735e-02 4.41285938e-01
-1.66829586e-01 -7.22634196e-01 1.01011968e+00 -6.91910684e-01
-3.01806659e-01 5.13843775e-01 -3.97486746e-01 -2.29307175e-01
1.65883470e+00 -6.51060343e-01 5.11174262e-01 6.45874590e-02
-9.37613964e-01 1.21410206e-01 1.31106719e-01 -2.86469340e-01
5.12684464e-01 -1.23668456e+00 -7.23609686e-01 9.48628187e-02
-2.79071122e-01 4.35766935e-01 3.22623461e-01 1.37029994e+00
-8.09583545e-01 6.09140217e-01 -2.65686749e-03 -8.95790696e-01
-1.47868323e+00 4.85202938e-01 8.70876849e-01 -8.17131758e-01
-5.48510551e-01 6.34515762e-01 1.19734392e-01 -3.43439400e-01
9.66897409e-04 -5.93058705e-01 -7.13792026e-01 -3.72150838e-02
-8.14808309e-02 3.05974692e-01 1.15558088e-01 -7.60320485e-01
-3.99398148e-01 6.25213027e-01 6.78812340e-02 -2.54497980e-04
9.54674184e-01 2.15048939e-02 -2.56381541e-01 1.81435123e-01
9.87793446e-01 -2.83376072e-02 -9.36961889e-01 -2.20930517e-01
4.94134650e-02 -1.84556171e-01 3.02232504e-01 -7.66002774e-01
-1.36993611e+00 7.24524438e-01 6.89262271e-01 1.78585574e-01
1.01094246e+00 -2.44101230e-02 6.20507181e-01 -8.59613791e-02
-6.38697892e-02 -7.16695905e-01 -4.74455319e-02 1.03313692e-01
7.10364461e-01 -1.17401910e+00 2.11537883e-01 -7.68971741e-01
-4.96369153e-01 1.30063343e+00 8.74257147e-01 -2.41358176e-01
7.97418714e-01 3.70465338e-01 -1.69288851e-02 -4.66123998e-01
-3.19281787e-01 -3.10226709e-01 7.20093012e-01 2.17551768e-01
6.64891064e-01 -3.58673744e-02 -5.14070928e-01 4.39300150e-01
-5.71639724e-02 1.68259814e-01 3.06506753e-01 8.08454335e-01
-7.14265108e-01 -7.14558482e-01 -7.05860257e-01 8.67427707e-01
-5.82079589e-01 6.88956231e-02 -4.33800757e-01 1.30651271e+00
3.65100265e-01 5.28839231e-01 1.02608100e-01 1.48961484e-01
1.83903262e-01 1.20161168e-01 3.33578229e-01 -7.95348942e-01
-1.03696215e+00 4.19681042e-01 -2.37888396e-01 -2.60324240e-01
-4.53987598e-01 -5.38324952e-01 -1.63228405e+00 3.80599707e-01
-6.56705916e-01 2.43422627e-01 6.01290762e-01 1.00068629e+00
-1.87247545e-01 9.13142681e-01 6.67113721e-01 -8.37362945e-01
-3.85108024e-01 -1.11547649e+00 -3.35796863e-01 2.88198125e-02
-4.08210196e-02 -6.02416694e-01 -5.31669319e-01 -3.38602781e-01] | [15.309381484985352, -2.1599204540252686] |
1635af0d-7b27-4077-a9bd-9c09097cab04 | unsupervised-domain-adaptation-for-cardiac | 2204.09334 | null | https://arxiv.org/abs/2204.09334v3 | https://arxiv.org/pdf/2204.09334v3.pdf | Unsupervised Domain Adaptation for Cardiac Segmentation: Towards Structure Mutual Information Maximization | Unsupervised domain adaptation approaches have recently succeeded in various medical image segmentation tasks. The reported works often tackle the domain shift problem by aligning the domain-invariant features and minimizing the domain-specific discrepancies. That strategy works well when the difference between a specific domain and between different domains is slight. However, the generalization ability of these models on diverse imaging modalities remains a significant challenge. This paper introduces UDA-VAE++, an unsupervised domain adaptation framework for cardiac segmentation with a compact loss function lower bound. To estimate this new lower bound, we develop a novel Structure Mutual Information Estimation (SMIE) block with a global estimator, a local estimator, and a prior information matching estimator to maximize the mutual information between the reconstruction and segmentation tasks. Specifically, we design a novel sequential reparameterization scheme that enables information flow and variance correction from the low-resolution latent space to the high-resolution latent space. Comprehensive experiments on benchmark cardiac segmentation datasets demonstrate that our model outperforms previous state-of-the-art qualitatively and quantitatively. The code is available at https://github.com/LOUEY233/Toward-Mutual-Information}{https://github.com/LOUEY233/Toward-Mutual-Information | ['Gaurav Gupta', 'Shen Zheng', 'Changjie Lu'] | 2022-04-20 | null | null | null | null | ['cardiac-segmentation', 'mutual-information-estimation'] | ['medical', 'methodology'] | [ 2.17094392e-01 -1.57179050e-02 -3.06291401e-01 -5.71052313e-01
-1.26204598e+00 -3.75228405e-01 1.80351406e-01 -4.91144806e-02
-3.65173459e-01 7.32004941e-01 2.25610226e-01 1.32977039e-01
-2.55605608e-01 -3.93149287e-01 -4.01218474e-01 -9.49278235e-01
2.47000307e-01 5.94717860e-01 2.76401609e-01 2.28877947e-01
7.61411786e-02 8.21595117e-02 -8.39078784e-01 1.50313169e-01
1.16342676e+00 7.81731009e-01 5.05084336e-01 3.77876908e-01
1.10175565e-01 3.69188935e-01 -4.10541855e-02 4.73330542e-02
2.87483662e-01 -7.14507878e-01 -9.65615571e-01 2.01665342e-01
7.64614940e-02 -1.88930392e-01 -1.94808781e-01 1.23504841e+00
6.99124157e-01 9.58401803e-03 9.56654489e-01 -9.71648455e-01
-3.71519387e-01 4.69954967e-01 -8.10417473e-01 3.83015722e-01
7.54184648e-02 -8.06037933e-02 7.41969407e-01 -6.66190028e-01
8.94313335e-01 7.92977154e-01 6.53664649e-01 6.07711375e-01
-1.51447463e+00 -7.77433634e-01 -9.52045247e-02 3.45369317e-02
-1.40995109e+00 -1.87319055e-01 1.00968969e+00 -8.43362987e-01
3.74192357e-01 9.32457857e-04 1.99518755e-01 8.98689032e-01
2.70333529e-01 8.43127906e-01 1.18325865e+00 -1.38479948e-01
6.41052946e-02 2.15808555e-01 1.07014380e-01 6.43752635e-01
6.09999113e-02 -4.65530567e-02 -3.44939411e-01 -2.89115638e-01
1.05603898e+00 -6.18040226e-02 -3.19574922e-01 -8.48059833e-01
-1.38398218e+00 7.04446912e-01 2.43625686e-01 4.70498264e-01
-3.31750125e-01 -3.27357471e-01 4.52799946e-01 1.52351499e-01
6.27914369e-01 2.27311566e-01 -4.49956030e-01 1.38702482e-01
-1.15561807e+00 3.68446112e-02 4.88229960e-01 9.30282533e-01
6.26830459e-01 -3.47690940e-01 -1.27679437e-01 1.10498357e+00
4.00735289e-01 3.39353681e-01 6.47492826e-01 -1.13786495e+00
3.48317951e-01 2.79273421e-01 -1.32376656e-01 -9.00713861e-01
-4.51645255e-01 -5.29520929e-01 -1.20185971e+00 -2.16138959e-01
4.97957855e-01 -1.85667872e-01 -6.64373934e-01 1.86008501e+00
6.32911921e-01 3.09493721e-01 -1.30145654e-01 1.09598505e+00
7.93733001e-01 4.22074407e-01 1.20846480e-01 -4.40154642e-01
1.35384941e+00 -8.25957358e-01 -7.67171919e-01 -2.23493859e-01
5.30582666e-01 -8.91006529e-01 8.57193649e-01 1.34102330e-01
-1.15333843e+00 -5.92138410e-01 -9.47225988e-01 -1.03289045e-01
2.31697053e-01 1.62502900e-01 1.84614480e-01 3.98941696e-01
-7.88470566e-01 6.13555789e-01 -1.13215613e+00 -2.25798994e-01
4.79790121e-01 2.53006548e-01 -3.92189920e-01 2.78409868e-02
-1.05676663e+00 5.85498512e-01 4.56503570e-01 -2.80084997e-01
-6.71597481e-01 -8.92559588e-01 -8.02808940e-01 -4.17599350e-01
1.70254350e-01 -8.64179313e-01 1.02678609e+00 -8.23836267e-01
-1.50598943e+00 1.21229947e+00 -1.66671678e-01 -2.90268868e-01
6.59585893e-01 -6.33270741e-02 -1.22476004e-01 3.00131619e-01
4.78691489e-01 5.96671164e-01 6.87334955e-01 -1.08248818e+00
-2.70231664e-01 -3.79609674e-01 -5.86909890e-01 3.76954824e-01
-1.44007072e-01 -6.98140189e-02 -5.65506637e-01 -9.20414388e-01
5.92320561e-01 -9.85624075e-01 -3.23638022e-01 -6.53051189e-04
-3.16948980e-01 8.92156065e-02 4.66504902e-01 -8.26345265e-01
1.24632287e+00 -2.29979467e+00 3.56959134e-01 9.15977880e-02
3.94243777e-01 -6.41460493e-02 1.30955398e-01 2.39623971e-02
-1.82235017e-01 -1.15180790e-01 -9.32556510e-01 -3.41120631e-01
-2.96828568e-01 4.44919392e-02 1.29985765e-01 7.35977471e-01
-4.45577502e-02 5.57022393e-01 -9.29290175e-01 -9.75550234e-01
2.34193355e-01 4.28818673e-01 -6.44444346e-01 2.88287431e-01
2.44042724e-01 1.14443457e+00 -6.86424732e-01 4.54065591e-01
1.00271630e+00 -6.34931326e-01 3.10129970e-01 -2.45672882e-01
3.38343605e-02 -1.54500706e-02 -1.17347968e+00 2.05247879e+00
-1.91394135e-01 3.58795524e-01 2.20736131e-01 -1.28782773e+00
8.69846344e-01 3.51472080e-01 1.13467824e+00 -5.07371962e-01
1.60332873e-01 3.75076532e-01 -2.54705071e-01 -3.76260489e-01
-5.89244924e-02 -5.03467500e-01 -9.32531878e-02 2.98164159e-01
2.31105790e-01 -1.73734114e-01 2.97196787e-02 1.54584229e-01
7.64397442e-01 1.69788316e-01 6.24622881e-01 -5.56075752e-01
6.44642472e-01 -1.50315613e-01 1.00313199e+00 6.14500999e-01
-6.31445169e-01 1.10336816e+00 3.80708307e-01 -3.68798710e-02
-1.01166272e+00 -1.25224614e+00 -6.29844904e-01 7.86810279e-01
3.83955598e-01 -6.72069341e-02 -9.48936284e-01 -6.22436404e-01
-2.08454713e-01 3.98609310e-01 -4.15218860e-01 -1.37596220e-01
-5.18480241e-01 -8.23604107e-01 4.52196956e-01 4.37054992e-01
7.62172580e-01 -7.23577559e-01 -3.89165789e-01 1.85684949e-01
-6.98517799e-01 -1.11509144e+00 -8.10890973e-01 1.26844235e-02
-1.20412314e+00 -8.00462425e-01 -1.15382779e+00 -8.87429178e-01
6.81130648e-01 -1.30585819e-01 1.10461116e+00 -4.37895238e-01
-2.97769874e-01 3.57947558e-01 -1.70589209e-01 4.93037561e-03
-3.29719931e-01 1.20820120e-01 -9.52373371e-02 -2.00143587e-02
2.17480600e-01 -6.27181351e-01 -8.79710257e-01 6.26287580e-01
-7.49483168e-01 2.54548609e-01 5.08749545e-01 1.03526473e+00
1.05946410e+00 -3.46916318e-01 4.05392945e-01 -1.04338908e+00
3.44543338e-01 -6.43557549e-01 -4.02593195e-01 6.52409941e-02
-6.35133266e-01 1.87943295e-01 1.94372505e-01 -3.28831643e-01
-1.21479034e+00 3.06996137e-01 -6.26652911e-02 -4.49109614e-01
-2.28412673e-01 2.82318681e-01 -1.51567221e-01 2.29926303e-01
6.76102340e-01 3.98423165e-01 1.36866301e-01 -5.25453508e-01
1.61642700e-01 6.16991580e-01 5.62757432e-01 -6.82401359e-01
5.00495851e-01 7.01542974e-01 -1.00745372e-01 -6.93549633e-01
-8.68731081e-01 -9.66130733e-01 -1.04361594e+00 -9.06732976e-02
1.07519794e+00 -9.91327167e-01 -7.96059370e-02 5.45726240e-01
-9.02775884e-01 -3.65654022e-01 -3.34970325e-01 6.75532877e-01
-9.09500957e-01 7.47778714e-01 -5.78718245e-01 -3.08300406e-01
-2.97736019e-01 -1.31520593e+00 1.03522611e+00 2.17431933e-01
-2.88860053e-01 -1.22605145e+00 3.96476179e-01 5.59081495e-01
1.82883352e-01 4.47901756e-01 6.03707731e-01 -5.60864270e-01
-3.47441554e-01 2.12628528e-01 -3.66410375e-01 5.90113044e-01
2.43856743e-01 -5.00951886e-01 -7.10793436e-01 -3.39560717e-01
3.11699480e-01 5.64164482e-02 9.16942239e-01 9.98534262e-01
1.10292840e+00 6.11660928e-02 -4.92377371e-01 8.61045957e-01
1.24975109e+00 1.17753103e-01 4.36584890e-01 2.21104622e-01
5.42295873e-01 3.99934292e-01 7.38722861e-01 5.82317829e-01
4.60799932e-01 7.24820495e-01 -1.24104582e-01 -2.16746584e-01
-2.12716341e-01 -3.08167830e-04 2.40522181e-03 1.16665757e+00
2.26165690e-02 1.78899214e-01 -1.08112788e+00 7.87115276e-01
-1.99648118e+00 -6.93421483e-01 1.71707988e-01 2.12194204e+00
1.14639747e+00 -7.10857436e-02 9.78929549e-02 -2.90368676e-01
8.74538004e-01 1.28300413e-01 -6.83085740e-01 8.59557837e-02
1.22141100e-01 -2.97095925e-02 5.16012013e-01 5.76572597e-01
-1.44422567e+00 7.01898098e-01 5.50946474e+00 9.47365582e-01
-9.10718977e-01 4.94747430e-01 6.50335193e-01 1.57059133e-01
-7.96729401e-02 -1.60543062e-02 -4.83658850e-01 5.41226804e-01
6.43015444e-01 -2.41337895e-01 6.43082485e-02 8.48685503e-01
1.58067256e-01 -3.73697244e-02 -9.24944222e-01 1.18616414e+00
-5.64215183e-02 -1.18662417e+00 -3.05597156e-01 -2.48291157e-02
7.81306267e-01 1.14276692e-01 1.34982303e-01 -2.00003255e-02
-8.22408870e-02 -5.56968033e-01 3.00125360e-01 7.65798450e-01
9.15927529e-01 -3.54630172e-01 6.50106609e-01 3.24290305e-01
-1.18545198e+00 2.70068169e-01 -2.64412284e-01 3.90455425e-01
3.07886571e-01 8.24048996e-01 -6.55055165e-01 6.34161949e-01
6.73590481e-01 1.10815501e+00 -1.79449379e-01 9.24679697e-01
8.41433331e-02 4.54740733e-01 -2.43872702e-01 5.21042526e-01
-8.92402679e-02 -4.92312193e-01 8.56560707e-01 1.35132051e+00
1.74936593e-01 1.62423298e-01 2.79060751e-01 8.34236920e-01
6.97144419e-02 3.38674217e-01 -3.79587680e-01 2.51667529e-01
2.56848752e-01 1.06399596e+00 -8.73344004e-01 -2.95423925e-01
-2.96284348e-01 1.08158076e+00 7.15900958e-02 2.81309485e-01
-9.31682765e-01 -1.51125476e-01 5.79786658e-01 2.46158823e-01
1.57942772e-01 -1.34329975e-01 -5.97540855e-01 -1.43869853e+00
3.38169858e-02 -6.52519763e-01 6.49196148e-01 -4.46187496e-01
-1.40897787e+00 3.66651624e-01 3.22013885e-01 -1.54795730e+00
-1.02379389e-01 -2.27032200e-01 -1.60162091e-01 7.85910845e-01
-1.35627949e+00 -8.31194282e-01 -1.98800445e-01 5.89368522e-01
5.90476692e-01 -1.58117071e-01 5.63982427e-01 5.58023155e-01
-4.37325776e-01 5.56861401e-01 5.25697827e-01 2.23803416e-01
1.04036868e+00 -1.14389753e+00 -6.32873401e-02 6.32523537e-01
-2.57759631e-01 6.32949054e-01 6.67546809e-01 -6.05122507e-01
-6.37992203e-01 -9.35685039e-01 5.65245152e-01 -4.76505578e-01
4.66305733e-01 -1.04084879e-01 -1.07887805e+00 6.85219586e-01
-1.54348733e-02 1.37883440e-01 8.09602261e-01 -1.15398794e-01
-1.33132741e-01 -7.87150338e-02 -1.31842232e+00 2.88031965e-01
9.38871980e-01 -4.78847295e-01 -5.06179750e-01 3.56554866e-01
4.83380944e-01 -6.24132097e-01 -1.25975072e+00 5.69253564e-01
6.36932671e-01 -9.72613156e-01 9.97823298e-01 -1.37783036e-01
4.73011047e-01 -3.34153235e-01 -1.38975263e-01 -9.65416193e-01
-3.96761000e-01 -3.82585227e-01 1.48014426e-01 1.09359610e+00
3.34187180e-01 -6.65473342e-01 6.66915834e-01 5.47545135e-01
-4.39081565e-02 -6.75571918e-01 -1.15536344e+00 -6.70520365e-01
4.22882438e-01 -2.96866626e-01 9.41133052e-02 1.11287272e+00
4.48497608e-02 1.17081240e-01 -2.90200353e-01 2.05559254e-01
1.06998980e+00 1.94851056e-01 4.02191043e-01 -1.09539819e+00
-4.38114941e-01 -3.65728170e-01 -2.58684933e-01 -1.17500138e+00
4.53410931e-02 -1.03493452e+00 -5.96890738e-03 -1.44224489e+00
7.55800724e-01 -3.09031844e-01 -5.84925056e-01 2.36999854e-01
-7.33965859e-02 1.56804603e-02 1.17201909e-01 4.88976419e-01
-6.98343217e-01 3.87309313e-01 1.42836440e+00 -2.95168404e-02
-3.11456412e-01 6.69386163e-02 -6.08112395e-01 8.70116234e-01
9.36154962e-01 -8.30698788e-01 -4.55961019e-01 -2.65404165e-01
-2.38857210e-01 3.62296313e-01 2.50922889e-01 -9.77720857e-01
2.82920271e-01 -6.34124205e-02 2.05218181e-01 -5.98339975e-01
4.03444245e-02 -6.71096504e-01 2.18307853e-01 3.91677976e-01
-4.48754400e-01 -3.80468875e-01 5.35333753e-02 5.68416476e-01
-4.13047343e-01 -1.19573899e-01 1.40323293e+00 -1.72700644e-01
-5.75944483e-01 3.91921788e-01 -1.18224449e-01 4.49586838e-01
9.24068093e-01 -2.24826559e-01 1.69841915e-01 -6.57095164e-02
-1.17516351e+00 2.66379446e-01 3.98046553e-01 1.40865043e-01
5.11980832e-01 -1.23415923e+00 -9.42938089e-01 1.78691402e-01
1.23412661e-01 1.88585714e-01 7.26363242e-01 1.24210203e+00
-3.84083182e-01 2.72150606e-01 -3.28784287e-01 -9.64321852e-01
-1.25583661e+00 2.08353832e-01 5.89812994e-01 -6.35378718e-01
-6.88474059e-01 8.13913643e-01 7.65997887e-01 -6.02952778e-01
-2.10717559e-01 -2.01380804e-01 -3.35077033e-03 -1.65565312e-02
2.03762263e-01 4.60939437e-01 -2.29006484e-01 -7.58332372e-01
-5.69517851e-01 9.06336546e-01 -2.13629752e-01 -1.05114043e-01
1.16522694e+00 -4.74648297e-01 -1.88656598e-02 3.89513910e-01
1.40561497e+00 -4.15149033e-01 -1.47984362e+00 -5.76428771e-01
-1.21060275e-02 -3.80969822e-01 1.64785311e-01 -5.40169775e-01
-1.11507332e+00 8.12080920e-01 1.00086284e+00 -2.38915920e-01
1.36390090e+00 2.66232371e-01 7.77762055e-01 -2.17107818e-01
1.25179961e-01 -1.25846684e+00 8.56892318e-02 3.45915973e-01
8.11272979e-01 -1.50863278e+00 2.55075246e-01 -5.51727355e-01
-9.98089731e-01 7.63225675e-01 2.91717708e-01 -2.46644065e-01
1.14073586e+00 1.99563578e-01 8.91705677e-02 -1.08924903e-01
-3.30109596e-01 -7.07202032e-02 5.01388133e-01 5.85603654e-01
6.18493795e-01 2.63012350e-01 -6.30500555e-01 6.92639887e-01
1.35058388e-01 1.77217141e-01 7.58761689e-02 6.80592537e-01
-1.06376290e-01 -1.11287892e+00 -1.27757356e-01 2.80472934e-01
-5.27235806e-01 3.62089798e-02 8.84354860e-02 6.38280332e-01
9.65401828e-02 5.46198606e-01 -1.44188210e-01 -2.26841727e-03
2.11503237e-01 -6.80802017e-02 3.86998981e-01 -5.52180290e-01
-1.39284268e-01 5.56659460e-01 -4.00890946e-01 -5.71713924e-01
-6.52406752e-01 -1.14816272e+00 -1.38057411e+00 1.58449292e-01
-1.49026692e-01 -4.76140641e-02 3.36025089e-01 7.43582070e-01
3.84996951e-01 4.65542078e-01 6.12212062e-01 -6.45409167e-01
-4.43911254e-01 -8.17432821e-01 -8.08716655e-01 5.78555226e-01
3.19699287e-01 -7.12377191e-01 -2.22549558e-01 4.01083380e-01] | [14.515846252441406, -2.0306899547576904] |
3e5ad759-2565-43e9-8bed-02ca4a2eb975 | a-weakly-supervised-method-for-instance | 1908.09891 | null | https://arxiv.org/abs/1908.09891v1 | https://arxiv.org/pdf/1908.09891v1.pdf | A Weakly Supervised Method for Instance Segmentation of Biological Cells | We present a weakly supervised deep learning method to perform instance segmentation of cells present in microscopy images. Annotation of biomedical images in the lab can be scarce, incomplete, and inaccurate. This is of concern when supervised learning is used for image analysis as the discriminative power of a learning model might be compromised in these situations. To overcome the curse of poor labeling, our method focuses on three aspects to improve learning: i) we propose a loss function operating in three classes to facilitate separating adjacent cells and to drive the optimizer to properly classify underrepresented regions; ii) a contour-aware weight map model is introduced to strengthen contour detection while improving the network generalization capacity; and iii) we augment data by carefully modulating local intensities on edges shared by adjoining regions and to account for possibly weak signals on these edges. Generated probability maps are segmented using different methods, with the watershed based one generally offering the best solutions, specially in those regions where the prevalence of a single class is not clear. The combination of these contributions allows segmenting individual cells on challenging images. We demonstrate our methods in sparse and crowded cell images, showing improvements in the learning process for a fixed network architecture. | ['Pedro D. Marrero Fernandez', 'Fidel A. Guerrero-Peña', 'Tsang Ing Ren', 'Alexandre Cunha'] | 2019-08-26 | null | null | null | null | ['contour-detection'] | ['computer-vision'] | [ 5.54652214e-01 2.65148461e-01 1.11462928e-01 -2.93401152e-01
-4.94079858e-01 -5.60212314e-01 2.52350688e-01 7.70733654e-01
-9.61566508e-01 1.09256232e+00 -2.50495136e-01 -8.57302994e-02
-1.41941667e-01 -5.59225738e-01 -7.63451636e-01 -1.27744555e+00
-2.20854161e-03 6.93298340e-01 6.19741321e-01 4.11145501e-02
3.04051220e-01 9.86740947e-01 -1.28544104e+00 2.89535642e-01
7.27510452e-01 8.32649827e-01 4.71038699e-01 6.43656552e-01
-3.12367171e-01 5.97190142e-01 -5.02880931e-01 -7.53965601e-02
1.81260720e-01 -2.48740509e-01 -8.26155782e-01 1.13581836e-01
2.31178612e-01 1.07132271e-01 3.29645932e-01 1.07617414e+00
4.87429857e-01 -1.84536397e-01 8.89262855e-01 -7.33196914e-01
1.49240583e-01 4.48409230e-01 -6.15068018e-01 4.14742142e-01
-1.97316170e-01 1.83601633e-01 6.68677330e-01 -6.82310700e-01
8.01104307e-01 8.50908101e-01 6.78097308e-01 3.95655483e-01
-1.62379003e+00 -2.22451061e-01 2.68782198e-01 -9.91764665e-02
-1.38124180e+00 -3.58799100e-01 7.92336226e-01 -6.79193318e-01
5.31774580e-01 9.46073160e-02 6.08489931e-01 7.70799458e-01
1.70123458e-01 5.80368042e-01 1.18677306e+00 -3.53618175e-01
4.71805751e-01 3.64802808e-01 5.27818911e-02 5.48068702e-01
3.01342666e-01 -4.86824661e-01 -2.22176820e-01 -1.31584004e-01
7.91810989e-01 -8.79430026e-02 -3.89920980e-01 -5.75138271e-01
-1.12674296e+00 6.62855446e-01 3.96051437e-01 6.80733681e-01
-3.79444689e-01 -2.49341905e-01 3.35886836e-01 -3.92821850e-03
5.49902320e-01 6.44706130e-01 -4.89599049e-01 2.22770497e-01
-1.34720147e+00 1.05224818e-01 6.46864235e-01 1.84077054e-01
9.36499894e-01 -2.75262296e-01 -2.20568240e-01 9.48738992e-01
7.20725954e-02 -1.39031172e-01 2.85107136e-01 -9.53436017e-01
8.16606954e-02 8.13219547e-01 2.23857015e-02 -6.78147316e-01
-7.73945630e-01 -5.89306653e-01 -8.32254350e-01 6.07319355e-01
1.19275963e+00 -1.93405718e-01 -9.54751313e-01 1.52654493e+00
2.46430382e-01 -1.46332443e-01 -2.26079762e-01 9.59782004e-01
4.70457375e-01 3.32449406e-01 1.81108624e-01 -3.10654014e-01
1.13919830e+00 -5.31135380e-01 -6.86007500e-01 -2.10632354e-01
7.68952549e-01 -5.52959681e-01 9.02167797e-01 3.71224344e-01
-1.04903340e+00 -2.58966506e-01 -9.82376993e-01 2.12836694e-02
-6.96034551e-01 9.59206745e-02 3.78388196e-01 3.29986960e-01
-1.01357841e+00 8.33927572e-01 -8.29965711e-01 -3.29999775e-01
1.09866059e+00 6.81741476e-01 -3.61442089e-01 1.47321329e-01
-7.28182614e-01 9.09389496e-01 4.00102973e-01 2.03504950e-01
-4.60987210e-01 -6.15617692e-01 -6.77862227e-01 8.47574174e-02
2.10667819e-01 -4.83052641e-01 3.89626086e-01 -1.04868329e+00
-1.22770309e+00 1.22524405e+00 -1.09661631e-01 -4.78750616e-01
8.87905777e-01 3.69544834e-01 3.19909364e-01 4.39589560e-01
-1.63490072e-01 1.07120252e+00 5.26620567e-01 -1.52316093e+00
-3.56609613e-01 -5.83970070e-01 -1.12693116e-01 6.52798042e-02
-2.93998867e-01 -2.69969046e-01 -1.67978272e-01 -5.32254755e-01
1.16932154e-01 -7.28338599e-01 -4.81476486e-01 3.47348183e-01
-4.16370898e-01 7.59297237e-02 9.82364416e-01 -5.53423226e-01
7.33568907e-01 -1.95089507e+00 1.92335874e-01 3.77603650e-01
2.80399352e-01 3.20095628e-01 4.49602008e-02 8.36877674e-02
1.43539160e-01 4.24507214e-03 -6.98024333e-01 -3.66437435e-01
-4.06722695e-01 3.05666447e-01 1.63854942e-01 7.43501127e-01
5.45105219e-01 7.21727610e-01 -5.90621352e-01 -8.21186244e-01
2.21591055e-01 7.44266570e-01 -3.95897895e-01 1.31391048e-01
-1.31774232e-01 8.23155165e-01 -2.44740456e-01 5.65050960e-01
6.17597640e-01 -3.52336705e-01 2.04028144e-01 -1.53034806e-01
-6.48554415e-02 -1.19334579e-01 -1.24733675e+00 1.53628194e+00
-1.54443353e-01 5.61122060e-01 5.21685719e-01 -1.44908118e+00
9.05537903e-01 1.72724485e-01 6.31487489e-01 -2.31999606e-01
3.21306318e-01 2.72219539e-01 1.00587226e-01 -5.56875169e-01
3.23135890e-02 -3.01046759e-01 4.44047570e-01 2.02859908e-01
1.71246648e-01 -8.98371637e-02 3.72557253e-01 -3.46124321e-02
9.09939408e-01 2.41239563e-01 -2.99409553e-02 -7.51621306e-01
7.04687297e-01 -1.06122032e-01 6.37289941e-01 5.96580386e-01
-4.09662694e-01 9.27456796e-01 9.48337734e-01 -4.25521404e-01
-1.01224148e+00 -7.78198481e-01 -4.16538417e-01 9.39233780e-01
4.35799211e-02 4.17845070e-01 -9.24962401e-01 -7.73052216e-01
-1.74892887e-01 6.97105974e-02 -6.60995543e-01 1.78673252e-01
-6.23561800e-01 -1.24838459e+00 4.40777361e-01 2.85570085e-01
1.17728353e-01 -1.08756530e+00 -8.35573912e-01 2.51105070e-01
2.61753276e-02 -1.03431916e+00 2.15097796e-02 9.84342158e-01
-9.85047936e-01 -1.04951203e+00 -1.12350535e+00 -1.07157958e+00
1.16184485e+00 -1.48857549e-01 8.11723769e-01 3.02357346e-01
-3.42053860e-01 -1.05995037e-01 -1.01558566e-01 -3.09734046e-01
-2.44138047e-01 2.21333504e-01 -2.88967401e-01 6.31051883e-02
1.74751952e-01 -5.11778831e-01 -6.41107500e-01 1.34151518e-01
-1.01736903e+00 -2.77991384e-01 4.62729871e-01 1.10286129e+00
8.12352538e-01 6.33406192e-02 5.29572666e-01 -1.24623537e+00
1.74910277e-01 -2.12544516e-01 -6.42862082e-01 2.07227077e-02
-3.41884375e-01 1.03011206e-01 5.90957284e-01 -5.21537244e-01
-7.55977094e-01 4.17917371e-01 -3.10673386e-01 -2.58810213e-03
-5.33151209e-01 3.75613600e-01 -9.90577266e-02 -4.18415278e-01
5.71182847e-01 4.47693747e-04 3.99154782e-01 -3.17112654e-01
-1.80553645e-01 2.68225938e-01 3.71399671e-01 -4.65377867e-01
3.99049193e-01 9.34911430e-01 2.26640448e-01 -9.48040962e-01
-5.66302538e-01 -4.77028698e-01 -9.66166198e-01 -2.70992815e-01
9.21478570e-01 -4.40873802e-01 -5.91729403e-01 3.57231230e-01
-1.04862964e+00 -7.16138601e-01 -4.33657348e-01 4.30492312e-01
-4.39864814e-01 4.15137261e-01 -7.25621521e-01 -6.88600183e-01
3.12415492e-02 -1.49044728e+00 1.05442536e+00 3.88532162e-01
-2.12038726e-01 -1.30838358e+00 -3.33316512e-02 2.62848735e-01
4.20645982e-01 3.72689784e-01 1.10336185e+00 -8.58903348e-01
-3.05457145e-01 -5.41333109e-02 -2.09319651e-01 2.56936997e-01
3.92053984e-02 1.60981804e-01 -1.23228490e+00 -2.06113204e-01
-1.80266991e-01 -3.40735227e-01 1.07415855e+00 7.32568979e-01
1.34837437e+00 -5.88836940e-03 -5.04351914e-01 5.03905475e-01
1.28091908e+00 8.11116844e-02 5.36673248e-01 3.16964149e-01
5.66551268e-01 1.25218701e+00 1.82632372e-01 1.04932956e-01
-5.48138842e-02 5.80623448e-01 4.74314779e-01 -7.12192476e-01
-1.25473768e-01 3.77862334e-01 -9.35455412e-02 2.51377016e-01
-1.21356688e-01 -2.00686753e-01 -7.81942427e-01 7.04757750e-01
-1.74803364e+00 -6.07035935e-01 -2.79195577e-01 2.18771076e+00
9.07141387e-01 4.51948047e-01 2.49365479e-01 3.45431805e-01
6.51521325e-01 4.40474898e-02 -4.73788261e-01 -7.36853927e-02
-5.33926368e-01 1.90579504e-01 4.31834877e-01 7.01781750e-01
-1.08207166e+00 6.22938275e-01 5.96017218e+00 7.38728225e-01
-1.31664550e+00 -2.87834462e-02 1.15904355e+00 -4.87633422e-02
-1.63646385e-01 -7.19457865e-02 -6.52116358e-01 5.63042521e-01
3.57830256e-01 6.12854123e-01 1.29089981e-01 3.65197092e-01
3.62897813e-01 -3.95768702e-01 -8.60659719e-01 6.97957098e-01
-1.17741592e-01 -1.38198996e+00 -1.31517187e-01 2.91550606e-01
6.14863932e-01 1.25181815e-02 -6.70768991e-02 -2.68338233e-01
-3.39850001e-02 -1.12429309e+00 5.11633873e-01 4.94043469e-01
3.10443193e-01 -7.21340239e-01 1.13186634e+00 4.52548891e-01
-5.83646119e-01 1.67760760e-01 -3.84586424e-01 1.22821696e-01
1.42550826e-01 9.95975971e-01 -7.60990262e-01 -3.43982838e-02
5.78610539e-01 3.42991948e-01 -5.99050641e-01 1.13558888e+00
2.01789185e-01 4.42851454e-01 -4.89781469e-01 4.06333245e-02
3.06637585e-01 -4.10039902e-01 5.54490209e-01 1.52172232e+00
4.15734537e-02 -2.34176055e-01 1.22738905e-01 9.79204893e-01
7.68848062e-02 1.53669238e-01 -4.72592711e-01 1.45994097e-01
5.93288429e-02 1.63162708e+00 -1.62574089e+00 -1.36739194e-01
-1.86765008e-02 8.31918895e-01 5.30375183e-01 5.49063385e-01
-4.15480137e-01 -2.20554262e-01 2.89859682e-01 5.37000358e-01
4.14506346e-01 -1.11284025e-01 -6.31165445e-01 -7.89880753e-01
-9.07310247e-02 -5.13356209e-01 4.42995131e-01 -3.52361172e-01
-1.04430366e+00 4.97096956e-01 -3.14718395e-01 -7.79653549e-01
2.08115399e-01 -8.48304987e-01 -5.09332061e-01 7.13548899e-01
-1.71004081e+00 -9.43455219e-01 -1.37426764e-01 3.18125606e-01
2.56130189e-01 3.98444235e-02 5.40325046e-01 4.23330367e-01
-4.67303753e-01 2.45760456e-01 -1.48130357e-01 -4.76947241e-02
6.12150669e-01 -1.52431130e+00 -4.76455182e-01 7.79865265e-01
1.09426394e-01 4.92791772e-01 8.22005510e-01 -4.50649709e-01
-6.63870275e-01 -8.84410977e-01 9.28341508e-01 -2.16554016e-01
4.47730601e-01 -5.25321484e-01 -1.41756344e+00 2.52476484e-01
-4.56996337e-02 3.60367596e-01 7.23706484e-01 -1.25615299e-01
5.25224991e-02 -1.35947010e-02 -1.36836493e+00 5.17723620e-01
4.74425048e-01 -3.57312113e-01 -1.43562391e-01 3.94302815e-01
5.28121591e-02 -2.49613926e-01 -7.77220130e-01 4.03076947e-01
3.69387090e-01 -1.07583189e+00 7.46818423e-01 -2.79853642e-01
1.72028244e-01 -5.64534545e-01 2.86628693e-01 -1.12996495e+00
-1.36111900e-01 -2.95583487e-01 1.72414526e-01 1.13272595e+00
6.61894739e-01 -4.91369873e-01 1.22829902e+00 1.40438527e-01
-3.76480520e-02 -9.18860734e-01 -1.00964642e+00 -3.90444994e-01
2.90294498e-01 4.55292724e-02 1.72166690e-01 9.00961936e-01
-4.20662127e-02 1.15497380e-01 1.13396540e-01 -3.19912173e-02
5.81492126e-01 -1.20954551e-01 3.50881130e-01 -1.39495718e+00
-2.06198201e-01 -7.60608137e-01 -3.71706933e-01 -7.83660710e-01
1.05504245e-01 -8.91734660e-01 2.09181964e-01 -1.39948773e+00
2.36751318e-01 -6.28415883e-01 -3.80836040e-01 4.15426046e-01
-2.86455840e-01 6.90794945e-01 -1.29932433e-01 1.75786048e-01
-4.90740478e-01 6.47308826e-02 1.28419328e+00 -2.92599946e-01
-2.42792770e-01 -8.10393244e-02 -5.14646947e-01 9.02722597e-01
6.26215398e-01 -5.05603492e-01 -1.55750543e-01 -1.75019756e-01
1.31827161e-01 -3.00902903e-01 4.66463149e-01 -1.01118267e+00
3.49955112e-01 1.78296685e-01 6.63711369e-01 -3.82081717e-01
2.76233852e-01 -9.13945735e-01 -1.14549920e-01 5.47113597e-01
-4.81926262e-01 -4.79090124e-01 2.52556447e-02 2.89340228e-01
-2.15079397e-01 -4.13393199e-01 1.31370020e+00 -2.68159658e-01
-2.59775549e-01 6.93968758e-02 -5.03079474e-01 -1.81528464e-01
8.24638307e-01 -4.39890981e-01 -1.44415975e-01 -1.01309314e-01
-9.29610729e-01 2.19772518e-01 7.03951657e-01 -1.74108922e-01
1.68860719e-01 -7.89948404e-01 -4.83745456e-01 1.26827151e-01
1.03429398e-02 2.60456204e-01 1.49548918e-01 1.25897503e+00
-8.35190535e-01 9.33163613e-02 -3.24792325e-01 -7.91430116e-01
-1.16938603e+00 4.09292966e-01 5.86345673e-01 -5.59589267e-01
-5.19441426e-01 9.37845826e-01 2.77919203e-01 -2.51227677e-01
3.57704103e-01 -3.48695844e-01 -7.07213461e-01 2.65907407e-01
3.72315407e-01 3.07508200e-01 3.23157459e-01 -7.61999488e-01
-4.72230345e-01 6.22554839e-01 -2.70631313e-01 1.76704794e-01
1.40438569e+00 -1.17845975e-01 -1.50218025e-01 5.01639605e-01
1.09873867e+00 -9.48758051e-02 -1.80484211e+00 2.28906069e-02
1.11455999e-01 6.03416823e-02 2.20495582e-01 -7.05132902e-01
-1.14747560e+00 1.01437867e+00 5.88348389e-01 2.80098259e-01
8.97033334e-01 -4.87334654e-02 4.72588480e-01 9.56135057e-03
1.38123617e-01 -1.27315462e+00 -1.11791112e-01 2.69964397e-01
3.45757365e-01 -1.33473015e+00 1.43353967e-02 -4.07894045e-01
-4.52749133e-01 1.28966260e+00 2.16667116e-01 -1.00146301e-01
5.28951108e-01 6.56523347e-01 2.60462612e-01 -3.75758380e-01
-2.93650061e-01 -3.97703499e-01 6.73837364e-02 6.37483537e-01
4.38895106e-01 -4.12899345e-01 -4.20602769e-01 2.28952557e-01
2.80089885e-01 -1.98470861e-01 4.80505705e-01 9.27068174e-01
-5.34449816e-01 -9.29331481e-01 -2.91316003e-01 4.84343112e-01
-9.06916916e-01 1.68981880e-01 -4.37275112e-01 6.33258402e-01
4.27518815e-01 6.72614753e-01 2.41406724e-01 2.99798518e-01
3.52813117e-02 7.75136203e-02 5.43234468e-01 -5.14474452e-01
-6.12741590e-01 4.44611102e-01 -3.14706802e-01 -2.54861653e-01
-6.88192129e-01 -5.70106566e-01 -1.46145666e+00 2.50482321e-01
-2.07912520e-01 1.40655041e-01 5.88308334e-01 1.16324592e+00
-2.30678804e-02 5.55131733e-01 2.35610843e-01 -1.12827551e+00
4.15992215e-02 -6.70862854e-01 -8.79966974e-01 4.27291483e-01
4.94154990e-01 -6.96583450e-01 -4.06272113e-01 4.62520331e-01] | [14.598539352416992, -3.1119754314422607] |
d081d659-75c4-4222-b4dd-7161e0f2b961 | reenactnet-real-time-full-head-reenactment | 2006.105 | null | https://arxiv.org/abs/2006.10500v1 | https://arxiv.org/pdf/2006.10500v1.pdf | ReenactNet: Real-time Full Head Reenactment | Video-to-video synthesis is a challenging problem aiming at learning a translation function between a sequence of semantic maps and a photo-realistic video depicting the characteristics of a driving video. We propose a head-to-head system of our own implementation capable of fully transferring the human head 3D pose, facial expressions and eye gaze from a source to a target actor, while preserving the identity of the target actor. Our system produces high-fidelity, temporally-smooth and photo-realistic synthetic videos faithfully transferring the human time-varying head attributes from the source to the target actor. Our proposed implementation: 1) works in real time ($\sim 20$ fps), 2) runs on a commodity laptop with a webcam as the only input, 3) is interactive, allowing the participant to drive a target person, e.g. a celebrity, politician, etc, instantly by varying their expressions, head pose, and eye gaze, and visualising the synthesised video concurrently. | ['Anastasios Roussos', 'Mohammad Rami Koujan', 'Michail Christos Doukas', 'Stefanos Zafeiriou'] | 2020-05-22 | null | null | null | null | ['video-to-video-synthesis'] | ['computer-vision'] | [ 1.81827918e-01 3.44893813e-01 3.04626048e-01 -4.55992579e-01
-4.77103859e-01 -3.59582812e-01 6.41401947e-01 -3.60817313e-01
-2.71754116e-01 4.78820860e-01 1.75754055e-02 1.70863122e-01
5.26923716e-01 -2.78871596e-01 -9.36861753e-01 -5.18373072e-01
6.34534806e-02 3.29393774e-01 2.63778389e-01 -1.49983624e-02
9.49652344e-02 5.01794219e-01 -1.95214200e+00 2.57117767e-02
1.14347972e-01 1.05857229e+00 1.31660625e-01 1.05545878e+00
6.42559350e-01 1.02027047e+00 -4.08922732e-01 -3.31129193e-01
4.27584618e-01 -5.53994179e-01 -3.68864149e-01 6.18356645e-01
9.59338307e-01 -3.42755944e-01 -7.35880196e-01 9.74731982e-01
4.86842483e-01 2.79637337e-01 2.35369071e-01 -1.72585213e+00
-1.32733569e-01 -1.64778858e-01 -5.26324272e-01 -1.69820130e-01
9.52307820e-01 4.44076747e-01 2.96824008e-01 -7.01428652e-01
8.20826828e-01 1.37824273e+00 1.09848052e-01 7.82772779e-01
-9.48516607e-01 -7.17707276e-01 1.15528360e-01 3.35264862e-01
-1.56155539e+00 -8.41284156e-01 6.84855461e-01 -4.22842145e-01
2.31523722e-01 3.01146597e-01 8.98004293e-01 1.27635109e+00
9.58806723e-02 4.90213841e-01 8.57461989e-01 -3.11874360e-01
1.86891392e-01 3.21233302e-01 -5.30880213e-01 7.11998045e-01
-3.76127809e-01 1.65781796e-01 -9.42202628e-01 5.24559207e-02
7.40472734e-01 -6.13199845e-02 -6.30608320e-01 -4.24568683e-01
-1.38909888e+00 2.96278119e-01 3.99838090e-02 -1.77825019e-01
-4.76659775e-01 1.66764721e-01 9.19713899e-02 2.10903034e-01
1.78180009e-01 -1.67723984e-01 -8.83649662e-02 -2.03019232e-01
-9.95793402e-01 5.41391611e-01 5.09386003e-01 1.32761621e+00
7.01655269e-01 1.08903334e-01 -1.64739281e-01 -3.03591713e-02
1.26109391e-01 7.91154861e-01 2.11112127e-01 -1.47469771e+00
1.48311660e-01 2.28262216e-01 5.73127389e-01 -9.69329238e-01
-1.91118941e-01 2.96782792e-01 -3.31048012e-01 4.52791542e-01
3.16722095e-01 -3.65088254e-01 -4.71238792e-01 1.88553441e+00
8.13617826e-01 3.01865786e-01 -1.08811744e-01 1.25582540e+00
6.06551886e-01 6.39270604e-01 1.62384883e-01 -4.47922230e-01
1.51505160e+00 -7.55554914e-01 -8.29227209e-01 -4.37559783e-01
6.36965930e-02 -6.93579793e-01 1.00980628e+00 3.00004989e-01
-1.45437872e+00 -7.71252811e-01 -7.08076715e-01 -1.91678762e-01
-3.88733461e-03 2.00379118e-01 -2.40598302e-02 4.69118059e-01
-1.01971066e+00 3.26191813e-01 -8.80640864e-01 -4.39899415e-01
-9.46535692e-02 2.63417065e-01 -8.73256981e-01 -1.84071511e-01
-1.00136745e+00 8.49001527e-01 2.80624002e-01 -1.98047180e-02
-9.77884412e-01 -5.34318924e-01 -1.10413671e+00 -1.17044367e-01
3.23415309e-01 -7.70956635e-01 1.17545080e+00 -1.67570710e+00
-1.97280824e+00 1.12187231e+00 -4.39715385e-01 -1.26722474e-02
8.45055282e-01 -1.17250934e-01 -5.10106921e-01 3.00861388e-01
-4.18102182e-02 1.02712023e+00 1.18035364e+00 -1.10164511e+00
-7.74290740e-01 -6.07802510e-01 1.15842214e-02 5.59116840e-01
3.73125775e-03 3.46197128e-01 -7.87462950e-01 -2.71429420e-01
-3.33200246e-01 -1.12524819e+00 3.07418346e-01 5.02301991e-01
-3.51197064e-01 1.13852620e-01 1.18994224e+00 -5.73257506e-01
6.54030859e-01 -2.40941000e+00 3.85917038e-01 -4.33046445e-02
-5.20094708e-02 1.75890669e-01 5.26052117e-02 6.04980662e-02
-2.97108710e-01 -8.53846252e-01 4.65689413e-02 -7.16762364e-01
-1.83592096e-01 -1.01349361e-01 1.61353111e-01 9.48127627e-01
-1.46053821e-01 5.90165794e-01 -8.66381526e-01 -6.57250762e-01
4.91792619e-01 8.37499917e-01 -3.76335740e-01 7.21562803e-01
-4.81149405e-02 7.75793314e-01 4.71608415e-02 2.33094856e-01
5.25030911e-01 2.28727311e-01 1.36228306e-02 -7.74998739e-02
-1.64367199e-01 -2.60854989e-01 -1.29743087e+00 1.65662587e+00
-3.39381665e-01 1.06779015e+00 4.08811241e-01 -3.62967044e-01
9.63016391e-01 4.30028260e-01 2.08569631e-01 -5.77309608e-01
4.91987348e-01 -1.89276576e-01 -3.91825318e-01 -7.88047075e-01
3.40451092e-01 -8.62808228e-02 4.80150916e-02 4.57158744e-01
-2.18388885e-01 -4.21125349e-03 1.04671560e-01 -6.48139790e-02
7.28726149e-01 3.55973482e-01 6.03118166e-03 2.34133564e-02
6.57635510e-01 -3.35750967e-01 2.17155233e-01 -6.67483266e-03
-2.63771385e-01 7.51136839e-01 3.92694980e-01 -3.70965362e-01
-1.03745115e+00 -9.86639023e-01 5.29734373e-01 1.22282934e+00
2.10672215e-01 -3.19041386e-02 -1.10785532e+00 -1.05712652e-01
-2.42760882e-01 8.87512922e-01 -6.01453543e-01 -1.42805263e-01
-6.47404730e-01 3.39011550e-01 2.74999827e-01 2.43462831e-01
3.93271089e-01 -9.95001793e-01 -1.21110284e+00 -1.74948603e-01
-4.71661001e-01 -1.35368800e+00 -1.11424112e+00 -6.19346797e-01
-2.64142603e-01 -9.10808384e-01 -8.28738451e-01 -6.90706551e-01
9.27396178e-01 2.10892469e-01 6.30694926e-01 -4.56056505e-01
-1.74050719e-01 5.12202740e-01 9.75943804e-02 -3.15501899e-01
-3.56634587e-01 -5.63485026e-01 2.89819270e-01 8.41705739e-01
1.15118146e-01 -2.24623814e-01 -6.58954322e-01 4.04860556e-01
-7.19293535e-01 4.36200142e-01 -1.38024002e-01 7.29861036e-02
2.93216646e-01 -1.89836115e-01 -1.58669353e-01 -5.13748169e-01
2.15125144e-01 -1.09622024e-01 -6.42931879e-01 1.53713509e-01
1.36145786e-03 -4.20968711e-01 5.25400817e-01 -6.13282800e-01
-9.64233398e-01 6.10753894e-01 2.88229585e-01 -9.54356194e-01
-3.88801932e-01 -4.15706873e-01 -4.88873154e-01 3.01397610e-02
6.20646238e-01 1.15037419e-01 3.77504379e-01 1.16001368e-01
3.36319029e-01 8.19106996e-01 1.27283692e+00 -1.79083213e-01
7.09256530e-01 6.92202806e-01 -2.43767444e-02 -9.24881876e-01
-3.84519845e-01 -2.43859902e-01 -8.81707788e-01 -7.21754670e-01
8.23601127e-01 -1.13680804e+00 -1.17136621e+00 6.15500808e-01
-1.09041178e+00 -3.77359837e-01 -8.57175067e-02 4.21173543e-01
-9.40080047e-01 3.73750031e-02 7.64264464e-02 -5.61322391e-01
-1.21266060e-01 -9.72738445e-01 1.31692672e+00 4.13071185e-01
-3.00973207e-01 -7.80579031e-01 -1.55339479e-01 4.00766522e-01
1.90931782e-01 6.58259869e-01 3.82773757e-01 1.23174436e-01
-4.56691235e-01 -4.99728620e-01 -9.44121405e-02 1.81354657e-01
7.60705918e-02 3.60297143e-01 -9.56420004e-01 -3.13908279e-01
-1.65161490e-01 -8.00418779e-02 -1.71647966e-01 4.30976361e-01
7.03489602e-01 -3.76456916e-01 -2.44939700e-01 5.74810624e-01
8.82559896e-01 1.13332853e-01 7.35265315e-01 -9.14840326e-02
6.52007163e-01 9.32167292e-01 6.27403200e-01 4.30599719e-01
6.97208524e-01 1.07272291e+00 3.22295249e-01 -2.05685198e-01
-2.13455454e-01 -3.84922802e-01 7.37623811e-01 1.10496078e-02
-5.63859567e-02 -1.98499471e-01 -4.89856333e-01 4.39826846e-01
-1.72669315e+00 -8.60051572e-01 4.69984934e-02 2.57015347e+00
4.30433035e-01 -2.34742254e-01 5.56664765e-01 3.66306901e-02
9.87958014e-01 -1.08831957e-01 -5.42570889e-01 -1.52018368e-01
3.20338726e-01 -1.82610929e-01 4.35790956e-01 5.85334897e-01
-7.84645438e-01 8.86746764e-01 5.22087479e+00 5.29052503e-02
-1.68172932e+00 -8.46344531e-02 4.62561220e-01 -5.29370725e-01
9.69917849e-02 -1.65206671e-01 -4.66397315e-01 5.10754049e-01
1.11022067e+00 -5.07901311e-01 4.82685328e-01 7.05344737e-01
6.28938973e-01 -7.34819844e-02 -1.45407486e+00 1.28877401e+00
5.74745595e-01 -1.06309140e+00 -3.21996808e-01 -1.34163678e-01
2.89187014e-01 -3.49708378e-01 1.36713281e-01 2.47583855e-02
-2.00005949e-01 -9.39182580e-01 1.26144779e+00 5.57361722e-01
1.34831345e+00 -8.04509163e-01 1.61499903e-01 4.46562856e-01
-7.79161274e-01 1.52962804e-01 2.02743128e-01 -4.16276008e-02
4.24478889e-01 -3.37790847e-01 -5.36888659e-01 -3.02425399e-02
6.88847244e-01 3.40763718e-01 -4.28386271e-01 5.67290068e-01
-1.97278842e-01 -1.40380472e-01 -2.35487774e-01 2.80601084e-01
-2.93818332e-04 3.07931118e-02 4.83620107e-01 1.02025509e+00
2.15060130e-01 4.30316448e-01 1.96654350e-02 5.07629156e-01
1.30385794e-02 1.58059206e-02 -5.92832983e-01 3.30917150e-01
1.03625387e-01 1.05313659e+00 -3.55020285e-01 -2.13263780e-01
-1.46628276e-01 1.15335011e+00 -1.17341794e-01 3.40989262e-01
-1.00279045e+00 -5.45496345e-01 8.61011624e-01 7.46366203e-01
1.39545426e-01 -5.37070520e-02 4.75595325e-01 -9.40244198e-01
7.12778494e-02 -9.25081551e-01 5.09519801e-02 -1.38548565e+00
-3.37272525e-01 8.12362134e-01 1.68255836e-01 -1.18053198e+00
-5.60145557e-01 -2.34401569e-01 -5.39246261e-01 7.81646132e-01
-1.11715484e+00 -1.04959345e+00 -9.78007555e-01 1.14550078e+00
6.44934237e-01 -6.25078306e-02 5.90689480e-01 3.21490556e-01
-4.00234491e-01 5.09252727e-01 -4.92146105e-01 4.09789756e-02
9.22418535e-01 -7.24963427e-01 5.06927609e-01 6.86032534e-01
-1.84269711e-01 9.54073071e-02 1.13628674e+00 -2.52511531e-01
-1.33196640e+00 -1.02323461e+00 9.64781106e-01 -4.49567020e-01
2.55535901e-01 -5.60154617e-01 -6.10127091e-01 9.44423199e-01
3.07111710e-01 3.90040249e-01 1.70328826e-01 -8.22249234e-01
-3.01937051e-02 -3.62841129e-01 -1.38447022e+00 7.78590083e-01
7.05909550e-01 -5.47734082e-01 -2.09641054e-01 3.40445518e-01
2.83136368e-01 -9.78386462e-01 -3.88647377e-01 -1.84966594e-01
8.16074848e-01 -1.11025941e+00 7.46160865e-01 -4.72847492e-01
2.19880521e-01 -6.26322150e-01 5.69389984e-02 -1.09744430e+00
3.57469916e-02 -1.10103929e+00 1.88624233e-01 6.63394511e-01
-1.67443916e-01 -2.60554999e-01 7.08784699e-01 1.08365870e+00
2.96708584e-01 -2.25705922e-01 -9.62804735e-01 -2.69906402e-01
-5.54170310e-01 -2.71267295e-01 4.60416287e-01 4.48317826e-01
9.39567760e-02 1.59707636e-01 -7.93741822e-01 3.77216756e-01
6.90627813e-01 -2.50221729e-01 1.30311155e+00 -8.48845541e-01
-7.48105049e-02 3.13651934e-02 -6.78840518e-01 -8.08803976e-01
3.28526556e-01 -2.34526962e-01 -1.51009988e-02 -8.68462682e-01
-1.65538386e-01 1.95603564e-01 4.12786424e-01 2.04012245e-01
8.57240334e-02 3.85311484e-01 3.44327569e-01 -1.68114901e-02
-4.00909692e-01 3.14120322e-01 1.20342660e+00 2.25579903e-01
-2.18886826e-02 3.29542816e-01 -3.00670564e-01 7.98377931e-01
3.31753880e-01 -3.70542794e-01 -5.19214630e-01 -3.28204334e-01
-1.40180081e-01 7.78295815e-01 5.98292470e-01 -1.02415311e+00
4.61664081e-01 -3.59978497e-01 3.71398181e-01 -1.05085902e-01
7.40043700e-01 -9.39269960e-01 6.55699074e-01 3.52436632e-01
-3.38883191e-01 2.20397621e-01 1.98514059e-01 3.62176359e-01
-6.07024804e-02 1.81571275e-01 1.10637093e+00 -1.83055788e-01
-5.50883651e-01 4.35910046e-01 -1.36968940e-01 6.83508022e-03
1.51137114e+00 -2.70236522e-01 -1.47960801e-02 -9.30963874e-01
-8.02843392e-01 5.67301326e-02 8.63305211e-01 6.42259777e-01
6.35062456e-01 -1.12793946e+00 -9.19950604e-01 7.37936199e-01
1.76065117e-01 -1.24125049e-01 5.34362793e-01 5.66616654e-01
-8.23541880e-01 2.31550619e-01 -4.42167342e-01 -7.46486783e-01
-1.74562275e+00 4.90514100e-01 4.65186447e-01 6.92460716e-01
-7.66506374e-01 7.43918002e-01 3.28426838e-01 3.59648839e-02
4.43900794e-01 2.25835696e-01 -1.17770629e-02 -1.28925338e-01
8.92224550e-01 5.93544781e-01 -1.33678749e-01 -1.29597199e+00
-3.32940608e-01 6.90786064e-01 2.88915157e-01 -3.83898228e-01
8.33199084e-01 -4.62295890e-01 2.19003364e-01 3.20073307e-01
1.37337303e+00 -1.12436362e-01 -1.82622194e+00 -1.60096288e-01
-6.19046330e-01 -8.67732584e-01 -2.40512341e-02 -2.48686492e-01
-1.25725913e+00 7.24751115e-01 7.15762854e-01 -4.22680736e-01
1.06527662e+00 -3.27435322e-02 6.92501962e-01 -1.54173083e-03
5.00475824e-01 -9.59463179e-01 1.47418883e-02 -1.05345687e-02
8.18177581e-01 -1.07524300e+00 -1.07120678e-01 -4.04221267e-02
-1.17636383e+00 9.95825052e-01 5.60680628e-01 -2.25537177e-02
4.11917180e-01 1.74662665e-01 3.91063780e-01 -8.67067650e-02
-9.63524520e-01 3.20519577e-03 1.34715110e-01 7.95199931e-01
9.90079064e-03 1.30333528e-01 4.94451791e-01 -3.42312485e-01
-3.84748280e-01 1.73882455e-01 7.45362639e-01 5.56489468e-01
-1.30724773e-01 -5.32721460e-01 -5.39782405e-01 -4.84792352e-01
-1.92446977e-01 4.76772279e-01 -1.66856602e-01 7.26946294e-01
2.12296754e-01 7.94657171e-01 4.91148710e-01 -3.56623195e-02
7.43817806e-01 5.62259778e-02 5.40771186e-01 -4.07559454e-01
-3.00471634e-01 -1.12914648e-02 -3.44360411e-01 -8.16283286e-01
-3.19466203e-01 -9.49558914e-01 -1.06788313e+00 -6.34393156e-01
3.89123261e-01 -1.46508530e-01 8.17644715e-01 6.86082661e-01
4.30550069e-01 -1.84748811e-03 9.00382519e-01 -1.39745796e+00
9.55747068e-02 -7.74204016e-01 -4.69508290e-01 5.46752751e-01
6.65128410e-01 -5.03289878e-01 -3.89830649e-01 6.00755334e-01] | [13.045905113220215, -0.4107978940010071] |
33b58ba4-d6fe-432c-b40e-ac71d0c83be4 | multi-level-multimodal-common-semantic-space | 1811.11683 | null | https://arxiv.org/abs/1811.11683v2 | https://arxiv.org/pdf/1811.11683v2.pdf | Multi-level Multimodal Common Semantic Space for Image-Phrase Grounding | We address the problem of phrase grounding by lear ing a multi-level common semantic space shared by the textual and visual modalities. We exploit multiple levels of feature maps of a Deep Convolutional Neural Network, as well as contextualized word and sentence embeddings extracted from a character-based language model. Following dedicated non-linear mappings for visual features at each level, word, and sentence embeddings, we obtain multiple instantiations of our common semantic space in which comparisons between any target text and the visual content is performed with cosine similarity. We guide the model by a multi-level multimodal attention mechanism which outputs attended visual features at each level. The best level is chosen to be compared with text content for maximizing the pertinence scores of image-sentence pairs of the ground truth. Experiments conducted on three publicly available datasets show significant performance gains (20%-60% relative) over the state-of-the-art in phrase localization and set a new performance record on those datasets. We provide a detailed ablation study to show the contribution of each element of our approach and release our code on GitHub. | ['Shih-Fu Chang', 'Brian Chen', 'Svebor Karaman', 'Hassan Akbari', 'Carl Vondrick', 'Surabhi Bhargava'] | 2018-11-28 | multi-level-multimodal-common-semantic-space-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Akbari_Multi-Level_Multimodal_Common_Semantic_Space_for_Image-Phrase_Grounding_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Akbari_Multi-Level_Multimodal_Common_Semantic_Space_for_Image-Phrase_Grounding_CVPR_2019_paper.pdf | cvpr-2019-6 | ['phrase-grounding'] | ['natural-language-processing'] | [ 2.46072456e-01 -9.33331028e-02 -2.04863116e-01 -1.77840486e-01
-1.17286682e+00 -7.44667053e-01 7.00361609e-01 4.56587911e-01
-6.50833964e-01 3.38734031e-01 4.60206330e-01 -2.78889909e-02
1.35477796e-01 -5.78161001e-01 -8.65859032e-01 -5.18273532e-01
7.51603544e-02 1.39888972e-01 4.39415090e-02 -1.54807046e-01
5.70518553e-01 2.81755954e-01 -1.32335126e+00 8.90053093e-01
3.02816600e-01 1.12264872e+00 4.59822327e-01 9.19631541e-01
-4.08345163e-01 3.68394941e-01 -5.27507126e-01 -5.05263448e-01
2.63553485e-02 -4.65540230e-01 -9.85391915e-01 1.59551486e-01
8.73301983e-01 -6.44859895e-02 -4.26916152e-01 1.25510156e+00
3.92076850e-01 1.38857290e-01 6.87922418e-01 -1.13329661e+00
-1.39358103e+00 4.02308434e-01 -6.60785079e-01 4.01632249e-01
6.53021634e-01 4.44186069e-02 1.63777447e+00 -1.37565339e+00
6.67567611e-01 1.33472776e+00 5.14083147e-01 3.12161535e-01
-1.41203725e+00 -2.70012617e-01 1.50391877e-01 2.41007447e-01
-1.57870257e+00 -2.64078617e-01 3.88033062e-01 -3.59463811e-01
1.26309311e+00 1.20661005e-01 5.00114322e-01 1.07101083e+00
4.48541135e-01 8.30639243e-01 1.02189100e+00 -6.60934150e-01
-3.17697562e-02 3.10289204e-01 -6.67627752e-02 1.05563211e+00
-2.19144791e-01 -3.81846607e-01 -1.08811402e+00 -2.75858361e-02
5.93295515e-01 -1.84240043e-01 -2.74049342e-01 -4.85187799e-01
-1.32563305e+00 9.16615009e-01 7.51314700e-01 4.42623794e-01
-6.97330758e-02 4.04926926e-01 2.95697093e-01 8.56695697e-02
3.64610910e-01 4.95657563e-01 -2.36772835e-01 -3.54479700e-02
-9.34772968e-01 1.54648691e-01 5.24452865e-01 1.00794041e+00
7.92149782e-01 -4.99699920e-01 -6.21627927e-01 8.07740629e-01
4.93319482e-01 5.73683560e-01 3.64429027e-01 -6.30053520e-01
7.01441109e-01 4.34579879e-01 -7.27714375e-02 -1.36237419e+00
-3.29949021e-01 -4.14168179e-01 -2.53936946e-01 -1.93694711e-01
2.16543302e-01 2.39269018e-01 -7.77131379e-01 1.74134660e+00
-1.21576987e-01 1.86939687e-01 -1.64378993e-02 9.69780505e-01
1.08963740e+00 7.41787970e-01 1.90902561e-01 2.29569793e-01
1.80687308e+00 -1.03867149e+00 -4.04079914e-01 -3.94013464e-01
5.19705415e-01 -7.66261578e-01 1.24617338e+00 -1.11885197e-01
-1.13179326e+00 -5.76787055e-01 -1.12596393e+00 -3.62999171e-01
-7.47582614e-01 6.58375099e-02 5.77299967e-02 2.64887929e-01
-1.25478637e+00 3.87485802e-01 -6.20514870e-01 -8.18019390e-01
4.05339181e-01 1.93417087e-01 -6.10206485e-01 2.53635109e-03
-1.16953373e+00 9.80292737e-01 2.36946762e-01 1.37065863e-03
-6.78977847e-01 -6.45699978e-01 -1.14229453e+00 2.42975444e-01
4.79881540e-02 -8.74969482e-01 1.01064575e+00 -9.00129199e-01
-8.34779024e-01 1.34407043e+00 -2.47830063e-01 -2.51856506e-01
-1.42379981e-02 -2.73623019e-01 -1.72786906e-01 5.91477573e-01
5.23941040e-01 1.14701486e+00 9.86421645e-01 -1.47684240e+00
-6.74419224e-01 -1.69599742e-01 2.58399665e-01 5.85253358e-01
-6.27873003e-01 2.34568030e-01 -7.81811476e-01 -5.65110147e-01
7.44712725e-02 -8.63421977e-01 1.78171560e-01 3.33526045e-01
-4.64221776e-01 -1.33656904e-01 5.82043469e-01 -7.08236992e-01
1.02053738e+00 -2.26963425e+00 4.78070945e-01 4.34316732e-02
2.80951858e-01 -3.70667845e-01 -5.65946698e-01 6.69501007e-01
-1.04476526e-01 2.90913463e-01 -1.33113399e-01 -6.61443174e-01
2.91916996e-01 3.22819799e-02 -3.11334133e-01 5.11654615e-01
4.69051927e-01 1.17966259e+00 -1.01125598e+00 -5.68731546e-01
1.12776496e-01 4.95355695e-01 -3.51491958e-01 1.76460087e-01
-1.07213832e-01 1.20198764e-01 -2.08115250e-01 4.54695940e-01
2.85061151e-01 -4.90761995e-01 -3.71442325e-02 -4.49664414e-01
1.80998132e-01 2.15436399e-01 -7.31606662e-01 2.15294456e+00
-5.88267505e-01 9.30881023e-01 -1.53137282e-01 -7.22917438e-01
7.38469601e-01 3.29746395e-01 1.16477773e-01 -7.57818699e-01
1.07112557e-01 -9.62213427e-02 -2.51787752e-01 -5.88374734e-01
8.22291434e-01 1.10284239e-02 -3.52329671e-01 4.32793438e-01
5.41528404e-01 -2.26325765e-02 4.14806493e-02 5.25363386e-01
8.71895969e-01 1.37070552e-01 1.43425196e-01 -3.31749320e-01
5.08120120e-01 -3.77211757e-02 -2.10063055e-01 9.40507770e-01
-1.17140733e-01 9.39624012e-01 7.11196482e-01 -2.05925733e-01
-1.09157658e+00 -1.17413104e+00 -8.58718976e-02 1.49213445e+00
2.63694942e-01 -7.55649507e-01 -6.33968294e-01 -6.03405833e-01
3.82921584e-02 5.27767599e-01 -1.03224456e+00 -7.83073753e-02
-2.24621460e-01 -4.51325864e-01 4.36873108e-01 7.21883714e-01
3.68399948e-01 -1.18304551e+00 -5.47744215e-01 -1.34330317e-01
-2.27776855e-01 -1.44450057e+00 -6.42575204e-01 1.34559482e-01
-3.32188219e-01 -8.43494892e-01 -6.23769701e-01 -1.13623106e+00
5.81406236e-01 3.00592393e-01 1.33440673e+00 2.30087936e-01
-4.70336139e-01 8.11835706e-01 -3.96421492e-01 -1.49193192e-02
-1.08504348e-01 3.36500183e-02 -1.04428567e-01 1.60067305e-01
5.17486751e-01 -5.50698638e-02 -7.09520400e-01 -9.27477404e-02
-8.54898691e-01 9.87649411e-02 4.22868013e-01 8.66642177e-01
4.85484183e-01 -4.60944444e-01 -4.26947838e-03 -2.12556481e-01
7.84535646e-01 -5.02476692e-01 -2.02419281e-01 3.57944310e-01
-2.14675203e-01 9.74326581e-02 1.57322109e-01 -1.59696028e-01
-5.00636280e-01 -1.40657440e-01 -1.21260099e-02 -4.62091178e-01
-1.64030746e-01 6.33063734e-01 1.80285171e-01 3.46160419e-02
4.63047594e-01 3.57073426e-01 -2.20032990e-01 -2.30005369e-01
9.02167737e-01 5.52253008e-01 6.64005339e-01 -6.36510372e-01
6.69883192e-01 4.04489815e-01 -2.11545274e-01 -8.10697317e-01
-1.03134477e+00 -5.01755774e-01 -8.48051667e-01 -6.08444288e-02
1.42141056e+00 -8.41756165e-01 -6.34010434e-01 -5.72139509e-02
-1.42103505e+00 1.32377610e-01 -4.91325222e-02 1.91084996e-01
-5.63096404e-01 3.69102418e-01 -5.39000750e-01 -4.91610020e-01
-3.32963556e-01 -1.18218029e+00 1.51453066e+00 1.19773813e-01
-3.03907603e-01 -1.12323868e+00 1.28296837e-01 2.74367511e-01
1.98827222e-01 -3.73360664e-02 1.09939909e+00 -7.03511953e-01
-4.91383582e-01 -1.50795141e-02 -7.32841372e-01 8.75357911e-02
-1.60439834e-01 -2.07552165e-01 -1.04265571e+00 -4.85815227e-01
-3.38384658e-01 -5.27313769e-01 1.12470639e+00 2.20302090e-01
9.43947732e-01 -3.15861702e-02 -3.36024225e-01 4.58482504e-01
1.65369821e+00 -4.18927699e-01 3.45106214e-01 6.64858043e-01
8.15442145e-01 5.58237374e-01 3.06172669e-01 1.78661585e-01
4.28983182e-01 7.57664919e-01 6.50220990e-01 -2.58212686e-01
-2.37875313e-01 -2.49522164e-01 3.10369402e-01 5.05923331e-01
4.22608823e-01 -3.42322260e-01 -9.87557888e-01 8.66566658e-01
-1.84467137e+00 -9.97054040e-01 6.53361827e-02 1.97957122e+00
6.70386791e-01 1.44972816e-01 1.78701226e-02 -3.53556037e-01
6.86505914e-01 2.51795709e-01 -6.66751787e-02 -6.83823347e-01
-1.32701382e-01 2.75501221e-01 2.59170890e-01 5.43066502e-01
-1.34992647e+00 9.36787248e-01 6.61199188e+00 6.60752237e-01
-8.43468666e-01 2.43544474e-01 2.93756068e-01 -1.89694285e-01
-4.31423128e-01 -2.93197781e-01 -6.82995439e-01 2.27028862e-01
7.62640357e-01 3.11063118e-02 3.10728252e-01 4.05284047e-01
-3.52373756e-02 3.04012396e-03 -1.36462069e+00 1.15677655e+00
6.27299249e-01 -1.49590445e+00 1.15241438e-01 -8.05328563e-02
7.37373054e-01 2.05309153e-01 3.88362676e-01 1.50999814e-01
-9.69725009e-03 -1.26689720e+00 8.15548539e-01 5.27456522e-01
8.18926454e-01 -6.90335929e-01 6.13642752e-01 -2.02198565e-01
-1.38787591e+00 8.07387456e-02 -3.15435827e-01 2.40239844e-01
-3.36329043e-02 -1.33654699e-01 -5.69109857e-01 5.28006732e-01
8.75559986e-01 7.70219028e-01 -7.83385396e-01 7.49432802e-01
-3.27480555e-01 8.76012966e-02 -1.65356882e-02 -1.84183747e-01
6.96783781e-01 2.26642534e-01 4.92207676e-01 1.62710667e+00
9.50842872e-02 -3.62165689e-01 2.41790622e-01 1.06136715e+00
-2.72455782e-01 2.83421963e-01 -7.09818065e-01 -6.60580099e-02
3.00183386e-01 1.53845489e+00 -6.90432608e-01 -2.27017954e-01
-8.66726935e-01 1.32674944e+00 6.34949207e-01 4.51464415e-01
-7.08532572e-01 -4.78628069e-01 6.28246844e-01 -3.04852515e-01
5.73466241e-01 -3.16281676e-01 -2.40462810e-01 -1.08922756e+00
-6.85528219e-02 -5.48167408e-01 4.33512747e-01 -1.12227368e+00
-1.47041655e+00 6.82209373e-01 -1.51776299e-01 -1.12083936e+00
2.79739946e-02 -8.75226259e-01 -3.80747676e-01 1.18658412e+00
-1.31650138e+00 -1.37691534e+00 -1.09606273e-01 5.75261831e-01
6.45369053e-01 -2.32193589e-01 1.13676310e+00 -1.52113944e-01
-1.68323174e-01 6.83311880e-01 4.17680405e-02 3.62869829e-01
6.41377628e-01 -1.38524055e+00 4.87653077e-01 6.87281370e-01
7.15221226e-01 7.00297058e-01 6.62701964e-01 -2.36278012e-01
-1.42784619e+00 -7.73195982e-01 1.07128882e+00 -8.99924338e-01
1.21863592e+00 -7.53766656e-01 -8.15470278e-01 6.26704991e-01
7.24528909e-01 -3.13930068e-04 8.20810556e-01 1.66966423e-01
-6.33654118e-01 2.49816999e-01 -8.37940156e-01 7.22337186e-01
7.33910859e-01 -1.11540461e+00 -8.94713879e-01 3.92358691e-01
7.64446557e-01 -3.88988346e-01 -8.55840802e-01 -1.84297413e-01
6.50939465e-01 -7.62842655e-01 1.05942929e+00 -8.20338428e-01
9.62195337e-01 -1.95318118e-01 -5.49530685e-01 -1.06462860e+00
-4.31775749e-01 -2.97483578e-02 9.93858501e-02 1.01138330e+00
6.27776325e-01 -1.20211922e-01 6.50990546e-01 2.04394579e-01
7.82891810e-02 -9.28270876e-01 -9.71211314e-01 -4.82484579e-01
1.53177783e-01 -5.02872109e-01 9.16096568e-02 7.88925588e-01
2.23288178e-01 7.19887197e-01 -9.19088721e-02 3.85660648e-01
4.68637466e-01 1.93235353e-01 3.95033240e-01 -6.40333533e-01
-2.80469239e-01 -4.87114131e-01 -6.73831165e-01 -1.13591158e+00
3.55027974e-01 -1.27247036e+00 2.00978085e-01 -1.89261997e+00
7.86488771e-01 1.81122139e-01 -6.80132627e-01 5.53499997e-01
-1.43364996e-01 8.32005262e-01 5.62647283e-01 1.20420299e-01
-8.85904610e-01 5.22344351e-01 1.12900567e+00 -4.50308770e-01
1.64322808e-01 -8.15467715e-01 -7.54575253e-01 4.68369335e-01
4.10784900e-01 -4.87347066e-01 -1.98739931e-01 -9.07135963e-01
4.11618084e-01 -4.84771132e-02 7.24281907e-01 -6.99829400e-01
1.82344228e-01 1.09978765e-01 6.08178020e-01 -6.26845539e-01
6.22058451e-01 -7.43605852e-01 -6.22092247e-01 1.64959446e-01
-7.19547212e-01 2.91392505e-01 3.63554239e-01 6.26885295e-01
-1.62538290e-01 -3.16657364e-01 5.37926316e-01 -1.29503623e-01
-8.49961162e-01 8.81558582e-02 -1.98124439e-01 -1.77979190e-02
7.80169547e-01 -2.05220327e-01 -4.47599500e-01 -4.23693180e-01
-9.56098378e-01 2.00164139e-01 4.04718250e-01 8.48491311e-01
8.21679950e-01 -1.42880177e+00 -7.59840310e-01 7.35218376e-02
5.88196635e-01 -7.42073953e-01 1.12233050e-02 7.74333000e-01
-2.81221181e-01 6.99945629e-01 -2.10053787e-01 -9.02381897e-01
-1.22371304e+00 6.36750698e-01 3.50545168e-01 -4.29361174e-03
-5.48119903e-01 1.18147635e+00 4.18863356e-01 -1.74443021e-01
4.46281470e-02 -1.18412122e-01 -2.32314616e-01 7.67665878e-02
5.71314096e-01 -2.30921701e-01 1.11350521e-01 -1.06776273e+00
-7.85832703e-01 9.17704642e-01 -8.12674314e-02 -4.68667388e-01
1.05612087e+00 -3.87630731e-01 -2.58469999e-01 5.70168316e-01
1.64000082e+00 -1.00397579e-01 -1.03057611e+00 -4.17559892e-01
-1.33258015e-01 -4.58863914e-01 5.45489565e-02 -7.21719146e-01
-7.90157378e-01 1.01277089e+00 6.94453835e-01 2.89779842e-01
8.67140472e-01 5.33778667e-01 3.97971541e-01 3.07032079e-01
-7.85017908e-02 -8.00215483e-01 5.06906867e-01 5.66678584e-01
1.07789731e+00 -1.32681382e+00 -3.02498098e-02 6.00053603e-03
-6.59771323e-01 1.18439221e+00 4.79357481e-01 -3.65159690e-01
3.42555106e-01 -3.11202854e-02 1.14580467e-02 -4.76207048e-01
-7.24600375e-01 -3.24449450e-01 6.84233367e-01 4.74560261e-01
4.91851628e-01 -2.24005446e-01 -1.59303233e-01 4.20405090e-01
-1.00262882e-02 -6.72203183e-01 1.35934755e-01 9.40201223e-01
-5.85629821e-01 -8.80221844e-01 -4.19539362e-01 2.07201406e-01
-4.50751007e-01 -5.23102701e-01 -3.57088864e-01 7.36211896e-01
-1.38126031e-01 9.60998058e-01 4.54568565e-01 -2.41515979e-01
2.52178133e-01 8.91850144e-02 7.14503050e-01 -9.18766797e-01
-4.92092967e-01 7.38427043e-02 4.30495851e-02 -6.31203711e-01
-3.51016998e-01 -5.21773517e-01 -1.20665288e+00 -8.27301890e-02
-7.74100423e-02 -1.12163447e-01 8.08484316e-01 1.18666267e+00
2.64840156e-01 5.12997389e-01 3.96468759e-01 -9.00663197e-01
-1.80516660e-01 -7.87733197e-01 -4.07642007e-01 5.56940198e-01
3.79225910e-01 -6.30069673e-01 -9.74216610e-02 1.40492901e-01] | [10.58759593963623, 1.6189011335372925] |
11306c66-27be-46eb-9d4c-f80417cf6173 | a-generalised-linear-model-framework-for | 2006.06267 | null | https://arxiv.org/abs/2006.06267v3 | https://arxiv.org/pdf/2006.06267v3.pdf | A Generalised Linear Model Framework for $β$-Variational Autoencoders based on Exponential Dispersion Families | Although variational autoencoders (VAE) are successfully used to obtain meaningful low-dimensional representations for high-dimensional data, the characterization of critical points of the loss function for general observation models is not fully understood. We introduce a theoretical framework that is based on a connection between $\beta$-VAE and generalized linear models (GLM). The equality between the activation function of a $\beta$-VAE and the inverse of the link function of a GLM enables us to provide a systematic generalization of the loss analysis for $\beta$-VAE based on the assumption that the observation model distribution belongs to an exponential dispersion family (EDF). As a result, we can initialize $\beta$-VAE nets by maximum likelihood estimates (MLE) that enhance the training performance on both synthetic and real world data sets. As a further consequence, we analytically describe the auto-pruning property inherent in the $\beta$-VAE objective and reason for posterior collapse. | ['Stefanie Schwaar', 'Robert Sicks', 'Ralf Korn'] | 2020-06-11 | null | null | null | null | ['systematic-generalization'] | ['reasoning'] | [-2.47534305e-01 2.31202245e-01 3.26772854e-02 -3.89597923e-01
-2.66154647e-01 -8.01784247e-02 3.98469716e-01 -2.10108561e-03
-4.38987613e-01 6.30969763e-01 -2.47050717e-01 -3.86531413e-01
-6.32202208e-01 -9.20242906e-01 -8.32716286e-01 -8.21780443e-01
-2.83121794e-01 4.55682427e-01 -1.22186013e-01 -1.50463998e-01
-1.70080021e-01 3.26026171e-01 -1.50539553e+00 -2.59511054e-01
8.80108833e-01 1.06630075e+00 2.70792335e-01 3.72344971e-01
-4.03339416e-02 1.14796795e-01 -4.73555475e-01 -5.41438401e-01
2.13739529e-01 -5.26700497e-01 -2.01787323e-01 -1.40477926e-01
1.09933242e-01 -1.16815515e-01 -2.03987524e-01 1.22242725e+00
1.81619957e-01 3.65327924e-01 1.37406445e+00 -1.26583052e+00
-5.53695321e-01 3.62460822e-01 -3.06576520e-01 3.03515106e-01
-2.51180023e-01 -2.30513707e-01 1.13993907e+00 -9.57294345e-01
4.95995671e-01 1.12516701e+00 7.39489794e-01 3.90872836e-01
-1.46206903e+00 -5.86439848e-01 4.30080965e-02 6.98765069e-02
-1.61005127e+00 -8.36294442e-02 8.37785125e-01 -4.97983426e-01
8.65724206e-01 -1.88889161e-01 6.28407300e-01 1.34718633e+00
4.90649253e-01 6.93077743e-01 6.35441601e-01 -5.39376318e-01
5.02599299e-01 4.88809675e-01 2.87553012e-01 6.37222290e-01
4.12202895e-01 3.58068585e-01 -3.96228015e-01 -1.02819890e-01
8.56478930e-01 -3.35529357e-01 -2.57888168e-01 -8.53243828e-01
-3.63404185e-01 1.21838796e+00 3.60384792e-01 3.24215382e-01
-3.77552629e-01 1.44564852e-01 1.85981676e-01 2.96386778e-01
3.43401372e-01 4.45991009e-01 -1.73852921e-01 1.25671089e-01
-9.14594352e-01 1.96385399e-01 4.98473972e-01 6.59527957e-01
7.17749417e-01 5.83845437e-01 3.17088604e-01 8.71549129e-01
6.08210564e-01 3.12399477e-01 3.89113635e-01 -1.00192189e+00
1.66112319e-01 5.13924003e-01 -6.35963008e-02 -8.60853493e-01
-3.41518283e-01 -9.22744453e-01 -1.02328730e+00 4.31959629e-01
4.29274857e-01 -1.89718112e-01 -4.63947475e-01 2.17164636e+00
5.77579178e-02 -7.65161514e-02 3.34487885e-01 6.42339885e-01
3.39460641e-01 6.10163927e-01 -1.39700510e-02 -4.03044999e-01
9.13658202e-01 -2.59738505e-01 -7.36606479e-01 -5.35018332e-02
4.43507969e-01 -2.19997875e-02 1.20933843e+00 3.47235560e-01
-1.08130598e+00 -5.92539489e-01 -1.39679968e+00 1.06415704e-01
-3.44297022e-01 9.29105133e-02 4.84005153e-01 6.03027225e-01
-7.40973473e-01 6.92113936e-01 -1.11494339e+00 -1.55311003e-01
3.46755058e-01 2.31350303e-01 -1.35752082e-01 3.41124594e-01
-1.07631898e+00 8.21506560e-01 3.50654453e-01 1.82371512e-01
-1.02269042e+00 -7.78123140e-01 -8.44884038e-01 3.72456074e-01
3.45123708e-02 -5.87777674e-01 8.48458529e-01 -7.44041324e-01
-1.27489436e+00 5.44225693e-01 -1.48194894e-01 -8.17635894e-01
3.18862557e-01 -1.91811502e-01 -2.79642969e-01 -7.34683359e-04
-1.99771509e-01 4.80083406e-01 1.04678357e+00 -1.26030648e+00
-1.13915600e-01 -3.20165455e-01 -3.78561646e-01 -1.04272202e-01
-4.00055975e-01 -5.06811023e-01 1.82289869e-01 -5.83098829e-01
2.47751594e-01 -5.29695630e-01 6.88898861e-02 -1.14113465e-01
-1.18118659e-01 -3.42619807e-01 6.31575227e-01 -4.95650083e-01
1.44189715e+00 -2.13484526e+00 3.51073354e-01 3.59513193e-01
2.07887143e-01 2.01332673e-01 9.23796073e-02 3.54605377e-01
-1.67227253e-01 -6.09099958e-03 -4.81720716e-01 -6.13403320e-01
2.26066917e-01 2.65105188e-01 -5.07900417e-01 6.07025921e-01
1.28592908e-01 6.10391557e-01 -4.30490315e-01 -2.29165882e-01
3.60496581e-01 6.64287686e-01 -7.35008597e-01 1.10316992e-01
-5.05519986e-01 1.01125330e-01 -2.97469795e-01 2.04289079e-01
4.77598101e-01 -2.69652069e-01 1.86217591e-01 4.64583235e-03
-1.36887943e-02 -6.09812960e-02 -1.10969996e+00 1.35340929e+00
-3.19821924e-01 8.09041500e-01 8.16537440e-02 -1.30415976e+00
1.11170053e+00 1.52944311e-01 3.50108236e-01 -4.10601228e-01
3.65550131e-01 1.93634108e-01 1.04751335e-02 -1.42217085e-01
3.42072487e-01 -4.72102195e-01 7.87941292e-02 3.61457378e-01
5.41625559e-01 4.23147529e-02 1.59422204e-01 4.98690382e-02
4.19279665e-01 1.56752005e-01 3.16945553e-01 -6.20983183e-01
2.82037556e-01 -4.84167725e-01 6.39437318e-01 8.06219399e-01
3.32764499e-02 5.38706899e-01 7.39382744e-01 -2.85993993e-01
-1.20010817e+00 -1.45159674e+00 -6.05836868e-01 5.27112126e-01
-7.10685179e-02 -3.73185754e-01 -7.67107666e-01 -2.69455016e-01
-9.76058245e-02 1.35359883e+00 -6.84231818e-01 -4.45241123e-01
-2.69008934e-01 -1.09955716e+00 3.22743684e-01 4.16922599e-01
4.43541229e-01 -7.68556774e-01 -6.56684220e-01 1.56221986e-01
5.92499748e-02 -8.22881579e-01 9.07514989e-02 3.63750607e-01
-8.88417065e-01 -8.74397516e-01 -4.84231442e-01 -4.14193749e-01
4.22920644e-01 -3.49307984e-01 1.09917474e+00 -3.65135163e-01
-2.28005052e-01 3.87935191e-01 -1.06098838e-01 -6.40946388e-01
-5.80048621e-01 -1.95342768e-02 3.06205034e-01 1.04740132e-02
6.18401706e-01 -1.01927674e+00 -4.73090768e-01 2.67223716e-01
-8.41225386e-01 -3.83652031e-01 3.06293041e-01 8.03951979e-01
8.05964768e-01 4.21512127e-01 5.99793673e-01 -2.44461849e-01
7.18985856e-01 -5.31752050e-01 -1.07996655e+00 2.61495680e-01
-9.59943771e-01 4.92952824e-01 6.89324975e-01 -4.52639371e-01
-1.00022995e+00 -3.78702432e-01 -2.96142012e-01 -8.28367829e-01
8.10109675e-02 4.79981869e-01 -1.33181676e-01 4.50081915e-01
7.42361367e-01 3.18699360e-01 5.97239248e-02 -5.15700340e-01
2.23449409e-01 3.41064513e-01 3.52007896e-01 -4.79887545e-01
5.34709096e-01 4.19276029e-01 3.32183480e-01 -1.16982520e+00
-6.36863291e-01 5.06474860e-02 -2.53758639e-01 -4.26254347e-02
8.91336560e-01 -6.12993240e-01 -8.83861303e-01 1.78965628e-01
-8.68794143e-01 -2.27195635e-01 -6.48308098e-01 9.02813792e-01
-7.61628628e-01 2.87125677e-01 -3.11087936e-01 -1.25904977e+00
-1.19633943e-01 -1.03852856e+00 5.58565319e-01 1.06590435e-01
-1.46466494e-01 -1.14283371e+00 2.77926058e-01 -1.28629476e-01
1.79906085e-01 9.23985541e-02 1.36112463e+00 -4.24548924e-01
-3.32959294e-01 -2.03100294e-01 5.03576137e-02 8.25268269e-01
-4.77198601e-01 1.13877989e-01 -8.56139123e-01 -3.70973676e-01
4.80386525e-01 -2.20247746e-01 1.10545814e+00 9.71326053e-01
1.01687491e+00 -1.15156949e-01 -1.51006162e-01 7.58952558e-01
1.50107300e+00 1.66302949e-01 5.36439180e-01 1.10199548e-01
-3.81422080e-02 6.13436937e-01 4.45079282e-02 4.82953697e-01
1.01453744e-01 5.46178043e-01 5.67649186e-01 2.98436224e-01
1.60662711e-01 -5.06652951e-01 2.96921283e-01 7.64321625e-01
1.51813909e-01 -3.99895132e-01 -5.91797829e-01 4.59867418e-01
-1.56265569e+00 -8.58599901e-01 5.21156676e-02 2.41339183e+00
3.52251530e-01 2.47504398e-01 1.06314950e-01 5.08781429e-03
5.18239856e-01 1.72673166e-01 -6.63492858e-01 -3.56269896e-01
-3.17294508e-01 1.58272162e-01 1.44600064e-01 6.42542005e-01
-9.19268131e-01 5.61519563e-01 6.50279760e+00 9.01183844e-01
-7.77812004e-01 -1.01853078e-02 3.00831318e-01 -1.43972844e-01
-4.21167254e-01 -2.21115425e-01 -1.05670846e+00 5.82818210e-01
1.05191910e+00 -7.67082274e-02 2.61907637e-01 1.03454852e+00
5.31690754e-02 2.67428514e-02 -1.10758007e+00 9.75733340e-01
-1.62299782e-01 -1.12719893e+00 5.28154597e-02 3.82353038e-01
6.07237577e-01 -1.12655528e-01 5.59458017e-01 5.03332615e-01
2.42701501e-01 -1.04501855e+00 5.56308866e-01 4.69115704e-01
6.26230896e-01 -8.15871537e-01 6.01200104e-01 6.17313266e-01
-9.31659281e-01 -1.63340122e-01 -8.11557949e-01 4.83183116e-02
2.26483136e-01 7.38433421e-01 -5.92313111e-01 3.29162747e-01
5.57429314e-01 3.32833111e-01 -2.05864653e-01 5.81789970e-01
-1.13029487e-01 5.37241340e-01 -6.24819636e-01 -6.13898411e-02
1.31201014e-01 -6.20304286e-01 7.77129412e-01 8.16316605e-01
6.66044652e-01 -2.46242926e-01 -4.65946794e-01 1.44161069e+00
9.56792980e-02 -5.27951196e-02 -4.57451493e-01 -2.49061301e-01
2.07777053e-01 6.25324667e-01 -3.81056190e-01 1.48728848e-01
-2.91933775e-01 4.92315352e-01 5.55817425e-01 5.17906725e-01
-7.16035128e-01 -2.55554587e-01 9.66981292e-01 6.41784221e-02
6.51962042e-01 -3.73160452e-01 -2.42667541e-01 -1.33920848e+00
1.22025542e-01 -5.29296935e-01 4.16442245e-01 -5.94683349e-01
-1.34714174e+00 6.60222471e-01 3.88423175e-01 -1.01703238e+00
-6.98194921e-01 -8.47743213e-01 -4.23861295e-01 7.11666524e-01
-1.18526542e+00 -4.73675638e-01 2.16133147e-01 5.39917111e-01
2.28705063e-01 -3.37353557e-01 9.87399936e-01 1.04977615e-01
-6.67696476e-01 7.01671422e-01 6.45213664e-01 -2.03378126e-02
-1.75506920e-01 -1.13282514e+00 5.49649410e-02 7.07771599e-01
3.10006857e-01 7.54416585e-01 1.10896182e+00 -3.76190126e-01
-8.33147049e-01 -6.50882602e-01 5.28153479e-01 -2.81675428e-01
4.56176728e-01 -5.04454315e-01 -1.05952895e+00 8.47818196e-01
-9.63977128e-02 -2.78547674e-01 7.80415595e-01 3.85201454e-01
-3.13680977e-01 -1.31660283e-01 -1.04907632e+00 5.00621915e-01
6.88027561e-01 -5.12981176e-01 -6.84975207e-01 1.35574359e-02
5.15277505e-01 1.79803535e-01 -8.59918535e-01 5.20509720e-01
6.06937706e-01 -1.29514384e+00 1.06938088e+00 -7.77552187e-01
5.06972313e-01 1.36363983e-01 -5.55547774e-01 -1.23163712e+00
-1.92610268e-02 -2.22481698e-01 -4.78905678e-01 8.00556660e-01
3.88806432e-01 -6.75128698e-01 7.30636597e-01 4.04966474e-01
-3.96887176e-02 -8.91072273e-01 -1.24533248e+00 -9.16429698e-01
5.29069483e-01 -6.62462711e-01 1.13102682e-01 4.72028702e-01
-2.24615231e-01 2.75201440e-01 -3.41104984e-01 1.28313273e-01
9.18752968e-01 -1.83741748e-02 3.87018442e-01 -1.40725696e+00
-6.39017105e-01 -6.56227767e-01 -3.36121410e-01 -1.24869990e+00
2.99216628e-01 -7.51672387e-01 -2.38194048e-01 -1.06464314e+00
-1.34088635e-01 -3.15080166e-01 -4.07143921e-01 -2.65809476e-01
3.29293936e-01 -1.81525633e-01 -5.13420999e-02 1.47632226e-01
-1.41863972e-01 1.07522631e+00 6.82346702e-01 2.06792325e-01
-3.82082611e-01 2.04333261e-01 -3.11940044e-01 9.69488621e-01
6.89454913e-01 -4.88920629e-01 -7.13372827e-01 -2.28707805e-01
5.45956850e-01 5.15296981e-02 4.53175813e-01 -9.29786325e-01
-2.70562544e-02 1.08705059e-01 3.37944835e-01 -5.91672242e-01
7.79714108e-01 -5.67082524e-01 1.03060298e-01 1.52934104e-01
-3.27184051e-01 -2.24398896e-01 1.34912189e-02 7.88494051e-01
-2.26566419e-01 -5.98051190e-01 9.38339829e-01 -9.67665389e-02
-3.04371983e-01 5.33538684e-02 -3.74090075e-01 -5.01807854e-02
8.08792710e-01 -1.45339444e-01 2.47537307e-02 -6.21088624e-01
-8.52837861e-01 -3.68485530e-03 2.61547983e-01 -4.80161048e-02
5.92705727e-01 -1.10652947e+00 -5.45650065e-01 6.58345222e-01
-1.65632457e-01 6.19805651e-03 4.67380881e-01 5.66678226e-01
-2.67955691e-01 4.03805792e-01 -9.52035487e-02 -6.02677047e-01
-6.55768812e-01 7.27719367e-01 6.45049453e-01 -3.08412880e-01
-5.22669315e-01 9.33089495e-01 5.29075027e-01 -2.53648341e-01
4.76623893e-01 -1.57283381e-01 1.26805122e-03 3.39785069e-02
2.94893980e-01 4.42215592e-01 -2.24641368e-01 -4.52938706e-01
-1.45197228e-01 4.54273701e-01 1.96601495e-01 -3.71586114e-01
1.43822753e+00 -1.71387091e-01 1.91480786e-01 6.66888654e-01
1.48619425e+00 -4.06886637e-01 -1.44839847e+00 -1.30011961e-01
-3.99045408e-01 -1.10438392e-01 2.43041292e-01 -3.63295943e-01
-8.25043023e-01 1.22983277e+00 6.13808334e-01 3.46853644e-01
8.43015373e-01 2.45342389e-01 2.82005787e-01 3.98711801e-01
4.08614166e-02 -1.16040516e+00 7.24253356e-02 2.76071280e-01
8.96206737e-01 -1.00048804e+00 -1.40524268e-01 6.67273924e-02
-3.97396654e-01 9.54890370e-01 3.48072797e-01 -2.95717508e-01
9.88169670e-01 6.67922944e-03 -1.76428750e-01 -2.13554129e-01
-6.12815261e-01 4.81000356e-02 2.55848408e-01 5.36986172e-01
4.63780388e-03 -5.98369166e-02 -3.65980953e-01 8.88619304e-01
-4.91049051e-01 -3.15859735e-01 2.42306083e-01 3.32454562e-01
-3.36473227e-01 -8.53957713e-01 -3.52534354e-02 2.70312518e-01
-2.23107502e-01 1.58253878e-01 5.46162091e-02 1.05963242e+00
-7.14974850e-02 6.58329666e-01 3.55589569e-01 -2.26681843e-01
9.84722655e-03 5.56407273e-01 3.73712480e-01 -8.44891444e-02
6.14695959e-02 1.24392666e-01 -3.71528953e-01 -2.14210808e-01
-8.27969834e-02 -7.44977415e-01 -8.94956648e-01 -2.58863866e-01
-3.43123436e-01 2.77642697e-01 8.58196676e-01 7.52662301e-01
2.10673958e-01 3.95385087e-01 3.62487376e-01 -4.12267894e-01
-1.01197124e+00 -1.06830156e+00 -1.00908768e+00 2.31086358e-01
2.72739559e-01 -1.01285827e+00 -8.34687829e-01 -3.20412129e-01] | [7.494840145111084, 3.8390302658081055] |
119e2241-8810-48d2-ade7-3af6d4560376 | robust-machine-learning-pipelines-for-trading | 2301.0079 | null | https://arxiv.org/abs/2301.00790v2 | https://arxiv.org/pdf/2301.00790v2.pdf | Dynamic Feature Engineering and model selection methods for temporal tabular datasets with regime changes | The application of deep learning algorithms to temporal panel datasets is difficult due to heavy non-stationarities which can lead to over-fitted models that under-perform under regime changes. In this work we propose a new machine learning pipeline for ranking predictions on temporal panel datasets which is robust under regime changes of data. Different machine-learning models, including Gradient Boosting Decision Trees (GBDTs) and Neural Networks with and without simple feature engineering are evaluated in the pipeline with different settings. We find that GBDT models with dropout display high performance, robustness and generalisability with relatively low complexity and reduced computational cost. We then show that online learning techniques can be used in post-prediction processing to enhance the results. In particular, dynamic feature neutralisation, an efficient procedure that requires no retraining of models and can be applied post-prediction to any machine learning model, improves robustness by reducing drawdown in regime changes. Furthermore, we demonstrate that the creation of model ensembles through dynamic model selection based on recent model performance leads to improved performance over baseline by improving the Sharpe and Calmar ratios of out-of-sample prediction performances. We also evaluate the robustness of our pipeline across different data splits and random seeds with good reproducibility of results. | ['Mauricio Barahona', 'Thomas Wong'] | 2022-12-30 | null | null | null | null | ['feature-engineering'] | ['methodology'] | [ 2.53135502e-01 -3.08118194e-01 -2.83170849e-01 -6.20846629e-01
-7.76956558e-01 -9.26205277e-01 8.11808407e-01 2.54873931e-01
-2.74840951e-01 8.67290258e-01 3.06872666e-01 -6.66845143e-01
-6.32288635e-01 -6.39499962e-01 -9.70658481e-01 -6.65426433e-01
-5.86705580e-02 5.73478162e-01 3.51884305e-01 -2.11682603e-01
-6.54196441e-02 5.23867965e-01 -1.45527470e+00 6.51604950e-01
9.28623080e-01 8.46590579e-01 -3.59135091e-01 7.30376840e-01
4.54018265e-01 7.23494768e-01 -2.27553800e-01 -6.03934705e-01
6.94135249e-01 -2.48761103e-01 -2.35272884e-01 -3.17467839e-01
6.10249400e-01 -2.55932391e-01 -3.36843044e-01 3.76520962e-01
8.27945530e-01 9.01400670e-02 7.20948100e-01 -9.62221324e-01
-1.57094553e-01 6.92748666e-01 -4.94153082e-01 2.52128720e-01
-1.40128613e-01 4.56377268e-01 9.67004180e-01 -4.08386648e-01
5.72001159e-01 1.28753388e+00 1.36328804e+00 -2.91881580e-02
-2.06636691e+00 -7.22764671e-01 4.61313933e-01 2.40559116e-01
-8.56160700e-01 -5.59913933e-01 5.17066479e-01 -7.35080481e-01
1.14672160e+00 3.24094713e-01 5.18933713e-01 1.28889608e+00
4.26367700e-01 3.36843431e-01 1.29467928e+00 -3.09561223e-01
2.14960530e-01 7.96941444e-02 2.89959073e-01 2.10679263e-01
2.53051102e-01 6.57391787e-01 -5.21626890e-01 -3.57593238e-01
2.95005739e-01 7.29006454e-02 1.06945835e-01 -1.82396337e-01
-7.64513671e-01 8.35600078e-01 3.12648654e-01 -1.97161078e-01
-5.09526849e-01 3.29153575e-02 6.12983942e-01 6.04107022e-01
8.82510424e-01 1.72289297e-01 -1.14012623e+00 1.15586914e-01
-1.08736026e+00 3.32681924e-01 6.60253704e-01 2.54831433e-01
5.18556833e-01 -5.30124940e-02 -4.93104070e-01 8.84750426e-01
-1.37912497e-01 5.36301374e-01 2.88182527e-01 -8.64830852e-01
4.20185000e-01 7.50024736e-01 2.26078868e-01 -8.19091082e-01
-9.69106257e-01 -8.67937624e-01 -9.73070562e-01 3.13188463e-01
5.38267314e-01 -5.31745076e-01 -1.11079085e+00 1.84945929e+00
4.35245633e-01 -7.64358193e-02 -1.69051915e-01 3.95459533e-01
3.75772029e-01 5.40867448e-01 3.25494617e-01 -3.88055533e-01
1.02383327e+00 -5.28057337e-01 -3.80935550e-01 -1.81499600e-01
7.83784568e-01 -2.33571708e-01 7.61748433e-01 3.64017099e-01
-5.78085124e-01 -6.18713558e-01 -8.94742906e-01 2.28678733e-01
-4.80803400e-01 5.20794317e-02 7.47709632e-01 6.97140992e-01
-9.05180752e-01 1.18240595e+00 -8.63071084e-01 -3.84627879e-01
4.62756872e-01 6.58681810e-01 -3.33557129e-01 1.95162237e-01
-1.20717263e+00 9.16419387e-01 5.19646525e-01 2.80964017e-01
-6.27714932e-01 -1.11266160e+00 -5.47230780e-01 2.10956052e-01
9.83112380e-02 -8.34385514e-01 1.06992161e+00 -1.39583874e+00
-1.49377465e+00 6.03998601e-01 -4.90171388e-02 -8.03307414e-01
8.65576923e-01 -2.22333655e-01 -4.27945077e-01 -5.13540387e-01
-2.52275348e-01 4.13747907e-01 6.88274741e-01 -7.97136009e-01
-7.82962561e-01 -3.47749949e-01 -3.65701616e-01 -4.56715189e-03
-8.29183608e-02 1.74281329e-01 1.93262592e-01 -4.97498214e-01
-1.44576013e-01 -1.11552930e+00 -2.61997819e-01 -6.34502232e-01
-8.92587155e-02 -8.37564841e-02 3.79023224e-01 -1.00174630e+00
1.31112552e+00 -1.89076936e+00 -6.67462125e-02 2.82309294e-01
-2.81463206e-01 2.34362572e-01 -1.24593951e-01 6.04227841e-01
-1.84614092e-01 1.23346299e-01 -1.91556603e-01 -7.62999877e-02
1.25152498e-01 -3.66309960e-03 -6.25731587e-01 4.00025517e-01
3.39081407e-01 1.10604274e+00 -5.29928982e-01 5.00271609e-03
2.70046473e-01 1.16512008e-01 -4.98525321e-01 -2.09735762e-02
-3.36228222e-01 4.34590578e-01 2.26951912e-01 3.37767452e-01
6.67991281e-01 -1.26633108e-01 6.82643533e-01 1.07586704e-01
-2.56849468e-01 5.32719076e-01 -9.06158328e-01 9.74240780e-01
-4.13761914e-01 7.18039453e-01 -1.95962787e-01 -1.00211990e+00
8.00283432e-01 5.73485531e-02 2.38110021e-01 -9.47696686e-01
-2.18038872e-01 2.74654478e-01 2.97194034e-01 -1.81139052e-01
8.56212229e-02 -2.14738667e-01 4.46737744e-02 3.42965305e-01
-7.98757002e-02 3.63547742e-01 1.48623347e-01 -2.35677868e-01
1.13355720e+00 4.27233398e-01 1.77848041e-01 -3.82815927e-01
-6.08303770e-02 1.27346694e-01 9.85924780e-01 1.11353076e+00
1.68113187e-01 3.59413266e-01 7.87977278e-01 -8.03498566e-01
-1.13704610e+00 -8.17609727e-01 -2.65609235e-01 1.81376910e+00
-7.30217993e-01 -6.97866380e-02 -3.78534868e-02 -6.52615428e-01
5.58130085e-01 8.06254685e-01 -8.79355669e-01 -2.00888470e-01
-5.87738335e-01 -1.60418177e+00 2.78910816e-01 6.33566141e-01
1.70752570e-01 -9.85442698e-01 -4.78477597e-01 3.60696048e-01
3.52290362e-01 -6.14108622e-01 2.13862419e-01 6.62701905e-01
-1.22546434e+00 -9.39615786e-01 -3.73983890e-01 -3.40838104e-01
4.87877131e-02 -1.62862420e-01 1.34151721e+00 -3.44908834e-01
2.18818694e-01 -2.34634057e-01 -2.50722736e-01 -5.96914649e-01
-5.66888034e-01 6.13377094e-01 5.24700917e-02 6.64227828e-02
2.25731850e-01 -7.81973183e-01 -7.32023656e-01 3.06014568e-01
-6.13858283e-01 2.08872586e-01 4.20521826e-01 1.05327916e+00
3.14523190e-01 -1.70297772e-01 9.09563959e-01 -1.16365945e+00
3.47976059e-01 -4.59539920e-01 -9.52589750e-01 4.27653760e-01
-9.61966634e-01 1.65772289e-01 6.17358148e-01 -4.02450979e-01
-1.24056566e+00 1.15839332e-01 1.50007576e-01 -2.22028792e-01
-7.55568370e-02 6.42422199e-01 2.30149463e-01 4.09929574e-01
1.00614667e+00 -1.35417029e-01 -1.33352026e-01 -6.86531544e-01
2.03671187e-01 3.48992437e-01 1.63749203e-01 -1.78172648e-01
7.11730123e-01 4.49559748e-01 2.00048849e-01 -2.75844693e-01
-8.94156694e-01 -5.31137437e-02 -8.51371884e-01 -1.47880852e-01
3.10569376e-01 -1.22159648e+00 -6.13614857e-01 7.06955791e-01
-5.19243896e-01 -1.11859226e+00 -8.84507298e-02 3.56846482e-01
-1.93206802e-01 -4.24239859e-02 -4.43835557e-01 -8.92105281e-01
-5.74159682e-01 -5.62321484e-01 8.12432587e-01 2.69211054e-01
-1.96260363e-01 -1.31101906e+00 5.04490495e-01 6.80930093e-02
5.09019971e-01 6.10594213e-01 1.07551694e+00 -8.41208577e-01
-1.30778939e-01 -2.66010046e-01 -3.43767554e-02 2.20085546e-01
7.79494494e-02 3.58470500e-01 -1.29463506e+00 -5.27044237e-01
-7.40758479e-01 -2.67145395e-01 1.49651682e+00 8.52030277e-01
8.73140872e-01 -2.69526273e-01 -4.01678622e-01 7.71774709e-01
1.19408560e+00 9.11652073e-02 5.08433759e-01 5.24682641e-01
3.55411202e-01 9.25595045e-01 3.62294853e-01 3.24700326e-01
5.86847551e-02 7.08252728e-01 7.09944665e-02 -3.36477727e-01
1.87007442e-01 -3.44909906e-01 5.02885163e-01 1.23680435e-01
1.06022414e-02 -1.41120836e-01 -9.79168355e-01 3.90648127e-01
-2.31231332e+00 -1.24746215e+00 -4.35796559e-01 2.40574193e+00
6.83884263e-01 3.88910800e-01 5.87235749e-01 -5.61793782e-02
4.42480028e-01 3.02359741e-02 -9.34340835e-01 -5.53936124e-01
-4.89288867e-01 2.16014281e-01 8.32761467e-01 3.94466162e-01
-1.28308749e+00 8.51361036e-01 6.84435511e+00 5.78419089e-01
-1.16259098e+00 7.26061836e-02 1.10271358e+00 -4.15691823e-01
-8.25058296e-02 2.33125210e-01 -8.80234122e-01 5.48388600e-01
1.46554160e+00 2.66403835e-02 3.01692009e-01 5.47329843e-01
6.27527416e-01 4.02599387e-03 -1.09077299e+00 4.36422735e-01
-5.23107827e-01 -1.41887867e+00 -3.11981916e-01 6.57212408e-03
9.67577577e-01 4.47670668e-01 1.08835809e-01 6.42988384e-01
8.81373346e-01 -8.75105441e-01 5.88490725e-01 7.20263779e-01
6.91078305e-01 -8.21846843e-01 8.57654512e-01 3.21263909e-01
-7.92069495e-01 -6.38715565e-01 -3.54547560e-01 -4.21345860e-01
-6.33748025e-02 8.77432764e-01 -8.20914209e-01 7.01259971e-01
7.91907787e-01 7.11736202e-01 -8.51916254e-01 9.40340340e-01
7.74757192e-02 1.09902036e+00 -6.47435129e-01 2.17279181e-01
-1.47302076e-01 -2.22893313e-01 1.23213686e-01 1.39658093e+00
1.81362808e-01 -2.19223112e-01 6.39175624e-02 3.16432536e-01
1.16577134e-01 -8.89844969e-02 -6.04664087e-01 4.69723523e-01
2.30822742e-01 1.09758496e+00 -4.33806390e-01 -3.06519061e-01
-1.25601858e-01 5.06138682e-01 4.99006569e-01 4.39424515e-01
-7.26614535e-01 -1.18974008e-01 4.36152726e-01 2.80751139e-01
3.94333541e-01 1.05282046e-01 -4.16615129e-01 -9.63105619e-01
-1.31556630e-01 -9.57466900e-01 9.76559520e-01 -5.53659141e-01
-1.54907525e+00 2.00886354e-01 1.90833896e-01 -8.77503097e-01
-4.82338846e-01 -6.11400485e-01 -7.13099718e-01 8.40445459e-01
-1.34063208e+00 -1.16689003e+00 -5.55021130e-02 2.22036734e-01
2.04820782e-01 -9.99576822e-02 4.46361452e-01 -1.73653010e-02
-9.22480166e-01 6.21345520e-01 8.29226911e-01 -2.35965356e-01
9.00045097e-01 -1.19199038e+00 5.43170691e-01 8.72105479e-01
-1.93974867e-01 4.16853040e-01 6.85018063e-01 -7.50311732e-01
-8.43934715e-01 -1.38087857e+00 8.13645244e-01 -5.01862705e-01
6.73458576e-01 -6.60103202e-01 -9.60586190e-01 8.87061834e-01
2.32949838e-01 -4.43184584e-01 7.89455533e-01 7.36170471e-01
-3.70885074e-01 -7.44512856e-01 -8.54563653e-01 5.03728330e-01
9.98792350e-01 -2.29081437e-01 -2.01575145e-01 5.09161711e-01
4.46513951e-01 -9.92305949e-02 -8.75047982e-01 7.71496654e-01
1.06263602e+00 -1.12089360e+00 5.83295226e-01 -1.17992818e+00
1.59320399e-01 5.92710776e-03 -5.61643988e-02 -1.34205687e+00
-8.12062442e-01 -8.23113739e-01 -5.74778183e-04 1.41143477e+00
9.15725231e-01 -9.10637498e-01 4.56234634e-01 7.01836884e-01
8.88702720e-02 -3.54049087e-01 -9.66199875e-01 -7.31358171e-01
3.84509385e-01 -3.11504453e-01 3.91782939e-01 8.33091557e-01
-3.36107373e-01 3.86378706e-01 -6.47540927e-01 3.68650407e-02
3.72320324e-01 3.64321560e-01 9.50712979e-01 -1.58935416e+00
-5.00100493e-01 -4.91091341e-01 4.42130752e-02 -3.80596459e-01
2.91438699e-01 -9.57277596e-01 -3.18202853e-01 -1.24730933e+00
3.11495781e-01 -3.70002687e-01 -6.75002277e-01 8.28821838e-01
-3.48572731e-01 8.97089913e-02 1.00042924e-01 9.83275250e-02
-2.65610486e-01 4.30955380e-01 4.98629361e-01 2.47722045e-02
-6.48616850e-01 3.39350611e-01 -4.97790039e-01 3.41117561e-01
1.01988935e+00 -6.66386306e-01 -1.43527985e-01 -1.92423016e-01
5.36001801e-01 -1.27649933e-01 5.08785129e-01 -8.44193757e-01
-2.57306129e-01 -1.52233675e-01 9.53752339e-01 -6.33026719e-01
-7.56337121e-02 -6.03123963e-01 7.86657870e-01 7.64594197e-01
-4.90860164e-01 3.28177989e-01 6.60788357e-01 7.03842282e-01
4.83602524e-01 1.28357753e-01 6.47636294e-01 8.40799883e-02
-9.21026319e-02 -3.09108533e-02 -4.29123342e-01 -4.05048490e-01
5.83784044e-01 6.74657822e-02 -4.20256823e-01 -2.92380482e-01
-7.49239743e-01 4.52933669e-01 4.07601267e-01 2.60545641e-01
-3.73644114e-01 -1.06318963e+00 -1.20529222e+00 1.05246805e-01
-1.69864446e-01 -5.68451703e-01 4.04821217e-01 1.08499515e+00
-3.59580278e-01 4.65295523e-01 -6.69617504e-02 -6.83100820e-01
-1.29478884e+00 4.78613466e-01 6.84820473e-01 -8.80758286e-01
-3.45904887e-01 4.33447748e-01 7.73271099e-02 -6.36297822e-01
-6.45254701e-02 -3.07669044e-01 3.31909433e-02 5.08769572e-01
1.76658839e-01 5.59255421e-01 3.49827498e-01 -1.15751959e-01
-3.68924469e-01 2.46139243e-01 -1.62108615e-01 -9.55178067e-02
1.69954598e+00 1.95211962e-01 1.66593894e-01 6.92963779e-01
7.56152213e-01 -2.45818436e-01 -1.80246711e+00 -4.89137396e-02
3.30112040e-01 -8.45668167e-02 4.91298102e-02 -1.24888551e+00
-7.00978041e-01 6.94037020e-01 9.83313680e-01 3.01168084e-01
1.24689925e+00 -4.44836050e-01 3.50530818e-02 3.84856820e-01
2.24525090e-02 -1.21236384e+00 -5.74194849e-01 4.29269820e-01
8.27647030e-01 -1.42767799e+00 3.29148740e-01 2.92426646e-01
-5.24153948e-01 9.26548779e-01 3.25009674e-01 6.08157786e-03
6.81220233e-01 -1.10191695e-01 7.43986815e-02 1.97889045e-01
-1.52181685e+00 -2.58104056e-01 3.49804163e-01 5.76138675e-01
2.77875394e-01 2.11159766e-01 -2.11657196e-01 4.92425054e-01
-1.82924315e-01 -1.41036818e-02 7.04010874e-02 4.52163041e-01
-2.22443894e-01 -9.74597931e-01 -2.18706310e-01 9.62655962e-01
-5.89009583e-01 -1.16962343e-01 -6.02469563e-01 8.35441411e-01
-2.89217699e-02 1.04227233e+00 1.94174394e-01 -5.27310610e-01
4.81606632e-01 6.34727061e-01 6.48264438e-02 -2.34627828e-01
-1.18148339e+00 1.72080994e-01 4.52772021e-01 -2.88122922e-01
-2.13126644e-01 -1.06498349e+00 -3.23123187e-01 -4.80203927e-01
-3.68963182e-01 -1.48323119e-01 3.56044471e-01 7.34552324e-01
9.34730649e-01 3.36829185e-01 9.00843859e-01 -9.14495170e-01
-7.33088136e-01 -1.39683378e+00 -2.60947675e-01 5.24521589e-01
2.48287171e-01 -5.99578857e-01 -4.52355236e-01 -4.65515628e-02] | [7.121635437011719, 3.38749623298645] |
c1a96da3-d056-4d64-95a9-f37a464c4630 | graph-model-for-chinese-spell-checking | null | null | https://aclanthology.org/W13-4416 | https://aclanthology.org/W13-4416.pdf | Graph Model for Chinese Spell Checking | null | ['Zhongye Jia', 'Peilu Wang', 'Hai Zhao'] | 2013-10-01 | null | null | null | ws-2013-10 | ['chinese-spell-checking'] | ['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] | [-7.286793231964111, 3.6747426986694336] |
f7d3ea00-d1ba-4ff4-b6a6-d7445ecab82f | multi-source-education-knowledge-graph | 2305.04567 | null | https://arxiv.org/abs/2305.04567v1 | https://arxiv.org/pdf/2305.04567v1.pdf | Multi-source Education Knowledge Graph Construction and Fusion for College Curricula | The field of education has undergone a significant transformation due to the rapid advancements in Artificial Intelligence (AI). Among the various AI technologies, Knowledge Graphs (KGs) using Natural Language Processing (NLP) have emerged as powerful visualization tools for integrating multifaceted information. In the context of university education, the availability of numerous specialized courses and complicated learning resources often leads to inferior learning outcomes for students. In this paper, we propose an automated framework for knowledge extraction, visual KG construction, and graph fusion, tailored for the major of Electronic Information. Furthermore, we perform data analysis to investigate the correlation degree and relationship between courses, rank hot knowledge concepts, and explore the intersection of courses. Our objective is to enhance the learning efficiency of students and to explore new educational paradigms enabled by AI. The proposed framework is expected to enable students to better understand and appreciate the intricacies of their field of study by providing them with a comprehensive understanding of the relationships between the various concepts and courses. | ['Hui Zhao', 'Xinning Zhu', 'Chunhong Zhang', 'Linya Cheng', 'Zeju Li'] | 2023-05-08 | null | null | null | null | ['graph-construction'] | ['graphs'] | [ 2.95441430e-02 5.71484193e-02 -1.45956337e-01 1.11325175e-01
7.25233257e-02 -6.80321932e-01 4.52757388e-01 1.12616372e+00
-8.65128115e-02 5.78759491e-01 -5.53546799e-03 -5.62788785e-01
-8.09358358e-01 -1.03612781e+00 -1.78792149e-01 -5.76541126e-01
-5.07315174e-02 3.12274583e-02 2.25631848e-01 -4.39067006e-01
6.02631807e-01 1.03577971e+00 -2.06200314e+00 4.07980904e-02
1.47582197e+00 5.98163545e-01 1.98601365e-01 1.86654747e-01
-8.26975882e-01 6.84603095e-01 -7.33742535e-01 -4.77064669e-01
-3.60848397e-01 -2.63512254e-01 -6.60548806e-01 -6.98562562e-02
4.09184277e-01 3.17357033e-01 -5.71280681e-02 1.16736770e+00
1.66172177e-01 2.79810756e-01 3.83812338e-01 -1.22337484e+00
-4.99073148e-01 4.49478239e-01 -4.88362283e-01 9.43000764e-02
6.34836733e-01 -3.11945736e-01 4.82413799e-01 -6.25511527e-01
7.82425642e-01 1.07117128e+00 1.35180518e-01 -2.15091825e-01
-7.73376763e-01 -4.71251249e-01 2.37458721e-01 7.90216088e-01
-1.39167571e+00 3.28681648e-01 1.03353322e+00 -6.82026505e-01
1.42938763e-01 1.11552089e-01 1.27922297e+00 3.27862710e-01
2.36136854e-01 6.80971980e-01 1.05628002e+00 -8.65837097e-01
1.64201692e-01 6.67542577e-01 5.45355380e-01 1.05233836e+00
7.37976253e-01 -3.28084558e-01 -8.27185750e-01 3.03196162e-01
4.43154514e-01 2.54237652e-01 -2.73468107e-01 -6.57626212e-01
-9.88722205e-01 6.10781074e-01 4.60928887e-01 7.15781808e-01
-2.55949140e-01 -6.15635216e-01 -9.44777858e-03 1.54764131e-01
-2.55028725e-01 6.11060083e-01 6.38124645e-02 -2.39018589e-01
-7.85343647e-01 1.11442544e-01 8.20294559e-01 7.90847301e-01
7.35855997e-01 3.89252231e-02 1.58034936e-01 1.70497626e-01
4.10591483e-01 8.76930654e-02 5.11865497e-01 -4.05082703e-01
2.35484049e-01 1.54515147e+00 -4.01706547e-01 -1.41609848e+00
-2.45399475e-01 -5.45641899e-01 -6.03426933e-01 3.88634264e-01
4.14491951e-01 8.20450187e-02 -5.70037782e-01 1.16853833e+00
4.89242136e-01 -7.33072832e-02 5.78374341e-02 4.59232420e-01
1.33050013e+00 2.26425469e-01 3.73632193e-01 7.94058666e-03
1.69401598e+00 -3.55069429e-01 -1.11318195e+00 1.76332697e-01
8.63134325e-01 -8.66801023e-01 1.06675994e+00 5.35000265e-01
-7.33070195e-01 -5.99437654e-01 -1.15742445e+00 -7.30439872e-02
-1.18318319e+00 8.77290517e-02 6.08247876e-01 8.43184352e-01
-7.64782310e-01 3.64254594e-01 -5.80953240e-01 -3.92738491e-01
4.18921620e-01 1.32187635e-01 -2.11834103e-01 -4.21670228e-01
-1.07948482e+00 1.00745392e+00 3.92277747e-01 -1.79713458e-01
-1.61859930e-01 -1.04004741e+00 -7.85432279e-01 3.81896377e-01
5.40041506e-01 -5.14721692e-01 4.84217763e-01 -4.00394887e-01
-1.02488124e+00 5.59864640e-01 2.45416909e-01 1.99887846e-02
2.43929207e-01 -1.33085251e-01 -4.58003163e-01 1.61107793e-01
-1.66430056e-01 1.17090091e-01 -4.40052040e-02 -1.04893184e+00
-6.10879838e-01 -6.74599826e-01 -1.13718815e-01 4.74422306e-01
-6.15173101e-01 -1.79266021e-01 -2.43059546e-01 -2.45924905e-01
2.51235723e-01 -4.30035323e-01 -8.78651589e-02 -2.02188417e-01
-4.94440570e-02 -3.10612231e-01 1.30125546e+00 -5.10485768e-01
1.45255160e+00 -1.99784636e+00 8.27285945e-02 6.78064048e-01
4.39499348e-01 3.16221476e-01 4.49858397e-01 6.15716696e-01
1.60759851e-01 8.02261159e-02 2.68617660e-01 3.53503138e-01
-9.39886943e-02 7.08903894e-02 -1.58460473e-03 -6.94825351e-02
-2.85297304e-01 5.68729341e-01 -1.03509498e+00 -8.69602203e-01
5.03192663e-01 6.04432583e-01 3.05205006e-02 1.94908068e-01
1.33789927e-01 2.93517143e-01 -7.00056553e-01 7.59272695e-01
2.41521761e-01 -1.21035822e-01 3.31693381e-01 1.06733054e-01
-5.63986063e-01 -1.10100061e-01 -1.41198063e+00 1.46112406e+00
-1.70085758e-01 9.99682903e-01 -1.20677337e-01 -8.89049292e-01
1.07447577e+00 3.13691765e-01 3.70434493e-01 -5.52908599e-01
1.73521191e-01 -1.88788936e-01 -5.49192429e-02 -6.52656078e-01
5.73835731e-01 2.53843457e-01 1.46069258e-01 2.08559334e-01
1.21732429e-01 -2.23077089e-01 6.92503750e-01 5.68229079e-01
6.94977939e-01 1.44891798e-01 5.93938649e-01 -2.68723607e-01
6.45935059e-01 1.29481599e-01 4.43733446e-02 3.11457247e-01
-5.60144894e-02 -3.61988902e-01 4.89663690e-01 -5.09507120e-01
-2.39710718e-01 -8.91210198e-01 8.70210230e-02 7.61526644e-01
1.62239298e-01 -4.30738807e-01 -4.04251695e-01 -2.59456247e-01
-6.42354563e-02 5.40842414e-01 -8.05788413e-02 -8.15276504e-02
-1.90669015e-01 -2.58309215e-01 1.17484294e-01 3.05217076e-02
4.26818132e-01 -1.08189058e+00 -8.51512909e-01 5.91059886e-02
5.72465993e-02 -1.02427363e+00 4.20542210e-01 -2.04713196e-01
-9.27886486e-01 -1.22064197e+00 -2.51929939e-01 -9.28731382e-01
9.20359731e-01 3.44283283e-01 1.06037915e+00 3.65827709e-01
-8.45907152e-01 6.84171438e-01 -2.90063351e-01 -8.53775620e-01
-1.98325574e-01 8.66917446e-02 -2.46798515e-01 -5.13661623e-01
5.60117304e-01 -6.42609656e-01 -4.16892916e-01 -1.10111587e-01
-1.07032430e+00 1.86579153e-01 6.25290453e-01 1.06782399e-01
4.59349066e-01 6.93331838e-01 4.44733471e-01 -8.90515268e-01
1.00045168e+00 -4.07972306e-01 -7.68364131e-01 7.47415006e-01
-8.55498135e-01 2.22031131e-01 3.87766480e-01 -1.30402148e-01
-1.20517492e+00 -3.10206532e-01 4.44990247e-01 -5.42106107e-02
-6.21709645e-01 9.79646087e-01 -2.75553852e-01 -5.58394372e-01
4.50580209e-01 1.60880044e-01 -3.02315593e-01 -3.55999142e-01
4.92714316e-01 2.92953372e-01 3.65573615e-01 -6.62787020e-01
6.19749367e-01 -1.26992315e-01 5.10704041e-01 -1.02400637e+00
-5.44323564e-01 -4.48027372e-01 -7.70117819e-01 -8.33960414e-01
7.04049349e-01 -5.93327343e-01 -1.30935907e+00 -2.10688174e-01
-6.94077909e-01 4.44585770e-01 -1.43915176e-01 6.97306395e-01
2.08799407e-01 5.14864862e-01 -1.06800973e-01 -8.27053905e-01
-3.01660672e-02 -1.03426933e+00 4.58529629e-02 1.07422137e+00
-5.65721132e-02 -1.05608571e+00 -6.95564076e-02 6.54957891e-01
2.52361536e-01 4.01796073e-01 1.21896076e+00 -4.60713029e-01
-7.42093205e-01 -3.55339974e-01 -2.84299608e-02 -3.13312888e-01
1.08951546e-01 4.48905110e-01 -5.08740664e-01 1.59223616e-01
-4.52864826e-01 4.63597029e-02 2.85929024e-01 -3.25671285e-02
1.11562908e+00 -1.42937332e-01 -4.35622633e-01 -4.90858853e-02
1.48323786e+00 5.77721655e-01 2.82269657e-01 4.65128630e-01
6.87655687e-01 1.06780684e+00 6.00340366e-01 4.46864665e-01
5.37050843e-01 4.19155210e-01 2.82404304e-01 -5.43647781e-02
-3.65314335e-02 -1.80715486e-01 -1.03606872e-01 1.10486221e+00
-2.88270921e-01 1.52441889e-01 -1.34401858e+00 5.34873188e-01
-1.63279068e+00 -6.55065298e-01 -2.65933841e-01 1.97687209e+00
5.42020321e-01 2.66765878e-02 -1.99384019e-01 4.41961139e-01
4.01711285e-01 -2.60652542e-01 1.09211117e-01 -4.92242277e-01
2.29178235e-01 3.01930845e-01 1.18126631e-01 3.09111625e-01
-5.19840121e-01 8.58985364e-01 4.90742064e+00 7.64087439e-01
-9.80987728e-01 -7.02438533e-01 2.09433556e-01 2.56951958e-01
-3.77063662e-01 1.34846708e-02 -8.46874952e-01 4.12774906e-02
7.77145684e-01 -7.00988591e-01 2.09352538e-01 7.38797307e-01
1.86894223e-01 -4.19519514e-01 -6.29575968e-01 5.84143460e-01
-2.72599813e-02 -1.37363935e+00 2.96080232e-01 3.62643413e-02
8.18088055e-01 -8.24151456e-01 7.91461244e-02 1.45370647e-01
5.49973667e-01 -9.50914919e-01 1.34574473e-01 6.33239925e-01
1.08241342e-01 -1.13072991e+00 6.88033164e-01 2.34705389e-01
-1.25743449e+00 -2.45926902e-02 5.76105490e-02 -2.79829323e-01
-3.29010099e-01 4.67431188e-01 -1.08472621e+00 1.09622097e+00
7.07754493e-01 2.57415473e-01 -8.20816219e-01 1.28511536e+00
-4.24417138e-01 2.53608167e-01 1.27224401e-01 -3.91481280e-01
-4.62154374e-02 -5.04555583e-01 2.75364518e-01 8.46498847e-01
5.04629195e-01 3.88492703e-01 2.17645511e-01 6.22309268e-01
1.81247354e-01 6.92710340e-01 -7.58796990e-01 -5.10067284e-01
5.45716107e-01 1.51264894e+00 -1.07852376e+00 -3.83201510e-01
-5.64275026e-01 5.71310632e-02 2.40390003e-01 3.96794349e-01
-9.10432711e-02 -7.32177913e-01 3.19328815e-01 3.08349758e-01
-1.52950451e-01 -4.55851585e-01 -5.22850931e-01 -7.78206468e-01
-3.78866732e-01 -6.77134871e-01 5.75695932e-01 -6.30356312e-01
-6.23227596e-01 2.50321418e-01 1.42810896e-01 -1.06036425e+00
1.13172373e-02 -5.72835207e-01 -6.01232588e-01 7.00213253e-01
-1.28672063e+00 -8.25001657e-01 -6.63298368e-01 6.44951165e-01
6.51355907e-02 -1.67422861e-01 9.03266549e-01 1.10994384e-01
-6.72748923e-01 8.56054425e-02 -4.57045473e-02 -3.16267088e-02
2.82765359e-01 -1.24068308e+00 -6.24438822e-01 8.15377295e-01
3.45137239e-01 5.76454401e-01 6.95172131e-01 -5.51704168e-01
-1.26353061e+00 -3.15664560e-01 9.60224330e-01 2.66047865e-02
6.61585391e-01 1.32288978e-01 -1.00254822e+00 2.61038482e-01
3.44515800e-01 -4.04722601e-01 1.24916804e+00 2.12419018e-01
-1.76878870e-02 -2.75451094e-01 -8.80289912e-01 8.65355670e-01
3.79099965e-01 -3.15707892e-01 -8.02601755e-01 -5.10126688e-02
4.18075979e-01 -2.65260100e-01 -1.18150187e+00 2.28980288e-01
4.17063177e-01 -8.39892626e-01 8.49627852e-01 -5.14425635e-01
2.28764445e-01 -3.48258376e-01 4.03448015e-01 -1.05848789e+00
7.55791217e-02 -1.22281335e-01 -1.06042013e-01 1.09496343e+00
1.05740782e-02 -2.05253854e-01 1.09234619e+00 8.58914316e-01
-9.50428173e-02 -7.26303697e-01 -5.41901410e-01 -1.09394334e-01
-3.89739901e-01 -1.91317424e-01 4.66745973e-01 1.29409087e+00
6.04401529e-01 6.37052357e-02 2.61806458e-01 1.95609376e-01
6.54172182e-01 5.80255449e-01 5.35828114e-01 -1.90780437e+00
3.16329867e-01 -7.42530823e-01 -7.94925928e-01 -8.96283239e-02
-1.82155907e-01 -5.61568975e-01 -6.15469217e-01 -1.82678163e+00
-1.21160313e-01 4.52106185e-02 -3.28296781e-01 4.80508268e-01
-1.66544154e-01 -2.29671419e-01 1.85177505e-01 -1.89329684e-01
-5.56739867e-01 2.02556551e-01 1.64653671e+00 2.35426679e-01
-5.22233367e-01 -1.26890838e-01 -7.47809291e-01 7.87473381e-01
6.89149916e-01 -2.14983389e-01 -5.92191696e-01 1.12049542e-01
4.98940736e-01 1.93170965e-01 -1.12302778e-02 -1.19541979e+00
7.89062798e-01 -4.61344272e-01 6.43173695e-01 -6.04995906e-01
-1.66142464e-01 -1.10265505e+00 6.99367598e-02 4.08618838e-01
-2.33282894e-01 2.18094781e-01 5.13539672e-01 1.78438932e-01
-7.00876296e-01 -1.60736918e-01 3.75176460e-01 -1.90619409e-01
-8.60823452e-01 -2.74239182e-01 -5.08149803e-01 -1.78854719e-01
1.33298016e+00 -2.29575455e-01 -4.11635369e-01 -2.51543909e-01
-7.49028921e-01 3.07760030e-01 -9.62471939e-04 3.84776831e-01
6.76069736e-01 -1.04926133e+00 -3.26198712e-02 2.39407793e-01
8.97080451e-02 1.37545496e-01 5.38897336e-01 5.10219574e-01
-7.74496436e-01 5.47486424e-01 -7.85478950e-01 -1.25943094e-01
-1.92424536e+00 3.88217241e-01 -3.52130741e-01 -4.90277946e-01
-2.91436195e-01 5.32325983e-01 -6.21181577e-02 7.48909218e-03
5.89864373e-01 -9.77718383e-02 -1.01013446e+00 3.54503363e-01
5.55037558e-01 7.15677619e-01 9.17136967e-02 -3.10630679e-01
-1.56206906e-01 4.37497318e-01 2.32316591e-02 2.22464815e-01
1.11831689e+00 -8.15215614e-03 8.87294263e-02 3.14823240e-01
4.29050475e-01 3.04535210e-01 -3.97249609e-01 -2.65253246e-01
4.34718221e-01 -3.78788143e-01 5.28932363e-02 -1.01413679e+00
-8.51888299e-01 1.06442249e+00 5.60280800e-01 2.74594069e-01
1.16095996e+00 -1.91412270e-01 -6.38951585e-02 3.90490115e-01
2.83110198e-02 -8.46016228e-01 9.23826918e-02 1.60459563e-01
3.84675294e-01 -9.00036752e-01 4.58650440e-01 -7.47321904e-01
-3.60321432e-01 1.48074496e+00 8.08669448e-01 6.31781161e-01
5.72679341e-01 1.47710621e-01 1.45344436e-01 -6.43050253e-01
-3.97074729e-01 -3.05668294e-01 7.27332473e-01 2.96256393e-01
7.97981739e-01 4.30649752e-03 -6.30674601e-01 2.77647883e-01
-3.25385600e-01 1.11388318e-01 5.73273063e-01 1.26336718e+00
-7.40107775e-01 -1.24683928e+00 -5.89472950e-01 1.17742941e-01
-3.84402186e-01 3.30269895e-02 -5.75228930e-01 1.05695057e+00
2.22628102e-01 8.24435592e-01 -2.86991984e-01 -7.86679462e-02
4.26187217e-01 3.36322695e-01 4.76957828e-01 -3.70269626e-01
-4.49072123e-01 -3.93273175e-01 -3.51812810e-01 -2.35084295e-02
-4.00844425e-01 -1.63759306e-01 -1.44996977e+00 -4.66514140e-01
-4.12120260e-02 5.30737221e-01 1.04816675e+00 8.74468684e-01
3.96801293e-01 9.30595577e-01 -9.12796557e-02 -1.28533661e-01
1.85257956e-01 -5.22351623e-01 -4.43931818e-01 9.72015560e-02
-3.51769328e-01 -7.32957661e-01 4.45100926e-02 1.06174443e-02] | [9.664751052856445, 7.790193557739258] |
ec144979-77dc-44f9-8b01-d3ad88c705c3 | sparse-representation-based-multi-sensor | 1702.03515 | null | http://arxiv.org/abs/1702.03515v1 | http://arxiv.org/pdf/1702.03515v1.pdf | Sparse Representation based Multi-sensor Image Fusion: A Review | As a result of several successful applications in computer vision and image
processing, sparse representation (SR) has attracted significant attention in
multi-sensor image fusion. Unlike the traditional multiscale transforms (MSTs)
that presume the basis functions, SR learns an over-complete dictionary from a
set of training images for image fusion, and it achieves more stable and
meaningful representations of the source images. By doing so, the SR-based
fusion methods generally outperform the traditional MST-based image fusion
methods in both subjective and objective tests. In addition, they are less
susceptible to mis-registration among the source images, thus facilitating the
practical applications. This survey paper proposes a systematic review of the
SR-based multi-sensor image fusion literature, highlighting the pros and cons
of each category of approaches. Specifically, we start by performing a
theoretical investigation of the entire system from three key algorithmic
aspects, (1) sparse representation models; (2) dictionary learning methods; and
(3) activity levels and fusion rules. Subsequently, we show how the existing
works address these scientific problems and design the appropriate fusion rules
for each application, such as multi-focus image fusion and multi-modality
(e.g., infrared and visible) image fusion. At last, we carry out some
experiments to evaluate the impact of these three algorithmic components on the
fusion performance when dealing with different applications. This article is
expected to serve as a tutorial and source of reference for researchers
preparing to enter the field or who desire to employ the sparse representation
theory in other fields. | ['DaCheng Tao', 'Yi Liu', 'Jungong Han', 'Qiang Zhang', 'Rick S. Blum'] | 2017-02-12 | null | null | null | null | ['infrared-and-visible-image-fusion'] | ['computer-vision'] | [ 6.00400329e-01 -5.93776047e-01 -2.28818655e-01 -1.86222732e-01
-8.35931361e-01 -1.62142083e-01 3.25375825e-01 1.16959594e-01
-3.32313366e-02 5.05719304e-01 1.11835755e-01 1.40753284e-01
-2.56605536e-01 -7.63994277e-01 -2.68598467e-01 -1.02858663e+00
3.94341499e-02 -2.29748547e-01 1.09664783e-01 -3.66666555e-01
1.83973640e-01 5.38054049e-01 -1.93188596e+00 9.90365744e-02
9.94032145e-01 1.18633819e+00 4.91076589e-01 3.28753442e-01
1.27622075e-02 7.91154623e-01 -5.09846807e-01 -7.49024749e-02
3.33598137e-01 -6.90802813e-01 -2.64125049e-01 4.08365101e-01
2.92265236e-01 1.31343693e-01 -4.38713014e-01 1.49153364e+00
6.26980722e-01 3.39388698e-01 4.40293461e-01 -1.05071521e+00
-8.33143115e-01 3.56558144e-01 -9.87307906e-01 4.33045030e-01
6.54782355e-01 -2.82938302e-01 4.70532894e-01 -1.04608774e+00
1.36312842e-01 1.13631928e+00 6.74871802e-01 2.70982436e-03
-9.52223659e-01 -5.85575461e-01 -1.03446484e-01 3.99605185e-01
-1.76327956e+00 -6.03763282e-01 9.76755559e-01 -2.15725794e-01
4.65273082e-01 4.45028365e-01 5.68656802e-01 4.31165069e-01
4.87792879e-01 6.72653019e-01 1.36549425e+00 -6.80755258e-01
2.01718524e-01 -1.40897542e-01 1.82790384e-01 6.09472454e-01
3.58623087e-01 2.56360024e-01 -6.79830611e-01 -2.93715686e-01
8.55712891e-01 4.49800491e-01 -5.80976486e-01 -1.74399778e-01
-1.39565027e+00 8.98104191e-01 4.63078111e-01 8.24948192e-01
-7.10923553e-01 -2.30208799e-01 5.27876504e-02 3.87787789e-01
2.77325481e-01 -1.29897855e-02 1.08434230e-01 4.78660882e-01
-1.16017485e+00 4.10143025e-02 3.73720944e-01 6.61935270e-01
8.97687137e-01 3.89374435e-01 7.17816129e-02 9.74893987e-01
4.34521556e-01 9.31076527e-01 5.37507892e-01 -1.05235350e+00
5.20167202e-02 3.06372195e-01 -7.99421743e-02 -1.36211896e+00
-2.09279627e-01 -2.87235767e-01 -1.24869108e+00 1.93247199e-01
-1.99179694e-01 8.50339457e-02 -7.72707701e-01 1.40538323e+00
2.14273170e-01 4.35571909e-01 3.24598819e-01 9.10001278e-01
1.23360980e+00 7.61728227e-01 -1.61433503e-01 -8.12807798e-01
1.48414946e+00 -7.52433598e-01 -1.26248109e+00 -1.18577220e-01
2.56105252e-02 -1.26972139e+00 3.22372019e-01 3.71768266e-01
-1.19031155e+00 -9.56168175e-01 -1.14160240e+00 2.54079640e-01
-2.72956163e-01 1.13881432e-01 6.19658053e-01 5.26745498e-01
-1.16469014e+00 2.18450040e-01 -8.35666656e-01 -5.53363502e-01
4.19564098e-02 3.19971919e-01 -3.32460165e-01 -4.59477335e-01
-1.16235375e+00 1.13120604e+00 2.34790996e-01 1.69353355e-02
-3.75936180e-01 -4.34973717e-01 -1.14327729e+00 -2.28517115e-01
1.61509216e-01 -7.36610234e-01 7.55678713e-01 -9.40496802e-01
-1.05222714e+00 6.76103473e-01 -6.39380872e-01 -2.60653824e-01
-2.86292166e-01 4.56428155e-02 -8.92634571e-01 4.42205191e-01
1.62278593e-01 3.63115460e-01 1.13310242e+00 -1.49097800e+00
-7.76568234e-01 -3.13421279e-01 -1.17736951e-01 3.28461677e-01
-7.34807700e-02 1.24727733e-01 -1.27024859e-01 -1.04809892e+00
6.07028306e-01 -5.64766765e-01 -4.19915497e-01 -1.78084910e-01
5.22434004e-02 6.25430793e-02 1.05268168e+00 -6.29350722e-01
1.25545430e+00 -2.47768593e+00 3.25620621e-01 1.97739676e-01
1.94015130e-01 2.08261997e-01 -1.26357347e-01 5.14315248e-01
-1.69469506e-01 -4.53976512e-01 -4.48693365e-01 -2.21442282e-01
-6.08818233e-01 2.31846675e-01 -2.86077172e-01 7.85173833e-01
-2.63246417e-01 5.81567109e-01 -8.67169321e-01 -7.05287516e-01
7.70211518e-01 7.23147750e-01 8.43392909e-02 5.87447062e-02
6.54155672e-01 6.79721892e-01 -4.29199666e-01 1.00092947e+00
7.90561140e-01 -2.22490355e-01 -1.66448385e-01 -8.05204034e-01
-2.06552863e-01 -4.18060660e-01 -1.45891213e+00 1.79226708e+00
-2.87556529e-01 3.41724664e-01 3.69668603e-01 -1.29968190e+00
1.04538190e+00 6.17051244e-01 1.08690643e+00 -7.25799143e-01
1.75689504e-01 3.18570644e-01 -1.78793311e-01 -3.39117497e-01
6.11931026e-01 -3.88340414e-01 1.15276091e-01 2.34154493e-01
1.91307724e-01 -2.69956738e-01 2.82187760e-01 1.41892523e-01
6.92979574e-01 -5.68706036e-01 8.40386808e-01 -2.58274913e-01
9.73864913e-01 -1.25769451e-01 5.61150193e-01 5.83048582e-01
-1.69308335e-01 6.87449455e-01 -7.27829933e-01 -3.81905735e-01
-5.16928315e-01 -1.13574004e+00 -1.83645487e-01 6.08651996e-01
7.19167233e-01 -3.58847529e-01 -2.50327677e-01 5.70115671e-02
-8.61312896e-02 3.13676894e-01 -2.56950945e-01 -2.49517426e-01
-3.61839145e-01 -8.86957407e-01 1.13179065e-01 2.52639472e-01
7.03937590e-01 -7.83073366e-01 -5.41979432e-01 1.13110371e-01
-5.56481779e-01 -1.09854794e+00 -3.21412116e-01 -1.55708313e-01
-9.34670269e-01 -1.17337430e+00 -1.02233374e+00 -7.89229274e-01
6.86434388e-01 1.31052351e+00 7.77136207e-01 2.71618426e-01
-1.83785930e-02 7.98839271e-01 -6.95789695e-01 -4.13172096e-01
-1.77294865e-01 -4.70601112e-01 2.23052800e-01 4.27495956e-01
4.43289161e-01 -6.72670305e-01 -4.68406677e-01 3.67869854e-01
-1.23313737e+00 -2.47397110e-01 5.69997370e-01 6.92127824e-01
9.38395798e-01 5.68407953e-01 4.35751110e-01 -1.42311841e-01
7.02402174e-01 -4.67730612e-01 -5.18431604e-01 4.79004949e-01
-3.89075398e-01 -4.08311546e-01 2.65521944e-01 -3.04295450e-01
-1.00382864e+00 1.04238823e-01 3.40208262e-02 -5.77498198e-01
5.20842373e-02 7.14789927e-01 6.92901313e-02 -8.67294312e-01
6.25665367e-01 6.89857662e-01 2.87891209e-01 -3.60445350e-01
1.94869891e-01 6.04500592e-01 7.69370794e-01 -4.03085709e-01
9.85151887e-01 7.81887293e-01 1.06331043e-01 -1.14409339e+00
-8.17363441e-01 -7.71918416e-01 -3.82611901e-01 -3.12316597e-01
7.58901834e-01 -1.20229065e+00 -3.08592260e-01 4.92732137e-01
-9.51970994e-01 4.82786089e-01 -3.94487947e-01 8.26727450e-01
-5.44685543e-01 8.66487086e-01 -3.50829542e-01 -8.61108720e-01
-2.59873986e-01 -1.37351465e+00 1.12257552e+00 6.82075024e-01
7.90114701e-02 -1.11730909e+00 4.77634510e-03 4.29853946e-01
5.81628859e-01 2.73146987e-01 3.14154208e-01 -1.15914315e-01
-3.25146377e-01 -1.34031802e-01 -2.99536041e-03 3.96376580e-01
6.61661506e-01 -2.65960097e-01 -7.96517789e-01 -5.51458001e-01
6.66570842e-01 4.02559079e-02 8.41930687e-01 8.03654194e-01
8.32844794e-01 1.17904663e-01 -3.53572726e-01 6.09028161e-01
1.56668663e+00 4.38643605e-01 6.46335244e-01 4.78655882e-02
4.40763831e-01 4.11277026e-01 8.02986085e-01 3.36699575e-01
3.37150514e-01 6.46562159e-01 2.68928677e-01 -4.28816557e-01
-3.10783267e-01 2.11864725e-01 4.22418922e-01 1.12756455e+00
-3.55215073e-01 5.85791394e-02 -4.83037710e-01 4.34706539e-01
-1.86982858e+00 -1.07177603e+00 -4.14904691e-02 2.21703482e+00
4.23838407e-01 -6.15608573e-01 -8.82530361e-02 4.36720908e-01
1.04282570e+00 4.64149624e-01 -3.67112041e-01 7.62332678e-02
-6.73830926e-01 3.78970027e-01 4.19763863e-01 5.08312404e-01
-1.11135221e+00 5.39943397e-01 6.99862289e+00 9.15231347e-01
-1.10085011e+00 3.30189437e-01 1.74282730e-01 2.50136793e-01
-2.78904200e-01 2.84006383e-04 -4.43989635e-01 3.70324850e-01
4.91084039e-01 -3.29035342e-01 5.19714773e-01 3.58010381e-01
8.96001980e-02 -4.76974696e-01 -5.88276982e-01 1.62758601e+00
4.01100695e-01 -1.36136055e+00 7.02226236e-02 -8.24732892e-03
9.38421905e-01 7.73738325e-02 -7.63351098e-02 -2.42172390e-01
1.61876786e-03 -8.45233023e-01 6.05032504e-01 6.31259918e-01
5.89507699e-01 -5.32824934e-01 9.40948904e-01 2.88807899e-01
-1.59932876e+00 -1.09852970e-01 -3.98311883e-01 -5.56745008e-03
5.58097005e-01 1.00885677e+00 3.92958343e-01 1.18118703e+00
6.74075186e-01 9.09450471e-01 -2.45471835e-01 1.09481370e+00
-1.00017246e-02 1.35998040e-01 -2.42454931e-01 6.00097656e-01
-9.32473168e-02 -4.76028949e-01 7.33458996e-01 9.05340850e-01
6.99783146e-01 4.39960808e-01 4.82347071e-01 4.51786608e-01
4.17226940e-01 -4.10373472e-02 -7.22983003e-01 1.42445832e-01
6.57109201e-01 1.23634207e+00 -7.21240759e-01 -4.27679837e-01
-8.05917203e-01 6.49122059e-01 -4.57394034e-01 5.17910957e-01
-4.72454846e-01 -2.58034527e-01 5.98815441e-01 -2.71255702e-01
3.74637783e-01 -3.35576862e-01 -4.23149556e-01 -1.39321804e+00
-1.49581060e-01 -1.09345579e+00 5.86825728e-01 -8.07494521e-01
-1.25792098e+00 6.00043178e-01 2.54946351e-01 -1.47377920e+00
-4.69461270e-02 -1.37452170e-01 -4.79614496e-01 9.30745661e-01
-1.77134752e+00 -1.10813570e+00 -4.95121837e-01 1.06414855e+00
4.14145678e-01 -4.23609704e-01 7.33896494e-01 4.31108356e-01
-2.28723437e-01 2.08623409e-01 1.37688026e-01 -2.22043425e-01
5.07790804e-01 -5.56320667e-01 -2.81130642e-01 1.23401487e+00
2.74105906e-01 6.31579161e-01 7.61786461e-01 -5.38808823e-01
-1.60813439e+00 -6.73755229e-01 6.37305737e-01 5.58893681e-02
3.96104187e-01 3.43774408e-01 -8.48700702e-01 4.96913195e-01
3.84913683e-01 2.15016529e-01 8.69985163e-01 -3.35028559e-01
-1.15905710e-01 -3.60849857e-01 -1.43869853e+00 1.93064243e-01
7.54207909e-01 -4.46312487e-01 -6.93991482e-01 2.34321043e-01
3.33080202e-01 -3.43151182e-01 -1.24766314e+00 6.99908197e-01
3.10050726e-01 -1.08886766e+00 1.33021438e+00 2.78799921e-01
-5.94594106e-02 -7.82374263e-01 -7.27632582e-01 -1.20025611e+00
-6.18521690e-01 -4.44932640e-01 -3.20053369e-01 1.04986668e+00
-3.16447139e-01 -7.88693368e-01 9.75754783e-02 -4.56300229e-02
-1.00660950e-01 -5.84889948e-01 -9.78965282e-01 -6.67938113e-01
-5.80413520e-01 -1.55944228e-01 5.95704079e-01 9.85881627e-01
-2.33553693e-01 1.17054909e-01 -4.86082643e-01 4.35108870e-01
1.13800979e+00 3.64415318e-01 4.75716889e-01 -1.08267570e+00
4.90778796e-02 -3.10828298e-01 -6.32240295e-01 -9.04020607e-01
-1.36265427e-01 -5.79646528e-01 -1.72813222e-01 -1.55259275e+00
1.42018124e-01 -3.02742362e-01 -5.19269586e-01 1.84323996e-01
-2.29393557e-01 4.62105781e-01 4.58161503e-01 6.74161851e-01
-4.68589097e-01 5.54733753e-01 1.21776605e+00 -2.49815047e-01
7.43086785e-02 6.78261742e-02 -1.06192172e+00 6.52606130e-01
7.14650273e-01 -2.58789897e-01 -5.32968044e-01 -3.70039582e-01
-2.15814635e-01 3.07734013e-01 3.16275299e-01 -1.18279636e+00
4.80926812e-01 -3.03391278e-01 2.07964465e-01 -6.92539215e-01
5.53963900e-01 -1.17307115e+00 5.14514506e-01 4.93573070e-01
2.04212919e-01 1.36787236e-01 -2.99403369e-02 4.61470008e-01
-8.36765587e-01 -2.44836763e-01 1.06626737e+00 -3.32063615e-01
-9.84491110e-01 2.11267009e-01 -3.58152598e-01 -4.92687136e-01
1.10327971e+00 -6.22095525e-01 -3.25030349e-02 -5.06629407e-01
-6.01142406e-01 2.39831489e-02 4.82898831e-01 2.61593372e-01
9.37365592e-01 -1.54938245e+00 -7.31168270e-01 4.57969815e-01
5.63021526e-02 -1.85771286e-01 4.50651318e-01 1.10383570e+00
-1.85250252e-01 2.31647223e-01 -2.16631398e-01 -7.44498372e-01
-1.29935169e+00 7.15703428e-01 2.31117249e-01 -1.73586030e-02
-4.63793874e-01 5.17101526e-01 2.58173823e-01 1.19938022e-02
2.49784850e-02 -1.22198962e-01 -3.35575014e-01 -9.53643993e-02
9.34149623e-01 4.91183698e-01 4.29350175e-02 -1.20303679e+00
-4.27775264e-01 1.10022008e+00 2.82593429e-01 8.48752856e-02
1.20695317e+00 -4.76996243e-01 -4.39216465e-01 3.40125084e-01
9.38547075e-01 -8.96231830e-02 -7.89733112e-01 -5.22671342e-01
-3.78820390e-01 -6.51835859e-01 3.34057391e-01 -4.03262258e-01
-1.50651765e+00 4.65303808e-01 8.38242292e-01 3.62214267e-01
1.78875208e+00 1.14369623e-01 7.88414240e-01 2.07582340e-02
6.82460308e-01 -6.99967027e-01 -9.47648212e-02 1.86686829e-01
9.19220865e-01 -1.23063207e+00 3.18575919e-01 -5.91272354e-01
-3.62456948e-01 1.04109919e+00 7.96728358e-02 -1.26607239e-01
9.56282496e-01 3.13749671e-01 1.58511013e-01 -3.36763918e-01
-1.91703334e-01 -3.31869841e-01 4.55787063e-01 8.70702386e-01
3.73587698e-01 -7.32843354e-02 -5.18876612e-01 1.07898138e-01
3.53981033e-02 1.92701146e-01 2.78408468e-01 1.15840006e+00
-7.19368160e-01 -1.09087408e+00 -1.17675745e+00 3.68861526e-01
-3.16091746e-01 7.29999393e-02 9.47048962e-02 3.78228575e-01
2.25022733e-01 1.40471733e+00 -2.54028857e-01 -4.30903018e-01
2.47355461e-01 -4.34438974e-01 6.38242841e-01 -3.58843237e-01
-2.73093522e-01 2.01689303e-01 -4.54758674e-01 -5.01918972e-01
-1.47925782e+00 -8.10237348e-01 -1.03107238e+00 -3.90535802e-01
-5.06068885e-01 3.24715763e-01 5.42400062e-01 9.64065433e-01
2.39404261e-01 3.54675442e-01 7.22402632e-01 -1.06662154e+00
-2.51742452e-01 -7.66775250e-01 -8.74637783e-01 3.14097822e-01
4.40474421e-01 -9.71891940e-01 -2.25289986e-01 2.21718982e-01] | [10.592103958129883, -1.8633339405059814] |
aa3a2f3c-f705-44b6-a50a-ee15ee36acb9 | mastering-2048-with-delayed-temporal | 1604.05085 | null | http://arxiv.org/abs/1604.05085v3 | http://arxiv.org/pdf/1604.05085v3.pdf | Mastering 2048 with Delayed Temporal Coherence Learning, Multi-Stage Weight Promotion, Redundant Encoding and Carousel Shaping | 2048 is an engaging single-player, nondeterministic video puzzle game, which,
thanks to the simple rules and hard-to-master gameplay, has gained massive
popularity in recent years. As 2048 can be conveniently embedded into the
discrete-state Markov decision processes framework, we treat it as a testbed
for evaluating existing and new methods in reinforcement learning. With the aim
to develop a strong 2048 playing program, we employ temporal difference
learning with systematic n-tuple networks. We show that this basic method can
be significantly improved with temporal coherence learning, multi-stage
function approximator with weight promotion, carousel shaping, and redundant
encoding. In addition, we demonstrate how to take advantage of the
characteristics of the n-tuple network, to improve the algorithmic
effectiveness of the learning process by i) delaying the (decayed) update and
applying lock-free optimistic parallelism to effortlessly make advantage of
multiple CPU cores. This way, we were able to develop the best known 2048
playing program to date, which confirms the effectiveness of the introduced
methods for discrete-state Markov decision problems. | ['Wojciech Jaśkowski'] | 2016-04-18 | null | null | null | null | ['2048'] | ['playing-games'] | [-5.66050150e-02 -5.44425324e-02 -2.62047559e-01 1.80769011e-01
-5.85676849e-01 -4.87147778e-01 6.14556849e-01 1.22742601e-01
-7.81858861e-01 8.97788167e-01 -2.04407349e-01 -6.63678110e-01
-4.06978548e-01 -8.23859632e-01 -4.97731745e-01 -8.02543223e-01
-8.10488582e-01 8.30300987e-01 8.61251235e-01 -3.62023294e-01
2.64955074e-01 3.16180229e-01 -1.75678277e+00 2.58610278e-01
3.70018482e-01 9.84333217e-01 -1.13296285e-01 1.02303565e+00
5.28217629e-02 1.30395377e+00 -4.36657488e-01 -1.20991841e-01
4.47065920e-01 -3.84361744e-01 -8.89602005e-01 -3.19027692e-01
-5.68393648e-01 -6.27459109e-01 -2.79231161e-01 5.70925772e-01
5.77134788e-01 1.39999747e-01 2.20876798e-01 -1.35962105e+00
6.45193636e-01 9.03774798e-01 -5.15380800e-01 6.86126426e-02
1.20246418e-01 6.06965482e-01 9.07293022e-01 -5.31474091e-02
5.33204436e-01 1.01291263e+00 6.02919042e-01 5.99880397e-01
-1.26304138e+00 -5.56115687e-01 -2.96047200e-02 4.18238312e-01
-1.01344907e+00 -3.28416526e-01 4.57454950e-01 4.12384085e-02
1.18154359e+00 2.08464414e-01 9.01250303e-01 8.87760103e-01
5.24989069e-01 8.72988164e-01 1.52461302e+00 -6.55602038e-01
7.85837710e-01 -4.40271884e-01 -6.42312765e-02 8.08066547e-01
2.87058353e-02 5.32151341e-01 -7.29878366e-01 -4.46560681e-01
7.39944458e-01 -3.14909786e-01 1.92398161e-01 -5.69823742e-01
-1.05588555e+00 6.21294975e-01 -2.52874911e-01 3.39107513e-02
-4.23171967e-01 8.67800534e-01 8.34436774e-01 5.95631957e-01
1.07697479e-01 2.57489145e-01 -4.33922201e-01 -9.86515999e-01
-9.11034524e-01 5.45832515e-01 1.18827605e+00 4.04674381e-01
7.00376332e-01 1.11451879e-01 -1.03407979e-01 4.51118276e-02
9.75164175e-02 2.46585160e-01 3.74444544e-01 -1.57019126e+00
2.08694905e-01 6.04469851e-02 2.51569390e-01 -4.12824154e-01
-5.02360344e-01 -9.64816287e-02 -8.14433634e-01 9.14481938e-01
6.36922717e-01 -4.01905984e-01 -3.65235388e-01 1.65695989e+00
3.77007484e-01 4.04750288e-01 7.88767412e-02 4.93743002e-01
-3.47230285e-01 6.90583646e-01 -1.32275820e-01 -4.53800648e-01
1.14149773e+00 -7.24399269e-01 -2.49204025e-01 1.72694072e-01
7.67542839e-01 -3.69372874e-01 6.46924555e-01 9.22300816e-01
-1.28686738e+00 -1.14196286e-01 -1.13268030e+00 4.75773543e-01
2.25344390e-01 -4.43463624e-01 1.09699321e+00 8.09937358e-01
-1.24474990e+00 1.07616639e+00 -1.37106836e+00 1.35802329e-01
4.84479442e-02 6.59920990e-01 -6.71614558e-02 2.14399129e-01
-1.20903790e+00 7.10330665e-01 6.77471995e-01 -1.16018571e-01
-1.10086429e+00 -4.90695864e-01 -2.12857768e-01 8.89645889e-02
9.81067121e-01 -7.82609642e-01 1.61014843e+00 -7.20943093e-01
-2.26665735e+00 4.55146402e-01 4.48025405e-01 -8.17168713e-01
6.71294749e-01 2.36604556e-01 1.87039882e-01 3.34993154e-01
-3.16573769e-01 3.45164865e-01 9.69378829e-01 -7.36770928e-01
-8.34580123e-01 -3.64552081e-01 3.86478096e-01 2.18781546e-01
2.75655668e-02 -3.76310162e-02 -3.74464132e-02 -1.08386397e-01
-1.15066491e-01 -1.24408603e+00 -6.81505382e-01 -3.48334849e-01
2.39367083e-01 -3.04100782e-01 3.57365578e-01 -1.66605353e-01
9.95328128e-01 -1.97622299e+00 1.49742678e-01 3.22302490e-01
2.48873129e-01 2.05534115e-01 2.11526360e-02 7.29900777e-01
1.51776865e-01 -2.54229635e-01 8.29720870e-02 -3.00791323e-01
2.43903756e-01 6.01338208e-01 -2.83779383e-01 4.85667855e-01
-1.16002627e-01 6.58245385e-01 -1.02942693e+00 -4.09960568e-01
3.39302458e-02 -2.09307715e-01 -6.36785626e-01 2.00051531e-01
-5.96080601e-01 9.22545940e-02 -4.74481910e-01 8.19864273e-02
2.89058685e-01 4.73043732e-02 6.15562379e-01 5.73624671e-01
-4.34340507e-01 3.86682510e-01 -1.60003936e+00 1.71963704e+00
-4.80134547e-01 2.18082458e-01 2.67740279e-01 -9.37047601e-01
5.46603262e-01 3.44000369e-01 6.13326371e-01 -8.13114166e-01
4.43046652e-02 3.41337055e-01 1.83009505e-02 -1.03689626e-01
6.21613920e-01 -2.29870245e-01 -2.89007485e-01 8.83237898e-01
-9.76765305e-02 -3.39911401e-01 4.03413922e-01 1.52330607e-01
1.49619424e+00 3.64293128e-01 2.67799795e-01 -2.12315187e-01
3.22434574e-01 2.61138398e-02 5.84579289e-01 1.09158993e+00
-9.76407379e-02 -1.59200773e-01 1.25073898e+00 -4.71118480e-01
-9.96576130e-01 -7.24838793e-01 5.43524683e-01 1.32075047e+00
-1.85289353e-01 -6.55831873e-01 -6.64161980e-01 -3.39788496e-01
-2.58309215e-01 4.92878586e-01 -4.51180696e-01 -7.36268088e-02
-8.12157094e-01 -5.41162789e-01 9.09949720e-01 3.80515069e-01
4.46399957e-01 -1.12713397e+00 -1.07781351e+00 8.61386836e-01
3.32520634e-01 -5.57050288e-01 1.47247240e-01 7.83671498e-01
-9.55290198e-01 -9.92754519e-01 -1.98536709e-01 -3.50626737e-01
-2.83100992e-01 1.32633895e-01 9.74711835e-01 -1.50413260e-01
1.20925151e-01 4.35436338e-01 -1.56287640e-01 7.41299912e-02
-6.78152740e-01 2.19942644e-01 6.04493590e-03 -4.05470431e-01
-2.47367501e-01 -1.04556704e+00 -4.21027452e-01 6.59230426e-02
-9.83476758e-01 1.94047377e-01 4.33980763e-01 1.08709085e+00
3.01313937e-01 2.87261546e-01 1.57487959e-01 -7.14856088e-01
6.76064491e-01 -1.33286625e-01 -1.06547809e+00 -6.26885742e-02
-7.96644807e-01 5.33065319e-01 7.91698337e-01 -2.54749268e-01
-9.13897514e-01 -1.59687046e-02 -2.59051740e-01 -2.24347226e-02
2.13775530e-01 4.38449681e-01 3.04943889e-01 -1.39512166e-01
5.64631641e-01 3.69980574e-01 4.50202256e-01 8.93230066e-02
4.99635279e-01 4.05138910e-01 3.24820012e-01 -1.24275696e+00
2.28026837e-01 4.00720388e-01 5.75401366e-01 -3.22300285e-01
-4.66054901e-02 -2.44788811e-01 -5.28854914e-02 -2.70918667e-01
2.55626202e-01 -7.33111978e-01 -1.84720826e+00 9.26294208e-01
-1.03901267e+00 -9.47252691e-01 -4.61801589e-01 1.94761723e-01
-1.15509665e+00 4.03515875e-01 -1.17471659e+00 -1.10444582e+00
-1.84234142e-01 -1.09831822e+00 5.73631406e-01 1.99000370e-02
-4.58507985e-03 -5.59863746e-01 4.91977632e-01 -4.57678549e-02
4.29498047e-01 3.61702554e-02 9.70453680e-01 -3.85771960e-01
-7.77516305e-01 1.35682926e-01 2.26372972e-01 2.40903005e-01
-4.26328987e-01 1.23940885e-01 -6.72087371e-01 -3.58397543e-01
2.59953320e-01 -4.44462210e-01 5.52218258e-01 6.48520812e-02
8.97751689e-01 -1.29113242e-01 1.59760758e-01 2.11553812e-01
1.40419137e+00 2.37341538e-01 7.03109324e-01 6.41385376e-01
1.08249374e-01 4.73565668e-01 6.21760666e-01 1.04835629e+00
3.09533447e-01 7.96702743e-01 6.16219878e-01 6.01076901e-01
2.58670747e-01 -9.58717242e-02 5.79167008e-01 8.37551475e-01
-5.37356257e-01 1.75355554e-01 -1.03129923e+00 1.73173398e-01
-2.13162088e+00 -1.00995195e+00 1.51252402e-02 2.36375737e+00
1.04088855e+00 6.71968758e-01 4.18115616e-01 2.63836741e-01
2.00602084e-01 2.05875844e-01 -3.15326214e-01 -6.81257248e-01
2.38894299e-01 8.26570451e-01 7.26647377e-01 3.91903251e-01
-6.43145621e-01 9.13414478e-01 6.03658628e+00 1.34376264e+00
-1.03292966e+00 1.24276750e-01 4.53729600e-01 -6.90089241e-02
-5.87809868e-02 2.70212233e-01 -6.84738696e-01 5.41356206e-01
1.54568756e+00 -1.98349163e-01 9.25084889e-01 8.65717888e-01
3.17948759e-01 -5.56827724e-01 -9.84374464e-01 6.74893260e-01
-7.03853786e-01 -1.47431314e+00 -5.04751086e-01 2.88494714e-02
5.29694080e-01 -9.83937979e-02 -6.19819723e-02 7.45145798e-01
7.32307494e-01 -7.77395725e-01 7.05091476e-01 1.89541265e-01
5.66586673e-01 -1.33579886e+00 6.31039500e-01 6.79703236e-01
-1.02255774e+00 -2.30673939e-01 -7.58019835e-02 -7.90207028e-01
-1.67788230e-02 2.66011655e-01 -5.74351430e-01 5.49458921e-01
4.06811565e-01 1.83533300e-02 -1.79265738e-02 8.48901749e-01
-2.07801372e-01 7.77565539e-01 -6.84686661e-01 -2.66298592e-01
5.91739357e-01 -2.18011722e-01 2.91883737e-01 7.31882215e-01
1.79936886e-01 1.56279400e-01 1.91244572e-01 2.47673348e-01
3.72334808e-01 -1.63745269e-01 -2.48866439e-01 7.40474090e-02
1.93417758e-01 9.85401213e-01 -8.70733321e-01 -2.67741680e-01
-1.00920573e-01 6.73888505e-01 2.72941679e-01 -3.59792146e-03
-8.67082417e-01 -5.92353009e-02 7.23908126e-01 1.43341171e-02
4.83243316e-01 -6.42402053e-01 -1.25630230e-01 -8.17422032e-01
-2.49450713e-01 -1.32923234e+00 1.56226188e-01 -4.75018471e-01
-7.14446068e-01 3.23622316e-01 -5.88943288e-02 -9.37307239e-01
-9.30328846e-01 -4.16980505e-01 -3.42876852e-01 5.04184067e-01
-1.11737287e+00 -6.38709605e-01 3.56488705e-01 6.59315944e-01
2.10056618e-01 4.17708419e-02 1.00757277e+00 -8.05169195e-02
-2.12309703e-01 2.51890838e-01 4.74382013e-01 -4.64759290e-01
3.93317431e-01 -1.30858958e+00 3.97834003e-01 5.93026936e-01
-2.06929937e-01 4.12302345e-01 7.26624787e-01 -2.98301488e-01
-1.83831489e+00 -4.10306543e-01 4.01670307e-01 2.23853573e-01
1.16442597e+00 -3.85591596e-01 -6.49877667e-01 3.33111018e-01
1.00648172e-01 -1.09010763e-01 2.83991158e-01 1.89240411e-01
-1.17029533e-01 -4.03676569e-01 -8.25407684e-01 7.55920291e-01
7.60255277e-01 -3.90252799e-01 -2.92321295e-01 1.22547723e-01
5.71679235e-01 -7.38287091e-01 -6.28457129e-01 6.05429290e-03
4.78720814e-01 -1.26697624e+00 6.18590653e-01 -4.63810503e-01
2.29156807e-01 -2.15928242e-01 1.73427507e-01 -1.07937801e+00
-2.40802214e-01 -1.45460725e+00 -4.34064090e-01 8.16141784e-01
-5.60057573e-02 -6.89041674e-01 1.18437076e+00 4.20170009e-01
6.29637837e-02 -7.76188791e-01 -1.50874984e+00 -8.65262449e-01
1.27180904e-01 -7.13457167e-01 5.28138816e-01 2.56887048e-01
3.13983172e-01 1.53229401e-01 -7.40580678e-01 -2.03336000e-01
5.89637458e-01 1.00944892e-01 7.07343996e-01 -8.50618243e-01
-1.28122318e+00 -3.09805125e-01 -2.77494580e-01 -1.16848242e+00
7.00680241e-02 -3.36936623e-01 2.44029149e-01 -6.81719661e-01
2.72426046e-02 -8.26439559e-01 -3.39703977e-01 4.74792510e-01
1.34485334e-01 -3.15387966e-03 2.91630000e-01 1.22089982e-01
-9.10024226e-01 3.95140231e-01 9.98262882e-01 1.40563443e-01
-5.04262038e-02 5.06828070e-01 -1.35823727e-01 3.90749037e-01
8.66700470e-01 -7.00912893e-01 -4.47908670e-01 9.46275741e-02
7.39128470e-01 8.21836293e-01 1.27663746e-01 -1.10291898e+00
5.28993845e-01 -3.00074428e-01 -4.14113343e-01 -2.68869489e-01
3.18972141e-01 -5.87634087e-01 3.13916147e-01 1.06011164e+00
-2.16512069e-01 2.98464268e-01 3.11628491e-01 4.90343541e-01
3.87745239e-02 -5.16869009e-01 5.17159522e-01 -2.09917650e-01
-9.13672447e-01 1.23976059e-01 -1.11397684e+00 4.87267636e-02
1.11024821e+00 1.29843121e-02 -1.40054032e-01 -3.58866870e-01
-6.50984645e-01 1.44472256e-01 4.67042029e-01 -3.75194132e-01
1.39138550e-01 -7.79008687e-01 -3.98764998e-01 -1.42010838e-01
-3.98639858e-01 -2.61828989e-01 5.28382719e-01 7.80416965e-01
-8.35269868e-01 2.74150401e-01 -5.45136809e-01 -4.04995501e-01
-1.21021402e+00 5.35608947e-01 1.98832437e-01 -1.04979813e+00
-5.06037354e-01 4.68815595e-01 -6.73060358e-01 -5.61576746e-02
5.81595562e-02 -3.50446671e-01 3.23732913e-01 -5.14511056e-02
4.56471980e-01 6.42843664e-01 1.45815983e-01 5.62296867e-01
-2.68472642e-01 -4.18522879e-02 3.55601721e-02 -6.43770039e-01
1.44510710e+00 7.25777820e-02 -3.92375052e-01 4.79456246e-01
4.31718141e-01 -1.62977442e-01 -1.50819802e+00 -6.58917353e-02
1.35607034e-01 -2.64234636e-02 -6.07048944e-02 -3.56015444e-01
-6.00047231e-01 6.18594289e-01 4.23784286e-01 4.40506309e-01
1.00278842e+00 -5.89279234e-01 7.16023028e-01 6.41889036e-01
1.24960005e+00 -1.15831268e+00 4.96241413e-02 8.80181313e-01
6.25682175e-02 -6.73510015e-01 -9.84536111e-03 -6.67058006e-02
-4.14976746e-01 1.40608335e+00 1.71378538e-01 -2.92577654e-01
1.23165071e-01 8.55627477e-01 -3.08604181e-01 2.36918956e-01
-1.36075962e+00 -1.59753427e-01 -7.70815969e-01 5.59106648e-01
-2.32431322e-01 2.02613920e-01 -5.31249106e-01 3.31779957e-01
6.63002059e-02 4.63170618e-01 8.92477393e-01 1.24133599e+00
-5.50205588e-01 -1.76605546e+00 -2.18875170e-01 1.18181951e-01
-4.40923423e-01 -1.13653623e-01 4.24127638e-01 9.56739306e-01
-1.51428804e-01 5.34234822e-01 -1.19895816e-01 -1.89562663e-01
-1.14611350e-01 -6.26823008e-02 9.57920492e-01 -4.07451451e-01
-1.01986563e+00 -1.90266520e-01 2.93056309e-01 -8.39833081e-01
-4.93463352e-02 -6.53236449e-01 -1.08821034e+00 -9.34530079e-01
1.72034979e-01 2.73750991e-01 6.22215807e-01 9.54991043e-01
2.77634442e-01 3.88567120e-01 4.80424553e-01 -8.83091271e-01
-1.33401704e+00 -5.19265056e-01 -8.07058275e-01 -3.48992497e-01
-1.76275417e-01 -4.80145514e-01 -1.06725708e-01 -4.96665567e-01] | [3.6899991035461426, 1.584936261177063] |
137ca74b-49dc-4a30-9bd5-81c487686231 | recognizing-musical-entities-in-user | 1904.00648 | null | http://arxiv.org/abs/1904.00648v1 | http://arxiv.org/pdf/1904.00648v1.pdf | Recognizing Musical Entities in User-generated Content | Recognizing Musical Entities is important for Music Information Retrieval
(MIR) since it can improve the performance of several tasks such as music
recommendation, genre classification or artist similarity. However, most entity
recognition systems in the music domain have concentrated on formal texts (e.g.
artists' biographies, encyclopedic articles, etc.), ignoring rich and noisy
user-generated content. In this work, we present a novel method to recognize
musical entities in Twitter content generated by users following a classical
music radio channel. Our approach takes advantage of both formal radio schedule
and users' tweets to improve entity recognition. We instantiate several machine
learning algorithms to perform entity recognition combining task-specific and
corpus-based features. We also show how to improve recognition results by
jointly considering formal and user-generated content | ['Lorenzo Porcaro', 'Horacio Saggion'] | 2019-04-01 | null | null | null | null | ['genre-classification'] | ['computer-vision'] | [ 1.71872243e-01 -3.04383576e-01 -2.09769413e-01 5.91475479e-02
-1.07522607e+00 -7.84711599e-01 9.06850219e-01 5.44585824e-01
-7.32204318e-01 6.27042413e-01 5.34187734e-01 2.36831248e-01
-5.03750741e-01 -9.33785737e-01 -4.99276370e-01 -2.70757705e-01
-5.49056865e-02 4.16309834e-01 1.71387538e-01 -2.33266801e-01
6.64306819e-01 4.00404751e-01 -1.83597171e+00 4.50523317e-01
5.50659239e-01 7.35664487e-01 1.63509130e-01 7.74613798e-01
-5.21368444e-01 8.07745337e-01 -1.05763471e+00 -4.87648606e-01
-2.45728612e-01 -3.25831771e-01 -8.13485563e-01 -3.25906992e-01
1.52847588e-01 1.88935146e-01 -1.95745021e-01 9.52356994e-01
6.12156987e-01 1.93690032e-01 7.66429424e-01 -1.01547265e+00
-7.44290724e-02 1.44569504e+00 -3.14798832e-01 1.61805805e-02
7.35587001e-01 -7.70539641e-01 1.38649178e+00 -6.92765713e-01
9.06128049e-01 8.03766370e-01 8.32781851e-01 6.62259832e-02
-8.94133985e-01 -9.54036772e-01 -4.06792670e-01 3.79860461e-01
-1.83117652e+00 -3.96232784e-01 9.11653936e-01 -4.96101886e-01
6.22302294e-01 7.02809811e-01 7.28388608e-01 9.19997811e-01
-3.61967295e-01 9.99769270e-01 7.05136180e-01 -6.84495866e-01
-1.07013196e-01 1.11146025e-01 -1.09292306e-01 2.26673052e-01
2.19261833e-02 -4.96837497e-01 -8.70799899e-01 -4.71732199e-01
5.82539916e-01 -2.73709089e-01 9.75341164e-03 1.45319119e-01
-1.53111017e+00 6.43849373e-01 -4.36300077e-02 1.05735803e+00
-3.75921071e-01 -7.99478069e-02 5.52299798e-01 2.77461410e-01
3.05845499e-01 1.04146671e+00 -3.55006158e-01 -4.58280414e-01
-1.39885366e+00 3.33691597e-01 1.17821920e+00 8.30656111e-01
8.67739081e-01 -2.37878531e-01 -3.72549407e-02 1.19045532e+00
-4.92672250e-02 4.34037328e-01 1.02388680e+00 -5.51997364e-01
3.12246501e-01 4.79248941e-01 2.44809724e-02 -1.41866779e+00
-5.99453032e-01 -6.89443171e-01 -5.77333808e-01 -6.69991612e-01
3.44485104e-01 1.90245509e-02 -1.15159780e-01 1.45203781e+00
1.59042507e-01 6.30744517e-01 -6.85640723e-02 4.47469264e-01
1.18570447e+00 4.45824921e-01 3.41891348e-02 -3.87418687e-01
1.35851240e+00 -6.31776571e-01 -6.28220677e-01 4.59215283e-01
6.84418023e-01 -1.33051169e+00 7.03915298e-01 4.51245278e-01
-9.38468158e-01 -5.17691672e-01 -7.99640000e-01 1.34401783e-01
-4.91971910e-01 4.19814676e-01 4.27355438e-01 5.69648385e-01
-3.92572373e-01 9.97197151e-01 -3.20358783e-01 -3.35063070e-01
1.90050885e-01 4.59006667e-01 -1.92429215e-01 5.75028658e-01
-1.17510557e+00 2.53737897e-01 4.59070325e-01 -3.89495999e-01
-1.09230228e-01 -6.96240962e-01 -5.24700582e-01 -9.04366672e-02
2.79695541e-01 -5.40091693e-01 1.35460675e+00 -9.46482301e-01
-1.55606401e+00 9.67351079e-01 3.88306007e-02 -4.03396934e-01
1.93668485e-01 -1.56383529e-01 -8.06001604e-01 -1.39296949e-01
1.88839003e-01 -2.85210758e-02 6.81213021e-01 -5.99813759e-01
-9.43670571e-01 -2.91976184e-01 -9.84817892e-02 1.15286574e-01
-7.24054575e-01 2.67667323e-01 -6.77621782e-01 -1.24640024e+00
2.34329641e-01 -1.10985351e+00 1.70077533e-01 -9.32807207e-01
-6.39125824e-01 -4.51286167e-01 3.22326064e-01 -5.70655107e-01
1.84334219e+00 -2.12544298e+00 5.07787913e-02 6.69034541e-01
-7.44682476e-02 4.08110693e-02 5.44636548e-02 7.38079607e-01
1.83312848e-01 5.10459505e-02 2.88534909e-01 -9.20097381e-02
2.29691729e-01 -7.27661178e-02 -3.09713542e-01 1.07433550e-01
-4.69533056e-01 7.08049178e-01 -9.17084515e-01 -5.66722095e-01
-4.17114981e-03 2.67782837e-01 -3.86099994e-01 -2.71282136e-01
-9.11414996e-02 5.40591180e-01 -7.77767539e-01 3.09797734e-01
3.11615597e-02 -1.27196208e-01 3.07680905e-01 -2.37499639e-01
-3.35927755e-01 6.79716051e-01 -1.59599149e+00 1.81143677e+00
-7.17370450e-01 9.19710755e-01 -1.83922768e-01 -6.57631576e-01
9.34562743e-01 4.51608807e-01 9.90420103e-01 -4.27772462e-01
2.59347230e-01 4.70823407e-01 -3.23943585e-01 -4.60973263e-01
1.18680954e+00 -3.77213000e-03 -4.69749272e-01 6.64750993e-01
-2.23632175e-02 2.06480533e-01 4.36804831e-01 2.11125091e-02
1.07420695e+00 -2.55402088e-01 6.52230859e-01 1.66177437e-01
8.82272184e-01 1.19882701e-02 1.60760880e-01 9.74741280e-01
2.45864019e-01 4.80421841e-01 1.59469508e-02 3.51499058e-02
-9.68977451e-01 -6.39672101e-01 -3.45242918e-01 1.66756999e+00
-4.07858528e-02 -1.06178892e+00 -5.03324449e-01 -3.09365541e-01
6.44442961e-02 3.34483415e-01 -2.58370936e-01 1.63636059e-01
-6.47701025e-01 -5.27664721e-01 1.05686402e+00 1.29164886e-02
1.19075797e-01 -1.35305464e+00 1.41727282e-02 5.88323653e-01
-4.92996603e-01 -8.80187452e-01 -2.49988437e-01 -2.49981388e-01
-5.10544360e-01 -9.66627419e-01 -7.54557788e-01 -7.14505255e-01
-1.45482615e-01 -6.42721578e-02 1.26622117e+00 4.11801524e-02
-3.38556767e-01 6.02590740e-01 -6.83339536e-01 -4.96043384e-01
-4.11558092e-01 8.17295313e-01 1.99908301e-01 1.23550348e-01
3.88064742e-01 -7.51484156e-01 -1.54766306e-01 3.17399293e-01
-9.04877245e-01 -2.31097028e-01 4.36443597e-01 5.21278918e-01
4.51016337e-01 2.57692009e-01 6.86541557e-01 -1.30172801e+00
7.40718305e-01 -4.60131317e-01 -1.64871737e-01 2.24127471e-02
-3.12247574e-01 -8.46348256e-02 3.86300057e-01 -6.90155804e-01
-5.51645041e-01 7.26666152e-02 -5.31922162e-01 9.71655082e-03
-2.64673352e-01 8.96545053e-01 4.26533520e-02 -1.59773733e-02
8.43618035e-01 2.17173710e-01 -7.90346086e-01 -7.87196815e-01
1.63228795e-01 1.09262252e+00 7.01557279e-01 -4.64923650e-01
8.14512491e-01 1.97929472e-01 -1.16415255e-01 -1.24013650e+00
-9.34316158e-01 -1.24518633e+00 -6.02970302e-01 -3.84152472e-01
6.54587090e-01 -8.84876668e-01 -8.73700023e-01 3.16809028e-01
-8.97028267e-01 4.16681409e-01 -4.12343860e-01 8.76352906e-01
-6.02004170e-01 2.89823890e-01 -4.05611038e-01 -9.40528929e-01
-3.34466606e-01 -3.98412317e-01 1.15303266e+00 1.76652849e-01
-6.66277826e-01 -7.30707467e-01 6.21058345e-01 3.46525401e-01
9.99233872e-02 1.05228588e-01 5.43216109e-01 -1.03392088e+00
-3.07896912e-01 -3.95585269e-01 6.50046617e-02 -2.38051638e-01
1.01095224e-02 -9.37629715e-02 -1.08432829e+00 7.81589746e-02
-6.08592451e-01 -1.51076159e-02 7.85087526e-01 -5.97213097e-02
1.07591093e+00 -3.35089535e-01 -3.50015998e-01 3.58474404e-01
1.04757321e+00 -2.05993697e-01 5.48352182e-01 7.74655282e-01
6.65264130e-01 4.84162420e-01 5.91988802e-01 9.72617328e-01
1.90971509e-01 1.07608819e+00 -1.27239421e-01 3.87089878e-01
1.10921308e-01 -4.46560025e-01 2.98039347e-01 1.31815279e+00
-3.99653047e-01 -3.08948439e-02 -7.12426960e-01 5.84672511e-01
-1.92951906e+00 -1.49531615e+00 -3.66444200e-01 2.18689299e+00
9.25289869e-01 -1.49473950e-01 4.92957473e-01 7.33077347e-01
7.03616023e-01 -1.21376915e-02 -5.10387169e-03 1.03003755e-01
-3.34466338e-01 6.04802787e-01 4.61486965e-01 3.69175039e-02
-1.43071365e+00 8.37422371e-01 5.31393337e+00 1.40441179e+00
-8.96765292e-01 1.00660093e-01 -3.83898795e-01 -1.49848968e-01
-1.19987018e-01 -3.71450186e-01 -8.51363778e-01 3.48912150e-01
8.66412699e-01 -3.64986569e-01 4.37120974e-01 6.93503916e-01
-4.94442880e-02 2.48160332e-01 -8.31625164e-01 1.36206722e+00
1.73916414e-01 -1.54129899e+00 -3.79459448e-02 6.51640119e-03
7.73357391e-01 1.36898011e-01 -1.55118123e-01 4.36024755e-01
-1.18635647e-01 -6.42654717e-01 8.77531826e-01 7.46732414e-01
5.64473271e-01 -8.17703903e-01 6.85268402e-01 3.73974770e-01
-1.44272792e+00 8.72572213e-02 -5.52838258e-02 4.64850217e-02
7.03173056e-02 5.73936999e-01 -1.02281904e+00 8.98317754e-01
5.45177341e-01 9.49909031e-01 -5.22815704e-01 1.40471160e+00
7.46098235e-02 6.56078458e-01 -5.73269725e-01 -3.15635085e-01
-2.11336371e-02 -8.71745870e-02 8.43091905e-01 1.41011882e+00
7.30158150e-01 -1.12487510e-01 1.04304124e-02 2.66062230e-01
-3.40760857e-01 1.01458013e+00 -3.92780602e-01 -4.37114447e-01
4.48510110e-01 1.34118271e+00 -7.91351676e-01 -3.23585868e-01
-1.54972151e-01 8.55653524e-01 4.57739545e-04 2.40259059e-02
-4.86227155e-01 -6.29301786e-01 3.45354408e-01 3.24302286e-01
4.00322020e-01 -2.53712505e-01 8.22085738e-02 -1.20622504e+00
-2.69930333e-01 -7.48883188e-01 5.46332240e-01 -5.44092655e-01
-1.33319938e+00 5.32350481e-01 -5.45632839e-01 -1.55171871e+00
-5.17460644e-01 -3.52172703e-01 -3.91682923e-01 3.17711174e-01
-1.12705755e+00 -1.17816710e+00 -3.12509239e-02 8.01180482e-01
1.78095222e-01 -4.96769071e-01 1.01104367e+00 9.51809347e-01
-1.76781312e-01 7.63197243e-01 5.72970152e-01 4.13523674e-01
1.13370049e+00 -1.19137466e+00 -1.45871967e-01 -2.28212588e-02
1.28523076e+00 5.00177622e-01 7.39431441e-01 -5.96283019e-01
-1.18297887e+00 -1.07091522e+00 1.52942657e+00 -4.66699064e-01
1.10758030e+00 1.64983254e-02 -4.33281779e-01 2.82521397e-01
-1.19693734e-01 -8.48534882e-01 1.32966328e+00 8.63109589e-01
-3.16071153e-01 -5.37449718e-02 -5.37061930e-01 5.62616289e-01
1.04572809e+00 -9.72051322e-01 -5.16492844e-01 4.91637349e-01
8.34398195e-02 -2.61139691e-01 -1.16261482e+00 2.32342228e-01
1.04634762e+00 -2.63333291e-01 1.09575713e+00 -5.97637594e-01
9.72983241e-02 -4.21582699e-01 -1.27827674e-01 -9.69196320e-01
-1.43168747e-01 -7.44692087e-01 1.09912030e-01 1.41239095e+00
6.34341657e-01 6.19296217e-03 8.31361353e-01 -8.43857750e-02
7.34466240e-02 1.49988100e-01 -8.08580756e-01 -6.82076097e-01
-4.26198095e-01 -9.70378101e-01 6.98509753e-01 1.35764837e+00
3.48928779e-01 6.97192430e-01 -6.27526283e-01 -1.71838671e-01
2.32008621e-01 6.06103420e-01 1.09221697e+00 -1.73439741e+00
-5.62579513e-01 -6.38661265e-01 -7.58431435e-01 -7.24928319e-01
2.28124425e-01 -1.26989782e+00 -3.04591388e-01 -1.05763841e+00
1.69394702e-01 -7.23416209e-01 -4.73560184e-01 1.98847085e-01
2.66406745e-01 8.85880828e-01 3.08464974e-01 6.26613617e-01
-1.11315286e+00 1.68809935e-01 9.20764208e-01 -1.29054368e-01
-5.56156516e-01 5.68130851e-01 -3.45038414e-01 7.62562811e-01
6.86623394e-01 -7.32167542e-01 2.23575681e-01 -6.94486201e-02
9.56886351e-01 -2.49986708e-01 -1.75776631e-02 -1.17445707e+00
5.84763885e-01 8.16726312e-02 1.25797719e-01 -6.94374681e-01
3.23644996e-01 -5.65182447e-01 4.74181801e-01 3.52208130e-02
-4.41299617e-01 -3.01718861e-01 7.83550367e-02 7.06643343e-01
-5.92395961e-01 -6.65162921e-01 2.30029643e-01 1.64558105e-02
-3.64718109e-01 5.43563403e-02 -6.30958438e-01 -1.94622483e-02
5.05655468e-01 1.57037869e-01 1.48149773e-01 -5.15417755e-01
-1.12036419e+00 -4.84806448e-01 2.97184378e-01 4.12238926e-01
9.82936770e-02 -1.54137468e+00 -7.29790926e-01 -1.41347200e-01
4.27866608e-01 -8.01239014e-01 1.39112234e-01 8.36695969e-01
-3.64518672e-01 4.71996665e-01 -2.11505074e-04 -4.04904515e-01
-1.76352060e+00 2.10525557e-01 -3.02848220e-01 -4.90417659e-01
-3.07669550e-01 6.99209809e-01 -2.54362732e-01 -5.07356882e-01
3.25527906e-01 -5.21094538e-02 -9.03355300e-01 6.79600239e-01
5.93936682e-01 4.62191761e-01 3.00607502e-01 -1.00568748e+00
-7.65909851e-02 5.83725750e-01 2.22061843e-01 -3.57603759e-01
1.30959845e+00 -1.19149938e-01 -1.22402288e-01 8.24892879e-01
9.66564417e-01 8.44128013e-01 1.06844909e-01 -5.05290210e-01
5.40614307e-01 -2.45702684e-01 9.24623311e-02 -5.11287928e-01
-6.62354112e-01 3.62304837e-01 2.11327001e-01 6.33916140e-01
1.01943159e+00 1.08710624e-01 9.39504445e-01 7.81855464e-01
5.77925146e-01 -1.42357409e+00 -2.62803823e-01 7.61323869e-01
5.51540375e-01 -8.87786388e-01 7.82080228e-04 -2.78092146e-01
-2.59656578e-01 1.12969482e+00 -2.25369692e-01 1.73932642e-01
9.39377129e-01 -3.13803069e-02 -5.40719293e-02 -9.09639150e-03
-3.16749990e-01 -6.47771835e-01 8.69703233e-01 7.52691031e-02
6.80091381e-01 2.42314503e-01 -7.16575205e-01 1.13085210e+00
-7.78027892e-01 -4.36591506e-02 4.07372177e-01 7.63822198e-01
-5.73746979e-01 -1.58524144e+00 -4.18447614e-01 6.75267696e-01
-1.01217330e+00 -2.23139286e-01 -7.00327754e-01 5.38272917e-01
2.78223455e-01 8.22809577e-01 -1.41405212e-02 -5.40656447e-01
2.90640324e-01 4.79905196e-02 5.24172068e-01 -6.92436755e-01
-1.11600411e+00 4.69668478e-01 5.32374501e-01 -2.64449418e-01
-9.19143021e-01 -1.02893078e+00 -1.10515440e+00 -4.30673271e-01
-4.78097260e-01 5.33817410e-01 8.63352120e-01 9.30742145e-01
1.85605809e-01 4.72863317e-01 7.55710185e-01 -7.36927629e-01
1.05822206e-01 -1.05970526e+00 -9.96208429e-01 5.50040007e-01
-1.31097764e-01 -5.35106242e-01 -1.15718514e-01 2.43778259e-01] | [15.889097213745117, 5.230820655822754] |
38d80f4e-ede5-4fe9-a3fb-7049c0e1c15d | graph-few-shot-class-incremental-learning | 2112.12819 | null | https://arxiv.org/abs/2112.12819v1 | https://arxiv.org/pdf/2112.12819v1.pdf | Graph Few-shot Class-incremental Learning | The ability to incrementally learn new classes is vital to all real-world artificial intelligence systems. A large portion of high-impact applications like social media, recommendation systems, E-commerce platforms, etc. can be represented by graph models. In this paper, we investigate the challenging yet practical problem, Graph Few-shot Class-incremental (Graph FCL) problem, where the graph model is tasked to classify both newly encountered classes and previously learned classes. Towards that purpose, we put forward a Graph Pseudo Incremental Learning paradigm by sampling tasks recurrently from the base classes, so as to produce an arbitrary number of training episodes for our model to practice the incremental learning skill. Furthermore, we design a Hierarchical-Attention-based Graph Meta-learning framework, HAG-Meta. We present a task-sensitive regularizer calculated from task-level attention and node class prototypes to mitigate overfitting onto either novel or base classes. To employ the topological knowledge, we add a node-level attention module to adjust the prototype representation. Our model not only achieves greater stability of old knowledge consolidation, but also acquires advantageous adaptability to new knowledge with very limited data samples. Extensive experiments on three real-world datasets, including Amazon-clothing, Reddit, and DBLP, show that our framework demonstrates remarkable advantages in comparison with the baseline and other related state-of-the-art methods. | ['Huan Liu', 'Ruocheng Guo', 'Kaize Ding', 'Zhen Tan'] | 2021-12-23 | null | null | null | null | ['few-shot-class-incremental-learning'] | ['methodology'] | [ 2.47781977e-01 1.63950875e-01 -4.28809732e-01 -1.53980345e-01
-7.97981992e-02 -2.36782715e-01 4.43442613e-01 3.47065926e-01
-1.70439169e-01 6.33636236e-01 -1.40536234e-01 -1.59865588e-01
-1.86248437e-01 -1.18259549e+00 -7.41261482e-01 -4.40282404e-01
-1.38975129e-01 4.63692725e-01 7.00695693e-01 -1.70456603e-01
1.60807312e-01 2.32680187e-01 -1.49362910e+00 3.59460920e-01
1.19338405e+00 8.78007233e-01 3.67215753e-01 4.08742756e-01
-3.30916196e-01 9.69856083e-01 -2.72670537e-01 -7.19625831e-01
5.74132986e-02 -2.37720236e-01 -8.30861747e-01 3.84233683e-01
3.39876980e-01 2.56150011e-02 -5.84510565e-01 1.01885068e+00
4.45444405e-01 5.30324340e-01 3.72234881e-01 -1.28945255e+00
-1.06676197e+00 7.95536578e-01 -7.84036636e-01 4.07259852e-01
2.19742417e-01 2.44022086e-01 9.00844634e-01 -9.77188408e-01
7.57786989e-01 1.15353453e+00 7.93273151e-01 3.94387692e-01
-1.12269044e+00 -5.23436844e-01 9.28264976e-01 6.59046531e-01
-1.40558088e+00 -1.83790267e-01 1.17228734e+00 -1.67440355e-01
7.92346954e-01 -2.29018144e-02 9.18724775e-01 8.89892817e-01
1.32650152e-01 8.66975546e-01 6.95846736e-01 -3.15322429e-01
2.53210515e-01 3.84643115e-02 3.00162762e-01 1.02103186e+00
4.97102469e-01 -4.06057060e-01 -3.30942363e-01 -2.16398351e-02
7.09688842e-01 4.51396018e-01 -1.70805529e-01 -7.16988921e-01
-8.99377882e-01 6.78560436e-01 8.44570458e-01 3.19023401e-01
-4.17486727e-01 1.45824850e-01 4.47299182e-01 3.58732164e-01
7.15965688e-01 1.45942956e-01 -4.30198073e-01 2.45630682e-01
-3.91733319e-01 -1.07002117e-01 5.22803485e-01 1.28302109e+00
9.34832633e-01 1.68314695e-01 -3.40484440e-01 1.07897675e+00
1.72199562e-01 4.56014387e-02 6.07736766e-01 -4.26669598e-01
4.17588443e-01 1.17868423e+00 -4.81859863e-01 -1.24582100e+00
-3.62588465e-01 -9.47878718e-01 -9.89858449e-01 -4.27026987e-01
-1.79656863e-01 1.62769631e-01 -1.21643984e+00 1.53301668e+00
6.81551397e-01 7.16130495e-01 -2.00360775e-01 4.71113622e-01
9.40589905e-01 5.64688385e-01 1.28188953e-01 -4.64660525e-01
1.04124486e+00 -1.33809853e+00 -5.05889654e-01 -3.90100569e-01
3.70021194e-01 -1.23347782e-01 1.27107680e+00 2.58766741e-01
-8.19958210e-01 -8.49726319e-01 -9.12554443e-01 1.40277877e-01
-5.62706769e-01 -3.89839977e-01 7.89623976e-01 3.02943915e-01
-8.05586755e-01 6.96152031e-01 -5.54690242e-01 -4.94885147e-01
6.75462723e-01 1.91542402e-01 -8.41714665e-02 -4.22730654e-01
-1.09483302e+00 4.02996391e-01 6.57452106e-01 1.84607416e-01
-7.41963983e-01 -5.25426269e-01 -8.62534583e-01 1.34396210e-01
8.88977468e-01 -8.23385835e-01 9.73256111e-01 -1.03445625e+00
-1.24131966e+00 5.36872387e-01 2.02773418e-02 -1.92873582e-01
1.98076412e-01 1.51587039e-01 -5.59533596e-01 -3.90937962e-02
4.64164019e-02 1.61992535e-01 1.17050719e+00 -1.33806241e+00
-5.65065265e-01 -4.82657880e-01 1.68963879e-01 4.50737238e-01
-6.79929018e-01 -7.05962777e-01 -8.10085952e-01 -7.28952289e-01
1.03188433e-01 -8.13672543e-01 -3.85783672e-01 -2.99197555e-01
-2.59725392e-01 -5.28664291e-01 7.87072957e-01 -3.26472789e-01
1.67634380e+00 -2.01379538e+00 3.12206954e-01 1.87883705e-01
5.80374897e-01 4.76509124e-01 -2.68761516e-01 3.79502624e-01
1.06016971e-01 1.74024682e-02 -2.24467665e-01 -7.60973617e-02
-3.20211887e-01 3.41889173e-01 2.77426876e-02 2.30801813e-02
3.89209436e-03 1.21981609e+00 -1.37172687e+00 -4.73553628e-01
-5.24290279e-03 1.95970982e-01 -4.78726089e-01 4.23393808e-02
-4.28925484e-01 -4.26409878e-02 -3.69688839e-01 7.20508993e-01
5.59203863e-01 -7.00794041e-01 4.92076486e-01 -2.23736688e-01
4.52190071e-01 -3.48655850e-01 -1.02372706e+00 1.86219776e+00
-3.27471048e-01 8.49996284e-02 -3.39226484e-01 -1.21319795e+00
8.67435455e-01 -2.14238554e-01 1.77844733e-01 -7.04285562e-01
1.03527062e-01 -1.53804943e-01 2.91174203e-02 -4.97299969e-01
5.02494037e-01 6.58102930e-02 -9.79815796e-03 2.18109503e-01
3.33230138e-01 1.52748540e-01 3.21226448e-01 5.47988057e-01
1.17406356e+00 -2.76049636e-02 3.46256882e-01 2.11880524e-02
4.58348662e-01 -1.10938072e-01 6.01228118e-01 7.64904976e-01
-1.99901953e-01 3.38504136e-01 3.74888867e-01 -6.53854191e-01
-5.81633687e-01 -1.04026234e+00 4.00159568e-01 1.43164265e+00
2.67172307e-01 -3.95912498e-01 -5.51908076e-01 -1.15500081e+00
2.46873200e-01 6.62906051e-01 -7.22832143e-01 -8.11314106e-01
-5.29453635e-01 -7.68897057e-01 -1.62681744e-01 6.03229463e-01
5.18815041e-01 -1.24631321e+00 9.64117274e-02 5.66985965e-01
1.09706171e-01 -9.62745130e-01 -7.87750363e-01 -1.65947378e-01
-1.05993164e+00 -1.26564133e+00 -5.03288627e-01 -1.23219347e+00
9.20346498e-01 6.88796818e-01 1.19530511e+00 4.57219571e-01
-3.46600711e-01 6.66133225e-01 -6.05068505e-01 -2.27490887e-01
1.74189471e-02 4.15572792e-01 -2.99347397e-02 3.80471736e-01
2.73284405e-01 -7.91096568e-01 -5.89683533e-01 1.45754054e-01
-8.33243668e-01 1.22510903e-01 6.94824755e-01 8.94205153e-01
7.57594347e-01 2.71922797e-01 1.05985272e+00 -1.40949869e+00
8.88413429e-01 -8.28190982e-01 -1.96155041e-01 6.65437520e-01
-9.61064577e-01 -7.61820897e-02 8.67915928e-01 -8.36199880e-01
-1.01010621e+00 -9.51468423e-02 2.97095060e-01 -6.28430963e-01
3.77651632e-01 9.05673862e-01 -1.03509977e-01 -1.48328483e-01
4.99516964e-01 4.92907554e-01 -1.19346566e-01 -3.36932510e-01
7.03189790e-01 4.34958279e-01 5.21564960e-01 -6.11960769e-01
8.06333601e-01 1.96893618e-01 -1.68830633e-01 -7.48904407e-01
-8.76831055e-01 -3.69490266e-01 -7.40854681e-01 -4.45498645e-01
3.39544713e-01 -7.31358230e-01 -4.38606232e-01 5.74200988e-01
-8.52902889e-01 -6.05166912e-01 -4.67171848e-01 -1.09914899e-01
-2.22611144e-01 3.96028101e-01 -4.97308820e-01 -6.06196404e-01
-5.14434814e-01 -6.13681138e-01 7.32545078e-01 4.18321997e-01
4.31526512e-01 -1.22493160e+00 5.46593070e-02 1.40728623e-01
4.59853470e-01 3.31988603e-01 1.12623799e+00 -6.47595465e-01
-6.85250103e-01 -2.17222542e-01 -2.99750984e-01 1.36498958e-01
2.55891383e-01 -2.46884853e-01 -5.25634110e-01 -6.12577021e-01
-4.37433720e-01 -3.38875949e-01 1.12307596e+00 -2.11776402e-02
1.49188733e+00 -3.73741359e-01 -5.66914678e-01 4.67211664e-01
1.33286166e+00 2.33472541e-01 4.98903126e-01 -3.63322310e-02
1.17527640e+00 1.99101359e-01 4.66051996e-01 3.56601685e-01
8.43778312e-01 4.10134673e-01 3.16398382e-01 1.58384621e-01
-5.18413246e-01 -6.15997851e-01 -9.69075784e-02 1.27441347e+00
-1.15382560e-01 -2.06067488e-01 -8.21538687e-01 6.06574655e-01
-2.18795252e+00 -9.66944098e-01 2.12655559e-01 2.20561004e+00
6.51492894e-01 4.33040172e-01 1.78976171e-02 -7.74100125e-02
9.69162822e-01 2.31996268e-01 -9.68274832e-01 -6.25181720e-02
2.77536929e-01 1.62987739e-01 -5.41079678e-02 3.29608321e-01
-9.07105148e-01 1.17361462e+00 5.42121077e+00 9.80461180e-01
-9.66410995e-01 1.79554820e-01 6.72116339e-01 -6.23906814e-02
-3.17197621e-01 -8.03022981e-02 -5.19718349e-01 4.82645959e-01
5.56224465e-01 -5.53749859e-01 7.43229568e-01 9.54764724e-01
-3.90991002e-01 3.39875519e-01 -8.25645506e-01 9.14424658e-01
4.00138229e-01 -1.30078030e+00 3.43038678e-01 -2.21388474e-01
8.70055616e-01 -6.32534623e-02 -1.12221695e-01 1.00905025e+00
3.23910505e-01 -4.14302677e-01 3.18229735e-01 6.79673374e-01
8.80100131e-01 -7.88179040e-01 4.30599779e-01 3.21685404e-01
-1.76403677e+00 -3.47840369e-01 -4.66796368e-01 -1.35844657e-02
-1.13295518e-01 5.53461075e-01 -8.13438118e-01 7.74657190e-01
5.50854445e-01 1.08160901e+00 -9.58798945e-01 1.10619855e+00
-1.52225569e-01 5.02839983e-01 4.36269082e-02 -2.03237653e-01
1.32229496e-02 -1.47292810e-02 3.33950311e-01 8.45631599e-01
4.06811126e-02 3.44877005e-01 5.28331220e-01 4.84063476e-01
-5.04604518e-01 1.95229650e-01 -6.52978539e-01 -6.63420185e-02
6.25963211e-01 1.27732670e+00 -9.76039648e-01 -5.16828120e-01
-4.92910087e-01 1.18616164e+00 1.03197110e+00 4.56577957e-01
-7.23636985e-01 -6.73791528e-01 1.26956329e-01 2.46445656e-01
4.65739816e-01 -1.31261244e-01 1.71454459e-01 -1.35071790e+00
1.00169688e-01 -5.79797447e-01 7.56685197e-01 -6.42399073e-01
-1.61450589e+00 6.85564280e-01 -6.52612969e-02 -1.00262427e+00
1.76083192e-01 -2.36327469e-01 -1.02844238e+00 1.52168363e-01
-1.45572293e+00 -1.37784970e+00 -7.31359661e-01 7.01141715e-01
7.23301768e-01 -2.67875165e-01 3.58907133e-01 3.58018219e-01
-8.23900640e-01 7.21350908e-01 -9.83064249e-03 8.52260739e-02
4.31810051e-01 -1.11419022e+00 7.34475553e-01 7.87359118e-01
2.00030446e-01 5.30224681e-01 1.80832699e-01 -9.97635067e-01
-1.55307996e+00 -1.57420146e+00 4.35945243e-01 -2.45575443e-01
6.78197265e-01 -4.53432143e-01 -1.20497501e+00 7.72611499e-01
-6.55956939e-02 5.08203208e-01 4.67106640e-01 3.03077281e-01
-3.02198887e-01 -5.44989407e-01 -9.95450556e-01 4.79462117e-01
1.73617697e+00 -2.41218761e-01 -5.44071496e-01 5.04443765e-01
1.22305334e+00 -2.47598678e-01 -6.84022665e-01 5.05533218e-01
3.78960669e-01 -5.57306886e-01 9.11389887e-01 -8.78104806e-01
6.46903412e-03 -1.73696369e-01 2.77790129e-01 -1.60365820e+00
-9.48688686e-01 -5.50122142e-01 -8.45512390e-01 1.26374412e+00
2.55428314e-01 -7.89666533e-01 9.67417717e-01 2.26687416e-01
-3.01833481e-01 -1.14644825e+00 -7.53225982e-01 -8.95526707e-01
-2.70626515e-01 -6.60463199e-02 7.00354099e-01 1.12266624e+00
1.18810497e-02 7.10708916e-01 -3.00239325e-01 -6.45103976e-02
7.58313477e-01 3.21465075e-01 7.72970259e-01 -1.49052477e+00
-3.77341062e-01 -4.31631386e-01 -4.71906006e-01 -9.06834006e-01
6.98632002e-02 -1.25199330e+00 -3.26322317e-01 -1.87678754e+00
4.00750130e-01 -6.43475175e-01 -7.48994172e-01 6.08729661e-01
-6.85521662e-01 9.82605387e-03 9.08597559e-02 8.63452330e-02
-1.26597500e+00 7.20993340e-01 1.40406811e+00 -4.40897018e-01
-4.91950899e-01 3.49460691e-02 -9.54026103e-01 4.48844731e-01
6.40613317e-01 -3.87632310e-01 -9.14626956e-01 -4.20950830e-01
3.16064417e-01 -2.75254518e-01 -2.41902918e-02 -9.32763100e-01
5.28591573e-01 -2.54789174e-01 3.19336921e-01 -3.21306854e-01
-4.06762399e-02 -7.53416121e-01 6.06157295e-02 5.48000455e-01
-9.79591012e-02 3.88476700e-02 2.68864185e-02 1.18168485e+00
1.57990158e-01 -2.66102906e-02 5.55218875e-01 -3.67767125e-01
-1.16626441e+00 9.90275741e-01 1.33538574e-01 3.06809455e-01
1.29218781e+00 -1.90973729e-01 -4.81620431e-01 -7.96273127e-02
-8.82066488e-01 6.07194364e-01 2.75540650e-01 6.93435609e-01
8.96043658e-01 -1.61507976e+00 -6.03926659e-01 3.33894670e-01
4.43807065e-01 2.00677872e-01 4.94992167e-01 5.37988484e-01
-1.89066276e-01 -2.61279106e-01 6.67261332e-02 -4.18621123e-01
-9.25154150e-01 1.26405430e+00 5.70926592e-02 -3.90459538e-01
-6.21738970e-01 8.46615791e-01 -1.82829723e-01 -4.91884410e-01
3.47335547e-01 -1.17811523e-01 -2.23452270e-01 1.37367770e-01
4.48151499e-01 4.97335285e-01 -3.38359922e-02 -1.78908542e-01
-2.36958385e-01 4.64049757e-01 -4.38284338e-01 7.36610532e-01
1.33189690e+00 -2.20469132e-01 -1.17143348e-01 5.89901567e-01
1.01323891e+00 -3.09030950e-01 -1.09108436e+00 -8.00897658e-01
-6.65830672e-02 -4.16751772e-01 -1.50245681e-01 -6.43764913e-01
-1.25888026e+00 5.88585973e-01 3.96522015e-01 4.42793727e-01
1.10366499e+00 3.47201452e-02 8.46585155e-01 5.77370465e-01
6.81694210e-01 -1.10619462e+00 5.24323463e-01 3.85889500e-01
7.87775815e-01 -1.19031966e+00 2.45174810e-01 -5.26949406e-01
-6.14995778e-01 9.35388923e-01 1.04442716e+00 -1.78068176e-01
8.26395035e-01 -3.29900295e-01 -5.17183185e-01 -3.28371465e-01
-9.42144454e-01 -3.12482595e-01 3.62132460e-01 7.27540910e-01
2.26169806e-02 5.88782541e-02 -1.40517578e-01 7.06743479e-01
3.00628632e-01 1.17179953e-01 2.72270024e-01 9.65391755e-01
-6.81451142e-01 -7.78998613e-01 1.99636921e-01 9.88482535e-01
1.66424170e-01 -1.52595252e-01 -2.90877461e-01 6.00175321e-01
1.45439267e-01 6.13792598e-01 2.52995137e-02 -7.13644385e-01
5.59042573e-01 7.37238452e-02 5.57787180e-01 -9.09528196e-01
-3.85417312e-01 -2.70432115e-01 -9.89701226e-02 -3.56860161e-01
-1.90003529e-01 -3.48483503e-01 -1.09394073e+00 -1.99761525e-01
-5.64927340e-01 1.34935513e-01 5.97174913e-02 6.47672474e-01
6.99103355e-01 9.40594196e-01 7.97531307e-01 -7.67056167e-01
-2.91054994e-01 -8.85670006e-01 -5.73402405e-01 4.89367902e-01
-8.66867676e-02 -8.25653255e-01 -1.64510325e-01 -1.41966954e-01] | [9.611858367919922, 3.648683786392212] |
60a21a86-3314-45b4-90a6-eb62d5fff2a0 | mapping-chatgpt-in-mainstream-media-early | 2305.1834 | null | https://arxiv.org/abs/2305.18340v1 | https://arxiv.org/pdf/2305.18340v1.pdf | Mapping ChatGPT in Mainstream Media: Early Quantitative Insights through Sentiment Analysis and Word Frequency Analysis | The exponential growth in user acquisition and popularity of ChatGPT, an artificial intelligence(AI) powered chatbot, was accompanied by widespread mainstream media coverage. This article presents a quantitative data analysis of the early trends and sentiments revealed by conducting text mining and NLP methods onto a corpus of 10,902 mainstream news headlines related to the subject of ChatGPT and artificial intelligence, from the launch of ChatGPT in November 2022 to March 2023. The findings revealed in sentiment analysis, ChatGPT and artificial intelligence, were perceived more positively than negatively in the mainstream media. In regards to word frequency results, over sixty-five percent of the top frequency words were focused on Big Tech issues and actors while topics such as jobs, diversity, ethics, copyright, gender and women were poorly represented or completely absent and only accounted for six percent of the total corpus. This article is a critical analysis into the power structures and collusions between Big Tech and Big Media in their matrix of domination. | ['Maya Karanouh'] | 2023-05-25 | null | null | null | null | ['chatbot', 'ethics', 'sentiment-analysis', 'chatbot'] | ['methodology', 'miscellaneous', 'natural-language-processing', 'natural-language-processing'] | [-3.42771262e-01 6.53681278e-01 -4.96334612e-01 2.06053749e-01
-2.22905770e-01 -6.84237897e-01 9.47707653e-01 4.86817598e-01
-3.63657683e-01 5.06049752e-01 9.27621722e-01 -3.58546376e-01
7.66291469e-03 -3.77399206e-01 -4.99506183e-02 -3.15376252e-01
4.72078353e-01 4.55178380e-01 -2.00806469e-01 -6.63483441e-01
8.80942643e-01 -2.38271594e-01 -1.07682395e+00 2.87454665e-01
7.71374404e-01 9.08155203e-01 7.98445940e-03 4.40981954e-01
-6.02818966e-01 1.57174110e+00 -1.14572680e+00 -8.56934369e-01
1.49138244e-02 -4.66512084e-01 -7.63682544e-01 -2.09959850e-01
4.80016805e-02 -1.67608410e-02 -3.23232204e-01 9.72518444e-01
2.76417971e-01 -4.11663324e-01 4.39778954e-01 -1.30942130e+00
-5.81646442e-01 1.09993196e+00 -7.02574313e-01 4.68803525e-01
3.29908401e-01 9.94240269e-02 1.44551456e+00 -4.74706531e-01
1.32321119e+00 1.11977136e+00 6.97879136e-01 -6.99628592e-02
-6.79990113e-01 -6.03578925e-01 -1.39592931e-01 -1.28734812e-01
-7.91522026e-01 -5.59166148e-02 6.18064106e-01 -1.10463297e+00
7.38846600e-01 3.09782505e-01 1.03358650e+00 8.57607663e-01
5.04215658e-01 4.18391049e-01 1.04401457e+00 -3.05086583e-01
1.73089013e-01 8.27831626e-01 5.93476593e-02 2.01899469e-01
2.16471538e-01 -1.00734758e+00 -7.19936728e-01 -4.82782483e-01
1.80502683e-01 -9.38650966e-02 1.58661008e-01 5.51994145e-01
-1.36093211e+00 1.13077939e+00 -1.61338523e-02 9.06992912e-01
-7.46268749e-01 -2.65983999e-01 8.42678308e-01 6.42510653e-01
1.15765607e+00 9.04749870e-01 -4.73120362e-01 -1.30436289e+00
-3.93628210e-01 2.83570915e-01 1.32964170e+00 7.44116008e-01
3.44707638e-01 -1.73497930e-01 1.42096356e-01 8.55256021e-01
1.23003557e-01 1.90353021e-01 3.93361568e-01 -1.14933348e+00
6.59200251e-01 1.11447537e+00 -1.31389484e-01 -1.64418042e+00
-2.45569974e-01 -3.32198977e-01 -3.22944880e-01 -3.60789716e-01
6.11496806e-01 -7.63867378e-01 6.57144701e-03 1.21316063e+00
4.06434759e-02 -9.40348148e-01 -2.85344690e-01 6.02355480e-01
8.37563634e-01 7.06469595e-01 1.57841399e-01 -3.49482030e-01
1.44082093e+00 -5.73381782e-01 -1.03832638e+00 -4.91317511e-01
5.42814434e-01 -1.07574427e+00 9.76440728e-01 4.00517195e-01
-1.19424570e+00 5.67122549e-02 -8.32223058e-01 -1.90400407e-01
-6.54813707e-01 -3.07268530e-01 6.59044921e-01 5.88599622e-01
-6.38372898e-01 4.18729603e-01 -4.04496081e-02 -6.95862055e-01
5.18752337e-01 -1.62915960e-01 -1.57231897e-01 3.85129929e-01
-8.66679072e-01 1.16924226e+00 1.03666425e-01 -4.40723181e-01
2.68261492e-01 -8.41972828e-01 -2.44690627e-01 -1.67516857e-01
5.79437733e-01 1.26550302e-01 1.29000580e+00 -1.25324595e+00
-1.29517531e+00 1.07601297e+00 3.44713390e-01 -2.04838976e-01
3.31418216e-01 -2.33057961e-01 -2.31719509e-01 5.21708615e-02
4.63032514e-01 -6.80560805e-03 2.07881510e-01 -7.15272248e-01
-9.25832391e-01 -2.88582832e-01 1.37295708e-01 -1.29459739e-01
-8.51568043e-01 8.59140456e-01 -2.64842138e-02 -4.87304419e-01
-1.20562471e-01 -7.06497192e-01 -5.52535020e-02 -5.87177753e-01
-1.97146848e-01 -3.81993175e-01 7.19634771e-01 -8.82411420e-01
1.50529253e+00 -2.19198847e+00 -4.60430933e-03 1.35642663e-01
9.39206541e-01 -4.73784834e-01 5.54725826e-01 1.07418060e+00
4.89901036e-01 5.37017584e-01 5.56647122e-01 2.83803165e-01
2.13746816e-01 -1.75103953e-03 -2.78901234e-02 4.17459220e-01
-1.90674469e-01 6.85629010e-01 -9.58526134e-01 -4.42221761e-01
-1.65672287e-01 5.98141178e-02 -3.81301343e-01 -4.81424958e-01
-2.45196506e-01 1.84284657e-01 -6.17686868e-01 7.84939289e-01
7.20856786e-02 -4.24358964e-01 4.30215925e-01 2.59913266e-01
-9.95713651e-01 5.44620693e-01 -1.57010090e-02 9.82000709e-01
-2.27370635e-01 1.48167956e+00 4.96151686e-01 -5.87221503e-01
9.55476642e-01 4.14903313e-01 7.74897933e-01 -7.80163467e-01
9.21077490e-01 3.46585363e-01 5.42439461e-01 -4.47726667e-01
8.62604916e-01 -3.69729009e-03 -6.91635966e-01 6.63774967e-01
-9.83655732e-03 -4.44401324e-01 1.60214558e-01 6.23829722e-01
1.18689656e+00 -4.38276023e-01 4.99457091e-01 -6.43554747e-01
-1.96144566e-01 6.03846371e-01 3.80492866e-01 3.30367059e-01
-2.18105167e-01 1.83642387e-01 1.36422014e+00 -3.82174075e-01
-1.19741595e+00 -5.34539074e-02 1.25382483e-01 1.39032400e+00
-4.55054007e-02 -8.08990479e-01 -7.99574614e-01 -3.09707105e-01
-7.71916509e-02 6.19896948e-01 -5.41086435e-01 2.99902558e-01
-3.61391485e-01 -4.83054668e-01 2.18801554e-02 -2.29937792e-01
6.54129803e-01 -1.15263331e+00 -5.27382612e-01 2.39690408e-01
-3.64565223e-01 -1.35990131e+00 -1.16985723e-01 1.07661545e-01
-3.63579392e-01 -8.22604775e-01 -5.95626295e-01 -6.53047144e-01
3.33871514e-01 -1.48227602e-01 1.16916180e+00 -1.07918635e-01
-1.32532179e-01 -4.28701751e-02 -5.29162586e-01 -1.13838768e+00
-6.51482284e-01 3.81950617e-01 -1.54030532e-01 -3.77771080e-01
5.58716416e-01 -7.43861020e-01 -2.10420385e-01 -8.82361606e-02
-2.56814420e-01 3.39195766e-02 5.01928329e-01 4.05311465e-01
-5.17175138e-01 -5.39021641e-02 7.85947025e-01 -1.32161486e+00
9.47066963e-01 -1.18099880e+00 7.43077248e-02 -4.63260204e-01
-6.37420475e-01 -8.15913677e-01 4.18700427e-01 -4.47872758e-01
-8.86951268e-01 -8.67022634e-01 4.85780127e-02 3.42216253e-01
3.23619455e-01 1.12110472e+00 3.59415889e-01 1.11477964e-01
8.77984345e-01 -4.63214785e-01 4.60529923e-01 -3.39401633e-01
9.65264067e-03 1.17583644e+00 1.90589756e-01 -2.64920682e-01
5.40901303e-01 3.33243161e-01 -7.49588966e-01 -1.28295565e+00
-7.29169488e-01 -5.86896539e-01 -4.66625541e-02 -9.40113425e-01
8.00311506e-01 -7.03453362e-01 -1.05852735e+00 4.92274851e-01
-1.04116893e+00 -1.97692290e-01 -4.99329448e-01 4.05997187e-01
-9.38721821e-02 -9.10267383e-02 -8.77139270e-01 -8.71395946e-01
-2.78423071e-01 -5.57246149e-01 1.62504941e-01 3.71934474e-01
-9.15146172e-01 -9.32390153e-01 6.93509802e-02 8.65936458e-01
6.96809649e-01 5.15786707e-01 9.97632325e-01 -1.04228342e+00
2.13117957e-01 -4.96073842e-01 -2.08710015e-01 1.60596937e-01
2.39592165e-01 1.70318231e-01 -4.38652873e-01 3.00809175e-01
2.41500542e-01 -3.89548928e-01 -1.36399552e-01 6.18680678e-02
2.98073411e-01 -9.09898698e-01 -1.78871140e-01 -5.91221571e-01
1.20372212e+00 5.36795676e-01 5.31799674e-01 8.33616138e-01
3.78120124e-01 9.32665884e-01 6.28521204e-01 1.07378685e+00
2.44071916e-01 2.08450377e-01 3.25490609e-02 3.11878055e-01
4.12439018e-01 -9.18845832e-02 3.87815744e-01 1.46845341e+00
-4.65632111e-01 -7.54270926e-02 -1.24582660e+00 7.59657383e-01
-1.71473038e+00 -8.86593759e-01 -2.57637233e-01 1.29847527e+00
6.56394780e-01 6.76126122e-01 3.88981611e-01 -2.33371351e-02
4.90075290e-01 1.77595362e-01 -1.17113546e-01 -6.65843725e-01
-1.05771124e-01 -7.93613195e-02 5.78656793e-01 -1.15219817e-01
-3.43914807e-01 7.46034741e-01 6.94204044e+00 7.02990055e-01
-9.30879176e-01 4.21509802e-01 9.84828174e-01 -1.40710056e-01
-2.36918837e-01 3.92041095e-02 -1.92301661e-01 5.85313559e-01
7.75518179e-01 -6.67803228e-01 8.68688971e-02 1.12880290e+00
3.52665424e-01 -4.32357520e-01 -2.43261635e-01 6.75337315e-01
1.04068287e-01 -1.60880411e+00 -7.25137830e-01 4.09221411e-01
9.57165122e-01 3.55819613e-01 2.45007444e-02 3.34927589e-01
3.38593423e-01 -5.88422120e-01 1.01239204e+00 -1.10950500e-01
3.89642268e-01 -5.17112851e-01 8.63520622e-01 4.37904894e-01
-3.43169898e-01 -3.15263480e-01 -1.56713277e-01 -9.59417284e-01
1.76295191e-01 5.85802913e-01 -6.50986195e-01 3.55880074e-02
7.89614081e-01 6.93227232e-01 -1.46039605e-01 3.01304787e-01
2.07806081e-01 9.84788537e-01 -1.83396548e-01 -6.75867558e-01
4.72519517e-01 -4.52396125e-01 8.03082049e-01 1.03578722e+00
-9.01250616e-02 2.32944369e-01 -2.24165335e-01 6.00394130e-01
-1.79567933e-01 2.76302606e-01 -6.07445836e-01 -1.10472965e+00
5.10405481e-01 1.50902486e+00 -1.13111460e+00 -1.63440585e-01
-6.36655927e-01 2.89254546e-01 -7.04372600e-02 4.25427556e-02
-5.24292231e-01 -6.11432135e-01 7.16726124e-01 5.87143302e-01
-1.45742655e-01 -4.68477830e-02 -7.98126280e-01 -6.56881154e-01
-2.19106674e-01 -1.13258994e+00 -6.77048489e-02 -4.61760461e-01
-1.35304570e+00 5.07920682e-01 -2.47421712e-01 -7.12117434e-01
2.53216047e-02 -3.04367095e-01 -7.28239894e-01 4.56218183e-01
-3.90532166e-01 -6.34399474e-01 -9.21059474e-02 -7.55872671e-03
3.85690659e-01 -4.55203623e-01 3.66516888e-01 3.51759821e-01
-2.32838064e-01 1.42387971e-01 1.75708920e-01 7.66459182e-02
3.36829960e-01 -9.07096386e-01 2.92026132e-01 -4.24678996e-02
-5.42448997e-01 3.91012311e-01 9.96778607e-01 -9.23720419e-01
-1.25163805e+00 -5.12937941e-02 1.44664240e+00 -4.83259529e-01
1.43341303e+00 -2.58018911e-01 -1.10660307e-01 5.81601918e-01
8.38839352e-01 -1.02786803e+00 7.40718305e-01 3.98218155e-01
3.39193866e-02 3.74962986e-01 -1.24898803e+00 7.26384461e-01
7.29303479e-01 -4.16556865e-01 -4.69042867e-01 5.97359538e-01
7.10053444e-01 -1.69827342e-01 -1.11579990e+00 -2.99517959e-01
6.82617724e-01 -5.56053400e-01 2.55562484e-01 -4.41636026e-01
1.03148425e+00 4.42914933e-01 1.11121997e-01 -9.84907687e-01
-3.27161312e-01 -1.36756587e+00 4.60466623e-01 1.37964463e+00
5.25646448e-01 -8.21341991e-01 1.11200762e+00 8.89551342e-01
-2.08260298e-01 -6.08722448e-01 -8.05042922e-01 -5.37561476e-02
1.41283721e-01 -2.20201373e-01 -8.22211877e-02 1.37020183e+00
1.04480958e+00 6.63012862e-01 -1.64177135e-01 -1.01586080e+00
2.44180355e-02 -2.23967671e-01 7.58227408e-01 -1.45617199e+00
6.14245981e-03 -7.74522543e-01 -6.60908103e-01 -5.60319185e-01
-5.10954320e-01 -3.48275542e-01 -1.30421758e-01 -1.65641189e+00
3.03521335e-01 -7.69228339e-02 4.74419415e-01 2.30985522e-01
4.45790231e-01 2.91533291e-01 3.82727981e-01 4.85718012e-01
-3.38555217e-01 -6.44236654e-02 1.22108269e+00 -4.26900797e-02
-2.42369667e-01 -1.33745059e-01 -1.37293518e+00 1.18025494e+00
8.36612761e-01 -2.99544573e-01 -3.36681753e-02 1.04761370e-01
1.27923179e+00 1.25864428e-02 -3.81929278e-01 -6.29591405e-01
4.14221622e-02 -3.93372744e-01 -2.60753363e-01 -3.33100408e-01
2.58881092e-01 -9.01681185e-01 1.27060667e-01 2.95047224e-01
-4.08416063e-01 2.25606993e-01 4.34922241e-02 -1.64948270e-01
-4.40394968e-01 -6.50877133e-02 4.08523321e-01 -1.43792748e-01
-6.89762831e-02 -4.12646085e-01 -1.39631021e+00 6.64715290e-01
9.60487843e-01 -4.87697929e-01 -6.41612828e-01 -9.79807675e-01
-3.64561707e-01 -6.44896254e-02 1.74993753e-01 4.46590126e-01
1.40893487e-02 -8.41623127e-01 -7.95974493e-01 -6.69528425e-01
-2.67939001e-01 -4.36449200e-01 1.10418223e-01 1.21407604e+00
-7.63209641e-01 2.41584554e-01 -4.28367168e-01 1.32832587e-01
-1.12101901e+00 5.40921353e-02 -4.52836394e-01 -1.60340294e-01
-6.41049087e-01 6.15857244e-01 -2.10935399e-02 -1.91384599e-01
8.63920823e-02 -1.92430153e-01 -4.08603072e-01 6.32445455e-01
3.72573584e-01 7.48816848e-01 -4.48069185e-01 -7.82697022e-01
-1.66173786e-01 6.72775060e-02 7.53865298e-03 -4.22288567e-01
1.57455814e+00 -7.44052306e-02 -6.18536651e-01 1.04066896e+00
1.22550035e+00 4.00909960e-01 -1.92681700e-01 4.36179936e-02
3.32177341e-01 -6.29840612e-01 -2.39052400e-01 -9.63388085e-01
-1.02389956e+00 2.71107942e-01 -4.86858279e-01 1.02752554e+00
4.76599813e-01 4.18805450e-01 7.18032539e-01 1.03713982e-01
7.98850879e-02 -1.60854352e+00 1.85436442e-01 7.78694630e-01
9.42563355e-01 -7.16523767e-01 1.54528961e-01 -5.58759868e-01
-1.08377290e+00 9.17559981e-01 7.45450795e-01 1.35512456e-01
9.34005141e-01 4.82952803e-01 2.11837575e-01 -7.72469461e-01
-7.28187382e-01 4.68575954e-01 -3.21235925e-01 -4.28042784e-02
6.24632478e-01 -5.46863228e-02 -1.02993107e+00 8.73739958e-01
-7.14017451e-01 -4.16421890e-02 8.36732626e-01 1.00842834e+00
-4.39419031e-01 -4.11371797e-01 -8.90800804e-02 7.79556751e-01
-1.26094723e+00 8.66047107e-03 -1.12864923e+00 1.00108111e+00
9.19672474e-02 9.16479170e-01 3.36234421e-01 -7.79785454e-01
2.37582568e-02 -6.60607219e-02 -2.37466693e-01 -4.45828378e-01
-1.42216551e+00 9.57643092e-02 7.71725059e-01 -2.90069114e-02
-3.48813683e-01 -8.90464962e-01 -9.65564728e-01 -9.29202735e-01
-4.47980404e-01 6.85547471e-01 1.00801861e+00 9.93728697e-01
4.01013702e-01 3.48020285e-01 3.22889626e-01 -3.54809046e-01
1.00504093e-01 -1.46051061e+00 -6.13460362e-01 -5.00943139e-03
-2.42734388e-01 -1.71216816e-01 -5.84557533e-01 -1.73134521e-01] | [10.666936874389648, 7.2535271644592285] |
f2d5b877-1aa8-4059-9939-c58dbd5f23d7 | attribute-specific-manipulation-based-on | 2302.0926 | null | https://arxiv.org/abs/2302.09260v1 | https://arxiv.org/pdf/2302.09260v1.pdf | Attribute-Specific Manipulation Based on Layer-Wise Channels | Image manipulation on the latent space of the pre-trained StyleGAN can control the semantic attributes of the generated images. Recently, some studies have focused on detecting channels with specific properties to directly manipulate the latent code, which is limited by the entanglement of the latent space. To detect the attribute-specific channels, we propose a novel detection method in the context of pre-trained classifiers. We analyse the gradients layer by layer on the style space. The intensities of the gradients indicate the channel's responses to specific attributes. The latent style codes of channels control separate attributes in the layers. We choose channels with top-$k$ gradients to control specific attributes in the maximum response layer. We implement single-channel and multi-channel manipulations with a certain attribute. Our methods can accurately detect relevant channels for a large number of face attributes. Extensive qualitative and quantitative results demonstrate that the proposed methods outperform state-of-the-art methods in generalization and scalability. | ['Furao Shen', 'Jian Zhao', 'Yuanjie Yan'] | 2023-02-18 | null | null | null | null | ['image-manipulation'] | ['computer-vision'] | [ 7.73929119e-01 5.14813364e-02 -2.68893450e-01 -4.42807794e-01
-4.84774917e-01 -8.89032066e-01 6.56969666e-01 -4.58184212e-01
-1.92565635e-01 5.20990312e-01 2.19445869e-01 2.22488478e-01
2.14515969e-01 -9.03958619e-01 -8.03359449e-01 -1.11963642e+00
-3.08355063e-01 -1.67968795e-01 -3.33384514e-01 7.55635500e-02
3.12330455e-01 3.86628181e-01 -1.31650794e+00 7.44957745e-01
2.53339201e-01 1.30177236e+00 -1.14174463e-01 6.83884144e-01
-9.69227627e-02 2.50225335e-01 -4.25836056e-01 -4.16877002e-01
4.15289491e-01 -8.61150742e-01 -3.76985759e-01 3.88092957e-02
3.82009536e-01 -5.95857948e-02 -3.56796980e-01 1.60159004e+00
5.85086644e-01 -8.31001937e-01 8.16851079e-01 -1.18091524e+00
-7.31745422e-01 7.29649842e-01 -4.32851225e-01 -1.38710663e-01
4.77532893e-01 2.01642677e-01 1.03950417e+00 -7.76738107e-01
8.67169380e-01 1.67939568e+00 2.05066696e-01 1.01453960e+00
-1.67029655e+00 -1.54423463e+00 1.47560045e-01 -3.47551592e-02
-1.33566368e+00 -7.18109429e-01 1.15209258e+00 -4.13068920e-01
2.37795070e-01 2.58341074e-01 5.42569220e-01 1.50515604e+00
3.12322587e-01 7.96411932e-01 1.73750818e+00 -4.01057482e-01
2.11952269e-01 5.01446307e-01 -5.91146052e-01 1.09476674e+00
-6.71619996e-02 2.62010545e-01 -1.02013576e+00 -2.45063618e-01
8.31402957e-01 -4.27922368e-01 -1.26002595e-01 -4.77788687e-01
-1.47273350e+00 9.84062374e-01 3.95143539e-01 -1.84819981e-01
4.50907886e-04 5.30806601e-01 1.88751161e-01 6.68932259e-01
2.21043229e-01 5.26764810e-01 -3.10621232e-01 1.54245481e-01
-4.60419863e-01 -2.16282964e-01 7.56607950e-01 1.41445291e+00
1.04212248e+00 -1.33415490e-01 -6.33778393e-01 4.41721767e-01
3.66168618e-01 8.95320237e-01 -1.96722314e-01 -8.32491577e-01
3.65804672e-01 5.61059415e-01 -3.18892986e-01 -9.57802832e-01
1.27162054e-01 -3.59532863e-01 -1.00045288e+00 2.12605283e-01
3.88667911e-01 -3.07306796e-01 -9.27281857e-01 2.01188660e+00
1.36297166e-01 -2.63994560e-02 -3.39203775e-02 8.04665267e-01
4.74546105e-01 5.55336773e-01 2.10663557e-01 -2.16630697e-01
1.41904736e+00 -4.49055344e-01 -8.28878462e-01 -1.49480700e-01
5.21000326e-01 -8.31583500e-01 1.10864127e+00 1.04103424e-01
-7.50371635e-01 -3.66343707e-01 -1.14790618e+00 4.53491718e-01
-3.83475542e-01 5.88911176e-01 9.17523205e-01 9.27762449e-01
-8.74278367e-01 1.67279601e-01 -2.11925194e-01 -8.29200968e-02
8.89199913e-01 6.35731876e-01 -3.76401812e-01 -2.63392180e-03
-1.40369701e+00 9.55264047e-02 -1.96019057e-02 3.63238119e-02
-1.39951479e+00 -2.92539299e-01 -5.72626770e-01 -6.24670684e-02
4.64002378e-02 -2.52551079e-01 5.04576027e-01 -1.14709222e+00
-1.72778940e+00 1.15634573e+00 -4.07914035e-02 2.76309818e-01
4.97753382e-01 1.09324433e-01 -3.44235033e-01 2.95031458e-01
1.18653692e-01 1.11607277e+00 1.53938580e+00 -1.09286523e+00
-5.71442306e-01 -4.00813460e-01 1.05657876e-01 -1.60191432e-02
-7.77456462e-01 2.65772969e-01 -5.13025522e-01 -6.48220837e-01
2.79201716e-01 -1.18652189e+00 2.64492463e-02 2.91654348e-01
-9.42573369e-01 1.44742236e-01 9.29442942e-01 3.82123962e-02
9.39888000e-01 -2.48470402e+00 4.27796453e-01 3.74887377e-01
2.84190595e-01 -1.18465178e-01 -3.05481642e-01 1.02693677e-01
1.20571844e-01 4.12150055e-01 -1.50895029e-01 -3.45937349e-02
-6.77050576e-02 -1.06182955e-01 -3.55338901e-01 7.16687500e-01
2.60281712e-01 7.12850094e-01 -6.77993774e-01 -5.95709026e-01
-1.83639601e-01 4.72259790e-01 -7.19155610e-01 3.68364096e-01
-2.01641008e-01 8.80290270e-01 -9.24241245e-01 8.48317802e-01
6.96773529e-01 -5.63920438e-02 1.77489102e-01 -1.80149481e-01
1.48298457e-01 1.57366637e-02 -9.97757256e-01 1.68477094e+00
-1.54928803e-01 7.97075272e-01 3.79250616e-01 -5.10978401e-01
9.96859252e-01 2.53828675e-01 6.14590757e-02 -3.47590655e-01
1.56465515e-01 2.08661392e-01 -5.13032787e-02 -3.98418784e-01
-1.22154169e-01 3.23307812e-02 -5.80081940e-01 4.43781942e-01
2.57054895e-01 4.00009453e-02 -3.66986245e-01 1.84252858e-01
9.36148822e-01 -1.16807602e-01 1.00059062e-03 -4.50801998e-01
5.74418485e-01 -6.42437160e-01 4.35455412e-01 1.11137807e+00
-2.74686128e-01 1.69480100e-01 1.00610495e+00 -1.52881548e-01
-9.97108519e-01 -8.53455842e-01 -2.88501382e-01 1.25985038e+00
2.10197702e-01 -3.21690410e-01 -9.93103445e-01 -7.79324472e-01
8.86013806e-02 3.78141478e-02 -8.47101688e-01 -3.33505362e-01
-2.42373541e-01 -7.08178461e-01 8.56414795e-01 1.03295438e-01
1.08886588e+00 -1.08818209e+00 -2.95850068e-01 -4.81102876e-02
1.05978258e-01 -1.31192720e+00 -2.51865745e-01 8.22142139e-02
-4.71877843e-01 -7.06301033e-01 -3.96363497e-01 -6.97598934e-01
1.05843627e+00 -4.71945673e-01 4.47288364e-01 -1.98461965e-01
-3.14884394e-01 3.47163230e-01 -1.79897621e-01 -3.36935312e-01
-3.91614288e-01 3.34949374e-01 -4.97240722e-02 6.22823060e-01
4.51857805e-01 -4.22446966e-01 -6.65904999e-01 3.60994190e-01
-4.95430499e-01 6.40731156e-02 1.03189480e+00 8.83394539e-01
3.70590955e-01 -2.35307634e-01 1.28665105e-01 -1.07091379e+00
6.07685804e-01 -1.03773713e-01 -7.76823997e-01 1.27137706e-01
-5.82026422e-01 5.92038572e-01 3.37304115e-01 -5.67846358e-01
-9.34986413e-01 6.04014337e-01 3.40438396e-01 -1.02758430e-01
-3.77378076e-01 -5.20325117e-02 -7.04459846e-01 -6.34049952e-01
5.85844576e-01 1.85699031e-01 -1.55854389e-01 -1.65363386e-01
5.31344116e-01 6.60298049e-01 1.83401089e-02 -7.55239308e-01
9.97307301e-01 7.36770689e-01 4.72537398e-01 -6.41944945e-01
-7.31449425e-01 -1.88689884e-02 -7.00378597e-01 -4.21452999e-01
1.07814586e+00 -9.11730289e-01 -8.40058327e-01 7.81707764e-01
-1.11434996e+00 -2.17855886e-01 4.37485762e-02 3.16539675e-01
-4.91517454e-01 -9.29507613e-02 -8.01101208e-01 -6.57413363e-01
-5.38228787e-02 -1.45591557e+00 1.23183692e+00 3.19819152e-02
8.97047892e-02 -6.56772792e-01 -5.00036299e-01 -2.69905567e-01
4.97847319e-01 1.54137745e-01 1.01491785e+00 -2.30852410e-01
-1.08997166e+00 -2.88422763e-01 -2.55423814e-01 1.84221610e-01
-8.14858750e-02 -2.79987246e-01 -1.47216845e+00 -3.70030046e-01
-1.33878112e-01 -5.03506899e-01 1.08513188e+00 1.16217338e-01
1.44152009e+00 -4.71659422e-01 -6.32038176e-01 1.09005892e+00
1.22635984e+00 7.53837964e-03 5.65781474e-01 3.96846272e-02
8.25699508e-01 4.98451948e-01 3.13115507e-01 4.39387918e-01
-2.79869705e-01 5.54758966e-01 3.90770555e-01 -1.12671874e-01
-7.33708367e-02 -4.89731401e-01 6.17879093e-01 4.95165497e-01
1.59712076e-01 -9.03744102e-02 -3.53532761e-01 1.12498544e-01
-1.27626574e+00 -6.15157425e-01 8.35939720e-02 1.87472951e+00
1.14132774e+00 3.93266588e-01 -5.52120864e-01 -1.12173511e-02
9.25634682e-01 3.62992734e-01 -4.89908576e-01 -3.40328962e-01
-3.00791591e-01 1.45170853e-01 8.29277217e-01 3.91231447e-01
-1.32417381e+00 1.22451091e+00 7.20870209e+00 9.45761204e-01
-1.24890649e+00 2.57865693e-02 8.33377898e-01 -1.34899793e-02
-4.71440405e-01 1.94594264e-01 -1.27285480e+00 4.93182331e-01
3.95607471e-01 5.24293125e-01 6.55618250e-01 7.97484815e-01
-3.85572106e-01 1.99419305e-01 -1.36862659e+00 9.85449135e-01
-1.79966837e-02 -1.26989985e+00 2.99927980e-01 3.98633242e-01
8.49613667e-01 -5.52034914e-01 7.27384031e-01 1.19394818e-02
8.76410306e-02 -1.09872758e+00 6.26690090e-01 5.29097021e-01
1.57178032e+00 -4.63557720e-01 2.56987840e-01 -1.38048679e-01
-1.02093256e+00 -3.06848377e-01 -4.56517577e-01 1.44603476e-02
-5.65217376e-01 6.06362581e-01 -5.81311405e-01 -4.27880377e-01
5.21611691e-01 6.59409404e-01 -5.60452461e-01 2.23176584e-01
-8.80942762e-01 8.93057525e-01 -1.26881853e-01 -3.02387208e-01
1.18647762e-01 -1.59883723e-01 6.95036948e-01 1.39428997e+00
3.27477396e-01 1.54352918e-01 -1.56648323e-01 1.30337656e+00
-5.10060012e-01 -9.87422094e-02 -8.39715004e-01 -3.75921220e-01
7.24253476e-01 1.37280977e+00 -6.05889022e-01 -2.98428684e-01
-2.57710367e-01 1.23811424e+00 -9.74875018e-02 5.51315725e-01
-4.93326426e-01 -4.31933552e-01 7.23075926e-01 -2.43441880e-01
3.16657454e-01 -6.88898489e-02 -4.95312452e-01 -9.45886135e-01
-2.10716441e-01 -8.63827765e-01 1.50209114e-01 -3.71325850e-01
-1.11086452e+00 3.08586329e-01 -2.80590385e-01 -1.01347113e+00
2.82731146e-01 -8.42974007e-01 -1.52703479e-01 9.26568985e-01
-1.52338922e+00 -1.22955239e+00 -2.95771569e-01 8.76622975e-01
5.34582846e-02 -7.11154938e-01 1.18675947e+00 9.65804160e-02
-4.06202316e-01 1.10732841e+00 -6.72715008e-02 5.70731461e-01
7.49849737e-01 -1.02222204e+00 8.12980980e-02 6.72179997e-01
-1.27760485e-01 7.03932285e-01 5.26259363e-01 -5.08148372e-01
-1.68366718e+00 -8.42436731e-01 6.40873849e-01 -3.51240247e-01
3.34530503e-01 -1.21621597e+00 -1.18201829e-01 6.08053684e-01
9.92893949e-02 2.65539944e-01 7.33199120e-01 5.26807047e-02
-7.20740497e-01 -4.49327707e-01 -1.07287955e+00 3.89989227e-01
1.25182581e+00 -1.18656623e+00 2.69934505e-01 2.10487291e-01
3.94040555e-01 -1.20177351e-01 -5.09386599e-01 2.49189422e-01
7.65024960e-01 -7.22504139e-01 7.11571991e-01 -4.00289536e-01
6.17087781e-01 -1.71935260e-01 -2.63596505e-01 -1.20973051e+00
-5.22850275e-01 -8.84215832e-01 3.00549090e-01 1.23894346e+00
5.99830449e-01 -4.85812813e-01 7.69217491e-01 3.11972171e-01
4.32054639e-01 -3.68105739e-01 -9.09454167e-01 -1.94588900e-01
-2.76719630e-01 -9.51136127e-02 7.79659331e-01 1.00206697e+00
-4.26810086e-02 2.29928672e-01 -5.56547999e-01 1.88162565e-01
9.32396412e-01 1.51090801e-01 6.36769235e-01 -1.02730286e+00
-1.46331891e-01 -2.79405683e-01 -8.02765608e-01 -1.12159455e+00
3.32009286e-01 -1.02601051e+00 -1.49232149e-01 -5.26211083e-01
4.79071409e-01 -7.16504097e-01 -4.61666197e-01 5.61105967e-01
7.09967017e-02 4.44444567e-01 -5.33258682e-03 2.72225469e-01
-3.00566792e-01 3.31409425e-01 1.44655049e+00 -3.35308760e-01
6.10283203e-02 -2.71337122e-01 -7.42646277e-01 3.65975171e-01
7.68976510e-01 -6.85743630e-01 -2.40356401e-01 -2.05466866e-01
5.89689434e-01 -2.47478053e-01 3.35182756e-01 -8.76605511e-01
6.83754683e-02 -3.97634692e-03 7.62945354e-01 1.29113942e-01
3.45003694e-01 -7.12210476e-01 -2.64565766e-01 4.01392609e-01
-1.15023804e+00 -5.32986581e-01 -1.24815442e-01 6.78417981e-01
-7.20947906e-02 2.71131396e-01 9.66334522e-01 -3.94244760e-01
-6.73170865e-01 4.56334382e-01 -4.97675419e-01 -2.38011882e-01
1.07789814e+00 -1.97175443e-02 -1.27245739e-01 -5.50352037e-01
-6.89806461e-01 -1.83614552e-01 2.19535366e-01 2.90024221e-01
6.99196875e-01 -1.57266724e+00 -4.90946531e-01 9.29789722e-01
3.95963758e-01 -4.93145674e-01 -3.42109829e-01 4.48803663e-01
-1.26454696e-01 2.78997004e-01 -7.13753045e-01 -8.01315248e-01
-1.25536370e+00 4.64830041e-01 3.04650545e-01 2.88285911e-01
-4.94832806e-02 1.45737803e+00 5.38103402e-01 -1.16089024e-01
2.40903020e-01 1.03137203e-01 -1.65409371e-01 1.80859804e-01
2.04755470e-01 -1.43214568e-01 -5.74176967e-01 -6.21966898e-01
-2.30715826e-01 7.65340030e-01 2.87704375e-02 -2.72591680e-01
7.93467820e-01 -1.26292601e-01 -4.89687800e-01 2.15153590e-01
1.49805391e+00 2.15659097e-01 -1.48187876e+00 -2.39557862e-01
-4.44429964e-01 -6.52406275e-01 1.14552081e-01 -7.54537284e-01
-1.28525102e+00 1.20658076e+00 1.04561794e+00 -7.16153383e-02
1.22008359e+00 2.41019085e-01 2.68932432e-01 1.74544334e-01
3.51653665e-01 -1.26179504e+00 2.69412547e-01 9.94583741e-02
6.15283668e-01 -9.58421707e-01 -1.86988056e-01 -7.77957916e-01
-5.16768336e-01 1.16278887e+00 5.71010888e-01 2.64669135e-02
8.22697639e-01 3.23592275e-01 1.27734661e-01 -4.29611564e-01
-5.06382465e-01 1.69555664e-01 4.95940447e-02 4.61722165e-01
3.27743560e-01 4.37460870e-01 1.91763174e-02 -9.92785916e-02
-8.32220763e-02 -3.89311641e-01 3.85815859e-01 5.34694493e-01
-2.11992741e-01 -1.20691454e+00 -3.68907839e-01 3.51633936e-01
-4.70308691e-01 -1.49121106e-01 -7.54604280e-01 3.04417759e-01
3.92836958e-01 7.54672706e-01 -1.38875723e-01 -7.38295615e-01
-3.35064121e-02 1.94593310e-01 5.72224796e-01 -4.13583368e-01
-1.42671511e-01 2.23053113e-01 -4.96114641e-02 -8.41969192e-01
-4.30877775e-01 -7.38569200e-01 -7.88664520e-01 -1.06291950e-01
-3.31279904e-01 -1.40554681e-01 7.25947917e-01 6.28647506e-01
5.16231954e-01 2.44199112e-01 1.06396425e+00 -4.75790530e-01
-5.98829091e-01 -8.55754673e-01 -7.58003235e-01 5.29352546e-01
4.85439271e-01 -6.29059255e-01 -6.25112772e-01 9.15407315e-02] | [11.972771644592285, -0.24241182208061218] |
cc1f186b-0068-4ac2-b7db-de7378b1fa16 | catch-you-and-i-can-revealing-source | 2302.12434 | null | https://arxiv.org/abs/2302.12434v1 | https://arxiv.org/pdf/2302.12434v1.pdf | Catch You and I Can: Revealing Source Voiceprint Against Voice Conversion | Voice conversion (VC) techniques can be abused by malicious parties to transform their audios to sound like a target speaker, making it hard for a human being or a speaker verification/identification system to trace the source speaker. In this paper, we make the first attempt to restore the source voiceprint from audios synthesized by voice conversion methods with high credit. However, unveiling the features of the source speaker from a converted audio is challenging since the voice conversion operation intends to disentangle the original features and infuse the features of the target speaker. To fulfill our goal, we develop Revelio, a representation learning model, which learns to effectively extract the voiceprint of the source speaker from converted audio samples. We equip Revelio with a carefully-designed differential rectification algorithm to eliminate the influence of the target speaker by removing the representation component that is parallel to the voiceprint of the target speaker. We have conducted extensive experiments to evaluate the capability of Revelio in restoring voiceprint from audios converted by VQVC, VQVC+, AGAIN, and BNE. The experiments verify that Revelio is able to rebuild voiceprints that can be traced to the source speaker by speaker verification and identification systems. Revelio also exhibits robust performance under inter-gender conversion, unseen languages, and telephony networks. | ['Wenyuan Xu', 'Xueluan Gong', 'Qianhao Miao', 'Yinan Zhong', 'Yanjiao Chen', 'Jiangyi Deng'] | 2023-02-24 | null | null | null | null | ['voice-conversion', 'voice-conversion', 'speaker-verification'] | ['audio', 'speech', 'speech'] | [ 3.88207972e-01 1.11039802e-01 2.07544893e-01 -5.40907048e-02
-9.51810360e-01 -1.04541492e+00 3.62456918e-01 -4.57034230e-01
-9.22208205e-02 4.71681476e-01 2.87512362e-01 -5.19396365e-01
2.71123767e-01 -3.48951906e-01 -6.76675498e-01 -6.82097912e-01
1.98587671e-01 2.00699493e-01 -1.55940294e-01 -1.97364181e-01
5.91268986e-02 8.14890385e-01 -1.38264060e+00 2.85528988e-01
3.37139785e-01 5.93648195e-01 -6.18188009e-02 1.01388383e+00
6.79996163e-02 4.64195102e-01 -9.94262457e-01 -3.63977134e-01
3.94276559e-01 -6.65470481e-01 -8.38498175e-01 4.52372767e-02
2.51179755e-01 -3.31162274e-01 -4.65466589e-01 1.05477011e+00
5.33611238e-01 -2.29622826e-01 6.65507436e-01 -1.44171190e+00
-3.50324750e-01 8.46556067e-01 -3.80391687e-01 1.82210669e-01
6.15406215e-01 -7.77098462e-02 7.63071895e-01 -7.57506073e-01
3.63253444e-01 1.32974482e+00 4.79248226e-01 8.02856982e-01
-1.15423989e+00 -1.22375011e+00 -2.50999093e-01 -2.78096478e-02
-1.57580113e+00 -1.10041773e+00 1.07926989e+00 -2.19225109e-01
4.15990263e-01 7.93247104e-01 2.04397470e-01 1.33299935e+00
3.87045108e-02 4.23327625e-01 7.88712263e-01 -6.05951607e-01
-8.41112062e-02 5.59657276e-01 -9.29728821e-02 3.19163263e-01
-3.62607926e-01 3.76023710e-01 -5.35143554e-01 -4.67908114e-01
3.21848750e-01 -4.02942210e-01 -7.85193384e-01 -8.74663983e-03
-7.17934668e-01 8.29965711e-01 -5.25205843e-02 3.89176667e-01
-1.62327573e-01 1.32950069e-02 2.55074650e-01 6.39298558e-01
-1.07874587e-01 3.80365789e-01 -1.04080535e-01 -1.73490047e-01
-9.47149396e-01 -1.44140065e-01 1.00820279e+00 8.69038224e-01
2.94840246e-01 6.11252785e-01 2.95112491e-01 6.63200915e-01
3.21466058e-01 5.80714941e-01 8.89264166e-01 -7.65401423e-01
3.21714371e-01 -6.53837547e-02 -1.07588165e-01 -9.95347142e-01
1.14198640e-01 -3.04857165e-01 -7.18227983e-01 7.84931481e-02
2.04419658e-01 -2.65739083e-01 -4.56125677e-01 1.65158391e+00
1.82762340e-01 2.49077290e-01 3.01410496e-01 7.23653674e-01
6.57751441e-01 8.61904383e-01 -5.41973948e-01 -3.34138304e-01
1.17255676e+00 -6.85974896e-01 -7.23761976e-01 -1.18210442e-01
1.22670203e-01 -1.05890715e+00 9.71741259e-01 4.09370691e-01
-6.86048687e-01 -5.18830359e-01 -1.19037223e+00 2.78013170e-01
6.64718822e-02 3.09133139e-02 1.79405123e-01 1.13779891e+00
-6.88482463e-01 3.50598902e-01 -5.90843618e-01 1.35402068e-01
9.82253402e-02 5.05947888e-01 -6.68829441e-01 2.43613943e-01
-1.22006154e+00 2.81647474e-01 6.43648803e-02 -3.53372097e-02
-1.07590997e+00 -6.03396416e-01 -7.08722293e-01 1.81979433e-01
6.20371476e-02 -9.87967849e-02 1.29424179e+00 -1.18399143e+00
-1.64258826e+00 5.21020055e-01 -2.16417685e-01 -4.46980536e-01
5.42136788e-01 7.46490657e-02 -1.12304544e+00 2.20176756e-01
4.44073044e-02 1.44603491e-01 1.74207628e+00 -1.34919989e+00
-5.89314938e-01 -4.03135002e-01 -5.08613825e-01 -1.71689436e-01
-3.73648554e-01 2.01927096e-01 -7.63852820e-02 -6.30931199e-01
2.25533266e-02 -1.01784968e+00 6.07528865e-01 -4.50982302e-01
-9.23660815e-01 3.10353070e-01 1.14871359e+00 -8.71082067e-01
9.16633844e-01 -2.73219132e+00 -1.20078661e-01 4.63656366e-01
-1.03780903e-01 5.91928959e-01 -2.08182976e-01 5.75652063e-01
-5.28165042e-01 2.10082144e-01 -8.26575682e-02 -3.85407627e-01
-2.15490893e-01 1.02549680e-01 -1.08881295e+00 5.97061992e-01
-7.64556974e-02 3.06931227e-01 -4.27343726e-01 -1.06327020e-01
-3.30657810e-01 7.73055255e-01 -5.34923494e-01 3.66163850e-01
4.63234097e-01 4.70700592e-01 -3.69691029e-02 3.67747545e-01
5.54361582e-01 5.39674461e-01 1.90018624e-01 -4.22647633e-02
1.21419929e-01 5.11274695e-01 -1.40366852e+00 9.12960172e-01
-5.13395607e-01 8.07059824e-01 7.10105658e-01 -5.44127882e-01
1.09502971e+00 7.68789113e-01 9.40274149e-02 -2.40623564e-01
2.23026872e-01 2.92435288e-01 1.65614784e-01 -2.65625775e-01
4.64935869e-01 -4.18837279e-01 -1.55152708e-01 6.90632284e-01
1.22383513e-01 -1.90205067e-01 -4.44096237e-01 3.16907167e-02
7.90573061e-01 -5.70304692e-01 1.08191542e-01 2.12760702e-01
7.77396202e-01 -5.66189826e-01 4.67340022e-01 5.21514297e-01
-1.00499094e-01 7.34006405e-01 3.76911193e-01 3.18618536e-01
-6.38881505e-01 -1.23596144e+00 1.45177513e-01 1.01735461e+00
-2.67645329e-01 -2.14761406e-01 -7.29363620e-01 -5.82798243e-01
-8.28580856e-02 9.73556459e-01 1.14429463e-02 -5.47923386e-01
-6.06631458e-01 -1.30882859e-01 1.48813939e+00 2.70404279e-01
2.47646764e-01 -8.02483976e-01 1.05401315e-02 3.28112990e-01
-4.46187079e-01 -1.07185829e+00 -1.03239071e+00 1.15195654e-01
-3.77564549e-01 -1.05072141e+00 -3.18042874e-01 -8.92640173e-01
3.02910000e-01 2.55479425e-01 3.09568882e-01 -1.19941346e-01
-1.31120324e-01 4.95152712e-01 -1.23359710e-01 -3.82973969e-01
-1.21858215e+00 5.31650335e-02 5.76952636e-01 4.29323763e-01
4.89284168e-04 -8.17008317e-01 1.87653989e-01 4.25878644e-01
-9.56927419e-01 -4.55718488e-01 1.32471263e-01 7.11744487e-01
9.78101492e-02 6.20320141e-01 7.19662011e-01 -7.38563299e-01
8.63952875e-01 -1.66943416e-01 -3.81573230e-01 -4.42799367e-02
-2.02796876e-01 -1.62526801e-01 8.66004944e-01 -7.95812607e-01
-8.53997171e-01 1.67350933e-01 -4.63172972e-01 -6.01041317e-01
-2.14685947e-01 1.26895621e-01 -7.78308988e-01 2.12978330e-02
5.70057333e-01 5.68305552e-01 1.18001521e-01 -7.60249436e-01
4.26802248e-01 1.41240442e+00 1.27465403e+00 -2.80737877e-01
1.23289382e+00 2.42910907e-01 -6.28013790e-01 -1.15046227e+00
-1.22557499e-01 -4.47568208e-01 -5.31603515e-01 -3.18045332e-03
2.79389590e-01 -8.17524910e-01 -8.19828808e-01 7.11401224e-01
-1.33083081e+00 4.11721677e-01 -1.51689842e-01 5.10418236e-01
-3.13319653e-01 4.81135339e-01 -4.75572079e-01 -6.97881997e-01
-4.14694011e-01 -9.79920983e-01 6.04914367e-01 4.07029167e-02
-3.81540775e-01 -6.70437276e-01 3.18212733e-02 5.83820283e-01
2.93069899e-01 -2.01545313e-01 1.11788559e+00 -1.18506086e+00
-1.59561217e-01 -4.81468618e-01 3.00141960e-01 7.89381206e-01
6.36502504e-01 1.58405006e-01 -1.47331357e+00 -3.94687951e-01
5.51162779e-01 7.48969540e-02 5.35020113e-01 -2.11244971e-01
5.99478424e-01 -8.02415729e-01 -7.84015134e-02 7.66194344e-01
7.75365710e-01 4.05606478e-01 5.78293979e-01 -1.91211343e-01
6.69915318e-01 5.38099766e-01 -2.28545815e-02 1.23522840e-01
-7.48534799e-02 5.04844069e-01 3.52152407e-01 2.58834600e-01
-2.62644202e-01 -6.88437104e-01 8.40847313e-01 1.14232302e+00
2.42904603e-01 -2.17277750e-01 -4.98246074e-01 4.15405452e-01
-9.15386856e-01 -1.08148289e+00 6.21010400e-02 2.18769121e+00
9.09546852e-01 8.20523351e-02 -5.30514419e-02 7.93626308e-01
1.06593418e+00 1.24415763e-01 -3.80334049e-01 -7.06467748e-01
1.56772405e-01 2.03698799e-01 1.92279071e-01 7.85552680e-01
-9.40637052e-01 8.28890562e-01 5.86182070e+00 7.66529739e-01
-1.54355705e+00 -1.64045379e-01 8.92518554e-03 5.07327579e-02
-2.89554000e-01 -3.41741256e-02 -7.28313267e-01 2.47731119e-01
1.09597886e+00 -3.98298413e-01 8.35835278e-01 6.23938382e-01
6.95166811e-02 6.24175489e-01 -1.37634027e+00 9.74340260e-01
3.28960925e-01 -8.33475709e-01 8.43993053e-02 2.35256061e-01
4.16244008e-02 -3.97467256e-01 1.95454299e-01 3.43448132e-01
4.23747674e-02 -1.20877516e+00 8.03245723e-01 1.45399675e-01
9.00546789e-01 -9.69581544e-01 4.27922696e-01 5.39111137e-01
-1.02244806e+00 -5.98505251e-02 -1.70094937e-01 9.20006037e-02
-1.20440774e-01 1.74253315e-01 -1.48892045e+00 3.77913833e-01
2.16499746e-01 7.77892396e-02 -4.25439328e-01 7.15801954e-01
-4.51083750e-01 9.85530078e-01 -1.72352776e-01 4.42003220e-01
-1.45823151e-01 1.64498553e-01 1.06407571e+00 9.96703863e-01
4.56658721e-01 -1.45717084e-01 -2.69790918e-01 8.06680143e-01
-4.90842521e-01 3.91878374e-02 -6.27790689e-01 -4.67176944e-01
8.21938455e-01 1.00379765e+00 -1.90464601e-01 2.13483244e-01
2.77460992e-01 1.15445971e+00 -2.47774124e-01 3.39120716e-01
-6.57242358e-01 -6.91414714e-01 7.37787426e-01 1.52579889e-01
2.78811395e-01 -4.01265062e-02 3.00319493e-01 -9.76605415e-01
-1.64957885e-02 -1.38717210e+00 -7.53169227e-03 -6.17615998e-01
-9.60417211e-01 9.71617997e-01 -4.83452201e-01 -1.26091444e+00
-8.36207867e-01 -1.67269215e-01 -7.55193949e-01 1.07024014e+00
-1.18219948e+00 -1.05696535e+00 3.17866147e-01 8.79617035e-01
2.60451317e-01 -5.97331226e-01 1.12468779e+00 1.66201964e-01
-4.14971441e-01 8.73640656e-01 2.27292161e-02 6.10461950e-01
7.46122420e-01 -6.76426530e-01 2.30197236e-01 8.00748348e-01
4.85725671e-01 7.23820508e-01 7.08317935e-01 -3.93787473e-01
-1.36064100e+00 -9.56744313e-01 1.13766205e+00 -5.85378595e-02
4.60279584e-01 -5.86826086e-01 -1.09643698e+00 7.66230047e-01
4.41765524e-02 -2.27659732e-01 1.03155828e+00 -3.58460277e-01
-7.24047661e-01 -2.16321662e-01 -1.10927939e+00 2.32580408e-01
4.49476898e-01 -1.24903202e+00 -9.55568433e-01 -1.26525953e-01
7.34445214e-01 -6.96624666e-02 -5.06205082e-01 -1.55820251e-01
6.22738302e-01 -6.56945944e-01 1.02640295e+00 -5.86023808e-01
4.44651879e-02 -2.23048091e-01 -3.50503385e-01 -1.37464893e+00
1.51280716e-01 -1.02463484e+00 1.69730172e-01 2.07433105e+00
2.13861480e-01 -7.70998299e-01 3.86708647e-01 3.20931345e-01
2.12682188e-01 2.69629627e-01 -1.27280378e+00 -9.66547132e-01
8.15187693e-02 -7.14438617e-01 9.81455207e-01 9.84959900e-01
-7.29092630e-03 4.54643518e-01 -7.26312578e-01 6.36901438e-01
4.05910552e-01 1.59012392e-01 8.33543241e-01 -1.07497513e+00
-5.19585192e-01 -6.60969242e-02 -2.89252162e-01 -8.24293435e-01
4.59551603e-01 -1.08466756e+00 4.04939093e-02 -6.32813871e-01
-4.66678679e-01 -1.84189022e-01 -2.50366908e-02 4.15377051e-01
2.86395550e-01 1.01260491e-01 4.42340851e-01 3.68734181e-01
4.68313247e-01 5.74226141e-01 9.07319427e-01 -4.34950083e-01
-3.94587487e-01 7.79880524e-01 -8.98144245e-01 8.87925744e-01
7.14191198e-01 -9.59465086e-01 -4.43619043e-01 -1.31603733e-01
-3.15585494e-01 5.21332383e-01 4.82938081e-01 -8.66244435e-01
3.66017759e-01 1.93127036e-01 7.63047114e-02 -1.85291648e-01
5.18002570e-01 -1.02774966e+00 2.72446930e-01 3.56702238e-01
-3.71037573e-01 -2.50263631e-01 6.65141121e-02 4.24403310e-01
-4.01274770e-01 -5.50312698e-01 7.56768823e-01 2.37635478e-01
-1.85647458e-01 -1.06888466e-01 -8.14343095e-01 -3.56747909e-03
5.68399966e-01 8.23261440e-02 -1.05913775e-02 -8.60183299e-01
-7.15532839e-01 -2.92100787e-01 2.33668700e-01 6.83347046e-01
7.62804210e-01 -1.14106977e+00 -6.92008972e-01 8.70409667e-01
-2.03670576e-01 -5.90048909e-01 -3.20071802e-02 3.26217771e-01
-1.52491465e-01 3.84300083e-01 -3.91158760e-02 -2.24820092e-01
-1.87058699e+00 5.22225618e-01 4.79296088e-01 2.79166222e-01
-4.43119556e-01 7.08530366e-01 -1.23785719e-01 -6.74822569e-01
3.66340339e-01 -2.43934952e-02 -1.71035945e-01 2.48134375e-01
6.84815943e-01 3.42420191e-01 1.42933041e-01 -1.08408916e+00
-5.29223025e-01 1.93569615e-01 -1.28943920e-01 -3.92675936e-01
1.12889230e+00 -2.82899618e-01 -2.98527367e-02 3.07333559e-01
1.67037117e+00 8.59834671e-01 -7.02529848e-01 -1.92255273e-01
-3.87186080e-01 -5.60563564e-01 5.68652637e-02 -4.69156653e-01
-1.17565310e+00 9.55194652e-01 4.96843278e-01 2.53711522e-01
9.90373552e-01 -2.54986793e-01 9.36851084e-01 2.68230677e-01
3.64322722e-01 -7.15398312e-01 -1.21327080e-01 3.48469943e-01
9.19014096e-01 -6.16978109e-01 -4.87682283e-01 -4.92057264e-01
-5.87485075e-01 1.26254582e+00 7.10926801e-02 6.12691417e-02
8.64460051e-01 3.64890218e-01 4.18364227e-01 3.50903809e-01
-4.66076255e-01 3.25326979e-01 2.04506144e-01 9.25061941e-01
-6.77377880e-02 9.93622169e-02 4.83877599e-01 9.67685819e-01
-9.71783519e-01 -3.34621519e-01 1.03141987e+00 5.58404565e-01
-2.21811026e-01 -1.24423349e+00 -9.57990229e-01 -8.03252682e-02
-6.65923536e-01 -2.30578203e-02 -7.84795344e-01 6.43774152e-01
-1.21442311e-01 1.41272163e+00 2.01007277e-02 -7.68179059e-01
3.04970950e-01 4.58753943e-01 3.70453559e-02 -5.29971957e-01
-6.09985173e-01 2.87549615e-01 5.85800149e-02 -8.84921476e-02
-5.84807582e-02 -6.07547462e-01 -1.42344987e+00 -4.97861117e-01
-3.38031769e-01 3.90673041e-01 7.88146198e-01 8.83206964e-01
1.12967655e-01 4.11235780e-01 1.23406672e+00 -4.65502918e-01
-9.50318158e-01 -6.47269845e-01 -9.88221109e-01 2.20196933e-01
7.94377685e-01 -2.80092210e-02 -7.97878921e-01 2.00215951e-01] | [14.076655387878418, 5.8930206298828125] |
87ecb998-6156-4fe0-922c-97c0a5005ff1 | kliep-based-density-ratio-estimation-for | 2105.12549 | null | https://arxiv.org/abs/2105.12549v1 | https://arxiv.org/pdf/2105.12549v1.pdf | KLIEP-based Density Ratio Estimation for Semantically Consistent Synthetic to Real Images Adaptation in Urban Traffic Scenes | Synthetic data has been applied in many deep learning based computer vision tasks. Limited performance of algorithms trained solely on synthetic data has been approached with domain adaptation techniques such as the ones based on generative adversarial framework. We demonstrate how adversarial training alone can introduce semantic inconsistencies in translated images. To tackle this issue we propose density prematching strategy using KLIEP-based density ratio estimation procedure. Finally, we show that aforementioned strategy improves quality of translated images of underlying method and their usability for the semantic segmentation task in the context of autonomous driving. | ['Federico Tombari', 'Artem Savkin'] | 2021-05-26 | null | null | null | null | ['density-ratio-estimation'] | ['methodology'] | [ 3.96939427e-01 6.83967948e-01 2.66610265e-01 -4.14193869e-01
-8.48551989e-01 -5.94749629e-01 8.53058279e-01 -3.56855750e-01
-6.94151998e-01 1.20229483e+00 -2.97515899e-01 -2.31392041e-01
1.18377224e-01 -7.36748517e-01 -1.19629097e+00 -5.78891337e-01
6.10942066e-01 7.86612093e-01 4.57307875e-01 -3.25121373e-01
1.52846221e-02 1.91623405e-01 -1.24253595e+00 -6.29528016e-02
1.15407956e+00 6.40377045e-01 1.12503089e-01 6.98352218e-01
-3.49448860e-01 5.30960500e-01 -7.43744612e-01 -4.69127208e-01
5.85141003e-01 -6.00961745e-01 -5.02082288e-01 -4.57846411e-02
4.70773727e-01 -2.25904420e-01 -3.73989910e-01 1.23497069e+00
2.05874234e-01 1.91032276e-01 1.01392245e+00 -1.42987645e+00
-5.65439820e-01 3.58994514e-01 -3.86464417e-01 1.67169079e-01
-1.49886236e-01 3.30510736e-01 1.80365518e-01 -7.16547132e-01
6.77953839e-01 1.20949757e+00 6.58228278e-01 6.28487587e-01
-1.19304395e+00 -5.06945193e-01 -3.50491375e-01 1.91917419e-01
-1.06472838e+00 -3.78338277e-01 9.54368472e-01 -3.19524556e-01
6.61425829e-01 -2.64569581e-01 2.41760880e-01 1.25503981e+00
4.45381366e-03 5.95089436e-01 1.49961877e+00 -3.30548346e-01
4.37399894e-01 4.46138412e-01 -4.10894006e-01 6.47159576e-01
2.88845181e-01 1.50430262e-01 -1.86125219e-01 3.69771779e-01
5.51610291e-01 -4.79532272e-01 2.12940395e-01 -5.03641188e-01
-7.86451578e-01 1.01065302e+00 4.85010177e-01 4.65717129e-02
-1.35663211e-01 2.31637344e-01 4.79099751e-01 3.04701149e-01
5.45690536e-01 4.78372127e-01 -3.80576700e-01 -9.38972980e-02
-1.11785233e+00 4.95596021e-01 5.41954994e-01 1.11857498e+00
6.63618684e-01 6.84858143e-01 2.41949372e-02 8.87816906e-01
5.92196882e-02 8.15438926e-01 5.45779228e-01 -1.10915864e+00
3.78498703e-01 4.26289588e-01 1.80042684e-01 -6.86953127e-01
-4.70005646e-02 -3.02587748e-01 -5.47461152e-01 5.29315650e-01
6.28451169e-01 -2.28807345e-01 -1.34244871e+00 1.69475436e+00
3.43594372e-01 2.08858669e-01 4.16594893e-01 6.60938859e-01
4.37257290e-01 4.28259254e-01 6.89929649e-02 1.14347741e-01
6.57550752e-01 -9.45514143e-01 -4.33085591e-01 -4.36728209e-01
9.76426303e-02 -6.31202281e-01 9.99846160e-01 2.18276873e-01
-9.29288983e-01 -6.67615414e-01 -1.25165367e+00 6.18102849e-02
-5.18194318e-01 -1.28089741e-01 1.33516699e-01 9.26587880e-01
-8.52151573e-01 6.60626769e-01 -9.31884229e-01 -3.33867222e-01
8.07093799e-01 3.21009606e-01 -3.68176490e-01 -1.24291673e-01
-1.10231018e+00 1.17422366e+00 6.60291195e-01 -8.34730268e-02
-1.07228994e+00 -5.88492334e-01 -1.01481116e+00 -3.29956055e-01
2.88531214e-01 -7.14778304e-01 1.04606187e+00 -1.17167330e+00
-1.55580163e+00 9.18440580e-01 2.46036813e-01 -1.01898921e+00
1.09555161e+00 -2.57020295e-01 -7.19569847e-02 1.54556870e-01
1.41838297e-01 1.02759755e+00 1.26790941e+00 -1.31629860e+00
-3.07048231e-01 -3.45987976e-01 -2.52137691e-01 1.66500330e-01
2.49791503e-01 -4.30071801e-01 1.36348400e-02 -5.57371020e-01
-1.52727097e-01 -9.10339057e-01 -2.55886227e-01 -9.65468958e-02
-3.01471621e-01 2.51471251e-01 1.16427863e+00 -7.04191983e-01
1.36557937e-01 -1.85381401e+00 1.16381191e-01 -5.13967350e-02
-1.44551083e-01 6.17777348e-01 4.68655862e-02 1.11427709e-01
2.25096494e-01 1.06072491e-02 -8.35344613e-01 -3.23310167e-01
1.96281821e-01 4.46658283e-01 -3.79511684e-01 5.20084202e-01
5.71069181e-01 1.04419446e+00 -7.20172167e-01 -5.82305729e-01
6.34608030e-01 5.79881668e-01 -4.50079560e-01 2.91215062e-01
-4.95861679e-01 7.17967093e-01 -1.70015022e-01 2.32262939e-01
1.03640652e+00 5.76507986e-01 -3.18581939e-01 8.12715571e-03
3.16693246e-01 -2.63559669e-01 -5.30478895e-01 1.80991673e+00
-5.97410023e-01 7.08046436e-01 -1.03182770e-01 -1.42311406e+00
9.94924903e-01 -1.21510103e-01 -9.26597975e-03 -6.64902985e-01
1.05471149e-01 1.82906866e-01 7.09987769e-04 -3.07092190e-01
4.00484115e-01 -4.29797977e-01 -7.83098936e-02 -1.40262088e-02
4.00819063e-01 -8.41835380e-01 -9.65840369e-03 4.88039032e-02
9.77578938e-01 7.48823702e-01 5.13462387e-02 -1.27038762e-01
4.51506585e-01 3.26324821e-01 2.07466304e-01 7.87508309e-01
-5.55427015e-01 8.43031287e-01 4.21406984e-01 -9.85767171e-02
-1.81603980e+00 -1.18469203e+00 -1.56467646e-01 5.73604286e-01
1.22477166e-01 4.46973294e-01 -1.17813766e+00 -8.90012264e-01
-1.06374808e-01 9.84687984e-01 -6.95949137e-01 -5.14966667e-01
-5.71900964e-01 -7.33772159e-01 8.28989744e-01 3.53840292e-01
8.96171272e-01 -1.05228722e+00 -6.03245795e-01 1.80723503e-01
1.82502359e-01 -1.59487247e+00 7.59738907e-02 2.04854935e-01
-8.16893399e-01 -9.96374726e-01 -8.58586788e-01 -5.13895810e-01
5.48134089e-01 -3.42661291e-01 9.58024085e-01 -6.09269559e-01
-3.54229599e-01 1.75109625e-01 -3.43326420e-01 -4.53017712e-01
-1.00717092e+00 1.82316110e-01 -2.10673422e-01 -2.08943278e-01
3.36656392e-01 -7.53465533e-01 -6.29883945e-01 2.02966243e-01
-9.54065382e-01 -1.30718276e-01 6.07201576e-01 8.37891340e-01
3.55432123e-01 -7.85849020e-02 8.79659593e-01 -1.13187754e+00
6.17974281e-01 -3.80532563e-01 -6.09772980e-01 -4.36610654e-02
-7.12788105e-01 4.50547308e-01 9.64010715e-01 -3.06844682e-01
-1.16763461e+00 1.46313161e-01 -3.55004877e-01 -6.45294189e-01
-4.37817454e-01 -2.56701112e-01 -2.61246394e-02 -1.57975614e-01
9.40786600e-01 4.36594605e-01 1.02721289e-01 3.06025501e-02
4.21931028e-01 3.97321910e-01 8.41495097e-01 -3.93056840e-01
9.44025636e-01 5.07283211e-01 2.14323074e-01 -6.36713684e-01
-4.98958528e-01 6.55441582e-02 -5.48146486e-01 -1.79084644e-01
9.75688159e-01 -7.54207075e-01 5.39407656e-02 5.49414456e-01
-9.16280150e-01 -4.48434740e-01 -3.38782370e-01 2.92544663e-01
-1.03405535e+00 1.70854047e-01 -1.98726416e-01 -8.67790818e-01
-6.28867820e-02 -1.19066775e+00 1.04666805e+00 3.12149018e-01
1.00067139e-01 -9.90363598e-01 1.34178787e-01 5.19314289e-01
5.23780048e-01 6.61619842e-01 7.82173276e-01 -7.86723971e-01
-3.56309175e-01 -1.66382015e-01 -2.93893516e-01 7.83891439e-01
-5.37240319e-02 -1.34574503e-01 -1.13605261e+00 4.44651768e-02
9.27319303e-02 -5.44092417e-01 1.00707734e+00 3.43613863e-01
7.82508016e-01 1.54653029e-03 -7.60640875e-02 5.08393049e-01
1.59816992e+00 1.00961901e-01 7.68415689e-01 2.15454146e-01
8.68514836e-01 6.06376946e-01 5.25233209e-01 -9.33724567e-02
1.48070887e-01 5.18300951e-01 4.57708180e-01 -2.74308166e-03
-3.87132436e-01 -5.06744802e-01 1.20391726e-01 1.02484934e-01
2.33213946e-01 -2.19888091e-01 -8.71104181e-01 7.36607969e-01
-1.86057770e+00 -5.76009512e-01 -3.37931626e-02 1.90306711e+00
5.31950116e-01 5.76146305e-01 4.85098325e-02 1.67314216e-01
7.07344770e-01 2.64441613e-02 -8.35721493e-01 -6.89947784e-01
-1.36818171e-01 4.65497494e-01 8.72903883e-01 6.23311281e-01
-9.60611522e-01 1.26761997e+00 6.36615324e+00 9.26886261e-01
-7.48517513e-01 4.09426361e-01 6.34208381e-01 2.79337853e-01
-3.28774989e-01 -1.12664051e-01 -3.31436723e-01 6.15714550e-01
9.65286374e-01 5.13879806e-02 1.94200709e-01 9.60900187e-01
1.56259567e-01 -4.50419515e-01 -7.69631028e-01 8.84187579e-01
2.00589463e-01 -9.58016396e-01 1.15601584e-01 -1.10119730e-01
9.71280992e-01 -5.10604158e-02 4.38706487e-01 4.35911357e-01
4.84848022e-01 -1.28784657e+00 4.96911079e-01 3.72863561e-01
6.86399817e-01 -9.94228780e-01 7.99935997e-01 3.19194019e-01
-4.13319975e-01 3.06439757e-01 -4.76782769e-01 2.31878728e-01
2.56419867e-01 3.90503913e-01 -1.19027066e+00 4.31608111e-01
2.06630677e-01 1.43973336e-01 -5.13212502e-01 5.76401591e-01
-9.34237391e-02 7.14697003e-01 -4.70069855e-01 2.45613590e-01
4.28417891e-01 -4.79202867e-01 7.22966254e-01 9.57749128e-01
1.65590525e-01 -5.60913801e-01 -2.63039410e-01 1.26675665e+00
3.69172804e-02 -2.63439059e-01 -1.14340532e+00 4.81995270e-02
1.43044055e-01 9.08896089e-01 -9.74367857e-01 -2.29202449e-01
-8.33818242e-02 1.25460565e+00 2.70874888e-01 3.65408361e-01
-1.07820129e+00 -2.32513607e-01 2.70907164e-01 2.64106572e-01
4.03083980e-01 -3.13076466e-01 -4.21802163e-01 -7.61063933e-01
3.91038246e-02 -5.98677218e-01 -2.60396153e-01 -8.05404365e-01
-1.03835475e+00 8.37398708e-01 2.67733485e-01 -8.76119018e-01
-4.93072659e-01 -5.49058735e-01 -5.29954374e-01 7.43102610e-01
-1.29424119e+00 -1.35048449e+00 -2.86516994e-01 5.81044674e-01
9.46489394e-01 -5.97258985e-01 4.52868968e-01 1.00608200e-01
-2.79547453e-01 4.97738123e-01 1.90861672e-01 -1.43246606e-01
6.64632082e-01 -1.43434334e+00 5.57727218e-01 1.10249984e+00
-1.05046679e-03 -9.62306038e-02 1.24437201e+00 -5.33852637e-01
-7.65113533e-01 -1.36836660e+00 4.10569347e-02 -5.11881053e-01
3.68244022e-01 -3.79916519e-01 -8.45559835e-01 5.05835176e-01
3.71625841e-01 1.33771703e-01 2.92750485e-02 -6.51690125e-01
-9.71551836e-02 -5.92162367e-04 -1.61523783e+00 3.98194283e-01
7.36196160e-01 -3.53580445e-01 -6.81635857e-01 2.36089855e-01
6.53892934e-01 -4.84087557e-01 -3.68981451e-01 6.10421896e-01
8.03518817e-02 -1.16465235e+00 8.79204631e-01 -5.05967796e-01
6.17507398e-01 -1.92051530e-01 -6.05485663e-02 -1.52144885e+00
5.15191555e-01 -4.32168573e-01 1.58684015e-01 1.09791124e+00
4.87934470e-01 -6.09446108e-01 1.01410031e+00 4.12787378e-01
-2.35338628e-01 -2.12905839e-01 -9.94949698e-01 -6.64285421e-01
4.94539291e-01 -2.14970201e-01 1.47830233e-01 5.44613242e-01
-6.94648862e-01 1.40212908e-01 -2.35058531e-01 -2.01444641e-01
8.08292210e-01 -4.81142133e-01 9.78560925e-01 -6.65162086e-01
-2.63871461e-01 -9.19288173e-02 -7.83929527e-01 -5.19042492e-01
5.25302589e-01 -6.51982546e-01 2.84531951e-01 -1.30022144e+00
-4.90393266e-02 -3.92991781e-01 1.50471404e-01 -1.05691172e-01
-1.47272319e-01 6.08860373e-01 2.87909545e-02 -1.54021829e-01
-3.45806420e-01 7.18733072e-01 1.16804564e+00 -2.20331475e-01
1.18868142e-01 -1.41130581e-01 -1.78928614e-01 6.68018639e-01
1.07377803e+00 -6.77661002e-01 -7.87843943e-01 -8.40203911e-02
-2.48344219e-03 -2.25325540e-01 5.67592740e-01 -1.32071841e+00
-1.35343835e-01 4.55118753e-02 4.06029046e-01 -2.28722885e-01
2.47804508e-01 -6.53781533e-01 1.01122521e-01 2.39970356e-01
-1.41801804e-01 -8.67130682e-02 4.15641308e-01 8.08144927e-01
-3.72173488e-01 -2.93723851e-01 1.25410020e+00 -3.24864209e-01
-6.41214848e-01 -3.54844406e-02 -2.01920405e-01 4.08603311e-01
1.29611182e+00 -3.27744365e-01 -1.39281988e-01 -4.24798369e-01
-8.08691800e-01 -1.80214152e-01 5.68626285e-01 2.76326030e-01
4.25214261e-01 -1.10193384e+00 -6.39666736e-01 4.48299460e-02
-8.11747015e-02 4.10885215e-01 1.39374584e-01 6.39918506e-01
-1.00987232e+00 1.60531595e-01 -6.15952373e-01 -6.14414334e-01
-8.03645790e-01 5.43001771e-01 6.57991171e-01 -8.85056928e-02
-5.97757638e-01 7.40613759e-01 1.68024406e-01 -4.72534120e-01
-1.81756318e-01 -3.41910660e-01 1.94770604e-01 -4.24966007e-01
-1.43997923e-01 2.28086293e-01 1.79096028e-01 -6.52743876e-01
-2.17205390e-01 3.50076884e-01 -1.96204931e-02 -5.76730847e-01
1.11131012e+00 -1.74635559e-01 4.10876364e-01 3.44350308e-01
1.21802402e+00 -2.68719316e-01 -1.72538984e+00 2.51114070e-01
-2.37536520e-01 -3.57247651e-01 8.96189436e-02 -7.93580472e-01
-8.91392946e-01 8.22915673e-01 8.70796382e-01 -6.60655438e-04
8.47245872e-01 -1.91466585e-01 7.91194141e-01 1.65122896e-01
1.99686661e-01 -1.15435994e+00 -1.04047887e-01 1.32753715e-01
6.57384813e-01 -1.76234889e+00 -3.14153582e-01 -2.01431975e-01
-8.62869382e-01 7.69360185e-01 6.89266205e-01 -7.56988227e-01
4.69433814e-01 4.04463053e-01 1.26401514e-01 1.14978790e-01
-3.13951999e-01 -3.07055175e-01 3.43004018e-02 1.09202075e+00
-1.17469400e-01 -1.51736953e-03 -2.68230319e-01 3.40962522e-02
-3.50440741e-01 6.29825965e-02 6.42262101e-01 7.00337052e-01
-2.44831756e-01 -1.07627034e+00 -2.44579643e-01 2.20865250e-01
-5.04347384e-01 -1.29069686e-02 -3.81394088e-01 1.02060354e+00
1.83413148e-01 6.58856869e-01 4.14086245e-02 -1.56707108e-01
1.96554706e-01 2.47496828e-01 9.39708114e-01 -2.72931606e-01
-2.39245638e-01 -2.53095895e-01 -4.36087176e-02 -2.87738711e-01
-6.29888892e-01 -4.59086865e-01 -1.05980623e+00 -1.06260784e-01
6.72462359e-02 -6.24079965e-02 9.71202075e-01 1.15125000e+00
7.89960963e-04 6.72213495e-01 2.87016571e-01 -5.44964552e-01
-5.07546008e-01 -1.19815946e+00 -4.31443363e-01 4.81237471e-01
3.35606515e-01 -7.43053734e-01 -3.67279977e-01 1.05002791e-01] | [9.839681625366211, 1.3023182153701782] |
238ee7cb-8dfb-480c-a514-c96ce831b0cd | lc-2-lidar-camera-loop-constraints-for-cross | 2304.0866 | null | https://arxiv.org/abs/2304.08660v1 | https://arxiv.org/pdf/2304.08660v1.pdf | (LC)$^2$: LiDAR-Camera Loop Constraints For Cross-Modal Place Recognition | Localization has been a challenging task for autonomous navigation. A loop detection algorithm must overcome environmental changes for the place recognition and re-localization of robots. Therefore, deep learning has been extensively studied for the consistent transformation of measurements into localization descriptors. Street view images are easily accessible; however, images are vulnerable to appearance changes. LiDAR can robustly provide precise structural information. However, constructing a point cloud database is expensive, and point clouds exist only in limited places. Different from previous works that train networks to produce shared embedding directly between the 2D image and 3D point cloud, we transform both data into 2.5D depth images for matching. In this work, we propose a novel cross-matching method, called (LC)$^2$, for achieving LiDAR localization without a prior point cloud map. To this end, LiDAR measurements are expressed in the form of range images before matching them to reduce the modality discrepancy. Subsequently, the network is trained to extract localization descriptors from disparity and range images. Next, the best matches are employed as a loop factor in a pose graph. Using public datasets that include multiple sessions in significantly different lighting conditions, we demonstrated that LiDAR-based navigation systems could be optimized from image databases and vice versa. | ['Hyun Myung', 'Woojoo Lee', 'Hyungtae Lim', 'Seungwon Song', 'Alex Junho Lee'] | 2023-04-17 | null | null | null | null | ['autonomous-navigation'] | ['computer-vision'] | [ 4.48354408e-02 -4.33701873e-01 -1.47652596e-01 -8.80398929e-01
-7.37914562e-01 -5.62948823e-01 4.35813934e-01 -9.97820869e-03
-7.01281726e-01 4.80914384e-01 -4.46968079e-01 -1.16606489e-01
-6.77132905e-02 -9.97949660e-01 -1.10565305e+00 -4.95148569e-01
1.56809583e-01 4.47142571e-01 1.94204152e-01 -2.44241655e-01
4.58070010e-01 8.92434359e-01 -1.65625107e+00 -3.31716269e-01
9.90590215e-01 9.70725060e-01 5.78470886e-01 2.17336148e-01
-3.52892578e-01 -4.40018475e-02 -3.49917173e-01 -1.78925231e-01
6.14951551e-01 4.53273244e-02 1.00541664e-02 -1.49730310e-01
1.01441848e+00 -4.50078368e-01 -4.61709261e-01 1.25183940e+00
6.54791951e-01 1.76672176e-01 3.76766264e-01 -1.40729630e+00
-6.56274974e-01 -5.49471863e-02 -5.96768677e-01 -2.99360424e-01
4.10971791e-01 4.41701077e-02 6.69170260e-01 -1.13032222e+00
5.92999279e-01 1.21411204e+00 7.60094583e-01 4.11920130e-01
-1.22389138e+00 -1.18141639e+00 2.86259335e-02 2.39315793e-01
-1.77067256e+00 -4.05645162e-01 9.59737420e-01 -3.93982112e-01
8.51525247e-01 -1.56057760e-01 6.65526867e-01 8.30070794e-01
3.45130831e-01 1.94769070e-01 7.98834503e-01 -1.84127331e-01
1.01398088e-01 2.31158137e-01 -2.99752235e-01 7.85123765e-01
5.39447725e-01 1.77062407e-01 -8.32609177e-01 2.26678014e-01
6.52671456e-01 4.08468753e-01 -1.57901317e-01 -1.10474610e+00
-1.18986881e+00 7.72693992e-01 1.02121854e+00 -4.54879785e-03
-1.52745575e-01 2.13804334e-01 -4.10715714e-02 2.30100796e-01
1.13711149e-01 3.39840770e-01 -2.16967419e-01 1.84391618e-01
-7.08183169e-01 1.03098214e-01 3.98030102e-01 1.30855799e+00
1.51901376e+00 -7.13821575e-02 5.44570863e-01 6.27909899e-01
5.65400898e-01 1.19900918e+00 3.45353365e-01 -9.91726875e-01
6.42004371e-01 8.27554405e-01 1.15056388e-01 -1.59105217e+00
-4.42285389e-01 -2.26016998e-01 -8.32684457e-01 3.83531809e-01
1.01110168e-01 2.64144212e-01 -8.66965652e-01 1.70709968e+00
2.54207641e-01 2.28841558e-01 -6.40176749e-03 9.11461890e-01
4.68577385e-01 3.09943527e-01 -3.92988980e-01 4.06821400e-01
9.73305881e-01 -5.46377659e-01 -4.79955107e-01 -6.46028519e-01
5.46508014e-01 -6.80481732e-01 9.12731349e-01 2.47951616e-02
-4.89576280e-01 -7.45903909e-01 -1.56899536e+00 -3.09292704e-01
-6.78183675e-01 3.69135067e-02 4.74050522e-01 4.57260698e-01
-1.06860375e+00 3.49810869e-01 -8.52358580e-01 -4.26376462e-01
2.71177799e-01 5.05393088e-01 -7.18454301e-01 -4.19766694e-01
-1.04238272e+00 1.06036043e+00 1.95480928e-01 2.70614803e-01
-5.22824109e-01 -4.57485408e-01 -1.34126031e+00 -3.41379881e-01
9.42850783e-02 -5.68791509e-01 7.33079016e-01 -3.15305650e-01
-1.32006443e+00 9.86435473e-01 -3.77420127e-01 -4.82087314e-01
2.99372405e-01 -3.64077002e-01 -2.83749491e-01 -1.96890831e-01
5.19923687e-01 9.12682116e-01 8.35954964e-01 -1.47808361e+00
-8.15905094e-01 -7.62032330e-01 3.76965962e-02 2.37541184e-01
7.06350133e-02 -7.99328566e-01 -5.90864122e-01 1.09068483e-01
9.76987898e-01 -9.73298490e-01 -1.72996238e-01 3.65378350e-01
-1.82023302e-01 1.41679198e-01 8.42118621e-01 -2.35973150e-01
5.36829948e-01 -2.26155806e+00 -1.22222483e-01 3.50717843e-01
1.00575380e-01 -2.27586970e-01 -1.97343886e-01 1.10839322e-01
2.09803239e-01 -5.19127473e-02 -7.09573328e-02 -5.62355995e-01
1.06467865e-01 3.57022822e-01 -3.61389399e-01 8.43119025e-01
1.96159422e-01 7.96528041e-01 -8.97822618e-01 -3.67370725e-01
5.85991144e-01 5.67309737e-01 -5.35363495e-01 -2.60709003e-02
1.40142605e-01 5.36949754e-01 -2.79928058e-01 7.71484256e-01
1.08924520e+00 6.25756755e-02 -2.79376775e-01 -5.08631349e-01
-3.85478318e-01 2.12025240e-01 -1.10720956e+00 2.29136515e+00
-7.25282013e-01 7.91118741e-01 -1.12562552e-01 -7.53920734e-01
1.39721060e+00 -3.53437930e-01 5.95446527e-01 -1.17364502e+00
8.03460404e-02 5.27568281e-01 -2.33449563e-01 -5.58401830e-02
6.67843223e-01 2.36933082e-01 -2.35082343e-01 -6.79499283e-03
-2.08938256e-01 -7.51304507e-01 -9.52202231e-02 -1.90912962e-01
7.64801502e-01 2.46559158e-01 4.36826870e-02 2.23791242e-01
5.13696849e-01 2.29350045e-01 6.07759237e-01 6.74470901e-01
-2.03504816e-01 7.27752328e-01 -2.33939543e-01 -5.41291296e-01
-1.10275543e+00 -1.30420458e+00 -2.72058755e-01 4.13380951e-01
8.93172562e-01 -1.83963954e-01 -2.05459118e-01 -3.02171230e-01
4.68644917e-01 2.88470596e-01 -1.78772122e-01 -3.71355742e-01
-7.50424385e-01 -9.23969746e-02 3.86288315e-01 2.88004637e-01
8.74294102e-01 -4.18034017e-01 -7.23068118e-01 1.07879825e-01
-6.26380071e-02 -1.15625095e+00 -4.79644477e-01 2.44326577e-01
-6.59328759e-01 -1.04992533e+00 -2.92168111e-01 -8.39808941e-01
7.56314754e-01 7.76190817e-01 6.64187253e-01 -1.41629249e-01
-1.23364463e-01 2.26389721e-01 -6.00825883e-02 -4.41468984e-01
1.46400303e-01 5.42186387e-02 4.34213400e-01 -2.33451247e-01
7.08147228e-01 -6.60371363e-01 -7.30025411e-01 4.81584162e-01
-4.97561365e-01 -1.55029207e-01 7.67668784e-01 8.81803870e-01
1.00478208e+00 -2.44284302e-01 4.29799408e-02 -3.31901848e-01
2.23764777e-01 -1.70272738e-01 -1.08806813e+00 -1.20842993e-01
-7.57636130e-01 1.05538808e-01 2.82072186e-01 -1.24972492e-01
-4.94234860e-01 6.47172928e-01 2.95707360e-02 -6.52495027e-01
-6.71039745e-02 4.18715984e-01 -2.92411417e-01 -5.33044875e-01
6.19969308e-01 2.76377857e-01 2.48561278e-01 -2.92363882e-01
3.36682051e-01 6.56081200e-01 7.16022074e-01 -2.53336996e-01
1.24309170e+00 7.82090068e-01 3.00952464e-01 -6.54666781e-01
-5.18117249e-01 -5.91250777e-01 -1.00848472e+00 3.51193212e-02
7.38317728e-01 -1.20393264e+00 -6.29324436e-01 2.38842204e-01
-1.19712162e+00 1.48633599e-01 6.48599640e-02 7.77442873e-01
-4.63676482e-01 2.69213706e-01 5.85622750e-02 -4.68737692e-01
-4.07007858e-02 -1.29212260e+00 1.20935321e+00 4.04793054e-01
2.07624227e-01 -5.55551231e-01 2.77357668e-01 1.14303589e-01
2.51700491e-01 1.72121689e-01 4.61447269e-01 -1.55204132e-01
-1.12393570e+00 -4.36435074e-01 -4.12812948e-01 1.34045303e-01
3.60774577e-01 -2.14327186e-01 -8.96671653e-01 -4.77576882e-01
-1.02824949e-01 -2.13355664e-02 6.45016015e-01 1.63305759e-01
7.80201733e-01 3.58971626e-01 -5.86652637e-01 1.20410132e+00
1.49006188e+00 4.20076400e-01 4.54766631e-01 8.25570643e-01
8.84438634e-01 4.44543540e-01 7.88301885e-01 1.35158375e-01
7.14610040e-01 7.49155045e-01 6.82564199e-01 1.20153138e-02
8.86829123e-02 -6.59303963e-01 2.79080212e-01 7.68899739e-01
3.72595310e-01 1.57179177e-01 -9.66054320e-01 4.10326928e-01
-1.63210094e+00 -5.93215942e-01 1.68950528e-01 2.36194491e+00
3.80570740e-01 9.19756666e-02 -6.70262575e-01 -3.11027199e-01
7.04438090e-01 1.78379416e-01 -8.77228498e-01 1.40791714e-01
-1.85897246e-01 3.35103758e-02 1.05185831e+00 5.44275761e-01
-1.01631653e+00 9.69452560e-01 5.09824657e+00 2.32432157e-01
-1.51527369e+00 -1.57262042e-01 -2.33253345e-01 6.15485571e-03
-3.01404059e-01 1.08780235e-01 -9.91353393e-01 3.85731637e-01
5.82126617e-01 -1.13254473e-01 2.40818992e-01 1.03208852e+00
3.78501089e-03 -8.28076676e-02 -1.20436895e+00 1.54126716e+00
1.52034774e-01 -1.12493336e+00 2.50003505e-02 1.65446207e-01
6.37241960e-01 4.16283846e-01 1.91122115e-01 3.29518437e-01
-4.98821400e-03 -8.16878259e-01 7.45910645e-01 4.49761927e-01
8.96931052e-01 -7.74392843e-01 7.49540687e-01 3.64842415e-01
-1.45724380e+00 1.70845896e-01 -7.93625772e-01 1.31820545e-01
1.58398300e-01 3.32257748e-01 -8.79366159e-01 6.47371948e-01
8.77485752e-01 9.92410302e-01 -6.64036155e-01 1.02357590e+00
-1.68339223e-01 -4.03156906e-01 -5.97829044e-01 1.40437514e-01
1.33981466e-01 -4.94064659e-01 3.38861763e-01 6.13440394e-01
7.77335048e-01 -3.88778448e-01 1.90049112e-01 9.89044309e-01
-9.85418558e-02 -6.86936229e-02 -1.21013689e+00 3.39802057e-01
9.51956391e-01 1.03029943e+00 -4.37125564e-01 7.77185559e-02
-4.06490535e-01 8.73746932e-01 3.66863579e-01 1.22536950e-01
-7.59606957e-01 -6.28296256e-01 9.60138023e-01 -5.70378229e-02
2.09861889e-01 -7.92730153e-01 -2.14727923e-01 -1.17031002e+00
1.08626373e-01 -4.38224345e-01 -2.42500126e-01 -8.72928381e-01
-1.00812268e+00 5.15435755e-01 -1.76390991e-01 -1.71207154e+00
-2.40690634e-01 -6.30279660e-01 -1.44447178e-01 1.11912549e+00
-1.88211071e+00 -1.00165474e+00 -8.35795879e-01 6.85711086e-01
2.76799053e-01 -3.45804058e-02 5.89226902e-01 6.40089393e-01
-2.88885325e-01 5.32688141e-01 2.11927518e-01 7.77997524e-02
1.00592971e+00 -1.04554725e+00 5.85266590e-01 8.54309499e-01
3.78146738e-01 9.11795616e-01 5.24864078e-01 -6.26270890e-01
-1.77737951e+00 -1.05365705e+00 7.05606878e-01 -4.67175901e-01
2.76492655e-01 -6.04315460e-01 -7.61326909e-01 6.22873783e-01
-3.18771422e-01 1.04018286e-01 2.28081062e-01 -1.65676862e-01
-3.74593198e-01 -5.58041990e-01 -1.03677785e+00 6.44391119e-01
1.25915694e+00 -8.71625066e-01 -3.60506147e-01 1.04624271e-01
7.80907452e-01 -8.93917739e-01 -5.64788580e-01 4.34827775e-01
6.91634595e-01 -9.21272397e-01 1.21265388e+00 1.83453918e-01
-3.15916725e-02 -6.83146536e-01 -5.90706110e-01 -1.22935057e+00
-2.33019553e-02 -6.49685711e-02 3.52357358e-01 9.58607614e-01
2.66686291e-01 -8.99621308e-01 9.99961495e-01 3.97372544e-01
-2.37049177e-01 -3.03815395e-01 -1.17495620e+00 -8.51287127e-01
-1.83813468e-01 -4.93877113e-01 7.84149230e-01 6.95438027e-01
-6.53158784e-01 2.80167490e-01 -2.27987051e-01 7.58451283e-01
8.50806057e-01 3.61679077e-01 1.29370999e+00 -1.28985476e+00
4.28591460e-01 -3.05613816e-01 -8.81286740e-01 -1.46238101e+00
3.97416294e-01 -9.36552763e-01 3.63814145e-01 -1.45884073e+00
-4.39787030e-01 -7.56551504e-01 -1.56230152e-01 1.38476208e-01
2.20729053e-01 3.30413967e-01 7.87524059e-02 4.65508878e-01
-2.70512462e-01 7.20477283e-01 1.04530621e+00 -4.54549462e-01
-2.38660246e-01 -5.73291555e-02 -2.78025329e-01 5.78973830e-01
8.43489230e-01 -4.51177150e-01 -4.68175441e-01 -8.98177862e-01
3.46557319e-01 -3.04749995e-01 4.26244169e-01 -1.28841674e+00
7.17845380e-01 -6.78810552e-02 4.61315095e-01 -1.28183413e+00
6.88891649e-01 -1.19732976e+00 2.50356466e-01 4.75420326e-01
8.95253122e-02 3.28732908e-01 7.47194663e-02 6.95018470e-01
-3.91527116e-01 -1.20606758e-01 5.15795112e-01 1.72477830e-02
-1.24198234e+00 6.58826590e-01 2.77804613e-01 -4.28440720e-01
8.91025782e-01 -6.28121018e-01 -1.68302432e-01 -2.90123820e-01
-5.93202375e-02 3.87038618e-01 8.64203572e-01 5.84777355e-01
1.07420802e+00 -1.57635295e+00 -1.91561356e-01 6.62997901e-01
5.16950011e-01 6.00917220e-01 1.45567864e-01 7.58753479e-01
-8.42124879e-01 5.15211701e-01 -3.62937659e-01 -1.14135063e+00
-9.92148161e-01 5.03813922e-01 4.50490534e-01 4.16452676e-01
-5.71296811e-01 6.55029416e-01 -2.68235267e-03 -9.72670674e-01
2.12531805e-01 -4.35230732e-01 8.32196400e-02 -5.87998517e-02
3.16294223e-01 3.93675901e-02 9.27311406e-02 -8.99098933e-01
-6.74346447e-01 1.22260785e+00 1.21526577e-01 -7.70037621e-02
1.14531660e+00 -6.94725931e-01 -4.59627137e-02 4.18252826e-01
1.54853642e+00 1.37734532e-01 -1.35283732e+00 -3.92690420e-01
6.20439909e-02 -8.61359537e-01 -6.42758235e-02 -2.79128198e-02
-1.02523947e+00 9.04110193e-01 1.04899371e+00 -2.44522378e-01
6.77653313e-01 -2.30070919e-01 7.08209932e-01 8.55439663e-01
9.03758526e-01 -8.68448257e-01 -5.35758436e-02 7.16810882e-01
7.35882759e-01 -1.79835188e+00 5.81243001e-02 -1.82444111e-01
-7.62568042e-02 1.10803473e+00 8.27343166e-01 -1.54790193e-01
5.41128159e-01 4.57901023e-02 3.11833858e-01 -2.65955031e-01
-7.91563615e-02 -2.51286328e-01 1.90649867e-01 9.36933219e-01
-9.86168385e-02 -1.71627685e-01 2.65396059e-01 -5.38345128e-02
-5.32994390e-01 -3.37119490e-01 6.56876266e-02 1.16766214e+00
-6.72002912e-01 -9.73062873e-01 -4.18428898e-01 1.07186101e-01
4.22012925e-01 7.89441243e-02 -1.59653232e-01 9.21119153e-01
3.69894892e-01 6.82203352e-01 2.73710757e-01 -8.28604817e-01
6.74275696e-01 -1.98976263e-01 4.62487668e-01 -6.09671593e-01
2.27923170e-01 -3.93467188e-01 -5.26637614e-01 -8.81344736e-01
-3.62537086e-01 -4.40337747e-01 -1.30655110e+00 -3.09703916e-01
-3.75910401e-01 -8.84563997e-02 1.28124964e+00 6.47729337e-01
5.24054110e-01 1.85057506e-01 7.70885587e-01 -1.25611770e+00
-4.75785881e-01 -5.22569537e-01 -4.37605351e-01 2.62863755e-01
5.11825204e-01 -1.01260340e+00 -4.52845693e-01 -2.36542732e-01] | [7.535645961761475, -2.207530975341797] |
606e06ec-5a62-4601-87f6-dafc1c6b4baf | tmu-japanese-english-multimodal-machine | null | null | https://aclanthology.org/2020.wat-1.7 | https://aclanthology.org/2020.wat-1.7.pdf | TMU Japanese-English Multimodal Machine Translation System for WAT 2020 | We introduce our TMU system submitted to the Japanese<->English Multimodal Task (constrained) for WAT 2020 (Nakazawa et al., 2020). This task aims to improve translation performance with the help of another modality (images) associated with the input sentences. In a multimodal translation task, the dataset is, by its nature, a low-resource one. Our method used herein augments the data by generating noisy translations and adding noise to existing training images. Subsequently, we pretrain a translation model on the augmented noisy data, and then fine-tune it on the clean data. We also examine the probabilistic dropping of either the textual or visual context vector in the decoder. This aims to regularize the network to make use of both features while training. The experimental results indicate that translation performance can be improved using our method of textual data augmentation with noising on the target side and probabilistic dropping of either context vector. | ['Mamoru Komachi', 'Masahiro Kaneko', 'Tosho Hirasawa', 'Hiroto Tamura'] | null | null | null | null | aacl-wat-2020-12 | ['multimodal-machine-translation'] | ['natural-language-processing'] | [ 7.95815825e-01 2.63932794e-01 1.35295004e-01 -4.73073572e-01
-1.30552685e+00 -5.64797163e-01 9.06358242e-01 -6.76737010e-01
-7.15246558e-01 9.76446688e-01 5.05752563e-01 -4.16634768e-01
8.38197589e-01 -3.04312348e-01 -9.80535746e-01 -7.53937542e-01
6.19236708e-01 7.65511751e-01 -5.89744262e-02 -3.04991722e-01
-2.28784740e-01 2.48251949e-02 -1.02872729e+00 7.88655162e-01
7.60070860e-01 8.82472336e-01 6.88742161e-01 6.81406260e-01
7.76530057e-03 3.57341290e-01 -5.55713654e-01 -6.77448571e-01
2.57295907e-01 -7.37417102e-01 -8.38738143e-01 3.53973866e-01
5.06871164e-01 -3.21640462e-01 -3.76519889e-01 9.07066643e-01
6.71324968e-01 3.72636057e-02 5.97935319e-01 -9.67435658e-01
-9.25126493e-01 7.20472455e-01 -4.72925603e-01 -1.61645845e-01
3.72015536e-01 3.37226987e-01 6.33784711e-01 -1.21766138e+00
1.00006354e+00 1.11010230e+00 2.02960029e-01 9.61025536e-01
-1.48898578e+00 -3.90223742e-01 -9.62055102e-02 1.22284099e-01
-1.24365962e+00 -7.72725701e-01 4.86867011e-01 -1.59131020e-01
9.49084401e-01 3.99120212e-01 1.86363146e-01 1.78856945e+00
1.10465549e-01 9.66426194e-01 1.08060586e+00 -7.89974809e-01
-8.32656622e-02 4.01825994e-01 -5.75965106e-01 6.04995012e-01
-3.62660140e-01 6.20542429e-02 -5.73558092e-01 2.14582026e-01
3.19881171e-01 -6.81221545e-01 -2.58043081e-01 -7.78297707e-02
-1.62138319e+00 5.32012641e-01 4.19958979e-01 3.04402977e-01
-2.86583036e-01 1.63237363e-01 4.01838362e-01 5.46808600e-01
3.78785700e-01 4.50750649e-01 -5.19430995e-01 -2.10137814e-01
-9.46996093e-01 -2.62088299e-01 3.52161407e-01 1.28211379e+00
7.46028841e-01 1.30746933e-03 -7.06902921e-01 1.02728415e+00
2.43223935e-01 9.75302100e-01 6.72938645e-01 -7.40921497e-01
1.15167916e+00 1.69247001e-01 1.59230009e-01 -3.68877769e-01
-1.17970072e-01 -1.65271610e-01 -7.66285837e-01 -2.27037370e-01
2.21603379e-01 -5.22262156e-01 -1.33525026e+00 2.03587508e+00
3.30867693e-02 -2.96896607e-01 3.10197383e-01 8.96310568e-01
5.97201824e-01 7.96809971e-01 -1.02305086e-02 -2.33298510e-01
1.01277828e+00 -1.21297204e+00 -1.07884562e+00 -3.42572331e-01
6.90072417e-01 -1.19268394e+00 1.30130267e+00 3.71259451e-02
-1.03636479e+00 -6.75685048e-01 -9.60509479e-01 -3.23835403e-01
-4.56864357e-01 6.52089179e-01 8.49164277e-02 5.17967999e-01
-1.23377883e+00 1.06417544e-01 -7.54221857e-01 -4.58787531e-01
1.57635912e-01 5.22263944e-01 -5.51643431e-01 -3.70420367e-01
-1.36298859e+00 1.32534778e+00 3.16443622e-01 3.49055052e-01
-7.60540783e-01 5.69632929e-03 -9.94257390e-01 -2.82443523e-01
3.82170677e-01 -8.74789774e-01 1.41464007e+00 -1.46833003e+00
-1.66524076e+00 7.48745263e-01 -4.44529295e-01 -2.18702048e-01
7.13856936e-01 -1.49544001e-01 -3.33306551e-01 2.21056089e-01
2.14176774e-01 1.24082899e+00 1.07174146e+00 -1.50356877e+00
-4.81859863e-01 -8.53118077e-02 -3.84809941e-01 5.40723026e-01
-2.49130324e-01 -1.68540273e-02 -8.19774747e-01 -6.16970479e-01
-2.29911566e-01 -1.24337733e+00 -3.37345637e-02 -4.21820343e-01
-5.04250646e-01 1.93612471e-01 8.37410033e-01 -1.06880939e+00
9.29108977e-01 -2.02843189e+00 7.67219305e-01 1.57690525e-01
-2.37008050e-01 1.72487959e-01 -8.21915865e-01 3.11305732e-01
1.34180579e-02 7.37074688e-02 -5.33903003e-01 -8.27037752e-01
-1.34468880e-02 4.11375493e-01 -1.85334221e-01 9.97135341e-02
5.59781671e-01 1.40203619e+00 -6.35152102e-01 -4.37139332e-01
-1.47153307e-02 4.63123590e-01 -1.87227070e-01 3.59267473e-01
-4.23199147e-01 6.55770183e-01 -1.44085124e-01 5.14967561e-01
5.84077835e-01 1.47714719e-01 3.15939486e-01 -3.55076522e-01
9.94503051e-02 1.21079192e-01 -5.88016868e-01 2.05209732e+00
-4.41231847e-01 8.04424942e-01 2.39354987e-02 -5.71024418e-01
5.26314139e-01 5.82120538e-01 1.85284466e-01 -1.01832056e+00
3.16240489e-01 3.64039689e-01 5.28331287e-02 -7.83607423e-01
7.65010178e-01 -2.62546301e-01 -2.00199887e-01 3.10670674e-01
2.48498648e-01 -2.55034357e-01 1.74638420e-01 1.85189277e-01
7.45438159e-01 5.19063473e-01 -3.24653357e-01 2.23409310e-01
4.04435247e-01 1.33676454e-01 8.31676926e-03 5.88294983e-01
7.04245642e-03 9.52983975e-01 2.73914784e-01 1.04498111e-01
-1.41849387e+00 -7.67210841e-01 1.79697901e-01 1.00920856e+00
-2.68067300e-01 -2.85102185e-02 -9.87633049e-01 -9.41125512e-01
-4.59847420e-01 9.83408570e-01 -7.32583761e-01 -3.22271407e-01
-6.61219358e-01 -8.79814863e-01 6.94740474e-01 5.14234781e-01
4.24016207e-01 -9.55611110e-01 -1.54588312e-01 1.91619396e-01
-9.22678173e-01 -1.40524852e+00 -9.55801249e-01 4.18070048e-01
-6.84569836e-01 -3.56413275e-01 -1.04035032e+00 -9.47440386e-01
8.72280240e-01 7.22866319e-03 9.46664274e-01 -1.19995326e-01
2.32607514e-01 3.40718806e-01 -4.55902934e-01 -1.19910138e-02
-7.75470138e-01 4.08126026e-01 -1.09599508e-01 1.37067258e-01
1.49776116e-01 4.00264673e-02 -1.07090445e-02 3.77047777e-01
-1.10647547e+00 5.14374912e-01 8.67528200e-01 1.20180249e+00
4.15392369e-01 -5.47442734e-01 3.85066390e-01 -5.88115931e-01
5.43416262e-01 -1.26360163e-01 -2.64539659e-01 4.94861007e-01
-2.33539626e-01 1.73089042e-01 2.89089203e-01 -6.49635971e-01
-1.24290311e+00 3.86700839e-01 -1.71434537e-01 -3.61236244e-01
3.39274965e-02 6.20224714e-01 -5.76207817e-01 9.77878198e-02
4.77845877e-01 4.12580937e-01 1.94813870e-02 -3.85886937e-01
6.85721815e-01 9.22880173e-01 5.43873906e-01 -7.19320416e-01
7.72092402e-01 1.45525690e-02 -1.80382743e-01 -5.84903300e-01
-4.82982635e-01 2.44845767e-02 -9.33921814e-01 -1.77244365e-01
1.03036535e+00 -9.45645809e-01 9.22414362e-02 2.56443173e-01
-1.32616115e+00 -5.46987414e-01 -2.61715446e-02 5.32751322e-01
-7.23131120e-01 2.28967324e-01 -5.91135919e-01 -4.22832221e-01
-4.90221232e-02 -1.57429647e+00 1.25637126e+00 -1.52950406e-01
-1.32836476e-01 -6.73894286e-01 -4.21361104e-02 7.59438872e-01
4.58683938e-01 -1.64585382e-01 1.01091385e+00 -5.13488591e-01
-5.12177169e-01 -1.56157464e-01 -3.23432088e-01 4.76054460e-01
6.24890020e-03 -1.86198428e-01 -1.04061341e+00 -2.69688308e-01
-5.07034212e-02 -5.54487824e-01 9.19901073e-01 3.03417854e-02
6.30642116e-01 -3.04036438e-01 -1.15239672e-01 1.95557222e-01
1.10843682e+00 1.32182434e-01 7.14230001e-01 1.46278381e-01
8.24987650e-01 7.07784176e-01 4.82656866e-01 -2.94686437e-01
3.51782054e-01 1.00002444e+00 3.15204531e-01 -3.74019146e-01
-5.13416350e-01 -1.53743401e-01 6.78127229e-01 1.01319826e+00
-9.50793251e-02 -4.24657643e-01 -7.95538902e-01 5.16875803e-01
-2.02087235e+00 -6.07781827e-01 5.98046638e-04 2.19117355e+00
1.24702239e+00 -7.38817751e-02 -1.22950986e-01 -5.33871114e-01
7.06015348e-01 -3.05030823e-01 -1.14714503e-01 -4.28889900e-01
-5.43954372e-01 -6.43739849e-02 4.20078516e-01 8.86179686e-01
-9.55297232e-01 1.28351533e+00 6.11515713e+00 6.71106398e-01
-1.15920234e+00 3.97523344e-01 6.99858248e-01 -1.55064791e-01
-2.90688962e-01 -2.26023585e-01 -5.33526540e-01 4.59560037e-01
1.19114709e+00 2.99195021e-01 7.47087121e-01 1.87064037e-01
1.98447138e-01 -2.01638013e-01 -1.26703882e+00 5.98782659e-01
4.83033031e-01 -9.24403965e-01 3.24727982e-01 7.92023316e-02
1.02022767e+00 1.80562928e-01 2.71415472e-01 4.79688138e-01
-1.19709939e-01 -1.08618438e+00 8.28881979e-01 6.52934849e-01
1.08484542e+00 -5.09644330e-01 1.16033518e+00 8.74785557e-02
-6.55778289e-01 3.35170209e-01 -1.12520993e-01 3.40307146e-01
3.04508090e-01 1.10868052e-01 -9.23232317e-01 6.79495215e-01
2.15325847e-01 2.59303331e-01 -8.03505838e-01 6.32165313e-01
-4.00808781e-01 4.69635010e-01 -2.75394976e-01 1.11556709e-01
4.41735208e-01 -1.25608876e-01 4.44889992e-01 1.31123543e+00
5.21914184e-01 -2.76016325e-01 1.37041286e-01 6.35932148e-01
-4.32328343e-01 -7.73431733e-04 -6.47816539e-01 -5.48967607e-02
-7.52881840e-02 1.10855842e+00 -3.83103669e-01 -6.89614117e-01
-5.81671953e-01 1.55138457e+00 1.67099342e-01 6.99457765e-01
-8.34200799e-01 -7.37911165e-02 2.15544067e-02 -3.71777147e-01
4.21137124e-01 -1.61685437e-01 -4.55962926e-01 -1.42154121e+00
1.20670222e-01 -1.04649341e+00 -1.04999743e-01 -1.22520292e+00
-9.29217219e-01 1.02015042e+00 -1.49612039e-01 -1.22001040e+00
-4.69181418e-01 -6.25247002e-01 1.02347985e-01 1.31362903e+00
-1.28197634e+00 -1.64208019e+00 3.52619857e-01 5.45222938e-01
6.59115314e-01 -2.53832012e-01 7.98973501e-01 3.83012325e-01
-4.85934377e-01 7.20263243e-01 1.79941401e-01 6.99302107e-02
1.21036553e+00 -1.12609446e+00 2.36067802e-01 8.80338013e-01
8.52703676e-02 3.33938122e-01 8.07750046e-01 -8.74678135e-01
-1.46019351e+00 -1.11807847e+00 1.30210245e+00 -7.92181075e-01
6.02844119e-01 -5.55456877e-01 -7.19238222e-01 8.40083241e-01
9.58666921e-01 -4.11217898e-01 4.33213979e-01 -2.67811894e-01
-1.63874879e-01 1.60011157e-01 -1.04334605e+00 8.57088268e-01
5.03359258e-01 -6.43426538e-01 -6.19698644e-01 2.16240123e-01
8.41731369e-01 -5.69409847e-01 -5.40823817e-01 3.02770972e-01
4.20312941e-01 -1.26377344e-01 5.30738115e-01 -7.41229951e-01
7.37806201e-01 -3.54571849e-01 -6.13618016e-01 -1.63901532e+00
2.01417077e-02 -5.56166887e-01 1.13195531e-01 1.09285450e+00
1.06479228e+00 -1.13799728e-01 4.64093834e-01 5.71521938e-01
-2.01916903e-01 -3.18110287e-01 -1.29851413e+00 -4.69115287e-01
1.18741490e-01 -3.57251257e-01 2.64815271e-01 7.46582985e-01
-2.30463110e-02 8.86098444e-01 -8.41111004e-01 2.38905549e-02
2.36326396e-01 -2.94995368e-01 7.03231514e-01 -3.17765594e-01
-8.93329009e-02 -2.33087555e-01 1.44018099e-01 -9.02779818e-01
-4.07226719e-02 -9.47492719e-01 5.90362966e-01 -1.40129447e+00
2.87680566e-01 2.21247196e-01 -8.13238136e-03 7.70750523e-01
-3.90470058e-01 6.39726400e-01 4.15123343e-01 2.10260689e-01
-4.94245261e-01 8.48122776e-01 1.52258360e+00 -4.09905374e-01
3.06846946e-02 -1.84387207e-01 -3.79569441e-01 1.77326962e-01
6.71009600e-01 -3.36395979e-01 -2.20674187e-01 -9.59558904e-01
1.11320548e-01 1.15836553e-01 2.06215099e-01 -5.02347171e-01
3.09711434e-02 4.41356748e-02 5.27216792e-01 -5.93576789e-01
5.83535910e-01 -1.15947735e+00 -2.83869915e-02 2.17346191e-01
-5.61690867e-01 3.36151212e-01 4.47406679e-01 2.50204235e-01
-2.49090761e-01 -1.11909084e-01 7.20790267e-01 -6.34281188e-02
-2.41649285e-01 -2.51909286e-01 -7.11852312e-01 -2.55161136e-01
5.21171391e-01 1.26446888e-01 -4.24406111e-01 -4.30742919e-01
-1.01708603e+00 3.15308839e-01 4.24741030e-01 5.63399851e-01
4.66278613e-01 -1.72545171e+00 -8.08312178e-01 1.32145569e-01
2.52565593e-01 -4.57865655e-01 6.05180636e-02 1.06547296e+00
-2.07168341e-01 4.57393050e-01 -2.53060102e-01 -5.31940162e-01
-1.26260924e+00 6.04291975e-01 1.63317665e-01 -1.75678775e-01
-1.44008711e-01 6.48760021e-01 2.08635684e-02 -6.25487208e-01
2.89996177e-01 -1.37854040e-01 6.67946860e-02 9.12058875e-02
3.48860413e-01 -7.60724694e-02 2.10245475e-01 -1.04476500e+00
-1.52060568e-01 2.89288431e-01 -1.91846102e-01 -9.80657399e-01
1.07187629e+00 -6.28672957e-01 -8.21726918e-02 4.82542366e-01
1.26691449e+00 -1.20277308e-01 -1.07696342e+00 -4.01531577e-01
-1.20139636e-01 -1.41826481e-01 -4.71730977e-02 -1.58813894e+00
-8.87234926e-01 9.96134877e-01 7.44923830e-01 -2.90470511e-01
1.15843809e+00 -1.11493282e-01 7.17560768e-01 5.60485840e-01
3.53446044e-02 -1.34996819e+00 2.24895496e-02 7.31373191e-01
1.25425112e+00 -1.49048305e+00 -3.97699863e-01 -1.58297136e-01
-1.07811368e+00 1.02673328e+00 6.52563334e-01 4.90744978e-01
-8.96365941e-02 1.38670698e-01 4.76742148e-01 3.14379573e-01
-8.96653652e-01 -2.56762177e-01 6.36489153e-01 6.49666905e-01
3.65821987e-01 7.11445883e-02 -3.46380830e-01 3.04557651e-01
-3.83440256e-02 -3.61234248e-02 4.80584949e-01 8.49576354e-01
-1.06976420e-01 -1.31772792e+00 -6.03115857e-01 2.05999359e-01
-1.73082352e-01 -4.99855310e-01 -7.93713033e-01 5.08258522e-01
2.93636560e-01 1.00721967e+00 -2.60546476e-01 -5.76465011e-01
4.24399257e-01 4.54815567e-01 5.88949740e-01 -6.19764566e-01
-5.31740844e-01 6.33527040e-01 4.67622012e-01 -2.76275873e-01
-4.33486223e-01 -7.50662446e-01 -6.81533158e-01 1.27283812e-01
-2.16731831e-01 2.38343067e-02 9.33556318e-01 1.09631968e+00
2.03847423e-01 5.60168266e-01 5.18121362e-01 -1.00081837e+00
-4.24954295e-01 -1.33850312e+00 3.47590111e-02 3.04749250e-01
3.25301349e-01 -3.22144091e-01 -7.52038583e-02 5.29939651e-01] | [11.462691307067871, 1.5247068405151367] |
f23c1b4b-a490-4c2c-bf3f-b99c0772222b | analysis-of-the-fed-s-communication-by-using | 2306.04277 | null | https://arxiv.org/abs/2306.04277v1 | https://arxiv.org/pdf/2306.04277v1.pdf | Analysis of the Fed's communication by using textual entailment model of Zero-Shot classification | In this study, we analyze documents published by central banks using text mining techniques and propose a method to evaluate the policy tone of central banks. Since the monetary policies of major central banks have a broad impact on financial market trends, the pricing of risky assets, and the real economy, market participants are attempting to more accurately capture changes in the outlook for central banks' future monetary policies. Since the published documents are also an important tool for the central bank to communicate with the market, they are meticulously elaborated on grammatical syntax and wording, and investors are urged to read more accurately about the central bank's policy stance. Sentiment analysis on central bank documents has long been carried out, but it has been difficult to interpret the meaning of the documents accurately and to explicitly capture even the intentional change in nuance. This study attempts to evaluate the implication of the zero-shot text classification method for an unknown economic environment using the same model. We compare the tone of the statements, minutes, press conference transcripts of FOMC meetings, and the Fed officials' (chair, vice chair, and Governors) speeches. In addition, the minutes of the FOMC meetings were subjected to a phase analysis of changes in each policy stance since 1971. | ['Tomochika Sawaki', 'Yasuhiro Nakayama'] | 2023-06-07 | null | null | null | null | ['natural-language-inference', 'sentiment-analysis'] | ['natural-language-processing', 'natural-language-processing'] | [-3.49999577e-01 3.17898914e-02 -4.23727393e-01 -2.18506977e-01
-3.62140208e-01 -7.51747429e-01 7.84022450e-01 5.90871513e-01
-5.40861845e-01 7.84512103e-01 9.57213521e-01 -1.05460191e+00
1.52794972e-01 -8.49671602e-01 -1.01461336e-01 -4.36273664e-01
2.15592980e-01 2.78541058e-01 -1.33510426e-01 -5.46122551e-01
8.53732824e-01 2.54908055e-01 -9.45373654e-01 3.50053281e-01
3.34436715e-01 8.59031677e-01 -6.56708330e-02 4.27399248e-01
-3.29470932e-01 1.12381852e+00 -8.92277420e-01 -5.50059676e-01
-4.16152645e-03 -3.00991654e-01 -9.03606594e-01 5.04675023e-02
-3.07307899e-01 -7.79940858e-02 1.36820823e-01 1.31433523e+00
1.58082500e-01 -1.74980998e-01 4.25994098e-01 -4.57640618e-01
-4.87985462e-01 1.13541579e+00 -4.93277848e-01 8.06537926e-01
2.30024219e-01 -1.87123194e-01 1.32905340e+00 -6.18932962e-01
4.76106495e-01 1.35525072e+00 2.36457035e-01 -7.01325759e-02
-9.28487539e-01 -7.27528036e-01 4.13680464e-01 9.29767638e-02
-8.11041653e-01 -7.25729287e-01 7.16956854e-01 -7.84226775e-01
9.46703374e-01 2.81111568e-01 7.75493681e-01 8.49648058e-01
5.64460754e-01 8.28739330e-02 1.02908075e+00 -8.48323405e-01
3.54059249e-01 6.03495121e-01 2.84043700e-01 -1.81794927e-01
5.92208803e-01 -2.55905837e-01 -2.38946408e-01 -4.42361742e-01
1.95912682e-02 -1.67019159e-01 -5.26943207e-01 5.85313261e-01
-1.12404919e+00 1.19036758e+00 -2.18105480e-01 8.08142006e-01
-5.55813253e-01 -2.57426053e-01 7.52540052e-01 3.82340848e-01
9.09692943e-01 3.16601813e-01 -5.47272086e-01 -6.15599692e-01
-7.87904263e-01 3.33086699e-01 1.18711436e+00 1.08123846e-01
1.74761102e-01 1.16377525e-01 3.96532744e-01 7.28060544e-01
6.29869759e-01 5.45028985e-01 8.07005405e-01 -7.61041045e-01
8.31613839e-01 4.69860822e-01 3.57883155e-01 -1.43347657e+00
-1.62599474e-01 -3.24515074e-01 -6.00006759e-01 3.08407843e-01
2.62181908e-01 -3.93402725e-01 -3.22634071e-01 1.27278054e+00
-4.48483713e-02 -9.85340595e-01 2.26650625e-01 5.98782659e-01
1.64264187e-01 8.54008138e-01 4.04299572e-02 -7.70064712e-01
1.52868807e+00 -2.65372634e-01 -9.10027504e-01 -7.26862252e-02
6.13905907e-01 -1.16484189e+00 6.90092981e-01 5.45060277e-01
-9.38120067e-01 -1.83270559e-01 -8.89065623e-01 4.78295952e-01
-2.19980538e-01 -1.82380930e-01 2.74163008e-01 5.99184573e-01
-5.71408451e-01 5.83950818e-01 -7.17713594e-01 -1.83447823e-01
1.13036215e-01 -2.56957918e-01 1.14974529e-01 6.86686099e-01
-1.29539859e+00 1.11498046e+00 4.66674119e-01 3.24974246e-02
2.46029571e-02 -1.57345697e-01 -6.02091730e-01 1.69417247e-01
2.06845537e-01 3.74599397e-02 1.53449297e+00 -1.07343817e+00
-1.31082892e+00 6.78968251e-01 -5.80721274e-02 -6.40621841e-01
2.24943385e-01 1.28234932e-02 -7.82127023e-01 5.31364307e-02
1.73620179e-01 -4.40617949e-01 2.90018171e-01 -5.43397546e-01
-1.01538050e+00 -4.92597938e-01 3.48125622e-02 -1.15571782e-01
-3.51256102e-01 7.30533838e-01 2.90393651e-01 -1.08321726e+00
8.71716887e-02 -5.24276137e-01 -1.15549020e-01 -9.22771275e-01
-5.98964728e-02 -2.26986259e-01 7.17728198e-01 -1.04449952e+00
1.71835613e+00 -2.24365330e+00 -4.46041793e-01 2.61406302e-01
1.20483428e-01 -7.52838552e-02 7.98480809e-01 6.32281363e-01
-3.91706496e-01 5.66469014e-01 9.44885835e-02 1.18645333e-01
6.36715516e-02 7.62563646e-02 -9.99836862e-01 3.63381505e-01
-2.95484334e-01 3.44924957e-01 -4.76191759e-01 -2.18311921e-01
-6.88308701e-02 -8.00208077e-02 -2.83477187e-01 -3.04823369e-01
-1.34085529e-02 -6.82454482e-02 -5.44046462e-01 5.91552019e-01
1.48533076e-01 -1.99464485e-01 4.70177352e-01 2.37600990e-02
-8.11351597e-01 1.12402499e+00 -8.86043131e-01 5.57763100e-01
-1.24874815e-01 9.70581532e-01 -6.86975243e-03 -1.21961784e+00
1.05026805e+00 7.51858830e-01 2.30245203e-01 -7.90840745e-01
2.19105676e-01 1.75859153e-01 1.63467780e-01 -3.38094741e-01
8.41663063e-01 -6.25426233e-01 -2.28385285e-01 8.65603268e-01
-6.35832787e-01 1.67021438e-01 4.42789346e-01 2.73284316e-01
4.12999481e-01 -3.62765014e-01 8.32234025e-01 -5.36387861e-01
4.06431884e-01 2.46771857e-01 8.33156288e-01 1.42460857e-02
-8.03873092e-02 2.10686997e-01 8.46791089e-01 -7.29839981e-01
-1.00496733e+00 -1.98013663e-01 -3.06385458e-01 9.03865695e-01
-6.29037201e-01 -4.42736506e-01 -5.43537676e-01 -1.18794866e-01
-1.05536655e-01 1.08545136e+00 -6.36001229e-01 2.99046278e-01
-4.40672636e-01 -1.04769838e+00 8.63950625e-02 1.01049773e-01
5.63665032e-01 -7.93333650e-01 -1.25083923e+00 7.37632334e-01
-4.01577204e-01 -9.19337451e-01 -1.92610279e-01 3.05731446e-01
-6.80549145e-01 -1.21787369e+00 -5.36825299e-01 -1.97832495e-01
3.54896486e-01 -8.89422074e-02 9.67879474e-01 -2.61911958e-01
3.24082822e-01 -1.03019647e-01 -7.11692929e-01 -1.08747280e+00
-1.03815281e+00 -9.60598439e-02 4.72119041e-02 -2.30477508e-02
3.80493999e-01 -3.00357223e-01 -1.59430712e-01 1.89988762e-01
-7.40130007e-01 -1.42812520e-01 1.11978678e-02 5.58646560e-01
-4.08908166e-02 4.13961053e-01 8.44086885e-01 -1.00458837e+00
8.71063232e-01 -5.23245335e-01 -7.79052973e-01 2.59823292e-01
-1.11060786e+00 -2.80953139e-01 2.30559543e-01 1.03045426e-01
-1.37157333e+00 -8.97811532e-01 7.93124214e-02 3.82386059e-01
8.35816637e-02 1.27018595e+00 1.58553168e-01 5.93178511e-01
6.07661784e-01 -1.08289450e-01 -1.20961137e-01 -6.14455938e-01
-1.07267924e-01 9.70094204e-01 5.53920567e-01 -4.75145787e-01
4.38852638e-01 3.77313226e-01 -6.40550554e-01 -8.49922299e-01
-7.20229387e-01 -4.93991792e-01 -3.53758276e-01 -2.84834296e-01
6.40608490e-01 -8.92679989e-01 -3.69239718e-01 5.04248202e-01
-1.16534185e+00 4.79320548e-02 -2.45960236e-01 7.91835368e-01
-8.70708898e-02 5.85718036e-01 -5.41133046e-01 -1.11977780e+00
-3.85377258e-01 -8.64634097e-01 -2.49606911e-02 1.05670094e-01
-8.84071767e-01 -1.36003280e+00 3.99775624e-01 3.31536859e-01
4.06220555e-01 4.32455093e-01 1.15091133e+00 -8.53659809e-01
3.16528201e-01 -1.28862813e-01 6.67783022e-02 4.60514724e-01
7.70544887e-01 5.24343550e-01 -7.20955491e-01 -1.25226215e-01
7.91493356e-01 1.18669309e-02 6.08166754e-01 3.48432153e-01
3.04423392e-01 -8.39204669e-01 -3.03488541e-02 -1.66001096e-01
1.20849669e+00 7.76932418e-01 3.39473367e-01 1.02152622e+00
-2.55214989e-01 7.49851882e-01 4.50940579e-01 9.44871426e-01
3.51355851e-01 4.29076970e-01 4.83571552e-03 3.22746843e-01
5.41464686e-01 2.34348133e-01 7.01312006e-01 1.29228222e+00
-3.22716922e-01 -4.58341502e-02 -1.10121381e+00 7.24593639e-01
-1.60276854e+00 -1.25965929e+00 -1.71516448e-01 1.91276062e+00
9.44171846e-01 7.91601062e-01 1.05986111e-01 5.01762390e-01
5.65890193e-01 3.31318796e-01 1.08025797e-01 -9.22232807e-01
-2.85615742e-01 -5.46823777e-02 4.98021603e-01 4.54877526e-01
-6.76454306e-01 4.64486837e-01 6.36590242e+00 6.42904043e-02
-1.25227165e+00 -2.93228745e-01 9.96154845e-01 2.19422840e-02
-4.06381011e-01 3.53667140e-01 -8.03110301e-01 6.43398225e-01
1.11361206e+00 -9.03497338e-01 -1.71511695e-01 6.43435895e-01
8.71477783e-01 -6.90211356e-01 -6.85701013e-01 4.27757800e-01
-2.43351087e-01 -1.61762261e+00 4.29850444e-02 4.95797902e-01
5.15566885e-01 -7.93931074e-03 -3.26374397e-02 -6.20565861e-02
4.18708801e-01 -5.37946105e-01 1.27700472e+00 2.60790080e-01
3.58877093e-01 -7.18729913e-01 1.07894087e+00 4.45669234e-01
-8.55871797e-01 -3.30148250e-01 -4.56840903e-01 -7.69200087e-01
3.11538666e-01 9.32628155e-01 -8.98729503e-01 3.80405128e-01
9.65940773e-01 4.16421741e-01 -1.35417327e-01 3.21903378e-01
-8.01874045e-03 1.24799478e+00 6.80648535e-02 -7.60371909e-02
4.51250613e-01 -3.64203066e-01 4.31226641e-01 1.14709938e+00
1.74965054e-01 5.00684977e-01 -2.48797923e-01 3.87394369e-01
1.12898842e-01 4.13141400e-01 -3.16862911e-01 -5.92573404e-01
2.71959692e-01 7.83277452e-01 -9.16888833e-01 -5.07143378e-01
-7.37794399e-01 1.79551300e-02 -1.74349040e-01 4.52505022e-01
-1.44874305e-01 -2.56064296e-01 8.12122405e-01 3.77652287e-01
1.46694005e-01 -2.07577392e-01 -5.67273438e-01 -9.92900789e-01
7.26052560e-03 -1.18363011e+00 4.40251410e-01 -3.48077714e-01
-9.00181472e-01 4.33068544e-01 8.27436298e-02 -7.32220411e-01
-5.49536407e-01 -5.30876458e-01 -1.01127315e+00 1.09979713e+00
-1.30898380e+00 -1.73234358e-01 5.62756836e-01 5.14857136e-02
6.23686373e-01 -5.02875328e-01 6.27606213e-01 -7.57481158e-02
-5.85892200e-01 -7.41060823e-02 3.88847113e-01 5.21375775e-01
6.20597839e-01 -1.06772816e+00 4.64772016e-01 7.64432132e-01
1.81941420e-01 6.43975139e-01 1.00590646e+00 -7.46360242e-01
-4.69450593e-01 -5.50436676e-01 1.56175578e+00 -1.96726963e-01
1.12807620e+00 1.58390537e-01 -9.23069775e-01 5.32789052e-01
5.64208388e-01 -9.31630552e-01 8.83100748e-01 9.42882225e-02
-6.24646060e-03 -8.94226953e-02 -4.96392399e-01 3.13154191e-01
-2.97026277e-01 -5.02981603e-01 -1.39092815e+00 1.95359722e-01
4.69493806e-01 2.08070129e-01 -6.39988601e-01 -8.76978040e-02
5.01992285e-01 -7.42599368e-01 4.18926418e-01 -6.68998480e-01
4.91143435e-01 9.89291742e-02 -4.56112176e-01 -1.29468238e+00
-4.03453678e-01 -5.86694419e-01 4.36782002e-01 1.36689425e+00
6.09944046e-01 -1.04250824e+00 4.83141005e-01 8.17206383e-01
8.45702067e-02 -5.21429300e-01 -9.42880332e-01 -3.44823152e-01
3.87032956e-01 -5.43045521e-01 5.69571733e-01 1.23124051e+00
6.96206033e-01 1.78758860e-01 -2.75357440e-02 -1.29635096e-01
3.23868573e-01 2.61356831e-01 4.99563754e-01 -1.22624648e+00
-2.89912254e-01 -9.61040020e-01 -5.23167104e-02 -4.96994495e-01
-1.97379723e-01 -4.87736106e-01 -5.89032769e-01 -1.28866315e+00
9.06270817e-02 -8.43001381e-02 -3.01496655e-01 2.33019322e-01
1.02951124e-01 -4.68878865e-01 2.41720870e-01 6.07898772e-01
3.31238568e-01 2.07451776e-01 7.58366227e-01 -2.56043971e-01
-1.12445347e-01 1.81496337e-01 -1.15872216e+00 1.01220691e+00
1.00403070e+00 -3.98846507e-01 -2.09445044e-01 -1.99524254e-01
8.22480917e-01 9.41360891e-02 1.55253828e-01 -4.80634004e-01
2.52257854e-01 -5.40652394e-01 2.18959540e-01 -6.96289182e-01
-3.96832287e-01 -5.47188759e-01 2.76000082e-01 5.92283010e-01
-2.88966238e-01 4.91239876e-01 1.93465427e-01 1.31919652e-01
-7.73385048e-01 -4.33835119e-01 6.37642741e-01 -4.03015971e-01
-2.05288500e-01 -2.88376033e-01 -1.02928972e+00 6.74445257e-02
7.23475754e-01 -3.55146378e-01 -3.64746213e-01 -6.17557108e-01
-5.92254341e-01 8.27690363e-02 4.43764776e-01 3.30739856e-01
3.98464948e-01 -9.31583583e-01 -1.04209650e+00 -7.27333948e-02
-3.58562410e-01 -2.70542085e-01 -2.32972071e-01 8.44078839e-01
-7.00547576e-01 7.87213743e-01 2.50360277e-02 1.69586733e-01
-1.07156503e+00 2.37445161e-01 2.56130576e-01 -3.84852290e-01
-7.40199864e-01 4.73506361e-01 2.65967727e-01 4.79431540e-01
-2.64572874e-02 -6.71212912e-01 -5.92486203e-01 8.58996451e-01
7.67207026e-01 3.09167266e-01 5.75738996e-02 -7.61049569e-01
-1.27690107e-01 4.02160436e-02 -3.38906109e-01 -5.53166091e-01
1.57817805e+00 -3.93116236e-01 -3.36472094e-01 1.07663739e+00
8.93284857e-01 1.75874501e-01 -5.23703694e-01 -2.77318388e-01
6.74799502e-01 -4.33279544e-01 2.36491144e-01 -6.84082508e-01
-8.16368818e-01 6.94162667e-01 -6.89431280e-02 6.83383644e-01
7.02674270e-01 -2.89747268e-02 4.76557225e-01 3.28300774e-01
1.99602276e-01 -1.49570954e+00 -4.27218348e-01 4.48459685e-01
1.06166017e+00 -8.85357440e-01 4.09851551e-01 1.97671920e-01
-4.47379351e-01 1.36293900e+00 -2.07959771e-01 5.91661572e-01
9.92238283e-01 2.21412405e-01 5.07284343e-01 -1.64597794e-01
-1.02489877e+00 7.14606285e-01 -3.06392014e-01 -1.69618338e-01
6.97487950e-01 3.33774388e-01 -8.03592861e-01 6.47023022e-01
-5.93683422e-01 -2.69048810e-01 9.44981039e-01 8.51639509e-01
-8.98833215e-01 -1.08944416e+00 -9.07291710e-01 3.48252028e-01
-1.24352491e+00 -1.58094615e-01 -4.68102038e-01 7.83901036e-01
-2.75088996e-01 1.08788753e+00 1.07243232e-01 -2.27511041e-02
1.43941551e-01 3.46540302e-01 -4.93534982e-01 -5.43192744e-01
-4.29348886e-01 3.01116824e-01 6.00723207e-01 -2.97398889e-03
-5.49098492e-01 -1.16148388e+00 -1.15573049e+00 -5.99205852e-01
-3.08735043e-01 9.66742516e-01 7.76152551e-01 1.10707176e+00
-2.42886588e-01 2.41857082e-01 9.51504052e-01 -9.84373689e-02
-1.00806975e+00 -1.16434765e+00 -9.05634463e-01 2.53951661e-02
1.45593122e-01 -3.34349215e-01 -5.33140242e-01 4.81694251e-01] | [4.506572246551514, 4.366347312927246] |
beabc446-6198-4871-b2dc-5800a574f7d7 | learning-similarity-between-scene-graphs-and | 2304.0059 | null | https://arxiv.org/abs/2304.00590v1 | https://arxiv.org/pdf/2304.00590v1.pdf | Learning Similarity between Scene Graphs and Images with Transformers | Scene graph generation is conventionally evaluated by (mean) Recall@K, which measures the ratio of correctly predicted triplets that appear in the ground truth. However, such triplet-oriented metrics cannot capture the global semantic information of scene graphs, and measure the similarity between images and generated scene graphs. The usability of scene graphs is therefore limited in downstream tasks. To address this issue, a framework that can measure the similarity of scene graphs and images is urgently required. Motivated by the successful application of Contrastive Language-Image Pre-training (CLIP), we propose a novel contrastive learning framework consisting of a graph Transformer and an image Transformer to align scene graphs and their corresponding images in the shared latent space. To enable the graph Transformer to comprehend the scene graph structure and extract representative features, we introduce a graph serialization technique that transforms a scene graph into a sequence with structural encoding. Based on our framework, we introduce R-Precision measuring image retrieval accuracy as a new evaluation metric for scene graph generation and establish new benchmarks for the Visual Genome and Open Images datasets. A series of experiments are further conducted to demonstrate the effectiveness of the graph Transformer, which shows great potential as a scene graph encoder. | ['Michael Ying Yang', 'Bodo Rosenhahn', 'Wentong Liao', 'Yuren Cong'] | 2023-04-02 | null | null | null | null | ['scene-graph-generation'] | ['computer-vision'] | [ 6.76571190e-01 -1.07896902e-01 2.49753613e-02 -4.21229243e-01
-5.75232208e-01 -6.40390575e-01 6.92839146e-01 1.51592001e-01
-9.24639683e-03 3.45776498e-01 1.41107187e-01 -1.46982029e-01
-7.90943503e-02 -1.09528291e+00 -8.95379663e-01 -7.01061010e-01
3.09126914e-01 -1.99807193e-02 1.95577279e-01 -1.03311844e-01
2.65658379e-01 2.06512183e-01 -1.48478389e+00 4.15247560e-01
9.38512385e-01 8.34065318e-01 7.62472808e-01 5.65352798e-01
-4.69443589e-01 9.09053683e-01 -5.46709776e-01 -5.76542020e-01
1.59126252e-01 -8.07141304e-01 -6.90811634e-01 4.46775436e-01
5.95333576e-01 -1.34505741e-02 -4.16248083e-01 1.31437111e+00
3.55534434e-01 5.96496761e-02 5.86597621e-01 -1.51052809e+00
-1.02058566e+00 5.14902830e-01 -3.21880817e-01 6.86823130e-02
6.61002934e-01 2.97235668e-01 1.15177321e+00 -6.65776730e-01
8.11095417e-01 1.18053555e+00 2.14376643e-01 2.28043109e-01
-1.05884051e+00 -3.61208707e-01 2.56224006e-01 3.37979436e-01
-1.39245653e+00 -2.42222100e-01 8.94196749e-01 -4.49158400e-01
7.63953388e-01 4.11774427e-01 7.23661721e-01 7.99754083e-01
8.59029964e-02 6.56880498e-01 1.00035608e+00 -4.38514829e-01
6.49189726e-02 -3.45319025e-02 -2.08065510e-01 1.04089022e+00
2.49269858e-01 -1.67662814e-01 -5.01891971e-01 3.22618842e-01
7.92709649e-01 1.55256102e-02 -5.22000372e-01 -6.42903745e-01
-1.34294081e+00 6.25467658e-01 1.00810373e+00 2.14850813e-01
-2.76684582e-01 3.88389118e-02 3.71295661e-01 1.03519008e-01
3.02951604e-01 5.19868970e-01 1.43803015e-01 2.18458742e-01
-5.36221981e-01 -3.08521241e-02 4.75491524e-01 1.34162259e+00
7.77204156e-01 -3.86761688e-02 -6.81539476e-01 6.18959844e-01
2.26864427e-01 5.74831903e-01 3.12436551e-01 -5.44208825e-01
4.07004654e-01 1.04329920e+00 -5.17305911e-01 -1.45020473e+00
1.39074773e-02 -1.86921149e-01 -1.12129140e+00 -5.78433573e-01
1.55305713e-02 6.29568458e-01 -7.89800823e-01 1.71476245e+00
2.02288046e-01 2.84469724e-01 1.10219754e-01 9.01431441e-01
1.13905525e+00 8.16500068e-01 2.83026099e-02 -3.90518337e-01
1.37239707e+00 -9.94247735e-01 -5.85144758e-01 -2.31646582e-01
6.39592409e-01 -7.26374507e-01 1.52262223e+00 -1.05806664e-02
-8.29283416e-01 -8.53609681e-01 -1.11682045e+00 -1.55033380e-01
-4.79994029e-01 3.95564325e-02 6.50381029e-01 3.39304626e-01
-9.84869599e-01 2.79845715e-01 -3.54085714e-01 -5.18217444e-01
3.40694189e-01 -2.31238350e-01 -4.32793766e-01 -4.93285954e-01
-9.07182872e-01 4.51675773e-01 6.62091851e-01 1.44541502e-01
-9.29061174e-01 -4.24232602e-01 -1.18733490e+00 1.15339436e-01
2.98354477e-01 -9.32728171e-01 6.63667858e-01 -8.11101377e-01
-1.09166539e+00 1.07423556e+00 -4.03260812e-03 -1.36726975e-01
1.62776455e-01 3.52301598e-01 -2.66848624e-01 2.57213086e-01
3.27105522e-01 6.54741883e-01 7.84402788e-01 -1.27319288e+00
-4.90436316e-01 -2.17326149e-01 2.24347711e-01 4.23032582e-01
-2.72524178e-01 -1.44036546e-01 -7.98363566e-01 -5.84381878e-01
1.51414916e-01 -7.82436490e-01 -2.89114237e-01 -2.52807468e-01
-8.26725364e-01 3.22817974e-02 4.76477206e-01 -5.63433826e-01
1.09353638e+00 -2.20147443e+00 2.41904140e-01 2.32741877e-01
4.21386063e-01 1.08249366e-01 -5.46405792e-01 5.38944542e-01
-2.57761538e-01 -1.09992381e-02 -3.43671590e-01 1.18423812e-01
-1.20043561e-01 1.48012087e-01 -4.07761216e-01 1.17221214e-01
2.88031489e-01 1.23562706e+00 -1.21447670e+00 -5.57470083e-01
2.33933195e-01 1.95868865e-01 -5.24458230e-01 6.77980423e-01
-4.01658028e-01 3.82964790e-01 -4.34735566e-01 3.37644339e-01
6.08044982e-01 -7.82539606e-01 3.36650550e-01 -7.05719888e-01
2.21692383e-01 8.37403908e-02 -7.94611275e-01 1.88787460e+00
-2.32048750e-01 3.88965279e-01 -6.91710413e-01 -1.11036980e+00
1.14511955e+00 -1.92767695e-01 2.25353613e-01 -9.62796390e-01
-2.46734768e-02 -6.59202486e-02 -4.02760834e-01 -4.59262818e-01
5.39799869e-01 1.16812490e-01 -1.88092008e-01 2.08001509e-01
1.30055949e-01 -4.90471303e-01 4.42671895e-01 6.25135243e-01
1.15106070e+00 1.58843815e-01 3.48997563e-01 -1.59918979e-01
7.10022211e-01 2.52351090e-02 2.37999558e-01 5.72116733e-01
1.10477187e-01 6.61064565e-01 5.84232330e-01 -4.39091533e-01
-9.79070842e-01 -1.17505789e+00 3.76308769e-01 8.63037825e-01
5.48811674e-01 -8.12647581e-01 -8.10292602e-01 -7.43038118e-01
-3.62923533e-01 5.48754573e-01 -3.26289743e-01 -5.96518219e-01
-2.04949379e-01 -5.66665411e-01 3.78512889e-01 2.24214837e-01
6.81676328e-01 -9.77469742e-01 -3.02223444e-01 -1.24350302e-01
-5.95220029e-01 -1.57676220e+00 -7.12647915e-01 -2.92122096e-01
-4.43845898e-01 -1.24879372e+00 -4.95644212e-01 -1.16990280e+00
1.00101399e+00 7.39457607e-01 1.30718493e+00 2.90600449e-01
-3.13880563e-01 4.67280209e-01 -6.12457216e-01 -3.49767692e-02
-6.14191413e-01 -1.47383288e-01 -3.00148040e-01 1.87189132e-01
-3.16676572e-02 -2.81333715e-01 -5.57567537e-01 1.09847233e-01
-1.13707232e+00 9.23787236e-01 6.02370143e-01 7.72645295e-01
9.08683896e-01 8.03048462e-02 8.89818966e-02 -8.83251250e-01
6.17602468e-01 -2.47477919e-01 -8.13201249e-01 8.04467142e-01
-4.62256521e-01 2.54956901e-01 7.11379230e-01 -5.58468513e-02
-9.26180363e-01 3.64378355e-02 2.28080407e-01 -4.75781709e-01
2.07850143e-01 7.53473043e-01 -4.78741884e-01 5.51611967e-02
5.15147984e-01 7.17601717e-01 -1.02092199e-01 -1.25101162e-02
7.90929198e-01 4.79629070e-01 8.11100245e-01 -3.87101233e-01
8.25127721e-01 2.35197887e-01 2.82718003e-01 -7.75000691e-01
-1.02023172e+00 -3.79530847e-01 -4.78982329e-01 -3.03013980e-01
9.98281419e-01 -9.62948382e-01 -3.60782802e-01 2.69964308e-01
-1.28504455e+00 -1.80019692e-01 -2.72307187e-01 1.09954841e-01
-7.38983631e-01 6.56867921e-01 -3.01928133e-01 -3.51633728e-01
-2.54415393e-01 -1.28312159e+00 1.39228582e+00 5.30437604e-02
3.21749508e-01 -7.54623592e-01 -4.64024991e-02 2.41581753e-01
1.37540758e-01 3.56107444e-01 1.17333829e+00 -2.38313645e-01
-1.05963182e+00 -3.57914641e-02 -7.71110952e-01 3.12119097e-01
3.57862025e-01 -6.67256257e-03 -8.03150594e-01 -3.22232485e-01
-2.33836442e-01 -2.07414731e-01 7.14697063e-01 1.34544492e-01
1.31961191e+00 -3.03735435e-01 -3.04289937e-01 7.75558054e-01
1.66554534e+00 3.56059559e-02 9.05769587e-01 8.57464820e-02
1.17354882e+00 5.51296771e-01 4.90192592e-01 2.20023751e-01
6.16323471e-01 7.62268305e-01 3.12143594e-01 -1.21293314e-01
-5.18880486e-01 -8.38420689e-01 4.04077083e-01 1.22289383e+00
1.38182521e-01 -6.33082449e-01 -6.95345283e-01 2.49337062e-01
-1.68288291e+00 -8.24670672e-01 -1.94375396e-01 2.12103152e+00
5.72853029e-01 -3.88204083e-02 -2.37288103e-01 -1.13164343e-01
9.17558610e-01 2.46716663e-01 -3.35177332e-01 -6.81344979e-03
-3.58886212e-01 -1.23144113e-01 2.93992728e-01 1.76052019e-01
-9.58026409e-01 1.15172577e+00 5.98165369e+00 9.95882332e-01
-1.02054811e+00 -2.50849932e-01 6.86808169e-01 5.05574167e-01
-5.32014608e-01 1.93621904e-01 -3.62080604e-01 5.01433134e-01
5.70967853e-01 -5.90960979e-01 5.43797135e-01 8.19489658e-01
-1.52026638e-01 3.55789244e-01 -1.13996124e+00 1.41022789e+00
2.63779938e-01 -1.42842519e+00 8.75387728e-01 -3.01390439e-02
6.97553098e-01 -3.52414280e-01 -8.74253586e-02 2.30454296e-01
1.92346185e-01 -8.49363327e-01 5.84552705e-01 6.91097140e-01
1.09591556e+00 -2.44164392e-01 5.01448929e-01 1.01116695e-01
-1.59107304e+00 2.15323374e-01 -6.97733104e-01 9.98239070e-02
9.82162878e-02 5.83474159e-01 -1.00401330e+00 1.09339583e+00
5.04399300e-01 1.05518520e+00 -1.01272047e+00 8.94432664e-01
-3.74040395e-01 2.20941812e-01 2.22386658e-01 -7.08977059e-02
6.79952502e-02 -4.20858860e-01 2.89347857e-01 1.09940803e+00
3.72012436e-01 -6.02504704e-03 2.94767946e-01 1.04176354e+00
-1.26956910e-01 4.25338835e-01 -1.11236703e+00 -4.09341663e-01
3.24227005e-01 1.41885853e+00 -8.71573865e-01 -5.30708671e-01
-4.54861939e-01 1.28598487e+00 4.34211254e-01 2.02055812e-01
-8.72602344e-01 -1.72232971e-01 2.16681078e-01 -9.48938578e-02
7.04987347e-02 -1.47478834e-01 2.06081063e-01 -1.33162510e+00
1.24574110e-01 -9.47269678e-01 4.17897373e-01 -1.26336002e+00
-1.24155962e+00 8.26486528e-01 1.20364428e-01 -1.50444388e+00
-1.42371967e-01 -4.74260688e-01 -3.45849514e-01 4.57020372e-01
-1.32935762e+00 -1.37316084e+00 -9.68743682e-01 5.84217191e-01
4.51350957e-01 -2.15326875e-01 6.67674780e-01 1.11360818e-01
-5.17070889e-01 5.56797504e-01 -2.44827300e-01 6.74606413e-02
3.33456129e-01 -1.20971179e+00 7.35899687e-01 1.12419724e+00
4.78256941e-01 4.12207633e-01 4.28378999e-01 -6.02133512e-01
-1.72786820e+00 -1.47883940e+00 5.96546352e-01 -3.65286559e-01
4.65978026e-01 -3.73967201e-01 -7.88818300e-01 5.14203548e-01
2.68151648e-02 1.07393436e-01 4.55661446e-01 -4.55046147e-01
-4.98560876e-01 -2.12714046e-01 -7.23215997e-01 8.42878163e-01
1.69917178e+00 -6.87325776e-01 -2.74533778e-01 5.87830007e-01
1.17221332e+00 -3.24508339e-01 -7.50112593e-01 6.15126312e-01
2.27573514e-01 -1.04257703e+00 9.92297471e-01 -5.09428382e-01
6.54263198e-01 -3.87732118e-01 -4.80215281e-01 -1.31808662e+00
-5.14209390e-01 -3.50297064e-01 2.47747838e-01 1.33648264e+00
1.15898035e-01 -3.24534118e-01 5.03133893e-01 8.50585550e-02
-2.16306508e-01 -4.37983990e-01 -3.98487419e-01 -8.00631702e-01
-5.87124884e-01 -1.95298478e-01 8.74851465e-01 8.68698537e-01
-2.57677943e-01 6.75560117e-01 -1.99828431e-01 1.47217736e-02
5.69408000e-01 5.07588744e-01 1.08417284e+00 -7.99406171e-01
-1.57269135e-01 -4.65845674e-01 -8.93955350e-01 -1.15187037e+00
1.53705671e-01 -1.32428527e+00 1.77708492e-01 -1.66293204e+00
6.07807994e-01 -1.85213879e-01 -3.08701128e-01 1.46412417e-01
-4.01079863e-01 1.81030020e-01 3.23212802e-01 1.19441092e-01
-9.59421635e-01 8.79257083e-01 1.66764104e+00 -3.56067270e-01
2.14107022e-01 -6.65664256e-01 -8.15793872e-01 4.06752408e-01
4.41383511e-01 -1.04618356e-01 -8.32098603e-01 -4.63317424e-01
3.87588888e-01 6.55482337e-02 4.96049076e-01 -1.04668331e+00
8.86048526e-02 -3.83062512e-01 1.39457941e-01 -3.44061434e-01
6.44786805e-02 -5.49711347e-01 4.01004493e-01 3.34007740e-01
-3.69773149e-01 2.29312092e-01 -1.76819727e-01 6.48482442e-01
-3.35393816e-01 2.00823948e-01 5.18538535e-01 -2.92604655e-01
-1.14426589e+00 5.97258866e-01 3.32201928e-01 1.58044677e-02
1.03929877e+00 -2.67327219e-01 -5.24810195e-01 -3.28285277e-01
-2.28100792e-01 -2.44648382e-02 6.57019019e-01 5.90943277e-01
1.03380775e+00 -1.54244149e+00 -6.72375500e-01 4.03430492e-01
8.18342626e-01 -6.16179816e-02 2.28549600e-01 4.15765733e-01
-7.16353118e-01 3.73243243e-01 -2.36155093e-01 -7.59764731e-01
-1.24073195e+00 9.49325204e-01 1.79938287e-01 -2.51125932e-01
-5.22669733e-01 7.06106186e-01 8.57938528e-01 -3.56972337e-01
-1.45121813e-01 -4.75373834e-01 -1.75622836e-01 -3.90314579e-01
4.61870193e-01 -1.24270968e-01 -4.92998138e-02 -8.74626040e-01
-1.83598682e-01 5.69767237e-01 2.06470490e-01 3.20703536e-01
1.05028546e+00 -2.25432888e-01 -4.59188968e-01 2.85788894e-01
1.22992015e+00 -3.03203851e-01 -8.93136919e-01 -2.35466287e-01
-5.42972572e-02 -6.61993206e-01 -2.29062065e-01 -2.71065146e-01
-1.06077242e+00 6.68660164e-01 5.05561292e-01 4.38664079e-01
1.42357290e+00 2.33148605e-01 6.95146143e-01 2.99815148e-01
4.83741313e-01 -4.57199842e-01 6.04294240e-01 3.11646521e-01
1.10329807e+00 -1.16095650e+00 -3.88841901e-04 -9.44667816e-01
-7.11087942e-01 8.34323227e-01 6.46480441e-01 1.63734481e-02
2.16939434e-01 -2.95863926e-01 -2.35172316e-01 -5.66777825e-01
-5.92233479e-01 -5.84291995e-01 6.63867891e-01 6.59315526e-01
3.14918518e-01 1.81321874e-01 -1.55006096e-01 3.64111871e-01
-4.35462803e-01 -1.55778483e-01 3.17672551e-01 5.67456782e-01
-2.80396134e-01 -9.37962472e-01 1.28263190e-01 5.04638791e-01
5.71964234e-02 -1.70065671e-01 -4.26882505e-01 4.37438935e-01
-6.38194755e-02 8.46950889e-01 1.76574327e-02 -6.60238564e-01
4.08515841e-01 -2.74516851e-01 6.75262511e-01 -7.84907222e-01
8.88680443e-02 -1.77252769e-01 -1.39824137e-01 -7.33959079e-01
-5.00594914e-01 -1.37229010e-01 -8.86727870e-01 -4.97665107e-02
-3.20399553e-01 4.04720893e-03 4.10086840e-01 7.76952863e-01
4.66212660e-01 6.85478568e-01 8.28411579e-01 -4.83030170e-01
-1.11031570e-01 -6.57674253e-01 -5.15753269e-01 9.97971296e-01
-1.12829193e-01 -4.92570013e-01 -4.55060117e-02 4.04560417e-01] | [10.45053768157959, 1.5051616430282593] |
5112b365-1f0d-4ce1-96dd-e66233827b49 | semantic-scene-completion-combining-colour | 1802.04735 | null | http://arxiv.org/abs/1802.04735v1 | http://arxiv.org/pdf/1802.04735v1.pdf | Semantic Scene Completion Combining Colour and Depth: preliminary experiments | Semantic scene completion is the task of producing a complete 3D voxel
representation of volumetric occupancy with semantic labels for a scene from a
single-view observation. We built upon the recent work of Song et al. (CVPR
2017), who proposed SSCnet, a method that performs scene completion and
semantic labelling in a single end-to-end 3D convolutional network. SSCnet uses
only depth maps as input, even though depth maps are usually obtained from
devices that also capture colour information, such as RGBD sensors and stereo
cameras. In this work, we investigate the potential of the RGB colour channels
to improve SSCnet. | ['Teofilo Emidio de Campos', 'Andre Bernardes Soares Guedes', 'Adrian Hilton'] | 2018-02-13 | null | null | null | null | ['3d-semantic-scene-completion'] | ['computer-vision'] | [ 5.70352077e-01 3.75537127e-01 2.44709462e-01 -6.60245001e-01
-3.12037110e-01 -5.54945111e-01 5.96128166e-01 1.18079796e-01
-5.60440779e-01 3.64351034e-01 3.54975283e-01 -1.34399384e-01
3.10963899e-01 -8.82544935e-01 -7.71002293e-01 -1.02460772e-01
1.20430350e-01 4.97825146e-01 5.81535161e-01 1.05068646e-01
5.80820255e-02 8.37367535e-01 -1.66509748e+00 4.38507378e-01
2.10020050e-01 1.36490166e+00 5.33230543e-01 8.99571061e-01
-3.40179801e-01 7.24410176e-01 -2.93817390e-02 -4.85120825e-02
3.63542289e-01 -9.17292163e-02 -8.89991581e-01 3.23644936e-01
5.14273524e-01 -5.36353290e-01 -5.77459276e-01 7.85055995e-01
3.18951160e-01 -1.16881303e-01 3.57622236e-01 -1.17814875e+00
-1.24002941e-01 1.13548286e-01 -3.67529213e-01 3.88771528e-04
4.38381523e-01 1.89788342e-01 5.76126993e-01 -8.52237046e-01
8.02559376e-01 1.22023666e+00 5.86377501e-01 5.21534085e-01
-1.08725059e+00 -3.38555455e-01 8.33839402e-02 -1.84929740e-04
-1.27687871e+00 -3.22400898e-01 9.51408923e-01 -3.95884544e-01
1.16353047e+00 1.98246732e-01 1.09746933e+00 9.18448150e-01
6.76991884e-04 7.61577666e-01 1.43010795e+00 -1.89025819e-01
5.70545256e-01 -3.56367946e-01 -3.41404051e-01 6.22850895e-01
7.20386878e-02 6.18704446e-02 -6.48865461e-01 3.68737318e-02
1.18038893e+00 3.51303458e-01 -5.29111642e-03 -9.83974695e-01
-1.16045666e+00 7.42874742e-01 9.20880198e-01 -7.04802051e-02
-4.81685102e-01 5.16604304e-01 3.59346777e-01 -1.93263561e-01
7.08886385e-01 2.45487958e-01 -5.90291798e-01 -5.36702313e-02
-8.03778708e-01 1.55203357e-01 3.57828647e-01 1.07383740e+00
7.93804646e-01 -2.94579983e-01 -6.27128333e-02 2.98508465e-01
5.17398357e-01 4.94739860e-01 -1.05747133e-01 -1.35304999e+00
3.96955371e-01 7.77356923e-01 2.83103660e-02 -3.79052937e-01
-8.56635511e-01 -1.02126367e-01 -5.91556907e-01 4.00254637e-01
2.13723555e-01 3.08068275e-01 -1.44771397e+00 1.21421230e+00
4.77575153e-01 3.36567461e-01 6.03786074e-02 1.26464903e+00
1.26088643e+00 4.63909432e-02 2.08491832e-01 4.15095121e-01
1.21029758e+00 -9.56325352e-01 -3.58054608e-01 -4.40972298e-01
4.75360513e-01 -5.57995021e-01 8.95572841e-01 4.18994337e-01
-1.07218778e+00 -3.04666728e-01 -1.06019735e+00 -6.04095519e-01
-4.26560462e-01 -3.50618005e-01 1.15141141e+00 5.69944203e-01
-1.27782178e+00 4.72451091e-01 -1.18042612e+00 -3.33820283e-01
1.05777800e+00 1.20324291e-01 -6.26742959e-01 -6.09639347e-01
-9.31928873e-01 7.09211290e-01 4.12645221e-01 7.85882920e-02
-1.32990086e+00 -7.74883449e-01 -1.04859340e+00 -1.97824821e-01
4.00851756e-01 -8.82052243e-01 1.41259992e+00 -5.74956238e-01
-1.41182578e+00 1.23886263e+00 -1.85171977e-01 -3.07512909e-01
5.22466242e-01 -2.07215324e-01 1.75229594e-01 5.81382215e-01
2.72460300e-02 1.03803861e+00 5.63114941e-01 -1.31629074e+00
-4.27332401e-01 -6.93360031e-01 3.45451981e-01 3.96753490e-01
3.95240009e-01 -2.81869769e-01 -7.31261611e-01 -8.93594846e-02
5.66382527e-01 -6.95271730e-01 -5.93585849e-01 2.89135844e-01
-5.76129198e-01 5.24972267e-02 6.00089133e-01 -6.19989991e-01
2.25838214e-01 -1.94123781e+00 2.43553355e-01 3.25179607e-01
5.49576700e-01 -8.64074230e-02 9.10422280e-02 1.13323107e-01
-1.84081662e-02 -8.47008899e-02 -6.82556927e-01 -8.48055661e-01
-1.60213381e-01 3.33793700e-01 1.38978623e-02 6.49348974e-01
1.49374023e-01 1.15815949e+00 -1.07485926e+00 -4.30153787e-01
8.63979220e-01 8.57369602e-01 -5.14026105e-01 1.81199864e-01
-5.61740398e-01 9.31897879e-01 -3.51101190e-01 7.01152921e-01
8.32331479e-01 -9.34884995e-02 -1.00158945e-01 -1.03650302e-01
-7.72610754e-02 3.31420213e-01 -9.49087918e-01 2.74534893e+00
-6.19966447e-01 5.09144485e-01 1.50692254e-01 -7.02302396e-01
5.99335670e-01 2.23281831e-01 7.03488767e-01 -9.50861037e-01
2.76026398e-01 1.02797292e-01 -7.62782037e-01 -2.32917130e-01
4.92883652e-01 -2.30231598e-01 -5.15494682e-02 3.19647014e-01
-9.46455970e-02 -9.38914001e-01 -3.71991962e-01 2.71627575e-01
1.27598250e+00 4.52553213e-01 8.20564479e-02 1.28326207e-01
2.92497903e-01 2.80532032e-01 1.75341591e-01 5.84087789e-01
-9.75296125e-02 1.11230278e+00 4.13801879e-01 -4.40685391e-01
-1.41493547e+00 -1.38606608e+00 -2.13640243e-01 6.06660306e-01
1.74878702e-01 -3.45646888e-01 -5.86640656e-01 -5.80761969e-01
8.35587680e-02 4.77542013e-01 -6.69455409e-01 5.81221618e-02
-4.74609852e-01 -4.65857312e-02 4.81577724e-01 6.66775346e-01
5.53273022e-01 -1.07367611e+00 -1.40199435e+00 6.39595166e-02
5.08259330e-03 -1.47924209e+00 6.64789528e-02 4.71890122e-01
-1.09518874e+00 -1.20957911e+00 -6.84561908e-01 -4.74663854e-01
6.53090715e-01 3.65796953e-01 1.20343208e+00 -1.91353690e-02
-5.54304838e-01 6.18951857e-01 -4.67870027e-01 -3.88514668e-01
-7.34526217e-02 5.32153295e-03 -4.04242009e-01 -3.56685519e-01
8.67986307e-02 -5.97223401e-01 -8.83753717e-01 -1.49718463e-01
-1.20192742e+00 7.35266626e-01 2.73098052e-01 2.06010029e-01
9.77890372e-01 -1.66509032e-01 -3.71936932e-02 -8.72632325e-01
-1.55967055e-02 -2.59189934e-01 -5.27118027e-01 -2.21550643e-01
-2.12432176e-01 -1.81458533e-01 2.74754941e-01 4.00534451e-01
-9.15408134e-01 6.88177645e-01 -4.61134017e-01 -7.70489812e-01
-6.50713861e-01 1.49222583e-01 -2.14025244e-01 -1.37507720e-02
3.55710387e-01 1.86559588e-01 -6.18872158e-02 -5.68641126e-01
6.71124995e-01 5.56333005e-01 7.35269368e-01 -1.78187087e-01
4.71883208e-01 1.20932341e+00 3.21521878e-01 -6.85822368e-01
-9.35728788e-01 -8.95658314e-01 -1.15086782e+00 -4.05160785e-01
1.16180205e+00 -1.24458158e+00 -5.42803288e-01 5.21218717e-01
-1.25927663e+00 -7.09560752e-01 -3.97592098e-01 3.67248297e-01
-9.32248533e-01 2.11455271e-01 -3.20939243e-01 -5.41445494e-01
-1.18454076e-01 -1.03347659e+00 1.79920423e+00 1.35806993e-01
-5.37273176e-02 -9.00605679e-01 -9.33150053e-02 3.12616736e-01
2.62381047e-01 6.75159872e-01 4.56752449e-01 5.02854362e-02
-7.17691123e-01 6.24217018e-02 -5.52143395e-01 2.18323112e-01
-1.60162732e-01 -6.35535657e-01 -1.54493320e+00 -1.93617910e-01
-1.20238468e-01 -4.27707702e-01 9.44075227e-01 5.13256490e-01
1.53180432e+00 4.29576874e-01 -2.71318704e-01 1.04247701e+00
1.71965861e+00 -1.36276394e-01 8.34858716e-01 3.57962638e-01
1.15894639e+00 6.49414301e-01 3.98335606e-01 3.69905233e-01
6.27380550e-01 4.63140845e-01 9.96054351e-01 -4.91476476e-01
-5.21107197e-01 -3.34160954e-01 -3.48117739e-01 2.38483429e-01
3.68004255e-02 6.80571198e-02 -1.08837771e+00 4.54917252e-01
-1.58453965e+00 -3.52887481e-01 -3.61130327e-01 1.96104074e+00
5.51607788e-01 1.49724960e-01 -9.88693535e-02 7.67011419e-02
3.32481384e-01 9.61918011e-02 -7.22075760e-01 -2.88306296e-01
-1.27508089e-01 5.63188791e-01 1.07766867e+00 5.00478148e-01
-1.07785153e+00 1.08499718e+00 6.17122316e+00 2.46451169e-01
-8.78087759e-01 3.87145638e-01 5.53930342e-01 -3.89271885e-01
-4.69783574e-01 -4.92033325e-02 -1.64841056e-01 3.16233411e-02
6.35975659e-01 5.35697281e-01 4.68603909e-01 8.37272048e-01
8.29612762e-02 -7.33859301e-01 -1.11268246e+00 1.29358447e+00
3.40162180e-02 -1.14757717e+00 -3.38454068e-01 1.72003195e-01
8.45279455e-01 6.78398371e-01 -3.21276814e-01 -1.18493028e-01
4.12143767e-01 -1.24600816e+00 9.17070150e-01 5.49943626e-01
9.39952314e-01 -7.01122582e-01 4.79666114e-01 1.53226480e-01
-1.32607210e+00 2.62781769e-01 -4.31289136e-01 -1.44595101e-01
4.07910198e-01 7.05977678e-01 -9.28486586e-01 6.27957940e-01
9.67500448e-01 7.94518590e-01 -7.55219698e-01 1.21621418e+00
-5.36123335e-01 2.96531379e-01 -5.13858438e-01 4.55922097e-01
3.17041904e-01 2.38043755e-01 2.76237011e-01 1.01571083e+00
2.94956490e-02 8.33065659e-02 1.69496968e-01 8.35072041e-01
5.67036085e-02 -3.68570030e-01 -7.85508752e-01 2.38065913e-01
1.97376162e-01 1.26755702e+00 -1.42532778e+00 -2.93103456e-01
-3.16323996e-01 1.35220432e+00 2.80075759e-01 2.49659076e-01
-6.15988493e-01 -8.80063102e-02 5.74386537e-01 1.46764442e-01
4.65707362e-01 -5.67739785e-01 -8.73766541e-01 -6.83265865e-01
-1.04353026e-01 7.56406114e-02 8.91739205e-02 -1.20465362e+00
-1.00293076e+00 2.76475608e-01 1.00227103e-01 -9.36210811e-01
1.10696480e-01 -6.37369931e-01 -2.17579588e-01 8.96066546e-01
-1.78104436e+00 -1.21193564e+00 -7.82494366e-01 8.26613307e-01
4.21336085e-01 3.84581387e-01 6.67080581e-01 7.81503469e-02
8.52056369e-02 -2.01688543e-01 -1.13858923e-01 -1.46222219e-01
1.24971449e-01 -1.32228959e+00 8.78505647e-01 4.69522268e-01
-1.26754105e-01 -5.29752579e-03 4.38730508e-01 -7.60782361e-01
-1.54458272e+00 -1.26913190e+00 5.51115632e-01 -6.55301452e-01
1.00397199e-01 -7.37818062e-01 -5.28910279e-01 5.46379864e-01
-1.63629293e-01 3.60878527e-01 2.98116237e-01 -3.33899200e-01
-2.57661909e-01 2.50491887e-01 -1.28685892e+00 2.97797173e-01
1.67244017e+00 -8.78979802e-01 -3.37159693e-01 2.66126543e-01
1.06328845e+00 -9.99345362e-01 -8.24009359e-01 3.64941627e-01
3.62816125e-01 -1.14144170e+00 1.33625305e+00 -1.71754479e-01
5.96273422e-01 -4.51499343e-01 -4.75412011e-01 -1.21737802e+00
2.98752654e-02 2.25736737e-01 4.14100550e-02 5.93697429e-01
-1.07235558e-01 -2.55592644e-01 1.09200454e+00 5.57096481e-01
-6.16572022e-01 -6.04193091e-01 -1.32905054e+00 -6.22051060e-01
-1.37329519e-01 -1.10764635e+00 8.19710135e-01 6.38166964e-01
-3.72020811e-01 3.99647988e-02 1.21569954e-01 1.10722326e-01
7.92196810e-01 -1.56195592e-02 6.95942342e-01 -1.29289234e+00
5.18410280e-02 -1.57732472e-01 -6.61876440e-01 -7.40248561e-01
1.05021000e-02 -1.24794960e+00 2.64953732e-01 -2.24845576e+00
1.67286750e-02 -4.21728104e-01 -1.12246916e-01 4.32717264e-01
2.41404831e-01 6.90697551e-01 1.81698993e-01 5.26406337e-03
-9.45642650e-01 6.79918110e-01 1.52420175e+00 2.18257792e-02
-1.42375439e-01 -4.54675615e-01 -4.20169920e-01 6.14461362e-01
8.64041150e-01 -4.27008569e-01 -5.44659317e-01 -6.55179381e-01
3.05354208e-01 1.04360476e-01 1.01328886e+00 -1.06577981e+00
5.97532317e-02 1.05927922e-01 7.94094980e-01 -9.50326145e-01
6.78762913e-01 -1.06464732e+00 1.95894375e-01 4.83404398e-01
-4.15060148e-02 -2.08588302e-01 2.47069314e-01 6.56841636e-01
1.45914644e-01 2.21227780e-01 5.62041759e-01 -5.39212704e-01
-1.01390743e+00 4.01814491e-01 -2.65014678e-01 -9.56199542e-02
1.04701567e+00 -4.36383337e-01 -1.17984958e-01 -8.05447176e-02
-8.11887980e-01 2.32693613e-01 7.88708329e-01 3.87183964e-01
1.15657353e+00 -1.36424708e+00 -3.79949898e-01 3.21195334e-01
2.79330879e-01 9.79352117e-01 4.39588279e-01 7.04403460e-01
-8.75728309e-01 7.39890635e-01 -2.79706180e-01 -1.06523263e+00
-8.63824844e-01 3.96301568e-01 3.17788959e-01 8.12323689e-02
-1.15113699e+00 9.10870671e-01 4.46876377e-01 -6.79172218e-01
2.78749704e-01 -5.50386429e-01 3.99594614e-03 -3.86369824e-01
3.71390492e-01 1.12485386e-01 3.27693611e-01 -5.73713601e-01
-5.93432307e-01 4.38061863e-01 3.59928995e-01 -4.72799271e-01
1.48161590e+00 -3.51362914e-01 -8.51552759e-04 4.65664357e-01
1.32710290e+00 -6.51806355e-01 -1.71294296e+00 -3.72975647e-01
-2.71685004e-01 -6.63523972e-01 5.53352654e-01 -7.58577406e-01
-1.19438136e+00 8.97119462e-01 8.02202761e-01 -1.67307839e-01
1.03154206e+00 4.98706102e-01 8.50820363e-01 1.62624680e-02
7.50564218e-01 -1.00297987e+00 -7.82343000e-03 5.36553442e-01
8.47689867e-01 -1.18473911e+00 6.83420599e-02 -5.92435598e-01
-5.41693747e-01 8.71542811e-01 5.95418394e-01 1.25050135e-02
7.23742247e-01 9.01326165e-02 -2.08263367e-01 -8.01895618e-01
-2.26811260e-01 -6.68224156e-01 1.28200442e-01 8.85562420e-01
6.20127358e-02 2.42386743e-01 3.88060004e-01 1.90817252e-01
-2.85556018e-01 1.85171694e-01 3.82895291e-01 1.27798390e+00
-4.42854881e-01 -6.64976716e-01 -5.13636768e-01 5.06210208e-01
4.29681651e-02 -1.09687343e-01 -5.97898483e-01 4.50915247e-01
3.97290111e-01 7.77806103e-01 4.23464984e-01 -2.76827604e-01
6.08840764e-01 -3.80441472e-02 8.02404404e-01 -9.70149696e-01
-2.03845114e-01 -1.29656568e-01 4.47828434e-02 -1.07486832e+00
-5.88423073e-01 -6.11891389e-01 -1.60834634e+00 -6.87989071e-02
4.18131888e-01 -5.75016797e-01 1.22853959e+00 9.41564202e-01
9.43302270e-03 8.14397514e-01 5.27194381e-01 -1.37212288e+00
4.67170417e-01 -7.62564242e-01 -8.00173402e-01 3.44544768e-01
4.52507108e-01 -7.35882640e-01 7.87784830e-02 -2.54120380e-02] | [8.452519416809082, -2.862555503845215] |
5ef6c791-6454-421d-9d4c-2c830afd065c | orthogonal-deep-features-decomposition-for | 1810.07599 | null | http://arxiv.org/abs/1810.07599v1 | http://arxiv.org/pdf/1810.07599v1.pdf | Orthogonal Deep Features Decomposition for Age-Invariant Face Recognition | As facial appearance is subject to significant intra-class variations caused
by the aging process over time, age-invariant face recognition (AIFR) remains a
major challenge in face recognition community. To reduce the intra-class
discrepancy caused by the aging, in this paper we propose a novel approach
(namely, Orthogonal Embedding CNNs, or OE-CNNs) to learn the age-invariant deep
face features. Specifically, we decompose deep face features into two
orthogonal components to represent age-related and identity-related features.
As a result, identity-related features that are robust to aging are then used
for AIFR. Besides, for complementing the existing cross-age datasets and
advancing the research in this field, we construct a brand-new large-scale
Cross-Age Face dataset (CAF). Extensive experiments conducted on the three
public domain face aging datasets (MORPH Album 2, CACD-VS and FG-NET) have
shown the effectiveness of the proposed approach and the value of the
constructed CAF dataset on AIFR. Benchmarking our algorithm on one of the most
popular general face recognition (GFR) dataset LFW additionally demonstrates
the comparable generalization performance on GFR. | ['Yitong Wang', 'Zheng Zhou', 'Wei Liu', 'Hao Wang', 'Zhifeng Li', 'Xing Ji', 'Dihong Gong', 'Tong Zhang'] | 2018-10-17 | orthogonal-deep-features-decomposition-for-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/yitong_wang_Orthogonal_Deep_Features_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/yitong_wang_Orthogonal_Deep_Features_ECCV_2018_paper.pdf | eccv-2018-9 | ['age-invariant-face-recognition'] | ['computer-vision'] | [-1.47028759e-01 -2.27395073e-01 5.93006052e-02 -8.13408017e-01
-1.86810344e-01 4.85065803e-02 3.63464683e-01 -2.97177643e-01
-2.02446952e-01 6.94220185e-01 2.56050617e-01 2.62484998e-01
-1.62954912e-01 -7.61028647e-01 -4.44316626e-01 -7.99500942e-01
-4.09088016e-01 4.86085042e-02 -3.89218450e-01 -2.83161253e-01
4.85616475e-02 7.99403787e-01 -2.14869118e+00 -4.29330990e-02
9.20681715e-01 1.52460289e+00 -5.49965978e-01 1.27646878e-01
1.48346007e-01 2.49683097e-01 -3.81504267e-01 -6.23835266e-01
4.62663114e-01 -4.59354185e-02 -4.76250827e-01 6.32654652e-02
1.11429787e+00 -4.07637835e-01 -5.91664910e-01 9.26594853e-01
9.26812589e-01 2.89313756e-02 5.64388335e-01 -1.73229456e+00
-1.13376641e+00 7.43703321e-02 -7.94212699e-01 3.67018022e-02
1.54285938e-01 -2.08113398e-02 4.82398540e-01 -1.20404768e+00
4.50231194e-01 1.91270626e+00 1.01496542e+00 1.07992435e+00
-7.72854090e-01 -1.24598682e+00 1.39725238e-01 4.37938988e-01
-1.43254125e+00 -7.99903631e-01 8.15333188e-01 -3.51215482e-01
2.91818738e-01 3.35274846e-04 3.98378700e-01 1.26433420e+00
1.90435871e-01 4.46183681e-01 1.25807381e+00 -2.99551398e-01
-3.86719331e-02 -1.81373700e-01 1.71749040e-01 9.53293562e-01
3.19859207e-01 1.73557505e-01 -8.47773015e-01 -8.37913677e-02
4.19470549e-01 1.72984362e-01 -3.58783424e-01 -1.35545492e-01
-6.37924254e-01 6.57086134e-01 2.86644965e-01 1.11459307e-01
-2.18545288e-01 -9.32385847e-02 5.78894496e-01 4.96002316e-01
9.25728023e-01 -1.35444254e-01 -6.29555643e-01 2.04183340e-01
-6.54598176e-01 2.71738440e-01 3.33800554e-01 6.02720976e-01
7.71090269e-01 2.13280305e-01 -3.18135947e-01 1.13209927e+00
4.27228957e-01 4.59322810e-01 5.24447978e-01 -6.68782651e-01
7.27003366e-02 7.05666721e-01 -3.06509346e-01 -1.13587797e+00
-4.30507600e-01 -3.46013248e-01 -9.63931024e-01 3.12431216e-01
4.32093650e-01 1.68940991e-01 -9.92393911e-01 2.26296616e+00
5.27791798e-01 3.19870591e-01 -3.99618922e-03 3.89660954e-01
9.65452790e-01 1.52329028e-01 3.19891632e-01 -3.16467673e-01
1.67354715e+00 -7.09611356e-01 -7.40290582e-01 9.02312100e-02
2.45644823e-01 -6.34599686e-01 7.61931539e-01 1.81093901e-01
-7.28486598e-01 -8.64002824e-01 -1.04513538e+00 -7.26882145e-02
-5.93820274e-01 4.08425808e-01 7.92178512e-01 8.62534046e-01
-1.30845523e+00 7.24539399e-01 -3.91727865e-01 -5.23330271e-01
9.95334029e-01 6.74466550e-01 -9.81957614e-01 -4.78122085e-01
-1.23492742e+00 6.84105396e-01 2.30546892e-01 3.74631017e-01
-8.48144948e-01 -1.19638705e+00 -9.31877315e-01 -2.40881965e-01
-3.17250937e-02 -5.04309833e-01 7.16822267e-01 -1.17238533e+00
-1.20563889e+00 1.12273872e+00 -2.14270532e-01 -1.29082184e-02
3.48756760e-01 -4.19901580e-01 -9.40146387e-01 3.57431732e-02
-4.36216919e-03 7.79619455e-01 1.24564672e+00 -1.11870372e+00
-3.37586492e-01 -1.19913042e+00 -1.78575560e-01 -2.39974841e-01
-1.12302589e+00 2.19506189e-01 -2.15440899e-01 -6.75879955e-01
-1.54153064e-01 -8.67416859e-01 4.54565972e-01 6.72225475e-01
4.26426269e-02 -7.39279389e-01 1.28130662e+00 -1.15093863e+00
1.18406105e+00 -2.22820497e+00 8.92043635e-02 -1.11788519e-01
2.99317777e-01 3.44859809e-01 -5.32134354e-01 -2.19769534e-02
-7.14268565e-01 -6.88990057e-02 -7.21993521e-02 -4.52329278e-01
-2.13581041e-01 3.36583182e-02 1.93726182e-01 6.90244317e-01
5.04565239e-01 6.57086313e-01 -5.56744516e-01 -3.86771798e-01
-2.80528843e-01 7.79593110e-01 -3.95935327e-01 2.60784775e-01
2.35040858e-01 2.81023920e-01 -2.23725945e-01 1.17784870e+00
1.16070139e+00 4.72649723e-01 -2.24991128e-01 -6.96177483e-01
1.20879248e-01 -7.42140949e-01 -7.99528122e-01 1.55541348e+00
-3.50822926e-01 3.13982606e-01 6.51272759e-02 -7.56492972e-01
1.18436337e+00 2.32420161e-01 6.18908167e-01 -6.62639260e-01
2.82566667e-01 1.95225880e-01 -6.57437146e-02 -4.29790944e-01
1.92319110e-01 -5.27530024e-03 4.92959827e-01 1.65162787e-01
4.20373678e-01 7.54861712e-01 1.49005234e-01 -1.43667996e-01
8.49659264e-01 6.53624386e-02 4.88928482e-02 -5.86584866e-01
9.72521782e-01 -1.02842021e+00 9.38763440e-01 -7.64995888e-02
-7.55235255e-01 4.62658048e-01 2.71076143e-01 -7.97881067e-01
-9.18866634e-01 -9.59757984e-01 -4.58050728e-01 9.71691370e-01
-3.42440069e-01 -3.83072168e-01 -7.68347502e-01 -9.69690740e-01
3.98773253e-01 -1.11184474e-02 -1.28519309e+00 -4.90966231e-01
-4.98062581e-01 -8.79824221e-01 5.72320282e-01 7.43267417e-01
8.92371893e-01 -9.42634106e-01 1.87032685e-01 -1.84906214e-01
2.22527847e-01 -9.93000329e-01 -6.94987655e-01 -5.46556950e-01
-7.27333248e-01 -1.38479912e+00 -9.29273009e-01 -7.89720774e-01
9.17596459e-01 8.19437504e-02 9.29041266e-01 2.87882715e-01
-6.55431628e-01 5.77432811e-01 -2.11258218e-01 -3.76280129e-01
-5.99858463e-02 -1.17402516e-01 6.65444791e-01 6.72605932e-01
4.31237340e-01 -5.73711634e-01 -8.28500926e-01 2.12694407e-01
-6.89982414e-01 -5.64032257e-01 4.16342616e-01 8.69156003e-01
1.28905565e-01 -6.10454194e-02 1.13022375e+00 -5.47225237e-01
5.34838796e-01 -4.10887510e-01 -2.29339138e-01 4.43286210e-01
-7.53596127e-01 -1.62882224e-01 3.29801649e-01 -4.07026142e-01
-1.23875391e+00 -1.34132236e-01 -8.92617255e-02 -4.90014136e-01
4.27095927e-02 2.04897597e-02 -6.87533677e-01 -4.76339787e-01
5.60733855e-01 -5.87483421e-02 4.73751724e-01 -4.92715389e-01
2.18187317e-01 7.22482502e-01 6.13404989e-01 -7.77415216e-01
1.08154345e+00 4.11437362e-01 2.06648454e-01 -7.50749052e-01
-7.49033034e-01 -1.09050488e-02 -5.52301288e-01 -3.70095164e-01
7.50444233e-01 -1.16933286e+00 -8.04968297e-01 1.15611613e+00
-9.96725500e-01 2.54028231e-01 6.19492643e-02 1.18305199e-01
-1.37632787e-01 6.00240529e-01 -6.14525974e-01 -8.81071866e-01
-8.46575022e-01 -9.36625123e-01 1.10306394e+00 5.96626163e-01
1.97320268e-01 -7.62305319e-01 -5.49691208e-02 3.04710656e-01
5.99305749e-01 5.80814958e-01 9.34850693e-01 -3.48053694e-01
-4.80572395e-02 -2.58982182e-01 -5.00095069e-01 7.55195796e-01
5.63643277e-01 3.80201876e-01 -1.26822376e+00 -6.81195498e-01
-3.33752453e-01 -4.72691894e-01 8.69801641e-01 7.08703399e-02
1.54302597e+00 -7.80627280e-02 -1.05648883e-01 6.21427059e-01
1.30633342e+00 8.28703195e-02 9.16091621e-01 1.76031500e-01
6.43784404e-01 8.14601421e-01 5.18019855e-01 5.26088595e-01
3.42462480e-01 6.70526147e-01 3.15505683e-01 -9.88016650e-02
-2.60353208e-01 -4.99149300e-02 4.49890643e-01 6.47861838e-01
-4.09994841e-01 2.48107001e-01 -5.52704632e-01 3.50176573e-01
-1.52539814e+00 -8.13775241e-01 1.72334537e-01 2.04440236e+00
7.62562752e-01 -5.07355332e-01 9.98164266e-02 1.33569434e-01
8.53731334e-01 6.98384866e-02 -5.57822108e-01 -1.74415171e-01
-4.02456969e-01 6.09880567e-01 6.51118979e-02 -7.08421767e-02
-1.19221807e+00 7.06617296e-01 5.45430851e+00 8.15245688e-01
-1.17973399e+00 9.47676674e-02 1.02741981e+00 1.78604543e-01
1.59674764e-01 -5.81571043e-01 -9.64329004e-01 4.76380974e-01
8.67064059e-01 -2.80637503e-01 3.86118829e-01 1.05713356e+00
-1.54073806e-02 3.27125937e-01 -1.06454635e+00 1.26143777e+00
4.86298114e-01 -8.07894707e-01 7.78182968e-02 5.64013794e-02
6.53674781e-01 -5.80718458e-01 4.58411604e-01 5.51841080e-01
-3.06150377e-01 -1.23009551e+00 3.57056648e-01 5.47670126e-01
1.22695458e+00 -9.19154286e-01 7.74578512e-01 -5.33842623e-01
-1.45911014e+00 -4.48459238e-01 -4.16816026e-01 2.48698264e-01
-4.66167897e-01 5.30107975e-01 -1.88476264e-01 6.86457634e-01
1.06876934e+00 9.73171830e-01 -1.18350315e+00 8.00891757e-01
1.95496947e-01 2.32454002e-01 1.74287096e-01 7.85365939e-01
-4.02725935e-01 -9.18216929e-02 3.19796167e-02 5.85781276e-01
5.79436600e-01 -9.75165367e-02 -2.13012919e-01 5.22961855e-01
-5.60470521e-01 1.14081264e-01 -4.41831976e-01 -1.32732868e-01
3.94930184e-01 1.52505612e+00 -1.71548724e-01 -6.94712549e-02
-4.95845646e-01 9.18340504e-01 3.32378298e-01 2.40121916e-01
-7.12019622e-01 -1.64274156e-01 1.24035764e+00 7.85773527e-03
7.35379905e-02 -8.05695541e-03 2.60849744e-01 -1.14502347e+00
2.66070187e-01 -1.22854781e+00 5.40838122e-01 -3.71869683e-01
-1.73442829e+00 7.19908476e-01 -2.04900607e-01 -8.34786296e-01
3.88402700e-01 -8.48461509e-01 -6.56033754e-01 8.72710168e-01
-1.75875568e+00 -1.68331039e+00 -7.79422700e-01 8.13495278e-01
4.30394948e-01 -6.06866181e-01 8.63474846e-01 8.20583642e-01
-1.06607604e+00 1.12986600e+00 -2.06032053e-01 2.14561880e-01
1.11232984e+00 -9.01009023e-01 1.15366690e-01 6.99793637e-01
-5.15895844e-01 8.22592616e-01 2.36735865e-01 -5.20316184e-01
-1.50166392e+00 -1.34554517e+00 5.75259924e-01 -2.67792970e-01
3.49521220e-01 -3.46532643e-01 -1.16551173e+00 4.75983113e-01
5.33830933e-02 5.03294885e-01 7.59076059e-01 1.92373499e-01
-9.81031656e-01 -8.10871005e-01 -1.54993141e+00 3.79453480e-01
1.46365416e+00 -5.23776591e-01 -1.90356091e-01 2.96696350e-02
5.60788155e-01 3.41136232e-02 -1.23880589e+00 9.15678561e-01
9.78207052e-01 -8.62362385e-01 1.05625522e+00 -7.73462534e-01
3.75249505e-01 -2.03094646e-01 -1.30850047e-01 -1.11132610e+00
-3.26397330e-01 -3.93533319e-01 -3.25849593e-01 1.80884099e+00
-2.06833377e-01 -7.52283692e-01 7.11955905e-01 5.44710636e-01
-3.84392515e-02 -8.80091369e-01 -1.08475947e+00 -8.44511986e-01
1.00301586e-01 1.46293104e-01 8.33822250e-01 9.85099196e-01
-8.06136787e-01 -1.05217837e-01 -4.14103121e-01 1.85336277e-01
8.41435909e-01 -3.93998504e-01 6.11432254e-01 -1.68258262e+00
3.59426528e-01 -2.37824649e-01 -8.28678966e-01 2.31808200e-02
6.65829897e-01 -7.17129588e-01 -2.62417912e-01 -8.08892310e-01
3.19003880e-01 -3.29555809e-01 -6.09207332e-01 6.39558971e-01
-3.40691090e-01 5.45991004e-01 -7.07871467e-02 -2.49882445e-01
-2.04266548e-01 1.08286703e+00 1.13255203e+00 -2.72302270e-01
2.46783733e-01 -1.96963102e-01 -7.85907090e-01 4.57172334e-01
7.43852615e-01 -4.82590683e-02 -2.73677617e-01 -7.06913099e-02
-3.11713964e-01 -5.62421620e-01 2.04568446e-01 -1.18691456e+00
-1.86940476e-01 1.59720212e-01 8.30438197e-01 -2.94863284e-01
3.53702039e-01 -6.49434924e-01 8.47371370e-02 4.97138232e-01
6.35784194e-02 3.25867087e-01 4.43187535e-01 4.95900959e-01
-1.80981278e-01 1.54212773e-01 1.07829738e+00 2.96886325e-01
-8.11805964e-01 1.22490823e+00 3.02890331e-01 -4.63892370e-02
1.08261538e+00 -2.28960365e-01 -4.34357792e-01 -5.58587862e-03
-4.74818587e-01 1.09537050e-01 3.47058237e-01 9.95288968e-01
6.47014201e-01 -1.76866210e+00 -9.03076649e-01 3.73129666e-01
4.51768458e-01 -5.85452735e-01 6.28923059e-01 6.53808117e-01
-1.50186673e-01 1.27275307e-02 -8.39740813e-01 -2.21783206e-01
-1.72904599e+00 6.19465649e-01 4.54097986e-01 5.58091365e-02
-1.70799866e-01 1.05083060e+00 2.65544713e-01 -3.52666438e-01
2.84101456e-01 2.93588430e-01 -3.62718433e-01 3.46557111e-01
8.87671590e-01 4.83554304e-01 1.19252414e-01 -8.93109500e-01
-4.40405339e-01 8.22270691e-01 -3.42919379e-01 5.93904436e-01
1.52250779e+00 5.30231148e-02 -5.35551488e-01 -5.17273583e-02
1.46919477e+00 -2.12616205e-01 -1.22083366e+00 -1.64681613e-01
-7.57798227e-03 -6.41126812e-01 -1.44334510e-01 -5.23283422e-01
-1.69187164e+00 8.30091357e-01 1.46707475e+00 -3.92001271e-01
1.40344834e+00 -3.45819116e-01 7.16270149e-01 -2.72530336e-02
2.84597397e-01 -9.89224255e-01 4.33169037e-01 8.90270770e-02
1.31369853e+00 -1.32619786e+00 1.85074117e-02 -5.08174121e-01
-1.68193690e-02 1.46662652e+00 1.16625488e+00 2.08956912e-01
8.29478920e-01 -2.67786175e-01 3.50310132e-02 -6.73168823e-02
-4.23575670e-01 4.17594425e-03 3.23680282e-01 9.17973936e-01
5.31658530e-01 -2.33223870e-01 -4.20662075e-01 5.72002709e-01
-4.23787124e-02 2.10071445e-01 -2.59921737e-02 6.99040830e-01
-1.25070199e-01 -1.35914135e+00 -2.59538710e-01 5.14419198e-01
-4.22274858e-01 1.71738684e-01 -2.47790635e-01 7.20201015e-01
6.41284466e-01 6.89565003e-01 -1.45621568e-01 -4.17264909e-01
2.76525110e-01 2.93479055e-01 6.98351204e-01 -2.81378210e-01
-4.19159204e-01 -7.35612273e-01 -1.04155503e-01 -6.52468741e-01
-4.85601366e-01 -6.37499332e-01 -6.40804589e-01 -4.66305614e-01
-9.87805501e-02 -3.08934063e-01 6.06454670e-01 6.38812900e-01
5.43983221e-01 3.95553797e-01 9.25228477e-01 -7.26384223e-01
-6.09494448e-01 -9.91032422e-01 -6.70416892e-01 6.98605299e-01
1.27338454e-01 -1.19534290e+00 -3.26452196e-01 -1.08458541e-01] | [13.399888038635254, 0.6728320121765137] |
94a163b8-7a23-419c-a52f-f85dfae07c5b | hyperspectral-image-segmentation-a | 2303.08252 | null | https://arxiv.org/abs/2303.08252v1 | https://arxiv.org/pdf/2303.08252v1.pdf | Hyperspectral Image Segmentation: A Preliminary Study on the Oral and Dental Spectral Image Database (ODSI-DB) | Visual discrimination of clinical tissue types remains challenging, with traditional RGB imaging providing limited contrast for such tasks. Hyperspectral imaging (HSI) is a promising technology providing rich spectral information that can extend far beyond three-channel RGB imaging. Moreover, recently developed snapshot HSI cameras enable real-time imaging with significant potential for clinical applications. Despite this, the investigation into the relative performance of HSI over RGB imaging for semantic segmentation purposes has been limited, particularly in the context of medical imaging. Here we compare the performance of state-of-the-art deep learning image segmentation methods when trained on hyperspectral images, RGB images, hyperspectral pixels (minus spatial context), and RGB pixels (disregarding spatial context). To achieve this, we employ the recently released Oral and Dental Spectral Image Database (ODSI-DB), which consists of 215 manually segmented dental reflectance spectral images with 35 different classes across 30 human subjects. The recent development of snapshot HSI cameras has made real-time clinical HSI a distinct possibility, though successful application requires a comprehensive understanding of the additional information HSI offers. Our work highlights the relative importance of spectral resolution, spectral range, and spatial information to both guide the development of HSI cameras and inform future clinical HSI applications. | ['Tom Vercauteren', 'Michael Ebner', 'Sebastien Ourselin', 'Conor Horgan', 'Luis C. Garcia-Peraza-Herrera'] | 2023-03-14 | null | null | null | null | ['hyperspectral-image-segmentation'] | ['computer-vision'] | [ 1.19355798e+00 -1.08797617e-01 -1.94338739e-01 -2.80542731e-01
-1.16290641e+00 -4.03813064e-01 6.75578089e-03 5.97699620e-02
-5.49358130e-01 4.11518127e-01 -1.42801434e-01 -4.76379693e-01
-1.41860500e-01 -4.70550328e-01 -1.87923431e-01 -1.32265139e+00
2.00739846e-01 2.30436146e-01 -6.08235113e-02 -5.19476682e-02
3.38342562e-02 6.38134360e-01 -1.66024578e+00 4.28848445e-01
1.02087700e+00 1.10826278e+00 3.18538815e-01 5.84962904e-01
4.90331389e-02 5.09501733e-02 -4.08761263e-01 3.31654251e-02
2.92321116e-01 -4.10713315e-01 -5.91235578e-01 3.17427814e-01
6.18786335e-01 -3.56825113e-01 -5.12081804e-03 1.02954328e+00
8.06865096e-01 9.90543980e-03 4.80515867e-01 -8.28085482e-01
-8.13387454e-01 1.71026304e-01 -5.63381195e-01 3.80182773e-01
3.11740786e-01 5.88587582e-01 7.14548767e-01 -3.12762588e-01
3.81875187e-01 5.13572156e-01 4.54038799e-01 6.57689512e-01
-1.08575618e+00 -4.39535916e-01 -3.35413218e-01 2.87441075e-01
-1.17344606e+00 -3.28331977e-01 8.46148849e-01 -2.64696568e-01
9.17033613e-01 5.38174152e-01 1.06427777e+00 8.33388865e-01
-9.19102579e-02 6.48421228e-01 1.92145336e+00 -4.99041498e-01
2.69822568e-01 1.20378733e-02 6.24959394e-02 5.78766465e-01
2.90836036e-01 1.35881409e-01 -3.78109604e-01 9.29369330e-02
6.16600633e-01 -6.84042508e-03 -5.73314428e-01 1.43278241e-01
-1.01107240e+00 3.67687553e-01 7.67068446e-01 6.27299130e-01
-4.18237478e-01 -2.19629914e-01 8.79376754e-02 -1.68103337e-01
4.52794731e-01 2.63654470e-01 -1.33132502e-01 2.37080619e-01
-1.05348623e+00 -6.43185377e-01 1.01624876e-01 2.64235109e-01
4.90639508e-01 2.32149158e-02 2.06720635e-01 1.08710742e+00
3.29196900e-01 6.50576055e-01 5.03484130e-01 -9.75597143e-01
-2.97084945e-04 4.86725837e-01 -4.61236805e-01 -4.04034406e-01
-7.40577221e-01 -3.49681199e-01 -5.70577919e-01 3.15566063e-01
3.40923816e-01 2.32151777e-01 -1.36844432e+00 1.06816649e+00
3.93092036e-01 -1.14396706e-01 7.95702115e-02 1.10617828e+00
9.80000556e-01 3.85830671e-01 3.03056598e-01 -3.81360233e-01
1.29985178e+00 -4.24984217e-01 -4.64907497e-01 -4.21629518e-01
3.06310713e-01 -5.63633442e-01 1.30786109e+00 6.38931572e-01
-7.66543984e-01 -9.91135538e-02 -1.06960201e+00 -1.24404870e-01
-5.54647803e-01 1.32801712e-01 8.45811665e-01 1.31155121e+00
-1.15642262e+00 2.35254675e-01 -1.05613792e+00 -5.10681152e-01
7.15263128e-01 4.95662987e-01 -4.61298794e-01 -5.29081762e-01
-8.56093228e-01 6.25060856e-01 2.79810011e-01 4.64468062e-01
-1.85801044e-01 -8.90147746e-01 -7.03959525e-01 -3.53635222e-01
7.06751421e-02 -4.40022320e-01 9.42568362e-01 -9.77670968e-01
-1.62829208e+00 1.33974361e+00 1.65177390e-01 2.04667926e-01
3.99540871e-01 3.72330636e-01 -4.52573419e-01 9.19965625e-01
-6.83159158e-02 6.12382948e-01 4.21681792e-01 -1.39096463e+00
-2.95729935e-01 -9.86714184e-01 -3.24199051e-01 3.36559862e-01
-1.55463994e-01 -9.99002457e-02 -5.11065423e-01 -1.78348571e-02
3.74921858e-01 -1.11932683e+00 -1.10819720e-01 1.89915165e-01
-7.13238180e-01 1.17205329e-01 5.24471760e-01 -9.30324435e-01
5.98427713e-01 -2.20251322e+00 -4.38083082e-01 2.01581329e-01
-1.11934736e-01 3.76201212e-01 -1.36881806e-02 8.39136541e-02
-1.76186413e-01 7.98128098e-02 -7.18675613e-01 2.19153520e-02
-4.12094206e-01 4.39479128e-02 4.23929095e-01 9.65090811e-01
-2.70918459e-01 7.96088040e-01 -8.44342470e-01 -6.54043555e-01
7.43585348e-01 8.85408401e-01 -1.70153961e-01 -1.82372242e-01
9.30940907e-04 8.27803016e-01 -3.09561908e-01 1.34661174e+00
7.58407116e-01 -3.81653130e-01 1.83577135e-01 -5.65087438e-01
-1.71964750e-01 -1.93229496e-01 -5.57040334e-01 1.79256272e+00
-2.82081515e-01 6.33906484e-01 2.54653960e-01 -7.27994621e-01
3.81523728e-01 2.76440531e-01 1.12899375e+00 -1.04406369e+00
9.63331088e-02 4.76571679e-01 -8.96840021e-02 -8.63587260e-01
4.20849323e-02 -7.04328656e-01 5.56264818e-01 2.31326967e-01
-1.85170636e-01 -3.71151090e-01 -1.49680778e-01 -4.24206942e-01
5.56704879e-01 -3.70667242e-02 1.31416947e-01 1.50524631e-01
1.99514911e-01 3.86920869e-01 1.19930811e-01 4.57894087e-01
-5.84185421e-01 1.00958383e+00 -5.14362305e-02 2.74364073e-02
-7.06789553e-01 -1.22295129e+00 -6.85858071e-01 8.71921003e-01
1.56336576e-01 5.10387957e-01 -5.80384314e-01 -2.44430661e-01
-5.92557155e-02 4.38210577e-01 -6.09578371e-01 1.23687245e-01
-4.06253934e-01 -1.41071367e+00 3.34694237e-01 3.14502865e-01
4.76291507e-01 -1.03782201e+00 -9.72478867e-01 6.93924949e-02
-2.19578609e-01 -1.19139183e+00 -8.82155634e-03 3.40898871e-01
-8.99686933e-01 -1.52477479e+00 -9.60033894e-01 -5.80850124e-01
4.53361601e-01 5.34934878e-01 7.21398652e-01 -4.56968732e-02
-9.99471605e-01 6.69121146e-01 -3.05078626e-01 -3.24280024e-01
-2.28964180e-01 -1.24750152e-01 -5.91121793e-01 -3.20432365e-01
3.30055952e-01 -3.30473006e-01 -1.10084105e+00 -4.25320901e-02
-1.32416379e+00 9.96343568e-02 8.15652609e-01 6.32054389e-01
8.78459215e-01 -6.41088337e-02 3.17049444e-01 -1.01850140e+00
1.49997130e-01 -2.93623686e-01 -2.48376831e-01 5.07070422e-01
-5.84069431e-01 -4.70373511e-01 8.15923437e-02 -7.67739266e-02
-1.32734847e+00 1.89988509e-01 -3.28809619e-01 1.33594215e-01
-6.02653086e-01 7.13297307e-01 1.28011182e-01 -3.29536557e-01
8.01102340e-01 2.25911021e-01 3.77183735e-01 -2.08912209e-01
3.19619924e-01 1.03229570e+00 7.69557834e-01 -1.59709066e-01
9.79905054e-02 1.05108821e+00 -3.11081763e-02 -1.37976360e+00
-7.58638799e-01 -9.85855222e-01 -6.65119708e-01 -4.31036025e-01
1.32196641e+00 -8.03761661e-01 -8.02219152e-01 5.28944552e-01
-4.96399432e-01 -6.07484460e-01 -7.65718287e-03 4.59203333e-01
-5.18423676e-01 4.97838736e-01 -7.77993500e-01 -7.89824903e-01
-6.04455888e-01 -1.44693780e+00 1.35675383e+00 4.12497252e-01
3.08524847e-01 -9.55953240e-01 -9.92438793e-02 1.25740862e+00
2.97787517e-01 6.32467926e-01 1.01469374e+00 -1.38030024e-02
-4.75613624e-01 -4.74067442e-02 -4.00903314e-01 3.54268044e-01
4.84707952e-01 1.00411355e-01 -1.44842434e+00 -5.54849580e-03
-3.48590547e-03 -2.90501654e-01 1.09602094e+00 9.26857054e-01
1.30848598e+00 3.53611410e-01 -2.83834875e-01 8.16450775e-01
1.75911164e+00 4.80716079e-01 6.24253571e-01 5.41140795e-01
8.25121403e-01 7.99949765e-01 4.76898044e-01 7.39987269e-02
1.01475306e-01 4.07749623e-01 7.00246751e-01 -8.02128911e-01
-3.63725096e-01 4.98719424e-01 -7.18830153e-02 3.24075878e-01
-1.46645039e-01 8.20324197e-03 -1.00092244e+00 3.05859417e-01
-9.72391188e-01 -6.88165665e-01 -1.82616636e-01 2.03339314e+00
9.90150452e-01 -4.45176125e-01 1.32297695e-01 3.10752153e-01
6.37578905e-01 8.31994265e-02 -9.46023881e-01 5.68014570e-02
-3.03181916e-01 5.38101912e-01 6.88567281e-01 2.41081744e-01
-9.73468661e-01 6.20591402e-01 6.75268602e+00 4.52422857e-01
-1.68346357e+00 2.88555194e-02 8.95144045e-01 -1.25955567e-01
-3.76674950e-01 -7.84934461e-01 -1.44700974e-01 1.16911270e-01
7.60862410e-01 5.60808003e-01 5.72398067e-01 3.81446481e-01
3.03106070e-01 -6.16971850e-01 -6.06196940e-01 1.02578604e+00
2.09913135e-01 -1.00192773e+00 -2.65579045e-01 2.86509842e-01
5.81191659e-01 3.99700403e-01 6.90395355e-01 -3.78784865e-01
-3.64436686e-01 -1.05781877e+00 3.34850430e-01 1.61648586e-01
1.27037811e+00 -3.89237195e-01 4.30031866e-01 -2.01388448e-01
-7.41370916e-01 -1.01103939e-01 -1.45882100e-01 7.40560472e-01
3.98693457e-02 5.99465430e-01 -1.03452182e+00 3.71654153e-01
6.76648259e-01 5.11091769e-01 -5.70346534e-01 1.19714773e+00
-6.92947432e-02 5.26285112e-01 -3.08517724e-01 4.36827868e-01
2.75878668e-01 -4.40381318e-01 5.06134778e-02 1.14142013e+00
3.42244834e-01 5.03535390e-01 3.43904272e-02 7.96621799e-01
3.67590010e-01 9.31736678e-02 -2.31724679e-01 -2.72624910e-01
1.88484281e-01 1.30905592e+00 -1.22492325e+00 3.08750905e-02
-6.04922116e-01 9.42943454e-01 -3.53908420e-01 7.24592626e-01
-4.99801427e-01 2.13489383e-01 5.74688554e-01 7.39468262e-02
-2.48537958e-01 -1.76401347e-01 -6.11140370e-01 -6.50995553e-01
-3.57820272e-01 -7.40165770e-01 6.75360620e-01 -1.16679680e+00
-1.07211256e+00 5.17014265e-01 -4.26839702e-02 -8.72743189e-01
2.81989485e-01 -1.06100619e+00 -5.22173233e-02 7.82247841e-01
-1.97248816e+00 -1.47256362e+00 -7.49721169e-01 7.80210376e-01
1.96600512e-01 3.63216966e-01 6.96294725e-01 1.90142214e-01
-6.97955608e-01 2.55841136e-01 3.34782898e-01 -2.35106096e-01
6.27910852e-01 -1.37369931e+00 -5.21465540e-01 2.61639565e-01
-4.12204973e-02 3.69310200e-01 4.78625894e-01 -5.24413288e-01
-1.49156702e+00 -7.92046249e-01 -1.62786141e-01 -1.04891405e-01
2.57786542e-01 3.67188364e-01 -6.40843809e-01 3.87510598e-01
5.40363304e-02 6.03368096e-02 1.44583237e+00 -2.71100134e-01
-2.47421771e-01 -8.65720958e-02 -1.48536909e+00 3.67808014e-01
5.11138558e-01 -7.89085150e-01 1.14267217e-02 6.65435731e-01
6.13982141e-01 -5.73144674e-01 -9.54792380e-01 5.28013468e-01
6.52407110e-01 -1.28473461e+00 1.35564506e+00 1.62181988e-01
2.39902154e-01 -1.56018108e-01 -1.49563149e-01 -9.49242830e-01
1.85828984e-01 1.97878331e-01 7.32883275e-01 5.29908001e-01
3.64467889e-01 -6.34785771e-01 1.17519975e+00 6.67410612e-01
-6.61513984e-01 -5.23745060e-01 -1.10436630e+00 -4.55388218e-01
-4.55403887e-02 -6.36605918e-01 2.87781179e-01 9.26270664e-01
-1.13773197e-01 -5.63652515e-01 3.50995779e-01 1.99099928e-01
7.85251975e-01 2.70518035e-01 -2.19500929e-01 -1.08210361e+00
-1.84860572e-01 -6.95785999e-01 -2.38606051e-01 -8.23023468e-02
5.12397811e-02 -9.36102808e-01 3.21902372e-02 -1.87691593e+00
3.84953588e-01 -4.54849869e-01 -3.51863444e-01 5.99451065e-01
-3.29212189e-01 9.03978705e-01 -5.21279685e-02 3.56352001e-01
-8.75026919e-03 1.15721822e-02 1.66981626e+00 -4.50393170e-01
-4.00741816e-01 7.86398258e-03 -6.41716361e-01 2.88560569e-01
7.21157432e-01 2.83374395e-02 -3.87168080e-01 -1.09306641e-01
-1.10844918e-01 1.82772979e-01 4.89467293e-01 -9.79672074e-01
3.60417701e-02 -2.90295273e-01 4.07721847e-01 -3.90951723e-01
5.55423081e-01 -9.24305618e-01 3.48152041e-01 6.03163004e-01
-1.21705674e-01 -6.43522859e-01 2.00384066e-01 4.13105577e-01
-1.43221304e-01 4.18621413e-02 9.45519328e-01 -5.06492913e-01
-8.74553323e-01 3.35626155e-01 -2.53495574e-01 -4.16781336e-01
8.97672474e-01 -9.04922247e-01 -5.98943889e-01 6.81795320e-03
-9.49834287e-01 -4.13309902e-01 6.24903500e-01 -2.32550904e-01
5.27154505e-01 -7.23578036e-01 -1.89623132e-01 5.42496750e-03
3.73148233e-01 -2.23132372e-02 8.36289465e-01 1.15695608e+00
-1.00245702e+00 4.74606663e-01 -3.42744410e-01 -9.37682390e-01
-1.26996660e+00 3.93181831e-01 5.56105077e-01 2.99058974e-01
-5.96152008e-01 7.15582192e-01 1.83899067e-02 -2.63964921e-01
-6.01761267e-02 -3.98139209e-01 -3.14506859e-01 1.02069043e-01
3.19246918e-01 4.08516914e-01 4.85645503e-01 -7.12343812e-01
-3.66254896e-01 7.20233500e-01 5.14814138e-01 -1.81199580e-01
1.43631649e+00 -3.34958792e-01 -2.14301243e-01 4.48668420e-01
1.24167979e+00 -4.20367628e-01 -1.14437973e+00 -2.47632619e-02
-3.94438326e-01 -3.52294505e-01 4.54954505e-01 -1.26569879e+00
-1.41840124e+00 1.10513914e+00 1.23131394e+00 1.69583261e-01
1.70030093e+00 1.40526429e-01 6.60268247e-01 -4.44156200e-01
3.30013573e-01 -9.75145698e-01 1.72336459e-01 -3.47241938e-01
4.50210005e-01 -1.62303460e+00 6.30343184e-02 -1.00072968e+00
-6.04858041e-01 1.15897477e+00 2.69254416e-01 6.19684994e-01
3.50022942e-01 2.24993583e-02 4.76453960e-01 -4.88518000e-01
2.01484337e-01 -6.31477594e-01 4.62965906e-01 1.07202375e+00
6.97516143e-01 2.96065241e-01 -2.22690906e-02 -2.17275023e-01
7.52971647e-03 -7.91478381e-02 4.07876313e-01 8.02711070e-01
-4.83204812e-01 -7.26806998e-01 -4.49089408e-01 4.40663368e-01
-5.02309620e-01 -2.01212928e-01 -2.31420144e-01 7.73174047e-01
2.35131662e-02 1.16663980e+00 -2.19256371e-01 -3.57704237e-02
1.11542298e-02 -1.82651896e-02 7.36587167e-01 -6.85853183e-01
-4.94144320e-01 4.32626039e-01 -1.31923661e-01 -6.67381942e-01
-9.26452279e-01 -7.75304258e-01 -1.34001756e+00 7.27042556e-02
-2.07906574e-01 -3.69665563e-01 1.27777839e+00 1.02005541e+00
-2.36037314e-01 5.74098408e-01 2.93040603e-01 -8.31714511e-01
-2.98832227e-02 -5.96329987e-01 -1.05928636e+00 2.18697011e-01
4.65510249e-01 -2.46291474e-01 -3.21966797e-01 8.78640264e-02] | [15.312244415283203, -2.9162139892578125] |
f6489a74-8649-4a1d-b706-fbdbebe1d7f8 | learning-to-select-nodes-in-bounded | null | null | https://openreview.net/forum?id=ztEQOAzM1cN | https://openreview.net/pdf?id=ztEQOAzM1cN | Learning to Select Nodes in Bounded Suboptimal Conflict-Based Search for Multi-Agent Path Finding | Multi-Agent Path Finding is an NP-hard problem that is difficult for current approaches to solve optimally. Research has shown that bounded suboptimal solvers, such as Enhanced Conflict-Based Search (ECBS), are more efficient than optimal solvers in finding a feasible solution with suboptimality guarantees. ECBS is a tree search algorithm that repeatedly selects nodes from a focal list to expand the tree. In this work, we propose to use imitation learning and curriculum learning to learn node-selection strategies for different grid maps and agent sizes. We then deploy the learned models in ECBS and test their solving performance on unseen instances drawn from the same distribution as the one used in training. Our approach shows substantial improvement over the baselines on different grid maps. | ['Sven Koenig', 'Bistra Dilkina', 'Taoan Huang'] | 2020-10-17 | null | null | null | neurips-workshop-lmca-2020-12 | ['multi-agent-path-finding'] | ['playing-games'] | [-1.32059976e-01 2.39835009e-01 -5.33960700e-01 1.95468292e-01
-1.06586385e+00 -8.84679377e-01 2.07460746e-01 -4.67717014e-02
-5.07102132e-01 1.45349789e+00 -1.42843038e-01 -4.78249669e-01
-4.63436782e-01 -9.31039810e-01 -8.78476560e-01 -7.00556993e-01
-5.82343280e-01 1.33321381e+00 3.74381304e-01 -1.80733338e-01
4.60856825e-01 3.44669461e-01 -1.04636276e+00 1.99828744e-02
1.09207761e+00 4.00641292e-01 3.92140508e-01 7.71698892e-01
-4.67643477e-02 5.34862638e-01 -8.16322744e-01 -7.76712596e-02
5.19265771e-01 -3.25854838e-01 -1.00790179e+00 -4.40080315e-02
2.94958651e-01 -4.42271411e-01 -2.46703535e-01 6.97446108e-01
3.24339390e-01 1.37853771e-01 4.45019424e-01 -1.63004553e+00
-1.95393145e-01 7.85083175e-01 -9.14631784e-01 4.24818635e-01
4.52327460e-01 2.51656175e-01 1.18256104e+00 -1.08646952e-01
1.13087082e+00 1.36706829e+00 4.42388058e-01 4.89003867e-01
-1.37812614e+00 -8.90353978e-01 5.38693011e-01 6.45587564e-01
-1.18779171e+00 7.54813924e-02 3.46273750e-01 3.71508934e-02
1.43228149e+00 1.30090758e-01 8.22265923e-01 1.12938797e+00
2.64523178e-01 6.08823419e-01 1.20888066e+00 -3.70314419e-01
4.84799713e-01 -1.09606020e-01 -3.83610934e-01 7.47052133e-01
4.24292713e-01 4.81103301e-01 -3.93908948e-01 -4.53666955e-01
8.10110450e-01 -7.13463902e-01 -1.59827679e-01 -6.14990413e-01
-9.88636792e-01 1.25975204e+00 5.22072315e-01 -4.89512458e-02
-5.20577431e-01 5.31033099e-01 2.42250904e-01 4.09292042e-01
6.23556450e-02 1.00864494e+00 -5.49101949e-01 -1.27599210e-01
-5.72503746e-01 9.10435379e-01 1.21190953e+00 9.60726023e-01
7.23946273e-01 -1.76718414e-01 -9.95125435e-03 5.24609685e-01
-1.57136872e-01 3.94179225e-01 -1.09138405e-02 -1.56222153e+00
5.93188763e-01 3.50357920e-01 4.71501052e-01 -1.04254270e+00
-7.08574414e-01 -4.60280567e-01 -8.54834020e-02 4.30626929e-01
4.35487717e-01 -6.66417539e-01 -8.08996499e-01 1.77212954e+00
6.32835448e-01 5.66281259e-01 1.03384033e-01 9.14528072e-01
3.37868392e-01 1.10718358e+00 -5.87245673e-02 -2.03619003e-01
7.92495191e-01 -1.45325685e+00 -1.90971419e-01 -4.26210076e-01
7.79603958e-01 -1.83745041e-01 5.65281272e-01 5.30209720e-01
-1.15992773e+00 3.12977314e-01 -9.57945585e-01 4.71858710e-01
-1.70141265e-01 -6.22549117e-01 8.09014797e-01 4.80940759e-01
-1.34912145e+00 4.76551056e-01 -8.32488954e-01 -4.06422019e-01
4.10573512e-01 4.08319533e-01 -1.17459260e-01 -3.09620291e-01
-9.04733658e-01 1.25878525e+00 5.74422240e-01 -3.96828145e-01
-1.39111030e+00 -5.68428934e-01 -7.58818209e-01 1.54245287e-01
1.12683165e+00 -7.12550282e-01 1.41289341e+00 -6.57799184e-01
-1.44826353e+00 3.96221101e-01 -2.53862161e-02 -6.90003574e-01
3.02801996e-01 3.68168771e-01 2.43050188e-01 2.62291789e-01
2.80681014e-01 9.31091487e-01 4.75304157e-01 -1.39197612e+00
-1.13099897e+00 3.37416567e-02 4.67202693e-01 5.21791279e-01
-1.29530653e-02 -8.26447308e-02 -4.02157605e-01 -4.22918573e-02
-8.71491358e-02 -1.19149911e+00 -8.55962932e-01 -4.55964953e-01
-3.81570548e-01 -5.71957827e-01 3.54998648e-01 -3.51261556e-01
9.78735149e-01 -1.35945511e+00 7.49066472e-01 4.81005311e-01
-1.48670580e-02 -2.42339090e-01 -8.05172861e-01 7.50054598e-01
5.37955344e-01 1.71190307e-01 9.65502784e-02 6.90552145e-02
1.27944991e-01 6.03780091e-01 1.05331456e-02 3.18943441e-01
-1.21888772e-01 7.83721685e-01 -1.08192468e+00 -5.37797272e-01
-1.18147559e-01 -2.49062046e-01 -9.24004436e-01 -1.09174855e-01
-8.27860594e-01 2.15856701e-01 -6.23526514e-01 6.65903509e-01
5.75474977e-01 -3.59898090e-01 7.82625020e-01 5.40291727e-01
-1.72597095e-01 2.09563747e-01 -1.25292945e+00 1.71158326e+00
-4.10102308e-01 5.28955698e-01 3.31057310e-01 -1.00624228e+00
4.99050856e-01 3.58963013e-02 4.47587192e-01 -1.03294683e+00
-1.96283385e-01 2.68943906e-01 2.91265577e-01 -4.15071577e-01
3.36205184e-01 3.22583199e-01 -1.61893174e-01 7.63156295e-01
-2.76642352e-01 -3.08617204e-01 6.23234749e-01 3.06847274e-01
1.66584396e+00 1.68187648e-01 7.82637149e-02 -1.59338087e-01
1.04128398e-01 8.06636989e-01 8.97100210e-01 1.08482540e+00
-1.30034223e-01 -5.73063157e-02 8.06734443e-01 -5.88382483e-01
-1.10078824e+00 -7.77399778e-01 1.85916021e-01 1.21915042e+00
5.35056174e-01 -4.39455152e-01 -6.91043496e-01 -7.55291939e-01
2.18980402e-01 8.78684163e-01 -5.14428735e-01 2.77382672e-01
-1.08781230e+00 -5.65321028e-01 6.19765073e-02 3.00021976e-01
3.71172696e-01 -1.26919830e+00 -9.60548997e-01 4.78916734e-01
-1.65752873e-01 -1.05411494e+00 -3.38425487e-01 1.33276060e-01
-5.90039194e-01 -1.34364164e+00 -4.80094105e-01 -9.65873182e-01
7.02176511e-01 4.31563109e-01 1.33463752e+00 4.00768757e-01
-3.03541452e-01 3.89439553e-01 -4.45981145e-01 -2.86019236e-01
-3.33707899e-01 6.74381495e-01 -2.69308418e-01 -9.73443985e-01
-7.49929547e-02 -6.12244189e-01 -5.58274329e-01 4.77747202e-01
-2.09976673e-01 -6.09198622e-02 5.07011950e-01 9.46024954e-01
5.44507027e-01 4.38756555e-01 6.29417658e-01 -6.70557082e-01
8.58461738e-01 -6.80171609e-01 -1.13922465e+00 3.72366130e-01
-6.71607971e-01 6.63219299e-03 6.29296362e-01 -4.30518836e-01
-7.48792231e-01 -6.47323355e-02 3.46543998e-01 -1.05978951e-01
-3.12322304e-02 6.63554668e-01 3.75499457e-01 -3.93348962e-01
5.90821981e-01 -2.42194571e-02 -1.94524303e-01 -9.99734923e-02
2.25410506e-01 1.97785208e-03 2.30891168e-01 -9.82401431e-01
8.14675212e-01 1.93093315e-01 5.76083176e-02 -2.79194146e-01
-4.82818693e-01 -8.57294053e-02 1.97920371e-02 -2.47019827e-02
4.79491174e-01 -6.77081168e-01 -8.58675659e-01 -8.10883567e-02
-1.07345581e+00 -1.02733588e+00 2.00416952e-01 1.41804039e-01
-7.35952675e-01 -4.08583581e-02 -5.06746233e-01 -6.14443600e-01
-1.09861121e-01 -1.11404049e+00 7.44407356e-01 5.29206157e-01
-1.68127999e-01 -9.19450462e-01 5.73912740e-01 2.69223183e-01
3.80177110e-01 4.32196796e-01 9.61625338e-01 -4.53770608e-01
-1.17593241e+00 1.51815891e-01 -4.46735770e-02 -7.20646441e-01
-3.81192982e-01 -1.62305683e-02 -9.75123271e-02 -7.38968551e-01
-7.19568312e-01 -5.63559175e-01 5.83391607e-01 7.57527649e-01
8.70305479e-01 -6.18949056e-01 -1.04753172e+00 6.30854905e-01
1.53835261e+00 3.50697160e-01 4.73697811e-01 1.20176935e+00
-1.63468793e-02 5.85061252e-01 9.39269602e-01 5.34888029e-01
8.40053618e-01 7.34264433e-01 7.41710722e-01 1.25965729e-01
2.55499899e-01 -1.02729194e-01 1.56587079e-01 -8.96360800e-02
9.50516164e-02 -4.84094650e-01 -1.11260164e+00 8.25123191e-01
-2.15048623e+00 -9.88679171e-01 2.45494545e-01 1.56537664e+00
7.94951499e-01 1.34189487e-01 4.16154832e-01 -4.39688712e-01
5.40645659e-01 -1.02250598e-01 -7.88943350e-01 -6.40464962e-01
-3.28769311e-02 3.01795185e-01 6.80736005e-01 7.91602671e-01
-8.86741221e-01 1.34752476e+00 7.10805321e+00 6.35539651e-01
-7.42081225e-01 -1.66646056e-02 3.68893117e-01 -5.08717954e-01
-1.97196171e-01 1.82033762e-01 -6.54439509e-01 2.44841009e-01
8.42827976e-01 -4.43878382e-01 1.08543622e+00 8.43826830e-01
1.52684465e-01 -4.69116002e-01 -1.12370622e+00 4.83099759e-01
-1.34145871e-01 -1.69147134e+00 -4.55955088e-01 2.84783006e-01
1.36362767e+00 1.73340708e-01 2.34358478e-02 4.36149955e-01
1.21452510e+00 -1.25191367e+00 4.21318769e-01 -2.86422133e-01
1.59886301e-01 -1.00988221e+00 6.41740561e-01 5.66443443e-01
-8.86118948e-01 -4.88535315e-01 -5.11755705e-01 -1.73324376e-01
2.99998432e-01 -3.54837865e-01 -1.16215384e+00 3.91583443e-01
9.04557288e-01 3.00306439e-01 -1.68416008e-01 1.58435595e+00
-3.82951975e-01 4.56773490e-01 -5.52902818e-01 -3.75011742e-01
8.18686306e-01 -2.16763720e-01 8.07368338e-01 8.43114495e-01
2.49189630e-01 1.90254837e-01 5.81789851e-01 8.87705863e-01
2.57197738e-01 -1.05026528e-01 -5.18505871e-01 1.66568071e-01
1.02859807e+00 1.28025591e+00 -8.85024130e-01 -3.79698612e-02
-1.30669430e-01 5.46032429e-01 8.26839507e-01 5.55034578e-01
-1.05730736e+00 -7.40345642e-02 5.82840800e-01 -3.09283912e-01
6.77440584e-01 -2.88023442e-01 -9.15285870e-02 -5.37477911e-01
-3.11201364e-01 -1.23089492e+00 8.09247673e-01 -5.39060056e-01
-9.36933815e-01 4.69554693e-01 3.16383272e-01 -7.59088159e-01
-5.12702703e-01 -3.85121346e-01 -8.12580287e-01 6.45574629e-01
-1.70603263e+00 -8.26232731e-01 -3.31002235e-01 2.08382875e-01
6.51544929e-01 -1.55918866e-01 7.31770992e-01 -2.05713600e-01
-8.15934658e-01 5.15174210e-01 5.55450767e-02 -3.09235930e-01
2.12553471e-01 -1.24280787e+00 3.34030807e-01 5.81832170e-01
-5.84984617e-03 1.76252753e-01 8.89463067e-01 -6.58608794e-01
-1.65139771e+00 -6.40486360e-01 2.42798030e-01 -1.86444640e-01
7.43164599e-01 8.01561691e-04 -5.76486647e-01 8.85317981e-01
5.94274461e-01 -4.47582453e-01 2.46874899e-01 2.06216350e-01
-7.11779594e-02 2.37368599e-01 -1.16376197e+00 7.61082470e-01
1.07224035e+00 3.54439050e-01 -2.31116951e-01 5.71744084e-01
6.49811387e-01 -7.94839144e-01 -5.05797148e-01 1.36772797e-01
3.35460156e-01 -7.58022845e-01 1.00082195e+00 -8.29135776e-01
4.11243707e-01 -8.58530328e-02 4.45729084e-02 -1.91030157e+00
-5.25463879e-01 -7.72109270e-01 -1.92086659e-02 5.77929080e-01
7.20585704e-01 -8.31627071e-01 1.28654647e+00 4.96056169e-01
1.36274800e-01 -9.20492351e-01 -1.20476127e+00 -9.44668770e-01
5.91473103e-01 2.30807573e-01 7.45419204e-01 6.79670393e-01
2.77562767e-01 1.39641777e-01 -1.09276406e-01 5.03444850e-01
8.34964514e-01 6.06945574e-01 1.00157833e+00 -9.64844584e-01
-4.79776502e-01 -6.73271775e-01 2.61283875e-01 -8.73825371e-01
5.89279652e-01 -8.25009048e-01 1.96341321e-01 -1.96137285e+00
3.11401576e-01 -9.66876447e-01 1.64013177e-01 5.21184325e-01
7.15610534e-02 -3.04221898e-01 2.93491274e-01 -3.24413553e-02
-9.99063134e-01 3.34864050e-01 1.47155499e+00 -3.18397939e-01
-3.88156474e-01 -2.38685757e-01 -4.31901962e-01 5.66389382e-01
1.06145298e+00 -7.23138630e-01 -2.85037160e-01 -6.23415112e-01
4.00148988e-01 5.77912092e-01 2.19004415e-02 -7.28030145e-01
4.87533569e-01 -8.97194326e-01 -5.90028390e-02 -5.60617089e-01
1.85889944e-01 -8.27690303e-01 9.19254050e-02 8.55723262e-01
-4.19288546e-01 3.51925164e-01 4.06947941e-01 6.12674892e-01
3.43491733e-01 -6.81976140e-01 1.97865278e-01 -3.78308922e-01
-8.80690098e-01 1.90376490e-01 -5.10220945e-01 1.84116691e-01
1.56872725e+00 -2.35289529e-01 -7.11446762e-01 -6.26352668e-01
-3.71102244e-01 1.20215690e+00 5.28006077e-01 4.93720360e-02
5.75773478e-01 -9.92421448e-01 -9.36113060e-01 -5.08522034e-01
-2.99794257e-01 1.45282388e-01 -2.89269853e-02 7.39434242e-01
-7.88154483e-01 4.43982035e-01 -3.47255409e-01 -2.59937316e-01
-1.27072394e+00 5.47031164e-01 6.00809813e-01 -6.85848475e-01
-7.10468829e-01 7.88042188e-01 -2.50656635e-01 -4.90359902e-01
4.14131790e-01 9.06793624e-02 -1.04454018e-01 -2.76194066e-01
3.55645299e-01 6.47862375e-01 -5.47330022e-01 2.17187461e-02
-4.53928232e-01 4.49065030e-01 -2.48413339e-01 -3.16247493e-01
1.53366268e+00 1.38472229e-01 -5.75727448e-02 -5.49307764e-01
6.37802422e-01 -1.76188618e-01 -1.33679080e+00 -3.43822762e-02
2.71922976e-01 -5.51083088e-01 2.83004660e-02 -1.11162412e+00
-1.03192782e+00 4.89503844e-03 1.04550816e-01 3.20178926e-01
6.79216385e-01 5.97001724e-02 7.55888343e-01 6.31740510e-01
1.03826010e+00 -1.26555300e+00 2.39255682e-01 4.04098421e-01
8.91545236e-01 -1.12596929e+00 2.12443665e-01 -6.09903634e-01
-5.07162809e-01 1.01523519e+00 1.11149538e+00 -5.39911091e-01
-4.79687713e-02 5.70222318e-01 -2.17255905e-01 -2.05490783e-01
-1.36794591e+00 -2.97509730e-01 -4.43716824e-01 9.24617827e-01
-4.78853673e-01 -4.37577582e-05 -3.15102071e-01 5.87880313e-02
-2.57832855e-01 -1.22509666e-01 7.64605403e-01 9.96432066e-01
-5.66826582e-01 -1.18466878e+00 -5.84542513e-01 3.16875994e-01
1.87774152e-02 2.65153974e-01 -4.72957164e-01 1.03390121e+00
-2.04163179e-01 1.06413412e+00 2.70899862e-01 -1.74604774e-01
1.03132293e-01 -4.50101316e-01 1.02806211e+00 -5.60276806e-01
-5.74471891e-01 -2.69763148e-03 5.88773727e-01 -8.29988122e-01
-7.88692832e-02 -7.20601320e-01 -1.57905531e+00 -5.21325052e-01
-2.67467648e-01 4.09273863e-01 3.51168334e-01 6.82937205e-01
3.73800516e-01 4.65075076e-01 5.15811563e-01 -9.50799465e-01
-7.23863423e-01 -4.95998442e-01 -1.43913582e-01 -8.20362195e-02
1.95573345e-02 -8.78857434e-01 -3.41008693e-01 -8.01755369e-01] | [4.962387561798096, 2.0564544200897217] |
5ddb7325-6b3b-4233-a62a-3cf01fd7255a | inductive-and-transductive-few-shot-video | 2207.10785 | null | https://arxiv.org/abs/2207.10785v1 | https://arxiv.org/pdf/2207.10785v1.pdf | Inductive and Transductive Few-Shot Video Classification via Appearance and Temporal Alignments | We present a novel method for few-shot video classification, which performs appearance and temporal alignments. In particular, given a pair of query and support videos, we conduct appearance alignment via frame-level feature matching to achieve the appearance similarity score between the videos, while utilizing temporal order-preserving priors for obtaining the temporal similarity score between the videos. Moreover, we introduce a few-shot video classification framework that leverages the above appearance and temporal similarity scores across multiple steps, namely prototype-based training and testing as well as inductive and transductive prototype refinement. To the best of our knowledge, our work is the first to explore transductive few-shot video classification. Extensive experiments on both Kinetics and Something-Something V2 datasets show that both appearance and temporal alignments are crucial for datasets with temporal order sensitivity such as Something-Something V2. Our approach achieves similar or better results than previous methods on both datasets. Our code is available at https://github.com/VinAIResearch/fsvc-ata. | ['Rang Nguyen', 'Binh-Son Hua', 'Khoi Nguyen', 'Quoc-Huy Tran', 'Khoi D. Nguyen'] | 2022-07-21 | null | null | null | null | ['video-classification', 'classification'] | ['computer-vision', 'methodology'] | [-1.18518593e-02 -6.93152726e-01 -5.61035395e-01 -5.00625014e-01
-9.15338516e-01 -5.76164305e-01 7.24435210e-01 5.53439707e-02
-2.14202031e-01 2.34436989e-01 3.81101258e-02 1.51656955e-01
-2.29021069e-02 -2.45224938e-01 -6.77264810e-01 -5.28509378e-01
-3.61141324e-01 1.60607606e-01 6.43249929e-01 -7.50859454e-03
2.88484186e-01 1.92008898e-01 -1.63293839e+00 5.01054168e-01
2.45184362e-01 1.26657557e+00 -4.70992960e-02 8.88456285e-01
3.56163353e-01 1.01328516e+00 8.62684175e-02 -2.34760031e-01
4.84474152e-01 -6.01641834e-01 -7.65328825e-01 2.70276427e-01
8.30224097e-01 -4.62916136e-01 -5.24174869e-01 8.44177961e-01
2.89633185e-01 6.03697717e-01 7.12885737e-01 -1.48413849e+00
-5.24004400e-01 5.10229953e-02 -6.59870088e-01 6.98226094e-01
7.23943472e-01 4.64221001e-01 1.13256514e+00 -1.23465824e+00
8.50681543e-01 9.51109350e-01 6.97421610e-01 4.42486405e-01
-1.20621979e+00 -4.59983885e-01 2.52551287e-01 6.39324844e-01
-1.41548812e+00 -8.60691667e-01 8.11351418e-01 -7.66715109e-01
8.75490725e-01 1.35797083e-01 9.24177587e-01 1.11462605e+00
2.91041266e-02 9.29471970e-01 8.17824662e-01 -2.13287428e-01
3.28345656e-01 -2.11737886e-01 1.67432338e-01 8.73730958e-01
-3.73473078e-01 1.53095737e-01 -6.42488658e-01 -1.02661185e-01
6.19242072e-01 3.13974768e-01 -1.28975928e-01 -7.24690139e-01
-1.21957505e+00 6.45076573e-01 6.51235729e-02 3.14910442e-01
-3.94639969e-01 3.89817119e-01 4.99883473e-01 5.72448134e-01
3.92651618e-01 2.15990961e-01 -2.64761031e-01 -4.20730233e-01
-1.11378336e+00 5.52847311e-02 5.31055331e-01 9.82139349e-01
8.97259831e-01 -5.68452999e-02 -3.11095417e-01 6.43779516e-01
1.50986835e-01 1.70170665e-01 4.44947034e-01 -1.29642034e+00
1.11734085e-01 2.53675699e-01 -4.22290564e-02 -1.03324342e+00
-6.20127507e-02 3.19091350e-01 -2.92686015e-01 -4.42540593e-04
4.50739443e-01 3.66929471e-01 -8.99588645e-01 1.76266611e+00
4.08881158e-01 9.66538131e-01 -2.14349225e-01 1.07738638e+00
9.42841470e-01 4.70275581e-01 1.24928460e-01 -3.91211987e-01
1.37340689e+00 -1.20246863e+00 -5.24773180e-01 2.48747095e-02
6.42616451e-01 -7.83072352e-01 1.03340554e+00 2.76935026e-02
-1.30929255e+00 -6.68781400e-01 -1.00882292e+00 2.82429419e-02
-1.04308017e-01 -4.99818213e-02 6.59354806e-01 2.88181812e-01
-1.01221418e+00 7.12782323e-01 -1.02816927e+00 -7.64267564e-01
4.35492694e-01 5.24646565e-02 -5.22454739e-01 -1.76901370e-01
-1.02951324e+00 4.76121724e-01 4.54384536e-02 -2.84624219e-01
-1.10499144e+00 -7.99612880e-01 -8.29257071e-01 -3.06442529e-01
4.92250532e-01 -6.22576118e-01 1.46010959e+00 -1.10647810e+00
-1.35173416e+00 9.81132388e-01 -2.78071940e-01 -3.28843594e-01
4.40863609e-01 -6.97973967e-02 -3.59259516e-01 5.28093874e-01
1.27124950e-01 7.08666146e-01 1.00701237e+00 -9.58374560e-01
-5.64801335e-01 -2.40803227e-01 1.81672677e-01 2.92206049e-01
-2.18092605e-01 1.72874480e-01 -9.78554547e-01 -6.33253455e-01
7.81025067e-02 -8.82114053e-01 7.09889606e-02 3.61421734e-01
3.45376879e-02 -1.19980119e-01 8.15462589e-01 -5.36861300e-01
1.07096136e+00 -2.25248861e+00 1.70862764e-01 8.12646225e-02
2.04899281e-01 1.93688646e-02 -3.97307694e-01 4.00878906e-01
-1.09974623e-01 -9.29797962e-02 9.07116234e-02 -3.79553080e-01
-2.52748609e-01 -3.76231968e-02 -1.03963293e-01 7.06903815e-01
2.03334600e-01 1.11459625e+00 -1.11305058e+00 -7.01957047e-01
2.96514750e-01 2.99341440e-01 -6.19465351e-01 3.01003277e-01
-1.50506809e-01 3.10483009e-01 -4.25917894e-01 1.01980174e+00
2.21803933e-01 -3.71294439e-01 1.33845150e-01 -3.67189556e-01
-4.89760488e-02 -3.65449414e-02 -8.31955612e-01 1.96199644e+00
-3.53054777e-02 7.46545613e-01 -2.36726552e-01 -1.18348658e+00
5.76052308e-01 5.27038872e-01 1.00646579e+00 -7.89201915e-01
2.03984469e-01 -3.27852964e-02 -3.39295328e-01 -6.83877230e-01
3.80844474e-01 -1.10052601e-01 6.50619864e-02 4.12533551e-01
2.86037773e-01 -5.38054258e-02 5.10661185e-01 4.71450955e-01
1.27823997e+00 4.45056885e-01 4.53951418e-01 7.08519369e-02
2.92487681e-01 8.49989280e-02 6.32946491e-01 6.27962589e-01
-7.07964480e-01 7.59611845e-01 2.76600063e-01 -5.83668709e-01
-1.17224228e+00 -1.18492186e+00 2.59362221e-01 1.20868158e+00
4.22372669e-01 -6.01024270e-01 -4.85327899e-01 -6.19253039e-01
-2.66798452e-05 2.84832686e-01 -8.76806259e-01 -2.12868467e-01
-5.16133249e-01 -9.70705748e-02 2.78467178e-01 6.50211453e-01
1.43292069e-01 -8.86918366e-01 -3.93456846e-01 4.87952568e-02
-1.63626537e-01 -1.23621702e+00 -8.57344747e-01 -1.77285075e-01
-9.28699493e-01 -1.20949614e+00 -4.68930662e-01 -7.74956882e-01
4.92619246e-01 6.79159284e-01 9.38780487e-01 1.59042358e-01
-5.37251174e-01 9.13668156e-01 -6.05043292e-01 2.03162208e-01
-5.86543493e-02 -4.31205869e-01 2.69303352e-01 7.22039491e-02
4.53533381e-01 -6.55840278e-01 -8.08702826e-01 7.14287043e-01
-6.18965387e-01 3.12216356e-02 3.53374839e-01 7.45883584e-01
8.27005923e-01 -4.56915587e-01 1.94919109e-01 -4.35736984e-01
1.47914231e-01 -6.12812638e-01 -3.30244780e-01 4.12539303e-01
-3.54067087e-01 -3.51886392e-01 4.02529150e-01 -8.14826131e-01
-7.89579153e-01 1.88579500e-01 3.24686050e-01 -1.14671922e+00
-1.95680290e-01 2.64812112e-01 3.28960866e-01 -5.25355488e-02
6.08293891e-01 2.69565910e-01 -6.88807620e-03 -1.98377907e-01
4.30897743e-01 3.04960757e-01 6.17267430e-01 -5.93869686e-01
8.78915250e-01 7.30659068e-01 -2.43796393e-01 -8.41004908e-01
-6.98808551e-01 -1.04000425e+00 -8.95118892e-01 -5.42339802e-01
1.07943308e+00 -9.40042257e-01 -5.52042186e-01 3.01035881e-01
-8.53562355e-01 -4.67046022e-01 -3.89482647e-01 7.75596261e-01
-1.03975403e+00 4.96854722e-01 -6.42233789e-01 -5.31648517e-01
-1.44778416e-01 -1.04183733e+00 9.96273220e-01 1.09387688e-01
-3.32866788e-01 -9.05191422e-01 4.08873707e-01 4.78773475e-01
9.30230916e-02 8.83717015e-02 3.45147073e-01 -5.90586245e-01
-6.50142133e-01 -1.92893386e-01 -6.62055984e-02 -4.83215824e-02
2.19197884e-01 3.17565441e-01 -7.82716751e-01 -4.72112179e-01
-2.49838918e-01 -6.19875729e-01 1.08847225e+00 5.30756414e-01
8.03672910e-01 -5.28106689e-02 -3.02204102e-01 7.42371023e-01
1.16200078e+00 3.94405037e-01 5.49502611e-01 1.08040571e-01
6.53960288e-01 3.63286406e-01 1.09935653e+00 5.03986835e-01
2.88803458e-01 1.02676582e+00 2.51420438e-02 6.61116093e-02
-2.82925725e-01 -1.32888645e-01 5.11299074e-01 9.62423742e-01
-3.42402458e-01 -3.52899171e-02 -6.68164849e-01 5.28242528e-01
-2.14664340e+00 -1.74031138e+00 2.85198182e-01 2.08874106e+00
6.36081338e-01 -1.36731073e-01 4.13703293e-01 -2.35977575e-01
7.53706515e-01 3.35415900e-01 -5.72725236e-01 -1.43174499e-01
1.31613150e-01 -1.87640205e-01 8.48034862e-03 2.05374032e-01
-1.42298281e+00 9.64017689e-01 6.20443630e+00 7.99478412e-01
-9.71668363e-01 2.96283156e-01 5.75933039e-01 -4.02267456e-01
7.29193166e-02 2.88839489e-01 -4.36714500e-01 3.61472040e-01
6.44086301e-01 -1.24676488e-01 5.57104051e-01 8.09247732e-01
3.17649722e-01 9.24410578e-03 -1.36714351e+00 1.03463387e+00
2.45751858e-01 -1.45299029e+00 -1.33985579e-01 -3.07880014e-01
8.10091317e-01 6.52059317e-02 6.03120252e-02 2.33903840e-01
5.51164337e-02 -5.38967013e-01 8.75768125e-01 7.23302603e-01
7.00543106e-01 -1.91171423e-01 3.12700331e-01 -2.00522333e-01
-1.79996657e+00 1.44029036e-02 -1.33358955e-01 4.70233411e-02
2.03073591e-01 1.36792764e-01 -5.19291699e-01 3.64137292e-01
8.13792348e-01 1.34640121e+00 -5.54967701e-01 1.11394119e+00
4.25427146e-02 4.85808700e-01 -7.99992532e-02 1.85668647e-01
9.65177044e-02 -2.18343079e-01 6.25035882e-01 1.14091098e+00
2.90395826e-01 3.86499703e-01 5.58311045e-01 4.86872286e-01
6.07540347e-02 2.12435111e-01 -6.57388091e-01 -2.52681196e-01
5.50749779e-01 1.33359456e+00 -8.41070414e-01 -5.53217351e-01
-8.59133780e-01 1.12500381e+00 3.76420826e-01 3.91715974e-01
-1.00108409e+00 8.21854845e-02 8.28413308e-01 1.17111795e-01
5.14399707e-01 -1.12786904e-01 2.75302827e-01 -1.47646058e+00
-4.59167771e-02 -7.74289250e-01 7.76865184e-01 -8.19459021e-01
-1.30214000e+00 2.06708416e-01 1.35558635e-01 -1.66601181e+00
-1.88386261e-01 -2.93184459e-01 -9.20788705e-01 -3.12773744e-03
-1.14123964e+00 -1.26125026e+00 -4.69666004e-01 8.22077215e-01
9.75010872e-01 -2.35977381e-01 5.15546262e-01 2.71339476e-01
-6.24878943e-01 4.63758975e-01 -2.68993806e-02 9.97724831e-02
9.63132858e-01 -9.09689188e-01 4.09903914e-01 7.88548470e-01
3.61686856e-01 6.12570822e-01 5.57325780e-01 -6.25821173e-01
-1.63187003e+00 -9.78104293e-01 3.13185215e-01 -3.46178383e-01
9.51401711e-01 -1.16664656e-01 -8.43010366e-01 7.63511419e-01
3.90054360e-02 3.98569375e-01 7.78713703e-01 5.75262569e-02
-6.75369203e-01 -6.40948787e-02 -8.07903647e-01 6.51452124e-01
1.34247780e+00 -7.36325741e-01 -4.92173940e-01 2.79461920e-01
5.14732301e-01 -1.58910930e-01 -1.03848255e+00 3.95603120e-01
9.66789305e-01 -9.88599837e-01 1.09559083e+00 -7.00767040e-01
3.43925267e-01 -4.66082811e-01 -4.86674309e-01 -1.01972449e+00
-5.27500749e-01 -6.11856937e-01 -3.99806887e-01 9.62708533e-01
2.28870332e-01 -2.95723408e-01 6.64524078e-01 4.13061351e-01
-9.25583243e-02 -7.68485785e-01 -8.58304203e-01 -1.11816180e+00
-4.49710399e-01 -5.32173991e-01 1.20012455e-01 1.03383648e+00
2.71730661e-01 1.61383629e-01 -5.93246341e-01 -7.77583122e-02
7.12937593e-01 3.83370668e-01 8.58307183e-01 -8.93877089e-01
-4.31450844e-01 -4.37918931e-01 -8.25873792e-01 -9.86445129e-01
9.17041227e-02 -7.46262431e-01 1.19119063e-01 -1.12286294e+00
6.82806194e-01 -1.11268617e-01 -6.04208410e-01 5.90264857e-01
6.19055284e-03 7.71112978e-01 2.02213958e-01 4.93461937e-01
-1.28981817e+00 4.24698979e-01 1.09384036e+00 -2.41561219e-01
-2.41779789e-01 -2.12355390e-01 -1.48640051e-01 6.01207197e-01
8.86929989e-01 -3.48892063e-01 -5.09540558e-01 -3.20788957e-02
-1.73735321e-01 2.44090423e-01 5.22037923e-01 -9.49631333e-01
3.36405814e-01 -4.32003736e-01 3.31025571e-01 -4.69615221e-01
6.96000695e-01 -4.69390422e-01 2.86445588e-01 4.54619259e-01
-4.09021646e-01 1.26760349e-01 2.03200765e-02 7.75455952e-01
-1.57024696e-01 9.50004160e-02 9.86285448e-01 -2.49388684e-02
-1.27440894e+00 7.02823818e-01 -4.21854973e-01 9.95529965e-02
1.45495725e+00 -5.39338350e-01 -3.98476154e-01 -5.92115283e-01
-7.61799037e-01 2.07637370e-01 7.97634125e-01 5.64108491e-01
9.51469183e-01 -1.64318395e+00 -5.11312306e-01 1.18577987e-01
5.30489743e-01 -8.13375413e-01 4.33299333e-01 1.45316195e+00
-1.59579068e-01 1.92827493e-01 -3.03616166e-01 -8.16205919e-01
-1.72051620e+00 7.97406018e-01 1.85101584e-01 1.93891436e-01
-6.09394252e-01 7.64167547e-01 2.50324279e-01 1.41173741e-02
1.11030839e-01 -3.91144343e-02 6.35036081e-03 1.18343666e-01
5.99190652e-01 4.44236487e-01 -3.37622672e-01 -8.65655899e-01
-5.67998648e-01 8.72971117e-01 -2.20378339e-01 -8.38750452e-02
1.19938624e+00 -1.87990546e-01 2.26824731e-01 6.35459781e-01
1.40562117e+00 -3.69843990e-01 -1.45466149e+00 -3.53835642e-01
-1.68846160e-01 -7.75982618e-01 -1.51852638e-01 -2.79367507e-01
-1.03537679e+00 6.00852430e-01 4.96496856e-01 3.55112851e-02
1.09362662e+00 2.77278900e-01 6.70041680e-01 3.30647260e-01
2.41412655e-01 -1.05568910e+00 7.67147064e-01 2.82606542e-01
6.00521564e-01 -1.33785939e+00 1.59668416e-01 -2.97322810e-01
-7.07498968e-01 8.98592472e-01 6.04984879e-01 -8.81803185e-02
5.27101278e-01 -6.49640849e-03 -2.23808251e-02 -4.13439512e-01
-1.21282029e+00 -3.80486757e-01 5.27146995e-01 4.15064573e-01
3.58424902e-01 -2.13505268e-01 -1.68582544e-01 9.62259695e-02
4.43776071e-01 6.12249188e-02 2.17464507e-01 1.04982471e+00
-4.27907884e-01 -8.68146539e-01 -1.30164670e-03 4.67197508e-01
-1.90799505e-01 -3.90112679e-03 -3.20210487e-01 6.78308010e-01
-2.69983113e-01 8.93386006e-01 2.51040220e-01 -7.34191358e-01
1.57559127e-01 9.07367766e-02 7.14273334e-01 -4.78872597e-01
-2.24329159e-01 1.47428304e-01 4.02990691e-02 -1.05261993e+00
-7.93494463e-01 -9.44572926e-01 -1.19287419e+00 -3.16525877e-01
-2.71148175e-01 4.79839891e-02 2.58659035e-01 7.62820899e-01
3.30621362e-01 -3.21632922e-02 9.13055718e-01 -1.15911734e+00
-2.33990982e-01 -6.38361692e-01 -5.21900952e-01 9.44100201e-01
1.53517559e-01 -8.69536757e-01 -4.09235120e-01 4.85363662e-01] | [8.645445823669434, 0.7589871287345886] |
f5289c30-7385-47bd-a634-4bf07215c84e | slap-improving-physical-adversarial-examples | 2007.04137 | null | https://arxiv.org/abs/2007.04137v3 | https://arxiv.org/pdf/2007.04137v3.pdf | SLAP: Improving Physical Adversarial Examples with Short-Lived Adversarial Perturbations | Research into adversarial examples (AE) has developed rapidly, yet static adversarial patches are still the main technique for conducting attacks in the real world, despite being obvious, semi-permanent and unmodifiable once deployed. In this paper, we propose Short-Lived Adversarial Perturbations (SLAP), a novel technique that allows adversaries to realize physically robust real-world AE by using a light projector. Attackers can project a specifically crafted adversarial perturbation onto a real-world object, transforming it into an AE. This allows the adversary greater control over the attack compared to adversarial patches: (i) projections can be dynamically turned on and off or modified at will, (ii) projections do not suffer from the locality constraint imposed by patches, making them harder to detect. We study the feasibility of SLAP in the self-driving scenario, targeting both object detector and traffic sign recognition tasks, focusing on the detection of stop signs. We conduct experiments in a variety of ambient light conditions, including outdoors, showing how in non-bright settings the proposed method generates AE that are extremely robust, causing misclassifications on state-of-the-art networks with up to 99% success rate for a variety of angles and distances. We also demostrate that SLAP-generated AE do not present detectable behaviours seen in adversarial patches and therefore bypass SentiNet, a physical AE detection method. We evaluate other defences including an adaptive defender using adversarial learning which is able to thwart the attack effectiveness up to 80% even in favourable attacker conditions. | ['Martin Strohmeier', 'Ivan Martinovic', 'Ivo Sluganovic', 'Giulio Lovisotto', 'Henry Turner'] | 2020-07-08 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 5.43016732e-01 2.47366220e-01 4.96729225e-01 2.75257349e-01
-2.81658471e-01 -1.12220252e+00 8.96581650e-01 -4.01428461e-01
-4.51348454e-01 5.90728164e-01 -3.40150744e-01 -4.02895361e-01
4.37089838e-02 -9.24565673e-01 -1.07578242e+00 -1.04710960e+00
-4.44616616e-01 1.44024417e-01 7.33773828e-01 -5.04113793e-01
-5.81717864e-02 1.19436288e+00 -1.57256866e+00 1.28249827e-04
3.65963817e-01 7.80912995e-01 -3.43905419e-01 8.64342332e-01
5.42797863e-01 5.90228915e-01 -1.32631946e+00 -4.51337904e-01
8.81748021e-01 -7.98710734e-02 -2.05289125e-01 -3.05890262e-01
6.85478687e-01 -1.11015536e-01 -5.45788348e-01 9.64814842e-01
6.95163548e-01 -6.45051152e-02 4.85536873e-01 -1.50146544e+00
-2.95136034e-01 1.46719426e-01 -2.82687843e-01 5.46029694e-02
4.68559921e-01 8.99862289e-01 4.42495674e-01 -2.08434016e-01
5.76094329e-01 1.11780119e+00 5.94841659e-01 8.18685174e-01
-1.30557156e+00 -9.45137024e-01 -1.37500614e-01 2.08419412e-01
-1.21881735e+00 -3.00950110e-01 8.89501929e-01 -9.37033743e-02
7.69851327e-01 8.54482353e-01 4.32723820e-01 1.74686396e+00
6.78502202e-01 1.42247707e-01 1.55808890e+00 -3.64401042e-01
4.56555456e-01 1.89827532e-01 -5.40284514e-01 3.22091341e-01
2.83850044e-01 8.67943764e-01 -2.32305795e-01 -3.22032392e-01
3.38023961e-01 -3.14963281e-01 -3.11286360e-01 -3.82081926e-01
-9.84882653e-01 6.53320193e-01 8.13813984e-01 6.35696799e-02
-4.47851121e-01 2.38552183e-01 1.42654970e-01 4.70846742e-01
-5.78729026e-02 8.25912595e-01 -1.06723882e-01 1.45481586e-01
-4.31276083e-01 2.32553154e-01 8.50749850e-01 3.71933043e-01
4.51847970e-01 3.97487223e-01 -1.34544775e-01 1.06461309e-01
-1.04595060e-02 1.19207740e+00 -6.69824630e-02 -4.63679135e-01
2.14662820e-01 2.29209453e-01 -2.19905414e-02 -1.21035731e+00
-3.90920132e-01 -2.90219069e-01 -6.42398238e-01 1.23948061e+00
4.92847145e-01 -4.47156250e-01 -9.98953938e-01 1.72493052e+00
5.32696903e-01 4.01822746e-01 6.39229864e-02 7.04187989e-01
2.93977916e-01 4.73326594e-01 -2.19473299e-02 4.04245369e-02
1.00852776e+00 -3.00424993e-01 -2.23665267e-01 -3.76242936e-01
9.18635577e-02 -6.63408399e-01 9.24809635e-01 5.38420439e-01
-5.43387294e-01 -2.57727057e-01 -1.26257455e+00 9.32801604e-01
-5.91414511e-01 -7.91684210e-01 3.36625636e-01 1.23639727e+00
-7.21842408e-01 4.26569074e-01 -7.53944039e-01 -2.68518150e-01
4.28128630e-01 4.99908864e-01 -5.26022613e-01 2.06062287e-01
-1.38105750e+00 1.22067785e+00 -7.31042176e-02 7.48947635e-02
-1.37983477e+00 -6.05931759e-01 -3.24389309e-01 -3.66056770e-01
3.63933027e-01 -3.31677049e-01 3.88357908e-01 -1.11570907e+00
-1.62581289e+00 7.14022696e-01 5.75953901e-01 -7.92216659e-01
8.18442702e-01 -1.30474031e-01 -8.87724400e-01 2.36939594e-01
-3.97228777e-01 5.02310753e-01 1.35494626e+00 -1.46363473e+00
3.36146052e-03 -2.06154063e-01 6.22062683e-01 -5.88096380e-02
-3.46701503e-01 1.51862904e-01 3.86850208e-01 -6.74259484e-01
-4.97200906e-01 -1.48257184e+00 -2.40458533e-01 1.93945959e-01
-5.90482295e-01 4.18531269e-01 1.37612176e+00 -2.26601630e-01
5.84544659e-01 -2.05768275e+00 4.41881269e-03 4.63621616e-01
-3.85160930e-02 7.89738476e-01 -1.33918598e-01 5.96687019e-01
-2.72465587e-01 -6.91139558e-03 -3.92692029e-01 1.51367396e-01
-4.21055481e-02 3.53601307e-01 -8.14456940e-01 8.85991275e-01
1.35464519e-01 7.15870261e-01 -6.95849299e-01 1.37981832e-01
4.26474959e-01 5.60508430e-01 -2.86886454e-01 1.37632608e-01
-6.25578314e-02 6.77130699e-01 -2.67742813e-01 5.58755577e-01
7.42550492e-01 8.57969463e-01 -4.00247276e-01 1.09653518e-01
7.75305182e-02 -2.45808840e-01 -1.16952872e+00 8.36702347e-01
-5.74499249e-01 8.88413429e-01 1.24975622e-01 -7.01902151e-01
8.81545901e-01 6.66871518e-02 9.41309184e-02 -7.30487704e-01
1.39776781e-01 8.96270871e-02 2.23585755e-01 -2.67993778e-01
2.23687552e-02 -2.92622328e-01 -3.12402725e-01 4.41458404e-01
-3.52637827e-01 -2.76012331e-01 -3.42809260e-01 7.07982630e-02
1.80917323e+00 -1.55078784e-01 -1.19660214e-01 -1.44958347e-01
5.31442940e-01 -2.01246813e-01 3.05468887e-01 8.70977044e-01
-2.64941841e-01 4.22898769e-01 3.92580271e-01 -3.86924505e-01
-7.45827556e-01 -1.48205340e+00 -2.26947814e-01 6.46122634e-01
3.95079076e-01 2.22750276e-01 -7.33881056e-01 -1.02933145e+00
1.40328079e-01 1.00034010e+00 -6.71356320e-01 -7.70833671e-01
-7.11023808e-01 -5.56879997e-01 1.23296666e+00 1.70877159e-01
5.63488662e-01 -1.34438586e+00 -1.01691437e+00 -9.48905274e-02
3.91043246e-01 -1.09725535e+00 2.31133159e-02 2.25619003e-01
-1.38122275e-01 -1.21890116e+00 -2.57880270e-01 -7.09340498e-02
8.10721278e-01 1.23348378e-01 8.49312901e-01 1.52436774e-02
-5.76790452e-01 5.95732331e-01 -3.39779556e-01 -7.15702176e-01
-8.57049346e-01 -6.06879115e-01 4.55794334e-01 3.20285618e-01
4.58733663e-02 -7.63143122e-01 -4.50377285e-01 7.69529998e-01
-1.08591843e+00 -6.38154387e-01 6.42902613e-01 5.17935395e-01
1.22542143e-01 3.50568533e-01 2.76678473e-01 -6.64875627e-01
4.23998356e-01 -2.57734299e-01 -7.46549487e-01 5.07689118e-02
-1.58299416e-01 -1.83608383e-01 1.01357818e+00 -9.30889726e-01
-7.17334151e-01 -4.49736342e-02 -1.50580257e-01 -4.97916132e-01
-5.05621612e-01 -3.50605786e-01 -4.53114033e-01 -9.98327255e-01
1.40356886e+00 7.60039687e-02 -2.18780518e-01 5.34979925e-02
4.87943828e-01 3.19012731e-01 5.20159245e-01 -3.73629510e-01
1.91100311e+00 9.63427901e-01 5.89567661e-01 -1.06993556e+00
-3.60687107e-01 1.54044062e-01 -4.17473316e-01 -5.40321290e-01
6.39460385e-01 -3.16445649e-01 -7.97182679e-01 7.04168916e-01
-8.53429198e-01 -4.32478726e-01 -5.45916140e-01 9.62421745e-02
-3.14397752e-01 3.42859715e-01 2.36012433e-02 -7.33030200e-01
1.42761664e-02 -1.00020576e+00 6.87895894e-01 1.38330847e-01
3.80659066e-02 -8.68828595e-01 2.24047750e-01 1.23838618e-01
6.80812359e-01 9.96630549e-01 4.45595264e-01 -6.70681000e-01
-6.90390825e-01 -7.01297164e-01 3.09681147e-01 5.83476543e-01
-1.33822769e-01 3.96628268e-02 -1.19953227e+00 -6.20776832e-01
9.20165181e-02 -3.02606285e-01 6.32245004e-01 -2.10097104e-01
7.76081085e-01 -5.11795640e-01 -3.43720973e-01 7.35592008e-01
1.27746999e+00 1.64907753e-01 1.03008342e+00 5.63577473e-01
6.36167824e-01 5.87797165e-01 3.56536061e-01 8.59450474e-02
-5.47621548e-01 9.58043754e-01 1.12461078e+00 -2.49212354e-01
-2.99650103e-01 -8.55794996e-02 7.86994755e-01 -3.47181618e-01
-1.07340872e-01 -4.98052418e-01 -7.58916080e-01 1.90858573e-01
-1.20059443e+00 -1.21121776e+00 -2.15575173e-01 2.51476765e+00
4.39690262e-01 7.32065797e-01 1.09927557e-01 2.51237422e-01
6.99862659e-01 3.33348960e-01 -4.81239855e-01 -5.99454105e-01
-3.11341703e-01 5.63362777e-01 1.06744611e+00 3.91447186e-01
-1.17032754e+00 7.83372641e-01 5.93686247e+00 5.96609354e-01
-1.51017559e+00 1.60542518e-01 -1.08443439e-01 -2.99491107e-01
-1.81455225e-01 2.35691831e-01 -5.49944580e-01 4.92550850e-01
9.67725933e-01 1.73623055e-01 4.56773371e-01 6.61450088e-01
-1.05777700e-02 1.35653690e-02 -7.54114032e-01 3.59489858e-01
2.82497346e-01 -9.60737586e-01 -8.54752511e-02 3.43399286e-01
5.16258299e-01 9.54133049e-02 3.87176156e-01 1.50301307e-01
4.76829618e-01 -1.07504249e+00 6.81307197e-01 3.71165514e-01
6.86458230e-01 -9.03482258e-01 5.04885316e-01 3.24757427e-01
-6.68517947e-01 -2.41488844e-01 -3.43773067e-01 6.63137585e-02
2.83503205e-01 3.62076372e-01 -8.24199498e-01 3.43120098e-01
5.11200070e-01 -4.84071188e-02 -7.80285478e-01 8.32331002e-01
-6.55904174e-01 8.20190310e-01 -7.99996912e-01 1.48755386e-01
3.74222517e-01 1.34558141e-01 1.37985551e+00 9.64241982e-01
-7.28515163e-02 -2.64521450e-01 -2.36421190e-02 7.30167270e-01
2.47141689e-01 -4.53459442e-01 -1.07969594e+00 4.23669249e-01
2.68464297e-01 1.12518430e+00 -5.99222302e-01 3.56694758e-01
8.27511996e-02 1.01084256e+00 -2.02666670e-01 3.52326900e-01
-1.33609319e+00 -5.89644313e-01 8.44964683e-01 3.17242712e-01
3.32823992e-01 -8.23847204e-02 2.52147373e-02 -8.18581939e-01
2.27561280e-01 -9.11955655e-01 2.37203017e-02 -5.94514966e-01
-1.17343926e+00 7.49322772e-01 8.80786031e-02 -1.34046483e+00
-1.52420793e-02 -6.95029140e-01 -1.09336782e+00 6.27959430e-01
-1.20128930e+00 -1.42647779e+00 -3.10161173e-01 8.61619473e-01
-3.10323620e-03 -3.50981385e-01 7.38037944e-01 5.87091446e-02
-3.35735500e-01 6.77310765e-01 -6.25102893e-02 1.52067337e-02
7.28506923e-01 -1.04091191e+00 4.30651039e-01 1.35946405e+00
3.44549298e-01 2.80585676e-01 1.09700799e+00 -6.81763172e-01
-1.34723818e+00 -1.12322068e+00 7.79849067e-02 -9.22906518e-01
8.48415196e-01 -7.98163772e-01 -6.74821734e-01 4.00988817e-01
9.53881070e-02 2.76852429e-01 2.40023091e-01 -2.18556404e-01
-5.69279313e-01 -3.06675136e-01 -1.52611005e+00 9.23485935e-01
8.86542499e-01 -4.18889433e-01 -4.71451461e-01 5.83400726e-01
5.44633269e-01 -1.87222674e-01 -3.98181617e-01 5.22930264e-01
4.28596973e-01 -1.15138495e+00 1.26289690e+00 -4.25720602e-01
-2.52519429e-01 -5.30704856e-01 1.67448670e-02 -1.29332662e+00
-6.44351095e-02 -1.11438465e+00 2.85214391e-02 1.13311803e+00
1.71823412e-01 -1.12802279e+00 5.11512995e-01 1.46434650e-01
-1.79205779e-02 -3.90039206e-01 -1.40821147e+00 -1.27912784e+00
1.27039999e-01 -5.64863682e-01 4.14670169e-01 7.45724201e-01
-4.80955362e-01 -1.61906049e-01 -4.42301273e-01 7.25406051e-01
7.72605002e-01 -5.25209725e-01 1.09515965e+00 -8.77576828e-01
-5.23418546e-01 -2.43794724e-01 -1.08067966e+00 -1.95556432e-01
2.40124434e-01 -6.43888950e-01 -1.42173618e-02 -6.17562115e-01
-5.75407863e-01 -6.20998681e-01 -1.37518495e-01 5.91462374e-01
5.49399629e-02 6.98528290e-01 2.33428329e-01 1.29050568e-01
-3.07527184e-02 1.68667167e-01 8.47249389e-01 -2.77736753e-01
5.91491386e-02 3.05700451e-01 -2.92437017e-01 7.05484271e-01
6.36618257e-01 -7.71567464e-01 -2.70902693e-01 8.02981332e-02
1.78096011e-01 -5.20293832e-01 1.03303564e+00 -1.54295635e+00
2.20659338e-02 -2.77660042e-01 1.39492825e-01 -4.16808203e-02
3.94898564e-01 -1.21931624e+00 3.70749801e-01 6.94337368e-01
1.22337505e-01 -2.34980732e-01 4.83039320e-01 6.86273932e-01
2.89643854e-01 -2.95118205e-02 1.19351578e+00 1.53462201e-01
-5.07125974e-01 5.64420596e-02 -3.60946029e-01 -9.30686221e-02
1.82546580e+00 -3.61012936e-01 -6.43961549e-01 -3.56077671e-01
-5.40115476e-01 -2.78957903e-01 7.91164160e-01 4.94555414e-01
3.38467032e-01 -1.02648497e+00 -6.15423739e-01 5.36750078e-01
-3.07896771e-02 -4.87824678e-01 1.97424203e-01 6.49266422e-01
-7.89318562e-01 8.46519619e-02 -4.37057674e-01 -5.00635087e-01
-1.33653915e+00 9.55939651e-01 5.31471491e-01 -7.83336349e-03
-6.85977340e-01 7.28933930e-01 1.90057218e-01 -2.51283556e-01
1.75912783e-01 3.69144112e-01 1.13289967e-01 -3.47182930e-01
3.67436886e-01 1.85604289e-01 1.77573144e-01 -7.36214519e-01
-5.58951735e-01 5.67350388e-01 2.45355651e-01 -5.91480024e-02
1.04901779e+00 4.66412157e-01 1.42157719e-01 -1.14689015e-01
8.87044251e-01 5.96484601e-01 -1.37716222e+00 2.61526525e-01
-4.59490001e-01 -6.63099408e-01 -4.93924171e-02 -9.36347663e-01
-9.67216194e-01 6.48975849e-01 8.87548089e-01 4.84696954e-01
1.09312570e+00 -1.22809745e-01 7.27433145e-01 4.06429529e-01
5.07400990e-01 -6.53224587e-01 1.81220308e-01 3.80754098e-02
9.65400338e-01 -6.55356467e-01 -2.64829285e-02 -2.23505810e-01
-6.23985767e-01 8.00166130e-01 4.79339063e-01 -5.67152381e-01
5.08356929e-01 5.15703797e-01 3.07691842e-01 -1.68506652e-01
-4.27347481e-01 -1.52098220e-02 1.44814521e-01 1.14856660e+00
-6.61839843e-01 4.26106267e-02 2.10385323e-01 -3.06164891e-01
-3.90733659e-01 -7.81007051e-01 7.50794411e-01 1.07675779e+00
-2.14237541e-01 -1.09955299e+00 -8.68687510e-01 -6.18242510e-02
-3.28678071e-01 2.28516176e-01 -8.06185126e-01 1.01345992e+00
3.91756356e-01 9.42019641e-01 -1.93224117e-01 -4.73182261e-01
6.51556134e-01 -8.83413926e-02 4.35111105e-01 -2.38517717e-01
-8.96084964e-01 -6.06606007e-01 3.22334543e-02 -8.32927525e-01
-1.84986144e-01 -6.83193684e-01 -7.24987149e-01 -3.37254971e-01
-4.96966302e-01 -2.77537167e-01 7.25231051e-01 8.39985669e-01
2.49777108e-01 6.11436725e-01 1.08956242e+00 -9.13873494e-01
-7.14560330e-01 -6.54710650e-01 -3.30985159e-01 5.40476322e-01
3.58421266e-01 -9.13311362e-01 -1.07626212e+00 -2.57490844e-01] | [5.469581127166748, 7.8476948738098145] |
b64ebed9-d89d-494a-ae6f-090f678b8e61 | the-nci-imaging-data-commons-as-a-platform | 2303.09354 | null | https://arxiv.org/abs/2303.09354v2 | https://arxiv.org/pdf/2303.09354v2.pdf | The NCI Imaging Data Commons as a platform for reproducible research in computational pathology | Background and Objectives: Reproducibility is a major challenge in developing machine learning (ML)-based solutions in computational pathology (CompPath). The NCI Imaging Data Commons (IDC) provides >120 cancer image collections according to the FAIR principles and is designed to be used with cloud ML services. Here, we explore its potential to facilitate reproducibility in CompPath research. Methods: Using the IDC, we implemented two experiments in which a representative ML-based method for classifying lung tumor tissue was trained and/or evaluated on different datasets. To assess reproducibility, the experiments were run multiple times with separate but identically configured instances of common ML services. Results: The AUC values of different runs of the same experiment were generally consistent. However, we observed small variations in AUC values of up to 0.045, indicating a practical limit to reproducibility. Conclusions: We conclude that the IDC facilitates approaching the reproducibility limit of CompPath research (i) by enabling researchers to reuse exactly the same datasets and (ii) by integrating with cloud ML services so that experiments can be run in identically configured computing environments. | ['André Homeyer', 'Andrey Fedorov', 'Ron Kikinis', 'Steve Pieper', 'William J. R. Longabaugh', 'William Clifford', 'Henning Höfener', 'David A. Clunie', 'Markus D. Herrmann', 'Daniela P. Schacherer'] | 2023-03-16 | null | null | null | null | ['whole-slide-images'] | ['computer-vision'] | [ 2.46586148e-02 -3.95532399e-01 -5.17586648e-01 -1.63817436e-01
-1.23684776e+00 -6.74931347e-01 3.54599267e-01 5.02027035e-01
-7.16946423e-01 6.65761828e-01 5.82955368e-02 -7.44695961e-01
-2.25272164e-01 -3.70411575e-01 -5.05805254e-01 -9.47934508e-01
-6.56401962e-02 6.31254256e-01 3.24521929e-01 4.21083152e-01
1.63102690e-02 8.80733848e-01 -1.07513952e+00 6.78233743e-01
5.82732737e-01 5.98086059e-01 4.15161729e-01 1.03684020e+00
-9.87886414e-02 7.95599341e-01 -4.66693670e-01 -4.88838963e-02
7.15782270e-02 -1.91035941e-01 -7.67840564e-01 -1.27605632e-01
3.05280775e-01 -1.20385863e-01 9.96370912e-02 7.09958076e-01
6.45507693e-01 -5.91161370e-01 3.35938811e-01 -1.41789973e+00
-7.02148154e-02 2.91835308e-01 -2.99211353e-01 4.43819642e-01
-6.17743805e-02 6.14348412e-01 6.51293397e-01 -4.47161555e-01
1.04437697e+00 4.48811144e-01 9.57556307e-01 4.70993400e-01
-1.36794901e+00 -5.31663358e-01 -3.51758510e-01 6.98753521e-02
-1.44200337e+00 -3.75141621e-01 -1.87052071e-01 -5.69883704e-01
8.30300093e-01 7.91679084e-01 6.33158088e-01 7.69018233e-01
8.13491821e-01 2.52037495e-01 1.28435659e+00 -2.70907760e-01
2.99746364e-01 4.66760427e-01 8.06931406e-02 5.49636841e-01
5.44073999e-01 -2.71467775e-01 -1.24643415e-01 -5.76583505e-01
3.44227254e-01 3.60253155e-02 -3.55603784e-01 2.65061725e-02
-1.27384508e+00 5.27629256e-01 3.72489035e-01 6.54213667e-01
-1.75039157e-01 8.69543627e-02 8.81143749e-01 2.38380149e-01
1.90988392e-01 4.46591467e-01 -2.26154044e-01 2.58270465e-02
-8.70281518e-01 3.30985226e-02 7.08383620e-01 7.86032677e-01
-1.20176300e-01 -6.99114561e-01 -1.24646813e-01 5.33213079e-01
1.56948209e-01 2.06777439e-01 9.00715172e-01 -1.09129763e+00
-1.78883016e-01 4.48406279e-01 -7.22119063e-02 -4.48131979e-01
-7.63140440e-01 -4.93362248e-01 -6.39537990e-01 1.15243107e-01
5.73010087e-01 1.92933127e-01 -6.66986346e-01 1.38557029e+00
1.08252898e-01 2.04033956e-01 -3.66118848e-02 5.78697681e-01
7.78880954e-01 -1.77729353e-01 6.45180106e-01 -2.14273870e-01
1.40204883e+00 -6.14417136e-01 -5.16510844e-01 5.41310489e-01
1.31771851e+00 -8.38559091e-01 1.13813198e+00 2.20777526e-01
-7.87184119e-01 2.17115562e-02 -7.83872545e-01 7.34841153e-02
-5.56074202e-01 6.57456368e-02 5.64089119e-01 7.89549828e-01
-1.35229301e+00 4.66060370e-01 -1.00648594e+00 -8.76969218e-01
4.89599437e-01 2.28148147e-01 -7.33853400e-01 -3.55883002e-01
-5.98322570e-01 9.19722736e-01 1.37021929e-01 -2.24294424e-01
-7.47198224e-01 -1.11574650e+00 -1.11794636e-01 -3.38445574e-01
-6.66192919e-02 -7.40905046e-01 9.89048600e-01 -5.08618653e-01
-8.82955611e-01 1.21297336e+00 -1.95145711e-01 -3.28933984e-01
6.20970726e-01 3.84117097e-01 -6.38052762e-01 3.14713776e-01
4.83267844e-01 3.28049392e-01 -2.82822460e-01 -9.47033823e-01
-8.44551086e-01 -5.45652211e-01 -5.08649111e-01 2.89367028e-02
-3.71826179e-02 2.68114835e-01 -5.99563062e-01 -9.90949422e-02
-4.52714302e-02 -1.07425594e+00 -4.50792313e-01 4.07833010e-01
-8.30461606e-02 -1.09423781e-02 6.88113809e-01 -6.33637905e-01
8.33963692e-01 -2.15485096e+00 -5.88691533e-01 2.32832178e-01
5.69649339e-01 8.65074173e-02 2.83893980e-02 1.63158759e-01
6.84502795e-02 6.84580326e-01 2.79217809e-01 -6.89439550e-02
-4.71724540e-01 2.05355287e-01 4.70195800e-01 7.97328174e-01
-9.64498147e-02 8.71563435e-01 -7.86958992e-01 -7.41767168e-01
-1.20438941e-01 1.63274348e-01 -3.13638836e-01 -3.99859771e-02
-1.53498473e-02 6.06866181e-01 -9.63107869e-02 7.52238393e-01
6.42408311e-01 -6.71120822e-01 8.14350486e-01 -1.00377537e-01
-2.95935899e-01 -1.37364164e-01 -6.04332268e-01 1.56709647e+00
-2.34213397e-01 5.44916451e-01 1.47255808e-01 -2.75588006e-01
3.79937947e-01 3.85859251e-01 7.63624966e-01 -4.01639342e-01
-1.28242642e-01 4.81892407e-01 3.71809244e-01 -5.44477940e-01
-6.55716658e-02 -2.90076643e-01 3.87318283e-01 4.56807196e-01
1.32920863e-02 -7.50860646e-02 1.56822190e-01 2.22781479e-01
1.64757502e+00 -4.91614193e-01 2.49815971e-01 -4.32131022e-01
3.71173918e-01 4.98967618e-01 4.54048991e-01 8.46573472e-01
-7.99739480e-01 4.45642173e-01 5.80542266e-01 -1.61441252e-01
-1.06891370e+00 -1.23127329e+00 -7.55992591e-01 7.76098073e-01
-4.14790958e-01 -3.31261784e-01 -2.25111112e-01 -7.63602257e-01
2.39569098e-01 5.68557918e-01 -7.84677684e-01 2.99586236e-01
-1.34544596e-01 -9.18243110e-01 9.75853145e-01 3.07000041e-01
8.70938525e-02 -3.88722450e-01 -4.86149400e-01 1.34783730e-01
-7.08399862e-02 -1.18465221e+00 -3.26559365e-01 2.28608117e-01
-7.71584690e-01 -1.42185616e+00 -6.73887908e-01 -3.80221456e-01
6.43945575e-01 3.81789535e-01 9.88723159e-01 4.37316954e-01
-8.52290332e-01 4.41355109e-01 -1.88268036e-01 -6.21336639e-01
-7.27583885e-01 -1.24933511e-01 -7.98759460e-02 -5.72218180e-01
4.47432637e-01 -2.22132817e-01 -5.59752524e-01 3.41471583e-01
-8.62097859e-01 -9.88168865e-02 9.66232657e-01 6.70561492e-01
9.77263272e-01 -6.31169602e-02 4.88117635e-01 -1.19902289e+00
3.84158820e-01 -6.64783001e-01 -4.28643107e-01 4.38291997e-01
-1.04742610e+00 -4.56688702e-01 5.08697689e-01 -4.49452512e-02
-6.80147350e-01 -1.55831560e-01 8.14641098e-05 -2.64072567e-01
-1.86561242e-01 5.72504580e-01 1.66682437e-01 -4.08064038e-01
6.97699547e-01 -8.39981623e-03 6.58491910e-01 -8.63593519e-02
-1.14013270e-01 7.10583031e-01 2.96124429e-01 -3.64921629e-01
2.21534342e-01 6.90477252e-01 3.29701930e-01 -6.07256651e-01
-4.82254297e-01 -9.57751632e-01 -4.49805796e-01 -2.87117660e-01
8.19346189e-01 -8.46506596e-01 -6.56013489e-01 3.33501577e-01
-4.65072662e-01 -5.04583240e-01 -1.47525683e-01 7.43276119e-01
-1.72499746e-01 5.13395220e-02 -8.70581031e-01 -1.73043758e-01
-6.15514040e-01 -1.42034817e+00 8.99646461e-01 1.53429314e-01
-2.99961001e-01 -1.02974057e+00 3.98439318e-01 2.40025267e-01
7.32822239e-01 5.53349078e-01 9.55282509e-01 -9.34166014e-01
-2.43373781e-01 -4.91599560e-01 -3.91372949e-01 -1.30424485e-01
3.45912308e-01 5.00874162e-01 -8.55975807e-01 -4.02620018e-01
-1.27495900e-01 -1.41911328e-01 3.92650247e-01 2.76993692e-01
1.33193147e+00 -1.04318671e-01 -8.26778531e-01 3.85072917e-01
1.73118901e+00 7.25789666e-02 7.10594654e-01 3.92542779e-01
1.48042113e-01 3.30960423e-01 2.38899559e-01 8.95447135e-02
2.18051225e-02 5.31572521e-01 -3.65915522e-02 -2.34881178e-01
-1.29175588e-01 -4.28698286e-02 -3.15535516e-01 4.06044483e-01
-6.01355545e-02 -6.57600164e-03 -1.21337891e+00 3.89112622e-01
-1.37783957e+00 -8.03352773e-01 -5.58390975e-01 2.27255559e+00
9.20762479e-01 2.21581101e-01 2.19502617e-02 -2.90820569e-01
5.09039283e-01 -4.77244109e-01 -4.18825388e-01 -2.68160999e-01
-1.74104404e-02 2.06872091e-01 6.98376477e-01 8.35693106e-02
-4.23277408e-01 3.62066448e-01 7.44137335e+00 6.23219252e-01
-1.39515686e+00 4.19261873e-01 8.87627900e-01 -2.91108429e-01
-1.26586422e-01 1.08948341e-02 -5.97064555e-01 4.20702636e-01
1.40008712e+00 -4.95315760e-01 -2.10142046e-01 6.75358534e-01
4.97827917e-01 -3.80335867e-01 -8.89832795e-01 4.44684148e-01
-3.00213426e-01 -1.85716307e+00 -2.25741163e-01 4.91593182e-01
4.14332747e-01 4.92785752e-01 -2.65361127e-02 1.14344761e-01
3.05760682e-01 -1.16272664e+00 -3.73822488e-02 7.68570364e-01
1.23006451e+00 -4.01917547e-01 1.16611397e+00 1.49350032e-01
-6.94930732e-01 3.53490621e-01 -3.48613054e-01 5.74718297e-01
-4.81509596e-01 7.98765004e-01 -1.74266219e+00 6.66388214e-01
7.69614637e-01 2.84259856e-01 -9.39865649e-01 1.15672696e+00
4.38847005e-01 7.22527385e-01 -6.65971115e-02 -4.09291871e-03
1.22456960e-02 2.46271476e-01 5.99635877e-02 1.46826160e+00
9.67011005e-02 -5.28191365e-02 -6.37831017e-02 5.67317903e-01
1.35208309e-01 2.94613332e-01 -2.71170944e-01 -2.20957458e-01
6.14273906e-01 1.68991446e+00 -8.42201233e-01 -3.38451177e-01
-5.90263307e-01 5.67531645e-01 2.75483370e-01 6.59487024e-02
-7.19477594e-01 -3.33499946e-02 4.13394362e-01 3.99556041e-01
-2.23592862e-01 1.11273862e-01 -3.45018089e-01 -5.80520689e-01
-3.98291409e-01 -8.57799530e-01 9.45039213e-01 -6.49033308e-01
-1.38815653e+00 2.61695504e-01 -4.38376069e-01 -1.20557690e+00
2.29532003e-01 -6.00160122e-01 -5.92015624e-01 1.17220283e+00
-1.26797962e+00 -1.09675002e+00 -5.59162974e-01 3.81725341e-01
-8.48035440e-02 1.50906006e-02 1.29921842e+00 4.49253097e-02
-5.97592294e-01 8.70738924e-01 6.04745269e-01 4.94353697e-02
1.08967185e+00 -1.24682868e+00 -3.89972210e-01 3.34936947e-01
-4.98935997e-01 8.05125654e-01 4.15672928e-01 -6.07053101e-01
-1.41780841e+00 -1.34445846e+00 7.41163194e-01 -5.02056837e-01
7.78406799e-01 8.51091817e-02 -8.04453909e-01 8.45948040e-01
-6.19887598e-02 2.20371842e-01 1.65497077e+00 5.94289228e-02
-2.01753125e-01 -2.48269290e-02 -1.68221188e+00 6.26077473e-01
4.51921016e-01 -3.60026687e-01 2.32974768e-01 7.16714084e-01
4.53378201e-01 -3.21220130e-01 -1.55141699e+00 2.77380615e-01
7.44479537e-01 -7.13904619e-01 4.33650166e-01 -8.68481100e-01
2.59622872e-01 -4.60440904e-01 -2.14810967e-01 -8.72979045e-01
-4.16458070e-01 5.38375601e-02 5.77883780e-01 8.70423615e-01
7.18183041e-01 -9.19446170e-01 8.20128977e-01 9.60794151e-01
-2.33380064e-01 -8.36143196e-01 -1.09103179e+00 -7.10859954e-01
3.22771937e-01 -2.84967393e-01 3.72443318e-01 9.95436907e-01
2.69427776e-01 -5.06658792e-01 5.48253775e-01 2.24329695e-01
6.23789012e-01 -2.70162553e-01 7.57594228e-01 -8.74224722e-01
-4.55181360e-01 -4.95048702e-01 -6.85701132e-01 2.16654927e-01
-2.99066901e-01 -1.23966396e+00 -3.10428619e-01 -1.24143112e+00
8.47971261e-01 -8.09910715e-01 -5.39762795e-01 4.70651597e-01
-2.17182860e-01 2.06810236e-01 -1.05726104e-02 7.87513733e-01
-4.30030257e-01 -5.18723071e-01 1.04801214e+00 1.37532905e-01
6.47469312e-02 -1.68585315e-01 -9.37629819e-01 1.62006468e-01
8.56693625e-01 -5.12765706e-01 -1.70303762e-01 6.06777780e-02
-2.75836736e-01 5.46132512e-02 4.84836161e-01 -1.24633205e+00
4.23051924e-01 -4.44568008e-01 8.39291990e-01 -3.79850566e-01
-3.12329143e-01 -6.58017874e-01 8.86755109e-01 9.71653581e-01
-3.14876109e-01 -7.26720989e-02 4.40869898e-01 4.38341200e-01
2.27832809e-01 -2.13311628e-01 9.17779267e-01 -3.21982145e-01
-3.89080733e-01 1.94765434e-01 -5.17034292e-01 -3.06624532e-01
1.29567635e+00 -2.88111605e-02 -9.00499523e-01 -2.75212079e-02
-7.82667100e-01 2.33884007e-01 8.43833923e-01 -1.92124560e-01
1.64643273e-01 -1.05504107e+00 -9.61100340e-01 -2.81623662e-01
5.31661391e-01 -3.39427739e-01 4.60052341e-01 1.48106658e+00
-8.20386171e-01 6.63076222e-01 -1.65273592e-01 -9.10485566e-01
-1.46069682e+00 6.03625238e-01 7.63574421e-01 -6.77360952e-01
-5.46051919e-01 6.35408700e-01 -2.26438910e-01 -5.11336863e-01
-9.20897163e-03 3.94637212e-02 1.21231787e-01 -4.00266409e-01
7.32823551e-01 1.42327711e-01 5.03060102e-01 -1.67274356e-01
-5.55111825e-01 -2.52393961e-01 -7.13695586e-01 -1.24418937e-01
9.96602058e-01 1.87825873e-01 -2.40560770e-01 6.36755943e-01
1.27787852e+00 1.69853002e-01 -5.27192771e-01 7.60708973e-02
-6.78514242e-02 -3.37448984e-01 -1.87030751e-02 -1.20245111e+00
-8.36456478e-01 9.61544961e-02 1.05102992e+00 -1.66218743e-01
9.90020037e-01 1.71427894e-02 2.28901312e-01 -5.49663790e-02
4.17839736e-01 -8.50607097e-01 -3.93661976e-01 -9.12335515e-02
6.20311677e-01 -9.86495256e-01 3.51145893e-01 -3.40045094e-01
-5.26811898e-01 1.14786458e+00 4.39128101e-01 3.09333235e-01
5.80336094e-01 6.33167744e-01 3.64185780e-01 -3.39774579e-01
-1.19246078e+00 2.22919047e-01 -2.22302645e-01 6.41153634e-01
8.66758466e-01 3.91431004e-01 -8.50850761e-01 4.60283726e-01
4.14329879e-02 5.98593831e-01 7.22141325e-01 1.22582221e+00
-2.16363877e-01 -1.07648122e+00 -3.65784377e-01 9.86027122e-01
-5.01659691e-01 1.64189599e-02 -3.63849372e-01 1.18346858e+00
-6.52662590e-02 7.89591134e-01 -5.32830767e-02 -1.32165074e-01
3.14576440e-02 9.71425772e-02 1.19643390e-01 -4.50227141e-01
-5.21118045e-01 -3.12434323e-02 2.15062704e-02 -6.61413252e-01
-1.75544932e-01 -8.21221530e-01 -1.31423283e+00 -4.36822474e-01
-2.60079473e-01 2.96473145e-01 9.02144670e-01 6.64322317e-01
5.68451226e-01 6.81883216e-01 6.89613342e-01 6.83803782e-02
-5.11069298e-01 -8.98663104e-01 -8.52716267e-01 2.63359845e-01
-3.08740914e-01 -2.55987018e-01 -5.75457811e-01 -1.27610089e-02] | [15.11402416229248, -2.9617979526519775] |
8a42a453-3d09-4ab8-95bb-1cdd37adcbd8 | ok-computer-analysis-an-audio-corpus-study-of | 2211.15834 | null | https://arxiv.org/abs/2211.15834v1 | https://arxiv.org/pdf/2211.15834v1.pdf | OK Computer Analysis: An Audio Corpus Study of Radiohead | The application of music information retrieval techniques in popular music studies has great promise. In the present work, a corpus of Radiohead songs across their career from 1992 to 2017 are subjected to automated audio analysis. We examine findings from a number of granularities and perspectives, including within song and between song examination of both timbral-rhythmic and harmonic features. Chronological changes include possible career spanning effects for a band's releases such as slowing tempi and reduced brightness, and the timbral markers of Radiohead's expanding approach to instrumental resources most identified with the Kid A and Amnesiac era. We conclude with a discussion highlighting some challenges for this approach, and the potential for a field of audio file based career analysis. | ['Nick Collins'] | 2022-11-29 | null | null | null | null | ['music-information-retrieval'] | ['music'] | [ 2.17478693e-01 -2.81432897e-01 -3.22842538e-01 1.81396022e-01
-1.27887499e+00 -8.44578207e-01 2.79464632e-01 3.66731703e-01
-3.36157709e-01 4.49216247e-01 8.66222560e-01 9.99196395e-02
-1.11316824e+00 -2.28447035e-01 -1.96096852e-01 -4.56486553e-01
-3.04630995e-01 7.09699169e-02 -8.27257410e-02 -2.99419940e-01
8.27112377e-01 1.93986773e-01 -1.75433350e+00 3.19601297e-01
4.36946116e-02 7.10786283e-01 -1.24388948e-01 7.85379648e-01
1.64514273e-01 6.88985050e-01 -1.04708946e+00 -3.55940670e-01
-1.32268786e-01 -4.43018347e-01 -7.12553918e-01 -3.82362574e-01
6.71445310e-01 -1.04583114e-01 -2.05079183e-01 4.92070049e-01
1.10756433e+00 2.58589208e-01 2.26791903e-01 -6.39708877e-01
-9.90779772e-02 1.26536691e+00 -6.44454420e-01 7.89341867e-01
9.26704764e-01 -2.68806726e-01 1.06079710e+00 -4.89797324e-01
8.24174941e-01 9.77059901e-01 1.16442466e+00 -7.54580200e-02
-1.14987075e+00 -1.17307472e+00 -4.82099682e-01 4.27521944e-01
-1.64540577e+00 -6.39346600e-01 1.04845321e+00 -8.28019202e-01
4.96426523e-01 4.25691992e-01 1.27273107e+00 6.86130941e-01
-1.39719054e-01 3.54218334e-01 9.32703972e-01 -5.10145962e-01
-1.78372934e-01 -2.88061589e-01 1.39419407e-01 -1.77551582e-01
-5.74700236e-02 2.76748347e-03 -1.52397346e+00 -4.67803657e-01
3.78008783e-01 -6.23886168e-01 -1.65149122e-01 4.21422154e-01
-1.17239308e+00 3.80182594e-01 -3.73905152e-01 6.23978496e-01
-6.78750128e-02 3.78596395e-01 8.06927919e-01 2.73323774e-01
3.55047137e-01 1.10588109e+00 -1.78111121e-01 -1.07332611e+00
-1.61315024e+00 5.99212885e-01 3.70065480e-01 4.01723385e-01
2.01715916e-01 1.71797827e-01 6.33601099e-02 1.10748684e+00
2.79261004e-02 2.55178846e-02 4.75144476e-01 -1.40237451e+00
5.16007356e-02 -1.31227016e-01 -4.45927769e-01 -1.05121195e+00
-5.50929606e-01 -6.49438858e-01 1.36199787e-01 -1.38784528e-01
4.09950316e-01 2.38200754e-01 3.00285872e-02 1.63408816e+00
1.87899128e-01 1.93996534e-01 -5.48223197e-01 6.13998592e-01
9.27838564e-01 2.08712861e-01 -1.30857229e-01 -6.73710346e-01
1.47144449e+00 -2.02459589e-01 -7.52616465e-01 3.28530461e-01
1.60550952e-01 -1.28134632e+00 1.03036785e+00 1.02171767e+00
-1.55162168e+00 -6.77233756e-01 -1.12524390e+00 2.79045943e-02
1.15933634e-01 -1.71426073e-01 5.20214736e-01 9.65952992e-01
-8.68643284e-01 9.58111823e-01 -5.02356708e-01 -2.73010910e-01
3.25268298e-01 2.83161849e-01 3.44365649e-02 4.90207165e-01
-9.65499282e-01 2.09442228e-01 1.39231235e-01 -3.33713651e-01
-4.90131497e-01 -1.14715970e+00 -2.76372015e-01 -2.25732401e-01
1.50296018e-01 -8.66066292e-03 1.44769490e+00 -5.56256056e-01
-1.24146771e+00 1.01970756e+00 3.57716918e-01 -2.50649750e-01
-4.10477594e-02 -3.79373610e-01 -7.94331253e-01 5.74921668e-01
2.55603075e-01 3.80038247e-02 5.96445620e-01 -6.37067795e-01
-6.86735511e-01 -3.17177117e-01 -2.21481636e-01 1.97505981e-01
-5.45818508e-01 8.39907765e-01 -3.15749139e-01 -1.04477453e+00
3.75046253e-01 -8.77648056e-01 5.77083647e-01 -7.26332068e-01
-1.61682591e-01 -9.97547805e-02 1.69700310e-01 -9.56457019e-01
2.10327768e+00 -2.39460588e+00 -1.08357795e-01 2.93896049e-01
1.23190746e-01 -5.67032337e-01 4.01152849e-01 7.66258419e-01
-2.43945837e-01 2.36267760e-01 4.21373099e-01 6.68965131e-02
-6.41318411e-03 -3.98417205e-01 -3.37148577e-01 7.37914741e-01
-7.14815080e-01 3.70248914e-01 -7.42767870e-01 -4.82256114e-01
-4.33359087e-01 2.93539494e-01 -5.38386881e-01 -4.61510122e-01
4.27597374e-01 4.83626068e-01 1.97208906e-03 1.01794600e+00
1.32062942e-01 6.42974615e-01 8.07996616e-02 -7.19808489e-02
-8.16089869e-01 9.91113186e-01 -1.02197170e+00 2.06976962e+00
1.64679717e-02 1.18211484e+00 1.91820860e-01 -4.49783117e-01
7.75067568e-01 6.02811098e-01 6.39046609e-01 -5.48761368e-01
-2.27559600e-02 3.83701742e-01 3.75484079e-01 -4.63262230e-01
1.14207673e+00 -4.28351283e-01 -2.91035801e-01 6.05986476e-01
9.93586779e-02 -2.69483387e-01 2.65350848e-01 1.18000992e-01
9.63063657e-01 2.23556563e-01 1.20711654e-01 -5.17258525e-01
4.52211536e-02 -2.73710564e-02 4.13233370e-01 4.88776237e-01
-1.56674087e-01 7.08530664e-01 3.65233690e-01 1.91361278e-01
-7.84871459e-01 -8.79755259e-01 -5.61640441e-01 1.58306539e+00
-4.30972844e-01 -1.14997768e+00 -3.07003200e-01 3.63636702e-01
-9.76319984e-02 5.06268322e-01 -5.95843852e-01 -1.57243222e-01
-5.76263249e-01 -3.08524579e-01 1.20801139e+00 8.60648304e-02
-1.81907669e-01 -1.06895626e+00 -5.35709918e-01 2.34381616e-01
-2.11949065e-01 -3.18426788e-01 -4.57336158e-01 5.90351969e-02
-8.34710956e-01 -8.19093227e-01 -6.30871296e-01 -5.95744669e-01
-4.34230596e-01 2.67203152e-03 9.92719054e-01 -1.94036081e-01
-5.54119825e-01 9.02838767e-01 -4.56952363e-01 -6.77920580e-01
-3.75916660e-01 2.14804962e-01 3.39034170e-01 -5.86925209e-01
2.48766653e-02 -9.66911912e-01 -4.56660002e-01 8.43778178e-02
-6.43170238e-01 -4.51600581e-01 -1.84217133e-02 4.61097866e-01
4.98510182e-01 2.63638943e-01 8.92841160e-01 -5.36629617e-01
8.22496891e-01 -5.84330976e-01 2.48661846e-01 -3.43419850e-01
-6.57244325e-01 -8.47228110e-01 -7.60789663e-02 -6.20092511e-01
-7.71601558e-01 -6.57366812e-01 -1.27424195e-04 -1.04646318e-01
6.60520047e-02 6.97637081e-01 1.81421503e-01 1.50130168e-01
9.43367362e-01 -1.48512334e-01 -2.89135277e-01 -8.17579806e-01
-3.17719541e-02 7.56108820e-01 1.25206733e+00 -1.00092947e+00
7.77461588e-01 3.34105045e-01 -1.44245043e-01 -1.16573954e+00
-6.81977332e-01 -6.79226458e-01 -3.95280838e-01 -9.57901537e-01
5.69334507e-01 -9.36425686e-01 -6.92836940e-01 2.59115219e-01
-4.29257631e-01 -1.90153807e-01 -7.72723079e-01 9.01785195e-01
-6.50819659e-01 2.93679208e-01 -6.84741557e-01 -1.00208139e+00
-2.15350658e-01 -4.03149724e-01 6.89564049e-01 4.52444881e-01
-1.10504758e+00 -5.70487320e-01 7.80674398e-01 6.79506063e-01
-7.77590498e-02 3.21658373e-01 6.54598534e-01 -4.14398283e-01
1.74154520e-01 -1.24969460e-01 6.41918957e-01 -1.42514259e-01
7.82710984e-02 1.43642560e-01 -1.08522034e+00 -3.37656438e-01
-2.62994952e-02 -4.19827014e-01 4.30608958e-01 5.16228497e-01
7.98946679e-01 5.44967018e-02 1.11318029e-01 4.66021180e-01
9.30607617e-01 3.51732016e-01 5.27726233e-01 8.85482490e-01
1.40413210e-01 7.38066673e-01 9.14966285e-01 7.94106543e-01
-1.09895878e-01 8.39940548e-01 -1.48951232e-01 5.40011346e-01
-5.24201214e-01 -3.84712368e-01 5.77909410e-01 1.30797219e+00
-3.98024589e-01 4.32941675e-01 -7.75868654e-01 9.30976272e-01
-1.12098169e+00 -1.33319604e+00 -1.80331618e-01 2.28671861e+00
9.31905925e-01 2.62027830e-01 7.70340383e-01 7.84226418e-01
7.44625270e-01 3.11634154e-04 -1.53824955e-01 -6.71598136e-01
-1.99687988e-01 6.87923133e-01 1.76533327e-01 -1.85317323e-01
-6.68574512e-01 4.47600901e-01 7.96493006e+00 1.20788896e+00
-6.31713331e-01 -1.75867639e-02 1.02640251e-02 -9.01150763e-01
-3.30273092e-01 2.92214632e-01 -4.00850475e-01 3.38957757e-01
1.11479557e+00 -4.05906975e-01 4.97620881e-01 3.52709413e-01
3.52372080e-01 -2.50404686e-01 -5.87185860e-01 9.50653076e-01
1.87163457e-01 -1.17190027e+00 -5.89250088e-01 6.67980835e-02
4.73263800e-01 -2.50992656e-01 1.90057263e-01 2.14626536e-01
-5.76529682e-01 -7.22591281e-01 1.41347349e+00 6.32130742e-01
1.19011354e+00 -1.03578258e+00 9.24627110e-03 -4.01907891e-01
-1.28385377e+00 -2.50844479e-01 9.05922204e-02 -5.26465416e-01
1.22001208e-01 3.56441498e-01 -6.72463357e-01 5.75170279e-01
1.00891423e+00 7.40874648e-01 -6.27554655e-01 1.38374507e+00
1.58223793e-01 1.33472359e+00 -3.44963729e-01 3.76771986e-01
-4.14582849e-01 -7.58201582e-03 1.05546641e+00 1.13858902e+00
7.09827065e-01 -1.02801897e-01 -3.81513834e-01 4.10890758e-01
3.44968110e-01 4.18998569e-01 -2.45962128e-01 -4.97758418e-01
9.24586892e-01 1.18005395e+00 -9.52553988e-01 4.01691124e-02
-1.20636947e-01 2.78991193e-01 -5.05907774e-01 -1.11794747e-01
-4.98964518e-01 -5.56452453e-01 5.41681528e-01 6.28960252e-01
-2.21679404e-01 -8.28465670e-02 -5.30772507e-01 -3.22924733e-01
-4.00609255e-01 -1.06368482e+00 6.69735372e-01 -8.08798552e-01
-8.14417720e-01 -6.88172281e-02 1.61306858e-01 -1.38772976e+00
-1.93732664e-01 1.53575867e-01 -6.64771855e-01 3.77501994e-01
-4.02359098e-01 -8.97103429e-01 3.26094359e-01 2.81682640e-01
3.95055175e-01 -3.99308532e-01 7.11580157e-01 7.38975465e-01
-1.87178418e-01 9.41704452e-01 9.52993408e-02 -4.88809347e-01
1.17801607e+00 -1.23022044e+00 -5.92103899e-02 1.62759289e-01
4.07046974e-01 7.24456728e-01 1.08804107e+00 -6.92323804e-01
-1.18330824e+00 -1.00008585e-01 6.70326293e-01 -3.12727660e-01
1.08587563e+00 1.01942323e-01 -3.87521952e-01 3.10011208e-01
1.63602188e-01 -1.26865745e+00 1.42489624e+00 7.40579486e-01
-7.53996149e-02 -5.55129945e-02 -6.38114691e-01 4.41860527e-01
1.05779016e+00 -9.78406608e-01 -6.68098271e-01 1.48174306e-02
2.81920344e-01 -3.15954417e-01 -1.37126124e+00 1.04020186e-01
1.43324661e+00 -8.72444391e-01 1.03221512e+00 -1.35202333e-01
3.42070431e-01 -1.98924780e-01 -2.96906605e-02 -7.87747443e-01
-4.63847250e-01 -1.35457265e+00 3.85056943e-01 1.92860508e+00
4.31314632e-02 1.32635340e-01 4.59654510e-01 -2.46298383e-03
-7.30161011e-01 -3.44598949e-01 -1.08742881e+00 -5.40277004e-01
3.27359401e-02 -1.00957680e+00 2.22875997e-01 1.04800606e+00
7.45396733e-01 2.57452250e-01 -4.06895757e-01 -5.19058347e-01
2.23034605e-01 -2.53985878e-02 7.15473592e-01 -1.43756044e+00
-7.93926239e-01 -8.75904620e-01 -6.33755624e-01 -2.14625552e-01
-3.05762738e-01 -9.97738659e-01 -3.07084024e-01 -9.91399884e-01
3.11552972e-01 -1.18248411e-01 -4.54667121e-01 1.31064117e-01
4.27926719e-01 8.10273886e-01 3.44041258e-01 4.22390491e-01
-3.91344249e-01 -2.39190131e-01 9.80893075e-01 -1.30552417e-02
-6.56347394e-01 2.22905219e-01 -1.10519409e+00 9.01466191e-01
7.00162947e-01 -6.15706384e-01 -4.32467043e-01 1.38907969e-01
1.04563022e+00 1.58353820e-01 -1.12836860e-04 -1.26488864e+00
4.32318263e-02 8.04975480e-02 1.01936556e-01 -7.93893158e-01
3.90986741e-01 -2.08848994e-02 8.23421478e-01 1.27892867e-01
-4.17982876e-01 6.48328438e-02 4.97970670e-01 1.78084746e-01
-1.91826537e-01 -2.73240328e-01 4.57062691e-01 2.54142553e-01
-1.14750244e-01 -3.82439375e-01 -8.01528335e-01 1.20902508e-01
4.58831578e-01 -5.26627541e-01 -8.19893554e-02 -5.50791860e-01
-1.06178784e+00 -2.06233412e-01 4.94054884e-01 3.32199275e-01
1.00752063e-01 -1.55870163e+00 -6.45310462e-01 -2.60203630e-01
1.55785769e-01 -7.43547499e-01 7.21981823e-01 1.30701268e+00
-4.09691900e-01 3.87398414e-02 -2.10747421e-01 -1.57384723e-01
-1.67957497e+00 2.89549865e-02 -1.79362655e-01 7.59039745e-02
-7.66049623e-01 9.43972826e-01 -1.31703407e-01 2.34208807e-01
4.09312397e-01 1.28542766e-01 -5.02126575e-01 8.99780095e-01
6.12602711e-01 1.10278320e+00 -7.20964521e-02 -7.79272914e-01
-3.78391474e-01 6.06451213e-01 3.05561602e-01 -7.65951872e-01
1.15721762e+00 -2.92990476e-01 -2.32132763e-01 1.33051968e+00
8.28518391e-01 9.00141180e-01 -5.32359779e-01 2.66466439e-01
1.22294202e-01 -3.28468829e-01 3.47235836e-02 -6.82303727e-01
-4.42567170e-01 4.88249034e-01 7.16298342e-01 4.58199233e-01
1.19620121e+00 1.87342152e-01 6.24464571e-01 -1.46695331e-01
-4.55086026e-03 -1.75769830e+00 1.07290849e-01 1.47334799e-01
8.93749297e-01 8.74502305e-03 6.29684031e-01 1.97854772e-01
-2.37796545e-01 1.17942405e+00 -2.27334127e-01 7.22260922e-02
6.79196775e-01 1.67321801e-01 -1.19247630e-01 -5.03286064e-01
-4.06821579e-01 -1.06510356e-01 4.03917104e-01 4.15281236e-01
9.41430151e-01 4.73608859e-02 -1.14301240e+00 1.28407717e+00
-1.34033823e+00 -2.23185435e-01 7.71229088e-01 7.63206840e-01
-4.11861867e-01 -1.12340963e+00 -8.51714671e-01 5.09850442e-01
-1.24499929e+00 -2.57797036e-02 -6.64885283e-01 7.28918910e-01
4.89158005e-01 1.01695466e+00 1.38771713e-01 -7.31708050e-01
1.12326145e-01 4.09030259e-01 6.92133784e-01 -4.65292126e-01
-1.00773036e+00 9.91346002e-01 4.71321464e-01 -2.43072152e-01
-6.78580046e-01 -1.59693778e+00 -1.13864565e+00 -2.84316599e-01
-4.30076510e-01 3.04532260e-01 8.11372101e-01 6.58406079e-01
-1.65044308e-01 9.85076725e-01 3.04824919e-01 -9.13024306e-01
3.72520648e-02 -1.18233609e+00 -1.43742287e+00 -8.26267600e-02
1.64833001e-03 -5.69708049e-01 -3.09322387e-01 1.04290642e-01] | [15.940689086914062, 5.278630256652832] |
dfa91f21-74ee-45d6-98f4-fad3d2280ec0 | pointwavelet-learning-in-spectral-domain-for | 2302.05201 | null | https://arxiv.org/abs/2302.05201v1 | https://arxiv.org/pdf/2302.05201v1.pdf | PointWavelet: Learning in Spectral Domain for 3D Point Cloud Analysis | With recent success of deep learning in 2D visual recognition, deep learning-based 3D point cloud analysis has received increasing attention from the community, especially due to the rapid development of autonomous driving technologies. However, most existing methods directly learn point features in the spatial domain, leaving the local structures in the spectral domain poorly investigated. In this paper, we introduce a new method, PointWavelet, to explore local graphs in the spectral domain via a learnable graph wavelet transform. Specifically, we first introduce the graph wavelet transform to form multi-scale spectral graph convolution to learn effective local structural representations. To avoid the time-consuming spectral decomposition, we then devise a learnable graph wavelet transform, which significantly accelerates the overall training process. Extensive experiments on four popular point cloud datasets, ModelNet40, ScanObjectNN, ShapeNet-Part, and S3DIS, demonstrate the effectiveness of the proposed method on point cloud classification and segmentation. | ['DaCheng Tao', 'Baosheng Yu', 'Jianzhi Long', 'Cheng Wen'] | 2023-02-10 | null | null | null | null | ['point-cloud-classification'] | ['computer-vision'] | [-2.11496800e-01 -3.42623681e-01 -8.18235651e-02 -2.74061292e-01
-4.68896002e-01 -3.72026771e-01 2.51119316e-01 2.06921518e-01
4.51932587e-02 6.55778348e-02 -4.00957346e-01 -4.15499628e-01
-1.62376329e-01 -1.10052407e+00 -6.36261642e-01 -5.91424525e-01
-3.02265048e-01 2.48648643e-01 2.58065701e-01 -1.82160903e-02
1.70609042e-01 1.02385521e+00 -1.29632092e+00 -8.78836066e-02
9.61057305e-01 1.05538201e+00 2.32023031e-01 1.77095041e-01
-5.05355418e-01 3.63360256e-01 -1.42642051e-01 4.39697597e-03
3.78867269e-01 3.57247479e-02 -4.46970999e-01 4.54416096e-01
4.02788341e-01 -1.06428407e-01 -6.03354752e-01 1.25557041e+00
3.25945914e-01 3.71835917e-01 4.55423802e-01 -1.23837852e+00
-7.12238252e-01 8.15594643e-02 -9.44803953e-01 1.81930199e-01
3.24483588e-02 2.21932575e-01 1.05686581e+00 -1.09472084e+00
4.35446978e-01 1.33138323e+00 8.60493302e-01 5.35737500e-02
-1.04653287e+00 -8.60000849e-01 1.72084212e-01 3.42639536e-01
-1.50064468e+00 1.92305624e-01 1.40487063e+00 -4.73478824e-01
1.13539815e+00 -1.66128322e-01 1.01946938e+00 5.40198803e-01
7.73514584e-02 5.43970823e-01 1.01648092e+00 -1.75970703e-01
1.71256676e-01 -5.88665724e-01 -4.08707932e-02 1.08336401e+00
1.10013023e-01 -3.00652962e-02 -3.01703006e-01 -1.97595134e-01
1.25536346e+00 4.04513657e-01 -2.93647200e-01 -6.26983404e-01
-8.58802915e-01 9.37102795e-01 9.63362634e-01 3.33785534e-01
-3.65418434e-01 5.05505621e-01 2.34687522e-01 2.57292092e-01
9.32670772e-01 -1.69550672e-01 -1.21454060e-01 4.31846222e-03
-7.56094098e-01 2.14800373e-01 3.51538301e-01 9.21930432e-01
1.27835882e+00 1.13050468e-01 4.43934172e-01 8.54736269e-01
5.73316157e-01 6.32111371e-01 -3.47532444e-02 -6.97523832e-01
4.10323620e-01 9.39443767e-01 -3.91870081e-01 -1.40034783e+00
-4.94349301e-01 -5.26423514e-01 -9.58282173e-01 3.59200656e-01
-3.41689661e-02 2.54225433e-01 -1.11034977e+00 1.10022473e+00
4.51826215e-01 8.63760650e-01 -3.49005908e-01 1.11530268e+00
8.56740236e-01 7.99738169e-01 -1.41854376e-01 2.45447263e-01
9.08157110e-01 -5.21125615e-01 -1.37470946e-01 -1.99098483e-01
3.56784523e-01 -4.84536052e-01 9.56140161e-01 9.34449676e-03
-6.96112216e-01 -6.43234015e-01 -1.17213547e+00 -6.40679896e-03
-1.33703560e-01 -2.29791850e-01 8.78108501e-01 1.83780685e-01
-8.28253329e-01 6.29776537e-01 -1.11914289e+00 -2.39242554e-01
7.89497316e-01 2.76910484e-01 -3.42298448e-01 -3.41319859e-01
-8.24972093e-01 4.24525470e-01 3.07962179e-01 3.49195898e-02
-5.12402296e-01 -7.72270739e-01 -8.72145951e-01 2.68403888e-01
1.07360281e-01 -6.66559041e-01 6.86604500e-01 -6.12993002e-01
-1.30127382e+00 1.06531191e+00 3.01304255e-02 -2.24572182e-01
1.69307724e-01 1.47668391e-01 -1.91013038e-01 4.21448827e-01
-2.09777635e-02 4.33963239e-01 1.15486038e+00 -1.33990932e+00
-4.24800932e-01 -6.18866146e-01 -4.55878489e-02 6.79035187e-02
-1.25536591e-01 -3.15495223e-01 -4.12100583e-01 -5.92031300e-01
7.25498855e-01 -8.26691389e-01 -2.16041639e-01 8.96132141e-02
-2.04098046e-01 -5.47795653e-01 9.96297300e-01 -4.29154366e-01
5.54142833e-01 -2.57952785e+00 1.18922152e-01 6.61226034e-01
6.47452712e-01 -2.73586642e-02 -1.75129831e-01 3.22420269e-01
-2.96484977e-01 8.41133390e-03 -3.95969540e-01 -3.18772912e-01
-1.11698598e-01 3.62118691e-01 -3.01717073e-01 7.25924373e-01
5.29045522e-01 1.00495851e+00 -7.90794313e-01 -3.60343546e-01
5.10065198e-01 5.17792284e-01 -5.67131162e-01 2.88260803e-02
-1.83462858e-01 3.10769439e-01 -8.07263374e-01 7.67443180e-01
1.09539163e+00 -4.16422427e-01 -4.16330069e-01 -1.01829678e-01
-1.29375264e-01 2.98231225e-02 -9.55287755e-01 1.97722840e+00
-3.40981364e-01 5.98608196e-01 1.45771012e-01 -1.45826566e+00
1.28112137e+00 -1.15286939e-01 9.30756092e-01 -5.32485425e-01
-4.66002710e-02 2.96273440e-01 -2.71397859e-01 -3.32141191e-01
-3.19320858e-02 -3.11327159e-01 2.92133838e-01 5.07764257e-02
2.35211030e-02 -5.89549065e-01 -4.29093271e-01 -1.03158839e-01
1.16872740e+00 -4.02674861e-02 -9.06095430e-02 -1.63051426e-01
4.48654085e-01 2.02351555e-01 3.32279503e-01 3.10014188e-01
-1.51898548e-01 6.06735408e-01 1.36615619e-01 -5.88217974e-01
-8.29840243e-01 -1.10363388e+00 -2.50264313e-02 5.15762687e-01
5.31201363e-01 -3.04870874e-01 -4.18396354e-01 -5.93004644e-01
5.19365549e-01 2.06486121e-01 -1.79706290e-01 -2.74826556e-01
-8.13506663e-01 -3.77253264e-01 2.98737735e-01 4.41431284e-01
7.09378958e-01 -8.22875500e-01 -4.20321524e-01 1.44491628e-01
1.52771890e-01 -1.21187866e+00 -3.59684408e-01 -8.89271721e-02
-1.36320710e+00 -9.31893408e-01 -6.18278205e-01 -9.96469736e-01
5.91840088e-01 8.28507185e-01 8.98598254e-01 4.33289438e-01
-3.73472363e-01 5.97917736e-01 -2.22980544e-01 -3.14503640e-01
1.20418191e-01 5.27183861e-02 -2.34548792e-01 4.67923209e-02
6.50306940e-01 -9.48341966e-01 -5.13439596e-01 -4.38284539e-02
-6.06274009e-01 6.55509308e-02 6.06055319e-01 7.59041369e-01
8.39595079e-01 2.97400266e-01 2.46333063e-01 -3.65679234e-01
5.46083689e-01 -4.65835959e-01 -8.05574894e-01 -1.71826184e-01
-2.93244839e-01 -4.09856997e-02 3.87500852e-01 -1.67182326e-01
-4.75419462e-01 1.23408295e-01 -2.61547118e-01 -1.16257238e+00
-1.38372317e-01 9.66392815e-01 4.91526127e-02 -7.86163330e-01
3.99515688e-01 3.97086084e-01 1.56996593e-01 -5.11092126e-01
3.16236764e-01 2.63449252e-01 4.21991646e-01 -5.81789374e-01
1.48698258e+00 6.60613596e-01 2.86452353e-01 -1.17895877e+00
-3.37977976e-01 -6.56802833e-01 -5.70984423e-01 -3.92150551e-01
9.30455565e-01 -1.04476213e+00 -8.22936833e-01 3.46344113e-01
-1.22119248e+00 -1.51804730e-01 -1.82061300e-01 5.24110973e-01
-5.55964470e-01 5.76718926e-01 -3.60080630e-01 -5.87226331e-01
-3.03921640e-01 -1.00640821e+00 1.51359451e+00 9.76938307e-02
3.49468410e-01 -1.08614886e+00 -2.09778249e-02 1.43416390e-01
1.23204045e-01 5.84446311e-01 1.26027775e+00 -1.98329419e-01
-1.20455945e+00 -2.34628022e-01 -5.59022129e-01 2.89650887e-01
1.23024136e-02 -9.85821411e-02 -6.59437656e-01 -3.53916615e-01
9.42803025e-02 -8.03137943e-02 8.79615486e-01 5.05688310e-01
1.40580618e+00 2.79408306e-01 -2.94664890e-01 1.08341825e+00
1.54493725e+00 -8.30815807e-02 2.61669308e-01 1.65452793e-01
1.05348504e+00 3.15515280e-01 2.68695742e-01 4.16277170e-01
3.43855441e-01 3.84537578e-01 7.74392724e-01 -3.80874664e-01
-7.04252645e-02 -3.86139780e-01 -3.69065106e-02 1.17342353e+00
-4.90174502e-01 3.18533659e-01 -1.25137925e+00 3.91179234e-01
-1.90089822e+00 -7.43405342e-01 -6.62504584e-02 1.74847317e+00
2.57695109e-01 2.27331564e-01 -7.01152757e-02 1.14676103e-01
5.93994439e-01 3.89768660e-01 -5.95480323e-01 1.59587041e-01
-1.55124692e-02 6.40697956e-01 5.86657166e-01 3.09210390e-01
-1.03459859e+00 1.02912843e+00 5.27148294e+00 8.38171661e-01
-1.41867769e+00 -1.47665134e-02 2.28018716e-01 4.04062152e-01
-5.26326120e-01 5.50860353e-02 -2.20966518e-01 1.88314185e-01
2.19873458e-01 -2.12346300e-01 5.19746780e-01 1.06888759e+00
1.78730965e-01 4.24797148e-01 -7.39238977e-01 1.58021867e+00
-2.20608279e-01 -1.53274643e+00 7.72678554e-02 2.32535630e-01
4.38274771e-01 4.75853086e-01 -5.67683466e-02 8.25730935e-02
1.39997378e-01 -9.78363514e-01 6.52603328e-01 3.47222865e-01
6.58057988e-01 -1.07021499e+00 2.73808837e-01 3.63851547e-01
-1.62754977e+00 2.36277003e-02 -7.59785295e-01 -2.91169733e-02
1.33458734e-03 8.17196488e-01 -6.95803046e-01 6.82905972e-01
8.73080492e-01 1.34447861e+00 -3.34995478e-01 1.22481656e+00
-1.48781061e-01 7.44400740e-01 -4.69921410e-01 4.02175859e-02
5.46078444e-01 -6.79361224e-01 7.71932244e-01 8.59436929e-01
5.15146196e-01 3.80301505e-01 6.25590682e-01 1.20077372e+00
-1.51673317e-01 2.69804057e-02 -8.04871619e-01 -1.97891548e-01
2.44271368e-01 1.10595667e+00 -8.65661442e-01 8.92057940e-02
-6.38131142e-01 6.39223814e-01 3.23561996e-01 5.84619582e-01
-5.95725119e-01 -3.91656518e-01 9.83677387e-01 1.11852787e-01
3.66327405e-01 -1.03577304e+00 -3.58736724e-01 -1.16838145e+00
2.56209802e-02 -4.94236380e-01 1.05588302e-01 -6.55732453e-01
-1.41271698e+00 2.20595911e-01 4.03017588e-02 -1.35587895e+00
3.01700711e-01 -6.88884974e-01 -7.32296288e-01 1.03793895e+00
-1.74173617e+00 -1.19901454e+00 -6.64741278e-01 8.07390332e-01
4.18731689e-01 -2.19223462e-02 2.87011057e-01 3.58943045e-01
-1.97859883e-01 1.48207203e-01 -5.40012047e-02 1.57882154e-01
1.40332177e-01 -1.03076470e+00 8.01327050e-01 5.94795167e-01
2.76313215e-01 4.10931975e-01 3.11296940e-01 -7.35308766e-01
-1.75446606e+00 -1.13713884e+00 3.14618140e-01 7.03416318e-02
6.94500983e-01 -5.52673876e-01 -1.37289095e+00 3.98639768e-01
-1.41893759e-01 4.02980834e-01 2.87985772e-01 -1.36671811e-01
-3.31640631e-01 -1.88481688e-01 -9.84403193e-01 2.61542678e-01
1.51102400e+00 -8.53751004e-01 -5.32382607e-01 5.43061316e-01
8.14335525e-01 -5.15498757e-01 -8.57266426e-01 6.38913274e-01
5.59435338e-02 -7.01627731e-01 1.24224269e+00 -4.37173426e-01
1.71636567e-01 -2.64076144e-01 -3.57666276e-02 -1.32148612e+00
-6.84620976e-01 -3.43980253e-01 -2.24134941e-02 6.72770083e-01
-5.10943942e-02 -8.81409466e-01 1.00342333e+00 -6.14495687e-02
-5.39155364e-01 -7.08040595e-01 -1.26371062e+00 -7.60050535e-01
-1.44041292e-02 -5.43380916e-01 8.16176832e-01 1.12935627e+00
-3.34252626e-01 3.61948274e-02 1.55370995e-01 4.86562133e-01
9.00525331e-01 6.01678073e-01 8.26053977e-01 -1.58790529e+00
-2.08647445e-01 -5.79548001e-01 -1.03335249e+00 -1.27682483e+00
4.33308393e-01 -1.17973506e+00 -9.24244821e-02 -1.53592920e+00
-1.62856936e-01 -6.31189466e-01 -2.12634087e-01 4.14497584e-01
-1.47798792e-01 5.41524030e-02 1.49006546e-01 4.00266439e-01
-1.76778689e-01 8.02298844e-01 1.38789105e+00 -3.86532128e-01
-1.00115649e-01 -6.08085021e-02 -8.58555138e-02 6.55435264e-01
4.78269339e-01 -3.74556452e-01 -5.04595518e-01 -6.41294956e-01
1.70300335e-01 -1.10343665e-01 5.74781835e-01 -1.18084240e+00
1.52774677e-01 -2.46342540e-01 3.02754253e-01 -8.00848901e-01
2.84628034e-01 -9.69785690e-01 1.34424135e-01 5.02457559e-01
3.14671069e-01 7.47073889e-02 3.80366027e-01 7.90017664e-01
-4.54950154e-01 1.33177519e-01 7.75861740e-01 -1.36663407e-01
-9.20437157e-01 1.01711559e+00 4.52469327e-02 -1.68870181e-01
9.14351881e-01 -3.91252011e-01 -1.54607985e-02 -1.87090829e-01
-5.10838389e-01 2.15870559e-01 5.97936153e-01 4.63450968e-01
1.07427001e+00 -1.57218254e+00 -6.35273337e-01 4.39960539e-01
2.13936731e-01 4.02427793e-01 2.63192236e-01 5.96733809e-01
-9.29257214e-01 3.42291631e-02 -1.34888574e-01 -1.15732074e+00
-9.72666502e-01 4.23103839e-01 4.10905182e-01 2.15622067e-01
-1.29306602e+00 8.66796672e-01 1.64551243e-01 -2.93906420e-01
-1.48022771e-01 -6.75958037e-01 9.27254632e-02 -3.31235558e-01
-8.57418701e-02 1.35087565e-01 1.53385699e-01 -6.54271960e-01
-4.29496020e-01 1.09832466e+00 3.57991964e-01 3.52828890e-01
1.62020862e+00 2.04667911e-01 -2.37898335e-01 3.01355422e-01
1.42904484e+00 -2.26520345e-01 -1.22869885e+00 -3.49377304e-01
1.49035156e-01 -7.16048002e-01 3.49079043e-01 1.96398512e-01
-1.35866320e+00 1.26437354e+00 6.33249938e-01 4.05148894e-01
1.09163332e+00 1.24435455e-01 1.02486587e+00 3.72711062e-01
5.03602624e-01 -6.41947746e-01 1.21696651e-01 5.20466208e-01
8.02907109e-01 -1.21259856e+00 5.57446852e-03 -7.15750992e-01
1.07091434e-01 1.28546560e+00 3.41876656e-01 -6.45398378e-01
9.28635657e-01 -1.86803281e-01 -1.33065253e-01 -7.14854777e-01
-2.11386159e-01 -2.88565338e-01 4.04897988e-01 5.43523908e-01
4.40802351e-02 3.41076143e-02 2.23398775e-01 6.93424419e-02
-4.83872816e-02 -1.63299099e-01 -1.00182250e-01 8.84140909e-01
-6.38628364e-01 -7.96600044e-01 -3.23784918e-01 4.42046791e-01
9.11945850e-02 1.18847504e-01 -3.92071158e-01 7.35686481e-01
-6.15395932e-03 5.34775734e-01 2.26186424e-01 -3.89426261e-01
5.48003674e-01 -4.09605950e-02 4.72029358e-01 -6.93659425e-01
-4.69132252e-02 8.95372480e-02 -4.47287589e-01 -6.54224157e-01
-4.49105024e-01 -4.98350382e-01 -1.62791014e+00 -3.10941935e-01
-3.10603082e-01 6.63997885e-03 7.84622014e-01 8.40714633e-01
4.92766380e-01 4.73280728e-01 8.23978782e-01 -1.17419899e+00
-3.63109648e-01 -5.42412698e-01 -6.76729381e-01 3.83203328e-01
4.36312973e-01 -9.30773020e-01 -3.34062815e-01 -3.48901093e-01] | [7.98133659362793, -3.633312940597534] |
a6db6bba-7b84-4563-a623-4385dee27d71 | m3er-multiplicative-multimodal-emotion | 1911.05659 | null | https://arxiv.org/abs/1911.05659v2 | https://arxiv.org/pdf/1911.05659v2.pdf | M3ER: Multiplicative Multimodal Emotion Recognition Using Facial, Textual, and Speech Cues | We present M3ER, a learning-based method for emotion recognition from multiple input modalities. Our approach combines cues from multiple co-occurring modalities (such as face, text, and speech) and also is more robust than other methods to sensor noise in any of the individual modalities. M3ER models a novel, data-driven multiplicative fusion method to combine the modalities, which learn to emphasize the more reliable cues and suppress others on a per-sample basis. By introducing a check step which uses Canonical Correlational Analysis to differentiate between ineffective and effective modalities, M3ER is robust to sensor noise. M3ER also generates proxy features in place of the ineffectual modalities. We demonstrate the efficiency of our network through experimentation on two benchmark datasets, IEMOCAP and CMU-MOSEI. We report a mean accuracy of 82.7% on IEMOCAP and 89.0% on CMU-MOSEI, which, collectively, is an improvement of about 5% over prior work. | ['Uttaran Bhattacharya', 'Dinesh Manocha', 'Rohan Chandra', 'Trisha Mittal', 'Aniket Bera'] | 2019-11-09 | null | null | null | null | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [ 1.64319724e-01 -1.91939697e-01 -1.37548856e-02 -4.11362827e-01
-1.33944941e+00 -3.55942219e-01 5.65466464e-01 -6.69027097e-04
-4.02149439e-01 5.62487602e-01 4.84609067e-01 6.01479113e-01
6.59328848e-02 -2.09715709e-01 -4.57463473e-01 -6.97397113e-01
-2.35569049e-02 -2.05445200e-01 -2.36268878e-01 -1.22855432e-01
-2.17832029e-01 2.15869412e-01 -1.71357846e+00 6.22055233e-01
4.94706392e-01 1.66974747e+00 -4.96772200e-01 6.81217432e-01
7.36131966e-02 8.47285986e-01 -4.17786092e-01 -3.14709306e-01
3.77276838e-02 -3.85072440e-01 -3.22714984e-01 4.48100418e-02
5.78886569e-01 -2.52932981e-02 -2.55583793e-01 8.14612508e-01
8.06436896e-01 2.53171504e-01 5.55588543e-01 -1.49062634e+00
-2.75859356e-01 6.23544157e-01 -6.68636680e-01 -2.30049580e-01
6.22384548e-01 -1.17307104e-01 1.02223468e+00 -1.47900903e+00
5.00459313e-01 1.46515608e+00 7.35651553e-01 6.88666403e-01
-1.32345974e+00 -7.37275958e-01 1.99465737e-01 7.68001750e-02
-1.49163401e+00 -1.05271971e+00 8.01300228e-01 -2.25130677e-01
8.95245552e-01 3.57677698e-01 2.22573392e-02 1.49701643e+00
-4.16486785e-02 9.84526992e-01 1.48705101e+00 -3.37804377e-01
3.96356463e-01 9.88462120e-02 2.04676762e-01 6.24176621e-01
-2.25347787e-01 8.02589674e-03 -1.41606081e+00 -5.33443689e-01
2.00521395e-01 5.10724857e-02 -1.69409618e-01 5.04018851e-02
-1.00357294e+00 3.64530087e-01 1.74681902e-01 1.67482957e-01
-4.59129095e-01 1.16351783e-01 2.94678599e-01 3.92398298e-01
3.57682765e-01 1.75876003e-02 -6.60067677e-01 -3.79485011e-01
-7.86125958e-01 -1.86682314e-01 8.19631219e-01 6.63049579e-01
6.20930374e-01 1.04370423e-01 -1.76974729e-01 1.23722625e+00
3.45077902e-01 9.30871248e-01 1.43768534e-01 -1.17638588e+00
3.95718813e-01 5.20792186e-01 6.32248968e-02 -8.76978576e-01
-4.68676239e-01 4.50374326e-03 -1.08347392e+00 1.47874907e-01
9.26095769e-02 -6.38920724e-01 -9.29337204e-01 2.16682482e+00
1.85858775e-02 2.36595601e-01 2.72087961e-01 9.49934542e-01
1.18654978e+00 4.71727550e-01 1.63492933e-01 -2.94957340e-01
1.17679954e+00 -7.52089679e-01 -9.56763625e-01 -1.34501144e-01
1.24213926e-01 -6.28248870e-01 7.05031574e-01 7.89040089e-01
-1.04797113e+00 -3.64288479e-01 -9.40794706e-01 7.07610697e-02
-4.75756109e-01 2.46721536e-01 5.18934608e-01 5.77988923e-01
-1.00389111e+00 4.50798482e-01 -7.66409457e-01 -2.94865876e-01
3.26307863e-01 3.14164639e-01 -8.49873245e-01 -4.81394120e-02
-1.01317894e+00 7.16904283e-01 7.75348172e-02 4.99644177e-03
-5.33891201e-01 -3.86689454e-01 -9.31618094e-01 9.49567333e-02
2.75895715e-01 -3.13471705e-01 8.64333332e-01 -1.12673092e+00
-1.64798367e+00 4.87315863e-01 -5.17587483e-01 -4.36324952e-03
1.15228802e-01 -4.20166701e-01 -9.48616803e-01 3.66798222e-01
-4.66463655e-01 8.13725173e-01 9.95646477e-01 -1.39040196e+00
-5.93613029e-01 -3.16691160e-01 -4.13665235e-01 1.86400250e-01
-5.95991373e-01 2.22873777e-01 -6.91778362e-01 -3.81495535e-01
2.94262439e-01 -9.21668470e-01 2.89251730e-02 7.38723129e-02
-2.54076034e-01 -4.40070257e-02 7.38630414e-01 -4.51496273e-01
1.09078491e+00 -2.30491090e+00 2.65882790e-01 5.21759629e-01
3.79799426e-01 1.77095446e-03 -4.75731015e-01 3.65808606e-01
-1.61043763e-01 8.34081620e-02 -1.41683847e-01 -8.68437827e-01
1.81665912e-01 3.60105969e-02 1.59222004e-03 2.95337766e-01
3.01576555e-01 7.77913690e-01 -6.50415540e-01 -3.53540719e-01
2.68082500e-01 7.84655571e-01 -2.70258218e-01 2.23727793e-01
2.24822357e-01 1.63111821e-01 7.16848671e-02 1.23379147e+00
7.09447622e-01 1.05697317e-02 5.81903569e-02 -4.14317399e-01
2.14031518e-01 -6.36766329e-02 -1.35453224e+00 1.68701351e+00
-4.02261555e-01 6.35966301e-01 4.73036796e-01 -4.57801133e-01
9.13199782e-01 5.62360942e-01 7.22832799e-01 -7.54281044e-01
3.76735240e-01 5.37817478e-02 -3.30378503e-01 -4.77850676e-01
4.16898847e-01 1.02918550e-01 -3.82373214e-01 3.05897683e-01
5.85346878e-01 2.29873285e-01 6.50690570e-02 2.53847957e-01
1.25036740e+00 -2.18563214e-01 1.78424880e-01 1.48031667e-01
3.18380177e-01 -7.72340834e-01 7.60249138e-01 9.90036011e-01
-3.86712313e-01 7.44990945e-01 3.72836620e-01 1.89446732e-01
-3.34873825e-01 -1.23096228e+00 5.78240491e-02 1.26347995e+00
-3.50565650e-03 -7.71994948e-01 -4.52573001e-01 -5.64306319e-01
7.06254318e-02 3.44012469e-01 -6.84945524e-01 -2.70170510e-01
1.10370450e-01 -9.36627388e-01 6.25522316e-01 6.77679479e-01
3.22388917e-01 -7.72658408e-01 -1.94226801e-01 7.79136494e-02
-4.81228858e-01 -1.26711929e+00 -1.59727931e-01 3.78272116e-01
-5.97429752e-01 -9.49454367e-01 -3.72582704e-01 -1.60370514e-01
5.02098441e-01 9.17133838e-02 1.03520751e+00 -2.04209492e-01
3.65702435e-02 8.14359605e-01 -5.26989341e-01 -4.56810147e-01
1.57850515e-02 -2.46060431e-01 4.17869806e-01 4.37246889e-01
6.50729001e-01 -6.42805159e-01 -2.64338762e-01 2.35890895e-01
-7.96875775e-01 -2.15380207e-01 5.21809757e-01 9.53492999e-01
5.20260334e-01 -2.26972222e-01 6.05538845e-01 -5.26109636e-01
5.33601940e-01 -6.66757941e-01 -1.13049902e-01 2.05127254e-01
-3.87410879e-01 -1.06477328e-01 2.41006836e-01 -6.22945845e-01
-1.03588068e+00 3.03296953e-01 -1.99601233e-01 -5.94072998e-01
-3.06271613e-01 6.18954659e-01 -2.84381002e-01 -9.75714549e-02
6.69138432e-01 -1.36320338e-01 -2.88246989e-01 -4.76127476e-01
3.90075684e-01 9.48658824e-01 7.70716786e-01 -6.66965723e-01
5.39677799e-01 5.92971444e-01 -1.53580695e-01 -8.93593669e-01
-6.89933717e-01 -5.50621688e-01 -3.19282293e-01 -4.85366851e-01
4.46011275e-01 -1.13825810e+00 -9.42787468e-01 7.26872623e-01
-8.96607935e-01 4.19128090e-02 -1.14544906e-01 5.13171136e-01
-1.87586173e-01 9.30886716e-02 -8.26371312e-01 -1.30809450e+00
-4.04843122e-01 -8.24251592e-01 1.30379713e+00 4.70571995e-01
-3.31909567e-01 -7.42832124e-01 -1.26166835e-01 2.45201826e-01
3.81431729e-01 3.60607862e-01 1.96471751e-01 -6.37199879e-01
-4.80304472e-02 -3.55616540e-01 -2.67649770e-01 5.36233127e-01
1.97000802e-01 1.59554064e-01 -1.69449067e+00 -1.91537142e-01
-2.05272973e-01 -7.96049297e-01 1.21905017e+00 4.75853644e-02
1.02853692e+00 1.90742370e-02 -3.73728834e-02 3.56127560e-01
1.25496423e+00 4.04488258e-02 6.00231826e-01 -1.77237943e-01
5.30781507e-01 4.13614839e-01 3.66556704e-01 6.91029251e-01
4.02445257e-01 3.90548229e-01 3.30713868e-01 -2.79752016e-01
1.38802215e-01 -4.31798920e-02 7.72457719e-01 1.06194711e+00
9.54599753e-02 -3.66784573e-01 -5.47931373e-01 4.60347801e-01
-1.95972717e+00 -8.44788432e-01 7.87019357e-02 1.86495018e+00
6.76585317e-01 -1.01106286e-01 1.02370255e-01 2.99302012e-01
3.92865390e-01 1.14435174e-01 -5.68560362e-01 -2.90671647e-01
-7.37537026e-01 2.68862993e-01 2.28033334e-01 2.87771642e-01
-1.24928164e+00 6.08492076e-01 7.02886581e+00 6.61568463e-01
-1.12449741e+00 1.77920178e-01 6.45951807e-01 -4.81367528e-01
-4.91268225e-02 -3.79244238e-01 -4.73080426e-01 4.36013609e-01
1.09965849e+00 1.87449664e-01 5.46236098e-01 6.75469100e-01
1.08104326e-01 -5.26576281e-01 -1.21422660e+00 1.59409320e+00
4.46775913e-01 -8.11313570e-01 -5.30201554e-01 -4.27452505e-01
7.41904616e-01 3.21059287e-01 6.36271387e-02 3.68122816e-01
2.31395602e-01 -1.00026584e+00 6.62291408e-01 8.37704241e-01
6.87338412e-01 -6.81869328e-01 8.54697108e-01 1.18797369e-01
-1.12497580e+00 -1.62767842e-01 6.88514188e-02 3.44694126e-03
-3.97596657e-02 8.17938089e-01 2.81198863e-02 5.85823238e-01
9.02794838e-01 5.71481943e-01 -4.62789416e-01 5.93159139e-01
-1.07565401e-02 6.59999013e-01 -7.63594806e-01 1.83628127e-01
-2.58513838e-01 1.74438596e-01 4.63200897e-01 1.52658236e+00
3.72693837e-01 1.28361940e-01 1.56004876e-01 4.31682289e-01
-4.62453008e-01 1.99612956e-02 -3.61025900e-01 -1.04487307e-01
5.90231240e-01 1.66474545e+00 -1.75870955e-01 -4.04687434e-01
-5.34853458e-01 1.03991818e+00 2.30189472e-01 4.95401055e-01
-5.85123420e-01 -2.19476894e-01 8.51414919e-01 -7.38185227e-01
1.59600839e-01 -7.86408335e-02 -2.66417712e-01 -1.49307072e+00
1.15697689e-01 -1.03053749e+00 5.94101012e-01 -9.83777821e-01
-1.68798757e+00 5.79358220e-01 -3.24193925e-01 -1.14638186e+00
-2.14292824e-01 -7.28625953e-01 -3.14624935e-01 7.97564030e-01
-1.37101901e+00 -8.88720810e-01 -4.48402017e-01 8.08808565e-01
5.48255965e-02 -9.29189697e-02 1.13176477e+00 5.32771349e-01
-6.58021450e-01 8.87403429e-01 2.86276825e-02 1.34602502e-01
1.15435147e+00 -1.15134203e+00 -2.48746425e-01 7.43734181e-01
1.52883172e-01 4.77105349e-01 5.43847859e-01 -3.76032352e-01
-1.66959786e+00 -8.31252217e-01 7.93683231e-01 -5.45162737e-01
5.80515325e-01 -6.27439857e-01 -7.85745382e-01 3.47955674e-01
2.85623670e-01 1.70493424e-01 1.01629245e+00 5.00115514e-01
-8.21240008e-01 -3.84887695e-01 -1.28319919e+00 5.22383392e-01
9.08557236e-01 -7.94487476e-01 -3.91662568e-01 -9.47421640e-02
2.75368184e-01 -4.88505453e-01 -1.08085930e+00 5.71354747e-01
8.02667975e-01 -1.04621339e+00 6.91753626e-01 -3.69130373e-01
3.85885835e-01 -1.89523563e-01 -7.27552891e-01 -1.44491041e+00
1.78012680e-02 -7.00604618e-01 -6.35575116e-01 1.54118073e+00
3.51499110e-01 -4.88332182e-01 4.55882788e-01 8.97737980e-01
3.61913502e-01 -5.76373577e-01 -9.58986938e-01 -5.61214387e-01
-4.15424883e-01 -9.84914124e-01 3.96813929e-01 9.54168439e-01
3.27342391e-01 3.11908305e-01 -8.88323307e-01 9.02323425e-02
5.97739220e-01 -2.27065101e-01 6.57783329e-01 -1.21320415e+00
-9.64299291e-02 -2.65609831e-01 -2.91757017e-01 -7.07350194e-01
7.37258717e-02 -6.32424474e-01 1.93976790e-01 -9.98325109e-01
3.07157665e-01 -7.75486976e-02 -1.01925814e+00 9.23953414e-01
-2.12038785e-01 5.70957303e-01 3.94797117e-01 -1.25532195e-01
-9.93357360e-01 6.54383957e-01 5.69485903e-01 -7.28406906e-02
-2.72063971e-01 -5.94737530e-01 -9.58034039e-01 8.34238589e-01
6.61535323e-01 -3.71061325e-01 -4.81151119e-02 -1.46090776e-01
1.54305771e-01 6.20374121e-02 2.65686810e-01 -1.11592937e+00
2.29825467e-01 -1.34960815e-01 7.50240088e-01 -5.59963763e-01
8.70121062e-01 -1.01706302e+00 1.02373756e-01 -2.44237453e-01
-4.00442690e-01 -2.45147720e-01 3.62656027e-01 3.75227422e-01
-3.27222645e-01 3.37840706e-01 7.39845634e-01 3.34289640e-01
-7.91238844e-01 1.91184774e-01 -4.90871817e-01 -1.19553380e-01
3.89671326e-01 1.67280912e-01 -4.34524089e-01 -7.27592647e-01
-9.09799457e-01 3.78165334e-01 8.22102800e-02 6.71408951e-01
7.52942979e-01 -1.61086905e+00 -5.72126806e-01 1.77328482e-01
3.91019821e-01 -6.69424593e-01 3.20842177e-01 1.02316821e+00
5.05767584e-01 -6.59808517e-02 -1.00492172e-01 -6.25885367e-01
-1.52719772e+00 7.89443105e-02 2.61718571e-01 8.48788768e-02
-2.90350970e-02 7.46325791e-01 -3.86359751e-01 -4.58260268e-01
5.12000561e-01 -1.60559431e-01 -4.88279834e-02 3.58060747e-01
7.71138430e-01 4.06350344e-01 1.16284348e-01 -8.54631901e-01
-6.54782832e-01 3.27576846e-01 2.38815799e-01 -4.50514853e-01
1.26319838e+00 -4.36821967e-01 -8.07604566e-02 8.78111124e-01
1.19613957e+00 1.31399766e-01 -1.03852594e+00 -6.28899276e-01
2.88006812e-02 -2.81761050e-01 2.00744495e-01 -1.33027864e+00
-1.14342868e+00 6.78682506e-01 9.17569935e-01 2.87009832e-02
1.67113328e+00 -4.52947617e-02 6.55021966e-01 5.41898012e-01
1.49688214e-01 -1.55381179e+00 1.42833829e-01 6.30711436e-01
8.55982602e-01 -1.67322969e+00 -1.51257068e-01 -3.42659146e-01
-6.58094049e-01 9.68775928e-01 5.30716479e-01 1.98488355e-01
7.74065793e-01 7.51445234e-01 4.16579157e-01 -1.13985986e-02
-1.11089504e+00 -4.29812342e-01 5.10135770e-01 4.53885615e-01
4.78710473e-01 1.08981691e-01 7.38937408e-02 1.02184069e+00
3.39110196e-01 1.95383608e-01 9.89610106e-02 1.04568851e+00
-1.92824721e-01 -9.45090473e-01 -7.33592153e-01 5.77902496e-01
-5.65432966e-01 -5.35893403e-02 -7.38486052e-01 5.02518177e-01
2.26770923e-01 1.54183805e+00 -4.61034924e-02 -8.83083582e-01
2.64265537e-01 4.06815588e-01 2.75355905e-01 3.91973928e-03
-5.23447871e-01 4.87543732e-01 3.91615391e-01 -9.40077901e-01
-9.00483668e-01 -7.21879125e-01 -1.06232178e+00 -2.94996560e-01
-1.62975192e-01 -1.30367860e-01 7.69600093e-01 9.71212506e-01
6.11511171e-01 4.43710804e-01 7.43162990e-01 -1.03343785e+00
-5.62094375e-02 -9.47842896e-01 -6.22330010e-01 7.02432215e-01
4.05197382e-01 -7.72033155e-01 -5.35371065e-01 -4.62384336e-02] | [13.238875389099121, 5.100305557250977] |
1cc22ab3-ae8e-409a-ae88-d1b9275afc63 | road-barlow-twins-redundancy-reduction-for | 2306.1084 | null | https://arxiv.org/abs/2306.10840v1 | https://arxiv.org/pdf/2306.10840v1.pdf | Road Barlow Twins: Redundancy Reduction for Road Environment Descriptors and Motion Prediction | Anticipating the future motion of traffic agents is vital for self-driving vehicles to ensure their safe operation. We introduce a novel self-supervised pre-training method as well as a transformer model for motion prediction. Our method is based on Barlow Twins and applies the redundancy reduction principle to embeddings generated from HD maps. Additionally, we introduce a novel approach for redundancy reduction, where a potentially large and variable set of road environment tokens is transformed into a fixed-size set of road environment descriptors (RED). Our experiments reveal that the proposed pre-training method can improve minADE and minFDE by 12% and 15% and outperform contrastive learning with PreTraM and SimCLR in a semi-supervised setting. Our REDMotion model achieves results that are competitive with those of recent related methods such as MultiPath++ or Scene Transformer. Code is available at: https://github.com/kit-mrt/road-barlow-twins | ['Carlos Fernandez Lopez', 'Marvin Klemp', 'Omer Sahin Tas', 'Royden Wagner'] | 2023-06-19 | null | null | null | null | ['motion-prediction', 'trajectory-prediction', 'trajectory-forecasting'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-1.98660329e-01 -6.80161417e-02 -2.86486566e-01 -4.74967718e-01
-7.06664681e-01 -2.62201339e-01 9.55547512e-01 -3.54480505e-01
-5.18813252e-01 3.37785840e-01 3.34172994e-01 -3.27950180e-01
3.96637395e-02 -9.38776791e-01 -8.01593602e-01 -6.25635147e-01
-2.18257576e-01 5.87928712e-01 7.96467423e-01 -4.13786381e-01
1.47030860e-01 5.04334271e-01 -1.88971770e+00 -7.40465969e-02
7.58903444e-01 7.34489620e-01 7.00388432e-01 9.68162417e-01
1.43906891e-01 1.11741972e+00 1.33501843e-01 -2.75977820e-01
3.76467645e-01 -1.18118394e-02 -7.13932097e-01 -1.79919988e-01
4.35565889e-01 -5.62166035e-01 -1.13222802e+00 4.67249453e-01
4.04798090e-01 5.81681788e-01 5.74230492e-01 -1.56832814e+00
-2.64437169e-01 1.71457052e-01 -3.38468909e-01 3.81522387e-01
1.22476213e-01 4.00856316e-01 8.45521510e-01 -1.03346944e+00
7.71904051e-01 1.14981115e+00 7.13589072e-01 5.66363096e-01
-1.02818406e+00 -5.60950518e-01 9.24222693e-02 9.61104631e-01
-1.56552386e+00 -8.89360130e-01 8.34285855e-01 -4.01528537e-01
1.18870890e+00 1.77288070e-01 5.85237145e-01 1.08750308e+00
2.36412302e-01 8.48148763e-01 5.63217998e-01 -6.33811532e-03
2.30287656e-01 4.16911021e-02 -5.72632439e-02 8.60979438e-01
-9.65727568e-02 3.83290708e-01 -5.24744391e-01 2.41874039e-01
3.67546201e-01 -1.80387035e-01 2.41594881e-01 -6.17593586e-01
-1.18895042e+00 8.18929434e-01 3.84037375e-01 -1.39314741e-01
-2.52939552e-01 5.01609623e-01 2.58600026e-01 4.25242297e-02
5.09922266e-01 -5.91484047e-02 -2.37966970e-01 -5.83779216e-01
-7.45343983e-01 4.80110854e-01 3.94806653e-01 1.12682831e+00
1.11492205e+00 8.26681331e-02 3.31479460e-02 5.87350070e-01
3.41889739e-01 7.17441320e-01 4.21918213e-01 -1.36423147e+00
6.53281271e-01 1.26105934e-01 1.32045358e-01 -9.25723314e-01
-4.73175585e-01 -3.52466553e-02 -3.53053063e-01 2.73984492e-01
1.47343084e-01 -6.99359179e-02 -7.60852635e-01 1.76538026e+00
6.05980515e-01 7.83094049e-01 6.26005605e-02 7.53480256e-01
7.55399823e-01 7.73310423e-01 2.70513119e-04 3.64216685e-01
9.40069735e-01 -1.53416193e+00 -4.57712293e-01 -3.19653034e-01
1.03106785e+00 -4.83461559e-01 9.57358241e-01 -2.32912088e-03
-8.63055229e-01 -7.67932653e-01 -8.33255231e-01 -3.11334282e-01
-5.14193773e-01 8.13848451e-02 5.54789901e-01 4.92524743e-01
-1.31677055e+00 5.02401888e-01 -1.01818061e+00 -4.84816164e-01
5.71490705e-01 1.95438400e-01 -5.51356256e-01 -3.03306073e-01
-1.01404285e+00 1.20417619e+00 1.55406862e-01 -3.82950082e-02
-1.16823721e+00 -6.90286875e-01 -1.21239817e+00 -4.70203072e-01
1.18073359e-01 -6.13112569e-01 1.21516192e+00 -3.01014900e-01
-1.51638937e+00 5.87220550e-01 -4.51987445e-01 -7.16841161e-01
6.58750892e-01 -5.12446821e-01 -5.03212571e-01 2.80841738e-01
3.44540477e-01 1.18678665e+00 6.68529689e-01 -1.12194502e+00
-9.83004570e-01 2.13852283e-02 -7.49278888e-02 3.79359394e-01
-2.34590411e-01 -2.18833223e-01 -7.61473060e-01 -3.82123381e-01
-2.71128237e-01 -1.14716566e+00 -5.29801190e-01 1.34850502e-01
-2.22607434e-01 -1.65244594e-01 1.26835370e+00 -5.34431577e-01
1.07298553e+00 -1.96159124e+00 -7.78914541e-02 2.11163703e-03
3.63750666e-01 2.45122403e-01 -5.44643283e-01 4.99022275e-01
5.63764684e-02 -2.34178439e-01 -1.87932938e-01 -8.03102732e-01
5.75428493e-02 5.63266098e-01 -1.43636674e-01 6.61426544e-01
3.75658154e-01 8.88205647e-01 -1.07231677e+00 -5.36575019e-01
8.15703750e-01 7.06212759e-01 -5.51198602e-01 2.39657402e-01
-4.26540524e-02 3.81801635e-01 -3.12599599e-01 2.83300847e-01
7.37269819e-01 1.34088352e-01 -1.62005782e-01 3.82644050e-02
-4.35997546e-01 6.54634118e-01 -9.08085883e-01 1.72191048e+00
-4.91640657e-01 1.11058295e+00 -5.35059273e-01 -8.30636799e-01
7.07128406e-01 8.50915629e-03 6.30113602e-01 -9.63127315e-01
-3.13904397e-02 -1.86700284e-01 -3.64378422e-01 -7.43569911e-01
9.54661727e-01 2.85168290e-01 -3.57215181e-02 3.28951389e-01
-1.57323822e-01 4.61314917e-02 3.23222339e-01 4.01923627e-01
1.40648651e+00 6.12880707e-01 -1.36412263e-01 -8.84229019e-02
3.79382610e-01 2.72011757e-01 6.75366938e-01 4.32678729e-01
-5.02386928e-01 6.21451855e-01 1.22339204e-01 -4.99327213e-01
-1.26181400e+00 -1.13799047e+00 -7.00309351e-02 1.09560132e+00
4.58305001e-01 -5.79133928e-01 -5.37565112e-01 -7.32249379e-01
3.22539024e-02 8.50201607e-01 -6.54681504e-01 -1.31965160e-01
-8.21302712e-01 -3.74536902e-01 4.80989248e-01 9.11634564e-01
5.28998256e-01 -6.88920438e-01 -7.14779854e-01 1.13458373e-01
-3.28254879e-01 -1.51138473e+00 -4.75093126e-01 -1.88524663e-01
-4.96243387e-01 -7.83459187e-01 -2.62708455e-01 -7.27136314e-01
4.99362469e-01 9.50115979e-01 8.32445920e-01 -1.00890763e-01
-1.46658748e-01 2.88764894e-01 -4.88692462e-01 -2.52178591e-02
-3.69251728e-01 1.35164231e-01 2.14902803e-01 -1.74974743e-02
4.34494644e-01 -6.50689542e-01 -7.42699921e-01 4.64876205e-01
-4.52160031e-01 4.10987318e-01 4.67569888e-01 3.28953594e-01
4.95338738e-01 3.07767745e-02 2.43105665e-01 -4.46085125e-01
-8.00767243e-02 -8.68787169e-01 -5.33500850e-01 -3.41275066e-01
-5.61785936e-01 1.34674028e-01 4.53587919e-01 -2.85768777e-01
-1.01454842e+00 3.40103120e-01 -3.63152117e-01 -6.61632121e-01
-3.02211881e-01 -4.12537418e-02 -1.22642189e-01 -1.08706713e-01
5.15966594e-01 1.38236627e-01 1.09331526e-01 -2.73926198e-01
7.22571254e-01 4.48713273e-01 6.43666446e-01 -1.80607095e-01
1.37064207e+00 8.22558582e-01 -1.03154220e-02 -9.00221169e-01
-2.44846016e-01 -6.98738992e-01 -8.52211058e-01 -5.26879966e-01
8.89994264e-01 -1.08083916e+00 -3.87782127e-01 3.12223643e-01
-9.96637344e-01 -8.90509307e-01 -2.93721318e-01 6.25418663e-01
-8.64860594e-01 4.56958830e-01 -3.73168707e-01 -5.86083829e-01
1.63214207e-01 -1.02607727e+00 1.11756337e+00 -4.54587489e-02
-2.01788992e-01 -1.05412722e+00 4.90811259e-01 5.67954183e-01
4.06181604e-01 1.24203920e-01 3.77111614e-01 -3.93702626e-01
-7.37718761e-01 -7.18091652e-02 -6.05135560e-02 7.74933249e-02
1.13106102e-01 4.99950647e-02 -9.72224355e-01 -4.06854749e-02
-6.03204429e-01 -5.60678579e-02 1.14343894e+00 2.25956142e-01
7.70598590e-01 -2.53173292e-01 -5.45744479e-01 6.60148740e-01
1.19989514e+00 -1.15796616e-02 9.71338391e-01 6.06721222e-01
1.10049498e+00 8.85728538e-01 8.67422283e-01 5.09445608e-01
1.40066361e+00 9.57749546e-01 5.90551794e-01 9.67272185e-03
-4.21075612e-01 -4.18736994e-01 6.57865822e-01 7.74850190e-01
-1.07787341e-01 -1.95217460e-01 -1.04444718e+00 1.02490938e+00
-2.16432142e+00 -1.36794472e+00 -4.07933980e-01 2.06367588e+00
2.88514376e-01 9.84032825e-02 5.01763105e-01 8.38794932e-02
4.34399843e-01 4.17852730e-01 -4.15273249e-01 -1.76497519e-01
-9.41078663e-02 -1.78152129e-01 7.88122475e-01 8.78419340e-01
-1.32914948e+00 1.41387355e+00 5.89853287e+00 7.44147599e-01
-8.85917604e-01 3.10169190e-01 3.25403512e-01 -1.36935681e-01
-3.65320861e-01 4.57547270e-02 -9.44750488e-01 5.16904831e-01
1.35689676e+00 1.84908565e-02 3.76340985e-01 9.02129114e-01
7.01858997e-01 -7.16642961e-02 -8.90910089e-01 8.67065430e-01
1.03937641e-01 -1.46731865e+00 -3.07770967e-01 1.28810570e-01
6.72535300e-01 6.88311100e-01 1.82938740e-01 3.52957517e-01
5.38180470e-01 -8.79174531e-01 9.91317809e-01 4.76627052e-01
4.63498116e-01 -6.62472129e-01 3.96398336e-01 2.60955065e-01
-1.65633678e+00 -1.14048414e-01 -3.05616349e-01 -1.89523902e-02
3.50994527e-01 1.17312968e-01 -1.00286078e+00 5.33275485e-01
7.33638525e-01 1.19185841e+00 -7.60672629e-01 9.13736284e-01
-6.18854947e-02 5.60408175e-01 -3.73155236e-01 2.69240916e-01
2.70856261e-01 5.59269823e-03 6.96224749e-01 1.43185961e+00
3.12568367e-01 -1.49857447e-01 1.09179355e-02 4.76512849e-01
3.11338305e-01 -2.25341707e-01 -8.77765298e-01 3.71058106e-01
7.18464673e-01 1.22792351e+00 -3.69174927e-01 -3.53012174e-01
-4.98256862e-01 8.75399768e-01 3.43681753e-01 3.23797494e-01
-1.24297416e+00 -1.82122305e-01 1.21378410e+00 2.71049857e-01
8.28780591e-01 -5.56637943e-01 -1.02913193e-01 -9.39498544e-01
-8.29439051e-03 -1.91888109e-01 5.90596162e-02 -7.03083694e-01
-7.18909502e-01 5.28044164e-01 1.96720406e-01 -1.73103976e+00
-5.11260211e-01 -4.46757913e-01 -7.39616811e-01 3.81560504e-01
-2.07183075e+00 -1.41270483e+00 -4.68807459e-01 5.88006556e-01
7.81844378e-01 -1.98456928e-01 4.51137990e-01 4.41711783e-01
-8.55363727e-01 4.67231989e-01 2.58891582e-01 -6.46006763e-02
3.91125649e-01 -1.08772838e+00 8.73827934e-01 1.05260098e+00
-2.04468743e-04 4.62962650e-02 8.00378799e-01 -4.59539413e-01
-1.52609944e+00 -1.71421695e+00 1.01308429e+00 -8.06504786e-01
6.79210067e-01 -3.63045871e-01 -7.19816208e-01 7.96140850e-01
6.03578836e-02 2.41220891e-01 4.71485704e-01 -3.60315382e-01
-3.95073861e-01 -3.43318582e-01 -1.00169766e+00 7.72234201e-01
1.39413655e+00 -4.82488304e-01 -3.86312068e-01 2.65864015e-01
7.45491922e-01 -3.73096049e-01 -6.18843079e-01 2.42566451e-01
6.23287976e-01 -8.83092046e-01 1.00217199e+00 -3.42449635e-01
3.50234479e-01 -4.93195266e-01 -4.30557936e-01 -1.06864810e+00
-4.85945135e-01 -7.02972233e-01 -4.01469499e-01 9.74409938e-01
3.46820503e-01 -6.29246116e-01 9.57021117e-01 4.51610982e-01
-4.90221679e-01 -5.98434269e-01 -1.03241062e+00 -9.88137960e-01
-4.54650223e-02 -1.02276433e+00 5.57031631e-01 7.06156015e-01
-3.15521479e-01 2.90927980e-02 -5.12900889e-01 4.71150845e-01
7.42784083e-01 -2.53430545e-01 1.24441206e+00 -9.08267260e-01
1.19766407e-02 -3.81678343e-01 -8.34789932e-01 -1.39603555e+00
4.11658287e-01 -7.72141278e-01 2.06793189e-01 -1.65971506e+00
-7.23628253e-02 -6.34080529e-01 -1.35212794e-01 4.53127623e-01
1.32894486e-01 4.07190919e-01 2.84832697e-02 2.82622010e-01
-9.03186321e-01 9.00936365e-01 8.51332307e-01 6.90469891e-03
-2.14818001e-01 -3.92844900e-02 -3.22031617e-01 6.53752208e-01
9.57377732e-01 -4.57757324e-01 -6.49305940e-01 -5.72865188e-01
-2.33608231e-01 -1.64432243e-01 6.37190521e-01 -1.18901753e+00
3.77078444e-01 -3.16546530e-01 -1.33064866e-01 -9.41571534e-01
8.01700830e-01 -5.56370616e-01 8.68192539e-02 2.17166305e-01
-9.28684920e-02 3.52408320e-01 2.75289863e-01 6.29102230e-01
1.58482105e-01 1.80416048e-01 4.52811182e-01 2.79654950e-01
-1.47117960e+00 4.73516941e-01 -7.22970128e-01 -7.75810704e-02
1.25420451e+00 -4.89011437e-01 -4.81963754e-01 -5.23765087e-01
-3.61978978e-01 6.37440860e-01 5.10495245e-01 7.76177824e-01
8.75500143e-01 -1.57723272e+00 -6.10102236e-01 6.34600297e-02
3.56572121e-01 -5.44596603e-03 5.15835762e-01 9.51081336e-01
-6.77792013e-01 5.48529446e-01 -1.83391944e-01 -5.90559423e-01
-1.23367155e+00 3.31832975e-01 8.76552835e-02 8.09993669e-02
-9.13032651e-01 6.57727420e-01 1.50818855e-01 -5.25734305e-01
-1.68758910e-02 -2.29165748e-01 -2.44888902e-01 -3.17848623e-01
5.82642853e-01 7.96059787e-01 -1.61866918e-01 -1.42368376e+00
-6.05130076e-01 6.73672736e-01 -5.60846180e-03 -3.79469961e-01
1.42682612e+00 -5.50880373e-01 3.86612177e-01 2.81613231e-01
1.41015375e+00 -1.05246142e-01 -1.80822372e+00 -2.95144826e-01
-1.20741270e-01 -5.84802449e-01 2.68062681e-01 -5.26155084e-02
-9.26347911e-01 7.90483296e-01 6.98725522e-01 -1.27252683e-01
8.76361072e-01 2.18720064e-01 1.04181373e+00 3.82135153e-01
4.33651149e-01 -1.10819292e+00 1.79204464e-01 5.35124660e-01
5.57013571e-01 -1.35407722e+00 -3.34498823e-01 -2.59457856e-01
-8.86933923e-01 7.39690304e-01 6.79245532e-01 -1.57230809e-01
6.11226320e-01 2.65637130e-01 5.09988517e-02 1.23967556e-03
-1.13871026e+00 -8.03350985e-01 1.28292322e-01 1.10526276e+00
-7.10221305e-02 4.06655632e-02 3.28036726e-01 1.14658158e-02
-4.24468905e-01 -1.93335056e-01 4.49216723e-01 8.96554708e-01
-5.47451019e-01 -1.08626664e+00 1.16655428e-03 7.31739402e-02
2.84333140e-01 1.30484700e-01 1.08730927e-01 8.16028893e-01
2.66329408e-01 1.09846270e+00 2.82719374e-01 -9.92594302e-01
2.36067176e-01 -2.67929822e-01 2.05217630e-01 -4.56355214e-01
6.04163408e-02 -2.70330995e-01 5.11309445e-01 -8.76251042e-01
-4.13396835e-01 -9.53844190e-01 -1.39500916e+00 -7.73985088e-01
7.10644722e-02 -2.18206421e-01 4.32144165e-01 1.00278652e+00
5.70254385e-01 2.81704813e-01 1.01855171e+00 -1.25888157e+00
7.01093897e-02 -6.05940759e-01 -1.06451381e-02 2.61016726e-01
6.37323141e-01 -8.27576399e-01 -1.93366379e-01 1.03921182e-01] | [6.162898540496826, 0.6310123205184937] |
6927ad29-9879-45cd-b860-dda900277d3e | deep-kernel-learning-for-mortality-prediction | 2212.00557 | null | https://arxiv.org/abs/2212.00557v1 | https://arxiv.org/pdf/2212.00557v1.pdf | Deep Kernel Learning for Mortality Prediction in the Face of Temporal Shift | Neural models, with their ability to provide novel representations, have shown promising results in prediction tasks in healthcare. However, patient demographics, medical technology, and quality of care change over time. This often leads to drop in the performance of neural models for prospective patients, especially in terms of their calibration. The deep kernel learning (DKL) framework may be robust to such changes as it combines neural models with Gaussian processes, which are aware of prediction uncertainty. Our hypothesis is that out-of-distribution test points will result in probabilities closer to the global mean and hence prevent overconfident predictions. This in turn, we hypothesise, will result in better calibration on prospective data. This paper investigates DKL's behaviour when facing a temporal shift, which was naturally introduced when an information system that feeds a cohort database was changed. We compare DKL's performance to that of a neural baseline based on recurrent neural networks. We show that DKL indeed produced superior calibrated predictions. We also confirm that the DKL's predictions were indeed less sharp. In addition, DKL's discrimination ability was even improved: its AUC was 0.746 (+- 0.014 std), compared to 0.739 (+- 0.028 std) for the baseline. The paper demonstrated the importance of including uncertainty in neural computing, especially for their prospective use. | ['Ameen Abu-Hanna', 'Miguel Rios'] | 2022-12-01 | null | null | null | null | ['mortality-prediction'] | ['medical'] | [-2.66741030e-02 4.97144699e-01 -3.77012715e-02 -5.71012974e-01
-8.22718799e-01 -2.07633018e-01 4.46130365e-01 4.78176326e-01
-5.08570850e-01 1.03055966e+00 4.14913714e-01 -4.76473033e-01
-4.16610271e-01 -7.28085756e-01 -8.91159177e-01 -6.58697724e-01
-2.62782812e-01 6.56735003e-01 -1.66185826e-01 2.28925645e-01
4.41744365e-02 3.30403388e-01 -1.10227406e+00 2.98043579e-01
6.87396824e-01 8.96939993e-01 -2.96491534e-01 5.86169064e-01
2.24007502e-01 8.02429199e-01 -6.90814734e-01 -4.68414754e-01
6.94526061e-02 -2.64259912e-02 -4.08782005e-01 -5.63292325e-01
9.15400609e-02 -1.75499514e-01 -2.26074859e-01 7.79502273e-01
6.78411722e-01 -4.39760275e-02 8.97676229e-01 -9.02528286e-01
-6.77673697e-01 8.40566933e-01 -1.09074578e-01 2.04807203e-02
7.97340497e-02 1.93966910e-01 4.56763476e-01 -4.63505566e-01
4.19978231e-01 9.60057497e-01 1.55726421e+00 5.78445971e-01
-1.39150953e+00 -5.42714059e-01 2.09628902e-02 -2.80562192e-01
-1.25941312e+00 -4.25409377e-01 2.85549134e-01 -6.68541670e-01
9.63646114e-01 -3.55325378e-02 4.90564495e-01 1.55155110e+00
9.44134355e-01 5.14420509e-01 8.70445311e-01 -1.99988142e-01
4.30684954e-01 6.50085807e-01 2.20211431e-01 1.98628500e-01
4.90587801e-01 6.19811714e-01 -2.19566196e-01 -4.65091944e-01
7.62784302e-01 3.25329840e-01 -3.71574908e-01 -2.19492212e-01
-1.12537515e+00 6.97356999e-01 2.22217754e-01 3.79393041e-01
-7.87468910e-01 1.32780910e-01 4.36562717e-01 2.15188384e-01
4.11186457e-01 8.50531220e-01 -7.94213176e-01 -4.41480458e-01
-1.10332656e+00 2.13474154e-01 1.04739487e+00 6.50936306e-01
-6.93836585e-02 2.50934847e-02 -4.55512255e-01 5.95099449e-01
-9.29322839e-02 3.08353335e-01 7.91657448e-01 -7.27060139e-01
8.12664479e-02 1.80449530e-01 2.22850397e-01 -1.15586758e+00
-9.97303963e-01 -8.45995605e-01 -1.22148561e+00 -8.60215724e-02
3.87255430e-01 -4.68438476e-01 -9.74130929e-01 1.75806713e+00
-3.04484040e-01 1.72986418e-01 5.58270514e-01 3.30697775e-01
3.47065300e-01 4.89099741e-01 1.75386921e-01 -2.66258210e-01
9.02965665e-01 -4.07492340e-01 -6.83105886e-01 9.34796631e-02
7.08926857e-01 -4.53332007e-01 7.67743409e-01 5.58982849e-01
-9.83877301e-01 -5.24182975e-01 -7.14421153e-01 6.98766232e-01
-2.03542173e-01 -7.68468827e-02 4.97156620e-01 7.57637918e-01
-1.11828136e+00 1.08877969e+00 -9.76289690e-01 -3.51872116e-01
5.51988304e-01 3.79442394e-01 -1.08673632e-01 1.81048304e-01
-1.34778309e+00 1.03651392e+00 4.98650283e-01 -4.42299386e-03
-4.44019258e-01 -1.10563827e+00 -4.98832017e-01 1.41162336e-01
7.18761459e-02 -8.88191581e-01 1.39663196e+00 -8.77773404e-01
-1.33917964e+00 2.52846718e-01 2.17894271e-01 -9.72079813e-01
7.94181049e-01 -2.29900941e-01 -5.47078013e-01 -3.47508758e-01
-1.60135180e-01 3.84911865e-01 5.26651084e-01 -1.11968768e+00
-5.25630295e-01 -3.44289243e-01 -6.89997375e-01 -2.46999547e-01
3.61388363e-02 -3.96531016e-01 4.85335402e-02 -5.40506184e-01
-5.93144484e-02 -8.92103493e-01 -5.46202600e-01 -5.76150954e-01
-3.80420744e-01 -7.03874826e-02 4.52799872e-02 -7.01565683e-01
1.48364687e+00 -1.98002446e+00 -5.42469144e-01 4.65560406e-01
2.40912974e-01 2.57636338e-01 3.76233280e-01 4.56581354e-01
-2.85254061e-01 1.86998889e-01 -1.63437918e-01 -1.65875450e-01
-2.07239076e-01 2.11563930e-01 -2.18424872e-01 3.37931752e-01
3.19434971e-01 9.36386943e-01 -7.37076342e-01 -7.62721077e-02
-5.61755374e-02 6.74509645e-01 -4.88154083e-01 2.20601577e-02
1.29508972e-01 3.02780181e-01 2.62535755e-02 5.17065585e-01
3.50106418e-01 -4.09469008e-01 -1.93929300e-02 -5.44241332e-02
-8.00945759e-02 -1.12271560e-02 -7.28240132e-01 1.17551601e+00
-3.60877246e-01 5.14875174e-01 -6.36466324e-01 -8.42059672e-01
1.12844527e+00 5.70707321e-01 6.43271089e-01 -3.93254697e-01
1.55712128e-01 2.05033258e-01 2.82755375e-01 -3.63825142e-01
2.85861373e-01 -3.89696240e-01 6.79450780e-02 1.79706141e-01
-3.00375164e-01 1.00390680e-01 -5.19953370e-01 -1.58457160e-01
1.18542862e+00 -1.11941598e-01 2.53709495e-01 -2.75743693e-01
-1.03135109e-01 -1.26753047e-01 7.00899839e-01 1.20151091e+00
-3.18169445e-01 7.21919179e-01 7.19972670e-01 -6.93358302e-01
-9.49494183e-01 -1.22340357e+00 -5.98228693e-01 2.70359904e-01
-5.23179173e-01 -2.32098047e-02 -4.51672912e-01 -5.03801942e-01
2.87518024e-01 1.20145237e+00 -9.39761341e-01 -6.72312617e-01
-3.24315339e-01 -1.13237298e+00 7.26252079e-01 8.48961830e-01
-3.13437171e-02 -8.69042754e-01 -8.96911144e-01 5.79093933e-01
2.98700571e-01 -5.89695454e-01 -1.60247147e-01 4.87254530e-01
-1.12271786e+00 -8.67275119e-01 -8.76474440e-01 -1.61534503e-01
3.85610223e-01 -7.13030636e-01 1.22329438e+00 -5.91383398e-01
4.83259149e-02 5.80176473e-01 -7.11352304e-02 -9.83108222e-01
-6.22921765e-01 1.50038794e-01 4.11153287e-01 -2.99088776e-01
4.73597825e-01 -4.22090501e-01 -7.36037731e-01 7.22936466e-02
-8.41022134e-01 -3.56478333e-01 7.43120074e-01 1.17983019e+00
4.22560632e-01 -2.13452224e-02 9.67280149e-01 -1.23827338e+00
1.12609255e+00 -6.20838881e-01 -4.33051229e-01 3.05547267e-01
-1.44026184e+00 2.88309723e-01 3.54999930e-01 -5.71950555e-01
-8.19354594e-01 -7.54251108e-02 5.74058015e-03 -6.04568124e-01
-1.38806343e-01 6.92901611e-01 5.44290662e-01 4.74832624e-01
8.88500631e-01 -4.60537262e-02 2.18472943e-01 -3.91836911e-01
9.07437876e-03 6.72693729e-01 6.52007520e-01 -3.50519210e-01
2.02866882e-01 5.11772297e-02 -1.72953069e-01 -4.22155470e-01
-5.23969769e-01 -1.36841401e-01 -4.45802033e-01 -2.32777633e-02
7.18266726e-01 -8.02175879e-01 -7.72573411e-01 4.24861670e-01
-1.01146209e+00 -3.49734902e-01 -6.48405790e-01 8.89819920e-01
-6.67549789e-01 -8.17420855e-02 -6.70193136e-01 -1.12883639e+00
-4.34764802e-01 -1.01767695e+00 5.72201073e-01 2.30380401e-01
-6.64930701e-01 -1.35066640e+00 2.26937190e-01 -2.06767619e-01
8.47064614e-01 5.29827356e-01 9.99430537e-01 -1.21828854e+00
1.20892294e-01 -5.98329723e-01 -8.52265731e-02 6.25237346e-01
2.10360587e-01 5.69309331e-02 -1.13032341e+00 -2.20351249e-01
1.51582256e-01 1.33595705e-01 7.58366942e-01 8.15435946e-01
1.02640676e+00 -3.61571133e-01 -5.22413969e-01 4.27167833e-01
1.16565096e+00 6.82612479e-01 5.63464701e-01 2.99856991e-01
1.94173142e-01 6.05003715e-01 2.45093286e-01 5.22837043e-01
6.01691157e-02 3.55550230e-01 1.23722255e-01 -1.70761824e-01
3.21529895e-01 -2.46082708e-01 3.37661177e-01 6.36829138e-01
9.65095162e-02 5.96211739e-02 -1.40075219e+00 3.90225142e-01
-1.92757463e+00 -7.57680893e-01 4.47118394e-02 2.44199252e+00
9.64009464e-01 4.76065099e-01 1.52711526e-01 -6.96274266e-02
5.39770424e-01 -4.41014737e-01 -7.62813568e-01 -7.03465939e-01
-2.38647163e-02 1.61937788e-01 8.41367424e-01 1.65911525e-01
-9.16327238e-01 3.11342567e-01 6.94335270e+00 3.03581864e-01
-1.07348526e+00 -1.29805833e-01 1.20112479e+00 -7.80303404e-02
-2.75317337e-02 -5.62751234e-01 -5.92804492e-01 6.25550568e-01
1.76470447e+00 -2.57319152e-01 -1.98171511e-01 8.84734273e-01
2.15047419e-01 -1.74308434e-01 -1.34423900e+00 8.96842301e-01
-7.95724094e-02 -1.41508484e+00 -1.37764573e-01 -3.38418409e-02
7.92683244e-01 3.55240367e-02 4.75385338e-01 5.72801411e-01
3.60837013e-01 -1.39297783e+00 3.17416072e-01 1.29423642e+00
5.31704903e-01 -9.80359137e-01 1.48126030e+00 3.51511568e-01
-2.20500425e-01 -1.28850359e-02 -2.80982554e-01 -7.92835560e-03
9.10069048e-02 9.66751814e-01 -1.49687994e+00 3.93813252e-01
7.72379100e-01 4.73464042e-01 -3.39593291e-01 1.16793704e+00
4.68243450e-01 8.52830470e-01 -1.71812221e-01 -3.45329419e-02
1.54252186e-01 4.47615199e-02 2.38669425e-01 1.25473571e+00
5.33935130e-01 -2.61975795e-01 -1.75788775e-01 7.79482305e-01
2.28777945e-01 4.94214357e-04 -6.98954284e-01 -1.41734272e-01
2.47776940e-01 4.53514248e-01 -3.56327266e-01 -3.79969895e-01
-1.77508861e-01 7.14624465e-01 -9.87698734e-02 3.22776198e-01
-6.78176403e-01 -3.26123327e-01 6.63618505e-01 2.41794601e-01
-4.21260595e-02 3.22925866e-01 -4.85284567e-01 -6.43421471e-01
-2.62973130e-01 -6.92813993e-01 3.88330072e-01 -5.85460305e-01
-1.75433815e+00 8.53194058e-01 -9.86973941e-02 -8.87981236e-01
-5.48034906e-01 -6.31609499e-01 -2.52358913e-01 1.03950989e+00
-1.17834520e+00 -5.54254472e-01 1.31283581e-01 2.81256258e-01
8.97197723e-02 -1.51278093e-01 1.12504709e+00 1.46018192e-01
-4.97469276e-01 8.43780875e-01 4.88310248e-01 2.07201857e-02
1.09899688e+00 -1.26929235e+00 3.48994285e-01 3.24482352e-01
-2.13345632e-01 7.79808760e-01 7.89402187e-01 -8.15863729e-01
-7.19225466e-01 -1.25651014e+00 7.02572405e-01 -9.32587862e-01
5.42303562e-01 2.62306511e-01 -1.24093223e+00 7.29799628e-01
-1.65016815e-01 -3.93656313e-01 9.55148876e-01 2.78864712e-01
-2.16651618e-01 -1.73977703e-01 -1.29274166e+00 3.39939535e-01
4.74560767e-01 -3.11326861e-01 -7.72365928e-01 2.09305674e-01
8.09440374e-01 -4.12807733e-01 -1.29754305e+00 8.41591477e-01
8.32140505e-01 -1.11151183e+00 8.67900372e-01 -1.03240108e+00
2.96267748e-01 2.40920529e-01 -1.22975312e-01 -1.44245374e+00
-4.21815991e-01 -4.42023695e-01 -3.36026624e-02 9.43953872e-01
8.82320404e-01 -1.06588817e+00 7.54037917e-01 1.12491310e+00
-1.05458274e-01 -1.29251373e+00 -8.05735826e-01 -8.69110346e-01
4.32560354e-01 -7.28093028e-01 5.50684035e-01 9.12621319e-01
-2.60188375e-02 -2.58018672e-02 -4.72650021e-01 2.36521184e-01
2.01787934e-01 -4.18278396e-01 3.21481913e-01 -1.46884930e+00
-4.65679437e-01 -5.11456668e-01 -5.39324641e-01 -2.97871470e-01
-2.71507800e-01 -5.82300901e-01 -2.91379727e-03 -1.30850947e+00
-3.66526172e-02 -5.90101004e-01 -8.21871638e-01 3.27285767e-01
-1.68958947e-01 -3.93432617e-01 -6.94560725e-03 2.28102237e-01
7.51241576e-03 1.56843647e-01 6.92779005e-01 1.20489635e-01
-6.45259619e-01 6.45256937e-01 -7.57017374e-01 6.98963761e-01
1.25209928e+00 -4.91106600e-01 -2.80070692e-01 4.44120541e-02
2.80356884e-01 2.18918800e-01 3.27290185e-02 -1.20684803e+00
2.59874314e-01 1.46384522e-01 7.55845785e-01 -3.99683923e-01
1.51941881e-01 -7.58931577e-01 5.74139655e-01 9.29369330e-01
-5.68366051e-01 4.46924537e-01 4.68767732e-01 7.62665272e-01
-2.85335649e-02 9.54500865e-03 5.65239012e-01 -1.41272843e-01
-6.06989712e-02 7.36776292e-02 -4.92591023e-01 -1.01665072e-01
9.30864453e-01 -1.74131006e-01 -8.93850923e-02 -4.31738406e-01
-1.01961958e+00 1.01338230e-01 2.50758708e-01 3.30490261e-01
5.41750908e-01 -9.34910953e-01 -5.89986801e-01 1.23954050e-01
7.05496222e-02 -1.67653374e-02 1.65672362e-01 1.12945724e+00
-2.68412828e-01 5.84621847e-01 1.11933187e-01 -6.38484478e-01
-8.49522054e-01 6.10328317e-01 6.84204996e-01 -3.98825258e-01
-5.85136354e-01 7.81319916e-01 -6.50497526e-02 -3.95175457e-01
5.66703796e-01 -8.05399239e-01 4.80606183e-02 1.41034171e-01
3.99491161e-01 3.12627822e-01 6.13012686e-02 4.27762941e-02
-2.42198870e-01 1.74201131e-01 -2.56589770e-01 3.23498063e-02
1.45220900e+00 1.45988107e-01 2.21801803e-01 1.14775777e+00
8.44299793e-01 -2.97266752e-01 -1.15143514e+00 -1.38245076e-01
3.86481345e-01 -1.13127735e-02 -4.75058965e-02 -1.34956360e+00
-6.19396806e-01 8.69673073e-01 9.30974245e-01 1.24686196e-01
9.47263002e-01 -1.32758871e-01 5.31844079e-01 3.44877630e-01
2.39452779e-01 -9.11219001e-01 -3.05263281e-01 2.36712933e-01
7.10819781e-01 -1.24569643e+00 -2.57081300e-01 1.74536705e-01
-7.94815838e-01 1.22338355e+00 1.60606742e-01 1.01157486e-01
8.85618389e-01 2.61243910e-01 2.09968299e-01 -5.41458204e-02
-9.12178755e-01 4.80122954e-01 1.30412251e-01 7.47277617e-01
4.52776611e-01 4.20841217e-01 1.33730903e-01 1.09499860e+00
-3.50237429e-01 3.29218656e-01 5.68578243e-01 5.75813890e-01
-4.90537100e-02 -7.33223855e-01 -3.31705183e-01 9.51547503e-01
-4.54675585e-01 -2.29427785e-01 1.18259341e-02 8.22276592e-01
8.57864097e-02 6.13424361e-01 2.46733084e-01 -4.43201661e-01
5.73794663e-01 5.65304935e-01 -2.88980342e-02 -4.35269415e-01
-6.90818667e-01 -1.63819045e-01 -1.02666616e-02 -6.01242721e-01
-5.84745593e-02 -8.90226185e-01 -1.14027107e+00 -1.85882598e-01
-4.82934788e-02 2.47052208e-01 6.14640534e-01 4.48077500e-01
6.18752956e-01 9.10721421e-01 1.99046373e-01 -3.28527391e-01
-1.01856732e+00 -9.74001765e-01 -4.96061325e-01 1.12934411e-01
4.51510489e-01 -4.40920860e-01 -3.74546379e-01 -2.20630616e-01] | [7.95538854598999, 6.1280903816223145] |
eb9216cf-9632-44a7-b34a-19a68d888ea2 | privacy-preserving-data-synthetisation-for | 2212.00484 | null | https://arxiv.org/abs/2212.00484v1 | https://arxiv.org/pdf/2212.00484v1.pdf | Privacy-Preserving Data Synthetisation for Secure Information Sharing | We can protect user data privacy via many approaches, such as statistical transformation or generative models. However, each of them has critical drawbacks. On the one hand, creating a transformed data set using conventional techniques is highly time-consuming. On the other hand, in addition to long training phases, recent deep learning-based solutions require significant computational resources. In this paper, we propose PrivateSMOTE, a technique designed for competitive effectiveness in protecting cases at maximum risk of re-identification while requiring much less time and computational resources. It works by synthetic data generation via interpolation to obfuscate high-risk cases while minimizing data utility loss of the original data. Compared to multiple conventional and state-of-the-art privacy-preservation methods on 20 data sets, PrivateSMOTE demonstrates competitive results in re-identification risk. Also, it presents similar or higher predictive performance than the baselines, including generative adversarial networks and variational autoencoders, reducing their energy consumption and time requirements by a minimum factor of 9 and 12, respectively. | ['Nitesh Chawla', 'Luís Antunes', 'Pedro Faria', 'Nuno Moniz', 'Tânia Carvalho'] | 2022-12-01 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 1.64705619e-01 7.83423483e-02 8.11882466e-02 -2.72917509e-01
-1.11240458e+00 -6.91983581e-01 4.92211312e-01 2.14815810e-01
-5.69871724e-01 8.04715812e-01 -2.01327316e-02 -3.49812329e-01
4.77674901e-02 -1.10248089e+00 -6.56307995e-01 -8.59029353e-01
2.08702669e-01 2.27538079e-01 -2.04889268e-01 -1.13244794e-01
-1.43356830e-01 4.39655870e-01 -1.16930950e+00 4.07397076e-02
1.10080838e+00 1.00290871e+00 -7.04514563e-01 1.68135136e-01
2.67707556e-01 3.02851528e-01 -6.60985172e-01 -1.16349769e+00
6.24514997e-01 -5.00158429e-01 -4.50027764e-01 -4.97146904e-01
4.75113183e-01 -7.59985685e-01 -7.31583297e-01 1.25136304e+00
8.26571941e-01 1.36054261e-02 5.68006396e-01 -1.48741949e+00
-1.02945542e+00 4.19127822e-01 -5.00602126e-01 -2.20673740e-01
1.56827748e-01 -4.12950926e-02 5.91065228e-01 -6.64840341e-01
2.28425890e-01 9.23539162e-01 9.64099348e-01 1.07580853e+00
-1.59504461e+00 -1.11490250e+00 -4.21038240e-01 1.02229796e-01
-1.58502889e+00 -5.87336004e-01 7.76933968e-01 -2.56660849e-01
4.96497393e-01 6.62661910e-01 3.22178751e-01 1.43609047e+00
-2.93158758e-02 6.38786495e-01 9.82450068e-01 -1.69784456e-01
3.45010042e-01 4.62675035e-01 -2.22822130e-01 6.09818697e-01
4.42297816e-01 4.30936575e-01 -3.05597574e-01 -7.16166973e-01
4.90735829e-01 2.90873080e-01 -3.50057423e-01 -4.22370940e-01
-6.69916630e-01 9.98455226e-01 2.02172026e-01 2.82117482e-02
-1.32168651e-01 1.94846749e-01 4.98232156e-01 1.24095112e-01
4.85057652e-01 2.12608010e-01 -1.01099819e-01 1.08764216e-01
-1.10509813e+00 3.07096422e-01 6.34120405e-01 8.23376954e-01
6.20521486e-01 1.26544029e-01 -3.21557820e-01 5.21824300e-01
8.99546146e-02 3.85060549e-01 4.85413611e-01 -7.10189939e-01
5.77151477e-01 4.01621372e-01 7.04620481e-02 -1.17669249e+00
1.24760807e-01 -3.46697330e-01 -1.28237665e+00 3.66725206e-01
3.60006511e-01 -3.94429684e-01 -8.16793561e-01 2.02283835e+00
3.58040124e-01 1.77161187e-01 2.29133576e-01 6.14715874e-01
5.93256950e-01 5.70425153e-01 1.51475549e-01 -1.31022915e-01
1.20717859e+00 -5.95615268e-01 -7.57739007e-01 1.79056272e-01
3.45428497e-01 -4.91603941e-01 9.15122926e-01 2.28275135e-01
-1.02309191e+00 -3.00482959e-01 -1.04267287e+00 -3.15480202e-01
-4.71291721e-01 -2.40040105e-03 5.06357968e-01 1.49119198e+00
-1.05835235e+00 6.99783266e-01 -7.49558270e-01 5.98204229e-03
8.06742191e-01 5.82257688e-01 -4.41656172e-01 3.62401426e-01
-1.37751305e+00 5.42218685e-01 2.00572878e-01 -2.83919901e-01
-1.00675786e+00 -1.19242299e+00 -8.78765762e-01 3.95426601e-01
2.65999526e-01 -6.81295276e-01 8.57225895e-01 -5.83143890e-01
-1.43291330e+00 7.13328302e-01 8.60931352e-02 -8.53610098e-01
1.07237601e+00 -2.74038374e-01 -6.30944788e-01 -2.13952795e-01
-3.58992547e-01 4.68083590e-01 9.45543230e-01 -1.23122692e+00
-3.87396663e-01 -4.43946660e-01 -2.50508428e-01 -2.26118788e-01
-9.91846681e-01 -1.11329712e-01 -2.71786243e-01 -9.75980937e-01
-5.02263486e-01 -9.91829038e-01 -2.81531662e-01 2.61720419e-01
-7.18798816e-01 1.23001099e-01 1.05658793e+00 -9.61439192e-01
1.33631980e+00 -2.42520809e+00 -1.53653085e-01 2.81995088e-01
4.50595170e-01 7.73206949e-01 1.69002190e-01 3.40810359e-01
5.40889762e-02 6.36952579e-01 -5.09577751e-01 -6.14920735e-01
2.23782316e-01 -5.84241822e-02 -5.04655540e-01 5.90395093e-01
-1.05314434e-01 8.55816245e-01 -6.10960245e-01 -3.78604501e-01
3.33455294e-01 9.34629917e-01 -8.35895836e-01 2.67920375e-01
1.32329330e-01 3.97076547e-01 -2.67496616e-01 3.49289924e-01
1.01189649e+00 1.43690258e-01 6.69757277e-02 -1.90448947e-02
4.15238440e-01 -5.79323210e-02 -8.90561759e-01 1.40067506e+00
-3.86905640e-01 5.30088782e-01 2.28754729e-02 -8.03481102e-01
8.35202813e-01 4.33407515e-01 4.94830638e-01 -4.98877972e-01
3.37355494e-01 4.63355780e-02 -4.61576760e-01 -1.74614444e-01
5.53256452e-01 1.00136265e-01 -3.87815475e-01 5.00998437e-01
-2.20271394e-01 4.71067488e-01 -3.90182555e-01 2.02112999e-02
8.29883635e-01 -1.36980847e-01 1.84154108e-01 1.94394402e-02
4.42505062e-01 -4.69664842e-01 8.66704047e-01 5.80654681e-01
-3.30636919e-01 5.46520531e-01 3.11134160e-01 -4.70761627e-01
-1.20579731e+00 -1.11018836e+00 -6.85724318e-02 7.03365147e-01
1.62412062e-01 -4.44069445e-01 -1.19141519e+00 -7.62011826e-01
3.41652453e-01 1.06151283e+00 -8.46434295e-01 -5.39122820e-01
-4.46252584e-01 -8.63401592e-01 1.16673875e+00 3.38376760e-01
7.66383648e-01 -6.23507857e-01 -3.93486857e-01 -2.10281108e-02
-2.01619089e-01 -7.82359302e-01 -7.91665316e-01 -3.01666886e-01
-7.15188503e-01 -8.19008589e-01 -7.05277205e-01 -4.54164714e-01
8.78046989e-01 9.77776200e-02 7.48103976e-01 -8.77245218e-02
-3.06400388e-01 -7.93364868e-02 -1.35865852e-01 -4.57128435e-01
-6.86918736e-01 -1.39457872e-03 1.24892548e-01 3.50387901e-01
5.04676461e-01 -6.54303312e-01 -7.91276991e-01 2.06869826e-01
-1.11355197e+00 -2.42422730e-01 2.92011738e-01 9.10698473e-01
6.64607704e-01 1.71623826e-01 4.32596087e-01 -1.05909753e+00
8.05768311e-01 -4.31114078e-01 -5.66920757e-01 4.35949147e-01
-1.02266300e+00 8.98851082e-03 1.05353844e+00 -4.53428060e-01
-1.09193063e+00 1.28237987e-02 -1.86275288e-01 -6.12099707e-01
-3.66319939e-02 -7.25460872e-02 -4.27218318e-01 -2.52356499e-01
8.15221906e-01 5.30234754e-01 1.28670052e-01 -6.78074896e-01
3.92049909e-01 5.76646686e-01 6.45349264e-01 -2.79705793e-01
1.16057944e+00 4.65432674e-01 9.04420391e-03 -3.25722128e-01
-2.42901713e-01 2.04434961e-01 -2.34112531e-01 2.50807911e-01
8.06622982e-01 -7.87083745e-01 -8.36447537e-01 7.03939974e-01
-7.96913683e-01 7.33207837e-02 -3.19433153e-01 3.14431936e-02
-2.53389776e-01 6.74382627e-01 -4.36086029e-01 -8.09043646e-01
-8.61603975e-01 -1.01868153e+00 7.81922579e-01 1.83047175e-01
-1.22248583e-01 -9.00668204e-01 -1.24500625e-01 4.43447530e-01
4.67275918e-01 9.18698549e-01 8.36828649e-01 -7.09978998e-01
-3.55488241e-01 -7.45856047e-01 6.00655703e-03 4.86047834e-01
2.74509937e-01 -2.37161577e-01 -1.10074890e+00 -7.68978775e-01
-4.42549363e-02 -1.39483005e-01 7.34574914e-01 3.29511091e-02
1.83571947e+00 -8.06853354e-01 -2.83072978e-01 1.13651085e+00
1.54449522e+00 4.21021163e-01 8.38993907e-01 1.11896642e-01
6.88089490e-01 4.25620437e-01 1.37142837e-01 6.44001663e-01
2.40188271e-01 6.29380524e-01 4.25649911e-01 -3.69481564e-01
1.91099107e-01 -6.58949137e-01 2.73147047e-01 1.03599094e-01
2.88091712e-02 -3.58682871e-01 -6.40667140e-01 4.92203206e-01
-1.69847906e+00 -1.24350500e+00 1.42699376e-01 2.53441715e+00
9.36038196e-01 -2.18118757e-01 3.28228176e-01 3.26900393e-01
6.50396168e-01 1.78391963e-01 -7.72621274e-01 -4.61776227e-01
7.12673888e-02 1.82982117e-01 8.15798521e-01 2.22950906e-01
-1.14007843e+00 6.29762650e-01 6.13060474e+00 9.57579494e-01
-9.83027697e-01 2.90103972e-01 9.29165661e-01 -4.31184471e-01
-5.24570465e-01 -3.10394585e-01 -6.34915948e-01 7.71497428e-01
9.57964301e-01 -5.21610320e-01 7.31940210e-01 9.91814017e-01
-2.49257490e-01 4.80985016e-01 -1.12636495e+00 1.05977821e+00
1.67590603e-01 -1.27616787e+00 6.22780249e-03 3.84255081e-01
6.30023420e-01 -4.89945710e-01 4.26753491e-01 1.51310101e-01
5.22482574e-01 -1.01511228e+00 6.01235271e-01 3.07012379e-01
1.13773751e+00 -1.31758726e+00 5.78197479e-01 2.50967652e-01
-7.52734780e-01 -3.81228030e-02 -3.41963321e-01 5.58858275e-01
3.04292645e-02 6.11303210e-01 -5.51543653e-01 5.91616452e-01
8.47223461e-01 -6.92641456e-03 -4.41204309e-01 7.16179729e-01
-9.20970291e-02 6.24422491e-01 -2.73588419e-01 9.01087001e-02
-1.67305648e-01 -1.34344891e-01 5.52355230e-01 9.95547056e-01
4.85637039e-01 3.54919070e-03 -1.00589402e-01 1.05500424e+00
-4.82412279e-01 -1.55668527e-01 -6.74999774e-01 -3.40804495e-02
7.99031258e-01 9.91797984e-01 5.47012966e-03 -2.28665784e-01
-7.00458437e-02 1.28108394e+00 1.54589057e-01 1.41869575e-01
-1.18311334e+00 -7.61486888e-01 8.83113861e-01 7.22578317e-02
5.53286746e-02 2.30332613e-02 -4.81612980e-01 -1.20380354e+00
2.06127763e-01 -8.74715686e-01 4.76435870e-01 -1.06283516e-01
-1.20063698e+00 6.84277534e-01 -3.18984836e-01 -1.39529347e+00
-2.79033273e-01 -6.21554516e-02 -3.82836580e-01 1.03153718e+00
-1.15076649e+00 -1.13732350e+00 -2.87255049e-01 8.31668258e-01
-4.99608032e-02 -4.87984568e-01 1.31183004e+00 5.56623936e-01
-8.12647104e-01 1.69884932e+00 6.10632122e-01 3.53085607e-01
6.29730225e-01 -1.10485637e+00 8.30509603e-01 1.06176877e+00
-4.14155200e-02 8.42670977e-01 5.27937829e-01 -4.00024921e-01
-1.39262426e+00 -1.31871998e+00 7.73911178e-01 -3.59301090e-01
1.20693028e-01 -7.66739130e-01 -9.39693749e-01 5.65006793e-01
1.02586225e-01 7.73129752e-03 1.23915935e+00 -1.42329589e-01
-6.68439090e-01 -2.69720256e-01 -1.92864811e+00 6.92238688e-01
9.24587846e-01 -6.95136607e-01 -2.12622106e-01 9.26276371e-02
6.89732432e-01 -3.53487253e-01 -1.11608601e+00 2.34818965e-01
6.01332009e-01 -1.04409480e+00 1.23952878e+00 -6.17070079e-01
2.91699022e-01 -8.86771381e-02 -5.58152646e-02 -1.14685130e+00
-4.13502902e-01 -1.15850687e+00 -3.77774328e-01 1.67018056e+00
3.26095551e-01 -8.78953099e-01 1.18276548e+00 1.20026374e+00
4.24580395e-01 -5.63464522e-01 -1.02244329e+00 -9.10182238e-01
1.72079533e-01 -4.24568318e-02 1.14449656e+00 1.24278045e+00
-3.02476972e-01 -3.36198956e-01 -9.44075763e-01 1.91389501e-01
9.99471247e-01 -1.59348339e-01 9.63213086e-01 -9.12430823e-01
-1.64282322e-01 -1.88694060e-01 -3.31330299e-01 -4.54477549e-01
3.93197425e-02 -8.89062762e-01 -3.52193207e-01 -1.12646461e+00
9.84668806e-02 -4.88158584e-01 -4.52741921e-01 7.58623779e-01
-2.06673801e-01 3.49510998e-01 1.11369155e-01 1.69587001e-01
3.57297286e-02 7.26972342e-01 7.95841753e-01 -1.86418921e-01
-2.40691230e-01 7.62336329e-02 -1.12633300e+00 4.15085137e-01
9.79617119e-01 -6.57665610e-01 -5.43423951e-01 -5.03145635e-01
-7.32345730e-02 -1.52557320e-03 5.38699806e-01 -1.00369835e+00
1.01636000e-01 -1.98967859e-01 5.48469067e-01 -3.78271431e-01
2.81578332e-01 -1.03505909e+00 6.21950865e-01 6.02727294e-01
-3.34522069e-01 -3.01629752e-02 2.95021564e-01 7.52704084e-01
-2.35935170e-02 3.07451412e-02 9.92115200e-01 2.24105924e-01
-3.18785459e-01 6.57836676e-01 -1.87921375e-01 -5.70223331e-02
1.34570718e+00 -3.00728381e-01 -3.61279488e-01 -3.74404371e-01
-4.51977164e-01 -2.67340578e-02 6.68926120e-01 5.11321545e-01
5.04502535e-01 -1.51925230e+00 -6.93323612e-01 4.58336085e-01
-6.45924062e-02 -1.69878706e-01 5.34854531e-01 2.21770808e-01
-4.32187378e-01 1.04169361e-01 -2.72850484e-01 4.64294069e-02
-1.41677344e+00 9.95009124e-01 2.79139400e-01 -2.15154216e-01
-6.22565329e-01 8.29017878e-01 1.68061614e-01 -3.48765969e-01
4.11334962e-01 2.69933671e-01 1.86019331e-01 -1.28126815e-01
6.08693600e-01 6.90970957e-01 -3.16943377e-02 -4.58288103e-01
-3.98133129e-01 2.37238139e-01 -1.68085933e-01 8.49591643e-02
1.26030982e+00 1.63975254e-01 1.16768576e-01 -5.36965132e-01
1.51717615e+00 2.39304245e-01 -1.27631438e+00 -2.60539651e-01
-3.41369182e-01 -7.66349912e-01 1.43142924e-01 -8.08004558e-01
-1.46089339e+00 7.75778532e-01 9.81820285e-01 2.86783397e-01
1.43533051e+00 -6.00683451e-01 1.33879244e+00 8.57510567e-02
4.31570172e-01 -9.39208269e-01 -3.61626625e-01 -2.41897807e-01
7.32539177e-01 -1.15349066e+00 1.14518665e-02 -3.60500127e-01
-5.97148716e-01 6.08725309e-01 4.50766116e-01 -1.10501997e-01
6.82567537e-01 2.57645309e-01 -5.80285080e-02 2.85874575e-01
-2.83760101e-01 5.21473587e-01 2.88137674e-01 8.60289156e-01
-3.99147756e-02 2.52866030e-01 -3.65567297e-01 7.30956793e-01
-4.74556506e-01 -1.27468660e-01 3.04365903e-01 6.30323887e-01
2.87450075e-01 -1.43004918e+00 -3.18027526e-01 4.71900612e-01
-9.89539206e-01 -3.74426059e-02 -5.13073146e-01 5.83361804e-01
1.94737911e-01 8.45663965e-01 -8.13366920e-02 -9.12859917e-01
1.97763786e-01 -5.61445244e-02 -3.53366323e-02 -1.24144722e-02
-8.43696833e-01 -3.43861848e-01 3.85202467e-02 -4.25804198e-01
-1.89073831e-02 -7.42406368e-01 -8.40232730e-01 -1.04091144e+00
-2.49923527e-01 2.73535005e-03 6.21828437e-01 3.62837553e-01
8.83163452e-01 2.94074297e-01 8.46467733e-01 -3.62733692e-01
-9.42016363e-01 -5.74560046e-01 -4.82626438e-01 6.82147026e-01
3.22496504e-01 -2.85573900e-01 -2.05138519e-01 -2.75621004e-02] | [6.038959980010986, 6.938927173614502] |
95af91e2-fdfb-46ed-b226-b09839901bfe | audio-anti-spoofing-using-a-simple-attention | 2211.09898 | null | https://arxiv.org/abs/2211.09898v1 | https://arxiv.org/pdf/2211.09898v1.pdf | Audio Anti-spoofing Using a Simple Attention Module and Joint Optimization Based on Additive Angular Margin Loss and Meta-learning | Automatic speaker verification systems are vulnerable to a variety of access threats, prompting research into the formulation of effective spoofing detection systems to act as a gate to filter out such spoofing attacks. This study introduces a simple attention module to infer 3-dim attention weights for the feature map in a convolutional layer, which then optimizes an energy function to determine each neuron's importance. With the advancement of both voice conversion and speech synthesis technologies, unseen spoofing attacks are constantly emerging to limit spoofing detection system performance. Here, we propose a joint optimization approach based on the weighted additive angular margin loss for binary classification, with a meta-learning training framework to develop an efficient system that is robust to a wide range of spoofing attacks for model generalization enhancement. As a result, when compared to current state-of-the-art systems, our proposed approach delivers a competitive result with a pooled EER of 0.99% and min t-DCF of 0.0289. | ['John H. L. Hansen', 'Zhenyu Wang'] | 2022-11-17 | null | null | null | null | ['voice-conversion', 'voice-conversion', 'speaker-verification'] | ['audio', 'speech', 'speech'] | [ 3.59899163e-01 -2.15278044e-01 -5.97366020e-02 -2.50250101e-01
-5.08394301e-01 -5.44272244e-01 5.38683534e-01 1.06129825e-01
-4.79670823e-01 2.73699015e-01 1.04780734e-01 -7.80437589e-01
1.59418806e-01 -4.16364610e-01 -5.23023665e-01 -6.68393791e-01
-4.91347834e-02 -2.93852240e-01 8.80991854e-03 -3.68906170e-01
3.07807386e-01 7.43626356e-01 -1.23244083e+00 1.40169695e-01
8.79594088e-01 1.25872862e+00 -3.73404436e-02 7.85681307e-01
2.67055541e-01 4.60112810e-01 -1.10605872e+00 -7.97279716e-01
7.97106139e-03 -1.56054527e-01 -4.75259095e-01 -4.68884021e-01
3.41359615e-01 -2.19247863e-01 -6.87421978e-01 1.42009079e+00
8.46833169e-01 3.89184020e-02 2.94228315e-01 -1.11567676e+00
-6.42363310e-01 6.43985510e-01 -1.18948705e-01 5.96462488e-01
2.87991226e-01 2.41239041e-01 9.24171567e-01 -7.14043617e-01
1.58258174e-02 1.21460688e+00 5.28263390e-01 7.50446916e-01
-9.80995715e-01 -1.10555661e+00 -4.30491343e-02 4.97049332e-01
-1.26419950e+00 -9.82749403e-01 1.17691898e+00 -5.09321801e-02
1.01290345e+00 4.58358824e-01 2.86920398e-01 1.39717960e+00
3.09142500e-01 5.66685438e-01 7.72820890e-01 -3.80237758e-01
1.58489291e-02 5.43637574e-01 1.37988508e-01 6.66220188e-01
3.70044082e-01 4.24018562e-01 -7.99421728e-01 -1.42240033e-01
2.06385896e-01 -3.17144752e-01 -4.13831681e-01 1.65362433e-01
-6.94542110e-01 8.38851690e-01 6.84923947e-01 2.14520007e-01
-2.24128500e-01 -7.87183046e-02 5.34321189e-01 1.89648345e-01
2.11903527e-01 5.78517556e-01 -2.75419444e-01 9.16722566e-02
-9.12332654e-01 -8.67700204e-03 7.33267963e-01 2.65207499e-01
2.17945874e-01 5.07945299e-01 -2.06188802e-02 8.87107491e-01
6.06472790e-01 5.98981142e-01 5.50148308e-01 -1.47489935e-01
6.67433441e-01 1.31849617e-01 -2.26847500e-01 -1.21798337e+00
-2.06348628e-01 -1.01292777e+00 -6.93513453e-01 1.36201456e-01
-1.19795263e-01 -2.86026359e-01 -7.95656800e-01 1.78396857e+00
3.37402016e-01 2.37167135e-01 8.68152156e-02 7.04658091e-01
5.31145513e-01 4.64770019e-01 -1.05321027e-01 -1.78414091e-01
1.45442045e+00 -8.58963609e-01 -1.07044756e+00 -2.82722205e-01
1.20239243e-01 -8.76233280e-01 7.66386151e-01 4.46537971e-01
-7.57394254e-01 -4.99258101e-01 -1.52223945e+00 2.29076892e-01
-6.11584425e-01 -1.47222236e-01 4.96178448e-01 1.57030940e+00
-7.46472418e-01 4.57500249e-01 -7.26542056e-01 4.48877700e-02
6.70063972e-01 5.82980275e-01 -1.99514359e-01 2.62093246e-01
-1.60136139e+00 9.94979501e-01 3.51237684e-01 1.92275375e-01
-1.15895665e+00 -5.56978345e-01 -7.80435205e-01 2.13817909e-01
-1.89529315e-01 -3.40637624e-01 9.87979472e-01 -6.52747273e-01
-1.83747709e+00 5.85999787e-01 -1.53751776e-01 -8.68730783e-01
3.03490669e-01 -3.03451985e-01 -1.16050923e+00 8.74611810e-02
-4.83412653e-01 3.11700165e-01 1.37835515e+00 -1.03582990e+00
-6.02402270e-01 -3.61787289e-01 -1.02755420e-01 1.86349638e-02
-9.97421026e-01 5.14438331e-01 -5.96413873e-02 -8.11320961e-01
9.97751355e-02 -6.63005710e-01 6.81561604e-02 -4.04582500e-01
-5.38031161e-01 9.26362164e-03 1.17336595e+00 -1.07198918e+00
1.54084504e+00 -2.24329782e+00 -1.21956944e-01 1.00714527e-01
-4.02747653e-02 9.28373456e-01 1.07921153e-01 7.46290758e-03
-4.16025221e-02 3.60040516e-01 -2.96490073e-01 -5.25432527e-01
-2.04473175e-02 -5.06650388e-01 -4.13849682e-01 8.04261148e-01
3.62604946e-01 4.35145795e-01 -6.01655543e-01 -1.73810914e-01
3.20684075e-01 1.03555405e+00 -7.18991160e-01 4.27692056e-01
3.20863128e-01 1.80626914e-01 -7.60544017e-02 6.77966774e-01
8.97347271e-01 4.05620664e-01 -1.80294052e-01 -1.18704587e-01
2.32136011e-01 7.64508307e-01 -9.07702804e-01 1.27758372e+00
-4.97950554e-01 9.67895567e-01 5.37044704e-01 -9.45108414e-01
1.10648882e+00 5.27766645e-01 7.74761960e-02 -3.69373709e-01
5.82561493e-01 3.30122977e-01 2.15863198e-01 -4.40055639e-01
3.14238012e-01 -2.03757539e-01 9.04033855e-02 3.29377353e-02
3.48688781e-01 1.34397252e-02 -4.64368582e-01 -1.32503927e-01
9.44249272e-01 -5.48526585e-01 -8.44020173e-02 -2.31249213e-01
1.01753724e+00 -7.94826329e-01 4.78979349e-01 6.52393341e-01
-7.64414549e-01 2.47599229e-01 -1.35346889e-01 -2.47880369e-01
-7.01620102e-01 -8.08149636e-01 -4.33272362e-01 9.17215884e-01
3.06958973e-01 -1.15049094e-01 -1.07810342e+00 -6.87475443e-01
-1.18507572e-01 6.99718356e-01 -3.23190987e-01 -5.88259220e-01
-6.67014360e-01 -8.35855782e-01 1.22757280e+00 5.49398512e-02
7.90648699e-01 -1.11504602e+00 -4.00474429e-01 3.07005167e-01
1.15354611e-02 -1.22628951e+00 -5.29356241e-01 3.88278097e-01
-4.77720857e-01 -6.64674461e-01 -4.88897860e-01 -8.91255915e-01
2.50224739e-01 1.55054331e-01 4.26828086e-01 7.48240948e-02
-6.32844344e-02 -3.05985749e-01 -1.51794285e-01 -5.89067101e-01
-6.96172893e-01 2.87337989e-01 5.04859149e-01 4.61806446e-01
4.58529294e-01 -4.72762913e-01 -5.04515707e-01 1.92617357e-01
-7.72185087e-01 -4.14967537e-01 3.21052432e-01 9.10168827e-01
-2.17894435e-01 3.10922772e-01 7.39531577e-01 -2.44795784e-01
9.61685300e-01 -3.66359591e-01 -4.30267334e-01 1.11289367e-01
-7.09226072e-01 1.16881900e-01 6.93829358e-01 -6.52362943e-01
-9.97120023e-01 -2.71011025e-01 -5.12895226e-01 -3.55048269e-01
-2.92957515e-01 9.67869908e-02 -5.78846395e-01 -3.98011953e-01
6.01832271e-01 3.87168169e-01 -1.12079531e-01 -4.67992872e-01
1.45234436e-01 1.25589097e+00 6.05799854e-01 5.20634986e-02
1.02141082e+00 2.27471173e-01 -3.90845954e-01 -1.13163650e+00
-4.12872046e-01 -3.63851011e-01 -1.40996799e-01 -5.89471199e-02
6.96492612e-01 -6.71015024e-01 -9.00841415e-01 9.46372151e-01
-1.35057342e+00 2.18832478e-01 5.16934037e-01 6.02450550e-01
-7.71751702e-02 5.29012978e-01 -7.76287079e-01 -1.17684078e+00
-8.04995656e-01 -1.47092628e+00 6.87098444e-01 3.31530094e-01
-2.39681527e-02 -8.10022295e-01 -5.70341311e-02 5.42685509e-01
8.35876167e-01 -1.81437254e-01 5.16571760e-01 -1.04548025e+00
-1.43929034e-01 -4.67671156e-01 -5.66967465e-02 7.65142500e-01
1.27080530e-01 -3.23311925e-01 -1.56940007e+00 -6.10921621e-01
4.57369030e-01 -1.43233743e-02 1.02135980e+00 1.69456407e-01
1.03963447e+00 -6.46731317e-01 -1.86202303e-01 1.01338947e+00
9.08099949e-01 4.58011985e-01 5.01204133e-01 2.12657943e-01
6.97473466e-01 4.94134039e-01 1.50214238e-02 2.10713327e-01
-1.04465835e-01 7.57170737e-01 6.65367007e-01 1.33598387e-01
-1.57537788e-01 -3.28663081e-01 7.02831566e-01 9.02584493e-01
4.41708982e-01 -3.87048125e-01 -6.51165307e-01 4.12246674e-01
-1.01701260e+00 -1.00905275e+00 5.72255373e-01 2.36679959e+00
5.89787960e-01 4.84189451e-01 -1.38007447e-01 6.56253874e-01
1.18154955e+00 6.05982363e-01 -4.79882360e-01 -7.78710485e-01
-2.26697087e-01 2.17502534e-01 4.97998863e-01 8.43366981e-01
-1.54648662e+00 1.13238418e+00 5.62658644e+00 9.39995289e-01
-1.64145744e+00 1.42007589e-01 5.68521798e-01 9.01298039e-03
-1.18599325e-01 -4.73777950e-01 -9.68181729e-01 5.75770557e-01
1.27101457e+00 1.50378779e-01 8.37075293e-01 5.70469916e-01
-3.25133884e-03 9.41280305e-01 -5.03976226e-01 1.03961194e+00
4.68147308e-01 -1.02824378e+00 -7.08075389e-02 2.01330766e-01
3.53312284e-01 -3.67126353e-02 4.83744860e-01 1.94077358e-01
2.09183805e-03 -1.25284302e+00 8.34182203e-01 -2.12619618e-01
6.38426542e-01 -1.00311720e+00 7.63110638e-01 1.09901480e-01
-9.81667578e-01 -4.73041952e-01 -1.88829452e-01 2.68165886e-01
2.65556455e-01 4.01120067e-01 -1.16467333e+00 3.36606741e-01
4.55061913e-01 1.13376200e-01 -1.76884681e-01 9.66462433e-01
-3.62095982e-01 9.39424813e-01 -3.03193271e-01 -2.52372533e-01
5.81058599e-02 5.35580695e-01 8.81529272e-01 1.48767447e+00
2.24276587e-01 -3.81137133e-01 -3.10846508e-01 5.55823863e-01
-3.24964762e-01 5.44194393e-02 -3.80230635e-01 -1.76206678e-01
8.73287797e-01 9.64564502e-01 -2.97890604e-01 -8.89647868e-04
6.83309212e-02 1.10195935e+00 2.93341011e-01 4.35867816e-01
-8.95799279e-01 -9.26466644e-01 1.03371274e+00 -2.09637076e-01
3.35057706e-01 -8.44145864e-02 -4.56186742e-01 -1.09322977e+00
-5.03873900e-02 -1.25899172e+00 1.56019688e-01 -2.35443749e-02
-8.29005420e-01 9.93926704e-01 -6.67877495e-01 -8.71552110e-01
3.53999473e-02 -7.38597512e-01 -6.07183158e-01 1.00271010e+00
-1.70747578e+00 -9.00730908e-01 1.35477915e-01 4.17011231e-01
5.72143555e-01 -5.38996637e-01 8.63021195e-01 4.10561562e-01
-9.53884065e-01 1.30604720e+00 -1.64756402e-01 3.01282257e-01
4.54919219e-01 -6.74475610e-01 7.14154363e-01 1.16031384e+00
3.33647788e-01 7.53287911e-01 8.19768667e-01 -2.56516308e-01
-1.36385226e+00 -8.03198040e-01 9.40433204e-01 -1.86007172e-01
4.95164961e-01 -4.43191528e-01 -9.25893784e-01 1.22504883e-01
2.48075128e-01 1.15787581e-01 6.11486733e-01 -7.63161108e-02
-7.62120485e-01 -3.05550337e-01 -1.50714731e+00 4.41046506e-01
7.09628284e-01 -9.97847080e-01 -5.02027750e-01 -7.66710415e-02
9.14378107e-01 -2.18720838e-01 -4.67143685e-01 4.22948033e-01
7.15199232e-01 -8.95044267e-01 1.11264598e+00 -6.95982635e-01
-1.50774777e-01 -1.19704902e-01 -3.16297740e-01 -1.23205674e+00
-2.66759992e-01 -1.07156074e+00 -5.02776980e-01 1.24252856e+00
5.88047445e-01 -8.26887310e-01 5.81177354e-01 6.10218309e-02
-1.74780309e-01 -5.79607725e-01 -1.34641826e+00 -6.00012898e-01
3.50120366e-02 -5.85982919e-01 7.76415169e-01 9.64785099e-01
2.12999701e-01 1.26918584e-01 -5.19071639e-01 8.66062582e-01
6.97160065e-01 -6.10958040e-01 2.80962884e-01 -9.56483424e-01
-2.62861311e-01 -8.09563160e-01 -6.03644192e-01 -1.02608287e+00
3.81667733e-01 -8.42660010e-01 -1.95003394e-02 -6.58823431e-01
-2.92490035e-01 -1.71175182e-01 -9.42813039e-01 1.74871068e-02
-4.51308757e-01 2.40145534e-01 1.90864161e-01 -1.44335702e-01
-3.26056629e-02 5.44954121e-01 6.79593444e-01 -4.14769232e-01
-1.38097823e-01 1.87577188e-01 -6.25393331e-01 6.17977262e-01
1.02316034e+00 -5.54716527e-01 -8.34220499e-02 -2.94638246e-01
-3.74831796e-01 -1.13977201e-01 9.17830020e-02 -1.28563499e+00
3.00326735e-01 1.55780256e-01 1.27202809e-01 -1.49011284e-01
5.41401982e-01 -8.21019173e-01 -3.86146814e-01 7.58328795e-01
-5.75368285e-01 -2.53723502e-01 2.81479239e-01 3.27134967e-01
-2.84930944e-01 -1.81834415e-01 1.08437288e+00 4.51920956e-01
-1.95991263e-01 2.66962409e-01 -2.14210212e-01 -3.33781689e-01
6.11287653e-01 -4.58641797e-02 -3.52107346e-01 -3.80134076e-01
-2.91163027e-01 -3.05986851e-01 2.00377349e-02 6.96113169e-01
7.38636494e-01 -1.13553596e+00 -7.13994920e-01 8.36330354e-01
-2.82135606e-01 -7.70229518e-01 2.24775031e-01 3.19832712e-01
-3.91113341e-01 6.91069424e-01 -1.04633607e-01 -2.29446888e-01
-1.43029082e+00 6.79952681e-01 4.69700485e-01 -1.25663385e-01
-1.81505933e-01 1.36347687e+00 -2.50773191e-01 -3.66077602e-01
7.25597799e-01 4.17967401e-02 -1.55507356e-01 -6.21567890e-02
9.76330400e-01 3.28949779e-01 3.74410272e-01 -1.02094758e+00
-6.68264031e-01 7.31762424e-02 -7.35130832e-02 -2.73098081e-01
9.76068437e-01 -1.22106411e-02 3.91369998e-01 -1.34701431e-01
1.53175163e+00 2.50920564e-01 -1.02482581e+00 -9.85204577e-02
-1.44167438e-01 -4.38985318e-01 4.94375169e-01 -7.51380146e-01
-1.03874338e+00 1.27176702e+00 9.60729003e-01 5.33550262e-01
9.75108802e-01 -3.64783227e-01 8.99870396e-01 1.17544331e-01
-4.31134272e-03 -8.08847189e-01 -3.75602469e-02 5.00947118e-01
8.61496806e-01 -1.29156291e+00 -4.11221266e-01 -2.27073237e-01
-3.30626339e-01 8.50425899e-01 2.80516058e-01 1.82306394e-01
7.41859496e-01 3.55742484e-01 2.64677882e-01 1.14148825e-01
-2.79239863e-01 2.60874093e-01 3.19618195e-01 6.98718786e-01
3.22183222e-01 1.25070468e-01 -3.39417495e-02 5.24553001e-01
-5.79646945e-01 -6.42094016e-01 -5.32991365e-02 5.37091315e-01
-6.09824538e-01 -9.22763050e-01 -7.19494402e-01 5.16542979e-02
-1.02889454e+00 -1.92297533e-01 -3.03693265e-01 4.93374933e-03
4.82834876e-02 1.37754917e+00 -1.48933008e-01 -9.28377450e-01
1.25767827e-01 2.25958958e-01 4.70015705e-02 -2.04994678e-01
-7.35175669e-01 -1.68382302e-01 1.84051730e-02 -1.78459153e-01
5.04688062e-02 -4.03921962e-01 -7.38822222e-01 -2.80872107e-01
-9.85174656e-01 1.02734499e-01 1.24163866e+00 7.64901876e-01
3.38533282e-01 5.59324086e-01 1.11606395e+00 -7.55133808e-01
-9.98612940e-01 -1.14327717e+00 -2.98605531e-01 4.19460475e-01
9.83509660e-01 -6.00281417e-01 -6.49438024e-01 -3.53678524e-01] | [14.079228401184082, 5.86139440536499] |
17d1c687-cfeb-4a13-b047-1f02ddbb7e41 | auto-focus-contrastive-learning-for-image | 2211.10922 | null | https://arxiv.org/abs/2211.10922v1 | https://arxiv.org/pdf/2211.10922v1.pdf | Auto-Focus Contrastive Learning for Image Manipulation Detection | Generally, current image manipulation detection models are simply built on manipulation traces. However, we argue that those models achieve sub-optimal detection performance as it tends to: 1) distinguish the manipulation traces from a lot of noisy information within the entire image, and 2) ignore the trace relations among the pixels of each manipulated region and its surroundings. To overcome these limitations, we propose an Auto-Focus Contrastive Learning (AF-CL) network for image manipulation detection. It contains two main ideas, i.e., multi-scale view generation (MSVG) and trace relation modeling (TRM). Specifically, MSVG aims to generate a pair of views, each of which contains the manipulated region and its surroundings at a different scale, while TRM plays a role in modeling the trace relations among the pixels of each manipulated region and its surroundings for learning the discriminative representation. After learning the AF-CL network by minimizing the distance between the representations of corresponding views, the learned network is able to automatically focus on the manipulated region and its surroundings and sufficiently explore their trace relations for accurate manipulation detection. Extensive experiments demonstrate that, compared to the state-of-the-arts, AF-CL provides significant performance improvements, i.e., up to 2.5%, 7.5%, and 0.8% F1 score, on CAISA, NIST, and Coverage datasets, respectively. | ['Q. M. Jonathan Wu', 'Hongyang Yan', 'Teng Huang', 'Guangcan Liu', 'Zhili Zhou', 'Wenyan Pan'] | 2022-11-20 | null | null | null | null | ['image-manipulation-detection', 'image-manipulation'] | ['computer-vision', 'computer-vision'] | [ 5.24195611e-01 -2.48372108e-01 -2.20614746e-01 2.81385016e-02
-6.31523550e-01 -3.47486079e-01 4.83353406e-01 -4.78540882e-02
1.14337452e-01 6.67171255e-02 -9.00403038e-02 1.03176214e-01
-2.00749725e-01 -7.99481153e-01 -8.95338833e-01 -7.72420228e-01
7.98495784e-02 1.82088949e-02 2.91325897e-01 -1.96055114e-01
2.84949154e-01 4.50039625e-01 -1.67895103e+00 4.21267658e-01
7.36468971e-01 1.23319697e+00 5.65357685e-01 3.17287475e-01
2.22784415e-01 8.45804453e-01 -7.57342637e-01 -8.14273059e-02
2.99328417e-01 -4.22058254e-01 -1.37392178e-01 4.20336664e-01
7.36724496e-01 -5.29463470e-01 -7.72243321e-01 1.32265961e+00
2.74886072e-01 6.18034340e-02 6.21744752e-01 -1.26009393e+00
-6.82220876e-01 4.74016070e-01 -1.15620017e+00 2.31787473e-01
3.12859237e-01 2.70500302e-01 6.82324171e-01 -1.01400554e+00
7.19717324e-01 1.32822931e+00 5.85664287e-02 4.03469026e-01
-1.19099712e+00 -9.54107523e-01 3.59173000e-01 3.81493896e-01
-1.49576592e+00 -2.94063270e-01 1.07944572e+00 -4.09142911e-01
5.53472757e-01 3.20766211e-01 4.06652600e-01 1.33168268e+00
2.88375854e-01 1.25454223e+00 9.80280459e-01 -4.17779118e-01
-3.71610336e-02 1.31432876e-01 8.97337720e-02 8.88678968e-01
9.99211669e-02 1.82745919e-01 -6.73678637e-01 5.08881286e-02
7.66518116e-01 6.41409401e-03 -6.45411193e-01 -7.02321410e-01
-1.10259724e+00 4.44452047e-01 3.83931696e-01 2.50292569e-01
-4.07456279e-01 -2.79048458e-02 4.03052002e-01 2.48377472e-02
8.97578299e-02 2.61912584e-01 -3.37900370e-02 1.95453271e-01
-8.05564344e-01 7.35030994e-02 2.98859060e-01 1.30758154e+00
7.08497941e-01 -1.19336285e-02 -3.85269821e-01 5.68888485e-01
1.28111914e-01 5.90718627e-01 8.64687413e-02 -8.08966815e-01
8.51663232e-01 8.21375430e-01 1.34453416e-01 -1.48883104e+00
-3.19620967e-02 -3.43998700e-01 -8.74918520e-01 1.51742831e-01
3.32870334e-01 2.08165273e-01 -9.01557803e-01 1.66322982e+00
2.60563165e-01 4.95379791e-02 -2.41222128e-01 8.37442577e-01
5.55675268e-01 5.96001804e-01 -1.14284568e-01 -2.36067325e-01
1.31260109e+00 -9.48858142e-01 -7.12862849e-01 -4.80254143e-01
1.71725780e-01 -8.05272102e-01 1.07839203e+00 6.03176236e-01
-1.08825898e+00 -9.07971144e-01 -1.27785790e+00 2.35663041e-01
-2.36732244e-01 6.46685421e-01 3.97512704e-01 2.84006506e-01
-5.46726108e-01 3.28207344e-01 -5.99768519e-01 -5.20145781e-02
4.52157766e-01 2.86853164e-02 -3.29807013e-01 -2.97419667e-01
-8.57346594e-01 6.29255831e-01 5.08266389e-01 2.00233191e-01
-1.31642365e+00 -5.32321453e-01 -8.73578191e-01 1.61047965e-01
1.01870799e+00 -2.80733615e-01 6.31778479e-01 -7.67080545e-01
-9.41457450e-01 6.83889627e-01 -2.18189165e-01 -4.04161848e-02
5.38566113e-01 -3.76608819e-01 -4.44931418e-01 2.68129498e-01
2.10285142e-01 3.83467972e-01 1.08677173e+00 -1.78816748e+00
-7.47277975e-01 -5.19933462e-01 1.03626050e-01 3.76154147e-02
-3.55470181e-01 -1.00111654e-02 -1.17288661e+00 -7.46428013e-01
2.81278878e-01 -8.36084366e-01 8.39796290e-02 9.49202925e-02
-6.76215172e-01 -1.29401028e-01 1.14311862e+00 -8.17826748e-01
1.38250816e+00 -2.30571008e+00 2.77896404e-01 2.37213731e-01
2.69204497e-01 4.95569676e-01 -3.05286407e-01 3.19464982e-01
8.64646427e-05 -1.11729436e-01 4.73494157e-02 -4.98360917e-02
-2.70094633e-01 -1.37105495e-01 -4.27028090e-01 6.53063178e-01
9.74128544e-02 8.11455429e-01 -1.07798135e+00 -2.33568877e-01
5.14243782e-01 3.97278428e-01 -1.54278651e-01 1.91237852e-01
-2.76523858e-01 3.11295152e-01 -4.05864358e-01 7.76283979e-01
7.90143430e-01 -1.71165451e-01 1.08366802e-01 -6.67985976e-01
6.25161547e-03 -8.74293298e-02 -1.23657775e+00 1.61434770e+00
-2.86922276e-01 5.14339149e-01 7.90640935e-02 -6.58322752e-01
8.37785661e-01 3.20398696e-02 4.98864263e-01 -9.55131590e-01
2.03699470e-01 -1.05432235e-02 -7.98444450e-02 -6.73323989e-01
3.93847793e-01 6.36121094e-01 -1.78301614e-02 3.66844475e-01
-1.53650073e-02 3.57157528e-01 3.83759737e-01 2.98491716e-01
8.78688812e-01 2.18184918e-01 1.50229499e-01 1.85381055e-01
7.02877760e-01 -3.33730549e-01 4.90236104e-01 9.80464339e-01
-3.59651446e-01 5.56838155e-01 5.86813629e-01 -1.78357452e-01
-8.07042658e-01 -1.16681015e+00 1.86960921e-01 8.97387624e-01
9.10814047e-01 -5.24412036e-01 -6.51994765e-01 -9.02403057e-01
-1.56006396e-01 8.35305929e-01 -6.50491416e-01 -4.65529263e-01
-8.47219288e-01 -2.40230605e-01 2.24894196e-01 6.22833669e-01
7.45546401e-01 -1.01586092e+00 -6.15562677e-01 -6.92645237e-02
-3.21816772e-01 -1.14985037e+00 -5.97954392e-01 -1.57539174e-01
-3.62386018e-01 -1.35281277e+00 -3.84959042e-01 -6.46128595e-01
7.35282838e-01 8.61437798e-01 8.14008594e-01 2.33362783e-02
-4.72406894e-01 1.97962508e-01 -2.53909111e-01 -2.58668661e-01
-3.38051379e-01 -3.40644687e-01 -2.45738700e-01 2.29239479e-01
2.85530478e-01 -2.91305870e-01 -7.04739630e-01 5.90448141e-01
-7.99954712e-01 4.74415496e-02 7.95416296e-01 9.29474652e-01
7.34033942e-01 4.00678217e-01 2.48838395e-01 -8.00431550e-01
2.90573418e-01 -3.07691902e-01 -6.12569511e-01 6.71130717e-01
-5.45552492e-01 -2.32455730e-01 3.06754559e-01 -5.82672656e-01
-1.11416340e+00 1.27474368e-01 5.21816432e-01 -1.10036540e+00
-1.51517779e-01 8.46826434e-02 -4.62626845e-01 -6.49546459e-02
6.60244703e-01 6.06662035e-01 -5.42911775e-02 -1.81714609e-01
4.57044393e-01 5.32780349e-01 5.76130986e-01 -3.31010342e-01
9.61682856e-01 6.24906480e-01 -1.82507411e-01 -5.32748044e-01
-9.01431978e-01 -7.60201812e-01 -5.21428287e-01 -6.12590134e-01
7.87211955e-01 -8.59086514e-01 -5.45662820e-01 5.53018093e-01
-1.08217084e+00 -3.72714177e-02 -7.54847154e-02 2.43699551e-01
-4.80404377e-01 6.02461934e-01 -5.03982425e-01 -8.97101820e-01
-1.69032365e-01 -1.31345546e+00 1.39903748e+00 2.26856649e-01
1.10529400e-01 -4.08366323e-01 -6.58535719e-01 5.97537220e-01
1.30400404e-01 3.16502243e-01 1.03251553e+00 -2.19094887e-01
-8.93508673e-01 -4.26358163e-01 -4.89798933e-01 3.29964161e-01
2.60549545e-01 -4.58725020e-02 -9.79713440e-01 -4.19444174e-01
1.59818932e-01 -2.32594356e-01 8.85915816e-01 1.14715829e-01
1.35790753e+00 -5.86179197e-02 -6.12383544e-01 3.67278904e-01
1.36001158e+00 5.04045784e-01 8.63783896e-01 1.31524682e-01
8.27826381e-01 5.95040500e-01 1.02096415e+00 3.78531963e-01
5.04726395e-02 8.83400857e-01 7.67765760e-01 6.70343414e-02
-3.19378644e-01 -3.42610240e-01 4.41000491e-01 4.62125272e-01
9.30570289e-02 -4.47558492e-01 -6.88210070e-01 4.52475131e-01
-1.92914104e+00 -9.54140604e-01 -6.15666248e-02 2.35827851e+00
3.23552638e-01 2.47060716e-01 -4.15136851e-02 1.82391763e-01
9.80323911e-01 4.95696008e-01 -7.61927009e-01 2.56038576e-01
-1.10525213e-01 -2.73142569e-02 4.70044583e-01 1.00121595e-01
-1.36053026e+00 7.99684405e-01 5.22147226e+00 1.19954705e+00
-1.03979015e+00 -1.12834655e-01 5.04031360e-01 -7.59310201e-02
6.44615814e-02 -2.09240749e-01 -8.55571568e-01 5.51781535e-01
1.91586852e-01 9.20883790e-02 5.56234300e-01 9.15453911e-01
1.68158978e-01 -2.58372307e-01 -1.02692366e+00 1.01301801e+00
6.26561165e-01 -9.24358010e-01 3.18507969e-01 2.03534495e-03
7.26150572e-01 -4.86891240e-01 1.94272682e-01 4.02881145e-01
-2.38839388e-01 -7.84774363e-01 8.97210777e-01 4.67087120e-01
6.48419321e-01 -8.78907740e-01 5.24217725e-01 6.08389318e-01
-1.36927652e+00 -3.11940283e-01 -2.15824693e-01 4.01881635e-01
7.34770522e-02 4.86662805e-01 -4.41562444e-01 7.40689993e-01
6.36446238e-01 5.26251376e-01 -5.36171019e-01 8.35605025e-01
-6.00461602e-01 2.95194983e-01 1.56336263e-01 2.80409157e-01
5.65308779e-02 -4.23675954e-01 7.59668350e-01 8.07344675e-01
1.34107858e-01 -2.44549543e-01 4.30451423e-01 1.18922842e+00
-7.20109716e-02 -1.59666076e-01 -5.39967477e-01 1.67565256e-01
6.41321421e-01 1.05181360e+00 -5.10876954e-01 -3.64091247e-01
-3.60563904e-01 1.06476521e+00 3.50063205e-01 4.63054210e-01
-1.08964324e+00 -4.36102659e-01 3.59929323e-01 2.05729127e-01
4.85564321e-01 -3.75436284e-02 2.17891037e-02 -9.48246062e-01
4.86436605e-01 -1.14288235e+00 2.86955327e-01 -7.72972822e-01
-1.12309146e+00 3.57247233e-01 4.77522053e-03 -1.44755328e+00
-3.69037911e-02 -5.59249282e-01 -4.09109473e-01 8.39713275e-01
-1.28415251e+00 -1.25200450e+00 -7.34861493e-01 4.28083926e-01
7.95099437e-01 -7.50356689e-02 5.43336630e-01 3.89151752e-01
-7.38731325e-01 7.06978858e-01 -1.43893644e-01 3.09804231e-01
6.11306727e-01 -9.62787509e-01 1.89295724e-01 9.80159283e-01
2.81851858e-01 5.32002389e-01 3.20576847e-01 -6.74869299e-01
-1.63804960e+00 -1.34669387e+00 3.83992761e-01 -2.85115868e-01
4.01163816e-01 -5.10330141e-01 -9.10310328e-01 5.71206748e-01
3.87549698e-02 6.12752698e-03 8.10181722e-02 -1.74528018e-01
-5.87001920e-01 -1.61667705e-01 -8.50898445e-01 6.91802263e-01
1.14503026e+00 -6.85987234e-01 -3.31413716e-01 3.52745682e-01
2.44936720e-01 -6.66940451e-01 -5.54454267e-01 7.58850873e-01
6.02652192e-01 -1.19246066e+00 1.08669877e+00 -1.87560737e-01
4.67937231e-01 -5.09855032e-01 -1.37382835e-01 -1.01855218e+00
-3.86203796e-01 -4.23252016e-01 -6.20574951e-01 1.23067391e+00
-5.71971666e-03 -1.39702767e-01 6.15540564e-01 6.76422343e-02
5.09485528e-02 -9.27527308e-01 -5.45019865e-01 -8.87917399e-01
-5.51662564e-01 -3.81431282e-01 2.30619878e-01 7.41442978e-01
-3.96487951e-01 2.12299526e-01 -6.07641459e-01 4.64126259e-01
7.45187700e-01 4.02409166e-01 9.07603681e-01 -6.96487546e-01
-3.45232338e-01 -4.05692905e-01 -5.31392515e-01 -1.36013651e+00
5.56249693e-02 -4.43219900e-01 1.40120685e-01 -1.30067945e+00
5.11404514e-01 -3.56734365e-01 -3.19288492e-01 3.12526941e-01
-4.54832047e-01 1.06778592e-01 3.16008061e-01 2.35045388e-01
-5.90662360e-01 3.43270063e-01 1.44907880e+00 -5.99358499e-01
-2.42764503e-01 -8.36203322e-02 -4.98588115e-01 7.27109253e-01
5.61522841e-01 -3.24473411e-01 -5.95368803e-01 -3.77334386e-01
-1.89673275e-01 1.42992571e-01 4.54877377e-01 -1.04417741e+00
2.06588954e-01 -1.98038518e-01 6.10371649e-01 -9.23670471e-01
3.93170685e-01 -6.60222232e-01 8.98309276e-02 4.51464474e-01
-2.73939043e-01 -4.68779773e-01 1.01338677e-01 9.45517957e-01
-2.08694920e-01 -2.61025757e-01 7.31596351e-01 -3.11881244e-01
-8.87133777e-01 2.09291399e-01 -1.85983017e-01 -2.44654953e-01
1.36268330e+00 -2.17634305e-01 -4.93753612e-01 -3.46309096e-01
-4.39994663e-01 3.10031891e-01 3.07724506e-01 8.17905843e-01
8.30373943e-01 -1.23191369e+00 -4.85873759e-01 4.55084175e-01
5.20080209e-01 9.10344198e-02 4.90130037e-01 7.45118618e-01
-1.63651451e-01 2.18984440e-01 7.19904229e-02 -7.65966415e-01
-1.38826549e+00 8.53359640e-01 1.49535701e-01 -2.89103299e-01
-6.31912172e-01 7.72063553e-01 6.67812824e-01 7.22085014e-02
5.70784390e-01 -2.57290825e-02 -2.72361517e-01 5.07001691e-02
6.99347615e-01 5.50714254e-01 -1.35176346e-01 -6.51340961e-01
-1.54914781e-01 5.82290828e-01 -3.28634143e-01 3.17971587e-01
8.83592665e-01 2.09519062e-02 4.71703820e-02 2.43229091e-01
1.08641326e+00 1.97134018e-01 -1.33516157e+00 -3.01885188e-01
-1.94165885e-01 -8.42746437e-01 7.93987885e-02 -8.02430153e-01
-1.41851413e+00 7.40844488e-01 6.46627069e-01 -7.89731517e-02
1.31793797e+00 2.91229449e-02 6.76713467e-01 1.57836303e-01
6.41280830e-01 -9.18413222e-01 5.77291727e-01 2.21033052e-01
1.07016146e+00 -1.02632868e+00 1.57034293e-01 -8.57508242e-01
-5.04522324e-01 1.00063968e+00 1.08536983e+00 -1.28792137e-01
1.98465779e-01 2.35468611e-01 -1.51998833e-01 -5.31877816e-01
-3.77552480e-01 -1.48491398e-01 4.17872339e-01 5.36112845e-01
-9.65984464e-02 -5.13339154e-02 2.33440191e-01 2.62152225e-01
3.99674416e-01 -2.80193001e-01 1.02924801e-01 9.36227858e-01
-3.14365387e-01 -7.85762012e-01 -4.47503388e-01 5.38886249e-01
2.46304255e-02 1.43761665e-01 -5.28318346e-01 9.40215111e-01
4.62735176e-01 9.84864235e-01 -6.84335753e-02 -7.40674555e-01
5.68729222e-01 -2.76188016e-01 6.43536448e-01 -5.09630382e-01
-3.56645852e-01 2.47471839e-01 -2.29756922e-01 -8.90721917e-01
-3.57841194e-01 -5.82958341e-01 -8.36557388e-01 4.67451476e-02
-9.04553294e-01 -3.62223923e-01 2.91314870e-01 7.36773491e-01
3.96241784e-01 7.64686584e-01 8.83359849e-01 -8.42159390e-01
-7.30402052e-01 -8.88469458e-01 -7.11667061e-01 6.37272835e-01
9.84370783e-02 -8.20176661e-01 -4.10129875e-01 -5.20887375e-02] | [12.134040832519531, 0.9041734933853149] |
9513fb18-d3f8-4a57-9a58-39ecb2b0f0be | invertible-rescaling-network-and-its | 2210.04188 | null | https://arxiv.org/abs/2210.04188v1 | https://arxiv.org/pdf/2210.04188v1.pdf | Invertible Rescaling Network and Its Extensions | Image rescaling is a commonly used bidirectional operation, which first downscales high-resolution images to fit various display screens or to be storage- and bandwidth-friendly, and afterward upscales the corresponding low-resolution images to recover the original resolution or the details in the zoom-in images. However, the non-injective downscaling mapping discards high-frequency contents, leading to the ill-posed problem for the inverse restoration task. This can be abstracted as a general image degradation-restoration problem with information loss. In this work, we propose a novel invertible framework to handle this general problem, which models the bidirectional degradation and restoration from a new perspective, i.e. invertible bijective transformation. The invertibility enables the framework to model the information loss of pre-degradation in the form of distribution, which could mitigate the ill-posed problem during post-restoration. To be specific, we develop invertible models to generate valid degraded images and meanwhile transform the distribution of lost contents to the fixed distribution of a latent variable during the forward degradation. Then restoration is made tractable by applying the inverse transformation on the generated degraded image together with a randomly-drawn latent variable. We start from image rescaling and instantiate the model as Invertible Rescaling Network (IRN), which can be easily extended to the similar decolorization-colorization task. We further propose to combine the invertible framework with existing degradation methods such as image compression for wider applications. Experimental results demonstrate the significant improvement of our model over existing methods in terms of both quantitative and qualitative evaluations of upscaling and colorizing reconstruction from downscaled and decolorized images, and rate-distortion of image compression. | ['Tie-Yan Liu', 'Zhouchen Lin', 'Chang Liu', 'Shuxin Zheng', 'Mingqing Xiao'] | 2022-10-09 | null | null | null | null | ['colorization'] | ['computer-vision'] | [ 8.70577514e-01 -2.51133144e-01 6.07226305e-02 -2.26649955e-01
-4.62277830e-01 -3.97385389e-01 4.09050375e-01 -6.68866336e-01
-1.68014586e-01 6.85681343e-01 5.43543160e-01 -2.30004132e-01
-7.82057792e-02 -9.51506436e-01 -7.90954053e-01 -9.49065387e-01
3.74045759e-01 -2.73076087e-01 -1.04841582e-01 -2.23101363e-01
1.37534738e-01 4.82212871e-01 -1.57331407e+00 3.77335757e-01
1.08138800e+00 7.71586001e-01 4.92143452e-01 6.33089960e-01
1.13648817e-01 6.25060797e-01 -5.59907615e-01 -1.75959229e-01
3.93583566e-01 -7.13016570e-01 -3.65220726e-01 3.70386899e-01
3.03572685e-01 -7.33906746e-01 -5.86026490e-01 1.42967701e+00
4.93597627e-01 3.46276425e-02 3.50726902e-01 -1.21730900e+00
-1.48705816e+00 1.61385760e-01 -7.11360991e-01 1.00352950e-01
2.90023148e-01 4.49720882e-02 4.60650831e-01 -7.85655379e-01
6.75804675e-01 1.55168986e+00 3.87716323e-01 5.38789868e-01
-1.47140694e+00 -6.14408731e-01 3.00478451e-02 3.21612537e-01
-1.20227659e+00 -5.14059603e-01 7.08035827e-01 -1.90417990e-01
3.72604817e-01 5.16149700e-01 4.49188501e-01 9.97609138e-01
1.42232761e-01 4.23899084e-01 1.38152194e+00 -4.40034121e-01
6.64394572e-02 4.62357915e-04 -3.18882257e-01 1.92422807e-01
3.42558205e-01 2.89613038e-01 -4.19796646e-01 2.34277308e-01
1.08077884e+00 3.92745852e-01 -7.82136261e-01 -3.78766656e-01
-1.26812553e+00 4.48272228e-01 5.67520618e-01 8.16613808e-02
-2.70383865e-01 7.94597319e-04 1.23667389e-01 6.22304618e-01
5.55662692e-01 1.50896996e-01 -7.12224143e-03 1.29481614e-01
-8.78170252e-01 -1.01517215e-01 2.89106727e-01 9.39293504e-01
7.57152498e-01 1.69753104e-01 -2.81743765e-01 9.40144062e-01
2.10430026e-01 5.13344467e-01 5.21148145e-01 -1.08089459e+00
4.96094555e-01 3.09086710e-01 2.36767426e-01 -1.06629574e+00
5.48080243e-02 -2.92398840e-01 -1.44056010e+00 5.87528527e-01
1.30494282e-01 2.86820620e-01 -1.04001677e+00 1.98905492e+00
9.83776376e-02 2.05241412e-01 1.22095868e-01 1.26982784e+00
3.85353059e-01 9.03473496e-01 -1.73667759e-01 -6.32923901e-01
1.32146502e+00 -9.75248218e-01 -1.20762968e+00 -1.07041467e-02
-1.33830711e-01 -8.09863269e-01 1.66549599e+00 3.72644752e-01
-1.19382715e+00 -7.17127144e-01 -1.26820159e+00 -5.16489267e-01
-1.02211818e-01 9.03519690e-02 5.20942397e-02 5.64372241e-01
-1.24642432e+00 6.14171624e-01 -6.06586337e-01 -2.66652197e-01
3.40914458e-01 4.72122058e-03 -5.04808307e-01 -5.75207770e-01
-1.23201144e+00 7.72527516e-01 5.57950735e-02 2.87776351e-01
-8.21415782e-01 -6.79850519e-01 -7.27464676e-01 8.13837424e-02
2.52753884e-01 -9.50399041e-01 5.57440281e-01 -1.24048948e+00
-1.62335968e+00 5.69072664e-01 -1.50615692e-01 -1.23582467e-01
8.22874725e-01 -2.14373261e-01 -6.29237711e-01 8.40557516e-02
1.20486088e-01 3.63954246e-01 1.23893595e+00 -1.54108822e+00
-3.08961570e-01 -3.01233083e-01 1.18603650e-02 5.21709323e-01
-6.31759942e-01 -1.38534009e-01 -7.02123284e-01 -1.10169065e+00
3.66810888e-01 -6.53550267e-01 9.59011093e-02 4.29153323e-01
-4.68576759e-01 6.76033497e-01 1.10986817e+00 -1.02516699e+00
1.27189314e+00 -2.45982957e+00 4.60798979e-01 -1.68223694e-01
2.49323949e-01 -2.91096438e-02 -4.46038812e-01 3.70951474e-01
-3.97892892e-01 2.35683471e-01 -5.94824553e-01 -4.85939860e-01
-1.40975609e-01 4.08065408e-01 -6.22640729e-01 6.79170251e-01
3.46162841e-02 6.01618588e-01 -8.27072442e-01 -1.68562636e-01
3.02023321e-01 9.53557849e-01 -3.98796141e-01 3.64078164e-01
2.90760010e-01 5.63180268e-01 3.60177481e-03 5.67340136e-01
1.24859047e+00 3.60642970e-02 3.47780101e-02 -5.85497618e-01
-1.45340383e-01 -1.67320892e-01 -1.10486734e+00 1.66964746e+00
-5.75936437e-01 5.37744999e-01 2.42517188e-01 -7.87796795e-01
7.74107814e-01 -3.01210470e-02 8.63115937e-02 -8.80463064e-01
-2.55401820e-01 2.93185115e-02 -3.03991705e-01 -5.58025599e-01
6.09749436e-01 -4.34742659e-01 2.67610937e-01 3.91753525e-01
-3.75301123e-01 3.41596529e-02 -5.33528402e-02 1.46039233e-01
7.51651347e-01 2.98495680e-01 4.41162847e-02 -1.04167990e-01
6.00641131e-01 -4.18691397e-01 4.47309762e-01 5.80378830e-01
-4.18836065e-02 1.07123983e+00 4.38196301e-01 -1.76533297e-01
-1.36390746e+00 -1.14535940e+00 -5.05716875e-02 9.49914098e-01
5.79094827e-01 -9.99498814e-02 -8.54149461e-01 -2.40567982e-01
-2.52290845e-01 6.86560094e-01 -4.89533186e-01 -4.01147544e-01
-5.67305028e-01 -8.31425369e-01 2.37196282e-01 1.76061064e-01
8.76263916e-01 -9.71310735e-01 -1.99434191e-01 -1.12110108e-01
-5.38246453e-01 -8.19978118e-01 -7.22370982e-01 -1.16199590e-01
-8.15329850e-01 -7.87501395e-01 -1.04937589e+00 -6.58170819e-01
9.39550042e-01 6.61126077e-01 6.76038623e-01 1.09719969e-01
-1.98090494e-01 7.50619173e-02 -3.06708694e-01 3.37837875e-01
-5.35986960e-01 -4.80054677e-01 1.21111363e-01 4.22441989e-01
-2.41908684e-01 -7.95840859e-01 -9.66664493e-01 3.61889064e-01
-1.67050123e+00 4.60659474e-01 5.82729459e-01 1.08286643e+00
7.93994069e-01 4.88613218e-01 3.56058091e-01 -6.19408727e-01
8.64135385e-01 -3.01033139e-01 -3.05890352e-01 2.86835998e-01
-7.59942234e-01 1.05711855e-01 8.70305717e-01 -6.68401897e-01
-1.30596995e+00 -1.55916572e-01 5.02219908e-02 -7.09259808e-01
1.43092558e-01 2.63896793e-01 -6.42424226e-01 -9.75759998e-02
4.65429634e-01 5.76065540e-01 1.82328522e-01 -7.42958605e-01
7.75591075e-01 6.37418628e-01 9.60711837e-01 -2.82582641e-01
1.19247591e+00 8.31442773e-01 -7.96707869e-02 -5.33820510e-01
-4.38798815e-01 7.28568807e-02 -3.22003037e-01 -1.19282492e-01
6.69584274e-01 -8.91642332e-01 -4.61830378e-01 5.89240313e-01
-9.49036598e-01 -2.47619763e-01 -5.09107411e-01 1.83645695e-01
-5.99212706e-01 8.09755564e-01 -7.17585802e-01 -5.99407971e-01
-8.96110013e-02 -1.24163675e+00 9.10041809e-01 1.84233725e-01
3.72211903e-01 -6.32298589e-01 -1.85377911e-01 1.49595991e-01
6.72427773e-01 1.53846458e-01 1.11273944e+00 3.80402595e-01
-7.21696436e-01 1.38390794e-01 -6.12780571e-01 7.05219746e-01
3.16805840e-01 -3.84443104e-01 -7.66122520e-01 -6.13678873e-01
3.59546870e-01 1.22553855e-03 1.03444004e+00 3.39062959e-01
1.30154085e+00 -5.49885452e-01 4.41085733e-02 1.14713275e+00
1.65894699e+00 -7.55004808e-02 1.23566651e+00 6.33521557e-01
6.18313372e-01 5.20057738e-01 4.56262887e-01 3.11860025e-01
1.68288857e-01 6.04601562e-01 5.29170096e-01 -5.48474550e-01
-6.76818848e-01 -5.22266686e-01 4.83519852e-01 7.56310403e-01
-1.26661077e-01 -3.54098678e-01 -3.28118056e-01 2.84912109e-01
-1.67907202e+00 -9.16549206e-01 1.35708213e-01 2.41322684e+00
9.00768399e-01 -3.20839792e-01 -3.96843076e-01 9.48201418e-02
1.12631989e+00 2.53825992e-01 -7.86955178e-01 -9.26094055e-02
-5.56412041e-01 -1.85870960e-01 5.43013096e-01 6.88814223e-01
-7.58452773e-01 6.17250025e-01 6.18783045e+00 9.29874539e-01
-1.06811118e+00 2.05909967e-01 8.73381078e-01 -5.06122634e-02
-5.80686271e-01 6.30980283e-02 -2.34509394e-01 6.40366912e-01
5.85127652e-01 -2.31141001e-01 1.16805565e+00 4.16909009e-01
5.29465854e-01 3.74332666e-02 -8.72656643e-01 1.13657928e+00
2.27158278e-01 -1.04754031e+00 4.35690552e-01 1.21071376e-01
6.47129595e-01 -5.56522667e-01 4.99261171e-01 -9.50521603e-03
-1.13741923e-02 -9.73929644e-01 8.61973882e-01 7.06462383e-01
1.42021561e+00 -4.06137168e-01 3.70175630e-01 2.14946553e-01
-9.13515866e-01 -2.16802835e-01 -5.25675535e-01 5.62578999e-02
4.10367191e-01 5.12754917e-01 4.69665565e-02 5.36136389e-01
9.09507632e-01 8.09407949e-01 -4.47497547e-01 6.87392712e-01
-3.28468651e-01 1.80054262e-01 -7.45311752e-03 8.51132929e-01
-3.88755053e-01 -7.17248440e-01 5.35621405e-01 9.61295903e-01
7.60237515e-01 8.58959705e-02 -3.16715240e-01 1.06306016e+00
-1.59367532e-01 -1.55085310e-01 -3.78401905e-01 2.66062558e-01
3.64409477e-01 1.13477302e+00 -4.92837042e-01 -4.01588380e-01
-3.45238596e-01 1.59304249e+00 -5.83656579e-02 9.02646422e-01
-8.81880045e-01 -3.64988834e-01 6.32717311e-01 2.06050411e-01
1.93817779e-01 -1.82555954e-03 -2.95479596e-01 -1.50483978e+00
2.77116060e-01 -9.57516491e-01 5.69066629e-02 -1.22590804e+00
-1.34668767e+00 6.36291564e-01 -2.04807762e-02 -1.56248438e+00
2.04092860e-01 -2.64626235e-01 -3.40123028e-01 1.09561276e+00
-2.06217837e+00 -1.10413826e+00 -6.44123554e-01 8.76228869e-01
4.12793130e-01 1.00313313e-01 4.17786002e-01 5.89172542e-01
-6.34761274e-01 4.21432108e-01 4.78120148e-01 -3.69650960e-01
1.03525043e+00 -1.06514001e+00 1.55642614e-01 1.31806874e+00
-3.93035859e-01 6.21329606e-01 7.83439159e-01 -6.18247509e-01
-1.42786324e+00 -1.21658158e+00 4.89382386e-01 1.00143656e-01
4.19845372e-01 -3.25479299e-01 -1.11425066e+00 4.39036548e-01
1.96556106e-01 1.08434252e-01 2.28015736e-01 -6.28353000e-01
-4.73900884e-01 -2.39685744e-01 -1.24309516e+00 7.53843963e-01
1.26774859e+00 -7.31016397e-01 -2.80769527e-01 1.97645932e-01
9.95144665e-01 -3.16074938e-01 -7.14506567e-01 2.10862845e-01
5.21182358e-01 -1.12869358e+00 1.30817485e+00 -1.06254376e-01
7.92087197e-01 -6.83226407e-01 -4.05280650e-01 -1.26624680e+00
-5.45001924e-01 -7.89810240e-01 -3.20893973e-02 1.38743353e+00
-1.86627563e-02 -7.96408594e-01 2.16754988e-01 5.14915466e-01
-6.12623468e-02 -2.65426338e-01 -9.17159438e-01 -6.89185917e-01
-9.98176262e-02 -5.29311039e-02 7.12570131e-01 8.40284228e-01
-3.88491362e-01 -8.79821479e-02 -9.36387479e-01 4.01460350e-01
6.80923402e-01 1.26472324e-01 5.30666530e-01 -5.56065083e-01
-2.77748346e-01 -1.07819162e-01 -1.56836435e-01 -1.38347840e+00
-2.50800669e-01 -5.56894839e-01 7.99100921e-02 -1.47346234e+00
4.32178378e-01 -2.60544747e-01 -4.02849078e-01 2.08214313e-01
-1.05781458e-01 6.33575976e-01 3.04612219e-01 7.13054359e-01
-1.27770230e-01 7.64081657e-01 1.73230159e+00 -5.08684143e-02
-1.60803184e-01 -3.79509926e-01 -1.03153503e+00 3.26965213e-01
4.38566238e-01 -2.99603224e-01 -5.85297108e-01 -6.24139369e-01
2.16448188e-01 2.78121769e-01 5.63537896e-01 -7.25071490e-01
5.86998388e-02 -2.58895993e-01 4.12160724e-01 -1.71127871e-01
3.67091447e-01 -8.74563217e-01 4.59581286e-01 2.41040796e-01
-2.71764189e-01 -8.05169996e-03 -7.58836046e-02 8.60614181e-01
-3.84426057e-01 9.57508907e-02 9.39132988e-01 3.41866910e-02
-5.21768928e-01 1.90831885e-01 -1.46866292e-01 -3.62333328e-01
8.67875636e-01 -5.77966809e-01 -6.16709471e-01 -6.15103662e-01
-7.67578483e-01 -4.09299612e-01 9.94880795e-01 4.93063003e-01
7.78023303e-01 -1.50907373e+00 -6.68854773e-01 6.63017869e-01
-1.42448574e-01 -2.78058261e-01 7.55160570e-01 7.75005102e-01
-5.75423419e-01 -1.67989731e-01 -4.67412472e-01 -3.96453559e-01
-1.09994984e+00 8.92956436e-01 2.23663688e-01 -2.71061454e-02
-9.28680658e-01 3.43044460e-01 6.98888659e-01 -4.45399992e-02
1.39725268e-01 -2.51733243e-01 -6.22006506e-02 -2.48185679e-01
6.91236854e-01 3.80623698e-01 -2.31150731e-01 -6.98149920e-01
5.18238684e-03 7.34827101e-01 1.85281187e-01 -1.95781514e-01
1.29113650e+00 -9.54147220e-01 -4.61109430e-01 1.21103123e-01
1.26921570e+00 -7.62348622e-02 -1.60646129e+00 -2.22567707e-01
-7.09068298e-01 -9.62251067e-01 2.37741813e-01 -7.73573101e-01
-1.23613715e+00 8.82763743e-01 9.42283273e-01 2.84226686e-01
1.81839597e+00 -3.57657254e-01 6.53721511e-01 -2.30385378e-01
5.75765371e-02 -8.51988494e-01 1.13842689e-01 -6.30470365e-02
1.32166171e+00 -9.23803210e-01 1.94620162e-01 -4.51772928e-01
-4.94721532e-01 1.16874754e+00 3.77730757e-01 -6.44513667e-02
3.55574131e-01 1.91778272e-01 2.63200328e-02 2.16270998e-01
-3.33098650e-01 1.78963825e-01 1.18462428e-01 6.61144853e-01
9.27559659e-02 -1.97361503e-02 -2.96434283e-01 3.31926107e-01
-2.24907175e-02 1.44115642e-01 7.62477875e-01 6.22860193e-01
-1.24097094e-01 -1.11355984e+00 -7.86692619e-01 1.67438593e-02
-9.20710117e-02 -3.36792797e-01 7.32285157e-02 4.59856212e-01
2.86451072e-01 9.37946796e-01 -4.57848795e-02 -3.41185868e-01
3.05798113e-01 -3.93831491e-01 2.37635493e-01 -1.88358918e-01
7.23149404e-02 3.29962194e-01 -4.30301756e-01 -6.72324955e-01
-4.96081650e-01 -3.18430871e-01 -6.07081294e-01 -4.89644408e-01
-2.79535443e-01 -4.37252671e-01 5.48768640e-01 4.32677656e-01
5.58990717e-01 7.04659402e-01 6.67954624e-01 -9.42738712e-01
-5.40078223e-01 -8.20544362e-01 -8.01321149e-01 8.42341363e-01
6.42719030e-01 -5.15230954e-01 -7.05004215e-01 4.90423411e-01] | [11.228896141052246, -2.045501947402954] |
702670cc-3468-470a-be47-74312aaa5f4c | overview-and-results-cl-scisumm-shared-task | 1907.09854 | null | https://arxiv.org/abs/1907.09854v1 | https://arxiv.org/pdf/1907.09854v1.pdf | Overview and Results: CL-SciSumm Shared Task 2019 | The CL-SciSumm Shared Task is the first medium-scale shared task on scientific document summarization in the computational linguistics~(CL) domain. In 2019, it comprised three tasks: (1A) identifying relationships between citing documents and the referred document, (1B) classifying the discourse facets, and (2) generating the abstractive summary. The dataset comprised 40 annotated sets of citing and reference papers of the CL-SciSumm 2018 corpus and 1000 more from the SciSummNet dataset. All papers are from the open access research papers in the CL domain. This overview describes the participation and the official results of the CL-SciSumm 2019 Shared Task, organized as a part of the 42nd Annual Conference of the Special Interest Group in Information Retrieval (SIGIR), held in Paris, France in July 2019. We compare the participating systems in terms of two evaluation metrics and discuss the use of ROUGE as an evaluation metric. The annotated dataset used for this shared task and the scripts used for evaluation can be accessed and used by the community at: https://github.com/WING-NUS/scisumm-corpus. | ['Min-Yen Kan', 'Dayne Freitag', 'Muthu Kumar Chandrasekaran', 'Dragomir Radev', 'Michihiro Yasunaga'] | 2019-07-23 | null | null | null | null | ['scientific-article-summarization'] | ['natural-language-processing'] | [ 1.22133754e-01 2.98704356e-01 -3.77102524e-01 1.62529662e-01
-1.55591726e+00 -9.34155583e-01 1.13603723e+00 8.41317832e-01
-5.09999216e-01 9.93709147e-01 1.01353478e+00 -2.48954207e-01
-3.48429114e-01 -1.86605096e-01 -5.54893136e-01 -3.63835633e-01
1.44384906e-01 5.64135432e-01 7.03643635e-02 -1.93418384e-01
1.05632544e+00 8.13506320e-02 -1.42286313e+00 6.71318293e-01
1.24825561e+00 4.13058341e-01 4.05254185e-01 9.29720938e-01
-2.93911994e-01 6.28972054e-01 -1.01387215e+00 -5.13883591e-01
-4.66543049e-01 -6.72789335e-01 -1.28910398e+00 -5.80907345e-01
7.19419777e-01 4.40409750e-01 -2.54606932e-01 1.01307762e+00
6.60907686e-01 2.16274112e-01 7.51040220e-01 -9.79571521e-01
-4.39484090e-01 1.28450310e+00 -4.03862566e-01 7.07800686e-01
6.85451508e-01 -4.79478419e-01 1.25947666e+00 -9.13237751e-01
1.16102910e+00 1.29145133e+00 3.35539430e-01 2.54251629e-01
-6.27842486e-01 -6.26652181e-01 5.12366928e-02 5.63983858e-01
-9.94464934e-01 -6.08766198e-01 5.43787599e-01 -4.59729940e-01
9.53163683e-01 7.47209251e-01 3.24357748e-01 1.19448805e+00
1.12223238e-01 9.36043680e-01 9.17609274e-01 -6.81355238e-01
1.71632543e-01 -1.04156598e-01 7.99332738e-01 1.34969309e-01
7.36700416e-01 -7.48073399e-01 -8.59845281e-01 -4.37189192e-01
-2.88567722e-01 -6.76712573e-01 -6.15430236e-01 4.87628818e-01
-1.51895428e+00 6.92531288e-01 5.48435338e-02 8.87757599e-01
-4.56511736e-01 3.14785577e-02 7.67187893e-01 2.96885639e-01
1.05681801e+00 7.40819573e-01 -1.91284612e-01 -3.56406093e-01
-1.08895802e+00 8.69986713e-01 1.16715360e+00 8.00045967e-01
-9.33931693e-02 -5.62481642e-01 -5.09447515e-01 9.13543344e-01
5.81568591e-02 4.12644297e-01 6.43147290e-01 -1.35840356e+00
8.98686826e-01 4.92602617e-01 9.47228447e-02 -8.64251673e-01
-3.39214832e-01 -4.80698049e-01 -5.47484636e-01 -4.01497573e-01
1.43578267e-02 -4.88789618e-01 -3.46128881e-01 1.33520472e+00
-2.95729912e-03 -8.43883008e-02 2.41647631e-01 5.24664581e-01
1.90115690e+00 9.20762479e-01 -1.51511252e-01 -8.65884840e-01
1.46557546e+00 -9.05679524e-01 -1.30648506e+00 9.11219046e-02
8.96187067e-01 -1.23519146e+00 5.10842919e-01 2.48230100e-01
-1.57620645e+00 -1.15781045e-02 -1.22023833e+00 -3.39389414e-01
-5.93891680e-01 1.62311345e-01 1.56024113e-01 -1.87679783e-01
-1.15673089e+00 5.88182032e-01 -3.98308367e-01 -5.69924116e-01
3.50138336e-01 -2.72052377e-01 -1.45666659e-01 -3.79799865e-02
-1.30840576e+00 1.20055854e+00 3.85722548e-01 -3.38123858e-01
-2.07607552e-01 -9.36519325e-01 -4.95096117e-01 -3.02448459e-02
4.07375336e-01 -5.49845397e-01 1.39006901e+00 -3.72252136e-01
-1.03038299e+00 1.28979671e+00 -2.18319595e-01 -4.26154882e-01
4.15482432e-01 -4.20453310e-01 -3.22449923e-01 1.00835636e-01
4.74257559e-01 9.79413986e-02 -1.98039357e-02 -1.12829375e+00
-5.89213014e-01 -2.40090907e-01 -2.24478543e-01 3.65416408e-01
-5.42064086e-02 5.90517759e-01 -4.10699129e-01 -8.04277897e-01
-1.31631181e-01 -7.87739277e-01 2.14707687e-01 -8.91546786e-01
-6.75173402e-01 -9.32130992e-01 6.00136578e-01 -9.01151299e-01
1.64012945e+00 -1.71700609e+00 4.75776464e-01 -3.79059047e-01
2.27532983e-01 2.80542672e-01 7.35221524e-03 1.06357455e+00
-1.17427647e-01 3.66906613e-01 -3.04069906e-01 -4.91929591e-01
2.41511259e-02 -4.95763004e-01 -5.41662753e-01 3.82919610e-01
-4.46760148e-01 8.39069664e-01 -1.24761796e+00 -6.36402309e-01
-3.61752361e-01 -2.74947472e-02 3.00657302e-01 -4.77370210e-02
-1.87568024e-01 1.57301471e-01 -6.36254311e-01 1.44510612e-01
3.35683346e-01 -1.37373596e-01 9.26933289e-02 -1.49946034e-01
-5.63240528e-01 9.23081994e-01 -6.24399066e-01 1.75270760e+00
-1.54302776e-01 1.07461858e+00 3.79723497e-03 -7.01280057e-01
5.86476386e-01 6.63772345e-01 4.64646310e-01 -3.63770306e-01
2.53114730e-01 5.65829456e-01 -7.42364153e-02 -3.71968985e-01
8.57413352e-01 5.70687532e-01 -3.45237285e-01 8.26925695e-01
1.69087559e-01 -5.46315789e-01 1.08807456e+00 1.02777112e+00
1.19467008e+00 -9.71197039e-02 4.95012820e-01 -6.66649640e-01
6.47561669e-01 2.14493796e-01 3.13835740e-01 7.00973988e-01
9.78293344e-02 4.90386844e-01 7.24541426e-01 -1.02809221e-01
-9.16110575e-01 -5.18401086e-01 -2.94821471e-01 9.64096487e-01
-7.94329569e-02 -9.21218634e-01 -7.33339190e-01 -2.86817342e-01
-1.97851494e-01 1.41127133e+00 -7.43738651e-01 1.40184030e-01
-7.20635235e-01 -7.34615028e-01 7.23690093e-01 -1.33317962e-01
4.03385103e-01 -1.21890783e+00 -2.87448734e-01 -3.79609801e-02
-9.01125193e-01 -9.38378990e-01 -5.90305209e-01 7.05366284e-02
-5.96169472e-01 -1.24924350e+00 -1.03862870e+00 -7.14883387e-01
1.85755730e-01 1.90820634e-01 1.13524711e+00 7.69255590e-03
-9.30977985e-02 2.57492691e-01 -6.21215343e-01 -9.57686126e-01
-6.64540470e-01 4.56121564e-01 -2.38221481e-01 -6.51388586e-01
2.10880592e-01 -2.38703907e-01 -3.54260921e-01 -5.26689827e-01
-5.48829973e-01 1.71766028e-01 1.67057976e-01 4.03750002e-01
2.83713669e-01 -6.83283031e-01 9.96032536e-01 -9.67175841e-01
1.33539093e+00 -6.60025775e-01 -1.13848314e-01 5.09011149e-01
-6.51805043e-01 -3.39882821e-01 8.13707784e-02 1.56196132e-01
-1.07726347e+00 -8.19871724e-01 4.42358926e-02 2.68803090e-01
4.04419005e-01 1.00845325e+00 1.09367743e-01 4.99012470e-01
7.77363062e-01 -1.11818455e-01 -3.76105726e-01 -6.67607903e-01
4.40003008e-01 9.69446421e-01 7.05173135e-01 -4.03758228e-01
2.68724710e-01 -7.47684315e-02 -2.09276956e-02 -7.66147435e-01
-1.34986138e+00 -6.44743264e-01 -3.50888699e-01 -4.10559475e-01
7.87636578e-01 -8.23169470e-01 -2.83969432e-01 3.29932302e-01
-1.56266558e+00 1.32431373e-01 -4.17444050e-01 3.65718037e-01
-2.56612867e-01 3.51910681e-01 -5.88744223e-01 -5.61273932e-01
-9.87338066e-01 -7.64738739e-01 8.49890769e-01 3.24350655e-01
-7.31532514e-01 -1.03069401e+00 3.90817732e-01 7.13056862e-01
1.76298797e-01 4.85153913e-01 7.55891562e-01 -1.15230143e+00
2.82824129e-01 -3.67985845e-01 -1.27230864e-02 -8.26037973e-02
-8.78941119e-02 3.16859752e-01 -7.22979665e-01 -1.31128982e-01
-2.29283631e-01 -4.97561485e-01 1.28290272e+00 5.36614358e-01
9.64721262e-01 -4.34135884e-01 -4.56722885e-01 -1.63493365e-01
8.95182014e-01 2.98765791e-03 4.68806714e-01 5.65344036e-01
3.94386202e-01 6.75900936e-01 5.46538115e-01 3.30411315e-01
6.16967380e-01 5.02588212e-01 9.86054093e-02 5.60435712e-01
-4.79461968e-01 1.38976768e-01 2.84388989e-01 1.68890500e+00
-2.17451558e-01 -8.15877080e-01 -1.02019382e+00 7.82000422e-01
-2.13583827e+00 -1.04023218e+00 -6.83942437e-01 2.00829053e+00
1.12898982e+00 -3.95043567e-02 -6.47292286e-02 -4.98067513e-02
7.58039355e-01 4.60258991e-01 -1.12868078e-01 -7.30193973e-01
-5.92728674e-01 4.77840342e-02 1.96863189e-01 6.81979060e-01
-9.26070690e-01 6.60181403e-01 5.41584635e+00 1.14083517e+00
-6.46950722e-01 3.98647368e-01 3.64936888e-01 -3.29025209e-01
-4.98950124e-01 2.67601199e-02 -6.90246642e-01 5.02337337e-01
1.36616111e+00 -1.37498271e+00 -1.93851978e-01 1.56734824e-01
1.37045205e-01 -4.38163340e-01 -6.75628901e-01 5.01310587e-01
4.68033493e-01 -1.79979193e+00 1.12410625e-02 -1.03145331e-01
9.70931232e-01 5.29704690e-01 -3.28132987e-01 1.64706364e-01
3.46923441e-01 -7.93249905e-01 7.72429585e-01 6.34588301e-01
6.91638649e-01 -5.99162161e-01 9.82860029e-01 2.86707729e-01
-4.64210540e-01 3.92434627e-01 -8.63245281e-04 1.21684507e-01
2.29928449e-01 7.52398491e-01 -2.79521674e-01 9.79688823e-01
6.19977772e-01 9.70699310e-01 -4.61268067e-01 1.13459027e+00
-1.84825391e-01 8.95592570e-01 -3.83828059e-02 -3.84612650e-01
2.60886580e-01 -5.03220595e-02 1.19538581e+00 1.73652577e+00
2.37908959e-01 2.57851869e-01 -1.14745446e-01 5.41014731e-01
-7.50584543e-01 2.02526614e-01 -2.55388498e-01 -3.51681828e-01
7.62366235e-01 1.28064215e+00 -6.73819482e-01 -5.57272851e-01
1.80121720e-01 6.53578281e-01 1.43254727e-01 2.19551668e-01
-4.15183961e-01 -7.21657932e-01 -1.17861433e-02 -3.11050862e-01
-4.89717387e-02 1.96450859e-01 -4.67952222e-01 -9.71589684e-01
-4.24096845e-02 -6.50156856e-01 4.92293209e-01 -8.32183480e-01
-1.24226224e+00 8.12515020e-01 3.74699920e-01 -9.93852198e-01
-3.46875250e-01 1.21140286e-01 -8.87871563e-01 1.01505053e+00
-1.06437624e+00 -1.04034281e+00 -1.89380139e-01 -1.03262119e-01
8.29396725e-01 -2.48799309e-01 8.63355279e-01 6.40712455e-02
-5.50536215e-01 2.42096052e-01 5.33997774e-01 -2.86047637e-01
1.10406065e+00 -1.35817659e+00 3.78371567e-01 6.16435647e-01
-1.89969748e-01 5.85079610e-01 1.11964846e+00 -8.22520494e-01
-1.06585550e+00 -8.98521185e-01 1.84614575e+00 -5.22889972e-01
1.00228298e+00 1.19995838e-02 -8.75629485e-01 5.32585382e-01
1.06751537e+00 -6.57973468e-01 6.97148681e-01 5.05781807e-02
-3.88494246e-02 3.21453720e-01 -7.62080610e-01 5.07687151e-01
8.10465574e-01 -1.17794268e-01 -1.10805976e+00 9.90782917e-01
8.64011943e-01 -7.23948598e-01 -1.01000309e+00 1.56009421e-01
2.55400062e-01 -2.06889406e-01 4.97431725e-01 -5.05045474e-01
1.19375813e+00 -2.11114790e-02 2.11422127e-02 -1.57543385e+00
-3.16462815e-01 -7.24504292e-01 -3.00253481e-01 1.58543813e+00
7.30572462e-01 -3.86917651e-01 -3.98912281e-02 8.54909942e-02
-6.66197360e-01 -7.56556034e-01 -1.08872890e+00 -4.15320903e-01
7.05563784e-01 -2.49673471e-01 -6.96663791e-03 9.39782798e-01
5.78760087e-01 5.60311496e-01 2.20560685e-01 -4.39116806e-01
7.45527327e-01 1.93233237e-01 3.09065133e-01 -1.26835108e+00
4.30023074e-01 -9.92740273e-01 9.72634926e-02 -3.54361832e-01
2.82660246e-01 -1.23720765e+00 -1.32750377e-01 -2.25310969e+00
5.22847891e-01 9.34065133e-02 -2.30940610e-01 3.29019308e-01
-4.17897701e-01 7.22751021e-02 2.91068733e-01 8.07545245e-01
-1.05865955e+00 2.51491010e-01 1.06539464e+00 -1.75182119e-01
1.12405933e-01 -1.82377160e-01 -1.12136614e+00 5.49755633e-01
5.68315864e-01 -4.42003220e-01 -5.90836070e-02 -2.64578968e-01
3.75174642e-01 2.46032998e-02 4.17274982e-03 -6.13333821e-01
5.79054117e-01 -9.28887352e-02 -1.62213743e-01 -1.02155089e+00
-1.32854488e-02 1.79821908e-01 -5.63983321e-02 3.48775178e-01
-8.88480723e-01 2.02427611e-01 3.80939752e-01 1.40963733e-01
-3.05672705e-01 -6.42359138e-01 3.22564423e-01 -1.81560472e-01
-1.74847528e-01 -2.98072577e-01 -5.07560670e-01 6.01942360e-01
7.75721192e-01 3.93924028e-01 -1.08027720e+00 -4.03205723e-01
-3.31626952e-01 4.99666452e-01 5.47653623e-02 6.51066959e-01
2.94797003e-01 -9.26262677e-01 -1.67130113e+00 -9.64395583e-01
1.58190906e-01 -1.63893908e-01 2.07812101e-01 1.06851745e+00
-5.06782651e-01 1.01874220e+00 2.53628165e-01 -8.50813289e-04
-1.32962155e+00 1.38874114e-01 -2.76887119e-01 -6.42155766e-01
-7.33881474e-01 7.18526959e-01 -2.16470450e-01 -3.71327437e-02
1.53628424e-01 5.02163321e-02 -8.38343799e-01 6.30336046e-01
9.26387489e-01 8.67233872e-01 4.50185657e-01 -6.35395885e-01
-3.94941509e-01 1.75042510e-01 -3.16483051e-01 -4.61134732e-01
1.62724590e+00 -3.26609552e-01 -6.73289180e-01 9.31241870e-01
1.19558918e+00 3.59301940e-02 -1.14382908e-01 -2.45019466e-01
4.73045588e-01 3.08403879e-01 5.52370965e-01 -1.32095885e+00
-6.08037114e-01 3.39022458e-01 -8.81148577e-02 3.55782151e-01
5.90392649e-01 2.03992963e-01 7.26375163e-01 2.72753656e-01
-1.15133166e-01 -1.22560799e+00 -3.98560375e-01 8.21714163e-01
1.81042683e+00 -8.19737077e-01 5.99737585e-01 -2.94781297e-01
-7.37185895e-01 1.23150051e+00 8.08629021e-02 -9.13299993e-02
4.42787409e-01 8.85730907e-02 -7.99931213e-02 -6.22553945e-01
-1.06957066e+00 1.36237204e-01 7.05713570e-01 -9.60781425e-02
1.09235346e+00 -1.05983593e-01 -1.54722548e+00 8.37867379e-01
-4.23184991e-01 -3.87906469e-02 9.20889378e-01 1.00393617e+00
-3.68149549e-01 -8.16607475e-01 -1.55284032e-01 7.15963244e-01
-7.12087154e-01 -1.27639621e-01 -9.94291604e-01 5.05585909e-01
-1.91778407e-01 1.09474874e+00 3.37624922e-02 -2.27893889e-02
2.93208063e-01 -5.10081761e-02 4.00023788e-01 -5.62437773e-01
-1.00992763e+00 -1.42306626e-01 8.07850897e-01 -1.84688911e-01
-6.27004921e-01 -1.22078454e+00 -1.15023959e+00 -5.90632677e-01
-2.09120795e-01 9.18703854e-01 9.66447711e-01 9.55316365e-01
4.51409012e-01 9.40522015e-01 2.05441967e-01 -8.79195929e-01
-2.40241632e-01 -1.51817286e+00 -3.91038775e-01 1.49893433e-01
4.82432470e-02 -4.23861891e-01 -6.53144717e-01 1.37142256e-01] | [12.384111404418945, 9.577133178710938] |
a674237e-943b-4fcc-ba66-e5f82de4bce7 | feature-robustness-and-sex-differences-in | 2204.01737 | null | https://arxiv.org/abs/2204.01737v3 | https://arxiv.org/pdf/2204.01737v3.pdf | Feature robustness and sex differences in medical imaging: a case study in MRI-based Alzheimer's disease detection | Convolutional neural networks have enabled significant improvements in medical image-based diagnosis. It is, however, increasingly clear that these models are susceptible to performance degradation when facing spurious correlations and dataset shift, leading, e.g., to underperformance on underrepresented patient groups. In this paper, we compare two classification schemes on the ADNI MRI dataset: a simple logistic regression model using manually selected volumetric features, and a convolutional neural network trained on 3D MRI data. We assess the robustness of the trained models in the face of varying dataset splits, training set sex composition, and stage of disease. In contrast to earlier work in other imaging modalities, we do not observe a clear pattern of improved model performance for the majority group in the training dataset. Instead, while logistic regression is fully robust to dataset composition, we find that CNN performance is generally improved for both male and female subjects when including more female subjects in the training dataset. We hypothesize that this might be due to inherent differences in the pathology of the two sexes. Moreover, in our analysis, the logistic regression model outperforms the 3D CNN, emphasizing the utility of manual feature specification based on prior knowledge, and the need for more robust automatic feature selection. | ['Maria Luise da Costa Zemsch', 'Melanie Ganz', 'Oskar Eiler Wiese Christensen', 'Anders Henriksen', 'Aasa Feragen', 'Eike Petersen'] | 2022-04-04 | null | null | null | null | ['alzheimer-s-disease-detection'] | ['medical'] | [ 3.19987208e-01 1.10651724e-01 -3.01703393e-01 -7.84343302e-01
-6.83270156e-01 -4.66590494e-01 5.78333557e-01 3.75879467e-01
-7.18410671e-01 6.07614100e-01 6.05089784e-01 -4.24494445e-01
-5.24500668e-01 -6.63760722e-01 -5.10986030e-01 -5.13714552e-01
-2.29520753e-01 6.23500228e-01 -2.20765501e-01 2.14392886e-01
3.26861739e-02 4.19452071e-01 -1.06911612e+00 1.02029704e-01
5.32425165e-01 8.94291580e-01 -1.87713161e-01 2.47097760e-01
2.79871345e-01 5.46373367e-01 -2.95123130e-01 -3.58676583e-01
3.48218739e-01 -2.12105468e-01 -7.01918542e-01 2.14170128e-01
9.09965217e-01 -5.97738862e-01 -4.62444484e-01 7.51060486e-01
6.62933707e-01 -2.73142368e-01 8.41128945e-01 -1.08200955e+00
-5.32967627e-01 7.54219770e-01 -6.12712920e-01 4.21425581e-01
-2.41507649e-01 4.19731200e-01 7.39912748e-01 -5.76573908e-01
7.31491089e-01 1.09343183e+00 1.05900288e+00 3.37672651e-01
-1.46549845e+00 -9.45127130e-01 4.32653055e-02 -1.14179976e-01
-1.22571278e+00 -6.09924138e-01 2.13943988e-01 -9.07477379e-01
5.28384924e-01 5.78942597e-02 5.39864182e-01 9.69220400e-01
3.64046216e-01 7.26062283e-02 1.30282533e+00 -6.84258435e-03
-4.84595262e-02 3.73420939e-02 2.90487766e-01 4.42052215e-01
6.67054474e-01 6.35047480e-02 -1.47141293e-01 -5.65566778e-01
7.68995285e-01 1.06253894e-03 -9.57220569e-02 -4.66920525e-01
-1.23112071e+00 8.17064941e-01 4.89073664e-01 3.33616495e-01
-3.61497164e-01 8.79528672e-02 6.53229356e-01 4.97861430e-02
5.03976226e-01 4.14026886e-01 -5.17469943e-01 7.99992010e-02
-1.27617526e+00 4.80809301e-01 3.09989035e-01 9.39799696e-02
6.01428561e-02 -6.99097514e-02 2.21755411e-02 8.98197651e-01
1.93009034e-01 2.30695367e-01 6.59677446e-01 -9.36490417e-01
3.76913637e-01 6.28984928e-01 -3.33109230e-01 -1.00625896e+00
-1.02289212e+00 -8.78457189e-01 -6.16414726e-01 3.22374284e-01
7.20551372e-01 -2.44925529e-01 -1.05439985e+00 1.94263577e+00
-3.49094979e-02 -4.37576860e-01 -2.52323478e-01 9.73901629e-01
7.29170978e-01 -3.55463207e-01 4.06955928e-01 6.50581867e-02
1.37930429e+00 -2.24993140e-01 -2.41483539e-01 -4.09148157e-01
7.10084975e-01 -4.53657299e-01 6.23455226e-01 2.69277275e-01
-9.05152798e-01 -2.82430083e-01 -8.70845854e-01 2.07772329e-02
-5.30329533e-02 4.04884703e-02 8.39832723e-01 8.64800453e-01
-9.37176406e-01 6.01950645e-01 -1.03074932e+00 -5.36082923e-01
9.16983485e-01 6.69359803e-01 -6.12534940e-01 -2.82790452e-01
-7.45549917e-01 9.83096778e-01 1.11926399e-01 5.03942929e-02
-4.91393656e-01 -1.08874583e+00 -7.59287834e-01 2.82142926e-02
-1.03753984e-01 -7.24266410e-01 9.84093547e-01 -1.15672839e+00
-4.59638387e-01 1.06777418e+00 1.04031488e-01 -4.90315825e-01
7.10923672e-01 2.61074275e-01 -1.49782032e-01 -2.94746216e-02
2.25633383e-01 8.01455557e-01 5.56199908e-01 -1.10373151e+00
-4.07165736e-01 -1.05348003e+00 -2.47715116e-01 1.51787058e-01
1.69147812e-02 1.67958096e-01 1.34194806e-01 -6.05722070e-01
3.75135630e-01 -8.61049831e-01 -3.39573592e-01 1.01952061e-01
1.78976497e-03 4.90181409e-02 4.57658738e-01 -8.30441177e-01
7.35660493e-01 -2.26666498e+00 -2.37227738e-01 3.37452531e-01
7.49417007e-01 -1.95092008e-01 1.58812910e-01 -2.38713071e-01
-5.58329105e-01 3.75997722e-01 -2.13140056e-01 -1.16681226e-01
-3.56906444e-01 -1.39620885e-01 2.15967819e-01 7.67815948e-01
4.66326416e-01 7.37087846e-01 -5.84477842e-01 -4.83639836e-01
-1.38202950e-01 4.15144414e-01 -6.79371834e-01 -5.51574349e-01
5.44511139e-01 5.20442367e-01 -1.39854908e-01 7.40141928e-01
6.86709762e-01 -2.89215416e-01 4.11260307e-01 -2.05004886e-01
1.15031511e-01 3.85604799e-01 -7.08199143e-01 1.24846947e+00
-8.02416503e-02 6.97099984e-01 2.64856815e-01 -9.54576731e-01
7.10518718e-01 1.89855307e-01 7.27745771e-01 -6.64827645e-01
3.57758701e-01 5.14324367e-01 9.62720215e-01 -1.77734360e-01
3.04379076e-01 -4.27402705e-01 9.05699655e-02 4.39308852e-01
8.39245319e-02 1.22751020e-01 -1.27098672e-02 -5.81280254e-02
1.14557576e+00 -3.93416375e-01 -4.35456820e-02 -3.81412864e-01
-2.19739586e-01 2.00327218e-01 5.87164104e-01 8.24709892e-01
-4.67516184e-01 1.07379961e+00 7.66509652e-01 -3.53595555e-01
-1.00712538e+00 -9.41562414e-01 -9.48539674e-01 7.37876713e-01
-3.92471403e-01 -1.37654483e-01 -3.15692544e-01 -7.27977812e-01
3.65830094e-01 4.29056972e-01 -8.81402910e-01 -2.93422222e-01
-5.32568216e-01 -1.42253852e+00 8.74116838e-01 6.77138627e-01
1.36225685e-01 -5.59858680e-01 -6.27324760e-01 1.23056687e-01
1.71935678e-01 -9.23362315e-01 -4.67727222e-02 2.39953697e-01
-1.19704115e+00 -1.16691101e+00 -7.00610280e-01 -5.31194270e-01
6.78515613e-01 -5.16460054e-02 1.09111464e+00 2.75506109e-01
-2.48809889e-01 3.79949778e-01 -8.04737061e-02 -4.23512727e-01
-3.41635913e-01 3.89207453e-01 -1.88853294e-02 -3.44933242e-01
2.74431825e-01 -4.78007048e-01 -8.05915713e-01 7.57816285e-02
-7.04741120e-01 -1.60682589e-01 7.69538522e-01 7.80673087e-01
1.85323939e-01 -1.36630192e-01 6.18954360e-01 -1.01548684e+00
3.21582913e-01 -8.24881434e-01 1.29779607e-01 -1.42961860e-01
-8.91435087e-01 -7.57020190e-02 -1.66847929e-02 -4.69264507e-01
-7.39600539e-01 -3.09018850e-01 8.45188722e-02 -3.89829203e-02
-2.40878835e-01 6.93206847e-01 3.04710448e-01 -1.59439534e-01
7.57086456e-01 -4.12492990e-01 5.18963516e-01 -3.66225451e-01
-2.52497017e-01 6.02488995e-01 3.07089239e-01 -4.36341077e-01
4.80535448e-01 4.55459744e-01 1.04809992e-01 -3.94321233e-01
-5.37265897e-01 3.64363790e-02 -6.93978429e-01 -2.55320594e-02
8.10790658e-01 -8.88993979e-01 -2.91720271e-01 4.24775958e-01
-4.31891143e-01 -5.39229572e-01 -8.16469267e-02 8.90132308e-01
-1.56926334e-01 2.22248305e-02 -4.85958457e-01 -3.28592926e-01
-3.81993383e-01 -1.57375944e+00 8.66059601e-01 7.03031570e-03
-6.19533241e-01 -1.01365960e+00 -1.94382340e-01 4.52137589e-01
5.97183108e-01 5.00009120e-01 1.45030904e+00 -1.00474298e+00
6.56244680e-02 -5.36080420e-01 -5.92160702e-01 1.07813664e-01
3.83742899e-01 -1.90554652e-02 -7.65173674e-01 -2.58601010e-01
-1.32471636e-01 -3.48931819e-01 1.03877032e+00 6.28508866e-01
1.14858580e+00 2.02963606e-01 -2.94917107e-01 4.07475293e-01
1.19627976e+00 1.24853559e-01 2.42367953e-01 4.61985320e-01
6.18329406e-01 5.60919404e-01 7.73497671e-02 2.70095933e-02
6.54601336e-01 4.80227172e-01 2.49399960e-01 -4.57318097e-01
-2.64661938e-01 2.00599596e-01 -4.44969051e-02 5.35910577e-02
-1.09271102e-01 3.49655151e-01 -1.33770466e+00 4.30213422e-01
-1.38492548e+00 -6.95122719e-01 -1.63873956e-01 2.31539941e+00
6.10537231e-01 5.33645451e-01 3.63702953e-01 8.26669484e-02
5.99845052e-01 2.79263165e-02 -6.47594213e-01 -2.39428386e-01
-1.82820424e-01 3.27667177e-01 9.73993957e-01 -2.93531269e-02
-8.73320818e-01 8.43287632e-02 7.06190777e+00 1.90860465e-01
-1.56855536e+00 1.67774856e-01 1.19491351e+00 -4.54889536e-01
2.95684580e-02 -1.07818492e-01 -3.79966319e-01 2.46420205e-01
9.33985293e-01 -5.07719480e-02 7.80423582e-02 5.33532739e-01
1.89038903e-01 -1.72238752e-01 -1.03945267e+00 5.56356251e-01
1.68949053e-01 -9.76538897e-01 -2.92546302e-01 3.96512568e-01
3.44579458e-01 3.45063478e-01 2.25598246e-01 3.57888281e-01
-2.24108249e-02 -1.39221156e+00 7.48880208e-01 1.28373399e-01
9.03500736e-01 -3.78219277e-01 1.02554035e+00 -1.32751569e-01
-4.30204511e-01 -2.84827679e-01 8.34238380e-02 -9.78351161e-02
-1.85432941e-01 4.56073225e-01 -9.77812052e-01 2.01353639e-01
8.11109304e-01 3.54184479e-01 -1.13299918e+00 1.15191650e+00
5.23167968e-01 8.22773397e-01 -2.81947941e-01 5.69438159e-01
2.42962047e-01 1.91828489e-01 1.56310529e-01 9.40512657e-01
8.07783455e-02 -9.86195751e-04 -7.95119479e-02 5.50719976e-01
-6.54222071e-02 6.50999323e-02 -3.62878412e-01 -1.63828969e-01
1.37291014e-01 1.44434702e+00 -1.07308805e+00 -1.42012075e-01
-7.13808239e-01 2.31175140e-01 1.89333245e-01 4.17515337e-02
-5.09170651e-01 5.37827797e-02 4.76920903e-01 6.64257169e-01
-5.36703244e-02 -2.78794974e-01 -8.59258056e-01 -9.35516834e-01
-1.52639657e-01 -1.00647199e+00 6.10911965e-01 -5.40974677e-01
-1.24330199e+00 2.67229050e-01 1.06872566e-01 -6.04971409e-01
-1.16472274e-01 -5.79130709e-01 -3.13385993e-01 8.93221796e-01
-1.02058816e+00 -1.04169679e+00 -1.23580977e-01 1.79112583e-01
1.05807140e-01 -2.90828384e-02 5.02483785e-01 5.09035230e-01
-4.49203700e-01 7.22602367e-01 -1.45025998e-01 3.69464308e-01
1.12443101e+00 -1.05809367e+00 1.13870524e-01 4.03788656e-01
-2.74192154e-01 7.97556579e-01 4.44737136e-01 -8.82993221e-01
-9.29124177e-01 -8.86028945e-01 6.38785481e-01 -5.18853724e-01
4.70571786e-01 -1.20208502e-01 -9.04998362e-01 9.95404661e-01
-1.66898653e-01 -7.08275139e-02 8.88482988e-01 5.60071468e-01
-3.46133709e-01 1.22372564e-02 -1.45564473e+00 4.56802011e-01
1.02599871e+00 -2.00041562e-01 -4.03021693e-01 7.66063407e-02
3.37601155e-01 -6.37070119e-01 -1.25937819e+00 7.57003605e-01
9.52769101e-01 -7.84040213e-01 7.98194885e-01 -6.58234656e-01
7.36311138e-01 1.54651731e-01 -7.07830861e-02 -1.02200031e+00
-4.51805860e-01 4.89001244e-01 5.50776720e-01 1.09979236e+00
7.50770926e-01 -4.92423147e-01 8.31415355e-01 1.20484233e+00
-1.95331182e-02 -1.03839254e+00 -9.83813226e-01 -3.28971535e-01
6.68433666e-01 -4.92941737e-01 4.46873635e-01 1.13512802e+00
-2.24204049e-01 2.74718646e-02 1.30136743e-01 7.76565745e-02
4.44338083e-01 -2.04000741e-01 3.61715019e-01 -1.34776723e+00
-1.81198314e-01 -7.26170003e-01 -6.69948101e-01 1.15689427e-01
-5.82826547e-02 -1.02797103e+00 -3.51510942e-01 -1.48216188e+00
5.71411073e-01 -8.30791354e-01 -3.70911270e-01 6.56084538e-01
-1.98892370e-01 5.66507280e-01 2.32105851e-01 2.04819873e-01
-5.59474602e-02 -6.24927096e-02 1.09241104e+00 5.68146445e-03
-2.44017720e-01 -9.50836390e-02 -1.17562211e+00 5.76287806e-01
8.98438871e-01 -6.04170084e-01 -1.99604422e-01 -6.17238224e-01
1.31157070e-01 -2.94687212e-01 7.96984076e-01 -9.38657582e-01
-2.37530515e-01 1.04338966e-01 1.13783574e+00 -2.16858134e-01
1.68670192e-01 -7.02098489e-01 1.52491435e-01 6.48948491e-01
-3.63333434e-01 2.92306691e-01 3.73337030e-01 2.14060284e-02
2.79443055e-01 -1.29367635e-02 6.83058143e-01 -1.97529510e-01
-4.30918597e-02 1.93602860e-01 -4.43191081e-01 8.29398707e-02
4.23332214e-01 -2.82255441e-01 -2.35568970e-01 -2.07273856e-01
-7.78489709e-01 1.22930229e-01 4.64150339e-01 4.74025756e-01
-1.61516629e-02 -1.19856977e+00 -7.20398188e-01 8.08634609e-02
2.29628701e-02 -1.72708899e-01 2.15813056e-01 1.36379623e+00
-3.11891675e-01 2.91310787e-01 -2.38089517e-01 -6.01900697e-01
-1.08045089e+00 -8.63888413e-02 6.66558146e-01 -9.93072912e-02
-4.59214568e-01 4.18934584e-01 1.66463286e-01 -5.61508119e-01
6.96721971e-02 -4.47117001e-01 -1.33928880e-02 2.86138505e-01
3.44846159e-01 4.00048494e-01 3.05614322e-01 -7.79569626e-01
-5.50670564e-01 1.56355828e-01 -4.15156215e-01 -1.85800180e-01
1.64703524e+00 7.37569481e-02 8.27690214e-02 3.28164041e-01
9.70117331e-01 -8.94621834e-02 -9.43269074e-01 -8.30450952e-02
-4.22808416e-02 -2.10841209e-01 2.40802988e-01 -9.22431052e-01
-1.34265935e+00 3.86200249e-01 1.04008830e+00 -6.66401908e-02
7.62135863e-01 1.13235898e-01 3.97225022e-01 -1.37320474e-01
8.50940570e-02 -7.23299503e-01 -5.63110232e-01 3.49629484e-02
7.87166953e-01 -1.26840723e+00 4.37509894e-01 6.41961256e-03
-6.09680951e-01 9.95105267e-01 6.47192359e-01 -1.42357022e-01
6.02660298e-01 1.44216835e-01 8.77114907e-02 -4.22535151e-01
-5.12144089e-01 1.85583364e-02 1.88164800e-01 5.76975644e-01
8.24322462e-01 -6.68695197e-03 -7.25142777e-01 8.83529544e-01
-6.60637081e-01 9.67778713e-02 5.35853088e-01 7.50284851e-01
-4.32203859e-02 -9.69599426e-01 -4.23200250e-01 1.24611032e+00
-8.81634176e-01 1.84295587e-02 -5.13215423e-01 1.16627574e+00
4.43254322e-01 5.95564842e-01 7.09587932e-01 -1.27261013e-01
3.41141790e-01 3.18247825e-01 5.62090397e-01 -6.68483019e-01
-8.00380409e-01 -6.00264855e-02 1.71983704e-01 -2.82374293e-01
-3.92312914e-01 -1.21519637e+00 -9.91503835e-01 -2.84946650e-01
-1.98483586e-01 -4.60770309e-01 6.89811528e-01 1.06563461e+00
3.71752501e-01 4.51782793e-01 6.43791929e-02 -6.32186890e-01
-4.87927526e-01 -9.88854766e-01 -5.40499985e-01 4.03962702e-01
3.73212427e-01 -9.42807078e-01 -1.81806520e-01 -2.61372328e-01] | [14.871918678283691, -2.1008336544036865] |
a2a63500-e524-4a6c-b020-5091082fda0b | creating-realistic-anterior-segment-optical | 2306.14058 | null | https://arxiv.org/abs/2306.14058v1 | https://arxiv.org/pdf/2306.14058v1.pdf | Creating Realistic Anterior Segment Optical Coherence Tomography Images using Generative Adversarial Networks | This paper presents the development and validation of a Generative Adversarial Network (GAN) purposed to create high-resolution, realistic Anterior Segment Optical Coherence Tomography (AS-OCT) images. We trained the Style and WAvelet based GAN (SWAGAN) on 142,628 AS-OCT B-scans. Three experienced refractive surgeons performed a blinded assessment to evaluate the realism of the generated images; their results were not significantly better than chance in distinguishing between real and synthetic images, thus demonstrating a high degree of image realism. To gauge their suitability for machine learning tasks, a convolutional neural network (CNN) classifier was trained with a dataset containing both real and GAN-generated images. The CNN demonstrated an accuracy rate of 78% trained on real images alone, but this accuracy rose to 100% when training included the generated images. This underscores the utility of synthetic images for machine learning applications. We further improved the resolution of the generated images by up-sampling them twice (2x) using an Enhanced Super Resolution GAN (ESRGAN), which outperformed traditional up-sampling techniques. In conclusion, GANs can effectively generate high-definition, realistic AS-OCT images, proving highly beneficial for machine learning and image analysis tasks. | ['Shady T. Awwad', 'Michèle Haykal', 'Elsa Lamah', 'Jeffrey Yammine', 'Timothy Archer', 'Cyril Zakka', 'Guillermo Amescua', 'Dan Z. Reinstein', 'Anthony Abou Mrad', 'Jad F. Assaf'] | 2023-06-24 | null | null | null | null | ['super-resolution'] | ['computer-vision'] | [ 6.05652332e-01 5.02652586e-01 2.90721208e-01 -2.67773539e-01
-1.37778819e+00 -4.46467817e-01 4.27996993e-01 -6.04671955e-01
-3.25153053e-01 1.11648822e+00 3.01529765e-01 -4.74393100e-01
2.61837512e-01 -7.12729633e-01 -6.83604717e-01 -6.99811399e-01
6.34517744e-02 3.35797399e-01 -5.37008792e-02 1.17651641e-01
-5.25986664e-02 5.41839361e-01 -1.49019849e+00 6.17849231e-01
8.71278226e-01 8.73122633e-01 -9.56176743e-02 1.12301290e+00
5.53260565e-01 8.87246788e-01 -9.23545778e-01 -4.17052597e-01
6.83685064e-01 -8.88955474e-01 -5.60108840e-01 -1.49277434e-01
9.78422940e-01 -7.59053051e-01 -5.67457564e-02 7.02964962e-01
9.09671664e-01 -3.47433805e-01 3.79047751e-01 -5.21613300e-01
-6.65704131e-01 2.45974615e-01 -5.39881289e-01 4.47785527e-01
1.75045609e-01 5.72682619e-01 5.40777981e-01 -4.73508775e-01
7.23966241e-01 7.18891978e-01 6.82245851e-01 7.62932062e-01
-1.54810536e+00 -9.15959895e-01 -9.03579533e-01 -4.68637079e-01
-9.38727736e-01 -3.20297569e-01 2.60784626e-01 -7.62875021e-01
5.77878952e-01 3.68766338e-01 1.35062528e+00 1.10718274e+00
6.40838683e-01 -2.72308514e-02 1.89196849e+00 -3.46274376e-01
-7.51027511e-03 -3.02991383e-02 -7.39086628e-01 4.85614359e-01
3.75383139e-01 6.39509559e-01 1.79652348e-02 -2.37812042e-01
1.45283568e+00 -5.57213783e-01 -5.74519217e-01 1.81117430e-01
-1.20477092e+00 8.20535898e-01 5.85273564e-01 2.18423396e-01
-6.41127527e-01 2.12672949e-01 1.48965627e-01 4.32136208e-02
3.77118498e-01 1.10920691e+00 2.04287112e-01 -2.32837602e-01
-9.05817807e-01 -8.90154094e-02 4.64331210e-02 4.22496766e-01
3.78765352e-02 4.83594447e-01 -3.31605822e-01 8.06659937e-01
-3.29065889e-01 4.29853857e-01 7.62579620e-01 -1.03618932e+00
-1.70125604e-01 2.19540551e-01 9.93008018e-02 -5.24250627e-01
-7.73010626e-02 -8.72348547e-01 -8.70493412e-01 8.34824800e-01
2.62754619e-01 -4.33637381e-01 -1.34441352e+00 1.38924980e+00
-1.80407185e-02 1.75558701e-01 -2.97448412e-02 1.02614045e+00
9.06317115e-01 2.76628643e-01 4.71656919e-02 -1.61072463e-01
1.20292068e+00 -5.01625836e-01 -4.35035676e-01 -8.97036865e-02
2.10495323e-01 -1.08604276e+00 1.06749523e+00 5.08473337e-01
-1.69968438e+00 -5.98766029e-01 -1.12162805e+00 1.51518598e-01
3.95480394e-01 1.64264187e-01 4.80375290e-01 8.04018021e-01
-1.33559370e+00 3.46775323e-01 -6.89908564e-01 1.06590666e-01
9.59847808e-01 3.81808341e-01 -5.29244363e-01 -1.90050885e-01
-7.21018314e-01 7.90862143e-01 4.82504331e-02 -3.80753279e-02
-6.14581525e-01 -1.03330421e+00 -6.94436610e-01 -3.39473903e-01
-5.02279282e-01 -1.13481510e+00 1.07978690e+00 -1.25083852e+00
-1.51022232e+00 1.37943685e+00 1.45478413e-01 -6.84668422e-01
7.07306921e-01 1.12399831e-01 -5.77533007e-01 5.76742709e-01
5.84465228e-02 8.51444781e-01 1.10964179e+00 -1.24945664e+00
-3.46035779e-01 -1.35060146e-01 -6.64104000e-02 -6.89073801e-02
4.47384864e-01 3.21912691e-02 1.19382657e-01 -8.64286721e-01
1.44293280e-02 -1.09281766e+00 -1.15751691e-01 -4.82961088e-02
-3.95509869e-01 6.74432218e-01 1.99343637e-01 -7.70799577e-01
6.43282473e-01 -1.90854752e+00 -4.52552110e-01 1.64198428e-01
6.82732224e-01 5.80976486e-01 -3.24457645e-01 -8.11668411e-02
-5.76689303e-01 3.37660313e-01 -1.53802127e-01 6.68919235e-02
-8.83438349e-01 -1.38062507e-01 -8.97426382e-02 4.05313045e-01
2.28481114e-01 1.30755770e+00 -9.08742428e-01 -2.45183289e-01
3.79134893e-01 7.85921097e-01 -5.47339320e-01 2.87273854e-01
1.56872541e-01 1.21094608e+00 8.79141018e-02 4.44270015e-01
6.21593237e-01 -3.73595893e-01 -7.82822892e-02 -3.95139188e-01
2.06502020e-01 -7.55132213e-02 -4.23759311e-01 1.17161429e+00
-6.30026579e-01 1.18019462e+00 -4.01379555e-01 -2.23351359e-01
7.94078171e-01 3.95606399e-01 5.01392961e-01 -9.91021574e-01
5.42507805e-02 4.50717866e-01 5.16167343e-01 -4.61744994e-01
-3.67294550e-02 -8.74006271e-01 4.70685095e-01 3.27272326e-01
-3.37310165e-01 -7.30732262e-01 -3.91382992e-01 -1.74633875e-01
1.05653811e+00 -1.13554403e-01 2.01097485e-02 1.57163143e-01
1.60964012e-01 1.36336103e-01 -1.61412749e-02 7.04999328e-01
9.41108987e-02 1.33725667e+00 5.28812051e-01 -5.12511730e-01
-1.37533331e+00 -1.07680249e+00 -5.31718612e-01 -2.15184629e-01
-3.29970747e-01 -1.78449489e-02 -7.33203292e-01 -4.16180938e-01
-4.40783590e-01 4.70246673e-01 -7.75182366e-01 -4.52289991e-02
-4.49306816e-01 -7.28998005e-01 5.37603557e-01 3.71267021e-01
5.59820235e-01 -9.56012368e-01 -7.12926686e-01 2.04925403e-01
-1.80444062e-01 -1.19820881e+00 -1.13484450e-01 -5.77914655e-01
-8.39707613e-01 -1.28580427e+00 -1.01242113e+00 -6.13406897e-01
7.93083429e-01 -3.12575638e-01 1.25015008e+00 -1.59456223e-01
-7.84930289e-01 2.05175713e-01 -1.80671349e-01 -4.67393398e-01
-9.73359585e-01 -5.02440333e-01 -2.45384797e-01 -9.70153362e-02
-1.24556288e-01 -7.39743650e-01 -1.24010611e+00 2.01891080e-01
-7.90943444e-01 2.31141970e-01 1.01425374e+00 1.18175721e+00
9.13456202e-01 -1.19109392e-01 2.76564479e-01 -1.04820466e+00
7.24118233e-01 8.06012657e-03 -6.17312729e-01 -3.02978873e-01
-5.36947906e-01 -3.04503649e-01 4.92862523e-01 -3.81278962e-01
-8.23217988e-01 -4.06392992e-01 -2.07627982e-01 -4.83850926e-01
9.18124709e-03 2.27601305e-01 6.33531451e-01 -5.72933316e-01
1.15356910e+00 1.01533912e-01 4.37546462e-01 4.25767861e-02
-5.42010441e-02 6.12656176e-01 7.06605792e-01 -4.56971154e-02
6.83366299e-01 5.64563334e-01 2.13734493e-01 -5.29987693e-01
-6.94687426e-01 1.54318959e-01 -1.32890433e-01 -1.56774685e-01
1.00561416e+00 -1.01210093e+00 -4.25281912e-01 5.78364015e-01
-7.23610282e-01 -5.99711061e-01 -6.91090643e-01 9.03930187e-01
-5.73552907e-01 -5.34388721e-02 -6.14033222e-01 -3.92369419e-01
-4.33818191e-01 -1.26108646e+00 1.08170605e+00 1.02977879e-01
-2.87132978e-01 -8.24608862e-01 2.15687647e-01 8.73694420e-01
7.63485074e-01 1.15208924e+00 1.08801389e+00 -2.27442626e-02
-5.66808343e-01 -4.05359238e-01 -3.98461521e-01 7.76211619e-01
3.50711167e-01 -7.14069158e-02 -1.12553525e+00 -5.78081727e-01
7.10574389e-02 -4.60040003e-01 5.55684090e-01 9.52703416e-01
1.22988391e+00 -3.64080399e-01 1.67398676e-01 9.86042202e-01
1.64404273e+00 2.40947902e-01 1.44271374e+00 1.40681744e-01
4.40914571e-01 2.47668594e-01 2.46064126e-01 7.22234845e-02
-3.78717095e-01 5.57480156e-01 4.24656183e-01 -8.31727862e-01
-8.02469611e-01 -2.77026713e-01 -1.48613155e-01 4.62031774e-02
-5.30053437e-01 -1.27353147e-01 -7.36670554e-01 4.38813418e-01
-7.10169733e-01 -8.94070983e-01 -1.89294294e-01 2.24365830e+00
7.89845228e-01 1.12864740e-01 1.52633423e-02 -1.65802002e-01
7.25804865e-01 -7.19718561e-02 -6.33730352e-01 -3.45243067e-01
-2.65261263e-01 1.15280175e+00 4.61508036e-01 3.59014332e-01
-6.28288209e-01 5.18942893e-01 7.09800816e+00 5.23962796e-01
-1.66974604e+00 2.12898865e-01 9.86159205e-01 -4.10389751e-01
-3.53707045e-01 -2.60936707e-01 -1.28182769e-02 5.59560061e-01
1.10150242e+00 1.54294656e-03 1.59991652e-01 3.03197920e-01
3.99462312e-01 -2.81288683e-01 -6.41498625e-01 1.13145852e+00
-1.47442035e-02 -1.76605320e+00 3.35087448e-01 3.91478300e-01
1.09133494e+00 1.96570054e-01 5.48877835e-01 -2.66240418e-01
1.49896696e-01 -1.85803890e+00 1.94533821e-02 5.70467353e-01
1.98679626e+00 -6.45669162e-01 9.74127412e-01 -2.60261208e-01
-4.01587248e-01 2.08340481e-01 6.34780452e-02 3.49795192e-01
1.17617495e-01 5.82890511e-01 -1.19085658e+00 1.80455759e-01
4.94769126e-01 3.97845209e-01 -6.34560823e-01 9.69861805e-01
-1.52427599e-01 6.38040125e-01 1.10114207e-02 6.06003582e-01
3.91450264e-02 -1.08979769e-01 6.58066094e-01 6.35569990e-01
4.11338031e-01 3.32377434e-01 -7.20824838e-01 9.57724273e-01
1.05587812e-02 -7.47122392e-02 -4.76941913e-01 -8.24215263e-02
1.95133016e-01 1.06242049e+00 -3.43324423e-01 -1.38725564e-01
-2.09041908e-01 6.68838441e-01 -2.99068928e-01 3.86758089e-01
-6.44727945e-01 -2.35315979e-01 3.85778159e-01 8.39702666e-01
3.72599401e-02 1.45332113e-01 -4.32523161e-01 -8.47812414e-01
1.93065301e-01 -1.15781546e+00 4.15199623e-02 -1.51462698e+00
-1.04450107e+00 1.15853214e+00 -3.93371135e-01 -1.74447298e+00
-3.96074116e-01 -3.60155404e-01 -5.73750377e-01 1.31612337e+00
-1.44998431e+00 -1.27636075e+00 -8.74359727e-01 3.74503762e-01
4.08470519e-02 -2.81071037e-01 9.11957979e-01 4.80454378e-02
3.91423367e-02 5.76969683e-01 -1.56775534e-01 1.36209890e-01
7.67112374e-01 -1.08435440e+00 2.95197308e-01 8.48410428e-01
-2.25663915e-01 4.97854203e-01 5.04377723e-01 -6.28244340e-01
-7.14273036e-01 -1.25723875e+00 2.56733954e-01 -3.73880297e-01
1.61423400e-01 3.00092578e-01 -4.71935123e-01 7.24809885e-01
2.48319030e-01 2.41780147e-01 8.59285116e-01 -6.67263150e-01
4.59472463e-02 7.52798840e-02 -1.58293462e+00 5.63114285e-01
6.59340203e-01 -3.30727309e-01 -1.88867882e-01 3.60429913e-01
4.35314596e-01 -9.64731216e-01 -1.28716493e+00 6.36461616e-01
6.64442122e-01 -1.58422160e+00 9.13968801e-01 -4.51496691e-01
1.20119369e+00 -3.38247158e-02 2.64524847e-01 -1.38177538e+00
-4.57023121e-02 -7.12986887e-01 3.96829695e-01 4.00148839e-01
3.45576942e-01 -1.15241051e+00 8.14969063e-01 5.91986418e-01
-9.00809690e-02 -8.71553361e-01 -8.78900886e-01 -4.09897029e-01
1.80999070e-01 -8.34861472e-02 5.09946585e-01 6.92298770e-01
-6.86602771e-01 -4.05414775e-02 -1.08584188e-01 -7.96646699e-02
5.05779684e-01 -7.76382461e-02 7.41236985e-01 -9.85724211e-01
-3.41918647e-01 -3.67359400e-01 -1.02704418e+00 -3.64256710e-01
-2.17533156e-01 -7.20287025e-01 -6.30607069e-01 -1.40377152e+00
6.34494275e-02 -6.13312781e-01 7.85459131e-02 3.06720018e-01
-2.37915777e-02 1.21529436e+00 -1.95297763e-01 4.61594790e-01
5.79107881e-01 3.67092341e-02 2.07657123e+00 1.61241665e-01
-9.59771052e-02 1.87874436e-01 -8.93421292e-01 2.96076924e-01
6.98345900e-01 -2.00035781e-01 -5.90648651e-01 -2.78775990e-01
3.11183576e-02 3.99439901e-01 8.56569469e-01 -1.36509573e+00
-2.68893123e-01 3.46235454e-01 8.61588299e-01 -8.64259526e-02
5.21903396e-01 -4.24803436e-01 7.29434490e-01 7.04553246e-01
-1.90796703e-01 -2.19598413e-01 3.89936835e-01 1.70266122e-01
-3.70856285e-01 2.80744910e-01 1.35772729e+00 -3.59762788e-01
-6.09815679e-02 2.45105356e-01 -3.06815058e-02 3.29857588e-01
1.03322220e+00 -6.52816832e-01 -4.28477049e-01 -7.04523563e-01
-8.12290430e-01 -6.09016180e-01 9.30736542e-01 -1.09216206e-01
8.95098925e-01 -9.59523320e-01 -1.18207157e+00 7.15559900e-01
1.32500634e-01 1.93873554e-01 4.52202678e-01 9.53291297e-01
-1.16862595e+00 1.39620185e-01 -5.52581728e-01 -8.07174742e-01
-1.19583988e+00 2.62071233e-04 8.83642673e-01 -2.51249801e-02
-8.32503080e-01 8.69901717e-01 8.28283490e-04 1.41824022e-01
-5.36669254e-01 -1.18538879e-01 5.03860973e-02 -4.10289884e-01
4.46705163e-01 -1.10284097e-01 2.56851792e-01 -4.11936611e-01
2.49594256e-01 8.38563740e-01 -1.61696561e-02 -2.49818832e-01
1.25087106e+00 3.12154204e-01 -1.04211040e-01 -1.12361677e-01
1.17525065e+00 3.76171768e-01 -1.08137202e+00 3.78796980e-02
-9.68186080e-01 -8.96213710e-01 3.50591004e-01 -1.11494398e+00
-1.36881030e+00 5.38316846e-01 9.47086811e-01 -2.81010680e-02
1.32235134e+00 -1.60549060e-01 8.36601138e-01 -5.60236454e-01
2.89954424e-01 -2.78779149e-01 7.85480961e-02 -3.83712411e-01
1.06258631e+00 -1.20878649e+00 -3.98356393e-02 -3.04395407e-01
-6.69158459e-01 1.08579421e+00 4.40184593e-01 -3.22086871e-01
1.85707435e-01 2.57876605e-01 4.46793169e-01 -2.71447569e-01
-2.48635143e-01 1.71319813e-01 5.25892615e-01 9.22383249e-01
3.88299674e-01 9.95809399e-03 -1.67648688e-01 -9.65088606e-02
-8.58585000e-01 3.26275527e-01 1.03718436e+00 5.15876710e-01
1.49395376e-01 -8.83917868e-01 -1.83321804e-01 9.83612835e-01
-6.79949939e-01 -4.07787293e-01 -3.92986946e-02 8.77433598e-01
1.90422863e-01 6.28884494e-01 2.90675104e-01 -1.38944164e-01
9.16084871e-02 -4.36355829e-01 6.82445765e-01 -7.11850822e-01
-5.37077427e-01 -1.61819756e-02 2.34772012e-01 -6.64258957e-01
-4.69427794e-01 -3.44254166e-01 -5.32776892e-01 -1.70323342e-01
-1.84162594e-02 -1.95613444e-01 6.48365736e-01 3.64209384e-01
3.16368192e-01 7.79200375e-01 7.01220393e-01 -5.44453144e-01
-1.46077067e-01 -9.39585447e-01 -5.09521723e-01 3.97022814e-01
7.05313563e-01 -2.25733206e-01 -6.62756026e-01 1.83460325e-01] | [14.290935516357422, -1.9654408693313599] |
b34bfd02-e81b-4981-aa20-7aa18d568bf0 | end-to-end-estimation-of-multi-person-3d | 2004.06239 | null | https://arxiv.org/abs/2004.06239v4 | https://arxiv.org/pdf/2004.06239v4.pdf | VoxelPose: Towards Multi-Camera 3D Human Pose Estimation in Wild Environment | We present an approach to estimate 3D poses of multiple people from multiple camera views. In contrast to the previous efforts which require to establish cross-view correspondence based on noisy and incomplete 2D pose estimations, we present an end-to-end solution which directly operates in the $3$D space, therefore avoids making incorrect decisions in the 2D space. To achieve this goal, the features in all camera views are warped and aggregated in a common 3D space, and fed into Cuboid Proposal Network (CPN) to coarsely localize all people. Then we propose Pose Regression Network (PRN) to estimate a detailed 3D pose for each proposal. The approach is robust to occlusion which occurs frequently in practice. Without bells and whistles, it outperforms the state-of-the-arts on the public datasets. Code will be released at https://github.com/microsoft/multiperson-pose-estimation-pytorch. | ['Wen-Jun Zeng', 'Hanyue Tu', 'Chunyu Wang'] | 2020-04-13 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/738_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123460188.pdf | eccv-2020-8 | ['3d-multi-person-pose-estimation'] | ['computer-vision'] | [-4.42958772e-01 -1.15197569e-01 1.58729777e-01 -4.65348214e-01
-9.36526060e-01 -6.99491501e-01 4.56046373e-01 -2.21701160e-01
-3.87124538e-01 5.74284375e-01 3.36039394e-01 2.70148695e-01
2.19282731e-01 -4.43692148e-01 -6.27974570e-01 -2.46598825e-01
5.08317053e-02 8.96451473e-01 -3.70729044e-02 1.21077053e-01
9.17378217e-02 5.51124036e-01 -1.18105185e+00 -1.99074671e-01
2.26227075e-01 6.66183412e-01 -2.67140418e-01 7.63324440e-01
4.22055066e-01 1.15562484e-01 -4.37168360e-01 -5.92757821e-01
7.26722062e-01 -1.15220025e-01 -5.10141075e-01 5.61599731e-01
1.13608825e+00 -5.98008633e-01 -3.07960302e-01 8.59062850e-01
6.70192719e-01 1.83461562e-01 3.68085742e-01 -1.17762184e+00
-1.41625598e-01 -1.63622007e-01 -9.73942041e-01 -1.13749430e-01
1.04135096e+00 -2.52165142e-02 7.31760621e-01 -1.41859508e+00
7.46485651e-01 1.67879558e+00 1.00993979e+00 4.73643690e-01
-1.15587795e+00 -6.04257345e-01 3.25872570e-01 -2.00826198e-01
-1.88100660e+00 -5.28580546e-01 6.49396598e-01 -2.29103267e-01
7.56263614e-01 1.89611733e-01 1.00429797e+00 1.22993565e+00
-3.41554396e-02 6.87188208e-01 9.93710518e-01 -3.24196965e-01
-2.73033589e-01 7.50367641e-02 -1.54915050e-01 8.79612982e-01
4.68762845e-01 6.65131770e-03 -7.27771997e-01 -3.14686984e-01
1.04171908e+00 4.64180917e-01 -1.45844638e-01 -9.53791261e-01
-1.27332211e+00 6.41639352e-01 3.73545736e-01 -3.69649678e-01
-3.41900289e-01 3.09453756e-01 5.58759496e-02 -1.73901722e-01
5.84553421e-01 1.59828752e-01 -4.41423804e-01 -2.38750905e-01
-8.08367908e-01 8.39197814e-01 7.76352763e-01 1.08563602e+00
6.75833225e-01 -4.25548911e-01 2.44849846e-01 6.42930388e-01
4.60259736e-01 6.24411821e-01 -2.28966549e-01 -1.40495849e+00
7.07753420e-01 5.51305234e-01 4.61309493e-01 -1.16495860e+00
-4.94960457e-01 -3.29686165e-01 -6.03583395e-01 1.09645106e-01
6.98145628e-01 -4.25201178e-01 -6.93535030e-01 1.48905659e+00
8.40079010e-01 6.69698492e-02 -3.18379968e-01 1.16630650e+00
6.16267025e-01 2.82589108e-01 -4.33489323e-01 1.52455598e-01
1.43358362e+00 -1.06008267e+00 -3.27180862e-01 -5.89890778e-01
1.44945294e-01 -8.72468114e-01 5.56696236e-01 5.25516510e-01
-1.26708460e+00 -6.48881376e-01 -8.24059784e-01 -1.17282711e-01
-1.54955357e-01 2.58110940e-01 3.18201929e-01 5.85390031e-01
-1.12989795e+00 3.85942668e-01 -9.68340755e-01 -7.89275408e-01
3.10041100e-01 5.58084667e-01 -7.28725970e-01 -2.66620725e-01
-6.04580045e-01 1.12400508e+00 4.07567918e-02 3.62542927e-01
-7.54905343e-01 -4.03835922e-01 -8.82733524e-01 -4.66711849e-01
7.58725703e-01 -1.05205703e+00 1.28543544e+00 -4.58011329e-01
-1.25160372e+00 8.36018920e-01 -3.42695177e-01 -9.37448069e-02
9.17910457e-01 -8.25418532e-01 6.07677735e-02 1.75552756e-01
2.68923044e-01 8.88688445e-01 6.00976467e-01 -1.46594357e+00
-5.05974174e-01 -7.68869817e-01 1.20284386e-01 7.79761791e-01
2.08247393e-01 6.16594329e-02 -1.14958894e+00 -4.15309131e-01
5.01409233e-01 -1.23449898e+00 -4.83979613e-01 3.35665375e-01
-7.33946085e-01 -8.00900906e-02 6.15306020e-01 -7.02637613e-01
6.91617489e-01 -1.72568464e+00 3.68745089e-01 3.52095783e-01
5.09527385e-01 -1.14529453e-01 1.58689588e-01 5.16279340e-01
9.36586559e-02 -2.44402170e-01 2.12759778e-01 -1.03100920e+00
7.68840313e-02 -1.38596907e-01 3.23248863e-01 1.05541587e+00
-1.75909221e-01 6.69138014e-01 -5.71245313e-01 -5.40415943e-01
3.33678216e-01 7.33690262e-01 -6.73648596e-01 2.01838568e-01
3.07678938e-01 4.59710062e-01 -4.55286622e-01 8.65003586e-01
7.50414371e-01 -3.10842812e-01 2.70016402e-01 -1.37375146e-01
-5.74116409e-02 -1.66142024e-02 -1.61855578e+00 1.85712314e+00
-1.28019169e-01 2.83143610e-01 2.52141118e-01 -5.39898455e-01
8.27962875e-01 3.16380203e-01 5.37320614e-01 5.87699749e-02
1.81589603e-01 -1.51831573e-02 -5.08461475e-01 -1.25305474e-01
4.33951139e-01 5.70242815e-02 -3.70943606e-01 2.58238316e-01
5.45351133e-02 6.26765564e-03 2.24261731e-03 6.97369203e-02
7.20856011e-01 4.74181890e-01 4.66546983e-01 6.89509064e-02
4.35906500e-01 -1.20743647e-01 5.69255292e-01 5.41311562e-01
-3.06602746e-01 1.02915967e+00 2.70571917e-01 -8.39590728e-01
-1.19694054e+00 -1.35433853e+00 1.96814686e-01 6.63381398e-01
3.03610593e-01 -7.31496215e-01 -7.16568351e-01 -7.74339557e-01
1.75724283e-01 1.95684820e-01 -4.37716603e-01 4.47962552e-01
-7.19720840e-01 -3.03166151e-01 2.31715903e-01 6.20442212e-01
5.07540941e-01 -3.19232970e-01 -5.74681938e-01 -8.06674920e-03
-4.55050945e-01 -1.31024599e+00 -9.11805272e-01 -4.47005779e-01
-5.62573195e-01 -1.21295202e+00 -8.33841145e-01 -4.22608912e-01
9.27786648e-01 5.23330212e-01 1.07589781e+00 -1.42322153e-01
-1.99987531e-01 7.84242868e-01 -4.94722687e-02 -3.56540829e-01
2.38808677e-01 -1.58447713e-01 5.30414104e-01 -1.11090764e-01
7.34771430e-01 -5.31937778e-01 -8.54357719e-01 4.58518416e-01
1.01443350e-01 -3.68650295e-02 3.81680518e-01 3.99792612e-01
5.70782483e-01 -2.66087353e-01 -2.25700200e-01 -4.74696338e-01
3.29556555e-01 -2.66323909e-02 -7.75998890e-01 -1.08348876e-02
-9.88293141e-02 -4.47959244e-01 1.07210934e-01 -2.82946646e-01
-6.39999270e-01 5.52171767e-01 7.96019379e-03 -7.39544868e-01
-3.96576613e-01 8.38352665e-02 -4.94919121e-02 -3.60024944e-02
5.82338452e-01 -1.35043412e-01 2.61079043e-01 -6.48275375e-01
2.44424403e-01 3.60403866e-01 4.39130574e-01 -4.76205885e-01
1.05796802e+00 6.92207932e-01 8.26505199e-03 -6.81607246e-01
-1.10099173e+00 -7.07855999e-01 -1.12420464e+00 -4.10374671e-01
9.28890049e-01 -1.39374483e+00 -1.03899193e+00 2.47542918e-01
-1.34178758e+00 1.71515957e-01 -4.27937321e-02 7.25754619e-01
-5.66459715e-01 4.87929672e-01 -3.85806948e-01 -9.31131124e-01
-2.31617570e-01 -1.12397575e+00 1.38797867e+00 1.52760312e-01
-5.51690638e-01 -7.64383614e-01 2.59092867e-01 5.62544346e-01
-1.04765065e-01 3.47293258e-01 -1.60429776e-01 -3.94901842e-01
-6.00032687e-01 -5.69912553e-01 -1.37264371e-01 7.34909698e-02
-1.25116259e-01 -2.46816859e-01 -7.61440098e-01 -5.98796129e-01
-1.58634365e-01 -2.18749389e-01 4.88326401e-01 6.88025177e-01
6.78664148e-01 -2.44681224e-01 -5.56400001e-01 6.09923065e-01
1.26862061e+00 -3.64261031e-01 1.98053569e-01 2.03814968e-01
7.88947642e-01 7.27295816e-01 4.84946311e-01 5.52382588e-01
9.68820214e-01 1.06090713e+00 4.45048809e-01 -1.57215402e-01
7.20384046e-02 -4.15557235e-01 3.31351668e-01 2.86387682e-01
-5.29213965e-01 -1.05520589e-02 -9.76188600e-01 2.75455415e-01
-1.91548777e+00 -8.85082901e-01 3.70326079e-02 2.16655087e+00
2.56097794e-01 2.33807340e-02 5.75407207e-01 -4.63322282e-01
8.99168253e-01 2.32023820e-01 -3.64765793e-01 2.31468469e-01
3.21089119e-01 -2.71539360e-01 6.39314532e-01 7.24905670e-01
-1.24708545e+00 8.54166627e-01 6.24643707e+00 2.11421415e-01
-4.30285305e-01 1.26080647e-01 4.29845154e-01 -4.85751688e-01
3.01919758e-01 5.24745695e-02 -1.39750361e+00 2.33336613e-02
4.81391549e-01 4.29993331e-01 4.88574773e-01 7.82925785e-01
2.35042766e-01 -1.97338700e-01 -1.12345147e+00 1.44772267e+00
4.00280207e-01 -1.11746323e+00 -3.04742277e-01 4.09652203e-01
5.07590413e-01 4.81442697e-02 -7.79710785e-02 8.44240561e-02
4.58307475e-01 -7.99564838e-01 7.81937778e-01 5.75551152e-01
6.31231129e-01 -8.25172126e-01 6.69277728e-01 4.73461330e-01
-1.32427716e+00 5.57188690e-02 -3.42522502e-01 -1.46499708e-01
4.23002303e-01 3.59649569e-01 -8.29440475e-01 3.99742395e-01
9.73715842e-01 7.04241097e-01 -5.77590525e-01 9.12553012e-01
-1.27949581e-01 -1.79168478e-01 -5.88498056e-01 1.03700608e-01
9.79361236e-02 -3.00306708e-01 6.75819755e-01 9.07585084e-01
5.99496841e-01 -4.55186069e-02 7.30334282e-01 5.54938138e-01
1.87739935e-02 -2.22214684e-01 -6.82953715e-01 4.38219011e-01
4.83797282e-01 1.34155715e+00 -6.79286659e-01 -2.09314167e-01
-4.82279420e-01 1.05967486e+00 3.79830658e-01 2.61208773e-01
-8.86949718e-01 1.37105629e-01 7.16052949e-01 3.80995661e-01
4.65404212e-01 -5.62624395e-01 1.41086727e-01 -1.42267370e+00
3.99137348e-01 -8.61047924e-01 2.64085621e-01 -6.96764588e-01
-1.40477264e+00 4.84425068e-01 3.82253706e-01 -1.40814817e+00
-3.80406529e-01 -6.02092743e-01 -2.74066031e-01 9.01437283e-01
-8.69118214e-01 -1.44381225e+00 -3.02233130e-01 6.43660188e-01
6.28198504e-01 1.34317512e-02 7.02302098e-01 1.27786428e-01
-4.18477654e-01 3.99225503e-01 -4.32220161e-01 3.41899633e-01
9.75673139e-01 -1.10825872e+00 6.83323205e-01 8.57948720e-01
1.42609999e-01 7.74318576e-01 7.71823764e-01 -6.72188222e-01
-1.35609376e+00 -8.84225368e-01 9.63858902e-01 -1.15727150e+00
1.97490141e-01 -7.77621806e-01 -1.78974971e-01 1.06405890e+00
1.17241919e-01 2.90167719e-01 5.41069627e-01 4.48577106e-01
-3.61640424e-01 -2.83425637e-02 -1.11897147e+00 7.01442242e-01
1.22807562e+00 -1.92989349e-01 -4.47091579e-01 4.39505965e-01
2.56685376e-01 -8.87231112e-01 -7.28211522e-01 9.13855210e-02
8.86395812e-01 -1.12592208e+00 1.44719267e+00 -2.85301983e-01
-1.72761381e-02 -4.48884517e-01 -3.44561487e-01 -1.11561465e+00
-2.14391142e-01 -6.81686699e-01 -1.66787192e-01 8.79856408e-01
2.73820460e-01 -4.87689823e-01 1.15962267e+00 5.43860376e-01
1.96706414e-01 -6.71512902e-01 -1.14274204e+00 -5.43058753e-01
-3.25106472e-01 -3.33291352e-01 3.86866152e-01 5.79353213e-01
-2.32217908e-01 4.10676509e-01 -7.03149676e-01 5.34193575e-01
1.00468326e+00 -1.00742586e-01 1.32708085e+00 -1.24620247e+00
-2.11226150e-01 -3.13662030e-02 -3.91996950e-01 -1.34198785e+00
-1.28410250e-01 -3.21994871e-01 3.69469300e-02 -1.43806553e+00
2.96775669e-01 -6.77993223e-02 3.13617617e-01 2.43778363e-01
-6.32692650e-02 6.32985950e-01 5.15388608e-01 2.41102010e-01
-9.13222134e-01 3.17692757e-01 1.12738991e+00 2.66061187e-01
-1.72263712e-01 2.70775050e-01 -6.93419516e-01 1.19608676e+00
7.43173420e-01 -4.18340862e-01 -7.08591565e-02 -4.19451088e-01
5.53423874e-02 2.27562830e-01 8.59751463e-01 -1.15375054e+00
2.84100950e-01 -2.56911814e-02 1.02430201e+00 -1.27999544e+00
1.15990746e+00 -9.02242839e-01 3.34470600e-01 2.95544803e-01
9.42551345e-02 5.63899040e-01 -1.60773814e-01 5.50348878e-01
1.70210257e-01 2.31723532e-01 6.35571003e-01 -6.20743811e-01
-2.93784827e-01 6.09343648e-01 2.21005157e-02 -1.23962805e-01
1.07672763e+00 -5.42643845e-01 -2.58589946e-02 -6.77608311e-01
-8.42335224e-01 2.73013562e-01 7.49603748e-01 3.56703937e-01
6.98066413e-01 -1.38697422e+00 -7.08119988e-01 1.81966856e-01
-1.17421657e-01 2.67702579e-01 3.22674304e-01 8.76727283e-01
-7.00464547e-01 4.35632437e-01 -6.57791691e-03 -8.62386405e-01
-1.44810879e+00 3.78109813e-01 4.84963804e-01 -2.97217548e-01
-5.90341032e-01 8.41669261e-01 -3.60738263e-02 -9.04778957e-01
2.12192759e-01 1.44427299e-01 8.86572152e-02 -9.50016975e-02
5.26187122e-01 5.55314600e-01 -3.20911646e-01 -9.97756481e-01
-6.00100338e-01 1.03218305e+00 -1.64344057e-01 -4.83138263e-01
1.41826892e+00 -5.21735013e-01 3.35778892e-02 1.76593333e-01
1.17290044e+00 2.99783975e-01 -1.71636844e+00 -2.44580567e-01
-4.22187686e-01 -7.94445276e-01 -3.03202599e-01 -4.38046634e-01
-8.88844073e-01 5.24468541e-01 5.19600093e-01 -9.08688307e-02
6.55016541e-01 2.21530125e-01 4.43226099e-01 4.95985955e-01
4.02763486e-01 -1.06699419e+00 1.67748451e-01 3.97142410e-01
9.01896417e-01 -1.17199552e+00 5.64248264e-01 -6.03385866e-01
-6.60935760e-01 1.06395590e+00 9.02721167e-01 -4.71201152e-01
5.00499845e-01 1.44107565e-01 1.72863320e-01 -3.77156615e-01
-4.77436841e-01 3.88450595e-03 3.01245958e-01 7.43338525e-01
2.61002272e-01 -8.22265819e-02 2.81910986e-01 6.84436262e-02
-3.17276955e-01 -2.82117635e-01 3.64251077e-01 9.19874191e-01
-2.07267150e-01 -1.10066485e+00 -9.59200382e-01 1.27252087e-01
-3.83569956e-01 1.93175256e-01 -4.58285838e-01 1.18866801e+00
1.84099540e-01 8.19835126e-01 1.18536368e-01 -2.65880883e-01
6.70992434e-01 -3.59133147e-02 6.36792302e-01 -6.69847488e-01
-3.47412378e-01 4.93704498e-01 2.92228371e-01 -9.18412924e-01
-5.84053218e-01 -1.09836411e+00 -6.79298282e-01 -5.42195141e-01
-1.64546430e-01 -1.14949234e-01 4.31870073e-01 7.18306184e-01
3.29372764e-01 -1.62993312e-01 4.13792878e-01 -1.54323804e+00
-4.95059341e-01 -8.44450533e-01 -4.19949979e-01 2.01887369e-01
3.77253175e-01 -8.33962917e-01 -1.50574312e-01 -1.06513605e-01] | [7.049137592315674, -0.9834847450256348] |
c9981b65-fa26-49c2-a160-861b2a71ad15 | tsdf-a-multi-object-formulation-for-dynamic | 2105.07468 | null | https://arxiv.org/abs/2105.07468v1 | https://arxiv.org/pdf/2105.07468v1.pdf | TSDF++: A Multi-Object Formulation for Dynamic Object Tracking and Reconstruction | The ability to simultaneously track and reconstruct multiple objects moving in the scene is of the utmost importance for robotic tasks such as autonomous navigation and interaction. Virtually all of the previous attempts to map multiple dynamic objects have evolved to store individual objects in separate reconstruction volumes and track the relative pose between them. While simple and intuitive, such formulation does not scale well with respect to the number of objects in the scene and introduces the need for an explicit occlusion handling strategy. In contrast, we propose a map representation that allows maintaining a single volume for the entire scene and all the objects therein. To this end, we introduce a novel multi-object TSDF formulation that can encode multiple object surfaces at any given location in the map. In a multiple dynamic object tracking and reconstruction scenario, our representation allows maintaining accurate reconstruction of surfaces even while they become temporarily occluded by other objects moving in their proximity. We evaluate the proposed TSDF++ formulation on a public synthetic dataset and demonstrate its ability to preserve reconstructions of occluded surfaces when compared to the standard TSDF map representation. | ['Juan Nieto', 'Roland Siegwart', 'Federico Tombari', 'Margarita Grinvald'] | 2021-05-16 | null | null | null | null | ['occlusion-handling'] | ['computer-vision'] | [ 4.26452845e-01 -1.03513353e-01 4.28007513e-01 -2.56182820e-01
-5.82238436e-01 -8.66996050e-01 8.75485897e-01 4.82447684e-01
-2.18673483e-01 5.97126544e-01 -2.85423040e-01 1.57665178e-01
-3.45971465e-01 -9.98043835e-01 -1.12831187e+00 -6.05699122e-01
2.47246877e-04 1.22394586e+00 1.08054829e+00 -4.43343073e-02
2.03027144e-01 1.02065086e+00 -2.07988691e+00 9.50427130e-02
8.23694289e-01 1.03122652e+00 7.87851691e-01 5.08689940e-01
-2.81840771e-01 4.51643050e-01 -5.14873922e-01 -1.53624579e-01
3.11521441e-01 9.92785320e-02 -5.13505578e-01 2.08147645e-01
1.00000215e+00 -2.10940331e-01 -2.78015584e-01 7.79661000e-01
1.34802833e-01 3.30741584e-01 4.64762986e-01 -1.02203238e+00
-9.13502276e-02 -2.23343581e-01 -5.57558477e-01 5.99638745e-02
3.07711929e-01 -2.32687011e-01 4.22458112e-01 -7.32475221e-01
1.06090152e+00 1.27732778e+00 6.25422180e-01 2.05247384e-02
-1.40307820e+00 -2.62908936e-01 3.59957308e-01 1.12684399e-01
-1.32075083e+00 -5.45361042e-01 6.26569986e-01 -5.70575714e-01
6.74263775e-01 5.19341469e-01 6.67028248e-01 3.33200812e-01
3.09466988e-01 4.67304409e-01 7.76371419e-01 -2.19770744e-01
2.32233718e-01 1.14107952e-01 1.96426421e-01 6.52195275e-01
5.22951245e-01 -1.99002624e-01 -6.39898419e-01 -2.34467775e-01
8.26229274e-01 -2.91272253e-02 -1.59535810e-01 -1.24686539e+00
-1.33854413e+00 3.79764855e-01 3.47068340e-01 1.18136980e-01
-3.90320778e-01 4.65130001e-01 7.33162910e-02 -6.05492434e-03
6.38674200e-01 2.14109588e-02 -2.66024619e-01 9.91367698e-02
-8.94232512e-01 5.86227179e-01 5.03930092e-01 1.07329774e+00
9.95927811e-01 -2.31666580e-01 -4.39518271e-03 4.80724961e-01
1.01355545e-01 6.79973483e-01 -3.12449306e-01 -1.12978649e+00
3.53114277e-01 5.01681387e-01 5.17592430e-01 -1.33585632e+00
-2.65094906e-01 -2.45986506e-01 -5.18116772e-01 5.60691893e-01
2.51703113e-01 5.52568853e-01 -8.91892374e-01 1.59475565e+00
8.71563792e-01 4.99211252e-01 -1.45042345e-01 5.83398700e-01
5.62784672e-01 6.77975893e-01 -2.69246697e-01 -1.43267021e-01
1.30903220e+00 -7.66771436e-01 -8.22113454e-01 -3.36966306e-01
8.67617726e-02 -6.86053336e-01 6.03325188e-01 1.88604131e-01
-1.15769219e+00 -4.11770254e-01 -9.82855976e-01 -3.41045409e-01
-3.02785337e-01 -2.40824968e-01 5.72838128e-01 3.48412871e-01
-1.08056056e+00 3.97433877e-01 -1.05292332e+00 -4.40463722e-01
1.65747181e-01 3.80420238e-01 -4.09271836e-01 -2.99236596e-01
-4.70781744e-01 1.05922580e+00 3.08004498e-01 1.11931190e-01
-8.27509463e-01 -7.47281492e-01 -7.04342961e-01 -5.80854341e-02
3.24489295e-01 -7.55733073e-01 9.88631189e-01 -4.24184531e-01
-9.92415488e-01 8.54194283e-01 -6.30603015e-01 -2.38704845e-01
7.68956661e-01 -3.78206939e-01 -3.23812872e-01 1.78839341e-02
2.40532905e-01 4.50352341e-01 6.06579661e-01 -1.77637970e+00
-8.11567843e-01 -5.15351236e-01 2.34849781e-01 4.05695438e-01
2.15296388e-01 -3.19349736e-01 -6.75330043e-01 -3.05743217e-01
6.15660965e-01 -9.54199791e-01 -1.01045437e-01 8.15847218e-01
-1.77366525e-01 2.60993809e-01 1.34549904e+00 -4.77241009e-01
7.55795360e-01 -2.09279895e+00 4.97536838e-01 1.89998984e-01
1.87651396e-01 7.80904386e-03 1.79338396e-01 3.83432239e-01
4.27115053e-01 -3.94299865e-01 -4.12229389e-01 -7.75014102e-01
-1.22329131e-01 5.83409071e-01 -3.27613235e-01 8.75378549e-01
-2.37089097e-01 6.11668766e-01 -8.16079974e-01 -4.08082277e-01
5.29988706e-01 8.26279283e-01 -4.98988748e-01 -1.90156773e-01
-4.73751605e-01 7.08368361e-01 -4.57771659e-01 4.71545875e-01
1.17105341e+00 -1.95276454e-01 4.99524325e-02 -1.59823537e-01
-4.83762056e-01 1.49183571e-01 -1.56483912e+00 1.98050797e+00
-2.08951682e-01 6.51412666e-01 3.76086563e-01 -4.80557501e-01
7.69291699e-01 1.22582175e-01 8.56244326e-01 -7.47049332e-01
-4.17856693e-01 2.82912254e-01 -4.92411613e-01 -1.21649243e-01
9.11973774e-01 -7.80953169e-02 4.27263789e-02 3.08548927e-01
-5.80325127e-01 -2.01095730e-01 3.54217067e-02 9.89952907e-02
1.14605582e+00 2.43005887e-01 6.18771240e-02 -3.43108982e-01
4.80314344e-01 3.42063010e-01 4.04188037e-01 6.16604626e-01
6.23258688e-02 6.06015444e-01 -1.87673911e-01 -4.75306630e-01
-9.94017601e-01 -1.40194440e+00 -5.23157239e-01 7.13942409e-01
9.24418867e-01 -1.09864369e-01 -2.78417826e-01 -1.48832828e-01
3.72056901e-01 5.66652358e-01 -5.51467776e-01 2.68338621e-01
-9.21639562e-01 -3.59177858e-01 -6.28004149e-02 -2.37980462e-03
3.83277416e-01 -7.87800133e-01 -1.22529030e+00 2.84884214e-01
-4.90269542e-01 -9.26658213e-01 -1.14378899e-01 -2.95858947e-03
-9.37003374e-01 -1.08917475e+00 -3.65483850e-01 -7.01365292e-01
7.59946406e-01 6.59518600e-01 1.07691479e+00 -8.13880265e-02
-2.45873094e-01 5.61361551e-01 -2.75322348e-02 -2.40973175e-01
-3.54491591e-01 -2.66560048e-01 -7.29900645e-03 2.00250316e-02
-2.84444153e-01 -5.75773180e-01 -4.21104282e-01 4.93617028e-01
-8.82798910e-01 2.84081757e-01 2.06275526e-02 2.08389089e-01
1.09585226e+00 2.25975469e-01 5.34287049e-03 -7.89642692e-01
-4.36622165e-02 -4.24138457e-01 -9.06806171e-01 4.84811187e-01
-1.09120414e-01 5.21866754e-02 -9.74034145e-03 -3.50735486e-01
-1.02562213e+00 4.82590437e-01 2.66042590e-01 -4.38257933e-01
8.46247301e-02 1.48740098e-01 -9.30924341e-02 -3.53930473e-01
3.09559703e-01 2.67995089e-01 -2.81044543e-02 -7.18880355e-01
3.90356123e-01 -5.56524284e-02 8.23853552e-01 -5.82578540e-01
8.50959837e-01 9.99785483e-01 4.28454369e-01 -7.47479200e-01
-4.48186785e-01 -5.57937801e-01 -9.71360028e-01 -5.07636368e-01
5.61242044e-01 -6.96985424e-01 -5.95696688e-01 4.19332534e-01
-1.32953835e+00 -7.80915767e-02 -3.67771506e-01 9.04257447e-02
-7.94423282e-01 2.87629545e-01 -6.27680402e-03 -7.89266825e-01
2.09772661e-01 -1.07892489e+00 1.41805387e+00 -1.70361578e-01
1.06157720e-01 -9.07289624e-01 3.30060571e-01 4.52469550e-02
4.53912199e-01 8.24828207e-01 7.27326691e-01 2.30460450e-01
-1.33623183e+00 -2.43895799e-01 -1.40829727e-01 -4.51220691e-01
3.95308495e-01 -8.51214007e-02 -8.83447528e-01 -4.63856131e-01
9.98087749e-02 2.83210397e-01 6.65762424e-01 4.39963758e-01
6.83062673e-01 -5.83741143e-02 -8.13895643e-01 5.35693765e-01
1.73824728e+00 3.23777705e-01 7.07257390e-01 5.13861358e-01
7.22662508e-01 8.16525578e-01 8.12530220e-01 3.22202832e-01
4.78650272e-01 1.21981549e+00 7.69318402e-01 6.02110103e-02
-2.98927695e-01 -1.67326033e-02 -1.30987717e-02 4.81786758e-01
-3.29918601e-02 -5.74033380e-01 -9.59739745e-01 4.71263379e-01
-1.99604452e+00 -8.39544833e-01 -7.52659976e-01 2.49121046e+00
1.42938808e-01 -1.17516160e-01 -2.36641005e-01 -4.66474593e-02
7.57006884e-01 2.06233323e-01 -6.93324506e-01 2.40905911e-01
-3.73606443e-01 -1.23901211e-01 5.65834403e-01 8.23606133e-01
-9.71780539e-01 6.59497797e-01 6.35496283e+00 4.54820156e-01
-8.84484947e-01 3.48945260e-01 -1.15128592e-01 -2.20936909e-01
-3.85237783e-01 5.76677695e-02 -8.24902177e-01 3.18361044e-01
7.59293556e-01 -2.20229745e-01 2.30316415e-01 5.08202493e-01
6.37710989e-02 -5.79972625e-01 -9.93690431e-01 8.21431756e-01
5.30853216e-03 -1.48018491e+00 5.96936531e-02 1.77254066e-01
6.18577361e-01 1.16000501e-02 3.22787166e-02 -2.76748896e-01
9.60959271e-02 -6.43075645e-01 1.39057851e+00 7.85432696e-01
7.62223899e-01 -5.43646157e-01 2.01883152e-01 5.89255333e-01
-1.62917185e+00 1.86222479e-01 -3.54724109e-01 1.40422046e-01
6.98384762e-01 4.37164754e-01 -7.39360094e-01 7.86801159e-01
6.60189211e-01 5.19515991e-01 -6.62733018e-01 1.37254822e+00
4.70256090e-01 -2.40650758e-01 -7.01395690e-01 5.31700194e-01
-1.12519562e-01 -1.94006696e-01 7.94841230e-01 6.82817280e-01
5.15214622e-01 -6.77489042e-02 2.17221290e-01 6.32351518e-01
3.05908054e-01 -1.22616507e-01 -6.95429802e-01 4.35285658e-01
5.15419781e-01 7.43517041e-01 -1.10525560e+00 -3.93860459e-01
-3.01978946e-01 1.03756881e+00 4.20932740e-01 1.64079502e-01
-9.83411252e-01 2.14659631e-01 7.88183331e-01 6.50201321e-01
4.06279057e-01 -6.41400337e-01 -3.24914902e-01 -9.21833098e-01
4.03715879e-01 -1.39597863e-01 5.85606284e-02 -7.65711486e-01
-7.32201278e-01 6.38149023e-01 1.93369806e-01 -1.44414616e+00
3.41868438e-02 -2.16634184e-01 1.29097980e-02 8.08527172e-01
-1.46205950e+00 -1.15655935e+00 -6.71791971e-01 6.26860797e-01
4.55383897e-01 4.29010004e-01 7.08351135e-01 6.04355991e-01
-8.10707808e-02 -1.95592791e-01 5.30996680e-01 -7.51200140e-01
4.20502812e-01 -9.09924626e-01 4.82025325e-01 9.60848153e-01
1.48068607e-01 4.65473145e-01 1.03335905e+00 -9.86332655e-01
-1.38590264e+00 -1.11268377e+00 6.60632789e-01 -7.82064915e-01
1.29663914e-01 -7.08473265e-01 -1.10499930e+00 8.23749065e-01
-3.78779024e-01 2.20351294e-01 -5.92159107e-02 -3.18971187e-01
-4.00341265e-02 -1.50964186e-01 -1.32351685e+00 1.92458749e-01
1.30656445e+00 -3.48701596e-01 -2.42968947e-01 4.76929337e-01
6.96586490e-01 -8.21328819e-01 -6.11800849e-01 5.89451909e-01
6.73065782e-01 -1.10879588e+00 1.36880898e+00 4.34987731e-02
-1.97779447e-01 -8.29236984e-01 -5.81151843e-01 -8.59272718e-01
-2.62919426e-01 -2.26963371e-01 -4.09916937e-01 9.89743412e-01
-1.43402755e-01 -5.88359177e-01 8.89765263e-01 6.69794917e-01
-4.80326742e-01 -3.45475703e-01 -1.26454091e+00 -8.51331949e-01
-5.02390921e-01 -2.99211890e-01 5.62197328e-01 7.05799341e-01
-8.64257216e-01 -2.55225241e-01 -4.14216548e-01 7.56208718e-01
8.55282247e-01 3.95276278e-01 1.04130816e+00 -1.74748623e+00
-4.30972464e-02 -1.02556773e-01 -5.51895559e-01 -1.24406588e+00
-1.11345634e-01 -7.53997386e-01 2.70187974e-01 -1.88897240e+00
4.84841364e-03 -1.10069907e+00 9.61074010e-02 1.21209279e-01
3.15576017e-01 2.57365912e-01 7.36877993e-02 5.52633107e-01
-6.49644554e-01 5.17554283e-01 1.25697434e+00 -4.53506820e-02
-8.20867270e-02 -8.55787098e-02 -9.19610411e-02 4.74494725e-01
2.21028298e-01 -6.22128308e-01 -5.17194748e-01 -8.56880128e-01
-1.91822741e-02 2.05247745e-01 6.37138128e-01 -1.30290866e+00
3.92980993e-01 -1.84135184e-01 1.30710244e-01 -1.18214703e+00
1.10408247e+00 -1.29998493e+00 1.36567545e+00 4.01803881e-01
2.50076026e-01 2.45148599e-01 4.49405998e-01 9.90812600e-01
-5.51813170e-02 -2.71837376e-02 8.07151794e-01 -1.66260615e-01
-9.04855072e-01 3.53091687e-01 -1.84018537e-02 -4.72116381e-01
1.46474874e+00 -5.27377427e-01 -4.03844982e-01 -2.81248428e-02
-6.68375790e-01 1.75353706e-01 1.21612680e+00 5.24634063e-01
6.49348378e-01 -1.11174631e+00 -3.36554766e-01 2.96139032e-01
1.08659327e-01 5.21939278e-01 5.13467550e-01 6.32671237e-01
-1.05327547e+00 4.51340705e-01 -3.39403957e-01 -1.07067752e+00
-1.45206130e+00 4.97792959e-01 2.71613419e-01 -3.10578849e-02
-1.16293991e+00 6.30636334e-01 5.78740895e-01 -3.20125908e-01
2.30208576e-01 -3.84955853e-01 1.85854614e-01 -6.19016364e-02
3.62271309e-01 5.28645813e-01 3.37128192e-01 -9.74144757e-01
-3.84228230e-01 9.33289587e-01 1.60338596e-01 -8.50856453e-02
1.45986140e+00 -5.40781736e-01 -2.91358620e-01 7.47792840e-01
6.99431300e-01 1.79782465e-01 -1.49448562e+00 -2.84883916e-01
5.55723906e-02 -9.50002611e-01 -1.13303721e-01 -4.74980175e-01
-7.43790925e-01 4.75508630e-01 7.84914494e-01 1.40019283e-01
8.43784213e-01 1.87651455e-01 5.69797575e-01 3.04595828e-01
1.34072936e+00 -4.51516151e-01 -1.19374484e-01 3.47716898e-01
9.77946460e-01 -8.40871096e-01 2.89913416e-01 -8.76553297e-01
-1.86419174e-01 8.76004755e-01 4.17874515e-01 -5.33961803e-02
5.14901161e-01 3.59167993e-01 -3.28350574e-01 -5.26497543e-01
-4.88579839e-01 5.37179410e-02 1.16605431e-01 6.31975591e-01
-1.85699135e-01 -9.75931063e-02 1.71685725e-01 -4.81675893e-01
-4.99221385e-02 -3.72214794e-01 3.94479841e-01 1.32433546e+00
-7.46630132e-01 -1.00974929e+00 -5.75083911e-01 3.94433558e-01
1.20369740e-01 4.58283335e-01 5.93252592e-02 7.52397239e-01
2.80190110e-01 5.21048486e-01 4.93084848e-01 6.56733811e-02
5.32259285e-01 -1.38505399e-01 6.78247809e-01 -6.56622171e-01
-2.39447668e-01 -9.72983986e-02 8.26238617e-02 -6.72456145e-01
-7.35458672e-01 -1.08390796e+00 -1.28991675e+00 -4.12967354e-01
-1.59657866e-01 -2.35541061e-01 9.72354054e-01 5.78479171e-01
4.01950806e-01 5.56495786e-01 1.03012465e-01 -1.23429084e+00
2.33479679e-01 -4.59611535e-01 -6.92244470e-01 2.84317911e-01
7.43956447e-01 -1.34192681e+00 2.05720086e-02 5.39430976e-02] | [7.353168964385986, -2.408747434616089] |
8a2d874a-95be-4e6a-aa8f-91d27c1e784b | crosskd-cross-head-knowledge-distillation-for | 2306.11369 | null | https://arxiv.org/abs/2306.11369v1 | https://arxiv.org/pdf/2306.11369v1.pdf | CrossKD: Cross-Head Knowledge Distillation for Dense Object Detection | Knowledge Distillation (KD) has been validated as an effective model compression technique for learning compact object detectors. Existing state-of-the-art KD methods for object detection are mostly based on feature imitation, which is generally observed to be better than prediction mimicking. In this paper, we show that the inconsistency of the optimization objectives between the ground-truth signals and distillation targets is the key reason for the inefficiency of prediction mimicking. To alleviate this issue, we present a simple yet effective distillation scheme, termed CrossKD, which delivers the intermediate features of the student's detection head to the teacher's detection head. The resulting cross-head predictions are then forced to mimic the teacher's predictions. Such a distillation manner relieves the student's head from receiving contradictory supervision signals from the ground-truth annotations and the teacher's predictions, greatly improving the student's detection performance. On MS COCO, with only prediction mimicking losses applied, our CrossKD boosts the average precision of GFL ResNet-50 with 1x training schedule from 40.2 to 43.7, outperforming all existing KD methods for object detection. Code is available at https://github.com/jbwang1997/CrossKD. | ['Qibin Hou', 'Ming-Ming Cheng', 'Xiang Li', 'Zhaohui Zheng', 'Yuming Chen', 'Jiabao Wang'] | 2023-06-20 | null | null | null | null | ['dense-object-detection', 'model-compression'] | ['computer-vision', 'methodology'] | [-1.91441283e-01 2.42401183e-01 -3.32251728e-01 -8.31908435e-02
-8.16059291e-01 -2.58520484e-01 3.67034137e-01 2.29146317e-01
-5.73175192e-01 5.12047887e-01 -2.91106045e-01 -2.77606249e-01
8.35177824e-02 -5.69402218e-01 -9.94441450e-01 -6.84473932e-01
1.60058498e-01 4.49829489e-01 8.87500405e-01 9.88877043e-02
7.92481601e-02 4.69979286e-01 -1.83680964e+00 2.49810383e-01
7.08121240e-01 1.04565120e+00 6.16091669e-01 9.01132762e-01
1.82099268e-01 8.94610524e-01 -6.87306762e-01 -6.84464514e-01
3.17208171e-01 -5.09572886e-02 -5.09935856e-01 -3.81534398e-01
7.47652352e-01 -7.28922844e-01 -8.03783298e-01 1.21551013e+00
9.45742249e-01 -1.01400226e-01 4.70368028e-01 -1.18935108e+00
-5.73564768e-01 7.11274385e-01 -7.83245385e-01 5.10229468e-01
-9.96178612e-02 2.11561203e-01 1.02643335e+00 -1.11172974e+00
2.57443011e-01 1.28278589e+00 6.24631941e-01 5.59305847e-01
-1.01215637e+00 -1.07501900e+00 -3.08195390e-02 5.05490839e-01
-1.80420208e+00 -3.55109662e-01 2.88117647e-01 -3.68377775e-01
8.83304894e-01 8.20475593e-02 6.12294972e-01 6.64889932e-01
-2.39243060e-01 1.34992981e+00 8.48491311e-01 -6.46781683e-01
3.61293484e-03 5.12220681e-01 1.76302850e-01 1.01271343e+00
5.48251390e-01 3.27109784e-01 -7.42933273e-01 -1.94255896e-02
4.07972455e-01 -8.93435106e-02 -2.10289896e-01 -2.49848753e-01
-7.26596892e-01 8.56533110e-01 6.59241319e-01 -7.62208253e-02
-1.23396508e-01 2.78891474e-01 3.01922917e-01 2.66808242e-01
3.88736963e-01 1.21821478e-01 -5.20259917e-01 -6.03768826e-02
-1.08484960e+00 3.34069490e-01 6.62473440e-01 1.08418512e+00
5.23753762e-01 -3.65529652e-03 -4.99964565e-01 5.96591055e-01
3.80559206e-01 5.98953247e-01 5.63598633e-01 -5.90290964e-01
2.57669806e-01 5.96538126e-01 -5.80752864e-02 -6.94989443e-01
-8.75601619e-02 -9.22843277e-01 -4.76013124e-01 3.38280834e-02
3.65339488e-01 -7.74459466e-02 -9.42448020e-01 1.57442033e+00
6.45222783e-01 6.59267604e-01 1.39068961e-01 8.61345291e-01
1.06695211e+00 4.53381419e-01 1.87455828e-03 2.99633127e-02
1.59870589e+00 -1.04214418e+00 -2.09876537e-01 -9.43883136e-02
1.05851328e+00 -8.22810829e-01 7.41736352e-01 4.67117250e-01
-1.02883387e+00 -5.49249530e-01 -1.23953843e+00 -4.17677313e-03
-2.18685791e-01 6.30866826e-01 4.58709449e-01 6.54972613e-01
-8.58978748e-01 4.90166456e-01 -7.91050136e-01 4.31253947e-02
8.51808190e-01 4.62679416e-01 2.72896215e-02 -1.35956317e-01
-1.04621232e+00 9.87542868e-01 5.87617815e-01 -4.32891190e-01
-1.11454272e+00 -1.06430864e+00 -3.78774107e-01 3.81460309e-01
5.24262667e-01 -3.28797728e-01 1.90450919e+00 -4.26739722e-01
-1.21632683e+00 7.82340407e-01 2.13729352e-01 -9.70213890e-01
7.44036973e-01 -4.52744067e-01 -4.27772999e-02 1.53146774e-01
-8.48125592e-02 9.78664517e-01 9.31960285e-01 -7.92533338e-01
-1.19405901e+00 -2.50701100e-01 -2.16594473e-01 2.63013214e-01
-3.13577294e-01 -1.79071710e-01 -8.08525980e-01 -4.53934401e-01
5.25463745e-02 -8.00934076e-01 -1.47762336e-02 1.45216554e-01
-5.53192794e-01 -5.76636910e-01 1.01431167e+00 -3.28764617e-01
1.27137125e+00 -2.43340802e+00 -4.00300533e-01 -2.90155560e-02
6.14871919e-01 8.65718067e-01 2.75286403e-03 -1.79658011e-01
1.58971936e-01 -3.33493412e-01 3.46427947e-01 -5.70600271e-01
1.98556613e-02 7.41636902e-02 -4.56597149e-01 5.80864489e-01
1.16197996e-01 8.12044799e-01 -8.08482707e-01 -5.26145935e-01
1.86020777e-01 4.66903567e-01 -5.84708631e-01 2.51597553e-01
-2.74344534e-01 -1.53456572e-02 -3.89618039e-01 5.16819060e-01
8.32816660e-01 -4.02465165e-01 -1.75483584e-01 -3.34626995e-02
-3.60426493e-02 5.70414603e-01 -1.23768890e+00 1.30544031e+00
5.26587181e-02 8.12513590e-01 -1.14896446e-01 -9.87414122e-01
7.60443211e-01 2.01094583e-01 1.59917220e-01 -6.08609974e-01
1.36973009e-01 1.99119657e-01 4.74235751e-02 -1.93145256e-02
5.16004980e-01 2.13499039e-01 1.51487738e-01 3.10769558e-01
3.28214228e-01 -2.99814511e-02 2.17197567e-01 6.48343146e-01
9.60450292e-01 -2.35168695e-01 2.32573897e-01 -1.09307636e-02
2.15314522e-01 -6.70183972e-02 5.28737605e-01 1.20817852e+00
-2.71913111e-01 3.25935632e-01 2.58210331e-01 -2.31353700e-01
-8.84258211e-01 -1.08731043e+00 -2.37482831e-01 1.25098097e+00
1.81841329e-01 -4.85450476e-01 -7.79719532e-01 -7.86315501e-01
3.52572441e-01 6.64238691e-01 -3.76941144e-01 -3.78144413e-01
-2.75913388e-01 -5.73337853e-01 9.71767187e-01 5.42605042e-01
6.67388737e-01 -5.34793079e-01 -8.76943290e-01 1.55845955e-01
4.06566188e-02 -1.14284503e+00 -3.55110675e-01 4.64279413e-01
-6.29386127e-01 -1.00427294e+00 -6.35615528e-01 -5.40049851e-01
5.33173561e-01 6.39967322e-01 1.06951439e+00 3.04750174e-01
-6.51148379e-01 2.13528961e-01 -2.72041857e-01 -9.65660810e-01
-5.16059995e-01 9.05732736e-02 3.08403105e-01 -6.11365676e-01
6.82479262e-01 -1.44186378e-01 -6.46731019e-01 2.05069825e-01
-7.12836742e-01 2.31243134e-01 7.76881397e-01 7.75864184e-01
6.52294159e-01 7.50172287e-02 3.59483570e-01 -7.06874192e-01
2.23848283e-01 -2.91553587e-01 -1.00902510e+00 1.88532040e-01
-8.89577448e-01 2.13348597e-01 3.62036526e-01 -7.82341838e-01
-7.59345114e-01 2.33740330e-01 1.19007775e-03 -9.00420129e-01
4.92547825e-02 -6.99813589e-02 2.25584894e-01 -3.33014786e-01
8.42289567e-01 4.37513411e-01 -1.73730493e-01 -5.47880888e-01
3.42436701e-01 5.77634513e-01 6.73878014e-01 -3.40064794e-01
7.54611850e-01 1.19088493e-01 -9.14906189e-02 -6.11651659e-01
-1.14538920e+00 -5.50918400e-01 -2.70073265e-01 -3.50522920e-02
4.19985801e-01 -1.41787446e+00 -9.10668731e-01 3.34627658e-01
-1.05752575e+00 -2.04878956e-01 -4.89959627e-01 6.93584204e-01
-2.41743609e-01 1.51723936e-01 -6.33396506e-01 -5.90151608e-01
-4.03754056e-01 -1.15585339e+00 8.41564417e-01 4.46510762e-01
2.37975582e-01 -3.22252363e-01 -1.71390638e-01 2.37553686e-01
1.94134235e-01 -5.64631283e-01 4.87517834e-01 -9.14606154e-01
-7.51703501e-01 -4.43059415e-01 -5.48029661e-01 4.80818629e-01
-4.08634990e-01 -1.46348193e-01 -1.27115226e+00 -4.04395282e-01
-2.66940892e-01 -3.50806177e-01 1.12895048e+00 3.76735210e-01
1.16587973e+00 -3.06477398e-01 -5.10498285e-01 4.06186312e-01
1.22094047e+00 -1.39258578e-01 2.80384868e-01 8.31838995e-02
4.65434283e-01 9.61913764e-02 6.56858742e-01 6.43430412e-01
4.10650432e-01 7.49295175e-01 5.51107824e-01 3.11666429e-02
-6.00607693e-01 -5.18727183e-01 3.06272715e-01 6.43049717e-01
3.40661377e-01 -1.68460399e-01 -7.83283114e-01 4.94683385e-01
-1.69172752e+00 -8.16510618e-01 -1.25200436e-01 2.31157684e+00
1.03752673e+00 3.25112790e-01 1.10777602e-01 1.48356155e-01
7.39999056e-01 -3.52612257e-01 -6.16223395e-01 -1.22874238e-01
1.27282351e-01 7.62174577e-02 9.16202605e-01 3.41041416e-01
-1.11531591e+00 1.08399558e+00 5.14235449e+00 1.43097425e+00
-1.10422421e+00 4.29409355e-01 4.20282930e-01 -4.21489328e-01
3.98090035e-01 -2.37448201e-01 -1.70059907e+00 5.32377601e-01
9.13698018e-01 -2.67079145e-01 9.27526876e-02 1.19135618e+00
-1.86219990e-01 -2.44396403e-01 -1.12184632e+00 9.50726926e-01
-1.29914850e-01 -1.34594715e+00 -1.28260897e-02 -5.39286397e-02
7.06044734e-01 2.72903264e-01 4.13104713e-01 7.42701054e-01
7.03455806e-01 -8.01756024e-01 8.40601206e-01 7.56040737e-02
7.24361837e-01 -6.52484000e-01 5.17506123e-01 7.01087713e-01
-1.10964739e+00 -1.32834375e-01 -6.14575744e-01 -2.58174967e-02
-3.87892634e-01 8.12577069e-01 -1.65895474e+00 1.05848797e-01
7.60627270e-01 2.27833048e-01 -6.99595273e-01 1.25730062e+00
-4.67285782e-01 9.62804914e-01 -6.85203552e-01 -6.37714490e-02
2.80990839e-01 5.16840935e-01 5.56710899e-01 1.46918380e+00
3.58584762e-01 3.37481387e-02 1.41160399e-01 6.53960586e-01
-5.00252843e-01 -7.44158700e-02 -3.88554126e-01 2.07550377e-01
8.26199234e-01 9.85406756e-01 -5.70478201e-01 -6.77515030e-01
-3.96690935e-01 7.65591264e-01 5.39025784e-01 -7.00594261e-02
-9.92977679e-01 -2.41146669e-01 6.48906887e-01 7.48112127e-02
7.53389299e-01 2.19411314e-01 -8.19649249e-02 -9.47286665e-01
7.46986037e-03 -6.59945726e-01 5.41389525e-01 -5.78471303e-01
-8.21053147e-01 1.46933720e-01 -7.75976926e-02 -1.07762301e+00
7.00141191e-02 -4.02707487e-01 -4.87893164e-01 6.23355210e-01
-1.66063190e+00 -7.34717965e-01 -1.90926790e-01 4.04787600e-01
5.46095729e-01 -1.53169394e-01 5.36157310e-01 5.37491202e-01
-6.24689341e-01 1.19641793e+00 1.38916314e-01 7.15193376e-02
6.83439493e-01 -1.20087039e+00 2.72330791e-01 8.90444696e-01
3.39215010e-01 1.94039389e-01 8.35651159e-01 -4.69871521e-01
-1.23589420e+00 -1.31753290e+00 7.17262805e-01 -1.84683532e-01
4.71045375e-01 -1.19672589e-01 -9.48976278e-01 4.35074151e-01
-3.29507679e-01 2.47289896e-01 4.04582202e-01 -2.20092788e-01
-5.25621355e-01 -1.31645262e-01 -1.31556702e+00 5.25522828e-01
7.69348979e-01 -3.86015296e-01 -4.56863195e-01 5.54062366e-01
8.51255357e-01 -7.35172093e-01 -5.11914849e-01 4.13544595e-01
4.58213329e-01 -7.57371724e-01 1.07025242e+00 -3.78139168e-01
2.64514416e-01 -2.43951842e-01 -2.09123358e-01 -9.37485039e-01
-2.67311931e-01 -3.34345788e-01 -7.55796790e-01 9.62498188e-01
2.96379924e-01 -4.08300728e-01 1.08901799e+00 2.23191068e-01
-5.84149472e-02 -8.99981380e-01 -1.05506432e+00 -9.55367982e-01
-5.75758405e-02 -5.73234677e-01 4.68210965e-01 6.29284620e-01
-2.29118720e-01 2.24535272e-01 -2.25325510e-01 5.15529811e-01
7.78371632e-01 -1.46239642e-02 7.57279098e-01 -1.15134907e+00
-6.22185946e-01 -5.29850423e-01 -6.11354291e-01 -1.63997209e+00
-1.38724059e-01 -9.73326921e-01 6.44961819e-02 -9.08539534e-01
4.54600841e-01 -5.83329618e-01 -3.08241099e-01 6.16608024e-01
-1.71415836e-01 3.36935401e-01 4.10940319e-01 2.21631199e-01
-8.01422358e-01 3.03251654e-01 1.26480579e+00 -1.20309055e-01
-2.01880649e-01 2.91769266e-01 -6.15163505e-01 7.49147058e-01
8.06759655e-01 -9.06582057e-01 -2.90586025e-01 -2.81637877e-01
-5.53488880e-02 -1.21312141e-01 3.92823040e-01 -1.20798337e+00
7.66259611e-01 3.13300073e-01 4.38864917e-01 -7.82811821e-01
3.94756258e-01 -5.73222995e-01 -3.73331755e-01 9.63478148e-01
-4.06129330e-01 -3.38989913e-01 3.45122248e-01 6.02385163e-01
4.19296771e-02 -4.69234526e-01 1.01511681e+00 1.07534252e-01
-7.06669629e-01 3.04575711e-01 -7.96863735e-02 1.27968483e-03
1.23935103e+00 -1.17368229e-01 -5.33215284e-01 -1.38687581e-01
-6.00026131e-01 3.17303747e-01 2.17424873e-02 9.76031497e-02
6.04260921e-01 -1.04179561e+00 -8.72816265e-01 2.80807793e-01
-8.34666640e-02 2.74525374e-01 1.19545542e-01 9.39529479e-01
-2.81119853e-01 6.97580636e-01 2.90367812e-01 -7.60035634e-01
-1.60818243e+00 5.46849728e-01 1.37937576e-01 -3.34089756e-01
-6.75057113e-01 1.44385195e+00 2.79734910e-01 -7.33465403e-02
7.07560241e-01 -3.21684569e-01 2.12457269e-01 -1.31469578e-01
6.71306491e-01 5.31051219e-01 1.19312264e-01 -3.42889726e-01
-2.12427288e-01 -5.86595945e-03 -6.83334947e-01 3.53267848e-01
9.88321960e-01 1.39877841e-01 5.60419142e-01 -9.95998159e-02
9.93305326e-01 -2.01733872e-01 -1.42222667e+00 -7.47419596e-01
-1.77638426e-01 -5.56999922e-01 3.40259999e-01 -7.01771915e-01
-1.05078566e+00 8.45432699e-01 9.52967823e-01 1.36688380e-02
9.70488548e-01 3.25099349e-01 6.81530356e-01 5.68412900e-01
2.80889511e-01 -9.82752919e-01 1.68406516e-01 4.32114035e-01
4.11221087e-01 -1.35015726e+00 2.81603873e-01 -4.71680731e-01
-4.37769204e-01 7.38571823e-01 8.38863850e-01 1.20603899e-03
5.18659472e-01 4.57347184e-01 -3.77928883e-01 -6.64649829e-02
-1.08530068e+00 -2.60202080e-01 2.64193982e-01 2.60071397e-01
4.03105728e-02 3.79433066e-01 9.22307074e-02 5.77919543e-01
-3.26814741e-01 -2.81455833e-02 3.52041483e-01 9.18675363e-01
-9.44751799e-01 -6.51040256e-01 -4.61930603e-01 7.23686159e-01
-7.14070499e-01 -2.67214328e-01 -1.02687381e-01 6.64830506e-01
2.75460333e-01 7.51826763e-01 5.15038446e-02 -4.37699348e-01
1.74628824e-01 6.71463646e-03 4.34623867e-01 -7.12329328e-01
-3.57623398e-01 -2.01034695e-01 -2.82335371e-01 -4.17226762e-01
1.71219110e-01 -3.01049948e-01 -1.35933912e+00 -6.18260443e-01
-9.24498022e-01 1.05389208e-01 6.93704367e-01 7.33226180e-01
5.29996336e-01 4.20039952e-01 3.91319752e-01 -4.82111037e-01
-1.16177237e+00 -9.46255684e-01 -4.84761715e-01 -1.87911332e-01
3.95396769e-01 -7.95534074e-01 -3.19935828e-01 -1.21501073e-01] | [9.267452239990234, 1.2898573875427246] |
17f5e634-08ae-4738-9d9d-5ad14c0d6958 | native-multi-band-audio-coding-within-hyper | 2303.08005 | null | https://arxiv.org/abs/2303.08005v1 | https://arxiv.org/pdf/2303.08005v1.pdf | Native Multi-Band Audio Coding within Hyper-Autoencoded Reconstruction Propagation Networks | Spectral sub-bands do not portray the same perceptual relevance. In audio coding, it is therefore desirable to have independent control over each of the constituent bands so that bitrate assignment and signal reconstruction can be achieved efficiently. In this work, we present a novel neural audio coding network that natively supports a multi-band coding paradigm. Our model extends the idea of compressed skip connections in the U-Net-based codec, allowing for independent control over both core and high band-specific reconstructions and bit allocation. Our system reconstructs the full-band signal mainly from the condensed core-band code, therefore exploiting and showcasing its bandwidth extension capabilities to its fullest. Meanwhile, the low-bitrate high-band code helps the high-band reconstruction similarly to MPEG audio codecs' spectral bandwidth replication. MUSHRA tests show that the proposed model not only improves the quality of the core band by explicitly assigning more bits to it but retains a good quality in the high-band as well. | ['Minje Kim', 'Inseon Jang', 'Darius Petermann'] | 2023-03-14 | null | null | null | null | ['bandwidth-extension', 'bandwidth-extension'] | ['audio', 'speech'] | [ 3.53025109e-01 1.06624521e-01 -5.12229443e-01 3.40643525e-02
-7.52588987e-01 -2.52876878e-01 -3.02098338e-02 -1.34694830e-01
-3.85860418e-04 6.34199381e-01 5.53761065e-01 -2.00608626e-01
-2.34563202e-01 -6.89829588e-01 -5.58341503e-01 -8.10509086e-01
-4.57988739e-01 -1.33176193e-01 4.79603708e-01 -1.56121165e-01
5.40050073e-03 2.80454963e-01 -1.97808313e+00 6.98218822e-01
5.02859890e-01 1.48603714e+00 4.66841787e-01 1.02909875e+00
-6.88719228e-02 8.57050896e-01 -6.48762167e-01 -2.81684045e-02
2.64210701e-01 -5.04841089e-01 -5.15312850e-01 1.66308016e-01
2.07636252e-01 -7.05764472e-01 -2.79322535e-01 1.00725150e+00
5.81098139e-01 1.11509666e-01 3.93118024e-01 -8.76799107e-01
-1.85026303e-01 8.69010985e-01 -3.66726369e-01 -1.05603859e-02
2.69827545e-01 -3.22923921e-02 1.41438794e+00 -3.26588094e-01
3.27413529e-01 9.28342462e-01 5.99900603e-01 4.17233795e-01
-1.29976928e+00 -6.76276207e-01 -1.18418626e-01 2.69381315e-01
-1.37627292e+00 -8.14552844e-01 1.09968674e+00 -2.03048512e-01
8.98373842e-01 3.83044869e-01 9.65580046e-01 6.73633337e-01
-9.20533314e-02 6.75958693e-01 4.30282891e-01 -5.24750113e-01
3.79727364e-01 -2.33946607e-01 -2.48712391e-01 2.65682310e-01
-2.75906205e-01 2.24259287e-01 -8.79765809e-01 1.06303439e-01
1.06380999e+00 -2.48919114e-01 -1.02219272e+00 -4.31913882e-01
-7.64224946e-01 5.48923373e-01 3.88794243e-01 5.61612070e-01
-4.74615842e-01 7.36921132e-01 6.08954489e-01 3.83565009e-01
3.21818292e-01 -9.35548991e-02 -3.63526285e-01 -4.17318791e-01
-1.38606787e+00 5.95834740e-02 5.67063391e-01 9.54319119e-01
5.89670599e-01 5.59339345e-01 -9.17045698e-02 1.15383422e+00
3.05023938e-01 5.24720326e-02 6.35719776e-01 -1.30226409e+00
4.10909444e-01 -1.42343968e-01 1.33383274e-02 -6.94546998e-01
-1.67767540e-01 -1.05469382e+00 -1.12285173e+00 2.43565798e-01
9.42623243e-02 2.46238299e-02 -5.81475973e-01 1.95482647e+00
9.28553008e-03 2.12322608e-01 -7.15087503e-02 1.08356345e+00
2.70066857e-01 1.03600001e+00 -7.77524188e-02 -4.33365405e-01
1.33476663e+00 -7.90886283e-01 -9.66990829e-01 1.26076654e-01
6.90097362e-02 -7.08074033e-01 8.92062962e-01 7.42634237e-01
-1.51685965e+00 -9.76587594e-01 -1.34742212e+00 1.02988561e-04
2.87400514e-01 1.64006338e-01 4.62945908e-01 8.42336297e-01
-1.36426020e+00 6.57619119e-01 -4.97233480e-01 3.75519693e-01
2.98391014e-01 2.55381018e-01 -2.62620896e-02 1.86359555e-01
-1.23548603e+00 2.28180349e-01 6.75126195e-01 -3.43038499e-01
-9.80700791e-01 -9.17832255e-01 -5.69418669e-01 7.60898232e-01
5.57210408e-02 -5.18909633e-01 1.39278495e+00 -1.20772696e+00
-1.68145001e+00 3.17383200e-01 1.61296755e-01 -8.20533752e-01
7.49143064e-02 -7.35104904e-02 -5.60250998e-01 7.15114176e-01
-2.83813030e-01 8.98313165e-01 1.28678095e+00 -1.09493852e+00
-7.10747421e-01 1.30143780e-02 4.91585881e-02 9.90754291e-02
-6.49865627e-01 -3.97766858e-01 -3.60431403e-01 -9.52969849e-01
2.15681449e-01 -3.28801632e-01 3.22759330e-01 4.60193068e-01
-4.48662676e-02 3.75327766e-01 5.96194029e-01 -6.48694158e-01
1.59843826e+00 -2.65071940e+00 3.85633856e-01 -5.52509911e-02
2.18847185e-01 1.37344226e-01 -1.29380450e-01 2.97291815e-01
-2.97171980e-01 -1.14014693e-01 -2.11423174e-01 -2.60876536e-01
1.30995188e-03 2.07996324e-01 -4.64441359e-01 3.05914700e-01
-9.11885723e-02 1.09283276e-01 -5.08586287e-01 -3.14495891e-01
6.91458434e-02 7.20035315e-01 -1.29596162e+00 1.99837148e-01
-1.39800608e-01 1.18275717e-01 3.78355421e-02 4.87739891e-01
8.35418940e-01 -5.95478266e-02 2.76342005e-01 -5.47130942e-01
-9.91587341e-02 2.76509017e-01 -1.24428368e+00 2.01604080e+00
-7.59528220e-01 7.06597567e-01 8.28890383e-01 -8.45984757e-01
8.70102406e-01 7.92459846e-01 4.54352230e-01 -8.02854478e-01
-4.12080400e-02 3.60771626e-01 1.61586568e-01 -3.16713303e-01
7.93366849e-01 -4.11135733e-01 4.02256548e-01 1.01904541e-01
2.85121709e-01 -1.89518496e-01 -2.16714088e-02 -7.88906142e-02
6.89028084e-01 -4.69523147e-02 4.20784831e-01 -3.44481379e-01
6.33604348e-01 -9.45216417e-01 3.98255914e-01 4.12470192e-01
-2.87862390e-01 5.95117390e-01 3.67053181e-01 2.56877020e-02
-1.41289139e+00 -1.14212084e+00 -3.13666075e-01 1.40692866e+00
-8.21593553e-02 -4.21530157e-01 -7.23828375e-01 2.56295562e-01
-1.83958262e-01 5.86699903e-01 -2.19257504e-01 -1.77009925e-01
-2.61153698e-01 -2.31895104e-01 6.43187284e-01 1.14294648e-01
6.08614385e-01 -6.70290828e-01 -1.00879645e+00 5.58082283e-01
-4.39749241e-01 -6.82525337e-01 -6.04634225e-01 5.64943731e-01
-9.45590794e-01 -5.68202674e-01 -1.08115923e+00 -5.81649244e-01
-1.37996390e-01 2.19544992e-01 9.39059138e-01 -9.09277573e-02
4.28146981e-02 7.35648796e-02 -6.12861991e-01 1.10865600e-01
-5.39542019e-01 1.00643709e-01 -2.48848289e-01 2.54213005e-01
-3.80011529e-01 -1.26381421e+00 -6.35795236e-01 8.37614685e-02
-1.14565015e+00 1.45476669e-01 3.93190444e-01 7.22161174e-01
3.53783071e-01 2.37273186e-01 9.12469983e-01 -1.43565880e-02
6.31914735e-01 -2.51713097e-01 -4.73855287e-01 -2.77643889e-01
-2.22488895e-01 -1.09658025e-01 8.33127439e-01 -3.66101205e-01
-7.93718398e-01 -1.87334359e-01 -6.67511046e-01 -5.10616481e-01
8.48932639e-02 1.85882807e-01 -1.70887709e-01 -1.69663932e-02
5.91760814e-01 5.75946033e-01 -1.45788252e-01 -6.43222570e-01
2.46021360e-01 9.45801377e-01 6.20291591e-01 -6.39367104e-01
2.14345425e-01 3.42511654e-01 -6.82737678e-02 -8.72027278e-01
-2.70019382e-01 -4.85850126e-01 -8.82245153e-02 -5.76602995e-01
6.63900018e-01 -1.11769331e+00 -8.34171176e-01 2.09603786e-01
-1.19079435e+00 -2.59743899e-01 -6.70636594e-01 5.74442506e-01
-1.02267444e+00 4.29138362e-01 -8.62270117e-01 -1.09513688e+00
-2.24144503e-01 -1.22127128e+00 1.02271235e+00 -8.24745670e-02
-4.95425276e-02 -4.21932191e-01 -5.97836711e-02 -1.73448011e-01
7.32625365e-01 -2.98808098e-01 8.82499933e-01 -3.18843238e-02
-6.25315487e-01 2.76080489e-01 -1.73032761e-01 4.23770875e-01
-1.77821554e-02 -2.99195379e-01 -1.34826732e+00 -4.09171313e-01
2.86721706e-01 -3.51075500e-01 1.01325095e+00 6.84426665e-01
1.33924091e+00 -3.93319070e-01 3.65538716e-01 8.81147504e-01
1.51274049e+00 2.79810190e-01 8.17680001e-01 6.85672611e-02
-9.57135335e-02 3.47183973e-01 7.83916339e-02 8.00796688e-01
-2.77004242e-01 9.09564316e-01 8.98506105e-01 2.36094296e-02
-5.25272369e-01 -1.44926414e-01 3.25399578e-01 9.93148327e-01
-6.87254369e-02 -2.23061323e-01 -3.23080391e-01 5.86977601e-01
-1.46359611e+00 -1.16191769e+00 2.83807009e-01 2.15857029e+00
1.30419827e+00 1.97954908e-01 1.61762878e-01 8.09433877e-01
7.70013571e-01 4.43000883e-01 -2.67041385e-01 -5.51065207e-01
3.52761112e-02 2.15069085e-01 3.18053603e-01 5.40363848e-01
-7.02977121e-01 2.72963434e-01 6.65652084e+00 1.44681275e+00
-1.02395105e+00 1.94015071e-01 5.25813460e-01 -4.71561611e-01
-3.56844455e-01 -3.20228189e-01 -3.96553725e-01 5.84515393e-01
1.27518308e+00 2.39797756e-02 9.46652770e-01 8.67193401e-01
2.54093349e-01 -1.50334660e-03 -8.16565156e-01 1.08588612e+00
-2.72186577e-01 -1.51197875e+00 2.39111319e-01 1.23116575e-01
3.86153311e-01 -4.15790945e-01 1.99697405e-01 1.13873586e-01
-3.36125106e-01 -6.27250493e-01 1.37189806e+00 2.29265153e-01
1.42170274e+00 -1.11061561e+00 2.78638512e-01 3.31763327e-01
-1.53006530e+00 -4.59171176e-01 -4.39089268e-01 -6.05773367e-02
1.97519019e-01 7.37817883e-01 -3.92736673e-01 4.01458591e-01
5.27487099e-01 4.25907224e-01 1.51224688e-01 1.05850089e+00
5.91319688e-02 6.43709540e-01 -7.39850104e-03 5.23995697e-01
1.98709428e-01 7.96909928e-02 5.42628646e-01 1.44290400e+00
6.00825489e-01 2.14805342e-02 -1.46173805e-01 7.34096587e-01
-2.15683594e-01 -1.13537177e-01 -1.85763255e-01 3.09467912e-01
4.88266528e-01 7.52060413e-01 -3.07145923e-01 -4.03513253e-01
-3.56446445e-01 7.26358831e-01 -4.02071700e-02 3.59657973e-01
-8.65992546e-01 -5.56574225e-01 8.69691253e-01 1.33818999e-01
6.98036849e-01 -1.11170888e-01 -4.86805849e-02 -8.20895970e-01
-3.05005640e-01 -8.69464695e-01 -1.30047705e-02 -1.06683457e+00
-5.47241449e-01 5.41869462e-01 -2.65558571e-01 -1.63033879e+00
-4.67440158e-01 -2.55067736e-01 8.52203816e-02 8.59662354e-01
-1.83574295e+00 -7.53025293e-01 7.86769912e-02 8.20918858e-01
6.60056710e-01 -1.43723562e-01 8.01370025e-01 8.34807277e-01
2.77685765e-02 6.48729682e-01 2.29822859e-01 -1.59783021e-01
4.34611619e-01 -8.63232732e-01 -3.11740786e-01 6.87225401e-01
-1.50570601e-01 4.75146532e-01 8.77280831e-01 -2.03176215e-01
-1.00967395e+00 -9.14129496e-01 3.99013519e-01 9.48855639e-01
6.30901098e-01 -1.57193422e-01 -8.85922253e-01 1.61936313e-01
3.64506155e-01 -2.51085162e-01 7.80984282e-01 -1.56413615e-01
-4.53552991e-01 -4.23510551e-01 -1.08074749e+00 3.61376405e-01
6.59071505e-01 -8.54136229e-01 -1.89689502e-01 -1.15300708e-01
1.08927405e+00 -2.03469947e-01 -8.52953970e-01 6.56035617e-02
6.09045923e-01 -1.59405220e+00 1.13323057e+00 -2.41536386e-02
8.08279812e-01 -2.71779686e-01 -7.30660439e-01 -1.12952077e+00
-3.72593194e-01 -6.35366142e-01 -4.06354606e-01 1.07252347e+00
2.56472696e-02 -2.41988942e-01 7.06322312e-01 -2.29618400e-01
-6.29787624e-01 -3.90800625e-01 -1.60233486e+00 -7.75208473e-01
-2.63229102e-01 -9.08709824e-01 6.77426636e-01 5.44135690e-01
3.57487708e-01 -4.55512628e-02 -7.55414009e-01 9.27046686e-02
5.07307887e-01 5.94338365e-02 1.18380025e-01 -1.00299060e+00
-9.84826446e-01 -6.74613178e-01 -3.07630599e-01 -1.60820568e+00
-2.16194674e-01 -7.25656211e-01 7.49577284e-02 -9.41693902e-01
-8.59732926e-02 -1.41763523e-01 -2.25095093e-01 4.23482656e-02
5.06294131e-01 5.08049965e-01 4.74796116e-01 1.77732706e-01
-1.71829224e-01 6.25919104e-01 1.42082143e+00 -1.92525312e-01
-1.42965436e-01 -6.43246546e-02 -4.52322900e-01 5.38155377e-01
6.58613503e-01 -1.73035383e-01 -6.83369279e-01 -3.79462957e-01
1.76978707e-01 8.97258043e-01 3.48563313e-01 -1.59368896e+00
2.97333330e-01 9.36580896e-02 1.52918845e-01 -4.26322401e-01
7.40645409e-01 -1.27368009e+00 4.00048673e-01 5.48407435e-01
-6.87196910e-01 -6.37943864e-01 3.34457874e-01 7.25610256e-01
-5.66381514e-01 -4.75591987e-01 1.18346751e+00 -4.65341359e-02
-3.79987895e-01 -2.95989923e-02 -6.78359807e-01 -3.19897979e-01
5.10667264e-01 -4.31483805e-01 7.26055354e-02 -7.36372948e-01
-1.02028644e+00 -4.77616876e-01 2.88555920e-01 -1.55092075e-01
6.13399506e-01 -1.35984683e+00 -5.25846899e-01 5.51912427e-01
-1.36502668e-01 -3.06068599e-01 6.42600358e-01 4.01767284e-01
-5.37615538e-01 5.49809813e-01 -5.98158121e-01 -6.67440593e-01
-9.23743427e-01 4.76665616e-01 5.02332509e-01 -2.23470405e-01
-5.00277162e-01 7.88621664e-01 1.28143355e-01 4.94515300e-01
4.47983474e-01 -4.83858436e-01 -4.84847911e-02 1.87324196e-01
7.75987327e-01 3.49613637e-01 1.32941017e-02 -5.46680808e-01
1.25026882e-01 5.71195126e-01 3.75715941e-01 -4.21322048e-01
1.25739253e+00 -3.62497091e-01 -5.53060621e-02 3.64439249e-01
1.34480190e+00 1.00751288e-01 -1.42490196e+00 3.60625982e-02
-6.38120890e-01 -6.13346756e-01 5.56606352e-01 -6.17436767e-01
-1.37532043e+00 1.16552806e+00 5.66849530e-01 6.43199325e-01
1.88240838e+00 -3.56595635e-01 7.83560872e-01 -1.21803850e-01
5.66968322e-01 -1.03714037e+00 1.67159706e-01 1.94773361e-01
8.86912286e-01 -1.73135206e-01 -1.36605889e-01 -2.73912132e-01
-1.50795206e-01 1.46531093e+00 -3.24802026e-02 4.65196557e-02
6.11104965e-01 4.57642734e-01 -4.84236926e-01 2.75273323e-01
-8.00530434e-01 -6.61881417e-02 6.68069441e-03 6.41315579e-01
5.83496749e-01 1.68519933e-02 -6.30858466e-02 4.16240126e-01
-2.91684806e-01 1.13185354e-01 6.75686538e-01 3.28999728e-01
-1.01156390e+00 -9.03937936e-01 -6.19130969e-01 2.68201143e-01
-5.49529076e-01 -3.11771989e-01 4.08346772e-01 4.26130831e-01
1.57494679e-01 9.58123028e-01 3.25573385e-01 -3.67617309e-01
2.53011975e-02 2.04127580e-01 2.53830045e-01 -2.23040730e-01
-6.03764653e-01 8.52586210e-01 1.12726666e-01 -6.62401497e-01
-3.90712917e-01 -2.01988935e-01 -1.02393818e+00 -4.56752598e-01
-2.56862164e-01 2.41249308e-01 6.38920844e-01 1.98867783e-01
2.52883226e-01 1.25426865e+00 5.89466572e-01 -1.06277883e+00
-5.54856658e-01 -8.12173247e-01 -1.21736443e+00 2.14962456e-02
1.06345439e+00 -2.98002720e-01 -4.23390448e-01 2.87988067e-01] | [15.490598678588867, 5.789776802062988] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.