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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
ddd16d49-9c69-42bd-b986-a5349a548117 | tracker-meets-night-a-transformer-enhancer | 2303.10951 | null | https://arxiv.org/abs/2303.10951v1 | https://arxiv.org/pdf/2303.10951v1.pdf | Tracker Meets Night: A Transformer Enhancer for UAV Tracking | Most previous progress in object tracking is realized in daytime scenes with favorable illumination. State-of-the-arts can hardly carry on their superiority at night so far, thereby considerably blocking the broadening of visual tracking-related unmanned aerial vehicle (UAV) applications. To realize reliable UAV tracking at night, a spatial-channel Transformer-based low-light enhancer (namely SCT), which is trained in a novel task-inspired manner, is proposed and plugged prior to tracking approaches. To achieve semantic-level low-light enhancement targeting the high-level task, the novel spatial-channel attention module is proposed to model global information while preserving local context. In the enhancement process, SCT denoises and illuminates nighttime images simultaneously through a robust non-linear curve projection. Moreover, to provide a comprehensive evaluation, we construct a challenging nighttime tracking benchmark, namely DarkTrack2021, which contains 110 challenging sequences with over 100 K frames in total. Evaluations on both the public UAVDark135 benchmark and the newly constructed DarkTrack2021 benchmark show that the task-inspired design enables SCT with significant performance gains for nighttime UAV tracking compared with other top-ranked low-light enhancers. Real-world tests on a typical UAV platform further verify the practicability of the proposed approach. The DarkTrack2021 benchmark and the code of the proposed approach are publicly available at https://github.com/vision4robotics/SCT. | ['Bowen Li', 'Guangze Zheng', 'Shan An', 'Ziang Cao', 'Changhong Fu', 'Junjie Ye'] | 2023-03-20 | null | null | null | null | ['visual-tracking', 'blocking'] | ['computer-vision', 'natural-language-processing'] | [ 9.09742340e-02 -4.59643394e-01 1.55866325e-01 1.07039846e-02
-3.99010330e-01 -8.95183206e-01 6.49176955e-01 -4.45055753e-01
-3.11498076e-01 6.76395655e-01 -2.30892539e-01 -1.88101724e-01
3.18569206e-02 -4.82837945e-01 -8.22309792e-01 -1.11405492e+00
1.64578050e-01 -2.57125646e-01 5.17382920e-01 -3.17275912e-01
-4.50321466e-01 2.67818362e-01 -1.69099653e+00 3.04444320e-02
7.20533013e-01 1.07894635e+00 5.44667006e-01 6.75551951e-01
5.75772285e-01 5.24711311e-01 -4.13251281e-01 -2.95338959e-01
8.63866568e-01 -1.34958729e-01 -2.05745220e-01 3.49966079e-01
8.27861011e-01 -4.19886351e-01 -4.22203511e-01 1.23138189e+00
4.21818763e-01 5.35163991e-02 1.32744223e-01 -1.36691010e+00
-4.21931893e-01 -2.98259735e-01 -6.04093134e-01 3.15629512e-01
6.07542917e-02 7.25965679e-01 6.41537488e-01 -8.13589811e-01
4.44341332e-01 7.54755855e-01 8.05857062e-01 3.79418552e-01
-9.25048828e-01 -9.46545303e-01 2.32252374e-01 7.89171010e-02
-1.21548092e+00 -2.83837855e-01 6.80815637e-01 -3.57950568e-01
5.37822545e-01 3.57643574e-01 1.00780904e+00 8.87115598e-01
4.31162924e-01 5.72899818e-01 1.10556853e+00 -8.37656781e-02
-1.31425634e-01 -9.34147909e-02 -3.14787596e-01 9.46772754e-01
4.80426729e-01 8.14667165e-01 -2.07691774e-01 6.77926540e-02
5.73439658e-01 1.63851649e-01 -6.32947922e-01 -4.38892722e-01
-1.45276737e+00 2.70313740e-01 8.68371964e-01 8.46632349e-04
-2.92360842e-01 2.02648267e-01 2.40541920e-01 1.37982354e-01
5.03823221e-01 1.42944798e-01 -4.27903026e-01 5.19715250e-01
-9.17452812e-01 1.80559665e-01 1.95230946e-01 1.34342182e+00
5.75320840e-01 4.27236468e-01 -6.57020330e-01 3.96415710e-01
5.12807608e-01 9.72937226e-01 -2.11630818e-02 -8.09974372e-01
1.95319548e-01 3.63046229e-01 5.90811789e-01 -8.09685886e-01
-5.34296751e-01 -9.30442810e-01 -6.57224715e-01 3.99998605e-01
3.88708591e-01 -3.15629512e-01 -9.18156564e-01 1.71993041e+00
6.89474881e-01 4.62195337e-01 1.59346774e-01 1.50289452e+00
9.56070542e-01 7.11906731e-01 1.81966960e-01 -3.96959215e-01
1.55714178e+00 -1.22517121e+00 -7.30132401e-01 -2.35758483e-01
4.17831302e-01 -1.06791997e+00 7.85453200e-01 3.01343858e-01
-7.46813834e-01 -7.56382525e-01 -1.06171310e+00 9.20479223e-02
-2.19608456e-01 7.50581443e-01 6.51475847e-01 5.98699033e-01
-1.00603044e+00 -3.18848863e-02 -6.90036416e-01 -3.34493041e-01
3.19261044e-01 1.09625332e-01 -1.41915366e-01 -2.36139327e-01
-1.05636573e+00 5.85269332e-01 3.66691232e-01 4.06627268e-01
-1.64597881e+00 -8.26736927e-01 -5.56662381e-01 -2.83301890e-01
6.01745486e-01 -9.17344749e-01 1.12347436e+00 -8.79116535e-01
-1.39401162e+00 8.30963194e-01 1.59514621e-01 -5.54783106e-01
3.63637745e-01 -2.05811590e-01 -4.56953794e-01 -7.42917694e-03
4.60239351e-02 8.74851048e-01 1.31133890e+00 -1.23145843e+00
-1.05508912e+00 -2.77909040e-01 2.65080452e-01 3.45700860e-01
-3.35980177e-01 -7.12359091e-03 -6.44926429e-01 -8.40975523e-01
-3.89555722e-01 -1.39681756e+00 2.71357782e-02 3.62202108e-01
-1.55124918e-01 2.80537736e-02 1.34727669e+00 -5.39419949e-01
9.52593207e-01 -2.10961175e+00 -5.61464168e-02 -5.69154382e-01
1.05043061e-01 7.03510523e-01 -2.10863948e-01 4.16657366e-02
2.40173172e-02 -7.59159327e-01 -3.77329320e-01 -2.84882337e-01
-9.67166796e-02 1.72664095e-02 -3.69455218e-01 8.81587446e-01
4.01239237e-03 1.01915014e+00 -9.99485195e-01 -1.32509768e-01
6.14730299e-01 6.02178037e-01 -4.19182628e-01 1.77322105e-01
-5.28338790e-01 7.25368023e-01 -3.28140527e-01 1.02114069e+00
9.66893733e-01 -1.19432122e-01 -3.04545779e-02 -5.56336641e-01
-6.96499825e-01 -5.31339884e-01 -6.08450949e-01 1.81033814e+00
-2.87447363e-01 9.46091890e-01 2.52445638e-01 -3.23596805e-01
7.78614938e-01 1.16162494e-01 3.49986613e-01 -6.56158864e-01
3.48000139e-01 -7.82430172e-02 -1.88907266e-01 -2.34209716e-01
5.98187149e-01 1.45046487e-01 1.77591164e-02 -3.43430609e-01
-1.13212191e-01 -1.18112639e-01 9.57386494e-02 -1.40049374e-02
9.87375259e-01 4.61545914e-01 5.03374338e-02 -4.91580009e-01
6.17897213e-01 3.42466384e-01 6.92998946e-01 2.53864586e-01
-5.03040671e-01 2.97727555e-01 -4.34210211e-01 -5.34925699e-01
-8.57449114e-01 -9.10571158e-01 -1.19361125e-01 1.07974613e+00
5.77584207e-01 -3.45612437e-01 -6.20060503e-01 -5.55745840e-01
-1.67331219e-01 5.33311844e-01 -4.63172704e-01 -1.26315027e-01
-1.36747792e-01 -9.16929543e-01 4.65200037e-01 2.78197974e-01
1.07389486e+00 -6.13444924e-01 -1.19588220e+00 3.14783938e-02
-3.49255979e-01 -1.52107418e+00 -5.26006877e-01 -1.10263228e-02
-5.11186957e-01 -1.13824785e+00 -8.63257825e-01 -7.04684675e-01
3.80726665e-01 1.10936034e+00 7.74793088e-01 -8.89067873e-02
-4.70088094e-01 5.95462799e-01 -2.41889551e-01 -6.30505621e-01
1.42492414e-01 -3.27765912e-01 7.29587451e-02 9.51565355e-02
1.79904010e-02 -7.05169067e-02 -9.01983559e-01 6.31393850e-01
-7.41811812e-01 8.33153352e-02 4.92174536e-01 9.09643650e-01
5.88316381e-01 3.29577833e-01 4.98967478e-03 -1.82633787e-01
-8.43273327e-02 -1.21914357e-01 -1.46836782e+00 2.39216149e-01
-4.50030357e-01 -5.76939821e-01 6.42927170e-01 -3.22389275e-01
-1.16868901e+00 4.31506187e-01 2.19759122e-01 -9.40856934e-01
-1.00250572e-01 -1.09764494e-01 -2.12859154e-01 -6.16847515e-01
4.38731372e-01 3.92132819e-01 -2.54856408e-01 -1.25401840e-01
2.76006788e-01 2.58488983e-01 7.96783447e-01 -2.01911539e-01
1.22307062e+00 9.20672178e-01 2.59776652e-01 -9.55303967e-01
-9.90180075e-01 -4.91549432e-01 -1.99557289e-01 -7.50843227e-01
1.03539133e+00 -1.55166280e+00 -8.34131896e-01 5.68107784e-01
-1.07773983e+00 -6.08887136e-01 -1.99999332e-01 5.86699247e-01
-4.32234287e-01 1.35273799e-01 -2.45358601e-01 -7.79431105e-01
-4.50281262e-01 -1.28000689e+00 1.35902119e+00 5.24613976e-01
6.29323840e-01 -6.96069956e-01 -5.18356189e-02 3.23932052e-01
4.24065262e-01 2.85648257e-01 3.47283073e-02 1.45295024e-01
-1.14964879e+00 -5.80484979e-02 -5.50784230e-01 2.48120293e-01
-1.94574392e-03 -1.55082345e-01 -1.25148988e+00 -7.65163064e-01
-9.69612747e-02 -5.78030646e-02 1.00098097e+00 5.96016169e-01
7.98500955e-01 6.07566722e-02 -4.70509797e-01 1.05452216e+00
1.46192217e+00 2.75865048e-01 5.03793716e-01 4.89291698e-01
8.34275305e-01 1.06537923e-01 1.35006106e+00 5.98524094e-01
4.12829876e-01 1.00086319e+00 1.11648107e+00 -4.62360650e-01
-4.62656707e-01 -1.71326771e-02 6.07576311e-01 2.98926532e-01
-2.57251143e-01 -3.69682103e-01 -5.60366929e-01 4.61925715e-01
-1.71851683e+00 -8.31504285e-01 -4.29790437e-01 2.22233272e+00
4.09407854e-01 -1.22365899e-01 2.46385634e-02 -2.18243912e-01
5.64485550e-01 2.97376990e-01 -5.33700585e-01 4.80778009e-01
-3.79065692e-01 -3.76705766e-01 1.01153219e+00 2.45763913e-01
-1.52321815e+00 1.14622998e+00 4.75985146e+00 8.49178135e-01
-1.23464704e+00 4.21635360e-01 -2.05142200e-02 -5.68567663e-02
1.83961332e-01 6.21249899e-02 -1.08355892e+00 5.15084207e-01
6.61961973e-01 -1.80375669e-02 4.55583841e-01 7.89391696e-01
3.78198475e-01 2.74838451e-02 -4.75983530e-01 1.01585758e+00
5.29923141e-02 -1.10747731e+00 -1.82213545e-01 3.20960372e-03
8.71350825e-01 2.39279270e-01 3.57394814e-01 1.63409114e-01
1.20538473e-01 -3.36018026e-01 9.12984848e-01 5.43337822e-01
8.17842364e-01 -6.85472190e-01 4.65337843e-01 8.52555707e-02
-1.75476241e+00 -3.89954060e-01 -7.47937441e-01 1.88672200e-01
4.28868867e-02 4.09140825e-01 -6.77363694e-01 9.21384096e-01
1.06347418e+00 1.12957239e+00 -7.28582263e-01 1.40933955e+00
-3.23446721e-01 5.00928462e-01 -2.64494985e-01 3.66405576e-01
4.83129174e-01 -2.23422498e-01 1.14487839e+00 1.05644274e+00
3.42435420e-01 3.04507077e-01 3.11825454e-01 6.55953586e-01
1.22685179e-01 -3.56904805e-01 -8.02916825e-01 1.64998531e-01
6.51586354e-02 1.68898654e+00 -6.28480732e-01 -9.93313119e-02
-4.73426253e-01 9.14402485e-01 -4.42515761e-01 4.28635359e-01
-1.46745872e+00 -1.86591782e-02 8.29558074e-01 3.77136171e-02
4.18108225e-01 -3.08016062e-01 3.91501963e-01 -1.02395082e+00
-1.31171525e-01 -6.96593225e-01 2.00778797e-01 -1.28974831e+00
-8.68418276e-01 8.95025253e-01 -1.54626980e-01 -1.92430460e+00
4.81011450e-01 -6.41237319e-01 -3.02106947e-01 4.64647830e-01
-1.90759277e+00 -1.75926232e+00 -1.08334959e+00 8.76844466e-01
8.02892447e-01 -2.02068731e-01 4.27715361e-01 6.00621819e-01
-6.68938220e-01 3.41187567e-01 5.64132184e-02 -2.49259055e-01
9.25099611e-01 -7.33020723e-01 2.52614319e-01 1.32373679e+00
-1.43290296e-01 1.75328851e-01 8.36646140e-01 -6.88717365e-01
-1.76429284e+00 -1.92204285e+00 1.66423187e-01 -4.93902475e-01
5.35394669e-01 -2.81326026e-01 -5.26410818e-01 5.75577080e-01
4.73567158e-01 4.70252484e-01 6.59728721e-02 -7.42629647e-01
4.35342528e-02 -3.21790785e-01 -7.77747333e-01 4.74061370e-01
1.25503194e+00 -1.20686986e-01 -1.89417034e-01 7.24434316e-01
1.00457942e+00 -5.13481081e-01 -6.35863721e-01 7.61269987e-01
5.34059167e-01 -9.43798006e-01 1.11888778e+00 -1.43513724e-01
-1.59016863e-01 -1.03266943e+00 -4.28335220e-01 -1.31818867e+00
-4.10294116e-01 -7.75350630e-01 1.13945596e-01 1.09480739e+00
-1.17350273e-01 -6.25850618e-01 5.14557779e-01 1.41689824e-02
-6.31869912e-01 -3.46365809e-01 -6.11701071e-01 -1.15272880e+00
-5.53041399e-01 -2.38148853e-01 4.64327127e-01 7.03981757e-01
-8.75933647e-01 4.71729971e-02 -7.47278512e-01 7.14171946e-01
1.17318785e+00 2.19612330e-01 9.64116216e-01 -1.16418457e+00
-4.65958528e-02 9.04392600e-02 -3.86310577e-01 -9.68967080e-01
1.05138272e-01 -6.87478304e-01 2.72423446e-01 -1.20498514e+00
-1.94153073e-03 -2.57384241e-01 -1.24327980e-01 3.97971600e-01
-3.61688957e-02 7.34283626e-01 3.70487720e-01 1.51497617e-01
-8.62407148e-01 8.64988327e-01 1.46259379e+00 -3.50592136e-01
4.06804644e-02 3.26538011e-02 -1.93172127e-01 7.02158213e-01
5.15800774e-01 -4.91062015e-01 -5.41156709e-01 -5.19440413e-01
-1.84419051e-01 -7.34776258e-03 9.43065166e-01 -1.42437685e+00
4.78967696e-01 -4.40096948e-03 5.87510586e-01 -7.43455529e-01
6.51154876e-01 -1.22741377e+00 5.32000422e-01 6.07092798e-01
2.31762901e-01 4.12564762e-02 6.96436644e-01 7.45293260e-01
-1.59236178e-01 4.13496792e-01 1.03018045e+00 2.28516102e-01
-1.11585867e+00 6.22413754e-01 -5.56775555e-02 -6.93733096e-02
1.41370523e+00 -1.26115322e-01 -7.20013440e-01 6.21672422e-02
-4.10714328e-01 2.34192967e-01 6.66658759e-01 4.63808954e-01
5.32843828e-01 -1.24958539e+00 -5.94401062e-01 4.17508721e-01
4.09104109e-01 -2.71259189e-01 5.32124758e-01 1.32417738e+00
-3.28466594e-01 7.41161168e-01 -5.63829064e-01 -8.61763537e-01
-1.41538978e+00 7.79948533e-01 3.46489549e-01 1.61588565e-01
-8.41429174e-01 7.53827989e-01 7.53646970e-01 9.11866035e-03
1.31576166e-01 -4.40335453e-01 6.79738224e-02 -2.07235307e-01
6.50984824e-01 2.77167767e-01 1.09031714e-01 -8.00182819e-01
-5.09856284e-01 7.84192324e-01 3.50658506e-01 4.16558117e-01
1.12996316e+00 -4.30075228e-01 1.88274667e-01 -1.81606442e-01
6.92657709e-01 -1.95811212e-01 -1.86051762e+00 -7.87798166e-02
-4.82139647e-01 -7.21139431e-01 4.14328247e-01 -7.32506812e-01
-1.46571815e+00 6.31584346e-01 1.10263121e+00 -3.37847993e-02
1.64872050e+00 -3.89987320e-01 8.13588858e-01 2.31348321e-01
6.38148606e-01 -5.56610048e-01 -1.09625205e-01 3.73228937e-01
8.40701878e-01 -1.31353068e+00 -3.81216444e-02 -4.36641604e-01
-6.72842741e-01 7.72141039e-01 6.82913363e-01 -4.12180051e-02
4.03729767e-01 3.14520478e-01 5.38866036e-02 -1.77725956e-01
-8.99854958e-01 -6.33231223e-01 4.48900968e-01 6.94657803e-01
-8.20712838e-03 -9.59644560e-03 8.75817090e-02 3.64264309e-01
1.74849898e-01 9.96461138e-02 2.99695492e-01 7.32911408e-01
-4.15366352e-01 -4.21409935e-01 -5.67336738e-01 -1.03896400e-02
-1.26752749e-01 -9.29189473e-02 -1.59136102e-01 7.81221867e-01
4.36908305e-01 1.02824581e+00 -3.06101799e-01 -5.07019043e-01
4.77008879e-01 -5.53978086e-01 5.18495560e-01 -9.97877792e-02
-5.63181639e-01 3.12012792e-01 -8.17115679e-02 -8.47394407e-01
-6.78519905e-01 -5.41112959e-01 -8.22552145e-01 -1.24133348e-01
-4.00042087e-01 -1.78700045e-01 6.64878666e-01 5.28640389e-01
3.10572356e-01 7.84714282e-01 4.45177615e-01 -1.10699236e+00
-5.74953668e-02 -4.75450903e-01 -4.89229232e-01 -1.54548973e-01
5.31653821e-01 -1.22166860e+00 -2.11203963e-01 2.86121100e-01] | [6.7493414878845215, -1.986563801765442] |
33990c85-7373-4e95-8c20-62b2a013cafa | predicting-crop-yields-with-little-ground | 2106.08720 | null | https://arxiv.org/abs/2106.08720v3 | https://arxiv.org/pdf/2106.08720v3.pdf | Predicting crop yields with little ground truth: A simple statistical model for in-season forecasting | We present a fully automated model for in-season crop yield prediction, designed to work where there is a dearth of sub-national "ground truth" information. Our approach relies primarily on satellite data and is characterized by careful feature engineering combined with a simple regression model. As such, it can work almost anywhere in the world. Applying it to 10 different crop-country pairs (5 cereals -- corn, wheat, sorghum, barley and millet, in 2 countries -- Ethiopia and Kenya), we achieve RMSEs of 5%-10% for predictions 9 months into the year, and 7%-14% for predictions 3 months into the year. The model outputs daily forecasts for the final yield of the current year. It is trained using approximately 4 million data points for each crop-country pair. These consist of: historical country-level annual yields, crop calendars, crop cover, NDVI, temperature, rainfall, and evapotransporation. | ['Nemo Semret'] | 2021-06-16 | null | null | null | null | ['crop-yield-prediction', 'crop-yield-prediction'] | ['computer-vision', 'miscellaneous'] | [-7.79771283e-02 5.74270077e-03 -3.61307532e-01 -7.39599690e-02
-3.11069220e-01 -8.62919092e-01 3.31567436e-01 6.02265179e-01
1.19902022e-01 1.01703846e+00 7.82503635e-02 -1.00902605e+00
7.07041547e-02 -1.42027450e+00 -4.25781667e-01 -5.84577501e-01
-3.42588365e-01 -3.98753248e-02 -7.91261718e-02 -7.82716930e-01
3.83606628e-02 4.75939840e-01 -1.38178825e+00 -2.85543323e-01
1.07685363e+00 8.68116438e-01 4.24672127e-01 7.61184335e-01
5.11039495e-02 2.86911100e-01 -2.31354803e-01 4.06537913e-02
2.39097223e-01 -1.80583239e-01 -3.79829228e-01 -5.87436222e-02
-2.40657359e-01 -5.64551055e-01 2.66709149e-01 7.55818427e-01
9.08036232e-02 -4.96225178e-01 5.42492568e-01 -1.09286237e+00
-6.29584372e-01 8.01662803e-01 -1.03066278e+00 -2.67693639e-01
1.98835403e-01 2.21092887e-02 6.25370443e-01 -5.48114359e-01
3.64726841e-01 9.24408257e-01 9.52140093e-01 -2.51143217e-01
-1.33498216e+00 -5.50927162e-01 -5.63412532e-03 -6.17052019e-01
-1.22978139e+00 -1.93558633e-01 -6.97160065e-02 -6.17076516e-01
1.18298924e+00 3.46228033e-01 1.05843151e+00 3.59436810e-01
7.31812000e-01 1.03570938e-01 1.07855523e+00 -5.03148913e-01
1.93661168e-01 -1.54042691e-01 -1.85259864e-01 4.27908510e-01
4.50766325e-01 5.45180440e-01 -9.13857296e-02 -2.21750483e-01
7.92119384e-01 -1.27560794e-02 -1.03133939e-01 -1.53722586e-02
-1.18035436e+00 8.35026860e-01 5.71920276e-01 1.41814619e-01
-8.09452295e-01 -2.80467272e-01 2.41047665e-01 5.37640035e-01
9.78681684e-01 1.64421275e-01 -1.12822747e+00 1.33642823e-01
-1.37199795e+00 6.10751808e-01 8.73382568e-01 9.20412183e-01
8.86791945e-01 4.71245617e-01 2.74141252e-01 5.27971983e-01
2.67011762e-01 1.22208416e+00 -6.02002442e-02 -7.40732312e-01
3.27953190e-01 5.39779603e-01 8.90592933e-01 -1.28539264e+00
-6.56342030e-01 -1.69322729e-01 -1.03763485e+00 1.31730884e-01
3.00485283e-01 -7.93833315e-01 -1.04576242e+00 1.49384785e+00
-2.49322178e-03 -1.29404068e-01 2.22811207e-01 7.77603507e-01
4.11830723e-01 1.02450669e+00 4.36932921e-01 -5.61318219e-01
1.09843934e+00 -3.38429570e-01 -5.99059463e-01 -2.01027349e-01
7.24013805e-01 -8.52022588e-01 4.09595400e-01 5.84781989e-02
-9.03921843e-01 -3.52464586e-01 -6.90039754e-01 6.43755376e-01
-9.44798112e-01 4.44880515e-01 9.23319757e-01 4.78685886e-01
-1.17715490e+00 8.08440804e-01 -9.94403183e-01 -8.52022946e-01
-3.10307533e-01 -9.66595113e-03 -4.15620834e-01 3.40754330e-01
-9.67160821e-01 1.41263247e+00 6.15187764e-01 4.21517938e-01
-5.33352852e-01 -9.52629924e-01 -8.88793230e-01 2.35037789e-01
-8.83215293e-02 -2.71785259e-01 8.96992147e-01 -5.77706754e-01
-1.44517791e+00 8.66557419e-01 -1.80930182e-01 -2.74999768e-01
-9.28762630e-02 -1.94595382e-01 -6.97028160e-01 -5.54724514e-01
4.39290851e-01 3.44521523e-01 2.10633531e-01 -1.07855666e+00
-1.09072471e+00 -6.84661865e-01 -3.81230921e-01 7.25793466e-02
1.40809655e-01 4.24075127e-01 3.85978669e-01 -7.69631803e-01
6.28108144e-01 -1.01867163e+00 -7.29681671e-01 -5.44788539e-01
-8.94478858e-02 5.28816521e-01 4.20682102e-01 -1.24388051e+00
1.11726856e+00 -1.88269782e+00 -2.27896050e-01 2.21005976e-01
-5.32291532e-01 -3.82558852e-02 1.92133486e-02 5.86280107e-01
-1.94816396e-01 1.58455074e-01 -4.44556057e-01 5.62813520e-01
-1.36243865e-01 2.63596475e-01 -6.43445492e-01 2.99144149e-01
6.15713120e-01 6.11119449e-01 -9.76977170e-01 2.11285695e-01
5.38629353e-01 1.13657750e-01 8.44218135e-02 8.26278850e-02
-1.95417359e-01 2.09066287e-01 -2.56055564e-01 1.16418111e+00
1.44542408e+00 2.21375272e-01 4.62462604e-01 1.81788966e-01
-1.03081131e+00 -2.08189249e-01 -8.12224507e-01 1.29057205e+00
-4.11006093e-01 2.53978819e-01 3.13589483e-01 -7.71428168e-01
1.26720321e+00 3.92324716e-01 4.74582732e-01 -3.39894712e-01
-2.66363204e-01 6.11880004e-01 -2.26817697e-01 3.34120393e-02
7.59160638e-01 2.27445498e-01 -2.00520784e-01 2.17430234e-01
1.54599160e-01 -4.57600176e-01 3.75148863e-01 -3.63838106e-01
5.70752561e-01 6.51604891e-01 6.73982739e-01 -9.54791367e-01
-5.31845819e-03 6.82956934e-01 7.12191641e-01 2.53822595e-01
-3.91526893e-02 3.70124996e-01 5.70499361e-01 -6.97257519e-01
-1.16380918e+00 -1.03346848e+00 -3.36457312e-01 1.31675231e+00
-3.85223925e-01 -1.25195116e-01 -3.41924936e-01 1.71881422e-01
4.48908627e-01 8.74776542e-01 -4.46298748e-01 4.11596030e-01
9.82125401e-02 -1.62604594e+00 3.05791825e-01 4.53109592e-01
4.58951920e-01 -1.01650834e+00 -7.54288435e-01 6.76860571e-01
-9.46607739e-02 -8.08601499e-01 4.12767589e-01 7.61956692e-01
-1.17091715e+00 -5.75595856e-01 -9.35157120e-01 -4.91179943e-01
4.50393796e-01 2.20058709e-01 1.58397877e+00 -3.75007868e-01
1.30876806e-02 -3.04632634e-01 -3.90154988e-01 -8.50321114e-01
-2.53980428e-01 1.50751129e-01 4.51543592e-02 -7.92241037e-01
5.19373298e-01 -6.89980268e-01 -3.71764600e-01 9.05139744e-02
-3.57484549e-01 1.39259934e-01 3.05649579e-01 6.18313730e-01
2.91559547e-01 2.28863172e-02 5.05077720e-01 -4.81051803e-01
1.68789700e-01 -9.94193733e-01 -1.12241983e+00 5.82058966e-01
-6.68958187e-01 -5.49430251e-01 3.28319699e-01 1.87389836e-01
-1.06537890e+00 5.09814024e-01 2.84872562e-01 5.67804098e-01
-5.45669734e-01 1.14843953e+00 2.65290320e-01 1.94722667e-01
4.96670902e-01 1.61379859e-01 -1.78347513e-01 -4.55035597e-01
3.95062476e-01 5.52330196e-01 8.79089177e-01 -3.80637079e-01
7.04203784e-01 -8.71892646e-02 -5.41348532e-02 -8.32054794e-01
-4.13695037e-01 -2.34543562e-01 -9.65594769e-01 -9.69901085e-02
4.16440964e-01 -1.45952654e+00 -3.30275089e-01 9.02438104e-01
-1.00804567e+00 -4.64121014e-01 1.14924349e-01 6.73235655e-01
-3.49124312e-01 -2.38136366e-01 -4.59494799e-01 -9.42934692e-01
-6.36929333e-01 -6.78988874e-01 7.24119365e-01 4.98893082e-01
1.27141606e-02 -1.06130910e+00 9.64018926e-02 -6.82594895e-01
7.29654312e-01 9.53087032e-01 9.11059976e-01 2.67706096e-01
2.27054656e-01 -1.36291385e-01 -7.15471148e-01 -2.57514026e-02
8.02459359e-01 7.99573421e-01 -7.72838771e-01 -2.66042113e-01
-5.40946424e-01 -1.21294349e-01 6.27243400e-01 1.05906463e+00
5.31848133e-01 -2.65628025e-02 -2.65116781e-01 6.37404561e-01
1.73231030e+00 1.38866365e-01 4.15142924e-01 3.34713221e-01
1.17394537e-01 7.09353089e-01 1.09082460e+00 7.49433517e-01
4.75539535e-01 1.75595954e-02 6.38376892e-01 -5.75176597e-01
7.11670399e-01 -1.78990103e-02 3.67300212e-01 5.62133491e-01
-5.71877956e-01 1.46101450e-03 -1.29269159e+00 1.03283322e+00
-1.65997219e+00 -8.55157793e-01 -4.64022487e-01 2.25629115e+00
8.64468038e-01 -2.19615787e-01 1.00405671e-01 4.43803100e-03
3.68420660e-01 3.45422894e-01 -2.99593478e-01 -7.30389357e-01
-3.71274799e-01 4.07674760e-01 1.45140243e+00 2.70512402e-01
-1.26175356e+00 1.37552106e+00 7.66177082e+00 -9.26025137e-02
-1.32070994e+00 -5.06871164e-01 6.48135662e-01 5.27133405e-01
-9.38175097e-02 4.49082047e-01 -6.39236748e-01 1.28583536e-01
1.26301396e+00 -2.92477369e-01 3.45242500e-01 7.12237954e-01
7.44797528e-01 -7.82729089e-01 -3.57750714e-01 5.67468591e-02
-7.78603792e-01 -1.20500350e+00 -3.40698868e-01 2.62551662e-02
9.42178547e-01 3.00894499e-01 -1.46951124e-01 1.79046586e-01
1.11914468e+00 -9.33292389e-01 4.07239974e-01 4.99698341e-01
9.71111953e-01 -8.36844444e-01 8.39762866e-01 3.89781773e-01
-1.43726766e+00 4.33668215e-03 -8.66934836e-01 -4.18359041e-01
-1.06784292e-01 8.60504448e-01 -3.11919302e-01 1.03694117e+00
1.09093654e+00 9.34056222e-01 -2.58858144e-01 5.68868399e-01
5.12435175e-02 8.02452445e-01 -6.83408618e-01 5.02357006e-01
5.05362868e-01 -5.10086358e-01 -2.21945703e-01 1.25454783e+00
9.72165644e-01 3.81729931e-01 1.78035453e-01 5.74536264e-01
5.36625564e-01 2.94373482e-01 -8.49896193e-01 2.18630075e-01
5.25915623e-01 1.05595696e+00 -5.21335065e-01 1.35057032e-01
-2.66608715e-01 6.64753258e-01 -2.50540882e-01 4.39385623e-01
-3.65559578e-01 -4.45178509e-01 8.13891888e-01 -2.71633267e-01
1.79142982e-01 -3.69256496e-01 -4.27978277e-01 -9.48400915e-01
-1.64505333e-01 -7.04084814e-01 -1.42198414e-01 -9.90254819e-01
-9.50047314e-01 -2.37510502e-02 -6.63982704e-02 -9.63703990e-01
-6.29237533e-01 -8.23056817e-01 -9.10250723e-01 1.77725697e+00
-1.86674166e+00 -1.23603833e+00 -3.42026561e-01 1.15120113e-02
4.11492847e-02 -2.87234902e-01 1.72635925e+00 -3.50581050e-01
-5.87391675e-01 6.53428435e-02 8.45332742e-01 -7.89998621e-02
3.89938742e-01 -1.20725429e+00 1.11824977e+00 6.78527594e-01
-6.98567331e-01 2.72881478e-01 6.40137792e-01 -7.76593149e-01
-1.20399022e+00 -1.35017407e+00 1.32920659e+00 1.04752243e-01
1.05097830e+00 2.60349512e-01 -6.09673858e-01 8.96940649e-01
1.65321976e-01 2.95264293e-02 5.41920781e-01 5.21569967e-01
8.80123954e-03 -2.29925662e-01 -1.21219587e+00 3.66434544e-01
3.21711808e-01 -7.41528049e-02 3.08940187e-02 3.03366989e-01
4.88085061e-01 -4.53215212e-01 -1.54736888e+00 5.63820839e-01
1.02698684e+00 -6.65782034e-01 6.48672283e-01 -5.69353700e-01
5.08649111e-01 -2.91023515e-02 -6.26997411e-01 -2.01783776e+00
-7.10855067e-01 -4.62747961e-01 2.95899868e-01 1.34815907e+00
5.65647364e-01 -4.96533453e-01 3.85744035e-01 5.51367879e-01
3.95877033e-01 -2.11437434e-01 -3.64266872e-01 -6.52283192e-01
5.83811641e-01 2.95247864e-02 1.18700051e+00 1.01907039e+00
2.82166097e-02 -3.36762160e-01 -2.42943078e-01 7.94572473e-01
5.07695794e-01 3.81524801e-01 6.17021799e-01 -1.50713706e+00
5.57074010e-01 -2.23664448e-01 -3.83424833e-02 -1.92840040e-01
-1.61104441e-01 -9.56031680e-02 -1.62150487e-01 -1.28256643e+00
-2.68100388e-02 -6.36318326e-01 -1.97513133e-01 1.06016469e+00
-2.62918413e-01 -1.57629311e-01 -3.58525291e-02 -1.70855850e-01
8.09010386e-01 1.25426203e-01 7.16137707e-01 1.09749734e-01
-4.86858964e-01 9.49924886e-02 -5.52925408e-01 7.46861815e-01
1.13734674e+00 -4.31945413e-01 9.08195078e-02 -7.30419934e-01
1.82558611e-01 6.96234763e-01 2.62703717e-01 -7.53631115e-01
-5.69697499e-01 -1.09853697e+00 6.18698299e-01 -1.17567945e+00
-3.16022426e-01 -7.19113708e-01 4.54569697e-01 4.07834291e-01
8.61962214e-02 3.73845369e-01 7.34615028e-01 2.12957356e-02
-2.71027945e-02 1.12259597e-01 3.30567569e-01 -2.71405011e-01
-8.52075279e-01 1.08306840e-01 -5.85857034e-01 -6.93024993e-01
8.01455259e-01 1.61865540e-02 -2.62649447e-01 -8.85526612e-02
-8.21691453e-01 6.45459354e-01 5.78978658e-01 3.48565161e-01
-3.99635360e-02 -1.12186027e+00 -1.58978426e+00 3.32174003e-01
5.16860224e-02 -4.44794744e-01 -2.66483158e-01 4.20851499e-01
-1.08128762e+00 7.12445319e-01 -8.05581093e-01 -4.69149292e-01
-7.23336875e-01 1.62215799e-01 1.81691036e-01 -3.83001298e-01
-7.59331211e-02 4.15829659e-01 -4.12235469e-01 -8.50946367e-01
-3.70203555e-01 -7.99652934e-01 -1.96175247e-01 3.24978441e-01
3.48852396e-01 3.44159417e-02 1.44935504e-01 -6.92314804e-01
-2.59777129e-01 3.76158863e-01 4.58409339e-01 -6.06501102e-02
1.60873830e+00 -2.89710432e-01 -3.99165183e-01 7.15910792e-01
4.62208003e-01 -5.23094594e-01 -1.26169670e+00 1.84348114e-02
3.84708077e-01 -3.91265482e-01 4.10980225e-01 -1.09075522e+00
-1.11426055e+00 5.19119680e-01 8.51081550e-01 3.01228762e-01
1.43969917e+00 -9.50720608e-01 4.44157869e-01 2.48471901e-01
5.14187515e-01 -9.32237267e-01 -1.35824728e+00 7.79262006e-01
7.88366318e-01 -1.47585011e+00 3.29188466e-01 -7.70399794e-02
-3.16458642e-01 1.18122888e+00 1.70943782e-01 -6.57598078e-02
1.04080379e+00 5.33448756e-01 1.99647173e-01 3.78874183e-01
-7.19532013e-01 -3.70036513e-01 -3.80372882e-01 9.58091915e-01
1.06023347e+00 1.03086936e+00 -3.22882175e-01 4.29940552e-01
-4.26526070e-01 2.98254907e-01 3.96672815e-01 1.04256523e+00
-6.02058411e-01 -7.54093826e-01 -7.99935222e-01 6.43833041e-01
-4.14931834e-01 -4.10473853e-01 -2.77030263e-02 8.14531326e-01
-1.19906321e-01 1.08888364e+00 1.71860605e-01 -1.33230910e-01
2.74404705e-01 -7.49474615e-02 1.74162343e-01 -3.74394059e-01
-7.54289269e-01 2.06746444e-01 -9.39278305e-03 -4.00883257e-01
-4.55797374e-01 -9.56111252e-01 -8.23509872e-01 -1.28258359e+00
-5.41850805e-01 -1.13107167e-01 9.93491232e-01 6.73544586e-01
8.63946453e-02 6.32360503e-02 1.22493517e+00 -1.23161864e+00
-5.79112887e-01 -1.19801509e+00 -1.32805717e+00 -5.27205408e-01
3.28422725e-01 -4.36084837e-01 -1.34621114e-01 8.94867778e-02] | [9.369826316833496, -1.6022987365722656] |
cd9464ec-f045-49bd-a928-f4fd4e2d9465 | sequence-to-sequence-lexical-normalization | 2110.02869 | null | https://arxiv.org/abs/2110.02869v3 | https://arxiv.org/pdf/2110.02869v3.pdf | Sequence-to-Sequence Lexical Normalization with Multilingual Transformers | Current benchmark tasks for natural language processing contain text that is qualitatively different from the text used in informal day to day digital communication. This discrepancy has led to severe performance degradation of state-of-the-art NLP models when fine-tuned on real-world data. One way to resolve this issue is through lexical normalization, which is the process of transforming non-standard text, usually from social media, into a more standardized form. In this work, we propose a sentence-level sequence-to-sequence model based on mBART, which frames the problem as a machine translation problem. As the noisy text is a pervasive problem across languages, not just English, we leverage the multi-lingual pre-training of mBART to fine-tune it to our data. While current approaches mainly operate at the word or subword level, we argue that this approach is straightforward from a technical standpoint and builds upon existing pre-trained transformer networks. Our results show that while word-level, intrinsic, performance evaluation is behind other methods, our model improves performance on extrinsic, downstream tasks through normalization compared to models operating on raw, unprocessed, social media text. | ['Liviu P. Dinu', 'Adrian Cosma', 'Ana-Maria Bucur'] | 2021-10-06 | null | https://aclanthology.org/2021.wnut-1.53 | https://aclanthology.org/2021.wnut-1.53.pdf | wnut-acl-2021-11 | ['lexical-normalization'] | ['natural-language-processing'] | [ 6.55573308e-01 5.81636056e-02 -1.45078674e-01 -3.26702207e-01
-1.23378813e+00 -8.12221587e-01 8.32917929e-01 2.90901244e-01
-9.05270278e-01 7.95256853e-01 7.61968613e-01 -6.16536438e-01
4.43368316e-01 -6.32102907e-01 -7.97489285e-01 -2.75565624e-01
7.44196713e-01 4.39483494e-01 3.18967514e-02 -7.81699181e-01
-8.85452796e-03 2.84232143e-02 -8.67953718e-01 5.01630664e-01
6.84206724e-01 4.22801822e-01 4.52510677e-02 6.27113879e-01
-6.12748384e-01 6.83408976e-01 -5.71265042e-01 -1.05202496e+00
2.99863160e-01 -5.69977283e-01 -9.09621239e-01 -1.68237448e-01
4.30073142e-01 6.34245574e-02 -1.01587050e-01 1.32756901e+00
4.43854541e-01 -1.32873610e-01 5.67725420e-01 -8.74785304e-01
-9.26241517e-01 1.26134956e+00 -3.21914434e-01 2.25841388e-01
2.77447909e-01 1.16427608e-01 1.21418393e+00 -9.14245129e-01
9.05487478e-01 1.17091274e+00 8.54102969e-01 6.95676565e-01
-1.43545353e+00 -4.89511162e-01 1.15707412e-01 -9.60539803e-02
-1.12452614e+00 -8.04971159e-01 4.44896221e-01 -4.48858082e-01
1.33206224e+00 7.58368969e-02 1.81223422e-01 1.74556208e+00
3.73875290e-01 6.67107880e-01 1.01664972e+00 -8.47334921e-01
-2.29712725e-01 1.38680404e-02 -5.69133461e-02 2.22529292e-01
2.15213269e-01 -2.46998876e-01 -6.61043286e-01 1.38672546e-01
1.78526103e-01 -3.32367182e-01 -2.32795447e-01 6.01835102e-02
-1.47932553e+00 7.51154363e-01 -6.80526495e-02 6.46864891e-01
-2.72915035e-01 -1.14821441e-01 8.39485466e-01 7.12636590e-01
7.86854029e-01 6.05699241e-01 -8.23500693e-01 -4.18531984e-01
-1.12730527e+00 1.41007751e-02 1.03254175e+00 1.08485627e+00
5.58815837e-01 -1.54286355e-01 -4.42036629e-01 8.28626156e-01
1.42494604e-01 5.70610523e-01 7.98480690e-01 -3.36253792e-01
1.01600778e+00 3.37955832e-01 -2.27578193e-01 -7.12774277e-01
-2.39952937e-01 -5.55443943e-01 -7.56809592e-01 -3.55341166e-01
5.81859231e-01 -2.09956631e-01 -9.54723537e-01 1.97925222e+00
-6.32314235e-02 -3.58767986e-01 2.95115352e-01 5.95542014e-01
6.65552139e-01 6.66020334e-01 3.35307300e-01 -2.52386034e-01
1.23613656e+00 -1.09815669e+00 -8.77624393e-01 -5.23531616e-01
7.81617045e-01 -1.23263180e+00 1.60848939e+00 3.01546752e-01
-1.07320797e+00 -2.81687587e-01 -1.10831881e+00 -5.30863762e-01
-6.31059527e-01 -4.11684103e-02 1.12863868e-01 6.91664219e-01
-1.01984322e+00 5.55273652e-01 -6.39396191e-01 -8.77563059e-01
9.53140110e-02 -9.96227339e-02 -4.85874534e-01 -3.82299758e-02
-1.54669631e+00 1.32270873e+00 2.73088187e-01 -6.74459115e-02
-4.70662087e-01 -7.61107266e-01 -8.81792665e-01 -1.31523728e-01
3.25876802e-01 -6.73184872e-01 1.65156031e+00 -1.18081117e+00
-1.70768452e+00 1.13425183e+00 -2.42426127e-01 -6.52000904e-01
7.04241097e-01 -4.44686741e-01 -5.99252939e-01 -3.16822529e-01
3.03701073e-01 2.43036196e-01 8.60121250e-01 -8.17592800e-01
-5.48395097e-01 -5.42907156e-02 -1.28041402e-01 4.68858555e-02
-6.07541323e-01 3.96090984e-01 -4.71573532e-01 -9.02103305e-01
-2.74116874e-01 -7.31273055e-01 -2.48560816e-01 -5.49680531e-01
-5.15083909e-01 -6.25759289e-02 4.26807135e-01 -8.57967436e-01
1.34868467e+00 -1.86789942e+00 1.55739650e-01 4.07022871e-02
-9.27357599e-02 3.71169508e-01 -4.02835488e-01 8.44896913e-01
-4.03683484e-02 4.87110615e-01 -3.88879716e-01 -7.45266497e-01
1.60854012e-01 8.18594396e-02 -6.30018294e-01 3.73712391e-01
4.67731625e-01 1.14533746e+00 -1.02724814e+00 -3.33445728e-01
-1.73684675e-03 4.17884499e-01 -3.86474550e-01 8.94735605e-02
-2.24930197e-01 4.11187053e-01 -1.35208756e-01 2.79367298e-01
3.74726057e-01 4.40699384e-02 1.54924095e-01 -2.76609957e-01
-2.08368331e-01 1.05448389e+00 -7.00424790e-01 2.12701321e+00
-6.51962399e-01 6.42930567e-01 3.37901637e-02 -9.32454169e-01
5.66537082e-01 3.98873985e-01 2.98310816e-01 -7.29695559e-01
3.03380072e-01 3.42023939e-01 -1.11351965e-03 -4.97179359e-01
7.44976223e-01 -4.29629445e-01 -4.30847615e-01 5.79439461e-01
5.27662098e-01 -2.64725894e-01 6.53026819e-01 2.17402101e-01
1.42928493e+00 3.77641201e-01 5.13419807e-01 -2.89572060e-01
5.24273276e-01 1.88152015e-01 4.42921817e-01 7.81754434e-01
-2.10786492e-01 8.14354718e-01 2.31156468e-01 4.02656570e-03
-1.42077100e+00 -9.72375214e-01 1.58958454e-02 1.41218388e+00
-4.27538455e-01 -8.38386595e-01 -1.04070938e+00 -7.16565788e-01
-2.80592620e-01 7.67677605e-01 -5.17003536e-01 -2.52287030e-01
-7.74483323e-01 -8.95833492e-01 9.70950782e-01 2.32919827e-01
1.16995215e-01 -1.01130760e+00 9.72353593e-02 7.21763134e-01
-4.48633283e-01 -1.48613036e+00 -8.02818358e-01 3.16183239e-01
-4.19461340e-01 -6.25788152e-01 -4.73016560e-01 -6.07189000e-01
3.02368551e-01 -5.05231209e-02 1.48943877e+00 -2.24818587e-01
9.56017748e-02 3.24124575e-01 -6.83289468e-01 -5.82392752e-01
-9.35560226e-01 7.29700506e-01 3.09622642e-02 5.23192585e-02
7.50211000e-01 -6.01361156e-01 -1.83618777e-02 -2.37120345e-01
-1.08614898e+00 1.20102368e-01 7.12321639e-01 7.82079756e-01
4.83919263e-01 -3.67694259e-01 5.91046691e-01 -1.29764259e+00
1.03635693e+00 -4.47199345e-01 -2.37040907e-01 2.58267850e-01
-5.19435406e-01 2.16544166e-01 9.95942056e-01 -4.78721678e-01
-9.83935893e-01 -5.55898733e-02 -5.57548821e-01 2.07997710e-01
-8.55227634e-02 7.27208972e-01 -2.84126014e-01 2.73344338e-01
9.38651979e-01 2.77382672e-01 -1.76850826e-01 -4.69072193e-01
6.91537619e-01 7.92263687e-01 6.75019622e-01 -6.10660434e-01
1.12211847e+00 1.85982198e-01 -4.49294627e-01 -7.04789877e-01
-1.36737812e+00 -4.62523997e-01 -7.80149817e-01 1.36182368e-01
8.81771564e-01 -8.69633734e-01 -1.08113475e-01 2.47656837e-01
-1.41983020e+00 -4.04849648e-01 -4.51835990e-01 2.53778189e-01
-3.37324262e-01 1.79096743e-01 -9.57637727e-01 -3.40990514e-01
-6.95157051e-01 -8.46988738e-01 1.08572853e+00 -3.59581202e-01
-5.73574901e-01 -1.10435677e+00 2.01738209e-01 3.64706904e-01
6.60110235e-01 -3.90561260e-02 8.61986995e-01 -8.45553517e-01
-6.91920817e-02 -1.98430896e-01 -1.29263103e-01 7.08292961e-01
2.14636087e-01 -1.06986323e-02 -9.97429371e-01 -2.49683410e-01
1.50813237e-01 -3.33586007e-01 7.50279248e-01 -9.71040800e-02
6.73520207e-01 -2.96905071e-01 3.90230375e-03 5.49715459e-01
1.26838851e+00 -4.74113792e-01 5.02338231e-01 5.47396183e-01
8.59349191e-01 5.13144791e-01 3.78466040e-01 3.55516071e-03
5.27533054e-01 4.16257560e-01 4.40956913e-02 -2.64969558e-01
-3.65281820e-01 -6.11885726e-01 8.06378841e-01 1.54565489e+00
2.77332693e-01 -5.07551789e-01 -1.01579905e+00 6.25060558e-01
-1.73882556e+00 -7.94870794e-01 -2.16106385e-01 1.86961460e+00
1.38016629e+00 4.02420521e-01 -1.78709954e-01 2.58368757e-02
6.26342118e-01 7.02590942e-02 -1.97324216e-01 -4.66057420e-01
-4.74369615e-01 3.32867831e-01 7.85288632e-01 4.77168232e-01
-1.07757926e+00 1.40335560e+00 6.27125263e+00 1.02356923e+00
-1.39360595e+00 5.74416041e-01 3.40914726e-01 -1.53196901e-02
-3.78975302e-01 -6.88958094e-02 -1.01986134e+00 3.83607209e-01
1.39722836e+00 -4.33679551e-01 5.05415499e-01 4.73799169e-01
3.07141393e-01 4.55566794e-01 -1.44390428e+00 7.84002066e-01
1.97744906e-01 -1.22526979e+00 2.91646361e-01 -2.24851340e-01
6.00213647e-01 5.50875843e-01 -2.06515461e-01 6.40665174e-01
5.50393105e-01 -1.07949376e+00 1.11535382e+00 1.82967201e-01
8.46483588e-01 -4.68317270e-01 8.32231462e-01 4.94143665e-01
-9.82008815e-01 4.54431504e-01 -2.22688556e-01 -1.54584244e-01
4.36357677e-01 8.63456607e-01 -7.70911455e-01 5.74344814e-01
5.20033717e-01 7.85103261e-01 -6.85813248e-01 4.29074407e-01
-5.26632011e-01 9.35838223e-01 -2.46906906e-01 7.51836449e-02
3.14905733e-01 -3.04349810e-02 5.92187464e-01 1.98642826e+00
2.73798168e-01 -3.71566236e-01 -1.14309229e-02 6.49445117e-01
-4.90592629e-01 6.82587624e-01 -6.63050354e-01 -4.21299547e-01
1.50993153e-01 1.36133492e+00 -5.25611579e-01 -3.96559179e-01
-7.48236656e-01 1.09564447e+00 7.26058900e-01 3.54937077e-01
-6.25375390e-01 -3.63991439e-01 6.72200978e-01 9.89135131e-02
6.07833974e-02 -2.80884385e-01 -4.03556973e-01 -1.59593105e+00
3.17820519e-01 -1.25785136e+00 3.20965767e-01 -4.10962105e-01
-1.83879364e+00 8.76421928e-01 -3.76097560e-01 -1.23692322e+00
-2.33757839e-01 -7.72322297e-01 -2.69869298e-01 9.38333750e-01
-1.67906439e+00 -1.28579319e+00 2.37048641e-01 5.17026126e-01
6.88263834e-01 -1.46385357e-01 8.65151882e-01 5.60615718e-01
-4.12374288e-01 7.99433410e-01 1.90940380e-01 4.49963748e-01
1.26867151e+00 -1.14461863e+00 9.04500604e-01 1.37209558e+00
3.65908772e-01 8.54035914e-01 8.55727434e-01 -6.09266758e-01
-1.39007127e+00 -1.34902430e+00 1.48227429e+00 -7.93916881e-01
1.23787773e+00 -9.27629590e-01 -8.90556395e-01 7.14623332e-01
5.53527296e-01 -1.27253681e-01 8.08329582e-01 1.43211350e-01
-4.90506560e-01 3.28450240e-02 -8.74244571e-01 9.39322829e-01
1.22079599e+00 -8.93989027e-01 -9.57140923e-01 3.20753276e-01
1.10483384e+00 -4.07881021e-01 -8.47942770e-01 2.30378941e-01
2.71286756e-01 -4.27949697e-01 6.34299159e-01 -9.22540128e-01
7.18958914e-01 -2.16400847e-01 -3.16637009e-01 -1.70524871e+00
-2.28254691e-01 -9.34625626e-01 3.13782007e-01 1.63326156e+00
1.00538242e+00 -5.16293406e-01 3.83267403e-01 3.06385696e-01
-2.84483343e-01 -2.13910520e-01 -8.39819908e-01 -7.62239218e-01
5.60109377e-01 -8.44278455e-01 4.52296853e-01 1.14007807e+00
1.60285562e-01 9.53744292e-01 -3.05998951e-01 -2.51041502e-01
3.54547113e-01 -2.91421294e-01 7.67421126e-01 -9.28052545e-01
-1.54370606e-01 -4.84853119e-01 -1.61312625e-01 -9.99030113e-01
4.50292110e-01 -1.25300884e+00 3.39916855e-01 -1.37940156e+00
4.53367904e-02 -8.72568637e-02 -1.96321279e-01 5.05129695e-01
-1.61093384e-01 4.97679502e-01 1.54856086e-01 -7.71402568e-02
-4.57407892e-01 5.59855938e-01 1.10379195e+00 -1.84353620e-01
-4.71987240e-02 -4.28456873e-01 -8.86902988e-01 6.33717060e-01
7.89571106e-01 -7.45214343e-01 -2.45749615e-02 -8.60760391e-01
6.49727285e-01 -5.54354727e-01 -1.71348572e-01 -7.14318454e-01
2.36074075e-01 -1.46249920e-01 -1.08442843e-01 -1.29479572e-01
6.30699247e-02 -7.06047833e-01 -2.93823928e-01 9.89342555e-02
-5.21890938e-01 5.48438616e-02 2.44952261e-01 1.69673622e-01
-3.34872127e-01 -1.88236237e-01 6.81423247e-01 -1.77280486e-01
-2.67762244e-01 1.73082665e-01 -5.61210215e-01 4.96796757e-01
4.66104805e-01 1.30639195e-01 -5.94734251e-01 -3.96064609e-01
-5.10288596e-01 -1.30574346e-01 4.65876281e-01 7.40811944e-01
-1.98306348e-02 -1.17206264e+00 -1.09201038e+00 9.20142829e-02
3.27623755e-01 -2.94110417e-01 -3.87760520e-01 8.31339180e-01
-2.74591416e-01 4.36901361e-01 5.25440723e-02 -3.06626201e-01
-9.18674171e-01 4.42065269e-01 1.59492299e-01 -5.07954895e-01
-5.15398741e-01 6.48855627e-01 -1.46063522e-01 -8.11864376e-01
-3.08650546e-02 -5.09553671e-01 -4.48059440e-02 1.73397556e-01
5.25439739e-01 -1.20748192e-01 6.93822265e-01 -8.70739400e-01
-2.05460519e-01 3.62494946e-01 -3.05742532e-01 -4.30709600e-01
1.44736290e+00 -3.98482800e-01 -3.77897084e-01 7.31684387e-01
1.24160528e+00 3.76927763e-01 -7.11777151e-01 -5.65241337e-01
3.74924541e-01 -1.44851357e-02 -1.14641272e-01 -7.22613811e-01
-6.59059823e-01 8.52838874e-01 6.31899163e-02 1.80802390e-01
9.76054192e-01 -1.39934957e-01 9.94059920e-01 6.74763858e-01
2.23167002e-01 -1.47255635e+00 -2.04606563e-01 1.21792185e+00
9.43008423e-01 -1.38013947e+00 -2.24743471e-01 -4.49676245e-01
-4.26158816e-01 1.06563580e+00 3.83522183e-01 9.72473696e-02
4.67201889e-01 6.30235791e-01 4.88630354e-01 2.66557068e-01
-7.43919313e-01 -2.90408075e-01 1.82843745e-01 4.74098951e-01
9.20287907e-01 -1.28029272e-01 -5.80226004e-01 8.48875582e-01
-7.87250936e-01 -9.89694819e-02 6.25347316e-01 8.96061361e-01
-1.05339035e-01 -1.62116146e+00 -2.29935914e-01 4.32384104e-01
-9.60387945e-01 -7.56624520e-01 -3.75548065e-01 5.14151871e-01
3.05681676e-02 1.05119503e+00 -1.93710893e-01 -2.48379499e-01
4.48134840e-01 3.46252054e-01 3.72033507e-01 -1.02852273e+00
-9.13366616e-01 -1.51170865e-01 3.35008740e-01 -4.60618436e-01
-3.87623787e-01 -6.17050588e-01 -9.72395957e-01 -3.11138391e-01
3.75913009e-02 3.43018398e-02 7.86392748e-01 1.32535255e+00
1.15672491e-01 5.24126828e-01 1.47418976e-01 -8.01536620e-01
-6.07103407e-01 -1.33377814e+00 -2.89792158e-02 7.83826768e-01
3.95205140e-01 -8.73299316e-02 -3.42131823e-01 5.43838084e-01] | [10.763033866882324, 9.88841438293457] |
b922f0ea-20da-4f77-b851-4c2d073ba1e3 | can-we-read-speech-beyond-the-lips-rethinking | 2003.03206 | null | https://arxiv.org/abs/2003.03206v2 | https://arxiv.org/pdf/2003.03206v2.pdf | Can We Read Speech Beyond the Lips? Rethinking RoI Selection for Deep Visual Speech Recognition | Recent advances in deep learning have heightened interest among researchers in the field of visual speech recognition (VSR). Currently, most existing methods equate VSR with automatic lip reading, which attempts to recognise speech by analysing lip motion. However, human experience and psychological studies suggest that we do not always fix our gaze at each other's lips during a face-to-face conversation, but rather scan the whole face repetitively. This inspires us to revisit a fundamental yet somehow overlooked problem: can VSR models benefit from reading extraoral facial regions, i.e. beyond the lips? In this paper, we perform a comprehensive study to evaluate the effects of different facial regions with state-of-the-art VSR models, including the mouth, the whole face, the upper face, and even the cheeks. Experiments are conducted on both word-level and sentence-level benchmarks with different characteristics. We find that despite the complex variations of the data, incorporating information from extraoral facial regions, even the upper face, consistently benefits VSR performance. Furthermore, we introduce a simple yet effective method based on Cutout to learn more discriminative features for face-based VSR, hoping to maximise the utility of information encoded in different facial regions. Our experiments show obvious improvements over existing state-of-the-art methods that use only the lip region as inputs, a result we believe would probably provide the VSR community with some new and exciting insights. | ['Yuan-Hang Zhang', 'Jing-Yun Xiao', 'Shiguang Shan', 'Xilin Chen', 'Shuang Yang'] | 2020-03-06 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [ 2.37151593e-01 2.81992584e-01 -4.76772666e-01 -3.81450802e-01
-9.20382261e-01 -5.12539804e-01 6.44175291e-01 -5.02368033e-01
-2.41191223e-01 2.44334519e-01 6.24288499e-01 -3.58970761e-01
2.26073354e-01 -1.44975737e-01 -4.25140083e-01 -7.82867968e-01
2.23101422e-01 -5.99454008e-02 -4.20755111e-02 -3.84798199e-01
3.14974308e-01 7.84548700e-01 -1.88422990e+00 5.22837996e-01
2.85573274e-01 1.10046077e+00 -1.53130088e-02 5.19319117e-01
-1.61551610e-01 3.32520813e-01 -5.39609671e-01 -5.18249869e-01
5.44703603e-02 -2.73390383e-01 -8.33707869e-01 2.45063826e-01
7.09775925e-01 -3.45145583e-01 -2.90226370e-01 7.41195858e-01
9.18167949e-01 9.90543216e-02 5.72634637e-01 -1.02200425e+00
-7.50262320e-01 1.65644333e-01 -9.35697913e-01 1.65023327e-01
6.05050266e-01 4.58979994e-01 1.13338363e+00 -1.00234377e+00
6.12714052e-01 1.57018578e+00 4.67066646e-01 1.09201050e+00
-9.86427605e-01 -4.39807802e-01 2.27126539e-01 2.80602217e-01
-1.43184757e+00 -1.36237812e+00 1.02265489e+00 -8.86287093e-02
9.39465046e-01 4.96561319e-01 3.28876853e-01 1.38444912e+00
-1.31316446e-02 1.05059290e+00 1.27605379e+00 -4.70968157e-01
-2.16818437e-01 1.77811399e-01 -2.11275667e-01 5.26050389e-01
-3.90310735e-01 2.34722748e-01 -8.53515029e-01 8.31685290e-02
2.47274980e-01 -3.30090404e-01 -5.75042963e-01 -1.81002796e-01
-8.32120478e-01 7.55814493e-01 1.24693550e-01 3.67502600e-01
-1.70171261e-01 -1.11462444e-01 4.67569441e-01 1.69033960e-01
6.24211252e-01 -1.82292029e-01 -5.11840999e-01 -2.16596320e-01
-1.15233624e+00 -2.17496291e-01 6.55497730e-01 4.42167670e-01
5.14157236e-01 -7.70651028e-02 -1.02104813e-01 1.44304335e+00
6.54556453e-01 4.33985800e-01 5.56885779e-01 -9.71092939e-01
2.92367399e-01 1.84940517e-01 -4.17718500e-01 -6.35837376e-01
-1.49605915e-01 -4.81434315e-02 -5.39494932e-01 3.85984033e-01
3.69743049e-01 -2.82424632e-02 -1.08076644e+00 1.67605245e+00
2.59533614e-01 -1.46089330e-01 1.11820020e-01 8.98147762e-01
1.17100036e+00 3.72097105e-01 -8.59160498e-02 -2.18396604e-01
1.49148571e+00 -9.35037553e-01 -7.74364829e-01 -3.44563812e-01
3.65435988e-01 -8.68957818e-01 1.18218446e+00 2.35383153e-01
-1.18451786e+00 -5.56140006e-01 -7.09591389e-01 -1.74331412e-01
-3.96696806e-01 2.42371172e-01 3.11050057e-01 8.19188416e-01
-1.53306365e+00 3.05826604e-01 -3.95525575e-01 -4.75207120e-01
5.43899655e-01 3.87967467e-01 -7.60925949e-01 3.49416398e-02
-1.05844963e+00 9.83074725e-01 -3.55388850e-01 2.10097089e-01
-5.92733026e-01 -2.75543958e-01 -1.06584311e+00 3.43801156e-02
4.14486647e-01 -3.28773856e-01 1.41490412e+00 -1.42554808e+00
-1.88500977e+00 1.33939683e+00 -9.19960141e-01 -1.80432796e-01
4.08725798e-01 -4.92938142e-03 -3.94234627e-01 3.26614767e-01
-3.91682237e-01 1.00249124e+00 1.20951557e+00 -1.36060989e+00
-1.26297399e-01 -4.75745857e-01 -1.39095902e-01 1.54498652e-01
-9.16378945e-03 3.80946517e-01 -4.11561728e-01 -5.61848462e-01
3.76078277e-03 -8.04141879e-01 4.91465360e-01 3.76920015e-01
-2.34274596e-01 -6.61782324e-01 8.59984756e-01 -9.27441120e-01
8.05721045e-01 -2.49772286e+00 -2.39080898e-02 -1.02203570e-01
1.34107023e-01 6.04849458e-01 -3.03643763e-01 2.49532595e-01
-3.22638065e-01 3.69024813e-01 -6.58913180e-02 -6.44163966e-01
-1.08334996e-01 7.04896525e-02 -1.64349839e-01 6.44413769e-01
2.55871773e-01 1.08120584e+00 -4.03614491e-01 -5.57499111e-01
2.50173956e-01 8.90514135e-01 -2.66423106e-01 2.87973396e-02
4.13455218e-02 1.59988537e-01 7.08043426e-02 7.84269333e-01
8.19450200e-01 -1.29293799e-02 -2.23586187e-02 -7.93762878e-02
1.31938830e-01 6.07779264e-01 -7.25549996e-01 1.47471821e+00
-4.38836873e-01 1.08036697e+00 4.85541701e-01 -9.74403799e-01
8.84967744e-01 4.99331355e-01 3.64950150e-01 -9.39092517e-01
9.91327316e-02 8.33424479e-02 7.23586157e-02 -6.90324426e-01
1.69977620e-01 -3.80168200e-01 5.71336031e-01 2.89581001e-01
9.15868133e-02 8.25509802e-02 -4.57652748e-01 -8.24389607e-02
5.63335478e-01 -6.98850453e-02 2.20683083e-01 -1.48583623e-02
8.48816335e-01 -7.77492285e-01 1.14601925e-01 1.75078452e-01
-6.27931297e-01 8.23351741e-01 4.20731843e-01 -4.48371954e-02
-5.58758974e-01 -1.04890013e+00 -4.14384067e-01 1.26499987e+00
-6.20744564e-02 -1.65203452e-01 -9.44629133e-01 -8.07582915e-01
-5.43874428e-02 5.36575139e-01 -7.11862922e-01 -1.18124112e-01
-3.43549162e-01 -1.93455935e-01 6.20268047e-01 4.30030316e-01
4.01554972e-01 -1.34338808e+00 -3.59957218e-01 -2.93908924e-01
-3.90922993e-01 -1.09549677e+00 -6.05294824e-01 -1.63786501e-01
-5.58591664e-01 -7.25573957e-01 -1.23327494e+00 -7.14153469e-01
3.59494716e-01 4.21560794e-01 9.74586070e-01 1.77768156e-01
-2.15187982e-01 6.85224533e-01 -3.46203446e-01 -3.64959121e-01
-5.50320923e-01 -8.59676376e-02 -2.10517906e-02 3.27564180e-01
5.05493462e-01 -2.03069538e-01 -4.64888930e-01 4.53326881e-01
-5.76670408e-01 -2.29476169e-01 5.44516802e-01 6.31850719e-01
2.18333364e-01 -4.31715697e-01 7.32704878e-01 -2.78770387e-01
5.75433016e-01 -2.95855522e-01 -1.02463417e-01 3.22382122e-01
-1.58976406e-01 -5.12811355e-03 3.96761149e-02 -4.38048005e-01
-1.01328743e+00 -1.57035649e-01 -7.98029304e-01 -4.28005069e-01
-4.62474912e-01 4.28121258e-03 -3.51975441e-01 -1.46280363e-01
3.86468887e-01 4.47440743e-01 5.95638990e-01 -4.27704573e-01
3.77360165e-01 1.33105707e+00 3.79132003e-01 -2.84886777e-01
3.61170769e-01 5.55320859e-01 -3.15405250e-01 -1.45174432e+00
-4.90829408e-01 -6.62278354e-01 -5.73786438e-01 -4.48661804e-01
7.95969605e-01 -7.09029615e-01 -1.06771123e+00 4.35356051e-01
-1.18063653e+00 -2.61915326e-01 -9.16818157e-02 8.43589380e-02
-6.41138077e-01 4.85483050e-01 -3.64669055e-01 -1.08096421e+00
-2.76982427e-01 -1.35539448e+00 1.58581078e+00 8.90873224e-02
-3.93421322e-01 -9.52254355e-01 -5.77411987e-02 8.37432265e-01
4.74591464e-01 -2.50671536e-01 5.33564508e-01 -6.02844238e-01
-3.44395548e-01 9.79708508e-02 -2.44085446e-01 7.50794888e-01
4.36406106e-01 1.09484568e-01 -1.59974563e+00 -2.51211494e-01
2.14819964e-02 -4.24990624e-01 1.17192841e+00 4.63278860e-01
1.12664258e+00 -3.22417617e-01 -3.96203905e-01 3.95543098e-01
8.63401473e-01 1.07659101e-01 9.24575925e-01 -1.51088372e-01
3.64059478e-01 1.11304092e+00 2.52085537e-01 -3.42929699e-02
2.28461176e-01 8.41755807e-01 3.73959094e-01 -2.43600309e-01
-6.06902480e-01 2.33678594e-02 8.38739038e-01 6.27005517e-01
-3.32516804e-02 -4.86886591e-01 -6.66008949e-01 6.15019977e-01
-1.05482674e+00 -1.03577948e+00 4.67879027e-01 1.99086034e+00
8.19918633e-01 -1.92280218e-01 2.27457881e-01 3.49788934e-01
7.01669812e-01 6.24996722e-01 -4.33541298e-01 -7.62760699e-01
-1.84812710e-01 1.79790109e-01 3.31200361e-02 6.92565084e-01
-8.53207529e-01 1.01226044e+00 6.60735035e+00 8.90681863e-01
-1.81311846e+00 -7.98981339e-02 7.66444087e-01 -2.62900710e-01
-2.72773921e-01 -4.96930838e-01 -9.40860808e-01 3.97395313e-01
1.02457333e+00 2.98231691e-01 6.20915353e-01 6.08135760e-01
3.60162377e-01 -8.75235200e-02 -9.64811683e-01 1.14526260e+00
4.45763290e-01 -1.19770324e+00 -1.21364415e-01 2.92062938e-01
1.82525933e-01 2.11512640e-01 3.92991990e-01 -4.84468155e-02
-3.33848506e-01 -1.39905214e+00 5.39542437e-01 4.06606555e-01
1.07656682e+00 -4.75183010e-01 4.51738387e-01 1.85190588e-01
-1.05020058e+00 9.53336582e-02 6.08557323e-03 4.03504789e-01
-6.62591383e-02 3.94639261e-02 -9.58548069e-01 2.43109643e-01
6.46402180e-01 5.89775205e-01 -4.40863043e-01 6.17657244e-01
-8.66201073e-02 5.81298947e-01 -2.65434057e-01 4.18266691e-02
1.09094821e-01 3.56879592e-01 5.58721960e-01 1.20769608e+00
-3.05161383e-02 -4.84566689e-02 -5.71057618e-01 6.71568215e-01
-2.75279999e-01 2.68621892e-01 -8.61148596e-01 3.12520266e-02
1.68759301e-01 1.11204290e+00 -3.36183965e-01 2.98084668e-03
-7.75543511e-01 9.23934042e-01 1.27506047e-01 4.79211539e-01
-5.22684753e-01 5.76720573e-02 9.89939868e-01 4.61735636e-01
4.30125773e-01 -2.89414190e-02 -1.10234350e-01 -9.21762109e-01
1.73491195e-01 -1.25425327e+00 5.41051365e-02 -8.48048985e-01
-1.04660428e+00 5.72058737e-01 -2.72333443e-01 -7.66962171e-01
-2.82031298e-01 -9.14188862e-01 -5.29204130e-01 9.27061796e-01
-1.93852115e+00 -9.57529485e-01 -1.36848800e-02 7.02361643e-01
9.09435928e-01 -9.30437595e-02 6.31227493e-01 5.94454780e-02
-2.88617849e-01 1.12838447e+00 -1.71021894e-01 4.15529042e-01
8.19451869e-01 -7.07124054e-01 4.05170888e-01 3.41244340e-01
3.96804959e-01 6.73910081e-01 3.05900812e-01 -2.82068998e-01
-1.27183533e+00 -5.06053567e-01 9.64958012e-01 -4.83896345e-01
2.05018237e-01 -4.40518200e-01 -9.93050814e-01 5.19996405e-01
4.91116017e-01 6.53798133e-02 6.14404619e-01 1.15846112e-01
-4.30631042e-01 -1.56600088e-01 -1.23389995e+00 5.71690142e-01
9.32784200e-01 -1.02586854e+00 -6.92620933e-01 2.93686166e-02
4.99297589e-01 -1.29644141e-01 -5.60576379e-01 5.26707888e-01
7.70509481e-01 -1.01338577e+00 1.11693966e+00 -4.36047554e-01
2.64764249e-01 5.12860380e-02 -2.29288578e-01 -1.21191430e+00
2.51340449e-01 -7.90573716e-01 -7.71107823e-02 1.37385869e+00
3.65620375e-01 -7.43773162e-01 6.93930686e-01 2.37101853e-01
5.06915785e-02 -9.23685312e-01 -1.23793232e+00 -6.58906281e-01
2.45185927e-01 -4.94911611e-01 3.24345291e-01 6.98137105e-01
-5.89465573e-02 3.98106575e-01 -2.37972841e-01 -6.08750731e-02
2.56722271e-01 -2.25722656e-01 5.88133216e-01 -1.14772582e+00
-4.29146597e-03 -6.18530869e-01 -3.47868055e-01 -1.18925452e+00
6.46812856e-01 -8.37323129e-01 1.24841891e-01 -1.36410582e+00
-2.28560884e-02 3.65712866e-02 -1.11094110e-01 5.90285778e-01
-2.58228406e-02 2.23777369e-01 1.83161676e-01 -7.51375854e-02
-2.03875676e-01 5.27989328e-01 1.56586945e+00 -1.56715736e-01
1.02805465e-01 8.40829238e-02 -7.77152300e-01 7.69577384e-01
7.06559360e-01 -6.15291148e-02 -3.16516429e-01 -1.88920125e-01
-2.82692879e-01 1.52768612e-01 5.00115514e-01 -5.34812510e-01
1.96006134e-01 -3.17925438e-02 2.94278830e-01 -4.00552690e-01
8.18292141e-01 -5.79656184e-01 -4.65208322e-01 -1.22347735e-02
-4.31224316e-01 -2.68443972e-01 4.14952964e-01 1.70419827e-01
-2.76287347e-01 -1.34428173e-01 1.21433592e+00 -8.68118033e-02
-6.63101017e-01 5.08606657e-02 -4.95293260e-01 1.84462234e-01
7.92002559e-01 -3.31762791e-01 -2.37509847e-01 -8.05968821e-01
-8.71894538e-01 -1.98131412e-01 4.01489347e-01 7.76044130e-01
8.76986384e-01 -8.98718059e-01 -7.70597219e-01 6.16787076e-01
4.06884402e-02 -3.95921081e-01 2.12372273e-01 9.12518442e-01
4.58901152e-02 5.58189094e-01 5.62179498e-02 -8.55523586e-01
-1.81034970e+00 5.68644524e-01 6.82916164e-01 3.99611831e-01
-5.31431794e-01 1.03852952e+00 2.18118712e-01 -3.99974853e-01
4.76938874e-01 -7.17102215e-02 -3.16154212e-01 4.16708887e-01
6.02006972e-01 1.81620076e-01 1.79290980e-01 -1.19678211e+00
-5.83092451e-01 9.00584280e-01 -8.54243785e-02 -2.11833641e-01
9.63326633e-01 -3.47811788e-01 6.22167066e-02 3.80815327e-01
1.61737370e+00 3.12034667e-01 -1.09614956e+00 -1.54357985e-01
-2.57456064e-01 -4.88592118e-01 1.74264520e-01 -9.70427871e-01
-1.27056599e+00 1.44979274e+00 7.42765784e-01 1.80616319e-01
1.18931985e+00 4.63789225e-01 7.03530550e-01 9.29850042e-02
9.73244533e-02 -9.62344170e-01 2.13305339e-01 2.28783056e-01
1.24483073e+00 -1.49532795e+00 -2.79846162e-01 -2.57405698e-01
-9.70739722e-01 1.05226254e+00 2.37387389e-01 3.63438815e-01
7.15656698e-01 2.40061209e-01 4.91583347e-01 3.84474434e-02
-8.66976142e-01 -4.00608212e-01 5.99496663e-01 6.54155672e-01
6.99199021e-01 -1.12473346e-01 1.10512845e-01 9.61546227e-02
-1.02856189e-01 -1.32474199e-01 1.32628053e-01 5.47138751e-01
-4.86240894e-01 -1.07181096e+00 -3.94736797e-01 2.58241057e-01
-6.28190517e-01 -2.03176051e-01 -7.75277495e-01 8.55481744e-01
-3.06300484e-02 1.15105891e+00 2.73355134e-02 -1.71262339e-01
1.62740052e-01 4.45852190e-01 4.70746249e-01 -4.35444146e-01
-2.36585036e-01 1.54186562e-01 8.15221518e-02 -6.99116409e-01
-4.59493876e-01 -8.96155715e-01 -1.06461525e+00 -3.45195860e-01
-4.07058209e-01 -1.99214250e-01 9.46490705e-01 1.07199907e+00
4.31754857e-01 1.92957401e-01 6.92559898e-01 -9.59728599e-01
-6.28583908e-01 -1.11587524e+00 -3.52811843e-01 2.10700408e-01
1.02316713e+00 -7.79891789e-01 -7.82269299e-01 -4.90625612e-02] | [14.327417373657227, 5.017500877380371] |
6fe4caf1-15f4-4b80-9bd5-62d1cca36a5b | clustering-word-embeddings-with-self-1 | null | null | https://aclanthology.org/2021.eacl-main.81 | https://aclanthology.org/2021.eacl-main.81.pdf | Clustering Word Embeddings with Self-Organizing Maps. Application on LaRoSeDa - A Large Romanian Sentiment Data Set | Romanian is one of the understudied languages in computational linguistics, with few resources available for the development of natural language processing tools. In this paper, we introduce LaRoSeDa, a Large Romanian Sentiment Data Set, which is composed of 15,000 positive and negative reviews collected from the largest Romanian e-commerce platform. We employ two sentiment classification methods as baselines for our new data set, one based on low-level features (character n-grams) and one based on high-level features (bag-of-word-embeddings generated by clustering word embeddings with k-means). As an additional contribution, we replace the k-means clustering algorithm with self-organizing maps (SOMs), obtaining better results because the generated clusters of word embeddings are closer to the Zipf{'}s law distribution, which is known to govern natural language. We also demonstrate the generalization capacity of using SOMs for the clustering of word embeddings on another recently-introduced Romanian data set, for text categorization by topic. | ['Radu Tudor Ionescu', 'Gaman Mihaela', 'Anca Tache'] | 2021-04-01 | null | null | null | eacl-2021-2 | ['text-categorization'] | ['natural-language-processing'] | [-3.09189379e-01 4.39685099e-02 -3.15026492e-01 -5.47605813e-01
-1.17969483e-01 -5.66841006e-01 8.99117053e-01 9.19216335e-01
-8.81903529e-01 2.29435384e-01 5.78863680e-01 -2.47084975e-01
-1.39440939e-01 -1.14234114e+00 -1.09853633e-01 -8.02020431e-01
-6.75526112e-02 6.11221731e-01 -1.01224430e-01 -5.11991799e-01
7.23232985e-01 3.14041764e-01 -1.90236843e+00 4.04507577e-01
7.62345552e-01 5.25422931e-01 9.84743424e-03 3.79980236e-01
-5.80689549e-01 4.17327255e-01 -4.85309720e-01 -5.19107401e-01
-5.12589961e-02 -1.48831189e-01 -6.64508700e-01 -2.94276010e-02
4.23786901e-02 3.27512771e-01 -1.64959460e-01 1.02597511e+00
2.67000049e-01 4.25090969e-01 1.16503692e+00 -8.81654263e-01
-1.25167441e+00 6.51856363e-01 -4.81386870e-01 3.16096097e-02
5.65388054e-02 -5.09921134e-01 1.51117146e+00 -1.16900384e+00
7.99369156e-01 1.54275513e+00 4.86459821e-01 5.62515080e-01
-1.03471160e+00 -1.88814834e-01 -6.88008144e-02 2.34482244e-01
-1.14342082e+00 -4.43131849e-02 7.26314425e-01 -6.25833690e-01
9.50034559e-01 -1.19111545e-01 6.01075292e-01 7.94093847e-01
3.79121095e-01 5.46965957e-01 1.12198412e+00 -8.62622142e-01
6.55283332e-01 8.03740084e-01 5.61284900e-01 2.08594009e-01
4.52157915e-01 -3.30159575e-01 -4.78629559e-01 -3.88753831e-01
-3.82491946e-02 2.48718828e-01 2.37118453e-01 -5.79583585e-01
-9.19843376e-01 1.44022191e+00 1.83127612e-01 7.53164470e-01
-4.67991859e-01 -2.57345110e-01 6.08228385e-01 2.55178988e-01
8.87700081e-01 6.56213164e-01 -6.21831179e-01 -1.79334983e-01
-6.76285684e-01 1.32106587e-01 6.82369649e-01 5.93158901e-01
6.50886655e-01 -1.49103254e-01 8.91022831e-02 1.41876090e+00
6.99109018e-01 5.13930678e-01 1.31322229e+00 -1.24280728e-01
1.84689865e-01 9.41874743e-01 -1.11346707e-01 -1.40171623e+00
-5.03187478e-01 2.11783182e-02 -4.32621360e-01 -3.79632451e-02
2.68313497e-01 -1.49253577e-01 -6.36726439e-01 1.31461895e+00
2.17942223e-01 -6.35509968e-01 4.88573343e-01 7.12099195e-01
7.67365515e-01 8.39896381e-01 1.30616784e-01 4.16070856e-02
1.78918386e+00 -8.81619096e-01 -6.06304049e-01 -1.39702767e-01
9.43053186e-01 -6.58644438e-01 1.31281698e+00 4.48573887e-01
-4.17467713e-01 -3.84274006e-01 -1.00879467e+00 -1.94637150e-01
-1.30898583e+00 -1.29185870e-01 6.88288927e-01 1.10323739e+00
-9.79166687e-01 5.17019629e-01 -4.83768135e-01 -8.45766068e-01
2.44175047e-01 1.29700020e-01 -5.97554684e-01 9.10860896e-02
-1.37370300e+00 9.78139758e-01 3.85778695e-01 -3.40868622e-01
-6.26414269e-02 -2.95290858e-01 -1.16754735e+00 1.55619839e-02
-3.07882100e-01 1.92344978e-01 6.48448408e-01 -7.75014520e-01
-1.38977110e+00 1.17069650e+00 -2.03011870e-01 -2.51816928e-01
-2.84022599e-01 9.17043313e-02 -6.65893853e-01 1.59243599e-01
1.65548638e-01 6.94631517e-01 6.41821861e-01 -1.25105965e+00
-5.77112079e-01 -7.70848274e-01 -1.95990011e-01 5.48470169e-02
-1.50730538e+00 1.20142177e-01 -1.66156709e-01 -7.02014863e-01
5.11167012e-02 -8.55729938e-01 -3.19796234e-01 -5.03199339e-01
6.98926002e-02 -7.69817948e-01 5.49215257e-01 -4.10806417e-01
1.43448114e+00 -2.44442487e+00 -7.29306415e-02 4.68677312e-01
1.99489012e-01 2.86913455e-01 -2.37824917e-01 6.94059074e-01
-2.18763545e-01 2.63273120e-01 -1.28590867e-01 -4.05985743e-01
1.69986799e-01 4.47497487e-01 -3.14865053e-01 4.46267605e-01
2.71223336e-01 5.12924433e-01 -1.08801281e+00 -4.44097459e-01
2.09909350e-01 4.47031796e-01 -5.04620016e-01 -4.13329929e-01
1.33530006e-01 -3.92892271e-01 -2.64805526e-01 4.57043350e-01
4.32762742e-01 2.23097160e-01 3.91834289e-01 1.67353675e-01
-3.13474417e-01 2.60139436e-01 -7.62113273e-01 1.16137242e+00
-6.19832456e-01 8.61373127e-01 -4.00743544e-01 -1.11966550e+00
1.42253351e+00 1.86741307e-01 4.00630414e-01 -5.71152449e-01
4.05834019e-01 1.33795321e-01 3.83040644e-02 -2.86989033e-01
1.08118212e+00 -2.15952560e-01 -4.73858923e-01 6.67071998e-01
3.45947713e-01 -3.05808008e-01 5.99491537e-01 2.84121931e-01
7.10577607e-01 -5.13808370e-01 2.80511230e-01 -9.87399161e-01
4.98456240e-01 2.34092236e-01 2.18504339e-01 4.64137256e-01
-1.97285533e-01 6.41195893e-01 8.88176501e-01 -4.62477684e-01
-1.12501585e+00 -7.51126885e-01 -7.11896539e-01 1.57918966e+00
-1.82255089e-01 -8.13490093e-01 -7.98427105e-01 -5.95989823e-01
2.54839480e-01 9.88959908e-01 -1.04670274e+00 -6.72920048e-02
-5.46577983e-02 -1.11375427e+00 3.43960583e-01 3.02898735e-01
-2.63583153e-01 -1.37252736e+00 -2.83176720e-01 8.86626989e-02
1.99815616e-01 -6.64773703e-01 -1.78139746e-01 5.57126820e-01
-8.07931066e-01 -7.69132137e-01 -5.38933754e-01 -1.08160460e+00
8.22255909e-01 1.60959542e-01 1.08198142e+00 -1.20366722e-01
-1.54540136e-01 2.73812771e-01 -8.62657845e-01 -8.16369891e-01
-3.51242751e-01 2.74440885e-01 5.00537634e-01 1.82579421e-02
1.35297883e+00 -2.22511441e-01 -1.99352935e-01 1.40659556e-01
-1.27001631e+00 -5.54619908e-01 1.78611904e-01 8.65904570e-01
1.77231342e-01 2.82252669e-01 6.82039678e-01 -1.23120975e+00
1.07990968e+00 -5.72236776e-01 -2.43310183e-01 -1.71912655e-01
-8.98280799e-01 1.18370220e-01 8.36854100e-01 -2.95077741e-01
-7.57029355e-01 -2.78152883e-01 -1.01026691e-01 2.26433307e-01
-2.47227266e-01 5.96863151e-01 1.90318242e-01 3.85089159e-01
9.43874836e-01 4.49138917e-02 8.07018802e-02 -3.25813442e-01
7.24108577e-01 1.42317092e+00 1.93955809e-01 -5.33250272e-01
5.36754191e-01 5.54286420e-01 -5.24765253e-01 -1.33479381e+00
-5.47377884e-01 -9.93746579e-01 -9.47493613e-01 4.68853116e-02
8.77704203e-01 -6.66627765e-01 -3.45008999e-01 2.59075075e-01
-9.72252667e-01 1.84039921e-01 -4.99392033e-01 5.29965997e-01
-2.42990628e-01 3.34733576e-01 -6.11055076e-01 -7.55567670e-01
-2.94936746e-01 -7.30987251e-01 7.90290594e-01 1.46446377e-01
-6.24069214e-01 -1.51753628e+00 4.93932515e-01 1.64785609e-01
2.10066751e-01 -1.26776591e-01 1.24467337e+00 -1.21470094e+00
7.81663001e-01 -4.28424895e-01 -2.60009002e-02 6.92146242e-01
1.80274025e-01 2.45523021e-01 -1.10670221e+00 -1.27451494e-01
1.55757397e-01 -1.56910017e-01 1.00949764e+00 1.43670052e-01
8.04571807e-01 -4.13756222e-02 -2.25184020e-03 -1.37990385e-01
1.37165380e+00 2.69760787e-01 7.10237682e-01 6.25747263e-01
5.17678499e-01 1.04225433e+00 5.37196279e-01 5.80887556e-01
5.75696647e-01 1.12358630e-01 -2.74247937e-02 1.53747192e-02
4.74201322e-01 -2.53540546e-01 5.04874825e-01 1.47808087e+00
1.85476363e-01 -3.15940268e-02 -1.16884422e+00 1.03442311e+00
-1.76030493e+00 -6.71135008e-01 -2.27997676e-01 2.02177215e+00
7.14779973e-01 1.25903055e-01 6.95459545e-02 5.35901248e-01
7.24600852e-01 1.37385219e-01 -1.13937393e-01 -1.07952011e+00
-1.57974571e-01 3.74209553e-01 4.14051950e-01 4.40835506e-01
-1.18474066e+00 1.18369460e+00 5.85186338e+00 8.27182174e-01
-9.43972170e-01 5.02556115e-02 4.25367564e-01 1.87977493e-01
-5.86842716e-01 -2.94672817e-01 -6.49332404e-01 4.25379395e-01
1.11923003e+00 -1.92381844e-01 -6.36346787e-02 1.16020405e+00
2.03262776e-01 -1.98845595e-01 -7.17519879e-01 7.93125570e-01
6.71522319e-01 -9.95479345e-01 8.71807337e-02 3.26325707e-02
8.29543769e-01 -6.77315295e-02 1.16239749e-01 4.90696460e-01
3.50851297e-01 -8.98589849e-01 4.77773607e-01 -6.26396313e-02
4.30377275e-01 -9.65620756e-01 1.04586518e+00 1.44433156e-02
-8.79874527e-01 -1.17958106e-01 -8.86888266e-01 -2.64062226e-01
-1.78001761e-01 7.26751685e-01 -6.37456834e-01 1.63688958e-01
8.97533178e-01 1.06113338e+00 -7.75188208e-01 3.99174243e-01
-6.82841167e-02 7.13567078e-01 -1.93006024e-02 -8.08318436e-01
4.50165182e-01 -5.76349795e-01 5.60255796e-02 1.49690664e+00
-2.10969988e-02 -2.48239622e-01 -2.01714367e-01 3.99276435e-01
9.26401652e-03 7.21031845e-01 -8.03482056e-01 -3.71653616e-01
3.66959274e-01 1.63047063e+00 -1.11431110e+00 -3.85900587e-01
-5.14327466e-01 5.75034618e-01 2.20219925e-01 7.52397850e-02
-2.51863480e-01 -1.14456630e+00 6.89733326e-01 2.77604852e-02
1.87421218e-01 -2.21438125e-01 -6.48062408e-01 -1.18825555e+00
-3.22222769e-01 -6.74962997e-01 2.73553461e-01 -5.17929733e-01
-1.77602899e+00 6.98895097e-01 -1.50941610e-01 -9.92901325e-01
-1.79602474e-01 -1.28020978e+00 -4.61664975e-01 6.92970335e-01
-1.30368948e+00 -6.66680276e-01 2.29313657e-01 4.00694937e-01
4.79414791e-01 -5.46689928e-01 1.18275964e+00 2.63858795e-01
-2.93338895e-01 3.46731484e-01 7.76554227e-01 3.82292032e-01
8.94630969e-01 -1.63156152e+00 4.08180177e-01 2.91376501e-01
2.31494352e-01 8.68630528e-01 5.55245996e-01 -2.83658952e-01
-1.13494754e+00 -7.99764156e-01 1.58523011e+00 -7.06160903e-01
1.29328120e+00 -7.08394527e-01 -7.21039772e-01 2.94539183e-01
1.70003623e-01 -3.64054084e-01 1.30631971e+00 5.96364796e-01
-4.09700543e-01 3.40559669e-02 -1.20123422e+00 6.46858454e-01
2.69411951e-01 -6.29688978e-01 -1.02678156e+00 3.76328975e-01
5.16977608e-01 5.10236382e-01 -9.89204526e-01 -1.84917480e-01
5.45010746e-01 -7.84249961e-01 6.38273120e-01 -6.31469727e-01
8.33605111e-01 -1.13641590e-01 -3.80552173e-01 -1.72963774e+00
-3.10031056e-01 6.71062544e-02 6.33179128e-01 1.29002988e+00
6.29697800e-01 -7.80070424e-01 7.32105374e-01 3.26119781e-01
8.17955360e-02 -6.35268152e-01 -6.48339689e-01 -4.95943427e-01
6.73772454e-01 -5.30981958e-01 4.23689902e-01 1.21084499e+00
8.12271714e-01 4.89218324e-01 6.72219992e-02 -3.59110296e-01
2.22596556e-01 -8.45217779e-02 5.16012132e-01 -1.44422984e+00
2.50399619e-01 -4.45397228e-01 -7.93970406e-01 -4.55185533e-01
3.36774379e-01 -9.57930148e-01 1.81011423e-01 -1.42269576e+00
4.05853800e-02 -2.76415259e-01 -2.90238976e-01 3.19500804e-01
1.55449167e-01 2.64799833e-01 1.52530104e-01 2.13068780e-02
-5.05361557e-01 6.86562538e-01 7.34176517e-01 -1.35310248e-01
-2.45194674e-01 -4.00759459e-01 -8.92867148e-01 1.08317876e+00
8.02739143e-01 -6.00899518e-01 -2.28162602e-01 -1.68912694e-01
4.41200852e-01 -6.19148791e-01 -4.19748157e-01 -6.37643576e-01
8.93593356e-02 1.25136701e-02 2.29479805e-01 -3.93659651e-01
2.39463132e-02 -7.55141020e-01 -7.70695031e-01 2.45820731e-01
-4.85723227e-01 3.66060853e-01 1.65538028e-01 2.44304836e-01
-4.89989936e-01 -5.99002600e-01 6.83728039e-01 8.73603746e-02
-7.07761943e-01 -9.73947048e-02 -1.06728351e+00 5.92133738e-02
8.11461687e-01 -7.12873861e-02 -3.23427081e-01 -6.38991967e-02
-5.41985571e-01 1.34718850e-01 4.79527652e-01 9.34504032e-01
3.78721684e-01 -1.30736983e+00 -6.25176907e-01 3.27594936e-01
6.47046208e-01 -4.74720985e-01 1.64071750e-02 2.09785491e-01
-7.47394383e-01 4.22357976e-01 -1.41459793e-01 -3.27586114e-01
-1.02032626e+00 6.57692492e-01 -3.26037824e-01 -1.64996445e-01
-1.93911612e-01 4.42962706e-01 -1.59552619e-01 -1.21203887e+00
-9.52821299e-02 -2.94011265e-01 -1.08194113e+00 9.30182636e-01
6.38675809e-01 3.30180854e-01 2.24451646e-01 -8.88235271e-01
-3.55073899e-01 6.52864218e-01 -1.25952154e-01 -3.73762906e-01
1.44779027e+00 8.32372438e-03 -6.35283411e-01 9.58817422e-01
1.22813272e+00 1.42720476e-01 -2.92798370e-01 -1.30111605e-01
4.05718327e-01 -3.38262200e-01 -3.34689952e-02 -2.13623077e-01
-7.18297899e-01 9.77464676e-01 4.85311449e-01 7.96073854e-01
5.63062787e-01 -1.02952577e-01 3.90097886e-01 6.25898659e-01
1.00792728e-01 -1.64918506e+00 -1.32349715e-01 9.30823624e-01
4.34004366e-01 -1.12282681e+00 -9.92990062e-02 2.06517354e-02
-8.80167663e-01 1.23815894e+00 2.52320379e-01 -3.62206280e-01
1.12760353e+00 3.60923633e-02 5.65916419e-01 -2.64386952e-01
-6.02787733e-01 -8.78931209e-02 1.17860161e-01 7.82646537e-01
7.64342725e-01 2.86087722e-01 -7.88529456e-01 8.55993867e-01
-6.99185193e-01 -3.89173269e-01 7.35322773e-01 7.27262735e-01
-7.19102979e-01 -1.25193763e+00 -4.24818844e-01 7.04921603e-01
-3.43058348e-01 -3.85608643e-01 -3.45703423e-01 6.39577031e-01
1.16250217e-02 1.36349273e+00 5.66146553e-01 -3.61500800e-01
2.21216500e-01 1.10670559e-01 -1.64124727e-01 -9.36325371e-01
-6.27414644e-01 -3.05834442e-01 -8.95474628e-02 6.12666570e-02
-3.40976834e-01 -7.27254093e-01 -1.19636476e+00 -4.54049677e-01
-1.80559725e-01 6.22234941e-01 1.02008486e+00 9.12600279e-01
2.29936421e-01 1.09577544e-01 7.72157967e-01 -8.85725200e-01
-4.00470197e-01 -1.15381479e+00 -1.16366363e+00 8.31527293e-01
-2.53578335e-01 -5.14422297e-01 -5.79816818e-01 3.36850807e-02] | [10.397171974182129, 8.51895523071289] |
214f933d-daea-4372-a86e-cd8c0f70485b | deep-graph-based-spatial-consistency-for | 2303.09950 | null | https://arxiv.org/abs/2303.09950v1 | https://arxiv.org/pdf/2303.09950v1.pdf | Deep Graph-based Spatial Consistency for Robust Non-rigid Point Cloud Registration | We study the problem of outlier correspondence pruning for non-rigid point cloud registration. In rigid registration, spatial consistency has been a commonly used criterion to discriminate outliers from inliers. It measures the compatibility of two correspondences by the discrepancy between the respective distances in two point clouds. However, spatial consistency no longer holds in non-rigid cases and outlier rejection for non-rigid registration has not been well studied. In this work, we propose Graph-based Spatial Consistency Network (GraphSCNet) to filter outliers for non-rigid registration. Our method is based on the fact that non-rigid deformations are usually locally rigid, or local shape preserving. We first design a local spatial consistency measure over the deformation graph of the point cloud, which evaluates the spatial compatibility only between the correspondences in the vicinity of a graph node. An attention-based non-rigid correspondence embedding module is then devised to learn a robust representation of non-rigid correspondences from local spatial consistency. Despite its simplicity, GraphSCNet effectively improves the quality of the putative correspondences and attains state-of-the-art performance on three challenging benchmarks. Our code and models are available at https://github.com/qinzheng93/GraphSCNet. | ['Kai Xu', 'Yuxing Peng', 'Changjian Wang', 'Hao Yu', 'Zheng Qin'] | 2023-03-17 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Qin_Deep_Graph-Based_Spatial_Consistency_for_Robust_Non-Rigid_Point_Cloud_Registration_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Qin_Deep_Graph-Based_Spatial_Consistency_for_Robust_Non-Rigid_Point_Cloud_Registration_CVPR_2023_paper.pdf | cvpr-2023-1 | ['point-cloud-registration'] | ['computer-vision'] | [-2.59819597e-01 -1.35804325e-01 -5.80825657e-02 -2.78364122e-01
-6.36797428e-01 -4.69487071e-01 4.74928528e-01 2.71013379e-01
-1.36497438e-01 7.39331618e-02 1.90408304e-01 1.15486883e-01
-5.40054977e-01 -6.29140019e-01 -8.25652838e-01 -6.81053758e-01
-9.28536654e-02 7.58272827e-01 2.21647829e-01 -1.91533566e-01
2.26048559e-01 9.27624941e-01 -1.27357793e+00 -2.43068010e-01
9.20769334e-01 5.67796528e-01 -2.49982700e-01 7.36017898e-02
2.27501497e-01 -1.54289216e-01 -9.84546244e-02 -1.53171793e-01
5.50032377e-01 -2.55855143e-01 -7.58543968e-01 -1.01379074e-01
8.76607716e-01 6.16294108e-02 -5.97705901e-01 1.35674167e+00
5.20282805e-01 2.40633503e-01 5.75805426e-01 -1.23305380e+00
-6.68620884e-01 1.57560363e-01 -6.64472699e-01 2.33731985e-01
4.85180855e-01 2.68739760e-02 1.04270947e+00 -1.16703546e+00
7.29872465e-01 1.13420653e+00 9.28271353e-01 2.95789272e-01
-1.26884186e+00 -5.02755702e-01 1.24612721e-02 6.96505234e-02
-1.60344410e+00 -3.97767991e-01 9.73194420e-01 -5.95087230e-01
7.13898838e-01 6.78674400e-01 6.17867112e-01 6.41290009e-01
3.16284001e-01 -4.21699090e-03 6.87180758e-01 -4.83118854e-02
1.13748945e-01 -6.09115362e-01 -4.70293360e-03 5.24073958e-01
4.58956629e-01 3.38061035e-01 -4.86289233e-01 -5.32801628e-01
1.00284648e+00 3.55167568e-01 -4.62380052e-01 -8.02180052e-01
-1.36472309e+00 6.60442829e-01 9.44210052e-01 5.40155113e-01
-3.02237839e-01 4.05405425e-02 2.11461544e-01 4.18868452e-01
6.79626703e-01 2.57891923e-01 -5.79624772e-02 1.21033438e-01
-8.27057242e-01 1.40632227e-01 5.90279579e-01 9.48349893e-01
7.10646868e-01 -1.40515298e-01 7.89994225e-02 6.59195125e-01
3.74604613e-01 2.62380242e-01 3.35239589e-01 -6.94075048e-01
4.05957401e-01 8.62991869e-01 -1.45835012e-01 -1.51559758e+00
-3.60953510e-01 -2.39500478e-01 -1.13884270e+00 2.72458911e-01
2.66774297e-01 6.47598684e-01 -7.75666773e-01 1.41783774e+00
5.33057570e-01 7.23887622e-01 -4.06717002e-01 1.08454943e+00
8.41383398e-01 2.81545728e-01 -2.25481421e-01 -4.71815541e-02
9.15299356e-01 -5.09961605e-01 -4.39415157e-01 1.02551490e-01
4.44129139e-01 -8.36783350e-01 9.07405972e-01 -3.15979451e-01
-1.04027855e+00 -3.24379236e-01 -6.92580283e-01 -1.59602642e-01
-3.36776264e-02 -2.49929398e-01 -9.18252207e-03 4.43640277e-02
-1.02784383e+00 9.74648952e-01 -1.11316502e+00 -3.53988022e-01
2.38251135e-01 6.58081651e-01 -7.79762328e-01 -1.44161746e-01
-6.76777184e-01 7.08740234e-01 -2.37431973e-02 2.92529970e-01
-9.38424841e-02 -6.45520270e-01 -9.33209777e-01 -8.02966654e-02
2.39734754e-01 -7.95148551e-01 4.37797964e-01 -6.50497139e-01
-1.06313717e+00 1.18279850e+00 -3.80939096e-01 1.59287024e-02
7.54284263e-01 -1.12763636e-01 -3.61437947e-01 -2.38381997e-01
1.87701389e-01 1.16897672e-01 7.79031277e-01 -1.31640971e+00
1.78415179e-02 -5.95259309e-01 -5.26855946e-01 1.65223166e-01
1.22930542e-01 7.28441849e-02 -6.38943791e-01 -9.13056791e-01
1.17361140e+00 -1.15891874e+00 -2.83379436e-01 3.65861952e-02
-5.29495895e-01 -4.14705515e-01 8.02863717e-01 -6.34544015e-01
1.03875339e+00 -2.44502854e+00 2.81733006e-01 8.27476025e-01
4.59363401e-01 -1.01894520e-01 -2.23473325e-01 3.73731762e-01
-5.12166023e-01 1.00973979e-01 -4.46481913e-01 -2.13245958e-01
5.10021672e-03 2.12332383e-01 -1.96394399e-01 1.16913819e+00
1.12171568e-01 7.95154214e-01 -8.64527345e-01 -1.40760392e-01
3.68955731e-01 6.05917215e-01 -5.90304554e-01 1.07458748e-01
2.34205708e-01 9.68766749e-01 -3.52999717e-01 7.14040995e-01
9.07235980e-01 -2.88534854e-02 -1.41849414e-01 -5.01676142e-01
-1.11958154e-01 8.17946866e-02 -1.42178702e+00 1.79431903e+00
1.45831227e-01 1.69666797e-01 -1.77972734e-01 -7.37931490e-01
1.04615545e+00 9.19731632e-02 9.48868990e-01 -5.72049201e-01
6.89201504e-02 4.69112545e-01 -1.74190495e-02 -1.56506881e-01
3.62938076e-01 9.17629600e-02 8.34578872e-02 1.59181550e-01
-3.25335711e-01 -1.25676438e-01 -8.61603171e-02 -1.44655004e-01
1.13833618e+00 1.05431333e-01 3.17635328e-01 -5.01909673e-01
5.42971671e-01 -1.89422369e-01 7.44814456e-01 3.25251073e-01
-1.08796388e-01 1.07096040e+00 9.17540416e-02 -7.90546417e-01
-1.02801466e+00 -1.29648113e+00 -2.97267675e-01 5.61713159e-01
6.02576315e-01 -5.49614191e-01 -3.74081552e-01 -4.85391378e-01
4.25048590e-01 1.37923077e-01 -4.04123694e-01 -3.05126816e-01
-9.98279512e-01 -4.20274764e-01 1.04887597e-01 3.29539180e-01
-2.27537863e-02 -7.68956125e-01 9.06767324e-02 -5.99171631e-02
-7.64525384e-02 -8.66352618e-01 -9.90006864e-01 -3.10781240e-01
-1.35134113e+00 -1.22964716e+00 -3.99697930e-01 -8.37631047e-01
1.28214347e+00 3.02216142e-01 1.08578527e+00 6.11193776e-01
-2.26706609e-01 3.07485491e-01 -7.02307373e-02 2.02121928e-01
-1.72210321e-01 -1.40276268e-01 4.49808806e-01 6.00241916e-03
3.60083520e-01 -7.88930833e-01 -6.09286368e-01 6.69718564e-01
-8.80988598e-01 -4.13031876e-01 2.07491875e-01 7.04643548e-01
1.26470065e+00 -1.38684526e-01 -9.85874683e-02 -6.66377783e-01
3.52804542e-01 -4.10154909e-01 -5.64098299e-01 3.78223509e-02
-3.36221039e-01 5.45187816e-02 2.89433062e-01 -2.71893412e-01
-1.99667320e-01 2.11114302e-01 -1.26633495e-01 -8.50128353e-01
-1.34763196e-01 3.44831407e-01 -1.58728421e-01 -5.31700850e-01
4.66636866e-01 -9.09560919e-02 -5.95804080e-02 -6.46454215e-01
2.87278622e-01 1.48984879e-01 7.11352289e-01 -6.58517778e-01
1.33950055e+00 7.41389811e-01 2.50580490e-01 -7.61071324e-01
-3.58571410e-01 -8.99772465e-01 -1.03727996e+00 7.00173751e-02
5.39960980e-01 -8.57927561e-01 -5.95139861e-01 6.79280460e-02
-1.04733729e+00 1.09542638e-01 -5.70600033e-01 5.09980142e-01
-6.26296699e-01 7.14992404e-01 -3.88877660e-01 -3.13503146e-01
-3.28824759e-01 -1.12116385e+00 1.19760060e+00 -1.38139009e-01
-3.92683744e-01 -9.37340200e-01 4.97401834e-01 -1.15249351e-01
1.55248225e-01 6.12487018e-01 7.03068256e-01 -6.04896247e-01
-7.48005509e-01 -5.37192822e-01 -4.10485752e-02 -5.71338013e-02
3.99596155e-01 6.04100823e-02 -5.63519299e-01 -7.53377438e-01
1.14243641e-01 4.47426975e-01 7.12553382e-01 3.02925169e-01
1.15218914e+00 -4.38099384e-01 -4.76366609e-01 1.03623164e+00
1.53456211e+00 -4.16280597e-01 7.58542955e-01 4.75130558e-01
1.23680532e+00 4.43357676e-01 4.62174863e-01 2.48081222e-01
2.37223282e-01 9.14958656e-01 6.97935462e-01 -1.00381702e-01
1.72436908e-02 -1.04517326e-01 2.32238248e-01 1.28277540e+00
-5.03274679e-01 1.52020127e-01 -1.02386618e+00 5.53363085e-01
-2.01015377e+00 -8.40781748e-01 -5.78266740e-01 2.63301063e+00
5.91042995e-01 -2.01103330e-01 4.06066980e-03 -1.37891963e-01
1.00365102e+00 8.45655128e-02 -3.43887329e-01 2.84857284e-02
-1.87615886e-01 1.03860937e-01 5.14985740e-01 6.48186922e-01
-1.13356113e+00 7.93576062e-01 5.01741934e+00 4.52661663e-01
-8.90267491e-01 1.11100435e-01 3.95085444e-05 3.33169401e-02
-4.20840144e-01 1.07846640e-01 -3.97565365e-01 5.47188759e-01
4.53992665e-01 -9.66454148e-02 2.48777598e-01 7.14608192e-01
1.81595683e-01 3.69640112e-01 -1.22542346e+00 1.10135758e+00
5.74634522e-02 -1.23423672e+00 1.43179372e-01 1.59276202e-01
6.97730482e-01 3.38413596e-01 -4.42304052e-02 -3.47405314e-01
1.12331949e-01 -1.02434027e+00 5.24419725e-01 7.01493204e-01
7.62794077e-01 -7.18696058e-01 7.14100063e-01 1.04772940e-01
-1.38522935e+00 3.57093960e-01 -6.11378610e-01 4.31050718e-01
1.20151214e-01 7.70450473e-01 -1.99584708e-01 9.58749890e-01
1.00111616e+00 1.01112509e+00 -5.00350714e-01 1.38711774e+00
8.10628932e-04 8.71548057e-02 -6.06957495e-01 8.62695515e-01
-1.28854930e-01 -6.91646755e-01 1.01522303e+00 8.51228297e-01
6.35521650e-01 1.32826284e-01 2.79489428e-01 8.42585206e-01
-1.27734259e-01 2.13443130e-01 -8.76442790e-01 5.31006515e-01
4.83407497e-01 9.38009977e-01 -5.92560351e-01 1.58925787e-01
-3.65015775e-01 1.09656239e+00 3.20884615e-01 1.97015658e-01
-5.71618736e-01 2.09154245e-02 1.00879431e+00 4.62807029e-01
-1.61883086e-02 -5.07775128e-01 -3.09814930e-01 -1.32258129e+00
4.69033122e-01 -7.65074968e-01 4.26092714e-01 -3.37867022e-01
-1.60213387e+00 6.43981516e-01 -2.24570736e-01 -1.65345263e+00
1.27695978e-01 -3.45521748e-01 -9.33186054e-01 9.05445814e-01
-1.30397367e+00 -1.02592874e+00 -6.84280157e-01 7.39716530e-01
1.10892773e-01 5.27805723e-02 6.56450868e-01 3.58470976e-01
-5.33035338e-01 5.78961730e-01 2.24202231e-01 1.99540928e-01
9.15388763e-01 -1.12948132e+00 7.82789826e-01 9.67766583e-01
2.53983885e-01 1.01694071e+00 6.19812727e-01 -1.03869462e+00
-1.47022867e+00 -1.20956552e+00 9.30231690e-01 -7.27529049e-01
5.33193171e-01 -5.07159866e-02 -1.38339543e+00 7.66552210e-01
-2.30576903e-01 6.55786216e-01 3.41863930e-01 -1.03352673e-01
-3.71287823e-01 -9.10479203e-03 -1.11125779e+00 6.40484452e-01
1.43409157e+00 -4.98595536e-01 -5.80811024e-01 4.04459655e-01
3.96351337e-01 -7.56772757e-01 -1.15615046e+00 7.43589997e-01
1.16932370e-01 -8.28917742e-01 1.31390345e+00 -5.02582192e-01
-1.56969935e-01 -7.77257025e-01 -3.92954126e-02 -1.13160312e+00
-6.22850776e-01 -7.93141186e-01 -4.14816961e-02 9.70119476e-01
-1.30952522e-01 -7.66053677e-01 7.87458658e-01 4.24597442e-01
-3.48421574e-01 -5.48742712e-01 -1.35588861e+00 -1.15013337e+00
4.64802841e-03 -4.93058823e-02 7.74392724e-01 1.28698146e+00
-3.86395693e-01 -2.82683760e-01 -5.08035831e-02 6.57030344e-01
7.72437871e-01 1.06826305e-01 7.70278096e-01 -1.66532576e+00
2.94135839e-01 -4.59698915e-01 -9.30318594e-01 -6.64585531e-01
3.03438544e-01 -1.29666007e+00 6.96017742e-02 -1.37283373e+00
-5.00085987e-02 -5.59297442e-01 -1.95888773e-01 3.35365951e-01
-1.00956708e-01 4.04913545e-01 -7.60631934e-02 9.33442116e-01
-3.40047687e-01 6.58480227e-01 1.01247859e+00 5.26884664e-03
-3.13531756e-01 6.84970841e-02 -2.18800858e-01 8.54404271e-01
7.71645904e-01 -7.47916937e-01 1.37196168e-01 -4.00569022e-01
2.25686640e-01 -4.60080564e-01 5.99176109e-01 -9.72040057e-01
4.51461554e-01 -7.36893639e-02 2.55420744e-01 -5.67822516e-01
5.57638891e-02 -1.25513852e+00 7.68848002e-01 3.30339491e-01
1.81189895e-01 6.98262393e-01 -5.77955805e-02 8.01457167e-01
-3.72579038e-01 4.98025455e-02 7.67851651e-01 2.47644838e-02
-4.96817410e-01 8.54281723e-01 5.32047153e-01 -8.06428418e-02
8.94362032e-01 -4.34244603e-01 -2.05380589e-01 7.17215762e-02
-8.75212431e-01 -7.45260715e-02 1.14246273e+00 4.44337517e-01
9.19427335e-01 -1.84244943e+00 -7.77130425e-01 5.00971913e-01
3.68746519e-01 3.15119714e-01 1.16227925e-01 1.27934957e+00
-6.75853968e-01 -3.31134647e-02 -2.49420870e-02 -9.65657949e-01
-1.33798063e+00 3.83645862e-01 4.56997156e-01 8.21721777e-02
-1.06279993e+00 5.84629655e-01 1.43963426e-01 -5.62494278e-01
1.21079899e-01 -5.13558090e-01 1.65188983e-01 -3.74697924e-01
5.26804067e-02 5.80090523e-01 3.83638889e-01 -1.23410928e+00
-6.64152384e-01 1.33846319e+00 2.59311706e-01 3.41963053e-01
1.36939263e+00 -1.37208886e-02 -7.45459020e-01 2.39416853e-01
1.25784409e+00 3.59899372e-01 -9.76030231e-01 -4.60464954e-01
2.53689706e-01 -8.73307705e-01 -2.12561324e-01 8.82486906e-03
-1.19610512e+00 3.20817620e-01 6.00929379e-01 -5.66187128e-03
8.97504807e-01 2.09046125e-01 7.14684963e-01 1.60862222e-01
2.55494207e-01 -7.73827493e-01 -6.92288876e-02 5.86648643e-01
1.42404521e+00 -1.28635633e+00 2.31127575e-01 -6.96196795e-01
-2.39984810e-01 1.06855094e+00 4.80544716e-01 -6.40246451e-01
7.90370107e-01 -1.12936668e-01 -8.73333141e-02 -6.61494970e-01
5.82572371e-02 -8.54642317e-02 9.41436291e-01 4.35726613e-01
4.15926635e-01 4.61902171e-02 -2.22633779e-01 1.32627368e-01
-3.00969064e-01 -5.25117874e-01 4.09376882e-02 7.22763181e-01
-1.50539890e-01 -1.12154007e+00 -5.31039298e-01 3.33014309e-01
-2.72064567e-01 6.56699836e-02 -5.18158972e-01 6.96197808e-01
-5.78076541e-02 5.47665894e-01 2.66790211e-01 -3.78777057e-01
7.65708685e-01 -2.68286109e-01 3.01903099e-01 -5.91093361e-01
-5.39355099e-01 2.74510115e-01 -4.73994493e-01 -1.01002038e+00
-4.72537726e-01 -8.00441384e-01 -1.22186100e+00 -5.96183538e-01
-4.27409410e-01 3.81362624e-02 4.18220788e-01 5.41624725e-01
6.12734020e-01 4.34164554e-02 6.82156563e-01 -9.34456825e-01
-3.72107208e-01 -5.68095863e-01 -5.15107870e-01 9.87729371e-01
4.34619963e-01 -6.27161980e-01 -6.15083814e-01 -1.32750243e-01] | [7.658512115478516, -3.0206754207611084] |
2f6b5078-9470-474e-b1cf-2df1b26b058d | conformal-prediction-intervals-for-markov | 2206.04860 | null | https://arxiv.org/abs/2206.04860v2 | https://arxiv.org/pdf/2206.04860v2.pdf | Conformal Prediction Intervals for Markov Decision Process Trajectories | Before delegating a task to an autonomous system, a human operator may want a guarantee about the behavior of the system. This paper extends previous work on conformal prediction for functional data and conformalized quantile regression to provide conformal prediction intervals over the future behavior of an autonomous system executing a fixed control policy on a Markov Decision Process (MDP). The prediction intervals are constructed by applying conformal corrections to prediction intervals computed by quantile regression. The resulting intervals guarantee that with probability $1-\delta$ the observed trajectory will lie inside the prediction interval, where the probability is computed with respect to the starting state distribution and the stochasticity of the MDP. The method is illustrated on MDPs for invasive species management and StarCraft2 battles. | ['Jesse Hostetler', 'Thomas G. Dietterich'] | 2022-06-10 | null | null | null | null | ['prediction-intervals'] | ['miscellaneous'] | [ 1.81594357e-01 4.76345688e-01 -3.00649703e-01 -3.33239108e-01
-6.36664569e-01 -4.41611052e-01 4.91802037e-01 2.77736127e-01
-3.19475323e-01 7.97990918e-01 -9.85751897e-02 -5.41778028e-01
-6.20465755e-01 -7.75767267e-01 -6.29784524e-01 -6.32656097e-01
-8.41115355e-01 7.41303384e-01 3.08876604e-01 -6.78745136e-02
2.00961843e-01 5.90491712e-01 -1.27283204e+00 -1.21606544e-01
6.06759012e-01 1.16874826e+00 -8.01923592e-03 1.08815145e+00
1.54074103e-01 4.95167255e-01 -7.77560353e-01 -1.16006687e-01
4.40773070e-01 -2.17681095e-01 -6.46353304e-01 -2.56601155e-01
-4.01638836e-01 -8.07748064e-02 -4.05115969e-02 9.45967495e-01
-1.22273512e-01 2.95339644e-01 1.24891555e+00 -1.71760488e+00
1.88537940e-01 5.20169437e-01 -2.63487130e-01 9.73047763e-02
1.35788694e-01 6.09399416e-02 6.13516033e-01 -8.00483599e-02
3.20754200e-01 1.01379216e+00 7.23479271e-01 2.99644977e-01
-1.62002647e+00 -3.60335678e-01 1.54245675e-01 -1.82958692e-01
-1.41004896e+00 2.48638615e-02 -7.87567534e-03 -7.62891412e-01
7.63145268e-01 -3.10643092e-02 6.06249928e-01 6.83356702e-01
1.55106866e+00 1.76478267e-01 9.11591589e-01 1.33493260e-01
7.84605622e-01 -1.60732239e-01 -1.53576508e-01 2.91237146e-01
-3.40766311e-02 6.92833126e-01 -1.22842006e-01 -6.11436844e-01
5.13778150e-01 3.41689050e-01 -6.04455881e-02 -5.18695831e-01
-7.91671932e-01 6.88667834e-01 -3.14436480e-02 -2.53424734e-01
-8.37775767e-01 3.43788952e-01 2.44637102e-01 3.52661043e-01
3.63104165e-01 9.25204530e-02 -6.22900665e-01 -3.66410941e-01
-8.54623139e-01 5.49067974e-01 1.17896307e+00 1.10198605e+00
5.59212744e-01 6.45129532e-02 -4.75348860e-01 -1.90829828e-01
1.08091742e-01 7.51145780e-01 -1.75509661e-01 -1.10167122e+00
1.18682705e-01 2.82494783e-01 7.41578043e-01 -4.51270610e-01
-3.30922812e-01 -3.74450646e-02 -2.73113012e-01 8.02967668e-01
5.37279844e-01 -4.48622316e-01 -5.81257939e-01 1.48234951e+00
2.66449064e-01 -5.55863865e-02 1.66794524e-01 5.20369411e-01
-7.38926411e-01 8.73629987e-01 3.38869095e-01 -5.24186254e-01
6.94140673e-01 1.34494051e-01 -4.39405710e-01 2.10503474e-01
2.22943917e-01 -3.64451557e-01 9.33616877e-01 5.55823028e-01
-7.66044080e-01 -3.07099104e-01 -9.20851529e-01 9.51257467e-01
2.19614893e-01 -4.25812274e-01 -5.94116300e-02 4.43676323e-01
-6.55646443e-01 9.15405452e-01 -1.36325419e+00 -5.83457761e-02
1.80636495e-01 2.83403099e-01 -4.78728078e-02 3.63020867e-01
-8.34851623e-01 1.00119233e+00 3.59523028e-01 -3.92816901e-01
-1.67993712e+00 -7.20952988e-01 -5.29898584e-01 6.52190372e-02
4.75003839e-01 -1.05668008e-01 1.46088958e+00 -6.09359145e-01
-1.45468390e+00 1.02282584e-01 1.00613207e-01 -9.21193361e-01
5.73710740e-01 -2.09827170e-01 -2.79479414e-01 -2.70640075e-01
3.54938686e-01 9.33933556e-02 1.12943578e+00 -8.69894981e-01
-9.61998880e-01 -3.91592860e-01 -5.98647371e-02 1.37914345e-01
3.63755316e-01 -3.83615941e-01 2.94673741e-01 -1.98470816e-01
-2.08994642e-01 -1.19755852e+00 -6.10952675e-01 -3.59249383e-01
-4.16245908e-01 -3.38593245e-01 6.10918283e-01 -5.36762536e-01
1.15777183e+00 -2.08397794e+00 3.78442794e-01 3.87519211e-01
-3.15256417e-01 -6.90315723e-01 4.25125033e-01 7.02101052e-01
-7.43011758e-02 -1.05663903e-01 -5.53092301e-01 1.33785665e-01
2.15367720e-01 2.67503142e-01 -7.67507076e-01 9.41618145e-01
-3.88476178e-02 1.69227317e-01 -4.25171047e-01 -1.74041346e-01
-2.42975838e-02 2.16379110e-03 -2.41935939e-01 4.24126536e-01
-6.49691880e-01 4.59303379e-01 -5.33192456e-01 2.48570085e-01
3.43882263e-01 3.75015348e-01 1.10467806e-01 6.50482178e-01
-4.72807825e-01 -2.14428678e-01 -9.76719201e-01 1.06299365e+00
-3.55834991e-01 1.84775352e-01 -1.73726976e-01 -7.98325241e-01
9.62516427e-01 3.20398271e-01 9.15271759e-01 -8.78501832e-02
2.46924326e-01 -2.88009912e-01 -3.63179110e-02 -4.55545075e-02
3.10912102e-01 -6.69568896e-01 -5.97338557e-01 5.28699934e-01
-5.32682985e-02 -4.47732240e-01 -3.45247149e-01 -8.39552283e-02
1.23407483e+00 4.40752923e-01 3.54872525e-01 -8.31736088e-01
8.52512568e-02 4.05581445e-01 6.26625299e-01 5.86570978e-01
-3.78003478e-01 -6.47539198e-02 1.24675608e+00 -3.58902693e-01
-1.24988246e+00 -1.45730424e+00 -4.22991186e-01 1.00133443e+00
2.88700730e-01 -1.62005529e-01 -9.50079978e-01 -5.62242448e-01
1.69616714e-01 1.37547410e+00 -9.42146242e-01 -3.41415614e-01
-4.71001528e-02 -3.30195010e-01 8.52513835e-02 3.44925582e-01
9.29875374e-02 -7.52685368e-01 -1.05938900e+00 3.21982473e-01
2.06166163e-01 -7.11907625e-01 -3.67452741e-01 4.34251726e-01
-7.74556935e-01 -1.17101395e+00 -2.36617729e-01 -5.24830446e-02
2.94787496e-01 -5.32818258e-01 7.43918002e-01 -8.39080095e-01
-1.27910465e-01 6.20952368e-01 5.40305972e-02 -7.37718642e-01
-5.91839671e-01 -1.26903579e-01 4.70345974e-01 -9.79236364e-02
9.58314836e-02 -3.17572683e-01 -5.65308630e-01 5.44984400e-01
-5.19261897e-01 -3.79409403e-01 1.31233275e-01 6.42767608e-01
1.06046677e+00 6.74337924e-01 3.90739709e-01 -4.15311903e-01
7.38555372e-01 -5.36379814e-01 -1.29230583e+00 2.46874988e-01
-7.05228567e-01 4.04639900e-01 6.15154684e-01 -5.03819525e-01
-9.12562847e-01 3.24897468e-01 5.05158603e-01 -4.80899066e-01
-3.63119662e-01 2.53503323e-01 -1.92380454e-02 5.45430779e-01
7.29996622e-01 7.47526661e-02 1.16651393e-01 -5.51894773e-03
7.67177492e-02 3.14277947e-01 4.94554669e-01 -7.25548327e-01
6.50524676e-01 3.84005815e-01 3.94044459e-01 -5.69625854e-01
-4.90361720e-01 -1.14472536e-02 -6.14330947e-01 -5.70081353e-01
1.07316804e+00 -4.75386232e-01 -1.37469196e+00 -1.93420187e-01
-7.33636796e-01 -7.65591502e-01 -7.67857134e-01 6.00568831e-01
-1.24417889e+00 -1.88198164e-01 2.14121286e-02 -1.48928726e+00
7.68994987e-02 -8.38225186e-01 9.51565027e-01 7.25393891e-02
-2.34126970e-01 -6.92457557e-01 3.90084386e-01 -5.00081897e-01
1.68881953e-01 6.48819268e-01 9.84163225e-01 -5.89506686e-01
-1.80186197e-01 -7.10773349e-01 5.31749725e-01 2.46155709e-01
-2.16665849e-01 9.23389047e-02 -5.20042062e-01 -3.60592902e-01
2.81442225e-01 2.42842957e-01 1.59491092e-01 7.44597852e-01
1.00084877e+00 -5.30924976e-01 -5.17654240e-01 1.07317857e-01
1.23687255e+00 4.87344831e-01 2.95968831e-01 1.98920310e-01
-1.73174590e-01 7.96917260e-01 1.18922460e+00 9.63733137e-01
9.06821862e-02 4.38883185e-01 7.39576101e-01 6.71070457e-01
9.03001606e-01 -4.19275612e-01 6.90837502e-01 -3.99437398e-01
1.45978499e-02 5.74712120e-02 -1.13269544e+00 4.91210520e-01
-1.76040137e+00 -9.26595807e-01 -4.47492786e-02 2.86958218e+00
3.79078627e-01 4.26364273e-01 4.76406276e-01 -1.87238574e-01
6.76091969e-01 -4.77095366e-01 -9.83370900e-01 -6.90291107e-01
5.15657485e-01 2.76783612e-02 9.53744471e-01 8.85049939e-01
-9.71665442e-01 5.22521496e-01 7.01919699e+00 5.93252897e-01
-5.61987102e-01 -5.82527556e-02 6.28986597e-01 3.76134925e-02
7.91300312e-02 3.02842885e-01 -9.43817735e-01 4.61634636e-01
1.50461149e+00 -7.48587370e-01 3.54192019e-01 1.21595716e+00
2.96906650e-01 -4.91702914e-01 -1.17181361e+00 1.47253186e-01
-7.13240266e-01 -6.36507213e-01 -5.41135550e-01 4.53159213e-01
4.73987758e-01 6.15707487e-02 2.73229212e-01 4.63837862e-01
9.48304236e-01 -9.54417825e-01 7.38774776e-01 1.07430911e+00
7.34154761e-01 -1.41780245e+00 7.01281846e-01 9.44298267e-01
-1.00761306e+00 -2.94962406e-01 -3.49221200e-01 -1.28927007e-01
1.97947249e-01 4.16192293e-01 -1.18563807e+00 1.34928092e-01
6.21469259e-01 8.17130432e-02 1.62795454e-01 6.92810357e-01
7.41428658e-02 7.81343043e-01 -4.18568045e-01 -6.46237656e-02
1.04674593e-01 -4.14954573e-01 7.27467179e-01 6.25711918e-01
4.56155866e-01 2.22149298e-01 4.75198179e-01 8.06831419e-01
8.50249946e-01 -1.00110121e-01 -9.78186667e-01 1.45839900e-01
2.32547268e-01 6.67984843e-01 -6.36804044e-01 -2.04719659e-02
1.53407808e-02 6.08294368e-01 -6.38067797e-02 4.40110236e-01
-9.79432762e-01 -4.49142903e-01 8.37610185e-01 2.06560969e-01
1.62386701e-01 -5.66985548e-01 -4.40139681e-01 -5.42285621e-01
-2.08681703e-01 -2.37886772e-01 4.03567106e-01 -3.38381886e-01
-1.05004632e+00 7.07439125e-01 4.02608395e-01 -1.39233112e+00
-7.24921703e-01 -4.15917307e-01 -5.65805554e-01 1.03512108e+00
-5.55855513e-01 -2.82271773e-01 2.15739816e-01 5.11699915e-01
3.86263549e-01 -2.85997033e-01 8.94766152e-01 -6.82924330e-01
-3.31060439e-01 -1.70030370e-01 2.20436409e-01 -4.94385779e-01
1.75362349e-01 -1.36619043e+00 1.81490853e-01 6.15111053e-01
-5.14922261e-01 7.16807693e-03 1.31411910e+00 -1.20557427e+00
-1.11647165e+00 -1.08055985e+00 2.55387425e-01 -7.26672590e-01
9.55988109e-01 -1.08864717e-01 -7.11955428e-01 8.77006769e-01
-8.29606652e-02 -2.40261763e-01 5.98132968e-01 8.00521858e-03
2.96735972e-01 -4.89068814e-02 -1.27276886e+00 4.28198367e-01
3.72762352e-01 -1.89849466e-01 -4.38676924e-01 2.32315272e-01
6.50234342e-01 -1.02639943e-01 -1.29258740e+00 5.01457155e-01
5.43227255e-01 -5.60552299e-01 3.52486879e-01 -8.89113128e-01
2.21627094e-02 -3.22383016e-01 -6.40521049e-01 -1.33985937e+00
-1.53033540e-01 -7.66625941e-01 9.80843306e-02 6.92310512e-01
4.25872952e-01 -4.55826461e-01 6.76874995e-01 8.63312066e-01
-7.17393905e-02 -5.36426306e-01 -1.52534676e+00 -1.02974355e+00
3.25927049e-01 -6.84296429e-01 5.68678856e-01 1.30831763e-01
1.42337725e-01 -4.80580628e-02 -2.51395524e-01 6.53507829e-01
9.31034565e-01 1.94269583e-01 6.77659214e-01 -1.17654741e+00
-3.30941021e-01 -1.38993770e-01 -4.68172312e-01 -5.77331722e-01
3.23803931e-01 -3.93808335e-01 6.35158300e-01 -9.55888987e-01
-7.47795478e-02 -3.83301228e-01 -1.21255040e-01 1.71159938e-01
3.12034547e-01 -6.50743902e-01 -1.56450257e-01 1.34678036e-01
-4.72023159e-01 8.95318747e-01 7.84582257e-01 2.55505800e-01
-2.94277281e-01 9.61842179e-01 -4.80834432e-02 7.88034439e-01
8.64418566e-01 -8.48536551e-01 -4.45174396e-01 4.48238641e-01
-9.39951688e-02 1.03430295e+00 1.11804299e-01 -1.02998459e+00
2.73091234e-02 -8.04416418e-01 1.40074462e-01 -7.52048612e-01
2.02289820e-01 -1.05722785e+00 4.87640083e-01 9.41630304e-01
-5.65174520e-01 3.19387645e-01 2.37613186e-01 1.20150375e+00
7.36977011e-02 -4.57926430e-02 6.93664074e-01 3.52523357e-01
-3.70092392e-01 5.15041828e-01 -9.65512276e-01 -1.73927516e-01
1.60651922e+00 1.85314149e-01 3.03046614e-01 -5.95849574e-01
-1.23254871e+00 5.80328524e-01 4.25242990e-01 2.28564367e-02
4.60040450e-01 -8.99747729e-01 -5.79672039e-01 6.19016699e-02
-3.94648723e-02 -2.13307545e-01 1.56361446e-01 5.85179090e-01
-2.56231725e-01 1.90007359e-01 -3.27690214e-01 -7.41050065e-01
-7.31732786e-01 6.46241009e-01 7.12831140e-01 -3.22758332e-02
-5.83505929e-01 3.51564497e-01 1.79222107e-01 -3.51732939e-01
1.38236642e-01 -2.87274212e-01 4.99803610e-02 -3.36188585e-01
3.10907006e-01 5.22701919e-01 -2.44644418e-01 -3.55155498e-01
-2.61076242e-01 1.89767867e-01 2.67317891e-01 -6.42712772e-01
1.09387684e+00 2.87387259e-02 1.18865378e-01 9.20194507e-01
4.65969831e-01 -4.17474538e-01 -2.21855831e+00 1.86305225e-01
3.16264898e-01 -2.30181754e-01 -1.60978392e-01 -7.04647481e-01
-3.14073861e-01 6.87692106e-01 7.34158635e-01 6.06146574e-01
8.52806985e-01 3.04555260e-02 1.72512811e-02 1.02779821e-01
7.23433912e-01 -1.06783235e+00 -3.58244509e-01 5.23306310e-01
7.80471504e-01 -6.97900176e-01 -8.12512115e-02 1.15655065e-01
-1.13289762e+00 9.37960029e-01 4.08832878e-01 -5.17908990e-01
1.13466549e+00 6.17843151e-01 -3.93374801e-01 1.47109732e-01
-1.07444847e+00 1.84821770e-01 5.38102277e-02 8.21946502e-01
-5.38400598e-02 7.37934172e-01 -8.83887038e-02 6.90962315e-01
-3.84013325e-01 -8.19449201e-02 4.46136564e-01 8.65136683e-01
-7.90104628e-01 -6.21277928e-01 -4.30076152e-01 5.06356239e-01
-3.54118437e-01 2.67926067e-01 -9.36429352e-02 8.15971911e-01
-2.19796941e-01 7.29596913e-01 3.96817505e-01 -2.94656843e-01
7.13021517e-01 2.04163417e-01 9.77973267e-02 -5.33072174e-01
-7.10370690e-02 -5.06772026e-02 -1.16355374e-01 -7.51880705e-01
4.56016332e-01 -1.19399738e+00 -1.43845880e+00 -4.22830939e-01
-6.91179335e-02 4.40876126e-01 7.95818329e-01 8.86610210e-01
2.01318711e-01 4.34282839e-01 7.88854182e-01 -6.14975154e-01
-1.19220102e+00 -9.11506116e-01 -8.68552685e-01 -3.97515357e-01
2.06697553e-01 -6.01057708e-01 -2.29224294e-01 -4.32452261e-02] | [4.589000701904297, 2.3097596168518066] |
defc7366-a067-4ef9-94a1-4f9e1ecd6dfc | large-scale-midi-based-composer | 2010.14805 | null | https://arxiv.org/abs/2010.14805v1 | https://arxiv.org/pdf/2010.14805v1.pdf | Large-Scale MIDI-based Composer Classification | Music classification is a task to classify a music piece into labels such as genres or composers. We propose large-scale MIDI based composer classification systems using GiantMIDI-Piano, a transcription-based dataset. We propose to use piano rolls, onset rolls, and velocity rolls as input representations and use deep neural networks as classifiers. To our knowledge, we are the first to investigate the composer classification problem with up to 100 composers. By using convolutional recurrent neural networks as models, our MIDI based composer classification system achieves a 10-composer and a 100-composer classification accuracies of 0.648 and 0.385 (evaluated on 30-second clips) and 0.739 and 0.489 (evaluated on music pieces), respectively. Our MIDI based composer system outperforms several audio-based baseline classification systems, indicating the effectiveness of using compact MIDI representations for composer classification. | ['Yuxuan Wang', 'Keunwoo Choi', 'Qiuqiang Kong'] | 2020-10-28 | null | null | null | null | ['music-classification'] | ['music'] | [ 2.45580629e-01 -4.26531106e-01 -4.21935506e-02 -7.24927858e-02
-1.06578243e+00 -1.09806204e+00 4.73586410e-01 -1.20661587e-01
-5.58785126e-02 2.95919091e-01 5.04390180e-01 2.89479587e-02
-1.86658397e-01 -6.13325834e-01 -4.70276505e-01 -2.40012154e-01
-1.10485300e-01 3.04948002e-01 -2.54002869e-01 -3.98682654e-01
3.90350342e-01 2.80493081e-01 -1.66927946e+00 8.94546449e-01
2.53325820e-01 1.56378698e+00 -3.18632126e-01 1.18466341e+00
2.90209323e-01 1.18368661e+00 -1.13112450e+00 -6.12396114e-02
1.90166041e-01 -6.67172670e-01 -7.88620710e-01 -4.52820927e-01
7.00685680e-01 7.99761247e-03 -4.01246101e-01 4.41200495e-01
8.23569477e-01 3.88163358e-01 8.55299413e-01 -8.54251325e-01
-6.68100297e-01 1.39340615e+00 -2.48575985e-01 1.92681551e-01
6.14059031e-01 -2.95570314e-01 1.54975116e+00 -7.70198882e-01
3.60397339e-01 7.58821905e-01 1.10178792e+00 3.07789832e-01
-1.08091450e+00 -1.18412304e+00 -1.67525783e-01 2.67144889e-01
-1.52372837e+00 -7.51970589e-01 1.10187495e+00 -5.92848063e-01
1.14674795e+00 5.48601329e-01 8.01067054e-01 1.04789793e+00
7.96888620e-02 8.46798718e-01 7.79330194e-01 -2.83148468e-01
7.79832900e-02 -6.54787362e-01 1.62974879e-01 2.26735294e-01
-6.80197597e-01 -1.05554745e-01 -1.09852052e+00 6.21613173e-04
6.46133423e-01 -1.88598365e-01 -1.84196457e-01 7.84138978e-01
-1.48756754e+00 7.29843318e-01 4.69082028e-01 5.09129763e-01
-3.26266378e-01 5.71324885e-01 8.16407084e-01 7.42715240e-01
5.22072017e-01 1.07584667e+00 -3.45413506e-01 -7.62883067e-01
-1.39164054e+00 3.58709842e-01 8.81236494e-01 6.98213995e-01
1.31494757e-02 5.48955023e-01 -2.62189299e-01 1.40438628e+00
-2.75982648e-01 2.15203971e-01 9.82425690e-01 -1.05406511e+00
2.60703564e-01 2.56654710e-01 -3.23222756e-01 -6.80217981e-01
-5.38833737e-01 -9.25741851e-01 -8.86128306e-01 3.48035060e-02
2.43273690e-01 1.60598174e-01 -7.35912561e-01 1.51506817e+00
-4.96400803e-01 2.75279045e-01 -8.29782486e-02 9.64069068e-01
1.33173692e+00 7.14996338e-01 -4.80190426e-01 -3.86103429e-02
1.25639176e+00 -1.04379046e+00 -4.14074570e-01 3.10439646e-01
1.28390417e-01 -1.14196110e+00 1.25092423e+00 9.34454143e-01
-1.06685090e+00 -1.08714128e+00 -1.29843760e+00 -7.71009400e-02
2.34203972e-03 7.01403081e-01 4.93050635e-01 4.65586632e-01
-5.21897972e-01 1.18606424e+00 -5.78801632e-01 8.11459795e-02
2.90619105e-01 2.74900198e-01 -1.52291819e-01 6.85060859e-01
-1.04250097e+00 8.76973793e-02 3.51857245e-02 -1.41807288e-01
-1.14585507e+00 -9.31896627e-01 -4.64317739e-01 -5.00579849e-02
-2.72644311e-01 -2.71298975e-01 1.66355979e+00 -1.09660196e+00
-1.87120271e+00 8.88557136e-01 2.02809587e-01 -5.69691420e-01
2.10674673e-01 -2.72753954e-01 -6.08636022e-01 1.46586727e-02
-7.58213475e-02 3.11204702e-01 6.86729908e-01 -6.41449928e-01
-7.18601763e-01 -1.53235674e-01 1.94548205e-01 2.42178470e-01
-1.27419829e-01 2.80456811e-01 6.69072941e-02 -1.22960222e+00
3.87685269e-01 -1.19379914e+00 3.58350158e-01 -7.04020500e-01
-6.04282677e-01 -3.55407953e-01 2.76739687e-01 -7.33754456e-01
1.58725429e+00 -2.29784918e+00 9.85943675e-02 -1.66613832e-01
1.02221489e-01 -6.16774289e-03 -3.57417136e-01 2.64788479e-01
-1.26416206e-01 -4.18066382e-02 1.36241198e-01 -2.26288393e-01
1.79943785e-01 -1.90177143e-01 -6.93331957e-01 2.34373674e-01
-2.20341533e-01 8.99923265e-01 -5.05954981e-01 1.61392763e-01
-2.85250396e-02 6.16799712e-01 -4.08144712e-01 2.64068365e-01
3.67061701e-03 4.38391745e-01 1.50387868e-01 8.28526795e-01
-5.47820069e-02 2.48508200e-01 -1.18361846e-01 -4.23431635e-01
-1.26044035e-01 1.20675254e+00 -9.10495520e-01 2.01510191e+00
-7.45315731e-01 1.02052200e+00 -2.65557587e-01 -7.37327993e-01
1.48297238e+00 7.01847255e-01 6.31355464e-01 -3.35001200e-01
1.92849413e-01 5.08772969e-01 4.07735467e-01 1.06795788e-01
7.16198742e-01 -3.38556319e-01 -5.41253686e-01 7.47504115e-01
3.11481982e-01 -2.41477847e-01 2.55236238e-01 -3.34390104e-01
1.19771183e+00 2.71889538e-01 8.01455230e-02 -4.59760539e-02
2.41592973e-01 -3.45001459e-01 5.38453758e-01 8.82423341e-01
9.71737429e-02 1.27163553e+00 2.25352883e-01 -8.20703328e-01
-7.66831219e-01 -1.02508759e+00 -6.87862486e-02 1.87718952e+00
-3.99328500e-01 -1.00747037e+00 -4.06430572e-01 -2.28765860e-01
-6.68061003e-02 4.56559926e-01 -6.49850905e-01 -1.02013566e-01
-9.03015852e-01 -4.79167789e-01 1.29372549e+00 8.11487615e-01
1.27428100e-01 -1.41911280e+00 -3.98031384e-01 6.13139808e-01
-3.97746235e-01 -4.66372073e-01 -7.15496123e-01 5.59895396e-01
-6.34708107e-01 -9.18065727e-01 -8.30964208e-01 -9.26622987e-01
-5.43569922e-01 -2.38087520e-01 1.46910346e+00 -4.93581176e-01
-2.04708561e-01 -8.18680525e-02 -6.44066215e-01 -3.64442408e-01
-5.51696777e-01 7.98510790e-01 3.21423620e-01 -9.62937400e-02
1.10871918e-01 -1.29778004e+00 -5.62347651e-01 3.19411874e-01
-3.70418072e-01 1.64146349e-01 1.28379300e-01 4.44727808e-01
7.23666847e-01 -3.81851047e-01 9.43551421e-01 -5.48081398e-01
8.04607332e-01 5.00475615e-03 1.56994518e-02 -5.27573563e-02
-3.23414862e-01 -3.70810211e-01 9.30946648e-01 -9.43107128e-01
-3.75695378e-01 -3.38676036e-05 -4.92996454e-01 -4.72219110e-01
-1.89600959e-01 5.02281845e-01 1.64527193e-01 2.80384451e-01
1.00692928e+00 1.54154971e-01 -5.99420249e-01 -9.04404640e-01
2.78185636e-01 1.15254736e+00 8.99075091e-01 -5.86306691e-01
3.22456598e-01 -2.94133201e-02 -3.69053602e-01 -7.31723726e-01
-1.18860126e+00 -4.51021999e-01 -5.23233831e-01 -3.90428334e-01
6.18370235e-01 -1.02342677e+00 -1.08759558e+00 4.42330122e-01
-9.25118744e-01 -2.56244987e-01 -6.23173475e-01 5.79916477e-01
-9.70656037e-01 -2.44402871e-01 -1.34391999e+00 -7.28974581e-01
-1.01354635e+00 -7.49938190e-01 1.11807871e+00 -7.05872178e-02
-9.76283967e-01 -1.98274150e-01 3.71130884e-01 4.05193478e-01
4.60383564e-01 4.24213886e-01 5.86176753e-01 -7.80705333e-01
-1.35262087e-01 -1.04418442e-01 1.92786589e-01 2.45642960e-01
2.84858584e-01 -1.51627600e-01 -1.41178322e+00 -1.27820447e-01
-2.21051157e-01 -6.61260247e-01 1.13333488e+00 3.62158626e-01
1.17195773e+00 -3.65775764e-01 3.37134451e-01 1.00888526e+00
8.10404778e-01 5.50263882e-01 4.29726064e-01 2.70416051e-01
8.61900151e-01 -8.29056725e-02 2.42704749e-01 5.43445408e-01
-7.83072561e-02 8.81890953e-01 -6.87528849e-02 3.18213671e-01
-6.17175221e-01 -4.62181628e-01 5.18179119e-01 1.37356007e+00
-6.53934062e-01 -5.13653643e-02 -7.64767230e-01 3.70764434e-01
-1.72226572e+00 -1.09799278e+00 4.10043932e-02 1.98330224e+00
9.54374909e-01 -3.20143923e-02 4.55039442e-01 8.66597235e-01
4.89074826e-01 2.94062376e-01 -5.13763487e-01 -4.82311219e-01
-1.82469502e-01 6.53810441e-01 -1.24553427e-01 -5.76313958e-02
-1.39737093e+00 8.50606978e-01 6.38901329e+00 9.44825292e-01
-1.37015021e+00 5.79839721e-02 3.98770630e-01 -6.53945982e-01
1.46199241e-01 -3.00659686e-01 -6.24865294e-01 3.80320609e-01
1.19199872e+00 2.06602991e-01 9.34635222e-01 6.86795831e-01
-2.66724914e-01 6.62840307e-01 -1.28109062e+00 1.55614316e+00
3.51113766e-01 -1.46545720e+00 2.11640354e-02 -2.10885882e-01
8.27585876e-01 3.58689219e-01 5.62297218e-02 5.85849166e-01
5.35240844e-02 -1.26324773e+00 1.34858286e+00 5.03398955e-01
1.09539759e+00 -9.66560662e-01 4.60235476e-01 -1.12019479e-01
-1.67962527e+00 -5.05979806e-02 -6.22450002e-02 -5.31164885e-01
6.30805194e-02 2.35965297e-01 -7.08963513e-01 3.33072752e-01
7.75385499e-01 1.22490454e+00 -4.70031768e-01 7.88255811e-01
-6.25605807e-02 1.24969971e+00 -3.31431717e-01 6.94037080e-02
-5.58093414e-02 -1.65885277e-02 4.22299534e-01 1.45385647e+00
4.77747947e-01 -2.47172341e-01 1.97886541e-01 4.91436005e-01
-4.74890411e-01 2.01230094e-01 -2.90952194e-02 -3.79859954e-01
2.45099217e-01 1.07302892e+00 -6.33850336e-01 -3.18674713e-01
2.28851959e-01 1.00267351e+00 1.56659409e-01 2.06569090e-01
-7.63621747e-01 -9.46431041e-01 8.11325252e-01 1.17136212e-02
1.51976511e-01 -1.19207829e-01 -4.53205407e-01 -1.19302428e+00
-2.25072920e-01 -1.16318476e+00 5.65447152e-01 -9.06241119e-01
-1.45401502e+00 1.10251963e+00 -8.18789721e-01 -1.57771337e+00
-4.87068832e-01 -5.14868200e-01 -6.27350032e-01 5.88525057e-01
-8.30824077e-01 -1.33017206e+00 -9.75210071e-02 3.61432880e-01
6.45554543e-01 -7.84862518e-01 1.19680691e+00 3.41543674e-01
-5.52106500e-02 8.48230839e-01 1.11313656e-01 6.86292231e-01
7.03003943e-01 -1.23777199e+00 4.75714624e-01 2.02724840e-02
1.19807398e+00 4.19160962e-01 3.38056445e-01 -7.05160201e-02
-9.44184959e-01 -9.60077584e-01 8.26624513e-01 -5.85937738e-01
6.32532537e-01 -4.96178329e-01 -6.05480731e-01 5.19814968e-01
1.02163002e-01 -3.84855866e-01 1.20805597e+00 8.79513979e-01
-9.03674662e-01 -5.18938184e-01 -3.41716111e-01 2.72274435e-01
1.21683383e+00 -1.17140627e+00 -7.34603703e-01 2.35268489e-01
5.85735142e-01 -3.72502685e-01 -1.40394628e+00 5.49104750e-01
1.36347997e+00 -7.60047436e-01 9.26684499e-01 -4.47753608e-01
6.43701851e-01 -4.69806582e-01 -3.50862890e-01 -1.24480557e+00
-4.95941550e-01 -7.42679834e-01 -1.89187959e-01 1.32247567e+00
5.70863307e-01 -1.97583251e-02 6.09498024e-01 -4.77915108e-01
-4.54294950e-01 -4.83454496e-01 -9.60054874e-01 -8.61032546e-01
8.23220983e-02 -9.08713043e-01 6.54145002e-01 1.01655769e+00
1.33628324e-01 7.13562787e-01 -4.33014125e-01 -5.25252700e-01
-3.74684902e-03 9.72596467e-01 7.11278737e-01 -1.36680496e+00
-7.83268929e-01 -8.84413362e-01 -5.95425308e-01 -6.33152723e-01
1.29089907e-01 -1.37240958e+00 -5.85106015e-02 -1.29377246e+00
4.38795276e-02 -3.69714200e-01 -1.04065764e+00 6.39646649e-01
5.16750634e-01 1.29357147e+00 4.05376971e-01 5.75869322e-01
-6.64231598e-01 3.54664862e-01 7.17118084e-01 -5.61878681e-01
-5.66353798e-01 4.01187241e-01 -6.02147639e-01 7.91820347e-01
1.08397865e+00 -4.06010956e-01 -4.90446128e-02 -1.83330387e-01
5.00422001e-01 3.07340920e-02 -7.69420564e-02 -1.59269965e+00
-3.99898477e-02 4.15957540e-01 5.26555717e-01 -6.03604078e-01
6.31759226e-01 1.43806204e-01 5.62075078e-01 2.06298634e-01
-1.02357221e+00 -2.15992630e-01 5.35372607e-02 -9.22448281e-03
-6.23943985e-01 1.34013565e-02 5.75806141e-01 -3.24323922e-02
-1.80486828e-01 1.76978365e-01 -5.64704835e-01 2.21349478e-01
6.87416121e-02 4.01323587e-02 6.62723929e-02 -5.19027770e-01
-1.10181129e+00 -7.23022163e-01 9.37622041e-03 8.41964781e-01
3.12288851e-01 -1.70161760e+00 -1.02870727e+00 2.30048135e-01
4.07060206e-01 -6.07077777e-01 5.19426689e-02 3.27226490e-01
-5.07365763e-01 1.17261484e-01 -1.80122003e-01 -4.38447535e-01
-1.39734495e+00 2.71568298e-02 2.98448533e-01 7.23774033e-03
-7.61516273e-01 1.02089751e+00 5.49764261e-02 -6.07375085e-01
4.30372655e-01 -5.59408307e-01 -1.72293290e-01 4.06494379e-01
7.46689796e-01 5.47349930e-01 1.76476717e-01 -7.04557598e-01
-3.86632055e-01 7.65822649e-01 2.13315532e-01 -3.66105676e-01
1.44626760e+00 3.77935112e-01 3.29624675e-02 1.25171554e+00
1.12385845e+00 2.33811587e-01 -7.90391564e-01 1.08894058e-01
-2.53834248e-01 5.10698631e-02 -1.45921245e-01 -1.14273190e+00
-9.45548594e-01 9.59164441e-01 5.15075207e-01 2.56045252e-01
1.18005788e+00 -4.08479339e-03 9.07453418e-01 6.01278067e-01
2.65048832e-01 -1.09628928e+00 -1.99769791e-02 8.75455558e-01
1.35864198e+00 -6.26696229e-01 -3.16967219e-01 2.15284899e-01
-6.13065064e-01 1.30098832e+00 7.81940371e-02 -5.67895412e-01
6.55758619e-01 2.87451267e-01 3.96796077e-01 1.44575611e-01
-7.74266005e-01 -1.06964834e-01 7.52641022e-01 1.71550930e-01
1.20542228e+00 5.45452416e-01 -3.78281102e-02 1.27614987e+00
-1.22004414e+00 7.35138506e-02 3.86565447e-01 3.95994574e-01
-2.36535475e-01 -9.29604053e-01 -3.11046720e-01 5.31368077e-01
-7.84852922e-01 -1.38597950e-01 -6.51171327e-01 1.54194430e-01
2.54933923e-01 1.15950871e+00 2.95840621e-01 -9.49914515e-01
2.75009632e-01 3.21934193e-01 5.04061937e-01 -5.35229445e-01
-1.37569284e+00 3.31697345e-01 2.79535711e-01 -3.95550072e-01
-3.55073690e-01 -1.56807646e-01 -9.61876988e-01 -9.24901292e-02
-1.93191841e-01 3.13271195e-01 6.08861148e-01 5.67744136e-01
2.99986690e-01 9.83699381e-01 6.43764317e-01 -1.06953824e+00
-2.90172428e-01 -1.46010756e+00 -1.04791200e+00 5.50456524e-01
2.84855872e-01 -2.29120687e-01 -2.52849549e-01 1.23096816e-01] | [15.8627290725708, 5.234228610992432] |
078c5680-5641-49ad-8f53-3dc96f5dbe89 | unsupervised-matching-of-data-and-text | 2112.08776 | null | https://arxiv.org/abs/2112.08776v1 | https://arxiv.org/pdf/2112.08776v1.pdf | Unsupervised Matching of Data and Text | Entity resolution is a widely studied problem with several proposals to match records across relations. Matching textual content is a widespread task in many applications, such as question answering and search. While recent methods achieve promising results for these two tasks, there is no clear solution for the more general problem of matching textual content and structured data. We introduce a framework that supports this new task in an unsupervised setting for any pair of corpora, being relational tables or text documents. Our method builds a fine-grained graph over the content of the corpora and derives word embeddings to represent the objects to match in a low dimensional space. The learned representation enables effective and efficient matching at different granularity, from relational tuples to text sentences and paragraphs. Our flexible framework can exploit pre-trained resources, but it does not depends on their existence and achieves better quality performance in matching content when the vocabulary is domain specific. We also introduce optimizations in the graph creation process with an "expand and compress" approach that first identifies new valid relationships across elements, to improve matching, and then prunes nodes and edges, to reduce the graph size. Experiments on real use cases and public datasets show that our framework produces embeddings that outperform word embeddings and fine-tuned language models both in results' quality and in execution times. | ['Paolo Papotti', 'Hansjorg Sand', 'Naser Ahmadi'] | 2021-12-16 | null | null | null | null | ['entity-resolution'] | ['natural-language-processing'] | [-1.44409360e-02 1.25884429e-01 -6.23376012e-01 -3.81614059e-01
-7.55048811e-01 -6.07785940e-01 6.06975496e-01 1.03694236e+00
-6.18770957e-01 5.15400648e-01 5.99682748e-01 -1.39941499e-01
-6.27256334e-01 -1.54083884e+00 -5.12230396e-01 -5.37704304e-02
-1.78247392e-01 1.00061905e+00 5.95458865e-01 -4.70529556e-01
2.80169666e-01 5.68214834e-01 -1.62022495e+00 2.87982523e-01
8.64793241e-01 7.40286946e-01 5.07526398e-02 4.74265188e-01
-9.73702192e-01 5.97386360e-01 -3.65593731e-01 -9.49011445e-01
3.73028725e-01 2.84791410e-01 -1.10816348e+00 -3.15052569e-01
4.64569598e-01 1.88613892e-01 -5.17801821e-01 1.03330386e+00
4.70562398e-01 1.51516035e-01 4.86771584e-01 -1.06955135e+00
-1.08354223e+00 9.90425646e-01 -4.91242349e-01 2.85599053e-01
7.31490970e-01 -6.14908040e-01 1.58182514e+00 -8.90364587e-01
9.97648835e-01 1.29063165e+00 7.38430858e-01 2.48696327e-01
-1.20287025e+00 -5.26032925e-01 -1.17424205e-01 2.62972593e-01
-1.60976970e+00 -3.17224771e-01 1.66524082e-01 -2.82047719e-01
1.42765999e+00 4.94197696e-01 1.20668575e-01 5.81248820e-01
1.17184140e-01 1.70333788e-01 3.65664691e-01 -5.02909482e-01
2.70223226e-02 4.44257081e-01 4.51103181e-01 4.76079494e-01
7.92398691e-01 -4.90495175e-01 -5.14097810e-01 -4.93924439e-01
4.53089267e-01 2.22413298e-02 -1.46680996e-01 -6.01828754e-01
-1.01074576e+00 9.08427298e-01 2.88035005e-01 7.95059383e-01
-3.27225626e-01 -1.75530612e-01 5.22195637e-01 5.96152365e-01
4.81576890e-01 8.47595930e-01 -5.07748008e-01 1.06448270e-01
-6.47680879e-01 4.03448075e-01 1.29966712e+00 1.36806846e+00
1.03979015e+00 -6.82314813e-01 -3.94156188e-01 1.03194439e+00
1.00726396e-01 1.61160678e-01 5.07610977e-01 -4.32975560e-01
9.37294483e-01 1.15999007e+00 -1.36186015e-02 -1.43701577e+00
-3.78052682e-01 -9.05905664e-02 -4.96992528e-01 -4.90611225e-01
1.81960404e-01 2.62792557e-01 -5.11973202e-01 1.48101497e+00
6.21575296e-01 -4.03797887e-02 1.39957592e-01 6.80735052e-01
1.04442859e+00 5.84146142e-01 1.39481127e-01 2.81558698e-03
1.78648221e+00 -7.36856878e-01 -8.98402333e-01 -2.61321902e-01
9.50643361e-01 -8.00753593e-01 7.41621196e-01 -2.65234858e-01
-9.44895864e-01 -4.19916600e-01 -9.49033260e-01 -6.32425368e-01
-1.00527668e+00 -3.24773639e-01 6.79048181e-01 6.97410405e-01
-9.65477109e-01 6.21743858e-01 -2.61593938e-01 -8.37987065e-01
1.05257615e-01 4.91113961e-01 -7.81079113e-01 -1.92201465e-01
-1.53783751e+00 1.06592441e+00 7.13417947e-01 -3.65075558e-01
2.58767128e-01 -8.62303138e-01 -1.15518010e+00 3.31511557e-01
6.51365519e-01 -7.72046149e-01 6.63407922e-01 -1.31841913e-01
-7.30495930e-01 1.04902184e+00 -1.86626688e-01 -6.31890237e-01
8.55635479e-03 -2.27009118e-01 -7.85051703e-01 -7.01544508e-02
3.89528304e-01 2.50698090e-01 2.48174563e-01 -9.10487890e-01
-7.18119919e-01 -3.95784706e-01 1.32869914e-01 1.37905300e-01
-9.14170563e-01 1.99930206e-01 -1.00232196e+00 -5.26046097e-01
9.47742388e-02 -6.17366970e-01 -2.02912092e-01 -1.82571843e-01
-2.85650909e-01 -6.54651701e-01 4.00240541e-01 -5.58525562e-01
1.82249701e+00 -1.96459162e+00 -5.88229522e-02 4.99837309e-01
4.19725537e-01 2.11165190e-01 -2.02043161e-01 1.01797664e+00
-1.07518829e-01 3.72861862e-01 -1.46410927e-01 -1.02646708e-01
2.85487622e-01 3.43856335e-01 -2.97467351e-01 1.87584966e-01
2.52480209e-01 1.10477102e+00 -9.34177279e-01 -8.15162480e-01
-1.63286522e-01 2.62333333e-01 -6.00718558e-01 2.56913543e-01
-7.65024722e-02 -5.80903709e-01 -6.02342725e-01 4.55953032e-01
5.97785115e-01 -2.71787167e-01 5.93778670e-01 -3.45324457e-01
-1.20137380e-02 5.49506783e-01 -1.67267561e+00 1.73318636e+00
-4.81426656e-01 3.64104092e-01 -2.14846894e-01 -1.07812393e+00
1.31609917e+00 2.35011637e-01 6.36708498e-01 -9.05583918e-01
-1.17457509e-01 3.17326099e-01 -2.33009398e-01 -9.51842308e-01
1.14728808e+00 2.35339627e-01 -3.73812199e-01 4.38208818e-01
1.01658039e-01 -2.07960382e-02 7.28371739e-01 4.52425778e-01
1.41948426e+00 -1.36664003e-01 5.09370565e-01 -2.59083420e-01
5.03846407e-01 2.11584657e-01 4.80607271e-01 6.08613014e-01
4.90367025e-01 3.94625336e-01 4.93950844e-01 -4.53093112e-01
-9.98565972e-01 -8.57080042e-01 -2.45285109e-01 1.33475459e+00
4.11959916e-01 -8.12979281e-01 -3.90397489e-01 -6.62545979e-01
6.60086930e-01 5.40872753e-01 -6.98464334e-01 -5.09408042e-02
-8.45775843e-01 -6.64814889e-01 3.29623729e-01 5.37231147e-01
-2.07963839e-01 -1.02082348e+00 -9.81627405e-02 5.37567139e-01
-9.73996241e-03 -1.33667529e+00 -4.66780603e-01 1.96261317e-01
-6.01104379e-01 -1.14681792e+00 -8.23576078e-02 -1.01311564e+00
4.37743247e-01 2.00116470e-01 1.49868882e+00 2.39557728e-01
-4.94267434e-01 2.85855353e-01 -4.72389728e-01 -1.94010869e-01
-1.84649125e-01 4.32166845e-01 -5.29059693e-02 -1.39722914e-01
8.67306232e-01 -4.54879940e-01 -2.04969913e-01 1.98779181e-01
-1.25387895e+00 -5.42505741e-01 4.47774649e-01 6.07214689e-01
5.91546059e-01 -5.30112209e-03 5.32999218e-01 -1.49337113e+00
9.66476142e-01 -8.62024665e-01 -4.79153037e-01 6.99193656e-01
-9.75368202e-01 6.73411906e-01 4.44789231e-01 -2.79603750e-01
-6.74028635e-01 -2.36642361e-01 -1.48611516e-01 -1.56495273e-01
1.60136238e-01 7.07953930e-01 -1.19418036e-02 2.00264201e-01
7.34511733e-01 -2.27829963e-01 -3.87860060e-01 -6.33408964e-01
7.03330219e-01 6.72950447e-01 5.74503362e-01 -7.68413246e-01
1.12624097e+00 2.75485188e-01 -1.72981679e-01 -4.46318716e-01
-5.79663813e-01 -9.47489798e-01 -6.38817489e-01 3.42149734e-01
8.26119661e-01 -7.25295305e-01 -4.68507975e-01 -6.27367914e-01
-1.20418882e+00 3.44380885e-01 -3.99299920e-01 2.98069984e-01
-8.69611204e-02 2.78398812e-01 -4.51710522e-01 -3.31198633e-01
-4.50886995e-01 -7.27995217e-01 1.01258612e+00 1.71529248e-01
-4.06600714e-01 -1.19444728e+00 4.89879280e-01 1.40787140e-01
3.50382179e-01 8.21332708e-02 1.33098650e+00 -1.26742160e+00
-6.56143308e-01 -5.37466586e-01 -6.14432156e-01 -3.01003575e-01
3.18571031e-01 -1.59233764e-01 -6.06708765e-01 -7.15347901e-02
-5.31270742e-01 -1.23819314e-01 7.01521635e-01 -2.86832452e-01
9.45199549e-01 -3.60010922e-01 -7.49199808e-01 5.21303296e-01
1.65606713e+00 -1.19447382e-02 7.37792969e-01 5.69480717e-01
7.57853985e-01 9.41984296e-01 6.17551684e-01 2.27997944e-01
6.20458841e-01 8.67089510e-01 2.75360286e-01 -2.26844028e-02
-7.41223395e-02 -4.14492846e-01 -1.55514970e-01 9.70842600e-01
2.26577550e-01 -4.46095854e-01 -8.35652530e-01 8.77554953e-01
-1.94722533e+00 -1.09080827e+00 -3.16109806e-01 2.13073945e+00
1.07232773e+00 5.77781387e-02 -1.93537056e-01 1.03926636e-01
7.49825954e-01 1.27418667e-01 -1.81142036e-02 -7.11472690e-01
-8.74995533e-03 6.85541451e-01 8.23822618e-01 3.51714373e-01
-8.85254145e-01 9.21034873e-01 5.71864080e+00 6.93467975e-01
-6.86492085e-01 3.00790146e-02 -6.86356798e-02 1.76231295e-01
-7.92828739e-01 1.14250034e-02 -1.06803322e+00 1.65700451e-01
9.26482201e-01 -5.80270112e-01 1.83200106e-01 7.91000664e-01
-3.65940362e-01 3.40155005e-01 -1.29413748e+00 8.70464027e-01
4.41883281e-02 -1.69403422e+00 1.18957400e-01 1.07074697e-02
5.43340266e-01 -2.52738297e-01 -2.74721503e-01 5.72000980e-01
3.43593091e-01 -1.00790834e+00 3.50259274e-01 4.30129290e-01
6.62421465e-01 -6.86308384e-01 8.58955085e-01 6.96871728e-02
-1.62080824e+00 4.05222662e-02 -7.24490285e-01 3.50667447e-01
-1.19250994e-02 5.52919984e-01 -9.55534995e-01 9.80839372e-01
7.53419518e-01 5.17817557e-01 -7.48954952e-01 9.30191219e-01
1.59898549e-01 -3.81313525e-02 -1.69331208e-01 -1.94006920e-01
8.20759684e-02 -2.53661454e-01 2.74340928e-01 1.66202986e+00
3.69404495e-01 -8.19526389e-02 8.84021968e-02 5.47728539e-01
-5.14402151e-01 5.71633041e-01 -8.03159237e-01 -9.24465135e-02
1.00046265e+00 1.37628031e+00 -5.47331452e-01 -3.60921651e-01
-6.75265908e-01 6.09733224e-01 8.76241088e-01 7.13517964e-02
-6.71851039e-01 -8.29814792e-01 8.20321262e-01 3.42936963e-01
5.28840184e-01 -4.38750349e-02 -2.08261102e-01 -9.94119227e-01
3.07013422e-01 -7.93873847e-01 8.96370351e-01 -1.97644070e-01
-1.41190159e+00 6.87019110e-01 1.58944745e-02 -1.00048018e+00
-2.90663242e-01 -4.29643571e-01 -1.84858367e-01 9.32717085e-01
-1.59416819e+00 -7.33349323e-01 -2.29493082e-01 5.10789156e-01
2.51753032e-01 2.83649918e-02 1.10782480e+00 8.64852488e-01
-5.06337225e-01 8.14179361e-01 1.50794923e-01 4.21718478e-01
9.50436950e-01 -1.37040448e+00 6.21760905e-01 6.31926775e-01
6.71894252e-01 1.02891409e+00 6.51105940e-01 -5.88664293e-01
-1.72706246e+00 -1.11031139e+00 1.78224969e+00 -6.30794048e-01
9.85076427e-01 -5.71568429e-01 -1.33963454e+00 5.33306658e-01
2.78631449e-01 2.85239667e-01 5.86181641e-01 4.88388777e-01
-7.77519882e-01 -2.74300605e-01 -1.16696823e+00 5.00136197e-01
1.28056324e+00 -5.98882973e-01 -9.29072499e-01 3.95071924e-01
9.75666702e-01 -4.67764586e-01 -1.40196383e+00 4.18568611e-01
3.94871622e-01 -6.69032574e-01 1.15728021e+00 -8.05404484e-01
3.00310105e-01 -3.22850823e-01 -2.48784199e-01 -9.75705624e-01
-4.29254889e-01 -3.99360687e-01 -2.88193554e-01 1.52425444e+00
7.75572777e-01 -7.27667093e-01 6.86082900e-01 7.07099199e-01
1.34029672e-01 -8.18559110e-01 -7.52289116e-01 -8.75618935e-01
-1.99294657e-01 -3.28641027e-01 1.16236734e+00 1.14670384e+00
1.41689256e-01 3.80726367e-01 7.90837184e-02 4.13428992e-01
2.46690482e-01 5.31044066e-01 8.93143237e-01 -1.44854772e+00
3.30598578e-02 -3.67726982e-01 -5.35677075e-01 -7.61028707e-01
1.80340737e-01 -1.29178882e+00 -2.60263473e-01 -2.00309634e+00
7.98724741e-02 -8.69640887e-01 -2.79462218e-01 2.75763661e-01
-2.57852793e-01 -4.20605615e-02 -1.78388357e-01 1.38468787e-01
-6.15927756e-01 1.06207594e-01 6.11972690e-01 -2.44339868e-01
-1.07276365e-01 -4.79641795e-01 -8.73425543e-01 1.51721224e-01
4.30166006e-01 -9.58340645e-01 -3.00057918e-01 -6.40759766e-01
5.62013626e-01 -8.21938142e-02 -1.59900770e-01 -6.80625021e-01
5.28142333e-01 -2.12737471e-01 -7.86216334e-02 -4.14871991e-01
8.67282897e-02 -9.57213461e-01 2.34977692e-01 7.36589544e-03
-4.81495053e-01 3.98148715e-01 1.05070613e-01 5.55650294e-01
-3.41756910e-01 -4.74450678e-01 3.18305701e-01 9.78268981e-02
-8.07941496e-01 5.42309463e-01 3.44553024e-01 4.86093223e-01
8.74278784e-01 -1.70202643e-01 -4.50949758e-01 9.78212208e-02
-3.85293335e-01 4.01616156e-01 4.78488207e-01 9.03392613e-01
4.18445528e-01 -1.54109001e+00 -7.53917038e-01 -2.47485028e-03
5.78570545e-01 -6.65241778e-02 -2.69441009e-01 3.85231704e-01
-4.54159677e-01 4.60336268e-01 1.39377788e-01 -1.49535984e-01
-1.21338797e+00 7.83105016e-01 -1.33415639e-01 -7.97981441e-01
-5.76462626e-01 7.18839884e-01 -1.65617496e-01 -5.33277631e-01
1.79280668e-01 -3.45907837e-01 -6.37675166e-01 5.15844762e-01
5.16965032e-01 2.21523479e-01 5.39700866e-01 -4.97502506e-01
-4.76040184e-01 5.62557697e-01 -2.69173533e-01 3.16548377e-01
1.49815679e+00 -1.10726058e-01 -4.49712813e-01 1.83980301e-01
1.43017006e+00 3.99859935e-01 -2.05861792e-01 -6.84547067e-01
8.60327244e-01 -5.63207030e-01 -3.60172898e-01 -1.23170726e-01
-9.53525424e-01 2.75188684e-01 1.03156127e-01 7.94190049e-01
7.51165748e-01 1.48785964e-01 9.55224872e-01 5.14365852e-01
3.51091594e-01 -1.10237730e+00 -1.84741408e-01 5.12506783e-01
6.42423272e-01 -9.96676624e-01 2.08657384e-01 -6.28304005e-01
-3.95265877e-01 1.15447676e+00 5.12460113e-01 -1.37173519e-01
5.42621851e-01 4.28328812e-01 -8.18470642e-02 -5.04596531e-01
-8.20081353e-01 -4.57720220e-01 5.73656380e-01 6.49308503e-01
6.71577990e-01 -1.22135758e-01 -6.33173466e-01 5.23260355e-01
-2.16525599e-01 -3.43194574e-01 1.66270062e-01 9.68621731e-01
-4.34706032e-01 -1.61952758e+00 -2.19116271e-01 7.47396171e-01
-4.16711032e-01 -1.90762177e-01 -3.39325696e-01 9.39146280e-01
2.86886305e-01 8.28723490e-01 3.34223896e-01 -2.86938310e-01
7.63854682e-01 1.22962296e-01 2.06801325e-01 -1.07405591e+00
-8.39670718e-01 -5.76391935e-01 5.28606713e-01 -5.30416191e-01
-2.78167307e-01 -4.04975802e-01 -1.34034038e+00 -4.43543255e-01
-4.14906174e-01 5.20444870e-01 4.26059991e-01 7.76690662e-01
4.98110831e-01 3.78448814e-01 5.59515357e-01 -4.18744460e-02
-3.55557621e-01 -6.70984447e-01 -5.64549088e-01 8.22611153e-01
-5.34411222e-02 -6.53537869e-01 3.57333571e-02 -1.78461045e-01] | [9.274896621704102, 8.171138763427734] |
9300b032-5273-4ed2-ada9-3f6e6e93ed59 | the-typical-behavior-of-bandit-algorithms | 2210.05660 | null | https://arxiv.org/abs/2210.05660v1 | https://arxiv.org/pdf/2210.05660v1.pdf | The Typical Behavior of Bandit Algorithms | We establish strong laws of large numbers and central limit theorems for the regret of two of the most popular bandit algorithms: Thompson sampling and UCB. Here, our characterizations of the regret distribution complement the characterizations of the tail of the regret distribution recently developed by Fan and Glynn (2021) (arXiv:2109.13595). The tail characterizations there are associated with atypical bandit behavior on trajectories where the optimal arm mean is under-estimated, leading to mis-identification of the optimal arm and large regret. In contrast, our SLLN's and CLT's here describe the typical behavior and fluctuation of regret on trajectories where the optimal arm mean is properly estimated. We find that Thompson sampling and UCB satisfy the same SLLN and CLT, with the asymptotics of both the SLLN and the (mean) centering sequence in the CLT matching the asymptotics of expected regret. Both the mean and variance in the CLT grow at $\log(T)$ rates with the time horizon $T$. Asymptotically as $T \to \infty$, the variability in the number of plays of each sub-optimal arm depends only on the rewards received for that arm, which indicates that each sub-optimal arm contributes independently to the overall CLT variance. | ['Peter W. Glynn', 'Lin Fan'] | 2022-10-11 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [-1.55701265e-01 3.47533792e-01 -5.52740097e-01 -1.63829759e-01
-9.59160745e-01 -1.08290291e+00 9.71430615e-02 1.08795196e-01
-6.28698707e-01 1.08812141e+00 1.37335137e-01 -8.65494728e-01
-7.96302795e-01 -5.50477386e-01 -1.09403503e+00 -8.62722695e-01
-1.91655457e-01 6.00351989e-01 -2.55321681e-01 2.25618362e-01
2.02623054e-01 2.85414517e-01 -1.05564320e+00 -9.26061347e-02
7.88565397e-01 1.61632299e+00 -2.34216496e-01 8.46519828e-01
-2.49901533e-01 7.29565322e-01 -4.12433982e-01 -5.64132988e-01
7.22119391e-01 -6.68554962e-01 -6.17188990e-01 -3.42885017e-01
-7.10451230e-02 -5.09917498e-01 -3.62123072e-01 1.28771698e+00
1.80477440e-01 1.78088367e-01 6.33357108e-01 -9.95229423e-01
-3.36057156e-01 1.15261376e+00 -9.74809289e-01 2.96319842e-01
3.35360132e-02 -3.45644206e-02 1.08386958e+00 2.73282677e-01
2.93135673e-01 1.15402603e+00 6.96772218e-01 4.91512924e-01
-1.16760457e+00 -6.48053944e-01 3.06437761e-01 -3.93166542e-01
-8.78209174e-01 -4.67727721e-01 1.66985244e-01 -3.62712026e-01
2.73616970e-01 5.68003178e-01 7.19902456e-01 5.84926963e-01
1.33925468e-01 1.01259685e+00 1.00956428e+00 -5.59948325e-01
6.00044668e-01 3.20228115e-02 2.88226664e-01 2.63777584e-01
6.53135240e-01 4.65890497e-01 -2.01496065e-01 -3.42270821e-01
8.66216183e-01 2.31270716e-01 -2.45574206e-01 -2.62746513e-01
-6.71679020e-01 7.60968149e-01 2.30997950e-01 -3.71805616e-02
-6.94256186e-01 8.10098886e-01 2.95779973e-01 6.14862084e-01
6.64292336e-01 2.94465795e-02 -5.55424631e-01 -5.02824545e-01
-8.96281838e-01 4.21045363e-01 9.21263158e-01 1.29436171e+00
3.03805292e-01 -4.29314941e-01 -6.67260528e-01 2.69899637e-01
-4.16080654e-02 7.52952218e-01 1.12217553e-01 -1.45620644e+00
7.56891251e-01 4.86649424e-02 1.09040427e+00 -1.29383534e-01
-3.14460605e-01 -7.97212601e-01 -5.83549917e-01 -4.50347885e-02
1.16590083e+00 -5.07240295e-01 -4.62706119e-01 1.88778269e+00
-2.61578569e-03 -6.07661903e-01 -1.50274351e-01 7.07060218e-01
-4.70265388e-01 3.46242100e-01 -2.19575197e-01 -8.80676866e-01
8.45884800e-01 -6.18411899e-01 -5.75111985e-01 -2.16142073e-01
6.72364116e-01 -5.77280819e-01 6.29653513e-01 1.75351307e-01
-1.53747761e+00 2.29699403e-01 -5.19146144e-01 7.25939214e-01
3.27053338e-01 -3.41817677e-01 6.65746927e-01 9.87024188e-01
-7.98706770e-01 8.46044421e-01 -6.60320461e-01 -4.65416051e-02
6.64553165e-01 1.11017995e-01 2.44714737e-01 -1.86412916e-01
-5.94537616e-01 4.23907638e-01 1.80636153e-01 1.62393749e-01
-7.26557076e-01 -7.53677905e-01 -8.01943764e-02 2.12497979e-01
4.92170006e-01 -4.97879207e-01 1.89085019e+00 -1.42350149e+00
-1.46559036e+00 4.06156838e-01 -1.00604668e-01 -7.57127285e-01
1.18052804e+00 -1.81013882e-01 2.91499406e-01 -1.40462533e-01
1.57640636e-01 -1.73174903e-01 6.58700109e-01 -6.09747589e-01
-8.25388849e-01 -6.05614722e-01 8.33300948e-02 1.70406476e-01
-2.48772316e-02 -3.01114231e-01 1.68197528e-01 -3.18282634e-01
1.02121560e-02 -8.78107786e-01 -3.37486356e-01 -1.03705913e-01
-2.73075968e-01 -3.35475393e-02 -1.94933847e-01 -2.13286191e-01
1.17613852e+00 -2.08924961e+00 -3.28454524e-01 3.26885670e-01
-1.73306093e-01 -3.36861551e-01 4.59304266e-02 5.62992573e-01
4.03504260e-02 2.32749075e-01 3.14401120e-01 4.58365232e-02
3.15408200e-01 -2.22886689e-02 -4.84575301e-01 1.01411760e+00
-7.37337530e-01 7.46555924e-01 -8.19829702e-01 6.06347620e-02
-1.64413482e-01 -6.39824867e-01 -6.65196776e-01 3.60467955e-02
-4.28878546e-01 1.90374956e-01 -6.48205400e-01 1.42929360e-01
7.16465414e-01 -3.47627223e-01 1.31698832e-01 4.60019499e-01
-3.24490905e-01 2.32346520e-01 -9.73708034e-01 1.03118443e+00
-1.20653287e-01 2.11783022e-01 3.26242179e-01 -1.06026578e+00
1.93500847e-01 2.80916184e-01 4.57967281e-01 -5.09979486e-01
4.34081167e-01 4.80481654e-01 -3.98758575e-02 -2.53117323e-01
1.47042543e-01 -6.94508374e-01 -4.14940082e-02 1.05645370e+00
-3.66686702e-01 2.71689326e-01 1.86150879e-01 2.66763985e-01
1.06636083e+00 -2.27648929e-01 1.81931749e-01 -4.80195016e-01
-1.04442179e-01 -9.75339338e-02 3.52962226e-01 1.58360851e+00
-3.78695101e-01 4.77332212e-02 1.07806313e+00 -3.17948431e-01
-1.16016519e+00 -1.09955585e+00 -9.16887522e-02 1.13655114e+00
1.10119067e-01 1.58618137e-01 -8.00324202e-01 -4.39176083e-01
5.89620411e-01 1.00987816e+00 -9.91015494e-01 -9.63687024e-04
5.33161648e-02 -7.18473017e-01 1.36876270e-01 2.42917925e-01
3.46880525e-01 -7.08041251e-01 -7.78690100e-01 2.40534201e-01
1.76644232e-02 -7.76152611e-01 -8.39568615e-01 2.03524679e-01
-9.85837877e-01 -1.00349867e+00 -7.94098139e-01 1.55795842e-01
5.12553096e-01 1.12376533e-01 8.18133056e-01 -4.88607466e-01
5.49456067e-02 5.68789244e-01 -3.49032342e-01 -7.78845847e-01
-2.32577667e-01 -2.14595661e-01 1.53389117e-02 -6.95063220e-03
-6.78410083e-02 -4.42879945e-01 -1.01434529e+00 2.24352509e-01
-4.46030498e-01 -4.26721752e-01 4.49787140e-01 5.30865729e-01
3.84397238e-01 -1.28315240e-01 5.91196477e-01 -6.48733318e-01
6.14805579e-01 -4.32216763e-01 -1.10989463e+00 3.32189620e-01
-6.27880335e-01 2.95386851e-01 4.71240968e-01 -4.96404767e-01
-9.67417955e-01 -4.23327684e-01 3.82277459e-01 -3.57347697e-01
4.51453090e-01 2.74300843e-01 4.84179318e-01 2.80069232e-01
4.79133308e-01 1.15375645e-01 2.26800248e-01 -6.09108090e-01
3.06937456e-01 6.49496496e-01 1.99390918e-01 -8.88629377e-01
2.73846745e-01 4.80497658e-01 5.61217405e-02 -3.14339697e-01
-1.40724206e+00 -1.34156525e-01 3.14553022e-01 -1.64532289e-01
2.05705911e-01 -5.30090094e-01 -1.66174960e+00 2.21231371e-01
-8.79930675e-01 -4.27743018e-01 -1.01075494e+00 6.91735208e-01
-1.16258800e+00 8.19393694e-02 -5.79763234e-01 -1.90717852e+00
-2.66003698e-01 -6.05571270e-01 5.11482537e-01 3.55471820e-01
2.49395566e-03 -6.84832990e-01 -1.42867882e-02 1.09150931e-01
6.02536678e-01 -8.56357254e-03 8.05529892e-01 -5.63174844e-01
-4.45286095e-01 -6.23487711e-01 -2.92859584e-01 2.63227433e-01
-1.10844940e-01 -3.83568972e-01 -4.96375263e-01 -5.08538961e-01
2.24316254e-01 -1.61802292e-01 6.36725545e-01 1.32704878e+00
9.89591897e-01 -8.22542787e-01 -3.16847980e-01 3.63869369e-01
1.40129995e+00 5.40764570e-01 2.01264471e-01 3.65983725e-01
-2.93577254e-01 3.34162325e-01 4.43090796e-01 9.68532622e-01
-2.94285178e-01 3.74448970e-02 4.96031225e-01 6.80257857e-01
6.91939771e-01 -3.18037510e-01 4.96242851e-01 4.02409676e-03
-1.64218936e-02 -7.18978420e-02 -4.97971624e-01 5.71361661e-01
-2.09747243e+00 -1.06876051e+00 1.54948965e-01 2.79302406e+00
8.96125913e-01 4.68321621e-01 6.30623221e-01 -3.80147457e-01
6.16055727e-01 -1.51377499e-01 -1.01778984e+00 -5.44823349e-01
1.22791477e-01 1.20435990e-01 1.50882196e+00 6.59859419e-01
-5.43007791e-01 4.16681081e-01 7.39899778e+00 1.00913107e+00
-4.76542681e-01 1.56619906e-01 7.93598831e-01 -9.61849809e-01
-3.67639542e-01 8.29693377e-02 -6.79954052e-01 7.15506792e-01
1.13779545e+00 -6.73983157e-01 8.23783636e-01 9.23776746e-01
2.40113825e-01 -4.34310049e-01 -1.09979975e+00 7.44660676e-01
-7.67269969e-01 -1.32311344e+00 -4.99583274e-01 2.48685464e-01
8.33860397e-01 -6.45305291e-02 1.31784037e-01 1.78333893e-01
9.11425471e-01 -8.74985397e-01 1.09126925e+00 6.03463411e-01
8.56915057e-01 -1.00436783e+00 7.53223658e-01 7.18511105e-01
-6.29972637e-01 -6.21782839e-01 -3.06831151e-01 -4.73845840e-01
4.60821353e-02 8.28337014e-01 -2.58837253e-01 4.34316009e-01
7.41000235e-01 1.88250929e-01 2.91301548e-01 1.01786363e+00
1.94637641e-01 5.65693915e-01 -6.32255852e-01 -2.85256803e-01
3.55983764e-01 -5.24307966e-01 5.58667719e-01 4.91204888e-01
4.43068177e-01 1.05368122e-01 -2.31249720e-01 8.35632801e-01
-7.45186657e-02 -5.82735799e-02 -9.64857191e-02 -4.61995035e-01
5.49899638e-01 5.96796870e-01 -4.22799945e-01 -1.47362381e-01
-2.63291411e-02 4.75469679e-01 2.26346493e-01 4.91378397e-01
-6.79284573e-01 -2.49628738e-01 8.60156476e-01 6.41491339e-02
5.43160856e-01 2.03794926e-01 -5.00118732e-01 -7.08612323e-01
2.60823339e-01 -3.71244848e-01 5.25251567e-01 -2.56511599e-01
-1.32629573e+00 1.91487167e-02 2.36138739e-02 -8.81131530e-01
-2.52965003e-01 -3.88844937e-01 -4.54611123e-01 9.31052148e-01
-9.62343812e-01 -2.47380778e-01 5.24483442e-01 4.84083951e-01
2.15169098e-02 8.23019370e-02 3.77542526e-01 -1.86597586e-01
-5.74844718e-01 7.40027428e-01 1.16442442e+00 -9.95680392e-02
1.17850989e-01 -9.44454908e-01 -4.35632206e-02 3.85222018e-01
-5.15385091e-01 5.05088329e-01 1.21111023e+00 -4.52777594e-01
-1.43489122e+00 -7.26013422e-01 2.25618273e-01 -2.58261919e-01
1.00243545e+00 9.94173884e-02 -1.87584013e-02 1.10914791e+00
-1.92466393e-01 -1.00027956e-01 3.62253010e-01 3.88148993e-01
-1.27205610e-01 -4.06275362e-01 -1.27155185e+00 4.54919428e-01
9.61541474e-01 -1.86984941e-01 -4.57319170e-02 4.76918489e-01
6.70759320e-01 -2.40293175e-01 -4.92770523e-01 5.12647070e-02
1.01830482e+00 -1.36677992e+00 3.33785951e-01 -9.17636514e-01
1.27399027e-01 5.47566831e-01 -2.68786490e-01 -1.07282090e+00
-8.67678300e-02 -1.33843744e+00 -2.83476897e-02 8.46195340e-01
3.11606944e-01 -9.05980825e-01 1.02694082e+00 8.25750709e-01
3.44748497e-01 -7.84024835e-01 -1.41635120e+00 -1.23440063e+00
4.34504837e-01 -4.80903268e-01 4.08382863e-01 1.09393053e-01
1.58517048e-01 -1.86652705e-01 -2.50747323e-01 -3.94002736e-01
1.03558016e+00 4.90832508e-01 3.53253663e-01 -8.97246897e-01
-5.51042020e-01 -8.53596985e-01 2.41923243e-01 -1.37787461e+00
-3.43611807e-01 -5.85769713e-01 4.99090031e-02 -1.07887840e+00
5.97742319e-01 -6.27112389e-01 -3.69869232e-01 -2.95503493e-02
2.26065382e-01 -6.00559413e-01 2.37054005e-01 1.67520314e-01
-7.44130135e-01 4.20667142e-01 1.21472895e+00 1.14330575e-01
-2.35167205e-01 7.00002491e-01 -9.66403246e-01 6.48908556e-01
6.07948124e-01 -7.06252396e-01 -1.93263873e-01 -3.56804356e-02
5.95158517e-01 8.28271985e-01 4.79776300e-02 -4.80731279e-01
4.22509853e-03 -4.68803078e-01 6.65891543e-02 -5.08789778e-01
-1.26596689e-01 -8.03150654e-01 2.18393236e-01 7.85156071e-01
-7.25178957e-01 -2.52030611e-01 -2.96681728e-02 1.02692723e+00
5.23869038e-01 -5.43044090e-01 7.93877482e-01 -2.91881770e-01
5.89589536e-01 3.89783531e-01 -5.03318071e-01 4.20132458e-01
1.08663702e+00 1.10714279e-01 -4.25213248e-01 -1.04368031e+00
-5.71743011e-01 5.04649580e-01 1.85765415e-01 -2.30164424e-01
-3.92411798e-02 -1.15036023e+00 -4.21600401e-01 -4.55333404e-02
-3.76874447e-01 -1.39082775e-01 4.52635437e-01 1.07900953e+00
-1.47164017e-01 5.28338253e-01 6.00061007e-02 -2.16183022e-01
-5.69657326e-01 6.29722595e-01 6.45496786e-01 -4.41869915e-01
-2.40902901e-01 9.84189153e-01 1.04889296e-01 1.66477963e-01
5.26154876e-01 -2.54659921e-01 7.01478660e-01 4.67217639e-02
6.84742033e-01 6.69122159e-01 -3.87106180e-01 2.68344492e-01
-1.28982723e-01 2.10438177e-01 -2.61297792e-01 -4.27644312e-01
1.26695836e+00 -2.39327088e-01 -2.28043526e-01 5.31511307e-01
1.04506588e+00 -6.58704937e-02 -1.42481053e+00 -2.93078482e-01
2.18682006e-01 -5.58336914e-01 -2.77998835e-01 -7.85979033e-01
-8.64442170e-01 4.97082323e-01 4.16003466e-01 7.45614469e-01
7.54791737e-01 1.54873505e-01 5.09359896e-01 1.51309833e-01
6.35016918e-01 -1.11097825e+00 -3.54280651e-01 4.65716630e-01
7.18740046e-01 -7.27866530e-01 -1.42647084e-02 2.43707895e-01
-4.09901828e-01 8.19899797e-01 -8.74644145e-02 -2.13671580e-01
7.45946527e-01 2.31044680e-01 -3.97994548e-01 1.67330474e-01
-7.54197776e-01 -1.20191090e-02 -2.50496924e-01 -3.58272977e-02
7.95800537e-02 4.48136866e-01 -5.57070374e-01 9.61923778e-01
-3.64216655e-01 1.24520503e-01 3.77266884e-01 9.04107749e-01
-6.81954622e-01 -8.14547539e-01 -4.08666790e-01 9.44422901e-01
-8.91404629e-01 3.00387405e-02 -9.31325629e-02 6.09028041e-01
-5.24007618e-01 8.08764875e-01 3.03763211e-01 1.86294034e-01
2.73841769e-01 1.50468171e-01 8.88414681e-01 -7.70663694e-02
-2.22401530e-01 2.81039149e-01 1.98494103e-02 -5.22398412e-01
-1.11686736e-01 -8.07819486e-01 -7.72819936e-01 -8.61561418e-01
-3.54912400e-01 5.42690516e-01 2.95540422e-01 1.07665837e+00
2.09742622e-03 5.78898154e-02 7.64634609e-01 -4.66025144e-01
-1.62954426e+00 -9.91502941e-01 -1.29477978e+00 2.45956019e-01
7.47792780e-01 -1.89667478e-01 -7.90847659e-01 -6.31690741e-01] | [4.5203094482421875, 3.296755790710449] |
57b1ba15-75b3-40d4-92b1-06a4e9f94f4c | styo-stylize-your-face-in-only-one-shot | 2303.03231 | null | https://arxiv.org/abs/2303.03231v2 | https://arxiv.org/pdf/2303.03231v2.pdf | StyO: Stylize Your Face in Only One-Shot | This paper focuses on face stylization with a single artistic target. Existing works for this task often fail to retain the source content while achieving geometry variation. Here, we present a novel StyO model, ie. Stylize the face in only One-shot, to solve the above problem. In particular, StyO exploits a disentanglement and recombination strategy. It first disentangles the content and style of source and target images into identifiers, which are then recombined in a cross manner to derive the stylized face image. In this way, StyO decomposes complex images into independent and specific attributes, and simplifies one-shot face stylization as the combination of different attributes from input images, thus producing results better matching face geometry of target image and content of source one. StyO is implemented with latent diffusion models (LDM) and composed of two key modules: 1) Identifier Disentanglement Learner (IDL) for disentanglement phase. It represents identifiers as contrastive text prompts, ie. positive and negative descriptions. And it introduces a novel triple reconstruction loss to fine-tune the pre-trained LDM for encoding style and content into corresponding identifiers; 2) Fine-grained Content Controller (FCC) for the recombination phase. It recombines disentangled identifiers from IDL to form an augmented text prompt for generating stylized faces. In addition, FCC also constrains the cross-attention maps of latent and text features to preserve source face details in results. The extensive evaluation shows that StyO produces high-quality images on numerous paintings of various styles and outperforms the current state-of-the-art. Code will be released upon acceptance. | ['Tiande Guo', 'Yinhan Hu', 'Congying Han', 'Xuecheng Nie', 'ZiCheng Zhang', 'Bonan Li'] | 2023-03-06 | null | null | null | null | ['one-shot-face-stylization'] | ['computer-vision'] | [ 3.17430019e-01 7.41672441e-02 -2.29015440e-01 -2.33936816e-01
-5.11646807e-01 -6.34890139e-01 6.92501307e-01 -7.51782775e-01
1.28482014e-01 5.97300947e-01 3.28350723e-01 3.77285749e-01
8.67425576e-02 -7.19357967e-01 -7.33394623e-01 -9.96569812e-01
4.99551326e-01 5.89072049e-01 -2.19688594e-01 -9.49202105e-02
1.94085129e-02 6.74858987e-01 -1.50640619e+00 3.02420229e-01
9.87135410e-01 9.15419281e-01 1.20986164e-01 5.13618946e-01
-4.16302443e-01 6.76816702e-01 -8.16291809e-01 -8.77590418e-01
2.10660517e-01 -6.25726581e-01 -4.33477968e-01 2.72130132e-01
7.42851257e-01 -4.59223419e-01 -1.45279899e-01 1.04691541e+00
6.67272627e-01 -2.68111438e-01 1.08668959e+00 -1.56909573e+00
-1.50440037e+00 5.14675736e-01 -9.93170977e-01 -1.96738526e-01
2.77194917e-01 3.31661969e-01 8.78086150e-01 -8.45635593e-01
7.56948411e-01 1.82372236e+00 2.79697120e-01 9.24687028e-01
-1.75758648e+00 -1.22201502e+00 2.14136302e-01 -1.17774516e-01
-1.31580818e+00 -5.84977984e-01 1.21474099e+00 -5.04501700e-01
3.75669271e-01 2.96768010e-01 6.07557416e-01 1.65355635e+00
1.59381643e-01 8.72171104e-01 1.12277758e+00 -1.83577314e-01
-1.28212534e-02 2.43678495e-01 -2.86724657e-01 7.75420249e-01
2.31017709e-01 2.83661097e-01 -7.08096921e-01 -1.86670218e-02
1.17980766e+00 -9.81735438e-02 -2.53617853e-01 -6.28216684e-01
-8.94607008e-01 8.61689031e-01 3.25382411e-01 6.12929687e-02
-2.08269179e-01 5.92196770e-02 7.01526403e-02 2.27561817e-01
4.42334741e-01 4.34470862e-01 -9.42369252e-02 1.29309252e-01
-8.00069749e-01 2.20512763e-01 4.91306037e-01 1.20741117e+00
8.73629034e-01 4.31443721e-01 -6.08124733e-01 1.01093650e+00
2.54821658e-01 8.70967269e-01 4.09884423e-01 -9.29758847e-01
4.33222026e-01 6.26943827e-01 -1.22974448e-01 -1.04340219e+00
1.11720443e-01 -2.04913199e-01 -9.14058089e-01 6.08057499e-01
4.45759743e-02 -2.00380266e-01 -1.13394892e+00 1.97513211e+00
1.56832382e-01 3.71075779e-01 5.69849182e-03 7.62889624e-01
1.07975817e+00 5.96056521e-01 2.04749346e-01 -1.35174260e-01
1.43375099e+00 -1.05376208e+00 -1.06930506e+00 -1.23466603e-01
4.82240729e-02 -8.52630794e-01 1.34900403e+00 3.04898292e-01
-1.26979911e+00 -7.85311043e-01 -1.05934489e+00 -2.72050977e-01
-2.27479532e-01 4.43898767e-01 3.73953938e-01 8.71101439e-01
-1.07262862e+00 4.13800359e-01 -4.41043228e-01 9.26824585e-02
7.26886034e-01 4.46103960e-01 -3.67854327e-01 6.97110668e-02
-1.02750731e+00 5.82324803e-01 4.36988585e-02 -6.21692091e-03
-9.54138994e-01 -9.85251486e-01 -9.46127057e-01 5.88411726e-02
3.38355213e-01 -1.03903425e+00 9.37051833e-01 -1.50637925e+00
-2.01679063e+00 9.74321783e-01 -2.50689089e-01 2.01098680e-01
5.83376348e-01 -1.87064365e-01 -5.42590022e-01 2.38959119e-02
1.60935506e-01 9.65234876e-01 1.69818187e+00 -1.81515455e+00
-2.34803483e-01 -3.51175189e-01 -1.29161879e-01 2.34370649e-01
-3.75376642e-01 -6.59058169e-02 -7.55286813e-01 -1.13755107e+00
-1.78293526e-01 -7.09645987e-01 4.13054168e-01 9.28376690e-02
-4.77232724e-01 1.26402497e-01 1.07016540e+00 -4.79722589e-01
1.09768701e+00 -2.19384408e+00 7.95596123e-01 -9.12033916e-02
4.72592741e-01 3.58842939e-01 -5.74844182e-01 1.37450933e-01
-3.72135460e-01 2.40261078e-01 -1.73341960e-01 -8.81272256e-01
-4.63371165e-02 1.90459982e-01 -4.46553737e-01 1.31171852e-01
5.86085200e-01 1.22905731e+00 -7.25310862e-01 -6.07744157e-01
1.23131201e-01 6.56653404e-01 -5.78163743e-01 3.36033612e-01
-3.80065382e-01 5.32846034e-01 -2.82111853e-01 6.18567705e-01
1.07009625e+00 -3.95236500e-02 -2.66917497e-02 -5.88600457e-01
3.50864172e-01 -3.66917342e-01 -1.18777096e+00 1.66300046e+00
-5.43318391e-01 4.60021883e-01 7.73632228e-02 -1.92031950e-01
1.05723393e+00 2.16059312e-01 1.72882617e-01 -6.39268577e-01
3.05725098e-01 -1.66511431e-01 -4.83889729e-01 -5.07919729e-01
3.07685882e-01 -3.20393175e-01 1.28162712e-01 4.74463046e-01
2.33855054e-01 -2.29533747e-01 -8.62832293e-02 2.17471540e-01
5.01266301e-01 3.59794140e-01 -1.00553878e-01 -2.33341865e-02
6.53399646e-01 -6.27066433e-01 5.17492473e-01 2.41361201e-01
1.55818298e-01 8.50467443e-01 7.90380061e-01 -1.41195357e-01
-9.26607192e-01 -1.51456714e+00 1.81118622e-01 8.41621935e-01
1.99498728e-01 -2.98287600e-01 -1.02170074e+00 -8.39517295e-01
1.45756572e-01 7.70529985e-01 -1.05949175e+00 -3.12482625e-01
-5.20939410e-01 -5.06088614e-01 5.51359594e-01 3.62909585e-01
5.20613134e-01 -1.31626964e+00 -2.49998868e-02 -2.96329379e-01
-2.07962524e-02 -7.79853463e-01 -9.05664861e-01 -2.17537418e-01
-5.81435919e-01 -7.89850891e-01 -7.69613028e-01 -6.84166193e-01
8.08034837e-01 1.02613971e-01 1.09062159e+00 -1.25051826e-01
-1.83906749e-01 1.27089605e-01 -2.49945059e-01 -2.16425836e-01
-3.32369506e-01 -2.64611356e-02 -1.04808640e-02 5.99326074e-01
1.17123626e-01 -7.00546741e-01 -4.44802850e-01 2.73768216e-01
-1.02765703e+00 2.13453695e-01 6.94099903e-01 9.40891862e-01
4.84537154e-01 -3.39869149e-02 4.77510571e-01 -1.13620937e+00
7.16201186e-01 -2.71910995e-01 -5.07667124e-01 3.27276140e-01
-5.32779753e-01 2.31677160e-01 7.22472548e-01 -8.11650813e-01
-1.46282589e+00 -9.54095274e-02 9.61226970e-02 -9.09255803e-01
-6.31744340e-02 -2.05144659e-01 -9.58963990e-01 -1.21865878e-02
4.66058522e-01 2.15636268e-01 1.45238087e-01 -4.94524926e-01
7.74278700e-01 4.75402832e-01 5.99891126e-01 -7.33079970e-01
1.11640215e+00 5.52806616e-01 -2.34944716e-01 -6.88570261e-01
-6.06291294e-01 3.33112806e-01 -7.33049631e-01 -1.30381584e-01
1.06157148e+00 -8.71584356e-01 -7.74063528e-01 8.02290976e-01
-1.17587614e+00 -3.71222883e-01 -5.16394675e-01 -4.57839109e-02
-5.49174845e-01 2.07011327e-01 -5.19308746e-01 -5.87367773e-01
-1.23298652e-01 -1.27696550e+00 1.44394827e+00 3.20170790e-01
-2.50925541e-01 -8.18248570e-01 -1.06618002e-01 3.34077299e-01
1.36223599e-01 2.98931718e-01 1.11266303e+00 1.53971510e-02
-7.50243783e-01 1.95208356e-01 -4.87312943e-01 2.50344843e-01
5.14616132e-01 3.18405002e-01 -1.20194554e+00 -3.04203421e-01
-2.82406285e-02 -2.32608154e-01 8.28593969e-01 3.17310631e-01
9.44201469e-01 -3.86416852e-01 -2.01539949e-01 1.07415164e+00
1.24008167e+00 2.52338797e-01 8.18323612e-01 3.78154330e-02
1.18853247e+00 7.64603078e-01 3.23155552e-01 2.56362140e-01
1.82068229e-01 7.03917027e-01 3.13935369e-01 -3.22100163e-01
-6.23361886e-01 -5.86075962e-01 4.74958450e-01 5.23975968e-01
1.76583916e-01 -3.75413269e-01 -2.90352106e-01 1.69848457e-01
-1.58173037e+00 -1.00480151e+00 1.93975225e-01 1.79023898e+00
9.02465641e-01 -3.82434689e-02 9.21078473e-02 -1.54413372e-01
6.90690577e-01 4.26750809e-01 -7.95776308e-01 -3.34419608e-01
-3.80620748e-01 2.12341756e-01 1.91295981e-01 6.07271850e-01
-7.29848981e-01 1.41191387e+00 5.60092497e+00 1.14314640e+00
-1.20457029e+00 1.56007707e-01 7.93696940e-01 -3.41017097e-01
-7.62936056e-01 -2.48070270e-01 -8.70282888e-01 5.93732953e-01
3.36101919e-01 6.42106533e-02 6.77752852e-01 6.54242337e-01
8.92991722e-02 1.28483787e-01 -1.13919008e+00 1.34076071e+00
4.11869407e-01 -1.29956043e+00 7.31252909e-01 -8.40018466e-02
9.05572772e-01 -8.74479949e-01 6.86879694e-01 1.80103809e-01
3.72553080e-01 -1.29683638e+00 1.10575020e+00 6.69209182e-01
1.40851343e+00 -7.88475633e-01 2.27071449e-01 -8.87923315e-02
-1.22394526e+00 -8.56638476e-02 -1.82161704e-01 3.81675899e-01
1.25211820e-01 2.06641048e-01 -2.14249626e-01 6.18883550e-01
5.27505100e-01 7.08452702e-01 -6.76660120e-01 3.45244884e-01
-4.46738839e-01 1.52739957e-01 1.75618514e-01 2.44047657e-01
-1.62473038e-01 -4.56500888e-01 5.63464522e-01 8.36891532e-01
1.44780830e-01 6.07455932e-02 -2.25123420e-01 1.42479932e+00
-2.44453877e-01 -5.99399880e-02 -4.94005203e-01 -2.29530577e-02
5.92998922e-01 1.18992150e+00 -4.59127694e-01 -3.67646456e-01
-1.45708919e-01 1.45628750e+00 4.46222663e-01 6.79900765e-01
-9.16833580e-01 -2.29473203e-01 8.55169237e-01 7.87388980e-02
3.90993297e-01 7.01864511e-02 -6.22800350e-01 -1.12208545e+00
-9.06857997e-02 -1.02014971e+00 7.18349367e-02 -1.06784952e+00
-1.42662513e+00 9.55628872e-01 5.40122166e-02 -9.83448386e-01
1.67980492e-02 -5.17264724e-01 -7.01069295e-01 1.32381654e+00
-1.44842231e+00 -1.69858539e+00 -4.43374753e-01 6.49398029e-01
6.35318518e-01 -3.39396149e-01 6.93355620e-01 2.34684959e-01
-9.16990757e-01 1.01206887e+00 -1.49884358e-01 -2.40272805e-02
1.03904724e+00 -1.20775378e+00 4.52984214e-01 6.71530843e-01
6.88706189e-02 6.47357523e-01 5.21384418e-01 -9.32689428e-01
-1.27575111e+00 -1.11373127e+00 4.70247239e-01 -6.27228260e-01
3.02375406e-01 -5.90960860e-01 -8.04030657e-01 5.67752481e-01
2.46758923e-01 -3.24497283e-01 5.02086580e-01 -2.77309716e-01
-7.08798885e-01 -2.65542328e-01 -1.16575086e+00 9.75052416e-01
1.09238553e+00 -6.32606268e-01 -3.66535544e-01 -1.90672383e-01
9.08669233e-01 -3.18891406e-01 -5.25431335e-01 1.89150870e-01
6.52050614e-01 -1.05763733e+00 1.02326322e+00 -3.43938708e-01
7.03028917e-01 -3.66025537e-01 1.35469317e-01 -1.29044914e+00
-5.59175253e-01 -8.46060097e-01 -2.08164930e-01 1.83086896e+00
2.08501637e-01 -4.56253350e-01 9.03170109e-01 4.04121220e-01
2.10424647e-01 -6.65327907e-01 -6.25395417e-01 -6.90166056e-01
1.01535685e-01 -2.60460237e-03 1.07766187e+00 8.82128894e-01
-6.33571446e-01 5.98765314e-01 -6.16011620e-01 -1.02939773e-02
6.53931379e-01 7.04585314e-02 8.63692462e-01 -1.10285068e+00
-2.14923188e-01 -6.48568749e-01 -1.33783016e-02 -9.62244630e-01
4.15639877e-01 -8.14134121e-01 -3.22953641e-01 -1.06377256e+00
1.94122359e-01 -4.74197030e-01 1.68303519e-01 3.12017798e-01
-3.01265091e-01 4.90242004e-01 3.62766922e-01 1.99445903e-01
9.27305371e-02 9.65842664e-01 1.71044314e+00 -2.75061458e-01
-3.41845453e-01 -2.51989752e-01 -1.00137973e+00 5.41149795e-01
5.40427029e-01 -3.06548178e-01 -6.99381769e-01 -5.34188092e-01
-5.21654636e-02 3.41767222e-02 3.52237195e-01 -6.75075114e-01
-1.87233791e-01 -1.98725611e-01 7.77965784e-01 -4.60574925e-01
6.80735826e-01 -8.10712576e-01 4.35068607e-01 1.03280947e-01
-2.54032850e-01 9.64362547e-03 2.23710969e-01 6.29032552e-01
2.31672339e-02 -1.29308373e-01 1.06893623e+00 1.59548715e-01
-4.40324396e-01 5.95966697e-01 8.13777819e-02 -4.84236889e-02
9.98400390e-01 -5.09979248e-01 -2.64171869e-01 -3.05605084e-01
-5.98107874e-01 -3.73184010e-02 7.42677748e-01 7.24088490e-01
6.09448373e-01 -1.68601143e+00 -6.91736460e-01 7.22562373e-01
-9.90408435e-02 -8.72433260e-02 5.62054276e-01 3.71339083e-01
-3.95246714e-01 -7.40614682e-02 -5.14709055e-01 -3.94471467e-01
-1.47404206e+00 9.17491138e-01 1.15515597e-01 -1.01883933e-01
-4.02521431e-01 1.07051349e+00 8.65824342e-01 -2.17309371e-01
2.04039156e-01 -3.42636928e-02 -3.29509258e-01 3.94056290e-01
5.44753850e-01 1.67266965e-01 -4.42861259e-01 -9.07582760e-01
5.39147370e-02 1.10299397e+00 -2.16973931e-01 -3.02568138e-01
1.08151257e+00 -2.84608871e-01 -2.32574552e-01 3.33718985e-01
1.33789980e+00 3.54586184e-01 -1.62323391e+00 -1.20941371e-01
-4.38014627e-01 -7.14277864e-01 -2.75391698e-01 -8.48348916e-01
-1.38443112e+00 8.89920235e-01 5.51821172e-01 -3.80200595e-01
1.34392416e+00 -1.46315888e-01 7.90958822e-01 -3.73646259e-01
-1.04219988e-01 -6.71752036e-01 7.12295890e-01 1.64323226e-01
1.28237772e+00 -8.89770329e-01 -2.43906870e-01 -7.00957179e-01
-1.01314688e+00 9.20704305e-01 8.66847634e-01 -1.04519702e-01
5.58781266e-01 3.51634294e-01 9.35795903e-02 -3.01161379e-01
-5.13137102e-01 7.50450790e-02 5.44887006e-01 8.70357573e-01
1.44991249e-01 9.72418711e-02 2.07408860e-01 7.73077309e-01
-4.16448474e-01 -2.48231471e-01 2.70515263e-01 3.10140938e-01
9.37245190e-02 -1.37121224e+00 -5.14796793e-01 2.14034498e-01
-7.79508650e-02 -1.27479106e-01 -7.85732269e-01 6.55036390e-01
5.49600005e-01 6.70906007e-01 4.75342013e-02 -5.69439173e-01
3.11719328e-01 -1.90136619e-02 6.43193066e-01 -6.43277287e-01
-4.95273232e-01 1.78861901e-01 -3.04107726e-01 -3.66474837e-01
-8.03910196e-02 -5.16688049e-01 -7.81293035e-01 -4.95719641e-01
-2.15162054e-01 -5.31232581e-02 2.78127968e-01 5.88554382e-01
4.43643868e-01 8.32599819e-01 7.44296432e-01 -1.06381619e+00
-2.75362760e-01 -7.64987707e-01 -7.37695932e-01 5.33790410e-01
3.86094660e-01 -9.76247311e-01 -2.29291871e-01 3.84044945e-01] | [12.23538875579834, -0.30464065074920654] |
f84ee9ad-02ac-4c1c-b111-db7e05652ee4 | odfnet-using-orientation-distribution | 2012.04708 | null | https://arxiv.org/abs/2012.04708v2 | https://arxiv.org/pdf/2012.04708v2.pdf | ODFNet: Using orientation distribution functions to characterize 3D point clouds | Learning new representations of 3D point clouds is an active research area in 3D vision, as the order-invariant point cloud structure still presents challenges to the design of neural network architectures. Recent works explored learning either global or local features or both for point clouds, however none of the earlier methods focused on capturing contextual shape information by analysing local orientation distribution of points. In this paper, we leverage on point orientation distributions around a point in order to obtain an expressive local neighborhood representation for point clouds. We achieve this by dividing the spherical neighborhood of a given point into predefined cone volumes, and statistics inside each volume are used as point features. In this way, a local patch can be represented by not only the selected point's nearest neighbors, but also considering a point density distribution defined along multiple orientations around the point. We are then able to construct an orientation distribution function (ODF) neural network that involves an ODFBlock which relies on mlp (multi-layer perceptron) layers. The new ODFNet model achieves state-of the-art accuracy for object classification on ModelNet40 and ScanObjectNN datasets, and segmentation on ShapeNet S3DIS datasets. | ['Gozde Unal', 'Alican Mertan', 'Yusuf H. Sahin'] | 2020-12-08 | null | null | null | null | ['3d-part-segmentation'] | ['computer-vision'] | [-1.09584488e-01 -1.39211431e-01 -1.16796896e-01 -5.55746794e-01
-3.91486019e-01 -5.83569288e-01 8.13755572e-01 5.76086402e-01
-4.34721380e-01 5.09615466e-02 -3.17817092e-01 -1.50072843e-01
-3.69226754e-01 -1.10125113e+00 -1.08209443e+00 -7.31964529e-01
-1.48669034e-01 9.14914191e-01 4.20817554e-01 4.17022146e-02
3.66047293e-01 1.55854440e+00 -1.80176890e+00 3.02390337e-01
4.86317694e-01 1.28871489e+00 1.92910507e-01 4.83950287e-01
-4.80597019e-01 1.71983168e-01 -4.72160667e-01 9.83098596e-02
3.29649001e-01 4.08257902e-01 -5.67562699e-01 -1.00942396e-01
7.05466807e-01 -4.60852198e-02 9.89009738e-02 9.48542178e-01
3.41099769e-01 1.55268818e-01 1.02344465e+00 -9.85807776e-01
-6.68363869e-01 2.21971244e-01 -2.27700606e-01 1.11626185e-01
7.44657665e-02 -3.00451051e-02 1.01300335e+00 -1.13136733e+00
5.45722067e-01 1.24718106e+00 8.75378609e-01 2.82791138e-01
-1.14990783e+00 -2.06814423e-01 1.52195483e-01 1.26439691e-01
-1.28449643e+00 1.64841965e-01 1.06729949e+00 -6.04048729e-01
1.16669571e+00 3.03411067e-01 1.06203508e+00 8.08317542e-01
1.28402978e-01 6.60876215e-01 9.51628923e-01 -4.20602649e-01
3.63191634e-01 1.41029237e-02 2.62528658e-01 3.25282961e-01
1.82585791e-01 1.11966059e-01 -5.77792786e-02 -9.17579979e-02
1.05100524e+00 4.10940111e-01 2.13410985e-02 -9.00445282e-01
-1.01387072e+00 9.20481682e-01 1.20943189e+00 6.37847602e-01
-5.44652462e-01 1.86722800e-01 1.44107178e-01 7.61841089e-02
6.67531908e-01 3.09699684e-01 -4.25721347e-01 2.12550536e-01
-8.59562099e-01 4.73052263e-01 7.52792239e-01 6.91239357e-01
1.23202956e+00 -1.24376051e-01 1.74020901e-02 7.65740871e-01
4.70366001e-01 5.96668839e-01 1.55713841e-01 -6.67354286e-01
1.89243019e-01 1.05339098e+00 -2.43525937e-01 -1.06899273e+00
-5.78487992e-01 -5.31730354e-01 -8.21104646e-01 6.82629406e-01
1.56556249e-01 4.36979383e-01 -1.07778668e+00 1.10300851e+00
4.76852000e-01 1.81345925e-01 -2.80342817e-01 9.49154735e-01
9.61854458e-01 6.45693898e-01 -2.31441647e-01 2.73547471e-01
1.13173974e+00 -3.83556247e-01 8.34009722e-02 2.74967756e-02
2.64691681e-01 -2.95462281e-01 5.75745642e-01 1.83095887e-01
-1.03973508e+00 -7.24285603e-01 -9.34611440e-01 -1.82036966e-01
-7.85989821e-01 -6.13656119e-02 4.22510058e-01 2.90978134e-01
-1.10229325e+00 8.26845527e-01 -9.52577472e-01 -2.59866118e-01
6.97776020e-01 5.27555525e-01 -2.56131023e-01 4.28071991e-02
-6.97644591e-01 9.03197527e-01 2.46970415e-01 1.74511150e-01
-5.19924819e-01 -8.19394827e-01 -8.20882201e-01 1.12028785e-01
-6.15180619e-02 -8.37769508e-01 8.09884548e-01 -6.41133726e-01
-1.19344318e+00 1.08847535e+00 -1.61261961e-01 -4.37275648e-01
3.49972814e-01 7.06245229e-02 5.52641265e-02 1.78641707e-01
-1.03279697e-02 8.22887182e-01 9.94957209e-01 -1.63785958e+00
-5.51837146e-01 -8.40462506e-01 5.69357909e-03 1.80687740e-01
3.20445225e-02 -2.59290606e-01 -2.12141842e-01 -2.94315219e-01
6.52055442e-01 -8.97667944e-01 -1.60630077e-01 1.32050827e-01
-2.42124602e-01 -8.16377163e-01 8.98287654e-01 -9.15606245e-02
4.44351465e-01 -2.08549285e+00 5.66424690e-02 6.27112031e-01
3.57172102e-01 3.27305585e-01 2.62475498e-02 3.95483486e-02
-1.63414165e-01 6.10185303e-02 -2.29186133e-01 -3.52217078e-01
1.11645125e-01 3.20949048e-01 -1.61019504e-01 5.06717443e-01
4.02393728e-01 9.98701811e-01 -5.83072364e-01 -1.35193422e-01
7.11268187e-01 8.31576526e-01 -6.05047464e-01 -3.25147100e-02
-3.93618315e-01 2.25491703e-01 -5.09799540e-01 6.90984845e-01
1.00270283e+00 8.15525744e-03 -5.02704024e-01 -2.23023623e-01
-3.12015265e-01 6.00292720e-02 -1.11930788e+00 1.55017948e+00
-5.42318225e-01 3.48048896e-01 -1.02081425e-01 -1.00210083e+00
1.54546833e+00 9.11455080e-02 7.58041799e-01 -3.79873961e-01
1.81685850e-01 4.47131336e-01 -3.15541029e-01 -2.00896487e-01
1.99898526e-01 -2.02677786e-01 2.41738915e-01 1.10085104e-02
1.52681798e-01 -5.25594473e-01 -3.78056467e-01 -6.16141319e-01
9.03544843e-01 -3.95151414e-02 3.61464620e-02 -9.73152742e-02
5.46439111e-01 1.66619325e-03 8.18268210e-03 7.64299273e-01
-5.99692203e-02 9.18101549e-01 1.61011279e-01 -9.17718470e-01
-1.04251683e+00 -1.21631062e+00 -5.53180516e-01 7.12749302e-01
6.65494949e-02 2.75445655e-02 -5.70719779e-01 -4.72115546e-01
5.11031687e-01 4.89546329e-01 -5.92886269e-01 3.19225863e-02
-7.87914455e-01 -1.85119405e-01 2.93394387e-01 5.55551410e-01
2.61808515e-01 -1.27531242e+00 -7.90596783e-01 1.45809814e-01
4.66644228e-01 -9.56216693e-01 7.33539909e-02 6.38119400e-01
-1.22957301e+00 -9.35637891e-01 -6.73138976e-01 -8.47334802e-01
6.04501665e-01 1.05412170e-01 1.13001561e+00 -1.79927051e-01
-1.50492042e-01 5.36769927e-01 -2.16038629e-01 -7.38803566e-01
-7.75579885e-02 1.89265579e-01 -1.11014634e-01 -3.76191386e-03
8.66631269e-01 -6.91388428e-01 -3.79085928e-01 8.74814689e-02
-7.43849933e-01 -5.21189213e-01 5.84159017e-01 6.01280749e-01
1.00054467e+00 -2.79045314e-01 1.53771251e-01 -4.31348652e-01
3.61843526e-01 -3.95414114e-01 -6.23173714e-01 -1.40869007e-01
-8.22547153e-02 -3.82839553e-02 3.55593532e-01 -1.47521585e-01
-5.09635150e-01 2.44490176e-01 -4.60791320e-01 -1.07717621e+00
-7.90050983e-01 4.62085992e-01 -3.61098312e-02 -4.56556052e-01
6.35403395e-01 2.19732061e-01 2.50535645e-02 -6.71075463e-01
3.45681250e-01 5.63271880e-01 2.99193084e-01 -5.50013006e-01
7.36818850e-01 6.95463538e-01 3.32343727e-01 -1.18530786e+00
-4.24353421e-01 -7.29652166e-01 -1.20473790e+00 -1.25696629e-01
9.62222636e-01 -6.21122658e-01 -9.00767088e-01 3.09732914e-01
-1.44606900e+00 1.79091468e-01 -7.77163744e-01 3.51413488e-01
-6.48174226e-01 -3.54284607e-02 -1.90806448e-01 -8.31402123e-01
-2.41045773e-01 -1.12102354e+00 1.46914434e+00 8.91209990e-02
5.76926582e-02 -1.15178692e+00 3.01844388e-01 -5.96599728e-02
2.71976680e-01 5.15324533e-01 1.03593361e+00 -8.28857243e-01
-8.53372276e-01 -5.51848114e-01 -2.75413781e-01 5.37968040e-01
-1.47088498e-01 -4.76272851e-02 -1.06690478e+00 -1.37279376e-01
2.83117324e-01 2.62741628e-03 1.00025570e+00 6.98552012e-01
1.35825491e+00 -4.00767522e-03 -4.75295931e-01 8.66871893e-01
1.56455660e+00 2.86334548e-02 3.59820038e-01 4.33602363e-01
8.90371680e-01 3.19021970e-01 1.84660196e-01 3.50849003e-01
3.17021996e-01 6.16630256e-01 9.39154327e-01 5.62971318e-03
9.11318213e-02 -1.13566332e-01 -1.71671271e-01 4.71708149e-01
-5.95998883e-01 7.62341097e-02 -1.14198577e+00 6.52655125e-01
-1.50589812e+00 -7.31954753e-01 -1.36961982e-01 1.99992943e+00
2.31927380e-01 2.03950182e-01 -6.18178919e-02 2.97841489e-01
5.60654223e-01 1.97873577e-01 -4.61890668e-01 -4.38784719e-01
-1.14303797e-01 4.83627766e-01 4.04598117e-01 2.45287582e-01
-1.30737960e+00 7.16517031e-01 5.44494200e+00 6.03404045e-01
-1.37678683e+00 -1.95827097e-01 2.39161417e-01 3.44843306e-02
-2.34439462e-01 -1.89658344e-01 -8.26679230e-01 1.91681594e-01
6.38109267e-01 3.56800795e-01 1.26333535e-01 1.13629246e+00
-3.80321778e-02 3.39287668e-01 -1.15368736e+00 1.12715816e+00
1.70064822e-01 -1.37279570e+00 3.31822872e-01 2.20770076e-01
4.51146156e-01 6.76997066e-01 1.23151913e-01 1.57517508e-01
-1.76166669e-01 -1.08650029e+00 8.24502409e-01 8.55939627e-01
5.80617666e-01 -8.49623859e-01 6.90993845e-01 6.11059308e-01
-1.13400578e+00 -9.07332152e-02 -7.97605217e-01 1.17897606e-02
-1.79390579e-01 7.74628639e-01 -1.05171359e+00 5.40702999e-01
9.69838917e-01 8.24515879e-01 -5.67594111e-01 1.33903646e+00
7.38606304e-02 2.72801369e-01 -8.36249113e-01 -2.41222680e-01
5.02290905e-01 -3.79477233e-01 9.29295301e-01 1.18577230e+00
4.86977398e-01 -2.55073398e-01 2.29359671e-01 1.20195210e+00
-2.72258595e-02 -6.42958879e-02 -9.93756294e-01 3.83918762e-01
3.18581730e-01 1.21907747e+00 -7.16011584e-01 -1.20114824e-02
-1.47020474e-01 3.72156113e-01 4.38445657e-01 2.04127297e-01
-2.75424927e-01 -2.73326546e-01 8.11147511e-01 2.71632046e-01
6.62519574e-01 -5.74623525e-01 -5.62904775e-01 -8.55765402e-01
1.71456739e-01 -3.16164374e-01 7.64998570e-02 -8.18152070e-01
-1.45182216e+00 6.98191643e-01 8.31255317e-02 -1.30683494e+00
-1.30144075e-01 -1.05374420e+00 -7.57348895e-01 9.92687285e-01
-1.58980238e+00 -1.45208526e+00 -4.09670234e-01 4.35780466e-01
2.17934951e-01 -4.72720340e-02 7.22513616e-01 1.46203523e-03
2.05743492e-01 8.10657665e-02 7.25838542e-02 2.28333939e-02
7.39040673e-02 -1.52390778e+00 4.20077503e-01 2.08138734e-01
4.79423463e-01 6.26836300e-01 3.12448621e-01 -5.09438753e-01
-1.28245211e+00 -1.18824291e+00 8.47223103e-01 -7.59744108e-01
3.33935618e-01 -4.83905494e-01 -1.21256530e+00 5.30776501e-01
-3.35017800e-01 4.89235222e-01 1.55669972e-01 7.32756853e-02
-2.00745896e-01 -1.20622620e-01 -1.25998402e+00 9.74701792e-02
1.09310746e+00 -6.08628094e-01 -7.66450405e-01 1.77671492e-01
6.23694658e-01 -5.72695911e-01 -9.81425583e-01 6.05556905e-01
1.11398533e-01 -1.10085547e+00 1.35555184e+00 -4.59634870e-01
1.71759173e-01 -3.01084280e-01 -2.65252709e-01 -1.31066418e+00
-4.65499282e-01 1.80956602e-01 -9.75128412e-02 9.31434870e-01
1.42631784e-01 -6.20284498e-01 9.65040445e-01 9.90035832e-02
-5.23388088e-01 -1.09169364e+00 -1.35829985e+00 -6.98669434e-01
4.23857480e-01 -8.25417638e-01 9.21179891e-01 7.39680290e-01
-5.58526635e-01 7.95665011e-02 4.57210660e-01 4.55296785e-01
5.81732035e-01 3.53928268e-01 5.44502318e-01 -1.70108557e+00
2.70479858e-01 -7.97773659e-01 -1.05208421e+00 -1.05706584e+00
2.40864143e-01 -1.31661046e+00 -2.12478966e-01 -1.78864789e+00
-3.51986021e-01 -7.36188829e-01 -2.53932059e-01 3.47890347e-01
3.64513844e-01 3.38316441e-01 3.02871346e-01 2.72590250e-01
-2.60505468e-01 6.82567418e-01 1.22767329e+00 -3.28775555e-01
-2.98298776e-01 3.62669706e-01 -9.46526788e-03 9.16201830e-01
5.16849279e-01 -3.24557871e-01 5.55978455e-02 -5.15257835e-01
1.84716791e-01 -3.67426217e-01 8.92965257e-01 -1.30380917e+00
3.99996221e-01 1.53672963e-01 7.38031268e-01 -1.23473358e+00
6.26377285e-01 -1.25474668e+00 -6.68067187e-02 1.10194482e-01
5.86474799e-02 -1.13532014e-01 1.57826945e-01 4.42625582e-01
-4.16183114e-01 -4.30830300e-01 7.74506986e-01 -2.69548565e-01
-5.41398466e-01 5.38589001e-01 1.04486346e-01 -4.95610058e-01
9.03650403e-01 -6.34413600e-01 2.75155604e-02 1.61219165e-01
-8.56140554e-01 -1.77779105e-02 5.35298169e-01 4.50160533e-01
9.29776430e-01 -1.50511539e+00 -4.98089045e-01 5.36533594e-01
1.61170736e-01 8.03682387e-01 3.62229832e-02 4.43253577e-01
-7.32038736e-01 5.75184464e-01 -1.95246637e-01 -1.50030839e+00
-9.83087718e-01 4.65669721e-01 6.72183156e-01 1.84338048e-01
-7.42160261e-01 8.66211593e-01 -6.45302162e-02 -8.51138115e-01
8.26519802e-02 -9.22094166e-01 -6.06686532e-01 1.62188321e-01
-2.30838805e-02 2.61749029e-01 4.69293654e-01 -7.99758673e-01
-4.97124135e-01 1.21632874e+00 2.64163852e-01 2.79284328e-01
1.68860018e+00 3.90995800e-01 -1.94814265e-01 7.61570752e-01
1.46419430e+00 -3.08450341e-01 -1.16923797e+00 -2.81423271e-01
7.20082074e-02 -4.96292323e-01 1.28759950e-01 -5.22689044e-01
-9.18290496e-01 1.03508997e+00 7.62565374e-01 5.51757097e-01
7.13993251e-01 3.55319768e-01 2.91970372e-01 5.35294354e-01
4.97219771e-01 -5.61101973e-01 -2.33106405e-01 9.54406321e-01
1.13933122e+00 -1.27251458e+00 -1.63049251e-01 -2.13258490e-01
-2.15441629e-01 1.40153062e+00 2.70627350e-01 -7.28768289e-01
1.01020896e+00 -8.51913691e-02 -7.81620368e-02 -6.13882780e-01
-3.40850949e-01 -2.05797210e-01 6.94853961e-01 7.51801908e-01
8.29064623e-02 1.85120821e-01 3.27171385e-01 4.89692718e-01
-4.14684802e-01 -1.81742236e-01 3.88158225e-02 8.14997911e-01
-5.54937720e-01 -7.25971341e-01 -6.12929642e-01 5.19628704e-01
1.22235022e-01 2.58383960e-01 -3.19013983e-01 7.53499687e-01
6.37245595e-01 2.79070437e-01 8.14277768e-01 -2.64869571e-01
6.89447463e-01 1.22359179e-01 5.50740778e-01 -6.87230527e-01
-6.21771276e-01 -5.53751588e-02 -5.14372110e-01 -3.24938625e-01
-4.78955209e-01 -7.35397160e-01 -1.26771176e+00 -6.13190047e-02
-2.52896637e-01 -4.51629609e-02 1.27187753e+00 7.97210872e-01
2.81970143e-01 4.76812750e-01 7.15775549e-01 -1.62222600e+00
-4.42679584e-01 -8.44874024e-01 -6.24444604e-01 3.32127690e-01
6.69695735e-01 -7.15702534e-01 -4.86591309e-01 -3.09894830e-01] | [7.970778465270996, -3.573122978210449] |
9fe9fd9a-494d-47e9-b7d1-9512243d504d | set-interdependence-transformer-set-to | 2206.03720 | null | https://arxiv.org/abs/2206.03720v1 | https://arxiv.org/pdf/2206.03720v1.pdf | Set Interdependence Transformer: Set-to-Sequence Neural Networks for Permutation Learning and Structure Prediction | The task of learning to map an input set onto a permuted sequence of its elements is challenging for neural networks. Set-to-sequence problems occur in natural language processing, computer vision and structure prediction, where interactions between elements of large sets define the optimal output. Models must exhibit relational reasoning, handle varying cardinalities and manage combinatorial complexity. Previous attention-based methods require $n$ layers of their set transformations to explicitly represent $n$-th order relations. Our aim is to enhance their ability to efficiently model higher-order interactions through an additional interdependence component. We propose a novel neural set encoding method called the Set Interdependence Transformer, capable of relating the set's permutation invariant representation to its elements within sets of any cardinality. We combine it with a permutation learning module into a complete, 3-part set-to-sequence model and demonstrate its state-of-the-art performance on a number of tasks. These range from combinatorial optimization problems, through permutation learning challenges on both synthetic and established NLP datasets for sentence ordering, to a novel domain of product catalog structure prediction. Additionally, the network's ability to generalize to unseen sequence lengths is investigated and a comparative empirical analysis of the existing methods' ability to learn higher-order interactions is provided. | ['Leon Derczynski', 'Mateusz Jurewicz'] | 2022-06-08 | null | null | null | null | ['relational-reasoning', 'sentence-ordering'] | ['natural-language-processing', 'natural-language-processing'] | [ 9.24695671e-01 1.61583070e-02 -4.39133346e-01 -5.79498649e-01
-4.50050861e-01 -8.80315006e-01 4.48577493e-01 3.09642404e-01
-1.92676440e-01 6.75321579e-01 1.25770003e-01 -7.05241323e-01
-7.67272234e-01 -8.46444130e-01 -1.17881572e+00 -3.71540785e-01
-4.99245971e-01 1.06540012e+00 -5.82653433e-02 -5.33522785e-01
3.49939138e-01 4.78948653e-01 -1.85579729e+00 8.23599637e-01
6.64297998e-01 9.35569465e-01 2.45139897e-01 4.36935484e-01
-2.15380058e-01 6.99372113e-01 -3.46946269e-01 -2.96027094e-01
5.16586483e-01 -1.08159102e-01 -1.15397310e+00 4.20117415e-02
6.28722966e-01 2.34480083e-01 -8.73713940e-02 9.40652072e-01
1.04624324e-01 2.45460987e-01 6.71522081e-01 -1.23252821e+00
-1.10037565e+00 1.02536535e+00 -3.11903954e-01 2.91701794e-01
3.80276144e-01 1.87626079e-01 1.96662092e+00 -5.20023227e-01
5.77054560e-01 1.35362971e+00 5.04526079e-01 2.22218618e-01
-1.71847260e+00 -5.83903968e-01 2.56027728e-01 5.64561963e-01
-1.23044097e+00 -1.82479203e-01 7.05682456e-01 -3.55720162e-01
1.61864710e+00 5.30558884e-01 5.25453508e-01 6.20390058e-01
6.10482581e-02 5.92491925e-01 8.59042048e-01 -4.68590707e-01
-2.77336240e-02 -3.95926386e-01 4.87171918e-01 6.80408895e-01
2.86540240e-01 -8.89304355e-02 -6.39526844e-01 -1.20019568e-02
4.29417163e-01 -3.34844857e-01 -6.55399859e-02 -4.59942937e-01
-1.13758183e+00 6.35947049e-01 4.98881429e-01 2.55865604e-01
-4.37880009e-02 3.83586496e-01 2.96300113e-01 4.93960053e-01
9.89074185e-02 1.14483821e+00 -9.40384448e-01 1.79573327e-01
-1.63719654e-01 2.98933595e-01 1.17558515e+00 1.08224356e+00
6.26576364e-01 -3.83764952e-01 -3.51697206e-01 9.74101841e-01
1.05665892e-01 2.14592759e-02 2.62647837e-01 -6.84639156e-01
6.50379956e-01 7.41365731e-01 -2.09855407e-01 -1.01389349e+00
-4.39186037e-01 -5.25447309e-01 -9.04712617e-01 -3.13586354e-01
1.56606749e-01 5.14032602e-01 -7.15291321e-01 2.01742816e+00
-9.01303142e-02 1.62214085e-01 -1.68489322e-01 3.61588180e-01
7.19672024e-01 6.54571176e-01 -1.78706333e-01 -2.28327438e-01
1.61182857e+00 -8.09690893e-01 -3.02416295e-01 -3.90526921e-01
7.45195746e-01 -2.84210622e-01 1.25241530e+00 3.52950156e-01
-1.20709848e+00 -5.54368913e-01 -1.17530715e+00 -5.50933108e-02
-5.10122240e-01 -2.22408623e-01 9.10944223e-01 4.41554755e-01
-9.74997818e-01 8.42731118e-01 -2.90863484e-01 8.65018964e-02
6.66188657e-01 1.05392170e+00 -3.25809866e-01 -3.10537033e-02
-1.41848516e+00 9.56661403e-01 9.69860017e-01 6.72455132e-02
-3.22081000e-01 -1.10038781e+00 -8.52109432e-01 5.63837409e-01
6.64740562e-01 -9.04307246e-01 1.20021689e+00 -7.45443225e-01
-1.26395214e+00 7.83587158e-01 -4.02129367e-02 -8.06971788e-01
-1.38458118e-01 -8.90070014e-03 -2.41956443e-01 -1.51255533e-01
-1.53098628e-01 7.56611824e-01 4.22608852e-01 -1.05932856e+00
-4.33446825e-01 -1.00361995e-01 1.27012759e-01 3.02224934e-01
-2.06909493e-01 -1.52107984e-01 -3.58500630e-01 -4.61567968e-01
1.43221781e-01 -1.06699753e+00 -3.71853948e-01 -5.39042652e-01
-6.68944836e-01 -5.85429728e-01 3.99358869e-01 -1.43628046e-01
1.22856951e+00 -1.83225238e+00 2.57163674e-01 4.25148845e-01
1.92188814e-01 1.78603679e-01 -5.32284498e-01 6.54126585e-01
-6.29501462e-01 4.76361364e-01 -4.43665057e-01 -3.07112802e-02
4.04996485e-01 5.10743380e-01 -3.21665436e-01 1.21625490e-01
6.40519381e-01 1.20985115e+00 -8.26192081e-01 -9.16353464e-02
4.41983901e-03 -1.44671038e-01 -1.00273812e+00 -1.34829506e-01
-7.52436876e-01 4.88146544e-02 1.24193333e-01 2.34284103e-01
4.32836264e-01 -6.97319031e-01 6.72134459e-01 -2.33514726e-01
3.29146564e-01 6.09843016e-01 -1.02732980e+00 1.34433377e+00
-3.08295548e-01 3.84048104e-01 -5.71549535e-01 -1.21441126e+00
9.43912506e-01 -1.74862280e-01 5.55058122e-01 -7.66272902e-01
-1.37054905e-01 1.54987678e-01 9.60315526e-01 -4.65627134e-01
5.57024539e-01 -6.30909801e-02 -3.20313722e-01 5.14773846e-01
-3.37783210e-02 -1.00657277e-01 7.09776938e-01 5.92320226e-03
1.34803391e+00 -1.54104382e-01 4.52811539e-01 -5.64123988e-01
7.43213475e-01 -1.78650051e-01 4.43195373e-01 1.08607924e+00
4.23080772e-01 2.64149785e-01 9.39089417e-01 -6.89842582e-01
-1.17717469e+00 -1.02554643e+00 -4.96518845e-03 1.44758570e+00
-2.53807288e-02 -3.64895701e-01 -2.99287736e-01 -5.60829163e-01
2.11682662e-01 9.59706366e-01 -5.69496036e-01 -3.84575665e-01
-9.96860981e-01 -6.87462926e-01 4.86345053e-01 5.39650261e-01
2.28987709e-01 -1.24431717e+00 -1.32655635e-01 2.82843798e-01
-5.85519825e-04 -1.04811227e+00 -7.75125504e-01 5.42841434e-01
-5.64872980e-01 -1.22810400e+00 2.84158200e-01 -1.23925900e+00
6.67768419e-01 -3.07433516e-01 1.48346925e+00 -1.11892000e-01
-1.91318288e-01 1.91214904e-02 -7.62034133e-02 -4.95828599e-01
-4.62701738e-01 3.76757532e-01 1.14920950e-02 -6.58133999e-02
4.33155954e-01 -6.98431730e-01 -1.81647390e-01 4.64051366e-01
-1.10967517e+00 2.64939480e-02 7.90198565e-01 1.10590100e+00
6.55815065e-01 1.99284747e-01 5.29518187e-01 -1.20202780e+00
8.76304865e-01 -2.54975498e-01 -6.40611649e-01 3.47924978e-01
-4.54596668e-01 6.70028627e-01 1.01979315e+00 -5.15557408e-01
-5.64640760e-01 1.64688200e-01 5.71138412e-02 -2.94387817e-01
7.47947097e-02 8.32513928e-01 -4.43114758e-01 3.43711406e-01
4.92971808e-01 4.39956009e-01 2.06205845e-01 1.54844075e-02
5.55343091e-01 6.38185069e-02 5.46302795e-01 -8.38242948e-01
5.63208282e-01 1.05072148e-01 4.93848264e-01 -5.18786550e-01
-1.04815781e+00 -3.65456134e-01 -8.77070725e-01 4.10133541e-01
4.30356324e-01 -3.18061262e-01 -1.27080286e+00 9.96465459e-02
-1.25686395e+00 -1.94291741e-01 -4.46787745e-01 3.82528082e-02
-9.32493210e-01 3.34202737e-01 -4.52156842e-01 -3.10642302e-01
-2.18616396e-01 -1.23858213e+00 8.46927166e-01 -3.32276195e-01
-5.29637516e-01 -9.12099719e-01 -9.33642238e-02 1.98607087e-01
-4.51000892e-02 -1.57801673e-01 1.89526772e+00 -1.28636980e+00
-9.08269584e-01 1.79547921e-01 -5.88519089e-02 3.37166339e-01
6.54123649e-02 -2.60490090e-01 -3.44350785e-01 -1.56022817e-01
-2.01809824e-01 -2.40948245e-01 1.02327812e+00 3.14922333e-01
1.84332228e+00 -8.24871719e-01 -3.10581088e-01 4.64524925e-01
1.30677652e+00 4.99842435e-01 7.12159157e-01 2.94393718e-01
6.02238536e-01 7.54520059e-01 3.25883269e-01 1.97932586e-01
1.58298954e-01 5.46400130e-01 5.84399521e-01 3.29252481e-01
1.09269179e-01 -2.94531405e-01 -4.17438187e-02 8.12693536e-01
2.09800661e-01 -6.25348091e-01 -9.41699684e-01 2.64372110e-01
-1.77408731e+00 -1.13122320e+00 1.35394529e-01 2.13943434e+00
1.02526712e+00 4.47834343e-01 -1.72963217e-01 2.52200246e-01
5.98269820e-01 2.30918273e-01 -5.41066945e-01 -6.40561461e-01
-1.66565135e-01 4.12601739e-01 7.06881642e-01 3.83448392e-01
-1.19458318e+00 8.63668144e-01 6.29125309e+00 9.44790781e-01
-4.86052483e-01 -3.91342580e-01 9.06357169e-01 -1.09357238e-01
-6.75820827e-01 -2.02974617e-01 -9.58744407e-01 2.65124261e-01
9.61155891e-01 1.14446923e-01 6.93879485e-01 4.24775779e-01
-2.12518275e-01 3.13645959e-01 -1.96869707e+00 8.87500465e-01
1.70099631e-01 -1.85997319e+00 5.10813177e-01 1.71281934e-01
6.23479128e-01 -2.56059170e-01 3.62540036e-01 4.75814790e-01
5.38746297e-01 -1.65144181e+00 4.96592730e-01 5.62785268e-01
7.13693798e-01 -9.65762496e-01 4.47756588e-01 3.41621578e-01
-1.10842645e+00 -5.11134148e-01 -2.29383275e-01 -2.71863252e-01
5.71538843e-02 1.51196688e-01 -1.26454163e+00 3.71503472e-01
4.80676413e-01 8.48799765e-01 -3.53270769e-01 6.34276271e-01
-7.38385767e-02 2.78954625e-01 -2.87964821e-01 -5.93684196e-01
2.73055553e-01 -2.68559188e-01 5.68399370e-01 9.88899112e-01
3.98816681e-03 1.99689284e-01 1.21348940e-01 9.22555506e-01
-5.31577945e-01 -1.76558495e-01 -7.56785095e-01 -2.56690085e-01
6.02735162e-01 7.86090910e-01 -6.44890606e-01 -1.00193240e-01
-1.07511580e-01 3.77037317e-01 6.51116312e-01 4.00365070e-02
-8.66141379e-01 -3.21881294e-01 7.82875955e-01 3.05262245e-02
7.26860046e-01 -2.38118276e-01 -5.22587180e-01 -4.86113876e-01
5.55370897e-02 -1.14853871e+00 5.03633380e-01 -5.30555725e-01
-1.57887471e+00 4.40518320e-01 2.82969058e-01 -1.08514285e+00
-2.28596374e-01 -9.95510101e-01 -2.32803285e-01 7.14287400e-01
-1.09182680e+00 -1.03673065e+00 4.50622648e-01 1.19099438e-01
6.28801644e-01 -3.93249512e-01 7.35423028e-01 1.42363966e-01
-3.40821594e-01 7.66386330e-01 7.40332380e-02 3.17982808e-02
1.51093245e-01 -1.34748423e+00 5.58487475e-01 4.45110083e-01
4.79750097e-01 9.27860737e-01 5.61842620e-01 -3.73065531e-01
-1.79671299e+00 -1.21970642e+00 9.92308497e-01 -6.42046392e-01
7.02269733e-01 -7.00843036e-01 -7.95990050e-01 9.21604514e-01
1.51047111e-01 -9.60265025e-02 8.70004952e-01 5.44436812e-01
-6.12124920e-01 -2.27093205e-01 -6.53260827e-01 8.97559524e-01
1.68932343e+00 -4.56680447e-01 -7.80519485e-01 5.31607211e-01
1.10555208e+00 -4.27345067e-01 -9.10859108e-01 7.10710049e-01
4.88707066e-01 -6.42601073e-01 1.33039045e+00 -1.27783227e+00
9.52432871e-01 -2.10391894e-01 -2.73469448e-01 -1.28160203e+00
-8.54531407e-01 -6.19279742e-01 -1.60903484e-01 7.69443810e-01
9.76770759e-01 -6.35335445e-01 9.60512877e-01 2.89760917e-01
-4.15179610e-01 -1.14843369e+00 -7.76704311e-01 -9.07485843e-01
-7.58316964e-02 -5.13797224e-01 8.85139942e-01 9.02724564e-01
-9.99985561e-02 9.66183722e-01 -1.78749427e-01 2.36343414e-01
2.24049166e-01 3.40543598e-01 4.87593949e-01 -1.17192984e+00
-7.90590048e-01 -8.87733102e-01 -2.43586168e-01 -1.25135565e+00
4.04462963e-01 -1.56299675e+00 9.75465178e-02 -1.42215347e+00
1.31994084e-01 -4.69316363e-01 -3.91804963e-01 5.12480795e-01
1.11250579e-01 -1.27701253e-01 2.15832770e-01 1.22037642e-01
-6.26006901e-01 4.68565315e-01 1.39310491e+00 -6.08029246e-01
-2.03142509e-01 -6.05846122e-02 -9.91181314e-01 4.11613435e-01
7.59011209e-01 -4.42487419e-01 -7.35846341e-01 -4.12211269e-01
6.41191185e-01 -1.03626102e-01 1.39857277e-01 -6.51028037e-01
2.42484510e-01 -3.29209685e-01 1.60065323e-01 -7.41183400e-01
4.57580090e-01 -8.40779245e-01 2.29487613e-01 6.64560020e-01
-9.45998192e-01 2.77706653e-01 3.60579103e-01 5.90073586e-01
-3.41707505e-02 -3.44544709e-01 4.43881094e-01 -2.49395266e-01
-7.62630224e-01 3.64369422e-01 -8.65280032e-02 1.38230324e-01
9.25269604e-01 -1.96308658e-01 -4.69507456e-01 -9.45646390e-02
-6.41099393e-01 3.40329826e-01 -1.02353297e-01 5.33420563e-01
5.57158709e-01 -1.10397494e+00 -6.41283691e-01 3.79831642e-01
1.71177670e-01 1.86474085e-01 8.74884278e-02 4.99394864e-01
-4.96145278e-01 8.72359574e-01 -2.04201296e-01 -7.21894562e-01
-1.34659481e+00 7.02447176e-01 1.69902310e-01 -7.75157273e-01
-6.72825351e-02 1.14064705e+00 2.72901207e-01 -8.52920473e-01
7.39856586e-02 -7.72360146e-01 -3.41147929e-01 -1.45858809e-01
3.00812513e-01 -1.75022762e-02 1.19654246e-01 -2.63797104e-01
-2.50266045e-01 2.25399211e-01 -4.68184561e-01 2.42051899e-01
1.42419672e+00 4.12615925e-01 -6.54959500e-01 3.77744675e-01
1.35956502e+00 -3.62232983e-01 -8.03468823e-01 -3.76652271e-01
4.53945130e-01 -2.27747470e-01 -8.06887746e-01 -6.67293549e-01
-4.97341514e-01 4.00798649e-01 -9.10943449e-02 3.68866235e-01
9.85810280e-01 3.02554667e-01 4.16625321e-01 1.06553936e+00
1.31537303e-01 -8.38991165e-01 4.72685337e-01 1.22821772e+00
1.18345404e+00 -8.96887064e-01 3.99142914e-02 -8.55812252e-01
-4.88604367e-01 9.18389738e-01 6.81736708e-01 -1.84896857e-01
7.21471190e-01 2.28136912e-01 -8.04275036e-01 -1.90965295e-01
-1.29563665e+00 -1.29264504e-01 6.76597655e-01 4.45462525e-01
2.98096448e-01 -3.66296731e-02 -2.95461863e-01 4.15523469e-01
-5.42818904e-01 -3.75413567e-01 3.00328612e-01 7.04338789e-01
-2.87819475e-01 -1.38167942e+00 2.29732394e-02 1.16505885e+00
-1.53964713e-01 -5.05139887e-01 -2.90323824e-01 8.66952658e-01
1.90525472e-01 4.10674006e-01 5.54937065e-01 -3.82769704e-01
3.05699468e-01 3.77088711e-02 6.11327529e-01 -1.06338680e+00
-6.45468771e-01 -4.00621206e-01 4.75572646e-01 -3.24114084e-01
-2.45309576e-01 -8.31153870e-01 -1.13629961e+00 -5.19972369e-02
-8.73982459e-02 -1.00040719e-01 2.11872533e-01 8.97986770e-01
4.95909542e-01 8.21678519e-01 4.77971017e-01 -4.30589169e-01
-9.55641329e-01 -7.49459922e-01 -4.80336070e-01 6.10636294e-01
7.19906837e-02 -4.09819067e-01 2.59315651e-02 -6.41315207e-02] | [9.402661323547363, 7.218233108520508] |
ed4081eb-3399-44f0-ac4e-e109ba274f0b | towards-content-independent-multi-reference | null | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/4865_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123700052.pdf | Towards Content-Independent Multi-Reference Super-Resolution: Adaptive Pattern Matching and Feature Aggregation | Recovering realistic textures from a largely down-sampled low resolution (LR) image with complicated patterns is a challenging problem in image super-resolution. This work investigates a novel multi-reference based super-resolution problem by proposing a Content Independent Multi-Reference Super-Resolution (CIMR-SR) model, which is able to adaptively match the visual pattern between references and target image in the low resolution and enhance the feature representation of the target image in the higher resolution. CIMR-SR significantly improves the flexibility of the recently proposed reference-based super-resolution (RefSR), which needs to select the specific high-resolution reference (e.g., content similarity, camera view and relative scale) for each target image. In practice, a universal reference pool (RP) is built up for recovering all LR targets by searching the local matched patterns. By exploiting feature-based patch searching and attentive reference feature aggregation, the proposed CIMR-SR generates realistic images with much better perceptual quality and richer fine-details. Extensive experiments demonstrate the proposed CIMR-SR outperforms state-of-the-art methods in both qualitative and quantitative reconstructions. | ['Shuguang Cui', 'Kun Yuan', 'Xu Yan', 'Weibing Zhao', 'Ruimao Zhang', 'Zhen Li'] | null | null | null | null | eccv-2020-8 | ['reference-based-super-resolution'] | ['computer-vision'] | [ 7.51375377e-01 -3.87013048e-01 -5.49643673e-02 -8.60423446e-02
-1.44207537e+00 -1.90597802e-01 3.56688887e-01 -3.74420702e-01
1.39574274e-01 7.44820297e-01 3.61382872e-01 5.43895721e-01
-5.33908963e-01 -8.74756336e-01 -5.72040617e-01 -9.76294816e-01
4.08212304e-01 1.42442107e-01 7.64835596e-01 -5.76455414e-01
4.73116666e-01 7.37456203e-01 -1.91373622e+00 6.42591417e-01
9.47494924e-01 9.13448393e-01 9.62668717e-01 5.01818717e-01
2.32955784e-01 6.53066814e-01 -1.23978905e-01 1.45638257e-01
2.65256792e-01 -5.43963432e-01 -6.40163839e-01 9.92333367e-02
7.68853009e-01 -1.41036987e-01 -1.62682027e-01 1.23997593e+00
5.13591766e-01 2.99582541e-01 2.77077585e-01 -1.48646131e-01
-9.32334840e-01 2.82995969e-01 -1.14297736e+00 7.47523725e-01
6.22529566e-01 -2.55890429e-01 5.93790948e-01 -1.28858745e+00
9.68794644e-01 1.54562521e+00 4.16541994e-01 3.37608516e-01
-1.52574492e+00 -5.67896068e-01 -2.50298381e-02 3.06619823e-01
-1.70042503e+00 -5.49365401e-01 1.05347741e+00 -1.36523061e-02
2.91553319e-01 6.13488913e-01 7.63249770e-02 7.84311652e-01
3.24128598e-01 -1.64715555e-02 1.70413697e+00 -4.25259829e-01
8.26077983e-02 3.26461121e-02 -1.95320025e-01 3.02584708e-01
1.50389299e-01 4.78124350e-01 -7.92469382e-01 -7.78415427e-02
1.65213609e+00 -1.54337035e-02 -6.82896674e-01 -2.70474732e-01
-1.35142350e+00 3.92261058e-01 3.48592252e-01 5.92499197e-01
-4.75393444e-01 -5.11095703e-01 -2.12740451e-01 7.08740205e-02
5.43387830e-01 3.77287686e-01 -1.31034285e-01 5.35897791e-01
-9.02756870e-01 1.30120561e-01 -3.19901891e-02 7.88332462e-01
8.99079382e-01 1.18257396e-01 -2.99030066e-01 1.21868360e+00
-1.05236694e-01 5.08058071e-01 3.74530405e-01 -1.31498611e+00
1.72965273e-01 2.04639196e-01 4.97874349e-01 -1.16466546e+00
-2.21120771e-02 -6.66036427e-01 -1.21280611e+00 3.11051309e-01
8.93302634e-02 7.21950412e-01 -6.13058090e-01 1.33448875e+00
7.47481763e-01 2.13237748e-01 2.11603642e-01 1.30457866e+00
9.53812897e-01 6.26168668e-01 -2.87145495e-01 -7.41084635e-01
1.25861061e+00 -6.74481392e-01 -7.73513317e-01 -1.47684410e-01
-4.97691900e-01 -1.06223857e+00 9.22013044e-01 4.08427507e-01
-1.44230032e+00 -1.17402101e+00 -9.13121223e-01 -1.84233040e-01
2.71596462e-01 4.16194648e-02 -7.18997642e-02 1.60897195e-01
-1.04570317e+00 6.20999932e-01 -2.52612859e-01 -1.20785810e-01
1.51657075e-01 2.06457591e-03 -5.87079525e-01 -8.24906111e-01
-1.08611405e+00 8.64278972e-01 1.47901163e-01 8.16447511e-02
-7.70695150e-01 -9.03026760e-01 -7.80035019e-01 -1.48457125e-01
3.39377999e-01 -6.03691041e-01 4.86108452e-01 -7.68861532e-01
-1.54684365e+00 8.50499988e-01 -4.76767659e-01 8.30759630e-02
7.68970251e-02 2.27270752e-01 -7.79420733e-01 5.74015975e-01
1.46640167e-01 1.28475726e-01 1.20552039e+00 -1.95648956e+00
-6.25675619e-01 -4.55943167e-01 -2.64801711e-01 4.95950490e-01
4.41682309e-01 2.94989318e-01 -1.63318872e-01 -8.35451186e-01
5.62063217e-01 -2.24196270e-01 -2.49040961e-01 -2.07092717e-01
8.23300704e-02 3.32270384e-01 7.46385038e-01 -8.82473528e-01
1.05657279e+00 -2.13001752e+00 4.41129714e-01 -6.35627285e-02
2.86201388e-01 7.51301646e-02 -2.96305239e-01 3.20103168e-02
-1.56204015e-01 -2.66383260e-01 -7.98213705e-02 7.47779235e-02
-7.58941114e-01 -1.39637655e-02 -2.08158255e-01 7.02345669e-01
1.03921287e-01 5.22570133e-01 -8.23070407e-01 -7.96921730e-01
4.58435059e-01 1.08225787e+00 -1.99982718e-01 2.72026271e-01
9.53668952e-02 1.14155185e+00 -5.10503352e-01 7.58583903e-01
1.18558514e+00 -2.58789033e-01 -1.80967841e-02 -7.54282713e-01
-5.31890035e-01 -3.74828309e-01 -1.58191681e+00 1.58094978e+00
-6.67647362e-01 1.85594529e-01 3.90258282e-01 -4.67415154e-01
1.38027787e+00 3.47696356e-02 3.42725545e-01 -1.10453212e+00
-3.75550240e-01 3.27545762e-01 -4.17125344e-01 -2.49322370e-01
6.95238888e-01 -4.02913332e-01 3.31275582e-01 -1.04077838e-01
-2.80950308e-01 8.29550326e-02 -2.28472114e-01 -1.90733731e-01
6.80080354e-01 1.69449642e-01 5.32658875e-01 -3.35803300e-01
1.07312238e+00 -2.80586272e-01 6.35198355e-01 7.16304064e-01
2.14879975e-01 1.19144523e+00 -4.58160788e-01 -3.63243461e-01
-1.34692669e+00 -1.26457846e+00 -5.21772265e-01 8.82747650e-01
6.50818288e-01 1.77651778e-01 -3.27409685e-01 1.90848172e-01
-4.08614427e-01 4.03547049e-01 -7.18424439e-01 1.86000407e-01
-8.56406987e-01 -5.04922628e-01 -4.12837595e-01 4.00770083e-02
7.39693224e-01 -9.58824813e-01 -5.10131299e-01 1.91105038e-01
-7.31082022e-01 -1.16099620e+00 -5.02715886e-01 -2.87030637e-01
-8.30445588e-01 -8.47912729e-01 -9.56281960e-01 -7.58639514e-01
5.61052620e-01 8.77384782e-01 1.23718905e+00 -1.02048934e-01
-3.16970766e-01 2.54480511e-01 -5.21845579e-01 6.54335201e-01
-4.12849516e-01 -5.04357040e-01 -2.11791009e-01 5.71130514e-01
-3.69711399e-01 -7.04280853e-01 -8.44949484e-01 7.19835341e-01
-8.88253748e-01 3.23644400e-01 9.15732503e-01 1.01696336e+00
1.48296142e+00 4.86510187e-01 4.33224201e-01 -6.12637341e-01
2.02128202e-01 -3.01190197e-01 -5.47339439e-01 4.05656248e-01
-3.83608162e-01 -6.00671135e-02 6.69721603e-01 -6.17394269e-01
-1.63044035e+00 -1.82576686e-01 1.65254369e-01 -6.73095584e-01
-1.88505515e-01 -2.35682707e-02 -3.53241801e-01 -6.53728545e-01
6.87696159e-01 7.71779239e-01 -2.53330559e-01 -9.08297718e-01
2.56650716e-01 4.98361439e-01 9.61263597e-01 -5.57209551e-01
9.71867561e-01 7.92795837e-01 2.42619872e-01 -1.01275992e+00
-1.13262904e+00 -5.86911023e-01 -7.51700878e-01 -1.77040532e-01
7.49383986e-01 -1.07808268e+00 -2.95831084e-01 1.42466918e-01
-8.76571715e-01 -4.54613753e-02 -3.79414231e-01 3.89942199e-01
-6.67400718e-01 4.93343204e-01 -4.72488403e-01 -7.51621842e-01
-2.70743936e-01 -1.10212874e+00 1.38309228e+00 4.86091882e-01
3.55339408e-01 -5.94934762e-01 5.76974861e-02 4.95307922e-01
8.69259655e-01 4.72020626e-01 5.89476705e-01 2.64750123e-01
-9.10025239e-01 3.77025843e-01 -6.08733475e-01 1.53184950e-01
2.53729284e-01 -4.58439738e-01 -6.33345604e-01 -4.39463168e-01
3.59967083e-01 1.62682438e-03 6.10205650e-01 5.63949704e-01
1.06089103e+00 -2.55142540e-01 -5.76877818e-02 8.47825348e-01
2.10956073e+00 -1.02549225e-01 1.03381813e+00 4.95672286e-01
6.21755838e-01 5.56891263e-01 1.23466873e+00 3.12658370e-01
1.46941677e-01 1.27524042e+00 2.19835237e-01 -2.89626747e-01
-7.17138886e-01 -8.45604688e-02 1.17936961e-01 4.20733750e-01
-4.06614035e-01 3.14309090e-01 -2.19394043e-01 4.12616938e-01
-1.54245830e+00 -1.29183054e+00 -1.93320334e-01 2.41782379e+00
9.90050316e-01 -4.11819994e-01 -2.29591012e-01 4.77788597e-02
1.16688466e+00 2.41395295e-01 -5.95314741e-01 3.14083137e-02
-6.78167760e-01 4.00697917e-01 3.06721151e-01 7.78436244e-01
-7.46128678e-01 7.72837698e-01 5.77539253e+00 1.41657555e+00
-9.34350669e-01 3.53417009e-01 6.00926399e-01 1.69077933e-01
-4.09277767e-01 -2.08390936e-01 -9.02288377e-01 4.15846974e-01
6.85043812e-01 -1.93334952e-01 7.85511971e-01 4.55558032e-01
4.17669713e-01 -3.50538462e-01 -5.62680781e-01 1.35217214e+00
2.86005914e-01 -1.38651931e+00 3.61254007e-01 -1.57914311e-01
8.75903964e-01 -3.76599133e-01 2.07507238e-01 -2.81248301e-01
-8.85397345e-02 -1.11889088e+00 5.90430975e-01 9.75683391e-01
1.37574852e+00 -8.64449501e-01 4.40722167e-01 1.05424352e-01
-1.57224607e+00 -1.03013478e-01 -7.30987906e-01 5.55320680e-01
1.84678972e-01 5.60858548e-01 2.37799987e-01 1.06185377e+00
9.98057008e-01 7.42198110e-01 -6.75710797e-01 7.42558062e-01
2.11324468e-01 -1.05937727e-01 8.13742802e-02 9.44916308e-01
-4.21887130e-01 -4.12131548e-01 8.39218736e-01 8.46684813e-01
4.49858367e-01 7.09404230e-01 1.29029274e-01 9.75051522e-01
4.13337886e-01 1.17593177e-01 -1.68582007e-01 7.38456249e-01
5.58920503e-01 1.48308158e+00 -5.24543166e-01 -1.15235187e-01
-1.86876133e-01 9.39748526e-01 -2.18024720e-02 5.24838746e-01
-5.36695719e-01 -6.35571126e-03 2.94906318e-01 6.31052792e-01
5.91303289e-01 1.65220592e-02 -9.12191644e-02 -1.15435600e+00
3.51955853e-02 -1.09772170e+00 2.99869716e-01 -1.32056272e+00
-1.21312726e+00 9.72258210e-01 4.84795533e-02 -1.53389823e+00
5.46576753e-02 1.12418458e-01 -1.61960647e-01 1.35148799e+00
-1.90171385e+00 -1.37905264e+00 -6.82802677e-01 9.34868217e-01
6.85142994e-01 1.70433819e-01 5.71672201e-01 1.99533284e-01
-2.00610682e-01 1.99479893e-01 2.20840648e-01 -5.46744764e-01
7.11530268e-01 -7.39229620e-01 -1.91140637e-01 9.91989136e-01
-2.79366553e-01 3.99841219e-01 9.11500216e-01 -6.55438602e-01
-1.33996749e+00 -1.09328020e+00 3.46297830e-01 -2.71989763e-01
1.12049155e-01 1.71896875e-01 -1.27449489e+00 1.57737166e-01
-2.22084209e-01 3.64118904e-01 2.90336192e-01 -3.98345023e-01
-4.27965909e-01 -4.51429188e-01 -1.66254139e+00 3.50414693e-01
9.80035305e-01 -4.56322104e-01 -4.91989076e-01 -1.03732884e-01
7.75434375e-01 -5.87899745e-01 -1.35610116e+00 6.42180204e-01
5.00129819e-01 -1.35007954e+00 1.62819076e+00 2.54858941e-01
5.33776104e-01 -8.11372876e-01 -5.65573514e-01 -9.88278151e-01
-8.94954920e-01 -4.63382542e-01 -3.55179533e-02 1.21010578e+00
-2.58096606e-01 -5.53025246e-01 1.33958772e-01 6.50110692e-02
2.14317013e-02 -6.41727746e-01 -9.65537488e-01 -6.58839881e-01
-2.28346094e-01 4.30688858e-01 5.99517584e-01 8.98701429e-01
-7.24538803e-01 1.01868272e-01 -5.32332420e-01 7.65724480e-01
1.49601924e+00 6.83570683e-01 3.32394153e-01 -1.27625120e+00
-3.44341129e-01 -1.33924857e-02 -1.68843120e-01 -6.98295295e-01
-2.51375228e-01 -4.54756469e-01 -1.07775986e-01 -1.54742956e+00
5.40750802e-01 -4.36550826e-01 -4.12826121e-01 -2.23378450e-01
-2.07548663e-01 7.10838974e-01 -5.83327413e-02 6.59091413e-01
-7.50777662e-01 4.04780269e-01 1.88265598e+00 1.93381295e-01
-2.16225922e-01 -1.87387198e-01 -8.83397341e-01 3.73634249e-01
4.10302132e-01 -1.73926026e-01 -9.87786874e-02 -1.31093383e-01
-7.15211630e-02 5.87769747e-01 5.87702513e-01 -9.57562864e-01
5.21034524e-02 -4.29239988e-01 7.59105265e-01 -7.59171784e-01
6.11040115e-01 -6.20663881e-01 6.65029347e-01 -8.71778876e-02
-4.32188034e-01 -3.48091513e-01 -1.22656509e-01 7.97783852e-01
-3.47519875e-01 2.04070002e-01 1.61196971e+00 -4.23479676e-01
-8.06021094e-01 2.96973825e-01 3.17254066e-01 -4.47919853e-02
7.40702331e-01 -5.67888319e-01 -5.01293838e-01 2.35313382e-02
-7.01319575e-01 -3.31061453e-01 9.84860837e-01 2.50884414e-01
1.24862063e+00 -1.35944557e+00 -1.10750139e+00 2.35177889e-01
1.29681617e-01 3.74781862e-02 1.06583977e+00 8.17567170e-01
-1.07776381e-01 1.01682588e-01 -5.57048976e-01 -6.22679353e-01
-1.53871500e+00 8.31709623e-01 4.17491257e-01 -4.17355746e-01
-1.05605435e+00 5.65851390e-01 5.52147269e-01 1.34095445e-01
-2.53876746e-01 1.12327479e-01 -6.30587161e-01 -4.70396787e-01
1.20548737e+00 5.12881994e-01 -1.42257735e-01 -1.18183541e+00
-1.74054131e-01 1.37510812e+00 -7.31496736e-02 3.60664278e-02
1.58530354e+00 -8.93874049e-01 -3.02895337e-01 3.19265485e-01
8.43006134e-01 1.80583626e-01 -1.37575066e+00 -6.79449439e-01
-4.43347484e-01 -1.24455416e+00 3.96679640e-01 -7.00832069e-01
-1.12080097e+00 4.06064421e-01 8.83806944e-01 -2.07367361e-01
1.65192544e+00 1.88770652e-01 4.57288742e-01 -4.17342901e-01
9.24732924e-01 -6.56707346e-01 1.49439633e-01 -3.19329761e-02
1.37868619e+00 -1.11446977e+00 3.44963700e-01 -7.56678462e-01
-4.23123360e-01 1.00057960e+00 4.16281968e-01 -2.72646129e-01
2.64917493e-01 1.17515482e-01 -3.16834539e-01 -1.40086889e-01
-6.33503318e-01 -2.37084672e-01 3.44555348e-01 9.20088410e-01
1.06137976e-01 -2.34338328e-01 -2.04709575e-01 4.23257858e-01
1.97260082e-01 -7.31434077e-02 6.65604472e-01 2.62226224e-01
-5.97104847e-01 -8.44939232e-01 -9.62128997e-01 1.37469202e-01
-4.06474978e-01 -1.57530352e-01 2.32580572e-01 3.83461535e-01
9.96957198e-02 1.04076183e+00 -1.07415155e-01 -1.90038025e-01
3.79390568e-01 -6.89294279e-01 8.89769852e-01 -4.78627324e-01
-1.89147249e-01 4.13329870e-01 -2.80017942e-01 -9.80549872e-01
-8.43634069e-01 -6.51680350e-01 -7.63279676e-01 -3.69075686e-02
-2.09309340e-01 -3.79512087e-02 2.36184657e-01 4.87484097e-01
3.70377392e-01 4.56331968e-01 9.34460104e-01 -1.32237077e+00
-3.01938802e-01 -7.93045759e-01 -1.05066597e+00 3.33757550e-01
5.84670603e-01 -8.22214365e-01 -6.03060782e-01 6.28882721e-02] | [10.990273475646973, -2.1004343032836914] |
b65de6ea-fc55-490b-9e71-e1aa1f81824c | mtvr-multilingual-moment-retrieval-in-videos | 2108.00061 | null | https://arxiv.org/abs/2108.00061v1 | https://arxiv.org/pdf/2108.00061v1.pdf | MTVR: Multilingual Moment Retrieval in Videos | We introduce mTVR, a large-scale multilingual video moment retrieval dataset, containing 218K English and Chinese queries from 21.8K TV show video clips. The dataset is collected by extending the popular TVR dataset (in English) with paired Chinese queries and subtitles. Compared to existing moment retrieval datasets, mTVR is multilingual, larger, and comes with diverse annotations. We further propose mXML, a multilingual moment retrieval model that learns and operates on data from both languages, via encoder parameter sharing and language neighborhood constraints. We demonstrate the effectiveness of mXML on the newly collected MTVR dataset, where mXML outperforms strong monolingual baselines while using fewer parameters. In addition, we also provide detailed dataset analyses and model ablations. Data and code are publicly available at https://github.com/jayleicn/mTVRetrieval | ['Mohit Bansal', 'Tamara L. Berg', 'Jie Lei'] | 2021-07-30 | null | https://aclanthology.org/2021.acl-short.92 | https://aclanthology.org/2021.acl-short.92.pdf | acl-2021-5 | ['moment-retrieval'] | ['computer-vision'] | [-4.35947984e-01 -4.86355573e-01 -9.30285037e-01 -3.61202359e-01
-1.85845673e+00 -1.04675627e+00 8.54047418e-01 -2.49958858e-01
-7.33389854e-01 5.83928764e-01 7.06427038e-01 -8.37356299e-02
-1.16699830e-01 -1.19623449e-02 -9.84657645e-01 -1.80094048e-01
-8.12919363e-02 4.57756668e-01 2.94820382e-03 -2.40956098e-01
-5.03421165e-02 -3.60780652e-03 -1.15664721e+00 5.30557394e-01
5.26897013e-01 9.55478251e-01 3.77358377e-01 8.71607482e-01
4.17477757e-01 1.26705432e+00 -2.55792022e-01 -8.19303930e-01
1.59599110e-01 2.07491871e-03 -8.87273908e-01 -3.86946619e-01
1.02165127e+00 -7.07446337e-01 -1.31394958e+00 5.53781152e-01
6.94376826e-01 2.56267697e-01 3.59085321e-01 -1.05531263e+00
-1.19823241e+00 1.08053505e+00 -3.82977903e-01 4.04599607e-01
7.00138509e-01 -2.18030840e-01 1.46739912e+00 -1.12164354e+00
1.24367952e+00 1.14982390e+00 3.73940259e-01 4.93270457e-01
-7.33138800e-01 -8.35743546e-01 -7.58550456e-03 4.59777832e-01
-1.98464680e+00 -7.98179328e-01 4.27152634e-01 -1.71756864e-01
1.04751468e+00 4.36278999e-01 5.15661240e-01 1.70980239e+00
-1.41479984e-01 1.58332109e+00 5.01842022e-01 -6.63346946e-02
-5.56819856e-01 -7.33791525e-03 -1.32085040e-01 6.16068542e-01
-5.55851281e-01 -1.99021101e-01 -9.50531006e-01 -2.07066797e-02
5.37693977e-01 -1.97648361e-01 -3.46859723e-01 -1.81294233e-01
-1.47864139e+00 5.98825514e-01 8.48231167e-02 2.25820452e-01
8.48875642e-02 3.64455700e-01 6.94178820e-01 5.99565446e-01
6.05179071e-01 1.64381772e-01 -6.71323657e-01 -6.02413893e-01
-9.50430095e-01 2.31071308e-01 5.88658512e-01 1.65769899e+00
4.88974303e-01 -2.77820438e-01 -1.94514453e-01 1.26221836e+00
1.32649273e-01 1.16233444e+00 4.47856337e-01 -1.16705966e+00
7.42839515e-01 -1.27892364e-02 -2.22358406e-01 -6.87916517e-01
-3.97180067e-03 -5.87726608e-02 -1.03652701e-01 -1.16260421e+00
3.87333892e-02 1.13702035e-02 -6.86141670e-01 1.67525947e+00
-1.37915716e-01 1.73474550e-01 1.98311776e-01 1.07540965e+00
1.23071885e+00 9.50833321e-01 -7.60487318e-02 9.48857963e-02
1.06751680e+00 -1.22244394e+00 -7.92598486e-01 1.09815598e-03
9.16777074e-01 -9.95758414e-01 1.29157805e+00 3.26195151e-01
-1.07905865e+00 -2.47069389e-01 -6.43897951e-01 -9.91238356e-01
-3.49474818e-01 7.20248759e-01 6.99066281e-01 -2.08155494e-02
-1.18519747e+00 -7.66382143e-02 -8.87736678e-01 -4.89060551e-01
-3.08649708e-02 6.68618083e-02 -6.12796664e-01 -4.23394978e-01
-1.35448790e+00 4.74340439e-01 2.49662161e-01 9.75894406e-02
-1.28627086e+00 -7.51721442e-01 -1.11005259e+00 -3.45466554e-01
5.01363099e-01 -4.38608527e-02 1.71233797e+00 -6.35509789e-01
-1.26311886e+00 1.08863723e+00 -2.72018045e-01 -4.36866879e-01
5.01863480e-01 -6.68020606e-01 -8.69184434e-01 4.48120624e-01
1.20507330e-01 1.05597556e+00 4.72064406e-01 -7.23527610e-01
-5.10073602e-01 1.33843571e-01 2.62052387e-01 5.20044386e-01
-2.82667100e-01 4.63409603e-01 -1.67494476e+00 -8.58865082e-01
-3.12884033e-01 -1.21005249e+00 3.88816297e-01 -4.14440304e-01
-3.56063843e-01 -2.17192799e-01 8.35595906e-01 -1.09678173e+00
1.42265940e+00 -2.40197515e+00 3.66099536e-01 -1.20172828e-01
7.88827538e-02 -1.96278989e-01 -5.72424352e-01 7.40038633e-01
1.87302813e-01 1.31560430e-01 2.93361127e-01 -4.93274748e-01
2.16195837e-01 1.76401570e-01 -5.54872036e-01 4.10167336e-01
-2.74069905e-01 1.26809621e+00 -9.08074677e-01 -7.34309256e-01
4.35109362e-02 6.89766049e-01 -6.07766747e-01 -5.45312203e-02
-1.82136998e-01 3.05183440e-01 -4.57861811e-01 1.14571381e+00
1.10021584e-01 -1.54249117e-01 1.18299514e-01 -4.15250957e-01
1.66404881e-02 2.82088190e-01 -4.30168539e-01 2.56639910e+00
-5.38194418e-01 1.34340894e+00 -5.74994311e-02 -1.98953912e-01
4.70188797e-01 5.41130960e-01 6.92927063e-01 -1.23871517e+00
9.92359500e-03 4.79856968e-01 -5.88213801e-01 -6.19222164e-01
1.28533053e+00 8.04387391e-01 -7.81720221e-01 4.29772317e-01
4.38970983e-01 6.10723458e-02 6.38621688e-01 9.09002960e-01
9.01978731e-01 2.57084519e-01 -2.00785920e-01 -5.74019514e-02
2.44893655e-01 -1.02595858e-01 5.00331759e-01 8.24254990e-01
-9.91342738e-02 5.58834434e-01 2.47467995e-01 -2.32699484e-01
-1.06500435e+00 -1.27826619e+00 -2.87305087e-01 1.50652230e+00
2.11756274e-01 -1.03432369e+00 -3.53874415e-01 -5.54153919e-01
-8.97473991e-02 2.84254223e-01 -3.36838663e-01 1.07295789e-01
-6.99775457e-01 -2.00772196e-01 1.00991690e+00 1.70848235e-01
2.95948207e-01 -8.00704360e-01 3.37266661e-02 -3.29413712e-01
-7.96842396e-01 -1.69335294e+00 -1.10414720e+00 -3.73889387e-01
-3.78466070e-01 -1.00091934e+00 -8.62206757e-01 -8.62775683e-01
1.38237670e-01 3.04844081e-01 1.48178720e+00 -1.69854119e-01
-2.64556468e-01 9.94023502e-01 -8.11001837e-01 1.47963809e-02
-8.29394609e-02 6.96572840e-01 2.37070203e-01 -3.94877642e-01
5.19262850e-01 -1.17174782e-01 -4.86622512e-01 2.75797129e-01
-1.00725675e+00 -3.04752421e-02 4.12265837e-01 4.58250105e-01
9.28948939e-01 -5.23257136e-01 3.01146984e-01 -6.62312031e-01
5.43318748e-01 -6.52329445e-01 -7.03884125e-01 5.95360160e-01
4.75911312e-02 -1.72116563e-01 3.02969608e-02 -5.15515506e-01
-8.63289654e-01 -2.00583652e-01 -4.93249632e-02 -8.86490822e-01
2.36675471e-01 7.63763428e-01 9.92566720e-02 7.11393636e-03
2.94947982e-01 2.65456319e-01 -6.28597736e-01 -5.94507515e-01
6.46848500e-01 8.72891605e-01 7.52864838e-01 -8.34606886e-01
4.38410461e-01 2.75730163e-01 -7.37196684e-01 -8.87228608e-01
-6.24647141e-01 -6.19963765e-01 -5.17161369e-01 -3.73201728e-01
7.84335136e-01 -1.73408246e+00 -4.12178665e-01 2.36908525e-01
-8.05810571e-01 -7.14679003e-01 3.68662626e-02 7.73128688e-01
-4.86759216e-01 1.65809289e-01 -1.30784607e+00 -2.84425080e-01
-1.98954210e-01 -1.16763318e+00 1.59372354e+00 -2.01711059e-01
-9.69720036e-02 -9.77827370e-01 2.25225121e-01 5.93887985e-01
9.43653136e-02 -2.64734119e-01 3.23092014e-01 -6.54911160e-01
-7.49396741e-01 -3.24874490e-01 -2.04085149e-02 -3.29498462e-02
-1.68276072e-01 2.06432536e-01 -6.45600140e-01 -5.91176331e-01
-6.81209564e-01 -1.10026121e+00 9.02904510e-01 2.44912073e-01
1.17519999e+00 -2.65632212e-01 -2.29698941e-01 8.54536474e-01
1.22615588e+00 3.79332267e-02 3.68375987e-01 3.03225726e-01
7.41720021e-01 1.41756579e-01 8.49297106e-01 2.79699653e-01
8.35481584e-01 8.50374997e-01 -1.51297733e-01 6.92612678e-02
-1.89945877e-01 -6.31097972e-01 7.49806166e-01 1.73196363e+00
1.33477196e-01 -5.10465801e-01 -8.87211502e-01 6.71939790e-01
-1.87425041e+00 -9.83692169e-01 3.80279958e-01 1.81191123e+00
8.82507324e-01 -4.30079430e-01 -8.27023759e-02 -7.21636176e-01
1.83049604e-01 5.53920150e-01 -5.03456652e-01 -8.03821459e-02
-7.52504289e-01 -1.88774735e-01 6.54743671e-01 6.60585105e-01
-1.24522424e+00 1.44804537e+00 6.42904139e+00 1.10861504e+00
-1.13010097e+00 4.03744221e-01 3.36689293e-01 -8.74954820e-01
-4.90929365e-01 -1.39747895e-02 -7.65917182e-01 2.22869873e-01
1.01066005e+00 -1.81655973e-01 8.37250233e-01 5.48002720e-01
6.59997994e-03 1.00765862e-01 -1.12109995e+00 1.36070240e+00
5.23203790e-01 -1.16865444e+00 9.44023579e-02 -1.30387411e-01
8.58064413e-01 1.02941287e+00 3.68923753e-01 6.35001600e-01
2.34615013e-01 -6.30478680e-01 1.09194899e+00 5.41119397e-01
1.25189602e+00 -5.81757963e-01 2.93285429e-01 6.69634789e-02
-1.42280889e+00 -3.56308594e-02 -1.97702095e-01 5.83972216e-01
4.08182144e-01 -9.02843773e-02 -3.53731275e-01 6.12869442e-01
1.08167589e+00 1.36693776e+00 -9.08853769e-01 6.92902148e-01
-2.36019552e-01 3.90032649e-01 -3.38970512e-01 3.68074864e-01
3.17145109e-01 -1.21346720e-01 5.61414421e-01 1.44396126e+00
3.37141633e-01 -2.17223242e-01 2.34911606e-01 1.20109051e-01
-6.71122789e-01 3.08200300e-01 -7.95790792e-01 -3.64902526e-01
7.99464583e-01 1.19610834e+00 -1.52169198e-01 -4.47283655e-01
-6.64564967e-01 1.19228768e+00 4.37912464e-01 8.39500785e-01
-1.16544831e+00 -2.50353396e-01 4.67083097e-01 -2.97076404e-01
7.60635212e-02 -2.90023386e-01 9.45003033e-01 -1.97464561e+00
1.04858950e-01 -1.24473155e+00 7.15031564e-01 -1.14817762e+00
-1.06653130e+00 6.07678890e-01 2.73571193e-01 -1.14153886e+00
-5.52229404e-01 -3.67196023e-01 3.95726115e-01 2.55337626e-01
-1.32471633e+00 -1.47301543e+00 1.16005905e-01 9.48554754e-01
9.18798804e-01 -3.10907930e-01 5.27390242e-01 1.09202933e+00
-5.46377838e-01 1.00935364e+00 3.09713781e-01 4.47565317e-01
1.30218601e+00 -8.58998656e-01 1.64844275e-01 7.52942324e-01
6.79785013e-01 6.62903666e-01 1.79648086e-01 -6.65824115e-01
-1.84418738e+00 -9.74734545e-01 1.09008253e+00 -9.30218041e-01
1.04496288e+00 -7.98538685e-01 -4.99965191e-01 1.39037466e+00
7.41576672e-01 -1.26896754e-01 6.68016732e-01 2.06936032e-01
-6.40513301e-01 -9.16887745e-02 -4.24696058e-01 8.09550941e-01
1.13669360e+00 -1.45140219e+00 -2.83170640e-01 6.03893995e-01
1.07422578e+00 -8.11732352e-01 -1.36630893e+00 4.52111214e-01
8.99135888e-01 -3.24303359e-01 9.78552520e-01 -4.35120732e-01
3.43022883e-01 2.14363281e-02 -7.13610590e-01 -9.29624081e-01
3.62530388e-02 -7.18210101e-01 -3.24932128e-01 1.14239097e+00
7.75917113e-01 -7.58479303e-03 3.14686358e-01 3.60408664e-01
-2.26600185e-01 -4.56117749e-01 -9.72091854e-01 -7.78331101e-01
4.29645404e-02 -8.57433677e-01 4.14447844e-01 1.28284705e+00
1.28643483e-01 3.28984141e-01 -8.99372220e-01 1.04879420e-02
2.58033007e-01 -6.76633120e-02 8.36491227e-01 -3.75478745e-01
-3.69085073e-01 -2.23301992e-01 -6.98569193e-02 -1.36134076e+00
5.88392854e-01 -1.12623239e+00 -1.53805137e-01 -1.12658453e+00
5.14436305e-01 -2.42088616e-01 -3.40767652e-01 4.26136494e-01
3.70203942e-01 4.85219032e-01 2.38748074e-01 5.09284019e-01
-1.52160001e+00 4.19482142e-01 1.06670785e+00 -2.54832178e-01
-4.44997139e-02 -6.47893071e-01 -2.16824099e-01 2.78775513e-01
5.46541929e-01 -2.40646273e-01 -6.12522840e-01 -1.45216429e+00
7.57160664e-01 3.34176868e-01 1.04860552e-01 -5.97020805e-01
4.20944691e-01 2.33465075e-01 9.72435772e-02 -8.34277272e-01
6.70437932e-01 -4.74311709e-01 2.40389720e-01 -1.66482866e-01
-7.94581234e-01 6.61767721e-01 2.86683977e-01 3.32081735e-01
-6.04910493e-01 3.45550865e-01 -7.34270960e-02 8.93990919e-02
-1.16240120e+00 7.21718609e-01 -2.62368232e-01 3.07577133e-01
5.67133904e-01 5.42343497e-01 -6.02641165e-01 -8.90492082e-01
-6.36426330e-01 6.08658493e-01 5.03850341e-01 1.21321929e+00
4.75690275e-01 -1.58763754e+00 -7.08428085e-01 -2.37102866e-01
6.35142446e-01 -4.62444156e-01 4.01249409e-01 9.40955460e-01
-6.05248809e-01 9.73721862e-01 2.72277951e-01 -5.50271749e-01
-1.19711912e+00 5.04425704e-01 9.69948247e-02 -1.74876731e-02
-4.58817363e-01 7.48763561e-01 -3.63013595e-02 -7.24525750e-01
3.87502670e-01 -1.89264163e-01 -4.88811471e-02 9.73645821e-02
5.09970844e-01 2.79941946e-01 -1.47785604e-01 -1.07036746e+00
-2.78771162e-01 5.76299965e-01 -4.30122405e-01 -5.98057866e-01
1.04062045e+00 -5.61161816e-01 -9.86469239e-02 5.35533905e-01
1.75095391e+00 4.98497486e-01 -9.42697346e-01 -5.90635896e-01
2.23724507e-02 -4.38669771e-01 9.47952494e-02 -6.32360399e-01
-1.17568409e+00 3.90343010e-01 3.46982032e-01 -5.27373791e-01
1.16851556e+00 5.56797504e-01 1.05085456e+00 9.02539015e-01
4.17445302e-01 -1.39508843e+00 -4.89165410e-02 7.72402346e-01
8.47049534e-01 -1.35587132e+00 -2.76654810e-02 4.06504385e-02
-7.02971816e-01 6.93563700e-01 4.58055615e-01 1.16903171e-01
5.65231383e-01 1.09987795e-01 3.08996409e-01 -1.79272726e-01
-1.16869915e+00 1.12502347e-03 5.85130870e-01 8.52184072e-02
6.29576206e-01 7.39524961e-02 -2.06276149e-01 4.75357860e-01
-4.62743610e-01 -2.69434731e-02 1.97244674e-01 9.53425944e-01
2.48703316e-01 -1.10481656e+00 1.59066558e-01 2.60168344e-01
-8.03395152e-01 -5.10358393e-01 -5.36931396e-01 1.04722106e+00
-3.57159108e-01 7.01811552e-01 1.91432431e-01 -5.09511173e-01
9.61192548e-02 -2.79985994e-01 3.90096039e-01 -1.27955005e-01
-2.16805920e-01 4.40090656e-01 3.45314860e-01 -9.62424576e-01
-4.12517905e-01 -8.23215544e-01 -7.40479171e-01 -3.41873586e-01
3.28745395e-02 3.59403878e-01 5.73592186e-01 6.01311028e-01
5.49505770e-01 7.19936416e-02 4.25055981e-01 -4.76994395e-01
1.15896575e-01 -8.49008441e-01 -3.33714068e-01 2.61187524e-01
3.72244298e-01 -2.16371953e-01 -8.91560689e-02 1.25957400e-01] | [11.074193954467773, 1.4457118511199951] |
ee8b89cc-0d34-4691-ab3b-97c492dca677 | speaker-guided-encoder-decoder-framework-for | 2206.03173 | null | https://arxiv.org/abs/2206.03173v1 | https://arxiv.org/pdf/2206.03173v1.pdf | Speaker-Guided Encoder-Decoder Framework for Emotion Recognition in Conversation | The emotion recognition in conversation (ERC) task aims to predict the emotion label of an utterance in a conversation. Since the dependencies between speakers are complex and dynamic, which consist of intra- and inter-speaker dependencies, the modeling of speaker-specific information is a vital role in ERC. Although existing researchers have proposed various methods of speaker interaction modeling, they cannot explore dynamic intra- and inter-speaker dependencies jointly, leading to the insufficient comprehension of context and further hindering emotion prediction. To this end, we design a novel speaker modeling scheme that explores intra- and inter-speaker dependencies jointly in a dynamic manner. Besides, we propose a Speaker-Guided Encoder-Decoder (SGED) framework for ERC, which fully exploits speaker information for the decoding of emotion. We use different existing methods as the conversational context encoder of our framework, showing the high scalability and flexibility of the proposed framework. Experimental results demonstrate the superiority and effectiveness of SGED. | ['Songlin Hu', 'Wei Zhou', 'Lingwei Wei', 'Qianwen Ma', 'Yinan Bao'] | 2022-06-07 | null | null | null | null | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [ 6.71724379e-02 2.90901195e-02 7.31263608e-02 -9.45144653e-01
-4.61701691e-01 -2.87796289e-01 4.07802135e-01 -1.88944191e-01
-9.60494727e-02 2.83357471e-01 7.63671696e-01 -1.42985180e-01
3.60018045e-01 -2.04499289e-01 -1.29071310e-01 -6.26468480e-01
-1.34530574e-01 2.51275022e-02 -9.50021446e-02 -2.66278476e-01
1.55510411e-01 1.16467662e-01 -1.66327941e+00 5.73741436e-01
7.90241063e-01 8.46485913e-01 3.60150605e-01 8.16579521e-01
-6.22636437e-01 9.66754854e-01 -6.49608552e-01 -3.86923552e-01
-6.18101656e-01 -7.38333464e-01 -7.67428517e-01 2.76122838e-01
-5.75851560e-01 -1.52349338e-01 1.35581251e-02 7.14916527e-01
5.36119998e-01 1.11468464e-01 4.03370649e-01 -1.36399484e+00
-1.73119009e-01 7.56654382e-01 -4.25253481e-01 1.70898065e-02
5.78660727e-01 -3.66331279e-01 8.77245486e-01 -8.72651696e-01
1.79759309e-01 1.51890922e+00 2.49111906e-01 9.26746905e-01
-6.24741733e-01 -7.41504848e-01 8.99942636e-01 4.19557482e-01
-1.26723540e+00 -7.55259693e-01 1.20493186e+00 -2.62697488e-01
9.18898463e-01 3.99550200e-01 5.52701831e-01 1.26154733e+00
1.19981049e-02 1.01768172e+00 9.86774862e-01 -4.35953349e-01
1.63651124e-01 5.67200422e-01 2.35424042e-01 2.30816260e-01
-8.51020575e-01 -2.23854482e-01 -8.47078621e-01 -1.75171718e-01
3.03351104e-01 -4.93088737e-02 -2.96157002e-01 2.48056173e-01
-9.29066002e-01 7.09209979e-01 -5.65029867e-02 3.84333581e-01
-2.98645347e-01 -1.92055300e-01 7.11788237e-01 3.63031536e-01
5.28248429e-01 -4.68362004e-01 -5.01543522e-01 -6.92234159e-01
-5.20447433e-01 -2.67831266e-01 1.11910069e+00 9.76565182e-01
5.13197839e-01 -1.37450293e-01 1.01721607e-01 1.15796113e+00
6.19865179e-01 2.30007648e-01 6.35494411e-01 -6.42355919e-01
4.33821768e-01 6.45786941e-01 -6.04451858e-02 -1.10913444e+00
-3.70210588e-01 -3.02116871e-01 -8.98364365e-01 -4.28688586e-01
-2.97119588e-01 -3.37867111e-01 -3.12994182e-01 2.03078628e+00
5.75264931e-01 3.61771345e-01 4.69234079e-01 7.33370662e-01
9.52348232e-01 8.48785102e-01 3.96555096e-01 -7.10696638e-01
1.38237786e+00 -1.10827827e+00 -1.16149235e+00 -3.76357764e-01
6.58155084e-01 -7.07354307e-01 1.02343488e+00 2.53169030e-01
-7.08107710e-01 -4.18933213e-01 -6.82494938e-01 -2.18118681e-03
-5.25940582e-02 -3.99736017e-02 5.26973367e-01 6.41936421e-01
-7.91637063e-01 -2.69402444e-01 -7.60465324e-01 -1.45632491e-01
-1.34008363e-01 3.02493095e-01 -1.17810808e-01 2.93008894e-01
-1.34595406e+00 6.47989810e-01 8.27906281e-02 6.25861049e-01
-6.57025456e-01 -1.01192199e-01 -8.98436964e-01 3.20479721e-01
3.24358195e-01 6.60970435e-02 1.50862730e+00 -1.39673626e+00
-2.02912354e+00 3.52023721e-01 -9.81052220e-01 9.55805033e-02
1.56667680e-01 -8.54436532e-02 -9.47230935e-01 -7.88121596e-02
-4.21736956e-01 4.13702071e-01 6.02402925e-01 -1.37919998e+00
-7.36736238e-01 -4.34360743e-01 6.13056906e-02 6.75853133e-01
-4.97884691e-01 4.98896569e-01 -6.97861493e-01 -3.45031232e-01
2.16370568e-01 -7.78679907e-01 -7.23958537e-02 -6.04055882e-01
-3.49207878e-01 -4.12522703e-01 1.07870698e+00 -6.79779530e-01
1.64709115e+00 -2.47937274e+00 1.59082875e-01 5.07904068e-02
-1.28802747e-01 -7.66421333e-02 -1.10042505e-01 6.48269832e-01
-8.62721279e-02 2.42652604e-03 -1.98236734e-01 -6.71673000e-01
1.47186354e-01 2.67543077e-01 -2.65525103e-01 -9.85233951e-03
9.33916345e-02 5.20094991e-01 -4.90523309e-01 -7.32466578e-01
1.26847118e-01 8.67157996e-01 -5.06685376e-01 6.14334226e-01
-1.81140993e-02 8.11578572e-01 -6.60077572e-01 4.74354684e-01
6.03504956e-01 4.96314093e-02 5.04743695e-01 -1.59646315e-03
-1.60963565e-01 5.19247890e-01 -1.27387190e+00 1.48104954e+00
-9.12336469e-01 3.69385421e-01 5.67093194e-01 -8.88905883e-01
1.07561946e+00 7.74209857e-01 1.80862591e-01 -5.08072257e-01
1.96683872e-02 6.38657287e-02 -5.90147525e-02 -8.56927037e-01
3.47417533e-01 -4.72184241e-01 -2.53481388e-01 5.75991273e-01
-1.82200760e-01 3.27105016e-01 -2.63858646e-01 2.39270434e-01
6.09086037e-01 -2.30166793e-01 2.74047852e-01 4.80224863e-02
1.08845317e+00 -8.29539299e-01 8.69826972e-01 2.59456262e-02
-6.04971886e-01 4.78468537e-02 6.59673095e-01 -6.22703508e-02
-1.94627956e-01 -3.36837411e-01 7.70230219e-02 1.51049650e+00
2.30185360e-01 -5.29688716e-01 -8.44833553e-01 -6.85872495e-01
-5.39436281e-01 8.05975676e-01 -5.12456775e-01 -3.16238962e-02
-4.85299528e-01 -4.65119064e-01 2.99840599e-01 4.59648848e-01
4.00660276e-01 -1.34536994e+00 -2.71694571e-01 4.67746943e-01
-7.85003603e-01 -1.19405818e+00 -7.16877759e-01 1.13689922e-01
-6.13423169e-01 -6.24371350e-01 -3.09920669e-01 -1.01329494e+00
3.37331593e-01 1.26195282e-01 8.93516600e-01 5.06985262e-02
2.64761239e-01 4.27321732e-01 -4.76552367e-01 -3.87767196e-01
-5.69955587e-01 2.82097352e-03 -1.06585994e-01 4.51331019e-01
6.34080231e-01 -7.50693798e-01 -5.27953207e-01 5.37069738e-01
-7.64110208e-01 3.03719282e-01 4.13796186e-01 5.76447785e-01
1.26324028e-01 7.13162869e-02 1.15118778e+00 -8.38780403e-01
7.81397939e-01 -4.27181274e-01 -8.25047046e-02 4.96156424e-01
-3.34870040e-01 -1.62245229e-01 5.18980980e-01 -2.87538856e-01
-1.72189486e+00 6.72189593e-02 -4.98001128e-01 9.98495072e-02
-3.89164835e-01 6.76602960e-01 -7.00610280e-01 2.73201138e-01
-1.74336851e-01 4.60285634e-01 -5.34538403e-02 -4.45803404e-01
2.51554549e-01 1.38985813e+00 1.91223413e-01 -5.87116897e-01
-2.01972737e-03 6.03969917e-02 -6.85536563e-01 -8.71067345e-01
-6.41590714e-01 -6.83938384e-01 -4.47611898e-01 -5.46938956e-01
8.01829696e-01 -1.13535810e+00 -1.17539787e+00 5.06303608e-01
-1.32792330e+00 -1.26334786e-01 4.23149556e-01 5.93074620e-01
-3.39913607e-01 5.19425869e-01 -7.98396230e-01 -1.52349615e+00
-3.34719390e-01 -1.36650717e+00 1.04025662e+00 2.73119807e-01
-2.75104493e-01 -1.15408111e+00 -6.08335882e-02 5.28976083e-01
3.67272139e-01 -1.64542660e-01 7.01177239e-01 -6.26929164e-01
-5.28217778e-02 3.21547240e-02 -2.94408053e-02 3.18085641e-01
2.25539923e-01 1.22492649e-02 -1.17149317e+00 2.24226918e-02
3.57043713e-01 -3.88219535e-01 3.72089505e-01 3.48821632e-03
1.27659822e+00 -3.10631752e-01 -2.06616104e-01 1.15998454e-01
9.16072547e-01 5.17907619e-01 5.67711294e-01 -1.27017900e-01
5.00241160e-01 1.18204737e+00 5.93214035e-01 6.13130569e-01
1.02144659e+00 5.20844340e-01 2.19446197e-01 1.21902585e-01
5.20763755e-01 -7.44876042e-02 7.65375197e-01 1.57682216e+00
4.95960936e-02 -3.65461290e-01 -6.59884572e-01 4.54254478e-01
-1.90284884e+00 -9.30787683e-01 -1.34096354e-01 1.82377493e+00
1.01475573e+00 -1.06111407e-01 2.86460947e-02 6.16397262e-02
8.85459840e-01 2.87799180e-01 -4.89092410e-01 -8.48680139e-01
1.46890253e-01 -3.26737881e-01 -3.92672032e-01 7.26035535e-01
-8.32554460e-01 8.44110727e-01 6.07728386e+00 5.30409217e-01
-1.19670951e+00 6.79442063e-02 7.06299603e-01 6.58612773e-02
-4.29024398e-01 -1.37462407e-01 -8.21152270e-01 6.27506077e-01
1.17361665e+00 -8.14449415e-03 4.68784213e-01 8.38529468e-01
4.65294033e-01 2.09870208e-02 -1.07265699e+00 1.18736315e+00
2.68407106e-01 -5.48498571e-01 -2.89611816e-01 -1.03601612e-01
2.90102333e-01 -3.77885103e-01 -2.98545778e-01 6.32606447e-01
-1.21852227e-01 -7.12799489e-01 5.56593895e-01 1.70774564e-01
4.77430552e-01 -1.04865813e+00 6.58219934e-01 7.18195677e-01
-1.43167615e+00 -9.70308855e-02 1.26655512e-02 -6.82896972e-02
4.44628298e-01 4.84982282e-01 -4.90596861e-01 3.85516018e-01
5.98549545e-01 6.29130185e-01 -1.11071698e-01 1.53414160e-01
-2.62747645e-01 8.78716707e-01 -1.67262778e-01 -1.78182274e-01
-1.81643069e-02 -2.21636966e-01 2.95616150e-01 1.55382466e+00
1.61315739e-01 4.64809775e-01 2.19356403e-01 3.38454813e-01
-1.05416425e-01 4.81573522e-01 -1.27384186e-01 -7.70133510e-02
6.97374582e-01 1.27694368e+00 -3.43929887e-01 -3.77327502e-01
-6.18227124e-01 1.19666350e+00 3.03076178e-01 4.56675589e-01
-1.00246406e+00 -2.66023695e-01 6.34164691e-01 -4.49769437e-01
2.26473704e-01 -7.20922798e-02 -6.97792470e-02 -1.24920619e+00
2.69557565e-01 -1.12630534e+00 3.03556472e-01 -4.21923608e-01
-1.15899622e+00 7.32907653e-01 -3.34566951e-01 -1.01771939e+00
-2.77901322e-01 -9.93705466e-02 -1.01498950e+00 7.60855377e-01
-1.59600699e+00 -1.09509742e+00 -2.92251974e-01 7.03553975e-01
8.31986547e-01 1.02757640e-01 9.81328249e-01 2.69816905e-01
-8.64294946e-01 5.71824551e-01 -1.96979329e-01 -4.45243493e-02
8.25481772e-01 -9.20004368e-01 -7.80372769e-02 4.26243097e-01
-2.69811600e-01 5.88949859e-01 5.86284220e-01 -4.14057165e-01
-1.31472349e+00 -7.62358069e-01 1.44193888e+00 3.55045758e-02
4.54291552e-01 -7.56882966e-01 -1.09337795e+00 7.86021769e-01
4.04992729e-01 -5.18028438e-01 1.38292170e+00 6.89979494e-01
-2.26100013e-01 -1.00685038e-01 -8.20204437e-01 4.96721804e-01
8.70843709e-01 -9.42891836e-01 -5.33753514e-01 -8.38638693e-02
7.88379610e-01 -1.50775373e-01 -7.58211255e-01 2.61905342e-01
6.62536919e-01 -1.09928334e+00 5.54373026e-01 -2.68809170e-01
4.12102371e-01 -1.03114255e-01 -1.61875084e-01 -1.11104906e+00
9.97425765e-02 -5.64436376e-01 -1.66740835e-01 1.77704906e+00
3.42065245e-01 -6.66153193e-01 5.24312973e-01 9.67188716e-01
-2.38282651e-01 -6.37495100e-01 -9.28245127e-01 -1.94353640e-01
-2.87574649e-01 -7.48009264e-01 9.27827656e-01 1.07274091e+00
6.80128098e-01 6.85659409e-01 -7.39667833e-01 3.11172038e-01
1.50850758e-01 2.14054897e-01 6.61908031e-01 -1.01831579e+00
-3.30889046e-01 -2.23955333e-01 -8.45837668e-02 -1.28138149e+00
4.68770146e-01 -4.52816904e-01 5.61585248e-01 -1.35689974e+00
2.46038511e-01 -2.87170887e-01 -4.30697739e-01 3.03354055e-01
-3.99183661e-01 -4.45487201e-01 -5.93502261e-02 1.16349444e-01
-8.61871004e-01 9.93782341e-01 9.69236732e-01 1.64725110e-01
-4.41127300e-01 -6.05940372e-02 -7.76300013e-01 7.31559515e-01
7.74498045e-01 -2.72473156e-01 -6.48048103e-01 -2.72424549e-01
-1.06728025e-01 5.60687959e-01 -4.93870601e-02 -5.47869742e-01
2.60383457e-01 -2.53870219e-01 -2.76733916e-02 -6.69566333e-01
4.44695473e-01 -8.63581777e-01 1.73563007e-02 6.92870319e-02
-6.12920165e-01 -1.91923857e-01 2.93671768e-02 6.93397105e-01
-6.74506962e-01 1.95063632e-02 3.49780917e-01 1.66315809e-01
-6.67512298e-01 1.60889909e-01 -7.17793107e-01 -1.90563396e-01
8.86072695e-01 9.42453966e-02 1.71928898e-01 -7.89190888e-01
-9.46585119e-01 5.53668380e-01 -5.25600500e-02 6.24348104e-01
4.87630367e-01 -1.14911425e+00 -5.64188004e-01 2.62500703e-01
3.19108993e-01 -1.16271146e-01 8.36890280e-01 8.49203229e-01
1.82691306e-01 2.36700818e-01 2.55208313e-01 -4.49513555e-01
-1.73500645e+00 3.33936960e-01 2.48128459e-01 -3.09138149e-01
-2.31896311e-01 8.90616655e-01 6.60905600e-01 -6.62562966e-01
6.07221723e-01 -1.02742404e-01 -5.74916542e-01 1.05336979e-01
7.15865135e-01 -4.00440488e-03 -1.55334607e-01 -9.63526249e-01
-5.34411252e-01 2.17068031e-01 -1.49979815e-01 -1.85834080e-01
1.20671880e+00 -8.22477221e-01 -3.01607937e-01 8.69389236e-01
1.26125181e+00 8.29623118e-02 -1.07493186e+00 -2.42020011e-01
7.00167716e-02 -1.22830771e-01 -3.57258283e-02 -7.71296680e-01
-7.61962175e-01 1.14081478e+00 3.86185825e-01 2.38485694e-01
1.31766129e+00 -6.22352492e-03 8.84379148e-01 2.59063393e-01
1.30388483e-01 -1.34450996e+00 4.42190059e-02 5.47935903e-01
9.68624294e-01 -1.24745536e+00 -6.25771642e-01 -7.51331925e-01
-1.16791916e+00 1.04969633e+00 7.03845918e-01 6.62946224e-01
9.10700440e-01 3.68300617e-01 5.39574385e-01 -2.21981760e-02
-1.16319168e+00 5.60270399e-02 1.58543102e-02 2.76519567e-01
1.04778814e+00 1.70703456e-01 -3.26782078e-01 1.03912604e+00
-1.16532192e-01 -2.37663500e-02 2.55933225e-01 9.40718889e-01
-3.90821308e-01 -1.43878603e+00 -2.03682706e-01 -1.41852483e-01
-3.84418339e-01 -1.09205851e-02 -4.67594117e-01 1.66443586e-01
-1.60946567e-02 1.59210634e+00 -1.19965717e-01 -5.15753508e-01
2.47909904e-01 5.29532194e-01 -2.09588274e-01 -4.22056168e-01
-6.59592569e-01 5.30031741e-01 3.86520743e-01 -4.67451274e-01
-7.76789963e-01 -6.91549897e-01 -1.44912517e+00 -7.51138479e-02
-3.47656101e-01 6.04257643e-01 9.33997333e-01 1.13486624e+00
5.31106591e-01 5.92207730e-01 1.24433768e+00 -5.04487693e-01
-2.50249654e-01 -1.15252388e+00 -6.26237690e-01 3.21098238e-01
3.06011647e-01 -3.80062342e-01 -5.18222809e-01 2.01989800e-01] | [13.050408363342285, 6.067041397094727] |
7d28f2a9-9354-4c81-9ddd-1313ba9a3d0e | gan-based-domain-inference-attack | 2212.11810 | null | https://arxiv.org/abs/2212.11810v1 | https://arxiv.org/pdf/2212.11810v1.pdf | GAN-based Domain Inference Attack | Model-based attacks can infer training data information from deep neural network models. These attacks heavily depend on the attacker's knowledge of the application domain, e.g., using it to determine the auxiliary data for model-inversion attacks. However, attackers may not know what the model is used for in practice. We propose a generative adversarial network (GAN) based method to explore likely or similar domains of a target model -- the model domain inference (MDI) attack. For a given target (classification) model, we assume that the attacker knows nothing but the input and output formats and can use the model to derive the prediction for any input in the desired form. Our basic idea is to use the target model to affect a GAN training process for a candidate domain's dataset that is easy to obtain. We find that the target model may distract the training procedure less if the domain is more similar to the target domain. We then measure the distraction level with the distance between GAN-generated datasets, which can be used to rank candidate domains for the target model. Our experiments show that the auxiliary dataset from an MDI top-ranked domain can effectively boost the result of model-inversion attacks. | ['Keke Chen', 'Yuechun Gu'] | 2022-12-22 | null | null | null | null | ['inference-attack'] | ['adversarial'] | [ 4.29971457e-01 3.27287257e-01 -2.67708212e-01 -2.02974513e-01
-9.04991984e-01 -1.30930972e+00 4.95101929e-01 -5.41424572e-01
-1.24929890e-01 7.74527907e-01 -1.81727216e-01 -6.85532510e-01
3.25838745e-01 -1.22121382e+00 -1.11221576e+00 -7.69465744e-01
3.92946243e-01 6.09468400e-01 9.36453193e-02 -1.27846792e-01
2.63891906e-01 6.63105488e-01 -8.41387928e-01 6.34231865e-01
7.19238162e-01 8.70243251e-01 -1.34662818e-02 6.67729735e-01
5.91376089e-02 8.70171130e-01 -1.40544236e+00 -7.10959315e-01
4.90718007e-01 -6.05105400e-01 -8.29677880e-01 -5.30765831e-01
4.60425802e-02 -5.45213699e-01 -3.86414319e-01 1.50010324e+00
3.13127905e-01 -4.64461178e-01 6.59301639e-01 -1.60816371e+00
-3.83171946e-01 6.72758520e-01 -1.32764325e-01 6.31140992e-02
1.24761641e-01 3.36570531e-01 4.11115974e-01 -3.38606536e-01
4.86093879e-01 1.23357999e+00 1.72713697e-01 1.02333391e+00
-1.20947146e+00 -1.46005797e+00 -2.01269880e-01 1.05576575e-01
-1.24839163e+00 -2.99515337e-01 1.25304055e+00 -2.86446035e-01
2.52382278e-01 3.78767669e-01 9.04675722e-02 1.77487302e+00
1.00818463e-01 6.17469370e-01 1.25512981e+00 -3.12413514e-01
5.13540149e-01 6.14792049e-01 3.61035727e-02 2.32926816e-01
2.81707436e-01 6.63547337e-01 -3.80931258e-01 -6.40602767e-01
5.34847438e-01 -3.18708509e-01 -2.37870350e-01 -1.04965739e-01
-5.83417416e-01 1.12152386e+00 3.28598976e-01 -4.54690941e-02
-2.56287813e-01 4.07816283e-02 3.27531785e-01 5.71055949e-01
1.04213715e-01 8.36568236e-01 -5.65148056e-01 -4.20990959e-03
-6.58559859e-01 1.20495349e-01 8.83433640e-01 9.68481004e-01
7.61084437e-01 3.92564118e-01 3.51267666e-01 1.80512562e-01
1.03791215e-01 7.20889986e-01 2.57322222e-01 -5.17999172e-01
6.56692445e-01 3.09021890e-01 -4.35382500e-02 -7.09824562e-01
4.22304899e-01 -1.76648378e-01 -7.02137172e-01 5.70689797e-01
7.18515456e-01 -5.14597714e-01 -8.15389693e-01 1.98904932e+00
1.50880411e-01 4.98265535e-01 3.49426419e-01 5.40647805e-01
5.85397005e-01 6.94694459e-01 -7.32147228e-03 1.88254327e-01
1.03769028e+00 -4.51258689e-01 -1.92870975e-01 -3.41787666e-01
7.63246775e-01 -5.03190458e-01 9.78345454e-01 5.31835258e-01
-8.30166757e-01 -6.06855392e-01 -1.32873940e+00 5.67441463e-01
-4.83858883e-01 -1.57306865e-01 1.59442604e-01 1.22386646e+00
-5.13892829e-01 4.31604832e-01 -5.31426728e-01 3.18905324e-01
6.98316813e-01 4.98235345e-01 -2.77198583e-01 -8.81304815e-02
-1.63047421e+00 7.83331931e-01 5.23822606e-01 -4.52698976e-01
-1.74645960e+00 -8.22988689e-01 -7.25397825e-01 3.99300344e-02
2.36099094e-01 -3.95968616e-01 9.33143497e-01 -1.28492558e+00
-1.31948042e+00 6.52450740e-01 -7.79602118e-03 -8.53527844e-01
4.24670845e-01 1.34433687e-01 -7.22805262e-01 8.13190863e-02
-1.82559893e-01 3.02642196e-01 1.43538117e+00 -1.42811477e+00
-4.41131562e-01 -2.62993783e-01 4.49583352e-01 -2.19357625e-01
-2.93581903e-01 6.63509965e-02 1.45311892e-01 -6.82417035e-01
-5.27556777e-01 -1.06551266e+00 1.13163022e-02 -1.63858160e-01
-9.61384118e-01 3.37927878e-01 1.25223708e+00 -6.93668306e-01
1.03526473e+00 -2.37398696e+00 -3.42090160e-01 7.00048685e-01
1.87337816e-01 7.26669729e-01 -1.33067310e-01 2.49583870e-01
-4.33312982e-01 5.76572776e-01 -1.40585795e-01 1.36409387e-01
-1.38530076e-01 2.80658782e-01 -1.19609952e+00 3.41669112e-01
1.99341968e-01 7.74381280e-01 -4.29914206e-01 -3.27829830e-02
-4.68774661e-02 3.22575212e-01 -5.76963902e-01 6.66347444e-01
-2.69089013e-01 6.50026739e-01 -5.35292447e-01 2.44733006e-01
1.07087755e+00 1.57212123e-01 6.58109263e-02 -1.84713185e-01
6.76531494e-01 6.03985608e-01 -8.82111490e-01 9.17706907e-01
-5.32001257e-01 5.26069105e-01 -1.01777211e-01 -1.04143727e+00
1.22500455e+00 1.28700197e-01 -3.77709389e-01 -4.39244807e-01
2.43401178e-03 1.30545631e-01 4.48995590e-01 2.69116499e-02
-1.97964981e-01 -1.13455400e-01 -3.55784774e-01 6.37550294e-01
-2.97666371e-01 -5.72851487e-02 -6.07599616e-01 9.56151485e-02
1.11346459e+00 -2.13775724e-01 1.95962429e-01 -2.46321693e-01
4.18198824e-01 1.30880140e-02 5.22400022e-01 9.14406538e-01
3.69736403e-02 2.11307049e-01 9.57007468e-01 -3.71642262e-01
-1.00047529e+00 -1.35211277e+00 2.39416093e-01 6.96434200e-01
2.11420134e-01 -9.26371955e-04 -8.81998301e-01 -1.31441998e+00
-1.75967827e-01 1.33474123e+00 -7.83005238e-01 -7.99229503e-01
-6.07748389e-01 -5.99377453e-02 1.16973877e+00 5.27855217e-01
6.43241346e-01 -1.05340743e+00 -3.94197941e-01 -1.58435374e-01
9.05530527e-02 -7.38764107e-01 -3.49910825e-01 3.99705112e-01
-6.35097325e-01 -1.13211668e+00 -2.03864738e-01 -6.93743289e-01
8.49899232e-01 -2.96876609e-01 1.01249802e+00 9.72590521e-02
4.82512861e-01 -1.96772113e-01 -1.14704862e-01 -6.76248968e-01
-1.28220117e+00 1.53541967e-01 -1.15449682e-01 -4.74778004e-02
4.99274880e-01 -6.08240008e-01 -1.12476818e-01 5.24797022e-01
-1.04118669e+00 -2.07782939e-01 2.94801086e-01 8.02588701e-01
2.05823109e-01 5.95623255e-01 7.28651285e-01 -1.43908501e+00
6.65101469e-01 -4.83416557e-01 -7.53905892e-01 2.62353808e-01
-2.42097348e-01 2.68640697e-01 1.32255590e+00 -1.05129516e+00
-7.95401931e-01 -1.79427341e-01 -2.71109432e-01 -8.12406838e-01
-4.02643204e-01 2.67041892e-01 -1.13895273e+00 -1.42468512e-01
9.53766882e-01 4.08314675e-01 -2.04006787e-02 -2.42399067e-01
1.10064872e-01 5.70044577e-01 4.87802237e-01 -7.31697023e-01
1.56395721e+00 2.03341618e-01 -1.91197470e-02 -2.79726833e-01
-5.38793147e-01 5.51685870e-01 -2.83249497e-01 1.21600859e-01
5.59189677e-01 -6.76607609e-01 -5.41238606e-01 6.58833146e-01
-1.42550349e+00 -4.93327051e-01 -5.03505953e-02 8.33453909e-02
-1.89751446e-01 3.87160294e-02 -2.59770393e-01 -7.07985997e-01
-1.66692495e-01 -1.53000462e+00 5.40368557e-01 -6.64268062e-03
-3.33992213e-01 -1.11225784e+00 -3.04904372e-01 1.33781284e-01
3.11359525e-01 2.99680859e-01 1.45027351e+00 -1.58432806e+00
-5.03560662e-01 -5.11489451e-01 1.86236545e-01 7.74708629e-01
1.05381161e-01 -2.48404205e-01 -1.30221045e+00 -3.62034023e-01
6.00959122e-01 -3.13623607e-01 2.69438118e-01 -3.85712981e-02
1.47273719e+00 -7.86143303e-01 -2.51547694e-01 6.40733480e-01
1.13921201e+00 8.09715629e-01 1.17045259e+00 2.23019883e-01
6.94859922e-01 1.97498366e-01 6.08632982e-01 4.60119918e-02
-3.60647142e-01 5.54837346e-01 6.36496782e-01 -1.68211460e-01
2.45767608e-01 -7.74812043e-01 5.25205612e-01 1.59272570e-02
6.82857215e-01 -2.17223287e-01 -7.94652581e-01 1.72785774e-01
-9.39618647e-01 -9.28963602e-01 3.88563216e-01 2.49134088e+00
9.69937980e-01 6.97221220e-01 2.07972489e-02 1.17727429e-01
9.97512758e-01 -9.06387940e-02 -8.78570795e-01 -6.32546186e-01
7.06485137e-02 6.26509309e-01 5.82217395e-01 5.37890792e-01
-9.31895375e-01 9.10680771e-01 5.88080359e+00 1.27871764e+00
-1.35552144e+00 2.18209364e-02 9.01788712e-01 8.69426727e-02
-6.25690818e-01 2.42087409e-01 -9.04899120e-01 1.04116297e+00
1.13881564e+00 -5.54140449e-01 6.73389494e-01 1.27302325e+00
-2.83170283e-01 3.99618268e-01 -1.48333657e+00 6.85348332e-01
-1.48501024e-01 -1.30914402e+00 2.49151543e-01 5.62267065e-01
3.35453331e-01 -5.33178926e-01 4.87164974e-01 5.10931909e-01
6.87939942e-01 -1.25410628e+00 5.73837459e-01 1.52600229e-01
1.03514194e+00 -1.34543288e+00 6.26101911e-01 8.03602219e-01
-6.51559889e-01 -7.15374947e-02 -3.88136268e-01 1.31748527e-01
-3.72894585e-01 2.99559414e-01 -1.13505673e+00 2.12308705e-01
1.67445168e-01 6.63271546e-03 -4.94920313e-01 3.05910051e-01
-5.79967499e-01 1.19381905e+00 -3.25942367e-01 2.11270362e-01
3.83333564e-02 -1.56524125e-02 5.48633695e-01 6.00880623e-01
2.72556901e-01 -1.57007024e-01 -3.85306738e-02 1.17607939e+00
-3.68626684e-01 -6.12426996e-01 -1.09330058e+00 -2.05126956e-01
1.00784457e+00 6.83721185e-01 -8.64074826e-02 -3.21198136e-01
-4.82119843e-02 8.81960213e-01 1.24691322e-01 3.03393066e-01
-1.25674045e+00 -4.17132378e-01 8.81993532e-01 7.23197386e-02
9.82949063e-02 1.35724589e-01 -5.28948843e-01 -7.71241665e-01
-3.62950447e-03 -1.57193899e+00 4.20739770e-01 -7.40495443e-01
-1.31458831e+00 8.43885422e-01 4.52146642e-02 -1.55191672e+00
-5.26387334e-01 -4.15698618e-01 -1.00962520e+00 1.44414806e+00
-1.11836481e+00 -1.17346239e+00 1.47661865e-01 8.80181491e-01
2.19915122e-01 -6.74545765e-01 1.03372002e+00 -7.68261328e-02
-2.81790406e-01 1.28902876e+00 -1.75864100e-01 6.71598613e-01
4.48579371e-01 -8.21137071e-01 7.71434307e-01 1.14187729e+00
1.95752457e-01 6.60050094e-01 5.53415895e-01 -7.60975182e-01
-1.17350256e+00 -1.37398124e+00 4.00617719e-01 -6.76108122e-01
5.80040395e-01 -5.43041527e-01 -1.01123130e+00 7.99721837e-01
-5.44161983e-02 -3.35666209e-01 8.55212390e-01 -4.15815920e-01
-7.91582644e-01 -4.49619107e-02 -1.67139721e+00 7.17982709e-01
5.56545973e-01 -9.67300355e-01 -5.42746305e-01 -1.14678675e-02
8.20667505e-01 -3.28053117e-01 -6.09368920e-01 2.39818409e-01
2.04097167e-01 -6.61324978e-01 1.04033864e+00 -1.13275790e+00
5.16836762e-01 -1.96161568e-01 -2.25216508e-01 -1.53075814e+00
-1.06567824e-02 -5.27761757e-01 -1.62055254e-01 1.24702656e+00
5.70976555e-01 -1.01039016e+00 1.05062854e+00 8.04576337e-01
2.83944726e-01 -2.89520293e-01 -1.01417136e+00 -9.16474164e-01
6.55093193e-01 -5.04681647e-01 1.18086922e+00 9.79504108e-01
-3.76994073e-01 2.24982843e-01 -5.91777325e-01 7.60923564e-01
6.06656551e-01 5.26925884e-02 1.23247778e+00 -1.06719565e+00
-4.87959176e-01 -1.29771769e-01 -3.54222000e-01 -1.06242514e+00
5.40044427e-01 -9.48902965e-01 -3.93642545e-01 -5.94170332e-01
-3.00590843e-01 -5.05742788e-01 -1.24080941e-01 5.48031092e-01
6.94439039e-02 1.37629742e-02 3.05689007e-01 2.29492739e-01
2.42318660e-01 1.58127546e-01 1.09613311e+00 -2.90744692e-01
8.76469444e-03 4.38206643e-01 -1.12772214e+00 6.23978555e-01
1.12017047e+00 -1.06294024e+00 -7.96096861e-01 -1.04236759e-01
-1.71091467e-01 1.52760401e-01 5.30144334e-01 -1.05423057e+00
6.97267503e-02 -2.81431735e-01 5.14394701e-01 -2.81429678e-01
2.79797554e-01 -1.24305987e+00 4.49473202e-01 7.34712422e-01
-6.16026163e-01 -3.53169978e-01 2.33471334e-01 4.10166830e-01
-4.59976718e-02 -5.72018147e-01 1.11992764e+00 9.23326015e-02
-3.76456916e-01 2.99520999e-01 -2.64710903e-01 1.51576117e-01
1.01849115e+00 -1.77842885e-01 -5.70166111e-01 -5.93154728e-01
-7.25696325e-01 -3.10415000e-01 6.36824369e-01 2.02085316e-01
6.18299842e-01 -1.38433468e+00 -5.71874499e-01 7.25349665e-01
-7.75516480e-02 -1.18385412e-01 -1.61255643e-01 -2.40743789e-03
-1.95234060e-01 -9.69599187e-02 -3.53451222e-01 -2.95769572e-01
-1.27664268e+00 1.01894104e+00 4.22481030e-01 -4.81467932e-01
-7.84655437e-02 6.92052901e-01 7.78607786e-01 -3.40764940e-01
1.10019289e-01 1.51325360e-01 2.29906678e-01 -5.56583941e-01
6.02326989e-01 -1.39702568e-02 -1.58577263e-01 -2.53672630e-01
-3.85919213e-01 -1.16140269e-01 -3.00028324e-01 -4.53349501e-02
8.11298192e-01 4.80221450e-01 1.73005208e-01 -7.10808933e-02
1.46501112e+00 1.29830346e-01 -1.07524061e+00 -2.80897141e-01
-3.76589686e-01 -6.85259700e-01 -3.93150836e-01 -1.06694508e+00
-1.13421929e+00 1.20885253e+00 4.24591392e-01 4.41310227e-01
1.29741168e+00 -1.53950557e-01 7.42917299e-01 3.14666003e-01
6.51615560e-01 -6.73745811e-01 5.37413582e-02 2.48227850e-01
7.55537987e-01 -7.22797513e-01 -4.03315187e-01 -4.24069166e-01
-9.33460474e-01 1.00739396e+00 8.36252391e-01 -3.49715382e-01
7.59128273e-01 7.01077282e-01 1.40406549e-01 1.77123114e-01
-4.94206220e-01 5.90188384e-01 9.43489224e-02 1.23741353e+00
-4.94135797e-01 7.92778134e-02 5.52855074e-01 1.02816701e+00
-5.63381791e-01 -2.46589020e-01 6.14668667e-01 6.56593919e-01
-1.46669537e-01 -1.30055058e+00 -5.75325489e-01 4.76592571e-01
-4.21772659e-01 -2.24306747e-01 -6.22400045e-01 7.27738619e-01
4.36973497e-02 9.48577881e-01 -8.75743330e-02 -8.89905989e-01
9.73728895e-02 1.99182883e-01 8.05436969e-02 -5.53780317e-01
-6.11169577e-01 -2.46306002e-01 2.07672045e-01 -3.65083337e-01
2.63838887e-01 -1.45286888e-01 -8.82671356e-01 -6.18796587e-01
-2.80329049e-01 2.65816659e-01 3.70561123e-01 8.56476307e-01
2.51750350e-01 1.46238506e-01 1.24749756e+00 -3.18298966e-01
-9.02657270e-01 -6.35006011e-01 -5.53122342e-01 5.24372756e-01
1.16947345e-01 -5.38981438e-01 -6.40892208e-01 7.90320113e-02] | [5.776897430419922, 7.695074081420898] |
23169d92-cc12-45f1-98ff-26684ccd9e8c | rethinking-image-deraining-via-rain-streaks | 2008.00823 | null | https://arxiv.org/abs/2008.00823v1 | https://arxiv.org/pdf/2008.00823v1.pdf | Rethinking Image Deraining via Rain Streaks and Vapors | Single image deraining regards an input image as a fusion of a background image, a transmission map, rain streaks, and atmosphere light. While advanced models are proposed for image restoration (i.e., background image generation), they regard rain streaks with the same properties as background rather than transmission medium. As vapors (i.e., rain streaks accumulation or fog-like rain) are conveyed in the transmission map to model the veiling effect, the fusion of rain streaks and vapors do not naturally reflect the rain image formation. In this work, we reformulate rain streaks as transmission medium together with vapors to model rain imaging. We propose an encoder-decoder CNN named as SNet to learn the transmission map of rain streaks. As rain streaks appear with various shapes and directions, we use ShuffleNet units within SNet to capture their anisotropic representations. As vapors are brought by rain streaks, we propose a VNet containing spatial pyramid pooling (SSP) to predict the transmission map of vapors in multi-scales based on that of rain streaks. Meanwhile, we use an encoder CNN named ANet to estimate atmosphere light. The SNet, VNet, and ANet are jointly trained to predict transmission maps and atmosphere light for rain image restoration. Extensive experiments on the benchmark datasets demonstrate the effectiveness of the proposed visual model to predict rain streaks and vapors. The proposed deraining method performs favorably against state-of-the-art deraining approaches. | ['Chao Ma', 'Yinglong Wang', 'Bing Zeng', 'Yibing Song'] | 2020-08-03 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2761_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123620358.pdf | eccv-2020-8 | ['single-image-deraining'] | ['computer-vision'] | [ 2.71399409e-01 -4.76061672e-01 5.17660916e-01 -4.70131665e-01
7.60422796e-02 -3.32342863e-01 4.54448640e-01 -4.54428792e-01
6.83573782e-02 7.25358248e-01 1.40494838e-01 -2.64819026e-01
4.98303771e-01 -1.27131021e+00 -1.10266483e+00 -1.23758805e+00
2.41878048e-01 -5.29434346e-02 2.16621131e-01 -5.47986925e-01
-1.22077994e-01 4.73090291e-01 -1.56441104e+00 2.90689230e-01
1.25692260e+00 7.28789985e-01 6.17260933e-01 7.98544049e-01
-2.74432272e-01 1.15456760e+00 -7.63740540e-01 -3.29569392e-02
4.20742750e-01 -6.34626150e-01 2.83852555e-02 1.37852386e-01
9.73075032e-01 -7.82981098e-01 -5.26815891e-01 1.02845442e+00
3.00199419e-01 -3.86078842e-02 6.42180204e-01 -7.84759104e-01
-1.24046552e+00 1.44803077e-01 -9.26180243e-01 5.59498549e-01
1.30209355e-02 3.22729349e-01 6.09722078e-01 -1.12390387e+00
4.63496357e-01 1.25552714e+00 3.35591972e-01 2.74514526e-01
-8.97953570e-01 -7.31909573e-01 3.73949021e-01 8.04963410e-02
-1.03326952e+00 -3.06422561e-01 6.43442094e-01 -3.14886928e-01
5.87654829e-01 5.58785200e-01 9.18566883e-01 7.98932791e-01
5.07158399e-01 7.18383670e-01 1.25646222e+00 -5.89335598e-02
4.81240377e-02 1.38388321e-01 1.45477176e-01 6.55538738e-01
5.37361503e-01 1.88588172e-01 -1.83578938e-01 4.32934463e-01
9.10426795e-01 2.28375539e-01 -9.09471691e-01 1.58673346e-01
-6.54594362e-01 5.61487913e-01 1.06521451e+00 -3.50258984e-02
-5.75077295e-01 4.35969867e-02 -2.96068668e-01 3.90590847e-01
8.44110131e-01 -7.42183551e-02 -2.05192670e-01 7.51001298e-01
-1.01250017e+00 5.08363366e-01 5.15931964e-01 5.41050136e-01
9.53985691e-01 7.26296842e-01 -3.56824398e-01 8.82955730e-01
5.00827551e-01 1.47388268e+00 -1.79410651e-02 -6.10080719e-01
5.55678725e-01 1.63404599e-01 5.80515504e-01 -9.74233329e-01
-1.32807806e-01 -5.97727895e-01 -1.38440955e+00 6.40334189e-01
-1.04193188e-01 -6.35662600e-02 -1.53729022e+00 1.36077571e+00
2.76510954e-01 7.09373951e-01 3.75748426e-01 1.39546156e+00
1.18092203e+00 1.40653336e+00 -1.92830130e-01 -4.04702663e-01
1.08266389e+00 -1.04921758e+00 -1.01762772e+00 -3.47258598e-01
-1.17994837e-01 -8.32461238e-01 7.73241222e-01 2.02306643e-01
-1.10771441e+00 -7.88008034e-01 -9.72509563e-01 -2.30007898e-02
-6.39626831e-02 -1.60516761e-02 4.45565015e-01 4.60593820e-01
-1.09347105e+00 2.91553348e-01 -6.83334470e-01 -5.67914471e-02
1.87361106e-01 -2.56397307e-01 1.06255487e-01 -5.14825881e-01
-1.19023693e+00 7.99406469e-01 -2.51565427e-02 9.34273481e-01
-1.11336827e+00 -7.54490316e-01 -7.58730888e-01 -4.47753780e-02
-2.69618511e-01 -9.62605059e-01 6.75965130e-01 -1.11914909e+00
-1.21186793e+00 4.97550875e-01 -4.22761768e-01 -2.20672950e-01
2.80652732e-01 -3.31062913e-01 -7.13799953e-01 1.34670034e-01
-1.52008981e-01 2.23866954e-01 1.25760758e+00 -1.98312414e+00
-6.72023535e-01 -1.32898256e-01 1.33227170e-01 6.29454494e-01
3.85134429e-01 -1.67284817e-01 -2.66078085e-01 -5.29861093e-01
1.16790332e-01 -5.30410528e-01 -8.87913331e-02 7.02356966e-03
-4.49333966e-01 5.71142793e-01 1.15569782e+00 -8.81953001e-01
9.24428940e-01 -1.93437934e+00 3.45577858e-02 -1.63779352e-02
4.45621222e-01 4.83660638e-01 -3.53966087e-01 -2.16796696e-02
-1.83332697e-01 -7.23358989e-02 -6.13886118e-01 -2.28633419e-01
-3.74538392e-01 4.88881916e-01 -8.58847558e-01 5.59088707e-01
3.25431496e-01 8.21838915e-01 -8.17116678e-01 -1.23600930e-01
4.34466988e-01 9.61623669e-01 -1.06960468e-01 6.21847153e-01
-3.18840533e-01 5.59725583e-01 -1.87295556e-01 6.72600210e-01
1.72236383e+00 1.69870064e-01 -2.96228170e-01 -3.16652298e-01
-3.23148161e-01 -1.51046038e-01 -8.67744386e-01 6.69487000e-01
-6.48958385e-01 5.23976445e-01 5.41927338e-01 -3.72263014e-01
1.17248356e+00 1.05671629e-01 -2.07295820e-01 -9.87991214e-01
-1.94075748e-01 2.67626643e-01 -1.08388104e-01 -7.13509798e-01
4.57574964e-01 -5.18873870e-01 8.98624718e-01 6.64360449e-02
-2.83914983e-01 -4.83450323e-01 -1.15552865e-01 1.42013669e-01
7.81777918e-01 1.31499574e-01 -3.94196630e-01 1.39488399e-01
3.95678759e-01 -5.13502955e-01 5.80898941e-01 6.82925642e-01
1.98055342e-01 9.64535236e-01 -1.71491235e-01 -6.72935605e-01
-1.07420182e+00 -1.48856354e+00 -1.25942767e-01 7.46420681e-01
5.63478231e-01 1.30039841e-01 -4.19657201e-01 -1.60254925e-01
-9.66313183e-02 6.87312245e-01 -6.64509594e-01 1.32982537e-01
-7.97348797e-01 -1.39019287e+00 1.20615721e-01 1.95475549e-01
1.10484231e+00 -1.38684106e+00 -3.76119256e-01 1.06229164e-01
-4.13052171e-01 -1.15682220e+00 -2.70934910e-01 -1.83116365e-02
-6.62400961e-01 -9.26094115e-01 -8.69808853e-01 -6.80615127e-01
6.12455666e-01 8.77018690e-01 1.35140634e+00 1.47101551e-01
-5.34456491e-01 -2.40223613e-02 -5.22414148e-01 -4.67918694e-01
-1.59937039e-01 -8.67828965e-01 -2.94578403e-01 3.88959497e-01
-1.11674361e-01 -8.65184188e-01 -1.09829295e+00 -8.01808983e-02
-1.24313378e+00 3.32143515e-01 5.66722035e-01 6.19209349e-01
5.61484456e-01 2.03195632e-01 -1.35517210e-01 -9.68204796e-01
3.98667157e-01 -7.12324917e-01 -6.34064853e-01 2.22877890e-01
-1.89845294e-01 -3.03475291e-01 6.71011448e-01 -3.89029607e-02
-1.53967404e+00 -3.98410141e-01 -2.11807285e-02 -6.82522893e-01
-5.88369444e-02 3.43599379e-01 -2.44535469e-02 -1.72950849e-01
4.40154940e-01 7.69459009e-01 -5.92081845e-01 -4.93449301e-01
4.44547236e-01 4.29204822e-01 5.68243921e-01 -1.34980142e-01
1.37920976e+00 7.51459360e-01 -2.98846122e-02 -1.05311048e+00
-9.35488105e-01 -2.12865591e-01 -1.39825121e-01 -3.05711359e-01
9.78977203e-01 -1.31298029e+00 -3.44154269e-01 8.53393734e-01
-1.35719228e+00 -4.83723104e-01 -1.29703701e-01 1.69153631e-01
7.51945302e-02 2.12954983e-01 -8.93443763e-01 -1.12961376e+00
-6.45240486e-01 -8.41895044e-01 9.93819714e-01 6.09266043e-01
7.58856893e-01 -9.03100193e-01 1.85827285e-01 3.11214805e-01
7.22373366e-01 4.00801867e-01 7.58871913e-01 5.19623220e-01
-1.31890392e+00 3.01200420e-01 -7.20734954e-01 5.15029430e-01
4.51045483e-01 3.48411113e-01 -1.12844062e+00 -2.77224571e-01
3.48079562e-01 -2.84096934e-02 1.43068182e+00 6.18869781e-01
7.99104452e-01 -5.25628388e-01 4.59957123e-02 1.17822635e+00
1.75878632e+00 2.64604986e-01 1.01858544e+00 2.53219247e-01
1.16395485e+00 3.44975293e-01 4.89139080e-01 4.14294720e-01
2.83550739e-01 3.43272954e-01 9.22680080e-01 -7.93440282e-01
-5.81510961e-01 1.73416093e-01 4.69036460e-01 8.73282731e-01
-6.31926537e-01 -6.40888512e-01 -4.52093214e-01 3.69315326e-01
-1.42995965e+00 -1.25379050e+00 -5.78353703e-01 1.87857723e+00
7.26249814e-01 -2.08282426e-01 -6.74120784e-01 -3.94034564e-01
4.78932142e-01 7.51290798e-01 -5.43831110e-01 -1.76719636e-01
-7.35346556e-01 5.70386648e-01 5.47945261e-01 9.62891996e-01
-9.51933920e-01 1.02681172e+00 5.27961922e+00 2.72537768e-01
-1.50581348e+00 7.70597300e-03 5.60093224e-01 1.19350985e-01
-7.32532144e-01 -1.95011631e-01 -6.36955380e-01 6.40854955e-01
4.50959980e-01 3.24763834e-01 5.67281902e-01 1.64036408e-01
7.18173683e-01 -1.04987025e-01 -3.51065606e-01 8.02229285e-01
6.42588884e-02 -1.03357244e+00 3.96298081e-01 -5.29439330e-01
9.34301138e-01 3.30884486e-01 3.16996157e-01 1.05170898e-01
5.47410727e-01 -1.07642937e+00 6.64351046e-01 9.43462849e-01
6.83309972e-01 -2.54675061e-01 8.07558954e-01 -1.02085602e-02
-1.16295779e+00 1.24294713e-01 -6.47835433e-01 1.65569767e-01
1.70044169e-01 1.13111091e+00 -4.13911432e-01 9.58918691e-01
8.86096895e-01 8.80638838e-01 -3.25457573e-01 1.09861851e+00
-6.66105986e-01 9.08799708e-01 -2.90805250e-01 7.15755522e-01
1.74750835e-01 -8.78977716e-01 5.90463519e-01 1.35520327e+00
4.25072163e-01 3.34890366e-01 -1.87687296e-02 1.13662601e+00
-7.93220922e-02 -1.66336715e-01 -5.30150175e-01 3.31363976e-01
5.59517294e-02 1.23134708e+00 -3.35423470e-01 -3.99619967e-01
-3.66628975e-01 1.35092616e+00 -2.47445673e-01 1.10463870e+00
-9.35658216e-01 -3.13926160e-01 9.06786501e-01 8.28013793e-02
3.93864512e-01 3.73855792e-02 5.12045138e-02 -1.34105146e+00
1.92077786e-01 -7.13616133e-01 -3.57673109e-01 -1.37491548e+00
-1.24646020e+00 9.91795480e-01 -4.80346590e-01 -1.21240711e+00
5.36803007e-01 -5.63316882e-01 -1.17737222e+00 1.40413773e+00
-2.40100598e+00 -1.22896683e+00 -1.12195396e+00 6.79152071e-01
5.52273273e-01 2.72142351e-01 5.36171794e-01 2.60356098e-01
-6.07210934e-01 -1.66091532e-01 4.62016433e-01 -1.21826045e-01
7.40797758e-01 -1.28716528e+00 3.46898735e-01 1.14929235e+00
-4.24307436e-02 4.01429355e-01 1.12332070e+00 -7.23491132e-01
-1.25809443e+00 -1.71323442e+00 3.76904458e-01 6.17570430e-02
3.57103914e-01 -3.18612546e-01 -1.21515870e+00 5.60533822e-01
2.16542542e-01 4.75668341e-01 8.75940174e-02 -3.46204698e-01
-6.62979782e-01 -5.51745296e-01 -9.39137101e-01 4.44539726e-01
6.16588652e-01 -3.10572177e-01 -3.75579923e-01 6.20971978e-01
8.24434936e-01 -6.81902766e-01 -1.70804843e-01 6.06433511e-01
2.35993519e-01 -1.43295205e+00 1.10304332e+00 -1.44723738e-02
8.86490762e-01 -7.02664554e-01 -3.83504719e-01 -1.60468709e+00
-4.44306254e-01 4.50731255e-03 -2.92387277e-01 9.67461348e-01
1.58889100e-01 -5.53117931e-01 4.20258343e-01 -1.37054220e-01
-3.36059868e-01 -4.49369222e-01 -4.20120865e-01 -4.52088892e-01
-1.14951350e-01 2.02965647e-01 5.16233563e-01 9.34256494e-01
-1.05733061e+00 2.77148545e-01 -7.47328699e-01 7.75981963e-01
9.10078168e-01 5.75421810e-01 4.73740995e-01 -1.09154510e+00
-2.81841218e-01 -1.20457297e-03 7.58070499e-02 -1.05760503e+00
-1.58105150e-01 -6.47121906e-01 4.31124985e-01 -2.00100708e+00
3.33416045e-01 -4.65716839e-01 -5.64225495e-01 2.82404661e-01
-5.35699189e-01 5.18028796e-01 2.42358550e-01 3.81432205e-01
-2.59486794e-01 8.84656370e-01 1.68220055e+00 -5.59056699e-01
-3.42202932e-01 4.03650627e-02 -2.68736064e-01 6.94523811e-01
8.60479653e-01 -1.77906871e-01 -3.33057344e-01 -9.87336397e-01
1.25817969e-01 1.19645320e-01 5.00149608e-01 -9.62845623e-01
-1.20803289e-01 -2.79013991e-01 6.91113472e-01 -5.34482479e-01
6.61224961e-01 -6.29997492e-01 1.70323372e-01 2.74775594e-01
2.44185984e-01 -4.18878675e-01 1.89932194e-02 6.15976393e-01
-4.83507514e-01 2.68928945e-01 8.98867130e-01 -2.13234201e-01
-4.25904304e-01 7.52411008e-01 -3.52359086e-01 -4.36143845e-01
5.79443038e-01 -2.12644175e-01 -8.24957252e-01 -4.37588215e-01
-4.85964090e-01 2.43685424e-01 2.75325954e-01 2.83876777e-01
1.11433768e+00 -6.99248314e-01 -1.18295765e+00 2.58613735e-01
-2.46598050e-01 1.56618476e-01 6.71723127e-01 6.79749966e-01
-1.01209819e+00 -2.50262115e-02 -1.88352346e-01 -2.13167697e-01
-1.39559877e+00 2.77221471e-01 6.20857120e-01 -1.64461449e-01
-1.00032425e+00 1.13737488e+00 1.07763600e+00 -2.87766576e-01
-1.30378336e-01 -6.08918428e-01 -2.58074254e-01 -2.40626201e-01
7.67839789e-01 1.70334876e-02 -1.09339438e-01 -4.02053297e-01
9.22596827e-02 6.59883857e-01 -6.84835762e-02 3.60339314e-01
1.48009706e+00 -2.77875215e-01 -5.42715669e-01 5.63452423e-01
7.27605581e-01 1.83926642e-01 -1.34163165e+00 -2.14367226e-01
-9.18240428e-01 -6.08716786e-01 2.64055550e-01 -9.79588389e-01
-1.67974305e+00 9.88969982e-01 7.75707185e-01 9.94242355e-02
1.53746176e+00 -4.29978490e-01 8.09646130e-01 1.93910152e-01
8.64402428e-02 -3.60258043e-01 -1.69249177e-01 7.17717052e-01
1.01567352e+00 -1.24475527e+00 -6.02736548e-02 -6.00273013e-01
-4.68931049e-01 1.06173706e+00 7.60873139e-01 -3.21938336e-01
6.57150328e-01 4.08874750e-01 5.07375598e-01 -1.92679271e-01
-5.31684279e-01 -2.45525956e-01 -1.90105274e-01 5.20821393e-01
3.24710310e-01 1.67913646e-01 7.85986558e-02 4.50460501e-02
-2.95308381e-02 -5.15213124e-02 7.65319705e-01 4.98889655e-01
-7.85227120e-01 -4.89856362e-01 -7.54772961e-01 4.21073020e-01
-2.07965806e-01 -6.17689788e-01 8.40013474e-02 2.51844138e-01
6.54740691e-01 1.06140804e+00 2.54846424e-01 -2.08156213e-01
2.20346063e-01 -4.22527194e-01 5.32214463e-01 -4.74852532e-01
-3.24494451e-01 3.08571197e-02 -3.12043309e-01 -3.59228909e-01
-6.47377729e-01 -1.29599318e-01 -8.40889513e-01 -2.64993101e-01
-1.91235300e-02 5.93389682e-02 5.06631792e-01 6.41379595e-01
-1.91197783e-01 8.32741737e-01 9.06941056e-01 -9.90076005e-01
2.44161963e-01 -1.08187044e+00 -1.28106594e+00 2.76627898e-01
1.08186793e+00 -4.00913179e-01 -5.96337914e-01 2.36584887e-01] | [10.919404029846191, -3.241971015930176] |
0c163e40-21b1-4910-9db0-f0cefee038d0 | smart-home-energy-management-vae-gan | 2305.08885 | null | https://arxiv.org/abs/2305.08885v1 | https://arxiv.org/pdf/2305.08885v1.pdf | Smart Home Energy Management: VAE-GAN synthetic dataset generator and Q-learning | Recent years have noticed an increasing interest among academia and industry towards analyzing the electrical consumption of residential buildings and employing smart home energy management systems (HEMS) to reduce household energy consumption and costs. HEMS has been developed to simulate the statistical and functional properties of actual smart grids. Access to publicly available datasets is a major challenge in this type of research. The potential of artificial HEMS applications will be further enhanced with the development of time series that represent different operating conditions of the synthetic systems. In this paper, we propose a novel variational auto-encoder-generative adversarial network (VAE-GAN) technique for generating time-series data on energy consumption in smart homes. We also explore how the generative model performs when combined with a Q-learning-based HEMS. We tested the online performance of Q-learning-based HEMS with real-world smart home data. To test the generated dataset, we measure the Kullback-Leibler (KL) divergence, maximum mean discrepancy (MMD), and the Wasserstein distance between the probability distributions of the real and synthetic data. Our experiments show that VAE-GAN-generated synthetic data closely matches the real data distribution. Finally, we show that the generated data allows for the training of a higher-performance Q-learning-based HEMS compared to datasets generated with baseline approaches. | ['Damla Turgut', 'Melike Erol-Kantarci', 'Hao Zhou', 'Mina Razghandi'] | 2023-05-14 | null | null | null | null | ['q-learning', 'energy-management'] | ['methodology', 'time-series'] | [-2.36158758e-01 1.02207832e-01 3.95753026e-01 -1.87451243e-01
-1.03609908e+00 -3.55312526e-01 3.48118484e-01 -2.36482069e-01
2.18585730e-01 1.14044166e+00 1.09509081e-01 -3.72044556e-02
2.46291980e-02 -1.37636888e+00 -5.26439011e-01 -1.05757749e+00
-1.12911403e-01 4.47842360e-01 -4.28259134e-01 -1.91076905e-01
-4.46520358e-01 2.96782166e-01 -1.37099099e+00 -6.42510355e-02
1.14737880e+00 1.01061690e+00 3.23936641e-02 6.24816895e-01
5.96098661e-01 7.25816011e-01 -1.00254571e+00 -8.07563588e-02
4.44514483e-01 -1.10685468e+00 -2.85883725e-01 -2.50840962e-01
-5.14417887e-01 -4.96419191e-01 -1.45827740e-01 1.01261139e+00
9.63278592e-01 1.40987933e-01 8.10943365e-01 -1.83029151e+00
-4.19379413e-01 4.25466508e-01 -2.69767046e-02 -9.55120102e-02
3.41122895e-01 6.47846580e-01 5.95431745e-01 -7.36862072e-04
6.06007874e-02 7.51258790e-01 5.86218596e-01 3.67530227e-01
-1.54183567e+00 -6.13386452e-01 -7.45778680e-01 5.03441751e-01
-1.48617375e+00 -6.23866841e-02 1.03108454e+00 -2.01953292e-01
1.11491942e+00 2.94863522e-01 1.31859410e+00 1.35024369e+00
3.21449608e-01 6.66225076e-01 1.32733858e+00 -2.84731895e-01
1.06509626e+00 1.23766772e-01 -6.80562139e-01 2.32630819e-01
5.89720048e-02 3.67567539e-01 4.65165377e-02 -1.36541784e-01
3.69545698e-01 -1.71800867e-01 -1.42306477e-01 -3.97529066e-01
-6.36815131e-01 1.00260258e+00 3.15463126e-01 4.18947667e-01
-6.83869660e-01 3.15071702e-01 4.11292762e-01 2.43401393e-01
4.34546620e-01 3.41033816e-01 -3.31756830e-01 -3.26507390e-01
-1.13800931e+00 1.99152976e-01 9.39814746e-01 8.77569258e-01
4.26053852e-01 6.26533508e-01 -4.04963195e-01 4.21488196e-01
2.62357980e-01 9.31074440e-01 7.86559582e-01 -9.94239926e-01
1.06300987e-01 5.06634116e-01 3.15404058e-01 -2.42077962e-01
-3.32242370e-01 -7.62031227e-02 -1.16651988e+00 3.57834697e-01
1.65906742e-01 -5.54727077e-01 -5.05851328e-01 1.66779935e+00
2.56567001e-01 -2.84903217e-02 1.47342980e-01 3.88104707e-01
2.70888451e-02 9.50891793e-01 5.10252174e-03 -5.13512254e-01
7.54826248e-01 -5.69199204e-01 -7.80728936e-01 5.96944034e-01
4.80610579e-01 -1.84942320e-01 1.03823602e+00 2.42737487e-01
-1.32650173e+00 -4.84719098e-01 -1.18014944e+00 6.54932022e-01
-4.30242807e-01 -3.01426053e-01 1.54194459e-02 1.09806502e+00
-1.16065764e+00 7.60996640e-01 -1.04784882e+00 -3.25606018e-01
6.24476135e-01 1.52168781e-01 3.75039250e-01 3.02979499e-01
-1.36344898e+00 9.35462773e-01 4.08506125e-01 -3.27313185e-01
-1.12211919e+00 -9.73327398e-01 -6.87844932e-01 2.47032478e-01
-9.07021984e-02 -6.92984045e-01 1.26518381e+00 -7.82250106e-01
-1.76853120e+00 1.38024509e-01 3.89108062e-01 -8.60811114e-01
7.29290903e-01 3.06591809e-01 -6.84367359e-01 -1.30198613e-01
-4.03424092e-02 3.07646096e-01 6.92436695e-01 -1.07390451e+00
-4.57755685e-01 -2.15978414e-01 -6.16172969e-01 -3.00384283e-01
-1.79372877e-01 -7.11594820e-01 7.32607603e-01 -6.76883936e-01
-6.75533891e-01 -6.65045142e-01 1.80819109e-02 -4.83741730e-01
-4.54062194e-01 -4.47076648e-01 8.73679519e-01 -8.54093552e-01
9.71749187e-01 -1.81451738e+00 -1.10980920e-01 3.73786151e-01
-4.85576510e-01 1.61500633e-01 6.35558292e-02 7.46604323e-01
-8.91514421e-02 1.56429797e-01 -3.90964419e-01 -2.49400306e-02
4.34621751e-01 2.53741026e-01 -1.73172839e-02 3.67371827e-01
3.71422097e-02 1.12258697e+00 -1.08973837e+00 -1.04339711e-01
6.93105578e-01 6.56190276e-01 -2.96600968e-01 5.53530276e-01
-3.84817570e-01 6.41687512e-01 -3.18932354e-01 3.35017622e-01
3.54735911e-01 -8.15596282e-02 2.86929697e-01 -3.94318759e-01
2.92310089e-01 -8.58652964e-02 -9.83146846e-01 1.50805080e+00
-8.57403755e-01 3.76331240e-01 -4.15588915e-01 -1.16427648e+00
8.19062948e-01 4.22859162e-01 1.05065691e+00 -1.41778064e+00
2.47789949e-01 -8.59033596e-03 -2.59734154e-01 -3.76760840e-01
-9.48943198e-02 -3.55648607e-01 -4.21520025e-02 6.73722744e-01
1.04678914e-01 -6.22999370e-01 1.54399619e-01 -2.22718358e-01
1.25616610e+00 2.09101260e-01 3.33423495e-01 -3.70292306e-01
3.60074192e-01 -3.17733139e-01 3.72031569e-01 1.37788609e-01
-9.12670717e-02 7.77139813e-02 2.99197048e-01 -2.30350018e-01
-1.51308894e+00 -1.48273373e+00 -8.88610408e-02 2.52871752e-01
-2.90619314e-01 -1.79422230e-01 -1.24332571e+00 -6.28812551e-01
-8.96236748e-02 1.86716211e+00 -5.37686825e-01 -4.22134221e-01
-2.75116473e-01 -1.07118297e+00 4.66699988e-01 5.64697504e-01
6.49465144e-01 -1.24461091e+00 -1.25410426e+00 4.08847332e-01
-6.16998300e-02 -7.22682178e-01 -1.15643956e-01 3.75152290e-01
-5.79394817e-01 -1.06120229e+00 -6.51173949e-01 -2.62341708e-01
4.50989962e-01 -7.16065407e-01 1.29781687e+00 -5.04671454e-01
-5.13760507e-01 5.48466921e-01 -4.71767247e-01 -4.16451335e-01
-9.91496503e-01 1.77948773e-02 1.42354900e-02 -2.47370332e-01
3.10444444e-01 -9.86331522e-01 -9.06710804e-01 3.03142488e-01
-9.35651362e-01 2.69907601e-02 1.26053527e-01 8.31658065e-01
6.20013714e-01 8.37964058e-01 9.74243939e-01 -3.94324005e-01
7.14637160e-01 -5.98205686e-01 -8.95827174e-01 3.54235739e-01
-1.08203018e+00 3.50308448e-01 1.15455651e+00 -2.86117762e-01
-9.68574882e-01 -7.85566047e-02 -3.15819949e-01 -1.95564032e-01
-1.58213452e-01 -2.10318461e-01 -6.10460043e-01 4.37654376e-01
3.70692790e-01 3.27955782e-01 -3.16243082e-01 -1.47199839e-01
3.89860094e-01 6.44484937e-01 3.63674670e-01 -4.55825925e-01
1.05867732e+00 1.78563833e-01 1.47345543e-01 -5.82119226e-01
-3.17588210e-01 3.50830764e-01 -4.04541582e-01 -2.43685111e-01
9.54621851e-01 -7.46099532e-01 -7.55774796e-01 6.03700995e-01
-7.24397242e-01 -9.61780846e-01 -1.09089625e+00 5.23812845e-02
-1.06475902e+00 -8.67500156e-02 -3.18959802e-01 -1.10013247e+00
-6.84935153e-01 -9.97270882e-01 8.50471318e-01 3.86359304e-01
-3.23334455e-01 -1.16326964e+00 4.04848993e-01 6.78862110e-02
6.13175392e-01 9.18834329e-01 1.08361113e+00 -3.88870865e-01
-3.36769909e-01 -9.30335149e-02 4.23439592e-01 8.42896521e-01
3.62979352e-01 -2.98109055e-01 -9.32089508e-01 -6.64942980e-01
2.75846541e-01 -3.17849517e-01 -1.97924282e-02 5.52878141e-01
1.16938901e+00 -5.88905275e-01 5.91051802e-02 4.69694078e-01
1.76152301e+00 7.55614758e-01 9.52094853e-01 1.77925557e-01
2.30687201e-01 1.90632135e-01 2.54714668e-01 8.73681426e-01
3.14669073e-01 5.30860901e-01 4.51813608e-01 1.77652463e-02
1.74946725e-01 -6.36228681e-01 5.09954333e-01 7.32134402e-01
2.09744647e-01 -4.49718714e-01 -5.51843286e-01 5.36432445e-01
-1.48002076e+00 -1.13522220e+00 2.99097925e-01 2.25244331e+00
7.04890728e-01 -7.47563392e-02 4.42535430e-01 5.14768362e-01
3.55630755e-01 -1.96080253e-01 -1.08294249e+00 -3.54987383e-01
-2.90093943e-02 7.63217330e-01 4.78112698e-01 4.00006995e-02
-6.17136061e-01 1.42119884e-01 5.90656281e+00 8.18562806e-01
-8.58901799e-01 3.14180732e-01 8.14959168e-01 -1.58810526e-01
-4.47952777e-01 -5.13822973e-01 -7.67612085e-02 1.05033755e+00
1.67935216e+00 -7.54298568e-01 8.86751831e-01 8.92581642e-01
5.89866579e-01 -1.48559779e-01 -1.39160323e+00 1.00503683e+00
-2.53682375e-01 -8.52267981e-01 -3.69834244e-01 1.92900911e-01
1.17370236e+00 -2.45135259e-02 -6.12908378e-02 3.29320192e-01
7.79445827e-01 -1.16502178e+00 5.46718836e-01 6.32709265e-01
8.26915145e-01 -1.11660802e+00 9.42647159e-01 4.03469294e-01
-1.27872825e+00 -1.16630778e-01 -1.75322592e-02 3.62481952e-01
3.69132400e-01 6.07613564e-01 -9.08489048e-01 6.64237261e-01
8.02341759e-01 2.71396250e-01 -4.44692224e-01 6.82893693e-01
-2.68390000e-01 8.29455376e-01 -4.36946213e-01 -3.35788727e-01
-2.83570498e-01 -5.72357714e-01 8.25669914e-02 7.32620299e-01
7.62665331e-01 -6.02729470e-02 -2.02960297e-01 1.37716210e+00
-5.35453781e-02 -1.39409423e-01 -4.68069106e-01 -1.24570727e-01
4.94463354e-01 1.11548734e+00 -4.45004106e-01 -2.96887726e-01
-1.58826515e-01 1.12482405e+00 -3.19740027e-01 3.41709644e-01
-1.10111451e+00 -1.56030938e-01 4.09922391e-01 1.92554086e-01
3.17865252e-01 8.54190961e-02 -1.03208773e-01 -8.74714136e-01
-1.10302202e-01 -7.70829976e-01 2.00672597e-01 -1.02617431e+00
-1.62927449e+00 3.11021477e-01 1.85477585e-01 -9.78883207e-01
-9.28692400e-01 -2.11444404e-02 -8.21866095e-01 9.82329428e-01
-1.18052769e+00 -8.31235230e-01 -4.74950761e-01 6.79929852e-01
5.56479633e-01 -1.89757660e-01 1.12738454e+00 1.54003248e-01
-4.33538556e-01 4.89501953e-01 7.26567328e-01 -5.86580411e-02
-2.37827469e-02 -1.55754197e+00 4.66304392e-01 5.54597676e-01
-1.24447107e-01 -4.36527371e-01 8.90589833e-01 -4.44979787e-01
-1.20942688e+00 -1.40268362e+00 -7.37590119e-02 -1.39844328e-01
5.56779027e-01 -2.72262514e-01 -6.37315035e-01 3.68162215e-01
7.44957864e-01 -2.92874098e-01 6.95234716e-01 -7.87401736e-01
3.37879986e-01 -3.78088534e-01 -1.94826198e+00 2.60492295e-01
5.93997419e-01 -4.78742093e-01 -2.78079599e-01 3.16250533e-01
2.82412648e-01 2.22770527e-01 -1.42331958e+00 3.58955652e-01
3.03305715e-01 -1.13127136e+00 7.16963530e-01 -2.05373734e-01
1.71597421e-01 -2.30698600e-01 -3.48196447e-01 -2.13910508e+00
3.91217618e-04 -6.12035573e-01 -5.08369148e-01 1.40447390e+00
-1.68989412e-03 -6.61600530e-01 8.34122896e-01 5.85923433e-01
3.15043718e-01 -7.99859047e-01 -1.11857009e+00 -9.10847545e-01
4.37929481e-01 -2.72556096e-01 1.18926084e+00 5.35377204e-01
-1.28648624e-01 4.72757518e-02 -1.27979457e-01 -4.89188842e-02
1.06474018e+00 -1.33997008e-01 5.51223338e-01 -8.57419491e-01
-2.77932078e-01 -1.76032245e-01 -3.63473177e-01 -1.94413915e-01
1.55061424e-01 -6.54884696e-01 -6.52164593e-02 -1.61826968e+00
-3.53921577e-02 -1.06927872e-01 -3.94934118e-01 2.06577793e-01
1.80293128e-01 -1.24908194e-01 1.11000247e-01 -3.72516632e-01
-1.36484519e-01 1.32293570e+00 8.30812454e-01 -3.91311973e-01
-2.35692099e-01 1.49317637e-01 -4.92263064e-02 3.62165034e-01
1.22350895e+00 -3.79287809e-01 -7.44098663e-01 2.91671962e-01
-1.80978905e-02 2.52102286e-01 4.42670256e-01 -1.41462874e+00
-1.21522516e-01 9.68933944e-03 6.65747344e-01 -6.18523061e-01
-4.89263982e-02 -1.11758506e+00 8.78837764e-01 8.42648566e-01
9.68462974e-03 2.55387098e-01 1.47043774e-02 3.52862298e-01
1.58850297e-01 1.67627677e-01 9.70799506e-01 1.00313416e-02
-1.80815354e-01 9.95422155e-02 -4.16763276e-01 1.14556305e-01
1.46939588e+00 1.72560401e-02 6.51681423e-02 -5.97418547e-01
-4.53953087e-01 2.58634061e-01 6.01572514e-01 1.44136995e-01
3.00662696e-01 -1.58895242e+00 -6.09315991e-01 3.32844317e-01
-2.21066996e-01 -1.09013781e-01 2.03986973e-01 3.09217960e-01
-2.50026882e-01 1.60807252e-01 -3.14601332e-01 -4.00214374e-01
-6.96138918e-01 6.25223279e-01 8.34626377e-01 -4.58813995e-01
-5.75997114e-01 -9.53394026e-02 -3.76545340e-01 -3.12378824e-01
-6.79284185e-02 -3.56853694e-01 3.66101146e-01 -1.23846997e-02
1.79744303e-01 8.62317204e-01 1.69144809e-01 -2.78602391e-01
-6.35273531e-02 2.51025129e-02 8.49676549e-01 -4.19991240e-02
1.63214576e+00 1.24354064e-01 2.65151531e-01 5.10105908e-01
1.29873979e+00 -5.45512378e-01 -1.33426535e+00 4.08707380e-01
-9.99245122e-02 4.65762764e-02 6.86492622e-02 -9.93794918e-01
-1.34931338e+00 5.20321250e-01 1.38059497e+00 8.55576158e-01
1.61921942e+00 -2.68452615e-01 1.02679431e+00 -3.94375576e-03
8.24450970e-01 -1.41529739e+00 -6.47484437e-02 -3.69209200e-01
5.80955207e-01 -8.93499732e-01 -2.61597991e-01 3.40571433e-01
-4.53134805e-01 7.69979239e-01 2.62936771e-01 -8.13857391e-02
7.53380179e-01 6.23205662e-01 -2.54148096e-01 1.96965203e-01
-5.21649778e-01 9.00160819e-02 -1.96678951e-01 1.06831145e+00
1.07116178e-01 4.18551564e-01 -1.83988060e-03 5.95630407e-01
-4.92869854e-01 2.58517355e-01 3.93515885e-01 7.30434537e-01
8.79169405e-02 -1.16559505e+00 -3.02931666e-01 5.89751661e-01
-1.49085686e-01 2.81511188e-01 2.97997206e-01 7.02020228e-01
1.56629354e-01 1.11634934e+00 -1.59819201e-02 -2.31928036e-01
5.61304390e-01 1.95386931e-01 5.29508412e-01 -7.35687912e-02
-4.67962921e-01 -1.99372008e-01 -3.48445743e-01 -5.48367977e-01
-3.13090920e-01 -7.28739679e-01 -1.13926041e+00 -4.65088487e-01
-9.71703976e-02 3.44771206e-01 7.99949050e-01 8.34909558e-01
7.90304169e-02 1.02369571e+00 1.23724055e+00 -6.20925248e-01
-9.68652129e-01 -1.03051531e+00 -9.64118540e-01 7.17585504e-01
-4.95557263e-02 -2.32689232e-01 -4.67451125e-01 5.01411874e-03] | [16.044097900390625, 7.5634446144104] |
54c83505-15aa-45fc-ad39-6c142f148b06 | relation-guided-pre-training-for-open-domain | 2109.10346 | null | https://arxiv.org/abs/2109.10346v1 | https://arxiv.org/pdf/2109.10346v1.pdf | Relation-Guided Pre-Training for Open-Domain Question Answering | Answering complex open-domain questions requires understanding the latent relations between involving entities. However, we found that the existing QA datasets are extremely imbalanced in some types of relations, which hurts the generalization performance over questions with long-tail relations. To remedy this problem, in this paper, we propose a Relation-Guided Pre-Training (RGPT-QA) framework. We first generate a relational QA dataset covering a wide range of relations from both the Wikidata triplets and Wikipedia hyperlinks. We then pre-train a QA model to infer the latent relations from the question, and then conduct extractive QA to get the target answer entity. We demonstrate that by pretraining with propoed RGPT-QA techique, the popular open-domain QA model, Dense Passage Retriever (DPR), achieves 2.2%, 2.4%, and 6.3% absolute improvement in Exact Match accuracy on Natural Questions, TriviaQA, and WebQuestions. Particularly, we show that RGPT-QA improves significantly on questions with long-tail relations | ['Kai-Wei Chang', 'Yizhou Sun', 'Ziniu Hu'] | 2021-09-21 | null | https://aclanthology.org/2021.findings-emnlp.292 | https://aclanthology.org/2021.findings-emnlp.292.pdf | findings-emnlp-2021-11 | ['triviaqa'] | ['miscellaneous'] | [-1.94657192e-01 5.58330715e-01 -1.04590409e-01 -3.13587368e-01
-1.51612723e+00 -7.49648273e-01 3.42341036e-01 1.86864257e-01
-1.71444416e-01 1.09382606e+00 4.72909719e-01 -5.59089243e-01
-4.68705267e-01 -1.25880039e+00 -9.51089501e-01 -1.14942767e-01
1.49183333e-01 1.13936567e+00 5.41541100e-01 -8.73514950e-01
-5.49883731e-02 -1.46563485e-01 -1.02150095e+00 6.84625626e-01
1.79800844e+00 8.86749208e-01 -2.70043075e-01 4.77023274e-01
-4.79397744e-01 1.22682416e+00 -6.44116282e-01 -1.26451468e+00
2.92892158e-02 -3.54557812e-01 -1.49883032e+00 -6.88946426e-01
6.59203231e-01 -1.54851273e-01 -6.20791197e-01 8.36139739e-01
4.76736605e-01 7.69470260e-02 8.06283772e-01 -1.09620738e+00
-9.60272372e-01 9.13633168e-01 -2.50595540e-01 1.60150796e-01
8.76936972e-01 -1.82681710e-01 1.77191174e+00 -7.31784463e-01
8.19094360e-01 1.22817433e+00 7.02297866e-01 4.54564661e-01
-8.43840778e-01 -5.07832229e-01 -3.84865880e-01 6.35541022e-01
-1.19044721e+00 -2.00558156e-01 4.33913440e-01 -5.57689928e-02
1.19658649e+00 3.67813915e-01 -1.14866439e-02 8.06303978e-01
5.44502363e-02 6.13553822e-01 9.42826092e-01 -4.66716379e-01
-3.73710454e-01 -1.47068813e-01 5.38804233e-01 7.12848544e-01
1.83750868e-01 -5.08731425e-01 -4.21930671e-01 -3.58107716e-01
9.49825644e-02 -6.16338372e-01 -4.63466883e-01 1.38614606e-03
-9.63749290e-01 7.29727387e-01 8.22453439e-01 -6.21081777e-02
-1.77041918e-01 -3.06224316e-01 3.63472104e-01 8.97424161e-01
4.03752148e-01 1.00155008e+00 -7.80191839e-01 -1.08906157e-01
-2.18909740e-01 3.30298275e-01 1.32333159e+00 1.01971114e+00
9.37874794e-01 -9.32109773e-01 -3.59350741e-01 1.14646685e+00
2.92175472e-01 6.05759978e-01 2.98486114e-01 -1.08470345e+00
1.37430108e+00 9.84076440e-01 9.09338668e-02 -1.06600273e+00
-1.93187654e-01 -2.65749633e-01 -5.59042573e-01 -7.62066722e-01
6.57241762e-01 -2.60784060e-01 -7.20106781e-01 1.47364616e+00
5.80943644e-01 -4.96079803e-01 5.51220715e-01 7.09618747e-01
1.45524228e+00 8.19871604e-01 3.47089581e-02 1.25864819e-01
1.53131425e+00 -1.30791712e+00 -9.47091699e-01 -2.52960950e-01
1.00645387e+00 -7.27241457e-01 1.32134569e+00 5.64367278e-03
-9.52694356e-01 -2.61942953e-01 -7.65765071e-01 -7.55931556e-01
-3.18432540e-01 -1.64882571e-01 3.94209146e-01 2.81916887e-01
-6.29453838e-01 3.35287005e-01 -2.64378279e-01 -2.55818784e-01
2.05756843e-01 7.26024956e-02 -4.64179128e-01 -5.60824692e-01
-1.97876537e+00 1.30860937e+00 2.19518587e-01 -6.75734505e-02
-6.14341438e-01 -9.19759512e-01 -7.98217356e-01 2.78475463e-01
6.75330341e-01 -9.73968506e-01 1.20047355e+00 -3.99173051e-02
-1.25138581e+00 7.85865963e-01 -2.26817653e-01 -3.26509088e-01
1.67360872e-01 -6.89760864e-01 -6.17064059e-01 2.47040033e-01
3.57373267e-01 3.03379238e-01 1.18851773e-01 -9.42389071e-01
-2.01517016e-01 -3.53594422e-01 5.83544314e-01 5.18432677e-01
-2.21218958e-01 -1.14318945e-01 -3.72663528e-01 -6.00016601e-02
1.21211790e-01 -5.92367232e-01 1.19976334e-01 -4.47586030e-01
-5.73406279e-01 -8.75201643e-01 2.41691172e-01 -1.00996387e+00
1.21519434e+00 -1.62948954e+00 8.78017917e-02 6.48581982e-02
3.74673724e-01 3.88316453e-01 -3.87783140e-01 8.10725093e-01
1.47359729e-01 -8.05676281e-02 -6.30296022e-02 2.53905177e-01
8.60489160e-02 4.61354911e-01 -6.60974026e-01 -2.45558038e-01
1.95396721e-01 1.22926819e+00 -1.06248319e+00 -7.71238983e-01
-6.85456872e-01 -1.40503213e-01 -4.96112227e-01 5.94119012e-01
-8.02727282e-01 3.99969444e-02 -6.88201487e-01 5.39243400e-01
4.63012546e-01 -5.65395951e-01 1.44089147e-01 -4.18813795e-01
5.34104466e-01 1.34756541e+00 -5.27473271e-01 1.83974159e+00
-5.15556157e-01 5.13063908e-01 -3.95731539e-01 -6.04968667e-01
1.09989822e+00 3.82918954e-01 1.34159401e-01 -1.07542634e+00
-1.96443617e-01 4.76973653e-01 1.64087623e-01 -1.07388830e+00
7.09644735e-01 2.95260772e-02 -6.93827271e-02 5.67871451e-01
3.81029874e-01 -2.30219766e-01 4.68173295e-01 8.07601690e-01
1.44889295e+00 1.21905478e-02 -5.28577864e-02 -3.84978205e-02
5.11984348e-01 3.85582507e-01 3.61792266e-01 4.99345869e-01
1.48892760e-01 4.73365128e-01 7.52079368e-01 -2.76686307e-02
-7.08548784e-01 -1.29255652e+00 -7.45759830e-02 1.06304538e+00
2.00371191e-01 -7.32924640e-01 -5.44772446e-01 -1.11832845e+00
-1.27017554e-02 7.91718900e-01 -4.79288727e-01 -2.78741211e-01
-8.88920844e-01 -5.56201637e-01 8.45233798e-01 2.78910726e-01
7.68765867e-01 -9.01459992e-01 2.07814917e-01 1.10321201e-01
-1.13027465e+00 -1.21863711e+00 -2.10484520e-01 -1.52774408e-01
-6.46078169e-01 -1.32698154e+00 -3.76869351e-01 -7.10292041e-01
5.06825864e-01 1.09245054e-01 1.92511618e+00 1.03337117e-01
3.41962695e-01 2.05534667e-01 -7.00140059e-01 3.97505835e-02
-2.78896302e-01 5.86908400e-01 -4.15968478e-01 -4.93904144e-01
7.13944316e-01 -5.40579855e-01 -6.11872196e-01 3.55375111e-01
-6.00103021e-01 -2.12284982e-01 4.93339211e-01 7.73689866e-01
2.99838096e-01 -1.72101602e-01 1.00103593e+00 -1.25448060e+00
1.06740284e+00 -8.38860512e-01 -2.92758346e-01 1.03328776e+00
-5.20076692e-01 3.17478836e-01 5.69737434e-01 -9.90968384e-03
-1.41856825e+00 -6.78185999e-01 -4.46498394e-01 2.67971516e-01
2.16230139e-01 8.94340456e-01 -4.33084488e-01 1.78262755e-01
1.15240526e+00 -1.73563495e-01 -3.31672311e-01 -4.75511938e-01
8.58060956e-01 9.17724550e-01 5.36112845e-01 -1.00965285e+00
9.41161275e-01 2.46918239e-02 -3.82431298e-01 -1.17468767e-01
-1.61201346e+00 -5.04075885e-01 -3.69493037e-01 2.17757627e-01
6.87976718e-01 -9.70402241e-01 -4.70417202e-01 -5.69084485e-04
-1.28746474e+00 -3.72739494e-01 -3.60117674e-01 1.84744045e-01
-2.48268411e-01 3.11305553e-01 -8.94422293e-01 -2.52680719e-01
-8.24829578e-01 -5.89025855e-01 7.09637344e-01 2.96381503e-01
-2.33083695e-01 -9.59826231e-01 5.49855947e-01 1.25030410e+00
2.39172891e-01 -2.40566045e-01 1.31389976e+00 -9.70838785e-01
-7.42839515e-01 -1.22679941e-01 -5.43297470e-01 1.77575424e-01
1.14290766e-01 -3.18367779e-01 -6.03951216e-01 1.90523520e-01
-2.20826551e-01 -1.02354360e+00 5.62860072e-01 -5.00145018e-01
7.06808686e-01 -4.21725929e-01 -1.73222885e-01 1.63487807e-01
1.07623065e+00 -1.58358589e-01 9.40705597e-01 4.03792769e-01
7.30002701e-01 8.76733661e-01 8.79989862e-01 -3.39416295e-01
1.15038955e+00 4.40369636e-01 1.92500442e-01 3.56351703e-01
-3.71965706e-01 -6.86939240e-01 7.92183727e-02 1.37030613e+00
1.17424153e-01 -3.23517084e-01 -1.00945199e+00 8.39518666e-01
-1.69605994e+00 -5.75216651e-01 -4.82945561e-01 1.69810641e+00
1.65958703e+00 -9.67956260e-02 -2.23675594e-01 -2.31571391e-01
3.75293612e-01 -1.50347918e-01 -3.61671895e-01 -1.62364736e-01
-2.24602088e-01 4.15512502e-01 3.53621654e-02 6.06296778e-01
-5.16350269e-01 9.91404295e-01 5.57048321e+00 9.14669394e-01
-2.66207039e-01 1.86352730e-01 2.77893901e-01 4.27379251e-01
-7.92707026e-01 1.37433514e-01 -9.01495516e-01 1.01921484e-01
9.64895964e-01 -2.76028216e-01 2.32307255e-01 4.52366024e-01
-6.83198810e-01 6.32483140e-02 -9.84304488e-01 4.34205204e-01
2.06225976e-01 -1.09121966e+00 2.46243894e-01 -3.08647603e-01
8.91950190e-01 1.40819520e-01 -1.99756712e-01 8.86943042e-01
8.41104209e-01 -1.12932336e+00 -9.58171859e-02 3.50336790e-01
6.70817196e-01 -4.50680494e-01 9.14494574e-01 3.98883760e-01
-8.03276539e-01 1.50984317e-01 -6.58329785e-01 2.62865927e-02
2.26402760e-01 6.94716096e-01 -8.31323624e-01 1.13432777e+00
7.21855462e-01 3.82904828e-01 -7.01319754e-01 9.96828794e-01
-8.65833044e-01 8.20620477e-01 -2.69865662e-01 -1.64361507e-01
4.04931083e-02 -2.32343838e-01 1.73153296e-01 6.87670290e-01
1.81587696e-01 4.41578060e-01 -4.20929939e-01 6.00152671e-01
-7.07546711e-01 1.61839142e-01 -2.81344593e-01 2.73964982e-02
7.24311292e-01 1.27694571e+00 1.46262065e-01 -4.52352494e-01
-5.24367511e-01 7.04041362e-01 1.14862943e+00 3.92937601e-01
-6.78855300e-01 -8.40538144e-01 3.03337932e-01 -9.05603096e-02
8.61400217e-02 1.33182108e-01 -5.76587045e-04 -1.54188907e+00
4.23137158e-01 -1.27481425e+00 1.01794899e+00 -9.97236431e-01
-1.83108449e+00 7.54486740e-01 -1.75903440e-01 -9.17242765e-01
-2.93192536e-01 -3.67590129e-01 -4.06109661e-01 9.86659765e-01
-1.70584261e+00 -1.05301452e+00 -3.46666068e-01 6.64778769e-01
1.32281244e-01 1.05807729e-01 9.04477596e-01 6.60081685e-01
-2.28673160e-01 6.93198979e-01 -4.70241122e-02 4.76688206e-01
1.08942091e+00 -1.47398758e+00 4.90119129e-01 4.33246970e-01
3.98905128e-01 9.16340649e-01 5.96258640e-01 -6.87827766e-01
-1.28902125e+00 -9.78596032e-01 1.50106823e+00 -9.86718297e-01
9.75519896e-01 -1.99132308e-01 -1.46660233e+00 8.90232980e-01
6.62436843e-01 -1.45829856e-01 7.95413494e-01 6.85131431e-01
-9.55214739e-01 -2.84061521e-01 -1.00169241e+00 6.36567891e-01
1.10004342e+00 -8.01816463e-01 -1.49396026e+00 7.78736532e-01
1.25650263e+00 -7.29845524e-01 -1.35870600e+00 7.19147444e-01
1.15308754e-01 -5.36187530e-01 9.99870121e-01 -1.03794253e+00
7.44108140e-01 -1.92141175e-01 -6.28245473e-02 -1.49399316e+00
3.05102039e-02 -4.48849261e-01 -5.58265328e-01 1.48215985e+00
1.03089225e+00 -6.86303973e-01 6.57246649e-01 6.90832555e-01
-2.15368107e-01 -7.99930155e-01 -8.67079616e-01 -5.96172333e-01
5.28246224e-01 -1.14337681e-02 6.68293595e-01 1.10265398e+00
4.62899417e-01 1.10640180e+00 -2.07090937e-02 1.86267167e-01
3.26447010e-01 4.77836639e-01 9.36544955e-01 -9.33166802e-01
-3.02827418e-01 2.22867563e-01 1.53359786e-01 -1.64755011e+00
1.14695318e-01 -9.02124107e-01 -4.28036116e-02 -2.02478576e+00
5.39317839e-02 -8.13717902e-01 5.93954362e-02 2.61206836e-01
-7.67189920e-01 5.42515665e-02 -3.20285320e-01 1.72782093e-01
-1.07869470e+00 7.86821783e-01 1.60013616e+00 -1.60141215e-01
2.69331455e-01 -1.25511378e-01 -8.41630340e-01 1.56750113e-01
6.30912185e-01 -5.68786561e-01 -6.34989083e-01 -9.19290900e-01
1.08081925e+00 3.00250649e-01 3.44364122e-02 -5.86477757e-01
4.98206705e-01 1.60308912e-01 -1.36204228e-01 -6.12754345e-01
4.33739245e-01 -5.55647135e-01 -3.65208179e-01 1.16027921e-01
-6.64402068e-01 8.22577104e-02 -1.17376268e-01 3.57042432e-01
-5.26243329e-01 -3.80933791e-01 -4.58830073e-02 -3.85916345e-02
-2.04331517e-01 1.46381646e-01 2.12783828e-01 1.13210297e+00
3.98213744e-01 5.86242557e-01 -1.30323100e+00 -6.00316465e-01
-4.89207715e-01 8.42728913e-01 -1.27830982e-01 2.99221843e-01
3.49611104e-01 -1.27092612e+00 -9.49211061e-01 -4.33417618e-01
2.97646582e-01 3.90625358e-01 3.65510672e-01 6.61056399e-01
-6.02561176e-01 6.67092919e-01 3.03235650e-02 -1.52074113e-01
-1.00219989e+00 3.91348213e-01 3.06639552e-01 -9.14965570e-01
-2.73891628e-01 1.07705665e+00 -1.77311778e-01 -1.16648030e+00
-1.27019331e-01 -1.14004277e-01 -4.57911372e-01 8.74112472e-02
3.33638906e-01 3.30353349e-01 3.30827951e-01 -3.41012269e-01
-4.14873920e-02 4.73042220e-01 -3.93394500e-01 1.28019452e-01
1.08913207e+00 -2.19366372e-01 -5.96731544e-01 9.58684906e-02
1.15483975e+00 2.62287736e-01 -6.52234197e-01 -7.74007261e-01
2.91290790e-01 -2.61993647e-01 -7.31994212e-01 -1.18343711e+00
-7.32232630e-01 9.09975350e-01 -2.07786486e-01 4.70537543e-01
9.00553524e-01 3.70976627e-01 1.43931639e+00 9.10845637e-01
4.08189327e-01 -6.24360859e-01 2.97179222e-01 8.44332755e-01
9.07653093e-01 -1.10489702e+00 -7.89836198e-02 -8.33949685e-01
-6.01439297e-01 7.81060338e-01 1.07182527e+00 8.69977623e-02
2.66379654e-01 -2.58556753e-01 4.15564716e-01 -6.27601027e-01
-1.05899727e+00 -2.99713612e-01 5.51730335e-01 5.69543958e-01
4.36041802e-01 -1.68704167e-01 -4.09150481e-01 5.02124071e-01
-6.75000191e-01 -3.10776383e-01 3.17563206e-01 5.53696871e-01
-4.00930703e-01 -1.20752120e+00 2.20607524e-03 6.02634251e-01
-4.56953019e-01 -5.11988163e-01 -4.99218643e-01 6.06805503e-01
-2.96736479e-01 1.14766955e+00 -1.94430172e-01 -3.63717556e-01
5.63556552e-01 2.04202875e-01 5.73169291e-01 -6.07343256e-01
-6.51826501e-01 -9.04638290e-01 9.26286817e-01 -2.98099667e-01
-2.59859592e-01 -1.23716727e-01 -1.17477846e+00 -2.07068428e-01
-7.25705206e-01 9.71773028e-01 1.21107802e-01 1.12821829e+00
5.93067110e-01 4.92336035e-01 2.36964807e-01 6.87661767e-01
-7.03686655e-01 -1.28013551e+00 -4.01837267e-02 6.26895547e-01
-1.58052444e-02 -4.55822378e-01 -3.92119884e-01 -3.77344489e-01] | [10.712722778320312, 7.96984338760376] |
84d340a6-8cf2-4329-a3c4-c21c9fe5ed31 | synscapes-a-photorealistic-synthetic-dataset | 1810.08705 | null | http://arxiv.org/abs/1810.08705v1 | http://arxiv.org/pdf/1810.08705v1.pdf | Synscapes: A Photorealistic Synthetic Dataset for Street Scene Parsing | We introduce Synscapes -- a synthetic dataset for street scene parsing
created using photorealistic rendering techniques, and show state-of-the-art
results for training and validation as well as new types of analysis. We study
the behavior of networks trained on real data when performing inference on
synthetic data: a key factor in determining the equivalence of simulation
environments. We also compare the behavior of networks trained on synthetic
data and evaluated on real-world data. Additionally, by analyzing pre-trained,
existing segmentation and detection models, we illustrate how uncorrelated
images along with a detailed set of annotations open up new avenues for
analysis of computer vision systems, providing fine-grain information about how
a model's performance changes according to factors such as distance, occlusion
and relative object orientation. | ['Magnus Wrenninge', 'Jonas Unger'] | 2018-10-19 | null | null | null | null | ['street-scene-parsing'] | ['computer-vision'] | [ 4.44408685e-01 4.67891321e-02 2.45478109e-01 -6.54048681e-01
-5.73462427e-01 -8.13472092e-01 8.06656182e-01 4.26210165e-02
-4.06757951e-01 5.51734865e-01 -1.65538132e-01 -6.19248450e-01
2.04760313e-01 -7.52553880e-01 -9.99288559e-01 -4.37908620e-01
-3.69547933e-01 5.56094050e-01 7.52963841e-01 -2.51384318e-01
1.64534092e-01 9.97093618e-01 -1.79704940e+00 3.20776910e-01
4.98242855e-01 7.04787910e-01 -1.23490147e-01 1.14655113e+00
3.24885070e-01 5.29766977e-01 -9.47486222e-01 -3.24063867e-01
5.33619583e-01 -5.33268005e-02 -5.08599520e-01 1.65052816e-01
9.69241500e-01 -2.85335720e-01 -3.61050814e-01 9.99296546e-01
2.97074884e-01 -2.23870948e-03 6.89063966e-01 -1.26925540e+00
-8.46704245e-02 1.84459269e-01 -3.62109661e-01 4.91966188e-01
8.70851651e-02 7.61243284e-01 4.52270448e-01 -3.61216366e-01
7.77227402e-01 1.34270442e+00 9.88583505e-01 2.44494632e-01
-1.49877644e+00 -4.74273086e-01 6.70582578e-02 -1.82689071e-01
-1.06220293e+00 -4.67779368e-01 5.20007789e-01 -5.78390062e-01
8.10397029e-01 2.86877662e-01 6.61036253e-01 1.17381144e+00
9.42854807e-02 5.39333940e-01 1.22301257e+00 -4.44691062e-01
3.24204862e-01 1.53959105e-02 7.62455836e-02 6.73037767e-01
2.98592210e-01 4.74719435e-01 9.42683965e-02 -2.55905967e-02
1.02688575e+00 -7.86131799e-01 -1.49730250e-01 -7.34955072e-01
-1.23810232e+00 5.59678197e-01 3.48262429e-01 -9.04125795e-02
1.90208077e-01 2.99822569e-01 5.42865634e-01 9.81921628e-02
2.52842069e-01 5.99147737e-01 -5.18632531e-01 8.89624190e-03
-8.36126626e-01 4.92965102e-01 7.83270776e-01 9.39796686e-01
5.44902265e-01 3.87272745e-01 1.48587292e-02 6.09311938e-01
6.88892752e-02 3.97805929e-01 4.19271663e-02 -1.41574872e+00
4.40654784e-01 3.74847710e-01 1.44878834e-01 -9.52687263e-01
-5.59127927e-01 -3.46615940e-01 -4.96126980e-01 8.52972329e-01
8.92825961e-01 -1.85443848e-01 -1.03908849e+00 1.50256264e+00
1.85144991e-01 2.12831393e-01 2.64814217e-02 5.30945957e-01
4.95417506e-01 3.25905293e-01 -1.26292035e-02 1.77549794e-01
1.02481484e+00 -7.72946954e-01 1.25717312e-01 -4.29306626e-01
5.54789782e-01 -8.03861678e-01 1.17962873e+00 3.51484805e-01
-9.72117245e-01 -8.08203638e-01 -1.16141093e+00 1.67518571e-01
-6.12602174e-01 1.07816115e-01 4.49969530e-01 9.02291179e-01
-1.01405919e+00 8.62282455e-01 -9.02943909e-01 -5.74549377e-01
5.35240710e-01 2.59148866e-01 -5.04769683e-02 2.07975551e-01
-7.59395003e-01 1.08287716e+00 3.54864925e-01 -1.57500673e-02
-8.93028319e-01 -8.21014225e-01 -8.56770694e-01 -2.25120813e-01
9.80139151e-02 -4.76040870e-01 1.58970404e+00 -6.58612311e-01
-1.24587297e+00 1.18497324e+00 2.82868147e-01 -7.34781742e-01
9.72174406e-01 2.14410961e-01 -7.07679212e-01 4.02230471e-02
8.82852525e-02 8.60671043e-01 4.19329792e-01 -1.70364094e+00
-7.01895773e-01 -3.08397621e-01 3.59434068e-01 1.79244261e-02
5.26311100e-01 -1.03646349e-02 -4.20391977e-01 -3.83617610e-01
-1.80776924e-01 -7.92884946e-01 -5.75134337e-01 2.94080168e-01
-6.95253491e-01 4.46774513e-01 9.91322219e-01 -5.60195088e-01
5.52149951e-01 -1.96156240e+00 -6.03451014e-01 3.27880651e-01
5.52463625e-03 4.61488068e-01 -1.83767766e-01 2.65357703e-01
-3.40834796e-01 3.14250499e-01 -2.44897202e-01 -1.12218454e-01
-2.36801691e-02 3.82339507e-02 -3.11468482e-01 5.49304962e-01
1.86921135e-01 7.06732869e-01 -7.69490719e-01 -4.54272807e-01
8.43241334e-01 2.45372608e-01 -1.66692436e-01 1.03116311e-01
-3.75798166e-01 6.34172082e-01 -2.28568643e-01 3.34085494e-01
7.35494554e-01 -9.20185372e-02 -5.73598519e-02 -2.14420080e-01
-2.04376787e-01 1.16172887e-01 -1.06103921e+00 1.18642509e+00
-5.68998456e-01 1.25259852e+00 -2.21246749e-01 -6.98271930e-01
8.37432683e-01 -2.70885974e-01 4.97131646e-02 -9.36153829e-01
4.17253673e-01 -1.55618116e-01 1.36143994e-02 -4.65604901e-01
5.12182832e-01 3.79586518e-01 1.13734566e-01 3.46705347e-01
-2.87096411e-01 -6.45128846e-01 4.78116959e-01 4.60184133e-03
1.14994681e+00 2.26995140e-01 5.85581250e-02 -3.37207347e-01
1.27306461e-01 4.15767998e-01 1.00059070e-01 8.33963573e-01
-1.60534099e-01 8.85948598e-01 6.68866575e-01 -8.33226085e-01
-1.69199789e+00 -1.22362483e+00 -3.87050807e-01 5.94507933e-01
4.07111347e-01 -8.89557973e-03 -9.97850180e-01 -5.18091857e-01
5.82316890e-03 1.07112443e+00 -1.08313262e+00 4.07641456e-02
-6.32771313e-01 -1.06396198e+00 8.70261788e-01 7.31702983e-01
6.89114630e-01 -9.92842674e-01 -1.27593195e+00 -8.79251659e-02
3.44342947e-01 -1.64807475e+00 2.13857085e-01 2.09906489e-01
-7.79398322e-01 -1.44907641e+00 -2.87994444e-01 -5.71849406e-01
6.01393998e-01 2.84189343e-01 1.62222290e+00 2.18892798e-01
-5.73083282e-01 2.77168959e-01 -1.54091632e-02 -3.30791205e-01
-8.90866220e-01 -1.11510880e-01 -3.45393330e-01 -4.62066442e-01
-1.00682564e-01 -6.17535651e-01 -6.66580975e-01 7.15470612e-01
-8.66642356e-01 2.54020095e-01 1.75349876e-01 3.91220868e-01
5.33593714e-01 7.66618177e-02 -2.08169103e-01 -1.08497751e+00
5.17853975e-01 1.60492748e-01 -1.22048640e+00 4.09271717e-01
-2.99020648e-01 8.31903815e-02 6.71653628e-01 -1.73792616e-01
-1.21184802e+00 -1.57282259e-02 -8.41586217e-02 -3.19074363e-01
-6.27188325e-01 -3.18764985e-01 -2.66209859e-02 -4.49174434e-01
1.11163628e+00 -3.01260918e-01 -4.06051964e-01 -1.82151407e-01
5.58520675e-01 2.71384746e-01 9.32189167e-01 -6.43104970e-01
8.73059094e-01 8.44984412e-01 1.11882389e-01 -8.66153121e-01
-7.73813188e-01 -1.08404510e-01 -9.36073422e-01 -1.65451124e-01
8.69447768e-01 -4.55428451e-01 -7.46848702e-01 6.06682718e-01
-1.07341826e+00 -1.05510867e+00 -5.97671807e-01 2.73540139e-01
-6.87272549e-01 8.54482427e-02 -3.92216384e-01 -6.57519877e-01
3.12008113e-01 -1.26679230e+00 1.28313756e+00 3.11277837e-01
-1.39098078e-01 -1.09953570e+00 2.64427096e-01 1.14354707e-01
2.35974446e-01 7.35031068e-01 1.01469719e+00 -3.67007911e-01
-5.90024352e-01 -1.08143434e-01 -6.82733953e-01 1.95071951e-01
-3.02783072e-01 7.27222979e-01 -1.20948708e+00 5.80786821e-03
-3.80196780e-01 -3.37974101e-01 6.72349751e-01 4.50231761e-01
1.38185573e+00 1.31031007e-01 -4.38901663e-01 7.68791854e-01
1.46762955e+00 1.72366768e-01 8.72897387e-01 6.92836702e-01
6.34519458e-01 6.46596134e-01 4.62866813e-01 6.10089526e-02
1.19302176e-01 5.67755997e-01 5.40028512e-01 -4.26207244e-01
-3.14449668e-01 -3.27031076e-01 -3.72403711e-01 -1.17142744e-01
2.07983740e-02 -4.15040702e-01 -1.19654667e+00 2.58289307e-01
-1.49319947e+00 -8.52445006e-01 -3.26711327e-01 2.51097631e+00
3.21753979e-01 6.51895881e-01 2.26953179e-01 1.68148190e-01
8.48544836e-01 3.07464689e-01 -6.53719306e-01 -5.03991306e-01
-3.27467173e-01 2.47849777e-01 1.02520096e+00 4.94810849e-01
-1.22647655e+00 1.04709661e+00 8.05191612e+00 8.32291543e-01
-1.23153412e+00 -2.94099271e-01 1.10799062e+00 4.36783701e-01
-1.24389172e-01 -4.10323218e-02 -4.27832037e-01 2.15652421e-01
9.77836490e-01 1.90851569e-01 2.47011513e-01 8.66765320e-01
3.38946939e-01 -5.90818644e-01 -1.19537628e+00 6.75706625e-01
2.10767835e-02 -1.53106439e+00 -1.29945263e-01 3.79459523e-02
7.93754041e-01 4.73429739e-01 -8.98700431e-02 3.06996442e-02
9.87704694e-01 -1.02312458e+00 7.64421821e-01 1.27796620e-01
8.40567052e-01 -3.99608642e-01 4.20532078e-01 1.94743067e-01
-9.80449617e-01 2.93366700e-01 -2.89429784e-01 9.62344371e-03
2.28290945e-01 1.56094790e-01 -9.72234666e-01 2.93640584e-01
6.15784168e-01 3.09073299e-01 -1.11395442e+00 1.23972166e+00
-3.45132053e-01 5.17235935e-01 -5.26950896e-01 2.67251641e-01
1.07973754e-01 -2.61579435e-02 2.95428127e-01 1.29179156e+00
-2.56333113e-01 -2.39638135e-01 9.85036884e-03 8.84614587e-01
2.10312515e-01 -3.32481265e-01 -7.36360788e-01 3.33137453e-01
3.46180022e-01 1.06876910e+00 -1.47409022e+00 -3.30559283e-01
2.03890949e-02 5.52760124e-01 2.23899275e-01 6.60230815e-01
-9.40715432e-01 -2.56960373e-02 4.09066319e-01 2.90735275e-01
1.97384879e-01 -4.92325991e-01 -5.29338300e-01 -9.16465223e-01
-9.48313549e-02 -7.32609749e-01 1.88011732e-02 -1.13841128e+00
-9.04304087e-01 8.53799045e-01 4.55277890e-01 -1.36749279e+00
-3.75960797e-01 -9.35474396e-01 -7.32164383e-01 4.95941460e-01
-1.25647199e+00 -9.91723895e-01 -5.62718272e-01 1.41085535e-01
5.43185711e-01 -9.32231545e-03 7.57161736e-01 -7.44043589e-02
-4.36003327e-01 3.52185249e-01 2.83813179e-01 3.08875978e-01
2.24162057e-01 -1.15914094e+00 1.40555370e+00 9.07439411e-01
3.36193800e-01 5.51002882e-02 1.11573243e+00 -2.64864802e-01
-6.52726889e-01 -1.07836139e+00 -1.23673081e-01 -7.02799439e-01
5.98502696e-01 -4.01833981e-01 -6.86443627e-01 6.73162937e-01
-4.26873006e-02 1.79938167e-01 8.30843747e-02 -1.75132528e-01
-5.94086766e-01 2.61240304e-02 -1.41268075e+00 8.82831752e-01
1.23261702e+00 -3.61609876e-01 -2.68077463e-01 3.47475588e-01
6.71452224e-01 -8.41621935e-01 -4.07968909e-01 6.06085718e-01
7.97371984e-01 -1.69471395e+00 1.09145987e+00 -7.22721159e-01
5.23459554e-01 -2.69141704e-01 -2.07510099e-01 -1.30615890e+00
4.53195013e-02 -3.37182611e-01 4.36949462e-01 8.21570456e-01
5.42550981e-01 -6.21948540e-01 1.12386131e+00 6.75215244e-01
-6.97539821e-02 -7.76735842e-01 -5.95977366e-01 -8.90567839e-01
6.61300048e-02 -6.49629235e-01 5.25964558e-01 7.42738545e-01
-8.63921165e-01 1.23846211e-01 2.05068558e-01 2.69540131e-01
6.65637136e-01 -6.96780439e-03 1.25333190e+00 -1.15348840e+00
-1.80526182e-01 -4.83226717e-01 -8.31755817e-01 -8.08492661e-01
-1.40272174e-02 -2.03075230e-01 -1.16910823e-01 -1.29851711e+00
-1.91551372e-01 -6.89176559e-01 4.40984547e-01 5.14145126e-04
2.32294708e-01 6.14988029e-01 1.37830168e-01 -2.24680360e-02
-5.04367709e-01 1.29606366e-01 1.31187832e+00 3.10174134e-02
-1.01141155e-01 1.56111762e-01 -2.12993547e-01 1.19796252e+00
9.31388795e-01 -2.16506988e-01 -4.44566965e-01 -6.18103921e-01
4.78175990e-02 -2.08723053e-01 7.76831448e-01 -1.46237791e+00
-1.05752960e-01 -1.39373243e-01 7.07397580e-01 -6.03972793e-01
4.15755212e-01 -7.80257165e-01 3.03952415e-02 2.06635147e-01
-3.72403830e-01 3.70421857e-01 6.04623318e-01 3.99189740e-01
1.50366426e-01 -1.75497413e-01 9.65429723e-01 -2.28862986e-01
-9.44988489e-01 2.32583098e-02 -2.13081464e-01 4.39774632e-01
1.15907741e+00 -6.96841657e-01 -5.94845593e-01 -3.24408621e-01
-7.63711751e-01 1.97561562e-01 8.92006159e-01 3.69394541e-01
1.08264513e-01 -8.09019744e-01 -4.07291055e-01 3.22150111e-01
2.19268069e-01 2.27645382e-01 1.96597606e-01 1.83173954e-01
-1.16799402e+00 -3.18557061e-02 -3.45777869e-01 -8.80896211e-01
-1.18636668e+00 2.93539107e-01 7.41835833e-01 -2.19758466e-01
-4.41313177e-01 6.16179705e-01 3.83335710e-01 -5.82333446e-01
1.52288005e-01 -7.61607051e-01 1.79779410e-01 -4.27780092e-01
8.63589644e-02 2.04099670e-01 3.25577617e-01 -6.46290600e-01
-1.44085035e-01 6.26991689e-01 2.04738647e-01 -3.27293873e-01
9.77676392e-01 6.90021366e-02 4.48464155e-01 4.96301293e-01
9.05234456e-01 -2.95029819e-01 -1.56904018e+00 1.79952905e-01
-2.38918573e-01 -6.96500242e-01 -2.39764735e-01 -8.06324184e-01
-9.45366442e-01 1.02662635e+00 8.04590642e-01 3.08952183e-01
7.71904826e-01 -1.29863381e-01 2.53044337e-01 4.41300124e-01
5.21622300e-01 -1.03548920e+00 -1.62353709e-01 2.19284937e-01
6.54194772e-01 -1.32323802e+00 5.86447567e-02 -7.76311040e-01
-3.44102263e-01 1.10990763e+00 8.00095141e-01 -4.10007060e-01
5.85334480e-01 7.34117448e-01 4.08722311e-01 -1.64661080e-01
-3.29403341e-01 -3.90072428e-02 -1.11256227e-01 9.02672350e-01
5.37538268e-02 7.12799877e-02 3.54310453e-01 -2.88389564e-01
-7.03619301e-01 -1.78951368e-01 6.00100577e-01 8.20295572e-01
-2.07464129e-01 -9.71661389e-01 -3.80573839e-01 4.60025281e-01
-1.67276114e-01 1.30565763e-01 -3.42234969e-01 1.59883034e+00
2.18739912e-01 6.65716112e-01 3.90083373e-01 -3.01269472e-01
6.02976441e-01 -4.69261169e-01 7.47082829e-01 -5.84603012e-01
-4.52535748e-01 -3.91084075e-01 4.95632917e-01 -5.83389938e-01
-3.83755326e-01 -6.44190133e-01 -9.53771591e-01 -1.84704795e-01
-1.71481818e-01 -1.87884971e-01 8.06423604e-01 9.94060934e-01
4.62890491e-02 6.58565044e-01 4.16719466e-01 -1.30013573e+00
-1.77064598e-01 -6.49910331e-01 -3.35267246e-01 4.31013763e-01
1.95401967e-01 -5.34880817e-01 -4.29336160e-01 1.90766603e-01] | [8.655253410339355, -1.4783626794815063] |
6c271a59-1310-4de8-ba63-d6e6c107d49d | statistical-model-for-describing-heart-rate | 2207.08165 | null | https://arxiv.org/abs/2207.08165v1 | https://arxiv.org/pdf/2207.08165v1.pdf | Statistical model for describing heart rate variability in normal rhythm and atrial fibrillation | Heart rate variability (HRV) indices describe properties of interbeat intervals in electrocardiogram (ECG). Usually HRV is measured exclusively in normal sinus rhythm (NSR) excluding any form of paroxysmal rhythm. Atrial fibrillation (AF) is the most widespread cardiac arrhythmia in human population. Usually such abnormal rhythm is not analyzed and assumed to be chaotic and unpredictable. Nonetheless, ranges of HRV indices differ between patients with AF, yet physiological characteristics which influence them are poorly understood. In this study, we propose a statistical model that describes relationship between HRV indices in NSR and AF. The model is based on Mahalanobis distance, the k-Nearest neighbour approach and multivariate normal distribution framework. Verification of the method was performed using 10 min intervals of NSR and AF that were extracted from long-term Holter ECGs. For validation we used Bhattacharyya distance and Kolmogorov-Smirnov 2-sample test in a k-fold procedure. The model is able to predict at least 7 HRV indices with high precision. | ['Yakov Bozhko', 'Konstantin Ushenin', 'Ilya Kotov', 'Nikita Markov'] | 2022-07-17 | null | null | null | null | ['heart-rate-variability'] | ['medical'] | [ 1.67383865e-01 -4.93601382e-01 8.81256908e-02 -3.17981988e-01
-1.77580453e-02 -8.08387697e-01 1.39586523e-01 4.60015923e-01
-4.00226474e-01 1.25228453e+00 1.44673921e-02 -4.15248185e-01
-6.18417382e-01 -7.02912271e-01 1.60870790e-01 -5.28233409e-01
-6.75343335e-01 6.51935339e-01 -3.81994516e-01 -3.60868908e-02
3.02545935e-01 8.73311043e-01 -1.27446496e+00 -1.40568525e-01
7.17390776e-01 7.88085818e-01 -3.26046109e-01 1.18959153e+00
3.38844448e-01 1.77505627e-01 -7.97353625e-01 3.62598032e-01
1.91995338e-01 -1.03987336e+00 -5.68576276e-01 -3.52938563e-01
-4.55829173e-01 1.88256428e-01 2.63809502e-01 4.65113074e-01
9.70837593e-01 1.25440378e-02 9.02170837e-01 -9.25540149e-01
-2.33692735e-01 2.62562305e-01 -1.64295733e-01 8.42018366e-01
2.79516727e-01 -8.04349780e-03 4.67899054e-01 -6.31496727e-01
3.15762430e-01 5.76295555e-01 1.02636969e+00 1.42136857e-01
-1.65399492e+00 -4.96117651e-01 -7.91928947e-01 1.80408344e-01
-1.69522834e+00 -1.36359811e-01 6.14850402e-01 -5.79145670e-01
6.99911177e-01 6.45780265e-01 1.32065463e+00 6.12195313e-01
6.34123027e-01 -4.33885157e-01 1.58718801e+00 -3.58434498e-01
1.26928911e-01 -1.76962707e-02 4.21252668e-01 7.95641840e-02
7.32616127e-01 2.99454421e-01 -9.43407491e-02 -4.86038595e-01
7.90652752e-01 -3.65475230e-02 -4.28692311e-01 2.75442544e-02
-1.40922201e+00 7.12026954e-01 -2.29376823e-01 7.43330657e-01
-5.21098435e-01 -3.39327067e-01 5.95856726e-01 7.67473936e-01
5.22312820e-02 3.90179843e-01 -5.81990421e-01 -4.40801352e-01
-1.11494195e+00 7.51898661e-02 9.17369485e-01 2.48778492e-01
4.25462991e-01 7.60957971e-02 -5.02282754e-02 6.50063634e-01
2.24248603e-01 7.37313747e-01 7.02930450e-01 -7.58629620e-01
-3.20090055e-01 4.11507189e-01 3.82189155e-02 -1.07866573e+00
-6.93652451e-01 -6.61050916e-01 -1.31343997e+00 1.63534150e-01
5.68850160e-01 -5.98275140e-02 -2.57819861e-01 1.27871013e+00
5.64130694e-02 2.37670273e-01 1.80620134e-01 7.15426505e-01
5.22085726e-01 2.60594279e-01 -1.79541156e-01 -1.16408479e+00
1.32122135e+00 1.28450811e-01 -8.50783050e-01 5.67467451e-01
2.32936069e-01 -7.46533096e-01 8.05087447e-01 6.58481896e-01
-7.79962718e-01 -7.02394783e-01 -9.32959616e-01 4.38720524e-01
-2.73072086e-02 -1.91156819e-01 2.50004116e-04 1.08346927e+00
-8.82309198e-01 1.06636012e+00 -6.49423540e-01 -4.48465586e-01
4.71557900e-02 1.66006550e-01 -3.81840706e-01 5.47368586e-01
-1.24514496e+00 9.33605433e-01 4.33969527e-01 3.61310750e-01
-2.26358354e-01 -6.28793240e-01 -5.96322954e-01 -3.09281707e-01
-4.21231419e-01 -8.35348666e-01 5.00843406e-01 -5.05310953e-01
-1.24489248e+00 9.77027714e-01 -3.73256773e-01 -4.16298747e-01
4.81802970e-01 8.93338025e-02 -7.26595223e-01 6.75232662e-03
-2.53820360e-01 -4.56672788e-01 7.01216102e-01 -1.00281596e+00
2.00663984e-01 -6.91234946e-01 -8.76175165e-01 -5.44920824e-02
3.87705207e-01 4.26043235e-02 6.71727359e-01 -7.05609500e-01
4.42035288e-01 -8.80472958e-01 -1.32337362e-01 -6.52908266e-01
-2.23338664e-01 2.95983762e-01 4.83985364e-01 -7.61283934e-01
1.80150867e+00 -1.95956206e+00 -9.77784470e-02 7.64360368e-01
1.85733885e-01 2.40295440e-01 4.06710535e-01 4.62071180e-01
-3.28244448e-01 3.28613937e-01 -4.47740048e-01 3.10416251e-01
-3.81763726e-01 4.24937129e-01 -1.33452222e-01 7.96759307e-01
-2.67220825e-01 6.02255225e-01 -6.45539522e-01 -4.28136617e-01
2.55018890e-01 6.90375507e-01 9.15176347e-02 5.90672195e-02
5.78615248e-01 1.14957035e+00 -1.87495481e-02 4.26915586e-01
5.37612677e-01 8.27822536e-02 1.60333320e-01 -2.08472401e-01
-2.85370708e-01 -1.34751841e-01 -1.03388155e+00 1.19121218e+00
-5.54900020e-02 5.83027720e-01 -6.07227206e-01 -1.04412138e+00
1.57243252e+00 6.95745707e-01 4.10145223e-01 -4.51939166e-01
2.84418672e-01 3.75959188e-01 4.99822736e-01 -3.88147295e-01
-3.21840137e-01 -4.71902013e-01 3.16974729e-01 4.25463676e-01
-3.45879048e-01 -1.20058283e-02 2.18526080e-01 -6.16408944e-01
7.55585134e-01 -2.26988882e-01 1.34280729e+00 -7.13290572e-01
9.10131872e-01 -3.19176048e-01 5.76879859e-01 7.28865921e-01
-3.82649034e-01 7.42968976e-01 6.10040545e-01 -9.34850693e-01
-8.96328986e-01 -1.20289576e+00 -7.91255116e-01 -1.20240405e-01
-2.31224447e-01 -2.04586178e-01 -3.05491805e-01 -9.15775448e-02
5.17501449e-03 5.49699903e-01 -5.56213617e-01 -7.64390752e-02
-5.74813247e-01 -1.04675770e+00 8.88426423e-01 -1.73105691e-02
1.81200534e-01 -1.23105192e+00 -1.05373061e+00 3.87901664e-01
-8.58048424e-02 -5.34900367e-01 1.15825072e-01 7.57174715e-02
-1.51711464e+00 -1.34523678e+00 -6.62286580e-01 -3.11630756e-01
2.31367961e-01 -4.90194291e-01 1.28944302e+00 1.20544866e-01
-9.09418106e-01 1.32593274e-01 -3.06993484e-01 -4.64715540e-01
-6.09715581e-01 -4.02032137e-01 3.14313889e-01 -1.25293881e-01
5.27273059e-01 -1.16080642e+00 -9.71667409e-01 5.19886732e-01
-3.43019545e-01 -7.55283773e-01 4.02055234e-01 5.54461479e-01
9.80862319e-01 5.99249341e-02 9.49457407e-01 -6.78604662e-01
8.56673956e-01 -3.43897998e-01 -3.94172698e-01 -2.43739299e-02
-1.27975976e+00 -3.48514289e-01 5.57721496e-01 -1.68781787e-01
-5.15289269e-02 -4.14504170e-01 1.28839135e-01 -2.01963678e-01
-5.74823976e-01 3.94075304e-01 1.38011783e-01 1.94496661e-01
1.01212609e+00 6.04540169e-01 2.89551169e-01 -1.79515243e-01
-1.45709842e-01 6.64384961e-01 5.18394113e-01 -2.25488514e-01
6.23520613e-01 3.14272881e-01 4.14493531e-01 -1.26049340e+00
-2.38775194e-01 -6.21544361e-01 -7.88646162e-01 -2.49603108e-01
9.74357486e-01 -5.63669682e-01 -1.01557267e+00 1.42887592e-01
-9.30259585e-01 9.54640284e-03 -5.09546459e-01 9.23746228e-01
-7.71126032e-01 4.48632866e-01 -2.67950177e-01 -1.13376093e+00
-8.70593667e-01 -6.29255950e-01 2.95844346e-01 -2.32515968e-02
-8.70708406e-01 -1.04200721e+00 8.12595665e-01 -1.33596420e-01
5.45850873e-01 9.02458310e-01 7.68765390e-01 -7.97489464e-01
3.29179615e-01 -3.33758116e-01 2.59130210e-01 5.28593779e-01
6.21599138e-01 1.22052319e-01 -8.21517169e-01 -2.15506777e-01
5.23290098e-01 4.39818740e-01 3.43860269e-01 7.61801422e-01
8.68045628e-01 -8.36432129e-02 -2.87289042e-02 6.04591429e-01
1.46345055e+00 8.34646702e-01 8.86841238e-01 2.24013522e-01
3.09802830e-01 2.74959534e-01 4.05161530e-01 6.72409296e-01
-6.91934973e-02 2.62637168e-01 -1.45331174e-01 1.35257989e-01
4.76466060e-01 3.32453907e-01 3.64079215e-02 8.47761393e-01
-8.65160465e-01 2.65037179e-01 -1.07258654e+00 2.16988817e-01
-1.37952423e+00 -1.32117546e+00 -5.93995333e-01 2.73036766e+00
7.69775093e-01 1.32763088e-01 4.71864969e-01 8.17376316e-01
6.84328794e-01 -1.93941832e-01 -1.73676118e-01 -8.91457319e-01
-3.46639633e-01 5.53087711e-01 1.52265534e-01 5.86591601e-01
-7.83728957e-01 1.00312298e-02 6.65195227e+00 -1.30554006e-01
-1.17184889e+00 -1.41799137e-01 5.77761471e-01 3.08217287e-01
-1.72332546e-03 2.94585004e-02 -2.84392416e-01 4.79559898e-01
1.43968809e+00 -5.37424028e-01 4.01468694e-01 4.36724782e-01
7.30271637e-01 -8.37193877e-02 -8.76372278e-01 1.39820635e+00
5.64148324e-03 -9.59532738e-01 -3.37018251e-01 3.22542572e-03
2.91137636e-01 -2.97240287e-01 -2.89569169e-01 -5.33867776e-02
-8.97779524e-01 -1.34934306e+00 3.00563574e-02 1.05516672e+00
1.06405818e+00 -9.08592820e-01 1.02690637e+00 2.01324791e-01
-1.23735523e+00 7.22203031e-02 -2.09696159e-01 -2.43386924e-01
2.21537799e-01 1.13797450e+00 -6.33491635e-01 5.37077963e-01
5.16820431e-01 7.93775856e-01 -3.79882187e-01 1.16043901e+00
1.24746844e-01 8.36017072e-01 -2.58643627e-01 2.52421111e-01
-5.42470217e-01 -7.58200347e-01 8.95618618e-01 9.55560982e-01
5.97260416e-01 1.72074005e-01 -2.22186401e-01 8.62191677e-01
6.77888632e-01 3.90135109e-01 -9.87366378e-01 -4.58417386e-02
5.79983652e-01 9.50142920e-01 -1.01809418e+00 -1.41414911e-01
-9.34693366e-02 7.23973691e-01 -5.89157939e-01 1.47061437e-01
-5.46249330e-01 -4.87559855e-01 4.19614881e-01 3.18392605e-01
-2.42417812e-01 -1.30104244e-01 -7.12753654e-01 -8.12879801e-01
-9.40479115e-02 -8.59539568e-01 4.58354741e-01 -3.00600618e-01
-1.04967523e+00 8.74582767e-01 8.30935985e-02 -1.47366917e+00
-3.92487437e-01 -2.52074033e-01 -7.80934751e-01 1.41061676e+00
-1.05999517e+00 -4.10797715e-01 -3.41462195e-01 6.74172938e-01
-3.08814440e-02 -1.66122526e-01 1.54535353e+00 2.36896247e-01
-1.59528479e-01 1.45690337e-01 -1.54223248e-01 -9.12085325e-02
7.55389810e-01 -1.53035808e+00 -1.02467962e-01 4.46339875e-01
-1.12144396e-01 8.33432913e-01 1.02162457e+00 -6.51400268e-01
-6.39671922e-01 -8.29039097e-01 1.23141372e+00 -5.41513681e-01
2.67324239e-01 2.10053027e-01 -7.54462659e-01 3.01666826e-01
-1.73415497e-01 4.48501885e-01 1.38248372e+00 -1.06256939e-01
-2.48462185e-02 -2.80621260e-01 -1.17357004e+00 1.32610574e-01
3.66714925e-01 -3.81318092e-01 -8.06742013e-01 -3.38536650e-02
-3.77555639e-02 1.53175980e-01 -1.58054626e+00 5.36046565e-01
9.76014853e-01 -1.44845295e+00 9.64277565e-01 -2.85441726e-01
-1.78565845e-01 -5.96683860e-01 8.93707126e-02 -1.04039538e+00
-1.20437935e-01 -1.12839198e+00 -5.15386984e-02 6.99145615e-01
2.04639211e-01 -9.85830367e-01 2.73487061e-01 -9.30785686e-02
1.93054557e-01 -5.74608743e-01 -8.90695870e-01 -9.37462926e-01
-2.03512251e-01 -1.76423177e-01 3.71131837e-01 1.04557824e+00
7.14150518e-02 3.48093778e-01 -2.15421140e-01 -4.90909480e-02
8.80745292e-01 1.71878070e-01 5.23005962e-01 -1.89409506e+00
-2.01334685e-01 -2.95683175e-01 -8.82120252e-01 3.88025820e-01
-2.46345505e-01 -8.21580350e-01 -4.13369805e-01 -1.28134882e+00
-1.37856811e-01 -3.78191561e-01 -6.87439859e-01 -8.67578760e-02
1.58590393e-03 4.49009061e-01 -2.35997438e-01 3.78133118e-01
3.40319335e-01 3.14183868e-02 9.42331970e-01 5.72453976e-01
-8.97170067e-01 5.71326852e-01 -2.57385284e-01 5.95178604e-01
1.16395819e+00 -6.24264956e-01 -6.67910933e-01 7.68700123e-01
1.80453047e-01 5.55131972e-01 2.53642350e-01 -1.15777314e+00
-3.70277673e-01 1.11385271e-01 6.86559796e-01 -5.97033203e-01
-1.69151813e-01 -8.95338297e-01 9.41321135e-01 8.54011893e-01
-2.76537929e-02 6.94429219e-01 1.87145229e-02 5.19017160e-01
-2.72115558e-01 4.07497287e-02 8.82310629e-01 -1.16648011e-01
9.17119235e-02 2.61320680e-01 -4.95065361e-01 1.49703577e-01
9.98380363e-01 -8.08993101e-01 1.14209518e-01 -1.64303914e-01
-1.43047071e+00 -3.51936042e-01 2.90490597e-01 -1.32760912e-01
7.71470964e-01 -1.17874718e+00 -9.05480325e-01 4.44453984e-01
1.30605221e-01 -3.62043619e-01 2.95842052e-01 1.70049989e+00
-1.01745224e+00 2.37580001e-01 -5.88700116e-01 -7.57203221e-01
-1.50657308e+00 4.35060561e-01 4.78799224e-01 -7.79226422e-02
-8.87846351e-01 1.38212100e-01 -6.31074429e-01 -8.82684588e-02
-1.29501000e-01 -3.42849523e-01 -7.69791067e-01 4.25785810e-01
6.89042747e-01 6.67817473e-01 -1.27453417e-01 -6.73161268e-01
-5.43290854e-01 9.15843546e-01 5.05180120e-01 -1.08276911e-01
8.73475671e-01 -2.86317259e-01 -6.93643987e-01 1.16725254e+00
1.13685536e+00 7.27875009e-02 -2.22130656e-01 5.10763466e-01
9.78137627e-02 -2.77858049e-01 -3.27408671e-01 -6.64150417e-01
-4.28140670e-01 6.12732291e-01 1.16532648e+00 4.98235434e-01
1.26796794e+00 -5.94521880e-01 3.18018377e-01 2.72938967e-01
2.29782194e-01 -6.62423432e-01 -6.41074300e-01 1.19834699e-01
9.17929471e-01 -9.11642790e-01 5.39029725e-02 -6.98541626e-02
-4.40891713e-01 1.52708447e+00 -3.46472114e-01 -2.71243572e-01
1.28722763e+00 1.09601274e-01 3.85146081e-01 3.96262901e-03
-3.95390570e-01 1.23060308e-02 1.19636096e-01 7.45261610e-01
8.28734815e-01 1.66277438e-01 -1.15061629e+00 5.63199759e-01
-4.51933056e-01 3.15618634e-01 4.98211801e-01 7.24553823e-01
-3.79938781e-01 -1.04895401e+00 -4.68381524e-01 6.09828115e-01
-6.51491165e-01 -2.10242299e-03 -5.79509092e-03 8.74462605e-01
5.97128645e-02 9.65146363e-01 3.74088343e-03 -1.54451713e-01
3.46493393e-01 4.36836213e-01 4.89874005e-01 -2.92962670e-01
-6.51960433e-01 1.45715699e-01 -1.04439445e-01 -5.05360961e-01
-5.61369836e-01 -7.61925638e-01 -1.22372627e+00 -2.22235650e-01
1.93398760e-03 5.48909545e-01 5.77053368e-01 7.32112885e-01
1.02117330e-01 4.24057364e-01 7.65452981e-01 -2.25406513e-01
-3.44488263e-01 -1.09716034e+00 -1.29000413e+00 2.46769354e-01
5.53989053e-01 -3.25241387e-01 -7.52084374e-01 2.91027963e-01] | [14.143200874328613, 3.179441213607788] |
375687f6-db92-422f-a2ee-be5bf85a01ed | what-comprises-a-good-talking-head-video | 2005.03201 | null | https://arxiv.org/abs/2005.03201v1 | https://arxiv.org/pdf/2005.03201v1.pdf | What comprises a good talking-head video generation?: A Survey and Benchmark | Over the years, performance evaluation has become essential in computer vision, enabling tangible progress in many sub-fields. While talking-head video generation has become an emerging research topic, existing evaluations on this topic present many limitations. For example, most approaches use human subjects (e.g., via Amazon MTurk) to evaluate their research claims directly. This subjective evaluation is cumbersome, unreproducible, and may impend the evolution of new research. In this work, we present a carefully-designed benchmark for evaluating talking-head video generation with standardized dataset pre-processing strategies. As for evaluation, we either propose new metrics or select the most appropriate ones to evaluate results in what we consider as desired properties for a good talking-head video, namely, identity preserving, lip synchronization, high video quality, and natural-spontaneous motion. By conducting a thoughtful analysis across several state-of-the-art talking-head generation approaches, we aim to uncover the merits and drawbacks of current methods and point out promising directions for future work. All the evaluation code is available at: https://github.com/lelechen63/talking-head-generation-survey. | ['Haitian Zheng', 'Ziyi Kou', 'Lele Chen', 'Guofeng Cui', 'Chenliang Xu'] | 2020-05-07 | null | null | null | null | ['talking-head-generation'] | ['computer-vision'] | [ 1.40441850e-01 -6.29429445e-02 -2.84753770e-01 -4.31132525e-01
-8.74049723e-01 -4.33896989e-01 8.21986318e-01 -3.10711920e-01
-1.73342526e-01 6.32553816e-01 4.94546741e-01 -5.25588766e-02
1.15699030e-01 -2.50806749e-01 -3.35065067e-01 -7.40218878e-01
6.17782697e-02 -7.23448768e-02 1.27987310e-01 -4.71621454e-02
3.78060311e-01 2.28780374e-01 -1.94592810e+00 9.27152857e-02
7.01496124e-01 1.01449358e+00 1.04775012e-01 6.22033894e-01
2.46151537e-01 8.53429735e-01 -6.93080366e-01 -8.03476870e-01
-3.81991901e-02 -6.97411895e-01 -8.28613877e-01 2.65813947e-01
4.75158840e-01 -3.01333070e-01 -2.27426678e-01 1.15154254e+00
1.00041246e+00 9.63908955e-02 2.90456980e-01 -1.66433418e+00
-6.45531416e-01 5.98270655e-01 -3.48160297e-01 1.95378542e-01
9.00045156e-01 4.75273013e-01 1.00563002e+00 -9.04151618e-01
7.69844532e-01 1.18834639e+00 4.84122127e-01 7.46402264e-01
-8.82523835e-01 -7.50376165e-01 -1.44782394e-01 6.63594604e-01
-1.53898585e+00 -1.28268874e+00 8.85169208e-01 -3.97350997e-01
3.73174161e-01 4.48990911e-01 4.46148515e-01 1.52987230e+00
-9.43337902e-02 9.47994292e-01 1.06098652e+00 -3.87200028e-01
1.82895914e-01 2.52345651e-01 2.00622659e-02 3.70614350e-01
1.18017055e-01 2.78777685e-02 -8.62008035e-01 9.87873226e-03
2.95335770e-01 -5.28357983e-01 -8.19472313e-01 -2.63671190e-01
-1.41105831e+00 6.75422668e-01 -7.05399588e-02 3.18925381e-01
-2.03953668e-01 -1.92459419e-01 5.19159794e-01 1.28496677e-01
4.50587779e-01 1.09929726e-01 1.97325572e-01 -6.56513512e-01
-1.23474896e+00 3.86244476e-01 8.30032706e-01 9.74179864e-01
3.58178437e-01 -1.09637573e-01 -2.89873093e-01 8.07437837e-01
6.30379990e-02 4.37365115e-01 5.31014562e-01 -1.32804465e+00
3.84061426e-01 6.42386032e-03 1.04762517e-01 -1.12135768e+00
-1.12371132e-01 3.05384700e-03 -8.17537546e-01 6.58486784e-02
3.18285137e-01 -2.09979951e-01 -2.34666958e-01 1.75561059e+00
3.20445508e-01 2.81665057e-01 -6.50197715e-02 1.10154879e+00
1.11637282e+00 5.57701051e-01 -1.49439052e-01 -5.14958680e-01
1.32865679e+00 -9.82221425e-01 -9.68364000e-01 2.17687756e-01
3.26545924e-01 -1.08912539e+00 1.18045485e+00 4.60147083e-01
-1.32493818e+00 -5.46850801e-01 -7.81100929e-01 1.98498927e-02
-2.04670280e-02 -3.97627540e-02 4.09043431e-01 9.46716905e-01
-1.36685765e+00 3.42143148e-01 -4.56170827e-01 -6.39441431e-01
1.81954131e-01 8.71530026e-02 -3.53842765e-01 -2.04960242e-01
-1.09966171e+00 6.95170522e-01 1.88515764e-02 -1.81808382e-01
-8.41972888e-01 -5.02566516e-01 -7.22056389e-01 -2.86271751e-01
5.05634904e-01 -7.73908257e-01 1.59151602e+00 -8.58037114e-01
-1.87486351e+00 1.03190351e+00 -3.87219518e-01 -3.91653776e-01
6.99265361e-01 -2.16898590e-01 -5.14015734e-01 1.99693799e-01
4.26730625e-02 7.59872854e-01 9.55100596e-01 -1.40109813e+00
-6.53062105e-01 -1.44584969e-01 3.99822928e-02 3.08299720e-01
-4.31328356e-01 4.34052229e-01 -6.79942071e-01 -7.45428741e-01
-2.94020057e-01 -1.06956494e+00 2.14653537e-01 -1.97133645e-01
-5.91382802e-01 -1.71879590e-01 8.20979774e-01 -5.38496614e-01
1.44508660e+00 -2.31492400e+00 2.12002937e-02 -3.40484589e-01
2.77806878e-01 2.62223303e-01 1.44889399e-01 3.12861621e-01
4.90263328e-02 2.04770729e-01 -1.43441975e-01 -6.89065456e-01
6.69341981e-02 -3.13721150e-01 -1.65706232e-01 6.72842741e-01
-5.27608544e-02 7.22473264e-01 -8.72445166e-01 -7.18158424e-01
3.90370667e-01 6.86498523e-01 -5.26468217e-01 2.99377978e-01
1.78733081e-01 5.30154645e-01 -2.34845892e-01 8.14962566e-01
5.57156801e-01 -6.82930052e-02 -8.41067359e-02 -2.42023095e-01
-2.51457274e-01 1.66711912e-01 -1.11218929e+00 1.55064487e+00
-2.32173160e-01 1.04563856e+00 1.57301128e-01 -7.93019414e-01
6.72254086e-01 6.19909465e-01 4.44251001e-01 -5.96338868e-01
2.78123379e-01 1.41569883e-01 -1.04404189e-01 -8.79225433e-01
7.03813851e-01 -1.45007372e-02 7.31008276e-02 2.73368865e-01
7.01263128e-03 -8.96728709e-02 2.94715464e-01 2.12784037e-02
7.88234770e-01 -1.15728423e-01 3.54385495e-01 -1.10522933e-01
6.93831384e-01 -4.15258825e-01 3.86427462e-01 4.91195232e-01
-6.67750835e-01 1.06734276e+00 3.16458493e-01 6.33608326e-02
-7.99791455e-01 -9.71639097e-01 -6.29653782e-02 1.15748298e+00
1.90589622e-01 -4.90473390e-01 -1.12660766e+00 -2.37995893e-01
-5.44206977e-01 7.03471601e-01 -4.02824551e-01 -7.19845369e-02
-3.94523233e-01 -5.21776080e-01 9.29364264e-01 2.64679998e-01
5.50829113e-01 -1.22805333e+00 -6.52972341e-01 -1.04373462e-01
-8.11104119e-01 -1.41507554e+00 -6.80082917e-01 -6.51874006e-01
-4.48185533e-01 -9.03143764e-01 -1.27097833e+00 -5.87336779e-01
3.07549596e-01 6.12845838e-01 1.18236351e+00 -2.67196838e-02
-3.26465592e-02 5.43342352e-01 -6.03484809e-01 -2.54928738e-01
-3.80748540e-01 5.09634316e-02 2.05321297e-01 3.60107034e-01
1.96413070e-01 -5.19503117e-01 -7.89882958e-01 5.68538427e-01
-7.38357127e-01 6.44603232e-03 3.14636230e-01 6.47592127e-01
1.92757502e-01 -1.32611305e-01 5.10457039e-01 -3.41558218e-01
1.03543103e+00 -4.40532744e-01 -3.03424627e-01 1.43645406e-01
-4.25320685e-01 -3.47203374e-01 5.28073549e-01 -5.15517950e-01
-1.16419959e+00 -2.87298292e-01 -1.57900676e-01 -6.07368708e-01
-4.19282436e-01 2.38163501e-01 -4.62982982e-01 1.02823265e-01
5.28032660e-01 4.35320348e-01 1.92843527e-02 -2.30313882e-01
4.51535076e-01 8.96210313e-01 7.45382369e-01 -5.67490637e-01
5.81461906e-01 5.10254979e-01 -4.00174707e-01 -1.17116785e+00
-4.85397160e-01 -5.16365528e-01 -2.41997078e-01 -6.66722775e-01
7.53467977e-01 -7.27622092e-01 -9.65629816e-01 6.54441714e-01
-1.19787943e+00 -2.56668538e-01 -6.17079251e-02 4.84973729e-01
-6.06026530e-01 6.49545372e-01 -3.35274428e-01 -8.58988762e-01
-4.07951951e-01 -1.42849898e+00 1.06338477e+00 2.57960290e-01
-4.73793656e-01 -7.95157135e-01 6.48809746e-02 7.45286286e-01
5.35479903e-01 4.39487249e-02 2.18713358e-01 -3.32915872e-01
-4.73324329e-01 -1.40676379e-01 -1.80934846e-01 3.52650821e-01
1.07648790e-01 2.77849764e-01 -1.27816701e+00 -3.27946961e-01
1.63979113e-01 -3.20149779e-01 5.15134275e-01 5.53070843e-01
1.05012763e+00 -3.78981411e-01 -1.39719084e-01 5.63885331e-01
8.65949929e-01 1.67436600e-01 7.67968178e-01 1.61316007e-01
4.12521988e-01 9.56946075e-01 5.12049139e-01 7.09312975e-01
7.00246215e-01 9.73552585e-01 1.88671947e-01 1.33100003e-01
-2.75323391e-01 -5.70197478e-02 5.98204136e-01 9.36750948e-01
-4.17858750e-01 -6.44382238e-01 -8.34088802e-01 5.50963581e-01
-1.67324996e+00 -1.14458191e+00 1.05223559e-01 2.31313658e+00
7.30135560e-01 9.16879550e-02 4.76666391e-01 3.73460919e-01
9.98216212e-01 3.24729323e-01 -2.81032801e-01 1.86314117e-02
-2.53129303e-01 -8.97748247e-02 3.77087668e-02 4.20947790e-01
-8.83374929e-01 8.57322812e-01 5.71701765e+00 8.66969287e-01
-1.38840973e+00 8.11244547e-02 6.72500253e-01 -1.41847730e-01
-4.70438488e-02 -1.26583099e-01 -7.12632954e-01 6.58623517e-01
8.73742521e-01 -6.29497468e-01 4.21356946e-01 6.99279785e-01
6.28198802e-01 -3.62290293e-02 -1.13976991e+00 1.46131659e+00
5.13026178e-01 -1.22222376e+00 -1.84115879e-02 1.18871249e-01
3.54213595e-01 -1.04760297e-01 2.50224710e-01 7.89648679e-04
-1.73559189e-01 -8.77528131e-01 9.87613618e-01 4.06424612e-01
7.67826498e-01 -4.85485703e-01 7.27445781e-01 6.93019554e-02
-1.11432612e+00 2.36718252e-01 -6.14791475e-02 2.46165656e-02
4.53633010e-01 5.21567047e-01 -5.26045740e-01 6.31186664e-01
8.36829543e-01 6.61381125e-01 -5.21354377e-01 1.19698799e+00
-8.55450258e-02 7.66526043e-01 -1.35506839e-01 -1.67714655e-02
-1.40792534e-01 2.45551788e-03 8.48110020e-01 1.30225790e+00
4.47940707e-01 3.22537050e-02 -5.66121228e-02 6.40429258e-01
-2.67230451e-01 2.95073658e-01 -7.32237756e-01 7.67221004e-02
6.07048750e-01 1.19497633e+00 -6.27315879e-01 -2.97467172e-01
-3.78901154e-01 8.28454435e-01 -2.00437605e-01 2.19170392e-01
-1.01266921e+00 -1.98361129e-01 7.09854424e-01 1.91930234e-01
-7.33654872e-02 -1.24198794e-01 -1.22382767e-01 -1.23447609e+00
2.15138689e-01 -1.12385440e+00 1.46479487e-01 -8.65410030e-01
-1.00039709e+00 7.23627329e-01 1.52679697e-01 -1.45204782e+00
-3.86431575e-01 -9.77823436e-02 -6.59663439e-01 3.95784467e-01
-1.31555021e+00 -8.26281965e-01 -5.99056125e-01 7.41157651e-01
8.28688085e-01 -1.14302054e-01 4.69853997e-01 5.16770303e-01
-7.27454901e-01 9.26440239e-01 -2.31605172e-01 -6.73983023e-02
1.02132511e+00 -7.30740070e-01 3.76401186e-01 8.19454193e-01
9.57057346e-03 4.94078338e-01 1.04098368e+00 -1.29592463e-01
-1.36736465e+00 -7.36698866e-01 9.26610589e-01 -3.88493776e-01
5.54019511e-01 -3.10245812e-01 -7.47744977e-01 3.36410493e-01
6.16981387e-01 -1.64466098e-01 6.81232452e-01 -1.79709762e-01
-7.32261166e-02 -7.75182769e-02 -9.44904745e-01 8.87808144e-01
1.08518386e+00 -3.18617880e-01 -4.36381906e-01 -8.39058906e-02
5.76728642e-01 -1.40879989e-01 -7.34821796e-01 2.98742652e-01
6.87354565e-01 -1.50370145e+00 8.06446135e-01 5.30871563e-02
4.11351323e-01 -1.61234051e-01 -1.54373586e-01 -1.11188209e+00
1.04876250e-01 -1.17306519e+00 1.02908886e-03 1.70175695e+00
1.80972710e-01 -5.27411282e-01 6.97786927e-01 5.29394925e-01
-1.16035350e-01 -5.27648211e-01 -7.80514479e-01 -7.78211772e-01
-2.32385576e-01 -5.38417637e-01 4.78950858e-01 1.04732883e+00
1.82911009e-01 4.63841021e-01 -7.37141490e-01 -2.36659721e-01
7.24678814e-01 -1.73657075e-01 1.17549729e+00 -1.00730574e+00
9.04310867e-02 -7.26296902e-01 -5.03975213e-01 -1.11051965e+00
3.58596623e-01 -4.84067500e-01 3.52661163e-02 -1.20106590e+00
2.86525726e-01 8.58388618e-02 -2.76186373e-02 1.12206072e-01
-1.42628804e-01 2.53643900e-01 5.22534907e-01 2.81737626e-01
-8.22292626e-01 7.16993392e-01 1.23911691e+00 6.10140041e-02
-6.12965636e-02 2.01619819e-01 -7.04818666e-01 6.55366242e-01
8.91398907e-01 -1.28186151e-01 -4.92273450e-01 -1.59952745e-01
-1.38620347e-01 2.82478809e-01 4.15766478e-01 -1.27046239e+00
3.70951831e-01 -1.04789048e-01 -3.46309960e-01 -3.83853734e-01
5.59020221e-01 -4.57250625e-01 7.16917366e-02 3.89486589e-02
-2.73779571e-01 8.00966024e-02 -1.61329150e-01 2.21269175e-01
-5.27439952e-01 -2.53987722e-02 8.06641102e-01 9.94181260e-02
-6.89271033e-01 1.20805755e-01 -2.98124522e-01 2.60415763e-01
8.75461757e-01 -4.72892076e-01 -2.50066251e-01 -9.92541432e-01
-3.43411535e-01 -1.83888443e-03 6.90031111e-01 6.56825721e-01
7.00368106e-01 -1.06072497e+00 -8.80802572e-01 -1.93078279e-01
1.69720322e-01 -3.77154380e-01 2.76933938e-01 9.78784144e-01
-2.94712126e-01 3.82228464e-01 -2.02780992e-01 -6.61302626e-01
-1.56210649e+00 4.39129680e-01 7.24419132e-02 6.43696338e-02
-5.68645895e-01 6.13148272e-01 2.64066219e-01 -6.82176352e-02
4.57866490e-01 -9.42200497e-02 -2.15323046e-01 1.29051968e-01
6.38974428e-01 6.71572328e-01 5.93998730e-02 -1.03191936e+00
-4.10113186e-01 4.06515628e-01 2.02662140e-01 -2.86383212e-01
9.18542802e-01 -5.21259725e-01 1.89915985e-01 3.76229584e-01
1.10873652e+00 1.46539393e-03 -9.13758636e-01 9.73517168e-03
-3.87837440e-02 -4.54549372e-01 -9.05893818e-02 -3.66168678e-01
-1.08627713e+00 8.23943317e-01 6.14507854e-01 4.24517900e-01
1.26790035e+00 8.24323744e-02 7.93387294e-01 8.14066455e-02
3.33720267e-01 -8.47982645e-01 1.09570749e-01 3.28808069e-01
9.36227560e-01 -1.41766679e+00 -1.36571690e-01 -4.76554066e-01
-8.54408860e-01 7.96085477e-01 3.52719426e-01 4.67327803e-01
5.90137243e-01 1.59616098e-01 2.13799462e-01 -6.98073953e-03
-6.77355528e-01 -2.83647627e-01 1.03032798e-01 6.83530569e-01
8.16679895e-01 6.74032271e-02 -3.85785580e-01 4.08524305e-01
-6.41792595e-01 2.46466309e-01 6.36531115e-01 7.18094647e-01
-1.57523453e-01 -9.02255416e-01 -6.53189123e-01 7.65274242e-02
-7.12164581e-01 -1.54167349e-02 -4.59157467e-01 7.49490559e-01
-1.00334466e-01 1.41132939e+00 -2.68564731e-01 -5.25340855e-01
3.69242877e-01 -3.04164644e-02 3.56597275e-01 -2.17182979e-01
-2.94075072e-01 4.26559746e-02 1.99726224e-01 -5.76792955e-01
-8.40737343e-01 -7.39693105e-01 -5.98764658e-01 -6.84692740e-01
-2.66468555e-01 6.00083545e-02 5.17895818e-01 7.52850294e-01
4.36560959e-01 1.34532675e-01 6.52848601e-01 -1.23387671e+00
-2.04353347e-01 -9.95348752e-01 -2.42913857e-01 6.33173466e-01
2.42041409e-01 -7.63965249e-01 -4.49532181e-01 5.69850683e-01] | [13.287247657775879, -0.30536720156669617] |
6234a4f0-0e63-4932-bbab-706dd4d8137e | augmenting-ego-vehicle-for-traffic-near-miss | 2301.02726 | null | https://arxiv.org/abs/2301.02726v1 | https://arxiv.org/pdf/2301.02726v1.pdf | Augmenting Ego-Vehicle for Traffic Near-Miss and Accident Classification Dataset using Manipulating Conditional Style Translation | To develop the advanced self-driving systems, many researchers are focusing to alert all possible traffic risk cases from closed-circuit television (CCTV) and dashboard-mounted cameras. Most of these methods focused on identifying frame-by-frame in which an anomaly has occurred, but they are unrealized, which road traffic participant can cause ego-vehicle leading into collision because of available annotation dataset only to detect anomaly on traffic video. Near-miss is one type of accident and can be defined as a narrowly avoided accident. However, there is no difference between accident and near-miss at the time before the accident happened, so our contribution is to redefine the accident definition and re-annotate the accident inconsistency on DADA-2000 dataset together with near-miss. By extending the start and end time of accident duration, our annotation can precisely cover all ego-motions during an incident and consistently classify all possible traffic risk accidents including near-miss to give more critical information for real-world driving assistance systems. The proposed method integrates two different components: conditional style translation (CST) and separable 3-dimensional convolutional neural network (S3D). CST architecture is derived by unsupervised image-to-image translation networks (UNIT) used for augmenting the re-annotation DADA-2000 dataset to increase the number of traffic risk accident videos and to generalize the performance of video classification model on different types of conditions while S3D is useful for video classification to prove dataset re-annotation consistency. In evaluation, the proposed method achieved a significant improvement result by 10.25% positive margin from the baseline model for accuracy on cross-validation analysis. | ['Koji Zettsu', 'Minh-Son Dao', 'Hilmil Pradana'] | 2023-01-06 | null | null | null | null | ['video-classification', 'unsupervised-image-to-image-translation'] | ['computer-vision', 'computer-vision'] | [ 1.82726696e-01 -6.09818147e-03 -2.73907602e-01 -4.14374471e-01
-7.46072590e-01 -2.90887475e-01 4.34792101e-01 -2.20914945e-01
-4.53456610e-01 5.52898526e-01 2.62306929e-01 -5.13471425e-01
5.05892672e-02 -7.05050588e-01 -7.67528832e-01 -5.45985937e-01
1.29613042e-01 2.21238241e-01 5.37245631e-01 -2.33726799e-01
-7.52018392e-02 5.31324863e-01 -1.71850216e+00 6.19066060e-01
4.32965338e-01 9.77214813e-01 4.65246253e-02 5.05643725e-01
1.09883167e-01 6.06551111e-01 -4.39074904e-01 -3.05360079e-01
4.40711260e-01 -3.13853770e-01 -6.55484796e-01 7.37492591e-02
3.58370304e-01 -6.05063200e-01 -8.40492070e-01 8.48766267e-01
4.50642616e-01 1.01526259e-02 6.03936553e-01 -1.75312877e+00
-1.66888177e-01 -3.31894346e-02 -3.51625592e-01 6.49842024e-01
3.31490785e-01 3.32644343e-01 2.07605794e-01 -6.66411221e-01
5.59715688e-01 1.08779955e+00 6.39203191e-01 5.17326236e-01
-5.15695095e-01 -9.35121834e-01 -1.33531705e-01 8.69960010e-01
-1.44098389e+00 -3.96330863e-01 8.85745883e-01 -5.60167789e-01
9.17736709e-01 3.18035781e-01 6.53558612e-01 1.42416608e+00
2.34078854e-01 4.83718097e-01 6.81007385e-01 -1.11061953e-01
-1.15840122e-01 2.96022911e-02 1.71795473e-01 5.41110754e-01
1.91398993e-01 2.72827893e-01 -2.33896822e-01 2.90167511e-01
4.09683228e-01 1.16286427e-01 -1.57185987e-01 3.90512720e-02
-1.29078197e+00 5.50005317e-01 -5.66222891e-02 8.04918036e-02
-2.28665143e-01 -2.02711225e-01 9.52925920e-01 2.14001611e-01
3.21155727e-01 -3.13723266e-01 -3.19718927e-01 -5.58281362e-01
-5.96054792e-01 1.42402232e-01 1.66245222e-01 1.14143074e+00
5.78131557e-01 2.54964828e-01 -4.95367289e-01 6.00723982e-01
-2.41408154e-01 5.98518968e-01 5.76483190e-01 -1.00299382e+00
6.83106482e-01 6.64030015e-01 -1.39069661e-01 -1.02776694e+00
-5.50992846e-01 -4.56902862e-01 -8.26560080e-01 2.11692989e-01
3.44291568e-01 -2.35745683e-01 -8.26434612e-01 1.67307448e+00
2.20939219e-01 6.31342769e-01 9.96123552e-02 1.10611725e+00
8.83050621e-01 5.28300166e-01 9.84281972e-02 -3.39140654e-01
1.53810644e+00 -8.22818220e-01 -8.98336351e-01 -5.25387898e-02
1.14016461e+00 -5.22542655e-01 8.08289230e-01 1.91128686e-01
-7.17101157e-01 -9.94963229e-01 -9.65256810e-01 9.49864089e-02
-4.34683591e-01 2.99805582e-01 1.01899453e-01 6.22865379e-01
-6.39282227e-01 1.09380290e-01 -4.83373612e-01 -5.11246979e-01
3.42286080e-01 1.76691473e-01 -8.12814534e-01 -7.83446282e-02
-1.51649272e+00 1.01222730e+00 6.03251159e-01 -2.94624791e-02
-9.10683930e-01 -6.01338565e-01 -8.94814610e-01 -2.64524758e-01
5.83203733e-01 -5.57459831e-01 9.59194303e-01 -9.76294160e-01
-9.16428745e-01 9.85364914e-01 -3.11792314e-01 -7.10492313e-01
6.61646187e-01 -7.90064335e-02 -1.02761281e+00 1.62210763e-01
2.57495672e-01 6.49368167e-01 6.07856154e-01 -7.84107447e-01
-1.03250241e+00 -2.04505250e-01 -5.93825951e-02 1.33877814e-01
-2.56138653e-01 2.07344025e-01 -5.55668533e-01 -4.34894472e-01
-1.35040894e-01 -1.08463824e+00 3.76994520e-01 -1.91183075e-01
-1.96453422e-01 -4.01379824e-01 1.60606372e+00 -7.99698770e-01
1.35438859e+00 -2.25576401e+00 -5.43844879e-01 -2.16175601e-01
3.24312225e-02 6.59145951e-01 -4.51978631e-02 2.83570737e-02
-6.44153476e-01 -7.83578902e-02 -8.31833929e-02 -6.73671439e-02
-4.34386849e-01 3.29538107e-01 -2.73530126e-01 4.73461151e-01
2.96121836e-01 6.84523761e-01 -7.96001732e-01 -7.06659734e-01
5.32142937e-01 1.69459596e-01 -3.57124627e-01 1.19856805e-01
4.26187634e-01 5.53206146e-01 -3.00451875e-01 4.81572419e-01
6.97746336e-01 5.53988934e-01 -5.67122281e-01 -5.67831218e-01
-2.35307112e-01 -8.52205232e-02 -1.06067300e+00 1.41218686e+00
-1.83436170e-01 1.03244174e+00 -3.76308829e-01 -1.31246257e+00
7.93078005e-01 6.10738575e-01 7.34421015e-01 -9.42781389e-01
9.48964730e-02 -8.53591338e-02 -1.23620160e-01 -1.29219770e+00
3.54238987e-01 1.53728470e-01 -1.08545959e-01 -5.37568629e-02
-1.64168984e-01 5.48718512e-01 3.22910488e-01 1.14749655e-01
1.08722055e+00 1.55789047e-01 -7.37230331e-02 4.31402065e-02
8.31959784e-01 1.68936104e-01 8.69967997e-01 4.24013257e-01
-7.05861807e-01 8.55745614e-01 4.59944993e-01 -6.66644633e-01
-1.27166426e+00 -9.48744595e-01 -1.75962910e-01 5.24398029e-01
1.82658538e-01 -1.62857980e-01 -8.55530083e-01 -8.71485651e-01
-5.01424611e-01 1.08630788e+00 -4.47252601e-01 -6.08441949e-01
-7.42832720e-01 -5.80194950e-01 7.97141969e-01 5.88306248e-01
1.00147545e+00 -7.99507022e-01 -3.86625588e-01 -2.25297622e-02
-7.67103493e-01 -1.60549033e+00 -5.39265096e-01 -4.38105583e-01
-3.62315923e-01 -1.41185975e+00 -5.14065266e-01 -6.05745614e-01
4.99844849e-01 5.10008395e-01 6.50272429e-01 -1.17125712e-01
-2.93873012e-01 4.16633815e-01 -3.17034662e-01 -4.69837189e-01
-6.36235416e-01 -2.44591966e-01 3.61304015e-01 3.62774193e-01
7.37702429e-01 -2.53638893e-01 -6.56699359e-01 8.05569828e-01
-6.57392502e-01 2.42485896e-01 3.93486381e-01 3.60193014e-01
3.27665508e-01 2.02077463e-01 6.25385940e-01 -2.61266649e-01
3.09340894e-01 -7.31313229e-01 -1.75749913e-01 -5.56135550e-02
-3.83195460e-01 -4.41297650e-01 4.03174281e-01 -3.06377590e-01
-1.10658681e+00 1.23584159e-02 -3.66417438e-01 -8.49620998e-01
-7.80658066e-01 3.79781239e-02 -1.73866540e-01 4.11151886e-01
5.04758835e-01 2.72244543e-01 1.06553636e-01 -4.78891097e-02
-1.15150042e-01 1.00011969e+00 9.91633594e-01 -1.92905068e-01
4.18440133e-01 5.09251654e-01 2.58996725e-01 -9.31878328e-01
-6.23732686e-01 -7.29005277e-01 -7.67840624e-01 -7.95644403e-01
1.36484218e+00 -1.12093854e+00 -7.79308498e-01 4.44429606e-01
-1.23902702e+00 1.61425963e-01 -1.08285002e-01 8.20964992e-01
-5.15781283e-01 6.32428408e-01 -1.16341487e-01 -7.31052995e-01
4.31404300e-02 -1.31041324e+00 9.99658823e-01 -3.04170791e-02
-4.19654101e-02 -4.84898597e-01 -9.38024223e-02 6.42880380e-01
3.33016664e-02 3.67303789e-01 7.39968717e-01 -6.25887096e-01
-2.32801899e-01 -5.35732329e-01 -3.69423151e-01 5.40410757e-01
1.11965269e-01 8.20137262e-02 -8.83153498e-01 1.04427338e-02
-9.20999870e-02 2.53315300e-01 7.53661275e-01 3.88254523e-01
1.26269031e+00 -7.96708316e-02 -5.03658772e-01 4.94675219e-01
8.58711243e-01 5.52444696e-01 1.14974380e+00 4.85477298e-01
7.43922770e-01 7.34924197e-01 7.51302838e-01 1.02109583e-02
4.84608889e-01 1.08658814e+00 5.90593278e-01 -1.35273173e-01
-4.99057531e-01 -1.15966029e-01 5.95552683e-01 5.05141497e-01
-3.01472366e-01 -2.96720564e-01 -8.79157364e-01 6.50525808e-01
-1.92045557e+00 -1.48303115e+00 -8.08085978e-01 2.39289737e+00
2.82809913e-01 3.36389869e-01 4.48352158e-01 3.92258555e-01
1.07818806e+00 -2.39269271e-01 -2.82551318e-01 -4.24044639e-01
-1.37630105e-01 -5.48832595e-01 6.56643033e-01 1.14144348e-01
-1.20903575e+00 5.57535052e-01 5.24564266e+00 1.08837831e+00
-1.05714989e+00 3.87099802e-01 7.50861168e-01 -8.07362944e-02
3.39990199e-01 -3.07170570e-01 -8.17531705e-01 7.68191516e-01
1.26502991e+00 1.05101220e-01 -1.77793398e-01 8.04960430e-01
8.59645367e-01 -9.02178418e-03 -9.77463543e-01 1.20962071e+00
2.70776391e-01 -1.15271258e+00 1.36335284e-01 -1.60172239e-01
3.63142163e-01 -1.56332508e-01 -2.17362687e-01 5.67924500e-01
-5.46526849e-01 -6.12968504e-01 7.62042165e-01 6.99869037e-01
1.14584088e+00 -7.72568822e-01 1.03589213e+00 3.85389358e-01
-1.25266278e+00 -2.36969858e-01 -9.11757573e-02 -7.05628246e-02
4.21745688e-01 1.58623308e-01 -7.28557169e-01 6.81714416e-01
8.58164847e-01 6.60093129e-01 -5.76474905e-01 8.95937979e-01
2.75294423e-01 6.94026709e-01 -1.53967336e-01 4.95033205e-01
3.29037219e-01 -4.73036021e-02 8.94112051e-01 1.29868066e+00
6.41217649e-01 2.58821964e-01 5.43740205e-02 5.44742465e-01
3.06212395e-01 -2.39770889e-01 -1.11845040e+00 5.31626284e-01
1.11888602e-01 1.00635910e+00 -5.37829399e-01 -4.02517974e-01
-7.10294366e-01 8.63097072e-01 -2.57330507e-01 3.41579050e-01
-1.47075808e+00 -3.45787793e-01 7.17126071e-01 5.94407260e-01
5.71295694e-02 -2.20821664e-01 -1.27588823e-01 -9.11408007e-01
9.46435109e-02 -3.88698369e-01 4.40407604e-01 -1.18316984e+00
-6.91251397e-01 5.26063085e-01 5.37959039e-01 -1.82147396e+00
-1.91005692e-01 -6.02090299e-01 -8.66345167e-01 4.66159523e-01
-1.27872741e+00 -1.12946570e+00 -6.18185759e-01 7.81096935e-01
9.27745104e-01 -4.29081947e-01 4.20484483e-01 7.66024232e-01
-8.93329501e-01 7.86731601e-01 -2.83160180e-01 4.09277111e-01
7.71009862e-01 -5.59766710e-01 1.16190135e-01 1.01633620e+00
-4.01687682e-01 -1.23510778e-01 5.98600984e-01 -7.71923065e-01
-9.77424204e-01 -1.51534235e+00 8.50805581e-01 -5.45389652e-01
4.00773793e-01 -1.88561212e-02 -6.66867435e-01 7.18981326e-01
3.39172706e-02 4.45559099e-02 1.76047608e-01 -5.98796964e-01
7.59497508e-02 -3.46675277e-01 -1.11654162e+00 6.54409885e-01
1.16168141e+00 -4.23878163e-01 -6.29298508e-01 3.54585707e-01
7.17376113e-01 -4.11809802e-01 -4.10008371e-01 6.29032969e-01
2.69392908e-01 -1.14428997e+00 9.82282221e-01 -5.66333294e-01
4.67850864e-01 -5.11400759e-01 4.49088439e-02 -7.66569853e-01
-2.55907506e-01 -4.12498504e-01 1.74179673e-01 1.24005508e+00
1.42039880e-01 -5.95077574e-01 6.01600528e-01 6.64661109e-01
-9.31810260e-01 -4.98051465e-01 -1.32150507e+00 -8.09818864e-01
-3.52103710e-01 -1.17792010e+00 5.02727985e-01 8.13238084e-01
-2.37477958e-01 1.00413792e-01 -4.63707179e-01 3.28983814e-01
4.41536933e-01 -6.20953858e-01 8.42976034e-01 -8.59228432e-01
4.18503761e-01 -3.39055538e-01 -9.26954925e-01 -7.93308914e-01
8.10680240e-02 -8.27490926e-01 -3.29000831e-01 -1.05466318e+00
2.06036612e-01 -2.26170905e-02 -1.45690367e-01 4.79645312e-01
1.22529544e-01 4.30227101e-01 -1.88465536e-01 1.79169938e-01
-4.32354182e-01 3.21693182e-01 1.11418736e+00 -1.68649688e-01
8.91793147e-02 1.57243580e-01 -5.47465757e-02 8.34411621e-01
7.59052098e-01 -3.64647210e-01 -4.73223358e-01 -1.67258829e-01
-2.65531857e-02 2.90901244e-01 7.63980150e-01 -1.27948213e+00
1.31702736e-01 -8.39743391e-02 2.12625802e-01 -1.02185404e+00
2.92568207e-01 -1.13799775e+00 2.94409633e-01 3.55092198e-01
-1.11349963e-01 2.56899655e-01 4.08558100e-01 7.36169100e-01
-2.57582605e-01 -2.56936133e-01 5.99922299e-01 1.25520289e-01
-1.14601529e+00 3.61399472e-01 -7.39922404e-01 -1.02298923e-01
1.63058662e+00 -7.99670517e-01 -3.93007398e-01 -4.45520729e-01
-6.94844544e-01 1.28413662e-01 3.46440226e-02 7.65718937e-01
5.70709884e-01 -1.58673263e+00 -7.83540905e-01 5.43291807e-01
3.71472895e-01 -2.03203768e-01 6.80669010e-01 1.16050124e+00
-5.36010265e-01 7.11112380e-01 -3.27452868e-01 -8.66574585e-01
-1.47216964e+00 6.88676059e-01 4.51390892e-01 2.45953217e-01
-8.02372456e-01 3.26749474e-01 3.02246362e-01 -8.56989343e-03
1.24664411e-01 -3.02100748e-01 -4.10456210e-01 -1.97731584e-01
4.83220279e-01 7.16596425e-01 1.50065407e-01 -1.29729819e+00
-3.41264665e-01 5.71397960e-01 1.30891278e-01 2.89064437e-01
6.68790042e-01 -4.10821766e-01 2.43374780e-01 2.39157438e-01
1.36455941e+00 -5.13066769e-01 -1.16848159e+00 6.40991256e-02
-4.90423948e-01 -3.53028774e-01 1.84023947e-01 -4.97562706e-01
-1.20886648e+00 8.92778993e-01 1.20276594e+00 2.29712315e-02
1.03132224e+00 -1.32471398e-01 1.10505116e+00 3.52804899e-01
1.42779186e-01 -1.20590889e+00 -1.54556692e-01 3.53810906e-01
8.84834349e-01 -1.46497321e+00 -3.78846765e-01 -2.34199002e-01
-8.66766751e-01 1.32376277e+00 7.28944778e-01 2.43500054e-01
5.19076884e-01 -5.24254777e-02 -7.83926845e-02 -1.00255422e-01
-3.79706889e-01 -1.44010738e-01 4.52508688e-01 8.06349754e-01
-1.04820736e-01 -7.29557052e-02 -2.46096730e-01 6.25166953e-01
-3.30970064e-02 7.75155500e-02 5.52260220e-01 5.94650328e-01
-2.91130006e-01 -4.79498327e-01 -4.13012564e-01 4.26481038e-01
-3.13877553e-01 2.80335754e-01 1.87842205e-01 9.41729963e-01
8.01842391e-01 1.17114139e+00 4.74544078e-01 -6.24240637e-01
5.73382556e-01 2.83882499e-01 -1.68924569e-04 -1.54137135e-01
-1.17092580e-01 -2.68103600e-01 4.20943618e-01 -6.57236814e-01
-4.18859303e-01 -7.44162679e-01 -1.18466043e+00 -4.13501740e-01
-1.64246812e-01 -1.36544779e-01 5.33027351e-01 1.19054675e+00
4.60610420e-01 8.01884532e-01 6.14770770e-01 -7.18739629e-01
1.76277474e-01 -8.32028866e-01 -1.66519135e-01 7.45957494e-01
3.34575981e-01 -8.51686835e-01 -3.90972883e-01 4.57890868e-01] | [7.676427841186523, -0.2301357537508011] |
aa62ee54-bdf7-4506-a5d1-91291aa7be81 | auxiliary-learning-as-an-asymmetric | 2301.13501 | null | https://arxiv.org/abs/2301.13501v2 | https://arxiv.org/pdf/2301.13501v2.pdf | Auxiliary Learning as an Asymmetric Bargaining Game | Auxiliary learning is an effective method for enhancing the generalization capabilities of trained models, particularly when dealing with small datasets. However, this approach may present several difficulties: (i) optimizing multiple objectives can be more challenging, and (ii) how to balance the auxiliary tasks to best assist the main task is unclear. In this work, we propose a novel approach, named AuxiNash, for balancing tasks in auxiliary learning by formalizing the problem as generalized bargaining game with asymmetric task bargaining power. Furthermore, we describe an efficient procedure for learning the bargaining power of tasks based on their contribution to the performance of the main task and derive theoretical guarantees for its convergence. Finally, we evaluate AuxiNash on multiple multi-task benchmarks and find that it consistently outperforms competing methods. | ['Ethan Fetaya', 'Gal Chechik', 'Kenji Kawaguchi', 'Neta Glazer', 'Aviv Navon', 'Aviv Shamsian'] | 2023-01-31 | null | null | null | null | ['auxiliary-learning'] | ['methodology'] | [ 1.94777712e-01 1.24725580e-01 -4.60522890e-01 -3.41855660e-02
-1.19689989e+00 -5.88616610e-01 3.22840124e-01 -9.57644135e-02
-8.65312278e-01 1.19269693e+00 7.25862831e-02 -4.17037904e-01
-5.43294728e-01 -1.85418889e-01 -6.06598616e-01 -1.14528942e+00
1.88357249e-01 6.78945541e-01 -2.27218077e-01 -3.63255113e-01
1.44710660e-01 8.62237662e-02 -1.20071089e+00 3.60713042e-02
1.30195522e+00 1.19800365e+00 3.74236435e-01 3.44691694e-01
3.21454741e-02 6.81129038e-01 -7.22427487e-01 -8.97476792e-01
5.01327932e-01 -2.34863177e-01 -1.14092696e+00 -2.09276602e-01
1.01748906e-01 9.54311863e-02 2.70984739e-01 9.53442216e-01
5.24120808e-01 3.71119410e-01 9.01690543e-01 -1.83561099e+00
-1.35816455e-01 8.05084646e-01 -6.01055503e-01 3.37101609e-01
-3.59990478e-01 -2.64684379e-01 1.34297562e+00 -4.87512827e-01
3.37743402e-01 1.08579051e+00 5.42612910e-01 6.86088324e-01
-1.53901386e+00 -7.58195579e-01 4.09373939e-01 8.79542083e-02
-1.01806426e+00 -5.25224209e-01 8.41858685e-01 -2.22463831e-01
4.52467322e-01 2.10002959e-01 1.26164407e-01 1.46759129e+00
2.78047770e-02 1.05241883e+00 1.32905960e+00 -2.46813729e-01
1.21221304e-01 2.42634878e-01 1.98589072e-01 3.56688768e-01
1.99425653e-01 -1.78810820e-01 -4.56521690e-01 -2.92343944e-01
3.06748152e-01 -3.72476667e-01 -1.80608377e-01 -5.55902898e-01
-1.02361453e+00 9.69639838e-01 3.76094013e-01 1.35461375e-01
-4.11714733e-01 1.34018943e-01 6.44729614e-01 6.11807466e-01
8.85362744e-01 9.28006709e-01 -8.92694354e-01 -6.26222789e-02
-5.03901660e-01 4.44251120e-01 6.86388195e-01 7.85855472e-01
4.17373031e-01 -6.15538917e-02 -4.01032507e-01 8.92700493e-01
-8.43034033e-03 1.97821334e-01 2.00002939e-02 -1.19765222e+00
9.99330044e-01 2.59834200e-01 2.00710759e-01 -3.58994514e-01
-6.66374505e-01 -7.99297154e-01 -1.12810361e+00 2.16573969e-01
6.84511960e-01 -4.89280105e-01 -5.73290646e-01 2.11718559e+00
9.05716345e-02 -1.85182050e-01 1.11082092e-01 7.30016112e-01
5.24009287e-01 3.05523545e-01 3.95909518e-01 -4.46270198e-01
1.21405566e+00 -1.24356329e+00 -6.36174977e-01 -5.79851031e-01
5.98009288e-01 -6.03005409e-01 1.04631960e+00 4.46983039e-01
-1.21501386e+00 -4.35542345e-01 -5.20848215e-01 1.36700496e-02
-5.93498647e-02 1.88066840e-01 8.09775889e-01 5.67079365e-01
-8.86911511e-01 4.59547758e-01 -3.99521261e-01 3.41539592e-01
6.12308204e-01 6.79533780e-01 -1.33366913e-01 3.62590700e-01
-1.20976484e+00 1.06147730e+00 4.97744769e-01 -2.54753023e-01
-6.44357800e-01 -9.67290580e-01 -7.22093165e-01 3.67670000e-01
8.82137179e-01 -7.84880161e-01 1.50678992e+00 -9.73302543e-01
-1.23916650e+00 6.42165601e-01 1.36693165e-01 -2.98332840e-01
7.52202272e-01 -2.13959888e-01 3.85930657e-01 -3.38708878e-01
2.82192320e-01 6.01202011e-01 8.96076620e-01 -1.42333663e+00
-1.00125217e+00 -4.77198690e-01 4.84839678e-01 5.94551861e-01
-5.65044582e-01 -6.05013557e-02 -2.52627224e-01 -8.54036808e-01
-3.18833560e-01 -1.02533948e+00 -6.27202094e-01 -3.73709112e-01
-2.55478233e-01 -7.65575826e-01 4.80507702e-01 -5.20836234e-01
1.08143103e+00 -2.19218779e+00 7.03872204e-01 1.75217707e-02
4.71615881e-01 2.47725308e-01 -2.50968337e-01 -7.84045979e-02
5.82374493e-03 2.77738959e-01 5.56254126e-02 -8.52941632e-01
1.87770411e-01 2.96709001e-01 -2.87318617e-01 1.05334334e-01
-5.56084104e-02 1.06967437e+00 -8.58346462e-01 -4.79182005e-01
-2.75344878e-01 -9.62489098e-02 -5.69247305e-01 2.29480073e-01
-5.52549548e-02 7.08052695e-01 -7.05310225e-01 5.38974822e-01
5.12268841e-01 -2.28378490e-01 2.28023678e-01 5.62715046e-02
2.63526380e-01 3.28852713e-01 -1.02294886e+00 1.29141819e+00
-7.21286952e-01 2.83824056e-01 7.74703860e-01 -1.58486152e+00
5.91551006e-01 2.19632953e-01 6.39046550e-01 -6.49246335e-01
1.96958363e-01 1.22163475e-01 8.90657008e-02 -2.24118039e-01
4.27796602e-01 -2.76657343e-01 -3.60986501e-01 4.98398930e-01
2.19281226e-01 2.51529180e-02 3.38030964e-01 -8.71491060e-02
7.04149127e-01 -8.43547881e-02 3.48692089e-01 -5.02778232e-01
3.64171445e-01 -2.11334988e-01 6.81278348e-01 9.66489315e-01
-3.18263859e-01 2.21715774e-02 1.12608051e+00 -2.17344329e-01
-9.67737973e-01 -1.00772440e+00 -1.37856767e-01 1.87218559e+00
4.75657694e-02 -2.06175297e-01 -4.11269695e-01 -1.26590264e+00
1.23973332e-01 5.68206847e-01 -7.88549900e-01 -1.33150920e-01
-4.88185436e-01 -9.87358570e-01 3.67497146e-01 7.44556129e-01
4.27592933e-01 -8.53422463e-01 -3.18098038e-01 -7.96185806e-02
-7.28036582e-01 -1.09099793e+00 -5.77244818e-01 8.29869628e-01
-8.00809920e-01 -1.11249638e+00 -8.98134530e-01 -7.62368381e-01
6.93075061e-01 3.49091560e-01 1.36734748e+00 1.75755267e-05
1.95390418e-01 1.14713065e-01 -3.65120888e-01 -8.00103426e-01
-2.85606176e-01 7.35479236e-01 1.06820583e-01 -4.02732998e-01
-1.28224596e-01 -4.01590943e-01 -2.86229968e-01 4.97953981e-01
-6.64522827e-01 6.24696352e-03 6.99850500e-01 1.15567303e+00
2.26777926e-01 6.38312623e-02 9.54351306e-01 -9.90841508e-01
1.05147457e+00 -5.55008233e-01 -5.65273523e-01 4.02067095e-01
-8.83476675e-01 3.28596711e-01 6.90003574e-01 -5.19568622e-01
-1.22448647e+00 -1.38138920e-01 -6.70916811e-02 -3.23761135e-01
3.52314472e-01 5.43266892e-01 -1.07577048e-01 -2.86672264e-02
6.55537844e-01 -1.06576517e-01 1.61622122e-01 -4.78459924e-01
1.04386806e-01 5.23300171e-01 1.69790357e-01 -9.30864811e-01
8.80599558e-01 1.34517595e-01 3.16246241e-01 -2.51677364e-01
-1.35290873e+00 -4.97447252e-01 -5.74370801e-01 -1.25405379e-02
5.99878013e-01 -8.03985178e-01 -9.86785710e-01 2.40308687e-01
-8.82487059e-01 -7.78358221e-01 -1.54826894e-01 3.83613825e-01
-7.59791553e-01 1.53459415e-01 -3.43445271e-01 -8.83279324e-01
-3.03210407e-01 -1.30534816e+00 1.00228035e+00 1.31413182e-02
4.60380167e-02 -1.16176629e+00 -9.91151705e-02 8.95632684e-01
3.41956198e-01 -7.67726898e-02 1.09888375e+00 -7.96133518e-01
-1.70082062e-01 2.84699500e-01 -8.32752734e-02 4.84969497e-01
-3.94760258e-02 -6.19981587e-01 -7.51863897e-01 -4.66949463e-01
-5.80491452e-03 -8.85416746e-01 1.06414175e+00 5.08817494e-01
1.48099136e+00 -4.84213024e-01 -2.78272688e-01 6.11226737e-01
9.32677031e-01 -5.61391301e-02 3.51400375e-01 5.90204716e-01
3.82885218e-01 7.29407728e-01 9.53691781e-01 4.83775795e-01
3.62780482e-01 8.35172653e-01 5.17509937e-01 -1.72669038e-01
3.61827701e-01 1.96139693e-01 3.61126184e-01 3.52023691e-01
-5.95473886e-01 -1.23024262e-01 -7.09316552e-01 2.45923042e-01
-2.13041639e+00 -7.63674080e-01 7.50939846e-02 2.06783986e+00
8.61022949e-01 2.79746324e-01 3.89298975e-01 1.29178166e-01
4.04428720e-01 1.48458555e-01 -5.30395985e-01 -3.59965116e-01
-8.19501132e-02 3.13066959e-01 5.70906997e-01 2.46586546e-01
-1.34683299e+00 7.72976100e-01 6.90178013e+00 1.10388577e+00
-8.02568614e-01 4.28007782e-01 9.59868789e-01 -1.86352134e-01
-9.00952816e-02 -3.04432303e-01 -7.62265682e-01 2.50270367e-01
4.10016447e-01 -3.41283053e-01 6.10664248e-01 1.04287398e+00
-1.06526278e-01 -1.58343464e-02 -1.05574846e+00 7.63194442e-01
-9.75749865e-02 -8.92684042e-01 -2.73019731e-01 1.92637697e-01
9.21620190e-01 -1.48664460e-01 3.84984732e-01 1.02364242e+00
4.65906352e-01 -9.97395039e-01 7.18382776e-01 -9.12536085e-02
5.70541024e-01 -1.01808965e+00 9.67209458e-01 8.06455314e-01
-8.93346369e-01 -5.63766181e-01 -3.40042681e-01 -1.53151363e-01
-1.85005128e-01 1.25564381e-01 -5.48402667e-01 5.66492438e-01
7.91364491e-01 1.47335082e-01 -5.50509036e-01 7.97670782e-01
-3.14322323e-01 4.30568755e-01 4.69206497e-02 -8.46714247e-03
3.75312567e-01 -3.65521848e-01 5.25484681e-01 7.88828015e-01
5.26506156e-02 -9.69039053e-02 3.30826521e-01 5.79481840e-01
-5.44794619e-01 2.08184928e-01 -4.75937992e-01 1.69323817e-01
1.80388331e-01 1.51161802e+00 -6.10249579e-01 8.59476402e-02
-3.29239786e-01 6.24783754e-01 7.50759661e-01 4.08346742e-01
-9.19790983e-01 -2.33256668e-02 7.61893213e-01 -2.04415500e-01
6.53957278e-02 -7.61834159e-02 -2.99430341e-01 -1.06190467e+00
1.25035167e-01 -1.10257077e+00 7.53507316e-01 -5.01398623e-01
-1.32683909e+00 3.92089128e-01 1.41635343e-01 -1.01223636e+00
-1.33100301e-01 -7.47677147e-01 -5.59124470e-01 8.00264239e-01
-1.58195913e+00 -1.17901611e+00 -1.91826299e-01 6.42150640e-01
6.12815678e-01 -4.06804144e-01 6.04515433e-01 3.28584790e-01
-5.66132069e-01 8.49359035e-01 3.62737924e-01 2.51664817e-02
8.92492294e-01 -1.60373533e+00 -7.40677938e-02 4.63430315e-01
-7.67157078e-02 1.05323292e-01 5.47492683e-01 -2.59383053e-01
-1.11044848e+00 -9.24914718e-01 6.63941562e-01 -3.49555790e-01
6.54887378e-01 -6.02313936e-01 -5.02908826e-01 7.60415733e-01
9.51646343e-02 -2.64532149e-01 7.08278120e-01 8.76857460e-01
-1.78983033e-01 -2.51254380e-01 -8.07758152e-01 6.63281679e-01
1.03042245e+00 -1.25950336e-01 -5.33122778e-01 5.47846675e-01
5.98046541e-01 -5.18808007e-01 -6.02851331e-01 5.55326283e-01
2.71775454e-01 -5.86267829e-01 1.10626674e+00 -1.10858059e+00
5.89244783e-01 3.90231490e-01 1.02489498e-02 -1.83722842e+00
-5.69686651e-01 -6.13644123e-01 -1.39999449e-01 1.12679529e+00
5.82288086e-01 -6.21636212e-01 9.54393685e-01 5.90517819e-01
5.25851957e-02 -1.02580488e+00 -1.15925610e+00 -9.57929194e-01
5.05519271e-01 -1.76060349e-01 2.20819667e-01 9.20177758e-01
-1.41801849e-01 6.19332135e-01 -6.41992748e-01 -2.76775748e-01
7.79856145e-01 1.99008922e-04 7.06042051e-01 -1.54408944e+00
-2.08576351e-01 -7.68167377e-01 2.26521090e-01 -8.72947812e-01
9.24543083e-01 -1.05375028e+00 2.89110690e-02 -1.23532295e+00
5.23926735e-01 -8.97565842e-01 -6.20385945e-01 6.50997460e-01
-6.42091215e-01 8.25890601e-02 3.62068892e-01 1.79854199e-01
-8.63060117e-01 7.12024450e-01 1.38833201e+00 -2.86910862e-01
-2.05950961e-01 7.96474636e-01 -1.41598451e+00 6.51585400e-01
1.01407015e+00 -6.42425239e-01 -4.90758508e-01 -2.31736854e-01
3.85831982e-01 7.42733330e-02 2.40517512e-01 -3.61523241e-01
9.83459651e-02 -4.49118912e-01 1.67359978e-01 -1.89158484e-01
2.49999017e-01 -9.19030845e-01 -3.65418196e-01 6.25661850e-01
-6.17102027e-01 4.36024815e-02 6.78189322e-02 5.82002640e-01
1.54690677e-02 -5.67728221e-01 7.88519681e-01 1.49416579e-02
-2.41364673e-01 2.95816213e-01 -5.09914458e-02 5.18145204e-01
1.15121543e+00 4.39296961e-01 -2.80677915e-01 -4.70089912e-01
-6.48294568e-01 7.08889008e-01 -3.98467593e-02 4.86966848e-01
5.39409854e-02 -1.28867388e+00 -9.01344299e-01 -3.26169580e-01
1.06870104e-02 7.17949262e-03 9.48652327e-02 1.18406141e+00
1.62691191e-01 4.27834272e-01 -2.88831979e-01 -2.97269195e-01
-1.63944304e+00 7.14338124e-01 1.22857071e-01 -1.09811938e+00
-2.66122632e-02 7.26719022e-01 4.72367138e-01 -6.85392678e-01
4.58016783e-01 2.44834963e-02 -4.09042716e-01 5.22033274e-01
1.08754016e-01 5.91295779e-01 8.63723680e-02 -3.65504831e-01
-1.33316085e-01 1.60114139e-01 -3.22970226e-02 5.75378574e-02
1.50278699e+00 -2.56007127e-02 -1.63981348e-01 1.39276668e-01
9.52073336e-01 -1.04491375e-01 -1.27677500e+00 -5.23908317e-01
1.54528305e-01 -3.32388967e-01 8.68378580e-02 -6.73841894e-01
-1.13396227e+00 5.38566411e-01 6.47852495e-02 3.13981920e-01
1.35076606e+00 1.26709208e-01 2.64806688e-01 7.33848572e-01
3.45935255e-01 -1.39220428e+00 3.64945918e-01 6.22787356e-01
8.84468853e-01 -1.42473936e+00 7.73720369e-02 -3.47013503e-01
-1.01428342e+00 8.83607745e-01 8.41072857e-01 2.25592509e-01
6.27540469e-01 1.55533627e-01 -8.21651891e-02 -1.08950660e-02
-9.09499347e-01 -2.59281099e-01 4.50891256e-01 4.87053305e-01
1.84708998e-01 -5.21080382e-02 -5.44100821e-01 1.03712428e+00
-8.90904590e-02 -3.41991633e-01 1.37690440e-01 7.41942585e-01
-1.37870282e-01 -1.35740793e+00 -3.94320250e-01 7.79454291e-01
-8.60437512e-01 -4.97678705e-02 -3.69084388e-01 7.69353092e-01
1.75665945e-01 8.24279606e-01 -1.81575403e-01 -3.09823249e-02
2.23536387e-01 1.14020877e-01 5.03409564e-01 -4.73784149e-01
-6.56754971e-01 3.23422253e-02 1.62076652e-01 -3.91576529e-01
-5.00201762e-01 -4.89730328e-01 -5.10269403e-01 -2.20931605e-01
-4.14424956e-01 4.80970889e-01 4.49370176e-01 1.02851284e+00
1.24544047e-01 6.70406163e-01 8.58602643e-01 -9.00728524e-01
-1.26411355e+00 -8.94077718e-01 -6.93473697e-01 4.94854778e-01
2.15390906e-01 -1.04724348e+00 -4.03700382e-01 -2.44736165e-01] | [9.294780731201172, 3.855907917022705] |
dfcbc7a2-b42f-4070-9b02-94200f14d519 | how-does-feedback-signal-quality-impact | 2205.05888 | null | https://arxiv.org/abs/2205.05888v1 | https://arxiv.org/pdf/2205.05888v1.pdf | How does Feedback Signal Quality Impact Effectiveness of Pseudo Relevance Feedback for Passage Retrieval? | Pseudo-Relevance Feedback (PRF) assumes that the top results retrieved by a first-stage ranker are relevant to the original query and uses them to improve the query representation for a second round of retrieval. This assumption however is often not correct: some or even all of the feedback documents may be irrelevant. Indeed, the effectiveness of PRF methods may well depend on the quality of the feedback signal and thus on the effectiveness of the first-stage ranker. This aspect however has received little attention before. In this paper we control the quality of the feedback signal and measure its impact on a range of PRF methods, including traditional bag-of-words methods (Rocchio), and dense vector-based methods (learnt and not learnt). Our results show the important role the quality of the feedback signal plays on the effectiveness of PRF methods. Importantly, and surprisingly, our analysis reveals that not all PRF methods are the same when dealing with feedback signals of varying quality. These findings are critical to gain a better understanding of the PRF methods and of which and when they should be used, depending on the feedback signal quality, and set the basis for future research in this area. | ['Guido Zuccon', 'Bevan Koopman', 'Ahmed Mourad', 'Hang Li'] | 2022-05-12 | null | null | null | null | ['passage-retrieval'] | ['natural-language-processing'] | [ 3.02327812e-01 -4.19165879e-01 -4.11663324e-01 -1.34896204e-01
-1.00625527e+00 -8.53769362e-01 8.03774714e-01 4.65709031e-01
-6.20160282e-01 5.97310722e-01 6.56421602e-01 -1.51494607e-01
-4.95032787e-01 -6.86575353e-01 -6.27361953e-01 -5.17587066e-01
-9.21417177e-02 3.96830618e-01 4.70862597e-01 -6.95838630e-01
7.10316658e-01 3.54220718e-01 -1.56720769e+00 4.25617248e-01
3.64108086e-01 9.33734119e-01 1.34156823e-01 8.15106452e-01
4.22159061e-02 6.88798547e-01 -8.81559670e-01 -1.63095579e-01
4.21505243e-01 -4.49760914e-01 -6.72392786e-01 -1.33542344e-01
4.19203758e-01 -4.53982741e-01 -5.03324449e-01 8.90080392e-01
4.77188319e-01 3.73617798e-01 7.93580651e-01 -7.65956819e-01
-4.26746547e-01 6.17168903e-01 -2.94733465e-01 7.80448675e-01
5.81429720e-01 -8.32727924e-02 1.65598166e+00 -1.06631279e+00
6.44974291e-01 1.32779360e+00 2.24229157e-01 2.12481558e-01
-1.23556519e+00 -6.81304336e-01 1.82213098e-01 3.24340999e-01
-1.22543287e+00 -6.07082903e-01 5.37662804e-01 -3.91770035e-01
4.43755358e-01 5.49224913e-01 6.88343823e-01 6.64648175e-01
2.62731284e-01 7.97793031e-01 1.01438320e+00 -6.77007437e-01
2.14011103e-01 3.22127551e-01 3.82139534e-01 5.37388511e-02
3.69833052e-01 4.51892525e-01 -8.74469757e-01 -5.06363928e-01
3.06437254e-01 -1.24879345e-01 -6.15971982e-01 -2.60623366e-01
-7.70680904e-01 1.12095547e+00 5.11596084e-01 6.07445061e-01
-3.82521898e-01 2.34914869e-01 1.00881100e-01 6.44266307e-01
3.77614558e-01 9.89859223e-01 -3.37506860e-01 -1.75644323e-01
-1.01884985e+00 5.64503312e-01 6.67678297e-01 2.91312903e-01
7.47659743e-01 -2.49046132e-01 -7.07628131e-01 8.16210926e-01
2.79593587e-01 4.38868433e-01 6.05203152e-01 -7.18769550e-01
1.90574706e-01 4.13973182e-01 4.07989711e-01 -1.37151659e+00
-5.23175932e-02 -4.41679090e-01 -8.44020280e-04 -4.72109653e-02
2.45044842e-01 1.84970751e-01 -6.42512679e-01 1.37465870e+00
-7.77691454e-02 -3.83487046e-01 -4.00612980e-01 1.32596958e+00
4.24847215e-01 6.08307004e-01 -3.43212485e-01 -2.44015381e-01
1.09597373e+00 -6.87493443e-01 -6.94530189e-01 -5.43254793e-01
4.33293223e-01 -1.03901148e+00 1.12297297e+00 4.07188654e-01
-8.57847512e-01 -3.78455281e-01 -1.14244521e+00 2.20930979e-01
-1.64510667e-01 -4.12323736e-02 5.72609961e-01 5.50130844e-01
-1.16211557e+00 7.16702759e-01 -2.38622978e-01 -1.93265483e-01
-1.11058149e-02 3.10177058e-01 -1.38149887e-01 -4.73674804e-01
-1.59693575e+00 8.99374962e-01 7.68008009e-02 -7.30245337e-02
-7.73545504e-01 -4.35397953e-01 -2.74432182e-01 3.00505936e-01
4.39280838e-01 -3.81057441e-01 1.56514275e+00 -9.68674660e-01
-9.05257940e-01 1.68570518e-01 -3.18767309e-01 -3.57904047e-01
3.51358205e-01 -4.48152691e-01 -1.43344462e-01 3.87471348e-01
1.08512893e-01 3.46808165e-01 9.06323731e-01 -1.15635848e+00
-6.46768510e-01 -4.93859053e-01 1.57965720e-01 4.64965880e-01
-2.67483711e-01 -7.03729168e-02 -5.05745351e-01 -5.52942395e-01
4.60209921e-02 -8.61108661e-01 2.10449984e-03 -1.38299778e-01
8.63755271e-02 -1.82763085e-01 3.61460835e-01 -3.10414433e-01
1.69607103e+00 -2.15074277e+00 -1.91271544e-01 5.66744566e-01
1.42951801e-01 1.33434072e-01 -4.09806371e-01 9.21657741e-01
1.39815629e-01 2.65941143e-01 2.70155877e-01 2.36020237e-01
-1.41429856e-01 -5.40931001e-02 -5.20048320e-01 3.97893846e-01
-5.43795992e-03 9.26466286e-01 -1.08080709e+00 -2.84633785e-01
1.02424650e-02 4.28909898e-01 -5.29068291e-01 1.90975964e-01
-8.20256323e-02 1.23003803e-01 -5.74020922e-01 1.80060282e-01
9.67422873e-02 -3.13747883e-01 8.61631557e-02 -1.58234105e-01
-6.75069913e-02 6.75060391e-01 -1.02297866e+00 9.44698274e-01
-1.77917331e-01 7.83408880e-01 -2.54058894e-02 -6.88285708e-01
6.74165964e-01 2.44725943e-01 4.08499897e-01 -1.14943910e+00
-8.50612298e-02 2.59468615e-01 2.78193146e-01 -1.24388404e-01
9.38246250e-01 -1.40703782e-01 2.83371776e-01 6.14283144e-01
-2.01968163e-01 -9.91882682e-02 5.61068535e-01 5.67747176e-01
1.22271121e+00 -4.29755360e-01 2.70215005e-01 -3.91232558e-02
1.41976215e-03 3.84114645e-02 2.38705963e-01 1.31466675e+00
-8.16255584e-02 6.81627393e-01 3.72000694e-01 1.70406550e-01
-6.72890604e-01 -6.59009159e-01 -1.95513234e-01 1.45409381e+00
1.11472622e-01 -4.92363065e-01 -1.45963252e-01 -5.32530308e-01
3.01176906e-01 6.11015081e-01 -7.65880704e-01 -5.33573687e-01
-2.49536261e-01 -5.46088040e-01 8.52955729e-02 2.50970334e-01
-2.46921882e-01 -8.35564196e-01 -3.91519815e-01 2.41484255e-01
-2.42694125e-01 -5.20753920e-01 -7.24525154e-01 2.17603296e-01
-8.87151122e-01 -1.21951032e+00 -7.03724027e-01 -9.73407328e-02
4.94466215e-01 8.94494534e-01 1.17062151e+00 2.47619405e-01
1.73275501e-01 6.15148187e-01 -7.34239697e-01 -2.87961364e-01
-1.69985771e-01 -4.71111238e-02 -2.81277150e-01 9.81021225e-02
3.89530927e-01 -1.91477656e-01 -7.52510369e-01 3.11888456e-01
-1.06644559e+00 -6.92268074e-01 7.59910166e-01 8.99817169e-01
3.73250782e-01 2.37546042e-01 4.70456243e-01 -9.93391037e-01
1.34534538e+00 -3.93643171e-01 -3.64677012e-01 9.39759314e-02
-1.02394497e+00 1.98681355e-01 1.68656081e-01 -5.33794940e-01
-4.80026066e-01 -6.15917861e-01 1.69338495e-01 -3.65139276e-01
5.69778383e-01 8.01728606e-01 3.74841690e-01 -1.27505884e-02
1.05457413e+00 1.39124572e-01 -2.62790918e-01 -3.67208987e-01
2.41476804e-01 5.72779655e-01 -1.33999780e-01 -1.40173197e-01
6.32168114e-01 3.06865275e-01 -2.69767433e-01 -6.60027802e-01
-9.23939228e-01 -9.24219966e-01 -9.21527371e-02 -3.45789701e-01
2.17512980e-01 -7.24739969e-01 -3.53642046e-01 -1.35446161e-01
-7.03083992e-01 -8.11894163e-02 -4.42943960e-01 6.12861574e-01
-5.81712611e-02 2.31073871e-01 -5.12679756e-01 -9.79590356e-01
-2.20534563e-01 -1.10651672e+00 9.51815963e-01 -1.82357077e-02
-5.01302421e-01 -6.73223019e-01 1.40115559e-01 2.89309502e-01
7.92477608e-01 -5.11785746e-01 9.72791076e-01 -8.74472678e-01
-6.23576522e-01 -5.33090830e-01 -1.74265891e-01 1.00855634e-01
1.51008233e-01 -2.15787724e-01 -1.01536167e+00 -3.52553159e-01
5.30430861e-02 -2.18426481e-01 1.30256712e+00 4.99370635e-01
4.56793398e-01 -5.32657981e-01 -3.50336403e-01 -3.01010430e-01
1.25520039e+00 1.50005743e-01 4.30624813e-01 2.64416873e-01
1.16175391e-01 6.27985835e-01 8.97892535e-01 4.29401368e-01
1.79256126e-02 8.66444051e-01 2.78320074e-01 2.96716422e-01
-9.94961150e-03 -3.21531504e-01 4.27830547e-01 7.40529001e-01
4.43452597e-01 -2.58919716e-01 -5.67893744e-01 5.26631236e-01
-1.58825016e+00 -9.88368511e-01 8.30572918e-02 2.54963589e+00
9.90326583e-01 3.32957000e-01 1.02241732e-01 6.24190211e-01
3.35997939e-01 2.71141380e-01 -2.67456114e-01 -2.59199888e-01
5.82751865e-03 1.62239388e-01 3.15378994e-01 6.17012799e-01
-8.57495964e-01 4.47970152e-01 6.98599339e+00 8.11361015e-01
-1.16947591e+00 -1.64234415e-01 3.74657869e-01 -2.26203427e-01
-6.53525352e-01 1.89145431e-01 -7.00009823e-01 3.30962390e-01
9.52135563e-01 -4.00252432e-01 5.79687119e-01 6.49806321e-01
2.40300536e-01 -5.92771471e-01 -1.09457612e+00 5.90179384e-01
5.61288111e-02 -8.30293715e-01 2.64249202e-02 1.70343176e-01
5.27323306e-01 -2.96480022e-02 -9.53884423e-02 5.36612868e-01
7.05868006e-02 -8.52838337e-01 5.91488838e-01 5.52627265e-01
1.96432173e-01 -5.46194375e-01 7.82130122e-01 2.16780931e-01
-8.78920853e-01 -1.73952386e-01 -3.38805526e-01 -9.39471498e-02
-1.49264514e-01 8.91984344e-01 -9.56014097e-01 2.00395614e-01
5.01011550e-01 3.07228744e-01 -7.25774884e-01 1.09774184e+00
-1.64135233e-01 9.37874198e-01 -3.65030378e-01 -4.92884219e-01
2.00575754e-01 1.55901879e-01 6.19546652e-01 1.03313959e+00
3.71264406e-02 5.97362332e-02 1.07731543e-01 3.48191053e-01
-5.48059680e-03 3.51149857e-01 -5.21526933e-01 -5.38917363e-01
6.51814818e-01 9.46906626e-01 -5.61022639e-01 -2.00451195e-01
-3.17662448e-01 4.66411829e-01 1.34412140e-01 4.60617185e-01
-1.85596775e-02 -1.54714897e-01 3.95049989e-01 5.29125333e-01
2.77324885e-01 1.21577993e-01 5.69083691e-02 -8.56915414e-01
2.06943288e-01 -1.11713994e+00 3.61061484e-01 -6.77899063e-01
-1.18490136e+00 1.74421415e-01 5.28885834e-02 -1.14375329e+00
-3.66468579e-01 -1.77622661e-01 -1.61017627e-01 7.78020799e-01
-1.56315470e+00 -3.01112324e-01 1.04963537e-02 2.56993115e-01
3.55852157e-01 1.66959152e-01 6.06134057e-01 1.89288497e-01
1.94298685e-01 4.33602333e-01 2.54820257e-01 -4.92925048e-02
8.95065725e-01 -1.05019808e+00 -1.21712089e-01 5.33467591e-01
4.46432233e-01 8.76174510e-01 7.95366228e-01 -6.38672650e-01
-1.67006767e+00 -5.93137026e-01 9.45753694e-01 -4.30961341e-01
5.41402280e-01 8.53082761e-02 -8.86540532e-01 1.52319714e-01
-9.25281867e-02 -3.11813742e-01 5.77404141e-01 4.06551123e-01
-3.69394481e-01 -2.63180822e-01 -7.68897235e-01 5.40957689e-01
5.58943152e-01 -6.65135086e-01 -6.31301999e-01 1.95841670e-01
5.63687861e-01 -5.41028753e-02 -3.72280449e-01 5.05562961e-01
7.28928387e-01 -1.06195712e+00 8.06185484e-01 -3.99306923e-01
2.45575488e-01 -1.33150131e-01 -5.18142104e-01 -1.37032223e+00
-5.30458152e-01 -4.83092368e-02 -1.93667799e-01 8.14025462e-01
4.79778916e-01 -3.81264687e-01 5.59185565e-01 5.43161988e-01
3.58836412e-01 -8.04898024e-01 -5.92081487e-01 -5.69003999e-01
-1.45484582e-01 -4.41545516e-01 2.26798624e-01 5.68673015e-01
-6.15520701e-02 7.26354778e-01 -1.58106253e-01 -2.82189965e-01
1.05742291e-01 1.23947218e-01 6.32450223e-01 -1.36981916e+00
-5.70057809e-01 -5.33472598e-01 -2.22466394e-01 -1.09974194e+00
-4.39672172e-01 -6.50687456e-01 6.19461276e-02 -1.49837947e+00
1.86653182e-01 -2.99484134e-01 -5.63276410e-01 1.77852884e-01
-5.10823429e-01 2.34667689e-01 3.04679871e-01 5.96622169e-01
-4.18400466e-01 4.86804754e-01 1.31158316e+00 -2.61649191e-01
-3.67374033e-01 2.84084260e-01 -1.12190211e+00 1.21104188e-01
4.35050458e-01 -6.49151385e-01 -5.81054688e-01 -9.27372202e-02
6.38221145e-01 1.15649417e-01 1.00432761e-01 -6.15140855e-01
2.78215557e-01 -1.85062662e-01 4.01653528e-01 -4.18757260e-01
4.42267776e-01 -7.49847651e-01 -2.03015640e-01 2.78609335e-01
-7.35621870e-01 1.70091957e-01 1.86493956e-02 7.49818861e-01
-4.37295020e-01 -5.11784732e-01 4.16973382e-01 -1.86331883e-01
-4.10515666e-01 -5.18142208e-02 -4.38290030e-01 2.53730625e-01
2.92889029e-01 -1.37247607e-01 -1.43699683e-02 -9.82483387e-01
-3.47689390e-01 1.89787164e-01 2.53316909e-01 5.70679724e-01
6.52499497e-01 -1.24526811e+00 -7.31209576e-01 1.16460666e-01
3.89446080e-01 -5.57120204e-01 -1.00944504e-01 6.88941419e-01
1.68464571e-01 7.73964703e-01 4.79941666e-01 -3.03741544e-01
-1.08803833e+00 3.60298038e-01 8.73886868e-02 -5.15562534e-01
-1.36832625e-01 7.47208476e-01 1.94493815e-01 2.00426131e-01
3.73254329e-01 -3.17473896e-02 -3.95212054e-01 5.69470823e-01
7.53592491e-01 1.31028980e-01 3.54642838e-01 -4.31335837e-01
-2.81485617e-01 1.73714697e-01 -3.11854482e-01 -7.32873440e-01
1.01637042e+00 -1.06424373e-03 -3.35854553e-02 8.37889075e-01
1.09540856e+00 3.95353675e-01 -5.66597760e-01 -3.62621397e-01
6.95440471e-02 -8.67620826e-01 3.43781054e-01 -8.80418897e-01
-8.24366331e-01 7.43600368e-01 4.96905237e-01 3.44178438e-01
1.12923646e+00 -1.24173835e-01 1.92240998e-01 3.36856991e-01
6.18765712e-01 -1.02673495e+00 2.63152778e-01 3.93965989e-01
1.03679562e+00 -1.02408957e+00 3.51464659e-01 6.20505810e-02
-4.19932306e-01 7.79458821e-01 2.66952395e-01 -7.95270726e-02
7.61543751e-01 -2.27562681e-01 9.46908593e-02 -1.11210361e-01
-1.13500941e+00 -3.40423405e-01 7.12276757e-01 1.10956557e-01
7.30028152e-01 -1.43308610e-01 -7.63200819e-01 3.14637929e-01
-2.03928247e-01 7.12268287e-03 2.46556237e-01 1.03438818e+00
-6.92380488e-01 -1.13427234e+00 -5.30098319e-01 9.11926389e-01
-6.89372122e-01 -2.72919267e-01 -8.86317611e-01 5.43919981e-01
-5.05866945e-01 1.34541154e+00 -3.24297160e-01 -5.86119711e-01
4.57980752e-01 5.49905039e-02 3.52832228e-01 -7.51010835e-01
-8.78676176e-01 3.80184352e-01 6.19692877e-02 -5.03588736e-01
-2.26016834e-01 -5.36130428e-01 -6.55241251e-01 -1.20774105e-01
-8.89437079e-01 7.39004433e-01 5.45113266e-01 7.92836964e-01
3.19064528e-01 2.20424488e-01 8.11168551e-01 -7.14981496e-01
-9.33799624e-01 -1.21757936e+00 -5.86374640e-01 4.19941694e-01
4.16669935e-01 -7.13657320e-01 -6.95349514e-01 -5.69448829e-01] | [11.453526496887207, 7.578171253204346] |
0c42f8d6-2897-4a36-885e-4a769e0f8606 | using-openwordnet-pt-for-question-answering | null | null | https://aclanthology.org/2018.gwc-1.13 | https://aclanthology.org/2018.gwc-1.13.pdf | Using OpenWordnet-PT for Question Answering on Legal Domain | In order to practice a legal profession in Brazil, law graduates must be approved in the OAB national unified bar exam. For their topic coverage and national reach, the OAB exams provide an excellent benchmark for the performance of legal information systems, as it provides objective metrics and are challenging even for humans, as only 20% of its candidates are approved. After constructing a new data set on the exams and doing shallow experiments on it, we now employ the OpenWordnet-PT to verify whether using word senses and relations we can improve previous results. We discuss the results, possible future ideas and the additions to the OpenWordnet-PT that we made. | ['Alexandre Rademaker', 'Gerson Zaverucha', 'Guilherme Paulino-Passos', 'Bruno Cuconato', 'Pedro Delfino'] | null | null | null | null | gwc-2018-1 | ['topic-coverage'] | ['natural-language-processing'] | [-3.04943383e-01 6.11409783e-01 -7.75911510e-01 -3.37488085e-01
-1.00404286e+00 -7.51022518e-01 4.62829739e-01 6.20821834e-01
-6.83757424e-01 1.01210558e+00 2.12170213e-01 -1.16589272e+00
-9.22293305e-01 -9.15142357e-01 -4.20058131e-01 1.88380212e-01
4.26559329e-01 5.43268502e-01 4.33741033e-01 -8.79726171e-01
4.67999369e-01 4.91969347e-01 -1.37907171e+00 2.16007698e-02
1.49057889e+00 2.82360375e-01 -3.97613257e-01 1.67590246e-01
-2.05595344e-01 7.72287905e-01 -9.01447415e-01 -1.30341887e+00
2.89844960e-01 5.72430305e-02 -1.17388809e+00 -8.36639702e-01
1.26751816e+00 -2.29654074e-01 -7.25189686e-01 1.05143821e+00
3.48015100e-01 2.89369702e-01 5.23137987e-01 -7.40756452e-01
-1.01161516e+00 1.01252151e+00 -2.44567171e-01 5.39617240e-01
9.37559009e-01 -1.90728396e-01 1.66536093e+00 -3.15716118e-01
1.17551184e+00 1.09548402e+00 5.26855767e-01 5.14627278e-01
-7.82263041e-01 -6.57961905e-01 1.87363237e-01 6.87803149e-01
-1.05788600e+00 -2.91516691e-01 2.40605846e-01 -5.01869559e-01
1.00833416e+00 6.67111278e-02 7.35090733e-01 9.95067000e-01
3.12609732e-01 3.91048491e-01 1.03363335e+00 -5.24189949e-01
-1.00979872e-01 1.02410121e-02 9.47638392e-01 4.70922709e-01
1.12462008e+00 -1.57577723e-01 -3.05340022e-01 -2.29772195e-01
3.47843379e-01 -4.86837387e-01 2.18223827e-03 1.26224101e-01
-5.92774034e-01 7.54580379e-01 1.46684125e-01 7.51103163e-01
-1.22953020e-01 1.66317020e-02 -4.64335494e-02 5.88053405e-01
-3.38240683e-01 1.05013907e+00 -3.20890456e-01 -5.79218268e-01
-8.11366737e-01 7.36198485e-01 8.51761878e-01 7.64813244e-01
3.46271455e-01 -6.00630641e-01 -4.40701455e-01 8.15621316e-01
3.38903427e-01 2.71461964e-01 2.68870443e-01 -1.20629251e+00
7.21097112e-01 1.10025072e+00 -9.19439942e-02 -9.73914146e-01
2.69117355e-02 -2.49004960e-01 1.08450048e-01 -2.49657482e-02
7.80915499e-01 -3.54544520e-02 -9.55474198e-01 1.32748568e+00
-4.89811972e-02 7.03634545e-02 -1.13796294e-01 3.07463348e-01
1.09988654e+00 6.28131628e-02 3.43071848e-01 1.31626293e-01
1.55307186e+00 -2.98408777e-01 -1.00803459e+00 -2.62359440e-01
8.56839418e-01 -9.30575788e-01 1.03545344e+00 4.32914495e-01
-1.21425080e+00 -3.94103259e-01 -1.08711541e+00 -6.05941236e-01
-6.64003730e-01 -3.70352179e-01 8.19359243e-01 1.24985015e+00
-7.81706214e-01 6.60574734e-01 -3.36652309e-01 -3.88352036e-01
6.91425562e-01 -7.71865249e-02 -3.03997636e-01 -4.63553995e-01
-1.63774621e+00 1.42516434e+00 5.08340076e-02 -3.66347641e-01
-3.31126302e-01 -7.60013223e-01 -1.15036273e+00 2.65783250e-01
5.46943247e-01 -6.45528138e-02 1.18211377e+00 6.40098929e-01
-7.78546035e-01 1.14677095e+00 4.66208048e-02 -2.45243326e-01
3.23973000e-01 9.01454967e-03 -6.47284746e-01 1.66632340e-03
5.40923417e-01 2.34796658e-01 -4.22377020e-01 -5.71403801e-01
-7.37939119e-01 -2.49012023e-01 6.80173695e-01 -2.73269206e-01
-5.30332565e-01 2.69053727e-01 -4.45101261e-01 -2.10835621e-01
1.79713473e-01 -6.33850515e-01 -4.47773844e-01 -4.09445673e-01
-1.36719435e-01 -7.57530749e-01 3.02119166e-01 -5.53510785e-01
1.70157123e+00 -1.73819923e+00 -5.42687774e-01 7.72423804e-01
2.55456656e-01 5.45964956e-01 -7.97658935e-02 6.25034332e-01
-1.57157868e-01 6.73222184e-01 1.76569238e-01 2.28071615e-01
1.84652478e-01 4.03935492e-01 -4.71051484e-01 2.87488639e-01
-3.72592539e-01 8.42517853e-01 -9.99022543e-01 -8.13612401e-01
-2.91971629e-03 7.83158392e-02 -6.29428983e-01 -4.78696704e-01
1.24415375e-01 6.95254002e-03 -6.51637793e-01 8.68348777e-01
3.73199463e-01 1.86043661e-02 3.68647933e-01 3.30847114e-01
-5.48708923e-02 9.37881112e-01 -9.50292528e-01 1.72688687e+00
-3.45176160e-01 8.80671620e-01 -3.61375809e-01 -8.79271686e-01
7.25881279e-01 6.09473765e-01 5.58907390e-01 -9.94162917e-01
1.90872326e-01 4.88443911e-01 5.35314798e-01 -6.25254214e-01
7.36698508e-01 -2.16377694e-02 -2.06806868e-01 4.07985032e-01
6.75612614e-02 -3.19288015e-01 1.17773402e+00 7.19055414e-01
1.46918666e+00 -1.12144217e-01 4.25142288e-01 -4.73489702e-01
4.90337580e-01 1.68862864e-02 6.37293458e-01 1.10085595e+00
-3.07320654e-01 1.50019839e-01 8.19359362e-01 -4.51791763e-01
-5.48166215e-01 -8.81417036e-01 -8.83557916e-01 7.53026128e-01
1.28896669e-01 -8.20014417e-01 -3.08704048e-01 -1.02680421e+00
2.95859396e-01 8.29709530e-01 -5.20508051e-01 7.83165544e-02
-5.30119956e-01 -4.25805785e-02 7.93980002e-01 2.34634385e-01
8.84375796e-02 -9.99450445e-01 -2.57689744e-01 1.88671872e-01
-1.21345364e-01 -1.44548237e+00 -7.98350573e-02 -4.84815866e-01
-3.74988586e-01 -1.63444877e+00 -1.88776270e-01 -7.14839101e-01
3.38321179e-01 -2.28482202e-01 1.09157622e+00 6.77983999e-01
-2.66626030e-01 3.85374963e-01 -3.39219391e-01 -6.05069816e-01
-4.39103484e-01 3.84497613e-01 -1.62093550e-01 -8.53633225e-01
9.50438797e-01 -6.52328134e-01 -2.93278456e-01 1.18185021e-01
-8.27387214e-01 -7.52916217e-01 3.20061743e-01 2.98719555e-01
-1.28874406e-01 -4.14692879e-01 7.90128469e-01 -1.55158985e+00
1.39113104e+00 -2.76143968e-01 -7.95079350e-01 7.50038922e-01
-9.00317609e-01 -3.22627217e-01 8.30178261e-02 1.44981012e-01
-7.89235413e-01 -7.46944249e-01 -5.29445350e-01 5.12721479e-01
9.37672041e-04 9.51087236e-01 6.43966049e-02 -2.56620347e-01
1.20846307e+00 -6.84030294e-01 -6.47134066e-01 -3.70353878e-01
3.67289245e-01 7.96142697e-01 5.26405036e-01 -1.18397772e+00
1.00398612e+00 -2.70533133e-02 -6.16827719e-02 -6.08388722e-01
-1.44538629e+00 -5.52259564e-01 -5.61705530e-01 -3.63044977e-01
8.79186988e-01 -5.87801099e-01 -1.26812136e+00 -5.33378541e-01
-1.21172404e+00 2.98625410e-01 -3.61352652e-01 6.51918471e-01
4.77339281e-03 6.14173710e-01 -7.08574057e-01 -7.26260900e-01
2.95226723e-01 -1.07587528e+00 5.08003592e-01 4.33892250e-01
-5.17538786e-01 -1.10524309e+00 2.47926131e-01 1.19146860e+00
1.06280312e-01 1.86100051e-01 1.21016121e+00 -9.95086432e-01
-1.91074371e-01 -6.36674881e-01 7.86313135e-03 4.12449360e-01
4.41574939e-02 1.59032077e-01 -7.08121359e-01 -3.01864278e-02
-4.49951291e-01 -3.04206520e-01 8.58193398e-01 2.05846995e-01
8.99484038e-01 4.47893217e-02 -3.61101866e-01 5.20851128e-02
1.23327255e+00 2.68113077e-01 7.74771571e-01 6.83813572e-01
2.72499263e-01 4.82440740e-01 5.90362072e-01 -7.02693453e-03
5.96643031e-01 3.59814584e-01 -1.77215502e-01 4.20892298e-01
-2.04558313e-01 -3.36879134e-01 5.09900190e-02 7.71273255e-01
-5.36998749e-01 -7.37340096e-03 -1.32082403e+00 9.23610747e-01
-1.65257323e+00 -1.07282615e+00 -2.62741268e-01 2.03835773e+00
1.05091083e+00 5.98071635e-01 -1.87072888e-01 2.70581275e-01
4.11906391e-01 -1.52427042e-02 -3.20550203e-02 -7.18291402e-01
-2.35830635e-01 1.06638992e+00 5.76992333e-01 7.68667936e-01
-5.89083314e-01 1.17578447e+00 7.76355267e+00 7.88935781e-01
-4.34089720e-01 -9.33873877e-02 2.27228194e-01 1.73090622e-01
-6.31955743e-01 4.03237134e-01 -1.10701025e+00 1.77671999e-01
8.97824109e-01 -6.36362553e-01 4.78398129e-02 8.13456178e-01
-1.98327839e-01 -4.53741290e-03 -1.07012105e+00 5.35567343e-01
-1.60844132e-01 -1.65327370e+00 1.03116527e-01 4.79735374e-01
1.02161407e+00 -4.23829537e-03 -1.01105072e-01 7.69545615e-01
8.70601952e-01 -1.34169269e+00 2.88827419e-01 2.85450637e-01
4.95872229e-01 -3.09916914e-01 8.49772811e-01 1.97318062e-01
-5.50656557e-01 -2.97160178e-01 -5.64878285e-01 -5.10727108e-01
2.49832526e-01 4.42187965e-01 -6.16918385e-01 7.93029547e-01
4.44203138e-01 6.97801828e-01 -6.62862360e-01 1.34038794e+00
-8.59803915e-01 7.30851352e-01 6.84903050e-03 -8.87481403e-03
3.08274120e-01 -2.24741936e-01 3.79650325e-01 7.85248160e-01
3.70144725e-01 1.69413567e-01 2.57840604e-01 4.99936938e-01
-3.01295012e-01 2.84960419e-01 -6.56623125e-01 1.09916873e-01
8.89925122e-01 1.01246011e+00 -3.12155843e-01 -4.57661718e-01
-5.83194554e-01 -1.07778989e-01 2.95768440e-01 3.28169644e-01
-5.62492549e-01 -8.61200869e-01 7.84810305e-01 3.59237581e-01
-1.01520419e-01 -1.64534017e-01 -4.16362226e-01 -1.12976587e+00
1.54969813e-02 -9.98315215e-01 8.75490129e-01 -5.28113961e-01
-1.36902404e+00 4.70793307e-01 9.48834717e-02 -8.90335023e-01
-7.38652647e-02 -1.01236212e+00 -4.21076298e-01 8.48394632e-01
-1.60360277e+00 -6.95322096e-01 -8.51893052e-02 3.14015031e-01
-3.70609872e-02 -3.31575394e-01 8.84486318e-01 9.09084320e-01
-3.86559874e-01 9.04519796e-01 -4.66027826e-01 3.49601358e-01
9.71366942e-01 -1.29718041e+00 2.97588438e-01 9.22035038e-01
4.54903513e-01 1.40473366e+00 6.23901725e-01 -9.33790088e-01
-7.17171252e-01 -2.23083004e-01 1.39488113e+00 -8.23490262e-01
9.79842365e-01 1.32817701e-01 -7.85372555e-01 1.03999841e+00
4.45405364e-01 -2.89294481e-01 1.04450440e+00 6.39930308e-01
-5.13830304e-01 -5.82482517e-02 -1.07301438e+00 6.69575870e-01
1.23616040e+00 -4.09960985e-01 -1.39662087e+00 5.86267471e-01
3.96136016e-01 -6.66455984e-01 -1.38758481e+00 2.39061311e-01
6.78148031e-01 -7.12981939e-01 8.54139268e-01 -1.30123520e+00
4.78895336e-01 -1.81505643e-02 1.63703710e-01 -7.75964916e-01
-1.33068621e-01 -7.34062731e-01 3.82274091e-01 9.86811340e-01
1.04318798e+00 -8.27185392e-01 8.14505458e-01 1.20456636e+00
-2.07683101e-01 -8.09124053e-01 -1.31695962e+00 -1.08713746e+00
4.90121245e-01 -7.05207527e-01 7.19365597e-01 1.31972659e+00
6.77248597e-01 2.26940721e-01 -4.75618429e-02 -1.24999717e-01
3.54397207e-01 8.16049501e-02 5.68209708e-01 -1.69309723e+00
1.06214523e-01 -6.58212364e-01 -6.51799083e-01 -8.83243680e-01
3.07596087e-01 -1.05162418e+00 -5.28006792e-01 -2.13205481e+00
-1.03326015e-01 -3.88416052e-01 -3.72727722e-01 1.01430905e+00
-1.88868910e-01 2.56012291e-01 8.82607996e-02 -1.50944918e-01
-5.57391822e-01 -1.34700119e-01 1.46564376e+00 -1.11857504e-01
1.41696930e-01 -8.32289830e-02 -1.34517598e+00 6.57054543e-01
4.00374740e-01 -6.14058018e-01 -2.44887322e-01 -4.13292587e-01
7.13118613e-01 -1.49548262e-01 -3.70564133e-01 -8.49752307e-01
5.64782619e-01 -4.42983866e-01 -1.48200780e-01 -2.42953151e-01
2.56885439e-01 -5.63142836e-01 -7.63172925e-01 2.47243389e-01
-4.66234565e-01 -7.22900778e-02 1.53506070e-01 -1.33350089e-01
-5.23316562e-01 -7.57029891e-01 2.79676467e-01 -6.18745387e-02
-2.17206344e-01 6.75422773e-02 -5.53835809e-01 3.89380842e-01
9.44595575e-01 -3.90051097e-01 -8.42135310e-01 -3.83221120e-01
-6.61108375e-01 5.80633640e-01 2.35596105e-01 3.80652398e-01
4.09417719e-01 -1.12336493e+00 -8.61817122e-01 -3.31449389e-01
2.71220237e-01 -2.64016867e-01 6.07203431e-02 5.08478582e-01
-7.15832174e-01 8.48483622e-01 -2.24638686e-01 2.33427018e-01
-1.17844582e+00 -9.77091864e-02 2.55291630e-02 -6.35269344e-01
-4.91759360e-01 5.65572560e-01 -8.12175155e-01 -6.65419042e-01
1.89203203e-01 -3.86109561e-01 -9.31527734e-01 4.75860499e-02
4.12911594e-01 2.22985491e-01 1.59945324e-01 -3.53881270e-01
-2.96884328e-01 5.61712205e-01 -6.57475367e-02 -1.31064579e-01
1.60094643e+00 6.01628006e-01 -2.83179849e-01 3.27279046e-02
4.60900605e-01 9.37462509e-01 2.12866440e-01 1.08350091e-01
6.31605566e-01 -8.52114856e-01 -2.98067927e-01 -1.00899911e+00
-8.37728262e-01 4.78654206e-01 -7.65147656e-02 5.47427297e-01
4.54598129e-01 -3.26738060e-02 7.24400163e-01 9.70376253e-01
1.55765489e-01 -1.23245764e+00 -4.84876841e-01 7.53172219e-01
4.83341932e-01 -8.77324641e-01 3.60969812e-01 -7.24628806e-01
-2.38062933e-01 9.11988974e-01 7.77159154e-01 -2.69068349e-02
7.83282816e-01 -8.93577263e-02 5.76730549e-01 -5.70348084e-01
-5.89241862e-01 -1.08134024e-01 6.31962478e-01 2.90056109e-01
8.45269084e-01 -2.67073691e-01 -1.36679459e+00 6.22719884e-01
-9.09410894e-01 -3.21957655e-02 1.00339782e+00 7.28777409e-01
-4.02304769e-01 -1.93289208e+00 -5.12360521e-02 6.68569088e-01
-9.02521074e-01 -1.95906535e-01 -5.59994578e-01 1.03372073e+00
1.30538940e-01 1.05711246e+00 -3.73827904e-01 -1.04155198e-01
8.46246779e-01 1.55012235e-01 4.86867785e-01 -1.09330261e+00
-7.90025592e-01 -6.71128809e-01 4.35842156e-01 -4.27086502e-01
-2.73285389e-01 -6.58494711e-01 -1.12083900e+00 -2.89834380e-01
-3.24916691e-01 7.89953828e-01 6.84736252e-01 1.20313776e+00
5.48631772e-02 5.71334362e-01 -1.23201497e-01 5.81539631e-01
-5.77455699e-01 -1.24529791e+00 -4.45810944e-01 5.99027991e-01
-3.40562612e-01 -5.46118081e-01 -5.49007431e-02 -4.66181606e-01] | [9.954362869262695, 9.210872650146484] |
b74799fe-7536-48af-802d-963816114de3 | a-unified-and-efficient-coordinating | 2303.05710 | null | https://arxiv.org/abs/2303.05710v1 | https://arxiv.org/pdf/2303.05710v1.pdf | A Unified and Efficient Coordinating Framework for Autonomous DBMS Tuning | Recently using machine learning (ML) based techniques to optimize modern database management systems has attracted intensive interest from both industry and academia. With an objective to tune a specific component of a DBMS (e.g., index selection, knobs tuning), the ML-based tuning agents have shown to be able to find better configurations than experienced database administrators. However, one critical yet challenging question remains unexplored -- how to make those ML-based tuning agents work collaboratively. Existing methods do not consider the dependencies among the multiple agents, and the model used by each agent only studies the effect of changing the configurations in a single component. To tune different components for DBMS, a coordinating mechanism is needed to make the multiple agents cognizant of each other. Also, we need to decide how to allocate the limited tuning budget among the agents to maximize the performance. Such a decision is difficult to make since the distribution of the reward for each agent is unknown and non-stationary. In this paper, we study the above question and present a unified coordinating framework to efficiently utilize existing ML-based agents. First, we propose a message propagation protocol that specifies the collaboration behaviors for agents and encapsulates the global tuning messages in each agent's model. Second, we combine Thompson Sampling, a well-studied reinforcement learning algorithm with a memory buffer so that our framework can allocate budget judiciously in a non-stationary environment. Our framework defines the interfaces adapted to a broad class of ML-based tuning agents, yet simple enough for integration with existing implementations and future extensions. We show that it can effectively utilize different ML-based agents and find better configurations with 1.4~14.1X speedups on the workload execution time compared with baselines. | ['Bin Cui', 'Feifei Li', 'Jian Tan', 'Jia Chen', 'Yang Li', 'Hong Wu', 'Zhuo Chang', 'Xinyi Zhang'] | 2023-03-10 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [-4.57467198e-01 -3.53211612e-01 -3.72881144e-01 -3.34197909e-01
-6.32157624e-01 -4.98885751e-01 6.23948202e-02 2.48814359e-01
-5.75773835e-01 8.94868255e-01 -4.14780229e-01 -3.97229195e-01
-2.65242636e-01 -1.06377721e+00 -7.19436288e-01 -8.78314495e-01
-3.14609289e-01 1.36478996e+00 6.80481970e-01 -4.86254036e-01
3.48801106e-01 5.82576215e-01 -1.55949664e+00 1.32136941e-01
6.86018884e-01 7.18299747e-01 5.27556479e-01 1.08822370e+00
-1.22118078e-01 6.46883607e-01 -1.14539313e+00 2.20102087e-01
2.32183099e-01 -4.08921808e-01 -6.61285162e-01 2.72818673e-02
-1.93202421e-01 -5.44057071e-01 1.65773228e-01 6.37537420e-01
6.66072667e-01 -2.14443188e-02 3.17573994e-01 -1.30326474e+00
8.95288065e-02 1.08978355e+00 -7.83950567e-01 4.01647329e-01
-5.27697243e-02 2.48947486e-01 9.09436643e-01 1.19987443e-01
4.32646394e-01 1.06834054e+00 -6.05150610e-02 2.34738857e-01
-1.33445621e+00 -5.16344190e-01 3.43531758e-01 4.95011061e-01
-1.31732476e+00 -4.53046858e-01 4.01118368e-01 -9.05650109e-02
1.04172397e+00 4.20982689e-01 4.04959440e-01 3.22370410e-01
3.87013435e-01 6.30427897e-01 9.15802419e-01 -7.80893624e-01
5.50957680e-01 3.16862881e-01 2.34726593e-02 3.58409941e-01
2.83237278e-01 -1.01012588e-02 -4.86904532e-01 -7.21633613e-01
7.06859231e-01 -2.31723040e-01 2.51095071e-02 -6.25046194e-01
-1.26969802e+00 9.65071738e-01 7.02184439e-02 1.29899472e-01
-5.65539598e-01 4.37549353e-01 2.40908965e-01 5.41618049e-01
-9.71940234e-02 6.24343455e-01 -6.90151691e-01 -3.29649597e-01
-7.12629080e-01 3.30211729e-01 1.32407379e+00 7.54943669e-01
9.31798458e-01 1.01052737e-02 -2.71264315e-01 7.86803305e-01
1.93779960e-01 5.93483269e-01 1.14603169e-01 -1.43872809e+00
3.14480782e-01 4.28830206e-01 4.30427611e-01 -6.25923574e-01
-4.75840092e-01 -2.34810099e-01 -4.98688102e-01 5.63347578e-01
3.27480614e-01 -4.20602858e-01 -2.62815207e-01 1.89981937e+00
7.36289978e-01 -9.04489830e-02 -2.56021887e-01 7.98873961e-01
-3.00457180e-01 5.77312410e-01 -2.13525683e-01 -6.20643854e-01
1.18758571e+00 -8.03697288e-01 -4.38751131e-01 1.54220134e-01
5.22807896e-01 -9.17428851e-01 9.76669967e-01 3.04947555e-01
-1.28987527e+00 -2.28526339e-01 -1.25624788e+00 9.08918619e-01
8.25900771e-03 -1.49663433e-01 5.50793231e-01 7.33323574e-01
-1.09335589e+00 5.20456791e-01 -1.20035613e+00 -3.04446936e-01
-2.02327132e-01 8.27676117e-01 4.01778698e-01 2.43908152e-01
-8.87321830e-01 8.87637913e-01 2.53688037e-01 -3.54132205e-01
-6.83490634e-01 -3.68853509e-01 1.36029115e-02 8.94249305e-02
1.04984796e+00 -1.05447638e+00 1.65082014e+00 -5.75611711e-01
-1.86825359e+00 2.62418896e-01 2.58635934e-02 -3.75339836e-01
5.70581794e-01 2.50936836e-01 -7.33635714e-03 -1.58440724e-01
5.24063669e-02 1.28768772e-01 6.63966358e-01 -1.26724529e+00
-8.42322171e-01 -2.78728813e-01 3.48947555e-01 4.69825804e-01
-3.88668507e-01 -6.37297034e-02 -6.88078046e-01 -3.15093786e-01
-3.16920280e-01 -1.23910010e+00 -4.02536869e-01 -7.30616212e-01
-4.16621864e-01 -2.53975660e-01 3.86767715e-01 1.67715207e-01
1.39673793e+00 -1.65768480e+00 6.78701252e-02 4.42227989e-01
5.83652891e-02 1.80948793e-03 -4.92696129e-02 8.22101772e-01
5.12761176e-01 6.03806563e-02 2.48185158e-01 5.40807359e-02
2.02290744e-01 4.92728770e-01 4.91606481e-02 2.04863504e-01
-3.93994123e-01 4.11123335e-01 -5.38805604e-01 -6.79589987e-01
1.14984378e-01 -2.85583854e-01 -9.50515807e-01 6.01146340e-01
-6.86613321e-01 1.47002444e-01 -7.82074034e-01 5.50929844e-01
4.98760015e-01 -6.39853477e-01 8.58713806e-01 4.25503291e-02
-8.09411481e-02 5.03583066e-02 -1.73705101e+00 9.66815889e-01
-4.21596676e-01 -5.93355764e-03 2.68099993e-01 -1.02645016e+00
6.12107277e-01 -1.06189251e-01 7.90973783e-01 -6.20809019e-01
-6.32585734e-02 2.02090055e-01 2.49677122e-01 -5.04088640e-01
4.48741198e-01 2.11697623e-01 -2.34495595e-01 1.36275506e+00
-5.26734412e-01 1.03148073e-01 5.27120173e-01 1.21405564e-01
1.54420674e+00 -2.32100278e-01 4.12441522e-01 3.23564298e-02
3.65878969e-01 3.34283352e-01 7.16169059e-01 1.36830270e+00
-1.63685292e-01 -3.41136381e-02 5.88168561e-01 -3.95353407e-01
-9.13009763e-01 -6.53082311e-01 1.96046740e-01 1.78423071e+00
3.60500216e-01 -2.01128915e-01 -7.51097083e-01 -4.64166433e-01
4.07375813e-01 6.29626572e-01 -3.84816557e-01 8.06493536e-02
-8.98379862e-01 -1.04195297e+00 7.45977759e-02 1.12767704e-01
4.42547172e-01 -1.24584174e+00 -1.09453297e+00 7.01874316e-01
5.96674997e-03 -8.17407608e-01 -6.65429354e-01 3.00269574e-01
-6.64639831e-01 -1.18105650e+00 -4.93262932e-02 -2.40268826e-01
3.52896065e-01 2.59626925e-01 1.31703269e+00 3.91017407e-01
-2.08028078e-01 4.37782615e-01 -1.40642524e-01 -3.07712942e-01
-7.10712790e-01 6.97484195e-01 5.91831915e-02 -2.32931465e-01
2.14824885e-01 -4.64361221e-01 -6.82778239e-01 8.65020812e-01
-8.44320297e-01 -1.65470734e-01 6.24526262e-01 7.60060251e-01
6.05305254e-01 2.25263387e-01 6.87725186e-01 -1.13476837e+00
9.86554801e-01 -3.80855054e-01 -8.00898492e-01 7.50081599e-01
-9.99253571e-01 4.22440469e-01 5.64260602e-01 -5.98465979e-01
-7.64272451e-01 -1.79147750e-01 3.62855375e-01 -2.07268968e-01
7.37272762e-03 2.91240424e-01 -8.47880021e-02 -1.36674363e-02
6.36326075e-01 6.65788054e-02 2.54027933e-01 -1.25519678e-01
1.48028523e-01 6.53105676e-01 2.33973321e-02 -1.18771183e+00
5.07338643e-01 6.31324500e-02 -2.69120902e-01 -4.83389795e-01
-2.84090549e-01 -1.94421530e-01 -1.60227031e-01 -1.66002274e-01
1.98304310e-01 -3.43641073e-01 -1.62360322e+00 1.80209219e-01
-9.11722779e-01 -6.02342010e-01 -1.10384807e-01 1.55698180e-01
-7.88471162e-01 1.00057587e-01 -6.02373064e-01 -7.55485654e-01
-2.96298295e-01 -1.58188260e+00 6.68528795e-01 3.55754256e-01
-2.05012903e-01 -5.18951535e-01 5.40854633e-01 4.88279879e-01
8.78001630e-01 -1.23803094e-01 1.19733238e+00 -7.19340742e-01
-1.13554788e+00 3.36811356e-02 1.57267451e-01 -7.98602849e-02
-1.57627799e-02 1.29660904e-01 -3.27585071e-01 -6.18913472e-01
-1.54912069e-01 -1.75919876e-01 3.00719619e-01 4.85939175e-01
1.06078947e+00 -3.98355275e-01 -5.07934988e-01 3.01528424e-01
1.36005688e+00 7.14362383e-01 2.12211072e-01 8.38160276e-01
1.18218571e-01 3.10494810e-01 8.17735255e-01 9.08317745e-01
6.30629539e-01 1.11879110e+00 4.61142778e-01 2.02147484e-01
3.61161381e-01 3.56770068e-01 3.78439933e-01 8.73673141e-01
3.61922793e-02 -2.46815100e-01 -8.51074040e-01 1.79153576e-01
-2.25755072e+00 -7.57742107e-01 5.16698241e-01 2.46444678e+00
1.14026713e+00 2.48336852e-01 4.32338893e-01 -1.83567122e-01
7.48657763e-01 -1.01993203e-01 -1.23790610e+00 -4.82518435e-01
2.57410973e-01 1.61912162e-02 6.50413513e-01 4.69406098e-01
-6.39354765e-01 8.65843952e-01 6.16857862e+00 7.07809746e-01
-9.44239795e-01 -4.33035428e-03 4.89019036e-01 -4.45697457e-01
-9.27295834e-02 9.57257077e-02 -1.02915096e+00 5.76472700e-01
1.21426535e+00 -3.79801154e-01 9.63381171e-01 7.76144028e-01
5.00189781e-01 -6.06117666e-01 -1.10546327e+00 6.46556199e-01
-3.46235543e-01 -1.45126522e+00 -2.16981933e-01 2.52805233e-01
6.11997068e-01 2.06295505e-01 -3.43481988e-01 3.95193309e-01
8.43623102e-01 -5.73597312e-01 3.47563446e-01 4.74256992e-01
9.54954103e-02 -1.06552422e+00 7.41073668e-01 6.69792354e-01
-8.89482558e-01 -9.09867138e-02 -3.86579186e-01 9.81098711e-02
4.65667322e-02 5.38960814e-01 -1.10685444e+00 2.37614810e-01
8.17788422e-01 -4.46355343e-01 -3.21739852e-01 9.35808122e-01
4.04364824e-01 4.44960803e-01 -6.14026010e-01 -4.66310889e-01
-1.07840128e-01 -2.14602292e-01 3.09612244e-01 7.57626653e-01
6.45567700e-02 6.21164031e-02 7.16179192e-01 6.75039053e-01
-4.53803223e-03 2.23528177e-01 7.76983351e-02 1.09304197e-01
1.09332228e+00 1.17604256e+00 -6.92082226e-01 -4.03260708e-01
-8.98410678e-02 4.52751011e-01 6.39178216e-01 5.25383770e-01
-8.94729674e-01 -4.22423542e-01 8.96410525e-01 3.10825586e-01
3.84256154e-01 -3.35431725e-01 -2.47726381e-01 -8.65114987e-01
-2.97077659e-05 -1.35692286e+00 4.76455092e-01 -2.50308365e-01
-1.10468030e+00 4.66059089e-01 -1.34437978e-02 -7.63077974e-01
-7.28945792e-01 6.42767595e-03 -3.60712737e-01 6.73230886e-01
-1.37565708e+00 -4.40645814e-01 -2.33549580e-01 6.58955753e-01
3.06036562e-01 -4.85451847e-01 6.86962187e-01 7.00991899e-02
-9.04911280e-01 6.02173209e-01 3.04565787e-01 -4.07167941e-01
1.00629020e+00 -1.29335213e+00 -2.58701861e-01 2.93190509e-01
-1.34092018e-01 7.08412051e-01 8.30886722e-01 -4.00869668e-01
-1.78493559e+00 -6.11022055e-01 2.76192009e-01 -1.30385429e-01
5.90380788e-01 8.77080671e-03 -8.26964974e-01 3.73600006e-01
3.63084882e-01 -1.69705585e-01 6.06840849e-01 2.09277272e-01
8.97306949e-02 -6.22507393e-01 -8.80598783e-01 5.91066957e-01
4.75806057e-01 1.91804580e-02 5.84304892e-02 4.48853761e-01
5.25863707e-01 -5.37255704e-01 -9.22724664e-01 3.54794413e-01
3.09438884e-01 -1.18766570e+00 8.17655981e-01 -5.06163061e-01
-2.73546547e-01 -5.91573656e-01 -2.38493964e-01 -1.26813471e+00
-1.82703272e-01 -8.30277681e-01 -2.24163741e-01 1.01940203e+00
4.30282623e-01 -8.99241507e-01 9.69896376e-01 6.86060548e-01
4.12134618e-01 -8.53976786e-01 -7.02166677e-01 -6.07349575e-01
-1.27007529e-01 -4.27417271e-02 8.28024566e-01 6.49956465e-01
-1.33889064e-01 4.29639965e-01 -4.29271489e-01 3.57107699e-01
8.20556819e-01 6.99513435e-01 1.31341636e+00 -8.22233319e-01
-1.15878391e+00 -4.82915044e-01 2.76507020e-01 -1.03856933e+00
-4.40515503e-02 -2.92746186e-01 -6.15649335e-02 -1.09451127e+00
4.09763575e-01 -1.11162889e+00 -3.50104928e-01 5.45220375e-01
-1.61084339e-01 -4.86040056e-01 2.73355275e-01 6.72901213e-01
-1.03152418e+00 2.80250371e-01 9.18934166e-01 -6.38760775e-02
-6.82428420e-01 5.12744069e-01 -5.97575188e-01 2.11371943e-01
9.81571078e-01 -6.60443902e-01 -4.01728809e-01 -3.10900271e-01
2.94196963e-01 6.16692126e-01 -2.19421819e-01 -8.16515863e-01
8.22300732e-01 -7.60181308e-01 -9.81061906e-02 -4.71318752e-01
-9.61382091e-02 -5.72772682e-01 5.02395213e-01 5.46923816e-01
-3.41795266e-01 3.93789828e-01 -1.39720485e-01 5.59886813e-01
-3.00119910e-02 -1.68822393e-01 7.78747201e-01 -2.28136659e-01
-2.37272009e-01 2.48905197e-01 -4.65343446e-01 1.47335589e-01
1.44712341e+00 1.40916929e-01 -4.30658489e-01 -5.01850665e-01
-5.52867532e-01 7.33271539e-01 6.24668181e-01 7.61926249e-02
8.11164007e-02 -1.00801635e+00 -6.34482086e-01 -1.79691613e-01
6.33318443e-03 -2.69873649e-01 7.44155496e-02 6.65628552e-01
-6.14144504e-01 6.65563717e-02 -1.12390891e-01 -7.64356256e-01
-1.31720769e+00 5.24169207e-01 5.18338501e-01 -6.89945877e-01
-5.63359298e-02 2.74903536e-01 -3.38045627e-01 -3.86728317e-01
3.23034257e-01 4.39645201e-02 6.43713176e-02 -2.87358854e-02
3.48445415e-01 3.98141831e-01 9.26487744e-02 2.45638996e-01
-3.36953163e-01 3.83264810e-01 -4.17064816e-01 -1.12074360e-01
1.46713579e+00 -3.53025556e-01 -3.02977383e-01 4.30741519e-01
6.40779734e-01 -3.85854058e-02 -9.24020767e-01 -4.79947358e-01
2.13401616e-01 -3.56866419e-01 -5.04792519e-02 -7.06066728e-01
-1.18812835e+00 2.25891650e-01 4.47332889e-01 6.28012061e-01
1.04181349e+00 -2.08900198e-01 5.22289097e-01 6.79807246e-01
9.04489219e-01 -1.49985886e+00 1.66112930e-01 2.02969909e-01
4.33234274e-01 -1.25707090e+00 1.49846315e-01 2.42209956e-02
-7.12015748e-01 1.18120813e+00 9.20791686e-01 1.00508809e-01
5.51591456e-01 6.71326518e-01 1.50967151e-01 -8.29036832e-02
-1.29683661e+00 -8.07527103e-04 -4.91844445e-01 2.56775260e-01
1.20855168e-01 4.86453772e-01 -3.45204204e-01 1.06130108e-01
1.65188278e-03 -2.37662587e-02 6.66984975e-01 1.32314193e+00
-6.89709306e-01 -1.78601921e+00 -6.69826090e-01 5.12025237e-01
-1.59395620e-01 4.67248619e-01 2.14869514e-01 6.78978264e-01
-2.26258099e-01 9.38038230e-01 5.34544028e-02 -2.70896196e-01
4.99695480e-01 -2.65355498e-01 2.94165462e-01 -5.36642551e-01
-7.89763451e-01 2.80864328e-01 1.38072088e-01 -6.39179289e-01
-2.36305922e-01 -6.02898180e-01 -1.14237010e+00 -9.76341188e-01
-2.91406721e-01 3.37896466e-01 6.32921278e-01 7.54952073e-01
6.62760258e-01 6.56307161e-01 1.05519354e+00 -7.04310954e-01
-1.19529092e+00 -5.83623886e-01 -7.12812066e-01 -1.36609524e-01
-6.64101774e-03 -7.13964880e-01 -1.40984163e-01 -3.79961401e-01] | [4.566728115081787, 2.6064038276672363] |
14996b16-a7aa-4f5a-852f-2c9fde50257b | scalable-spatiotemporal-graph-neural-networks | 2209.06520 | null | https://arxiv.org/abs/2209.06520v2 | https://arxiv.org/pdf/2209.06520v2.pdf | Scalable Spatiotemporal Graph Neural Networks | Neural forecasting of spatiotemporal time series drives both research and industrial innovation in several relevant application domains. Graph neural networks (GNNs) are often the core component of the forecasting architecture. However, in most spatiotemporal GNNs, the computational complexity scales up to a quadratic factor with the length of the sequence times the number of links in the graph, hence hindering the application of these models to large graphs and long temporal sequences. While methods to improve scalability have been proposed in the context of static graphs, few research efforts have been devoted to the spatiotemporal case. To fill this gap, we propose a scalable architecture that exploits an efficient encoding of both temporal and spatial dynamics. In particular, we use a randomized recurrent neural network to embed the history of the input time series into high-dimensional state representations encompassing multi-scale temporal dynamics. Such representations are then propagated along the spatial dimension using different powers of the graph adjacency matrix to generate node embeddings characterized by a rich pool of spatiotemporal features. The resulting node embeddings can be efficiently pre-computed in an unsupervised manner, before being fed to a feed-forward decoder that learns to map the multi-scale spatiotemporal representations to predictions. The training procedure can then be parallelized node-wise by sampling the node embeddings without breaking any dependency, thus enabling scalability to large networks. Empirical results on relevant datasets show that our approach achieves results competitive with the state of the art, while dramatically reducing the computational burden. | ['Cesare Alippi', 'Filippo Maria Bianchi', 'Ivan Marisca', 'Andrea Cini'] | 2022-09-14 | null | null | null | null | ['temporal-sequences'] | ['reasoning'] | [-3.09986435e-02 5.71616702e-02 -4.37842667e-01 2.98426002e-02
-1.60869937e-02 -5.25215864e-01 7.76410818e-01 3.54793668e-01
-1.45789221e-01 2.33131886e-01 2.58987486e-01 -6.34222150e-01
-1.08776174e-01 -1.13895762e+00 -6.58872902e-01 -6.32216871e-01
-5.39415717e-01 1.18068546e-01 3.27540696e-01 -3.12393010e-01
1.01921663e-01 5.70334375e-01 -1.34416211e+00 -1.10022672e-01
5.83935320e-01 1.00560284e+00 1.00229949e-01 7.72339642e-01
-6.79900199e-02 9.03508544e-01 -2.05358759e-01 -1.23711139e-01
5.61275482e-02 -4.68500674e-01 -4.57356602e-01 -6.83139637e-02
-1.75560880e-02 -1.20670563e-02 -1.02599359e+00 7.24518895e-01
2.95795053e-01 3.64084572e-01 3.81052285e-01 -1.17316973e+00
-8.52331698e-01 5.57791293e-01 -2.73272067e-01 4.75545228e-01
1.10504307e-01 -8.22655857e-02 1.11753714e+00 -5.26806951e-01
7.57938921e-01 8.38699281e-01 6.49027586e-01 2.35858679e-01
-1.28772509e+00 -4.01713133e-01 5.28467238e-01 2.21603796e-01
-1.25832212e+00 -1.95844665e-01 1.14234447e+00 -4.45134848e-01
1.31812179e+00 -1.23713650e-01 9.08828437e-01 1.13489521e+00
4.27736700e-01 3.44801873e-01 5.01196861e-01 -1.91573761e-02
3.67451280e-01 -3.44842017e-01 -8.56188238e-02 7.67499804e-01
-2.05628052e-02 -7.26436898e-02 -6.03770673e-01 -1.37422338e-01
8.85509789e-01 4.67898279e-01 -8.19697604e-02 -2.28193268e-01
-1.18865049e+00 9.23102558e-01 9.52464938e-01 5.53111315e-01
-7.67692208e-01 5.77824116e-01 5.71576118e-01 5.48783839e-01
9.27034259e-01 2.06527531e-01 -2.43928492e-01 -1.00713730e-01
-8.67851675e-01 1.14550635e-01 8.22177708e-01 5.69217384e-01
7.10499823e-01 2.93835402e-01 1.89382896e-01 4.27064002e-01
5.93546368e-02 4.10551548e-01 5.44315577e-01 -6.44886911e-01
5.37164211e-01 8.41797531e-01 -2.53083497e-01 -1.73626220e+00
-5.83272755e-01 -5.52926123e-01 -1.22089458e+00 -4.30011839e-01
2.04842731e-01 -1.66625500e-01 -6.50459766e-01 1.77247036e+00
3.78903508e-01 7.26271331e-01 -1.51659697e-01 7.55204141e-01
2.29071870e-01 1.22925794e+00 -1.72883093e-01 -1.17323801e-01
9.73249316e-01 -8.57808352e-01 -4.85691726e-01 -1.05749890e-01
8.62418413e-01 -1.24223188e-01 7.41928935e-01 -1.23069324e-01
-7.73432791e-01 -3.89296621e-01 -8.98525655e-01 -2.53555756e-02
-6.89805686e-01 -2.09555909e-01 5.61712444e-01 1.71436206e-01
-1.25494552e+00 7.91801631e-01 -1.14599442e+00 -4.54052597e-01
1.29443750e-01 3.31048399e-01 -2.76269406e-01 1.19009772e-02
-1.31190777e+00 5.63029230e-01 2.98082590e-01 2.80881703e-01
-6.12396419e-01 -6.16397321e-01 -9.53959227e-01 2.03325585e-01
2.37764746e-01 -6.46387577e-01 6.55495405e-01 -6.09026253e-01
-1.47538507e+00 9.03361440e-02 -3.67438197e-01 -8.06380391e-01
1.04598530e-01 1.88369498e-01 -6.32840991e-01 1.08015463e-01
-7.40345344e-02 1.64006189e-01 8.30991864e-01 -5.49653292e-01
-3.74091774e-01 -2.39783227e-01 1.79788962e-01 -5.21562509e-02
-7.29448676e-01 -3.12303692e-01 -3.50237936e-01 -8.38066280e-01
1.18580103e-01 -1.16636205e+00 -6.37455642e-01 -1.98115945e-01
-3.51920694e-01 -3.70305240e-01 9.51768577e-01 -6.32282376e-01
1.49821115e+00 -2.07181525e+00 5.42426407e-01 3.44915688e-01
3.56493384e-01 6.32911399e-02 -2.76870966e-01 1.07364857e+00
8.15200806e-02 -7.93011766e-03 -1.33361146e-01 -2.79334396e-01
-6.94857240e-02 2.47768506e-01 -6.86925590e-01 6.18700981e-01
1.63995311e-01 1.22760463e+00 -1.15178704e+00 -4.57673520e-02
1.43012628e-01 6.92119539e-01 -4.26328659e-01 1.72909766e-01
-3.17273051e-01 4.17636275e-01 -6.63568377e-01 8.86358395e-02
-1.25887515e-02 -8.04810405e-01 4.36769813e-01 2.16368828e-02
-3.57836448e-02 6.29238129e-01 -8.21211040e-01 1.63814414e+00
-7.25691497e-01 6.99074984e-01 -3.88748258e-01 -1.40441883e+00
9.64837134e-01 4.09890443e-01 6.95615888e-01 -7.39351869e-01
-1.72765166e-01 8.75088200e-02 -1.56332165e-01 -3.04518521e-01
4.82738376e-01 7.21067935e-02 -2.68740982e-01 7.01161087e-01
-8.47607851e-02 2.35122994e-01 1.99286640e-01 3.20874751e-01
1.34855258e+00 -7.14153424e-02 2.61602495e-02 1.45790562e-01
3.15967441e-01 -1.63215026e-02 3.31141621e-01 3.10060680e-01
1.88917309e-01 1.53905809e-01 8.70145142e-01 -8.08145106e-01
-1.21730649e+00 -8.08964849e-01 5.77841640e-01 1.08160269e+00
-8.27754438e-02 -7.44270265e-01 -5.25763571e-01 -5.85701048e-01
1.47049725e-01 3.54770929e-01 -7.40986347e-01 -2.93086827e-01
-9.08179879e-01 -3.41094762e-01 3.23332042e-01 6.00391865e-01
1.04696102e-01 -9.80038404e-01 -5.34311175e-01 5.77227056e-01
1.05411686e-01 -1.26628029e+00 -4.72486794e-01 1.50191281e-02
-1.04279721e+00 -7.58938849e-01 -5.12432635e-01 -7.36547828e-01
5.80400705e-01 2.60627210e-01 8.24856281e-01 -3.15813571e-02
-1.37863100e-01 2.46846929e-01 -2.80193716e-01 2.03011587e-01
-1.82783693e-01 5.28381824e-01 1.22332700e-01 2.95861244e-01
-6.28203433e-03 -1.20002735e+00 -7.00538754e-01 3.08040138e-02
-9.48847651e-01 1.20240271e-01 2.90627062e-01 6.94930494e-01
4.78164405e-01 -3.02879959e-02 7.91292369e-01 -7.61097491e-01
7.38479912e-01 -9.51580763e-01 -7.62925565e-01 4.42540012e-02
-5.64555883e-01 1.64915025e-01 1.23886681e+00 -6.97973371e-01
-3.28611106e-01 -1.70112923e-02 1.98433742e-01 -7.46742845e-01
3.11906397e-01 1.03600311e+00 5.45360565e-01 1.66220404e-02
2.46616870e-01 5.24184525e-01 -9.06847883e-03 -2.94275016e-01
7.06559181e-01 2.50799686e-01 2.42077857e-01 -1.54945776e-01
8.07941139e-01 4.12227601e-01 3.32165748e-01 -9.59117115e-01
-4.70539212e-01 -2.05450654e-01 -5.72098374e-01 -2.37262592e-01
6.53418839e-01 -7.30175436e-01 -6.00746870e-01 1.61038071e-01
-1.00357389e+00 -5.38967848e-01 -3.17192346e-01 4.16632742e-01
-4.04140919e-01 1.32306129e-01 -9.54337478e-01 -6.18047297e-01
-1.55251116e-01 -7.48694897e-01 8.74130011e-01 -1.15428127e-01
-1.18684210e-01 -1.42401886e+00 2.95381844e-01 -3.03966939e-01
7.08300352e-01 4.88147110e-01 1.09867251e+00 -5.78979373e-01
-6.71529949e-01 -6.33889973e-01 -1.22056283e-01 -1.64590687e-01
9.19482335e-02 -1.41343400e-01 -5.18437982e-01 -2.33521208e-01
-3.02236050e-01 1.23051234e-01 8.02105963e-01 3.51766527e-01
1.01815867e+00 -5.09327292e-01 -4.92162794e-01 5.28780460e-01
1.25822699e+00 -1.02660656e-01 2.30916962e-01 -2.82733254e-02
1.08850694e+00 5.34949899e-01 3.88208255e-02 4.96787220e-01
7.30447412e-01 5.53317785e-01 5.03068805e-01 2.04678208e-01
4.55708988e-02 -6.45161569e-01 4.47871059e-01 1.44704258e+00
-1.93755433e-01 -3.78047436e-01 -1.08975101e+00 8.35316062e-01
-2.02148128e+00 -9.93009567e-01 9.18013155e-02 1.95571411e+00
2.08330721e-01 -2.02005487e-02 2.35932618e-01 7.45639652e-02
5.97709417e-01 9.17931497e-01 -7.03851461e-01 -4.39627588e-01
1.37867510e-01 2.51097325e-02 5.59291005e-01 3.74024659e-01
-9.09287274e-01 9.46809769e-01 5.93490171e+00 3.86158347e-01
-1.67343795e+00 1.24005936e-01 5.10722339e-01 -2.91135192e-01
-3.89450252e-01 4.87083122e-02 -3.10543239e-01 4.66007411e-01
1.67752755e+00 -3.61790478e-01 7.51703203e-01 6.36916578e-01
3.50475878e-01 5.35068810e-01 -9.10443068e-01 7.88471222e-01
-1.37077779e-01 -1.67983961e+00 4.88729551e-02 2.70576864e-01
7.20811188e-01 4.84241128e-01 1.62264377e-01 1.98422521e-01
2.43577734e-01 -1.00553620e+00 4.09308046e-01 4.46792662e-01
5.88639021e-01 -7.01273322e-01 3.23344558e-01 5.01143217e-01
-1.72454000e+00 -2.11241022e-01 -4.02522415e-01 -3.45291644e-01
4.95575905e-01 7.64493704e-01 -6.66013062e-01 6.85914338e-01
4.90974039e-01 1.30686295e+00 -2.45275483e-01 5.08093536e-01
-7.15119019e-02 8.14438403e-01 -4.78204042e-01 -3.18535656e-01
5.88163376e-01 -4.13453370e-01 4.32600498e-01 9.47681785e-01
5.85182786e-01 -7.29949325e-02 1.76159456e-01 6.60003662e-01
-1.66756034e-01 -4.14719544e-02 -1.08939672e+00 -6.24863565e-01
4.94411349e-01 1.03402054e+00 -7.15458989e-01 -2.68146336e-01
-6.01736426e-01 1.03403878e+00 7.52937913e-01 6.96222603e-01
-6.95502102e-01 -4.12779123e-01 5.89877665e-01 3.11759919e-01
6.30129993e-01 -7.39354253e-01 5.00768907e-02 -1.09132838e+00
2.55875587e-01 -3.05734694e-01 3.81465822e-01 -3.18729699e-01
-1.24371386e+00 7.77816534e-01 -2.36539721e-01 -1.07270670e+00
-7.21442819e-01 -3.02504629e-01 -6.85752690e-01 7.71289825e-01
-1.37905395e+00 -1.24963653e+00 8.24264064e-02 5.33580899e-01
2.18243331e-01 6.90716133e-02 7.92798162e-01 2.57417470e-01
-7.72323728e-01 2.02317938e-01 1.59318849e-01 1.32284358e-01
9.97004583e-02 -9.41946626e-01 1.00971401e+00 9.37457740e-01
4.12988305e-01 5.76656759e-01 5.00591159e-01 -6.35392845e-01
-1.86329019e+00 -1.42780912e+00 1.25391722e+00 -5.60939945e-02
1.28045762e+00 -6.15802169e-01 -1.00169718e+00 8.80783498e-01
8.14336091e-02 6.60133123e-01 3.59975815e-01 1.43812373e-01
-5.91252327e-01 -1.21042266e-01 -4.51250076e-01 7.55198896e-01
1.13169825e+00 -1.04683542e+00 1.32343024e-02 3.25578600e-01
9.58705246e-01 -1.95365310e-01 -1.20756435e+00 8.41113627e-02
4.22679245e-01 -5.93132019e-01 7.87183762e-01 -7.94678986e-01
4.47186649e-01 -2.13727802e-01 -3.30073684e-02 -1.38223147e+00
-4.83228892e-01 -7.16691315e-01 -7.20211864e-01 7.61172354e-01
4.81688917e-01 -1.03533697e+00 8.14139128e-01 2.93260604e-01
5.94369173e-02 -1.13873625e+00 -9.68878269e-01 -6.64479196e-01
-7.33272508e-02 -5.18034816e-01 6.69637084e-01 9.77127194e-01
2.37864032e-01 3.17269176e-01 -5.78754783e-01 2.72357851e-01
1.72898591e-01 4.86188352e-01 5.67939579e-01 -1.13108575e+00
-1.64669827e-01 -5.00728011e-01 -6.92409456e-01 -1.21306407e+00
3.45251590e-01 -1.05011761e+00 -3.09353948e-01 -1.57457376e+00
-3.01370114e-01 -3.26797098e-01 -4.51680213e-01 3.57499510e-01
6.73609674e-02 1.42587617e-01 3.26963365e-02 2.76342481e-01
-5.30849218e-01 7.42059290e-01 9.42146778e-01 -6.77978480e-03
-2.38392040e-01 -1.26181483e-01 -2.12758824e-01 3.15824151e-01
7.73433685e-01 -3.99213135e-01 -6.84652388e-01 -4.08490330e-01
4.90740895e-01 2.91409969e-01 4.05282021e-01 -9.45538282e-01
5.72646797e-01 -1.15043677e-01 1.65618389e-04 -4.06249255e-01
4.19844002e-01 -7.59575248e-01 2.54973203e-01 4.12549704e-01
-4.39957082e-01 6.39021516e-01 -2.57624872e-02 1.19160533e+00
-2.57734925e-01 5.55341601e-01 1.34394377e-01 2.05622777e-01
-4.80229557e-01 8.18606496e-01 -4.41196352e-01 -1.86242774e-01
1.01367640e+00 -1.89929586e-02 -1.03414565e-01 -6.58769846e-01
-6.47587359e-01 6.71097934e-02 4.56605285e-01 6.66028023e-01
4.66311425e-01 -1.39728522e+00 -4.02627170e-01 2.64592618e-01
-2.39892267e-02 -2.00115919e-01 2.77012527e-01 9.96948659e-01
-3.14784974e-01 6.25005960e-01 1.46832094e-01 -5.38757741e-01
-5.29947281e-01 8.17871094e-01 2.60417640e-01 -5.31928778e-01
-1.04190016e+00 4.57842112e-01 -9.28441733e-02 -2.82953829e-01
3.59336846e-02 -5.08746028e-01 -1.65729553e-01 5.84601238e-02
3.51283550e-01 3.45079392e-01 -1.12984240e-01 -8.49380374e-01
-2.98781931e-01 6.26525342e-01 2.55891412e-01 2.21242290e-02
1.60616374e+00 -1.40504122e-01 -2.89675027e-01 7.60095954e-01
1.57486451e+00 -3.30800205e-01 -1.23041928e+00 -5.01632690e-01
1.91412598e-01 -6.20430633e-02 4.48422730e-02 2.87057441e-02
-1.43662548e+00 8.97961080e-01 1.30017236e-01 7.83042967e-01
1.00099766e+00 3.64096388e-02 1.12091470e+00 3.40799749e-01
4.51799661e-01 -6.85931802e-01 -1.13033745e-02 6.17590725e-01
4.97055918e-01 -7.67275333e-01 -3.03288698e-01 -2.43847400e-01
-3.92349750e-01 1.11060894e+00 -1.24721462e-02 -7.79334664e-01
1.04402876e+00 1.58720523e-01 -3.63968730e-01 -3.61930370e-01
-1.14501011e+00 6.00476153e-02 3.69173229e-01 3.45024288e-01
3.78810912e-01 1.36640415e-01 -9.59072262e-03 3.25141311e-01
2.14788839e-02 -2.72231065e-02 2.06698626e-01 7.61958599e-01
-1.63915500e-01 -9.97088015e-01 3.41167212e-01 6.10915840e-01
-2.15833709e-01 2.40554027e-02 -8.43147337e-02 5.44699669e-01
-4.05533999e-01 7.73648202e-01 3.91594499e-01 -6.53397024e-01
2.16179356e-01 5.86719587e-02 2.86818426e-02 -4.48515236e-01
-5.12929559e-01 -2.25321516e-01 -3.81907597e-02 -9.87718701e-01
-2.25843892e-01 -5.24547577e-01 -1.10196650e+00 -5.56171060e-01
8.59009400e-02 1.53925136e-01 5.73109150e-01 9.23518479e-01
9.46307600e-01 6.41889036e-01 7.42979944e-01 -1.15636349e+00
-2.44620323e-01 -7.05515802e-01 -4.85217333e-01 2.90052325e-01
5.30866444e-01 -4.38587576e-01 -2.64780074e-01 -9.60820243e-02] | [6.813604831695557, 2.8231217861175537] |
99fa4b14-c224-4447-b0f0-34891c8e24c0 | context-aware-deep-feature-compression-for | 1803.10537 | null | http://arxiv.org/abs/1803.10537v1 | http://arxiv.org/pdf/1803.10537v1.pdf | Context-aware Deep Feature Compression for High-speed Visual Tracking | We propose a new context-aware correlation filter based tracking framework to
achieve both high computational speed and state-of-the-art performance among
real-time trackers. The major contribution to the high computational speed lies
in the proposed deep feature compression that is achieved by a context-aware
scheme utilizing multiple expert auto-encoders; a context in our framework
refers to the coarse category of the tracking target according to appearance
patterns. In the pre-training phase, one expert auto-encoder is trained per
category. In the tracking phase, the best expert auto-encoder is selected for a
given target, and only this auto-encoder is used. To achieve high tracking
performance with the compressed feature map, we introduce extrinsic denoising
processes and a new orthogonality loss term for pre-training and fine-tuning of
the expert auto-encoders. We validate the proposed context-aware framework
through a number of experiments, where our method achieves a comparable
performance to state-of-the-art trackers which cannot run in real-time, while
running at a significantly fast speed of over 100 fps. | ['Jongwon Choi', 'Sangdoo Yun', 'Jiyeoup Jeong', 'Hyung Jin Chang', 'Yiannis Demiris', 'Jin Young Choi', 'Tobias Fischer', 'Kyuewang Lee'] | 2018-03-28 | context-aware-deep-feature-compression-for-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Choi_Context-Aware_Deep_Feature_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Choi_Context-Aware_Deep_Feature_CVPR_2018_paper.pdf | cvpr-2018-6 | ['feature-compression'] | ['computer-vision'] | [-1.57988310e-01 -5.01710057e-01 -1.61080994e-02 -1.86401978e-01
-7.24382579e-01 -3.62312824e-01 4.81754065e-01 -1.34324953e-01
-6.28159702e-01 3.99560630e-01 -2.60800850e-02 1.26515165e-01
-1.37288868e-01 -4.54557657e-01 -8.05108607e-01 -6.57070577e-01
-2.09382102e-01 4.16135311e-01 4.84221101e-01 1.21184856e-01
-3.88638750e-02 5.29232144e-01 -1.67462075e+00 -2.14441508e-01
7.75047779e-01 1.39730322e+00 4.17012841e-01 7.70993769e-01
2.79329717e-01 5.02695739e-01 -5.59488416e-01 -5.95432758e-01
6.60561979e-01 -2.50164419e-01 1.49190560e-01 9.82309133e-02
8.65685046e-01 -2.32231110e-01 -6.44076526e-01 1.31238592e+00
6.08830392e-01 1.88012552e-02 2.69891620e-01 -9.76018548e-01
-2.61920273e-01 5.24606630e-02 -4.81557727e-01 2.82647073e-01
-1.47690237e-01 2.61168629e-01 8.81079972e-01 -7.06698120e-01
3.22951615e-01 1.10967743e+00 9.01562035e-01 4.65128481e-01
-1.10866988e+00 -1.01837611e+00 1.10720187e-01 2.26755783e-01
-1.37023342e+00 -4.66911942e-01 5.79631567e-01 -5.33134639e-01
3.33508909e-01 -2.18401462e-01 8.99076879e-01 6.94030046e-01
5.28591931e-01 4.29316133e-01 8.74261796e-01 -1.71626836e-01
2.43491814e-01 -1.81582212e-01 -1.40498936e-01 8.11728239e-01
4.66666609e-01 6.94352925e-01 -8.07161212e-01 -2.43746750e-02
1.00064182e+00 -8.99944529e-02 -1.14174701e-01 -4.91361737e-01
-1.26723707e+00 8.22440624e-01 5.96172750e-01 1.20797716e-01
-4.91874307e-01 5.31231463e-01 4.73732233e-01 1.57003224e-01
2.14043349e-01 5.91275729e-02 -2.75408357e-01 -3.41586709e-01
-1.31679738e+00 2.90309399e-01 5.51942825e-01 9.49571311e-01
5.90480566e-01 2.18300149e-01 -4.90483910e-01 3.74889940e-01
4.63186473e-01 1.05896318e+00 3.22507977e-01 -1.00945508e+00
3.50325733e-01 2.14746267e-01 3.86453629e-01 -9.29755211e-01
-1.06812231e-01 -1.32544482e+00 -5.58269620e-01 4.49130595e-01
3.32615018e-01 -2.81949162e-01 -7.40123034e-01 1.87366498e+00
6.45498514e-01 6.62766516e-01 -9.78011638e-02 1.15137196e+00
3.55597705e-01 3.09537798e-01 9.57993045e-02 -3.15807849e-01
1.67153835e+00 -9.64170754e-01 -8.74195933e-01 -2.69903839e-01
1.37688145e-01 -9.77766037e-01 2.99092233e-01 3.47026259e-01
-8.36937010e-01 -1.08442259e+00 -1.07086015e+00 3.74584019e-01
7.64784813e-02 8.65528166e-01 5.92453599e-01 6.21294022e-01
-7.58707047e-01 5.89753509e-01 -1.13134396e+00 -1.67547584e-01
3.23414981e-01 5.07717133e-01 -2.61362970e-01 3.10686380e-01
-8.84954929e-01 8.72147381e-01 2.96763122e-01 8.83157644e-03
-9.86341059e-01 -7.67146230e-01 -6.28293157e-01 7.94490650e-02
5.08883417e-01 -8.48342717e-01 1.09478116e+00 -7.94580519e-01
-1.51036239e+00 3.61408204e-01 -1.47719368e-01 -9.09271061e-01
3.86269122e-01 -5.56649208e-01 -5.23245096e-01 1.04359642e-01
7.37878159e-02 5.78977227e-01 1.33716917e+00 -9.95177031e-01
-1.07698512e+00 -2.59940416e-01 -1.53032333e-01 2.05865458e-01
-3.38234246e-01 3.60431075e-02 -8.54376078e-01 -8.87350619e-01
-7.48840421e-02 -1.11098301e+00 1.46551449e-02 4.13930178e-01
1.38498291e-01 -5.02493419e-02 1.11934614e+00 -6.41702414e-01
1.19346297e+00 -2.21462369e+00 1.12779029e-02 -1.13313153e-01
1.78525388e-01 6.63843751e-01 -1.28775332e-02 -3.50757129e-02
3.24315965e-01 -8.45366299e-01 1.90091193e-01 -5.64268231e-01
7.46136457e-02 1.52535602e-01 -2.02094480e-01 7.71930516e-01
3.52024399e-02 6.76895082e-01 -9.77708697e-01 -6.53493226e-01
5.16167819e-01 7.46352136e-01 -5.13700604e-01 2.45585069e-01
-1.56888947e-01 6.40332639e-01 -4.46031243e-01 4.79592770e-01
7.69011199e-01 -1.62570074e-01 -4.23519686e-03 -6.63140357e-01
-3.58877510e-01 9.38450471e-02 -1.25553703e+00 1.88657141e+00
-3.67605358e-01 7.18354464e-01 2.69580752e-01 -7.18856037e-01
8.92204046e-01 1.30820781e-01 7.35943437e-01 -5.54854274e-01
1.92153379e-01 1.95037931e-01 2.11329952e-01 9.41487476e-02
7.96582818e-01 2.03208476e-01 1.42408356e-01 6.66142255e-02
3.93942297e-01 5.35889208e-01 3.30597669e-01 6.59936592e-02
1.04526401e+00 5.40354967e-01 1.55753572e-03 -2.95509607e-01
7.20693648e-01 -1.83352903e-02 1.05380392e+00 7.30818987e-01
-3.50951672e-01 4.02991995e-02 -2.73214370e-01 -5.20926595e-01
-8.84496272e-01 -1.01226032e+00 -1.97003894e-02 9.78212774e-01
2.70830810e-01 -5.87366581e-01 -5.29518843e-01 -5.68485200e-01
1.33696631e-01 1.13595888e-01 -5.68729818e-01 -1.35134861e-01
-5.13423324e-01 -3.12130928e-01 3.84276897e-01 5.10852277e-01
6.96023226e-01 -5.55576146e-01 -1.06471109e+00 4.87958640e-01
2.68080272e-03 -1.31104207e+00 -7.55117536e-01 5.88461943e-02
-1.06835890e+00 -1.08082628e+00 -5.89824140e-01 -5.28756499e-01
5.44685423e-01 2.80238599e-01 8.85399997e-01 -2.17843696e-01
-1.82686403e-01 2.87323713e-01 -2.33759642e-01 -1.68506399e-01
-4.22842354e-02 -2.04560637e-01 2.95597047e-01 2.15048850e-01
3.48820210e-01 -2.65552461e-01 -9.10960913e-01 4.41226542e-01
-4.72209185e-01 -1.66110665e-01 8.22947502e-01 9.21604514e-01
7.69483209e-01 1.62059546e-01 6.60011619e-02 -1.52519152e-01
6.98688924e-02 8.96975324e-02 -1.31612873e+00 1.36243507e-01
-4.03829366e-01 2.87029564e-01 6.19149208e-01 -7.23538637e-01
-9.07508373e-01 5.96643984e-01 -2.54462920e-02 -9.00707006e-01
3.20176393e-01 1.08868346e-01 1.35497600e-01 -7.24170685e-01
4.44166571e-01 3.31095219e-01 1.24723837e-01 -5.58963716e-01
3.20175856e-01 2.34419882e-01 1.04785645e+00 -4.23869908e-01
1.31475973e+00 5.14227808e-01 2.56174147e-01 -5.48876822e-01
-8.69152069e-01 -6.63497150e-01 -5.08264542e-01 -5.25154471e-01
6.80262208e-01 -1.46201408e+00 -1.02458465e+00 2.41904557e-01
-9.81256664e-01 7.47759715e-02 -3.09668630e-01 1.05214131e+00
-4.87632155e-01 3.42631012e-01 -3.23796511e-01 -8.48383427e-01
-5.47268271e-01 -1.16442084e+00 1.22117734e+00 5.00735164e-01
3.87245893e-01 -6.70765281e-01 2.35159203e-01 6.20640069e-02
6.38513744e-01 1.20559089e-01 -1.12695009e-01 -3.76844823e-01
-8.06820810e-01 -2.13744536e-01 -1.93063170e-01 1.87205657e-01
-3.87772508e-02 -2.72167116e-01 -7.10198164e-01 -7.66034365e-01
-1.08810686e-01 -9.39213336e-02 8.28532755e-01 5.23817837e-01
8.19193840e-01 -5.89945726e-02 -5.85485518e-01 7.88327157e-01
1.51709449e+00 1.68358743e-01 3.11870992e-01 2.67591625e-01
6.19735897e-01 -1.25565410e-01 1.13600338e+00 6.11631632e-01
2.44871303e-01 1.19004345e+00 4.17217821e-01 2.05585822e-01
-4.50127661e-01 -2.52525061e-01 4.23212796e-01 6.89331532e-01
-1.53905168e-01 2.10780010e-01 -1.97938666e-01 5.17002761e-01
-2.00331068e+00 -1.14123476e+00 1.02069423e-01 2.68823647e+00
5.34350991e-01 1.99581102e-01 2.48278514e-01 -3.08228612e-01
8.40740085e-01 5.85923484e-03 -6.24475837e-01 3.87081832e-01
2.22821444e-01 1.59507707e-01 8.17484438e-01 3.92095029e-01
-1.45979142e+00 9.70933735e-01 5.52433729e+00 1.04483271e+00
-1.24379301e+00 2.08144039e-01 -2.54000038e-01 -2.00810045e-01
4.52063918e-01 3.16998363e-02 -1.29296184e+00 7.37149894e-01
1.03907311e+00 -1.34635881e-01 3.78277302e-01 1.04179323e+00
1.17416613e-01 1.05041623e-01 -8.75133932e-01 1.14079797e+00
-2.00919807e-02 -1.23511124e+00 -5.36415279e-01 2.46932149e-01
5.01141429e-01 2.03919902e-01 1.33903027e-02 3.20900351e-01
3.30939531e-01 -2.99822628e-01 9.26963031e-01 6.03596210e-01
6.03641272e-01 -7.74216473e-01 5.55638611e-01 8.34286287e-02
-1.97814965e+00 -1.36892959e-01 -6.63813829e-01 3.37520063e-01
1.06570423e-01 8.48345339e-01 -3.91133547e-01 8.35281432e-01
8.30058932e-01 7.69325078e-01 -5.34856319e-01 1.50664079e+00
-1.63809389e-01 4.31510836e-01 -5.26035070e-01 2.62242883e-01
7.41547495e-02 6.23674430e-02 8.41578543e-01 1.21634007e+00
4.81784195e-01 -9.61477086e-02 4.64477122e-01 4.76495385e-01
1.15865231e-01 -3.40172321e-01 -1.97580367e-01 1.37302324e-01
7.46596396e-01 1.46283734e+00 -4.31082666e-01 -3.60378146e-01
-4.15643603e-01 7.74456024e-01 1.85666919e-01 -3.12724896e-02
-1.22011626e+00 -3.93534005e-01 7.96363115e-01 -5.52525334e-02
9.89614785e-01 -4.51346010e-01 2.06882879e-01 -1.17625189e+00
-2.78302487e-02 -7.09726691e-01 2.56466717e-01 -3.13591093e-01
-1.13366008e+00 6.62178695e-01 -2.85145760e-01 -1.93366444e+00
-3.10910225e-01 -4.01611924e-01 -3.37831467e-01 7.83749282e-01
-1.45138991e+00 -1.24687946e+00 -3.98210526e-01 6.38505638e-01
2.46054009e-01 -4.49996799e-01 4.10279572e-01 6.23872459e-01
-1.89179093e-01 8.44721019e-01 2.40356028e-01 1.73301458e-01
8.86146188e-01 -1.01545584e+00 3.32269311e-01 1.20708847e+00
2.05919534e-01 6.89548433e-01 8.09047520e-01 -8.53792846e-01
-1.80856967e+00 -1.34924328e+00 7.03177750e-01 -1.86795592e-02
6.31495118e-01 -2.55900353e-01 -4.94889349e-01 4.83861595e-01
4.99127880e-02 3.86777371e-01 4.48390901e-01 1.00832686e-01
-2.32537448e-01 -5.30970812e-01 -1.00430655e+00 2.76223689e-01
9.47105467e-01 -3.72380435e-01 -5.57558596e-01 2.62712836e-01
4.77785885e-01 -6.77295506e-01 -1.01074433e+00 3.51201177e-01
8.66700470e-01 -7.92400777e-01 1.10663199e+00 3.50043438e-02
-2.86616772e-01 -1.02517998e+00 1.16941430e-01 -1.09857273e+00
-7.64509678e-01 -7.03493536e-01 -7.09653020e-01 1.02710295e+00
-2.11659044e-01 -4.56378371e-01 8.52604628e-01 -6.99865893e-02
-1.48901179e-01 -5.01427412e-01 -1.22648215e+00 -1.09463680e+00
-5.94758511e-01 -2.18582109e-01 2.98024118e-01 3.08149934e-01
-5.60733974e-01 1.48957476e-01 -7.17072904e-01 5.22729874e-01
1.35927403e+00 3.50092947e-01 9.87745345e-01 -1.44566512e+00
-7.30900824e-01 -1.13207046e-02 -7.25866139e-01 -1.46302986e+00
-1.35380998e-01 -3.94997269e-01 2.98588127e-01 -9.76204276e-01
8.17520022e-02 -3.95105332e-01 -2.63962269e-01 2.31826097e-01
-1.44955739e-01 3.87806445e-01 5.22347808e-01 3.46179336e-01
-9.39831793e-01 7.38664865e-01 1.08246255e+00 -7.07509518e-02
1.91143842e-03 2.02740088e-01 -4.41734344e-01 3.89829963e-01
3.54794890e-01 -6.75823510e-01 -1.15108721e-01 -4.40229058e-01
-3.63541901e-01 1.04700394e-01 4.24861223e-01 -1.74125457e+00
7.37980068e-01 3.74431461e-01 6.80733800e-01 -9.22966719e-01
5.63226998e-01 -1.20214522e+00 3.38992715e-01 8.54592025e-01
6.46338835e-02 1.19945044e-02 2.88725555e-01 8.68877411e-01
-1.92656174e-01 2.36232358e-04 9.10385251e-01 2.79042393e-01
-5.85445046e-01 5.65623999e-01 -5.10142669e-02 -1.87834799e-01
9.12752986e-01 -1.57314494e-01 -1.77954942e-01 -2.20106661e-01
-5.60818970e-01 2.04980716e-01 4.99349892e-01 3.90059799e-01
2.42386729e-01 -1.66483676e+00 -4.93072003e-01 1.54709280e-01
-7.75298029e-02 -3.93493503e-01 1.93281919e-01 8.64246428e-01
-1.75859153e-01 5.46158075e-01 -3.22199494e-01 -9.10392642e-01
-1.32873106e+00 4.75249022e-01 2.93183535e-01 -5.15693903e-01
-7.52825618e-01 5.34312308e-01 9.65057239e-02 1.16744161e-01
5.45919061e-01 -4.24226224e-02 3.17004099e-02 -3.78913164e-01
7.75727153e-01 2.59322077e-01 -1.67384297e-02 -9.08682287e-01
-5.17037153e-01 9.19325590e-01 -5.22297285e-02 -1.17601573e-01
9.34069932e-01 2.09391434e-02 4.17932272e-01 -2.13534355e-01
7.48033881e-01 3.03839911e-02 -1.95599282e+00 -3.99585485e-01
-3.26277852e-01 -8.15328360e-01 4.24582541e-01 -6.79857016e-01
-1.51760876e+00 5.38025022e-01 1.16121185e+00 -3.97542894e-01
1.35907400e+00 -3.38598162e-01 8.25644910e-01 2.91619956e-01
5.69545984e-01 -1.14549387e+00 2.96271350e-02 5.76605797e-01
4.19214368e-01 -1.11190152e+00 2.91975230e-01 -3.13575506e-01
-4.61544842e-01 9.95557249e-01 5.89809835e-01 -4.86097991e-01
5.80550730e-01 3.66067529e-01 -1.29605740e-01 -9.47179869e-02
-6.70580268e-01 -5.98316193e-01 5.93889475e-01 6.35350823e-01
2.87231833e-01 -1.29535004e-01 -3.91330153e-01 3.21873218e-01
-5.02291098e-02 1.91697180e-01 -1.77698955e-01 8.78261924e-01
-5.56318939e-01 -1.10753107e+00 -6.42294765e-01 2.63214529e-01
-5.04373610e-01 1.70715436e-01 1.62474573e-01 5.65246820e-01
3.16872030e-01 8.90102983e-01 1.44729286e-01 -6.71672881e-01
2.16446608e-01 -2.76098162e-01 7.71067739e-01 -1.58925697e-01
-8.29090953e-01 4.16583091e-01 -1.73434258e-01 -8.48234475e-01
-5.44434726e-01 -7.19044387e-01 -9.14234877e-01 -2.53000319e-01
-5.20769656e-01 3.57752860e-01 8.94605815e-01 8.14532816e-01
6.50915921e-01 6.39441967e-01 6.54717028e-01 -9.47266936e-01
-5.82718074e-01 -8.70966494e-01 -3.24033082e-01 2.46767998e-01
3.31747800e-01 -1.16956508e+00 -1.40097886e-01 1.56918973e-01] | [6.316038608551025, -2.1330602169036865] |
56ba78e7-c77b-4c16-9dc9-a88c037f749e | amodal-instance-segmentation | 1604.08202 | null | http://arxiv.org/abs/1604.08202v2 | http://arxiv.org/pdf/1604.08202v2.pdf | Amodal Instance Segmentation | We consider the problem of amodal instance segmentation, the objective of
which is to predict the region encompassing both visible and occluded parts of
each object. Thus far, the lack of publicly available amodal segmentation
annotations has stymied the development of amodal segmentation methods. In this
paper, we sidestep this issue by relying solely on standard modal instance
segmentation annotations to train our model. The result is a new method for
amodal instance segmentation, which represents the first such method to the
best of our knowledge. We demonstrate the proposed method's effectiveness both
qualitatively and quantitatively. | ['Ke Li', 'Jitendra Malik'] | 2016-04-27 | null | null | null | null | ['amodal-instance-segmentation'] | ['computer-vision'] | [ 3.21456909e-01 7.88672149e-01 -2.79995024e-01 -3.49562973e-01
-9.84693587e-01 -8.95457923e-01 7.24517047e-01 4.57029603e-02
-2.04671964e-01 6.84133947e-01 -2.57353753e-01 -3.74872178e-01
1.38878584e-01 -3.93458754e-01 -7.82126784e-01 -5.99534452e-01
3.83725286e-01 7.12059200e-01 2.37989441e-01 1.57383904e-01
4.14451152e-01 4.01400328e-01 -1.50628030e+00 3.77951831e-01
1.09535778e+00 1.12087095e+00 1.49189219e-01 4.44282532e-01
-3.26567441e-01 3.87742430e-01 -4.59928066e-01 -5.06164312e-01
1.07821181e-01 -3.78986418e-01 -1.43792152e+00 5.04255295e-01
6.92398906e-01 -6.69195578e-02 -1.02355152e-01 9.74819422e-01
1.14761889e-01 5.01896590e-02 9.43013012e-01 -1.29312050e+00
-6.48583829e-01 7.11544693e-01 -7.45821416e-01 1.50427684e-01
1.89584777e-01 2.00224236e-01 1.32294822e+00 -8.76272500e-01
8.58578503e-01 8.91075969e-01 3.88112962e-01 6.66837394e-01
-1.55174470e+00 1.01154111e-02 5.14464676e-01 -1.23361610e-02
-1.10677397e+00 -3.53317529e-01 9.50090170e-01 -6.22844636e-01
7.51226187e-01 2.77439773e-01 6.96402609e-01 8.10588360e-01
-3.62523407e-01 1.42747748e+00 1.35102177e+00 -5.04851222e-01
2.54672796e-01 2.59090960e-01 5.01610935e-01 4.84989017e-01
1.34837806e-01 -2.44910225e-01 3.59565485e-03 1.03486061e-01
5.11141300e-01 -3.98650467e-01 -2.91942924e-01 -4.47092295e-01
-1.09488618e+00 6.25594497e-01 4.84555006e-01 3.83369178e-01
-2.55762070e-01 1.19023852e-01 2.32826516e-01 -6.95098415e-02
5.59713960e-01 3.90947133e-01 -4.30685639e-01 2.46038772e-02
-1.11845279e+00 1.74061149e-01 7.55146623e-01 9.35592413e-01
7.48034656e-01 -1.13777801e-01 -4.83468659e-02 7.62309015e-01
3.47138405e-01 4.33569700e-01 2.25581303e-01 -9.41256225e-01
3.82301718e-01 7.49943614e-01 6.67366385e-02 -6.05714083e-01
-3.60899478e-01 8.56217220e-02 -2.35463619e-01 1.75592631e-01
9.10763323e-01 -3.19283940e-02 -1.35351872e+00 1.52451193e+00
3.07276338e-01 9.49367285e-02 4.81509790e-02 1.04817986e+00
9.51566339e-01 8.04922760e-01 1.48290306e-01 1.69420801e-02
1.27191973e+00 -9.52695906e-01 -6.29526913e-01 -4.24677372e-01
4.39389527e-01 -7.49497771e-01 1.17613256e+00 3.44043523e-01
-1.25115156e+00 -3.36775422e-01 -8.55541766e-01 -3.05489413e-02
-5.80563128e-01 1.05722688e-01 8.23429346e-01 5.26041627e-01
-9.15430307e-01 2.01063603e-01 -8.58000934e-01 -4.58536386e-01
5.95300376e-01 2.60378689e-01 -4.56857741e-01 3.73337083e-02
-9.28329229e-01 8.42089772e-01 6.52282059e-01 3.35731715e-01
-6.37958825e-01 -4.09526974e-01 -8.89880896e-01 -2.11145550e-01
6.06805921e-01 -6.15937352e-01 1.49315858e+00 -1.48824430e+00
-1.22262287e+00 1.12613153e+00 -1.98026642e-01 -3.66845101e-01
3.51610154e-01 -6.22144192e-02 -1.63858328e-02 2.52580971e-01
-4.53716144e-03 1.05126309e+00 7.58537114e-01 -1.88370955e+00
-5.56114018e-01 -5.37523031e-01 3.94070745e-01 1.93253800e-01
9.25972611e-02 -3.34722042e-01 -7.77375579e-01 -6.54604077e-01
2.55798697e-01 -8.00225914e-01 -3.37342292e-01 -2.98839509e-01
-9.47044194e-01 -3.22465569e-01 9.80808079e-01 -7.57884622e-01
9.62618589e-01 -2.07371068e+00 4.17628288e-01 1.49256587e-01
1.98871881e-01 -2.60222256e-02 -1.94923696e-03 2.84817487e-01
6.19687662e-02 2.69580603e-01 -7.83181369e-01 -5.03908038e-01
1.87659815e-01 3.54518414e-01 -5.00120401e-01 4.76877183e-01
3.08583707e-01 1.16684949e+00 -5.91496527e-01 -8.99537206e-01
2.30554432e-01 4.58047062e-01 -2.29622304e-01 2.73960382e-01
-7.42977440e-01 4.40846056e-01 -4.35472608e-01 1.00510526e+00
6.05922341e-01 -2.88022608e-01 -1.04756214e-01 -1.34789929e-01
-1.47931829e-01 -3.85042801e-02 -1.04364979e+00 1.80639803e+00
-3.39406775e-03 7.26476014e-01 3.84816639e-02 -1.36625350e+00
4.41233963e-01 3.49931180e-01 6.32338047e-01 -5.34154594e-01
1.64803013e-01 3.02778125e-01 -1.97794005e-01 -5.38834035e-01
5.24905860e-01 -3.78734648e-01 -1.92525655e-01 4.04372066e-01
1.53799817e-01 -2.54426003e-01 3.98724020e-01 2.09686995e-01
4.52374101e-01 4.11169022e-01 2.99421400e-01 -2.09390998e-01
3.99082810e-01 4.83046353e-01 1.64732963e-01 8.04193795e-01
-3.53797913e-01 8.78904879e-01 6.50043309e-01 -1.34414390e-01
-9.47227001e-01 -1.19032562e+00 -4.08983767e-01 8.17559063e-01
3.44659716e-01 9.69972908e-02 -1.11245358e+00 -9.47359264e-01
-1.45388385e-02 6.75442159e-01 -7.33178437e-01 4.66315627e-01
-5.02223670e-01 -5.40985405e-01 3.40548635e-01 6.74910307e-01
2.32264429e-01 -1.28722048e+00 -4.94551778e-01 -1.42652482e-01
-3.43187362e-01 -1.19082785e+00 -1.89029396e-01 9.24603045e-02
-1.01634264e+00 -1.09743416e+00 -1.01973140e+00 -8.81723166e-01
8.14366162e-01 -1.62384629e-01 1.22174203e+00 -4.57143635e-02
-3.25973392e-01 6.33224905e-01 -2.90330827e-01 -2.46663168e-01
-5.70388734e-02 1.73639789e-01 -6.44725025e-01 -7.05483258e-02
7.37550929e-02 -1.52537227e-01 -5.52949190e-01 5.76394089e-02
-1.09162509e+00 2.57664144e-01 6.27301455e-01 5.88008761e-01
1.03422368e+00 -2.78010905e-01 5.77040851e-01 -1.31095374e+00
3.39287400e-01 -3.55576098e-01 -5.57116687e-01 3.16824555e-01
-4.29846227e-01 -1.62637204e-01 4.47319388e-01 -3.38858128e-01
-1.14876223e+00 3.92085731e-01 -1.11520916e-01 -1.55644625e-01
-6.58796132e-01 6.17983341e-01 -2.23231345e-01 1.48887128e-01
3.49005550e-01 2.53138900e-01 -2.51766503e-01 -4.85323846e-01
8.34918559e-01 4.74388719e-01 7.19044924e-01 -6.44761622e-01
4.61478770e-01 7.09452331e-01 -2.74014622e-01 -9.28433418e-01
-8.34096670e-01 -4.16693389e-01 -8.27423513e-01 -3.04729104e-01
1.33024240e+00 -4.89361107e-01 -3.09826195e-01 3.85351956e-01
-1.25607419e+00 -4.45177138e-01 -2.74234444e-01 1.49937525e-01
-8.83185804e-01 4.67519492e-01 -6.14577174e-01 -8.90934825e-01
3.72644737e-02 -1.27756178e+00 1.27857018e+00 2.09451795e-01
-5.25071584e-02 -9.98664439e-01 7.27137551e-02 6.14544868e-01
-1.10189445e-01 4.17208999e-01 6.87298000e-01 -6.60316229e-01
-7.76748717e-01 -3.40381175e-01 -2.77526796e-01 2.08261326e-01
-1.28924266e-01 8.43702629e-02 -1.29000020e+00 -2.71485355e-02
-3.94946814e-01 -3.74061197e-01 9.98989165e-01 7.88409293e-01
1.34222806e+00 1.22412862e-02 -5.23999453e-01 2.77368993e-01
1.45540524e+00 5.33629023e-02 5.83910584e-01 1.53590098e-01
7.74709761e-01 7.24575877e-01 7.66984344e-01 1.17639735e-01
2.67538726e-01 6.54843867e-01 6.05424166e-01 -4.06378925e-01
-1.59355327e-01 4.86003086e-02 -9.78669524e-02 1.91235274e-01
-8.48169625e-02 -5.36812007e-01 -1.17711473e+00 8.18938315e-01
-1.89264929e+00 -7.17116892e-01 -1.34056479e-01 1.83418548e+00
5.13003945e-01 1.70683712e-01 2.67013937e-01 1.58604860e-01
6.09849811e-01 1.44206256e-01 -4.01287436e-01 -3.84721100e-01
1.41985238e-01 -5.01666032e-02 1.73835948e-01 6.51151717e-01
-1.38669932e+00 1.04018891e+00 6.71409369e+00 5.48506737e-01
-7.69824445e-01 -2.37938583e-01 8.00185382e-01 3.28255773e-01
-2.62654185e-01 1.17417410e-01 -6.26528502e-01 4.53995347e-01
3.92702520e-01 2.45203912e-01 2.29290903e-01 9.79866266e-01
-1.24557503e-01 -2.19749287e-01 -1.22892344e+00 7.04328418e-01
1.96885645e-01 -1.07403684e+00 -1.11814644e-02 -4.15554345e-02
6.64914191e-01 -3.65269840e-01 2.66713947e-01 2.29316726e-01
-5.54545075e-02 -1.21250451e+00 8.63426685e-01 4.58153248e-01
3.83215874e-01 -6.38668001e-01 6.17739856e-01 3.22711051e-01
-1.19047189e+00 1.46100327e-01 4.31448631e-02 2.24026665e-01
5.50389349e-01 4.77477372e-01 -6.87215745e-01 4.94340658e-01
4.73393142e-01 6.17573619e-01 -4.09478426e-01 1.15711975e+00
-2.13095799e-01 7.28912890e-01 -2.40231901e-01 4.34228152e-01
3.06552470e-01 -3.30327749e-02 5.14233589e-01 1.32350028e+00
1.02285005e-01 1.28015801e-01 4.57254618e-01 1.17500961e+00
4.38220613e-02 1.53927565e-01 -7.52730072e-01 -4.12182868e-01
1.15629852e-01 1.20956945e+00 -1.23352444e+00 -3.73073220e-01
-7.03033566e-01 1.00637734e+00 3.98790926e-01 6.43373013e-01
-8.83448601e-01 -2.47954577e-01 1.46924511e-01 5.34557104e-02
3.54552388e-01 -2.09685713e-02 -8.94883096e-01 -1.02851808e+00
-2.06709970e-02 -5.87444782e-01 6.81786954e-01 -7.93645263e-01
-1.52514648e+00 4.93323565e-01 3.11487287e-01 -9.33335781e-01
-1.20303959e-01 -7.92778313e-01 -5.94252229e-01 7.68326879e-01
-1.33575332e+00 -1.39338720e+00 -1.59953877e-01 2.43076041e-01
6.47392392e-01 1.76451221e-01 4.66819704e-01 9.00636092e-02
-6.87949657e-01 1.58122644e-01 -3.02480072e-01 -7.08647678e-03
3.71638209e-01 -1.66381574e+00 1.78651974e-01 8.65177631e-01
2.47819334e-01 3.81923914e-01 9.19050276e-01 -5.89270353e-01
-1.07439983e+00 -6.00312352e-01 8.11054826e-01 -6.74142361e-01
6.41354144e-01 -2.21093848e-01 -9.11385834e-01 9.41142023e-01
2.34562188e-01 9.73794237e-02 4.33760434e-01 1.83932796e-01
-2.31795087e-01 1.76006347e-01 -1.07077360e+00 5.62357068e-01
6.05986118e-01 -3.52936238e-01 -8.10155630e-01 1.06264487e-01
6.06073260e-01 -5.41700661e-01 -8.05188715e-01 5.90590715e-01
2.86224902e-01 -8.83104503e-01 1.08433795e+00 -3.86906117e-01
5.42636871e-01 -2.15872169e-01 -7.31114522e-02 -1.09792125e+00
2.29177028e-01 -3.41989025e-02 -3.94416213e-01 1.25409567e+00
7.66455531e-01 -5.28829336e-01 9.75084364e-01 6.85963750e-01
-4.57847536e-01 -9.28814411e-01 -9.25149858e-01 -5.19761801e-01
2.50054359e-01 -5.66969216e-01 3.36306393e-01 8.59575748e-01
9.62725580e-02 1.48646265e-01 -8.19195807e-03 2.71576643e-01
5.51667809e-01 6.57922029e-01 6.35672569e-01 -1.12466681e+00
-3.79494756e-01 -6.36043906e-01 -3.21333408e-01 -1.26856530e+00
3.76807570e-01 -8.68361115e-01 3.39952767e-01 -1.93456638e+00
4.40787852e-01 -2.60923803e-01 -1.44921064e-01 6.17152691e-01
-1.65267497e-01 5.05407572e-01 2.58524150e-01 9.35619026e-02
-6.68359041e-01 1.13609850e-01 1.34049320e+00 -2.67491519e-01
-2.32185528e-01 -3.35946120e-02 -7.38613904e-01 6.78730786e-01
1.06521535e+00 -1.69047460e-01 -6.01995468e-01 -2.68776745e-01
-1.22924723e-01 -4.02145311e-02 5.26723266e-01 -4.75037158e-01
-1.64346680e-01 -3.03903043e-01 2.56944567e-01 -8.42472017e-01
4.76337880e-01 -9.18236375e-01 -6.23049885e-02 1.84883982e-01
-2.73148745e-01 -1.34870335e-01 4.24711049e-01 6.34564221e-01
-4.23575699e-01 -2.46218696e-01 7.82286704e-01 -2.22847134e-01
-1.02601898e+00 1.53515220e-01 -2.80126005e-01 2.91970998e-01
1.32213163e+00 -5.25284231e-01 -5.62911332e-01 -1.88331768e-01
-1.02381718e+00 1.84659749e-01 5.61270595e-01 1.60068840e-01
5.77524483e-01 -1.05534256e+00 -3.96155387e-01 -9.20956731e-02
2.10165858e-01 2.16104195e-01 2.59220093e-01 1.12702954e+00
-3.12432408e-01 5.39598882e-01 6.40836284e-02 -7.88078487e-01
-1.22568882e+00 7.03575432e-01 3.91042918e-01 -3.63424644e-02
-8.19343090e-01 7.15904951e-01 5.54718912e-01 -3.92451674e-01
2.82549828e-01 -2.86306351e-01 -3.76938522e-01 7.29668140e-02
1.29873306e-01 1.76499873e-01 -1.79367974e-01 -8.97287607e-01
-2.90807337e-01 4.78027046e-01 -1.17012613e-01 -2.99925506e-01
9.81976092e-01 -1.13754742e-01 -1.14343375e-01 7.58697510e-01
8.33201945e-01 -3.23605597e-01 -1.22624791e+00 2.01861024e-01
1.17592968e-01 -3.92021954e-01 -1.00281417e-01 -9.21756983e-01
-1.10949039e+00 8.76005113e-01 4.77452070e-01 5.38157940e-01
1.04016137e+00 5.09731472e-01 7.90556371e-01 2.19619453e-01
4.47423048e-02 -1.21989882e+00 1.66100532e-01 2.53337204e-01
6.70553029e-01 -1.48041761e+00 -1.70833856e-01 -6.79419518e-01
-8.98779035e-01 7.63560295e-01 8.38476837e-01 2.06744317e-02
5.03359437e-01 6.78790584e-02 2.41410792e-01 -3.78152966e-01
-2.74321735e-01 -5.41283488e-01 8.18975449e-01 6.29581571e-01
6.00290596e-01 1.02920204e-01 -4.11414057e-01 4.37748641e-01
2.20013991e-01 -1.46832824e-01 1.91922858e-01 1.04838407e+00
-6.08713627e-01 -9.65150416e-01 -4.21686471e-01 3.60251039e-01
-4.48028952e-01 3.57600711e-02 -9.15656626e-01 1.18156815e+00
6.98683262e-02 7.37704396e-01 -7.41318762e-02 6.01223521e-02
3.55657399e-01 2.99849868e-01 6.37894750e-01 -4.62888986e-01
-2.34613016e-01 1.26654789e-01 -4.80870828e-02 -2.73357511e-01
-5.08462012e-01 -6.08412802e-01 -1.53912103e+00 1.52165219e-01
-2.61854142e-01 -8.64846334e-02 5.66606224e-01 9.80727613e-01
-9.98118296e-02 2.69553006e-01 1.43279403e-01 -1.15920365e+00
-1.85521409e-01 -6.68533087e-01 -6.27604961e-01 6.36017382e-01
2.94255018e-01 -6.21522486e-01 -3.10886383e-01 2.61538535e-01] | [9.590301513671875, 0.4930320978164673] |
c5e55c5d-10ed-4965-8587-b970dbefd1b1 | constraint-based-sequential-pattern-mining | 1811.06086 | null | http://arxiv.org/abs/1811.06086v1 | http://arxiv.org/pdf/1811.06086v1.pdf | Constraint-based Sequential Pattern Mining with Decision Diagrams | Constrained sequential pattern mining aims at identifying frequent patterns
on a sequential database of items while observing constraints defined over the
item attributes. We introduce novel techniques for constraint-based sequential
pattern mining that rely on a multi-valued decision diagram representation of
the database. Specifically, our representation can accommodate multiple item
attributes and various constraint types, including a number of non-monotone
constraints. To evaluate the applicability of our approach, we develop an
MDD-based prefix-projection algorithm and compare its performance against a
typical generate-and-check variant, as well as a state-of-the-art
constraint-based sequential pattern mining algorithm. Results show that our
approach is competitive with or superior to these other methods in terms of
scalability and efficiency. | ['Willem-Jan van Hoeve', 'Amin Hosseininasab', 'Andre A. Cire'] | 2018-11-14 | null | null | null | null | ['sequential-pattern-mining'] | ['natural-language-processing'] | [ 4.03307229e-01 -2.42405951e-01 -6.51592731e-01 -4.47695851e-01
1.95230782e-01 -4.57215667e-01 3.05771500e-01 4.79161263e-01
-1.84569150e-01 6.21509492e-01 -2.62205005e-01 -6.43956065e-01
-5.10576546e-01 -1.16893566e+00 -2.03074872e-01 -1.89925402e-01
-5.36982536e-01 9.62040961e-01 7.76821852e-01 -1.86825711e-02
2.44921386e-01 6.08713031e-01 -1.81921434e+00 8.22144508e-01
5.27830005e-01 1.18872333e+00 -1.78484544e-01 7.09296614e-02
-2.68780291e-01 4.90677774e-01 -2.68294305e-01 -2.95434475e-01
4.25618082e-01 -5.08334339e-01 -6.98548317e-01 2.57534921e-01
-3.40414912e-01 1.31786063e-01 2.04681084e-01 7.55180120e-01
-3.58817309e-01 -2.12693863e-04 2.86080152e-01 -1.42137778e+00
3.63412611e-02 5.67760289e-01 -7.71849394e-01 3.03181052e-01
9.25985575e-01 -3.88775200e-01 1.10718620e+00 -1.02804756e+00
8.66476595e-01 8.31309795e-01 4.64073509e-01 2.82329507e-02
-1.37460065e+00 -5.86191952e-01 5.69896400e-01 5.62158585e-01
-1.68450546e+00 -1.50794789e-01 4.67918605e-01 -3.23471688e-02
1.39042282e+00 5.60927451e-01 9.52001512e-01 2.86107510e-01
1.33615106e-01 6.64351165e-01 8.98068130e-01 -8.95652235e-01
5.14856696e-01 1.20251931e-01 5.13317704e-01 5.48206568e-01
9.33387220e-01 2.08233640e-01 -7.03979909e-01 -8.01927924e-01
1.14560217e-01 1.76985577e-01 9.10977945e-02 -7.25798965e-01
-9.88241971e-01 7.63893425e-01 -6.05899394e-01 2.78561622e-01
-4.11914617e-01 -5.26580393e-01 3.60923886e-01 6.99154019e-01
-1.04554877e-01 1.10439010e-01 -6.20768011e-01 1.28310323e-01
-9.45026755e-01 6.80640101e-01 1.23728120e+00 1.35612190e+00
6.19759142e-01 -1.94720224e-01 -2.19987556e-02 4.00295675e-01
2.83531338e-01 2.26718113e-01 7.94917420e-02 -4.09331948e-01
2.11081147e-01 1.12236488e+00 1.35385960e-01 -8.63935232e-01
-4.79605794e-01 -7.12653771e-02 -5.55195510e-01 1.09952979e-01
2.46757120e-01 1.36177972e-01 -5.40560186e-01 1.29360414e+00
5.06380916e-01 -2.21830577e-01 -7.36339539e-02 5.33759832e-01
1.55964466e-02 4.10969764e-01 -2.67469168e-01 -1.16867507e+00
1.25318658e+00 -3.92247349e-01 -7.33576357e-01 1.30496398e-01
4.40970421e-01 -4.47621584e-01 4.75939453e-01 1.07846594e+00
-1.03130352e+00 -1.57409981e-01 -1.23851240e+00 7.61353135e-01
-1.91416308e-01 -3.59662086e-01 6.72410846e-01 9.34673548e-01
-3.95579755e-01 2.53021300e-01 -8.81503403e-01 -3.03646624e-01
9.53173861e-02 6.24115705e-01 -2.70827293e-01 -3.20361942e-01
-8.00656438e-01 7.40780950e-01 7.73273468e-01 -5.31760633e-01
-3.96933481e-02 -5.94471157e-01 -6.84113026e-01 1.17579237e-01
8.55878651e-01 -4.90594953e-01 1.07004797e+00 -5.69642901e-01
-8.39839697e-01 6.13378108e-01 -5.26752591e-01 -5.82862973e-01
1.41920984e-01 2.36065567e-01 -1.20902717e+00 -5.90336556e-03
-7.41095766e-02 -3.46803963e-01 2.48775005e-01 -7.97884166e-01
-1.22182381e+00 -3.67304653e-01 -1.47351503e-01 -1.61307976e-01
-1.32053584e-01 4.31492835e-01 -4.13672507e-01 -4.85507220e-01
3.85082841e-01 -8.25820863e-01 -6.36252224e-01 -1.74149349e-01
-4.13506806e-01 -9.13303494e-02 8.53446484e-01 3.96345779e-02
2.11786175e+00 -1.94783401e+00 -2.00893402e-01 1.28410888e+00
-1.88300177e-01 -1.07848421e-01 3.56871605e-01 8.78805935e-01
-6.87589794e-02 2.93512139e-02 -4.22810286e-01 3.58326524e-01
-1.95984289e-01 4.72357750e-01 -1.55870304e-01 4.73729968e-01
1.14735983e-01 3.63073885e-01 -4.93420839e-01 -4.97140139e-01
-1.15067000e-02 -4.38154608e-01 -6.85701072e-01 -6.37556985e-02
-5.14548182e-01 -2.66732633e-01 -3.20971340e-01 1.04425204e+00
7.32406676e-01 -3.27490121e-02 1.23301554e+00 3.72786343e-01
-4.59536970e-01 2.85938382e-01 -2.01215601e+00 1.11474228e+00
9.18028876e-02 -2.74229556e-01 -1.98889256e-01 -1.05320811e+00
1.03809893e+00 2.10685536e-01 7.37025678e-01 -8.42870712e-01
-3.22222292e-01 3.69963109e-01 3.13186288e-01 -3.11664701e-01
2.32841805e-01 -2.21632972e-01 -1.93679646e-01 9.09176230e-01
-5.34319162e-01 5.96521735e-01 7.41486251e-01 1.65707036e-03
1.42614293e+00 -4.68813866e-01 1.18603563e+00 -5.93235016e-01
7.56584048e-01 3.47564816e-01 1.11093092e+00 6.87206030e-01
4.65138465e-01 1.36660606e-01 6.10603988e-01 -8.96993160e-01
-6.88140810e-01 -7.56108403e-01 -3.59011322e-01 7.47774005e-01
1.72952637e-02 -8.30268562e-01 5.12855947e-02 -6.56668365e-01
6.41628623e-01 6.52318060e-01 -4.60017592e-01 2.51231015e-01
-6.98107243e-01 -8.89749527e-01 5.78503422e-02 6.55504048e-01
2.33647996e-03 -8.92878413e-01 -1.13040018e+00 6.02187395e-01
3.72104406e-01 -8.86633873e-01 -1.92418024e-01 4.64387596e-01
-6.97109461e-01 -1.58407819e+00 4.59006518e-01 -6.64448380e-01
5.48563838e-01 2.61883736e-01 1.12585056e+00 2.33930826e-01
-2.88782507e-01 -1.04897760e-01 -7.23698020e-01 -8.01634192e-01
-2.28133500e-01 -1.11847684e-01 4.37174350e-01 2.49668539e-01
9.48129833e-01 -6.38763845e-01 -1.81940600e-01 6.96350336e-01
-9.57958043e-01 -3.61745268e-01 4.86929327e-01 5.95953465e-01
9.96198475e-01 5.64954758e-01 5.52875757e-01 -1.36648738e+00
6.05139613e-01 -7.15541720e-01 -7.77037382e-01 5.48982203e-01
-1.38936663e+00 -4.49684858e-02 4.95947003e-01 -4.80582505e-01
-9.07812119e-01 5.14356613e-01 2.02816516e-01 -1.24212183e-01
-1.07132047e-01 1.01076567e+00 -4.83079255e-01 2.29281753e-01
4.15031463e-01 2.59605795e-01 -9.69377682e-02 -3.93355966e-01
-5.88368438e-02 5.56234777e-01 2.96738982e-01 -6.05116427e-01
7.01206625e-01 5.48123538e-01 3.44656855e-01 -5.20358860e-01
-1.80331036e-01 -8.01387191e-01 -6.25445724e-01 1.58626005e-01
9.78741422e-02 -3.86063337e-01 -7.54940689e-01 6.73138052e-02
-8.07309687e-01 2.08044499e-01 -4.69954133e-01 2.98393369e-01
-3.66129160e-01 4.05473411e-01 -3.30016524e-01 -9.85225022e-01
-1.47316560e-01 -4.79468763e-01 4.55087647e-02 -3.04078370e-01
-7.49403954e-01 -4.02050823e-01 3.89189303e-01 -4.10930157e-01
1.27772033e-01 4.70074594e-01 1.32277215e+00 -1.31534958e+00
-2.99541205e-01 -3.85063320e-01 9.70783904e-02 -3.04644167e-01
1.96004182e-01 5.90641089e-02 4.97032925e-02 -3.24185371e-01
5.85353561e-03 2.94202641e-02 2.67425746e-01 2.19557658e-02
1.13724482e+00 -4.46247369e-01 -7.83934116e-01 2.73766249e-01
1.60252595e+00 9.00115252e-01 4.81901824e-01 5.21903038e-01
4.90284292e-03 6.48832560e-01 1.16691315e+00 1.12809169e+00
2.24916726e-01 8.69903684e-01 1.48509696e-01 2.36755371e-01
7.00158238e-01 7.37717724e-04 -2.03273147e-01 2.17275083e-01
1.44013762e-01 -1.57658771e-01 -9.81704712e-01 6.45540774e-01
-2.07500482e+00 -1.20528543e+00 -5.60101211e-01 2.19116879e+00
7.54430771e-01 5.37103474e-01 7.82772779e-01 1.07268763e+00
5.64214051e-01 -3.66483212e-01 -4.75370228e-01 -1.10993171e+00
-1.50400281e-01 5.55243134e-01 2.72960752e-01 2.48697281e-01
-8.23098958e-01 3.38228911e-01 7.45447302e+00 2.94628590e-01
-5.63293517e-01 -2.84870088e-01 -5.57190441e-02 -2.50685930e-01
-5.09363830e-01 1.14401855e-01 -8.65870535e-01 2.53364205e-01
1.04064274e+00 -6.07232988e-01 6.31149858e-02 9.84002948e-01
3.06592062e-02 -2.59068489e-01 -1.41803694e+00 6.11409724e-01
-4.03633304e-02 -1.57361627e+00 1.11415863e-01 4.39155996e-01
6.24766409e-01 -5.22793889e-01 -4.53534573e-01 -1.54643118e-01
9.17246342e-02 -7.43132055e-01 7.96068668e-01 2.49572590e-01
7.57100940e-01 -1.12306988e+00 6.23741150e-01 4.38005120e-01
-1.51117277e+00 -5.11752844e-01 -2.52009351e-02 -5.08585334e-01
2.96256453e-01 8.56757045e-01 -5.30255318e-01 7.82006025e-01
8.28134716e-01 3.31748605e-01 -7.34544126e-03 1.03826928e+00
1.06752694e-01 4.41897452e-01 -6.80552006e-01 1.24656092e-02
-1.13075465e-01 -1.15261272e-01 3.10709238e-01 1.49227762e+00
1.40504166e-01 4.99688566e-01 5.49703419e-01 4.96154785e-01
3.81320477e-01 2.39414632e-01 -6.64287090e-01 2.93753207e-01
9.39598083e-01 6.17075920e-01 -7.87165940e-01 -3.14713985e-01
-1.04626358e+00 2.67247111e-01 -9.82189849e-02 -1.07333265e-01
-3.82665336e-01 -3.21401089e-01 6.25857115e-01 6.52517557e-01
9.06301498e-01 -2.19856679e-01 -6.71945572e-01 -8.13162208e-01
5.97484708e-01 -1.35441637e+00 1.14663959e+00 2.67295986e-01
-1.04890490e+00 4.31137025e-01 4.64813709e-01 -1.34081078e+00
-4.55259860e-01 -3.48504663e-01 -6.77035630e-01 6.68042004e-01
-1.18044746e+00 -8.14959645e-01 2.24732876e-01 9.54752862e-01
2.83814490e-01 -2.87103623e-01 9.29641843e-01 6.57390431e-02
-4.75815803e-01 6.96818411e-01 -3.07295442e-01 -3.74099404e-01
9.60127562e-02 -8.68897200e-01 3.81523281e-01 1.13276291e+00
2.34346002e-01 8.11681271e-01 6.54692113e-01 -8.12133789e-01
-1.61817443e+00 -8.11669886e-01 1.09611130e+00 -6.39979988e-02
5.17032981e-01 -3.10067117e-01 -9.56739247e-01 7.49675512e-01
-1.46765307e-01 2.49175467e-02 1.27392995e+00 4.91887927e-01
-4.08702731e-01 -2.28719443e-01 -1.37496984e+00 2.10461885e-01
1.33640444e+00 4.82472815e-02 -7.92960048e-01 -1.96319818e-01
2.40523875e-01 -6.73468634e-02 -8.87441993e-01 7.49902904e-01
1.04348278e+00 -1.11714470e+00 5.76764047e-01 -7.93260574e-01
-6.87189284e-04 -5.79788148e-01 -4.54742759e-01 -5.36952376e-01
-7.46354759e-01 -4.85379457e-01 -6.53845429e-01 7.42625117e-01
9.10770655e-01 -5.88698387e-01 1.06293535e+00 6.52053535e-01
1.92413740e-02 -1.17708993e+00 -1.00157976e+00 -1.14292538e+00
-4.75942522e-01 -6.21920764e-01 9.94456887e-01 1.03388655e+00
8.01580131e-01 2.81595830e-02 -3.65495563e-01 9.81152058e-02
5.11581361e-01 9.50392187e-01 5.20323336e-01 -1.57776439e+00
-6.68619692e-01 -1.84260845e-01 -3.34098756e-01 -4.15179193e-01
-2.06002921e-01 -7.72029102e-01 -5.58352590e-01 -1.13101113e+00
1.32278755e-01 -6.51206553e-01 -4.50418085e-01 7.14043200e-01
3.50134552e-01 -1.28500536e-01 -1.16413355e-01 3.19079161e-01
-6.09119475e-01 -1.81381866e-01 4.18328375e-01 4.18865122e-02
-6.06408954e-01 2.76721895e-01 -7.15075731e-01 4.58444983e-01
6.03749394e-01 -7.15532064e-01 -4.62573528e-01 1.61182404e-01
4.85864133e-01 3.55226159e-01 -5.10114908e-01 -7.70630717e-01
4.46686894e-01 -8.58785331e-01 1.95523024e-01 -9.57522750e-01
-2.83263475e-01 -1.07190835e+00 8.86521459e-01 1.04594886e+00
-1.14769064e-01 4.30457324e-01 1.52685687e-01 6.35423303e-01
-2.70069778e-01 -1.32480308e-01 5.44247568e-01 -6.12887228e-03
-7.39117384e-01 2.01769531e-01 -8.15369368e-01 -3.84776115e-01
1.48277640e+00 -7.25968540e-01 1.04937591e-01 2.48699024e-01
-6.94963038e-01 3.27964634e-01 5.51893055e-01 2.58925885e-01
8.33876848e-01 -1.17879820e+00 -5.43947518e-01 7.10717499e-01
6.57978892e-01 -1.09241396e-01 -3.38740468e-01 6.19169593e-01
-3.38550568e-01 6.31701708e-01 -3.96014273e-01 -4.15761441e-01
-1.49299812e+00 1.10704708e+00 -2.62701780e-01 -7.14439273e-01
-6.23477221e-01 2.19409078e-01 -4.27608103e-01 1.27194189e-02
1.06063478e-01 -3.77545416e-01 6.12127483e-02 9.96869355e-02
8.34485590e-01 4.51657861e-01 4.27187681e-01 8.94024000e-02
-1.06318200e+00 1.75865725e-01 -2.08604127e-01 -8.92747045e-02
1.11760664e+00 1.25114799e-01 -2.53136635e-01 3.73221517e-01
4.48704749e-01 9.24496204e-02 -3.76630425e-01 -4.80384141e-01
7.74699509e-01 -5.50272465e-01 -6.03057206e-01 -7.82503963e-01
-7.62115061e-01 -2.46826589e-01 9.74687412e-02 4.71119285e-01
1.42980468e+00 -1.14235744e-01 4.96771306e-01 6.34912491e-01
9.86223936e-01 -1.06341505e+00 -4.33239520e-01 2.61889637e-01
6.82084084e-01 -7.66865671e-01 4.71165806e-01 -8.70494187e-01
-6.22451007e-01 9.94028926e-01 4.09462124e-01 -1.20104335e-01
1.05722141e+00 9.14110899e-01 -3.73836160e-01 -1.17448717e-01
-1.38984728e+00 -2.81511694e-01 1.93383619e-02 5.31891108e-01
4.22875397e-02 1.54015243e-01 -1.08640325e+00 8.35010767e-01
-2.07037274e-02 1.31086960e-01 5.23867190e-01 1.67552125e+00
-4.72099930e-01 -1.80087411e+00 -3.59710127e-01 7.91790485e-01
-4.53944117e-01 1.93124920e-01 -5.74695170e-01 9.74903047e-01
2.47565225e-01 1.09570193e+00 8.46072733e-02 -5.89747846e-01
8.91300261e-01 3.28837246e-01 4.34011698e-01 -8.61543000e-01
-6.82943583e-01 3.27709801e-02 6.13562524e-01 -7.28295445e-01
-4.44153160e-01 -1.21683383e+00 -1.32810163e+00 -2.50997990e-01
-4.21266735e-01 3.18675518e-01 2.22438425e-01 9.38007534e-01
2.05418110e-01 -1.50659084e-01 8.47844005e-01 2.14064762e-01
-3.41544271e-01 -5.21020651e-01 -8.96023452e-01 1.85678542e-01
-1.27479091e-01 -6.98740602e-01 1.17712803e-01 -1.14350788e-01] | [8.30202579498291, 6.306829929351807] |
57aa0028-e395-4656-94ad-a1be4510f7b0 | co-grounding-networks-with-semantic-attention | 2103.12346 | null | https://arxiv.org/abs/2103.12346v1 | https://arxiv.org/pdf/2103.12346v1.pdf | Co-Grounding Networks with Semantic Attention for Referring Expression Comprehension in Videos | In this paper, we address the problem of referring expression comprehension in videos, which is challenging due to complex expression and scene dynamics. Unlike previous methods which solve the problem in multiple stages (i.e., tracking, proposal-based matching), we tackle the problem from a novel perspective, \textbf{co-grounding}, with an elegant one-stage framework. We enhance the single-frame grounding accuracy by semantic attention learning and improve the cross-frame grounding consistency with co-grounding feature learning. Semantic attention learning explicitly parses referring cues in different attributes to reduce the ambiguity in the complex expression. Co-grounding feature learning boosts visual feature representations by integrating temporal correlation to reduce the ambiguity caused by scene dynamics. Experiment results demonstrate the superiority of our framework on the video grounding datasets VID and LiOTB in generating accurate and stable results across frames. Our model is also applicable to referring expression comprehension in images, illustrated by the improved performance on the RefCOCO dataset. Our project is available at https://sijiesong.github.io/co-grounding. | ['Shih-Fu Chang', 'Zongming Guo', 'Jiaying Liu', 'Xudong Lin', 'Sijie Song'] | 2021-03-23 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Song_Co-Grounding_Networks_With_Semantic_Attention_for_Referring_Expression_Comprehension_in_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Song_Co-Grounding_Networks_With_Semantic_Attention_for_Referring_Expression_Comprehension_in_CVPR_2021_paper.pdf | cvpr-2021-1 | ['video-grounding'] | ['computer-vision'] | [ 2.46123478e-01 -1.95103213e-01 -2.54028648e-01 -5.11513114e-01
-7.80833781e-01 -3.22820753e-01 2.37205774e-01 -1.42756879e-01
-2.77002603e-01 4.06336457e-01 3.29559982e-01 1.03441730e-01
-1.16740458e-01 -4.62282062e-01 -8.53611052e-01 -3.46376896e-01
1.98953748e-01 -1.72172114e-02 5.75250536e-02 -4.09752548e-01
2.30080500e-01 2.47383639e-01 -1.51666272e+00 5.86520374e-01
7.88019419e-01 1.07149780e+00 3.87573481e-01 4.57489371e-01
-3.36003542e-01 1.07682538e+00 -5.37315786e-01 -4.65762198e-01
6.01773709e-02 -5.78898013e-01 -8.96171749e-01 3.74118716e-01
9.37674522e-01 -2.02702060e-01 -3.14505279e-01 1.17984712e+00
3.99433762e-01 3.18342626e-01 5.51330261e-02 -1.70750916e+00
-7.64566481e-01 1.61744520e-01 -6.84568346e-01 2.69330561e-01
9.03931975e-01 -2.96771526e-03 1.11719251e+00 -1.08598995e+00
8.01795900e-01 1.57596457e+00 5.17415583e-01 5.64400077e-01
-1.02174568e+00 -8.24756145e-01 7.20492482e-01 7.66765535e-01
-1.43080652e+00 -5.54066837e-01 9.04831171e-01 -5.40459633e-01
6.93819404e-01 3.50012891e-02 7.17464268e-01 1.18138361e+00
-1.24616452e-01 1.06256378e+00 1.06231797e+00 -3.40958536e-01
-7.03690052e-02 -4.41315800e-01 1.06559142e-01 9.58802581e-01
-1.63542122e-01 -6.61588982e-02 -1.08613265e+00 2.31646761e-01
8.59630883e-01 1.45850509e-01 -6.42055452e-01 -3.26861501e-01
-1.44748902e+00 5.77799618e-01 5.59998572e-01 3.20384681e-01
-2.94192076e-01 4.40485209e-01 2.99424350e-01 2.01931193e-01
4.82000411e-01 2.75511593e-01 -4.09427822e-01 -1.69257626e-01
-8.68051529e-01 3.86040717e-01 3.84760439e-01 1.36388624e+00
8.01929593e-01 -3.35297473e-02 -4.36416507e-01 7.20528662e-01
1.87749714e-01 4.50059712e-01 1.97235450e-01 -1.42348611e+00
5.47935247e-01 4.80211079e-01 1.22312345e-01 -1.33172560e+00
-4.60158914e-01 -2.17291191e-01 -6.98166966e-01 -2.39706039e-02
5.01857579e-01 1.45386299e-02 -6.63480937e-01 2.10816360e+00
3.74973536e-01 4.01680976e-01 -1.58830538e-01 1.25373638e+00
1.16529167e+00 5.82971334e-01 3.48176569e-01 -1.76946625e-01
1.46771324e+00 -1.43773353e+00 -1.24028492e+00 -1.48469403e-01
6.86119080e-01 -8.39148641e-01 1.11841643e+00 2.02520654e-01
-1.14816606e+00 -8.32724273e-01 -6.74217224e-01 -3.99027556e-01
-9.87046435e-02 1.00798972e-01 7.79515922e-01 -6.88740611e-02
-1.08417940e+00 4.54181075e-01 -6.47562683e-01 -5.94295919e-01
7.16657639e-01 1.70779645e-01 -6.17773950e-01 -2.54777044e-01
-1.14702785e+00 7.60590613e-01 2.08360314e-01 2.24812552e-01
-4.71330911e-01 -7.24162579e-01 -1.04079759e+00 -5.30770011e-02
6.66470587e-01 -1.00184977e+00 1.24728894e+00 -1.52934432e+00
-1.40269399e+00 9.46457028e-01 -5.93440950e-01 -1.15517333e-01
4.72221643e-01 -6.84973419e-01 -2.50036448e-01 4.65699911e-01
3.06105703e-01 9.86985028e-01 7.61916399e-01 -1.06408226e+00
-5.16950965e-01 -2.42662579e-01 4.42185014e-01 3.34740549e-01
7.52625316e-02 2.39213169e-01 -7.61994600e-01 -8.96577239e-01
3.14138174e-01 -8.42879653e-01 -6.75032437e-02 4.41655427e-01
-9.19694453e-02 -3.50745887e-01 7.91385472e-01 -7.63313413e-01
1.15966964e+00 -2.20455551e+00 4.12999421e-01 -3.80499721e-01
2.03585789e-01 4.42299098e-02 -2.13465303e-01 1.85342342e-01
-2.42712498e-01 -3.14749815e-02 -3.70066687e-02 -3.43523800e-01
-1.09801337e-01 3.46110791e-01 -2.25543052e-01 4.76035088e-01
6.05923176e-01 1.15781379e+00 -1.22510672e+00 -8.07290494e-01
9.77277309e-02 3.47440958e-01 -6.00324392e-01 5.39034426e-01
-3.54650944e-01 7.63481319e-01 -5.57491779e-01 9.66500461e-01
5.95219195e-01 -4.14732218e-01 -2.03606218e-01 -6.01433873e-01
1.10966787e-01 -2.45829046e-01 -1.01301181e+00 2.48078513e+00
-3.96266222e-01 6.66752934e-01 4.27718051e-02 -1.08281541e+00
8.87840450e-01 2.03551278e-01 6.42399848e-01 -8.68570209e-01
4.33810577e-02 -2.53184559e-03 -1.50893480e-01 -8.56868982e-01
5.18562913e-01 4.38890085e-02 -1.65143415e-01 1.88849464e-01
3.83615583e-01 -2.35775411e-02 1.49255469e-01 3.81374270e-01
5.78233540e-01 9.39670801e-01 3.92579317e-01 -2.15379566e-01
6.70582771e-01 -6.63069040e-02 7.85263598e-01 5.20279884e-01
-4.65679854e-01 7.66831100e-01 4.31353092e-01 -4.91670758e-01
-6.12345040e-01 -7.77461648e-01 9.04397294e-02 1.41774404e+00
7.08567202e-01 -7.72728324e-01 -7.04500735e-01 -5.72898507e-01
-2.01060608e-01 4.01082516e-01 -6.25911534e-01 -5.48897646e-02
-6.28925562e-01 -2.86601245e-01 1.33722946e-01 7.79025257e-01
7.40893006e-01 -9.47144628e-01 -5.86505234e-01 9.59316567e-02
-8.34061205e-01 -1.47399080e+00 -7.68737674e-01 -2.14694366e-01
-5.74815691e-01 -1.10262525e+00 -7.36642003e-01 -5.71958840e-01
5.62315345e-01 5.43131709e-01 1.31328392e+00 4.06261116e-01
-1.82743475e-01 8.24292660e-01 -5.93280852e-01 -2.01355740e-01
1.93549976e-01 -2.54185855e-01 -2.85174221e-01 2.30435908e-01
3.66904885e-01 -2.58225739e-01 -6.42509401e-01 4.77914155e-01
-6.19083226e-01 4.18985516e-01 2.00824291e-01 1.02847290e+00
8.06770384e-01 -5.85357428e-01 4.30860698e-01 -5.48678100e-01
1.28316969e-01 -4.43802267e-01 -4.07184511e-01 3.76538515e-01
-2.08773136e-01 -1.50257111e-01 1.23406053e-01 -4.34647143e-01
-1.14626622e+00 4.34882864e-02 2.56078225e-02 -8.79198670e-01
-1.11693472e-01 2.21499667e-01 -3.69108289e-01 -9.01526883e-02
2.19066456e-01 -1.19120032e-01 -3.24722901e-02 -2.41025746e-01
4.87047613e-01 7.89871439e-02 7.11064696e-01 -8.93842638e-01
5.02327144e-01 5.74743867e-01 1.71635881e-01 -5.26968718e-01
-1.41608727e+00 -6.94817722e-01 -8.41674149e-01 -5.32906532e-01
1.15024650e+00 -1.31530416e+00 -7.17216730e-01 3.57600987e-01
-1.49739408e+00 -2.30841994e-01 -1.61366507e-01 3.09381515e-01
-9.77652073e-01 4.91251200e-01 -3.70379567e-01 -5.77735066e-01
-1.73308566e-01 -1.04023373e+00 1.45173526e+00 2.95974016e-01
-4.26764458e-01 -8.55544209e-01 -2.78653711e-01 4.89639342e-01
8.40730220e-02 4.52191174e-01 3.87331754e-01 -2.09039882e-01
-9.18949842e-01 2.64911264e-01 -4.66255844e-01 -4.60363030e-02
-1.17673561e-01 6.35052146e-03 -9.45861459e-01 -9.20505524e-02
-2.51623392e-01 -3.51506114e-01 8.47231686e-01 2.83711851e-01
1.40545022e+00 -8.89978260e-02 -1.66124970e-01 1.05392087e+00
1.19170237e+00 8.55776072e-02 5.72955430e-01 5.02938867e-01
9.79174078e-01 8.98915410e-01 1.32026041e+00 3.57105196e-01
6.24319375e-01 1.14741194e+00 5.45229733e-01 -3.66107702e-01
-4.08114135e-01 -2.67606467e-01 3.02797526e-01 5.64589918e-01
-2.34819829e-01 -2.95612842e-01 -7.84656346e-01 4.58670318e-01
-2.34664536e+00 -1.24248719e+00 -2.21777812e-01 1.61822557e+00
4.82612282e-01 -4.74644929e-01 -7.79177819e-04 -2.28495374e-01
8.37474525e-01 2.76911110e-01 -2.26807177e-01 -1.43870547e-01
-4.19869930e-01 7.01819081e-03 -7.45591102e-03 5.03492296e-01
-1.06936550e+00 1.22001326e+00 5.20801830e+00 7.21571445e-01
-9.04003680e-01 3.31967086e-01 6.70637488e-01 7.45500578e-03
-1.70414463e-01 5.84215708e-02 -6.09166563e-01 3.78844231e-01
2.29905099e-01 -8.48409086e-02 2.31100038e-01 7.32615054e-01
2.69078761e-01 -2.30178431e-01 -1.20191514e+00 1.52151871e+00
3.94431412e-01 -1.31623912e+00 -1.09691909e-02 -4.51343060e-01
6.18085921e-01 -3.87429565e-01 -1.45782173e-01 2.06975624e-01
-1.56207308e-01 -8.01825464e-01 1.15678895e+00 7.94952273e-01
7.42913485e-01 -4.38726872e-01 6.33087039e-01 -1.44828469e-01
-1.31559551e+00 -2.91991159e-02 -6.48624673e-02 -1.61737114e-01
5.07739425e-01 1.34012252e-01 -1.53344139e-01 7.95698762e-01
9.78828192e-01 1.24728000e+00 -5.20694315e-01 8.40188444e-01
-3.48800153e-01 1.28682792e-01 -5.34654669e-02 2.69944876e-01
2.02305317e-01 -2.64913023e-01 4.20216173e-01 1.24639463e+00
3.03283423e-01 4.45104539e-01 3.31786782e-01 9.33098197e-01
1.11541942e-01 5.35306074e-02 -4.12947267e-01 1.45281553e-01
3.22744220e-01 1.16741049e+00 -4.27629501e-01 -3.40846479e-01
-5.62428474e-01 1.15581095e+00 4.00421858e-01 6.39347911e-01
-1.07246256e+00 -3.08007710e-02 6.55069828e-01 1.68944240e-01
2.76362836e-01 -2.79162675e-01 1.30843222e-01 -1.43654084e+00
7.20485672e-02 -7.85461664e-01 6.43240213e-01 -1.36981940e+00
-1.10883796e+00 5.04108071e-01 1.78609997e-01 -1.33410060e+00
-1.79722726e-01 -5.94663143e-01 -5.43541014e-01 7.77452290e-01
-1.78703213e+00 -1.43228781e+00 -9.35019732e-01 1.00718701e+00
7.62197793e-01 2.09199563e-01 6.30230844e-01 3.98240775e-01
-6.08680129e-01 4.68275785e-01 -6.36699796e-01 2.84890383e-02
9.67119277e-01 -9.68216181e-01 -9.35628563e-02 7.94806838e-01
2.97832191e-01 4.95389909e-01 4.95904207e-01 -3.77551973e-01
-1.38718987e+00 -1.04054797e+00 9.14240837e-01 -4.40162599e-01
7.55681455e-01 -2.67921776e-01 -9.98097122e-01 7.53076255e-01
2.17152387e-01 3.81737024e-01 6.61314547e-01 8.92416090e-02
-4.57860589e-01 -2.89504305e-02 -6.86700463e-01 5.17898440e-01
1.62927330e+00 -6.09633148e-01 -4.93461102e-01 3.86418015e-01
9.03524756e-01 -8.05970490e-01 -8.20712090e-01 3.82211298e-01
4.44529712e-01 -9.64979768e-01 1.00142229e+00 -7.57727802e-01
4.93842363e-01 -4.84135300e-01 -4.02920842e-01 -8.42578053e-01
-3.45527172e-01 -8.85350347e-01 -2.65486259e-02 1.40535831e+00
-1.53325826e-01 -1.45963132e-01 4.78126913e-01 4.12718326e-01
-1.12436756e-01 -7.25132704e-01 -8.77998352e-01 -5.84796071e-01
-3.04186612e-01 -3.86276305e-01 5.46581388e-01 1.04772818e+00
1.36915706e-02 2.78744221e-01 -5.21166623e-01 6.72097281e-02
5.38550436e-01 5.42495370e-01 1.02030015e+00 -8.23669910e-01
-2.41989240e-01 -2.77303636e-01 -5.85345328e-01 -1.32762611e+00
5.89431643e-01 -7.31951833e-01 1.02959853e-02 -1.35309947e+00
1.63940921e-01 -2.27508426e-01 -2.52153754e-01 5.03749013e-01
-5.03007889e-01 1.59434199e-01 7.52702475e-01 1.97431087e-01
-1.25301230e+00 6.66534007e-01 1.56543136e+00 -1.13969684e-01
1.87496677e-01 -4.95150417e-01 -4.88351792e-01 8.26349795e-01
5.73161900e-01 -2.67694533e-01 -2.52206683e-01 -7.02448547e-01
2.06526201e-02 2.85777658e-01 7.29570627e-01 -7.85222769e-01
2.17691168e-01 -1.94269776e-01 2.73655534e-01 -6.28292203e-01
4.55632210e-01 -8.30585361e-01 1.24733269e-01 9.57661495e-02
-2.50900686e-01 3.89706701e-01 3.09862256e-01 6.17454052e-01
-4.75044847e-01 -2.87439339e-02 5.48300505e-01 -1.34078205e-01
-1.32299888e+00 5.11907518e-01 -4.66053607e-03 2.52112389e-01
9.77173626e-01 -3.22449118e-01 -3.51436704e-01 -6.00744605e-01
-9.74453330e-01 4.10032511e-01 3.81373465e-01 5.82372367e-01
6.09128416e-01 -1.57298911e+00 -5.94435632e-01 -6.54142722e-02
3.91092420e-01 6.80486560e-02 5.42238891e-01 1.13428211e+00
-2.45783195e-01 1.17894664e-01 -3.77403021e-01 -1.03084219e+00
-1.34395695e+00 6.01658583e-01 3.88402313e-01 1.60145536e-01
-5.73986709e-01 1.05672574e+00 6.94177032e-01 1.31702155e-01
1.08252518e-01 -1.82911217e-01 -1.64842814e-01 1.06635302e-01
4.41052169e-01 2.19560817e-01 -2.95141339e-01 -1.07850671e+00
-3.46582890e-01 1.20143414e+00 6.18637092e-02 1.93955734e-01
1.00471878e+00 -4.97316122e-01 -1.07276261e-01 2.58658171e-01
1.12655210e+00 -2.82851934e-01 -1.50878644e+00 -3.53852659e-01
4.11086380e-02 -8.87819231e-01 -1.56921037e-02 -4.62389618e-01
-1.07733858e+00 9.37561452e-01 3.99260551e-01 -4.43775147e-01
1.34919775e+00 1.67946592e-01 5.70978105e-01 2.95827955e-01
4.25767869e-01 -7.75972664e-01 6.41594350e-01 3.61444950e-01
1.17930591e+00 -1.41404390e+00 3.48504409e-02 -6.41832769e-01
-7.67678320e-01 1.20643568e+00 9.56889927e-01 -2.25015823e-02
2.60490626e-01 9.70756188e-02 1.72963813e-01 -2.60543466e-01
-7.20963717e-01 -4.86004114e-01 2.36272857e-01 6.29590988e-01
4.08179641e-01 -2.56423533e-01 -1.55152947e-01 6.79374814e-01
1.70454115e-01 1.75077885e-01 2.79356807e-01 8.15640748e-01
-1.83125660e-01 -9.77269828e-01 -4.26170230e-01 -2.02633902e-01
-3.68161470e-01 -7.57278502e-02 -2.42951527e-01 9.97911274e-01
2.76134282e-01 9.35048163e-01 2.12500036e-01 -1.20203048e-01
5.18250585e-01 -9.11259502e-02 7.31427312e-01 -3.83976549e-01
-3.89565915e-01 2.69424945e-01 1.40471071e-01 -1.24477899e+00
-1.15016651e+00 -7.53925860e-01 -1.26025760e+00 -5.01807407e-02
-3.59077573e-01 -1.73918940e-02 1.89420998e-01 7.44178295e-01
5.52785516e-01 6.63204670e-01 3.39746267e-01 -9.00215566e-01
-9.17712972e-02 -5.80622733e-01 -1.93659246e-01 8.78701508e-01
3.57332587e-01 -1.08330357e+00 -1.07113153e-01 4.57839400e-01] | [9.920746803283691, 0.7465674877166748] |
d585437d-114c-4dca-b6c3-3585c8cb5717 | a-remote-sensing-image-dataset-for-cloud | 1901.00600 | null | http://arxiv.org/abs/1901.00600v1 | http://arxiv.org/pdf/1901.00600v1.pdf | A Remote Sensing Image Dataset for Cloud Removal | Cloud-based overlays are often present in optical remote sensing images, thus
limiting the application of acquired data. Removing clouds is an indispensable
pre-processing step in remote sensing image analysis. Deep learning has
achieved great success in the field of remote sensing in recent years,
including scene classification and change detection. However, deep learning is
rarely applied in remote sensing image removal clouds. The reason is the lack
of data sets for training neural networks. In order to solve this problem, this
paper first proposed the Remote sensing Image Cloud rEmoving dataset (RICE).
The proposed dataset consists of two parts: RICE1 contains 500 pairs of images,
each pair has images with cloud and cloudless size of 512*512; RICE2 contains
450 sets of images, each set contains three 512*512 size images. ,
respectively, the reference picture without clouds, the picture of the cloud
and the mask of its cloud. The dataset is freely available at
\url{https://github.com/BUPTLdy/RICE_DATASET}. | ['Kun fu', 'Xian Sun', 'Guangluan Xu', 'Xiaoke Wang', 'Daoyu Lin', 'Yang Wang'] | 2019-01-03 | null | null | null | null | ['cloud-removal'] | ['computer-vision'] | [ 2.35173330e-01 -7.35032380e-01 4.76867527e-01 -1.25256002e-01
-3.28332454e-01 -4.21954751e-01 4.22472149e-01 -1.42322868e-01
-3.49499285e-01 7.07790256e-01 -2.33578622e-01 -3.80525351e-01
-4.17425446e-02 -9.92119253e-01 -3.92455101e-01 -1.16853046e+00
1.38600260e-01 7.07859322e-02 5.52781969e-02 -8.74380693e-02
1.49516150e-01 8.14731956e-01 -1.64600134e+00 1.80479646e-01
1.00465596e+00 9.31473851e-01 7.49045074e-01 5.80630720e-01
-2.12781891e-01 6.08865023e-01 -3.85673970e-01 2.12621734e-01
6.41987145e-01 -3.42161596e-01 -3.61199856e-01 2.48395279e-01
3.68666202e-01 -5.79535246e-01 -5.27574897e-01 1.47296858e+00
7.52394617e-01 8.74458626e-02 4.21971530e-01 -1.09182024e+00
-9.21379447e-01 2.98057288e-01 -9.47665036e-01 5.95820367e-01
-5.89908779e-01 1.89010918e-01 4.70910192e-01 -1.07403910e+00
1.93316370e-01 8.53787482e-01 4.34351355e-01 1.65168107e-01
-5.09085774e-01 -9.11989033e-01 2.15200931e-02 3.70168298e-01
-1.72094262e+00 -2.98592120e-01 5.12236655e-01 -6.52500153e-01
5.34020245e-01 3.66052568e-01 8.76111746e-01 1.81022078e-01
2.38529667e-01 4.28781122e-01 1.59977806e+00 -3.41421753e-01
-7.84284901e-03 2.30409816e-01 1.57848746e-01 -2.75945496e-02
5.48319161e-01 6.54576123e-02 1.10849649e-01 2.43991107e-01
9.06512082e-01 8.16389441e-01 -5.97361743e-01 2.68899351e-01
-7.18571842e-01 6.77693903e-01 7.59427786e-01 6.07545555e-01
-4.47044730e-01 -1.42951876e-01 -1.37585610e-01 3.52575630e-01
8.42437148e-01 -1.49549916e-01 -4.75624353e-01 5.24860919e-01
-8.34153891e-01 3.30897719e-01 1.20368302e-01 7.84326851e-01
9.77068424e-01 4.15152073e-01 3.72556448e-01 8.10371935e-01
8.51363614e-02 9.65252936e-01 4.06096756e-01 -5.91353476e-01
2.45869383e-01 5.38046837e-01 2.47656211e-01 -1.10745895e+00
-1.86912388e-01 -4.73466486e-01 -1.35844576e+00 4.46901083e-01
-1.54338896e-01 -2.85976946e-01 -9.97348845e-01 7.99395323e-01
3.20705950e-01 2.86039829e-01 5.22342399e-02 1.08188343e+00
1.11862099e+00 9.80044305e-01 -1.70439139e-01 -2.36355424e-01
1.13913214e+00 -7.91312993e-01 -6.43629014e-01 -1.62961453e-01
1.36613339e-01 -8.72146249e-01 9.06227887e-01 3.28623891e-01
-5.82541406e-01 -5.70751309e-01 -8.30146909e-01 5.86851835e-02
-7.58906007e-01 3.87741297e-01 5.57137907e-01 2.75093049e-01
-1.12864232e+00 3.59000534e-01 -3.99271101e-01 -3.28547269e-01
4.60334450e-01 1.10135987e-01 -2.34704301e-01 -3.34532589e-01
-9.47889328e-01 7.31130302e-01 5.55346012e-01 7.32001543e-01
-8.20951343e-01 -3.33399802e-01 -3.88011485e-01 1.47892267e-01
2.67728865e-01 -1.79208726e-01 7.90676415e-01 -1.37612987e+00
-9.32296157e-01 9.15656984e-01 1.11532874e-01 3.22373584e-03
4.37405467e-01 -3.78053188e-01 -5.99887609e-01 -8.47251192e-02
5.69183612e-03 2.09313020e-01 9.36320484e-01 -1.49959183e+00
-6.83504999e-01 -6.13850117e-01 -2.81895936e-01 2.41012022e-01
-9.90069434e-02 2.06203356e-01 -2.80774415e-01 -5.03477454e-01
2.14188382e-01 -8.49646449e-01 -2.71155059e-01 -1.36573300e-01
-2.60740340e-01 2.08579838e-01 1.11611366e+00 -9.74017620e-01
7.55351663e-01 -2.20969105e+00 -5.21598458e-01 7.40877464e-02
2.54837811e-01 7.72839725e-01 -3.84878218e-01 4.43670183e-01
-3.97773564e-01 3.54526490e-01 -4.65475351e-01 6.96868822e-02
-5.86026609e-01 -5.62080070e-02 -3.44047278e-01 6.96909368e-01
1.65239200e-01 4.53768253e-01 -4.96540427e-01 -3.95222157e-01
4.38675702e-01 6.54370010e-01 1.98612846e-02 2.46804446e-01
-1.06464587e-01 7.71551251e-01 -5.11785924e-01 6.70376599e-01
1.63840294e+00 -1.24033943e-01 -6.10998012e-02 1.15000069e-01
-5.89936376e-01 -2.14011848e-01 -1.27029943e+00 8.93948555e-01
-1.08635239e-01 8.42667699e-01 1.65107176e-01 -7.13294923e-01
1.08431578e+00 1.87842757e-01 4.80492711e-01 -7.28851914e-01
2.23093778e-01 1.37966380e-01 -3.49766761e-02 -8.08567762e-01
5.67477286e-01 -1.16077438e-01 6.85145617e-01 2.95933515e-01
-6.66475415e-01 -3.96376103e-01 -8.15088749e-02 -1.56520829e-01
4.26003337e-01 -4.19490114e-02 5.34300767e-02 3.63815166e-02
5.07423162e-01 2.91486919e-01 6.03615642e-01 5.71658969e-01
-7.60776103e-02 7.97742724e-01 -2.25311533e-01 -8.21938992e-01
-1.17411375e+00 -7.30589926e-01 -2.55806327e-01 7.25249887e-01
2.39854544e-01 3.28933686e-01 -4.60207552e-01 7.47273192e-02
2.41053253e-02 3.03062469e-01 -4.31520134e-01 3.54090720e-01
-4.85043406e-01 -1.02810204e+00 1.84340343e-01 1.00777149e-01
1.13533318e+00 -1.38024759e+00 -2.81687975e-01 -9.18723494e-02
-3.86040121e-01 -7.61123121e-01 -4.11653193e-03 -1.94007382e-01
-1.00093007e+00 -1.16521418e+00 -6.58513606e-01 -6.57689750e-01
6.21241987e-01 1.26596594e+00 1.04752219e+00 5.93506217e-01
-4.53046590e-01 -3.28200310e-01 -4.96762991e-01 -7.53854096e-01
-3.68710421e-02 -1.05170205e-01 -1.67191267e-01 2.39196531e-02
4.97816741e-01 -6.85289025e-01 -7.87558973e-01 1.44146299e-02
-1.25511503e+00 2.07242034e-02 8.28808010e-01 4.45316494e-01
6.73648953e-01 6.20099962e-01 -3.98517810e-02 -7.44527280e-01
1.12226397e-01 -6.17399514e-01 -1.03147900e+00 -8.07915479e-02
-4.24686253e-01 -7.58986831e-01 3.43248874e-01 8.36776868e-02
-1.09500861e+00 4.36156690e-02 -3.11511662e-03 -4.46900576e-01
-4.81621474e-01 7.29846299e-01 -1.62528336e-01 -5.58436625e-02
3.72232109e-01 5.53792775e-01 -1.30118668e-01 -7.81633496e-01
-3.33274505e-03 1.10693026e+00 3.59501004e-01 1.31521091e-01
1.01527548e+00 6.57062888e-01 -4.09876347e-01 -1.23391795e+00
-5.09024024e-01 -7.15229869e-01 -8.09594870e-01 -1.78947419e-01
9.37727869e-01 -1.30751002e+00 -3.39019924e-01 8.93383563e-01
-1.01145947e+00 -2.92375177e-01 1.76655084e-01 5.36470532e-01
5.89814223e-02 2.29788929e-01 -3.85011077e-01 -9.91289914e-01
-7.08212733e-01 -6.98659241e-01 6.44271553e-01 5.31651556e-01
7.93764412e-01 -5.36992729e-01 4.09973152e-02 3.65036875e-01
5.63277960e-01 3.16641122e-01 5.43098748e-01 -6.34739324e-02
-8.80903721e-01 -8.79566446e-02 -7.14812219e-01 7.25296617e-01
4.79746372e-01 4.39688027e-01 -1.14136589e+00 -4.48169678e-01
1.36093050e-01 2.47992240e-02 1.10325587e+00 6.32387280e-01
1.30962825e+00 -1.85197383e-01 -1.96897328e-01 7.35210896e-01
2.14805984e+00 4.00697738e-01 9.99795556e-01 5.44604659e-01
8.26895535e-01 4.88541156e-01 4.72283572e-01 4.04872388e-01
1.22650482e-01 1.21253148e-01 7.70765781e-01 -6.29732847e-01
2.55083907e-02 2.78453708e-01 4.53228019e-02 7.77583420e-01
-5.32647192e-01 -2.00454563e-01 -1.12844932e+00 6.00288808e-01
-1.48107982e+00 -1.22280931e+00 -6.81546867e-01 2.09652495e+00
4.73224699e-01 -5.21139264e-01 -2.60863662e-01 9.57999378e-02
9.24276590e-01 4.01171058e-01 -5.74062705e-01 2.11430460e-01
-3.97358328e-01 1.25246048e-01 6.49217069e-01 4.04870003e-01
-1.34864235e+00 9.86645997e-01 5.06309795e+00 5.49288988e-01
-1.66447628e+00 2.38177106e-01 5.29661834e-01 1.09034769e-01
-1.42718270e-01 -4.60587405e-02 -6.53937101e-01 6.00306034e-01
4.13649857e-01 1.83864057e-01 4.37364250e-01 6.18067682e-01
7.51942813e-01 -3.44719738e-01 -1.88873589e-01 1.02399516e+00
-6.85709119e-02 -1.18340099e+00 1.89759731e-01 1.05610311e-01
9.47528005e-01 6.19774461e-01 8.27105418e-02 1.11314782e-03
3.70293260e-01 -1.05720973e+00 4.07440633e-01 6.46816790e-01
9.23015833e-01 -5.00636816e-01 9.56305087e-01 4.63272125e-01
-1.25252831e+00 -1.20463446e-01 -8.62366796e-01 -4.50418830e-01
-4.31506276e-01 9.56294000e-01 -4.78496343e-01 6.58337295e-01
1.20387030e+00 8.01024437e-01 -5.56843758e-01 1.29570615e+00
1.95111446e-02 5.45156717e-01 -2.12863773e-01 3.19231242e-01
1.76976845e-01 -8.65935743e-01 2.74752140e-01 8.92484069e-01
4.25827533e-01 5.64438820e-01 2.23010540e-01 7.67772377e-01
6.66683465e-02 5.93785197e-02 -8.44627976e-01 -3.65076922e-02
7.04852700e-01 1.34575820e+00 -6.53210223e-01 -2.96137989e-01
-2.88470626e-01 7.02372968e-01 -2.27415234e-01 5.04109740e-01
-6.01054013e-01 -3.51105273e-01 6.39782131e-01 9.24818218e-02
3.96116853e-01 -3.05772632e-01 -2.11624846e-01 -1.29949224e+00
5.72448820e-02 -9.67974067e-01 6.63617477e-02 -1.20595407e+00
-1.07452488e+00 5.77330112e-01 -2.11487800e-01 -1.50913191e+00
4.09218162e-01 -4.58967477e-01 -8.72280777e-01 1.33070481e+00
-2.05370784e+00 -1.22776437e+00 -1.28430951e+00 6.77580178e-01
4.36201602e-01 -6.06699064e-02 5.96671760e-01 5.12736976e-01
-5.99698603e-01 -1.75946668e-01 8.08588624e-01 3.40160549e-01
4.28282261e-01 -8.46960545e-01 2.47323036e-01 1.16678977e+00
-2.08121121e-01 2.23901063e-01 5.52802444e-01 -5.34118772e-01
-1.05855978e+00 -1.48021996e+00 6.20922625e-01 1.01605795e-01
2.32476801e-01 4.17575575e-02 -1.10748708e+00 6.83283269e-01
2.12269485e-01 1.36793494e-01 4.98838365e-01 -5.28892100e-01
3.50412801e-02 -4.98064280e-01 -1.01207089e+00 2.44957462e-01
4.49600667e-01 -3.35132688e-01 -5.95919676e-02 6.01997197e-01
4.02036518e-01 -3.05131167e-01 -4.73603159e-01 2.00874135e-01
2.67535865e-01 -1.17735040e+00 7.34993398e-01 -6.87723458e-02
6.31626129e-01 -6.98289514e-01 -1.98266640e-01 -1.15089273e+00
-6.14724874e-01 1.80174306e-01 8.37110817e-01 1.19242704e+00
-3.81787270e-02 -7.93257236e-01 6.20208681e-01 6.69023916e-02
-1.73271075e-01 -2.34833494e-01 -5.61157763e-01 -5.08645177e-01
3.52684677e-01 1.45303592e-01 8.46259594e-01 1.34937000e+00
-1.13993204e+00 -9.23116431e-02 -3.76931965e-01 6.65424407e-01
5.09107530e-01 4.98745918e-01 9.74380732e-01 -1.44520843e+00
4.29479256e-02 -3.37230057e-01 -6.14492707e-02 -4.98409957e-01
-3.00861061e-01 -4.57881570e-01 -1.72137748e-02 -1.75105774e+00
4.50115800e-01 -8.55013251e-01 -1.24555640e-01 5.90322733e-01
-3.41170162e-01 4.35278416e-01 3.04105878e-01 8.40797663e-01
1.17321178e-01 5.67334652e-01 1.31071568e+00 -3.97063345e-01
-2.01667517e-01 6.08612262e-02 -4.40765440e-01 6.59166276e-01
1.32373703e+00 -5.17258823e-01 -1.33571237e-01 -9.54536080e-01
-1.04361430e-01 -2.35546768e-01 4.70187187e-01 -1.09894526e+00
-5.09017962e-04 -4.62987840e-01 7.01060295e-01 -1.02040911e+00
2.35481799e-01 -9.78964925e-01 5.91202438e-01 5.89529514e-01
2.16741353e-01 1.54596180e-01 2.56624550e-01 2.75655478e-01
-4.02055889e-01 -2.26061121e-01 8.77703190e-01 -6.60237968e-01
-1.01852465e+00 7.00091183e-01 -2.47447371e-01 -4.90119398e-01
9.26741660e-01 -2.82755405e-01 -5.13375878e-01 -2.10957572e-01
-5.28078139e-01 4.72673066e-02 4.86394703e-01 1.50477469e-01
7.56157815e-01 -9.50753272e-01 -1.15833163e+00 2.33894646e-01
5.99430948e-02 3.42866451e-01 6.46559894e-01 7.16221690e-01
-1.16349387e+00 1.10744767e-01 -4.42887485e-01 -4.59072053e-01
-1.45682371e+00 4.09975618e-01 7.24786997e-01 1.65132582e-01
-6.81426406e-01 7.38620460e-01 2.10483462e-01 -4.45464581e-01
-2.89292574e-01 -1.19615473e-01 -4.68822896e-01 -1.34786069e-01
8.13650608e-01 2.64908344e-01 2.74527580e-01 -7.02406466e-01
-1.86797038e-01 6.71458840e-01 -6.30873665e-02 1.29414275e-01
1.64873827e+00 -3.81521404e-01 -6.69469297e-01 4.18476850e-01
9.03249800e-01 -3.61221060e-02 -1.03411937e+00 -4.55518454e-01
-4.62390363e-01 -9.96118128e-01 3.08104485e-01 -6.62541747e-01
-1.54515052e+00 1.11615098e+00 1.13170171e+00 1.82841346e-01
1.30428576e+00 -4.27745908e-01 5.65238178e-01 3.78702641e-01
1.34299591e-01 -7.49351859e-01 -2.81344295e-01 6.42506897e-01
8.82398665e-01 -1.64106977e+00 2.25592911e-01 -3.99518549e-01
-4.30548996e-01 9.03448641e-01 4.48001087e-01 -2.22855166e-01
8.48965168e-01 4.23491895e-02 5.74912786e-01 -2.88651437e-01
-2.67709821e-01 -4.65326875e-01 -2.42753163e-01 7.08920002e-01
4.34885740e-01 3.03051353e-01 -5.73342144e-02 2.64082048e-02
-1.36943236e-01 1.17669560e-01 7.20169306e-01 9.18436885e-01
-7.24159837e-01 -6.16396546e-01 -7.30836868e-01 5.78424454e-01
-4.63318199e-01 -5.19622207e-01 -2.70206481e-01 7.45384693e-01
5.02210557e-01 1.05107677e+00 3.42045337e-01 -3.54714066e-01
-1.61887005e-01 -4.06617939e-01 -2.10811552e-02 -5.52941561e-01
-4.76235718e-01 7.44684860e-02 -4.50050056e-01 3.88854407e-02
-5.86625218e-01 -4.90262508e-01 -9.54399288e-01 -7.83004403e-01
-3.75739932e-01 1.74755931e-01 8.24134588e-01 7.29643822e-01
1.85475662e-01 1.99822366e-01 9.33964074e-01 -1.05415201e+00
-1.39014900e-01 -1.16373742e+00 -1.15299904e+00 1.85457736e-01
4.93231714e-01 -3.97354275e-01 -3.75139743e-01 3.69457036e-01] | [9.850919723510742, -1.7753591537475586] |
979a43b4-358b-4414-8678-a62c517a61fd | ad-kd-attribution-driven-knowledge | 2305.10010 | null | https://arxiv.org/abs/2305.10010v1 | https://arxiv.org/pdf/2305.10010v1.pdf | AD-KD: Attribution-Driven Knowledge Distillation for Language Model Compression | Knowledge distillation has attracted a great deal of interest recently to compress pre-trained language models. However, existing knowledge distillation methods suffer from two limitations. First, the student model simply imitates the teacher's behavior while ignoring the underlying reasoning. Second, these methods usually focus on the transfer of sophisticated model-specific knowledge but overlook data-specific knowledge. In this paper, we present a novel attribution-driven knowledge distillation approach, which explores the token-level rationale behind the teacher model based on Integrated Gradients (IG) and transfers attribution knowledge to the student model. To enhance the knowledge transfer of model reasoning and generalization, we further explore multi-view attribution distillation on all potential decisions of the teacher. Comprehensive experiments are conducted with BERT on the GLUE benchmark. The experimental results demonstrate the superior performance of our approach to several state-of-the-art methods. | ['Rui Wang', 'Qifan Wang', 'Xiaojun Quan', 'Hongzhan Chen', 'Siyue Wu'] | 2023-05-17 | null | null | null | null | ['model-compression'] | ['methodology'] | [ 1.85619876e-01 3.27857256e-01 -6.17206633e-01 -3.87449086e-01
-3.51367623e-01 -4.25557494e-01 6.17224157e-01 1.14587530e-01
-5.14723957e-01 7.92518854e-01 2.24538073e-01 -4.98889685e-01
-6.21286444e-02 -7.96678960e-01 -6.51636362e-01 -5.87517321e-01
5.97609878e-01 5.20155847e-01 4.36637700e-01 -1.19703732e-01
1.34162456e-01 2.79079556e-01 -1.18675649e+00 2.18135267e-01
1.28498173e+00 7.26990342e-01 -5.41864075e-02 3.34998339e-01
-2.68532038e-01 1.37151754e+00 -4.20016497e-01 -8.71249855e-01
-2.91743308e-01 -4.49040979e-01 -1.01236045e+00 -2.16022402e-01
5.11064887e-01 -4.16787714e-01 -5.44149637e-01 1.27190542e+00
3.56682837e-01 3.96822155e-01 7.65303910e-01 -1.22498393e+00
-1.24152231e+00 1.14223897e+00 -4.16043162e-01 3.00811201e-01
-4.14889269e-02 2.45066993e-02 8.13535988e-01 -9.00403798e-01
3.79603386e-01 1.27994645e+00 4.87683296e-01 7.02078402e-01
-1.29495966e+00 -9.18719947e-01 5.95339537e-01 7.83508539e-01
-1.40282488e+00 1.60401817e-02 1.02477264e+00 -5.28906643e-01
7.36397207e-01 -1.97444230e-01 5.21721125e-01 1.09807229e+00
-3.14390928e-01 1.44577932e+00 1.51646972e+00 -4.45603758e-01
1.15398556e-01 5.76942801e-01 3.85400623e-01 1.02908766e+00
1.48773700e-01 7.75693879e-02 -7.20459163e-01 -7.37283891e-03
3.97925407e-01 1.09686822e-01 -3.46033096e-01 -6.12739503e-01
-1.07385993e+00 8.27929318e-01 5.21794975e-01 -3.02207749e-02
-2.62194663e-01 9.53563452e-02 3.33418429e-01 1.89061180e-01
5.89752793e-01 3.78725290e-01 -6.81005299e-01 -1.44976452e-01
-1.07726920e+00 1.51836872e-01 8.39552164e-01 1.00995958e+00
9.29393053e-01 2.11865693e-01 -4.39362496e-01 7.53361404e-01
3.00653547e-01 3.68306130e-01 8.19570541e-01 -7.30396628e-01
4.00123984e-01 8.34823191e-01 -4.10362422e-01 -6.45982027e-01
1.10804565e-01 -6.16949558e-01 -6.84305429e-01 -1.40643746e-01
5.95007576e-02 -1.24551311e-01 -9.51571882e-01 1.68853641e+00
4.61792678e-01 6.74428403e-01 4.09723222e-01 5.74902058e-01
9.93962288e-01 1.22189783e-01 3.68542939e-01 4.73161601e-02
1.23682547e+00 -1.56730270e+00 -7.06738651e-01 -1.37909770e-01
6.83760405e-01 -4.18277860e-01 9.04913783e-01 3.68148208e-01
-1.17749834e+00 -5.01177669e-01 -7.86395907e-01 -1.40046403e-01
-7.19448030e-01 4.18758020e-02 6.58296704e-01 4.85490650e-01
-8.52055490e-01 4.99869496e-01 -7.31458545e-01 -1.17147774e-01
8.15881014e-01 1.31989211e-01 7.30048195e-02 -2.15266094e-01
-1.49263656e+00 1.18832302e+00 7.00125217e-01 -2.00312480e-01
-1.24688101e+00 -1.14873505e+00 -8.07903588e-01 3.56383353e-01
6.48232281e-01 -8.34036529e-01 1.70097029e+00 -7.61631727e-01
-1.87271857e+00 7.67798305e-01 -1.95037678e-01 -4.19271648e-01
7.62947321e-01 -3.68578732e-01 -9.94997099e-02 -3.80244218e-02
-6.61911070e-02 6.43617511e-01 7.14157939e-01 -1.20578349e+00
-7.19118237e-01 -1.35173887e-01 2.39036262e-01 3.34699184e-01
-6.83071852e-01 -2.53334016e-01 -4.32966113e-01 -7.07618296e-01
-1.28848359e-01 -7.35759020e-01 3.29041891e-02 -2.00252429e-01
-5.07497251e-01 -6.44733906e-01 7.87553370e-01 -4.17270362e-01
1.48660970e+00 -1.84011972e+00 2.88705647e-01 -6.40426055e-02
3.16029727e-01 7.21670628e-01 6.68330267e-02 1.94172412e-01
-1.14858210e-01 1.06981620e-01 -1.77418932e-01 -4.19736505e-01
1.29881695e-01 5.28461814e-01 -8.12607288e-01 7.58900270e-02
5.13642803e-02 1.15260851e+00 -1.32216549e+00 -6.84601486e-01
-5.32012768e-02 3.24114442e-01 -7.12737381e-01 5.39135754e-01
-4.19135451e-01 2.37914577e-01 -6.59849942e-01 5.29412031e-01
5.51398218e-01 -3.73134911e-01 8.90865549e-02 -1.30667076e-01
7.73773491e-02 5.89108050e-01 -8.25924873e-01 1.73758435e+00
-4.16063339e-01 3.13722342e-01 -2.73076117e-01 -1.18958306e+00
6.92918301e-01 2.10350305e-01 -5.47876861e-03 -2.92717397e-01
-5.58027811e-02 -4.52049337e-02 1.32305503e-01 -4.58985299e-01
4.35455561e-01 -2.72299975e-01 3.45996737e-01 7.05738127e-01
2.14735329e-01 -1.64939418e-01 -9.43869948e-02 4.88233030e-01
6.75479770e-01 4.34885502e-01 1.93579495e-01 -1.57505512e-01
5.75360715e-01 1.74993686e-02 5.56617141e-01 7.28789449e-01
-3.85991305e-01 -1.64507441e-02 3.31477940e-01 -1.21252507e-01
-3.58474731e-01 -1.14430594e+00 1.78396106e-01 1.29687536e+00
2.64846683e-02 -3.23876977e-01 -7.50678241e-01 -1.20941865e+00
2.70919412e-01 1.18261588e+00 -9.01585996e-01 -6.65167987e-01
-2.82595962e-01 -5.90500355e-01 8.21582675e-01 9.30725157e-01
9.98095393e-01 -7.65242100e-01 -3.14701200e-01 6.48123845e-02
-4.38042693e-02 -1.09723222e+00 -2.39079386e-01 8.13185126e-02
-9.60715175e-01 -8.77215266e-01 -4.64205891e-01 -5.37065506e-01
8.91161323e-01 3.62283975e-01 1.09561861e+00 4.75965925e-02
1.45254031e-01 6.52350605e-01 -1.82492077e-01 -4.95568663e-01
-3.00740153e-01 3.44400436e-01 1.45281494e-01 -1.87083930e-01
8.77782285e-01 -4.76001859e-01 -3.01884800e-01 6.48463592e-02
-7.22378075e-01 3.77278179e-01 6.04605079e-01 8.97660792e-01
4.90674376e-01 1.66971028e-01 5.98379195e-01 -1.15677202e+00
7.08864570e-01 -6.28924727e-01 -3.44523042e-01 6.38907671e-01
-1.19026423e+00 4.05298084e-01 6.45040154e-01 -6.82409823e-01
-1.61627531e+00 -4.04037803e-01 3.97660136e-01 -9.63392913e-01
-6.65460452e-02 7.17454255e-01 -3.56127210e-02 -4.18507978e-02
3.07862818e-01 5.38207769e-01 -2.82125950e-01 -6.74424052e-01
8.09988022e-01 3.59115720e-01 4.83859599e-01 -1.16402793e+00
8.13398480e-01 3.56781334e-01 -3.41958314e-01 -4.34024692e-01
-1.56554604e+00 -2.09070429e-01 -6.28443062e-01 1.81421544e-02
6.59811854e-01 -1.06838667e+00 -8.02109003e-01 3.79329562e-01
-1.12424994e+00 -5.31421959e-01 -4.20591116e-01 6.21177435e-01
-3.10510665e-01 2.12155208e-01 -5.55022180e-01 -4.97884482e-01
-9.15541053e-02 -1.24102569e+00 2.84549206e-01 4.93745863e-01
-1.99294239e-02 -1.51483691e+00 2.03247488e-01 7.84967959e-01
6.77974343e-01 -5.67486942e-01 1.30648386e+00 -1.15809631e+00
-5.32175481e-01 1.69385016e-01 -3.65474582e-01 4.92976457e-01
5.19841388e-02 -3.73853259e-02 -1.31617510e+00 6.67867064e-02
-1.75892383e-01 -8.40431988e-01 1.20848858e+00 -1.82958379e-01
1.49540091e+00 -3.84452403e-01 -3.82383496e-01 5.41672230e-01
1.06439579e+00 -1.84588134e-01 1.56946242e-01 3.13175201e-01
1.09954011e+00 3.98205310e-01 3.13825130e-01 1.00345299e-01
9.79678988e-01 4.07978594e-01 3.61099660e-01 1.38724335e-02
-3.12159389e-01 -8.33725989e-01 4.17017847e-01 1.14880085e+00
-1.17227107e-01 7.29758143e-02 -9.16181147e-01 6.73293173e-01
-1.91677105e+00 -8.32223356e-01 2.57890671e-01 1.92779875e+00
1.47905493e+00 -5.33692874e-02 -2.89223939e-01 -2.48169154e-01
5.03539205e-01 1.18925974e-01 -8.81866634e-01 -5.37929893e-01
1.44961908e-01 2.18062520e-01 3.38886023e-01 6.48876250e-01
-8.57404292e-01 1.43616927e+00 6.26230764e+00 9.55438077e-01
-1.07504225e+00 2.21349731e-01 3.47426564e-01 6.79375231e-02
-4.59005475e-01 9.74366963e-02 -1.21007180e+00 1.38508499e-01
8.43311012e-01 -4.32384461e-01 3.41465116e-01 1.05722523e+00
-2.77834892e-01 -7.29914755e-02 -1.22287071e+00 4.63184655e-01
2.49314815e-01 -1.08193564e+00 4.43713337e-01 -1.42184287e-01
1.19898343e+00 -5.15724812e-03 3.05971652e-01 9.44046855e-01
8.36353779e-01 -8.41350019e-01 4.81937379e-01 6.81913793e-01
3.19193333e-01 -7.10220516e-01 5.22945642e-01 4.89366651e-01
-7.99654424e-01 1.46439686e-01 -2.12277383e-01 -1.91368312e-01
-3.16785902e-01 4.30430442e-01 -1.21749496e+00 4.47850436e-01
4.57256019e-01 8.36298347e-01 -6.65331841e-01 4.89502192e-01
-1.23860753e+00 1.19191408e+00 3.85249220e-02 -4.38555367e-02
3.87632400e-01 -3.68703939e-02 9.49373543e-02 1.37598193e+00
-1.32828951e-01 2.38676161e-01 3.27441216e-01 1.28926325e+00
-5.86843193e-01 -1.70688510e-01 -4.76956636e-01 -2.49735326e-01
6.37537420e-01 1.03836548e+00 -1.79792345e-01 -9.51550305e-01
-6.59992814e-01 9.42052722e-01 9.35283899e-01 6.16353571e-01
-7.76763320e-01 -1.22062497e-01 7.46407449e-01 -3.33079070e-01
5.31088650e-01 -8.39717090e-02 -1.38952613e-01 -1.41950607e+00
-2.72362322e-01 -6.87212348e-01 4.69893068e-01 -8.07636380e-01
-1.34666169e+00 4.70109470e-02 3.04747432e-01 -6.49750948e-01
-1.91068843e-01 -5.56550920e-01 -8.03235710e-01 1.01009977e+00
-2.20009232e+00 -1.22939599e+00 -1.88966870e-01 6.24663949e-01
4.48632717e-01 -2.71757931e-01 8.93728733e-01 -5.87209547e-03
-8.20977628e-01 8.91897023e-01 1.55559883e-01 3.42159092e-01
7.64706969e-01 -1.37517250e+00 -5.52670285e-02 6.14910066e-01
6.69009192e-03 9.67942953e-01 4.57297385e-01 -6.46502912e-01
-1.37430882e+00 -1.16049755e+00 1.03628683e+00 -5.95023453e-01
9.98923063e-01 4.53004315e-02 -1.30337083e+00 1.12468576e+00
4.83996660e-01 -5.30792587e-02 1.12769675e+00 2.68546373e-01
-8.51322532e-01 4.52663936e-02 -9.72440600e-01 6.98514581e-01
6.19750381e-01 -7.71039665e-01 -1.25007880e+00 1.52798325e-01
7.32541084e-01 -3.79709631e-01 -9.45957839e-01 2.49143049e-01
3.25176358e-01 -7.00812817e-01 9.00447786e-01 -1.22071469e+00
7.27250993e-01 -2.28465587e-01 7.82604516e-02 -1.67324734e+00
-3.21020067e-01 -3.11565518e-01 -6.42378986e-01 1.44351435e+00
3.94769609e-01 -5.74060559e-01 8.21400166e-01 8.43640864e-01
1.03031099e-02 -9.52709734e-01 -6.16947234e-01 -7.74973273e-01
4.14787382e-01 -3.17754000e-01 5.84063292e-01 1.55756509e+00
2.19279617e-01 5.03633916e-01 -1.25025347e-01 9.55895036e-02
6.66978419e-01 3.88646901e-01 6.04528129e-01 -1.00866830e+00
-3.69333863e-01 -5.34089923e-01 -1.29716760e-02 -1.20228612e+00
7.23774314e-01 -1.33713436e+00 -1.84837386e-01 -1.40227592e+00
5.34175277e-01 -3.78035039e-01 -7.49176860e-01 9.37405109e-01
-7.79888451e-01 -2.26305097e-01 6.78698299e-03 7.48795271e-03
-7.40927577e-01 8.65022957e-01 1.44139266e+00 -3.01746041e-01
9.78399664e-02 -2.96029866e-01 -8.51678550e-01 9.67558622e-01
7.38345146e-01 -6.50886357e-01 -7.76152372e-01 -6.40838563e-01
1.58021808e-01 -5.38115561e-01 4.52841699e-01 -4.92109627e-01
6.57289624e-01 -4.38497066e-01 1.21252239e-01 -3.33412290e-01
1.39513701e-01 -5.70980608e-01 -5.78537643e-01 4.74314719e-01
-7.26525605e-01 -3.09914947e-01 3.42848837e-01 6.35016322e-01
-2.71888554e-01 -2.99970508e-01 6.92888737e-01 -2.17547581e-01
-8.53763103e-01 3.15504879e-01 -2.61331871e-02 4.45544362e-01
7.55725086e-01 1.77846447e-01 -5.98098814e-01 -4.74667661e-02
-4.81701314e-01 5.65280735e-01 2.63260186e-01 4.32693154e-01
5.58101535e-01 -1.32719064e+00 -7.13585138e-01 7.47507736e-02
2.56796628e-01 2.17958078e-01 1.60369143e-01 9.54519093e-01
8.62639546e-02 5.72717190e-01 1.21146739e-01 -2.87956417e-01
-1.03793037e+00 7.95637429e-01 2.84958512e-01 -4.00762618e-01
-3.52547199e-01 1.28786588e+00 4.15547967e-01 -7.38568008e-01
2.51405478e-01 -4.82138038e-01 -1.55059651e-01 6.42593205e-02
5.27846396e-01 4.92982626e-01 -1.89782888e-01 -2.32139885e-01
-2.28179097e-01 3.18372607e-01 -6.94494605e-01 1.02555454e-01
9.36682880e-01 2.74628997e-01 6.92180172e-02 4.60023135e-01
9.60765600e-01 -2.99893796e-01 -1.35723722e+00 -9.24755871e-01
-4.84745018e-02 -2.89295316e-01 3.77042323e-01 -1.22646153e+00
-1.14552486e+00 1.42964244e+00 3.47733349e-02 -2.65522122e-01
8.04590106e-01 -2.38394216e-02 3.75644058e-01 6.42196953e-01
3.82046103e-02 -1.12906241e+00 1.54746085e-01 7.64621854e-01
4.52904344e-01 -1.22568297e+00 1.70751154e-01 -1.88401982e-01
-7.37598777e-01 9.84221220e-01 1.13241756e+00 3.55230719e-01
6.33642793e-01 -5.46581186e-02 4.41491306e-02 -4.53703813e-02
-1.02077961e+00 -1.79158807e-01 2.19510213e-01 3.38192582e-01
6.00402415e-01 2.80824691e-01 -3.20708891e-03 8.20617497e-01
-6.89699799e-02 3.64360422e-01 3.07730108e-01 1.10157681e+00
-4.75755513e-01 -9.79456067e-01 -6.70988634e-02 1.14666194e-01
-2.01210544e-01 -4.94619250e-01 -5.35856128e-01 6.65544510e-01
1.52443096e-01 6.21286929e-01 -4.24851567e-01 -3.54419112e-01
3.84634256e-01 5.42221308e-01 4.95105863e-01 -7.11121500e-01
-6.80736005e-01 -5.43048918e-01 -1.94646224e-01 -3.47546577e-01
-4.08902943e-01 -3.58947754e-01 -1.25718307e+00 -2.58730531e-01
-3.78841937e-01 3.94414127e-01 4.82030421e-01 1.24404693e+00
3.86088639e-01 8.03654075e-01 1.87850788e-01 -3.64279687e-01
-1.05717087e+00 -9.59896624e-01 -4.10369694e-01 2.44001374e-01
1.58978015e-01 -9.03113663e-01 -4.23247576e-01 2.09931970e-01] | [9.515458106994629, 3.448256254196167] |
96e0c243-1e2f-422b-a454-8333d638b211 | sentiment-analysis-for-arabic-language-a | 1809.02782 | null | http://arxiv.org/abs/1809.02782v2 | http://arxiv.org/pdf/1809.02782v2.pdf | Sentiment analysis for Arabic language: A brief survey of approaches and techniques | With the emergence of Web 2.0 technology and the expansion of on-line social
networks, current Internet users have the ability to add their reviews, ratings
and opinions on social media and on commercial and news web sites. Sentiment
analysis aims to classify these reviews reviews in an automatic way. In the
literature, there are numerous approaches proposed for automatic sentiment
analysis for different language contexts. Each language has its own properties
that makes the sentiment analysis more challenging. In this regard, this work
presents a comprehensive survey of existing Arabic sentiment analysis studies,
and covers the various approaches and techniques proposed in the literature.
Moreover, we highlight the main difficulties and challenges of Arabic sentiment
analysis, and the proposed techniques in literature to overcome these barriers. | ['Hossam Faris', "Mo'ath Alrefai", 'Ibrahim Aljarah'] | 2018-09-08 | null | null | null | null | ['arabic-sentiment-analysis'] | ['natural-language-processing'] | [-4.23803836e-01 -1.93633631e-01 -4.46415663e-01 -4.06098574e-01
-1.50369570e-01 -8.17364275e-01 4.83726382e-01 7.78214872e-01
-4.81231213e-01 6.53507888e-01 9.70957950e-02 -1.83286279e-01
2.06769720e-01 -8.92650843e-01 2.95196623e-01 -2.79007971e-01
2.26503313e-01 8.02796185e-02 9.63414833e-02 -1.16787612e+00
9.14075971e-01 1.53519511e-01 -1.59619451e+00 2.94678211e-01
9.18145955e-01 9.84693527e-01 1.27803475e-01 4.77682114e-01
-7.94577241e-01 6.50592446e-01 -7.85176337e-01 -1.13445961e+00
-1.60910055e-01 -3.11797887e-01 -4.42710042e-01 1.67644441e-01
-1.91431686e-01 9.43204165e-02 3.61170799e-01 1.19875133e+00
2.99615771e-01 -6.77634552e-02 7.07938552e-01 -1.28098750e+00
-9.24912453e-01 7.56341100e-01 -8.99290919e-01 -1.27913700e-02
6.76238775e-01 -8.67639899e-01 9.99341130e-01 -1.14066601e+00
4.42238271e-01 7.55496085e-01 4.33017403e-01 8.84931087e-02
-1.38689369e-01 -4.12826210e-01 2.70566642e-01 4.37195487e-02
-9.08857226e-01 -2.18525920e-02 8.29122126e-01 -4.22775894e-01
9.12084222e-01 7.39465058e-02 8.52977157e-01 4.60471630e-01
5.83980799e-01 4.30399925e-01 1.36140418e+00 -8.75096262e-01
1.67380214e-01 8.26717913e-01 5.96732914e-01 4.93053049e-01
5.12289345e-01 -1.00473166e+00 -6.80129945e-01 -1.86579287e-01
-6.82132840e-02 1.10257685e-01 2.85275310e-01 2.09946074e-02
-5.83460093e-01 1.02516186e+00 -1.02255970e-01 5.17855406e-01
-3.82495791e-01 -7.47226000e-01 8.70610416e-01 4.16485757e-01
7.90637851e-01 3.37665856e-01 -6.78405344e-01 -1.29241422e-01
-5.57065547e-01 -9.37401131e-02 9.71913099e-01 8.65457475e-01
5.58564901e-01 -1.04491025e-01 8.37049842e-01 1.04782450e+00
7.99623191e-01 6.50424182e-01 9.19129908e-01 -2.46343523e-01
4.45998400e-01 9.20843065e-01 1.59132302e-01 -1.69614971e+00
-3.79711568e-01 -2.15022326e-01 -4.79717284e-01 -1.32818505e-01
2.13260688e-02 -5.01986504e-01 -4.52603519e-01 8.79680812e-01
3.28535795e-01 -8.88006687e-01 4.43783373e-01 4.08145338e-01
1.02630389e+00 6.88991785e-01 2.12501474e-02 -4.30264860e-01
1.53988504e+00 -1.08179426e+00 -1.14769197e+00 -2.91891158e-01
5.62434018e-01 -1.53108883e+00 1.02171814e+00 7.99218893e-01
-9.36890602e-01 -2.45665446e-01 -1.18794346e+00 8.39204416e-02
-1.04622054e+00 2.36377999e-01 7.20993757e-01 1.15964389e+00
-8.02350402e-01 1.80418357e-01 -5.68734109e-01 -7.79250741e-01
-9.99737605e-02 3.52373898e-01 -1.40640527e-01 2.33019412e-01
-1.25376487e+00 1.15814352e+00 -2.31786855e-02 2.14104995e-01
4.53668326e-01 2.33445048e-01 -8.31376493e-01 -4.90259260e-01
1.72124088e-01 1.25833482e-01 1.08929718e+00 -1.40481865e+00
-1.69744229e+00 7.58791149e-01 -2.86895812e-01 9.27676335e-02
-3.93525772e-02 -2.73554564e-01 -9.72123981e-01 1.86995730e-01
2.94859171e-01 -4.95377928e-02 5.24513185e-01 -8.74420702e-01
-1.02777290e+00 -5.00841081e-01 3.69391173e-01 3.18921983e-01
-9.71554279e-01 7.37670422e-01 -3.27502131e-01 -7.38294721e-01
4.49013971e-02 -8.89638543e-01 -2.34152302e-01 -6.29289806e-01
4.27874457e-03 -3.09904218e-01 6.62476540e-01 -4.52023685e-01
1.45629799e+00 -1.98721397e+00 -8.09852928e-02 5.02640545e-01
-6.75722733e-02 2.24701881e-01 2.04006165e-01 1.06790698e+00
2.59031683e-01 3.84560496e-01 1.21555708e-01 -7.69715309e-02
-4.98129241e-02 -5.06742820e-02 -2.83204347e-01 1.84651107e-01
-1.29977763e-01 3.94907713e-01 -8.77764761e-01 -5.62525213e-01
-1.25471503e-01 4.25685883e-01 -3.64735238e-02 -2.56637841e-01
2.53470331e-01 -1.35640889e-01 -8.02495718e-01 9.39784825e-01
4.88054693e-01 -2.00382859e-01 6.07738078e-01 -3.09367105e-02
-4.74326819e-01 3.94198805e-01 -1.02437973e+00 9.85726178e-01
-5.91445684e-01 4.24646378e-01 -8.91308784e-02 -8.28901052e-01
1.15444756e+00 4.56614107e-01 5.15984237e-01 -4.52593297e-01
5.87814569e-01 7.03536451e-01 -1.36265084e-01 -4.50591624e-01
1.13535917e+00 -4.71747778e-02 -2.96519309e-01 7.16469467e-01
-1.83567941e-01 -4.08952475e-01 8.44362438e-01 4.31890249e-01
3.14537734e-01 -1.31785378e-01 7.89548635e-01 -2.80734897e-01
9.98469472e-01 3.28275025e-01 -6.30171373e-02 6.63750470e-02
-2.28962392e-01 2.43102968e-01 5.82639456e-01 -4.32295829e-01
-7.76273549e-01 -3.55211020e-01 -9.29792691e-03 1.17913055e+00
7.53482953e-02 -1.01016092e+00 -6.69809997e-01 -7.95297742e-01
-3.04899126e-01 9.78719890e-02 -4.79384422e-01 4.30447370e-01
-1.24929890e-01 -9.45654273e-01 -1.22387350e-01 1.99844897e-01
2.66193181e-01 -1.08888829e+00 -2.32321322e-01 2.35386714e-01
-2.43478939e-01 -8.73956621e-01 -1.34013101e-01 8.86597633e-02
-7.45942235e-01 -1.01406884e+00 -5.91631591e-01 -1.00203073e+00
8.68252337e-01 5.89310110e-01 9.69551325e-01 2.99102992e-01
4.47764546e-01 3.48649651e-01 -1.31071770e+00 -8.66225421e-01
-2.28194356e-01 5.70293248e-01 2.25679725e-01 -2.23511130e-01
8.90707731e-01 -3.75415161e-02 -2.61370271e-01 2.95236409e-01
-9.70545471e-01 -3.90436202e-01 1.47783443e-01 4.08895969e-01
1.31148443e-01 3.86130393e-01 1.30261815e+00 -1.17298162e+00
1.31863654e+00 -8.25603426e-01 -3.84222716e-01 1.83056489e-01
-9.81625915e-01 -4.67780799e-01 7.64312863e-01 3.08492333e-02
-8.42674017e-01 -1.83959126e-01 -2.14454696e-01 9.92035627e-01
3.16957086e-01 1.35733664e+00 2.79181540e-01 -2.30540782e-01
5.88187039e-01 -1.93702564e-01 1.33307934e-01 -3.55561972e-02
1.85572654e-01 1.27181435e+00 -3.85280281e-01 -1.28130391e-01
3.70604396e-01 2.20689639e-01 -4.32674408e-01 -9.22090828e-01
-1.20466828e+00 -5.50008535e-01 -5.96841097e-01 -5.62253833e-01
6.16376221e-01 -6.67090237e-01 -3.27223718e-01 5.87402463e-01
-8.02286386e-01 3.24563593e-01 1.04478940e-01 4.54873919e-01
8.15010890e-02 6.20359242e-01 -7.39266336e-01 -8.85183275e-01
-6.22541904e-01 -1.18669534e+00 1.53028160e-01 5.54293036e-01
-6.29370034e-01 -1.11058068e+00 1.28978461e-01 5.48338890e-01
4.49350715e-01 -9.37319770e-02 4.38037097e-01 -6.29844069e-01
5.73987067e-01 -6.50849223e-01 5.09979762e-02 6.36408925e-01
5.41203916e-01 5.70148766e-01 -4.82845485e-01 -2.12931801e-02
1.55445069e-01 -3.01508069e-01 2.01453820e-01 1.40086368e-01
3.44686210e-01 -2.26470053e-01 -3.33689377e-02 -4.29888725e-01
1.30168033e+00 4.64456707e-01 4.68723595e-01 9.18113947e-01
2.10090399e-01 9.97655928e-01 1.16681838e+00 5.23866236e-01
6.41603112e-01 5.62981963e-02 1.83205634e-01 2.83232570e-01
7.27126777e-01 3.03992212e-01 6.01784229e-01 1.84023106e+00
-2.37992436e-01 -4.21321064e-01 -8.41978967e-01 4.50665385e-01
-1.65098524e+00 -5.34673333e-01 -3.62432539e-01 1.84377110e+00
7.19903648e-01 2.44307548e-01 1.58942148e-01 4.66611564e-01
6.78624094e-01 7.48631805e-02 2.03716140e-02 -1.07259500e+00
-3.15594137e-01 3.97311777e-01 2.83460528e-01 4.28876430e-01
-1.01432645e+00 8.84043813e-01 6.78213692e+00 3.03875357e-01
-1.23075056e+00 1.11639954e-01 6.13988221e-01 2.63717324e-01
-2.95963138e-01 -1.65027902e-01 -8.58326018e-01 3.65236968e-01
8.05964589e-01 -2.98065752e-01 4.96083908e-02 7.63845026e-01
1.55538052e-01 -4.55529273e-01 -8.59448090e-02 5.50069273e-01
7.26808071e-01 -9.36068177e-01 4.00628150e-02 -1.43160224e-01
9.87296283e-01 -3.62268882e-03 2.47775748e-01 -2.23989375e-02
1.17324962e-04 -5.83182156e-01 5.24881721e-01 1.03706405e-01
3.20046425e-01 -1.09578776e+00 1.34126019e+00 -5.71723990e-02
-1.11014223e+00 8.55459943e-02 -2.45582849e-01 -6.44530296e-01
3.10240120e-01 5.98516166e-01 -4.48599190e-01 5.71254730e-01
8.87241960e-01 1.19017220e+00 -6.90219760e-01 5.73865354e-01
-4.38206047e-01 4.15421993e-01 -1.41665235e-01 -9.01636183e-01
3.67299318e-01 -8.25518191e-01 2.11499841e-03 1.13013732e+00
4.66493487e-01 7.49764591e-02 -7.52773881e-02 -4.64278758e-01
2.70721823e-01 1.24397433e+00 -4.50908482e-01 -4.29911911e-01
1.45660877e-01 1.62060773e+00 -1.34393740e+00 -5.16207695e-01
-9.35874045e-01 6.17667735e-01 3.84536944e-02 -1.13501223e-02
-4.04924601e-01 -8.95593166e-01 1.64889663e-01 1.67573765e-01
1.35200461e-02 -3.95334065e-01 -2.52270192e-01 -1.34552133e+00
-2.59498656e-02 -1.17391431e+00 4.05167907e-01 -7.68252790e-01
-1.43741739e+00 1.03987336e+00 -3.19114089e-01 -1.33685899e+00
-3.72412018e-02 -9.32720184e-01 -2.34648660e-01 4.72998798e-01
-1.52738822e+00 -7.71663070e-01 7.47797042e-02 3.76191616e-01
5.70759296e-01 -5.81058562e-01 8.84909391e-01 4.81075525e-01
-4.59863186e-01 1.85865521e-01 2.66830027e-01 -9.84047204e-02
1.18492174e+00 -1.02836072e+00 -3.40909883e-02 6.04780674e-01
-2.35390335e-01 8.85632932e-01 7.30803072e-01 -7.42944360e-01
-1.13087809e+00 -1.97350532e-01 1.47315633e+00 -2.93245435e-01
1.27385604e+00 7.47908726e-02 -4.02286053e-01 3.89470786e-01
7.21240103e-01 -7.02272534e-01 1.42397761e+00 2.29523152e-01
-1.76502645e-01 -1.36859134e-01 -1.05407894e+00 7.75566816e-01
4.45148982e-02 -2.79253215e-01 -5.28497934e-01 4.47866827e-01
2.25289360e-01 -1.48575455e-02 -8.02148044e-01 -4.12392579e-02
9.03679550e-01 -7.09232926e-01 4.21611011e-01 -2.34318122e-01
8.29881132e-01 -3.68410498e-01 -1.89678892e-01 -1.29128122e+00
1.73042879e-01 -3.11936259e-01 2.94902712e-01 1.25246978e+00
8.79092813e-01 -8.90886605e-01 7.34498084e-01 5.99074364e-01
9.51773748e-02 -6.69790566e-01 -1.75901443e-01 9.10066813e-02
-6.73584491e-02 -4.55687314e-01 3.49891365e-01 1.03712773e+00
9.64998662e-01 5.87946653e-01 -2.62545884e-01 -2.36969620e-01
2.04367936e-02 -1.34017348e-01 5.09486496e-01 -1.27287090e+00
3.68010789e-01 -4.88050908e-01 -3.26522171e-01 -6.17307663e-01
-1.26106203e-01 -4.68789369e-01 -5.01978397e-01 -1.66590083e+00
-9.71146524e-02 -3.92778397e-01 -2.78863460e-01 2.01544687e-01
-4.96947381e-04 6.41942620e-01 2.12316707e-01 2.19170749e-01
-6.55171156e-01 1.46248834e-02 1.12820029e+00 1.15313232e-01
-3.24867249e-01 -4.69591208e-02 -1.27213717e+00 1.10275376e+00
1.18144631e+00 -5.49785316e-01 -4.34653491e-01 -1.63267627e-01
1.65035343e+00 -4.21456903e-01 -8.69892120e-01 -4.17990565e-01
2.44573444e-01 -2.35039935e-01 4.30691242e-02 -8.57356906e-01
6.41126931e-02 -9.29406226e-01 -3.95708710e-01 7.75952786e-02
4.41454984e-02 8.43116999e-01 -1.74921993e-02 1.89423740e-01
-6.66674674e-01 -8.62810135e-01 4.06088382e-01 -3.70836705e-02
-4.62671518e-01 -1.74097210e-01 -1.24914372e+00 -3.21864039e-01
9.86442208e-01 -1.45021006e-01 -1.46457955e-01 -5.74551284e-01
-6.15873754e-01 1.76698402e-01 4.80633110e-01 5.52134752e-01
5.31063557e-01 -9.76540208e-01 -4.47343230e-01 -1.38087958e-01
2.75407553e-01 -4.81109947e-01 -6.14754707e-02 6.13762677e-01
-8.86895776e-01 1.03956811e-01 -4.31409776e-01 1.03548437e-01
-1.31348395e+00 3.15605730e-01 -1.63588360e-01 -1.87714502e-01
2.20061526e-01 3.21828157e-01 -5.04361808e-01 -5.18165946e-01
-1.66465282e-01 -1.91716589e-02 -1.36485684e+00 8.83554399e-01
7.49071538e-01 3.22238296e-01 3.11728209e-01 -1.07541299e+00
-2.55444139e-01 8.33553135e-01 -3.56823832e-01 -4.55142647e-01
1.32807040e+00 -6.03468895e-01 -8.44718039e-01 7.99863100e-01
6.60698891e-01 8.21782947e-01 9.53600332e-02 1.32999584e-01
6.01830706e-02 -2.78538734e-01 -7.48226196e-02 -7.74134696e-01
-1.12500715e+00 4.09131914e-01 8.75989571e-02 1.01378834e+00
1.14354122e+00 -3.39810997e-01 6.82225704e-01 5.94998360e-01
2.16758728e-01 -1.71895075e+00 1.93363912e-02 8.47240925e-01
5.66076875e-01 -1.34258294e+00 5.01070678e-01 -6.58044517e-01
-1.10877180e+00 1.46271324e+00 3.02957982e-01 -1.51454896e-01
1.26280034e+00 2.59379774e-01 7.47267008e-01 -2.73167580e-01
-2.56490171e-01 7.48409554e-02 6.79000840e-02 4.88101035e-01
1.26088130e+00 -1.27840087e-01 -1.30709231e+00 8.23525608e-01
-5.41558981e-01 -2.06282273e-01 1.10466146e+00 1.27507699e+00
-6.42441630e-01 -1.68220341e+00 -4.15366173e-01 5.39532125e-01
-1.11478758e+00 -1.75869733e-01 -5.82749546e-01 5.17256021e-01
-2.97681779e-01 1.66402042e+00 -2.77997673e-01 -3.01160008e-01
2.57736325e-01 -6.75918087e-02 -1.35099456e-01 -7.10197866e-01
-9.04646993e-01 1.73546910e-01 3.12220603e-01 2.33201966e-01
-1.14704156e+00 -7.94120431e-01 -1.28843153e+00 -3.87872368e-01
-8.07963610e-01 8.20911825e-01 1.18654835e+00 8.74966681e-01
1.92412779e-01 1.06899701e-01 7.85152912e-01 -3.85617286e-01
1.18762940e-01 -9.82251167e-01 -8.54861557e-01 -5.58148399e-02
-9.52084735e-02 -4.51451123e-01 -4.04807329e-01 1.44174352e-01] | [11.0110502243042, 6.894454002380371] |
91bea4ed-85c1-4a51-beed-afc9a196139d | bel-a-bag-embedding-loss-for-transformer | 2303.01377 | null | https://arxiv.org/abs/2303.01377v1 | https://arxiv.org/pdf/2303.01377v1.pdf | BEL: A Bag Embedding Loss for Transformer enhances Multiple Instance Whole Slide Image Classification | Multiple Instance Learning (MIL) has become the predominant approach for classification tasks on gigapixel histopathology whole slide images (WSIs). Within the MIL framework, single WSIs (bags) are decomposed into patches (instances), with only WSI-level annotation available. Recent MIL approaches produce highly informative bag level representations by utilizing the transformer architecture's ability to model the dependencies between instances. However, when applied to high magnification datasets, problems emerge due to the large number of instances and the weak supervisory learning signal. To address this problem, we propose to additionally train transformers with a novel Bag Embedding Loss (BEL). BEL forces the model to learn a discriminative bag-level representation by minimizing the distance between bag embeddings of the same class and maximizing the distance between different classes. We evaluate BEL with the Transformer architecture TransMIL on two publicly available histopathology datasets, BRACS and CAMELYON17. We show that with BEL, TransMIL outperforms the baseline models on both datasets, thus contributing to the clinically highly relevant AI-based tumor classification of histological patient material. | ['Carsten Marr', 'Nassir Navab', 'Francesco Paolo Casale', 'Ario Sadafi', 'Daniel Sens'] | 2023-03-02 | null | null | null | null | ['whole-slide-images', 'multiple-instance-learning'] | ['computer-vision', 'methodology'] | [ 4.43219304e-01 2.53041744e-01 -3.08717757e-01 -3.88588905e-01
-1.44769263e+00 -2.73967296e-01 5.43482900e-01 7.62452841e-01
-4.04410660e-01 8.19584846e-01 2.53418505e-01 -4.40587364e-02
-2.41273031e-01 -7.74369895e-01 -8.41001570e-01 -1.34678209e+00
-6.22112211e-03 5.20935595e-01 2.91015863e-01 -8.08510333e-02
-1.89912573e-01 4.93045896e-01 -1.00079834e+00 7.03075469e-01
3.81517142e-01 1.11956441e+00 -1.12614930e-01 5.95843852e-01
2.02835798e-02 8.32410097e-01 -4.05484587e-01 -3.17839384e-01
6.24947250e-02 -8.80054682e-02 -8.34632993e-01 8.34795274e-03
7.06228256e-01 -5.92970252e-02 -2.73487836e-01 5.51514566e-01
4.31642771e-01 -2.88662612e-01 8.86628866e-01 -1.22829473e+00
-3.75868549e-04 3.99099499e-01 -5.25238276e-01 4.02144343e-01
-3.59063983e-01 4.26485166e-02 1.42755902e+00 -7.11169243e-01
6.78336859e-01 7.78257430e-01 7.97723174e-01 4.37291652e-01
-1.71669817e+00 -3.91774118e-01 3.50587107e-02 9.33484361e-02
-1.28729892e+00 -2.96534240e-01 5.48959076e-01 -4.48866993e-01
9.03498590e-01 5.03528118e-01 7.17279017e-01 9.88895237e-01
5.23469329e-01 1.04777634e+00 1.21650076e+00 -4.08417672e-01
1.11612082e-01 2.36470163e-01 4.50453192e-01 7.86149919e-01
1.38022169e-01 -3.27206612e-01 -6.33459508e-01 -2.84909606e-01
4.59060609e-01 2.56319761e-01 -2.78940916e-01 -4.59900618e-01
-1.17764378e+00 7.97184825e-01 8.03979874e-01 4.15892988e-01
-8.28022733e-02 4.05814320e-01 6.22815192e-01 9.86383706e-02
7.07463980e-01 3.05442572e-01 -2.04097092e-01 3.21026087e-01
-9.03738260e-01 1.56815916e-01 4.42270011e-01 5.26587784e-01
5.11195123e-01 -4.18805093e-01 -3.64508569e-01 7.74847806e-01
1.08079195e-01 1.39005790e-02 5.07104218e-01 -1.26483798e-01
9.13079977e-02 8.60998333e-01 -4.90962893e-01 -6.92252576e-01
-4.60784137e-01 -7.68252194e-01 -9.21939671e-01 -1.27442256e-02
4.87723976e-01 5.60477972e-01 -8.63910139e-01 1.63958120e+00
2.84074008e-01 2.41633192e-01 -1.16347335e-01 4.43334848e-01
9.71501529e-01 3.44230533e-01 2.18216076e-01 1.08439833e-01
1.46270764e+00 -7.72619843e-01 -6.51460290e-01 -2.97952354e-01
1.12405026e+00 -2.04679132e-01 9.15867984e-01 2.50668079e-01
-7.93901861e-01 7.34628960e-02 -9.33643818e-01 -1.32362366e-01
-5.97193241e-01 2.02661484e-01 6.47431135e-01 2.54433841e-01
-1.12507987e+00 3.46739501e-01 -1.05715561e+00 -1.83649287e-01
1.00738275e+00 5.50914824e-01 -7.76585877e-01 -1.52397707e-01
-9.16440010e-01 8.15852404e-01 9.48270410e-03 2.40195803e-02
-1.00570142e+00 -1.17042220e+00 -8.69597971e-01 2.33476892e-01
-1.30959945e-02 -4.84001756e-01 7.53347397e-01 -6.43648863e-01
-9.80641007e-01 1.22656775e+00 -7.61744380e-02 -5.37721694e-01
3.59143823e-01 4.14150387e-01 1.54678553e-01 2.71874517e-01
6.63930252e-02 6.98765337e-01 6.04582727e-01 -1.08132231e+00
-4.16287184e-01 -6.01559281e-01 6.58728182e-02 -6.45182803e-02
-6.24684751e-01 -4.89684790e-01 -1.27884507e-01 -5.18542886e-01
9.48354825e-02 -9.25250947e-01 -2.27162614e-01 1.37218803e-01
-4.56484854e-01 -2.73736656e-01 5.63138485e-01 -6.40149951e-01
9.33324933e-01 -2.35509992e+00 3.70047569e-01 1.86310247e-01
5.03637314e-01 -8.49975273e-02 -8.62246007e-02 1.78014666e-01
-1.61972806e-01 4.61901464e-02 -1.67537943e-01 -6.83887064e-01
-8.90864208e-02 4.05179769e-01 -2.53301054e-01 6.51812434e-01
3.31102461e-01 9.29975510e-01 -7.53525615e-01 -7.04598963e-01
2.47682389e-02 4.74247456e-01 -6.74387157e-01 1.63572520e-01
3.37422416e-02 2.52865523e-01 -1.20324410e-01 6.37242615e-01
3.09731066e-01 -5.86454868e-01 2.39536941e-01 -5.52967191e-01
4.86620426e-01 2.07907438e-01 -4.08110023e-01 1.51681054e+00
-4.75139439e-01 7.28599429e-01 4.78250496e-02 -1.25011599e+00
4.31522548e-01 4.18124229e-01 7.41343439e-01 -4.24793005e-01
-4.92751636e-02 1.84294745e-01 -9.53488275e-02 -1.01675324e-01
-2.37666175e-01 -5.72840869e-01 -1.08897343e-01 7.40820915e-02
4.45179313e-01 -1.88793346e-01 3.46498564e-02 3.52887243e-01
1.41245961e+00 -4.17284906e-01 3.43652457e-01 -3.70007575e-01
3.10543001e-01 -2.70105377e-02 5.09996712e-01 6.15002751e-01
-2.36504033e-01 5.33610761e-01 9.00387883e-01 -4.62194979e-01
-7.83813059e-01 -1.17387724e+00 -7.30701923e-01 9.38988447e-01
-2.03014746e-01 -4.44723308e-01 -3.52494121e-01 -1.05211020e+00
1.73473835e-01 2.99296796e-01 -1.13544619e+00 -2.37609476e-01
-4.61428106e-01 -1.20551217e+00 5.36361158e-01 6.05025351e-01
-1.02864370e-01 -4.60289240e-01 -3.65472406e-01 1.88033074e-01
-2.52234489e-01 -1.09775448e+00 -2.12820277e-01 6.49794877e-01
-9.01234865e-01 -1.06211877e+00 -4.57787961e-01 -6.90104961e-01
1.04778183e+00 4.54102941e-02 1.09689426e+00 1.60266548e-01
-1.04959953e+00 3.51079524e-01 -1.37807950e-01 -5.45879126e-01
-2.37561345e-01 1.15233846e-01 -3.75894517e-01 5.78136630e-02
3.74452502e-01 -3.37123543e-01 -6.42606139e-01 8.49692225e-02
-1.07818842e+00 2.86312222e-01 6.36854887e-01 1.48926079e+00
8.57026577e-01 -2.07125559e-01 5.33565700e-01 -9.14258540e-01
6.34947643e-02 -3.76239747e-01 -1.93357795e-01 3.48129213e-01
-2.48201579e-01 -2.08960511e-02 4.48304981e-01 -1.09633811e-01
-4.80143219e-01 -1.78892732e-01 -9.19321924e-02 -1.84575409e-01
1.72377944e-01 4.82393712e-01 2.00498894e-01 -3.80979061e-01
5.50915539e-01 1.59975320e-01 3.21712166e-01 -1.95437595e-02
-1.92642599e-01 5.06075263e-01 3.76170985e-02 -3.45226139e-01
2.07375154e-01 8.23016405e-01 3.52824509e-01 -6.24942482e-01
-1.26250887e+00 -7.46059597e-01 -5.33042967e-01 -1.63180724e-01
7.77726650e-01 -8.38472903e-01 -6.57982826e-01 2.59074152e-01
-6.24887407e-01 -4.91920084e-01 -4.59073454e-01 2.26315826e-01
-5.46580970e-01 -4.81039146e-03 -9.64858174e-01 -3.09960008e-01
-3.83799762e-01 -1.46429467e+00 1.59779799e+00 -2.50961185e-01
-1.97607934e-01 -1.32149649e+00 2.09132284e-01 5.30140162e-01
2.89299309e-01 6.73411429e-01 1.27748096e+00 -7.73584723e-01
-4.42052811e-01 -6.71896040e-01 -1.75189495e-01 3.01505983e-01
2.25861862e-01 -8.58857110e-02 -1.17413783e+00 -5.48616171e-01
-2.45226502e-01 -6.11216247e-01 1.28320956e+00 4.01756823e-01
1.46856189e+00 -3.18237156e-01 -7.43558407e-01 6.15602434e-01
1.63106143e+00 -4.22440529e-01 3.77988845e-01 2.26684168e-01
5.76792777e-01 5.53089499e-01 2.67399251e-01 1.00523040e-01
2.68719792e-01 7.50671685e-01 5.78684807e-01 -4.22634810e-01
-1.75003067e-01 -5.28427251e-02 1.13748414e-02 5.90996265e-01
1.80749610e-01 -2.46614739e-01 -1.04331529e+00 6.98642850e-01
-1.72000539e+00 -6.91452384e-01 1.25125989e-01 2.02174616e+00
1.05644810e+00 2.27558464e-01 -1.82583183e-01 3.43213677e-01
2.53848396e-02 1.08352415e-01 -2.32152984e-01 8.56566057e-03
-9.75802317e-02 2.29846463e-01 4.47927803e-01 4.52405155e-01
-1.21881497e+00 5.12141883e-01 6.08671045e+00 1.04350841e+00
-1.23037732e+00 3.42682719e-01 1.08487558e+00 -4.65586662e-01
-2.09785298e-01 -3.19308102e-01 -9.49843287e-01 3.56056869e-01
7.93654978e-01 1.42146409e-01 -1.96840346e-01 4.06114161e-01
-1.54418677e-01 -7.68731087e-02 -1.41413283e+00 7.55285084e-01
2.15392083e-01 -1.71260977e+00 1.85067087e-01 3.58240485e-01
4.93462145e-01 -6.07898384e-02 1.58983156e-01 3.89543176e-01
8.96375179e-02 -1.32685840e+00 1.91833213e-01 3.20094973e-01
7.75662482e-01 -4.07841921e-01 1.14402926e+00 1.18846960e-01
-7.84303367e-01 3.47110257e-02 -3.85512739e-01 4.18667108e-01
-4.10662472e-01 7.16688573e-01 -1.41391945e+00 4.19917643e-01
4.76155788e-01 8.00564229e-01 -7.44137049e-01 7.51115620e-01
1.71318442e-01 7.19535649e-01 -3.98081452e-01 1.15956388e-01
4.14970070e-01 1.74696743e-01 2.33544067e-01 1.32458651e+00
-3.04726586e-02 -1.07707530e-01 1.46530345e-01 5.47974169e-01
-6.19268604e-02 6.75451979e-02 -3.62040401e-01 -9.55411512e-03
3.71168368e-02 1.58750641e+00 -6.69429779e-01 -1.60508707e-01
-1.69819504e-01 6.06329381e-01 6.60323024e-01 1.40671149e-01
-5.50671577e-01 -3.40897474e-03 5.50830841e-01 4.69254255e-01
-4.05288562e-02 2.21214160e-01 -3.54559541e-01 -9.28234100e-01
-1.19406246e-01 -6.22150600e-01 7.96823502e-01 -2.55122811e-01
-1.47115123e+00 3.39183390e-01 -1.23472750e-01 -1.14698780e+00
2.19863862e-01 -9.69216347e-01 -5.49127698e-01 4.70336765e-01
-1.77012563e+00 -1.72159839e+00 -2.93338835e-01 3.18966836e-01
2.80380547e-01 -4.15322036e-02 1.07048321e+00 1.92193031e-01
-6.29025578e-01 9.11013186e-01 2.60251433e-01 2.41010323e-01
8.75144780e-01 -1.48778307e+00 -5.22458971e-01 1.50518283e-01
-2.03807186e-02 4.29015905e-01 5.04362762e-01 -1.63703591e-01
-1.24505329e+00 -1.13930893e+00 8.21074903e-01 -6.80225849e-01
7.85660446e-01 -6.74691319e-01 -8.32961798e-01 9.11030829e-01
-2.23733947e-01 7.15490043e-01 1.45292389e+00 1.05997562e-01
-4.84935641e-01 -5.77025115e-01 -1.33180189e+00 3.57416987e-01
5.26145995e-01 -5.19681990e-01 -1.36572257e-01 7.53585041e-01
3.49489123e-01 -3.17134291e-01 -1.30089188e+00 4.76053327e-01
4.16333318e-01 -7.26699173e-01 1.00148261e+00 -6.93487465e-01
5.51165044e-01 -4.57861368e-03 -1.01673089e-01 -1.43982387e+00
-2.99449831e-01 1.99837998e-01 2.31729954e-01 9.45133209e-01
3.83652389e-01 -5.87856591e-01 7.06193328e-01 3.58466864e-01
-2.36034617e-01 -1.41216671e+00 -1.41655421e+00 -4.27831978e-01
3.13374758e-01 6.75301850e-02 2.56368130e-01 7.47852683e-01
4.79893684e-01 4.68836613e-02 9.36895162e-02 7.63322711e-02
8.21391165e-01 1.67754404e-02 4.50619191e-01 -1.03199697e+00
-3.90262723e-01 -4.80688125e-01 -9.77713168e-01 -1.60463557e-01
2.93806434e-01 -1.32139730e+00 2.27734391e-02 -1.41158092e+00
7.10350275e-01 -7.02086687e-01 -6.81118071e-01 8.34937334e-01
-2.45582774e-01 6.28007293e-01 -1.62601873e-01 2.42869079e-01
-6.87577605e-01 4.43770230e-01 1.02091312e+00 -6.37197793e-01
4.81403202e-01 -3.16921055e-01 -5.54655910e-01 4.90618020e-01
4.96331006e-01 -4.83607918e-01 -1.25779971e-01 -1.56479388e-01
2.13761419e-01 -2.54604761e-02 6.90853715e-01 -9.08792794e-01
1.51024282e-01 6.70482591e-02 5.24962306e-01 -3.57691109e-01
5.21915376e-01 -7.34054983e-01 3.72901335e-02 5.14453709e-01
-6.43179893e-01 -6.96455717e-01 3.35197479e-01 6.04313493e-01
-4.68245894e-01 3.23610269e-02 1.00796497e+00 -3.17116268e-03
-1.51477233e-02 5.87826550e-01 -1.80335894e-01 -2.06666127e-01
1.22504258e+00 -1.45042136e-01 -4.84533966e-01 1.91482738e-01
-8.43039572e-01 2.68750876e-01 2.43020445e-01 6.40748953e-03
4.64350820e-01 -1.30115759e+00 -8.55027854e-01 3.14559013e-01
5.83810151e-01 3.21865767e-01 4.98953611e-01 1.22059822e+00
-3.80938679e-01 5.62135637e-01 -7.66307935e-02 -8.90885174e-01
-1.35048175e+00 1.01485468e-01 6.00060761e-01 -9.73329961e-01
-5.09374499e-01 1.22988582e+00 7.97355831e-01 -4.54360694e-01
1.57876611e-01 -1.99291244e-01 -1.42903626e-01 9.41974446e-02
4.69073743e-01 -1.25701968e-02 4.97930855e-01 -2.74921328e-01
-4.64028418e-01 9.37703773e-02 -6.06896877e-01 2.63901412e-01
1.59111214e+00 4.32740867e-01 -3.65131795e-01 6.22957408e-01
1.42602539e+00 -2.38879845e-01 -1.11478651e+00 -4.12269980e-01
-5.60437702e-02 -3.30413848e-01 3.73328567e-01 -6.36788487e-01
-9.08425629e-01 1.01934409e+00 6.01578116e-01 -3.92526463e-02
8.50903869e-01 2.68870115e-01 5.05769491e-01 -5.04573509e-02
2.88906515e-01 -7.12415159e-01 3.62830013e-01 1.68867260e-01
8.91680121e-01 -1.24824500e+00 -2.49895900e-02 -4.32561666e-01
-2.26540565e-01 1.09058273e+00 5.87304831e-01 -5.78543395e-02
5.61176300e-01 5.67042649e-01 -1.13536112e-01 -4.66721565e-01
-1.03197539e+00 2.34172732e-01 2.72465795e-01 2.80465722e-01
6.39656365e-01 2.12609187e-01 -4.55119759e-02 6.01109743e-01
3.27661075e-02 -1.60302788e-01 4.00519013e-01 7.96450019e-01
-1.86691985e-01 -9.60796416e-01 -2.29651872e-02 1.05633020e+00
-6.05740130e-01 -1.28918692e-01 -2.37345099e-01 6.83352709e-01
-1.44797504e-01 4.11244839e-01 3.13587278e-01 -1.72084421e-02
-2.74927523e-02 -9.63054691e-03 5.38675547e-01 -7.11472392e-01
-6.54532731e-01 6.42794184e-03 -1.18741542e-01 -3.59370708e-01
-2.54985511e-01 -5.75498939e-01 -1.02352107e+00 5.79779223e-02
-3.33848745e-01 6.02211431e-02 3.94149929e-01 8.56972277e-01
1.46783605e-01 8.22964311e-01 5.50145388e-01 -6.03099763e-01
-6.26926005e-01 -8.40305567e-01 -7.18666196e-01 5.51516175e-01
6.26599967e-01 -6.26472235e-01 -6.77403033e-01 -7.96100423e-02] | [15.119107246398926, -2.867824077606201] |
2589bb94-6c63-47f4-863c-61d2856f700b | cap-robust-point-cloud-classification-via | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Ding_CAP_Robust_Point_Cloud_Classification_via_Semantic_and_Structural_Modeling_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Ding_CAP_Robust_Point_Cloud_Classification_via_Semantic_and_Structural_Modeling_CVPR_2023_paper.pdf | CAP: Robust Point Cloud Classification via Semantic and Structural Modeling | Recently, deep neural networks have shown great success on 3D point cloud classification tasks, which simultaneously raises the concern of adversarial attacks that cause severe damage to real-world applications. Moreover, defending against adversarial examples in point cloud data is extremely difficult due to the emergence of various attack strategies. In this work, with the insight of the fact that the adversarial examples in this task still preserve the same semantic and structural information as the original input, we design a novel defense framework for improving the robustness of existing classification models, which consists of two main modules: the attention-based pooling and the dynamic contrastive learning. In addition, we also develop an algorithm to theoretically certify the robustness of the proposed framework. Extensive empirical results on two datasets and three classification models show the robustness of our approach against various attacks, e.g., the averaged attack success rate of PointNet decreases from 70.2% to 2.7% on the ModelNet40 dataset under 9 common attacks. | ['Min Yang', 'Wenxuan Li', 'Mi Zhang', 'Yuanmin Huang', 'Erling Jiang', 'Daizong Ding'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['3d-point-cloud-classification', 'point-cloud-classification'] | ['computer-vision', 'computer-vision'] | [-1.24489807e-01 -1.90440118e-01 3.79386634e-01 -2.23060958e-02
-4.45423096e-01 -9.06094253e-01 5.20659626e-01 -1.25004929e-02
-2.57436037e-01 4.58344579e-01 -3.15374494e-01 -4.24265027e-01
-1.73365436e-02 -8.77974451e-01 -8.10791612e-01 -8.87262583e-01
-2.04228893e-01 6.13250351e-03 3.78694892e-01 -3.70776743e-01
4.44308907e-01 1.22111273e+00 -1.09299684e+00 -3.04445112e-03
7.99472332e-01 1.30494678e+00 -4.08861995e-01 2.92435050e-01
5.75239472e-02 5.01840472e-01 -1.03598416e+00 -6.85403109e-01
6.53582215e-01 2.54271120e-01 -4.19755429e-01 -3.60254079e-01
4.77772951e-01 -4.60846037e-01 -8.72879684e-01 1.47012508e+00
5.79943538e-01 -2.04163194e-02 3.08757246e-01 -1.64754081e+00
-6.73487067e-01 1.70399189e-01 -6.17633522e-01 3.91607702e-01
-4.50422652e-02 4.70318079e-01 5.98686874e-01 -5.61547756e-01
1.42912373e-01 1.24029720e+00 3.17962736e-01 6.25967503e-01
-6.25677109e-01 -1.20866573e+00 2.56182998e-01 2.42853150e-01
-1.34919977e+00 -3.60118188e-02 9.09620464e-01 -2.06494987e-01
7.01079249e-01 3.11203033e-01 2.61120349e-01 1.29969037e+00
4.68872994e-01 4.61117655e-01 8.14301431e-01 2.57118911e-01
1.81128129e-01 -2.66011655e-02 -2.40650579e-01 2.91351974e-01
3.90989840e-01 3.55044693e-01 5.38618630e-03 -4.89038914e-01
6.11607492e-01 1.74406484e-01 -3.05571765e-01 -2.37608761e-01
-8.15816164e-01 7.33812451e-01 1.04163146e+00 6.91208616e-02
-1.78142697e-01 1.05342768e-01 4.85899955e-01 2.46597975e-01
3.32197189e-01 4.25836056e-01 -2.66890883e-01 3.72587055e-01
-2.11963847e-01 3.62973273e-01 4.74350721e-01 8.52475166e-01
2.98239768e-01 2.57787704e-01 2.66429633e-01 2.43113920e-01
2.42385179e-01 7.05001593e-01 2.52283588e-02 -5.07530510e-01
7.09558427e-01 5.16565681e-01 7.81036913e-02 -1.62107003e+00
-8.81154388e-02 -7.16839612e-01 -1.11076534e+00 5.77694178e-01
1.88924536e-01 -1.13651685e-01 -6.45540774e-01 1.74122036e+00
3.67003232e-01 6.80473924e-01 1.44501314e-01 8.30092728e-01
5.24088979e-01 5.88418365e-01 1.97159573e-01 3.30336392e-01
8.22736979e-01 -4.78972644e-01 -3.08396101e-01 -9.40312445e-02
6.76100031e-02 -5.63525498e-01 7.41407394e-01 2.77343035e-01
-7.72316039e-01 -5.49598455e-01 -1.38621044e+00 4.51257139e-01
-4.89567965e-01 -5.69442987e-01 4.29013610e-01 7.29273260e-01
-4.16427702e-01 6.33854866e-01 -8.03413749e-01 1.31255850e-01
7.75666475e-01 5.14519155e-01 -4.00054008e-01 1.13131225e-01
-1.36751926e+00 7.73363650e-01 2.53014982e-01 1.91327661e-01
-1.05131328e+00 -7.85892725e-01 -3.99276137e-01 1.94381297e-01
1.49875477e-01 -4.32820946e-01 7.89314389e-01 -6.80893302e-01
-1.28616333e+00 6.93482816e-01 4.90254194e-01 -6.46298528e-01
7.53456235e-01 -4.62511122e-01 -4.74592149e-01 3.38197857e-01
-1.23134382e-01 3.25605035e-01 1.04030383e+00 -1.28408325e+00
-5.14068067e-01 -5.67089379e-01 4.30334151e-01 5.81245311e-02
-5.76259971e-01 3.93820331e-02 -1.41013801e-01 -6.85790658e-01
-7.63623640e-02 -1.00525737e+00 -3.36851954e-01 2.78659791e-01
-5.29789031e-01 -9.29420218e-02 1.16243243e+00 -1.38588473e-01
6.95709825e-01 -2.60083938e+00 5.75109273e-02 2.96086222e-01
3.61134052e-01 5.49571097e-01 -5.44044673e-02 1.29838079e-01
-2.78086722e-01 4.59416062e-01 -3.42278540e-01 9.45995227e-02
-4.89126053e-03 4.23362441e-02 -1.08090508e+00 8.03507924e-01
4.89379048e-01 7.15249717e-01 -7.40853369e-01 -1.16821110e-01
3.82994652e-01 3.36213291e-01 -4.82588142e-01 2.48545229e-01
-4.00986522e-03 5.14028907e-01 -7.72193134e-01 5.87656140e-01
1.14100242e+00 2.15725396e-02 -4.78328496e-01 8.72588158e-02
2.42843404e-01 -1.00877903e-01 -8.51208746e-01 1.20563674e+00
-2.54131705e-01 3.79526019e-01 -5.21336496e-03 -8.26182783e-01
9.48482215e-01 1.66267097e-01 4.39737648e-01 -4.56706703e-01
4.23815966e-01 1.53534561e-01 1.57270014e-01 -1.76035449e-01
8.52651298e-02 -5.73785417e-02 -3.37233514e-01 2.04371825e-01
-2.72528440e-01 -1.37920633e-01 -6.11406922e-01 2.44423330e-01
1.29984641e+00 -4.12785321e-01 -2.68675297e-01 -9.15521085e-02
7.37516820e-01 -1.97681129e-01 4.89955664e-01 7.04138458e-01
-5.04812062e-01 4.79229748e-01 4.61356819e-01 -5.84406972e-01
-8.40488732e-01 -1.28303981e+00 -1.71926320e-01 1.84262097e-01
7.67247438e-01 -3.59547287e-02 -5.63848019e-01 -1.06164873e+00
2.89935410e-01 5.99586368e-01 -5.81431985e-01 -8.04600477e-01
-6.08800411e-01 -6.63884997e-01 1.07378113e+00 3.09060156e-01
9.40551400e-01 -9.02117848e-01 -4.28774983e-01 -8.35823417e-02
8.16175342e-02 -1.41329908e+00 6.22001439e-02 -2.85788745e-01
-6.69791818e-01 -1.16803014e+00 -2.72717893e-01 -3.81229728e-01
5.27099252e-01 4.74733412e-01 7.63252974e-01 5.02041340e-01
-2.30855066e-02 -6.52260799e-03 -2.94555932e-01 -6.82281971e-01
-2.70886838e-01 5.19018397e-02 2.99469918e-01 1.31830648e-01
2.51148432e-01 -7.88810909e-01 -5.51973701e-01 4.67701048e-01
-1.01235616e+00 -6.31480217e-01 4.94439572e-01 4.37221408e-01
2.53422976e-01 3.45391035e-01 6.36998713e-01 -4.33883399e-01
5.60699701e-01 -8.26932967e-01 -7.07778871e-01 -1.53480813e-01
-1.05388083e-01 -3.19816738e-01 9.89257991e-01 -4.80157495e-01
-5.73439121e-01 -2.16226265e-01 -4.09138322e-01 -1.08418250e+00
-3.31738055e-01 5.19637391e-02 -6.24619603e-01 -6.68635368e-01
6.70220315e-01 8.79087523e-02 -1.45522431e-01 -2.21391782e-01
1.45529240e-01 4.44947422e-01 6.42937422e-01 -5.17446935e-01
1.63423193e+00 7.44384944e-01 3.08650106e-01 -7.50676394e-01
-5.42119443e-01 1.70149460e-01 -1.84950560e-01 -1.38542965e-01
6.60080194e-01 -6.54341996e-01 -1.01067996e+00 8.23170960e-01
-1.27621090e+00 2.14221165e-01 7.96820596e-02 1.52786240e-01
-3.09743822e-01 7.21812069e-01 -4.15564060e-01 -4.52366054e-01
-4.30688798e-01 -1.29879594e+00 7.86383688e-01 1.74509943e-01
3.62204790e-01 -5.65776944e-01 -2.09588200e-01 1.10521413e-01
4.08978045e-01 7.65350342e-01 9.92695689e-01 -8.74285817e-01
-9.21824336e-01 -5.68491518e-01 -2.70156950e-01 5.08950770e-01
-4.42182049e-02 1.57448143e-01 -9.87379313e-01 -4.67429817e-01
4.24847960e-01 -1.27867341e-01 4.59500313e-01 -1.96507171e-01
1.51190114e+00 -1.53363109e-01 -1.60928339e-01 8.55454624e-01
1.36854649e+00 3.08131248e-01 8.90540719e-01 4.73361671e-01
6.31233096e-01 2.24662438e-01 4.76870477e-01 2.34207496e-01
-2.82666415e-01 5.45415998e-01 1.33298635e+00 1.20871641e-01
3.87407213e-01 -1.21826001e-01 1.18762277e-01 1.07430011e-01
1.40407428e-01 -4.73906189e-01 -9.73649383e-01 2.91483253e-01
-1.51865494e+00 -1.00673091e+00 1.17800795e-01 2.02849483e+00
1.62389293e-01 6.91221178e-01 -2.58603603e-01 1.23687282e-01
7.93142915e-01 3.82819116e-01 -8.19848716e-01 -1.85107976e-01
-1.04429252e-01 2.47338593e-01 5.74463665e-01 2.64800005e-02
-1.34578753e+00 9.84842122e-01 5.34469986e+00 7.42321193e-01
-1.37016177e+00 -1.01350583e-01 4.00762677e-01 -7.28494525e-02
1.40538782e-01 -2.47968718e-01 -3.84696722e-01 7.45115817e-01
6.78397059e-01 -3.41296703e-01 1.29878014e-01 1.06185913e+00
-1.39902055e-01 5.48679173e-01 -7.04863727e-01 7.77983725e-01
-5.62480055e-02 -1.21713769e+00 3.34317327e-01 2.76925355e-01
3.66256952e-01 2.14275107e-01 5.61663508e-01 1.64393798e-01
3.76199603e-01 -9.30599391e-01 6.52886689e-01 6.23407774e-02
4.83312070e-01 -1.20986450e+00 8.82442653e-01 4.52664644e-01
-8.45212400e-01 -1.55414253e-01 -5.99413455e-01 7.94888958e-02
2.81266794e-02 1.66333079e-01 -3.74530166e-01 7.82640457e-01
9.39594150e-01 4.98549610e-01 -4.75634396e-01 1.07337904e+00
-5.20146072e-01 3.32501680e-01 -5.47419250e-01 1.75905377e-01
4.62255538e-01 2.27467149e-01 1.00655615e+00 5.05979061e-01
5.54291010e-02 1.37853041e-01 -9.37790871e-02 7.44853854e-01
-2.83651948e-01 -1.73129886e-01 -9.65520442e-01 1.94950595e-01
6.22523129e-01 9.85086560e-01 -4.68390197e-01 2.00692117e-01
-2.93021411e-01 7.81947851e-01 2.95453042e-01 1.44010514e-01
-1.20460701e+00 -5.72677553e-01 1.25629663e+00 -2.04325095e-01
2.54797220e-01 -4.28301066e-01 -3.37822050e-01 -1.11808765e+00
1.43892154e-01 -9.03178215e-01 1.38804123e-01 -4.72984582e-01
-1.44911015e+00 9.15483594e-01 -1.05482511e-01 -1.53432608e+00
2.90212512e-01 -6.72782421e-01 -1.01320839e+00 7.27484167e-01
-1.39280462e+00 -8.95519197e-01 -4.74327922e-01 8.60812604e-01
1.85529411e-01 -4.38503653e-01 6.11599386e-01 4.28459734e-01
-6.54582322e-01 8.89396548e-01 -1.84794843e-01 4.40254658e-01
3.81976992e-01 -7.70246387e-01 9.78429019e-01 1.13026226e+00
-2.85317190e-02 6.40640616e-01 5.44090509e-01 -6.11069560e-01
-1.27146769e+00 -1.19705963e+00 5.87358326e-02 -5.47954261e-01
8.85543227e-01 -3.90538603e-01 -1.10552585e+00 4.77366656e-01
-7.49681727e-04 2.24955559e-01 4.47576731e-01 -5.12750626e-01
-5.45944154e-01 -4.59116101e-02 -1.39232290e+00 8.21261406e-01
8.76042902e-01 -4.43890840e-01 -5.58837593e-01 2.64295369e-01
1.02172947e+00 -5.02135754e-01 -6.51524007e-01 6.14730179e-01
2.56250292e-01 -8.93545270e-01 1.31196380e+00 -8.82033527e-01
4.48805928e-01 -4.31529373e-01 -2.88041979e-01 -1.10715032e+00
-2.08542407e-01 -4.62697238e-01 -8.37669820e-02 8.60614240e-01
6.30393401e-02 -9.81933415e-01 8.29911709e-01 4.20249969e-01
-2.11709604e-01 -6.07330620e-01 -1.33066356e+00 -7.92264640e-01
4.20707047e-01 -3.60541701e-01 9.18708444e-01 9.65256512e-01
-5.76130867e-01 -4.87877913e-02 -2.90277392e-01 1.10799170e+00
8.37483108e-01 -1.09273538e-01 9.82218444e-01 -1.11585104e+00
9.03548077e-02 -3.36166561e-01 -1.02339864e+00 -7.40598679e-01
3.87488782e-01 -7.34220743e-01 -3.55803758e-01 -7.80090153e-01
-2.82604337e-01 -6.33745193e-01 -5.98213077e-01 3.18291843e-01
-2.00553328e-01 2.33081415e-01 4.30989742e-01 3.08973014e-01
-2.28222728e-01 7.43220150e-01 9.58578229e-01 -4.03826594e-01
2.59873569e-01 3.47271591e-01 -6.35708570e-01 6.57609344e-01
1.02337313e+00 -7.92469084e-01 -3.69995773e-01 -6.30063295e-01
2.51213424e-02 -2.59001017e-01 6.88942373e-01 -1.27087843e+00
1.37937143e-01 -1.97524816e-01 1.93448082e-01 -5.22252262e-01
3.38357657e-01 -1.38009274e+00 1.39625771e-02 6.20061994e-01
8.17147642e-02 2.17211783e-01 6.02408886e-01 8.30990791e-01
-1.93476930e-01 4.22579125e-02 1.03083456e+00 9.13125053e-02
-4.92769927e-01 7.04291344e-01 8.59785154e-02 -2.47006826e-02
1.42349565e+00 1.40848449e-02 -6.26694441e-01 -2.06789166e-01
-4.50356036e-01 3.45479846e-01 5.13262451e-01 7.60368466e-01
8.25048745e-01 -1.35182035e+00 -6.84408486e-01 4.17594582e-01
-1.05394162e-02 1.27722502e-01 4.27665204e-01 3.00339669e-01
-7.28703737e-01 4.85253781e-02 -5.69387317e-01 -5.26447058e-01
-1.00902879e+00 1.04834878e+00 5.91789722e-01 4.57972325e-02
-6.59179986e-01 6.78247273e-01 2.97793180e-01 -1.28999949e-01
2.84012169e-01 1.30410090e-01 7.77293518e-02 -5.17361522e-01
5.06801724e-01 1.42966300e-01 7.73621947e-02 -6.00428879e-01
-6.70048594e-01 4.66038048e-01 -3.06024402e-01 3.21028024e-01
1.26929975e+00 3.04189354e-01 6.28319569e-03 -2.85902500e-01
1.15238523e+00 4.64995429e-02 -1.01390839e+00 -5.55752702e-02
-3.54545325e-01 -8.50448847e-01 -9.94182229e-02 -3.41801018e-01
-1.38335240e+00 1.06386352e+00 5.46264589e-01 4.75461215e-01
9.78609145e-01 -2.80973285e-01 9.35460210e-01 3.98393780e-01
4.38407034e-01 -2.30046034e-01 1.50119156e-01 3.63468081e-01
8.87964487e-01 -9.59107339e-01 -1.19927019e-01 -5.09309888e-01
-2.75757283e-01 7.85390317e-01 8.53071690e-01 -7.45620489e-01
7.95898736e-01 1.65558204e-01 1.19805068e-01 -2.43345231e-01
-4.85860348e-01 4.25172806e-01 -3.50145735e-02 5.19198954e-01
-5.63428581e-01 -2.44437829e-01 2.55495638e-01 4.81317163e-01
-2.31325746e-01 -5.03603578e-01 3.66363913e-01 8.78183544e-01
-2.13647574e-01 -7.65775144e-01 -3.73592079e-01 -8.16734210e-02
-7.27183044e-01 2.03843772e-01 -6.05137110e-01 9.75947440e-01
-1.11304820e-01 8.70444834e-01 -4.07737866e-02 -8.04888666e-01
5.67029834e-01 -3.84527624e-01 5.22463769e-02 -2.89674133e-01
-7.25551188e-01 -5.07957935e-01 -5.88633657e-01 -5.82104504e-01
-3.38576883e-02 -2.76414722e-01 -9.58977997e-01 -6.64279163e-01
-2.38727972e-01 2.57563651e-01 5.54605186e-01 6.95270121e-01
5.36437631e-01 4.86755073e-01 1.12110829e+00 -7.54084826e-01
-9.57586467e-01 -5.77782750e-01 -2.73011506e-01 5.02781332e-01
4.62742746e-01 -8.98900449e-01 -8.11008692e-01 -7.46802330e-01] | [7.700197696685791, -4.470991134643555] |
8403935e-0805-4b95-b1f9-ff38c1c7d9d9 | polardet-a-fast-more-precise-detector-for | 2010.08720 | null | https://arxiv.org/abs/2010.08720v1 | https://arxiv.org/pdf/2010.08720v1.pdf | PolarDet: A Fast, More Precise Detector for Rotated Target in Aerial Images | Fast and precise object detection for high-resolution aerial images has been a challenging task over the years. Due to the sharp variations on object scale, rotation, and aspect ratio, most existing methods are inefficient and imprecise. In this paper, we represent the oriented objects by polar method in polar coordinate and propose PolarDet, a fast and accurate one-stage object detector based on that representation. Our detector introduces a sub-pixel center semantic structure to further improve classifying veracity. PolarDet achieves nearly all SOTA performance in aerial object detection tasks with faster inference speed. In detail, our approach obtains the SOTA results on DOTA, UCAS-AOD, HRSC with 76.64\% mAP, 97.01\% mAP, and 90.46\% mAP respectively. Most noticeably, our PolarDet gets the best performance and reaches the fastest speed(32fps) at the UCAS-AOD dataset. | ['ShiLiang Pu', 'Ye Ren', 'Wenming Tan', 'Yingjia Bu', 'Zhenshen Qu', 'Pengbo Zhao'] | 2020-10-17 | null | null | null | null | ['object-detection-in-aerial-images'] | ['computer-vision'] | [-8.10300335e-02 -4.71364468e-01 3.76161151e-02 -4.27600704e-02
-4.63652790e-01 -8.98957014e-01 3.68009269e-01 -1.77602351e-01
-1.96548864e-01 2.89614111e-01 -3.18943292e-01 -2.36439541e-01
-1.28808275e-01 -1.01203775e+00 -4.81859148e-01 -6.06696844e-01
-2.27632150e-01 1.89722434e-01 7.58197904e-01 -1.28892794e-01
5.76018989e-02 7.71664023e-01 -1.35539782e+00 1.49111837e-01
7.45489061e-01 1.26164436e+00 3.65401804e-01 9.88422215e-01
4.24471468e-01 3.96188498e-01 -5.86791039e-01 -2.22090781e-01
7.86169648e-01 1.05527945e-01 -4.77817327e-01 1.23708390e-01
8.60519171e-01 -7.03510106e-01 -2.19853774e-01 1.23354864e+00
6.89860761e-01 -3.11680496e-01 5.77677310e-01 -1.02086079e+00
-6.78041995e-01 2.80349642e-01 -1.27401769e+00 5.70460320e-01
1.85901776e-01 1.44604653e-01 8.02167535e-01 -1.11661661e+00
3.01923066e-01 1.21559763e+00 8.80297959e-01 -2.57205486e-01
-8.85886371e-01 -6.81170046e-01 2.29349285e-01 -5.89236394e-02
-2.03812861e+00 -1.93342179e-01 7.26953298e-02 -4.24226522e-01
7.96091080e-01 4.69905227e-01 4.77158070e-01 2.20077634e-01
-5.66672459e-02 7.25674748e-01 1.11792982e+00 -1.32292315e-01
-8.56351107e-02 -5.64061515e-02 -1.48924217e-01 9.42411304e-01
6.72891915e-01 -1.29845694e-01 3.49897593e-02 -9.64418054e-03
8.25344682e-01 1.89013481e-01 -4.53543037e-01 2.76312176e-02
-1.17344582e+00 4.82475162e-01 9.24732506e-01 -2.02244356e-01
-6.00320876e-01 4.45310511e-02 2.78277826e-02 -2.62196690e-01
3.25537175e-01 5.03982067e-01 -2.86179811e-01 3.77394259e-01
-9.70145166e-01 2.68904030e-01 3.04979801e-01 1.36367428e+00
2.83072144e-01 1.92786679e-01 -2.51508772e-01 6.45461977e-01
3.42112511e-01 1.50131810e+00 -1.12353064e-01 -8.42423379e-01
4.70621020e-01 5.99331677e-01 3.88629496e-01 -1.63625467e+00
-4.10320222e-01 -5.86633325e-01 -8.19251239e-01 3.06482911e-01
1.22014768e-01 -2.48729765e-01 -1.01825666e+00 8.98779273e-01
5.05186498e-01 4.44136607e-03 2.01405417e-02 1.04686415e+00
8.90066326e-01 1.19069171e+00 7.09438697e-02 -2.40235254e-02
1.66729748e+00 -8.05282533e-01 -3.64964873e-01 -4.59286153e-01
2.98714161e-01 -8.46882164e-01 6.72690511e-01 3.56125355e-01
-8.50588977e-01 -4.83262330e-01 -1.21928716e+00 1.93500537e-02
-2.82763571e-01 8.95907342e-01 4.83651459e-01 5.86071074e-01
-7.02917337e-01 1.87649414e-01 -6.61342621e-01 -5.72536230e-01
7.49896646e-01 1.82114452e-01 -9.25194249e-02 -2.01966032e-01
-6.06742024e-01 5.70647240e-01 4.65828151e-01 1.69740826e-01
-7.56607592e-01 -5.08946598e-01 -5.61991632e-01 6.18323274e-02
5.72347879e-01 -4.20813113e-01 9.72016037e-01 -2.97557741e-01
-1.13029790e+00 8.22884798e-01 1.12752199e-01 -5.37515640e-01
4.16550398e-01 -3.97296876e-01 -4.66855526e-01 3.57960224e-01
3.25568855e-01 8.34229410e-01 8.22935939e-01 -8.98957908e-01
-1.17148054e+00 -7.98954725e-01 1.12675026e-01 3.87808591e-01
-1.83118924e-01 2.61100858e-01 -4.98608410e-01 -4.25437003e-01
5.80643117e-01 -1.04581630e+00 1.23774894e-02 5.62048078e-01
-3.42067331e-01 -8.80188718e-02 1.15864921e+00 -5.46753943e-01
1.28887582e+00 -2.29318666e+00 -3.00278038e-01 -1.57722458e-01
1.13460802e-01 7.17200875e-01 8.15673843e-02 -6.36693835e-02
1.44145787e-01 9.59719196e-02 -1.92694888e-01 2.83248484e-01
-2.63142467e-01 -1.55196130e-01 -6.10861003e-01 6.27181053e-01
2.34877467e-01 5.15946507e-01 -7.96830773e-01 -5.94116032e-01
2.79858738e-01 2.82051742e-01 -2.43524015e-01 1.61675327e-02
1.04830809e-01 -3.97885352e-01 -6.62014484e-01 1.34054053e+00
1.33555984e+00 -2.37101480e-01 -1.17678061e-01 -5.41918099e-01
-6.17942750e-01 -2.36605689e-01 -1.46009123e+00 1.07796645e+00
-4.35236003e-03 8.95225704e-01 1.53454855e-01 -4.66673821e-01
1.03654373e+00 -6.62656175e-03 1.59832865e-01 -5.12889266e-01
-4.89390595e-03 1.62893608e-01 -3.32821727e-01 -4.88148838e-01
8.84638488e-01 2.50157923e-01 5.28336614e-02 -1.06209092e-01
-3.72200131e-01 -4.33719844e-01 2.62229651e-01 4.02084226e-03
7.28239536e-01 4.21737544e-02 7.30736613e-01 -4.66247290e-01
2.41240203e-01 4.73452210e-01 4.02330667e-01 6.33377850e-01
-3.17460239e-01 9.07336116e-01 2.09097639e-01 -6.61925912e-01
-7.96798408e-01 -8.86146367e-01 -6.97883904e-01 8.78577113e-01
4.43511724e-01 -4.04961795e-01 -4.50973928e-01 -5.61013341e-01
1.09869525e-01 2.51890302e-01 -2.85136759e-01 4.24693018e-01
-2.24833816e-01 -1.30680990e+00 6.64535999e-01 6.64055824e-01
1.08871675e+00 -4.74074721e-01 -8.91605914e-01 -1.29713386e-01
-4.88687083e-02 -1.40810919e+00 -8.45376104e-02 -1.44431397e-01
-9.26948607e-01 -9.96995151e-01 -9.22682106e-01 -5.31484127e-01
5.81965029e-01 1.09701097e+00 8.97608936e-01 -3.10498029e-01
-5.33083975e-01 -2.20060259e-01 -2.65863359e-01 -6.77514434e-01
5.62151015e-01 -1.45678982e-01 4.68916520e-02 -3.49890530e-01
2.17066973e-01 -6.29449338e-02 -8.00895572e-01 6.37125432e-01
-7.52433836e-01 -5.62959649e-02 6.35067642e-01 4.33187127e-01
7.46947110e-01 3.74030083e-01 -1.14893652e-01 -5.32111585e-01
-9.15060714e-02 -3.33624423e-01 -1.26229441e+00 9.86673906e-02
-5.97840905e-01 -3.60737115e-01 4.33091253e-01 -1.23650119e-01
-7.07429409e-01 3.87556404e-01 3.23948562e-01 -1.39441162e-01
-1.57775491e-01 2.19516084e-01 1.40242800e-01 -1.47955358e-01
1.08418262e+00 1.32595137e-01 -5.21747708e-01 -2.92363286e-01
1.24210678e-02 1.03975344e+00 6.35419786e-01 -1.71152145e-01
9.38734114e-01 8.90853882e-01 1.31462380e-01 -1.17637742e+00
-1.04247367e+00 -6.03825986e-01 -5.27721584e-01 1.09696565e-02
7.78899491e-01 -1.53555834e+00 -6.11822844e-01 5.88125706e-01
-1.07628989e+00 -1.74526516e-02 3.76937799e-02 5.14854372e-01
1.81849405e-01 2.83129096e-01 -1.82000294e-01 -9.14568484e-01
-5.67080200e-01 -9.64953899e-01 1.30401099e+00 5.26043832e-01
4.08491194e-01 -1.75384864e-01 -1.75054684e-01 6.53996412e-03
3.66830409e-01 3.97044092e-01 1.34375736e-01 1.31412551e-01
-9.02281404e-01 -4.39471364e-01 -1.11415803e+00 1.82600051e-01
-1.00469835e-01 4.47677195e-01 -9.02393579e-01 -1.55823708e-01
-4.83657688e-01 6.73353747e-02 7.58112907e-01 5.44984281e-01
1.04079163e+00 -2.18387187e-01 -6.18294120e-01 1.01183796e+00
1.69017839e+00 2.64507562e-01 6.84443235e-01 3.08803529e-01
5.24588287e-01 5.04781120e-02 1.12570083e+00 6.76979661e-01
3.43499243e-01 7.07830369e-01 7.21452653e-01 -2.57827729e-01
-1.28586739e-01 3.82215865e-02 2.91804969e-01 -1.06416289e-02
-6.43144488e-01 -5.47368109e-01 -1.06198931e+00 5.76733887e-01
-1.60279155e+00 -1.04577041e+00 -7.10731328e-01 1.89035499e+00
4.58715379e-01 -1.98174734e-02 1.44522660e-03 -1.34031102e-01
7.20571697e-01 5.40476024e-01 -2.23507226e-01 3.40252519e-01
-3.74445975e-01 -1.62401080e-01 1.39872003e+00 7.33326226e-02
-1.69025600e+00 1.11068869e+00 6.55486298e+00 6.10421181e-01
-1.17946136e+00 -2.81069100e-01 3.28534961e-01 1.45608932e-01
7.87017643e-01 -2.36922458e-01 -1.34951091e+00 4.69082624e-01
4.02725369e-01 4.76412587e-02 1.14875473e-01 1.47555208e+00
6.45890981e-02 -4.22538966e-01 -3.30136627e-01 1.23757970e+00
1.57692477e-01 -1.03205633e+00 -3.18792574e-02 -1.06204480e-01
7.44663537e-01 2.79413044e-01 1.35815352e-01 2.76559088e-02
3.40395480e-01 -9.54349816e-01 7.08206713e-01 1.01083763e-01
1.14718866e+00 -6.33674920e-01 9.70709741e-01 2.34879091e-01
-1.57367659e+00 -2.54432529e-01 -9.78907168e-01 9.84614342e-03
6.37266487e-02 4.56640214e-01 -9.68421638e-01 4.21051353e-01
1.10142112e+00 8.70532751e-01 -8.50473881e-01 1.28572237e+00
-2.91060358e-01 5.48080504e-01 -8.15959811e-01 6.52610734e-02
3.79894435e-01 -3.26999947e-02 6.33448124e-01 1.33565664e+00
8.48586857e-01 5.98954082e-01 3.99383873e-01 4.53958511e-01
-1.98762189e-03 -1.48784012e-01 -7.33413160e-01 6.28641546e-02
4.86135781e-01 1.47561812e+00 -7.76174307e-01 -3.85942608e-01
-2.21241042e-02 9.02085900e-01 -3.38499397e-01 6.78853467e-02
-1.12540436e+00 -6.20470345e-01 4.96545374e-01 2.60560989e-01
5.45619071e-01 -3.95743310e-01 -1.66311145e-01 -9.50967133e-01
-2.82888468e-02 -7.24507630e-01 3.61160517e-01 -1.15240169e+00
-6.57042682e-01 8.24895561e-01 1.57639787e-01 -1.42983377e+00
3.78035516e-01 -1.09236169e+00 -2.98558295e-01 7.03832805e-01
-1.63793719e+00 -1.22377503e+00 -1.20951653e+00 3.05246919e-01
5.81350148e-01 -1.88691363e-01 7.28233695e-01 4.08674121e-01
-7.70915508e-01 1.00334510e-01 1.40365556e-01 2.74709135e-01
6.15576148e-01 -1.15278864e+00 5.27763247e-01 1.24066794e+00
-3.80642153e-02 1.44320935e-01 6.12593055e-01 -3.99040788e-01
-1.44037831e+00 -1.46279764e+00 4.69428420e-01 -4.60843801e-01
3.71866822e-01 6.22061975e-02 -6.11176491e-01 6.04425073e-01
-1.37406647e-01 5.98733604e-01 3.00868805e-02 -3.22602093e-01
-1.90274715e-01 -4.18117464e-01 -1.06871879e+00 3.69736701e-01
1.05154467e+00 -2.24069789e-01 -1.86845049e-01 7.14781761e-01
4.62767571e-01 -7.38358259e-01 -6.42821074e-01 5.65034389e-01
5.63450992e-01 -8.38141322e-01 1.31448328e+00 -5.01850294e-03
8.56219009e-02 -1.18274856e+00 -3.84835571e-01 -8.65186095e-01
-7.60395646e-01 -2.92037964e-01 7.83064291e-02 8.59908998e-01
2.58605361e-01 -5.49822450e-01 3.99447650e-01 -9.02579948e-02
-1.00074776e-01 -2.60450482e-01 -3.34655881e-01 -8.63153040e-01
-6.76168799e-01 -1.16508521e-01 6.82123244e-01 7.51276255e-01
-6.36169851e-01 2.76620120e-01 -3.12434494e-01 1.07885218e+00
6.65217638e-01 4.17294741e-01 8.79623652e-01 -1.26279449e+00
6.96349069e-02 -5.37250154e-02 -5.74759424e-01 -1.34118319e+00
-7.43049741e-01 -4.15785283e-01 9.99926180e-02 -1.52699542e+00
1.35067314e-01 -5.14923155e-01 2.51069814e-01 5.19883275e-01
-1.68433562e-01 8.00722182e-01 1.46952271e-01 3.33798021e-01
-6.74507797e-01 1.85850844e-01 9.50886965e-01 -1.35070518e-01
-4.50477712e-02 -5.42885512e-02 -4.59567994e-01 9.77503121e-01
7.89483309e-01 -6.60423338e-01 -7.90746063e-02 -9.27402079e-01
2.25577608e-01 -2.04028398e-01 5.45760512e-01 -1.44814396e+00
2.40185693e-01 -2.36556113e-01 7.98782051e-01 -1.07172322e+00
3.67816150e-01 -7.76456237e-01 2.28829235e-02 5.73897600e-01
3.57092470e-01 2.11698040e-01 3.26807976e-01 5.48196673e-01
-1.13041095e-01 -4.60284799e-02 1.09537053e+00 -1.15056597e-01
-1.00742602e+00 3.77624869e-01 -1.10097276e-02 -1.28556788e-01
1.37900603e+00 -2.68505275e-01 -6.72975659e-01 1.24383979e-01
-2.46807367e-01 2.61137754e-01 3.82775515e-01 1.78558920e-02
5.24997115e-01 -1.07748914e+00 -8.84134591e-01 1.74350604e-01
2.79381841e-01 5.23775697e-01 1.18063852e-01 8.85771275e-01
-1.29193342e+00 6.48750484e-01 -2.68437088e-01 -8.54862928e-01
-1.50210702e+00 2.73964196e-01 3.32144797e-01 1.97167128e-01
-5.46167552e-01 9.18065071e-01 2.65099853e-01 7.21340850e-02
-1.45671576e-01 -5.48850417e-01 -1.93246260e-01 2.14731202e-01
8.55274320e-01 6.43093348e-01 -1.06879361e-01 -5.62882423e-01
-5.25089383e-01 9.74529624e-01 2.29287148e-01 2.29450092e-01
1.26874232e+00 -8.02746564e-02 8.84383395e-02 -1.92543387e-01
6.40122116e-01 5.92677817e-02 -1.33170128e+00 -1.21536121e-01
-4.17889357e-01 -1.04442346e+00 4.45598006e-01 -6.57008111e-01
-8.74967217e-01 8.93491149e-01 1.06839287e+00 4.10244077e-01
1.12499559e+00 -1.88287467e-01 4.52224761e-01 6.25475049e-01
3.73520881e-01 -9.32050109e-01 -2.51747012e-01 5.70872247e-01
8.88838112e-01 -1.42383754e+00 7.96470344e-01 -6.37680948e-01
-6.21356189e-01 1.17472899e+00 4.66885120e-01 -2.23955706e-01
2.93829501e-01 1.47933364e-01 -6.65155202e-02 -3.96242291e-01
-2.23900750e-01 -4.12136853e-01 3.71097177e-01 6.28572702e-01
6.77999780e-02 2.67685920e-01 1.61731139e-01 -9.36685354e-02
3.06866653e-02 -1.52976438e-01 3.98253739e-01 8.31240714e-01
-9.24058974e-01 -1.06885597e-01 -7.23731816e-01 2.05142796e-01
-5.28636754e-01 -1.52564019e-01 -1.15947805e-01 9.63996053e-01
1.13407767e-03 5.77639699e-01 1.59652486e-01 -1.95726588e-01
5.55304706e-01 -4.89672065e-01 7.31696654e-03 -5.38718462e-01
1.53880678e-02 -6.17863908e-02 -7.28916526e-02 -4.83426213e-01
-2.73258477e-01 -5.49790800e-01 -1.17399549e+00 -4.18712318e-01
-6.63232684e-01 -2.07836136e-01 8.39542627e-01 3.06024045e-01
5.85182309e-01 2.94438630e-01 5.62084317e-01 -6.53357685e-01
-5.98290205e-01 -9.06191885e-01 -6.99621141e-01 -2.00612664e-01
1.25122964e-01 -6.27794683e-01 -2.67493486e-01 1.68471754e-01] | [8.751679420471191, -0.8428593873977661] |
c5997fa0-74fb-464f-b87d-6c561947342f | wavelet-feature-maps-compression-for-image-to | 2205.12268 | null | https://arxiv.org/abs/2205.12268v4 | https://arxiv.org/pdf/2205.12268v4.pdf | Wavelet Feature Maps Compression for Image-to-Image CNNs | Convolutional Neural Networks (CNNs) are known for requiring extensive computational resources, and quantization is among the best and most common methods for compressing them. While aggressive quantization (i.e., less than 4-bits) performs well for classification, it may cause severe performance degradation in image-to-image tasks such as semantic segmentation and depth estimation. In this paper, we propose Wavelet Compressed Convolution (WCC) -- a novel approach for high-resolution activation maps compression integrated with point-wise convolutions, which are the main computational cost of modern architectures. To this end, we use an efficient and hardware-friendly Haar-wavelet transform, known for its effectiveness in image compression, and define the convolution on the compressed activation map. We experiment with various tasks that benefit from high-resolution input. By combining WCC with light quantization, we achieve compression rates equivalent to 1-4bit activation quantization with relatively small and much more graceful degradation in performance. Our code is available at https://github.com/BGUCompSci/WaveletCompressedConvolution. | ['Eran Treister', 'Maor Ashkenazi', 'Yair Zohav', 'Shahaf E. Finder'] | 2022-05-24 | null | null | null | null | ['neural-network-compression', 'neural-network-compression'] | ['methodology', 'miscellaneous'] | [ 3.74092251e-01 -1.76609769e-01 -1.16760321e-01 -3.85549664e-01
-7.21618772e-01 -1.03048280e-01 2.40043804e-01 1.63287342e-01
-9.64742184e-01 4.94069517e-01 -5.55903502e-02 -4.11035389e-01
2.50229686e-01 -9.89999890e-01 -8.79680812e-01 -6.67206764e-01
-8.95747617e-02 -1.31308690e-01 4.64254230e-01 -3.36039886e-02
8.25221688e-02 5.52291095e-01 -1.72341383e+00 5.55858850e-01
6.40743077e-01 1.46879756e+00 4.12542433e-01 7.92996466e-01
-5.22237532e-02 6.16365075e-01 -4.87172425e-01 -6.73438609e-01
1.78785279e-01 -2.73388892e-01 -6.18335247e-01 -2.92893976e-01
3.83420140e-01 -7.15035200e-01 -5.85432529e-01 1.19391549e+00
3.00271153e-01 -1.15796305e-01 3.39367509e-01 -8.35032344e-01
-4.34083819e-01 7.99879670e-01 -5.16210735e-01 3.61152709e-01
-2.60371864e-01 1.83400977e-02 7.68100798e-01 -7.70647109e-01
3.29525381e-01 1.22345710e+00 7.43030846e-01 5.93521535e-01
-1.04385138e+00 -7.14298964e-01 -3.92949641e-01 3.80710542e-01
-1.56283557e+00 -6.32041812e-01 6.30872786e-01 1.18197851e-01
1.03157806e+00 2.47425780e-01 7.77011693e-01 7.91294575e-01
3.04913163e-01 5.54925561e-01 7.86546946e-01 -3.12032312e-01
4.72831547e-01 -3.35460573e-01 -7.65285045e-02 7.55610347e-01
4.09247011e-01 -4.02283706e-02 -5.85951209e-01 2.28514731e-01
9.83008265e-01 1.70407221e-01 -3.93744081e-01 3.47989559e-01
-8.37206483e-01 8.33544552e-01 7.87358165e-01 1.90747187e-01
-5.37922978e-01 8.13003063e-01 5.08952677e-01 2.97306687e-01
3.80164236e-01 7.81046897e-02 -2.76862144e-01 -4.70865637e-01
-1.12176979e+00 2.23764822e-01 4.69262689e-01 8.23853076e-01
6.54699087e-01 2.51632750e-01 9.17098895e-02 8.96681428e-01
-5.70491850e-02 2.23703638e-01 7.47792244e-01 -1.28370714e+00
2.72931039e-01 2.02324867e-01 -3.09952468e-01 -7.81418025e-01
-1.43323973e-01 -4.96168524e-01 -1.33905900e+00 3.94734859e-01
2.11714789e-01 6.95926994e-02 -9.74891722e-01 1.52741408e+00
-6.78998083e-02 2.20833823e-01 1.78392962e-01 1.01754820e+00
5.60266972e-01 7.45494843e-01 3.96242626e-02 -9.48487073e-02
1.58887732e+00 -7.04454243e-01 -6.70763731e-01 -1.26198202e-01
4.43458796e-01 -6.50072515e-01 1.08904850e+00 4.21959072e-01
-1.38549042e+00 -6.13559723e-01 -1.46030319e+00 -5.04136503e-01
-3.36123616e-01 -1.25453100e-01 5.77377260e-01 6.06190681e-01
-1.19472337e+00 1.11930716e+00 -1.21883869e+00 1.74099877e-01
8.32998753e-01 2.94118285e-01 -2.02591509e-01 -2.29632571e-01
-9.42366660e-01 5.75615346e-01 5.58148563e-01 -2.84348696e-01
-5.13422966e-01 -7.48374224e-01 -7.09801078e-01 4.83591437e-01
-2.53517888e-02 -3.17954421e-01 1.42420971e+00 -7.69925535e-01
-1.29308009e+00 5.54770231e-01 -6.58347756e-02 -1.09641027e+00
2.97799677e-01 -3.26087862e-01 -7.64033720e-02 6.67837799e-01
-3.44063401e-01 1.06869435e+00 1.03395259e+00 -7.22773552e-01
-5.35252571e-01 -3.37753475e-01 -1.39657527e-01 -5.33615006e-04
-8.21864367e-01 -5.46634942e-02 -5.92148721e-01 -9.17101502e-01
3.11453432e-01 -5.48434556e-01 -1.54953822e-01 4.97056365e-01
1.56521276e-01 -6.20891564e-02 8.68860364e-01 -6.56752348e-01
1.18132257e+00 -2.51890063e+00 -2.36454561e-01 -1.80507064e-01
3.02818388e-01 5.32578766e-01 7.68387988e-02 -1.27871707e-01
9.23876762e-02 1.69489905e-01 -4.77872849e-01 -5.69862247e-01
-2.07404956e-01 3.84779096e-01 -2.34681413e-01 4.33402151e-01
4.27703798e-01 1.01092196e+00 -5.58509409e-01 -4.58398253e-01
1.08159363e-01 9.49889302e-01 -6.35868788e-01 -5.04110456e-02
-1.64622903e-01 -1.85759068e-01 -1.32919684e-01 7.17078507e-01
7.47581065e-01 -2.61077553e-01 -1.27471030e-01 -4.50372428e-01
-1.10591970e-01 2.02035144e-01 -7.02854037e-01 1.75793397e+00
-4.81353223e-01 9.25710678e-01 1.24645703e-01 -9.33892787e-01
8.05569410e-01 2.88489640e-01 2.44319022e-01 -9.19003069e-01
3.26951921e-01 3.50439519e-01 -1.59563258e-01 -1.03490978e-01
6.88589454e-01 -2.74348836e-02 2.10497156e-01 1.99905366e-01
4.52866917e-03 -2.26226598e-01 2.17849135e-01 -6.44396618e-02
1.14539504e+00 -2.97273248e-01 1.76551253e-01 -1.17645569e-01
4.40006331e-02 -1.72234863e-01 4.43792939e-01 4.29535091e-01
-1.36444762e-01 7.75560737e-01 4.66852993e-01 -4.92426515e-01
-1.40460503e+00 -6.81299865e-01 -3.09514403e-01 8.62200737e-01
-6.07146882e-02 -5.19809067e-01 -1.08021688e+00 4.11152132e-02
-2.78629810e-01 3.91935468e-01 -1.65040895e-01 -1.87368557e-01
-6.84462488e-01 -5.43812275e-01 9.22982395e-01 7.63681173e-01
1.01085651e+00 -9.08142507e-01 -1.45525837e+00 2.66848207e-01
-1.24770002e-02 -1.19931984e+00 -2.99071193e-01 5.05403817e-01
-1.20137060e+00 -6.29451811e-01 -1.08562648e+00 -8.03480744e-01
4.44732159e-01 2.81944513e-01 8.93697441e-01 1.58497393e-01
-4.84538257e-01 -2.27580473e-01 -4.85928804e-01 -3.11624855e-01
-1.61204115e-01 -1.78139228e-02 -3.45288664e-01 -2.32817262e-01
2.17592105e-01 -8.32832515e-01 -8.45171988e-01 -1.80306777e-01
-1.25973082e+00 2.76698500e-01 7.99984515e-01 7.98598707e-01
8.84271026e-01 3.30875337e-01 1.66946977e-01 -7.02107430e-01
6.36703312e-01 4.17048261e-02 -7.15323985e-01 -1.82046205e-01
-5.87603390e-01 1.09021217e-01 8.07625413e-01 -4.64682460e-01
-5.73634207e-01 9.52418000e-02 -6.31178141e-01 -9.22421813e-01
-1.12343423e-01 4.34294909e-01 2.42449462e-01 -2.47879833e-01
7.74402976e-01 2.32433215e-01 9.30838808e-02 -5.38961828e-01
2.52845556e-01 6.26263678e-01 9.37711477e-01 -3.08464110e-01
4.88176465e-01 4.77644563e-01 8.67624804e-02 -8.14341843e-01
-3.62640381e-01 -1.67656317e-01 -2.59619415e-01 4.64778617e-02
9.60147440e-01 -9.59052801e-01 -6.32162809e-01 4.70990449e-01
-1.28105664e+00 -5.28978646e-01 -3.03845704e-01 4.23987478e-01
-4.47933912e-01 1.54864639e-01 -1.08090973e+00 -6.00329280e-01
-7.32025266e-01 -1.17877233e+00 7.94422209e-01 2.73278803e-01
1.48506612e-01 -4.90484446e-01 -6.43529713e-01 -1.75042465e-01
8.03373933e-01 2.41074502e-01 8.72502923e-01 -6.15719259e-02
-5.43952763e-01 -1.55771673e-01 -5.34676194e-01 6.35620177e-01
-2.68020660e-01 -2.68836230e-01 -1.10319209e+00 -2.89409220e-01
8.66806805e-02 -5.09552538e-01 1.34108031e+00 4.50340927e-01
1.84479284e+00 -4.97151673e-01 1.44804135e-01 1.12474442e+00
1.69197464e+00 1.55150205e-01 9.44889367e-01 7.75651261e-02
4.38128412e-01 1.28101319e-01 2.89720535e-01 5.47675073e-01
1.94772407e-02 4.36573863e-01 6.41506135e-01 -1.90757722e-01
-5.38906157e-01 -8.79602358e-02 5.93849607e-02 9.34858501e-01
-2.23959401e-01 -3.93847041e-02 -7.06833780e-01 3.17954630e-01
-1.41976941e+00 -8.25842798e-01 1.52612120e-01 1.99023044e+00
1.06184447e+00 4.57046986e-01 -2.32145697e-01 6.41199768e-01
4.08901989e-01 1.61765948e-01 -5.23255348e-01 -4.98152524e-01
-9.20252502e-02 8.81895602e-01 8.83872867e-01 3.30973864e-01
-9.45717871e-01 7.71967411e-01 5.46635914e+00 1.11671519e+00
-1.38437533e+00 2.60492980e-01 9.02529180e-01 -2.35911548e-01
5.23433350e-02 -4.48481619e-01 -6.46298647e-01 4.77396846e-01
1.22456026e+00 -3.30440998e-02 6.09461546e-01 9.76935267e-01
-2.50208210e-02 -7.54386634e-02 -7.83884883e-01 1.38311362e+00
-2.11872652e-01 -1.71350598e+00 9.63343084e-02 8.16260874e-02
5.41061580e-01 7.17781335e-02 3.16070825e-01 -3.27595212e-02
-2.44455546e-01 -1.12904608e+00 9.92047489e-01 1.77757796e-02
1.43021536e+00 -1.03609395e+00 8.42477083e-01 1.52960390e-01
-1.26034999e+00 -1.62876725e-01 -9.43702161e-01 1.45538166e-01
-9.02965888e-02 7.13358879e-01 -2.24971727e-01 -7.94430915e-03
9.75032985e-01 4.92122293e-01 -3.12670439e-01 8.85885119e-01
-1.45912822e-02 6.46307826e-01 -3.22089523e-01 -5.35955001e-03
2.65084177e-01 1.02293976e-01 -4.76958416e-02 1.36087680e+00
7.02933133e-01 5.26045442e-01 -3.84872645e-01 7.32620716e-01
-5.53829193e-01 -2.28915319e-01 -3.01137239e-01 -6.78120032e-02
6.36937737e-01 1.03620648e+00 -9.46973681e-01 -5.08462250e-01
-4.21034157e-01 1.27109802e+00 2.57959306e-01 1.61693215e-01
-8.32177937e-01 -8.22820425e-01 7.34239221e-01 2.08164603e-01
4.84616697e-01 -3.28553617e-01 -5.82931578e-01 -8.03011000e-01
1.08965732e-01 -6.19157076e-01 1.03081658e-01 -5.78785300e-01
-5.36927998e-01 7.55145013e-01 -2.35637203e-01 -1.08823323e+00
5.19548915e-03 -7.45703459e-01 -3.41644615e-01 8.59092653e-01
-1.87232471e+00 -6.42116666e-01 -5.76425850e-01 5.95942199e-01
8.11353862e-01 7.11803660e-02 8.59487474e-01 5.59046090e-01
-3.52896482e-01 8.07997882e-01 -2.36385390e-02 8.42945650e-02
3.14517766e-01 -8.44949782e-01 6.22958839e-01 8.17249596e-01
-7.94309229e-02 4.22281295e-01 5.34537196e-01 -2.54495800e-01
-1.61104488e+00 -1.26109147e+00 6.48222804e-01 6.25986397e-01
3.26786995e-01 -2.25241140e-01 -1.14704192e+00 1.95833609e-01
1.04880117e-01 3.23791295e-01 3.18122119e-01 -5.39750636e-01
-4.71204132e-01 -2.54104108e-01 -1.18932819e+00 6.00568831e-01
8.14161777e-01 -3.94662887e-01 -1.28092706e-01 1.62924677e-01
1.05500841e+00 -5.47778785e-01 -8.89268041e-01 2.75511831e-01
4.55774963e-01 -1.11882007e+00 1.12337554e+00 1.93609446e-01
9.41866815e-01 1.42308846e-02 -4.39119995e-01 -9.59837615e-01
-2.83100188e-01 -4.37967777e-01 -3.50448936e-01 6.68896258e-01
2.03868419e-01 -4.57735956e-01 9.44803894e-01 1.76006645e-01
-3.80669147e-01 -1.00704968e+00 -1.10256660e+00 -7.90907145e-01
6.95874542e-02 -5.15720069e-01 5.55401564e-01 5.58786273e-01
-8.93256366e-02 -8.27249065e-02 -1.80354610e-01 4.44820598e-02
7.30551600e-01 -3.31936747e-01 1.59247398e-01 -8.49582255e-01
-4.15132135e-01 -7.53618002e-01 -5.16998172e-01 -1.16597974e+00
-1.66422784e-01 -7.22223580e-01 1.20296709e-01 -1.25472438e+00
-1.66790038e-01 -2.90379107e-01 -1.39789283e-01 6.46258414e-01
2.38461033e-01 7.72678375e-01 2.17241362e-01 2.61486799e-01
-2.78470963e-01 4.83217210e-01 1.07990694e+00 -8.66629258e-02
1.16600180e-02 -3.78331214e-01 -4.21760291e-01 5.82579315e-01
1.18659937e+00 -4.03317243e-01 -3.59561622e-01 -8.85441005e-01
-1.45396039e-01 1.56256258e-01 4.83691454e-01 -1.38823056e+00
3.62295002e-01 1.62508726e-01 4.90892619e-01 -4.98867482e-01
6.88519418e-01 -5.78523040e-01 7.24601299e-02 9.34285402e-01
-4.18399036e-01 2.03669384e-01 3.08800131e-01 2.80009568e-01
-4.09053594e-01 -3.82158637e-01 1.25012779e+00 -2.79402345e-01
-7.18819499e-01 4.16441172e-01 -2.29773551e-01 -1.58812791e-01
6.76354587e-01 -2.09133536e-01 -3.95015091e-01 -1.98978409e-01
-3.43374550e-01 -3.47445309e-01 4.40505028e-01 -1.06232159e-01
1.21003771e+00 -1.20038748e+00 -6.75981760e-01 4.15988177e-01
-2.87346661e-01 1.91949278e-01 1.26456887e-01 5.32404542e-01
-1.18454182e+00 4.51904297e-01 -3.92562121e-01 -4.06081259e-01
-1.33375311e+00 3.12500507e-01 9.08394605e-02 1.11208841e-01
-1.05907917e+00 1.08767068e+00 -2.26905480e-01 6.10622644e-01
5.31817019e-01 -4.95198011e-01 1.67619646e-01 -2.37145126e-01
9.11579490e-01 2.76362836e-01 2.18765751e-01 -2.25059658e-01
-2.39368379e-02 4.93183672e-01 -1.46381050e-01 -1.91227764e-01
1.37951839e+00 2.79597610e-01 -1.10018685e-01 7.73092508e-02
1.47003198e+00 -6.92211211e-01 -1.49580073e+00 -1.65954322e-01
-2.10866645e-01 -5.10079384e-01 5.17975688e-01 -2.15127215e-01
-1.39783049e+00 1.29529309e+00 8.11154544e-01 2.32356250e-01
1.53998339e+00 -1.92563251e-01 1.21177006e+00 3.15263063e-01
3.85386884e-01 -9.55016017e-01 9.03341696e-02 4.16150391e-01
7.37256646e-01 -8.10600102e-01 4.07064892e-02 -4.18100476e-01
-1.87868670e-01 1.28568316e+00 3.92976075e-01 -2.60417521e-01
6.53829396e-01 7.31420040e-01 -2.05317363e-01 1.62311062e-01
-8.54185224e-01 -1.03960216e-01 -1.97443068e-01 4.37005460e-01
4.07211632e-01 1.31088525e-01 -3.56694221e-01 2.81022936e-01
-4.48794872e-01 6.64359890e-03 3.92443150e-01 1.07182634e+00
-6.98298633e-01 -7.68104613e-01 -2.64195621e-01 5.72071373e-01
-6.92498446e-01 -3.81942332e-01 2.99883097e-01 2.27019668e-01
1.39264107e-01 6.91941798e-01 4.76954579e-01 -5.43013275e-01
-3.00054178e-02 -1.09161608e-01 3.99918079e-01 -2.03623652e-01
-3.30948740e-01 2.23703891e-01 -2.58157462e-01 -8.73050451e-01
-1.42884359e-01 -2.03557476e-01 -1.51462245e+00 -7.97214568e-01
5.28810285e-02 -2.53316760e-01 1.00852859e+00 4.18936610e-01
2.23514900e-01 7.28411496e-01 3.04012507e-01 -8.92496288e-01
-6.49960697e-01 -7.42173135e-01 -4.97192562e-01 2.97379702e-01
4.34904337e-01 -1.52689159e-01 -2.68803775e-01 3.20427030e-01] | [8.572839736938477, 2.927731513977051] |
b7fc4609-7fce-4ad3-b0cd-9c03743643fe | sparse-signsgd-with-majority-vote-for | 2302.07475 | null | https://arxiv.org/abs/2302.07475v1 | https://arxiv.org/pdf/2302.07475v1.pdf | Sparse-SignSGD with Majority Vote for Communication-Efficient Distributed Learning | The training efficiency of complex deep learning models can be significantly improved through the use of distributed optimization. However, this process is often hindered by a large amount of communication cost between workers and a parameter server during iterations. To address this bottleneck, in this paper, we present a new communication-efficient algorithm that offers the synergistic benefits of both sparsification and sign quantization, called ${\sf S}^3$GD-MV. The workers in ${\sf S}^3$GD-MV select the top-$K$ magnitude components of their local gradient vector and only send the signs of these components to the server. The server then aggregates the signs and returns the results via a majority vote rule. Our analysis shows that, under certain mild conditions, ${\sf S}^3$GD-MV can converge at the same rate as signSGD while significantly reducing communication costs, if the sparsification parameter $K$ is properly chosen based on the number of workers and the size of the deep learning model. Experimental results using both independent and identically distributed (IID) and non-IID datasets demonstrate that the ${\sf S}^3$GD-MV attains higher accuracy than signSGD, significantly reducing communication costs. These findings highlight the potential of ${\sf S}^3$GD-MV as a promising solution for communication-efficient distributed optimization in deep learning. | ['Namyoon Lee', 'Chanho Park'] | 2023-02-15 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-2.13629231e-01 -1.48539349e-01 -5.76376207e-02 -4.61482882e-01
-1.00460458e+00 -2.55487829e-01 -9.39389095e-02 1.56452373e-01
-9.15986836e-01 6.92114353e-01 -3.48125368e-01 -4.75534707e-01
-4.64629531e-01 -7.14687347e-01 -7.89104342e-01 -1.23971367e+00
-3.57065111e-01 2.59118676e-01 -4.37847935e-02 1.24735236e-01
1.57323644e-01 5.08444071e-01 -1.34993565e+00 -2.32479554e-02
5.06657004e-01 1.63889337e+00 1.24971546e-01 5.99110186e-01
-1.20737463e-01 6.12191379e-01 -8.74137104e-01 -3.81035656e-01
6.76540375e-01 -3.32697093e-01 -4.49666709e-01 -6.80708215e-02
6.16717577e-01 -5.54446161e-01 -2.83364266e-01 1.11915410e+00
9.26317334e-01 2.33134478e-01 2.27690265e-01 -1.15027595e+00
1.50070563e-01 5.81414282e-01 -6.88768685e-01 2.82570064e-01
-3.30855459e-01 2.42353380e-02 9.45800543e-01 -8.36982071e-01
3.30111146e-01 8.62116218e-01 7.62107134e-01 3.10113013e-01
-1.07942653e+00 -1.06659400e+00 2.94302136e-01 1.02520831e-01
-1.50451922e+00 -3.94857228e-01 7.79814005e-01 -2.05734655e-01
1.02795601e+00 1.77135572e-01 5.37548363e-01 1.93081245e-01
4.21063751e-02 6.70758724e-01 1.01383018e+00 -4.17910457e-01
8.14939618e-01 -1.88629314e-01 9.36214253e-02 1.12858272e+00
2.72800595e-01 -3.39338094e-01 -1.08485162e+00 -4.14474845e-01
6.61754787e-01 1.32750154e-01 -1.30401745e-01 -2.16453999e-01
-9.07545090e-01 8.77189279e-01 3.64587277e-01 1.69842348e-01
-4.91713852e-01 6.13754570e-01 4.22708184e-01 5.58738947e-01
5.86745858e-01 -6.04159534e-02 -6.53133392e-01 -2.84606516e-01
-1.02144170e+00 1.56224549e-01 7.92127430e-01 5.98928154e-01
1.05453968e+00 2.03596935e-01 1.29036037e-02 7.44383812e-01
3.16097140e-01 8.46077204e-01 2.31120557e-01 -1.26166582e+00
8.17702055e-01 3.86459112e-01 -8.89640972e-02 -9.57460642e-01
-3.25709134e-01 -7.06102490e-01 -1.13882208e+00 3.99084300e-01
4.00246531e-01 -5.92882335e-01 -4.87755358e-01 1.86946905e+00
6.36784494e-01 -8.91062152e-03 -1.37926981e-01 7.80843258e-01
4.74365979e-01 3.88025254e-01 -1.90317035e-01 -1.26141608e-01
1.11830580e+00 -8.11988175e-01 -3.20229530e-01 -7.34518990e-02
9.95238185e-01 -6.72304869e-01 9.19050515e-01 4.16595429e-01
-1.20388377e+00 -2.42461309e-01 -1.00084698e+00 2.15381280e-01
1.42398566e-01 4.04535949e-01 8.06267619e-01 8.14725876e-01
-1.30951893e+00 7.08635390e-01 -9.75452244e-01 4.27362621e-01
7.86118746e-01 8.62518966e-01 -1.04735337e-01 -5.19485138e-02
-5.62450945e-01 2.87010789e-01 -1.24100968e-01 1.65737301e-01
-7.45666742e-01 -5.93236625e-01 -3.54343653e-01 2.36795485e-01
9.40471739e-02 -5.11285543e-01 1.16075122e+00 -6.62311554e-01
-1.42330289e+00 6.10576093e-01 -3.48178357e-01 -3.50586951e-01
5.25351763e-01 6.58951178e-02 2.69890159e-01 2.11750418e-01
-1.04989357e-01 3.12057078e-01 1.01575696e+00 -7.15219080e-01
-8.14377785e-01 -9.04088378e-01 -2.40987971e-01 2.46561468e-01
-8.49634588e-01 -1.38442278e-01 -3.96703303e-01 -4.86770838e-01
5.02942681e-01 -1.00114083e+00 -2.62708396e-01 3.17047507e-01
-1.00224443e-01 -3.00539702e-01 6.26131594e-01 -3.28047812e-01
1.19656837e+00 -2.34091330e+00 -1.07624985e-01 7.29980707e-01
6.45851493e-01 3.45798969e-01 1.14489950e-01 8.66742581e-02
4.47767138e-01 -2.37108350e-01 -3.74580398e-02 -8.66375804e-01
4.79845405e-02 3.13096702e-01 -9.22751240e-03 7.48085082e-01
-5.30127585e-01 3.78200203e-01 -5.44199467e-01 -2.82247424e-01
-2.98577666e-01 4.74743068e-01 -6.33387983e-01 -1.86980724e-01
4.19395417e-02 2.05483455e-02 -7.75559962e-01 2.20355213e-01
6.58888996e-01 -6.87959373e-01 4.28140819e-01 -1.95791367e-02
1.25210017e-01 3.45986664e-01 -1.69431770e+00 1.39692140e+00
-4.02461112e-01 5.52438617e-01 5.27868927e-01 -1.28936207e+00
9.80613768e-01 1.96997494e-01 6.87991977e-01 -6.83533490e-01
1.54638439e-01 4.95672256e-01 -2.45974883e-01 -7.17966333e-02
1.00984268e-01 1.91687979e-02 1.72754407e-01 9.83917475e-01
-3.68137844e-02 -3.01443245e-02 1.52968720e-01 2.42298573e-01
1.26822460e+00 -5.89686334e-01 -3.14067811e-01 -1.60345778e-01
2.87738621e-01 -4.14337039e-01 6.44365191e-01 8.50884080e-01
-2.24832818e-01 7.30142295e-02 4.59507227e-01 -4.06780213e-01
-7.59449124e-01 -8.57518435e-01 1.80103332e-01 1.43173623e+00
-1.05145052e-01 -3.62786531e-01 -8.90357375e-01 -6.82322741e-01
1.43797323e-01 -6.42201444e-03 -2.77254760e-01 -5.85433422e-03
-6.26063406e-01 -6.99473500e-01 5.71397901e-01 5.92498422e-01
8.37613404e-01 -6.46267235e-01 -6.48107886e-01 2.34892920e-01
2.34097075e-02 -8.13902736e-01 -5.75806737e-01 5.68606317e-01
-1.17705119e+00 -8.03812325e-01 -8.24767947e-01 -8.84569705e-01
1.22335100e+00 4.78881836e-01 6.52121246e-01 2.54243493e-01
-1.97613716e-01 3.57360125e-01 1.03552146e-02 -3.73200446e-01
4.66387011e-02 5.49534410e-02 -6.80832341e-02 6.34566769e-02
1.34145111e-01 -6.27537429e-01 -1.02543187e+00 1.44522205e-01
-7.94434547e-01 -2.31425941e-01 4.58594739e-01 7.34977245e-01
8.29985201e-01 1.98039696e-01 3.17220300e-01 -6.10651016e-01
5.97816706e-01 4.55395877e-02 -8.89578342e-01 -2.50051022e-02
-8.89309883e-01 3.03038329e-01 9.18325782e-01 -4.23308820e-01
-5.98671854e-01 1.33050472e-01 -6.26949221e-02 -4.53135520e-01
3.61979038e-01 4.08527583e-01 1.58922344e-01 -4.41291541e-01
5.95043898e-01 3.09379786e-01 1.00792937e-01 -5.25340080e-01
1.39635682e-01 5.81897080e-01 5.47449663e-02 -5.11648238e-01
3.74156833e-01 7.15873301e-01 3.20670813e-01 -6.77824020e-01
-4.06540811e-01 -3.17697436e-01 3.64387222e-02 -4.90223691e-02
3.13596338e-01 -9.56825078e-01 -1.14338434e+00 6.43821120e-01
-8.24297309e-01 -6.88725770e-01 -4.11111623e-01 7.13846803e-01
-2.92979211e-01 3.95213038e-01 -7.43489861e-01 -7.43921161e-01
-7.75487542e-01 -1.24328935e+00 9.08717036e-01 1.84677076e-02
2.74106581e-03 -7.08341360e-01 -3.06165010e-01 5.42085707e-01
5.56956649e-01 -2.19006360e-01 8.27656150e-01 -6.04229093e-01
-6.73704565e-01 -5.08183181e-01 -2.42914259e-01 6.99041486e-01
-1.19010232e-01 -3.76951724e-01 -5.73289871e-01 -6.62586749e-01
1.15263000e-01 -3.50478649e-01 7.08650947e-01 4.29893613e-01
1.26283014e+00 -5.59310973e-01 -6.06361218e-02 7.06378400e-01
1.30637085e+00 6.69981986e-02 6.30394295e-02 1.68840945e-01
3.85835052e-01 3.79788540e-02 1.82363719e-01 1.12061191e+00
3.52226883e-01 2.99114257e-01 3.31173241e-01 -9.34502780e-02
-1.21245243e-01 7.84389526e-02 2.70612210e-01 1.00682461e+00
-4.77484837e-02 5.78508303e-02 -6.80068016e-01 2.63340116e-01
-1.52524006e+00 -6.03103399e-01 8.65547732e-02 2.41791677e+00
1.09455001e+00 8.02605296e-04 -1.30468622e-01 4.89315957e-01
3.60160500e-01 2.56681770e-01 -6.62590623e-01 -3.07444900e-01
-7.31830075e-02 6.95997655e-01 8.64251494e-01 3.44076186e-01
-6.10368311e-01 5.86177409e-01 4.92692518e+00 1.19078994e+00
-1.45898688e+00 3.30659211e-01 7.82334685e-01 -6.24709249e-01
-1.62620515e-01 -2.30371460e-01 -1.22908068e+00 6.63082302e-01
7.11633205e-01 1.62766591e-01 6.37929380e-01 1.07963359e+00
2.55703598e-01 -2.63357162e-01 -8.22135627e-01 1.30945659e+00
-2.97597777e-02 -1.44067276e+00 -4.28111911e-01 2.93228090e-01
1.02511644e+00 3.56755197e-01 2.31137142e-01 -1.57267883e-01
4.65280920e-01 -6.07396126e-01 6.21680439e-01 -3.48693617e-02
6.99755490e-01 -8.07289600e-01 5.21969855e-01 4.48256433e-01
-1.13441968e+00 -2.77605295e-01 -3.03901464e-01 6.61492124e-02
-1.33747295e-01 1.09505785e+00 -7.39321709e-01 -1.02284793e-02
1.05385566e+00 6.52958006e-02 -1.29478663e-01 7.31681108e-01
-1.98876247e-01 7.22575605e-01 -9.14962888e-01 -4.35776293e-01
2.76024818e-01 -2.71001309e-01 7.54796639e-02 9.33922827e-01
5.42973220e-01 1.20337144e-01 -1.18316328e-02 2.41301343e-01
-4.37814713e-01 1.95857197e-01 -1.20371610e-01 1.46350190e-01
6.68795168e-01 9.42764819e-01 -8.41466963e-01 -5.19386172e-01
-1.43053219e-01 7.88553238e-01 5.59385955e-01 3.03636551e-01
-5.60610831e-01 -4.29953158e-01 7.68621147e-01 -9.69709679e-02
7.56083131e-01 -6.41821504e-01 -6.54332161e-01 -8.97245407e-01
3.63193393e-01 -8.94124091e-01 4.86230582e-01 -1.28112316e-01
-7.68510401e-01 3.50863516e-01 -3.86624128e-01 -7.34522462e-01
7.18425438e-02 -3.87126118e-01 -2.35247761e-01 5.30006170e-01
-1.42197955e+00 -4.41777170e-01 -2.88170069e-01 9.73815084e-01
-2.53037140e-02 -1.59705549e-01 6.86887920e-01 5.05539715e-01
-5.19495130e-01 1.10533094e+00 8.80318046e-01 4.04039137e-02
3.80054861e-01 -8.53855252e-01 -9.48612541e-02 4.52226430e-01
2.20268369e-01 5.04131615e-01 3.95543784e-01 -2.92596132e-01
-1.52880013e+00 -7.66730249e-01 1.03361487e+00 4.39245731e-01
3.53336722e-01 -2.02900186e-01 -6.18666530e-01 3.91373575e-01
-2.53441930e-01 3.52408469e-01 9.15061772e-01 -1.13037519e-01
-2.59203434e-01 -8.15196991e-01 -1.32225049e+00 5.44191420e-01
1.05100727e+00 -3.95114034e-01 2.71685570e-01 6.11287236e-01
7.60034546e-02 -3.42253238e-01 -9.63596284e-01 5.18046767e-02
2.65179783e-01 -9.30876493e-01 8.87363553e-01 -1.63860112e-01
-7.47649819e-02 -2.59507000e-02 -2.95340627e-01 -9.73894119e-01
4.04604115e-02 -8.86124313e-01 -2.07324371e-01 5.98292828e-01
4.26987648e-01 -9.94149387e-01 1.32983625e+00 8.80666018e-01
8.25617537e-02 -9.45428908e-01 -1.50930977e+00 -6.37902081e-01
-1.10758007e-01 -5.39136708e-01 5.21261454e-01 5.00733912e-01
-2.85795689e-01 -1.13199502e-01 3.31879519e-02 7.04826266e-02
7.96364307e-01 8.75195190e-02 9.32247043e-01 -8.30067337e-01
-5.64874828e-01 -4.61570472e-01 -2.04307273e-01 -1.39491367e+00
-3.42437774e-02 -7.43406177e-01 -6.32824004e-02 -1.27812994e+00
-1.08909875e-01 -1.00536239e+00 -2.22246140e-01 8.98864746e-01
1.36919305e-01 2.81755716e-01 1.76330596e-01 2.60307193e-01
-7.83575475e-01 5.03442287e-01 1.14475572e+00 3.16120945e-02
-2.28175044e-01 7.15235472e-02 -3.49408388e-01 6.56633854e-01
7.05119014e-01 -7.24559546e-01 -3.24099958e-01 -8.12465549e-01
5.23073435e-01 1.13596320e-01 1.41317442e-01 -8.86374950e-01
5.82963943e-01 6.88787326e-02 3.08224618e-01 -2.68372387e-01
4.76047009e-01 -6.79293752e-01 -1.79330006e-01 7.11340725e-01
-3.21849078e-01 1.16914511e-01 -1.08168907e-01 4.98540103e-01
3.32537964e-02 -1.61669478e-01 7.66735256e-01 -2.04804633e-02
-1.84510484e-01 4.27441567e-01 -3.72172922e-01 8.76702834e-03
9.28278804e-01 -1.84723437e-01 -8.99695307e-02 -4.73878860e-01
-4.77798194e-01 8.12235177e-02 -2.35717613e-02 -3.46624583e-01
5.87355256e-01 -9.46176827e-01 -4.15454626e-01 5.41524172e-01
-5.96190631e-01 4.73716110e-01 3.09703439e-01 1.02787089e+00
-5.97486079e-01 1.64936185e-01 3.62769336e-01 -5.57839215e-01
-1.46839583e+00 -1.58395618e-01 2.23027930e-01 -4.69664574e-01
-5.55309415e-01 1.45352685e+00 -3.66712779e-01 -5.97612001e-02
6.86927795e-01 -3.40482175e-01 5.71644485e-01 -5.16961701e-02
4.23920274e-01 8.03308010e-01 4.88555282e-01 -1.66809827e-01
-2.41751343e-01 4.54766124e-01 1.06469505e-02 -2.63229638e-01
1.47797716e+00 -1.12581842e-01 -1.88816294e-01 -2.10965171e-01
1.63293207e+00 -3.81149314e-02 -1.60361731e+00 -5.72532475e-01
-4.39505368e-01 -4.14545566e-01 2.54636884e-01 -3.71864706e-01
-1.77095187e+00 8.56539130e-01 7.55750239e-01 -1.21661782e-01
1.08125377e+00 -6.11683428e-02 1.05498409e+00 6.20922506e-01
6.77207291e-01 -1.36419940e+00 3.17744136e-01 4.52685267e-01
3.04373056e-01 -9.23364997e-01 1.22799747e-01 5.94183914e-02
-3.60622585e-01 9.32198286e-01 2.86990076e-01 -1.86481297e-01
9.05668437e-01 5.16727448e-01 1.01702191e-01 -1.79168552e-01
-5.45599163e-01 2.30075061e-01 -1.88470155e-01 1.02007270e-01
6.89371601e-02 -1.45560261e-02 -4.44139987e-01 4.81288254e-01
-2.90231347e-01 8.26867893e-02 -1.43965259e-01 1.32927287e+00
-5.18248379e-01 -1.07869411e+00 -3.12475204e-01 5.57648838e-01
-6.23129129e-01 -3.51571254e-02 1.11638948e-01 2.79759020e-01
1.45210102e-01 9.09708917e-01 6.16410673e-02 -1.79534942e-01
-1.19383549e-02 -2.51251999e-02 4.82482582e-01 -2.22428709e-01
-7.96639264e-01 1.30876690e-01 -1.84445783e-01 -6.35006547e-01
-1.02163486e-01 -3.83513898e-01 -1.75549912e+00 -6.75905287e-01
-2.04418883e-01 2.29893044e-01 1.20622158e+00 9.49926436e-01
7.76194870e-01 7.16302842e-02 9.05169606e-01 -6.39712811e-01
-1.39912045e+00 -5.56518018e-01 -8.12692702e-01 1.93254039e-01
6.06392138e-02 -1.63527101e-01 -7.21973300e-01 -2.07267404e-01] | [6.306149005889893, 4.861766338348389] |
a3a5c11c-4087-4422-bb3c-1b01155d86e9 | artifacts-mapping-multi-modal-semantic | 2307.01121 | null | https://arxiv.org/abs/2307.01121v1 | https://arxiv.org/pdf/2307.01121v1.pdf | Artifacts Mapping: Multi-Modal Semantic Mapping for Object Detection and 3D Localization | Geometric navigation is nowadays a well-established field of robotics and the research focus is shifting towards higher-level scene understanding, such as Semantic Mapping. When a robot needs to interact with its environment, it must be able to comprehend the contextual information of its surroundings. This work focuses on classifying and localising objects within a map, which is under construction (SLAM) or already built. To further explore this direction, we propose a framework that can autonomously detect and localize predefined objects in a known environment using a multi-modal sensor fusion approach (combining RGB and depth data from an RGB-D camera and a lidar). The framework consists of three key elements: understanding the environment through RGB data, estimating depth through multi-modal sensor fusion, and managing artifacts (i.e., filtering and stabilizing measurements). The experiments show that the proposed framework can accurately detect 98% of the objects in the real sample environment, without post-processing, while 85% and 80% of the objects were mapped using the single RGBD camera or RGB + lidar setup respectively. The comparison with single-sensor (camera or lidar) experiments is performed to show that sensor fusion allows the robot to accurately detect near and far obstacles, which would have been noisy or imprecise in a purely visual or laser-based approach. | ['Arash Ajoudani', 'Nikolaos Tsagarakis', 'Andrea Zunino', 'Gennaro Raiola', 'Federico Rollo'] | 2023-07-03 | null | null | null | null | ['scene-understanding'] | ['computer-vision'] | [ 5.02899766e-01 2.11167168e-02 5.44705808e-01 -3.84890348e-01
-4.13702726e-01 -4.74785656e-01 5.45804143e-01 4.68265951e-01
-7.70161808e-01 5.51550388e-01 -4.22591239e-01 -7.00536221e-02
-3.64980340e-01 -1.15624905e+00 -6.77882731e-01 -6.31033242e-01
1.66784346e-01 6.82981730e-01 6.70761704e-01 -3.29222023e-01
5.06658256e-01 8.44116211e-01 -2.18372750e+00 -2.74064481e-01
6.58524990e-01 1.15026832e+00 1.01959610e+00 7.34987915e-01
-2.27217451e-01 3.09437186e-01 -3.58261913e-01 2.66441166e-01
2.80112326e-01 5.67941405e-02 -2.72864282e-01 9.00350437e-02
3.65018040e-01 -8.43199119e-02 2.39340849e-02 1.24232268e+00
3.69414479e-01 1.24874532e-01 4.40807939e-01 -1.22833526e+00
8.17247778e-02 -2.54025817e-01 -3.19884837e-01 -3.05361032e-01
6.99813604e-01 -4.33093533e-02 2.02392429e-01 -8.55582535e-01
4.52144414e-01 1.32289350e+00 5.20470798e-01 -3.55590694e-02
-8.47396672e-01 -4.18741465e-01 -5.40041924e-02 4.27152455e-01
-1.55092788e+00 -2.63917774e-01 7.12747037e-01 -4.63510305e-01
5.87582827e-01 2.96679251e-02 4.38764423e-01 4.83185619e-01
4.46943909e-01 8.96518305e-02 1.29375970e+00 -5.80448925e-01
5.78547299e-01 2.92364478e-01 3.41997072e-02 5.55977643e-01
6.73451900e-01 2.74479985e-01 -5.11205196e-01 2.72013485e-01
6.00801647e-01 1.36022687e-01 -2.20938012e-01 -8.75832975e-01
-1.23061931e+00 6.43913269e-01 6.85487568e-01 3.00071865e-01
-5.70457399e-01 1.16530545e-02 4.34769690e-02 3.85826416e-02
-7.33771548e-02 2.07675733e-02 -2.89730608e-01 -1.12547047e-01
-5.07405400e-01 -9.77604389e-02 5.87379456e-01 1.23840714e+00
1.36456919e+00 -2.42910966e-01 7.21566260e-01 2.61027098e-01
5.83264709e-01 1.14820468e+00 3.89640212e-01 -9.43386316e-01
4.04084474e-01 7.35055327e-01 3.83350700e-01 -1.14183748e+00
-6.72890663e-01 4.73845266e-02 -5.09452879e-01 9.66200948e-01
2.85467416e-01 2.77311355e-01 -9.36670244e-01 1.14581859e+00
4.57270145e-01 7.25665987e-02 3.74209642e-01 8.61301780e-01
7.48156130e-01 2.96779871e-01 -3.39795291e-01 -1.63698290e-03
1.35076928e+00 -4.20283526e-01 -7.61376858e-01 -8.11032593e-01
3.93496871e-01 -7.29914725e-01 6.79089427e-01 5.93448937e-01
-3.12229544e-01 -7.76633203e-01 -1.36270130e+00 -1.77838653e-02
-9.06350195e-01 7.62512386e-02 3.45145911e-01 6.04194403e-01
-1.03475690e+00 1.20157264e-01 -8.68153572e-01 -7.77817488e-01
-3.93033326e-02 3.01381260e-01 -8.41108680e-01 -5.99090457e-01
-8.63685012e-01 1.39467406e+00 7.42799699e-01 3.04501742e-01
-6.25703514e-01 -1.79072246e-01 -1.10538185e+00 -2.98897952e-01
4.73487616e-01 -4.11734998e-01 7.61887491e-01 -2.10683510e-01
-1.28886008e+00 5.86830497e-01 -3.22364479e-01 -3.48402143e-01
4.37259376e-01 -4.65787232e-01 -1.83330745e-01 1.88967228e-01
2.55485952e-01 4.49981570e-01 5.09142518e-01 -1.65049815e+00
-1.10216224e+00 -9.86867905e-01 -9.72303972e-02 4.44402993e-01
2.81375259e-01 -5.25639534e-01 -1.77315906e-01 3.52488399e-01
1.04332113e+00 -9.13019300e-01 -2.92428136e-01 1.14413597e-01
-4.52818349e-02 3.26139390e-01 1.17272782e+00 -2.86173075e-01
4.76598442e-01 -2.07027292e+00 -8.35892782e-02 3.50704312e-01
-1.28732011e-01 -6.07822351e-02 3.07145089e-01 4.92673129e-01
4.10142988e-01 -3.82234097e-01 -2.79639244e-01 -4.33349490e-01
-1.93494081e-01 6.31848931e-01 -1.80079602e-02 6.90911531e-01
-4.09939617e-01 3.43192786e-01 -9.45143580e-01 -3.39028865e-01
1.05547595e+00 7.16218710e-01 1.40011817e-01 -1.02676786e-01
1.62458718e-01 6.41271770e-01 -3.28774154e-01 7.31730580e-01
9.35623288e-01 6.14153981e-01 -2.93015572e-03 -6.44916445e-02
-5.82620859e-01 -6.48773089e-02 -1.82341981e+00 2.02365494e+00
-7.37290859e-01 6.49611235e-01 3.75358254e-01 -6.99991465e-01
1.33730578e+00 7.97494128e-02 8.92560109e-02 -9.50958192e-01
7.43850246e-02 6.07200742e-01 -4.70906287e-01 -5.21765232e-01
7.40990400e-01 5.53357191e-02 -1.20457992e-01 -7.36612082e-03
-2.61184335e-01 -5.55204690e-01 -7.10529685e-02 -1.96414113e-01
1.18001938e+00 2.63972193e-01 4.26363707e-01 -1.20343983e-01
7.98581421e-01 2.66521633e-01 2.46395752e-01 7.08068490e-01
-9.06819850e-02 3.65420461e-01 -3.80268276e-01 -2.19146952e-01
-7.33392000e-01 -1.23857784e+00 -1.23675480e-01 4.77747798e-01
9.92501974e-01 1.34488955e-01 -2.53874063e-01 -9.71670151e-02
1.47696897e-01 4.87639219e-01 -3.50149363e-01 3.96695128e-03
-4.26058620e-01 -2.58964747e-01 1.51386892e-03 2.44205803e-01
7.41074145e-01 -8.16849589e-01 -1.63268650e+00 3.31063390e-01
-1.24068260e-01 -1.16756749e+00 7.29986906e-01 4.78778750e-01
-8.58327389e-01 -1.30531061e+00 -4.62188795e-02 -5.48384488e-01
7.23559201e-01 8.00658882e-01 5.99206984e-01 -2.11086199e-01
-3.67597878e-01 7.00152695e-01 -6.20971382e-01 -5.44203520e-01
-1.62695944e-01 -4.26703244e-01 7.57989883e-02 -1.97918668e-01
4.93689656e-01 -5.14653623e-01 -4.15945083e-01 4.79286522e-01
-7.49532580e-01 -3.91688086e-02 8.42844069e-01 2.15910077e-01
6.75688982e-01 5.01481056e-01 3.30988737e-03 -4.63422298e-01
1.20876729e-01 -3.26959223e-01 -9.01086867e-01 -9.01414528e-02
-4.66948539e-01 -1.76095873e-01 -1.42632723e-02 -3.05041187e-02
-8.90429616e-01 6.72308207e-01 9.77936834e-02 -1.31798834e-01
-7.35877573e-01 4.59791541e-01 -4.73607302e-01 -4.37457174e-01
5.72398841e-01 3.00359517e-01 3.54581997e-02 -5.28621554e-01
3.15773487e-01 8.62007320e-01 8.52120578e-01 -2.39364520e-01
8.21751297e-01 8.85680318e-01 4.37654257e-01 -1.00288522e+00
-1.98632181e-01 -9.17277098e-01 -1.12311566e+00 -6.02603495e-01
7.01340497e-01 -9.29975212e-01 -6.98878884e-01 4.05723095e-01
-1.23350513e+00 4.44402359e-02 -6.95670173e-02 8.28836918e-01
-6.69122994e-01 2.53738612e-01 1.69669792e-01 -1.18228269e+00
2.10616380e-01 -1.20393109e+00 1.28397381e+00 2.08704799e-01
1.13551252e-01 -7.06856549e-01 2.28720363e-02 1.84086263e-01
3.06498349e-01 6.26200318e-01 3.11321080e-01 -1.51425511e-01
-8.29976559e-01 -4.14412826e-01 -1.90220803e-01 -2.79617123e-02
3.04931223e-01 -4.25777555e-01 -1.06562841e+00 -6.58260956e-02
2.23939121e-01 1.88999042e-01 4.81648713e-01 1.22635894e-01
1.87786609e-01 4.39245999e-01 -5.71637750e-01 2.16206729e-01
2.05827403e+00 5.20547032e-01 6.30206406e-01 7.42454588e-01
5.08135378e-01 7.40628302e-01 1.29712439e+00 3.46710593e-01
4.85072285e-01 7.71249890e-01 1.19836426e+00 1.92996457e-01
1.15618862e-01 -5.72871231e-02 1.03517883e-01 3.21261376e-01
1.96708292e-02 6.27462473e-03 -1.11016083e+00 4.44236189e-01
-1.98931766e+00 -6.15550280e-01 -7.09867835e-01 2.45931649e+00
-6.45795390e-02 3.44323695e-01 -4.14401472e-01 7.36376762e-01
8.20006251e-01 -2.54733741e-01 -4.51168150e-01 -2.20020682e-01
-2.35044733e-02 6.01631738e-02 1.12465882e+00 9.07473564e-01
-1.01512480e+00 7.22511351e-01 4.97407484e+00 1.81514964e-01
-1.03774774e+00 7.89407715e-02 -5.02334237e-01 4.85803396e-01
8.89211595e-02 -9.25075635e-03 -6.74380183e-01 3.24144393e-01
8.67009103e-01 1.39355436e-01 1.75733685e-01 1.01115048e+00
3.24550271e-01 -1.19480419e+00 -9.82849121e-01 1.12690425e+00
1.75184608e-01 -8.07302713e-01 -4.85466152e-01 2.94436127e-01
2.60897368e-01 1.37771785e-01 -3.24383467e-01 -5.48271164e-02
1.21647097e-01 -7.31058359e-01 1.04341507e+00 8.70185852e-01
4.16859061e-01 -6.41264856e-01 1.15836513e+00 8.91897440e-01
-1.51208103e+00 -1.86686039e-01 -4.56107765e-01 -5.27898967e-01
3.07013929e-01 7.12152183e-01 -1.06014252e+00 9.31436181e-01
9.37155783e-01 5.20830989e-01 -6.45835221e-01 1.26095498e+00
-4.36744720e-01 -3.59869838e-01 -6.67743146e-01 -4.46663909e-02
-1.10571343e-03 -3.45702499e-01 4.26322758e-01 7.52979279e-01
6.77638590e-01 -6.81350529e-02 2.60931641e-01 5.55766761e-01
6.87280476e-01 -1.39356539e-01 -8.93659413e-01 7.59273767e-01
6.47092760e-01 1.14333332e+00 -8.97771001e-01 -2.02846229e-01
-2.42868558e-01 8.36164355e-01 -1.72081143e-01 1.18167989e-01
-4.75124270e-01 -6.23349071e-01 3.96787077e-01 -3.26688425e-03
2.21158251e-01 -7.00967789e-01 -5.76373577e-01 -5.32328069e-01
2.28824884e-01 3.88658717e-02 -1.53117940e-01 -1.27220237e+00
-5.29261112e-01 4.08579081e-01 -4.62066606e-02 -1.46869409e+00
-1.62558392e-01 -9.21975195e-01 -4.25170548e-02 9.90867615e-01
-1.66016901e+00 -1.05163264e+00 -1.11660409e+00 5.55842221e-01
3.80213797e-01 2.57201761e-01 8.14726293e-01 2.30798051e-01
2.05406725e-01 -4.27188754e-01 1.68886662e-01 -3.21827024e-01
4.83439565e-01 -1.16434991e+00 -2.82662928e-01 9.67502236e-01
-9.78863910e-02 3.10899913e-01 9.51139510e-01 -7.34290481e-01
-1.58177543e+00 -8.60385418e-01 6.46685421e-01 -3.70949268e-01
2.85061747e-01 -4.95160729e-01 -6.79568708e-01 5.25514364e-01
-3.97498876e-01 -2.79770140e-02 1.27363950e-01 -3.92168790e-01
1.15843937e-01 -2.63537318e-01 -1.60788584e+00 9.00978521e-02
9.49263155e-01 -4.60529268e-01 -7.50243008e-01 -3.82018164e-02
4.42873061e-01 -5.74359298e-01 -6.09089851e-01 7.86931634e-01
5.81577718e-01 -1.34852254e+00 9.15332139e-01 5.35061717e-01
-2.38480702e-01 -9.06045318e-01 -6.12640083e-01 -1.18095517e+00
1.34286746e-01 1.57889187e-01 2.10110277e-01 9.93036926e-01
-1.10813394e-01 -7.67599881e-01 8.41256201e-01 3.61534446e-01
-3.79153907e-01 3.28387246e-02 -1.23279524e+00 -7.88012803e-01
-6.02211952e-01 -9.15906906e-01 4.66015607e-01 3.96170050e-01
-2.68789351e-01 1.34412653e-03 1.39951304e-01 8.99977446e-01
7.48547196e-01 -1.88456476e-02 1.13611007e+00 -1.71067524e+00
4.61392760e-01 -1.34391814e-01 -1.27744043e+00 -8.27066004e-01
-9.17330608e-02 -3.80921513e-01 5.98399043e-01 -2.19867754e+00
-5.33850014e-01 -6.30706370e-01 9.45809409e-02 3.03978007e-02
3.71446937e-01 2.69990444e-01 -2.14563459e-02 2.45019302e-01
-4.88933295e-01 3.81038517e-01 7.65112162e-01 2.16335636e-02
-1.02576949e-01 1.92839026e-01 -1.38902083e-01 8.44987392e-01
5.40352523e-01 -3.99445564e-01 -3.64822000e-01 -3.96225154e-01
2.52596289e-01 1.40893087e-01 5.67962229e-01 -1.75808859e+00
7.62668252e-01 1.87088437e-02 2.55279094e-01 -1.17472196e+00
7.89159596e-01 -1.65268993e+00 4.50017899e-01 7.57342875e-01
4.19794977e-01 1.49188899e-02 1.44961551e-01 7.21997023e-01
-3.60369295e-01 -4.42043871e-01 6.19735360e-01 -2.93885827e-01
-1.49688268e+00 -2.12329701e-01 -4.44840044e-01 -7.32760429e-01
1.38797355e+00 -8.25949371e-01 -2.54774511e-01 -2.65373260e-01
-7.76690304e-01 1.23566911e-01 7.65392601e-01 3.53984743e-01
1.01016212e+00 -1.21397591e+00 -2.11366266e-01 3.21468264e-01
4.48175699e-01 5.89236140e-01 1.33512229e-01 6.49664283e-01
-9.58075702e-01 6.08574510e-01 -3.83931607e-01 -1.06717622e+00
-1.25855553e+00 5.78305364e-01 1.52076632e-01 3.79206985e-01
-4.81968433e-01 4.64506149e-01 -1.68496504e-01 -7.15642512e-01
2.61286199e-01 -4.42156911e-01 -2.11209908e-01 -1.50377542e-01
4.65866208e-01 7.11777091e-01 3.95455420e-01 -1.00680649e+00
-6.27312541e-01 1.14621520e+00 7.83637524e-01 -3.46776277e-01
1.09413946e+00 -7.19285131e-01 -1.26405239e-01 7.62383878e-01
8.60681713e-01 -5.01323454e-02 -1.01713848e+00 -2.32948080e-01
2.87974119e-01 -6.21145189e-01 8.30596462e-02 -5.92064619e-01
-3.73903036e-01 1.06680596e+00 1.12165177e+00 1.74033940e-01
1.07516396e+00 -1.55338719e-02 1.80453956e-01 6.07591987e-01
1.34955764e+00 -9.95807230e-01 -2.18465060e-01 6.10677540e-01
7.50494719e-01 -1.32616913e+00 1.24845192e-01 -6.61310613e-01
-1.30208999e-01 1.31101072e+00 4.60395843e-01 4.31493223e-02
6.07162654e-01 3.25058013e-01 8.51923600e-02 -2.28268579e-01
-4.48575579e-02 -6.41491771e-01 -2.24204838e-01 1.05800152e+00
-3.57938588e-01 2.24668551e-02 1.46143809e-01 -2.26046219e-02
-3.04328442e-01 -1.06421635e-01 7.98923790e-01 1.43815732e+00
-1.33515024e+00 -6.90197527e-01 -1.05652940e+00 -7.48323137e-03
3.29164922e-01 3.38677704e-01 -8.58638585e-02 9.98181820e-01
6.91133559e-01 1.23890889e+00 2.17960104e-01 -4.05623734e-01
7.30314314e-01 -7.08999857e-02 4.52148885e-01 -6.79624856e-01
1.26264587e-01 -1.24890052e-01 -2.07095131e-01 -7.98194647e-01
-7.08850682e-01 -6.59133196e-01 -1.43916488e+00 5.55080473e-02
-3.18260670e-01 -2.08284408e-02 1.68534696e+00 9.46347058e-01
3.17678377e-02 4.49542910e-01 3.05182010e-01 -1.30099297e+00
9.08164531e-02 -9.10336137e-01 -6.83413327e-01 -1.23161718e-01
3.20241779e-01 -1.12468910e+00 -4.79791880e-01 -1.59189746e-01] | [7.315219879150391, -2.053781270980835] |
bef79f0b-ef03-4106-a5c5-98cc066f51d6 | stochastic-nonsmooth-convex-optimization-with | 2303.12277 | null | https://arxiv.org/abs/2303.12277v3 | https://arxiv.org/pdf/2303.12277v3.pdf | Stochastic Nonsmooth Convex Optimization with Heavy-Tailed Noises: High-Probability Bound, In-Expectation Rate and Initial Distance Adaptation | Recently, several studies consider the stochastic optimization problem but in a heavy-tailed noise regime, i.e., the difference between the stochastic gradient and the true gradient is assumed to have a finite $p$-th moment (say being upper bounded by $\sigma^{p}$ for some $\sigma\geq0$) where $p\in(1,2]$, which not only generalizes the traditional finite variance assumption ($p=2$) but also has been observed in practice for several different tasks. Under this challenging assumption, lots of new progress has been made for either convex or nonconvex problems, however, most of which only consider smooth objectives. In contrast, people have not fully explored and well understood this problem when functions are nonsmooth. This paper aims to fill this crucial gap by providing a comprehensive analysis of stochastic nonsmooth convex optimization with heavy-tailed noises. We revisit a simple clipping-based algorithm, whereas, which is only proved to converge in expectation but under the additional strong convexity assumption. Under appropriate choices of parameters, for both convex and strongly convex functions, we not only establish the first high-probability rates but also give refined in-expectation bounds compared with existing works. Remarkably, all of our results are optimal (or nearly optimal up to logarithmic factors) with respect to the time horizon $T$ even when $T$ is unknown in advance. Additionally, we show how to make the algorithm parameter-free with respect to $\sigma$, in other words, the algorithm can still guarantee convergence without any prior knowledge of $\sigma$. Furthermore, an initial distance adaptive convergence rate is provided if $\sigma$ is assumed to be known. | ['Zhengyuan Zhou', 'Zijian Liu'] | 2023-03-22 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [ 1.55290738e-01 1.50730601e-02 1.78819105e-01 -1.40067324e-01
-1.27247787e+00 -5.52322805e-01 -2.53302783e-01 7.67114619e-03
-6.36266172e-01 9.62119401e-01 -3.43630642e-01 -2.58926421e-01
-4.42476094e-01 -6.78394616e-01 -9.99338806e-01 -1.16866827e+00
-1.92696825e-01 3.09892260e-02 -4.84196655e-02 -2.47713640e-01
1.05527706e-01 1.26197496e-02 -1.29652238e+00 -7.18926370e-01
1.08005774e+00 1.35738528e+00 2.73539066e-01 6.38172328e-01
2.80665845e-01 2.60544837e-01 -3.27911884e-01 -6.08441770e-01
6.43686473e-01 -3.74368429e-01 -3.94920945e-01 2.29657948e-01
1.63718790e-01 1.48959711e-01 -2.00191677e-01 1.78142726e+00
7.16986895e-01 3.79857004e-01 1.69659331e-01 -9.54349339e-01
-4.14339840e-01 5.50419688e-01 -8.92301738e-01 1.82891592e-01
-1.16221175e-01 9.48828235e-02 9.35779274e-01 -8.33236277e-01
2.23265350e-01 8.10675561e-01 7.93728054e-01 3.94493848e-01
-8.14843118e-01 -7.75311112e-01 3.37606221e-01 -4.38847356e-02
-1.60019934e+00 -2.66137302e-01 5.53830445e-01 -3.34352583e-01
3.89388859e-01 2.44623363e-01 2.96020627e-01 4.85815018e-01
9.90257710e-02 6.04182959e-01 1.20797610e+00 -3.01443040e-01
3.89280856e-01 1.40455604e-01 6.22292794e-02 7.99848318e-01
4.58982944e-01 -1.42362177e-01 -3.31149936e-01 -4.52204347e-02
3.18592340e-01 -3.10370445e-01 -7.14363158e-01 -1.24515817e-01
-9.22855556e-01 7.87140250e-01 -1.64315384e-02 3.00453931e-01
-4.11797911e-01 3.11910868e-01 2.06506073e-01 2.86829859e-01
6.87562823e-01 1.05200686e-01 -5.61075389e-01 -4.72086459e-01
-9.17076647e-01 2.21572503e-01 8.08144450e-01 1.29255819e+00
4.83949602e-01 3.94554317e-01 -3.55910540e-01 5.70069015e-01
4.74148653e-02 1.17271948e+00 -6.67769462e-02 -9.82647240e-01
8.30251932e-01 -1.33760616e-01 6.75631881e-01 -1.13847446e+00
-1.84720054e-01 -9.65767384e-01 -9.49882925e-01 -1.08698584e-01
8.50834727e-01 -5.45697033e-01 -4.22362298e-01 2.05274224e+00
2.21116453e-01 3.24080855e-01 -2.49554127e-01 9.91772175e-01
3.95236276e-02 6.00592732e-01 -4.69944268e-01 -8.49480391e-01
1.14512873e+00 -6.38252914e-01 -9.11576807e-01 -1.09071560e-01
4.42442656e-01 -6.40723944e-01 1.39797783e+00 6.13805473e-01
-1.43494511e+00 -6.53924271e-02 -9.83320177e-01 4.83734161e-01
4.33491357e-02 -2.00008154e-01 1.86888918e-01 1.04697275e+00
-9.86121893e-01 3.98281366e-01 -7.23380625e-01 1.05366401e-01
3.46120059e-01 2.28610456e-01 4.90698963e-02 -3.26129287e-01
-1.10129392e+00 4.48614508e-01 -1.98315188e-01 4.43505675e-01
-6.81044042e-01 -7.58222699e-01 -6.36600077e-01 4.81173620e-02
8.87616873e-01 -6.07845664e-01 1.04061687e+00 -6.48242950e-01
-1.60908306e+00 6.12399161e-01 -4.62761730e-01 -4.37928438e-01
8.77582431e-01 -1.81875169e-01 -3.17912027e-02 8.77080187e-02
8.95687044e-02 -4.52271760e-01 9.05401886e-01 -1.01027393e+00
-6.78577840e-01 -5.92508376e-01 1.39001086e-01 1.47520110e-01
-4.36229110e-01 1.57461613e-01 -6.84907019e-01 -6.20306075e-01
-4.55491915e-02 -1.04397583e+00 -6.18899643e-01 -1.05532825e-01
-4.09732401e-01 1.04137920e-01 3.91644448e-01 -5.06323934e-01
1.34655523e+00 -2.19099927e+00 -3.01201511e-02 2.81342417e-01
1.36100799e-02 1.23542204e-01 2.38582984e-01 2.03627810e-01
1.53884873e-01 3.40179086e-01 -7.41211653e-01 -5.59051931e-01
2.55477756e-01 2.55324282e-02 -2.31540382e-01 1.10071373e+00
-3.76250714e-01 5.50352097e-01 -9.18407619e-01 -1.56033173e-01
8.20979849e-03 3.65242988e-01 -4.49242324e-01 -2.18437955e-01
-3.75486277e-02 4.08952266e-01 -7.01029956e-01 4.16300118e-01
1.04435897e+00 -3.53403836e-01 -1.46668315e-01 3.08355600e-01
-1.73557967e-01 -4.97907013e-01 -1.55371451e+00 1.36753356e+00
-6.34744167e-01 4.03083831e-01 8.93685639e-01 -1.60121918e+00
6.14471555e-01 3.19654554e-01 5.84838748e-01 -3.40828925e-01
3.27674687e-01 4.79880214e-01 -3.07922900e-01 -3.12221497e-01
2.40609169e-01 -7.89806724e-01 -9.74578634e-02 3.87692079e-02
-4.76959318e-01 -2.36700609e-01 1.93000019e-01 -1.36102378e-01
1.03480852e+00 -3.81502301e-01 8.93600211e-02 -6.18215680e-01
5.83885491e-01 -3.82387191e-01 7.96103776e-01 1.04447615e+00
-4.44166809e-01 6.08983696e-01 4.48549807e-01 3.78674269e-01
-6.45156920e-01 -8.91563833e-01 -3.09169561e-01 9.55787718e-01
3.74238014e-01 -7.56068295e-03 -7.99460948e-01 -4.59380358e-01
-8.98148865e-02 6.94267511e-01 -6.41428947e-01 9.90693718e-02
-4.19252634e-01 -1.29484105e+00 1.78199708e-01 2.53580511e-01
5.88883400e-01 -4.11160558e-01 -2.67339796e-01 1.33384809e-01
-9.72257704e-02 -1.26475608e+00 -9.03748393e-01 2.00347528e-01
-6.87940180e-01 -1.02133417e+00 -1.13837588e+00 -4.57025856e-01
7.56686270e-01 4.10139054e-01 9.40649152e-01 -1.74241841e-01
4.64183725e-02 6.59454405e-01 -2.57842898e-01 -6.18780911e-01
3.82561624e-01 -4.12308246e-01 1.38176203e-01 3.59945744e-01
-1.59223720e-01 -4.06826288e-01 -6.58300519e-01 5.06992579e-01
-8.89954567e-01 -6.81145072e-01 2.69566983e-01 8.26706350e-01
8.93734157e-01 6.56025410e-01 5.12828708e-01 -7.27324903e-01
5.96165836e-01 -5.18655777e-01 -9.76360798e-01 1.15645207e-01
-6.00983799e-01 -1.48848370e-01 1.00452232e+00 -3.04279476e-01
-8.09143782e-01 -2.98152179e-01 -1.79531142e-01 -5.28585732e-01
4.15406048e-01 6.25419676e-01 -2.06807688e-01 -1.65029541e-01
3.52964431e-01 3.94881427e-01 -9.98761430e-02 -2.46730238e-01
2.51644194e-01 3.75477672e-01 3.36848229e-01 -8.15117121e-01
1.01770747e+00 8.48901093e-01 3.00488293e-01 -9.03104961e-01
-1.07289767e+00 -4.77750361e-01 -1.34801701e-01 -5.36865443e-02
4.22869027e-01 -7.59309471e-01 -1.11452353e+00 3.76677185e-01
-6.08139217e-01 -2.90208727e-01 -5.48816085e-01 6.72842085e-01
-9.01388466e-01 5.47847569e-01 -3.55306298e-01 -1.48097253e+00
-1.81500599e-01 -1.07992661e+00 6.20039105e-01 1.42936409e-01
4.78408068e-01 -1.07914376e+00 -2.79587448e-01 2.64052659e-01
6.43496096e-01 3.75960529e-01 3.30636472e-01 -2.75994778e-01
-4.08830553e-01 -4.44238752e-01 -1.67259693e-01 6.57967150e-01
-1.37822881e-01 -3.60097796e-01 -4.39466476e-01 -5.40231287e-01
6.67359412e-01 5.43356091e-02 6.14296973e-01 9.70022082e-01
1.40018833e+00 -5.11575818e-01 -1.68075860e-01 7.05602765e-01
1.61991370e+00 1.52791455e-01 2.26643726e-01 2.06505150e-01
1.72079220e-01 4.51819807e-01 7.74824321e-01 9.44698155e-01
2.03893051e-01 4.56943393e-01 6.49519324e-01 1.93585739e-01
5.02423108e-01 1.09576881e-01 3.61060351e-01 8.08380961e-01
-1.69185266e-01 -5.17910957e-01 -6.40228093e-01 6.91064954e-01
-1.71593118e+00 -8.13177705e-01 -2.92269915e-01 2.77121139e+00
8.79723608e-01 1.13156721e-01 7.16887973e-03 1.00172162e-01
9.06987846e-01 1.14880688e-01 -7.06879377e-01 -6.64655045e-02
-3.98775250e-01 4.51224208e-01 1.20506310e+00 7.04312801e-01
-8.74123156e-01 4.88972604e-01 5.43702555e+00 1.27665162e+00
-9.26294863e-01 2.95547009e-01 6.98113322e-01 -3.92428726e-01
-2.47550681e-01 -1.77729428e-01 -7.47817576e-01 8.95950735e-01
7.80709624e-01 -5.72294950e-01 5.23301482e-01 9.83245134e-01
6.22554779e-01 -1.34791732e-01 -4.84888554e-01 1.18346596e+00
-8.51478800e-02 -8.08129370e-01 -7.81760335e-01 2.68495589e-01
8.47655237e-01 -1.11605994e-01 4.51598465e-01 2.29098126e-01
4.09790605e-01 -9.34627533e-01 5.51270187e-01 4.79042530e-01
6.84369922e-01 -1.04765821e+00 1.00758600e+00 5.89874029e-01
-1.29216444e+00 -2.21263885e-01 -4.37077463e-01 -8.69836435e-02
5.31854093e-01 1.16410434e+00 -1.00849047e-02 7.05199540e-01
9.03124988e-01 3.45892638e-01 1.63505569e-01 1.21119475e+00
-1.38494417e-01 9.08878505e-01 -6.87436104e-01 -1.61771134e-01
3.79373759e-01 -4.38091218e-01 8.38891149e-01 1.08899665e+00
9.27058995e-01 3.84191185e-01 3.12557280e-01 4.31334972e-01
-2.13786177e-02 4.49574172e-01 -2.98012584e-01 1.64182141e-01
3.06926936e-01 8.85117292e-01 -5.10455072e-01 -1.71988681e-01
-4.83920157e-01 8.18903685e-01 -9.15370584e-02 5.70385039e-01
-1.05828190e+00 -6.75652564e-01 6.87804520e-01 1.68630421e-01
5.59431851e-01 -5.59401453e-01 -4.88739967e-01 -1.18520856e+00
5.91664255e-01 -4.70713556e-01 3.81554395e-01 -2.09900081e-01
-1.32179749e+00 3.31916898e-01 -2.08117664e-01 -1.01488519e+00
1.49776205e-01 -5.13926387e-01 -4.19080824e-01 1.03040469e+00
-1.53496456e+00 -4.68676120e-01 7.93134272e-02 7.53523290e-01
3.66098136e-01 2.12826923e-01 2.66955316e-01 3.55046690e-01
-7.47761786e-01 7.37927437e-01 7.77598262e-01 -8.92446414e-02
4.78973567e-01 -1.27755594e+00 -4.20244545e-01 1.07070148e+00
-2.69518435e-01 6.08356118e-01 1.21996260e+00 -3.46576869e-01
-1.55913150e+00 -9.88071442e-01 6.62083268e-01 -2.52265632e-01
9.10622716e-01 -2.16964856e-01 -8.09687376e-01 3.05873662e-01
8.69746879e-02 5.44247627e-01 5.92043519e-01 -5.29493280e-02
1.44366041e-01 -3.03974420e-01 -1.29252386e+00 4.56987470e-01
1.12650967e+00 -1.26405835e-01 3.83903868e-02 4.72303450e-01
6.14031672e-01 -4.44403172e-01 -8.39302301e-01 6.02354467e-01
2.13629991e-01 -7.78578222e-01 7.77677059e-01 -5.06973863e-01
-2.62274984e-02 -3.09962720e-01 -6.34853303e-01 -1.15614891e+00
1.39674157e-01 -1.13643277e+00 -5.93050942e-02 9.36706126e-01
5.02008796e-01 -8.05991650e-01 8.52113187e-01 6.65237010e-01
-2.76615292e-01 -1.07478631e+00 -1.37397468e+00 -1.30599678e+00
3.79106432e-01 -9.07540262e-01 -1.71119254e-02 6.14597023e-01
-1.21051423e-01 -1.80001080e-01 -5.58164418e-01 5.05633712e-01
8.89124632e-01 -2.59488802e-02 4.67492521e-01 -7.21976399e-01
-5.34234941e-01 -5.22867382e-01 -1.94025645e-03 -1.43976760e+00
1.10161886e-01 -6.20563686e-01 4.51019764e-01 -1.20962727e+00
6.37799352e-02 -5.99375904e-01 -4.90875721e-01 -1.27131388e-01
-4.12217170e-01 1.27450123e-01 1.25182018e-01 -5.63275218e-02
-6.94569588e-01 8.55698764e-01 1.42526019e+00 1.12021910e-02
-1.14435986e-01 5.94251037e-01 -9.45181012e-01 9.10339534e-01
6.44301832e-01 -3.54581207e-01 -4.46496189e-01 -2.96416342e-01
4.23261642e-01 3.59856248e-01 1.09454885e-01 -8.75000179e-01
1.45160541e-01 -3.58378232e-01 -4.61021394e-01 -3.35434467e-01
2.23165482e-01 -8.21706653e-01 -2.63046533e-01 2.56027579e-01
-9.91406292e-02 -1.42836526e-01 -8.55294317e-02 1.13003159e+00
-1.49293348e-01 -4.32131737e-01 1.05129457e+00 -2.53614113e-02
-1.05844900e-01 7.29395092e-01 -2.55303413e-01 5.80011368e-01
1.12703431e+00 -1.29659697e-02 -8.41818303e-02 -8.75815690e-01
-8.49507153e-01 3.94401193e-01 1.94099292e-01 -1.59594327e-01
3.62261981e-01 -9.59229529e-01 -7.87679374e-01 -4.13004458e-01
-2.39883572e-01 9.59416926e-02 6.65624499e-01 1.34435809e+00
-9.25773159e-02 3.36004555e-01 8.44675124e-01 -4.91330773e-01
-7.52418816e-01 7.80621648e-01 2.20932767e-01 -3.00502807e-01
-3.50516140e-01 1.18386447e+00 1.58246204e-01 -5.48352487e-03
3.67324561e-01 -3.74265254e-01 3.28796953e-01 -1.71994835e-01
4.66958851e-01 5.41249394e-01 -9.63691249e-02 -4.50117409e-01
-2.88223267e-01 8.56521666e-01 4.65320736e-01 -3.42081755e-01
1.18581426e+00 -5.77317834e-01 4.92781848e-02 3.44978690e-01
1.32264817e+00 6.61495984e-01 -1.47257090e+00 -4.90841657e-01
-1.51771694e-01 -6.39236271e-01 -1.58152491e-01 -4.26016599e-01
-1.41613662e+00 8.98018420e-01 4.17964518e-01 2.99046516e-01
1.25865626e+00 -1.27791151e-01 8.42279375e-01 2.46786177e-01
6.84044600e-01 -1.35683310e+00 -1.07054114e-01 6.03620946e-01
7.21862912e-01 -1.20586610e+00 -2.46485740e-01 -6.23737991e-01
-4.70010519e-01 8.37897718e-01 2.43542522e-01 -1.67087570e-01
1.06933272e+00 3.98824483e-01 -3.61589253e-01 1.13621630e-01
-2.93519408e-01 -4.18133140e-01 -1.06100307e-03 3.00619453e-01
3.24454457e-01 6.62895218e-02 -8.01083684e-01 9.43829477e-01
-3.01709384e-01 1.12306699e-02 4.05371070e-01 7.84916401e-01
-6.57486200e-01 -7.05890238e-01 -4.73075449e-01 3.66859913e-01
-1.02161062e+00 -2.31596604e-01 4.14552778e-01 7.06532776e-01
-1.03522260e-02 1.26466191e+00 -3.91969144e-01 -1.83975007e-02
4.39754248e-01 -2.79197752e-01 2.10668266e-01 -3.57562900e-01
-1.52397007e-01 9.66850221e-02 -9.45747718e-02 -4.31601703e-01
-3.44288141e-01 -7.45618641e-01 -9.99386609e-01 -4.04475570e-01
-5.04052818e-01 4.87054586e-01 6.44466698e-01 1.01068306e+00
7.83394054e-02 2.83260643e-01 9.09506738e-01 -3.58611941e-01
-9.78973567e-01 -7.71197379e-01 -1.05867898e+00 5.05827889e-02
2.85682827e-01 -5.49200177e-01 -8.34588885e-01 -1.37190327e-01] | [6.583136558532715, 4.526710033416748] |
c801981c-d745-4718-96ba-edcc7bd71c25 | mutr-multi-stage-transformer-for-hand-pose | null | null | https://www.mdpi.com/1424-8220/23/12/5509 | https://www.mdpi.com/1424-8220/23/12/5509 | MuTr: Multi-Stage Transformer for Hand Pose Estimation from Full-Scene Depth Image | This work presents a novel transformer-based method for hand pose estimation—DePOTR. We test the DePOTR method on four benchmark datasets, where DePOTR outperforms other transformer-based methods while achieving results on par with other state-of-the-art methods. To further demonstrate the strength of DePOTR, we propose a novel multi-stage approach from full-scene depth image—MuTr. MuTr removes the necessity of having two different models in the hand pose estimation pipeline—one for hand localization and one for pose estimation—while maintaining promising results. To the best of our knowledge, this is the first successful attempt to use the same model architecture in standard and simultaneously in full-scene image setup while achieving competitive results in both of them. On the NYU dataset, DePOTR and MuTr reach precision equal to 7.85 mm and 8.71 mm, respectively. | ['Marek Hrúz', 'Jakub Straka', 'Matyáš Boháček', 'Zdeněk Krňoul', 'Ivan Gruber', 'Jakub Kanis'] | 2023-06-12 | null | null | null | sensors-2023-6 | ['pose-estimation', 'hand-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-1.37283549e-01 -7.13425875e-02 1.79986745e-01 1.69483751e-01
-1.08799684e+00 -6.99656248e-01 2.28260994e-01 -3.46161515e-01
-5.53591073e-01 6.07502997e-01 1.62405148e-01 1.55140366e-02
-1.23570681e-01 -3.48923922e-01 -5.95655620e-01 -5.78040004e-01
2.81564891e-01 1.01360738e+00 6.31524563e-01 -7.14122653e-02
2.35143527e-01 8.35893452e-01 -1.35392499e+00 1.40452728e-01
8.35517123e-02 8.67565572e-01 5.51323891e-01 7.74545491e-01
3.95788521e-01 6.58366203e-01 -5.14608145e-01 -4.68224585e-01
3.77597332e-01 2.04456940e-01 -9.22820508e-01 -8.09419528e-02
8.97514403e-01 -8.69044125e-01 -5.90759635e-01 2.77476341e-01
1.44751215e+00 -2.73059100e-01 5.23010433e-01 -6.43197060e-01
1.27518829e-02 5.25203764e-01 -7.66120791e-01 -2.33767107e-02
6.84159100e-01 1.81530535e-01 9.43099618e-01 -9.15442467e-01
1.04642880e+00 1.11373818e+00 8.51631522e-01 3.91885012e-01
-1.13909864e+00 -4.96759951e-01 9.98765826e-02 -7.51600713e-02
-1.50723159e+00 -3.37843895e-01 6.69621706e-01 -5.66117465e-01
1.25242627e+00 2.77171373e-01 6.49607301e-01 1.13522887e+00
8.80794078e-02 1.06010580e+00 1.23653400e+00 -5.49901366e-01
-1.14657693e-01 -1.05485179e-01 -2.26015061e-01 6.96680069e-01
5.61641864e-02 -3.77045311e-02 -7.41067052e-01 -1.77870896e-02
9.69437659e-01 -1.79418653e-01 -2.99800754e-01 -3.59291703e-01
-1.31308663e+00 2.05405116e-01 4.83876944e-01 5.45849323e-01
-3.97046655e-01 3.55521530e-01 2.44316414e-01 -7.69382939e-02
2.10834607e-01 4.14835930e-01 -5.97041011e-01 -3.28486890e-01
-1.04347277e+00 4.81463641e-01 4.97781247e-01 1.05423653e+00
3.72578613e-02 -4.26444978e-01 -5.09438813e-01 6.69947088e-01
3.28816533e-01 7.66796768e-01 5.22390828e-02 -5.77303767e-01
7.54261374e-01 2.87513524e-01 1.56851634e-01 -6.22898459e-01
-8.11456859e-01 -8.77639413e-01 -3.47391278e-01 1.05412550e-01
1.06441629e+00 -1.21169962e-01 -1.06435537e+00 1.29409122e+00
2.15678617e-01 -1.89370051e-01 -6.09653771e-01 1.10021806e+00
8.50736141e-01 4.53656167e-02 -2.52614707e-01 1.37465164e-01
1.40998125e+00 -1.00383580e+00 -3.63842577e-01 -2.06550837e-01
3.60595524e-01 -1.26431191e+00 1.08231199e+00 8.95557404e-01
-9.99026954e-01 -1.39348418e-01 -8.07920039e-01 -9.63188559e-02
9.11090374e-02 8.73339951e-01 6.52753294e-01 7.70531952e-01
-1.05653274e+00 5.46263635e-01 -8.44466209e-01 -5.45611024e-01
3.88457328e-01 5.30402601e-01 -4.87659633e-01 -2.77627110e-02
-5.97365737e-01 9.14545298e-01 -7.56520554e-02 1.79728732e-01
-7.07687318e-01 -7.03639388e-01 -2.55797267e-01 -3.25951189e-01
4.85577464e-01 -6.13340974e-01 1.54704988e+00 1.33616760e-01
-1.78178477e+00 9.37296748e-01 -2.76667595e-01 -9.84731391e-02
1.01782310e+00 -5.98372340e-01 1.94518879e-01 2.95160264e-01
4.82861884e-02 4.78063524e-01 8.38304222e-01 -1.30470586e+00
-4.02104974e-01 -8.55731368e-01 8.06869194e-02 4.36432362e-02
-2.48488829e-01 1.76400095e-02 -8.34687650e-01 -7.04107761e-01
3.89496267e-01 -1.27863503e+00 6.67484924e-02 2.14251410e-02
-7.13168800e-01 -5.98334186e-02 7.64048457e-01 -9.39306498e-01
1.02395797e+00 -1.52321112e+00 4.72201258e-01 9.95962024e-02
4.03377771e-01 3.01628232e-01 1.05567314e-02 5.67990541e-01
2.59068102e-01 -4.90656435e-01 1.85968622e-03 -8.47885132e-01
-5.05451337e-02 -2.94125348e-01 -4.04602140e-02 5.04889727e-01
-3.68407637e-01 9.75454509e-01 -5.48095167e-01 -4.43789274e-01
3.61492813e-01 7.60015368e-01 -5.12870014e-01 9.37125366e-03
-4.44242023e-02 7.87696302e-01 -3.56709898e-01 8.86965811e-01
8.03628743e-01 3.07012517e-02 2.78431267e-01 -4.59899396e-01
-2.84063876e-01 1.33700326e-01 -1.19031703e+00 1.94356585e+00
-5.67796171e-01 6.61124229e-01 3.44802206e-03 -3.98504078e-01
5.95852673e-01 5.33780515e-01 6.03130758e-01 -5.00659168e-01
4.49159682e-01 4.80389386e-01 -1.39154375e-01 -3.59626144e-01
3.77993375e-01 -2.18476709e-02 3.33706766e-01 3.98538768e-01
1.66183978e-01 -2.80085355e-01 -8.72783661e-02 -4.51521464e-02
9.98359799e-01 4.27393407e-01 -9.58263874e-02 -1.77978650e-01
4.13909227e-01 -2.69808471e-01 -9.52020809e-02 7.07940638e-01
1.45868473e-02 1.04749453e+00 3.51372868e-01 -2.17469409e-01
-1.02353394e+00 -1.18170559e+00 -1.26734957e-01 9.02261257e-01
-2.74838448e-01 -5.41324198e-01 -1.00764716e+00 -6.20967150e-01
3.18429798e-01 2.81753540e-02 -5.90206206e-01 7.20169663e-01
-7.45272100e-01 -4.42040265e-01 5.93062997e-01 9.95848417e-01
5.26440263e-01 -7.65047908e-01 -7.74473250e-01 1.82955593e-01
-1.78810865e-01 -1.35260201e+00 -5.26745081e-01 6.38545826e-02
-6.09651148e-01 -1.01431739e+00 -1.29140317e+00 -6.17283940e-01
2.48702317e-01 2.25002300e-02 8.44289601e-01 -2.48796850e-01
-7.49810159e-01 3.60412598e-01 -2.21866548e-01 -2.89837807e-01
3.67639452e-01 7.07128406e-01 -2.46237119e-04 -4.09186393e-01
-3.24362010e-01 -6.73150063e-01 -1.00517464e+00 5.49517870e-01
-1.54388174e-01 -2.33521238e-01 7.60937989e-01 5.63557386e-01
4.82956946e-01 -4.09204364e-01 2.50588953e-01 -4.77663010e-01
2.63867795e-01 4.07491088e-01 -6.18044615e-01 7.05294982e-02
-5.16202092e-01 3.49134877e-02 4.24905181e-01 -3.92047942e-01
-9.79920447e-01 4.71892536e-01 -4.82892841e-01 -2.47783720e-01
7.35206529e-02 8.17957819e-02 -9.52953771e-02 -4.26301628e-01
5.40889800e-01 1.96465388e-01 -1.11323968e-01 -9.03662980e-01
1.26159281e-01 8.64023566e-01 4.88418609e-01 -4.95569974e-01
5.74519992e-01 7.94219792e-01 3.28435421e-01 -7.93310583e-01
-6.66065812e-01 -6.21004641e-01 -1.02333272e+00 -2.78202415e-01
7.16619909e-01 -6.21713817e-01 -1.03269076e+00 1.09113836e+00
-1.26782513e+00 -4.62277204e-01 4.63662334e-02 6.27925754e-01
-6.34485185e-01 2.98193812e-01 -5.59220076e-01 -8.77705932e-01
-4.54623312e-01 -1.40211320e+00 1.62458467e+00 -2.45498806e-01
-2.45622054e-01 -6.35937035e-01 -1.42354503e-01 5.90050280e-01
3.21313500e-01 1.18318439e-01 2.68983990e-01 -8.90781358e-02
-6.70959651e-01 -4.91020828e-01 -3.04636091e-01 -1.84086248e-01
8.68845358e-02 -2.53376245e-01 -1.41130650e+00 -6.13418341e-01
-3.92390430e-01 -1.49989769e-01 7.42838740e-01 7.71553934e-01
9.16180551e-01 3.49306792e-01 -5.27874410e-01 4.32582110e-01
1.29729342e+00 -7.45700672e-02 6.33262038e-01 5.33783436e-01
9.04163718e-01 3.85702759e-01 4.51599240e-01 6.25272870e-01
4.49046642e-01 1.38278651e+00 4.12828594e-01 3.67535986e-02
-6.60792232e-01 -1.58681259e-01 -1.11597687e-01 3.32254559e-01
-7.49190152e-01 -1.54256284e-01 -1.10926020e+00 3.69083136e-01
-1.71476042e+00 -6.54371738e-01 -1.50929451e-01 2.12873626e+00
3.52240831e-01 -1.75972171e-02 5.97849250e-01 4.51471508e-01
2.95026600e-01 7.19281733e-02 -4.22417909e-01 1.53298900e-01
1.85409382e-01 5.89695871e-01 6.37884557e-01 5.33189178e-01
-1.21186256e+00 1.06598711e+00 6.39023781e+00 5.74012995e-01
-1.21862686e+00 4.11634773e-01 -5.25077619e-02 -3.30839604e-01
2.86365658e-01 -2.84091622e-01 -1.00035155e+00 1.59587078e-02
1.39666470e-02 6.05165422e-01 4.63603258e-01 7.95725703e-01
-5.75663745e-02 -3.43801737e-01 -1.02490640e+00 1.23253155e+00
2.51012206e-01 -9.14372027e-01 -3.31149697e-01 1.43094137e-01
3.68333727e-01 1.28082884e-02 1.87202618e-01 -2.72531837e-01
-3.79306853e-01 -9.93598282e-01 1.12008345e+00 4.30884808e-01
9.61253345e-01 -6.41887307e-01 8.58108461e-01 2.75358796e-01
-1.65894651e+00 -1.14822630e-02 8.35069120e-02 8.30709469e-03
2.58984894e-01 3.57438356e-01 -9.30260658e-01 5.91559052e-01
8.11710179e-01 4.39199060e-01 -7.30871201e-01 9.36758399e-01
-3.38394195e-01 2.95288116e-01 -5.00521123e-01 -4.79708202e-02
-5.31794168e-02 5.07736623e-01 7.74316728e-01 1.17265904e+00
1.65285528e-01 -2.39657551e-01 -1.15450107e-01 4.34936434e-01
1.95000738e-01 2.68960204e-02 -2.38919139e-01 3.66973698e-01
4.53283906e-01 1.23504806e+00 -8.46007705e-01 5.87424226e-02
1.63521878e-02 1.36304188e+00 2.97434807e-01 -9.33788251e-03
-6.37889266e-01 -3.68776619e-01 2.52069563e-01 6.65366352e-01
5.36007106e-01 -6.80890560e-01 -4.45212841e-01 -9.96291697e-01
5.81846118e-01 -6.54346406e-01 1.18322775e-01 -7.51551270e-01
-7.84495592e-01 8.00414860e-01 2.29091104e-02 -1.17228258e+00
-2.55676091e-01 -1.10424769e+00 -8.34973529e-02 9.94449139e-01
-1.33618402e+00 -1.61271834e+00 -6.12627506e-01 6.02425694e-01
4.14349347e-01 -1.91654995e-01 9.35719013e-01 3.73488665e-01
-3.18230748e-01 1.07833433e+00 1.86797548e-02 5.55344112e-02
6.86807573e-01 -1.22851527e+00 4.97320473e-01 5.27283549e-01
1.03360415e-01 6.61844373e-01 6.25959337e-01 -4.66708392e-01
-1.81743622e+00 -5.14644384e-01 6.69249058e-01 -9.50195849e-01
3.14614326e-01 -5.69118500e-01 -1.47201523e-01 6.23282611e-01
-1.56725645e-01 -8.01297501e-02 3.95191908e-02 3.59973341e-01
-3.91709745e-01 -7.92693794e-02 -1.36997283e+00 5.02018154e-01
1.44496441e+00 -5.35488129e-01 -2.91054666e-01 4.77254033e-01
2.12332532e-01 -8.98220956e-01 -8.84687185e-01 4.12308425e-01
1.37020242e+00 -1.18356729e+00 1.22373343e+00 3.48869748e-02
1.88827604e-01 -1.02735974e-01 -3.18277538e-01 -1.01020801e+00
-2.00490475e-01 -6.11006260e-01 -1.57796875e-01 9.85367656e-01
2.22227126e-01 -8.34311485e-01 1.18448746e+00 1.97241153e-03
1.01180732e-01 -1.03768325e+00 -1.21162868e+00 -1.05133986e+00
4.53480780e-02 -5.11472642e-01 4.67098802e-01 1.73951894e-01
-1.99402660e-01 3.08251590e-01 -3.01983356e-01 1.46974280e-01
9.16350722e-01 8.12437609e-02 9.56632793e-01 -1.21245706e+00
-4.08380270e-01 -3.29268217e-01 -5.19134164e-01 -1.23085070e+00
-1.11570969e-01 -5.81510425e-01 -6.78818226e-02 -1.90360677e+00
2.06120536e-01 -3.76055598e-01 3.56381983e-01 4.98415053e-01
1.91886306e-01 7.11039603e-01 4.09191579e-01 2.20563620e-01
3.35165672e-02 1.76171616e-01 1.46361876e+00 3.67489755e-02
-8.54049474e-02 2.97669182e-03 -2.73949891e-01 7.32292295e-01
7.09422171e-01 -9.08712000e-02 -2.62612492e-01 -6.44435883e-01
-2.37827957e-01 2.80680205e-03 5.02395868e-01 -1.30146301e+00
1.24637604e-01 3.84132594e-01 3.49602938e-01 -9.65650022e-01
6.04437053e-01 -8.23952198e-01 6.23930357e-02 6.59827650e-01
2.28300631e-01 -1.49477631e-01 2.72178411e-01 -1.48997521e-02
4.01765227e-01 5.59952259e-02 8.13808441e-01 6.91163912e-02
-4.44129437e-01 3.43145430e-01 -5.48437648e-02 -2.58697122e-01
8.35673332e-01 -3.75627071e-01 -3.01224798e-01 -3.84556592e-01
-7.80940890e-01 -1.08828031e-01 1.64791837e-01 4.19044971e-01
5.27984440e-01 -9.89443123e-01 -7.33383238e-01 4.08103652e-02
-1.30108697e-02 -4.03889595e-03 2.42511466e-01 1.12605190e+00
-7.44482636e-01 9.67959940e-01 -5.32214753e-02 -8.34652841e-01
-1.58653569e+00 2.80871093e-01 4.10052598e-01 -4.48514044e-01
-8.73640418e-01 1.10152555e+00 -1.06263123e-01 -6.31346524e-01
3.60785544e-01 -3.29567730e-01 2.25576267e-01 6.09376989e-02
3.97933036e-01 7.03497231e-01 6.53177798e-01 -6.56918466e-01
-8.46516371e-01 1.24809849e+00 -1.51841760e-01 -5.09435654e-01
1.29456270e+00 1.15685254e-01 8.64119753e-02 3.81431431e-02
9.27603126e-01 4.43097621e-01 -1.29287434e+00 -1.06658433e-02
-2.83768833e-01 -7.11896837e-01 2.62469471e-01 -1.31010187e+00
-1.21710575e+00 9.50300157e-01 9.46389675e-01 -3.93712223e-01
9.86110389e-01 3.45431566e-01 8.90262425e-01 2.94170856e-01
9.53245223e-01 -9.39578176e-01 2.71149009e-01 3.71238679e-01
1.29458964e+00 -8.58423591e-01 2.96225637e-01 -8.18895578e-01
-3.53895187e-01 1.00508964e+00 6.24090850e-01 -3.30755152e-02
5.20573437e-01 7.32991934e-01 1.61144003e-01 -2.77103484e-01
-3.97401303e-02 -3.17476600e-01 4.54982996e-01 6.60736084e-01
6.91984057e-01 1.05912760e-01 -2.79703051e-01 2.92169809e-01
-4.04030621e-01 1.81962892e-01 -2.28915870e-01 1.14786017e+00
-6.25356883e-02 -1.35990357e+00 -5.94632089e-01 2.67766982e-01
-3.39909464e-01 5.56517132e-02 -5.16709805e-01 8.62214148e-01
-1.48816053e-02 7.67511725e-01 -3.04981560e-01 -6.13145709e-01
8.37130487e-01 -2.22725049e-01 1.49684119e+00 -4.44233119e-01
-7.49498844e-01 3.13885361e-01 1.37275492e-04 -6.42041385e-01
-1.02763437e-01 -7.22048104e-01 -1.02566934e+00 -2.49491125e-01
-5.42896450e-01 -4.49272424e-01 6.34601116e-01 9.73878145e-01
2.33919322e-01 5.93171299e-01 3.39033723e-01 -1.45423150e+00
-5.97697496e-01 -1.26500368e+00 -8.61217678e-01 -1.51007265e-01
2.75991291e-01 -1.26730955e+00 4.53726500e-02 -4.45231944e-01] | [6.60228157043457, -0.7454303503036499] |
a76104d1-e2c1-42ee-934e-dd086fc743f0 | nested-named-entity-recognition-via | null | null | https://aclanthology.org/2021.acl-long.275 | https://aclanthology.org/2021.acl-long.275.pdf | Nested Named Entity Recognition via Explicitly Excluding the Influence of the Best Path | This paper presents a novel method for nested named entity recognition. As a layered method, our method extends the prior second-best path recognition method by explicitly excluding the influence of the best path. Our method maintains a set of hidden states at each time step and selectively leverages them to build a different potential function for recognition at each level. In addition, we demonstrate that recognizing innermost entities first results in better performance than the conventional outermost entities first scheme. We provide extensive experimental results on ACE2004, ACE2005, and GENIA datasets to show the effectiveness and efficiency of our proposed method. | ['Taro Watanabe', 'Yuji Matsumoto', 'Hiroyuki Shindo', 'Yiran Wang'] | 2021-08-01 | null | null | null | acl-2021-5 | ['nested-named-entity-recognition'] | ['natural-language-processing'] | [-1.39244109e-01 2.88109064e-01 -5.74150085e-01 -5.64091206e-01
-8.72886539e-01 -5.58558226e-01 4.65556860e-01 2.49013960e-01
-4.22291219e-01 7.56723762e-01 5.18747509e-01 -5.09001851e-01
-2.17550974e-02 -8.35573316e-01 -7.68793821e-01 -3.63261551e-01
-4.66493875e-01 3.69183093e-01 6.48186207e-01 1.39800712e-01
5.23813069e-02 6.96699739e-01 -1.07851565e+00 4.78412032e-01
7.07647920e-01 5.94446957e-01 -1.46092385e-01 4.19558167e-01
-3.55839461e-01 9.01126921e-01 -4.86845285e-01 -6.63463056e-01
1.64817035e-01 1.82697494e-02 -1.12306798e+00 -2.61860073e-01
-4.13808366e-03 -1.19295619e-01 -4.88414288e-01 7.08050847e-01
2.68216163e-01 2.27810770e-01 1.29034653e-01 -9.44033861e-01
-3.23232919e-01 1.31017733e+00 -3.17858368e-01 3.47453147e-01
3.31548542e-01 -3.50085974e-01 1.32018793e+00 -1.10123932e+00
7.50800788e-01 1.00072396e+00 6.35975599e-01 4.39067394e-01
-9.20979261e-01 -7.73919165e-01 8.26046944e-01 8.77084509e-02
-1.56055796e+00 -6.96416497e-01 3.48274529e-01 -7.71901831e-02
1.40942180e+00 1.48191303e-01 1.59454241e-01 7.36575007e-01
-2.55435556e-01 1.15250754e+00 7.43043363e-01 -3.42943162e-01
1.12353861e-01 -1.13209441e-01 8.08480859e-01 9.82918382e-01
4.48762596e-01 6.36315495e-02 -4.12362069e-01 -6.49183452e-01
2.97496855e-01 -1.31236374e-01 -1.32612497e-01 -3.36675256e-01
-1.13747239e+00 4.81222808e-01 3.79052073e-01 2.80891031e-01
-4.33193654e-01 -5.15042283e-02 1.40972957e-01 -1.87110752e-02
1.47068381e-01 2.12577060e-01 -7.46771872e-01 -1.71802536e-01
-8.02635670e-01 -2.56553501e-01 1.29875910e+00 1.17667174e+00
6.89774334e-01 -1.21869214e-01 -3.00408721e-01 6.19019926e-01
5.47203243e-01 1.56651512e-01 1.61507934e-01 -4.44734842e-01
5.96925020e-01 7.91200280e-01 2.56677240e-01 -8.33276451e-01
-4.20591176e-01 -4.93787110e-01 -4.78447139e-01 -5.32810152e-01
2.60957815e-02 -2.94119745e-01 -1.13039815e+00 1.72293460e+00
5.86500227e-01 5.00129461e-01 3.17835510e-01 3.78651917e-01
6.62374020e-01 8.76532435e-01 3.01258087e-01 -1.41155243e-01
1.03686059e+00 -1.20886934e+00 -6.26949251e-01 -2.32173815e-01
8.06151152e-01 -2.21070528e-01 1.25692114e-01 -2.98756175e-03
-9.50761497e-01 -2.45552152e-01 -9.57969069e-01 1.69813126e-01
-5.12871623e-01 2.40532830e-01 9.08015370e-01 9.02771294e-01
-8.75751317e-01 4.84733462e-01 -1.31213498e+00 -1.36574671e-01
1.12847738e-01 5.08985102e-01 -4.74384606e-01 3.01088929e-01
-1.46986234e+00 7.68135786e-01 8.83171678e-01 4.21295553e-01
-6.46195829e-01 -3.73585552e-01 -1.10564733e+00 2.44434223e-01
4.47001576e-01 -2.42700025e-01 1.25711238e+00 -1.98062852e-01
-1.32102025e+00 4.01547402e-01 -7.66259313e-01 -6.86758816e-01
3.02023925e-02 -3.42146814e-01 -8.34812760e-01 -2.17144787e-01
-1.34966979e-02 2.74393916e-01 -1.48976047e-03 -1.23913121e+00
-8.43577147e-01 -2.89206088e-01 1.17265217e-01 -3.32157165e-02
-3.04225117e-01 6.39990494e-02 -9.99055684e-01 -4.37168956e-01
2.98581988e-01 -8.11111093e-01 -5.24903059e-01 -6.11028433e-01
-7.42117882e-01 -2.90097535e-01 3.98226023e-01 -5.37985981e-01
1.88380980e+00 -2.15771127e+00 -4.20173481e-02 5.69320142e-01
1.63386673e-01 3.53962004e-01 -1.08311795e-01 5.98401010e-01
-6.55553937e-02 5.64684093e-01 -1.23557642e-01 -3.71010840e-01
4.19010082e-03 2.74345458e-01 -2.46198401e-01 3.37048888e-01
1.52064353e-01 8.34538937e-01 -7.53761768e-01 -5.57088912e-01
-2.26526648e-01 4.34088796e-01 -4.63755190e-01 -1.28370851e-01
9.99866202e-02 -1.58631265e-01 -6.92541420e-01 7.82160163e-01
5.22028506e-01 -6.42455935e-01 7.34021306e-01 -7.47046713e-03
-2.86634207e-01 8.09882998e-01 -1.36507607e+00 1.15437663e+00
-2.17702359e-01 8.53417143e-02 -8.69533494e-02 -5.46047509e-01
8.87531579e-01 2.87135839e-01 2.40311816e-01 -4.95755672e-01
-2.63978750e-01 5.93694076e-02 -1.71920687e-01 -4.99560125e-02
6.15638852e-01 2.14232013e-01 -1.90367311e-01 1.32118955e-01
-1.98452510e-02 1.00988972e+00 2.37888843e-01 3.07032138e-01
1.22791350e+00 -6.57988787e-02 6.69884205e-01 9.03012380e-02
6.67744100e-01 -1.23351410e-01 1.16413379e+00 1.01726687e+00
-3.76178324e-01 1.92500725e-01 4.45137143e-01 -3.35795641e-01
-7.57880628e-01 -9.71119761e-01 -7.14731961e-02 1.12861395e+00
1.81814343e-01 -7.21499562e-01 -2.20574841e-01 -1.15131640e+00
2.95806751e-02 7.96094656e-01 -5.48740566e-01 1.17459707e-01
-8.63112390e-01 -7.41618812e-01 1.10530388e+00 8.70194674e-01
7.08999395e-01 -9.15921330e-01 -6.09468333e-02 3.37817580e-01
-2.45395318e-01 -1.24271107e+00 -4.23450410e-01 2.99640268e-01
-8.34707677e-01 -8.70544791e-01 -3.35534841e-01 -1.04808867e+00
5.38125932e-01 9.05577932e-03 8.52493107e-01 3.72242555e-02
3.12738687e-01 -6.10208586e-02 -3.86207283e-01 5.15078939e-02
-1.88016921e-01 6.48246527e-01 -2.12926731e-01 -2.35234778e-02
3.45500320e-01 -1.66278094e-01 -2.27346182e-01 3.92508417e-01
-5.57348788e-01 -1.43300861e-01 8.79501641e-01 8.89691710e-01
4.68715847e-01 -8.74365270e-02 4.34716165e-01 -1.33537591e+00
2.15987831e-01 -7.67123342e-01 -6.30454481e-01 6.79894269e-01
-7.87091613e-01 3.70671183e-01 3.33059400e-01 -5.05201183e-02
-1.50025749e+00 3.18111479e-01 -2.47104898e-01 1.90675594e-02
-2.54947662e-01 9.62241530e-01 -4.85490859e-01 2.21998408e-01
2.17122734e-01 3.40796679e-01 -6.91757560e-01 -6.46141529e-01
3.11911345e-01 5.41208923e-01 3.97283286e-01 -5.23252487e-01
7.03321338e-01 4.72120196e-01 -4.34523970e-01 -5.76532900e-01
-7.28168368e-01 -7.68618464e-01 -6.28752410e-01 1.83450907e-01
5.51080227e-01 -9.37130928e-01 -7.87541509e-01 4.24876601e-01
-1.12764680e+00 -1.28313363e-01 1.07925624e-01 4.00668472e-01
2.31338143e-01 5.90473115e-01 -9.58383501e-01 -9.34293032e-01
-2.67697275e-01 -6.50166631e-01 8.61348748e-01 2.64576465e-01
6.66607693e-02 -9.70542490e-01 3.87571812e-01 8.64864066e-02
-6.69199461e-03 -1.61496729e-01 8.18575144e-01 -1.24957752e+00
-5.64551830e-01 -3.30646366e-01 -1.20247088e-01 -2.42135853e-01
-6.29554922e-03 5.49009703e-02 -7.12292612e-01 -1.33607671e-01
-5.54985642e-01 4.40304242e-02 1.11653626e+00 -1.97790340e-01
6.84090912e-01 -4.29712772e-01 -9.96120155e-01 5.28966546e-01
1.19556451e+00 5.73332071e-01 7.48573303e-01 5.40933669e-01
6.24814212e-01 2.61373848e-01 4.97048229e-01 3.60679626e-01
7.65237033e-01 4.50143397e-01 -4.76111807e-02 -4.97427210e-02
2.78496534e-01 -4.53990161e-01 5.14535069e-01 8.88802052e-01
1.91644073e-01 -5.74453533e-01 -1.25954795e+00 8.51171017e-01
-1.85686326e+00 -1.15857720e+00 1.06260786e-02 2.16241050e+00
7.28912890e-01 2.41411462e-01 -1.12029277e-01 -2.40979627e-01
9.47518408e-01 2.82473236e-01 -4.16846514e-01 -2.08270937e-01
-3.10226887e-01 2.30877638e-01 7.05057204e-01 5.48828065e-01
-1.34780359e+00 1.31739843e+00 7.34736490e+00 4.54071343e-01
-8.21378112e-01 -1.40672147e-01 3.51703793e-01 2.43059307e-01
-2.95222104e-01 2.31072009e-01 -1.52356565e+00 2.75081635e-01
1.19639242e+00 -2.47481927e-01 -1.30094618e-01 8.40151191e-01
-2.07824141e-01 1.42556146e-01 -9.74760115e-01 4.13482368e-01
-1.42236024e-01 -1.32390833e+00 1.66785270e-01 6.57368079e-03
5.95394850e-01 3.10891211e-01 -2.85268754e-01 5.75052500e-01
8.04437876e-01 -6.55932367e-01 5.32222033e-01 3.57371002e-01
1.82677791e-01 -8.37450504e-01 5.93947470e-01 3.11808527e-01
-1.60691381e+00 -1.48436755e-01 -2.88738478e-02 1.86835557e-01
4.36551988e-01 3.81043553e-01 -9.91592646e-01 9.72834527e-01
4.44674104e-01 4.33734328e-01 -5.16870320e-01 1.23726320e+00
-3.61633092e-01 1.01783276e+00 -5.60811818e-01 -1.55293047e-01
3.56713891e-01 2.90462196e-01 5.93141615e-01 1.70751202e+00
-4.48913127e-02 4.21643108e-01 4.03775781e-01 3.52343053e-01
-3.35879028e-01 1.79863796e-02 -4.21029210e-01 -1.88503549e-01
9.33723330e-01 8.73985112e-01 -6.06210530e-01 -5.33140838e-01
-6.76369309e-01 8.37632775e-01 7.12499380e-01 4.31338549e-01
-8.32032084e-01 -5.19438267e-01 4.66584176e-01 -3.51192355e-01
8.33115697e-01 -3.92010123e-01 5.41682169e-02 -1.47663033e+00
3.16488855e-02 -6.78810716e-01 9.68911588e-01 -2.22775385e-01
-9.85752285e-01 8.31151783e-01 -2.31088027e-02 -7.21612751e-01
-1.38717204e-01 -4.47597563e-01 -4.06693995e-01 6.91086411e-01
-1.41014600e+00 -1.07712841e+00 2.29498446e-01 3.55115980e-01
2.54317760e-01 1.99579254e-01 9.51681435e-01 4.49469388e-01
-1.05405116e+00 8.41106474e-01 5.29398359e-02 8.25374782e-01
3.97780061e-01 -9.79146898e-01 9.96610343e-01 1.28328180e+00
3.82061809e-01 1.16087103e+00 1.92299634e-01 -1.01638424e+00
-1.21511889e+00 -1.14047217e+00 1.41601241e+00 -2.96103030e-01
6.69574738e-01 -2.26685733e-01 -9.84189332e-01 1.30797791e+00
-1.82106402e-02 -1.80851698e-01 9.23937738e-01 6.96810186e-01
-5.68257213e-01 2.25352630e-01 -8.23899150e-01 6.15416527e-01
1.19067669e+00 -5.14848471e-01 -8.25203180e-01 -1.61507174e-01
8.56134117e-01 -4.23189014e-01 -9.08887982e-01 7.51757562e-01
6.38186336e-01 -6.12432659e-01 7.95524478e-01 -9.90679622e-01
-2.65750766e-01 -3.38330179e-01 -1.22992612e-01 -8.46515417e-01
-6.03697598e-01 -5.95795155e-01 -5.98382652e-01 1.29101443e+00
1.26832187e+00 -8.14306021e-01 9.81982052e-01 8.18560362e-01
-7.37394951e-03 -6.95991397e-01 -7.27255046e-01 -9.56978381e-01
-2.30080321e-01 -4.56255198e-01 6.86951995e-01 1.04954803e+00
8.74662027e-02 1.78647906e-01 -3.14857453e-01 9.43823278e-01
4.22844321e-01 2.77877748e-01 5.04378617e-01 -8.54039550e-01
-3.45765024e-01 -1.28397778e-01 -4.55918372e-01 -1.53576612e+00
3.62838924e-01 -8.45866561e-01 7.34183267e-02 -1.48116100e+00
4.05985743e-01 -6.03599846e-01 -9.70303476e-01 8.81410837e-01
-3.98541898e-01 -3.07281673e-01 4.37399223e-02 8.64271671e-02
-9.52097654e-01 3.46493781e-01 4.99636620e-01 -2.43678633e-02
-3.52727383e-01 -2.68843751e-02 -6.55120254e-01 5.32342434e-01
6.30199373e-01 -6.08679831e-01 -9.83279571e-02 -4.82464224e-01
2.54714757e-01 1.38848573e-01 -1.10484116e-01 -7.73165107e-01
8.42350066e-01 -6.02250956e-02 3.29462588e-01 -9.91675198e-01
2.43494689e-01 -4.60345984e-01 1.79120570e-01 4.00559932e-01
-4.16057825e-01 1.98405702e-02 2.58542299e-01 8.20307970e-01
-3.13842148e-01 3.50392722e-02 3.54407877e-01 1.42772898e-01
-1.20966780e+00 3.01937342e-01 -2.60378748e-01 1.60543304e-02
8.38021994e-01 -1.80916879e-02 -2.65103847e-01 -1.14145778e-01
-1.13036203e+00 6.88835263e-01 1.17341481e-01 5.12425661e-01
5.80500603e-01 -1.07683527e+00 -5.27903438e-01 2.22000293e-02
2.24746764e-01 -3.19414556e-01 4.72406857e-02 6.04842901e-01
-9.24670771e-02 6.51793838e-01 3.51307094e-01 -2.55391777e-01
-1.34196448e+00 5.66510081e-01 3.03330541e-01 -7.09057033e-01
-4.51203585e-01 8.07869732e-01 7.72714466e-02 -6.77432418e-01
3.41623396e-01 -2.06242695e-01 -3.44601870e-01 -1.04887381e-01
5.55479288e-01 4.99737233e-01 9.25487876e-02 -6.36223614e-01
-7.40015805e-01 -1.68833341e-02 -6.09139442e-01 -2.75229216e-01
1.18959355e+00 -5.51884100e-02 3.05671543e-02 4.21253920e-01
9.88349676e-01 4.23177689e-01 -6.41060412e-01 -3.99629772e-01
6.77139401e-01 -4.52674180e-02 -1.06907122e-01 -7.98084199e-01
-9.27304268e-01 1.95069924e-01 5.57615496e-02 3.30571570e-02
7.95111656e-01 -4.13762778e-02 8.14728141e-01 9.34008539e-01
5.81346750e-01 -6.93627238e-01 -7.03350008e-01 8.77568305e-01
1.38588995e-01 -8.29355419e-01 -2.35816553e-01 -7.25345671e-01
-5.47914684e-01 6.46783531e-01 5.96399546e-01 2.61515588e-01
8.12220454e-01 5.26990831e-01 -7.10005760e-02 -5.09799235e-02
-1.09573090e+00 -4.03101981e-01 1.89056173e-01 1.22085102e-01
3.77101809e-01 7.43297487e-02 -3.81498903e-01 8.13492835e-01
3.40414762e-01 -1.45750985e-01 1.65113464e-01 1.36858106e+00
-3.91412675e-01 -1.33541071e+00 -3.77202258e-02 1.90414160e-01
-7.19581544e-01 -3.69357556e-01 -5.78021944e-01 9.86409724e-01
-2.03247353e-01 8.85401011e-01 -8.22580531e-02 -4.61083144e-01
5.48439443e-01 4.96944129e-01 2.79814121e-04 -7.10447192e-01
-7.04049170e-01 -1.12011693e-01 6.27935290e-01 -6.18701518e-01
-1.85477898e-01 -7.03741610e-01 -1.73295414e+00 -2.00037122e-01
-7.12245405e-01 6.13123119e-01 2.47015119e-01 7.27597713e-01
9.58869040e-01 1.88798591e-01 6.27489448e-01 3.57294939e-02
-3.17011476e-01 -6.75844610e-01 -2.52224445e-01 1.46274507e-01
1.41944006e-01 -5.34845293e-01 -1.70532092e-01 -1.72619790e-01] | [9.593649864196777, 9.492574691772461] |
dc8d8d0d-7304-49e9-bb55-5d27001dea76 | mixformer-end-to-end-tracking-with-iterative-2 | 2302.02814 | null | https://arxiv.org/abs/2302.02814v2 | https://arxiv.org/pdf/2302.02814v2.pdf | MixFormer: End-to-End Tracking with Iterative Mixed Attention | Visual object tracking often employs a multi-stage pipeline of feature extraction, target information integration, and bounding box estimation. To simplify this pipeline and unify the process of feature extraction and target information integration, in this paper, we present a compact tracking framework, termed as MixFormer, built upon transformers. Our core design is to utilize the flexibility of attention operations, and propose a Mixed Attention Module (MAM) for simultaneous feature extraction and target information integration. This synchronous modeling scheme allows to extract target-specific discriminative features and perform extensive communication between target and search area. Based on MAM, we build our MixFormer trackers simply by stacking multiple MAMs and placing a localization head on top. Specifically, we instantiate two types of MixFormer trackers, a hierarchical tracker MixCvT, and a non-hierarchical tracker MixViT. For these two trackers, we investigate a series of pre-training methods and uncover the different behaviors between supervised pre-training and self-supervised pre-training in our MixFormer trackers. We also extend the masked pre-training to our MixFormer trackers and design the competitive TrackMAE pre-training technique. Finally, to handle multiple target templates during online tracking, we devise an asymmetric attention scheme in MAM to reduce computational cost, and propose an effective score prediction module to select high-quality templates. Our MixFormer trackers set a new state-of-the-art performance on seven tracking benchmarks, including LaSOT, TrackingNet, VOT2020, GOT-10k, OTB100 and UAV123. In particular, our MixViT-L achieves AUC score of 73.3% on LaSOT, 86.1% on TrackingNet, EAO of 0.584 on VOT2020, and AO of 75.7% on GOT-10k. Code and trained models are publicly available at https://github.com/MCG-NJU/MixFormer. | ['LiMin Wang', 'Gangshan Wu', 'Cheng Jiang', 'Yutao Cui'] | 2023-02-06 | mixformer-end-to-end-tracking-with-iterative | https://arxiv.org/abs/2302.02814 | https://arxiv.org/pdf/2302.02814.pdf | null | ['visual-object-tracking'] | ['computer-vision'] | [-3.25513542e-01 -3.41627061e-01 -3.34973931e-01 -1.17653444e-01
-8.11585963e-01 -8.58851790e-01 6.47693038e-01 -3.05252045e-01
-3.67795765e-01 2.99310118e-01 -1.14116184e-01 -1.11417308e-01
1.88170031e-01 -4.36904281e-01 -7.46819198e-01 -5.88727474e-01
-8.98680985e-02 3.18067640e-01 6.85887516e-01 1.55188829e-01
-2.61050791e-01 6.87979341e-01 -1.50133848e+00 -4.93929312e-02
6.64242208e-01 1.39623177e+00 1.87456891e-01 7.35930681e-01
1.52027547e-01 4.96612430e-01 -7.17360437e-01 -5.36453724e-01
5.28984725e-01 1.64423287e-02 -1.39284134e-01 -9.20382887e-02
7.84564137e-01 -3.75523120e-01 -5.22158206e-01 1.01747847e+00
5.68666279e-01 -2.69847870e-01 5.25787294e-01 -1.51393580e+00
-2.97789156e-01 5.21568954e-01 -6.81927264e-01 4.13450450e-01
-4.81104031e-02 6.42923117e-01 8.67074251e-01 -9.84592259e-01
4.10245866e-01 1.07838356e+00 1.12380791e+00 4.56393093e-01
-1.00565171e+00 -1.17130065e+00 4.15659606e-01 -2.67865863e-02
-1.45056713e+00 -5.14108479e-01 2.75174797e-01 -8.08300138e-01
5.88550448e-01 2.51241356e-01 7.28400052e-01 1.00145364e+00
3.74635100e-01 8.70360792e-01 7.56364524e-01 1.23502612e-01
-1.04505152e-01 -5.04898466e-02 3.06288838e-01 6.31873250e-01
4.98303294e-01 6.42250597e-01 -4.35406893e-01 -9.86155272e-02
6.19737148e-01 1.54478520e-01 -2.19093382e-01 -3.77966732e-01
-1.30459785e+00 6.48065388e-01 6.87891245e-01 4.40734588e-02
-1.94650352e-01 4.49787676e-01 2.92220294e-01 6.25415221e-02
2.07566589e-01 4.98944931e-02 -3.33145320e-01 1.49827495e-01
-9.84766066e-01 1.93863809e-01 4.32739139e-01 1.19676673e+00
4.98608321e-01 2.53511906e-01 -6.84631526e-01 3.25387806e-01
6.94642067e-01 1.06357932e+00 2.53028929e-01 -4.50135589e-01
2.68180192e-01 5.50009787e-01 3.37730438e-01 -6.15062177e-01
-4.71105933e-01 -1.01712203e+00 -4.23551470e-01 2.24314436e-01
2.38886312e-01 -2.55459011e-01 -1.18714714e+00 1.82273734e+00
7.09274769e-01 3.43094379e-01 -1.79947734e-01 1.05778289e+00
9.56881225e-01 4.08641040e-01 2.17642084e-01 6.29341137e-03
1.47755277e+00 -1.22684014e+00 -5.71152866e-01 -2.75792688e-01
4.39026177e-01 -8.38242173e-01 5.06482482e-01 -6.23581894e-02
-6.28660142e-01 -9.89632308e-01 -1.04757190e+00 2.24750504e-01
-3.92954767e-01 7.36147463e-01 5.63620865e-01 5.91719389e-01
-8.37587297e-01 3.11025739e-01 -9.57971275e-01 -4.12954122e-01
4.39492911e-01 5.15287936e-01 -1.14859648e-01 3.55196655e-01
-9.07787800e-01 7.61458993e-01 3.30682695e-01 2.33163312e-01
-1.30845690e+00 -8.02690029e-01 -6.36169076e-01 -4.00745533e-02
5.69282174e-01 -6.14636242e-01 1.41481197e+00 -5.01303077e-01
-1.35776722e+00 6.00506723e-01 2.21327618e-02 -5.89480460e-01
4.47908551e-01 -4.94050235e-01 -5.08799016e-01 -3.02399993e-01
2.66009808e-01 8.00207913e-01 9.70341802e-01 -9.51446354e-01
-9.64924872e-01 -9.77890119e-02 -9.92783755e-02 -1.14440940e-01
-6.50426373e-02 2.02653751e-01 -8.80696654e-01 -7.01882660e-01
-2.17979953e-01 -9.17445838e-01 -3.50236632e-02 2.35450417e-01
-3.84792596e-01 -1.28259540e-01 1.05810690e+00 -3.75128180e-01
1.32268345e+00 -2.14871240e+00 -8.42062235e-02 -3.56110893e-02
4.49012607e-01 6.42447174e-01 -2.21319094e-01 1.23491831e-01
1.63953677e-01 -4.51712012e-01 2.22573459e-01 -5.07280827e-01
2.53702551e-01 -9.14828107e-02 -3.75545532e-01 5.85569978e-01
1.92516744e-01 1.21182811e+00 -7.66700387e-01 -5.28743029e-01
3.31058085e-01 3.48869264e-01 -4.26341027e-01 3.81610036e-01
-3.42033595e-01 2.99105495e-01 -5.54026306e-01 1.19693637e+00
7.39681661e-01 -3.62985611e-01 -4.97649945e-02 -5.59797943e-01
-5.57332695e-01 1.96121559e-02 -1.10089338e+00 1.33151448e+00
-3.77807096e-02 6.35371506e-01 1.93205982e-01 -2.12059140e-01
8.15943599e-01 2.80336849e-03 6.92475736e-01 -3.76401573e-01
5.31680286e-01 2.07369067e-02 1.57947883e-01 -1.37516018e-02
5.41240275e-01 3.69844079e-01 -1.28121942e-01 1.58447355e-01
3.87370616e-01 5.32250881e-01 2.21435562e-01 1.50719360e-01
1.20917332e+00 3.33488762e-01 1.82275429e-01 -1.37309015e-01
4.51084107e-01 1.67343646e-01 8.65658045e-01 8.49288046e-01
-4.86873418e-01 2.66300142e-01 6.01138212e-02 -3.49772513e-01
-5.42987108e-01 -1.09455717e+00 -1.21911213e-01 1.23265815e+00
3.43729168e-01 -6.90774024e-01 -5.03176212e-01 -8.97972107e-01
3.84402424e-01 1.62660778e-01 -5.88553131e-01 -1.77138895e-01
-4.45276886e-01 -5.83723605e-01 7.80180573e-01 7.45093703e-01
4.90392685e-01 -8.88994336e-01 -8.06154430e-01 2.92882234e-01
1.48416951e-01 -1.17275834e+00 -9.78798389e-01 1.31505668e-01
-4.20027524e-01 -1.16834223e+00 -4.94449437e-01 -3.34311187e-01
4.04787332e-01 4.16956693e-01 8.55160177e-01 -3.05845886e-02
-2.17453241e-01 4.84518021e-01 -3.02396446e-01 -6.16689026e-01
1.69401705e-01 1.91770196e-01 2.92218119e-01 2.07313791e-01
1.99006826e-01 -4.03908342e-02 -4.55133200e-01 7.53403604e-01
-4.16672826e-01 -7.91080743e-02 8.83869410e-01 7.11365581e-01
6.84757113e-01 -6.40823245e-01 1.40869021e-01 -1.39134124e-01
-4.66620130e-03 -4.63250965e-01 -1.37545848e+00 3.35565805e-01
-3.00687224e-01 -2.32485756e-01 4.12317902e-01 -8.40104759e-01
-6.75512373e-01 5.25271177e-01 -3.86859998e-02 -1.05410802e+00
3.00600588e-01 1.03960633e-01 -4.17784631e-01 -6.01305723e-01
4.19269800e-01 1.11859567e-01 -1.45971835e-01 -6.17611229e-01
2.71550715e-01 4.83958542e-01 6.83457315e-01 -2.91287780e-01
1.37158787e+00 3.39858592e-01 -3.25182229e-01 -4.03925687e-01
-9.76359665e-01 -5.03052652e-01 -4.45006430e-01 -5.48660457e-01
8.95079494e-01 -1.07246518e+00 -1.01085627e+00 5.86697459e-01
-1.01888585e+00 -4.06516314e-01 -1.93865806e-01 7.04224885e-01
-6.06993139e-02 1.01000607e-01 -2.87199795e-01 -7.89812982e-01
-5.59609354e-01 -1.25362337e+00 1.37276101e+00 5.42814434e-01
1.98823795e-01 -5.47056079e-01 2.05049574e-01 6.79323077e-02
5.40980399e-01 1.79662973e-01 -3.01354975e-01 -7.81172752e-01
-1.07051229e+00 -3.06059986e-01 -4.93246734e-01 1.67367291e-02
-5.86243346e-02 -5.11216745e-02 -1.01595056e+00 -6.92095697e-01
-4.18385416e-01 -1.55971006e-01 1.01314151e+00 4.85246480e-01
7.23015070e-01 -8.28343537e-03 -9.33172047e-01 1.01811194e+00
1.12829816e+00 3.27269882e-01 1.56630918e-01 2.54189879e-01
8.17795455e-01 -5.36728203e-02 9.96103942e-01 2.23530710e-01
4.82642561e-01 1.16675317e+00 5.77446342e-01 2.78257411e-02
-3.48726630e-01 -2.33924344e-01 8.36512327e-01 5.16888559e-01
2.16278657e-01 -2.49695748e-01 -7.86058307e-01 4.35591638e-01
-1.88529062e+00 -9.21124339e-01 -1.18912488e-01 2.19873190e+00
3.86857122e-01 2.37193361e-01 3.78758103e-01 -5.42704582e-01
7.82042980e-01 1.95927233e-01 -6.38209462e-01 3.95360053e-01
1.66269541e-01 -2.97034183e-03 9.26755548e-01 4.39034969e-01
-1.60224235e+00 1.27165341e+00 5.65720510e+00 9.47115839e-01
-1.25988925e+00 3.03619146e-01 -9.66667980e-02 -2.48247594e-01
3.37068468e-01 -1.21989036e-02 -1.66364706e+00 8.15898657e-01
8.37654054e-01 -1.59234498e-02 -2.99923681e-03 1.06662881e+00
-1.94713160e-01 1.77555054e-01 -9.66712356e-01 9.42587137e-01
-2.28007257e-01 -1.23017097e+00 -3.24970573e-01 1.25790209e-01
2.53504068e-01 5.68629324e-01 -9.64680538e-02 7.47225702e-01
4.56774086e-01 -7.30886102e-01 9.65327024e-01 3.89212012e-01
5.59009552e-01 -2.49006093e-01 7.08329737e-01 3.81384417e-02
-2.09890199e+00 -8.12382400e-02 -3.20345521e-01 4.05684382e-01
1.87612370e-01 3.06020200e-01 -5.36127031e-01 9.18036282e-01
8.93580675e-01 7.47945666e-01 -8.26225042e-01 1.65313923e+00
-1.48061782e-01 4.35149878e-01 -6.71309292e-01 4.87211198e-02
1.23189032e-01 2.78698087e-01 7.89416373e-01 1.21664274e+00
2.00391158e-01 -1.31243438e-01 6.13675714e-01 7.01019228e-01
1.20581232e-01 -3.26180190e-01 -1.78819925e-01 8.65801722e-02
7.67498910e-01 1.69932032e+00 -6.01973295e-01 -3.80661428e-01
-4.05146182e-01 3.65831286e-01 2.26770669e-01 1.36969388e-01
-1.74849772e+00 -1.71923250e-01 9.02141511e-01 -3.85472104e-02
8.35480571e-01 -1.20427944e-01 1.84941113e-01 -1.25480556e+00
-1.10846777e-02 -7.94346392e-01 4.76479918e-01 -5.11726737e-01
-1.06789362e+00 9.61544752e-01 9.49877501e-02 -1.67501795e+00
9.66427475e-02 -5.96178532e-01 -7.94951677e-01 7.11491942e-01
-1.24459982e+00 -1.57919478e+00 -6.34381831e-01 4.68352258e-01
3.58180523e-01 -2.29874372e-01 1.74720898e-01 7.12044537e-01
-9.10056055e-01 8.98874342e-01 -2.73761183e-01 3.22325915e-01
7.51039326e-01 -8.70240390e-01 5.75919986e-01 1.11800337e+00
6.79208860e-02 5.63348234e-01 4.06264871e-01 -9.14817452e-01
-1.54297745e+00 -1.53246045e+00 4.78841037e-01 -6.03240907e-01
8.84151399e-01 -4.80863631e-01 -5.64221680e-01 9.77794170e-01
2.60430947e-02 3.94518584e-01 2.98785508e-01 7.13535445e-03
-3.80864024e-01 -3.40712368e-01 -8.24686944e-01 3.92101288e-01
1.20028532e+00 -8.30383897e-02 -5.53598642e-01 1.86279222e-01
6.95641756e-01 -8.12534451e-01 -9.67544138e-01 6.70647681e-01
7.41746902e-01 -6.78124607e-01 1.01960003e+00 -1.96253866e-01
-5.97890675e-01 -1.03240669e+00 -9.49661806e-02 -9.00866568e-01
-6.04977250e-01 -7.17305243e-01 -5.48271537e-01 1.43225503e+00
2.26224914e-01 -7.24948406e-01 7.28328168e-01 3.89105752e-02
-5.14057040e-01 -7.31538177e-01 -9.53054070e-01 -1.14842451e+00
-5.33091784e-01 -3.15601230e-01 5.05411983e-01 5.91533244e-01
-6.65879488e-01 1.67343229e-01 -5.66463411e-01 4.79965448e-01
1.01363635e+00 2.19790578e-01 1.19026518e+00 -1.21498740e+00
-3.72243136e-01 -3.54343414e-01 -3.60948771e-01 -1.33803892e+00
-1.55475736e-01 -7.88613677e-01 1.30325750e-01 -1.18702841e+00
6.40897229e-02 -4.20984775e-01 -2.95878708e-01 8.77846718e-01
-1.07434668e-01 2.65305072e-01 5.79210520e-01 4.47441310e-01
-9.44190741e-01 7.03233719e-01 9.61648226e-01 -1.90474480e-01
-1.38909787e-01 1.36814848e-01 -4.62593585e-01 5.66422641e-01
6.53066039e-01 -7.46584833e-01 3.65480334e-02 -3.06375265e-01
-5.35458863e-01 -2.53271461e-01 6.76652193e-01 -1.47484827e+00
5.32989860e-01 6.25789016e-02 6.42968953e-01 -1.20016384e+00
3.51186723e-01 -8.35224152e-01 4.86119986e-01 7.60778010e-01
2.64415294e-01 9.27568078e-02 5.12837946e-01 4.08760935e-01
-5.26024252e-02 1.72879070e-01 7.77493775e-01 3.47134978e-01
-8.46066117e-01 7.01591730e-01 -1.64315909e-01 -9.05834958e-02
1.22258639e+00 -2.26646230e-01 -5.34959793e-01 7.06507266e-02
-5.65093756e-01 7.83800364e-01 3.13170195e-01 6.77461326e-01
2.30650619e-01 -1.63769424e+00 -3.83816719e-01 1.56931877e-01
2.00259238e-01 -3.04592669e-01 9.98830274e-02 1.26696086e+00
-1.40457481e-01 5.75393200e-01 -2.17817396e-01 -9.75816607e-01
-1.37413776e+00 5.46863198e-01 5.14122009e-01 -4.21729803e-01
-5.15636027e-01 8.68462145e-01 4.28407311e-01 -2.49300852e-01
5.50422072e-01 -5.55695355e-01 7.48267099e-02 6.50090352e-02
6.66368067e-01 1.72306791e-01 -2.80789226e-01 -7.76564360e-01
-8.03434193e-01 8.07393074e-01 -1.73536703e-01 2.71200150e-01
8.21920753e-01 2.11642459e-01 3.46167505e-01 -1.43735588e-01
5.98814785e-01 8.17974806e-02 -1.71436834e+00 -1.22792371e-01
7.40194768e-02 -3.50643665e-01 1.25434371e-02 -6.96052969e-01
-1.37914348e+00 4.38287050e-01 9.62284446e-01 4.42402959e-02
1.03872693e+00 7.35272095e-02 6.46077037e-01 1.64918795e-01
3.84984821e-01 -6.68802857e-01 -8.37639794e-02 6.74192250e-01
7.33803153e-01 -1.11806417e+00 2.10191365e-02 -2.88439304e-01
-3.83973569e-01 7.02560544e-01 1.22283280e+00 -9.41732004e-02
5.62306523e-01 6.32215977e-01 8.74274746e-02 -2.73490101e-01
-7.72368014e-01 -7.81803191e-01 7.47154295e-01 4.52226520e-01
1.26627877e-01 -1.69235900e-01 -3.93214375e-02 7.61449397e-01
3.84568945e-02 -6.62597865e-02 -1.43709987e-01 9.19754505e-01
-5.29514670e-01 -9.63170946e-01 -7.03331292e-01 2.72329122e-01
-3.55706394e-01 5.96715920e-02 -5.51754475e-01 1.10668480e+00
3.64791125e-01 6.37342095e-01 -3.53656262e-02 -8.91032517e-01
4.47984993e-01 -2.73871750e-01 3.39612812e-01 -3.60366315e-01
-1.04639912e+00 4.06380951e-01 7.30897561e-02 -8.67867947e-01
-3.47297609e-01 -5.81937611e-01 -9.43632126e-01 -2.18154952e-01
-6.75566018e-01 9.09451470e-02 4.14111495e-01 7.89068282e-01
5.68952501e-01 7.88375199e-01 3.43132466e-01 -1.15755486e+00
-4.76044357e-01 -9.98210907e-01 -2.67432295e-02 -2.06493631e-01
3.96352947e-01 -1.29951096e+00 -7.21679851e-02 -2.88182348e-01] | [6.2905778884887695, -2.122229814529419] |
9c359358-f448-44f2-92c6-fc3ec8655d2a | ambiguity-in-solving-imaging-inverse-problems | 2305.19774 | null | https://arxiv.org/abs/2305.19774v1 | https://arxiv.org/pdf/2305.19774v1.pdf | Ambiguity in solving imaging inverse problems with deep learning based operators | In recent years, large convolutional neural networks have been widely used as tools for image deblurring, because of their ability in restoring images very precisely. It is well known that image deblurring is mathematically modeled as an ill-posed inverse problem and its solution is difficult to approximate when noise affects the data. Really, one limitation of neural networks for deblurring is their sensitivity to noise and other perturbations, which can lead to instability and produce poor reconstructions. In addition, networks do not necessarily take into account the numerical formulation of the underlying imaging problem, when trained end-to-end. In this paper, we propose some strategies to improve stability without losing to much accuracy to deblur images with deep-learning based methods. First, we suggest a very small neural architecture, which reduces the execution time for training, satisfying a green AI need, and does not extremely amplify noise in the computed image. Second, we introduce a unified framework where a pre-processing step balances the lack of stability of the following, neural network-based, step. Two different pre-processors are presented: the former implements a strong parameter-free denoiser, and the latter is a variational model-based regularized formulation of the latent imaging problem. This framework is also formally characterized by mathematical analysis. Numerical experiments are performed to verify the accuracy and stability of the proposed approaches for image deblurring when unknown or not-quantified noise is present; the results confirm that they improve the network stability with respect to noise. In particular, the model-based framework represents the most reliable trade-off between visual precision and robustness. | ['James Nagy', 'Elena Loli Piccolomini', 'Elena Morotti', 'Davide Evangelista'] | 2023-05-31 | null | null | null | null | ['deblurring'] | ['computer-vision'] | [ 1.91706911e-01 -3.29618931e-01 2.44729251e-01 1.03946090e-01
-2.60720700e-01 -1.60757884e-01 3.60348463e-01 -2.91359276e-01
-5.51597893e-01 7.91243792e-01 6.52743205e-02 -9.00582224e-02
-3.76792431e-01 -7.12942004e-01 -7.06576586e-01 -1.12190187e+00
2.06701830e-01 1.18976288e-01 6.69300258e-02 -1.98058710e-01
4.39927578e-02 6.65691078e-01 -1.61353278e+00 -2.22974494e-01
1.11260033e+00 1.03115475e+00 2.68383384e-01 5.11943161e-01
1.85518473e-01 8.41809332e-01 -4.57761079e-01 -3.88513990e-02
1.18442476e-01 -5.78216016e-01 -6.22813702e-01 6.20602407e-02
2.70284235e-01 -6.12549961e-01 -3.53114337e-01 1.35642946e+00
5.57188809e-01 9.55578759e-02 6.93256021e-01 -7.26043522e-01
-6.40638769e-01 2.77739465e-01 -5.42770922e-01 2.47102484e-01
4.00129035e-02 1.19481832e-01 2.83039570e-01 -7.77563155e-01
5.55439532e-01 8.93440902e-01 9.99745727e-01 3.22151124e-01
-1.17837167e+00 -3.54326993e-01 -4.06669348e-01 4.12825316e-01
-1.45034730e+00 -5.04260242e-01 9.87347007e-01 -4.54625428e-01
4.62959945e-01 2.75061667e-01 6.39635801e-01 8.80592704e-01
4.02761221e-01 2.57734090e-01 1.25842690e+00 -4.74227339e-01
3.06378484e-01 -3.40227894e-02 1.30938485e-01 3.58967751e-01
5.00600696e-01 2.47078314e-01 -5.19175194e-02 2.16453280e-02
9.37782586e-01 -1.50901392e-01 -8.27766538e-01 -3.03397804e-01
-9.24356759e-01 5.96117496e-01 5.14219105e-01 9.01674569e-01
-6.56201959e-01 2.30875716e-01 3.21680218e-01 2.52035856e-01
5.35549283e-01 2.93828875e-01 2.53710784e-02 4.56249788e-02
-1.43290174e+00 1.51921734e-01 6.06274307e-01 1.60055026e-01
6.46449506e-01 4.35328096e-01 1.65188834e-01 7.17118502e-01
1.90398052e-01 5.81990361e-01 7.81821668e-01 -7.74891257e-01
-3.01159769e-02 3.32097560e-02 2.69109100e-01 -1.50049996e+00
-3.83976251e-01 -8.41941297e-01 -1.60327053e+00 5.89559138e-01
4.64345425e-01 -1.30981877e-02 -6.90377653e-01 1.72381103e+00
2.86297262e-01 2.82405794e-01 1.00047648e-01 1.31606698e+00
6.42365515e-01 7.99717903e-01 -2.19407752e-01 -6.16146922e-01
1.15571940e+00 -6.34610355e-01 -1.17114615e+00 -3.34020816e-02
1.85442507e-01 -8.18536699e-01 6.94619298e-01 5.56421995e-01
-1.18044794e+00 -6.83088005e-01 -1.18413270e+00 9.41330492e-02
-8.15032944e-02 2.51413763e-01 8.85478258e-02 5.71558058e-01
-1.35438037e+00 9.00740266e-01 -7.51643300e-01 -2.02979252e-01
5.02632894e-02 1.63577363e-01 -3.03422332e-01 1.15131840e-01
-1.39473200e+00 1.20479131e+00 3.35427254e-01 7.52769470e-01
-7.35315561e-01 -4.08447355e-01 -5.08769274e-01 2.75690347e-01
4.97663133e-02 -7.27469444e-01 8.28796029e-01 -1.44089699e+00
-1.63936675e+00 6.50376379e-01 -5.52087054e-02 -4.05648500e-01
9.73306954e-01 -2.20703125e-01 -4.52137172e-01 2.38753736e-01
-2.74622381e-01 3.82406376e-02 1.29127872e+00 -1.48304462e+00
1.42011926e-01 -1.06210008e-01 -2.60230482e-01 4.58082557e-02
-3.67238909e-01 -2.24744335e-01 -1.64278746e-01 -8.37180138e-01
3.55639428e-01 -6.06587648e-01 -1.93991095e-01 1.04772836e-01
-1.92349046e-01 3.77357066e-01 6.99268460e-01 -9.73796725e-01
1.16056395e+00 -2.14156556e+00 2.59404093e-01 1.69415891e-01
1.81037247e-01 5.66839755e-01 -5.16988821e-02 2.24980444e-01
-4.93774146e-01 -8.88792500e-02 -4.81635213e-01 -2.32261896e-01
-4.06578213e-01 7.11828172e-02 -3.12088609e-01 8.78525138e-01
-8.42260271e-02 6.93503499e-01 -5.54542840e-01 -2.57015914e-01
4.23505455e-01 8.80311191e-01 -2.61366904e-01 2.00525448e-01
-2.04561446e-02 5.16944170e-01 -2.64031261e-01 1.68505058e-01
1.08474302e+00 -1.93954438e-01 4.56817597e-02 -6.69116437e-01
-2.86266327e-01 -3.67756128e-01 -1.43386328e+00 1.39792800e+00
-5.16039848e-01 7.54120946e-01 6.02347851e-01 -1.30227661e+00
7.54585743e-01 5.00170052e-01 4.98625785e-01 -6.17265046e-01
4.30352688e-01 5.09022295e-01 -2.16376975e-01 -6.99042022e-01
4.73040044e-01 -3.00173491e-01 5.81857979e-01 4.42378074e-01
-2.65781343e-01 -1.73815280e-01 2.41471380e-02 -2.13835299e-01
6.59547389e-01 4.75716852e-02 3.98007661e-01 -5.33978999e-01
9.16216612e-01 -7.62958378e-02 3.02574754e-01 6.46778762e-01
-1.67942774e-02 8.67000759e-01 3.26728314e-01 -4.51526552e-01
-1.04195607e+00 -5.20330846e-01 -1.99182600e-01 2.43637741e-01
2.93193161e-01 3.05407882e-01 -1.13984406e+00 -1.86994188e-02
-5.27185142e-01 4.65195179e-01 -4.79339093e-01 -1.70502961e-01
-6.99248612e-01 -1.10874200e+00 4.74233389e-01 4.24341708e-02
7.25722551e-01 -8.51214886e-01 -6.48488939e-01 3.19562435e-01
-4.06040668e-01 -9.33265686e-01 -1.08998574e-01 1.87200624e-02
-1.12646902e+00 -1.01267827e+00 -1.30003417e+00 -6.34163857e-01
6.73997104e-01 3.64877224e-01 9.38760519e-01 3.28552097e-01
2.72167716e-02 3.41105796e-02 -1.79187924e-01 3.13820899e-01
-8.03657889e-01 -2.25192413e-01 1.04210436e-01 2.35502750e-01
-2.81920522e-01 -9.57569420e-01 -6.09637856e-01 3.86684000e-01
-1.27908218e+00 4.72531430e-02 7.48422742e-01 1.02526200e+00
3.93160671e-01 3.37335825e-01 3.42939466e-01 -4.32146817e-01
7.81825364e-01 -2.03970492e-01 -7.11953521e-01 1.07422806e-01
-7.25940466e-01 1.34419471e-01 8.45231414e-01 -5.54963529e-01
-1.01575279e+00 3.94334719e-02 -3.54357272e-01 -5.94161689e-01
-2.00188547e-01 6.97012186e-01 -6.85376525e-02 -3.55192840e-01
8.98565531e-01 6.51830196e-01 3.35421950e-01 -5.11953235e-01
1.92707762e-01 4.42500412e-01 6.20128095e-01 -3.21058899e-01
9.41354692e-01 4.49867606e-01 1.27344728e-01 -1.02560186e+00
-3.86169374e-01 -8.86226073e-02 -3.64871383e-01 -4.07008380e-01
7.83865929e-01 -7.46323049e-01 -7.52586067e-01 1.08729875e+00
-1.39076769e+00 -2.26248115e-01 -2.02751040e-01 4.98383969e-01
-4.89489585e-01 8.33851874e-01 -7.48683989e-01 -6.08017504e-01
-5.05899847e-01 -1.29790401e+00 3.46499026e-01 2.88277984e-01
2.36245498e-01 -9.76095140e-01 5.99261709e-02 1.16783746e-01
9.71469760e-01 3.38242888e-01 7.37513483e-01 -2.35967990e-02
-4.32450593e-01 -7.72244781e-02 -4.23191339e-01 6.58533812e-01
-5.47393085e-03 -1.29201338e-01 -9.85491574e-01 -4.50429738e-01
6.07651770e-01 -7.02342689e-02 8.85906219e-01 7.76983857e-01
9.29855168e-01 -3.73250127e-01 5.64823896e-02 6.96802258e-01
1.70754933e+00 -1.08448692e-01 8.52536082e-01 3.24798524e-01
4.48076814e-01 4.17956948e-01 6.26222119e-02 2.30237871e-01
-1.54817626e-01 7.40762651e-01 7.24708796e-01 -4.92597103e-01
-4.19670373e-01 2.60299772e-01 3.45087737e-01 9.32351470e-01
-4.51389879e-01 -3.36679399e-01 -7.06642151e-01 4.33481544e-01
-1.75474882e+00 -9.64839220e-01 -5.67546666e-01 2.23989630e+00
8.02677631e-01 -8.72576162e-02 -2.71817178e-01 3.85738790e-01
7.78735638e-01 3.07070583e-01 -2.92956859e-01 -4.17538024e-02
-3.07413787e-01 -1.70235988e-02 4.53390956e-01 7.85860002e-01
-1.01236618e+00 5.26185036e-01 5.88372135e+00 8.14168215e-01
-1.58070111e+00 3.43786925e-01 6.21256948e-01 3.15551043e-01
-1.05615139e-01 -2.80705988e-01 -4.74805050e-02 5.00614405e-01
6.88073575e-01 7.51999319e-02 5.33406973e-01 6.28487170e-01
5.88266134e-01 -2.88175970e-01 -6.30322635e-01 1.05025625e+00
1.61747515e-01 -1.31661415e+00 -1.41374260e-01 -9.26925614e-02
7.27261007e-01 -1.49530604e-01 5.58415838e-02 -3.91839474e-01
-2.19410717e-01 -1.06580973e+00 8.46008599e-01 8.12958300e-01
6.83526695e-01 -6.52032733e-01 1.05410361e+00 4.49810714e-01
-7.22619593e-01 1.93457216e-01 -3.88721168e-01 -8.44526142e-02
4.13037837e-01 1.13446021e+00 -8.67551938e-02 7.43779421e-01
5.97773194e-01 6.04750752e-01 -2.22550526e-01 1.11768103e+00
-2.83607543e-01 4.94441897e-01 -2.29566410e-01 1.40090972e-01
1.23777620e-01 -4.61446255e-01 7.66937673e-01 1.02287793e+00
6.01293445e-01 1.10884480e-01 -4.09209192e-01 1.11158061e+00
2.92435944e-01 5.42384014e-02 -3.10589671e-01 2.64082134e-01
-1.39652202e-02 1.11733103e+00 -7.70475328e-01 -2.97419757e-01
-2.02774368e-02 1.12111807e+00 -1.63277328e-01 6.76234663e-01
-8.37871671e-01 -2.27912396e-01 3.54946047e-01 2.82468706e-01
1.37722522e-01 -2.56033540e-01 -2.74457961e-01 -1.33384645e+00
9.37358961e-02 -9.02972460e-01 -1.30804896e-01 -9.68881011e-01
-1.08242476e+00 9.21578467e-01 -2.00113639e-01 -1.37800109e+00
-3.05264354e-01 -5.14878690e-01 -4.14237618e-01 1.10544050e+00
-1.70767105e+00 -9.28727984e-01 -6.70361578e-01 5.05074382e-01
1.53122574e-01 1.94262698e-01 6.62348568e-01 6.63167417e-01
-4.93690968e-01 2.73820698e-01 4.95308667e-01 -9.66372266e-02
6.09607399e-01 -7.18121290e-01 -1.21421278e-01 1.31013095e+00
-3.45646024e-01 5.55394709e-01 1.19718170e+00 -5.11411250e-01
-1.26591027e+00 -6.97151601e-01 6.08830690e-01 2.70980060e-01
4.32128727e-01 1.38987854e-01 -1.26891422e+00 1.47628546e-01
2.48860016e-01 1.65110350e-01 -3.25042419e-02 -4.05149877e-01
-3.15556712e-02 -2.23927125e-01 -1.15760505e+00 2.32626751e-01
4.03946579e-01 -3.61148357e-01 -5.10857403e-01 1.59223050e-01
2.90965796e-01 -4.00517821e-01 -5.99371910e-01 2.20967501e-01
4.38010246e-01 -1.46751046e+00 1.13781166e+00 4.02419381e-02
7.22722948e-01 -5.19135833e-01 1.76929876e-01 -1.38167429e+00
-3.65023404e-01 -5.16005933e-01 -1.22326575e-01 1.00859976e+00
-3.83995101e-02 -8.70119810e-01 5.37957609e-01 1.82677642e-01
-1.15253199e-02 -5.09484828e-01 -1.05783284e+00 -7.73951828e-01
7.75969848e-02 -2.65691817e-01 2.34819114e-01 1.08243918e+00
-4.82436329e-01 -9.75745097e-02 -8.48981917e-01 3.32006544e-01
7.79800832e-01 -6.07920550e-02 4.40955967e-01 -1.14757597e+00
-3.49871457e-01 -6.25977695e-01 -3.14686418e-01 -1.07202911e+00
-8.00204352e-02 -2.95293242e-01 1.80361778e-01 -1.44261193e+00
-4.63600084e-02 -3.57906282e-01 -1.58095017e-01 1.02719702e-01
-7.41176903e-02 4.75037426e-01 -4.02977951e-02 5.45755804e-01
1.78474829e-01 5.49271762e-01 1.26343000e+00 -6.50621951e-02
-1.22021779e-01 -2.87252050e-02 -1.51894361e-01 6.65444195e-01
6.16167188e-01 -4.49052721e-01 -3.10267746e-01 -5.93735754e-01
2.90167212e-01 2.15092197e-01 5.78995287e-01 -1.24082792e+00
3.44683021e-01 7.67682167e-03 2.68317103e-01 -3.24343324e-01
2.66524166e-01 -1.08899188e+00 6.69520795e-01 7.90315688e-01
1.55353621e-02 -2.35928699e-01 7.09197819e-02 3.20806354e-01
-4.90324527e-01 -5.22192299e-01 1.26394534e+00 -1.73556749e-02
-5.02689481e-01 -5.96053191e-02 -5.65854311e-01 -3.60661596e-01
7.05157638e-01 -3.49564821e-01 -1.46052569e-01 -7.02884555e-01
-5.87602079e-01 -4.65263098e-01 6.16053581e-01 -7.90990144e-02
3.70235682e-01 -1.10958695e+00 -6.78038120e-01 2.11781904e-01
-4.71011013e-01 -1.97294563e-01 7.19418705e-01 1.19277620e+00
-1.02408266e+00 5.36418967e-02 -3.67649049e-01 -4.86254483e-01
-1.06219089e+00 6.94709957e-01 7.45485008e-01 -3.11062783e-01
-5.89383125e-01 5.54390490e-01 -5.74925952e-02 2.83059850e-02
6.32333681e-02 -3.77270401e-01 -3.05426806e-01 2.91974861e-02
6.75792158e-01 4.82072979e-01 2.64539033e-01 -8.76171947e-01
-1.56629279e-01 9.26689804e-01 3.62890840e-01 -8.06163996e-02
1.39332163e+00 -2.49100894e-01 -5.65585792e-01 -1.05008529e-02
9.98822629e-01 -5.49274944e-02 -1.12875462e+00 -1.35898635e-01
-3.98687601e-01 -1.93247765e-01 6.57189548e-01 -6.18621707e-01
-1.35893512e+00 9.64995205e-01 9.25629973e-01 3.95381004e-01
1.50768805e+00 -5.67779899e-01 7.22275436e-01 8.26782286e-02
2.38697324e-02 -7.98760831e-01 -8.74433517e-02 3.46124679e-01
1.07853401e+00 -8.95105124e-01 2.16388121e-01 -3.62565845e-01
1.59472395e-02 1.45417166e+00 2.57126391e-01 -1.61079749e-01
5.83472908e-01 3.63261431e-01 2.11443186e-01 7.07779974e-02
3.02364249e-02 9.44712833e-02 2.59320617e-01 3.32507908e-01
2.76305348e-01 -2.61110038e-01 -6.55995071e-01 4.04949039e-01
1.32887140e-01 3.48636270e-01 6.89733803e-01 4.58734185e-01
-4.26142424e-01 -7.81570077e-01 -9.41404223e-01 -1.14650048e-01
-4.23315048e-01 -2.43792571e-02 1.71939760e-01 6.96457088e-01
1.34983003e-01 7.94258952e-01 -1.80467725e-01 -1.08981043e-01
1.07242495e-01 -2.30271265e-01 3.14109981e-01 1.97124600e-01
-4.24872667e-01 2.04125807e-01 -2.66281486e-01 -5.17927229e-01
-7.53711522e-01 -2.98356175e-01 -8.91949713e-01 -3.70480657e-01
-4.92123127e-01 1.78085431e-01 7.41497338e-01 1.01196051e+00
2.28454307e-01 5.97875357e-01 3.96756709e-01 -1.16406190e+00
-6.56110883e-01 -9.87952650e-01 -4.73301947e-01 2.93047994e-01
6.29510224e-01 -5.20384073e-01 -6.41449690e-01 2.47783333e-01] | [11.711026191711426, -2.520247220993042] |
34c8cade-9e55-4d96-92e8-9a021e151579 | followme-vehicle-behaviour-prediction-in | 2304.06121 | null | https://arxiv.org/abs/2304.06121v1 | https://arxiv.org/pdf/2304.06121v1.pdf | FollowMe: Vehicle Behaviour Prediction in Autonomous Vehicle Settings | An ego vehicle following a virtual lead vehicle planned route is an essential component when autonomous and non-autonomous vehicles interact. Yet, there is a question about the driver's ability to follow the planned lead vehicle route. Thus, predicting the trajectory of the ego vehicle route given a lead vehicle route is of interest. We introduce a new dataset, the FollowMe dataset, which offers a motion and behavior prediction problem by answering the latter question of the driver's ability to follow a lead vehicle. We also introduce a deep spatio-temporal graph model FollowMe-STGCNN as a baseline for the dataset. In our experiments and analysis, we show the design benefits of FollowMe-STGCNN in capturing the interactions that lie within the dataset. We contrast the performance of FollowMe-STGCNN with prior motion prediction models showing the need to have a different design mechanism to address the lead vehicle following settings. | ['Christian Claudel', 'Linda Ng Boyle', 'Jundi Liu', 'Abduallah Mohamed'] | 2023-04-12 | null | null | null | null | ['motion-prediction'] | ['computer-vision'] | [-2.45468885e-01 4.09770519e-01 -3.82063866e-01 -6.02030456e-01
-7.16628581e-02 -4.93569762e-01 8.29697311e-01 -4.36206520e-01
-3.15562874e-01 1.05067022e-01 4.70337480e-01 -9.14040267e-01
-2.29235440e-01 -8.27993989e-01 -9.38557148e-01 -3.39176267e-01
-6.39836341e-02 3.04582149e-01 4.55489367e-01 -5.99358261e-01
2.73932163e-02 6.60860658e-01 -1.49686432e+00 -2.62017306e-02
5.13395667e-01 6.32764220e-01 2.45854899e-01 8.73080552e-01
2.60221183e-01 8.63851964e-01 3.16783637e-01 -2.58230090e-01
1.60426766e-01 -1.84611484e-01 -8.04410279e-01 5.84298857e-02
3.47707033e-01 -4.84994233e-01 -1.17625034e+00 4.31289762e-01
-1.05283178e-01 4.01427060e-01 6.53628170e-01 -2.13831306e+00
-6.61287084e-02 3.22573215e-01 -5.26172593e-02 4.22397047e-01
2.12226391e-01 6.20211661e-01 7.09633887e-01 -3.47129881e-01
8.78947794e-01 1.23686266e+00 7.65348196e-01 5.11283815e-01
-9.98593390e-01 -4.32076365e-01 3.83719563e-01 6.79511428e-01
-1.37541044e+00 -7.64472127e-01 7.08762705e-01 -6.33375227e-01
9.36592579e-01 -3.55709977e-02 7.70117581e-01 1.21369755e+00
6.81805789e-01 8.27269495e-01 5.27715325e-01 3.39410931e-01
1.00431912e-01 3.39763314e-02 5.46393871e-01 8.33428323e-01
-1.12466544e-01 4.71896380e-01 -3.48273605e-01 2.10532859e-01
1.58713251e-01 -1.31742328e-01 -9.56200138e-02 -5.55607200e-01
-7.01050937e-01 8.41825426e-01 4.44261342e-01 -1.35482684e-01
-4.32662249e-01 5.60182393e-01 1.13890193e-01 2.45272726e-01
1.33252725e-01 -2.81328976e-01 -2.33790070e-01 -4.26072240e-01
-7.31402040e-01 5.50624192e-01 8.28836799e-01 1.39586723e+00
1.14804578e+00 -9.63428989e-03 -2.11461931e-01 -6.28223419e-02
4.01720613e-01 1.91818520e-01 -1.11862309e-01 -1.10246134e+00
3.88891876e-01 4.80568498e-01 2.58926988e-01 -1.12576079e+00
-7.37682581e-01 -3.24619234e-01 -3.12977463e-01 1.04656406e-01
4.19705272e-01 -2.27061167e-01 -8.81300032e-01 1.92279887e+00
4.05696422e-01 6.84296548e-01 1.08354904e-01 7.74969339e-01
9.00811374e-01 6.40043914e-01 5.79005964e-02 2.31048122e-01
1.09586418e+00 -1.04036713e+00 -6.63442075e-01 -6.49993062e-01
1.17970872e+00 -1.14946947e-01 7.86113381e-01 -3.37086499e-01
-8.75846446e-01 -5.81914306e-01 -8.96214306e-01 -1.96522400e-01
-4.07939523e-01 -1.07193738e-01 5.01429915e-01 4.67746168e-01
-1.44385099e+00 4.40425366e-01 -9.74728644e-01 -4.79922295e-01
3.28458279e-01 4.03694242e-01 -6.38230205e-01 -2.92120099e-01
-1.11236906e+00 1.15043116e+00 -8.08318481e-02 2.65686810e-01
-1.13591337e+00 -8.49654496e-01 -1.15071750e+00 2.40496080e-02
4.03122157e-01 -7.58216262e-01 1.31790280e+00 -3.09699178e-01
-9.55076396e-01 8.34920108e-01 -8.22156489e-01 -6.43967330e-01
8.61477256e-01 8.43440220e-02 -6.67462349e-01 -3.02577943e-01
2.57492065e-01 7.79768050e-01 5.87837160e-01 -1.01393878e+00
-9.65618968e-01 -3.30562681e-01 5.09687513e-02 2.21168369e-01
4.19718027e-01 -4.15752590e-01 -8.72394204e-01 2.93694466e-01
-1.58509985e-01 -1.53362107e+00 -1.85977548e-01 -3.09482992e-01
-5.52380800e-01 -6.95441782e-01 1.20947909e+00 -2.57413864e-01
1.13732374e+00 -2.48546028e+00 -2.32127279e-01 2.17359975e-01
5.77245772e-01 7.70540535e-02 -4.41632986e-01 6.47174239e-01
1.63994968e-01 1.23819806e-01 1.85968384e-01 -6.03354216e-01
1.96457654e-01 3.79832149e-01 -1.32022217e-01 7.59157240e-01
-7.63523504e-02 1.33453119e+00 -7.05830276e-01 -3.23298685e-02
6.70182481e-02 4.93102729e-01 -5.38265407e-01 9.61354524e-02
-7.77069554e-02 2.84797281e-01 -5.55247426e-01 1.26863375e-01
7.00657248e-01 -7.38858059e-02 -1.89237714e-01 2.52833776e-02
-2.91481316e-01 4.83579397e-01 -7.65502453e-01 1.18854320e+00
-2.20472276e-01 1.28050601e+00 2.22880080e-01 -5.54383874e-01
5.83673894e-01 -5.21115623e-02 3.81405383e-01 -9.08915997e-01
2.85596818e-01 -2.12634996e-01 3.11835527e-01 -6.71144843e-01
8.21363032e-01 3.69236827e-01 -8.39370042e-02 4.40887451e-01
-5.50355971e-01 3.27195376e-01 2.33381227e-01 4.96471077e-01
1.25895870e+00 7.43300095e-02 -1.69531360e-01 -3.43105882e-01
2.11257935e-01 4.23910022e-01 6.92205310e-01 7.95628726e-01
-7.87415981e-01 3.00470531e-01 6.17200077e-01 -6.66924655e-01
-8.46380293e-01 -8.00985277e-01 1.44650027e-01 9.36150730e-01
5.88379502e-01 -4.38276112e-01 -7.98477769e-01 -5.07854640e-01
2.49176726e-01 1.11509097e+00 -9.08791423e-01 -5.27478874e-01
-7.22023427e-01 -2.09339723e-01 4.06072587e-01 4.50561285e-01
5.05150497e-01 -7.15071082e-01 -6.27278864e-01 1.64601997e-01
-2.62820631e-01 -1.63207090e+00 -8.10381174e-01 -3.22817415e-01
-2.44047120e-01 -1.25154567e+00 1.06195867e-01 -6.48755908e-01
4.09139127e-01 7.80326009e-01 1.01670063e+00 2.54087448e-01
3.81490886e-01 8.03953588e-01 -7.00237975e-02 -4.68514055e-01
-6.76476777e-01 2.67145574e-01 1.68017000e-01 3.40393603e-01
8.32747936e-01 -4.90711302e-01 -1.00725901e+00 6.54778600e-01
-3.33651602e-01 3.73256147e-01 8.26556385e-02 4.73487638e-02
3.30727488e-01 2.39889100e-01 3.73424500e-01 -7.17980921e-01
4.66187924e-01 -8.52518976e-01 -7.03003287e-01 -3.21671665e-01
-8.53108287e-01 -1.71924397e-01 3.13161612e-01 -1.41513199e-01
-6.37473404e-01 2.64341563e-01 -2.84757793e-01 -4.10892725e-01
-2.92102516e-01 3.86765510e-01 -2.65715748e-01 5.86312301e-02
1.99748233e-01 3.64188224e-01 4.30736721e-01 -3.71829756e-02
4.99324232e-01 1.81986451e-01 5.42964518e-01 2.61449397e-01
7.40181625e-01 7.34427333e-01 3.92735779e-01 -9.55032766e-01
-2.57604659e-01 -4.45513338e-01 -4.85139668e-01 -5.20791948e-01
1.02489078e+00 -1.07939386e+00 -1.31747341e+00 3.03752959e-01
-1.07776332e+00 -9.25850272e-01 1.16097696e-01 3.53295743e-01
-7.89145112e-01 1.78418905e-01 -2.94473469e-01 -6.05557144e-01
2.55867392e-01 -1.38943779e+00 8.72225404e-01 6.42259642e-02
-3.22483927e-01 -1.21679461e+00 -1.07492305e-01 3.82008523e-01
2.88456529e-01 1.29633859e-01 7.45912492e-01 -5.33867300e-01
-1.00104070e+00 -2.79654682e-01 -1.41235933e-01 -3.93552274e-01
-7.08661526e-02 -8.22198913e-02 -6.86350167e-01 -3.38599801e-01
-4.31852162e-01 3.89950126e-01 8.99941504e-01 7.44364262e-01
6.83517396e-01 -2.30764955e-01 -8.80993783e-01 8.94401550e-01
9.89116669e-01 2.32656732e-01 8.71475697e-01 3.93313736e-01
9.81522739e-01 9.90809381e-01 5.34085453e-01 -6.61673024e-02
1.45276511e+00 7.22970307e-01 7.35212266e-01 1.21303163e-01
-2.28450030e-01 -8.09043229e-01 3.34114045e-01 1.67591274e-01
3.59892219e-01 -6.43127501e-01 -1.11541605e+00 9.28309560e-01
-2.14359331e+00 -1.14245641e+00 -8.92060220e-01 1.81119335e+00
-2.52228111e-01 1.35239556e-01 4.12557691e-01 -2.18370497e-01
3.39874297e-01 1.22981757e-01 -7.51662076e-01 -3.56506675e-01
5.42864278e-02 -8.65591824e-01 9.10466492e-01 9.33230937e-01
-9.20429468e-01 1.22167099e+00 6.56967688e+00 4.92165059e-01
-9.94115412e-01 -1.00532643e-01 5.84018826e-01 4.77110520e-02
-4.24860507e-01 2.38369271e-01 -1.34751487e+00 5.16564965e-01
1.28278947e+00 -3.71381044e-01 3.97124738e-01 6.10672355e-01
1.02311623e+00 -1.97158143e-01 -1.44368362e+00 9.01097000e-01
-2.47980103e-01 -1.47000492e+00 -2.91474372e-01 5.86114585e-01
4.42007273e-01 4.92089719e-01 1.37186617e-01 4.02524233e-01
4.55905259e-01 -1.13412213e+00 7.79188752e-01 5.24220288e-01
3.60662460e-01 -8.50227535e-01 3.74240100e-01 7.95014203e-01
-1.39701629e+00 -1.42395169e-01 -2.88720597e-02 -1.97529539e-01
5.76626003e-01 -4.83729839e-02 -1.13582468e+00 2.15699360e-01
4.75160211e-01 8.93972635e-01 -6.85396910e-01 7.03876138e-01
-6.52027130e-02 6.64947450e-01 -7.73085058e-02 2.25451589e-03
7.32853115e-01 -4.44043666e-01 8.67971361e-01 9.97060001e-01
3.36970568e-01 6.42872974e-02 1.90166179e-02 9.86665726e-01
8.88940766e-02 -4.33537513e-01 -1.35197699e+00 3.92711908e-02
3.53046328e-01 1.01932728e+00 -3.75183016e-01 -4.16002981e-02
-5.24133265e-01 6.91417336e-01 2.91762054e-01 6.90240920e-01
-7.86957145e-01 7.34704360e-02 1.24268270e+00 6.74067020e-01
3.00774902e-01 -4.47558701e-01 -1.82782263e-01 -5.57136595e-01
-8.21926072e-02 -2.84716070e-01 8.97179265e-03 -9.68376994e-01
-5.90011656e-01 5.88513732e-01 8.58822986e-02 -1.06213582e+00
-6.52236998e-01 -4.05008882e-01 -9.52573001e-01 9.29971099e-01
-1.75983846e+00 -1.36686933e+00 -5.07722795e-01 7.33573079e-01
4.42885727e-01 -1.57402411e-01 1.21748736e-02 -3.96060757e-03
-6.48235083e-01 3.93882006e-01 -2.15487003e-01 -6.74831569e-02
5.41502535e-02 -8.47128272e-01 9.83585835e-01 9.09876645e-01
-2.91897684e-01 3.53186488e-01 9.03092742e-01 -7.93691635e-01
-1.87410557e+00 -1.58239019e+00 1.13970780e+00 -1.08259213e+00
5.48679113e-01 -3.85260910e-01 -8.30017269e-01 1.28113639e+00
1.73994124e-01 -4.52227816e-02 3.80000234e-01 -6.47987351e-02
1.50193095e-01 -1.97736677e-02 -7.22825944e-01 1.05920386e+00
1.46493232e+00 -6.87460423e-01 -1.35709748e-01 -6.54127002e-02
7.24217296e-01 -1.92329243e-01 -1.60305172e-01 3.06401283e-01
6.17256165e-01 -1.04622495e+00 6.76373839e-01 -7.65697002e-01
2.34337062e-01 -3.21383893e-01 -3.82464305e-02 -1.40943217e+00
-6.56342447e-01 -8.03107619e-01 -5.00393324e-02 8.08361888e-01
4.40042049e-01 -6.53601706e-01 1.23543477e+00 1.22857356e+00
-3.07854533e-01 -6.03867352e-01 -8.12689066e-01 -7.07849264e-01
9.43303853e-02 -9.22939658e-01 6.96153045e-01 4.89151835e-01
-3.25307637e-01 3.44404936e-01 -4.13366407e-01 5.80175757e-01
5.28811634e-01 -1.48962080e-01 1.34568751e+00 -1.04141521e+00
3.67729425e-01 -5.04008651e-01 -6.83322012e-01 -1.64315677e+00
6.04124904e-01 -9.65852320e-01 7.81302452e-02 -1.93954170e+00
4.28040512e-02 -2.23073632e-01 2.34047204e-01 6.77616671e-02
3.30785476e-02 -1.42625690e-01 5.64130917e-02 -1.22609828e-02
-3.73371124e-01 5.26840687e-01 1.35490239e+00 -1.17330596e-01
-4.32788283e-01 3.25051516e-01 -6.90665901e-01 5.66389978e-01
6.84582829e-01 -3.05503517e-01 -9.03539002e-01 -4.11268413e-01
3.53986561e-01 2.64820874e-01 6.50403082e-01 -6.28987253e-01
7.92137086e-01 -1.78488344e-01 -3.37847501e-01 -1.06802499e+00
5.60100198e-01 -7.88200855e-01 3.76789331e-01 2.50743091e-01
-1.45103216e-01 4.24334973e-01 3.90824735e-01 9.77170110e-01
1.52798146e-01 2.69712597e-01 4.18194264e-01 2.87318021e-01
-1.31639135e+00 9.52083826e-01 -1.03201747e+00 -5.63980490e-02
1.21291769e+00 -6.51884675e-01 -2.95473784e-01 -1.00169265e+00
-7.73066640e-01 9.23280656e-01 5.30592263e-01 8.25172782e-01
8.11158776e-01 -1.27833104e+00 -7.58602321e-01 2.23183304e-01
2.92814016e-01 -3.36788416e-01 5.89759588e-01 9.78686273e-01
-3.04908872e-01 6.68875217e-01 2.32058913e-02 -6.75661147e-01
-1.17120969e+00 6.68821096e-01 5.41846693e-01 2.51793146e-01
-9.25182521e-01 4.70110476e-01 6.22068763e-01 -4.01409179e-01
-1.25329852e-01 -1.54057026e-01 -4.44693029e-01 -2.84240812e-01
2.70967007e-01 6.55896783e-01 -1.26403183e-01 -1.36450565e+00
-4.10156667e-01 3.49829108e-01 -9.88202542e-03 -2.77384464e-02
9.71028388e-01 -7.89439678e-01 3.91387016e-01 2.77535349e-01
1.40820277e+00 -2.67334759e-01 -1.65054619e+00 3.14589287e-03
-4.25387966e-03 -3.25521857e-01 3.57524872e-01 -2.45249063e-01
-9.20623302e-01 7.35708237e-01 5.18000007e-01 2.61850357e-01
5.10044098e-01 6.46229014e-02 9.03261662e-01 2.96515107e-01
2.71690160e-01 -9.10028279e-01 -4.88404304e-01 6.62729204e-01
6.75629318e-01 -1.43863451e+00 -5.39467096e-01 -5.67465961e-01
-6.46000385e-01 6.42989933e-01 6.98854685e-01 8.32484215e-02
1.04225671e+00 -7.81289190e-02 1.43846318e-01 -5.54902613e-01
-1.01393139e+00 -4.88604844e-01 4.73795682e-01 6.71337306e-01
-7.05254823e-02 1.80394322e-01 1.41808659e-01 1.99539766e-01
-5.85204482e-01 -7.15373382e-02 8.26306462e-01 6.87461793e-01
-1.08300731e-01 -7.53865302e-01 1.84748694e-01 4.37548429e-01
1.10083163e-01 2.72838056e-01 -3.61503839e-01 9.87417877e-01
-8.79040211e-02 1.53392839e+00 4.20923650e-01 -7.52762914e-01
4.36882645e-01 -6.17725030e-02 -1.30826727e-01 -2.75425345e-01
-3.37332040e-02 -3.73476535e-01 4.35187429e-01 -9.84461248e-01
-3.25927474e-02 -8.53385925e-01 -1.28486872e+00 -1.01906455e+00
2.50796705e-01 -9.74762142e-02 5.35136819e-01 1.21683931e+00
8.86135757e-01 5.18863380e-01 7.61538565e-01 -6.84627175e-01
1.12382919e-01 -5.15993536e-01 -5.36060989e-01 3.98816884e-01
9.44271922e-01 -5.29585302e-01 -3.12599748e-01 -2.94121772e-01] | [5.970358371734619, 0.8524004220962524] |
243c95e5-7a34-42b2-9fc8-87e40a697d53 | reco-a-dataset-for-residential-community | 2206.04678 | null | https://arxiv.org/abs/2206.04678v2 | https://arxiv.org/pdf/2206.04678v2.pdf | ReCo: A Dataset for Residential Community Layout Planning | Layout planning is centrally important in the field of architecture and urban design. Among the various basic units carrying urban functions, residential community plays a vital part for supporting human life. Therefore, the layout planning of residential community has always been of concern, and has attracted particular attention since the advent of deep learning that facilitates the automated layout generation and spatial pattern recognition. However, the research circles generally suffer from the insufficiency of residential community layout benchmark or high-quality datasets, which hampers the future exploration of data-driven methods for residential community layout planning. The lack of datasets is largely due to the difficulties of large-scale real-world residential data acquisition and long-term expert screening. In order to address the issues and advance a benchmark dataset for various intelligent spatial design and analysis applications in the development of smart city, we introduce Residential Community Layout Planning (ReCo) Dataset, which is the first and largest open-source vector dataset related to real-world community to date. ReCo Dataset is presented in multiple data formats with 37,646 residential community layout plans, covering 598,728 residential buildings with height information. ReCo can be conveniently adapted for residential community layout related urban design tasks, e.g., generative layout design, morphological pattern recognition and spatial evaluation. To validate the utility of ReCo in automated residential community layout planning, two Generative Adversarial Network (GAN) based generative models are further applied to the dataset. We expect ReCo Dataset to inspire more creative and practical work in intelligent design and beyond. The ReCo Dataset is published at: https://www.kaggle.com/fdudsde/reco-dataset. | ['Yu Ye', 'Yao Zhang', 'Tao Sheng', 'Haofen Wang', 'Siqi Wang', 'Yun Xiong', 'Xi Chen'] | 2022-06-08 | null | null | null | null | ['layout-design'] | ['computer-vision'] | [-2.10445791e-01 -3.04249078e-01 2.97572345e-01 -3.35559994e-01
-5.63574970e-01 -3.32280457e-01 4.22016889e-01 -2.06003472e-01
1.51354283e-01 7.10653245e-01 7.24265277e-01 -6.25571549e-01
-3.84743989e-01 -1.55324340e+00 -3.16650987e-01 -8.55975449e-01
6.00756407e-02 5.96841455e-01 -2.81598538e-01 -4.58878994e-01
-3.78272869e-03 7.06075907e-01 -1.40392649e+00 1.23464480e-01
1.06319427e+00 7.64671028e-01 1.72946587e-01 2.25379542e-01
3.11110802e-02 5.00476718e-01 -5.29513955e-01 -1.41587973e-01
1.21790528e-01 -1.77300543e-01 -5.62895954e-01 1.44755363e-01
-1.05302170e-01 -4.05433893e-01 -6.73564613e-01 7.13567674e-01
1.39948165e+00 1.71182573e-01 8.05475175e-01 -1.46237254e+00
-1.02534187e+00 8.50695193e-01 -3.44281495e-01 -9.89560932e-02
2.07530856e-01 4.98429239e-01 1.05810201e+00 -7.41900861e-01
2.74785161e-01 9.80653107e-01 6.60894156e-01 2.43495062e-01
-1.22290945e+00 -6.37861788e-01 6.18972955e-03 2.74291575e-01
-1.86648941e+00 -3.25677603e-01 1.19162941e+00 -5.50931692e-01
7.20175922e-01 4.69554484e-01 8.42472792e-01 1.20169449e+00
-3.69127810e-01 9.48485255e-01 8.29204023e-01 6.73084185e-02
4.86471504e-01 -3.53127301e-01 -4.26286131e-01 2.81771213e-01
2.02171147e-01 1.92784235e-01 3.20429236e-01 2.61095613e-01
9.21658516e-01 -5.02563119e-02 -1.43968642e-01 -2.62486309e-01
-9.30483818e-01 8.61056268e-01 1.13599741e+00 4.96460736e-01
-2.68221706e-01 1.28871724e-01 2.22851336e-01 -7.27554858e-02
9.65668857e-02 2.84180999e-01 4.86995764e-02 -1.55105710e-01
-9.14417982e-01 3.93713951e-01 3.03483278e-01 1.22976851e+00
4.72127914e-01 4.72646147e-01 -7.76895955e-02 1.05407476e+00
4.69860107e-01 8.00859094e-01 3.31468761e-01 -4.92665678e-01
7.68281639e-01 7.97682106e-01 -2.68762946e-01 -1.22709143e+00
-4.92479444e-01 -4.23722386e-01 -1.47604275e+00 -1.78952500e-01
-1.12523742e-01 -1.34896830e-01 -4.91450429e-01 1.40632224e+00
1.12868600e-01 -2.09623843e-01 -1.68754399e-01 7.06348956e-01
1.41809535e+00 6.86053932e-01 -1.20319098e-01 4.38258886e-01
9.66358781e-01 -6.73638880e-01 -1.71211094e-01 -1.56280026e-01
4.64529812e-01 -5.58326781e-01 1.17897534e+00 -1.96321197e-02
-5.29137671e-01 -5.32764137e-01 -9.80191588e-01 1.01316045e-03
-5.06174326e-01 2.21096784e-01 9.06839788e-01 7.65795648e-01
-9.50712681e-01 3.03176302e-03 -3.99418682e-01 -5.60100019e-01
1.12146902e+00 2.26262599e-01 -6.20703064e-02 -1.84895873e-01
-9.86287773e-01 3.89398336e-01 2.34329805e-01 4.46175575e-01
-6.67921722e-01 -6.28672481e-01 -7.53388286e-01 1.91592425e-01
-1.35469129e-02 -7.61516154e-01 8.13792288e-01 -3.30553353e-01
-8.34653676e-01 4.34066057e-01 4.46340978e-01 -1.71004370e-01
4.73544180e-01 2.87417173e-01 -7.96568394e-01 -5.74777842e-01
3.82340550e-01 5.94298661e-01 2.07811385e-01 -1.34154844e+00
-2.67808229e-01 -2.61745751e-01 -1.17839172e-01 -1.96697533e-01
-2.28195980e-01 -5.36176801e-01 -1.53169796e-01 -8.91872346e-01
6.61668461e-03 -6.59636915e-01 -4.41776693e-01 -4.24606591e-01
-7.97222674e-01 -1.39294758e-01 7.68334389e-01 -6.50661111e-01
1.63989949e+00 -2.14392066e+00 -3.73902559e-01 4.45736408e-01
-6.26930743e-02 2.32283577e-01 -3.21453780e-01 7.25739956e-01
-1.80977702e-01 3.14615577e-01 -3.19396019e-01 2.61695907e-02
3.58777642e-01 1.00038238e-01 -2.57140487e-01 2.42269844e-01
-2.12684497e-02 1.59364486e+00 -7.75172114e-01 -3.34192485e-01
5.38575768e-01 5.75223327e-01 -5.96678674e-01 -1.95012540e-02
-1.25395983e-01 5.12763977e-01 -8.39795709e-01 9.03194070e-01
8.15834701e-01 -1.16136767e-01 -2.10372619e-02 -1.99077487e-01
-8.18810090e-02 5.55634350e-02 -1.32039714e+00 1.51822138e+00
-5.11995316e-01 8.30601096e-01 -3.15809727e-01 -1.11292565e+00
1.17899847e+00 8.51644129e-02 6.86148465e-01 -1.25400174e+00
3.56990010e-01 1.19634010e-01 -2.10063547e-01 -5.41255951e-01
4.78979886e-01 3.12556028e-01 -5.55106521e-01 3.50531191e-01
-6.35567427e-01 -2.09913850e-01 1.85244873e-01 1.75534480e-03
1.17857301e+00 -8.12988579e-02 1.13860674e-01 -1.18163690e-01
3.67592990e-01 -2.25912288e-01 3.64512891e-01 2.38873988e-01
-9.98374000e-02 7.90006638e-01 8.74515474e-02 -4.70512450e-01
-1.38228667e+00 -1.26450443e+00 -2.70231396e-01 4.49790180e-01
-2.29614705e-01 -2.10440874e-01 -6.60305440e-01 -2.69928396e-01
7.13929534e-02 6.76564515e-01 -4.18603659e-01 2.31463015e-01
-7.25069821e-01 -8.12848806e-01 7.70984173e-01 9.46973562e-01
8.73054564e-01 -1.39115477e+00 -1.89391315e-01 4.16694134e-01
-3.45535874e-01 -6.94920182e-01 -4.21920329e-01 -1.88845113e-01
-5.46544254e-01 -1.07182860e+00 -6.14477098e-01 -9.88173783e-01
7.68330753e-01 2.19481632e-01 1.10260999e+00 2.15966225e-01
-2.86556840e-01 1.97924241e-01 -3.17600697e-01 7.01293871e-02
-1.46524280e-01 5.22831321e-01 -3.47166151e-01 -1.06307939e-01
1.23550616e-01 -1.21870255e+00 -7.15982378e-01 5.24221838e-01
-8.07374656e-01 1.61707550e-01 5.30202329e-01 7.55335510e-01
4.17006731e-01 6.56900525e-01 8.70391846e-01 -4.44825798e-01
6.86653197e-01 -6.46330416e-01 -6.57050252e-01 5.16018718e-02
-4.45101321e-01 -1.45568252e-01 6.64826691e-01 -1.91540495e-02
-8.61324131e-01 -1.28947079e-01 -8.32554936e-01 2.01438501e-01
-5.05343914e-01 6.09081388e-01 -1.04532719e+00 4.73486274e-01
5.29870749e-01 1.30328879e-01 -6.86854064e-01 -4.86494064e-01
4.34251964e-01 8.65946949e-01 4.90477920e-01 -7.39034235e-01
1.26654780e+00 1.14157163e-01 -1.18664429e-02 -1.06652904e+00
-1.38055876e-01 -2.28229910e-01 -6.23322070e-01 -3.02415103e-01
7.54995883e-01 -7.94256985e-01 -6.45563185e-01 5.58570981e-01
-8.74149144e-01 -6.02506638e-01 -4.81908232e-01 -2.72643358e-01
-4.31042731e-01 7.66623393e-02 -3.64012904e-02 -7.03115165e-01
-3.71878058e-01 -1.18789387e+00 1.01674044e+00 2.82800168e-01
-3.02603617e-02 -9.32834446e-01 -1.00528963e-01 5.47719896e-01
6.86929464e-01 6.74282134e-01 1.17799735e+00 1.17640160e-02
-1.03840661e+00 -8.76432210e-02 -2.81467825e-01 1.65517598e-01
2.46922508e-01 9.37804580e-02 -9.12261605e-01 -1.82754807e-02
-6.93233669e-01 -1.32547310e-02 3.92634869e-01 4.07244623e-01
1.33509171e+00 -2.98357457e-01 -2.41194144e-01 6.87207401e-01
1.50976741e+00 4.47096646e-01 1.23859477e+00 6.49041712e-01
1.01552784e+00 4.04485077e-01 3.43130857e-01 6.28538311e-01
6.86574519e-01 6.14222229e-01 4.61232305e-01 -2.26290524e-01
-1.75197691e-01 -7.11632073e-01 -6.58833459e-02 7.42660344e-01
-2.69848555e-02 -5.17800748e-01 -1.20801020e+00 7.47738421e-01
-1.80953169e+00 -1.37349033e+00 -2.79690892e-01 1.75914538e+00
4.02082264e-01 -8.36513415e-02 1.88635170e-01 5.26808262e-01
6.17938042e-01 3.43702197e-01 -4.54946935e-01 1.80728789e-02
-4.11506921e-01 2.66357630e-01 3.47558409e-01 1.80629045e-02
-1.20838165e+00 6.53371513e-01 5.03797483e+00 1.08602059e+00
-9.70033884e-01 -1.03471518e-01 9.06544805e-01 1.71507180e-01
-8.20002615e-01 -2.57424861e-01 -4.70070302e-01 7.58193076e-01
5.31651139e-01 -1.15999460e-01 6.60002589e-01 1.11570120e+00
5.77146351e-01 2.00085342e-01 -8.37730646e-01 1.16257489e+00
-4.34598386e-01 -1.49225581e+00 -2.66986161e-01 4.51275468e-01
1.10606015e+00 -2.09853891e-02 2.88798779e-01 4.60606456e-01
3.03515136e-01 -1.31762171e+00 6.50109649e-01 3.53041202e-01
8.91424596e-01 -1.02690613e+00 6.01530015e-01 2.98789263e-01
-1.84351635e+00 -2.67798662e-01 -1.06156871e-01 1.35122195e-01
4.61287946e-01 5.95382690e-01 -7.63817430e-01 6.91274047e-01
6.55126154e-01 7.07013130e-01 -9.21976745e-01 1.23540807e+00
-3.33265483e-01 6.39384449e-01 -3.22491735e-01 -1.36380300e-01
2.39544839e-01 -3.14129442e-01 1.04771800e-01 6.66681886e-01
5.16883910e-01 -1.06198013e-01 6.61422238e-02 1.02062297e+00
-1.39235646e-01 1.28688350e-01 -7.90198147e-01 -1.00608125e-01
8.18129301e-01 1.07191896e+00 -7.14251697e-01 1.37599692e-01
-2.16041163e-01 5.79349935e-01 4.32616472e-02 4.54162180e-01
-1.14214563e+00 -1.34137139e-01 5.01890242e-01 5.87852180e-01
3.67055625e-01 -3.60325873e-01 -6.81197643e-01 -8.02776337e-01
3.77096720e-02 -7.55995095e-01 -2.16794554e-02 -8.54266644e-01
-1.45243597e+00 4.11904514e-01 -1.05829544e-01 -1.30829883e+00
3.92901748e-02 -2.53387958e-01 -1.01150787e+00 6.75451398e-01
-1.05201244e+00 -1.62828934e+00 -7.23234296e-01 3.32694501e-01
5.85776985e-01 -3.56335253e-01 7.12769806e-01 9.98084247e-01
-8.76532197e-01 6.88653350e-01 3.43427151e-01 6.21897578e-01
1.55867904e-01 -1.04593837e+00 1.08454740e+00 7.58339882e-01
-7.28005543e-03 4.06556606e-01 1.46236151e-01 -5.84083438e-01
-1.18309152e+00 -1.59960282e+00 7.51562178e-01 -3.07064295e-01
4.77946728e-01 -4.31921512e-01 -4.60544556e-01 5.91435969e-01
1.07241310e-02 -4.02267903e-01 6.83482528e-01 -8.69584531e-02
1.01011492e-01 -3.06403011e-01 -1.20616126e+00 7.97838390e-01
1.47650731e+00 -3.83008152e-01 -1.52394623e-01 1.77846342e-01
4.64863151e-01 -4.60437573e-02 -8.97442341e-01 3.71484518e-01
4.22605336e-01 -9.91880596e-01 1.17465961e+00 5.27714789e-02
5.32775342e-01 -5.23225904e-01 -4.56797481e-01 -1.09321010e+00
-5.75507283e-01 -1.52352244e-01 2.69636244e-01 1.88270450e+00
3.57176781e-01 -6.10394359e-01 8.85627449e-01 5.28363824e-01
-3.10080320e-01 -9.82850194e-01 -8.80421102e-01 -5.76237440e-01
6.24735355e-02 -6.81503654e-01 1.75638163e+00 6.99713171e-01
-7.61205375e-01 3.82665277e-01 -1.27111837e-01 8.90084133e-02
4.59421784e-01 2.08070695e-01 1.03144121e+00 -1.09682083e+00
6.64917678e-02 -8.69006276e-01 -6.44405544e-01 -1.11940789e+00
9.25782043e-03 -9.97914314e-01 -1.16009548e-01 -2.18748617e+00
-2.29537562e-01 -1.02765131e+00 1.89045176e-01 3.48968416e-01
3.35833758e-01 2.23433703e-01 -1.88004419e-01 -1.62824169e-01
-1.86739936e-01 1.08855700e+00 1.30220282e+00 -7.35512793e-01
-2.04358846e-01 1.27886236e-01 -9.41784441e-01 3.55432361e-01
9.77113008e-01 1.34698287e-01 -7.11552560e-01 -5.25019467e-01
2.44565994e-01 -3.49028021e-01 5.66699803e-01 -1.24577951e+00
1.16075143e-01 -8.41690525e-02 5.27301252e-01 -1.24727166e+00
-4.14680690e-02 -1.02183473e+00 7.60439098e-01 2.18113020e-01
1.67631313e-01 2.36639753e-01 -8.23315606e-02 1.85361370e-01
3.11487298e-02 2.04025671e-01 5.09458482e-01 1.90136954e-01
-1.05635345e+00 6.08373761e-01 -3.81512016e-01 2.02593207e-02
9.74335611e-01 -4.36799943e-01 -3.46746236e-01 -2.95651704e-01
-5.28389156e-01 3.92006785e-01 5.81078231e-01 5.61513126e-01
7.53530204e-01 -2.03501439e+00 -5.35903156e-01 5.45487106e-01
2.46485516e-01 5.40175021e-01 6.18504703e-01 3.39231044e-01
-7.63573468e-01 4.50039744e-01 -2.16293067e-01 -4.17035162e-01
-8.22852850e-01 5.05581617e-01 2.05123410e-01 -9.80982557e-02
-7.63697684e-01 4.31939989e-01 -4.20462824e-02 -6.15492404e-01
1.30164668e-01 -3.35445195e-01 -2.18488976e-01 2.83882152e-02
2.07733497e-01 7.88035572e-01 1.17043830e-01 -9.53114331e-01
-3.43250722e-01 4.69456643e-01 6.15197837e-01 1.89648777e-01
1.72639310e+00 -7.83617571e-02 1.11246809e-01 -9.13918093e-02
1.12700284e+00 -1.62609398e-01 -8.93473804e-01 2.54684538e-01
-1.22598752e-01 -5.09539485e-01 -6.08858764e-02 -6.99928224e-01
-1.35220492e+00 6.72174752e-01 8.20209146e-01 1.42169774e-01
1.19411421e+00 -1.05244838e-01 1.10274112e+00 2.16629744e-01
8.54044795e-01 -9.57317889e-01 8.75325054e-02 2.94621617e-01
1.17543149e+00 -1.08107162e+00 -3.34624387e-02 -2.17958525e-01
-3.83793890e-01 6.19750857e-01 5.92345893e-01 4.81836088e-02
7.77769387e-01 3.87227118e-01 -1.75330460e-01 -1.86083928e-01
-4.00434360e-02 -4.38048691e-01 1.78321376e-01 1.00560021e+00
3.31876397e-01 5.78106284e-01 1.40996218e-01 7.84522891e-01
-7.91450083e-01 -3.04294288e-01 2.43598670e-02 7.72808373e-01
-1.85009122e-01 -1.42932642e+00 -4.05004561e-01 5.82227111e-01
4.69536066e-01 -2.27882609e-01 -1.92992371e-02 6.50314033e-01
4.28800941e-01 1.09772897e+00 -1.15715861e-01 -5.05602956e-01
5.58128357e-01 -3.91984582e-01 2.24561885e-01 -2.74071246e-01
-3.98314625e-01 -1.67057216e-01 5.16978949e-02 -2.52040178e-01
-8.73470604e-02 -7.55650401e-01 -1.11742651e+00 -9.67595994e-01
-6.69955537e-02 -7.35626668e-02 4.26859379e-01 5.47194421e-01
3.43094349e-01 7.39746571e-01 8.94810259e-01 -8.63927543e-01
6.60683680e-03 -7.93825686e-01 -6.56959832e-01 3.18246156e-01
-1.65151536e-01 -7.75317490e-01 8.51503536e-02 -2.42366090e-01] | [8.500718116760254, -1.6491787433624268] |
d5ed7461-542c-4e41-92eb-20fb5ca18a67 | efficienthrnet-efficient-scaling-for | 2007.08090 | null | https://arxiv.org/abs/2007.08090v2 | https://arxiv.org/pdf/2007.08090v2.pdf | EfficientHRNet: Efficient Scaling for Lightweight High-Resolution Multi-Person Pose Estimation | There is an increasing demand for lightweight multi-person pose estimation for many emerging smart IoT applications. However, the existing algorithms tend to have large model sizes and intense computational requirements, making them ill-suited for real-time applications and deployment on resource-constrained hardware. Lightweight and real-time approaches are exceedingly rare and come at the cost of inferior accuracy. In this paper, we present EfficientHRNet, a family of lightweight multi-person human pose estimators that are able to perform in real-time on resource-constrained devices. By unifying recent advances in model scaling with high-resolution feature representations, EfficientHRNet creates highly accurate models while reducing computation enough to achieve real-time performance. The largest model is able to come within 4.4% accuracy of the current state-of-the-art, while having 1/3 the model size and 1/6 the computation, achieving 23 FPS on Nvidia Jetson Xavier. Compared to the top real-time approach, EfficientHRNet increases accuracy by 22% while achieving similar FPS with 1/3 the power. At every level, EfficientHRNet proves to be more computationally efficient than other bottom-up 2D human pose estimation approaches, while achieving highly competitive accuracy. | ['Aneri Sheth', 'Hamed Tabkhi', 'Steven Furgurson', 'Christopher Neff'] | 2020-07-16 | null | null | null | null | ['2d-human-pose-estimation'] | ['computer-vision'] | [-1.07149504e-01 -9.02232677e-02 6.01463467e-02 -9.69753712e-02
-5.56382000e-01 2.44722161e-02 1.35674670e-01 1.94657952e-01
-6.91709638e-01 4.85035956e-01 -4.59441766e-02 1.52484506e-01
1.41795203e-01 -7.94101179e-01 -4.46899205e-01 -8.58862773e-02
-1.10402167e-01 8.01661015e-01 5.09835601e-01 -1.24432303e-01
-4.00134742e-01 5.33923805e-01 -2.10160279e+00 -4.82243933e-02
4.51556534e-01 1.27499998e+00 -2.01907426e-01 7.48358667e-01
5.11536658e-01 1.49437800e-01 -7.02725589e-01 -4.51762378e-01
2.66121924e-01 1.06674790e-01 -2.73585141e-01 -3.40236604e-01
9.67170537e-01 -6.36264503e-01 -4.11767393e-01 4.51821834e-01
1.14617383e+00 -1.76590934e-01 -1.44663537e-02 -1.38585198e+00
2.34584004e-01 2.29535978e-02 -4.14774388e-01 5.78689054e-02
9.32318926e-01 1.29858345e-01 4.80054349e-01 -7.05457091e-01
4.03596193e-01 1.35664678e+00 1.17662609e+00 6.33137047e-01
-8.64529550e-01 -7.14808643e-01 -1.24880180e-01 3.35406885e-03
-1.69419181e+00 -4.31440592e-01 2.54442006e-01 -7.47388527e-02
1.34202766e+00 4.11674023e-01 1.26568699e+00 1.04344535e+00
3.83527070e-01 5.66975296e-01 8.78591239e-01 -7.52353519e-02
1.99452132e-01 -2.51537442e-01 -7.86894262e-02 6.48298085e-01
8.99403572e-01 4.85527702e-02 -8.24799299e-01 -3.32837790e-01
1.00821733e+00 2.49652490e-02 1.13512389e-01 -3.28538418e-01
-1.15907049e+00 5.38394272e-01 5.05512238e-01 -3.25989202e-02
-4.47225809e-01 6.10552132e-01 6.24421358e-01 -2.73893714e-01
3.94075066e-01 1.97715849e-01 -4.52163100e-01 -7.02385783e-01
-1.07189429e+00 7.86538959e-01 6.95470512e-01 1.02578020e+00
1.49193630e-01 5.65816946e-02 4.23995145e-02 3.82592350e-01
2.73043811e-01 9.86175418e-01 3.24641705e-01 -6.86088204e-01
3.60976994e-01 6.62953317e-01 3.00116122e-01 -1.05320525e+00
-1.10894144e+00 -4.88978744e-01 -8.11535537e-01 1.16936676e-01
5.47775269e-01 -1.05521806e-01 -6.07923865e-01 1.27472138e+00
7.63665557e-01 -1.56595930e-01 -5.10568023e-01 9.45733070e-01
9.15491939e-01 3.92942190e-01 3.02260756e-01 -3.60778384e-02
1.81819129e+00 -7.11236656e-01 -4.84415978e-01 -4.37470943e-01
4.90454227e-01 -7.36557722e-01 8.38627040e-01 6.45066619e-01
-1.11623478e+00 -7.40524530e-01 -1.22264040e+00 -1.52745709e-01
-3.96895967e-02 2.00983375e-01 1.04852486e+00 1.20797324e+00
-8.15470755e-01 6.93463504e-01 -1.19067276e+00 -5.56102276e-01
3.27304929e-01 7.49377668e-01 -3.03335220e-01 9.98126492e-02
-7.99316883e-01 8.99797022e-01 8.09796974e-02 -6.99156001e-02
-2.05264688e-01 -8.70753706e-01 -7.56921589e-01 -1.39906049e-01
2.72123963e-01 -1.00246501e+00 1.21225846e+00 -4.03438836e-01
-1.52953565e+00 7.03061879e-01 -6.02101460e-02 -6.67027295e-01
9.40894842e-01 -8.56570899e-01 -6.94830060e-01 2.66829610e-01
-4.31626737e-02 7.53512084e-01 7.36815214e-01 -5.53272843e-01
-5.40906608e-01 -5.87527394e-01 -2.52889208e-02 2.00595364e-01
-3.82549793e-01 -4.68564481e-02 -4.68130618e-01 -2.75699675e-01
8.02938268e-02 -1.33519781e+00 -3.62508148e-01 5.17969310e-01
2.30038017e-02 -2.21307397e-01 9.51962411e-01 -5.72783589e-01
1.16037905e+00 -1.77268517e+00 -2.51407325e-01 -5.11257723e-02
4.19760734e-01 5.86668372e-01 3.64165574e-01 1.49766192e-01
4.16876912e-01 -4.02972192e-01 3.17033738e-01 -5.33703029e-01
4.52371687e-02 6.93528680e-04 3.94014508e-01 7.40828753e-01
-3.04155856e-01 8.91978085e-01 -5.87093472e-01 -4.80629086e-01
6.08832479e-01 1.07159507e+00 -6.87736750e-01 -1.03098825e-01
2.69926757e-01 2.02299938e-01 -3.37932229e-01 6.88144386e-01
7.65557170e-01 -2.47956201e-01 1.24466419e-01 -4.68615144e-01
-5.69848046e-02 6.83590770e-02 -1.69794714e+00 1.58622205e+00
-4.13686246e-01 4.07889485e-01 2.15817359e-03 -3.95097792e-01
8.29872608e-01 7.32427463e-02 6.93834543e-01 -7.89025366e-01
4.98343945e-01 2.98221797e-01 -1.34780467e-01 -5.48737086e-02
7.75294244e-01 8.33501741e-02 -2.79423296e-01 1.84698924e-01
-2.40174755e-01 -1.15238711e-01 1.18559659e-01 -2.97588985e-02
9.63287532e-01 2.39141315e-01 5.21473885e-01 -2.97018588e-01
1.60841689e-01 -2.94118762e-01 3.98121178e-01 5.77550113e-01
-4.54542011e-01 5.40001392e-01 -2.05185324e-01 -1.03908563e+00
-1.12945485e+00 -1.11142302e+00 -2.02538386e-01 9.37651992e-01
1.93557769e-01 -9.06483710e-01 -9.16992426e-01 -2.00132608e-01
2.26350799e-01 -3.79249975e-02 -2.55303860e-01 8.02380890e-02
-7.45828390e-01 -7.03914821e-01 7.89597869e-01 8.69036138e-01
8.38887751e-01 -6.26343012e-01 -1.76041675e+00 3.85852933e-01
4.36164998e-02 -1.52061880e+00 -1.02622487e-01 -3.82746458e-01
-9.14399266e-01 -1.02935696e+00 -7.42088199e-01 -2.87279576e-01
3.66676271e-01 4.91705313e-02 1.29196894e+00 1.40486822e-01
-8.39056909e-01 3.08312982e-01 -2.39047915e-01 -5.67413926e-01
3.89452189e-01 2.19009191e-01 6.90426648e-01 -4.34096932e-01
3.63217920e-01 -3.39038312e-01 -1.09880066e+00 2.65431553e-01
-1.75367936e-01 1.45180777e-01 3.19107503e-01 3.93327028e-01
6.38884544e-01 -1.63935736e-01 1.79529399e-01 -3.99948895e-01
7.63674974e-02 1.58152685e-01 -5.42676568e-01 -2.03983277e-01
-4.93256480e-01 -2.45675892e-01 5.81495106e-01 -5.17626941e-01
-6.73803687e-01 4.02070791e-01 -3.97182792e-01 -3.37902084e-02
-1.58404652e-02 -2.38645539e-01 1.26168042e-01 -3.05102944e-01
7.60161519e-01 -1.52520776e-01 4.70433570e-02 -4.65946436e-01
1.36852235e-01 6.05925918e-01 4.75433558e-01 -2.93787658e-01
4.94649947e-01 8.27690780e-01 2.50682145e-01 -1.12654626e+00
-7.09664285e-01 -3.95542353e-01 -7.59224296e-01 -3.80173445e-01
6.53288722e-01 -1.17502570e+00 -1.15769279e+00 6.06678069e-01
-8.53131533e-01 -2.52639242e-02 -3.03924203e-01 5.08762658e-01
-5.18231750e-01 1.65172696e-01 -4.10641193e-01 -9.18559015e-01
-8.04690123e-01 -1.00340533e+00 1.54874432e+00 2.06656501e-01
-6.62221372e-01 -3.95811826e-01 -1.81903049e-01 4.68046635e-01
5.93599856e-01 4.82798934e-01 -3.76770832e-02 -3.34800519e-02
-1.31388381e-01 -7.06217647e-01 -1.91925421e-01 -2.77650386e-01
-3.51236671e-01 -2.76276171e-01 -9.04957950e-01 -6.75211489e-01
-3.98241431e-01 -1.97617814e-01 1.38459921e-01 4.57899660e-01
9.09779727e-01 1.02751128e-01 -6.65052533e-01 5.59473157e-01
1.25738156e+00 -2.79257715e-01 4.64058042e-01 2.05787510e-01
7.27228940e-01 1.44407660e-01 5.53754091e-01 9.19834375e-01
5.17821908e-01 1.31226385e+00 2.21608475e-01 -1.17209084e-01
-3.85708898e-01 -2.35934630e-01 -6.30847691e-03 6.16450906e-01
-4.70321685e-01 1.83136955e-01 -9.66852844e-01 -3.07768472e-02
-1.85696828e+00 -7.60498822e-01 -4.01165903e-01 2.44599605e+00
2.37062693e-01 4.00049359e-01 7.82341599e-01 3.71794432e-01
4.60320204e-01 3.85710038e-02 -4.78791624e-01 -3.24657321e-01
1.92337528e-01 5.27696133e-01 6.12473965e-01 6.29563257e-02
-1.30346990e+00 6.64587438e-01 6.84564877e+00 6.03371441e-01
-9.61902082e-01 2.85482407e-01 1.91578180e-01 -8.36158991e-01
5.53644478e-01 -3.88097286e-01 -1.28515339e+00 4.44579303e-01
1.25663781e+00 -5.73019683e-02 5.09268120e-02 1.31852937e+00
1.00911468e-01 -4.17949080e-01 -1.04699492e+00 1.67045605e+00
2.32693136e-01 -9.19100702e-01 -8.12897235e-02 2.07813650e-01
2.61784792e-01 -1.22708470e-01 -3.25966597e-01 3.37295622e-01
-3.35312814e-01 -1.10082126e+00 8.11548829e-01 2.61625677e-01
1.05484271e+00 -1.16756713e+00 8.78955007e-01 3.73416364e-01
-1.80481136e+00 4.73052971e-02 -4.65630144e-01 -6.24256194e-01
4.94042724e-01 7.12776184e-01 -3.66079777e-01 2.57038862e-01
1.15034997e+00 4.08984989e-01 -7.20075548e-01 9.44923401e-01
1.66279614e-01 3.78197283e-02 -9.41160679e-01 -5.14343202e-01
-3.02713126e-01 5.26406288e-01 2.24562958e-01 1.13577163e+00
4.21211958e-01 1.53491557e-01 4.48133558e-01 2.02107504e-02
1.26680359e-01 1.10671595e-01 -3.85798782e-01 5.64764977e-01
5.45408607e-01 1.22623575e+00 -8.80438209e-01 -4.89305168e-01
-2.76117831e-01 1.13555694e+00 2.15843514e-01 -5.28720975e-01
-1.17893159e+00 -3.03042322e-01 8.19241524e-01 6.96483076e-01
1.03010327e-01 -6.36235416e-01 -3.20150703e-01 -9.36914623e-01
2.70840108e-01 -6.70899034e-01 2.02810943e-01 -3.61362219e-01
-8.44462097e-01 6.89239740e-01 1.49523523e-02 -1.37693501e+00
-2.34517843e-01 -7.87324071e-01 3.70079838e-02 2.94718325e-01
-8.09755623e-01 -1.30005491e+00 -7.86192536e-01 4.56169158e-01
3.96922767e-01 2.33629495e-01 1.16713202e+00 6.19338155e-01
-4.05239165e-01 9.76068556e-01 -3.48829359e-01 -1.86384216e-01
3.26087654e-01 -7.27912188e-01 7.73388684e-01 5.84581435e-01
-1.60777375e-01 5.47811627e-01 7.41950154e-01 -7.47071743e-01
-1.82412076e+00 -7.44130611e-01 7.80299962e-01 -5.99672496e-01
1.27674013e-01 -7.79451847e-01 -3.43711883e-01 4.13129389e-01
-4.30901051e-01 5.49317598e-01 6.10348523e-01 2.38200203e-01
-3.00531358e-01 -2.54927754e-01 -1.44828856e+00 6.51210725e-01
1.46239769e+00 -9.04535726e-02 -2.66053557e-01 4.19392765e-01
3.60351980e-01 -9.25834179e-01 -1.14091253e+00 5.00738442e-01
1.24425912e+00 -1.05057573e+00 1.31921506e+00 6.76569864e-02
-2.12889150e-01 -1.10685408e-01 1.58532578e-02 -6.71542466e-01
-4.85450149e-01 -5.93800604e-01 -5.45033753e-01 6.75815284e-01
-1.24001138e-01 -7.16168523e-01 1.18791306e+00 8.11411083e-01
3.25841069e-01 -9.58955586e-01 -1.48492491e+00 -1.03569782e+00
-3.49549025e-01 -6.74012721e-01 6.48118615e-01 1.84019372e-01
6.84709400e-02 2.33400255e-01 -5.06015122e-01 -3.54058295e-02
9.47794855e-01 -2.29432449e-01 1.15052938e+00 -1.43265963e+00
-9.94751006e-02 -1.78914472e-01 -9.97954011e-01 -9.92640316e-01
-4.17660296e-01 -2.91816801e-01 -2.66604304e-01 -1.45506966e+00
2.87119020e-02 -3.98567587e-01 3.34464401e-01 3.44019443e-01
2.76493747e-02 8.75984669e-01 4.89255816e-01 -2.96352692e-02
-8.55589867e-01 4.17438775e-01 7.79380918e-01 2.40241855e-01
-1.09887771e-01 -3.69829349e-02 -2.68524408e-01 9.97853994e-01
5.58293223e-01 -2.05871806e-01 -1.39617950e-01 -2.97293633e-01
1.89621329e-01 -2.38752350e-01 5.18004477e-01 -1.96667898e+00
3.15267295e-01 4.51710582e-01 9.51748192e-01 -6.22617245e-01
9.84020293e-01 -7.77096629e-01 4.96651441e-01 1.03843272e+00
4.27792192e-01 3.90562087e-01 3.87048364e-01 1.15873545e-01
3.18577290e-01 3.45194548e-01 7.98246324e-01 -1.60753548e-01
-7.12926924e-01 4.40560788e-01 -1.99733406e-01 -7.51001469e-04
1.19491911e+00 -5.31326294e-01 -1.96574226e-01 -1.12919748e-01
-5.35017669e-01 -7.55429789e-02 4.06308591e-01 5.69338024e-01
5.78490674e-01 -1.38823700e+00 -5.57862401e-01 3.21349651e-01
-4.82942127e-02 -1.41858652e-01 5.69693446e-01 5.36338985e-01
-9.42339480e-01 5.87018669e-01 -3.53546560e-01 -7.61234760e-01
-1.59637129e+00 1.26118228e-01 1.80676788e-01 -3.09812307e-01
-1.04170954e+00 6.89207733e-01 -4.15667415e-01 -2.57470161e-01
1.25059023e-01 -3.69248129e-02 1.44869104e-01 -2.69148126e-02
9.38176274e-01 9.83273447e-01 4.07128036e-01 -8.12540829e-01
-9.17735636e-01 9.23219740e-01 2.21054524e-01 8.04379508e-02
1.10296500e+00 1.01509772e-01 4.42736745e-01 -8.37969966e-03
8.81988049e-01 -1.05118178e-01 -1.25732028e+00 3.33709866e-01
-3.82968932e-01 -5.98367214e-01 5.53165078e-02 -5.04572034e-01
-8.29658687e-01 5.91411769e-01 1.18757868e+00 -1.24813341e-01
8.36427450e-01 -1.63300246e-01 1.31261086e+00 1.24659397e-01
1.28137183e+00 -1.25671530e+00 8.64779204e-02 2.10873440e-01
6.11088395e-01 -1.00142467e+00 8.19903433e-01 -7.38058209e-01
-2.00005919e-01 9.49703276e-01 8.39510024e-01 -4.62933958e-01
6.99893713e-01 6.88243389e-01 -1.78773060e-01 -2.08380848e-01
-2.28243262e-01 4.30853292e-02 2.62178451e-01 7.52605796e-01
4.92581666e-01 4.40515727e-01 -5.08732498e-01 5.14603674e-01
-8.07753623e-01 1.16797648e-01 4.08340245e-02 1.09402788e+00
-3.12229514e-01 -8.73522401e-01 -5.21465957e-01 5.15786648e-01
-4.46932673e-01 3.09985012e-01 1.52883828e-01 1.05231929e+00
1.71425939e-01 9.02007341e-01 3.14218014e-01 -3.63313109e-01
8.19604397e-01 -2.68768877e-01 8.46678376e-01 -3.91334772e-01
-8.53869736e-01 -1.35277420e-01 2.00173616e-01 -1.22610331e+00
-1.51176587e-01 -7.78472781e-01 -1.13185167e+00 -7.93682158e-01
-5.38394712e-02 -4.15017575e-01 7.71683872e-01 6.79060698e-01
7.36562788e-01 3.98375243e-01 -5.72756752e-02 -1.49561656e+00
-3.52276862e-01 -7.96736836e-01 -4.09564346e-01 3.08554560e-01
-1.76746339e-01 -1.01442802e+00 1.53951347e-01 -4.33026761e-01] | [7.161841869354248, -0.758607029914856] |
2bda02fc-dd5a-4a49-9604-da153b41e8e4 | cross-view-action-recognition-via-contrastive | 2305.01733 | null | https://arxiv.org/abs/2305.01733v1 | https://arxiv.org/pdf/2305.01733v1.pdf | Cross-view Action Recognition via Contrastive View-invariant Representation | Cross view action recognition (CVAR) seeks to recognize a human action when observed from a previously unseen viewpoint. This is a challenging problem since the appearance of an action changes significantly with the viewpoint. Applications of CVAR include surveillance and monitoring of assisted living facilities where is not practical or feasible to collect large amounts of training data when adding a new camera. We present a simple yet efficient CVAR framework to learn invariant features from either RGB videos, 3D skeleton data, or both. The proposed approach outperforms the current state-of-the-art achieving similar levels of performance across input modalities: 99.4% (RGB) and 99.9% (3D skeletons), 99.4% (RGB) and 99.9% (3D Skeletons), 97.3% (RGB), and 99.2% (3D skeletons), and 84.4%(RGB) for the N-UCLA, NTU-RGB+D 60, NTU-RGB+D 120, and UWA3DII datasets, respectively. | ['Mario Sznaier', 'Octavia Camps', 'Balaji Sundareshan', 'Dan Luo', 'Yuexi Zhang'] | 2023-05-02 | null | null | null | null | ['action-recognition-in-videos'] | ['computer-vision'] | [ 2.49998122e-01 -2.11285874e-01 -1.80809591e-02 -8.42368305e-02
-5.96774161e-01 -4.25081700e-01 4.38526213e-01 -2.66947955e-01
-4.22894448e-01 5.55588663e-01 3.08938950e-01 2.13258892e-01
1.62781298e-01 -3.87896657e-01 -6.24507904e-01 -8.05346668e-01
-3.67632657e-02 2.88975894e-01 4.32215661e-01 -1.15360796e-01
-3.40557806e-02 8.85902584e-01 -1.67403865e+00 4.76684213e-01
4.28776294e-02 1.16207027e+00 -2.32044667e-01 1.10091031e+00
4.92794663e-01 7.06847966e-01 -4.60772991e-01 -2.33159810e-01
6.21800840e-01 -2.65574068e-01 -6.37988567e-01 5.72531819e-01
6.91855550e-01 -7.81524777e-01 -6.43802643e-01 5.63116372e-01
3.98387492e-01 1.66289970e-01 6.10795677e-01 -1.43915951e+00
-3.79461199e-01 -4.12651956e-01 -1.25262725e+00 2.77799040e-01
9.15064037e-01 2.13603660e-01 4.36251074e-01 -6.25512958e-01
6.14773512e-01 1.27265728e+00 3.02542597e-01 6.76032960e-01
-7.29127169e-01 -3.23194414e-01 1.60387978e-01 1.92247540e-01
-1.19389617e+00 -6.29157007e-01 5.69262445e-01 -4.39020634e-01
1.04436457e+00 3.47955346e-01 9.85502601e-01 1.29430318e+00
1.76092431e-01 1.11165631e+00 9.70972776e-01 -4.59401935e-01
1.84704080e-01 -4.17546511e-01 -4.40407656e-02 9.31784630e-01
2.03322366e-01 -9.23983380e-03 -6.59770072e-01 -6.48740306e-02
1.05853009e+00 4.88366157e-01 -3.70903909e-01 -5.37768841e-01
-1.14939737e+00 3.53993922e-01 1.46281004e-01 2.61728931e-02
-4.57799643e-01 2.71494299e-01 3.25197428e-01 1.64345745e-02
2.70733476e-01 -6.68414608e-02 -6.64663851e-01 -5.67341208e-01
-6.11597419e-01 -1.88927844e-01 2.10556462e-01 1.17396438e+00
2.67795801e-01 1.56343639e-01 1.11195087e-01 6.36886716e-01
1.69709057e-01 9.52111125e-01 4.13154393e-01 -1.35552537e+00
4.46579039e-01 8.21798980e-01 1.95369199e-01 -8.11638176e-01
-6.65395677e-01 1.16934083e-01 -7.49595761e-01 5.85144043e-01
5.83594859e-01 -4.66628857e-02 -1.06209838e+00 1.22417688e+00
6.36816978e-01 -1.28908172e-01 2.93284237e-01 7.74447680e-01
8.66615832e-01 4.13838536e-01 -2.82811552e-01 -2.60818839e-01
1.24808216e+00 -9.60408986e-01 -5.42223930e-01 -4.23703432e-01
3.39649260e-01 -5.99288642e-01 8.44004214e-01 7.03782976e-01
-1.27322221e+00 -6.35715544e-01 -8.71459365e-01 4.31670807e-02
-1.21574402e-01 3.87491316e-01 4.34638619e-01 5.35851836e-01
-8.15695047e-01 2.35368267e-01 -1.23117840e+00 -7.16376364e-01
4.44194198e-01 2.68439949e-01 -1.05192137e+00 -6.70160472e-01
-3.98375899e-01 6.76419199e-01 8.83265957e-03 -3.30891237e-02
-8.50457489e-01 -1.86342165e-01 -8.57367933e-01 -4.64871228e-01
4.24119771e-01 -4.97673810e-01 1.01889527e+00 -7.81555712e-01
-1.48003423e+00 1.01874042e+00 -5.84791787e-02 -4.94808108e-02
7.95181215e-01 -6.02688789e-01 -5.26864409e-01 7.17777967e-01
-9.79415700e-02 5.59133589e-01 7.28626788e-01 -9.73463714e-01
-8.75233531e-01 -1.03077817e+00 1.76295131e-01 3.32555264e-01
8.46791714e-02 6.60424829e-02 -6.03191376e-01 -3.92583013e-01
4.25048620e-01 -1.15825117e+00 6.75711110e-02 6.59364522e-01
-1.13550872e-01 9.66711901e-03 1.02509379e+00 -8.46654296e-01
4.46137309e-01 -2.19205427e+00 2.25982010e-01 -3.05581272e-01
-9.46492553e-02 5.36080658e-01 8.69982466e-02 1.61336511e-01
-1.83196720e-02 -4.00476545e-01 1.79003939e-01 -1.14895612e-01
-4.27516580e-01 3.60277236e-01 4.09727275e-01 9.40140128e-01
-1.57414868e-01 6.05559766e-01 -9.21330452e-01 -5.19562244e-01
6.12485349e-01 5.60124397e-01 -3.16970140e-01 1.73609823e-01
3.29350501e-01 5.02472460e-01 -3.79785836e-01 1.20019281e+00
4.04792339e-01 -2.00550318e-01 5.39090633e-02 -3.39166224e-01
1.96025357e-01 -4.32421595e-01 -1.31102753e+00 1.68436468e+00
-3.80309857e-02 7.11633742e-01 -4.25222740e-02 -9.13992703e-01
8.40302944e-01 4.10485357e-01 9.17738914e-01 -6.69994295e-01
3.85887409e-03 -1.56365305e-01 -2.61681169e-01 -8.26484680e-01
2.20556960e-01 -1.76735893e-02 1.49345752e-02 2.76496708e-01
2.90307384e-02 2.59992033e-01 8.10198113e-02 1.01854941e-02
1.47415519e+00 3.93720359e-01 8.13592255e-01 2.69804448e-01
4.97351766e-01 -2.78707534e-01 4.30472314e-01 2.74162322e-01
-6.25973165e-01 8.57821107e-01 2.12837428e-01 -6.79242134e-01
-8.14507961e-01 -1.20160592e+00 1.75141662e-01 9.24174011e-01
5.96664473e-02 -8.02132189e-02 -5.03720105e-01 -7.68479526e-01
-1.31564476e-02 3.30284774e-01 -7.47636199e-01 -1.42252833e-01
-6.07477546e-01 -1.90301552e-01 3.23308289e-01 1.13554883e+00
7.65358508e-01 -9.18544650e-01 -1.35405755e+00 -2.04106182e-01
-2.99453735e-01 -1.47828996e+00 -4.19620305e-01 -1.29579037e-01
-1.08824265e+00 -1.40778315e+00 -8.23156714e-01 -1.91442356e-01
7.55222797e-01 6.38800621e-01 7.53544033e-01 -1.66679174e-01
-6.08497083e-01 1.05591953e+00 -6.11139774e-01 -4.39597033e-02
2.16385629e-02 -6.74175620e-01 3.66526037e-01 2.20522061e-01
1.57983959e-01 -2.95485735e-01 -6.24069512e-01 5.80218613e-01
-7.71909416e-01 -1.48241976e-02 5.40604651e-01 4.82661992e-01
8.01403642e-01 -1.31158099e-01 -2.25587144e-01 -2.65744239e-01
-3.04350078e-01 -1.93492845e-02 -3.88379842e-01 5.75487375e-01
-1.18586838e-01 -3.60893309e-01 3.01514238e-01 -5.49117088e-01
-8.79319131e-01 5.88358223e-01 2.47726098e-01 -9.57140028e-01
-6.77488029e-01 -1.92837045e-01 -2.78391361e-01 1.28561839e-01
5.63569963e-01 1.84130311e-01 5.73446713e-02 -4.25838262e-01
2.58462518e-01 8.11580777e-01 7.75411487e-01 -1.76750481e-01
5.06555676e-01 1.08892763e+00 9.19662118e-02 -1.09054458e+00
-7.53763258e-01 -8.34375203e-01 -1.33647680e+00 -5.24793446e-01
1.11930931e+00 -9.52898324e-01 -4.79203552e-01 7.22369552e-01
-8.62709582e-01 -1.80390209e-01 -4.01851207e-01 6.59043312e-01
-8.63385677e-01 6.86538398e-01 -2.52672136e-01 -9.43372965e-01
-3.71525317e-01 -9.86616671e-01 1.49266648e+00 2.72881985e-01
-3.69087785e-01 -6.72626972e-01 -1.16312839e-01 9.46039498e-01
-1.75609160e-02 9.60984528e-01 1.94183022e-01 -1.37101725e-01
-2.33272165e-01 -5.23141623e-01 -7.24229217e-02 4.62733716e-01
5.35700619e-01 2.24755242e-01 -9.11183298e-01 -2.85727739e-01
-2.46454149e-01 -4.45944071e-01 4.26368743e-01 3.57145101e-01
7.60186791e-01 -1.91625178e-01 -2.75673389e-01 5.25027692e-01
1.04901850e+00 5.76699734e-01 7.24394500e-01 2.44827971e-01
6.25956714e-01 2.97868669e-01 7.16806531e-01 8.49709094e-01
1.67183593e-01 8.95613909e-01 4.91475552e-01 -1.31870061e-01
-2.28481576e-01 1.09228618e-01 6.77504241e-01 2.19502300e-01
-5.96887231e-01 -5.51120862e-02 -9.66489673e-01 3.75802338e-01
-1.82585645e+00 -1.17944980e+00 -2.38032684e-01 2.12672591e+00
3.14666539e-01 -1.45554453e-01 4.07274455e-01 3.74981374e-01
5.10138333e-01 -7.11925626e-02 -7.20091105e-01 -6.31201491e-02
-8.56404100e-03 1.10932775e-01 5.08783817e-01 1.88540865e-03
-1.28786778e+00 4.70138729e-01 5.67294598e+00 1.11949712e-01
-7.99702525e-01 -8.26977193e-02 3.75892460e-01 -5.04675865e-01
6.74306214e-01 -5.81459641e-01 -5.64874172e-01 2.67593294e-01
5.51785350e-01 3.46302122e-01 1.80435181e-01 1.11924767e+00
2.49120090e-02 -5.81781685e-01 -1.29404759e+00 1.44258893e+00
5.09157300e-01 -7.94561505e-01 -4.17733431e-01 3.01520284e-02
4.79018062e-01 1.21479675e-01 -1.28036797e-01 5.70460744e-02
1.54933184e-01 -8.36896777e-01 5.53854108e-01 6.35557234e-01
9.34553504e-01 -6.08102977e-01 6.09325767e-01 3.88564378e-01
-1.13963878e+00 4.94954064e-02 -2.01707706e-01 4.91471663e-02
1.66649297e-01 3.48004960e-02 -3.28507781e-01 5.45173645e-01
1.21154308e+00 1.03906310e+00 -7.29703665e-01 7.79871941e-01
-3.21538121e-01 7.85555094e-02 -3.61105621e-01 3.50685328e-01
5.63518032e-02 6.30707145e-02 2.11629540e-01 9.43755686e-01
2.57343858e-01 6.63674951e-01 3.34920794e-01 -2.29019504e-02
2.40691543e-01 -2.87132651e-01 -6.85626507e-01 3.23703140e-01
-3.65300894e-01 1.01035619e+00 -6.71765745e-01 -2.38260761e-01
-5.95323443e-01 1.48913300e+00 9.97729152e-02 2.32868955e-01
-8.40788245e-01 5.11179268e-02 7.16979563e-01 -4.47301529e-02
6.01950467e-01 -3.18370402e-01 3.89021009e-01 -1.27403057e+00
1.34711653e-01 -8.28741014e-01 7.76898742e-01 -1.22597849e+00
-1.01590061e+00 4.62069571e-01 3.15499395e-01 -1.59593284e+00
-2.78560907e-01 -8.82495940e-01 -1.09291703e-01 2.64747620e-01
-1.16234446e+00 -1.04283464e+00 -8.76877725e-01 1.10005724e+00
6.63829684e-01 -2.94513136e-01 8.49247873e-01 3.50193754e-02
-5.06937981e-01 4.49834853e-01 1.46775603e-01 2.92009443e-01
6.10566974e-01 -7.90293336e-01 4.29610200e-02 7.27568626e-01
2.28361875e-01 -2.16129925e-02 2.96175778e-01 -4.27801102e-01
-1.71243489e+00 -9.93910193e-01 2.55210698e-01 -9.13350463e-01
6.13524355e-02 6.15051249e-03 -5.17475069e-01 7.95020938e-01
-1.78744331e-01 8.10004354e-01 8.08134377e-01 -4.13580596e-01
-3.45377415e-01 -1.49464861e-01 -1.50142443e+00 4.25952852e-01
1.24979651e+00 -1.16887979e-01 -4.60938692e-01 4.53493357e-01
1.37523189e-01 -7.41065562e-01 -1.06889069e+00 3.39071870e-01
1.16686845e+00 -1.21267533e+00 1.13732898e+00 -6.90542579e-01
2.91826397e-01 -3.81941885e-01 -8.24491739e-01 -7.35875547e-01
-1.55720159e-01 -3.22394699e-01 -5.08462608e-01 6.16561174e-01
-9.65173617e-02 -3.19822818e-01 9.10696328e-01 6.16858304e-01
-5.29639944e-02 -7.40904510e-01 -1.22698784e+00 -9.47893202e-01
-5.31804919e-01 -4.83653247e-01 1.38163447e-01 6.41737640e-01
-6.36795089e-02 -2.69620251e-02 -5.82184613e-01 2.07831666e-01
5.16404808e-01 1.13664366e-01 1.21290052e+00 -9.84214902e-01
-2.67579049e-01 1.65353622e-02 -1.00779164e+00 -8.16892147e-01
-3.16738129e-01 -2.08304316e-01 -3.59766275e-01 -1.79370844e+00
1.90716729e-01 9.08445939e-02 -2.71627456e-01 7.81964183e-01
6.15674369e-02 2.90993929e-01 2.89166629e-01 1.86071217e-01
-7.51875877e-01 4.77369785e-01 1.17857897e+00 -1.19488470e-01
2.24226899e-02 2.35380173e-01 -2.10217983e-01 1.01629674e+00
6.92385018e-01 -3.07649851e-01 -3.68022740e-01 -4.54233825e-01
-4.42620844e-01 3.78641039e-01 4.00215656e-01 -1.26322830e+00
5.09808883e-02 -4.30174887e-01 9.88549292e-01 -8.04702938e-01
8.63756895e-01 -1.18209839e+00 2.01110244e-01 6.89349592e-01
9.97531638e-02 1.51896909e-01 9.67713445e-02 6.99291646e-01
1.51588336e-01 2.13629827e-01 9.94040251e-01 -4.76230115e-01
-8.79941881e-01 2.72312343e-01 -2.86288738e-01 1.16239376e-01
1.58858478e+00 -7.67182529e-01 -2.83261031e-01 -3.97776991e-01
-7.26107776e-01 9.93466973e-02 6.20572031e-01 5.91623068e-01
1.10594273e+00 -1.33824658e+00 -4.86600429e-01 4.03610170e-01
3.25476199e-01 -1.34230973e-02 2.97404021e-01 1.04309916e+00
-6.78541958e-01 3.57918650e-01 -7.36375868e-01 -6.69687569e-01
-1.82471323e+00 5.28645515e-01 1.93741947e-01 -4.37335484e-02
-8.69496405e-01 7.51291573e-01 6.28364533e-02 -9.97123569e-02
4.22803611e-01 -2.22351179e-01 -1.10707562e-02 -7.65093118e-02
7.72707582e-01 1.01375890e+00 -7.87054934e-03 -1.07717323e+00
-6.56082988e-01 9.54010129e-01 1.46105811e-01 -1.15402392e-03
1.41776395e+00 4.77155522e-02 4.29958314e-01 4.14176255e-01
1.15715265e+00 -4.40115154e-01 -1.48634005e+00 -2.51645833e-01
-4.48048919e-01 -8.67934406e-01 -2.19338268e-01 -7.31228471e-01
-1.37796521e+00 9.52234983e-01 1.08417237e+00 -1.30841583e-01
1.33164561e+00 2.31869400e-01 5.85401356e-01 4.27943558e-01
7.53527999e-01 -1.13612509e+00 6.13296926e-01 9.69305560e-02
1.00412083e+00 -1.31765890e+00 5.93941689e-01 -1.09470941e-01
-9.05964494e-01 1.33884025e+00 6.70156002e-01 6.14507385e-02
4.89430934e-01 1.13891222e-01 9.28327069e-02 -2.35633880e-01
-4.12034422e-01 -2.52862096e-01 2.71147907e-01 9.11469400e-01
1.37036592e-01 1.23693742e-01 2.64533907e-01 -6.23602793e-02
2.21082360e-01 -1.29632756e-01 7.40119398e-01 1.46719003e+00
-1.95776761e-01 -7.02528358e-01 -5.63216865e-01 3.15280765e-01
-4.30275530e-01 5.86790442e-01 -4.81846482e-01 1.27100885e+00
2.82075778e-02 1.09081709e+00 -4.64762794e-03 -3.49013716e-01
6.62639022e-01 4.00591455e-02 8.38712454e-01 -5.55391490e-01
-3.10374983e-02 2.55026877e-01 8.25638976e-03 -1.03579044e+00
-9.54940438e-01 -1.28457785e+00 -1.21927106e+00 -5.64593934e-02
-1.59427077e-01 -5.16404510e-01 5.74183166e-01 8.24485600e-01
2.46929511e-01 2.96260148e-01 4.69497532e-01 -7.79450655e-01
-1.85097232e-01 -7.56960928e-01 -7.34049320e-01 3.51580292e-01
5.00455678e-01 -8.71262968e-01 -1.66316688e-01 4.26206142e-01] | [7.816623210906982, 0.3257697820663452] |
50a62586-b701-4475-af80-01d6377583ad | effects-of-creativity-and-cluster-tightness | null | null | https://aclanthology.org/P16-1062 | https://aclanthology.org/P16-1062.pdf | Effects of Creativity and Cluster Tightness on Short Text Clustering Performance | null | ['Rui Zhang', 'Catherine Finegan-Dollak', 'Xiangyi Ye', 'Reed Coke', 'Dragomir Radev'] | 2016-08-01 | null | null | null | acl-2016-8 | ['text-clustering', 'short-text-clustering'] | ['natural-language-processing', '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] | [-7.296499729156494, 3.6718928813934326] |
65436597-db02-40f3-9858-517d93e00fc1 | no-reference-color-image-quality-assessment | 1812.10695 | null | http://arxiv.org/abs/1812.10695v1 | http://arxiv.org/pdf/1812.10695v1.pdf | No-Reference Color Image Quality Assessment: From Entropy to Perceptual Quality | This paper presents a high-performance general-purpose no-reference (NR)
image quality assessment (IQA) method based on image entropy. The image
features are extracted from two domains. In the spatial domain, the mutual
information between the color channels and the two-dimensional entropy are
calculated. In the frequency domain, the two-dimensional entropy and the mutual
information of the filtered sub-band images are computed as the feature set of
the input color image. Then, with all the extracted features, the support
vector classifier (SVC) for distortion classification and support vector
regression (SVR) are utilized for the quality prediction, to obtain the final
quality assessment score. The proposed method, which we call entropy-based
no-reference image quality assessment (ENIQA), can assess the quality of
different categories of distorted images, and has a low complexity. The
proposed ENIQA method was assessed on the LIVE and TID2013 databases and showed
a superior performance. The experimental results confirmed that the proposed
ENIQA method has a high consistency of objective and subjective assessment on
color images, which indicates the good overall performance and generalization
ability of ENIQA. The source code is available on github
https://github.com/jacob6/ENIQA. | ['Xiaoqiao Chen', 'Manhui Lin', 'Guangyi Yang', 'Qingyi Zhang', 'Chu He'] | 2018-12-27 | null | null | null | null | ['no-reference-image-quality-assessment'] | ['computer-vision'] | [-4.78308089e-02 -8.98808360e-01 1.37258455e-01 -2.28207216e-01
-8.08535159e-01 -1.84989691e-01 1.80661187e-01 -1.28796771e-01
-1.47853836e-01 3.20392370e-01 2.49078259e-01 8.52269232e-02
-4.80962396e-01 -8.18003416e-01 1.44175157e-01 -1.00783062e+00
-1.27957404e-01 -4.89870936e-01 1.61390364e-01 -7.98082873e-02
4.31792885e-01 3.11319083e-01 -1.61776912e+00 1.45092621e-01
1.32698655e+00 1.62502861e+00 1.67636082e-01 6.58096671e-01
2.82686859e-01 6.56609178e-01 -4.93291169e-01 -2.08258212e-01
2.26162866e-01 -6.53247714e-01 -4.52853322e-01 3.32997590e-02
-2.22352371e-01 -3.76624465e-01 -4.88292456e-01 1.23005140e+00
7.38833249e-01 1.89253569e-01 6.25446260e-01 -1.15050900e+00
-8.56686652e-01 -1.82923794e-01 -4.51488972e-01 4.28197056e-01
4.09689158e-01 2.60340303e-01 7.05531061e-01 -1.04937363e+00
1.94223464e-01 1.07243037e+00 3.23956490e-01 5.00844531e-02
-9.68354046e-01 -8.50176036e-01 -3.75407577e-01 8.17729831e-01
-1.59368920e+00 -2.01865837e-01 8.20094585e-01 -5.56735814e-01
4.31714177e-01 3.98031563e-01 8.61288071e-01 1.92603856e-01
4.73488092e-01 5.35057545e-01 1.50964034e+00 -1.72392517e-01
2.11006865e-01 1.62139192e-01 -4.70279455e-02 7.74369121e-01
5.77953644e-02 3.99577320e-01 -2.00866669e-01 1.63996089e-02
6.63223624e-01 3.61779295e-02 -5.65617025e-01 -6.37619048e-02
-1.32808471e+00 7.36929119e-01 5.91487169e-01 5.42911708e-01
-3.59537691e-01 -4.72762823e-01 5.38657427e-01 2.85281211e-01
3.03163171e-01 4.21396494e-02 3.10360398e-02 -2.10151181e-01
-8.91868114e-01 -3.05391192e-01 4.57695544e-01 4.90002871e-01
6.02468312e-01 1.46178856e-01 -1.82145178e-01 1.02512562e+00
3.87096375e-01 9.31126297e-01 8.12907755e-01 -1.02028358e+00
2.47114673e-01 5.00733852e-01 -8.39339744e-04 -1.70709395e+00
-2.70996511e-01 -1.80903167e-01 -1.30203497e+00 4.98124570e-01
6.98850974e-02 1.94827691e-01 -5.25034904e-01 1.21543181e+00
9.10052732e-02 -5.20342179e-02 1.99666530e-01 1.20911407e+00
1.07385325e+00 1.10428572e+00 -8.04964676e-02 -5.89648664e-01
1.18600559e+00 -5.43426991e-01 -1.02733314e+00 3.55539232e-01
2.03591250e-02 -1.01901245e+00 9.95785415e-01 7.14456737e-01
-1.08753550e+00 -9.18147922e-01 -1.23211253e+00 1.87658563e-01
-2.16703996e-01 5.46600938e-01 2.51299560e-01 5.79328716e-01
-9.36811566e-01 5.16870081e-01 -4.77703452e-01 -3.32135893e-02
3.53374481e-01 -2.99866591e-02 -4.76575166e-01 -3.43877017e-01
-1.32717633e+00 8.28609526e-01 3.35255027e-01 1.26707196e-01
-4.60289747e-01 -4.55588311e-01 -8.66799891e-01 -2.27172375e-02
-4.22711670e-03 -1.98332280e-01 5.55953979e-01 -9.66438949e-01
-1.50629997e+00 5.11963248e-01 -1.86464712e-01 1.98510811e-01
1.05862357e-01 2.39281729e-01 -9.78013515e-01 5.13321638e-01
-7.45575652e-02 3.75555418e-02 5.47585785e-01 -1.22536850e+00
-4.83468056e-01 -5.37721872e-01 -3.51052612e-01 2.39371046e-01
-2.94509262e-01 -8.43699090e-03 -3.64870578e-01 -7.04875469e-01
5.04926801e-01 -5.06772816e-01 3.22852373e-01 1.39464915e-01
-2.59192467e-01 1.04822539e-01 6.56580627e-01 -1.23537016e+00
1.54220748e+00 -2.49582815e+00 6.85149282e-02 5.58412671e-01
-6.35631457e-02 3.65409553e-01 -1.75502256e-01 2.00053200e-01
-9.23882723e-02 1.57738745e-01 -2.94574201e-01 5.27042806e-01
-2.88071275e-01 -2.76878476e-01 2.68551826e-01 5.92109680e-01
5.75526617e-02 4.53393877e-01 -7.99740970e-01 -6.45294547e-01
4.86526489e-01 6.50606811e-01 -6.17585853e-02 3.41449142e-01
6.26013398e-01 3.25721711e-01 -2.25299746e-01 6.66276693e-01
1.05054021e+00 -2.00609595e-01 -2.02589720e-01 -7.61560202e-01
-9.06836241e-02 -1.82211980e-01 -1.55085194e+00 1.32958376e+00
-5.21383882e-01 6.41886234e-01 -1.09908640e-01 -6.87087297e-01
1.15899324e+00 3.69766533e-01 6.18665755e-01 -1.28678536e+00
1.39879808e-01 3.26179206e-01 6.76802248e-02 -8.56934667e-01
1.20143875e-01 7.02702999e-02 2.66469449e-01 2.24257976e-01
6.10540435e-02 -1.20899372e-01 2.21071273e-01 9.81033668e-02
5.41724861e-01 -1.29184887e-01 4.81876612e-01 -1.97278619e-01
9.99921262e-01 -3.63857925e-01 6.22110248e-01 9.43860635e-02
-4.62318838e-01 6.06370211e-01 7.18601570e-02 -8.99915248e-02
-1.01454651e+00 -1.11037564e+00 -3.77729803e-01 3.14714760e-01
7.18832731e-01 -2.57326990e-01 -4.68559086e-01 -2.10522905e-01
-1.80736095e-01 3.76036465e-01 -3.24221969e-01 -3.42343122e-01
-9.87078846e-02 -7.07011878e-01 -1.95648540e-02 9.37006772e-02
1.14317846e+00 -8.71390820e-01 -3.17654788e-01 -1.45066464e-02
-5.40671766e-01 -6.56645536e-01 -5.13429642e-01 -6.18639648e-01
-7.30326712e-01 -1.08666146e+00 -9.81783628e-01 -6.46373332e-01
4.52161580e-01 3.35601449e-01 7.52958536e-01 1.70756966e-01
-4.16739017e-01 3.44612896e-01 -3.46590340e-01 1.41891852e-01
-2.75374860e-01 -7.76546180e-01 5.86825013e-02 4.44082588e-01
1.65934876e-01 -5.69970608e-01 -1.11632967e+00 5.16657054e-01
-8.45506966e-01 -4.51799482e-02 7.59531319e-01 9.66997564e-01
8.21992576e-01 6.19738400e-01 4.25110459e-01 -2.43969727e-02
6.86046183e-01 -3.11444670e-01 -6.88128114e-01 1.51483133e-01
-9.40182447e-01 -2.75301158e-01 6.63874149e-01 -2.10105896e-01
-1.13086724e+00 -4.53395128e-01 -1.52762726e-01 -1.90598994e-01
3.98839004e-02 6.17448151e-01 -3.93675566e-01 -1.57952234e-01
4.21045393e-01 6.91335440e-01 9.37818363e-02 -4.26027834e-01
1.23604380e-01 8.66984665e-01 5.72798371e-01 -2.73359325e-02
8.36683929e-01 2.53253013e-01 8.12586099e-02 -6.74268842e-01
-2.79589266e-01 -6.02988839e-01 -3.58373970e-01 -4.73641247e-01
8.88806641e-01 -9.96729314e-01 -8.23231876e-01 8.40236247e-01
-6.87184751e-01 2.00417206e-01 2.30169302e-04 9.81676459e-01
-4.53515738e-01 6.19810939e-01 -5.71017325e-01 -8.91797423e-01
-4.95773733e-01 -1.25228786e+00 4.46106344e-01 6.41388655e-01
3.22858304e-01 -8.28116417e-01 -1.60513401e-01 1.88094392e-01
5.56653798e-01 2.87373900e-01 8.87708783e-01 -5.09080663e-02
-4.95241016e-01 -2.46847421e-01 -4.91229236e-01 7.78985023e-01
3.75255674e-01 1.12841196e-01 -6.18290961e-01 -3.85232240e-01
1.37758955e-01 -1.45772502e-01 5.88746488e-01 4.38772261e-01
1.30850101e+00 -4.87737566e-01 3.55662048e-01 7.40035772e-01
1.97336054e+00 6.18769050e-01 9.54068303e-01 3.36128831e-01
3.17515403e-01 2.33979419e-01 9.15168762e-01 6.33428633e-01
3.37180525e-01 7.92499363e-01 3.25917929e-01 -2.67938465e-01
-1.28617719e-01 4.63416288e-03 3.81024599e-01 1.41334963e+00
-2.63777047e-01 7.46895522e-02 -6.76252365e-01 4.25409108e-01
-1.32086647e+00 -1.27590239e+00 -1.50413603e-01 2.25794935e+00
7.93673813e-01 -1.95275709e-01 3.37710828e-02 6.91919088e-01
6.05629981e-01 4.25289832e-02 -5.99437773e-01 -2.89167017e-01
-2.18989715e-01 -1.01075985e-01 2.80039757e-01 3.94205958e-01
-1.00758874e+00 -7.80826882e-02 5.20501995e+00 1.03238237e+00
-1.17778790e+00 -1.75374635e-02 8.02197993e-01 1.30089283e-01
-1.45910740e-01 -3.00947756e-01 -7.01908246e-02 8.92335773e-01
8.19901705e-01 -6.08663917e-01 6.93576813e-01 6.06522083e-01
5.31776011e-01 -4.28776979e-01 -3.77427071e-01 1.45886767e+00
9.16770473e-02 -9.84775364e-01 -1.04865782e-01 -5.97680511e-04
6.56801224e-01 -2.64858961e-01 3.93417954e-01 -1.47618026e-01
-1.27397358e-01 -7.71303296e-01 3.49788606e-01 1.03480673e+00
1.09376919e+00 -9.06376421e-01 1.03344798e+00 2.13298183e-02
-1.34668303e+00 -3.32392067e-01 -4.73578364e-01 3.28059793e-01
-5.66439554e-02 8.41202736e-01 1.36316657e-01 9.70253527e-01
9.57860768e-01 9.89637077e-01 -7.28650868e-01 1.39943814e+00
-7.67948627e-02 5.61769843e-01 1.35888636e-01 2.23264575e-01
-2.16796041e-01 -6.13116086e-01 5.06979525e-01 9.38092291e-01
7.17363179e-01 5.70976138e-01 -2.09380686e-01 5.91153681e-01
1.76852867e-01 5.29987693e-01 -7.15640560e-02 1.88467443e-01
3.83258760e-01 1.26487207e+00 -4.03984487e-01 -4.12767857e-01
-3.90307426e-01 9.66894031e-01 -5.58588386e-01 3.29976976e-01
-6.71214104e-01 -8.86875570e-01 4.32481617e-01 -2.17289671e-01
1.94583535e-01 7.03328699e-02 -3.32983285e-01 -1.37387598e+00
2.91200072e-01 -9.09369469e-01 4.66097862e-01 -1.15870309e+00
-1.29577053e+00 5.95892429e-01 -2.40023836e-01 -1.93799174e+00
8.98723006e-02 -3.69348735e-01 -5.85177541e-01 1.29104626e+00
-1.64988005e+00 -5.07369161e-01 -6.68939114e-01 8.76449823e-01
1.58868581e-01 -2.86923617e-01 8.08306158e-01 4.39974785e-01
-6.23967707e-01 5.66490233e-01 6.96935356e-01 2.07247034e-01
5.13106167e-01 -1.03048694e+00 -3.67190421e-01 9.40439582e-01
-2.66303420e-01 2.53707081e-01 3.64334852e-01 -3.34444821e-01
-1.33645511e+00 -8.92743826e-01 5.27534127e-01 2.98615545e-01
2.33402833e-01 2.06310362e-01 -1.00616693e+00 -1.36399165e-01
1.20524295e-01 2.96474189e-01 7.99461246e-01 -5.90324879e-01
-1.55521661e-01 -6.28899157e-01 -1.47567821e+00 4.31854241e-02
3.78117859e-01 -6.56272769e-01 -1.53970748e-01 6.75945580e-02
4.50795442e-01 -7.49456063e-02 -1.60158718e+00 3.60819131e-01
7.20209777e-01 -1.23002720e+00 9.83151436e-01 2.94402540e-01
6.23384476e-01 -7.23912299e-01 -3.85020822e-01 -1.38269413e+00
-6.23805881e-01 6.35765344e-02 2.06655692e-02 1.26351273e+00
1.22485399e-01 -6.75768912e-01 -5.81278801e-02 3.30920905e-01
3.42545122e-01 -9.28145230e-01 -8.64168465e-01 -8.70626390e-01
-2.74699181e-01 -9.66709554e-02 6.46230519e-01 6.67029500e-01
9.35075209e-02 -2.16222912e-01 -4.46887076e-01 2.30062082e-01
7.20762372e-01 8.93806443e-02 1.91615745e-01 -1.01466370e+00
-1.38586015e-01 -4.30867106e-01 -8.79736662e-01 -5.46683192e-01
-5.03468812e-01 -8.58962715e-01 -1.76558629e-01 -1.72550499e+00
3.63642365e-01 -1.91221058e-01 -7.58682668e-01 8.97994917e-03
-3.02960843e-01 5.82544446e-01 3.74717385e-01 4.58957553e-01
-3.82254869e-01 8.17402780e-01 1.40148008e+00 -3.36528480e-01
-1.32291257e-01 2.18292102e-02 -5.37594140e-01 3.75175297e-01
7.62951970e-01 -1.25586048e-01 -3.07149440e-01 -8.04246739e-02
-1.07726954e-01 3.41954619e-01 2.93423772e-01 -1.28367734e+00
-5.58009297e-02 -1.97631747e-01 8.29428673e-01 -4.19109792e-01
2.07443163e-01 -9.43595827e-01 1.98547453e-01 6.76156998e-01
-2.06914812e-01 -1.09793812e-01 1.22942626e-02 3.31268072e-01
-7.26208687e-01 1.29210413e-01 1.15285909e+00 3.53384838e-02
-8.51554513e-01 4.87799168e-01 -1.15150809e-01 -2.13352948e-01
9.55335975e-01 -3.43057245e-01 -2.83811539e-01 -5.40555537e-01
-7.22186148e-01 -1.51657224e-01 4.84111249e-01 1.34784967e-01
1.17194879e+00 -1.89069867e+00 -9.11960602e-01 2.95568794e-01
2.76528358e-01 -6.45084798e-01 7.68867970e-01 9.13538933e-01
-4.83166158e-01 8.79407600e-02 -5.40891647e-01 -5.55162728e-01
-1.32148397e+00 5.84535420e-01 5.01491964e-01 -1.66166246e-01
-3.38521212e-01 3.15971017e-01 -9.16960090e-02 6.33760020e-02
-1.59567386e-01 1.04137965e-01 -6.50167465e-01 3.69671686e-03
9.52518165e-01 6.70249879e-01 6.32082596e-02 -1.22730660e+00
-3.46162081e-01 9.16649461e-01 4.15486604e-01 -1.58725858e-01
1.18026221e+00 -5.24247289e-01 -3.49243462e-01 3.57473075e-01
1.64258242e+00 -9.29651633e-02 -9.23893988e-01 -2.97149777e-01
-4.49908316e-01 -1.05586934e+00 4.01747793e-01 -1.19441998e+00
-1.45929253e+00 1.03099859e+00 1.48406065e+00 2.22689494e-01
2.00393867e+00 -4.16516274e-01 6.85118437e-01 -2.18108162e-01
2.58572727e-01 -9.97961640e-01 5.98904528e-02 2.42792647e-02
1.01944447e+00 -1.26664031e+00 2.49444813e-01 -1.81968138e-01
-7.51304746e-01 1.24104559e+00 4.16722298e-01 2.78164633e-02
9.19242501e-01 -3.39536935e-01 1.72333613e-01 1.01195656e-01
-2.80803740e-01 1.04088325e-03 5.18669367e-01 6.42358124e-01
2.26148993e-01 9.90385488e-02 -6.09678864e-01 6.94250226e-01
-4.53301221e-02 1.23043559e-01 4.33293432e-01 3.95757049e-01
-3.25946867e-01 -6.56491458e-01 -4.30807501e-01 5.33721566e-01
-4.24130082e-01 5.06862774e-02 3.55560005e-01 2.97931731e-01
2.55987078e-01 1.38082588e+00 -1.56736210e-01 -7.95899510e-01
4.40175205e-01 -3.67949039e-01 1.51744783e-01 1.32024795e-01
-8.30628872e-02 -1.22421890e-01 -3.81663293e-01 -9.60175216e-01
-5.26195288e-01 -4.67972934e-01 -9.08154190e-01 -3.30507159e-01
-3.64243060e-01 1.89752117e-01 7.64082551e-01 5.42545199e-01
2.66272038e-01 3.91128898e-01 1.43551171e+00 -5.25089562e-01
-1.43152475e-01 -8.96015167e-01 -8.93584073e-01 6.30796075e-01
4.40203160e-01 -4.82188195e-01 -4.71122056e-01 2.02993155e-01] | [11.715927124023438, -1.929663062095642] |
788b93b9-6ff0-4b63-bddc-5e5c7cbba3b4 | scene-text-retrieval-via-joint-text-detection | 2104.01552 | null | https://arxiv.org/abs/2104.01552v1 | https://arxiv.org/pdf/2104.01552v1.pdf | Scene Text Retrieval via Joint Text Detection and Similarity Learning | Scene text retrieval aims to localize and search all text instances from an image gallery, which are the same or similar to a given query text. Such a task is usually realized by matching a query text to the recognized words, outputted by an end-to-end scene text spotter. In this paper, we address this problem by directly learning a cross-modal similarity between a query text and each text instance from natural images. Specifically, we establish an end-to-end trainable network, jointly optimizing the procedures of scene text detection and cross-modal similarity learning. In this way, scene text retrieval can be simply performed by ranking the detected text instances with the learned similarity. Experiments on three benchmark datasets demonstrate our method consistently outperforms the state-of-the-art scene text spotting/retrieval approaches. In particular, the proposed framework of joint detection and similarity learning achieves significantly better performance than separated methods. Code is available at: https://github.com/lanfeng4659/STR-TDSL. | ['Wenyu Liu', 'Jing Wang', 'Shenggao Zhu', 'Mingkun Yang', 'Xiang Bai', 'Hao Wang'] | 2021-04-04 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Wang_Scene_Text_Retrieval_via_Joint_Text_Detection_and_Similarity_Learning_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Wang_Scene_Text_Retrieval_via_Joint_Text_Detection_and_Similarity_Learning_CVPR_2021_paper.pdf | cvpr-2021-1 | ['text-spotting', 'scene-text-detection'] | ['computer-vision', 'computer-vision'] | [ 4.63761449e-01 -5.41883945e-01 -1.11562692e-01 -5.11667013e-01
-1.39827979e+00 -5.51419258e-01 9.39734578e-01 1.92067385e-01
-5.15190184e-01 -8.25632215e-02 1.02720514e-01 1.24920830e-01
2.29006889e-03 -4.46205497e-01 -7.20388234e-01 -5.97171009e-01
6.23579621e-01 7.01641679e-01 1.53131545e-01 2.82454163e-01
6.58122957e-01 1.42879039e-01 -1.43916261e+00 3.87146831e-01
7.03433335e-01 9.96576548e-01 7.09627390e-01 6.55797899e-01
-2.58203536e-01 7.81926811e-01 -3.50717574e-01 -1.10131651e-01
6.62287250e-02 -3.46813679e-01 -8.46270084e-01 3.71065110e-01
1.01289511e+00 -3.14239472e-01 -6.33965015e-01 1.14299512e+00
4.07508641e-01 3.18547368e-01 6.07728839e-01 -9.02534366e-01
-6.58421338e-01 4.35603529e-01 -7.22372949e-01 1.69038121e-02
5.77221751e-01 -6.75252229e-02 1.21544647e+00 -1.38758790e+00
4.90221173e-01 1.19938672e+00 9.56620276e-02 3.90535474e-01
-1.04158676e+00 -5.87615013e-01 1.41969860e-01 1.26858905e-01
-1.79532027e+00 -7.08082974e-01 7.88066447e-01 -4.11220610e-01
5.41959584e-01 2.72324622e-01 9.88727212e-02 6.97282612e-01
-5.05994409e-02 1.32404196e+00 5.37323296e-01 -5.12958109e-01
-4.80618700e-02 8.25520307e-02 1.56904191e-01 9.36752677e-01
4.67047170e-02 -4.58215833e-01 -8.09314966e-01 -2.69467495e-02
3.83242667e-01 4.34428751e-01 -4.54239219e-01 -5.65981030e-01
-1.50871217e+00 6.02220237e-01 3.85257512e-01 4.21385109e-01
-2.57946283e-01 2.69938171e-01 4.98748839e-01 6.90248087e-02
6.12327814e-01 1.67816237e-01 -7.94546232e-02 1.85227200e-01
-1.25031972e+00 7.04015493e-02 5.54259539e-01 9.94034231e-01
8.72630298e-01 -2.22778961e-01 -3.99190694e-01 1.07331586e+00
5.76218367e-01 9.79180396e-01 6.39190972e-01 -2.39564031e-01
7.92978466e-01 7.56093383e-01 -1.58238113e-02 -8.52336824e-01
6.20554611e-02 -2.86482722e-02 -7.09490776e-01 -3.41581672e-01
1.81671426e-01 2.07121402e-01 -1.09891355e+00 1.22836530e+00
4.18755293e-01 3.52246910e-01 1.41275942e-01 1.10185874e+00
9.98316467e-01 9.51138020e-01 -3.53123806e-02 2.47498080e-01
1.38902640e+00 -1.25161767e+00 -5.51570594e-01 -6.11390233e-01
7.07935691e-01 -1.35362530e+00 1.29025042e+00 3.61921303e-02
-8.04358125e-01 -4.81803894e-01 -7.32543170e-01 -2.35105664e-01
-3.17845821e-01 7.78670132e-01 8.34920630e-02 2.04711884e-01
-8.36596668e-01 1.64265394e-01 -8.08064878e-01 -7.76475310e-01
3.06632251e-01 1.11612767e-01 -2.17755914e-01 -4.82516170e-01
-6.91728294e-01 3.92716110e-01 4.95199174e-01 1.05205894e-01
-1.20551240e+00 -3.94312143e-01 -1.02622867e+00 2.90257726e-02
5.07401586e-01 -6.20805979e-01 1.34812629e+00 -9.11924541e-01
-1.15838051e+00 1.31133497e+00 -4.41195786e-01 -1.65994644e-01
4.50146526e-01 -5.87199569e-01 -1.57452211e-01 3.47700596e-01
6.06527984e-01 5.75592339e-01 1.21316445e+00 -1.19008183e+00
-7.52309144e-01 -4.55090433e-01 -2.79002458e-01 5.57449639e-01
-3.55437815e-01 2.73321152e-01 -1.07167029e+00 -5.68071485e-01
2.50185430e-01 -7.98213363e-01 1.74986422e-01 1.80626720e-01
-7.66377628e-01 -5.10434866e-01 1.21543765e+00 -5.13267100e-01
8.26353371e-01 -2.21615863e+00 -1.10421814e-01 -1.24411108e-02
6.98849484e-02 1.69342875e-01 -3.27999443e-01 6.62033439e-01
1.24136023e-01 -2.33553916e-01 -1.18949227e-01 -7.14830518e-01
1.93739578e-01 -4.13927615e-01 -5.94598055e-01 9.03539002e-01
-1.40570521e-01 9.46374118e-01 -7.78622031e-01 -8.66908431e-01
6.12177372e-01 2.75394082e-01 -2.07457761e-03 4.44251359e-01
-3.98077995e-01 1.11213125e-01 -8.72288764e-01 7.61928082e-01
4.99191135e-01 -5.20586669e-01 -2.32961416e-01 -1.19145840e-01
-2.62604351e-03 1.69891432e-01 -9.75285888e-01 2.06276560e+00
-3.55250508e-01 1.00622058e+00 4.09114622e-02 -1.00075138e+00
9.32528496e-01 1.39107823e-01 3.19251120e-01 -8.69908690e-01
9.31440666e-02 8.26738626e-02 -7.82968998e-01 -5.60959876e-01
7.32997298e-01 3.03907007e-01 -1.73981637e-01 6.95611060e-01
-1.11679003e-01 -3.77306372e-01 2.45762855e-01 4.16162252e-01
8.02263439e-01 -3.51259671e-02 7.86369592e-02 -7.57520795e-02
6.52514398e-01 1.21663406e-01 -2.12850273e-02 1.00173473e+00
-1.77033581e-02 5.85781157e-01 -7.22510740e-02 -1.77530974e-01
-1.03838933e+00 -9.58180428e-01 -4.11207462e-03 1.36328197e+00
6.08817399e-01 -4.55130100e-01 -6.09690845e-01 -5.66953361e-01
-6.13560714e-02 5.02649188e-01 -4.26493019e-01 -4.22910638e-02
-3.11273903e-01 -1.40371159e-01 5.35384059e-01 1.96877792e-01
6.90104604e-01 -9.57606971e-01 -4.63337362e-01 -2.56466210e-01
-3.42605710e-01 -1.24714184e+00 -1.02353859e+00 -3.67546156e-02
-6.58105373e-01 -9.16611850e-01 -9.43906486e-01 -1.35747755e+00
7.34487236e-01 1.00763583e+00 9.05989647e-01 1.78315684e-01
-5.19153237e-01 7.39792585e-01 -4.29237038e-01 5.72744757e-03
-2.41592228e-01 9.74679440e-02 -2.03462973e-01 3.73188645e-01
3.98163646e-01 1.52982861e-01 -7.34262884e-01 3.38937700e-01
-1.23984122e+00 2.31273010e-01 6.00882709e-01 7.54440010e-01
8.58615220e-01 -7.50603562e-04 1.34074129e-02 -6.09612644e-01
4.44473743e-01 -1.82192191e-01 -7.07583427e-01 5.85337639e-01
-3.72219741e-01 -3.51484343e-02 4.34194952e-01 -2.79225290e-01
-8.81288111e-01 5.14274359e-01 3.57906640e-01 -7.92752564e-01
-4.91474390e-01 4.74535614e-01 -6.24542721e-02 3.82903405e-02
4.74906176e-01 8.66985261e-01 -4.53707576e-01 -3.75808179e-01
4.66300547e-01 9.45671916e-01 5.30414760e-01 -5.17009079e-01
9.71713960e-01 7.10384011e-01 -5.34399688e-01 -1.04782653e+00
-1.08107495e+00 -1.23536849e+00 -6.98230386e-01 -2.26935491e-01
9.09195185e-01 -1.20681536e+00 -2.50382572e-01 5.67655981e-01
-1.08038664e+00 -4.37817901e-01 7.08900020e-02 4.63259935e-01
-5.42006195e-01 6.81397200e-01 -3.65851164e-01 -6.27513945e-01
-7.30027556e-01 -9.54139650e-01 2.07741857e+00 1.85526788e-01
2.37833887e-01 -8.72892201e-01 8.18629563e-02 5.67246914e-01
1.83808077e-02 -2.26036221e-01 6.28495693e-01 -9.63929534e-01
-7.54324734e-01 -6.28592670e-01 -5.41282594e-01 -6.38872087e-02
1.81490257e-01 -1.31441444e-01 -9.66681421e-01 -5.32686710e-01
-2.22751737e-01 -4.88913089e-01 9.92349446e-01 2.68259764e-01
1.01273084e+00 -7.92204589e-02 -5.79237938e-01 5.61147451e-01
1.55228007e+00 3.77414370e-04 2.56950498e-01 3.25336605e-01
9.48579848e-01 4.39422905e-01 1.00699377e+00 4.26713407e-01
1.62473649e-01 6.68508649e-01 2.27354378e-01 -2.25235134e-01
-9.27593559e-02 -3.76506299e-01 3.08517456e-01 4.23400491e-01
9.67437327e-01 -5.69436848e-01 -1.17008841e+00 6.36177480e-01
-2.10288787e+00 -7.65979409e-01 -1.04212068e-01 2.17314196e+00
5.23115933e-01 -1.79447204e-01 -2.42003679e-01 -2.64590651e-01
1.21226823e+00 4.72792655e-01 -6.32864237e-01 3.12577993e-01
-9.28080231e-02 -2.27826193e-01 3.26491237e-01 4.02062744e-01
-1.43052208e+00 1.54287887e+00 4.82750130e+00 1.12004089e+00
-1.24895287e+00 -4.26944606e-02 4.94910777e-01 -5.43423407e-02
1.18382990e-01 3.11397165e-02 -7.48925567e-01 1.74485192e-01
3.83820593e-01 -5.37493587e-01 3.62855434e-01 8.60203445e-01
3.41889858e-01 -1.34187043e-01 -1.25073469e+00 1.31940591e+00
6.90995872e-01 -1.10275114e+00 3.24584454e-01 -3.49149495e-01
7.97660232e-01 3.28267813e-01 9.30305012e-03 5.61262257e-02
4.12014350e-02 -6.89276159e-01 8.16924751e-01 4.49752301e-01
8.10327291e-01 -3.94053251e-01 4.18073326e-01 2.50505209e-01
-1.54139018e+00 2.00161517e-01 -4.38024521e-01 5.13629079e-01
-3.75533141e-02 4.81175214e-01 -1.03287220e+00 3.66075993e-01
6.41137242e-01 1.09554720e+00 -7.74943233e-01 1.30751014e+00
-2.27247119e-01 5.20088613e-01 -3.09477597e-01 -1.81215599e-01
3.78414959e-01 -7.93024898e-02 5.20271599e-01 1.44615543e+00
2.67434686e-01 -2.57531285e-01 5.00458002e-01 9.37099516e-01
-3.79298866e-01 4.50581819e-01 -7.86095381e-01 -1.97144777e-01
4.56098765e-01 1.39444983e+00 -9.13355291e-01 -5.30529261e-01
-3.06392044e-01 1.38885164e+00 1.41096249e-01 4.87999767e-01
-6.18049681e-01 -6.80119038e-01 2.05101281e-01 -3.60406309e-01
3.62438411e-01 2.62048040e-02 3.72523186e-03 -1.40712202e+00
3.21172029e-01 -7.53258765e-01 4.57014769e-01 -1.27144611e+00
-1.24865949e+00 3.78557384e-01 -2.77796805e-01 -1.36348259e+00
3.94054390e-02 -4.35079098e-01 -6.82761312e-01 7.27247298e-01
-1.36386788e+00 -1.46042764e+00 -6.56589448e-01 1.01993763e+00
1.16172659e+00 -5.35199828e-02 4.17957962e-01 1.68474704e-01
-5.51671743e-01 5.29544175e-01 5.10598242e-01 4.13232744e-01
1.03426504e+00 -1.03814781e+00 4.49462473e-01 1.09624112e+00
5.36828458e-01 3.08083147e-01 4.59975988e-01 -6.26975894e-01
-1.66265929e+00 -1.47991896e+00 7.38726974e-01 -3.55379224e-01
8.57949138e-01 -6.20850027e-01 -9.12906408e-01 6.30235791e-01
3.06636512e-01 -1.57109544e-01 7.79198185e-02 -3.08799118e-01
-3.95571381e-01 -1.76190257e-01 -6.27889514e-01 7.44416714e-01
6.90537512e-01 -1.02124310e+00 -5.36342204e-01 9.11408901e-01
6.39277101e-01 -3.91764879e-01 -3.68395030e-01 1.16405115e-02
2.63889015e-01 -4.21602368e-01 8.83750498e-01 -2.09639534e-01
3.95990252e-01 -4.35445607e-01 -3.12764078e-01 -7.93220878e-01
8.47568884e-02 -4.04945940e-01 3.38137746e-01 1.35237169e+00
2.12410718e-01 -3.37238669e-01 8.14346731e-01 3.30855310e-01
-9.08305421e-02 -2.33998716e-01 -8.40596855e-01 -5.87131143e-01
-1.34393692e-01 -2.57522702e-01 1.78052887e-01 9.83608723e-01
-8.55742022e-02 6.25402749e-01 -3.59120935e-01 3.39605331e-01
8.18679810e-01 6.84939742e-01 9.92542207e-01 -9.39347029e-01
6.86924160e-02 -5.05620480e-01 -3.89156342e-01 -1.61290550e+00
5.00302613e-01 -1.09554732e+00 6.36699021e-01 -1.67707825e+00
7.18987763e-01 -2.35024214e-01 -1.03738360e-01 3.67237240e-01
-3.11360270e-01 1.50479153e-01 2.21442834e-01 5.51345170e-01
-1.37687361e+00 7.12681055e-01 1.00622213e+00 -5.50532162e-01
1.95661485e-02 -1.29267260e-01 -3.28713179e-01 4.79595840e-01
8.03714454e-01 -6.24602556e-01 -2.46234298e-01 -8.26113880e-01
-8.43200609e-02 1.68684185e-01 5.56254506e-01 -9.03999746e-01
8.79122615e-01 -8.91035944e-02 2.50876695e-01 -1.07166517e+00
3.50318968e-01 -7.49875188e-01 -3.66276413e-01 2.80543804e-01
-7.59987473e-01 -1.14208169e-01 6.48342520e-02 7.42638350e-01
-2.62294531e-01 -4.55074817e-01 6.34799540e-01 1.66688800e-01
-7.64148772e-01 3.36188197e-01 -3.20410520e-01 8.81981403e-02
7.61298299e-01 4.61951457e-03 -5.17530382e-01 -3.93022567e-01
-2.61487424e-01 3.75847727e-01 5.73154628e-01 5.88000596e-01
9.52332854e-01 -1.00622535e+00 -8.05904508e-01 -1.87088009e-02
7.32222795e-01 1.44688785e-01 1.42926186e-01 6.27893746e-01
-3.71888459e-01 7.90461123e-01 5.21163583e-01 -1.04191709e+00
-1.63209474e+00 5.97570419e-01 4.33323532e-01 -9.98976007e-02
-7.01823890e-01 7.78306067e-01 6.30130827e-01 -4.60774511e-01
4.55612510e-01 -6.50433227e-02 2.16947123e-01 -3.01536471e-01
4.18474168e-01 -1.91221163e-02 -6.54722527e-02 -8.01712692e-01
-4.17353332e-01 1.01579273e+00 -3.68218333e-01 -1.00988090e-01
9.58328307e-01 -4.26974237e-01 -1.26118511e-01 4.35298920e-01
1.39397645e+00 -2.17142507e-01 -9.25164282e-01 -8.26573670e-01
1.23899475e-01 -5.81153095e-01 2.03274891e-01 -5.67569554e-01
-1.04682553e+00 8.28075767e-01 8.05372179e-01 -3.49856876e-02
1.04008818e+00 3.70313317e-01 7.01669812e-01 1.04863238e+00
9.09713656e-02 -9.84120786e-01 5.07113576e-01 4.34035778e-01
6.42637193e-01 -1.60301888e+00 5.11201546e-02 -1.51596054e-01
-4.29666489e-01 9.63929296e-01 4.92310196e-01 -2.08710551e-01
3.31811070e-01 -2.00122148e-01 1.77858710e-01 -5.16182423e-01
-7.52068400e-01 -4.45694298e-01 4.28290099e-01 6.59390464e-02
4.24212009e-01 -7.53012523e-02 2.22922951e-01 -2.36366719e-01
2.38051608e-01 -2.43820935e-01 1.16173945e-01 9.51458693e-01
-6.66999459e-01 -7.46384621e-01 -4.67519194e-01 4.26330537e-01
-4.23390508e-01 -4.64898080e-01 -7.47017503e-01 5.70078552e-01
-6.79550171e-01 1.08480906e+00 7.42175281e-02 -1.11945383e-01
3.11284900e-01 4.99081090e-02 7.65142739e-02 -7.88252473e-01
-2.82771349e-01 4.09767717e-01 -3.24826837e-01 -4.23193365e-01
-3.40793669e-01 -7.01684356e-01 -1.16591620e+00 6.26564492e-03
-6.37311101e-01 1.94652453e-01 8.28228056e-01 8.55972767e-01
3.36439878e-01 1.67738214e-01 9.07229304e-01 -9.04436290e-01
-2.34525263e-01 -8.60363245e-01 -4.91599649e-01 5.98775625e-01
3.90098065e-01 -1.59501091e-01 -4.35828447e-01 4.25210267e-01] | [11.729063034057617, 2.0629615783691406] |
4f526571-5476-494b-9586-24ff027e9a1e | variational-linearized-laplace-approximation | 2302.12565 | null | https://arxiv.org/abs/2302.12565v1 | https://arxiv.org/pdf/2302.12565v1.pdf | Variational Linearized Laplace Approximation for Bayesian Deep Learning | Pre-trained deep neural networks can be adapted to perform uncertainty estimation by transforming them into Bayesian neural networks via methods such as Laplace approximation (LA) or its linearized form (LLA), among others. To make these methods more tractable, the generalized Gauss-Newton (GGN) approximation is often used. However, due to complex inefficiency difficulties, both LA and LLA rely on further approximations, such as Kronecker-factored or diagonal approximate GGN matrices, which can affect the results. To address these issues, we propose a new method for scaling LLA using a variational sparse Gaussian Process (GP) approximation based on the dual RKHS of GPs. Our method retains the predictive mean of the original model while allowing for efficient stochastic optimization and scalability in both the number of parameters and the size of the training dataset. Moreover, its training cost is independent of the number of training points, improving over previously existing methods. Our preliminary experiments indicate that it outperforms already existing efficient variants of LLA, such as accelerated LLA (ELLA), based on the Nystr\"om approximation. | ['Daniel Hernández-Lobato', 'Simón Rodríguez Santana', 'Luis A. Ortega'] | 2023-02-24 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [-2.57526278e-01 2.25781217e-01 1.17116801e-01 -2.97296375e-01
-8.58427763e-01 -2.36610487e-01 6.37071550e-01 -1.09609663e-01
-3.94176751e-01 9.42680418e-01 -6.59368262e-02 -2.47365415e-01
-2.57636577e-01 -8.56074989e-01 -9.83087420e-01 -9.20747697e-01
7.78778568e-02 7.11009741e-01 1.54247776e-01 3.67649704e-01
5.42078577e-02 5.63718736e-01 -1.29028261e+00 -4.08949226e-01
8.62155914e-01 1.25336766e+00 3.50816287e-02 3.11080247e-01
-2.02989176e-01 5.15354156e-01 -3.15265387e-01 -4.56397146e-01
2.36988217e-01 -2.23696843e-01 -3.73445928e-01 -1.70882091e-01
1.51179358e-01 -5.82450449e-01 -4.26625758e-01 1.02293468e+00
3.63550365e-01 5.59168458e-01 8.70639741e-01 -1.02312756e+00
-3.32978427e-01 6.19144201e-01 -7.18918145e-01 -2.61055261e-01
-2.18340635e-01 -2.57873535e-01 7.84850001e-01 -9.93645787e-01
3.29230517e-01 1.42913330e+00 1.19042671e+00 1.73449188e-01
-1.51772678e+00 -4.28018600e-01 7.95685872e-02 -3.82828191e-02
-1.59799898e+00 -5.67674994e-01 6.89090610e-01 -1.76982611e-01
7.32501030e-01 -2.09348172e-01 4.72507656e-01 1.09790051e+00
-1.78488977e-02 9.72490907e-01 8.23988318e-01 -4.08577740e-01
7.88269579e-01 -5.44734336e-02 -1.09775260e-01 8.12331796e-01
3.75166267e-01 -1.58878103e-01 -4.95763332e-01 -7.02678204e-01
1.00216067e+00 -1.91013113e-01 -2.66613036e-01 -6.69702590e-01
-7.55318701e-01 1.16285157e+00 1.99024692e-01 1.09895378e-01
-4.34507757e-01 6.18118644e-01 2.60805190e-01 -3.24809730e-01
6.11845016e-01 1.69509679e-01 -4.32679236e-01 -3.16670716e-01
-1.43435574e+00 4.66151685e-01 1.07881951e+00 7.05458403e-01
8.71894121e-01 3.87826622e-01 4.18008082e-02 8.83270800e-01
7.67440021e-01 7.02805400e-01 4.27110672e-01 -1.39519358e+00
2.66975552e-01 -1.26736745e-01 2.24927336e-01 -9.82445359e-01
-3.23133320e-01 -5.88462651e-01 -1.05270457e+00 -1.04657328e-02
6.00834727e-01 -2.22198725e-01 -9.33871150e-01 1.76036334e+00
3.15369487e-01 4.72568661e-01 2.92364694e-02 5.17748475e-01
4.70743418e-01 8.46694887e-01 -3.01264822e-01 -2.70280749e-01
8.96556020e-01 -6.29703820e-01 -6.51019275e-01 -8.09452906e-02
5.15189707e-01 -5.52367508e-01 7.90006340e-01 6.33999646e-01
-1.19538939e+00 -4.26879674e-01 -8.61046970e-01 -6.07923381e-02
-1.12217993e-01 3.47209424e-01 7.11419642e-01 8.38347435e-01
-1.17279685e+00 1.05641329e+00 -1.35801196e+00 -5.77536747e-02
4.08202380e-01 3.62845391e-01 -1.20926760e-01 -1.77837629e-02
-1.07892466e+00 6.85820103e-01 5.25525093e-01 4.00625527e-01
-5.42619109e-01 -6.30392790e-01 -9.74405885e-01 3.31222266e-01
3.69946867e-01 -7.22849250e-01 1.16000235e+00 -3.21131885e-01
-2.17095470e+00 3.47816795e-02 -2.90439695e-01 -7.16886759e-01
5.90494037e-01 -4.87983733e-01 2.03887597e-01 2.68696129e-01
-6.96358308e-02 5.35831153e-01 1.15993595e+00 -1.14774692e+00
-2.60699481e-01 -3.45775783e-01 -2.54858971e-01 -8.51257667e-02
-1.53348535e-01 -4.18925583e-01 -6.02068782e-01 -5.83714843e-01
4.83451992e-01 -1.09756935e+00 -4.33350146e-01 -4.26738895e-02
-4.22346711e-01 -1.75623834e-01 4.95194733e-01 -8.42128754e-01
9.83601272e-01 -2.15517211e+00 1.99773729e-01 6.07910156e-01
4.38265428e-02 1.61424190e-01 7.81838596e-02 4.04389203e-01
2.07856029e-01 -1.34953111e-01 -6.00416660e-01 -9.89961267e-01
2.56617785e-01 6.90413594e-01 -4.32924688e-01 6.06328726e-01
-4.31146622e-02 6.10414088e-01 -7.45644331e-01 -4.41632390e-01
1.36587054e-01 7.45938301e-01 -6.11924291e-01 -1.05579361e-01
-2.60473520e-01 3.29599619e-01 -3.20349991e-01 2.85971731e-01
7.59960830e-01 -2.29847342e-01 2.89657488e-02 -2.97106266e-01
1.33985505e-01 1.22067653e-01 -1.59676099e+00 1.59497881e+00
-6.98967159e-01 4.59583640e-01 1.56793192e-01 -1.13856769e+00
8.67119074e-01 2.83536226e-01 3.80798817e-01 1.32432684e-01
5.25742769e-02 4.56279099e-01 -4.54061389e-01 4.44896854e-02
4.36983764e-01 -1.75472111e-01 2.65412599e-01 2.52258718e-01
1.53965667e-01 -1.55337602e-01 3.63616854e-01 1.95820183e-01
7.03585923e-01 5.90983570e-01 3.00740123e-01 -2.45223150e-01
3.49964291e-01 -4.88002241e-01 6.13205433e-01 1.04329896e+00
2.03499392e-01 7.14224339e-01 5.95847070e-01 -3.16551507e-01
-8.38829279e-01 -1.16632879e+00 -3.99827033e-01 6.47737920e-01
-3.67577255e-01 -3.30900431e-01 -8.58492672e-01 -4.12436128e-01
5.97636178e-02 1.11415160e+00 -3.63111883e-01 7.12602064e-02
-4.49384898e-01 -8.76844466e-01 6.15436137e-01 5.69255054e-01
5.68978906e-01 -4.94097143e-01 -2.30029091e-01 4.36706871e-01
-6.13945834e-02 -1.25508344e+00 3.48673612e-02 2.46131063e-01
-1.22733402e+00 -4.48738694e-01 -9.06288803e-01 -5.12972362e-02
5.12826681e-01 -2.80801296e-01 7.00136542e-01 -5.17217755e-01
3.40069205e-01 1.72110751e-01 3.18958797e-02 -2.23682627e-01
-3.41321856e-01 1.07442699e-01 3.66559029e-01 1.04480706e-01
7.96837639e-03 -1.10281062e+00 -2.89914191e-01 -1.34515896e-01
-8.18041801e-01 -2.05912054e-01 6.93705797e-01 9.51768041e-01
6.29031360e-01 2.04878479e-01 2.13882938e-01 -7.55796909e-01
6.59519911e-01 -2.41710216e-01 -9.83149707e-01 -5.61334155e-02
-7.16084599e-01 6.41067863e-01 7.25922585e-01 -3.60838056e-01
-1.22911024e+00 1.53346077e-01 -3.26248080e-01 -8.25161934e-01
5.87318353e-02 7.65498281e-01 1.28484845e-01 -1.82891637e-01
5.17202854e-01 3.21898311e-01 6.61599711e-02 -6.12624228e-01
5.68812191e-01 3.74073684e-01 5.48049390e-01 -6.90520585e-01
7.35930562e-01 6.49662852e-01 4.95290518e-01 -8.18758726e-01
-9.19741452e-01 -1.44356355e-01 -4.49444354e-01 1.19857155e-01
5.19070148e-01 -7.97119439e-01 -6.50532544e-01 3.79898965e-01
-1.12149382e+00 -1.14090085e-01 -2.75044680e-01 8.65687072e-01
-6.78399205e-01 7.70755172e-01 -8.84927809e-01 -1.00577545e+00
-1.37657493e-01 -1.19329536e+00 1.13177860e+00 -5.13369683e-03
-1.32310659e-01 -1.15552390e+00 -9.66574810e-03 1.18874080e-01
3.70687038e-01 1.57259926e-01 8.70879054e-01 -5.95410883e-01
-5.10614812e-01 -3.07023674e-01 -1.91630349e-01 6.77724779e-01
-2.19393462e-01 2.70485040e-02 -8.53127956e-01 -2.48432830e-01
4.94061798e-01 -1.27467424e-01 9.20856059e-01 8.83356869e-01
9.70074594e-01 -3.73000950e-01 -2.38072336e-01 7.76170671e-01
1.20118320e+00 -1.42733395e-01 4.34446067e-01 1.45163730e-01
6.64922416e-01 2.17766568e-01 1.48428068e-01 6.16401315e-01
2.47243941e-01 3.63905907e-01 2.09252357e-01 4.07655388e-01
3.23116839e-01 -2.65826166e-01 3.38484764e-01 9.31873262e-01
-3.87145162e-01 1.53668284e-01 -8.65555882e-01 2.05027699e-01
-2.03132725e+00 -5.99448264e-01 -6.92151785e-02 2.34445024e+00
7.33973205e-01 1.11509621e-01 -1.66240573e-01 1.68628231e-01
3.03875029e-01 2.04615578e-01 -5.43056071e-01 -2.61197358e-01
5.99560924e-02 2.69743800e-01 7.09726334e-01 5.89951992e-01
-9.85816181e-01 7.83504188e-01 6.41258717e+00 1.35052419e+00
-7.69498944e-01 2.73160368e-01 6.17160976e-01 -1.46095410e-01
-1.78722978e-01 7.30166063e-02 -9.45810854e-01 3.58495116e-01
1.13489318e+00 1.62805974e-01 5.31579018e-01 1.11869919e+00
1.69650540e-01 -4.10066485e-01 -8.75942528e-01 1.13616276e+00
-5.64248860e-02 -1.25874615e+00 -2.11150181e-02 2.18551069e-01
8.08658481e-01 8.25241655e-02 9.12011135e-03 5.35279095e-01
4.27424490e-01 -7.22589612e-01 6.45913899e-01 5.76856971e-01
3.43846172e-01 -1.04719293e+00 8.59056950e-01 3.91504675e-01
-8.24008405e-01 1.81568623e-01 -7.03354955e-01 9.85895321e-02
3.57792258e-01 1.17266226e+00 -6.19672298e-01 6.79807186e-01
6.28151059e-01 2.61835337e-01 -1.09262586e-01 7.48715878e-01
-2.27504283e-01 7.55494118e-01 -1.00840259e+00 1.21643238e-01
4.67209548e-01 -7.26445794e-01 7.38355577e-01 7.79097319e-01
8.36822808e-01 -1.80034786e-01 -2.10561771e-02 1.10593867e+00
1.13265000e-01 -9.62449983e-02 -2.31840432e-01 -1.53348714e-01
2.39395276e-01 1.06602919e+00 -8.28110278e-01 -2.77153730e-01
-4.08822566e-01 6.69633746e-01 2.96213686e-01 7.17528760e-01
-7.74898231e-01 -1.02325335e-01 4.25645918e-01 -4.37660776e-02
7.12478518e-01 -5.29351354e-01 -4.85914014e-02 -1.15921748e+00
-1.44791277e-03 -6.67680740e-01 7.11427853e-02 -6.48759782e-01
-1.13530254e+00 4.95901823e-01 2.43412435e-01 -8.53816152e-01
-7.33391702e-01 -5.95676243e-01 -9.88224670e-02 8.19675207e-01
-1.26173842e+00 -1.04415035e+00 3.15739989e-01 3.33866000e-01
2.10338667e-01 -1.63717791e-01 6.09776080e-01 -5.61061734e-03
-4.94765282e-01 3.97732645e-01 7.94185102e-01 -2.36554101e-01
4.52397764e-01 -1.15479863e+00 2.45774597e-01 8.54978263e-01
3.70973766e-01 7.22217917e-01 9.59507704e-01 -4.89894211e-01
-1.20061135e+00 -7.64972389e-01 6.73422635e-01 -8.23115185e-02
7.33219624e-01 -2.38581985e-01 -9.60807562e-01 7.25314558e-01
-3.08922201e-01 -2.89472868e-03 5.78946829e-01 3.82312387e-01
-1.82471618e-01 -9.28249806e-02 -9.98196840e-01 6.31242812e-01
5.60755134e-01 -4.02423531e-01 -3.70426804e-01 3.47119331e-01
4.86499339e-01 -5.46261072e-01 -9.39654589e-01 4.84761238e-01
4.11139518e-01 -1.20684624e+00 8.90631497e-01 -9.11601558e-02
2.63627052e-01 -2.98836946e-01 -2.31337801e-01 -1.25015020e+00
-2.39501014e-01 -7.08240092e-01 -8.58267486e-01 1.05313647e+00
8.90893489e-02 -7.87727535e-01 7.92261839e-01 7.40121841e-01
2.12832056e-02 -9.41447854e-01 -1.21446013e+00 -9.89627838e-01
-4.25801352e-02 -7.65811741e-01 4.69859868e-01 5.79401970e-01
-4.79133308e-01 1.29576221e-01 -5.29468238e-01 4.46104169e-01
8.55585158e-01 2.43381541e-02 6.83051646e-01 -1.38234544e+00
-7.80130565e-01 -3.03837836e-01 -2.19447047e-01 -1.24874485e+00
3.51538420e-01 -5.46497166e-01 9.03774649e-02 -1.34928691e+00
-1.69431940e-01 -2.63093919e-01 2.30507236e-02 1.85284600e-01
4.32103425e-02 5.48574254e-02 5.82293347e-02 2.26874501e-01
-2.41127327e-01 9.70688820e-01 7.24250853e-01 3.06123823e-01
-2.98468411e-01 2.43958473e-01 -1.93395600e-01 1.26449203e+00
5.90422571e-01 -5.27517676e-01 -4.53980654e-01 -2.24080428e-01
3.45169336e-01 1.46456778e-01 1.61356196e-01 -1.26951051e+00
2.75721759e-01 2.23983869e-01 4.06573892e-01 -7.19377637e-01
7.22370803e-01 -5.35920799e-01 2.49386668e-01 2.73779273e-01
-5.23759127e-02 -3.59336466e-01 1.61295190e-01 8.52301240e-01
-3.75515431e-01 -7.85349607e-01 7.36728191e-01 -3.34509462e-02
-1.73144847e-01 3.19960892e-01 -4.99344409e-01 -2.90281832e-01
3.87279630e-01 -5.16682304e-02 3.17518771e-01 -8.58276725e-01
-8.46771181e-01 -1.47059277e-01 2.33873382e-01 -3.94556135e-01
3.98270875e-01 -1.24240315e+00 -2.84003824e-01 9.30304229e-02
-4.78730142e-01 4.69363183e-01 2.39678368e-01 1.03755736e+00
-5.39780378e-01 5.52319467e-01 3.83924365e-01 -5.98798931e-01
-6.43142343e-01 3.65571439e-01 1.16839848e-01 -5.46791255e-01
-5.13592184e-01 8.06579590e-01 1.16643488e-01 -5.80895782e-01
3.14255774e-01 -4.63202804e-01 6.88030720e-02 -2.02155840e-02
3.33903879e-01 5.61586022e-01 -1.01579785e-01 -4.74408865e-01
7.14328662e-02 3.20128232e-01 1.41712338e-01 -3.50794643e-01
1.39441085e+00 -2.47092381e-01 -3.56375098e-01 6.90299153e-01
1.09044409e+00 8.05001054e-03 -1.39972878e+00 -4.62803721e-01
-9.94973853e-02 -2.13116065e-01 4.29430574e-01 -2.36994796e-03
-1.01975262e+00 9.09788847e-01 3.25558782e-01 -8.67765397e-03
8.21831107e-01 -2.40475059e-01 7.91455865e-01 9.53768551e-01
3.70595664e-01 -1.14415419e+00 -2.95715988e-01 5.76141179e-01
7.46292233e-01 -9.32007253e-01 2.32498065e-01 -2.48483941e-01
-2.09141403e-01 1.25484312e+00 1.17201572e-02 -1.90914258e-01
8.26650023e-01 6.49853125e-02 -1.75344422e-01 2.06733942e-01
-2.79335439e-01 -2.84665488e-02 1.18500561e-01 3.95054966e-01
8.08778182e-02 -3.08217287e-01 -1.97900876e-01 4.99670476e-01
-2.30528161e-01 6.07701018e-02 2.90029258e-01 5.78496754e-01
-1.87748492e-01 -1.03618658e+00 -4.83032376e-01 5.20992219e-01
-4.71476465e-01 -2.77623832e-01 4.22679126e-01 7.43851542e-01
-2.34296143e-01 5.71012914e-01 -4.48164903e-02 2.17623964e-01
-2.21286178e-01 3.47275436e-01 4.57797617e-01 -2.57367879e-01
2.10730910e-01 3.65704417e-01 5.23802266e-02 -6.29338503e-01
-4.32585657e-01 -7.99639106e-01 -1.01302707e+00 -3.25622767e-01
-4.19707566e-01 1.17565572e-01 7.83706844e-01 1.32077837e+00
2.98853010e-01 9.95537639e-02 8.13361332e-02 -1.19472969e+00
-1.06956863e+00 -1.03203666e+00 -9.25835729e-01 -1.26222323e-03
-5.42113446e-02 -7.97673106e-01 -5.22484362e-01 -1.11948177e-01] | [7.115846633911133, 3.8220083713531494] |
396d59fe-c88f-44a4-9544-1f5e625df077 | depth-quality-inspired-feature-manipulation | 2107.01779 | null | https://arxiv.org/abs/2107.01779v2 | https://arxiv.org/pdf/2107.01779v2.pdf | Depth Quality-Inspired Feature Manipulation for Efficient RGB-D Salient Object Detection | RGB-D salient object detection (SOD) recently has attracted increasing research interest by benefiting conventional RGB SOD with extra depth information. However, existing RGB-D SOD models often fail to perform well in terms of both efficiency and accuracy, which hinders their potential applications on mobile devices and real-world problems. An underlying challenge is that the model accuracy usually degrades when the model is simplified to have few parameters. To tackle this dilemma and also inspired by the fact that depth quality is a key factor influencing the accuracy, we propose a novel depth quality-inspired feature manipulation (DQFM) process, which is efficient itself and can serve as a gating mechanism for filtering depth features to greatly boost the accuracy. DQFM resorts to the alignment of low-level RGB and depth features, as well as holistic attention of the depth stream to explicitly control and enhance cross-modal fusion. We embed DQFM to obtain an efficient light-weight model called DFM-Net, where we also design a tailored depth backbone and a two-stage decoder for further efficiency consideration. Extensive experimental results demonstrate that our DFM-Net achieves state-of-the-art accuracy when comparing to existing non-efficient models, and meanwhile runs at 140ms on CPU (2.2$\times$ faster than the prior fastest efficient model) with only $\sim$8.5Mb model size (14.9% of the prior lightest). Our code will be available at https://github.com/zwbx/DFM-Net. | ['Qijun Zhao', 'Keren Fu', 'Zhuo Wang', 'Ge-Peng Ji', 'Wenbo Zhang'] | 2021-07-05 | null | null | null | null | ['rgb-d-salient-object-detection'] | ['computer-vision'] | [ 3.29148561e-01 -1.38318673e-01 -1.85225189e-01 -3.53613168e-01
-7.22259521e-01 -3.11905202e-02 3.14657599e-01 -2.10091360e-02
-5.27698100e-01 1.85813218e-01 1.66081369e-01 -1.70786187e-01
1.81789711e-01 -8.81608903e-01 -5.62319458e-01 -7.47775793e-01
2.54236132e-01 -1.30218804e-01 6.76167667e-01 -2.67832100e-01
1.30663976e-01 3.29595327e-01 -1.70353436e+00 8.59658346e-02
9.50331092e-01 1.47696066e+00 5.79893172e-01 4.37659681e-01
-2.47360431e-02 6.11652076e-01 -1.43542722e-01 -4.26573545e-01
3.46889675e-01 -1.79487988e-01 -4.64071572e-01 4.16127071e-02
2.72750884e-01 -6.24563217e-01 -6.62512362e-01 1.04549181e+00
8.08407128e-01 -1.01531157e-03 1.90115109e-01 -1.13525426e+00
-4.16223645e-01 1.02694787e-01 -9.33187008e-01 2.46800587e-01
1.20619610e-01 3.82348508e-01 9.03655887e-01 -1.01415265e+00
3.28474253e-01 1.14835370e+00 4.21463847e-01 5.90615511e-01
-9.00253117e-01 -8.15727115e-01 3.56245399e-01 3.04169416e-01
-1.39977992e+00 -5.00036180e-01 8.73341560e-01 9.85834748e-02
9.25927222e-01 1.35800377e-01 9.76497650e-01 7.14758337e-01
4.87069739e-03 1.11533606e+00 9.04389203e-01 -2.62104183e-01
2.24998325e-01 -2.66847998e-01 -1.18240930e-01 8.91652286e-01
2.89144188e-01 3.10636293e-02 -8.64171326e-01 1.94508761e-01
8.30350459e-01 1.62603557e-01 -3.77167881e-01 -3.50034714e-01
-1.04329216e+00 6.15854859e-01 9.18482661e-01 1.01191979e-02
-1.89075455e-01 5.02897918e-01 1.32094055e-01 -1.10783279e-01
3.48212779e-01 -3.08124423e-02 -3.64511222e-01 -3.40035737e-01
-7.92245865e-01 1.78149581e-01 1.26738071e-01 9.67371464e-01
8.84982765e-01 -1.78670436e-01 8.08731988e-02 6.47647858e-01
5.86836636e-01 6.93267941e-01 3.32936525e-01 -9.62623537e-01
5.77679753e-01 8.85783792e-01 9.52927163e-04 -9.82783735e-01
-5.79320312e-01 -4.82221365e-01 -8.95207524e-01 1.23338811e-01
2.46085629e-01 2.32347384e-01 -8.48092258e-01 1.59912765e+00
6.04735672e-01 9.20266286e-02 -1.54532418e-01 1.21173024e+00
9.10739183e-01 5.98211706e-01 -4.39764522e-02 -3.69747728e-02
1.44715571e+00 -9.07993972e-01 -4.70545471e-01 -5.66189110e-01
6.02688015e-01 -6.64816022e-01 1.28558075e+00 4.44176495e-01
-1.16137314e+00 -5.52154720e-01 -1.17642498e+00 -6.68444395e-01
3.08876042e-03 1.55295610e-01 9.30745780e-01 5.50656378e-01
-9.88517046e-01 2.97203928e-01 -1.21949220e+00 -1.37160212e-01
6.91127658e-01 6.36274517e-01 -4.21191156e-02 -2.92813689e-01
-9.67170179e-01 5.64638317e-01 1.54740110e-01 3.34250093e-01
-5.52147746e-01 -6.16728067e-01 -8.19531262e-01 -9.73562524e-02
3.39367241e-01 -8.37150812e-01 1.30918622e+00 -5.88234723e-01
-1.52443385e+00 7.45215356e-01 -4.04404402e-01 -2.63806373e-01
3.97301018e-01 -3.68599087e-01 -2.27388203e-01 3.68994743e-01
-7.89559931e-02 9.20794129e-01 7.26577938e-01 -1.02574694e+00
-1.00680268e+00 -5.71320713e-01 3.08374703e-01 3.91193211e-01
-6.22369051e-01 -2.45736584e-01 -1.08466160e+00 -4.69332218e-01
5.63985765e-01 -8.11275721e-01 -3.04550201e-01 5.83294988e-01
-1.84907272e-01 2.64381245e-02 5.89974999e-01 -3.41788411e-01
1.34994054e+00 -2.30932426e+00 -3.85745466e-02 -2.46269275e-02
4.72649217e-01 3.01401764e-01 8.65172520e-02 8.67119357e-02
4.83461231e-01 -2.41361246e-01 -3.15350324e-01 -6.78478718e-01
4.27474901e-02 1.54802114e-01 -6.56348541e-02 6.68530285e-01
1.52323529e-01 9.93203640e-01 -9.01041389e-01 -5.06923974e-01
4.87714231e-01 6.86376274e-01 -7.35639930e-01 -3.26682553e-02
-6.07527792e-02 2.42314860e-01 -6.44341052e-01 9.18807805e-01
9.85845506e-01 -4.98403668e-01 -5.61906546e-02 -5.62529743e-01
-2.52070516e-01 3.98376793e-01 -1.19604087e+00 2.10337400e+00
-4.98508424e-01 3.94996732e-01 1.00106737e-02 -5.78298211e-01
7.82817662e-01 -2.38062114e-01 3.74768466e-01 -1.21156919e+00
3.82596761e-01 4.23790693e-01 -4.29517478e-02 -2.22000152e-01
6.14874005e-01 3.09545584e-02 -3.69798392e-02 1.93764880e-01
-2.60620058e-01 -1.97353452e-01 -1.56580061e-01 2.51437277e-01
1.02715695e+00 1.46120012e-01 1.87497139e-01 -4.68304381e-02
4.06773686e-01 -1.84450120e-01 6.80743635e-01 4.16202813e-01
-3.34809303e-01 6.29866779e-01 2.18406379e-01 -1.99517235e-01
-6.19413555e-01 -8.90242338e-01 -1.16821170e-01 8.03054571e-01
8.65541101e-01 -4.83093113e-01 -4.97321159e-01 -3.23433787e-01
-3.56201530e-02 2.74285823e-01 -4.22570795e-01 -4.61985409e-01
-5.11759698e-01 -8.32706630e-01 2.34703019e-01 6.88778818e-01
9.46662664e-01 -6.14618897e-01 -1.13443339e+00 1.72493339e-01
-2.20639467e-01 -1.14200044e+00 -2.94140816e-01 2.90602297e-01
-9.93448257e-01 -8.32911193e-01 -5.61608076e-01 -5.64836442e-01
5.52074313e-01 8.22506070e-01 8.76904249e-01 3.02302003e-01
-2.39001274e-01 6.91558421e-02 -4.13366616e-01 -3.01135600e-01
3.24163437e-01 3.26940507e-01 -1.44129932e-01 -2.15950627e-02
4.53842163e-01 -6.36167943e-01 -1.30406010e+00 2.24729314e-01
-1.05666471e+00 5.05356252e-01 7.65856624e-01 6.38540745e-01
7.57144690e-01 -1.25949159e-01 2.36275420e-01 -3.52718949e-01
-1.05682388e-01 -1.78687513e-01 -5.98634064e-01 -1.81307986e-01
-5.51006854e-01 -3.63726541e-02 1.96265250e-01 -2.17246845e-01
-9.37568784e-01 2.41611794e-01 -4.12140191e-01 -3.45601737e-01
2.56723851e-01 1.57227501e-01 -4.71445471e-01 -2.70576000e-01
1.85574144e-01 2.39247948e-01 -6.60464913e-02 -4.28260267e-01
3.03204656e-01 6.86094820e-01 4.25938755e-01 -3.24921787e-01
6.96548462e-01 9.08601820e-01 9.44603011e-02 -6.29797161e-01
-8.38311672e-01 -4.88199025e-01 -2.86222607e-01 -1.53094232e-01
5.32527745e-01 -1.34834528e+00 -9.10941184e-01 6.76983654e-01
-9.46128130e-01 -4.50046211e-01 -5.91891818e-02 4.50915009e-01
-3.34090203e-01 3.29437882e-01 -6.28367066e-01 -7.49530971e-01
-4.52663243e-01 -1.21732318e+00 1.36620927e+00 3.78304780e-01
1.81649283e-01 -4.90351200e-01 -4.79056448e-01 3.77177328e-01
4.57233489e-01 5.29837646e-02 5.41069150e-01 2.84209549e-01
-9.75489140e-01 3.22699957e-02 -7.01867282e-01 7.35936761e-02
-3.06620691e-02 -3.18766832e-01 -1.15805233e+00 -2.34952450e-01
2.66306866e-02 -1.73204362e-01 1.00791359e+00 3.96476805e-01
1.21848512e+00 1.88115388e-01 -4.31051940e-01 9.25542355e-01
1.60782671e+00 -1.49099238e-03 6.29612446e-01 4.79339153e-01
9.09397006e-01 2.22688153e-01 8.75539184e-01 7.52702177e-01
8.56665552e-01 8.26187253e-01 8.97606313e-01 -2.66568899e-01
-3.27377230e-01 -1.88188270e-01 3.04541260e-01 8.39290142e-01
3.90563980e-02 -2.13448063e-01 -7.96280503e-01 4.45303410e-01
-1.83483124e+00 -5.33563197e-01 -2.70116031e-01 2.11127329e+00
8.30226839e-01 4.84702587e-01 1.97844375e-02 3.51585269e-01
3.46761376e-01 2.60807157e-01 -8.06326091e-01 1.97668783e-02
-1.61845192e-01 2.44432345e-01 7.06956804e-01 4.39189345e-01
-1.04220915e+00 8.21868122e-01 4.62458658e+00 9.88667607e-01
-1.17439497e+00 2.67055839e-01 5.48343837e-01 -5.49524903e-01
-3.63491178e-01 -4.68593799e-02 -9.83418345e-01 5.47916353e-01
5.15631616e-01 1.68038547e-01 1.88715339e-01 7.69481957e-01
3.75246137e-01 -5.09780109e-01 -7.89953530e-01 1.43180192e+00
-8.81705210e-02 -1.10491920e+00 -1.44070774e-01 1.92937598e-01
5.02561867e-01 1.66970819e-01 1.06313854e-01 1.56280622e-02
-1.71534345e-01 -5.70704222e-01 1.04565430e+00 2.65470237e-01
7.97894180e-01 -8.74028087e-01 6.02694154e-01 2.03476965e-01
-1.52059352e+00 -1.61844626e-01 -4.52013016e-01 -2.02815533e-01
2.25171089e-01 9.25366938e-01 -1.35069236e-01 4.32895392e-01
9.92269337e-01 8.62324536e-01 -5.51883280e-01 9.80335236e-01
-2.14925215e-01 1.56484365e-01 -5.71945488e-01 -8.65226686e-02
2.33311817e-01 1.41146198e-01 1.42840117e-01 9.08576131e-01
3.76290530e-01 3.22694421e-01 -6.17859699e-02 5.11348605e-01
-9.24111679e-02 -1.29616663e-01 -8.08642954e-02 3.65911990e-01
6.02144301e-01 1.22879982e+00 -8.96570802e-01 -2.02138737e-01
-6.86653316e-01 1.16658890e+00 2.01303422e-01 -1.43638009e-03
-1.10764408e+00 -3.99268985e-01 8.78908753e-01 3.21166128e-01
5.60673833e-01 -2.96195656e-01 -3.81830931e-01 -1.20524573e+00
3.47808629e-01 -5.65555573e-01 9.17944089e-02 -8.31437230e-01
-8.24724138e-01 5.46829522e-01 -3.95509988e-01 -1.53401005e+00
4.00676638e-01 -6.11783743e-01 -5.51074073e-02 6.29781127e-01
-2.11658454e+00 -1.20627594e+00 -8.16320896e-01 6.76964998e-01
3.81450176e-01 5.49797833e-01 3.56902301e-01 6.63851142e-01
-5.58132291e-01 6.50546134e-01 -1.05474547e-01 -2.03660175e-01
5.43427646e-01 -8.50204945e-01 5.35493076e-01 9.72878993e-01
-1.49366900e-01 3.86856765e-01 4.36062753e-01 -3.44807833e-01
-1.92317796e+00 -1.02935398e+00 6.82586491e-01 -1.48560047e-01
4.13307279e-01 -5.09788036e-01 -6.43997848e-01 1.08327843e-01
-3.00016552e-01 4.45502311e-01 4.22510624e-01 -3.68947595e-01
-1.47444382e-01 -4.26851213e-01 -9.20455515e-01 6.55744851e-01
1.61525035e+00 -5.00177264e-01 -1.21933788e-01 -2.45337822e-02
8.81376445e-01 -6.96472764e-01 -6.30754411e-01 4.66691136e-01
6.36174798e-01 -1.30484200e+00 1.06570065e+00 3.40839744e-01
4.40177858e-01 -6.41673028e-01 -4.75901663e-01 -6.37584209e-01
-2.32779339e-01 -4.93367583e-01 -5.66466510e-01 1.11092079e+00
8.43643304e-03 -4.76040602e-01 8.53241563e-01 6.31424308e-01
-2.77785748e-01 -1.27748990e+00 -1.04007876e+00 -4.63460147e-01
-4.93372232e-01 -8.66397142e-01 6.51222885e-01 4.27483141e-01
-2.45662734e-01 1.38486192e-01 -1.55182108e-01 3.52645874e-01
5.78105569e-01 2.83774912e-01 8.03735435e-01 -9.69005823e-01
-2.12638155e-01 -5.36847830e-01 -5.57865322e-01 -1.80647135e+00
-4.52458620e-01 -5.67061841e-01 4.01832722e-02 -1.65958905e+00
8.78407434e-02 -7.09307075e-01 -2.83265263e-01 4.53429192e-01
-2.96072185e-01 6.40758574e-01 2.41293624e-01 1.57955989e-01
-8.63416851e-01 9.72103059e-01 1.26925254e+00 7.36260116e-02
-2.06888333e-01 -1.50136188e-01 -9.52828109e-01 7.28780568e-01
6.07276499e-01 -3.24520350e-01 -4.18920875e-01 -7.28486776e-01
3.69405776e-01 -1.66377679e-01 5.18055439e-01 -1.18943393e+00
3.21709275e-01 9.40036476e-02 4.65515465e-01 -7.26120174e-01
6.51903510e-01 -8.21923316e-01 -2.31812105e-01 5.30630410e-01
2.65804768e-01 1.15674017e-02 2.05888003e-01 5.17167687e-01
-2.10065678e-01 1.52388379e-01 7.68740714e-01 6.38409480e-02
-1.13258529e+00 5.68591475e-01 5.48083298e-02 -1.48318723e-01
8.71667385e-01 -5.18147171e-01 -3.14476609e-01 -1.96574211e-01
-2.19245285e-01 2.97925353e-01 7.96733081e-01 3.18523645e-01
7.73580074e-01 -1.33858144e+00 -2.99103081e-01 2.89754122e-01
2.32323915e-01 4.69822109e-01 5.41749239e-01 1.07518077e+00
-5.31002641e-01 2.29250386e-01 1.12338595e-01 -8.06750357e-01
-9.48623180e-01 3.64833623e-01 1.43928677e-01 -1.32627347e-02
-7.29614377e-01 1.13812602e+00 2.81055897e-01 9.35499221e-02
3.48525763e-01 -6.79546893e-01 3.39528650e-01 -2.21679099e-02
6.39001429e-01 2.93053448e-01 2.42106885e-01 -5.28224051e-01
-6.14034116e-01 8.12191129e-01 3.27087566e-02 7.69704133e-02
1.28951371e+00 -5.49160123e-01 1.23931669e-01 2.23755762e-01
1.26733494e+00 -1.41453519e-01 -1.65573680e+00 -4.56500679e-01
-3.99945706e-01 -7.57049859e-01 5.68445206e-01 -4.14632142e-01
-1.29016674e+00 9.98196185e-01 8.53516400e-01 -1.88511848e-01
1.66776466e+00 3.04971021e-02 1.20284486e+00 3.10263969e-02
6.15594923e-01 -9.18503881e-01 1.80935785e-01 1.88844651e-01
4.84530687e-01 -1.42110598e+00 2.53121018e-01 -6.09310031e-01
-3.98855329e-01 8.41796696e-01 7.34926939e-01 4.22740442e-04
6.21817350e-01 3.02319616e-01 -1.19288087e-01 -1.70807496e-01
-4.95485872e-01 -4.47078258e-01 1.06245443e-01 3.17298770e-01
1.73385769e-01 -1.63437814e-01 -1.68124557e-01 6.28552735e-01
-5.84048368e-02 1.14346765e-01 3.34036574e-02 1.18815768e+00
-4.70120847e-01 -9.50244009e-01 -6.19554035e-02 4.00169164e-01
-2.58521855e-01 -2.43567035e-01 2.29921564e-01 7.17199266e-01
4.33191180e-01 8.08397651e-01 7.04716668e-02 -6.07945025e-01
3.50427568e-01 -5.22790015e-01 5.50322413e-01 -3.28451931e-01
-4.08494741e-01 1.56235412e-01 -2.97720343e-01 -9.58152354e-01
-7.47701466e-01 -6.02095842e-01 -1.48995352e+00 -5.25509298e-01
-4.69268799e-01 -3.94946843e-01 5.87431669e-01 7.68238902e-01
5.86156487e-01 4.50217068e-01 6.29507959e-01 -1.05694461e+00
-1.15691990e-01 -6.73528194e-01 -4.39242989e-01 3.94952781e-02
4.44274276e-01 -8.87456417e-01 -1.86605006e-01 -2.41991803e-01] | [9.582967758178711, -0.8852068185806274] |
e1ebc651-fdb1-4e32-adf2-978a5d0c6015 | improving-language-understanding-by | null | null | https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/language-unsupervised/language_understanding_paper.pdf | https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/language-unsupervised/language_understanding_paper.pdf | Improving Language Understanding by Generative Pre-Training | Natural language understanding comprises a wide range of diverse tasks such
as textual entailment, question answering, semantic similarity assessment, and
document classification. Although large unlabeled text corpora are abundant,
labeled data for learning these specific tasks is scarce, making it challenging for
discriminatively trained models to perform adequately. We demonstrate that large
gains on these tasks can be realized by generative pre-training of a language model
on a diverse corpus of unlabeled text, followed by discriminative fine-tuning on each
specific task. In contrast to previous approaches, we make use of task-aware input
transformations during fine-tuning to achieve effective transfer while requiring
minimal changes to the model architecture. We demonstrate the effectiveness of
our approach on a wide range of benchmarks for natural language understanding.
Our general task-agnostic model outperforms discriminatively trained models that
use architectures specifically crafted for each task, significantly improving upon the
state of the art in 9 out of the 12 tasks studied. For instance, we achieve absolute
improvements of 8.9% on commonsense reasoning (Stories Cloze Test), 5.7% on
question answering (RACE), and 1.5% on textual entailment (MultiNLI). | ['Tim Salimans', 'Ilya Sutskever', 'Alec Radford', 'Karthik Narasimhan'] | 2018-06-11 | null | null | null | preprint-2018-6 | ['cloze-test'] | ['natural-language-processing'] | [ 5.13840377e-01 1.95960701e-01 -1.43730208e-01 -8.07454705e-01
-1.37328804e+00 -8.70564997e-01 9.72830296e-01 3.50002795e-01
-5.93756616e-01 7.10492194e-01 4.31590497e-01 -5.20246625e-01
1.20440885e-01 -6.56164646e-01 -8.82490635e-01 -1.95169374e-02
5.11925340e-01 8.90799999e-01 1.08756661e-01 -4.70832080e-01
4.64914620e-01 -7.55722150e-02 -1.20967162e+00 5.99560320e-01
1.09597909e+00 9.11941469e-01 -5.01162224e-02 7.22670078e-01
-2.60882020e-01 1.13218212e+00 -5.07409215e-01 -6.48435950e-01
-1.34425133e-01 -4.61276531e-01 -1.36870515e+00 -1.03806578e-01
7.23197758e-01 -5.49055338e-01 -1.74270943e-01 7.00300574e-01
3.20231020e-01 3.49528253e-01 1.07603395e+00 -8.90155792e-01
-1.02122378e+00 7.97983527e-01 -3.29684049e-01 2.97951341e-01
4.49289083e-01 1.63538992e-01 1.58443201e+00 -1.00574446e+00
4.62565750e-01 1.33427334e+00 4.78174806e-01 6.29149735e-01
-1.33439553e+00 -5.07594049e-01 5.30260091e-04 3.33588839e-01
-8.93494964e-01 -6.40624583e-01 5.42941272e-01 -4.01779532e-01
1.33777344e+00 1.17571153e-01 -2.10892409e-01 1.11101973e+00
-4.88678366e-02 9.77455914e-01 1.03539169e+00 -4.62186962e-01
1.15939021e-01 4.77128811e-02 5.60629368e-01 5.99561810e-01
9.08933729e-02 -3.90453994e-01 -2.99638897e-01 -3.83847617e-02
1.71354786e-01 -9.41166803e-02 -7.82035217e-02 -1.24396503e-01
-1.22140110e+00 1.05255759e+00 4.46075886e-01 2.63971806e-01
-1.02278836e-01 2.00295791e-01 6.57518506e-01 6.31510735e-01
6.29958212e-01 8.76130104e-01 -4.63806540e-01 -2.25502580e-01
-8.06597054e-01 4.14049655e-01 1.10393405e+00 7.50011206e-01
7.99717665e-01 -8.54046196e-02 -3.74466836e-01 9.49817479e-01
-1.12103343e-01 5.19363225e-01 5.93221426e-01 -9.81252968e-01
9.36220229e-01 7.41359055e-01 -5.41345179e-02 -6.86097026e-01
-3.18761349e-01 -3.36182624e-01 -4.66861993e-01 -2.54712254e-01
6.57683909e-01 -1.37255594e-01 -9.31419373e-01 1.78793252e+00
-7.41425827e-02 -1.29322127e-01 3.13961297e-01 5.34003615e-01
9.53327656e-01 6.65904939e-01 2.47860253e-01 2.51634568e-01
1.37342393e+00 -1.23382747e+00 -3.66944790e-01 -7.80915618e-01
9.44429457e-01 -8.49691689e-01 1.65164983e+00 2.32812002e-01
-1.07306576e+00 -4.17299300e-01 -8.97791445e-01 -6.04578197e-01
-2.70442724e-01 -2.03674555e-01 6.09352827e-01 3.94549727e-01
-7.73680627e-01 2.03483149e-01 -4.24325645e-01 -3.32926661e-01
5.74196100e-01 8.32230523e-02 -2.48664021e-01 -6.48124635e-01
-1.12964427e+00 1.22142339e+00 3.51080626e-01 -4.55474794e-01
-1.19039214e+00 -8.81361723e-01 -8.96748424e-01 2.99539715e-01
4.79198128e-01 -9.19595361e-01 1.70154774e+00 -8.24020505e-01
-1.20604098e+00 1.26391780e+00 -1.88777670e-01 -6.81627691e-01
5.18503010e-01 -6.05239332e-01 -7.18410239e-02 4.40129004e-02
2.86430895e-01 5.32718658e-01 6.15494549e-01 -8.17155480e-01
-3.49619448e-01 -2.03835025e-01 5.02817392e-01 3.60535711e-01
-5.41468322e-01 1.24802338e-02 -1.24372348e-01 -5.30980766e-01
-2.85887897e-01 -7.83935547e-01 -4.22806628e-02 -3.52005482e-01
-1.97276458e-01 -5.80654681e-01 7.31271386e-01 -7.47623980e-01
8.51151526e-01 -1.84477699e+00 5.88593073e-02 -3.11972797e-01
1.80001363e-01 1.67830586e-01 -4.68212575e-01 4.25268412e-01
8.43445733e-02 1.26873642e-01 -5.02233207e-01 -3.53867859e-01
3.08835834e-01 2.74213582e-01 -5.57915926e-01 7.22179413e-02
4.42771703e-01 1.32049036e+00 -9.81913686e-01 -1.52083293e-01
-8.46756771e-02 -6.23826347e-02 -9.04712379e-01 4.16635811e-01
-7.30793715e-01 2.82413632e-01 -4.91239130e-01 3.74532044e-01
1.44074470e-01 -5.65481544e-01 -9.92735550e-02 6.26847297e-02
6.11050367e-01 8.98184538e-01 -6.38477981e-01 1.90253615e+00
-9.31871891e-01 5.97011387e-01 -3.31376672e-01 -1.36093187e+00
8.28528941e-01 3.35945904e-01 -1.54820651e-01 -8.33703756e-01
7.42675513e-02 2.88565576e-01 1.89131781e-01 -6.73586071e-01
4.08743143e-01 -4.02668208e-01 -3.10274094e-01 9.99984086e-01
1.74493626e-01 -5.47528505e-01 3.19801599e-01 4.79088336e-01
1.24034011e+00 -2.04643965e-01 3.88773978e-01 -2.48288691e-01
5.35966575e-01 1.99278280e-01 -4.34176847e-02 7.38704264e-01
-6.17218204e-03 5.52096725e-01 5.14331281e-01 -3.75294566e-01
-1.06847155e+00 -1.01409733e+00 5.88726364e-02 1.74199092e+00
-1.01782002e-01 -2.11915284e-01 -7.56881654e-01 -8.83069158e-01
1.54948205e-01 1.26522911e+00 -5.99040151e-01 -4.15593028e-01
-6.95180774e-01 -6.73510611e-01 7.75794506e-01 7.67863154e-01
5.26452899e-01 -1.10603821e+00 -2.62406617e-01 1.36886895e-01
-4.93975878e-01 -1.43818426e+00 -4.26498443e-01 1.87364087e-01
-9.25913751e-01 -1.03534031e+00 -3.76500458e-01 -7.16188192e-01
4.56206083e-01 1.41814575e-01 1.78027368e+00 -4.39752117e-02
-1.37561232e-01 4.42017674e-01 -3.12588066e-01 -1.99040040e-01
-5.91897070e-01 3.79866272e-01 -2.83424795e-01 -1.75190508e-01
6.09643400e-01 -5.08506298e-01 -4.42557186e-01 1.55312181e-01
-9.64940131e-01 -1.20900899e-01 5.56349933e-01 1.15012968e+00
1.28746018e-01 -5.71033716e-01 9.61114287e-01 -1.26405656e+00
9.40541744e-01 -7.00565815e-01 -1.26893833e-01 3.92989248e-01
-6.01471603e-01 3.63312125e-01 7.74890721e-01 -3.58096182e-01
-1.18591809e+00 -4.63981032e-01 -1.44153535e-01 -4.66331206e-02
-2.79529631e-01 5.18473864e-01 -1.13005415e-02 3.34396213e-01
1.23986506e+00 1.63724124e-01 -1.43985167e-01 -2.77435899e-01
7.08096206e-01 6.18257701e-01 7.09011674e-01 -9.65356648e-01
8.16521764e-01 2.09667474e-01 -4.85245019e-01 -5.18338203e-01
-1.52506959e+00 -4.39649940e-01 -4.04729098e-01 3.63683283e-01
9.34752822e-01 -9.48164225e-01 -4.64287668e-01 1.04658604e-01
-1.19760895e+00 -5.89370966e-01 -1.80254743e-01 3.08733016e-01
-5.19825757e-01 1.62318468e-01 -6.85172260e-01 -3.63100529e-01
-6.48268580e-01 -9.50258911e-01 1.01213014e+00 -9.41693932e-02
-7.97736585e-01 -1.30493248e+00 1.55414999e-01 1.08198762e+00
6.36067271e-01 1.70583483e-02 1.41296363e+00 -1.17912006e+00
-4.18168575e-01 -2.28108197e-01 -3.89121681e-01 6.26264155e-01
6.80939779e-02 -4.83038038e-01 -1.14582467e+00 -1.84563518e-01
8.45567435e-02 -1.27652526e+00 1.09420264e+00 7.47544691e-02
1.09804165e+00 -1.20198071e-01 -1.45505276e-02 3.09555501e-01
1.03522003e+00 -6.23011999e-02 4.59216535e-01 3.13385397e-01
7.00252831e-01 5.66246986e-01 4.04951721e-01 5.16926982e-02
6.79362357e-01 4.37373042e-01 2.67999172e-01 1.37539431e-01
-7.84064606e-02 -1.80139661e-01 2.86620200e-01 4.86522615e-01
4.36016098e-02 -2.11023167e-01 -1.21704054e+00 7.42357910e-01
-1.73267329e+00 -9.05394197e-01 1.03446878e-01 1.86923373e+00
1.10208249e+00 2.71599293e-01 -1.34857506e-01 -9.89282783e-03
3.74020904e-01 2.36534998e-01 -8.46227467e-01 -5.73225319e-01
6.56759441e-02 4.60817933e-01 -6.21278472e-02 6.56487405e-01
-1.00587893e+00 1.12510824e+00 6.26996565e+00 7.42198646e-01
-8.91328454e-01 1.87770620e-01 8.09660852e-01 -6.72848821e-02
-7.49873936e-01 -3.31484452e-02 -6.20392263e-01 3.56264301e-02
1.05077386e+00 -2.87998319e-01 3.99379522e-01 9.04211283e-01
-1.83590740e-01 1.58255659e-02 -1.57976675e+00 8.09973598e-01
4.31525677e-01 -1.17466867e+00 2.41260946e-01 -4.37032193e-01
9.97400939e-01 2.22766623e-01 6.80312514e-02 7.98658669e-01
5.15630603e-01 -1.36743581e+00 4.99980837e-01 4.56890427e-02
6.37744069e-01 -7.08936632e-01 6.18829191e-01 4.51891154e-01
-5.95070660e-01 1.20911725e-01 -3.44479799e-01 -3.75920445e-01
8.53976831e-02 5.58545768e-01 -1.00696278e+00 1.45203516e-01
2.49606282e-01 5.55482984e-01 -6.02673471e-01 4.00211871e-01
-6.09707594e-01 1.02625191e+00 -1.20323211e-01 -1.50631398e-01
5.21988332e-01 2.62712747e-01 2.11478114e-01 1.35460949e+00
-7.98419937e-02 -1.30003123e-02 1.30385265e-01 9.25805330e-01
-7.39686668e-01 1.73286691e-01 -5.65880895e-01 -2.64701307e-01
3.59541237e-01 1.14314473e+00 -2.30119571e-01 -7.32857347e-01
-5.46733677e-01 1.00987065e+00 7.07841158e-01 3.85369629e-01
-8.53150189e-01 -4.70337808e-01 4.07905042e-01 -9.84129980e-02
2.41302088e-01 -4.64338809e-02 -6.57032907e-01 -1.40800762e+00
9.75705311e-02 -1.18846726e+00 6.90338016e-01 -6.56698823e-01
-1.57688129e+00 6.12479568e-01 -1.55099452e-01 -6.88059270e-01
-7.72361696e-01 -6.35288477e-01 -7.86660671e-01 9.13833618e-01
-1.50034690e+00 -1.10977781e+00 -3.30547601e-01 6.99165761e-01
8.98279011e-01 -2.87025660e-01 9.91499424e-01 1.67222336e-01
-2.61340410e-01 6.62776709e-01 -4.42291126e-02 1.21319033e-01
1.07125199e+00 -1.40415657e+00 6.87784553e-01 7.38150001e-01
3.64199907e-01 7.02241063e-01 6.18306875e-01 -2.93686897e-01
-1.26509261e+00 -1.01137900e+00 1.09342146e+00 -8.44564140e-01
9.58293080e-01 -4.10375059e-01 -1.10531771e+00 1.03509760e+00
4.24672723e-01 -1.22704022e-01 8.45983863e-01 5.54479957e-01
-1.15444505e+00 5.30253686e-02 -9.81905162e-01 6.21980309e-01
1.00019896e+00 -9.36624587e-01 -1.26909161e+00 5.57888091e-01
7.14163065e-01 -4.02216643e-01 -7.38036215e-01 3.14889282e-01
3.30381542e-01 -6.14674509e-01 9.78039503e-01 -1.26345158e+00
1.21833289e+00 1.68532297e-01 -2.87878871e-01 -1.38270557e+00
-2.35910118e-01 -3.47425342e-01 -3.46351787e-02 1.06445920e+00
7.73747385e-01 -5.90213597e-01 6.43817425e-01 8.16843510e-01
-1.94575056e-01 -6.05725527e-01 -4.18813169e-01 -6.35377169e-01
5.99790692e-01 -5.14138281e-01 2.33618602e-01 1.08626139e+00
1.92610517e-01 1.23607826e+00 -9.82601866e-02 -2.38675684e-01
3.83573174e-01 4.92339671e-01 9.23656642e-01 -8.89567375e-01
-4.97253507e-01 -5.22105336e-01 -1.58224832e-02 -1.25985479e+00
5.38127482e-01 -1.36112273e+00 -1.26908934e-02 -1.70713007e+00
5.70064366e-01 -1.67255566e-01 -9.33572724e-02 5.98011255e-01
-6.91381752e-01 3.21989864e-01 -4.83881980e-02 5.52060157e-02
-7.65039086e-01 5.12600124e-01 1.15760469e+00 -3.58799398e-01
2.13166162e-01 -1.87620103e-01 -1.09933591e+00 7.85143018e-01
8.15844178e-01 -3.34203959e-01 -7.33433366e-01 -9.61445689e-01
1.69578344e-01 -1.73565045e-01 3.66675943e-01 -6.98529363e-01
9.34149325e-02 -5.33158332e-02 1.40165329e-01 -2.16030478e-01
2.62709737e-01 -2.84770966e-01 -6.63756847e-01 3.31135899e-01
-8.91848624e-01 -1.31017249e-03 3.23231459e-01 3.44364673e-01
-3.79721075e-01 -3.84367913e-01 7.68201292e-01 -9.49101672e-02
-7.61017621e-01 -4.27356251e-02 -1.18164659e-01 1.07613325e+00
7.02262998e-01 2.62760371e-01 -6.21920824e-01 -6.86757684e-01
-3.70834202e-01 3.17417473e-01 1.27077028e-01 4.41722244e-01
5.27961552e-01 -1.01838982e+00 -1.16615129e+00 -2.96671540e-01
3.97706121e-01 4.77432944e-02 1.19239978e-01 5.68606317e-01
-3.60325485e-01 6.17582917e-01 -7.93430507e-02 -4.91875798e-01
-9.07451689e-01 4.15525049e-01 1.65841147e-01 -7.24248469e-01
-2.84463853e-01 9.97471035e-01 5.80295861e-01 -7.54233360e-01
1.91057231e-02 -4.01227653e-01 -2.49014166e-03 -1.89993367e-01
6.39423370e-01 1.80412903e-01 2.53731787e-01 -1.81724027e-01
-1.96863279e-01 3.63200933e-01 -4.99778628e-01 -4.36463840e-02
1.36835086e+00 4.93576564e-02 -3.20898220e-02 3.17777216e-01
1.41848195e+00 -1.89661533e-01 -9.22309220e-01 -5.90532720e-01
2.07248077e-01 -9.26948676e-04 -2.70931005e-01 -1.11090589e+00
-5.87630451e-01 1.10789943e+00 -1.43067330e-01 1.21162627e-02
8.75533640e-01 2.15049848e-01 9.95477438e-01 1.09075630e+00
1.22950658e-01 -7.27073729e-01 4.96225059e-01 1.01572144e+00
1.05446422e+00 -1.53553200e+00 -2.98405796e-01 -1.80006877e-01
-7.60646880e-01 8.85687172e-01 6.21696889e-01 -2.57513523e-01
1.27180651e-01 1.81306452e-01 2.03602128e-02 -1.82322189e-01
-1.02433538e+00 -4.88995761e-02 5.08352339e-01 2.78343171e-01
8.08725417e-01 -4.93985601e-02 -7.60990828e-02 4.96860445e-01
-4.19459462e-01 -2.19076395e-01 2.10792154e-01 7.50264287e-01
-5.18541455e-01 -9.15947080e-01 -1.62258700e-01 6.31876290e-01
-5.25016069e-01 -5.53766966e-01 -3.95343065e-01 6.43278062e-01
-5.13281167e-01 9.33132887e-01 5.34427576e-02 -1.06815271e-01
3.37323844e-01 4.86653537e-01 5.94572604e-01 -9.62724090e-01
-5.92188835e-01 -5.56227744e-01 6.27381802e-01 -4.63554859e-01
-1.43419132e-01 -3.88285488e-01 -1.21426642e+00 -2.74226367e-01
-3.07960585e-02 8.75370577e-02 2.71175772e-01 1.39330280e+00
5.19950911e-02 4.94229555e-01 2.08808079e-01 -3.30026001e-01
-1.06980813e+00 -1.10617149e+00 -3.85580212e-02 8.68210375e-01
1.46019816e-01 -3.89505804e-01 -3.14566165e-01 3.78462136e-01] | [11.026488304138184, 8.404294967651367] |
cad0fe39-3f50-402d-ab09-98281675973f | system-neural-diversity-measuring-behavioral | 2305.02128 | null | https://arxiv.org/abs/2305.02128v1 | https://arxiv.org/pdf/2305.02128v1.pdf | System Neural Diversity: Measuring Behavioral Heterogeneity in Multi-Agent Learning | Evolutionary science provides evidence that diversity confers resilience. Yet, traditional multi-agent reinforcement learning techniques commonly enforce homogeneity to increase training sample efficiency. When a system of learning agents is not constrained to homogeneous policies, individual agents may develop diverse behaviors, resulting in emergent complementarity that benefits the system. Despite this feat, there is a surprising lack of tools that measure behavioral diversity in systems of learning agents. Such techniques would pave the way towards understanding the impact of diversity in collective resilience and performance. In this paper, we introduce System Neural Diversity (SND): a measure of behavioral heterogeneity for multi-agent systems where agents have stochastic policies. %over a continuous state space. We discuss and prove its theoretical properties, and compare it with alternate, state-of-the-art behavioral diversity metrics used in cross-disciplinary domains. Through simulations of a variety of multi-agent tasks, we show how our metric constitutes an important diagnostic tool to analyze latent properties of behavioral heterogeneity. By comparing SND with task reward in static tasks, where the problem does not change during training, we show that it is key to understanding the effectiveness of heterogeneous vs homogeneous agents. In dynamic tasks, where the problem is affected by repeated disturbances during training, we show that heterogeneous agents are first able to learn specialized roles that allow them to cope with the disturbance, and then retain these roles when the disturbance is removed. SND allows a direct measurement of this latent resilience, while other proxies such as task performance (reward) fail to. | ['Amanda Prorok', 'Ajay Shankar', 'Matteo Bettini'] | 2023-05-03 | null | null | null | null | ['multi-agent-reinforcement-learning'] | ['methodology'] | [-1.16358884e-01 -1.48830697e-01 1.40525308e-02 5.62566578e-01
-6.86263070e-02 -8.45912337e-01 7.79334307e-01 2.52212971e-01
-5.47490180e-01 1.02766454e+00 -2.49028276e-03 -1.29640773e-01
-4.61920440e-01 -5.02621174e-01 -6.45084798e-01 -1.14280415e+00
-4.47132200e-01 2.63862997e-01 -2.80230287e-02 -5.97462535e-01
2.70966887e-01 4.27215129e-01 -1.82929671e+00 -3.22906733e-01
1.02611125e+00 2.60605276e-01 3.82468790e-01 7.31728017e-01
5.57432950e-01 8.63228559e-01 -1.07232487e+00 1.05357260e-01
4.44319576e-01 -7.33556628e-01 -5.68565905e-01 5.43371253e-02
-7.76314512e-02 5.03501594e-02 -1.48677468e-01 9.26089466e-01
7.83409774e-01 3.51640940e-01 8.07936847e-01 -1.36427689e+00
-6.60463810e-01 6.94049954e-01 -3.21789801e-01 4.23138231e-01
5.97109608e-02 5.78987241e-01 7.73251772e-01 1.67079587e-02
6.74762249e-01 1.33559024e+00 6.64490938e-01 6.40027821e-01
-1.62878752e+00 -3.76839429e-01 2.29252040e-01 -2.44657516e-01
-8.75732005e-01 -4.24976438e-01 4.51999575e-01 -8.27607512e-01
9.27085698e-01 1.79707780e-01 9.94469106e-01 1.22093415e+00
7.12204993e-01 3.43551040e-01 1.51184821e+00 -2.38777637e-01
6.17815435e-01 4.87221219e-02 -3.26124132e-02 5.05016744e-01
6.94136202e-01 6.54397070e-01 -3.78565580e-01 -3.03170562e-01
8.22831631e-01 -1.35593668e-01 -2.28089228e-01 -3.80836993e-01
-1.10026872e+00 6.76558912e-01 2.10435972e-01 4.70890939e-01
-5.73813558e-01 2.43004695e-01 3.85292768e-01 1.07414711e+00
2.02830434e-01 1.27294993e+00 -3.09910983e-01 -1.78745791e-01
-4.71290886e-01 5.27824283e-01 8.19805741e-01 3.21928591e-01
5.84341228e-01 4.45435762e-01 -3.66237536e-02 6.79897130e-01
-2.08268374e-01 3.72490734e-01 7.19004095e-01 -1.35852611e+00
-1.15492895e-01 5.84025443e-01 9.41943899e-02 -6.10896885e-01
-5.96764207e-01 -8.06501508e-01 -6.38894796e-01 8.06054592e-01
4.07497406e-01 -6.08195305e-01 -2.83945292e-01 2.06504631e+00
2.42840409e-01 -1.93231136e-01 3.38653117e-01 6.59382045e-01
-5.06838076e-02 4.17804927e-01 -2.13567451e-01 -4.98905271e-01
8.09547901e-01 -7.27061212e-01 -2.95588195e-01 -8.45180899e-02
6.05783165e-01 -2.91275591e-01 1.03250921e+00 1.60761431e-01
-1.22982109e+00 -2.55296260e-01 -9.61840689e-01 8.87174785e-01
-2.14760885e-01 -7.22137809e-01 3.84940535e-01 7.04284728e-01
-1.32725370e+00 7.51066685e-01 -9.47411060e-01 -6.88309491e-01
3.23960967e-02 2.51579106e-01 -2.97005828e-02 4.44816589e-01
-8.87124598e-01 1.25118160e+00 4.50705066e-02 -5.62204838e-01
-1.31413972e+00 -7.02110410e-01 -3.08787733e-01 1.78313136e-01
5.27153611e-01 -1.03930068e+00 1.18874955e+00 -1.36696970e+00
-1.62065053e+00 3.06599289e-01 4.12905335e-01 -2.44070649e-01
4.86615628e-01 4.22808498e-01 6.93534538e-02 -1.83671907e-01
1.03929207e-01 1.33236259e-01 7.79481888e-01 -1.56191516e+00
-4.48503941e-01 -4.12696928e-01 3.68853867e-01 4.87157613e-01
-3.77687603e-01 -1.32136777e-01 7.21152723e-01 -4.76863801e-01
-6.41976118e-01 -1.41957438e+00 -2.74707407e-01 -6.77483320e-01
1.63632794e-04 -7.75863752e-02 4.02944565e-01 1.96178034e-02
8.49111676e-01 -1.83338666e+00 6.62090600e-01 -2.18018536e-02
4.24043208e-01 -6.40028249e-03 -3.86471689e-01 8.68566215e-01
1.79627553e-01 4.14448231e-01 -2.34070376e-01 -1.60837488e-03
1.12780936e-01 1.73945919e-01 -8.04490317e-03 5.15899181e-01
1.94032654e-01 6.59197450e-01 -8.57729673e-01 -2.10561622e-02
-1.46141514e-01 1.18583880e-01 -6.66333556e-01 8.72928724e-02
-1.41930833e-01 6.03717625e-01 -4.75287825e-01 3.95288020e-01
5.67385405e-02 -1.50801212e-01 4.58657175e-01 7.30846226e-01
-5.32417893e-01 -1.51113749e-01 -1.05688441e+00 8.84885073e-01
-4.56244916e-01 6.76593542e-01 2.04630390e-01 -1.23708642e+00
5.31703949e-01 1.60549372e-01 6.82717144e-01 -8.81795883e-01
2.73451895e-01 3.23948711e-01 7.80134439e-01 -4.10078794e-01
4.01984125e-01 -9.88464206e-02 -5.00539541e-02 8.51110339e-01
1.44154266e-01 -2.61914730e-01 4.95163709e-01 -9.32432637e-02
1.66650271e+00 -1.91187263e-01 2.47017160e-01 -8.39391291e-01
6.71764761e-02 1.95298776e-01 6.10009968e-01 1.09963644e+00
-4.19435263e-01 -1.25666305e-01 8.48360956e-01 -8.71443599e-02
-1.33457851e+00 -1.01023459e+00 -5.63480593e-02 1.33359754e+00
1.41824096e-01 4.13506143e-02 -7.61363149e-01 -8.08268338e-02
4.02747363e-01 4.08720374e-01 -1.01267362e+00 -5.13285100e-01
-5.48205554e-01 -1.03996587e+00 5.39745629e-01 -2.38738284e-02
2.88772970e-01 -1.14709318e+00 -1.20523989e+00 1.59826487e-01
3.93569589e-01 -4.05178040e-01 -2.47404903e-01 4.74213958e-01
-9.02177691e-01 -1.11834311e+00 -8.23968589e-01 -4.58415478e-01
4.65504795e-01 3.84785205e-01 1.09581983e+00 4.68263626e-01
-1.75169185e-01 7.37474680e-01 -2.23147988e-01 -2.67323405e-01
-7.46618867e-01 1.09782971e-01 7.07128763e-01 -4.69953835e-01
-4.61043298e-01 -8.74963641e-01 -4.54261869e-01 3.82755399e-01
-9.79297519e-01 -3.41436714e-01 5.99876583e-01 1.03668606e+00
1.47056788e-01 1.00500546e-01 9.43649590e-01 -4.58693862e-01
1.29085803e+00 -6.99939013e-01 -5.40352762e-01 3.88597220e-01
-8.80024374e-01 2.15987787e-01 8.24422657e-01 -9.00616109e-01
-7.28510261e-01 -5.44876575e-01 6.78650379e-01 -7.69262835e-02
1.93511233e-01 6.15253747e-01 3.53299886e-01 -3.10581267e-01
9.99998212e-01 2.43145898e-01 4.72986519e-01 -4.66940925e-02
1.56333204e-02 1.84983253e-01 -9.20698121e-02 -9.33192790e-01
6.38454914e-01 2.67358068e-02 1.29087880e-01 -9.52121735e-01
-7.49661252e-02 3.16003203e-01 -1.11342013e-01 -3.78413916e-01
4.15205747e-01 -7.41518259e-01 -1.11835623e+00 4.76799905e-01
-5.80744267e-01 -9.78651166e-01 -5.67549944e-01 3.84627640e-01
-9.93463635e-01 -2.12091595e-01 -4.28029299e-01 -1.06586039e+00
1.89294934e-01 -1.18904459e+00 4.16646570e-01 5.24956882e-01
-2.08918061e-02 -1.11033273e+00 7.62750804e-01 -2.16069981e-01
7.26786673e-01 3.24461728e-01 1.06496716e+00 -6.75326765e-01
-3.05125445e-01 3.57119888e-01 6.07105434e-01 1.22733243e-01
4.64255102e-02 3.26677561e-01 -4.70104426e-01 -5.67129374e-01
1.62713096e-01 -3.63543302e-01 6.30406320e-01 5.94202697e-01
3.90749902e-01 -4.46962267e-01 -1.35288551e-01 2.83166170e-01
1.31181335e+00 5.18466830e-01 1.52155325e-01 8.78122509e-01
1.58016011e-01 8.89763892e-01 2.16764435e-02 4.65639502e-01
3.92783493e-01 4.95176971e-01 3.58954906e-01 2.54832834e-01
1.16374545e-01 2.07091540e-01 8.80428612e-01 8.40054274e-01
-3.64416033e-01 -2.95843035e-01 -9.52901363e-01 3.58061045e-01
-1.99693620e+00 -1.39856017e+00 3.23103368e-01 2.29457021e+00
8.81680131e-01 2.39930451e-02 8.17145944e-01 -1.92283228e-01
6.89520121e-01 -3.70030813e-02 -1.02996325e+00 -6.42103374e-01
-4.57112759e-01 -1.91904649e-01 4.48772937e-01 5.51764727e-01
-3.74106795e-01 7.46073186e-01 7.20154142e+00 3.61286610e-01
-1.07310891e+00 1.20535888e-01 4.28954571e-01 -5.22073090e-01
-2.45871440e-01 -9.29165706e-02 -2.61110604e-01 4.78568822e-01
1.11541247e+00 -8.43823493e-01 1.01859784e+00 4.52854872e-01
3.88878524e-01 -3.49953651e-01 -1.10060430e+00 2.45738134e-01
-1.74745083e-01 -9.54924226e-01 -5.44033289e-01 3.50965142e-01
1.13369429e+00 5.62659577e-02 4.29291785e-01 6.10061646e-01
9.80734646e-01 -9.23950911e-01 7.01702833e-01 5.57399929e-01
2.33671188e-01 -7.46655226e-01 3.56539786e-01 6.55425549e-01
-6.45281315e-01 -5.47727466e-01 -2.99992919e-01 -6.30750835e-01
-2.99805075e-01 2.41501763e-01 -5.01916111e-01 1.70132741e-01
4.87619787e-01 3.50695699e-01 -6.70989752e-01 8.74104857e-01
2.76244134e-01 3.62401456e-01 -5.59747033e-02 -2.53478855e-01
1.16711304e-01 -3.29774052e-01 9.21123087e-01 5.65669239e-01
2.87582844e-01 -8.61989185e-02 1.52679309e-01 9.15769458e-01
1.46944955e-01 -1.00316979e-01 -9.29513812e-01 -3.68703395e-01
6.35345519e-01 1.01688325e+00 -6.77637279e-01 -2.49610201e-01
3.74308391e-03 5.66197574e-01 5.43057382e-01 4.65196818e-01
-6.61214709e-01 -7.82473534e-02 1.14511728e+00 -2.54302382e-01
-1.25002980e-01 -4.83329386e-01 -2.65466124e-01 -1.11233640e+00
-2.85945684e-01 -1.35271680e+00 -1.48810312e-01 -3.25686842e-01
-1.43524110e+00 3.98966640e-01 2.41196156e-02 -8.88543785e-01
-4.86757487e-01 -2.57708251e-01 -6.27757251e-01 5.42269051e-01
-1.06921709e+00 -4.32510763e-01 -9.64992642e-02 4.41803634e-01
3.89514685e-01 -5.17930210e-01 5.07581115e-01 -2.09850937e-01
-1.08718073e+00 3.82687539e-01 7.62547255e-01 -5.91742992e-01
5.49034059e-01 -1.31110251e+00 1.16145007e-01 6.63651109e-01
-2.10544616e-01 8.04319978e-01 1.03108680e+00 -7.49877095e-01
-1.58793676e+00 -6.64170206e-01 -2.41649989e-02 -5.14900386e-01
8.74674559e-01 -1.38117224e-01 -7.49777198e-01 3.59024227e-01
4.09612089e-01 -4.88123894e-01 4.81356114e-01 1.81330025e-01
-6.13509677e-02 1.41709283e-01 -9.15264845e-01 1.07489109e+00
1.22436893e+00 -3.50840360e-01 -3.47666502e-01 2.61354536e-01
7.85982311e-01 7.71534815e-02 -8.90418291e-01 1.09644711e-01
6.90129042e-01 -1.11725116e+00 6.73737705e-01 -8.14707220e-01
4.83762532e-01 -7.20205083e-02 -2.41155848e-01 -1.98647630e+00
-6.30850554e-01 -7.18769670e-01 1.23253092e-01 9.60508943e-01
4.16023552e-01 -1.16477275e+00 2.86719829e-01 4.16675627e-01
-2.72825658e-01 -5.94122171e-01 -7.97703087e-01 -1.37010598e+00
7.52693772e-01 5.32748878e-01 3.99250895e-01 1.18959570e+00
2.88119972e-01 1.05297901e-01 -1.43710310e-02 -4.89107184e-02
5.64568460e-01 -2.30855644e-01 7.30876327e-01 -1.28328323e+00
-4.18786496e-01 -1.17171121e+00 -2.75303274e-01 -1.49820998e-01
4.98604923e-01 -6.00519598e-01 -2.63256393e-02 -1.11642623e+00
2.93000460e-01 -4.96962994e-01 -2.29910776e-01 2.49506503e-01
-1.82512388e-01 -2.39292443e-01 4.94314790e-01 5.57160854e-01
-4.15640175e-01 6.89558804e-01 1.22688484e+00 6.59747645e-02
-6.68813229e-01 -2.45394737e-01 -7.97176182e-01 4.35251236e-01
1.09041023e+00 -6.69934511e-01 -5.14265776e-01 -2.20260024e-01
2.41768032e-01 1.51095763e-01 3.63718599e-01 -1.11557937e+00
1.05133586e-01 -7.79345989e-01 -6.34580925e-02 6.21064246e-01
-9.94625688e-02 -5.50542116e-01 4.07727599e-01 1.07469654e+00
-5.05865693e-01 6.74805164e-01 9.06067044e-02 4.75806177e-01
2.48023152e-01 -2.08379015e-01 8.04985821e-01 -1.68120936e-01
-1.47320524e-01 -6.16383366e-02 -1.00401950e+00 3.85543495e-01
1.31550312e+00 -1.59492478e-01 -9.14807856e-01 -2.30377570e-01
-4.63379502e-01 4.16992724e-01 1.18408966e+00 2.01500222e-01
5.79159521e-02 -9.94027615e-01 -7.79878736e-01 -2.01535091e-01
-1.50105372e-01 -7.46618569e-01 2.64854670e-01 1.01144874e+00
-3.67124289e-01 2.07912326e-02 -8.47133040e-01 -4.11365747e-01
-9.33647156e-01 5.34323335e-01 8.31682384e-01 -1.20251857e-01
-3.14730257e-01 4.96587545e-01 2.75878817e-01 -1.88269868e-01
-8.86271223e-02 5.08050025e-02 -2.08412800e-02 3.68685484e-01
1.58461690e-01 5.52882850e-01 -1.78910926e-01 -3.29017907e-01
-3.05005878e-01 3.15358669e-01 2.09353730e-01 -5.39940178e-01
1.34667909e+00 -1.86974823e-01 -5.39713539e-02 7.77102113e-01
5.13747275e-01 -1.62507206e-01 -1.43767524e+00 1.15677394e-01
-1.71543241e-01 -2.68660486e-01 -2.14432999e-01 -7.05990434e-01
-7.30256438e-01 4.96266305e-01 5.11932552e-01 9.59107459e-01
9.81699884e-01 -2.46106103e-01 -1.99938659e-02 5.69555998e-01
4.34669554e-01 -1.26740825e+00 5.66364765e-01 6.91446126e-01
9.47650015e-01 -1.01892626e+00 -1.23844169e-01 5.25342643e-01
-9.00017977e-01 7.38130093e-01 7.58706868e-01 -4.99078244e-01
2.19267651e-01 3.42906386e-01 -2.68817574e-01 -5.17069362e-02
-1.37326431e+00 -4.41916525e-01 -2.85923392e-01 8.45449150e-01
2.34995559e-01 1.54770002e-01 -3.86552155e-01 1.85177520e-01
-4.12038237e-01 -5.44803917e-01 1.15410292e+00 1.00730252e+00
-7.58339405e-01 -1.00839818e+00 -3.54903698e-01 4.66124773e-01
-2.98379153e-01 9.83323008e-02 -7.80474484e-01 1.02680242e+00
-7.28380084e-02 1.00760627e+00 1.51020037e-02 -4.39683616e-01
2.67021716e-01 -2.41119452e-02 7.11417854e-01 -5.32323956e-01
-1.13661540e+00 -2.62200475e-01 -1.42006829e-01 -2.59995222e-01
-4.47216392e-01 -1.01850033e+00 -8.74337375e-01 -7.61615992e-01
-2.11872250e-01 1.40648395e-01 5.36889255e-01 6.28845394e-01
4.54594523e-01 9.07410085e-01 9.08249497e-01 -9.92466569e-01
-8.82063031e-01 -7.80897498e-01 -7.95034468e-01 2.61938184e-01
6.82773650e-01 -1.03840923e+00 -6.68276548e-01 -1.67896867e-01] | [3.9833285808563232, 2.1749398708343506] |
ff98c8f6-a4f8-4006-ac26-4a248a51e776 | panns-large-scale-pretrained-audio-neural-1 | 1912.10211 | null | http://arxiv.org/abs/1912.10211v5 | http://arxiv.org/pdf/1912.10211v5.pdf | PANNs: Large-Scale Pretrained Audio Neural Networks for Audio Pattern Recognition | Audio pattern recognition is an important research topic in the machine
learning area, and includes several tasks such as audio tagging, acoustic scene
classification, music classification, speech emotion classification and sound
event detection. Recently, neural networks have been applied to tackle audio
pattern recognition problems. However, previous systems are built on specific
datasets with limited durations. Recently, in computer vision and natural
language processing, systems pretrained on large-scale datasets have
generalized well to several tasks. However, there is limited research on
pretraining systems on large-scale datasets for audio pattern recognition. In
this paper, we propose pretrained audio neural networks (PANNs) trained on the
large-scale AudioSet dataset. These PANNs are transferred to other audio
related tasks. We investigate the performance and computational complexity of
PANNs modeled by a variety of convolutional neural networks. We propose an
architecture called Wavegram-Logmel-CNN using both log-mel spectrogram and
waveform as input feature. Our best PANN system achieves a state-of-the-art
mean average precision (mAP) of 0.439 on AudioSet tagging, outperforming the
best previous system of 0.392. We transfer PANNs to six audio pattern
recognition tasks, and demonstrate state-of-the-art performance in several of
those tasks. We have released the source code and pretrained models of PANNs:
https://github.com/qiuqiangkong/audioset_tagging_cnn. | [] | 2020-08-23 | panns-large-scale-pretrained-audio-neural | null | null | null | ['audio-tagging'] | ['audio'] | [ 2.93289185e-01 -6.63976252e-01 1.56717077e-01 -2.98211396e-01
-1.08060420e+00 -4.10482943e-01 -1.81372866e-01 2.22106948e-02
-4.90087211e-01 2.45924935e-01 9.62775499e-02 1.35705695e-01
-1.02101132e-01 -5.10117948e-01 -5.42835474e-01 -4.45332289e-01
-4.07143742e-01 2.57480815e-02 2.95764953e-01 6.90222532e-02
-8.41948669e-03 1.84992477e-01 -1.77946770e+00 6.99262500e-01
1.78078219e-01 1.59413314e+00 8.12159479e-02 1.00935447e+00
2.42028996e-01 5.88293314e-01 -8.85291696e-01 8.35159700e-03
-1.29831791e-01 -2.57037252e-01 -7.60127246e-01 -3.32998484e-01
4.45862442e-01 7.92695656e-02 -4.48356450e-01 1.01537323e+00
1.12399423e+00 2.90549040e-01 1.01570003e-01 -1.56110311e+00
-5.21798551e-01 1.02213109e+00 -1.96298420e-01 5.66917002e-01
2.62954205e-01 -9.71568003e-02 1.15637159e+00 -1.02995539e+00
-6.64602071e-02 1.04035854e+00 9.90281045e-01 3.41802984e-01
-7.63616681e-01 -1.20903277e+00 -2.45956600e-01 7.40727961e-01
-1.85651124e+00 -6.38568223e-01 9.69152749e-01 -3.81002158e-01
1.07227349e+00 3.34080547e-01 6.12426281e-01 9.44959164e-01
-1.63119845e-02 8.80947173e-01 6.78063750e-01 -5.28886378e-01
1.67604685e-01 -2.17525929e-01 1.58800900e-01 4.11745310e-01
-5.63333869e-01 3.69075239e-02 -1.08800578e+00 -2.20331281e-01
4.81834531e-01 -2.43711516e-01 -3.64587694e-01 4.11994308e-01
-1.02565134e+00 6.00028813e-01 3.17527950e-01 5.84001899e-01
-2.57325858e-01 2.81100571e-01 8.92274559e-01 5.31549275e-01
5.51988780e-01 4.72785473e-01 -7.68355727e-01 -6.61983430e-01
-9.41757023e-01 1.41639084e-01 7.43308306e-01 6.82531774e-01
2.72626400e-01 5.17906904e-01 -1.47611439e-01 1.56587005e+00
-5.50873093e-02 3.67057830e-01 8.33408654e-01 -6.63446844e-01
3.22415203e-01 1.44713044e-01 -4.10097718e-01 -9.48692083e-01
-5.27039111e-01 -4.53705251e-01 -9.72113490e-01 -2.68388689e-01
7.76093528e-02 -2.61230528e-01 -6.77054822e-01 1.62593579e+00
6.03848025e-02 7.65870750e-01 -1.89230144e-01 9.78134513e-01
1.13919604e+00 1.10457957e+00 6.16545230e-03 -5.04317321e-02
1.45241141e+00 -1.05197525e+00 -6.95473373e-01 -9.95340347e-02
2.54956514e-01 -1.08079672e+00 1.28415883e+00 8.49225402e-01
-9.37904000e-01 -9.71142948e-01 -8.47363949e-01 2.30201155e-01
-3.67810160e-01 5.17966568e-01 3.46298307e-01 5.37200332e-01
-1.03397739e+00 6.74079299e-01 -7.40382552e-01 -4.74453509e-01
3.88870835e-01 5.33892691e-01 -1.07873850e-01 3.90644759e-01
-1.41828072e+00 2.30145916e-01 6.07424557e-01 1.72560975e-01
-1.04352546e+00 -8.96417141e-01 -6.15242541e-01 3.36041600e-01
2.67188072e-01 -1.03573292e-01 1.76612306e+00 -8.14015687e-01
-1.72902393e+00 5.88307619e-01 6.11854978e-02 -6.88964486e-01
-3.24363440e-01 -4.31143761e-01 -1.20708561e+00 5.50884530e-02
-2.12864861e-01 9.13271487e-01 8.03732395e-01 -4.84177083e-01
-8.80784988e-01 1.31175056e-01 -3.25426757e-01 -5.02503291e-02
-8.42757285e-01 6.19864404e-01 -3.69584978e-01 -9.17726040e-01
-2.67629504e-01 -8.41041327e-01 8.67178887e-02 -2.90681630e-01
-3.61885607e-01 -5.14971316e-01 7.61100233e-01 -5.32622039e-01
1.62527847e+00 -2.64823008e+00 -2.32768700e-01 -1.63690150e-01
-2.59769619e-01 5.76858878e-01 -5.24214625e-01 2.57919818e-01
-2.37845182e-01 -1.39416531e-02 7.62223732e-03 -1.80106178e-01
3.38538498e-01 -3.91654521e-02 -6.95408523e-01 -3.82609144e-02
2.36390889e-01 5.95754385e-01 -7.16718018e-01 -2.66167581e-01
5.49393706e-02 4.87503678e-01 -6.28432989e-01 3.50180954e-01
3.74128856e-02 1.98660076e-01 -1.01944311e-02 7.51390874e-01
3.47302794e-01 8.53734985e-02 -1.76872894e-01 -1.58885464e-01
2.58216355e-02 6.66183591e-01 -1.24540424e+00 1.93398201e+00
-5.13378441e-01 8.19803178e-01 -2.21962039e-03 -8.53732467e-01
1.01625252e+00 8.62523735e-01 7.05369830e-01 -3.59243244e-01
1.61541268e-01 3.44731659e-01 3.45017761e-01 -5.12769520e-01
6.01812959e-01 -7.97057301e-02 -2.71779984e-01 2.94253945e-01
5.17349601e-01 -1.18600495e-01 5.05854711e-02 -3.11108947e-01
1.19638622e+00 -4.76609141e-01 1.57409355e-01 1.31434262e-01
4.84172642e-01 -3.17809761e-01 6.93177164e-01 5.40790975e-01
-2.66141564e-01 7.16052532e-01 -4.94241416e-02 -5.32909930e-01
-7.25620031e-01 -8.58157754e-01 -3.01259011e-01 1.62096035e+00
-4.11564648e-01 -9.66742396e-01 -6.16304815e-01 -1.96189836e-01
-1.90606311e-01 1.32293984e-01 -2.36121744e-01 -1.93977654e-01
-6.13146067e-01 -6.21238768e-01 1.53504252e+00 7.16805637e-01
5.00730336e-01 -1.74209464e+00 -3.01924318e-01 5.88953495e-01
-2.85644174e-01 -1.09685552e+00 -5.69109738e-01 3.25415045e-01
-4.73968595e-01 -8.21584046e-01 -4.18851405e-01 -1.06402135e+00
-3.24762166e-01 -1.53677389e-01 1.06472564e+00 -1.75234243e-01
-2.72535264e-01 3.00635964e-01 -5.65044045e-01 -9.24261928e-01
2.35699844e-02 4.39248800e-01 5.80778480e-01 1.53127670e-01
5.64792871e-01 -8.60528350e-01 -3.12299937e-01 4.11524802e-01
-8.45076084e-01 -4.59655464e-01 3.63076299e-01 7.70512581e-01
7.48101175e-01 2.69811064e-01 8.87388885e-01 -2.69090980e-01
7.85432816e-01 -2.83914179e-01 -2.49558613e-01 -9.68943313e-02
-6.84106052e-02 -7.26454258e-01 6.84238791e-01 -9.70338523e-01
-4.63518322e-01 6.48865998e-02 -7.95914769e-01 -8.64383340e-01
-7.65253246e-01 6.70450449e-01 -1.96337730e-01 1.12401448e-01
4.72949982e-01 1.10101156e-01 -4.51844066e-01 -7.87639320e-01
-2.79240757e-02 1.13658154e+00 5.67932308e-01 -4.93155122e-01
3.54471117e-01 -1.28134470e-02 -4.30435807e-01 -1.08199298e+00
-1.15470982e+00 -6.41398549e-01 -4.16082531e-01 -2.89310634e-01
7.35160828e-01 -1.01481891e+00 -6.99535668e-01 7.50672996e-01
-1.01776576e+00 -3.92744809e-01 -4.48966712e-01 7.18642056e-01
-4.39757168e-01 -4.08592857e-02 -8.49959671e-01 -7.56144524e-01
-5.45549572e-01 -8.13109279e-01 1.06369150e+00 8.90734568e-02
-4.95327026e-01 -5.86602986e-01 4.29041564e-01 1.24758314e-02
5.50555468e-01 -4.02091414e-01 4.86194879e-01 -1.01016247e+00
-1.57247558e-01 -1.36048421e-01 1.15699589e-01 5.45916378e-01
3.58579122e-02 -2.21628144e-01 -1.65656471e+00 -1.95637703e-01
-1.33721069e-01 -6.02431357e-01 9.47823822e-01 5.49213767e-01
1.79059589e+00 -2.09682733e-01 -9.52463821e-02 6.11131370e-01
7.95665503e-01 7.14780927e-01 5.18292964e-01 2.18670592e-01
6.06192648e-01 1.89755291e-01 6.36410296e-01 5.27436733e-01
1.55848628e-02 9.05799687e-01 5.62489629e-02 1.09662227e-01
-1.04419388e-01 -2.07605347e-01 5.71658671e-01 1.54388177e+00
1.65146455e-01 -3.61256123e-01 -8.94195855e-01 6.10635340e-01
-1.70678675e+00 -9.59475338e-01 1.97374165e-01 1.68941379e+00
9.29090440e-01 4.24775332e-02 2.90396333e-01 8.94800782e-01
8.05525839e-01 2.67921180e-01 -1.65290281e-01 -3.25780302e-01
1.26951233e-01 8.15620065e-01 -2.83380836e-01 4.31131423e-02
-1.57141721e+00 9.39372778e-01 5.69877672e+00 1.33039188e+00
-1.45981157e+00 3.68163198e-01 3.98610383e-01 -3.20608079e-01
6.24785721e-01 -4.53606248e-01 -7.94729471e-01 4.63680446e-01
1.49341428e+00 1.78367153e-01 3.91664952e-01 9.15390074e-01
4.06760834e-02 4.02939528e-01 -1.10381901e+00 1.46829319e+00
1.01016752e-01 -1.11476505e+00 -1.22391030e-01 -2.24964231e-01
3.15906554e-01 2.03721926e-01 1.53214082e-01 7.02262700e-01
-1.91573992e-01 -9.39499438e-01 7.68157244e-01 2.04392016e-01
9.74796951e-01 -8.79905045e-01 8.23045611e-01 -9.77113172e-02
-1.74325776e+00 -2.26261869e-01 -3.80711228e-01 -2.71675736e-01
2.61291951e-01 5.45328140e-01 -9.35929000e-01 2.88682908e-01
1.25572121e+00 8.69411826e-01 -2.58578867e-01 1.50951314e+00
-4.45311032e-02 1.46096420e+00 -5.51143408e-01 -1.81130767e-02
-2.22511590e-02 3.56940210e-01 5.66951275e-01 1.50287092e+00
6.12073481e-01 -6.18246980e-02 3.57422501e-01 3.81861657e-01
-3.99779916e-01 3.32107961e-01 -2.07077801e-01 -2.51925558e-01
6.89440787e-01 1.34986353e+00 -4.92129028e-01 -1.91793203e-01
-8.64404142e-02 3.63830119e-01 -4.58675101e-02 8.69283378e-02
-8.07040513e-01 -7.94112921e-01 8.67762923e-01 -2.41788268e-01
4.32067364e-01 2.43322458e-02 -1.27932370e-01 -8.97863865e-01
3.36564109e-02 -8.29286277e-01 6.77935958e-01 -9.29139018e-01
-1.44694984e+00 1.00810277e+00 -3.07397455e-01 -1.55714786e+00
-1.91041484e-01 -6.32597685e-01 -8.89733970e-01 4.60486889e-01
-1.31521499e+00 -9.35022950e-01 -4.02388945e-02 7.52035022e-01
7.74070323e-01 -5.73797405e-01 1.18666923e+00 9.66143668e-01
-5.32419980e-01 7.42933512e-01 -3.51358831e-01 5.56588411e-01
9.33203638e-01 -1.03820050e+00 2.47774571e-01 6.20952785e-01
8.61093462e-01 1.29819393e-01 1.99719623e-01 -8.86614323e-02
-9.38247323e-01 -1.42995417e+00 9.73990440e-01 -1.04619339e-01
1.01004457e+00 -4.53625530e-01 -9.94941235e-01 4.80602264e-01
3.73694181e-01 1.03273958e-01 1.14410281e+00 4.72856194e-01
-3.32880437e-01 -6.46270275e-01 -5.39649427e-01 7.69437626e-02
8.77431631e-01 -9.41923320e-01 -4.26875770e-01 1.46319732e-01
8.46910655e-01 -4.84097540e-01 -1.04587257e+00 4.90658253e-01
5.43462336e-01 -3.74209106e-01 9.52231288e-01 -5.47435582e-01
2.25579858e-01 -4.41252172e-01 -4.04940903e-01 -1.39118969e+00
-2.84605950e-01 -6.52258337e-01 -4.04405259e-02 1.58786201e+00
4.10853505e-01 -2.25940287e-01 4.21572030e-01 -3.47983390e-01
-6.79714620e-01 -5.64216137e-01 -1.15013850e+00 -8.96663427e-01
-2.60747999e-01 -1.28276992e+00 7.09009469e-01 1.05532777e+00
1.16059408e-01 4.11074877e-01 -6.16563559e-01 3.36064667e-01
-5.60701406e-03 1.59074515e-01 6.50979519e-01 -1.31781685e+00
-4.09993827e-01 -3.74130666e-01 -7.36660063e-01 -8.95240307e-01
1.67723507e-01 -9.34561431e-01 2.95441836e-01 -9.72300768e-01
-1.76018387e-01 -3.44540417e-01 -1.01764202e+00 9.65012848e-01
2.55499035e-01 7.67985165e-01 2.82386243e-01 -2.24045403e-02
-8.04827929e-01 5.58216095e-01 6.51207209e-01 -3.42513621e-01
-4.24494416e-01 1.69800013e-01 -2.41353899e-01 8.55915368e-01
1.21634066e+00 -6.30353153e-01 -1.68804720e-01 -4.17983234e-01
7.00029135e-02 -7.82912374e-02 4.28610921e-01 -1.63757706e+00
4.36058223e-01 2.79772431e-01 2.16431230e-01 -7.44835496e-01
7.50557482e-01 -5.99287987e-01 1.41997099e-01 1.12965509e-01
-4.85968441e-01 -4.62597795e-02 6.15948856e-01 2.07645193e-01
-9.22720075e-01 -1.03316091e-01 5.86173654e-01 2.95773238e-01
-7.90187061e-01 4.59025770e-01 -6.21579945e-01 8.55769776e-03
3.84879649e-01 3.34504724e-01 -1.18276782e-01 -2.60573894e-01
-9.92397845e-01 -2.05440193e-01 -4.58722323e-01 6.88089967e-01
7.53016114e-01 -1.66123390e+00 -5.77260613e-01 2.79900581e-01
8.67008865e-02 -2.29936764e-01 4.71288025e-01 7.33735383e-01
-1.17421329e-01 5.34028947e-01 -2.76056796e-01 -8.59249473e-01
-1.57720900e+00 2.17656270e-01 2.76429892e-01 -4.76744287e-02
-3.67600471e-01 1.13142502e+00 -5.75073250e-02 -5.22544205e-01
7.65995443e-01 -7.39626944e-01 -2.64213741e-01 7.79897347e-02
7.50733256e-01 2.72230774e-01 3.22703898e-01 -4.04358536e-01
-3.64128828e-01 5.83074689e-01 1.09784983e-01 -2.62814872e-02
1.62186062e+00 2.88264602e-01 -8.23426396e-02 8.06132555e-01
9.96224403e-01 -5.56078441e-02 -6.52370393e-01 -3.52486014e-01
-5.00078239e-02 -1.79524258e-01 1.48540035e-01 -7.62884557e-01
-1.28342080e+00 1.34438384e+00 9.56800997e-01 4.55644071e-01
1.60317421e+00 4.97534114e-04 1.04367828e+00 5.47932565e-01
1.77303240e-01 -1.08681405e+00 2.07216427e-01 9.81859803e-01
9.07750785e-01 -9.70630229e-01 -5.74325085e-01 -1.57527193e-01
-6.29013240e-01 1.07728171e+00 6.15896642e-01 -1.55927181e-01
1.18467736e+00 5.27147353e-01 2.52377898e-01 -4.62151542e-02
-9.61427748e-01 -2.13439628e-01 6.31545126e-01 4.05214220e-01
5.68988562e-01 1.99642479e-01 3.32488537e-01 1.28287756e+00
-7.18848407e-01 1.77774921e-01 1.23302065e-01 8.07659328e-01
-3.81024331e-01 -1.10958159e+00 -3.83179367e-01 3.23220819e-01
-8.22023809e-01 -3.55940133e-01 -3.89931887e-01 3.60799789e-01
3.63059074e-01 9.96969461e-01 3.44265163e-01 -9.50370431e-01
4.40266937e-01 4.27732795e-01 1.85279280e-01 -6.35258734e-01
-9.40500319e-01 5.18373311e-01 8.06686655e-02 -3.39656383e-01
-6.08036757e-01 -4.41415548e-01 -9.92698848e-01 7.30923265e-02
-3.37720066e-01 4.76200104e-01 4.13828105e-01 6.90623522e-01
4.46258843e-01 9.48956430e-01 5.24902761e-01 -9.20961916e-01
-8.35071132e-02 -1.52366853e+00 -8.87305319e-01 5.21734655e-02
2.30718702e-01 -4.86888260e-01 -8.27078670e-02 2.83176005e-01] | [15.19260311126709, 5.179794788360596] |
87867575-9063-49d4-9c8f-02bedea54126 | layoutlmv3-pre-training-for-document-ai-with | 2204.08387 | null | https://arxiv.org/abs/2204.08387v3 | https://arxiv.org/pdf/2204.08387v3.pdf | LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking | Self-supervised pre-training techniques have achieved remarkable progress in Document AI. Most multimodal pre-trained models use a masked language modeling objective to learn bidirectional representations on the text modality, but they differ in pre-training objectives for the image modality. This discrepancy adds difficulty to multimodal representation learning. In this paper, we propose \textbf{LayoutLMv3} to pre-train multimodal Transformers for Document AI with unified text and image masking. Additionally, LayoutLMv3 is pre-trained with a word-patch alignment objective to learn cross-modal alignment by predicting whether the corresponding image patch of a text word is masked. The simple unified architecture and training objectives make LayoutLMv3 a general-purpose pre-trained model for both text-centric and image-centric Document AI tasks. Experimental results show that LayoutLMv3 achieves state-of-the-art performance not only in text-centric tasks, including form understanding, receipt understanding, and document visual question answering, but also in image-centric tasks such as document image classification and document layout analysis. The code and models are publicly available at \url{https://aka.ms/layoutlmv3}. | ['Furu Wei', 'Yutong Lu', 'Lei Cui', 'Tengchao Lv', 'Yupan Huang'] | 2022-04-18 | null | null | null | null | ['document-image-classification', 'document-layout-analysis', 'document-ai', 'semantic-entity-labeling', 'key-information-extraction'] | ['computer-vision', 'computer-vision', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 5.40986836e-01 2.67550386e-02 -3.17349672e-01 -5.82522273e-01
-9.87538874e-01 -7.52077162e-01 1.02724159e+00 2.44102553e-01
-2.01687321e-01 1.24280853e-02 3.38091880e-01 -6.88941777e-01
2.51297057e-01 -4.23328102e-01 -9.49949145e-01 -6.24448717e-01
4.80812997e-01 7.56326973e-01 -3.62969488e-01 -1.71527579e-01
2.95251250e-01 4.18228768e-02 -1.23437250e+00 1.32722008e+00
6.10793054e-01 8.62847030e-01 4.08859134e-01 1.19824159e+00
-6.66846693e-01 1.06666481e+00 -4.43108410e-01 -4.45693374e-01
-3.02229911e-01 -5.97536385e-01 -1.18994045e+00 4.02970374e-01
1.07762003e+00 -3.36808115e-01 -4.67202514e-01 7.67685831e-01
4.12654549e-01 -1.88068107e-01 1.06301832e+00 -1.19645441e+00
-1.30431879e+00 6.75589979e-01 -8.15370381e-01 -3.52151915e-02
4.70981061e-01 1.50965095e-01 1.24279475e+00 -1.27544177e+00
6.33530140e-01 1.36858630e+00 2.05967814e-01 6.37856364e-01
-1.30173516e+00 -2.34299317e-01 4.61721092e-01 3.76718342e-01
-1.31769276e+00 -4.87721711e-01 8.50576341e-01 -5.15496552e-01
9.33267117e-01 4.22000468e-01 1.69833675e-01 1.32073677e+00
6.14014342e-02 1.60027564e+00 1.00338745e+00 -8.01231205e-01
-2.12966785e-01 2.68646717e-01 1.76046237e-01 1.09402525e+00
-2.65119970e-01 -5.55346668e-01 -6.27717376e-01 1.84017897e-01
5.11301339e-01 -8.12681764e-02 -2.66764194e-01 -4.53281313e-01
-1.33664024e+00 8.03680539e-01 4.23059672e-01 4.28021133e-01
6.22877106e-02 3.32698882e-01 4.23525780e-01 2.31720701e-01
2.33390003e-01 1.56552196e-01 -8.93897265e-02 2.24309087e-01
-9.64359701e-01 -8.18442404e-02 4.22778070e-01 1.01194191e+00
8.60448301e-01 9.36025903e-02 -4.84698415e-01 1.02421856e+00
4.62424189e-01 9.59996045e-01 4.82284635e-01 -6.67208791e-01
9.17942822e-01 8.56562674e-01 -4.65884149e-01 -8.65452051e-01
-2.91035920e-01 -2.15877816e-02 -1.07371449e+00 2.17669969e-03
3.23646456e-01 3.11820477e-01 -1.43306172e+00 1.50284588e+00
-2.31008798e-01 -6.86048448e-01 1.98917419e-01 9.21361685e-01
1.29566693e+00 1.00004327e+00 5.48946895e-02 1.97811216e-01
1.53101873e+00 -1.30737257e+00 -6.84908211e-01 -5.32796323e-01
9.20365751e-01 -1.07412803e+00 1.58455086e+00 1.84222996e-01
-1.24503899e+00 -7.37769783e-01 -8.93524110e-01 -6.10710442e-01
-6.27896249e-01 6.09447598e-01 4.62076873e-01 5.71013868e-01
-1.11995149e+00 -4.83373612e-01 -4.05578375e-01 -4.95911807e-01
4.11210418e-01 1.46718472e-01 -5.03870666e-01 -5.65216780e-01
-6.26304507e-01 6.85198903e-01 1.98084116e-01 1.33421838e-01
-1.03669763e+00 -3.54097188e-01 -1.06700623e+00 4.85713780e-02
2.19446734e-01 -8.29298258e-01 1.13503385e+00 -1.25961626e+00
-1.09181213e+00 1.45404994e+00 -6.14854991e-01 -7.48968795e-02
2.09449217e-01 2.87230425e-02 -3.32034528e-01 4.03225839e-01
-3.92970182e-02 1.10216880e+00 1.13847470e+00 -1.85698891e+00
-1.49404258e-01 -4.99932259e-01 -1.31055132e-01 4.35611814e-01
-5.13318360e-01 -1.88028783e-01 -9.28772449e-01 -6.09672129e-01
3.31072062e-02 -7.08853185e-01 2.68445522e-01 -1.48211839e-03
-6.43691480e-01 -2.68538743e-01 1.10006416e+00 -6.48227096e-01
9.18003500e-01 -2.04479623e+00 4.26426023e-01 2.18176365e-01
4.28693034e-02 7.67604485e-02 -8.22561622e-01 7.01436758e-01
-5.51803708e-02 5.65654971e-02 -3.60164911e-01 -7.84361541e-01
1.49570495e-01 2.32982278e-01 -6.68682635e-01 3.05970103e-01
2.39827007e-01 1.41294503e+00 -4.40744489e-01 -8.09446514e-01
3.04198235e-01 4.89650816e-01 -2.95211852e-01 2.89208889e-01
-7.30302691e-01 2.23491147e-01 -6.80777058e-02 9.53129768e-01
5.79285800e-01 -6.07431889e-01 1.78427905e-01 -4.85567242e-01
2.13118538e-01 -3.62675264e-02 -6.66943312e-01 2.18574214e+00
-5.10787487e-01 1.13005376e+00 9.78349969e-02 -1.15396893e+00
7.27623045e-01 9.99275222e-02 1.68871433e-01 -1.14065647e+00
2.41817087e-01 -1.51981756e-01 -3.05150837e-01 -6.09714568e-01
7.49491513e-01 3.52403730e-01 -1.01425117e-02 7.80233622e-01
2.55399317e-01 -3.53595942e-01 3.32705051e-01 7.13199139e-01
6.28567457e-01 7.60747418e-02 -2.83885479e-01 -1.45403091e-02
8.00288498e-01 1.24764092e-01 -6.05037034e-01 8.33880007e-01
3.72015536e-01 8.84113371e-01 3.73111874e-01 -4.92565110e-02
-1.02668548e+00 -1.05312932e+00 1.18830860e-01 1.65782070e+00
2.02150598e-01 -6.21026874e-01 -5.70389688e-01 -7.26828873e-01
2.09604166e-02 6.55624747e-01 -6.61960900e-01 -7.83338323e-02
-5.50144613e-01 -4.42441911e-01 7.32861161e-01 6.89536512e-01
4.16614801e-01 -1.01271653e+00 2.83028260e-02 -4.42895204e-01
-3.82444024e-01 -1.12777829e+00 -7.34465957e-01 1.70089230e-01
-5.52599728e-01 -8.40426922e-01 -9.63987350e-01 -1.33154881e+00
1.21399772e+00 4.89313930e-01 1.10502481e+00 2.52655149e-01
-4.44107860e-01 1.20659387e+00 -2.56508857e-01 -2.09063351e-01
-3.54516357e-01 1.08923979e-01 -5.94001114e-01 1.43136904e-01
2.04386339e-01 7.11419657e-02 -5.25213957e-01 1.79921642e-01
-1.17941809e+00 3.37319970e-01 7.26625443e-01 8.58164251e-01
5.84664464e-01 -5.79948068e-01 -4.51795980e-02 -5.96862078e-01
7.19992161e-01 -1.22824401e-01 -1.98024690e-01 8.70254457e-01
-4.25306231e-01 4.89087068e-02 3.94883811e-01 -4.90544200e-01
-1.01654541e+00 -9.92261060e-03 9.41038430e-02 -3.33765894e-01
-2.67538220e-01 5.89965582e-01 -3.00829202e-01 2.00001732e-01
4.65612710e-01 6.26815200e-01 7.65994787e-02 -3.77795935e-01
9.43062484e-01 6.67469680e-01 7.00402617e-01 -6.94674015e-01
7.02218413e-01 6.37305140e-01 -1.55908182e-01 -1.05645382e+00
-7.35365927e-01 -4.97545928e-01 -6.99210167e-01 -2.38612816e-01
1.16374445e+00 -9.32398736e-01 -8.24850857e-01 3.84743929e-01
-1.23938608e+00 -5.52940249e-01 -1.20319589e-03 -7.59126768e-02
-5.45426607e-01 5.97277522e-01 -3.97779137e-01 -7.06729352e-01
-4.58125353e-01 -1.07321465e+00 1.50181293e+00 -7.34453872e-02
-1.22911483e-01 -1.19942772e+00 -8.60399827e-02 9.50396776e-01
1.59491315e-01 -2.51029462e-01 1.34344280e+00 -2.70287216e-01
-6.08474374e-01 -4.14497256e-02 -7.09409773e-01 3.44627023e-01
-4.98392433e-02 -1.34421244e-01 -1.12647581e+00 -4.27852452e-01
-5.28907359e-01 -8.21301281e-01 1.33748496e+00 2.96423912e-01
1.45593929e+00 -1.84123933e-01 -2.55817026e-01 6.62931740e-01
1.17628717e+00 -5.15973866e-02 5.75268269e-01 1.87651739e-01
1.21658075e+00 5.97413719e-01 1.94133103e-01 -3.77054401e-02
6.86545491e-01 5.12007773e-01 4.82383490e-01 -6.43648863e-01
-4.56903964e-01 -3.07152957e-01 3.57649505e-01 7.67578244e-01
2.54842401e-01 -8.42227101e-01 -1.25567532e+00 6.18577898e-01
-1.95565546e+00 -8.16811204e-01 -1.85648605e-01 1.62276137e+00
7.11312175e-01 -2.91659117e-01 -2.86321435e-02 -9.31025222e-02
3.67336005e-01 4.54210430e-01 -4.02734548e-01 -6.15624309e-01
-6.75213277e-01 -1.23332977e-01 2.25576982e-01 6.59610927e-01
-1.12955189e+00 1.14625287e+00 6.00986385e+00 7.04774201e-01
-1.05547178e+00 -1.88751724e-02 7.18471766e-01 9.09106061e-02
-6.91235542e-01 -1.66776881e-01 -6.52522385e-01 4.29115444e-03
4.80429441e-01 2.68534958e-01 1.92314461e-01 5.54011047e-01
-2.62243450e-01 -1.80907577e-01 -1.32269907e+00 1.26018465e+00
9.13178980e-01 -1.50247657e+00 8.77174675e-01 -1.58626754e-02
6.90593183e-01 -2.27134019e-01 6.72895730e-01 1.52084291e-01
-1.21300995e-01 -1.53311849e+00 6.77807033e-01 5.32516897e-01
9.78470922e-01 -4.88140523e-01 3.92680794e-01 7.23917410e-02
-1.13062441e+00 4.53342162e-02 -1.92704931e-01 3.55560899e-01
1.59421474e-01 9.31636542e-02 -8.33503902e-01 4.75987822e-01
5.44409633e-01 7.34913111e-01 -9.94085670e-01 4.05851245e-01
-1.09755926e-01 5.28955877e-01 1.34486288e-01 3.56307477e-02
5.44219971e-01 -9.55182239e-02 2.65396118e-01 1.57091761e+00
-6.61385730e-02 -2.70947486e-01 1.71418652e-01 8.37666571e-01
-2.15402767e-01 2.31945008e-01 -6.29508734e-01 -5.14603555e-01
-2.12691709e-01 1.08958292e+00 -4.58873838e-01 -3.59036416e-01
-5.49581110e-01 1.35475945e+00 2.07736880e-01 7.14712262e-01
-5.15902638e-01 -2.66153038e-01 1.99826285e-01 -2.34207455e-02
3.29910159e-01 -5.00956833e-01 -4.28887904e-01 -1.17210126e+00
4.11891378e-02 -1.05052602e+00 5.91605902e-01 -1.22306240e+00
-1.19787288e+00 5.28118491e-01 -4.94164377e-02 -8.25514138e-01
-2.61274189e-01 -1.18286335e+00 -4.96354669e-01 7.17227280e-01
-1.36551380e+00 -1.97767878e+00 -4.34952080e-01 1.06505978e+00
8.07204604e-01 -5.22001088e-01 8.87518287e-01 2.84819435e-02
-3.51917833e-01 8.56009126e-01 4.84697595e-02 3.24063748e-01
1.02238703e+00 -1.28422344e+00 7.44049177e-02 6.41972840e-01
6.41821623e-01 6.06085837e-01 3.28239799e-01 -4.32124287e-01
-1.99300361e+00 -8.57140124e-01 7.22736597e-01 -7.59389102e-01
5.36281109e-01 -7.54776895e-01 -9.80043471e-01 7.96969831e-01
9.36626673e-01 -3.74737620e-01 7.59889126e-01 6.76663816e-02
-8.59534562e-01 6.75162897e-02 -4.85945582e-01 8.07918131e-01
7.98135400e-01 -1.16056859e+00 -5.05521238e-01 7.01229632e-01
4.54808503e-01 -3.35181952e-01 -5.45593143e-01 2.29977056e-01
5.07482946e-01 -5.90518475e-01 1.14526618e+00 -7.27085531e-01
7.39230990e-01 -2.35814437e-01 -4.73494053e-01 -7.96335340e-01
-2.32745502e-02 -2.27799147e-01 -2.27887377e-01 1.30958343e+00
6.39897883e-01 -1.05493739e-01 6.40616655e-01 7.43771866e-02
3.34177800e-02 -5.08856654e-01 -6.07731879e-01 -4.15587902e-01
2.52714694e-01 -3.52266401e-01 6.22478873e-02 8.69995058e-01
2.24806089e-02 7.47746170e-01 -4.38638210e-01 5.91840968e-03
4.81264859e-01 3.65281314e-01 1.00913095e+00 -5.72343230e-01
-1.39909655e-01 -7.10888386e-01 -7.83739425e-03 -1.56659925e+00
3.86424333e-01 -1.27251816e+00 -4.87766080e-02 -2.02434945e+00
5.66499531e-01 2.19171226e-01 -1.20351324e-02 6.29447401e-01
-9.54573881e-03 3.65420103e-01 3.26375932e-01 2.04045787e-01
-8.45036864e-01 5.52353621e-01 1.36402822e+00 -9.93114769e-01
8.61571878e-02 -5.27599752e-01 -6.14439428e-01 4.30941314e-01
5.92324972e-01 1.77870870e-01 -5.85558474e-01 -1.21729577e+00
3.12638700e-01 -2.69509673e-01 5.55355728e-01 -3.72121125e-01
3.84656310e-01 3.12563181e-02 8.51002812e-01 -1.20637941e+00
6.58378899e-01 -7.13902056e-01 -7.64323890e-01 1.03550762e-01
-7.03959227e-01 2.39684775e-01 4.23873425e-01 4.01702762e-01
-4.47854370e-01 -1.30871177e-01 3.32423061e-01 -5.27025461e-02
-7.93601334e-01 4.71944176e-02 -6.05843663e-01 -2.87049916e-02
4.85948324e-01 -1.97472036e-01 -9.39618647e-01 -7.98067153e-01
-5.84998965e-01 4.59573776e-01 4.17641342e-01 6.72694743e-01
9.88993347e-01 -1.12268114e+00 -6.95024192e-01 2.00261340e-01
6.31124198e-01 -1.86505020e-01 3.86606485e-01 7.74816632e-01
-2.40747064e-01 7.93546736e-01 -3.59546556e-03 -1.14319098e+00
-1.62140739e+00 5.10965586e-01 7.30456561e-02 -1.96496602e-02
-3.99482101e-01 9.25378323e-01 6.47930503e-01 -5.61490178e-01
5.70544600e-01 -1.42116845e-01 -2.45778169e-02 -1.42893186e-02
5.41732311e-01 -1.93001285e-01 -1.79381177e-01 -7.75466859e-01
-3.60347271e-01 9.30119634e-01 -3.04821759e-01 -4.08423036e-01
7.82133341e-01 -3.58148396e-01 -2.80123085e-01 5.22891045e-01
1.45731068e+00 -2.13088095e-01 -6.37701213e-01 -2.81769633e-01
-1.16917342e-01 -1.39447182e-01 8.19758698e-02 -1.02261639e+00
-8.24288845e-01 1.30270684e+00 5.81558049e-01 -1.24808490e-01
1.10348332e+00 3.93340945e-01 4.55736637e-01 7.75137544e-01
-4.07267690e-01 -9.50086534e-01 7.51684546e-01 7.62248456e-01
1.27980018e+00 -1.55938375e+00 3.52141708e-02 -8.21471512e-02
-9.99944985e-01 1.11137605e+00 7.94299543e-01 3.63089591e-01
3.03828925e-01 2.32300279e-03 4.26351845e-01 -4.30985421e-01
-6.39222622e-01 -4.17740107e-01 9.89705086e-01 5.55836618e-01
7.17631996e-01 -1.50230631e-01 1.69397265e-01 2.00285122e-01
-1.44072687e-02 -7.35653400e-01 1.15763441e-01 1.11146212e+00
-2.28884593e-01 -1.10800481e+00 -5.32084525e-01 1.21047601e-01
1.12384044e-01 -3.04173529e-01 -1.09585381e+00 7.81150401e-01
-2.35152379e-01 9.19144392e-01 3.50143760e-01 -1.95805654e-01
4.54653911e-02 3.84926289e-01 8.64680588e-01 -4.12934810e-01
-3.63086611e-01 1.60884097e-01 2.89138667e-02 -3.54375035e-01
-3.96870732e-01 -2.97209442e-01 -1.24693060e+00 -4.04379144e-02
-6.46255612e-02 -1.81719631e-01 9.23135996e-01 9.26406205e-01
3.23992491e-01 4.53968823e-01 2.86764771e-01 -7.35710800e-01
6.62825704e-02 -9.05339539e-01 -2.71547940e-02 4.96200353e-01
4.62570757e-01 -1.88513532e-01 -5.80678843e-02 4.94363964e-01] | [11.214029312133789, 1.9756054878234863] |
f7d1b72c-7401-46e0-a3e7-a43f9f900198 | probabilistic-sequential-matrix-factorization | 1910.03906 | null | https://arxiv.org/abs/1910.03906v3 | https://arxiv.org/pdf/1910.03906v3.pdf | Probabilistic sequential matrix factorization | We introduce the probabilistic sequential matrix factorization (PSMF) method for factorizing time-varying and non-stationary datasets consisting of high-dimensional time-series. In particular, we consider nonlinear Gaussian state-space models where sequential approximate inference results in the factorization of a data matrix into a dictionary and time-varying coefficients with potentially nonlinear Markovian dependencies. The assumed Markovian structure on the coefficients enables us to encode temporal dependencies into a low-dimensional feature space. The proposed inference method is solely based on an approximate extended Kalman filtering scheme, which makes the resulting method particularly efficient. PSMF can account for temporal nonlinearities and, more importantly, can be used to calibrate and estimate generic differentiable nonlinear subspace models. We also introduce a robust version of PSMF, called rPSMF, which uses Student-t filters to handle model misspecification. We show that PSMF can be used in multiple contexts: modeling time series with a periodic subspace, robustifying changepoint detection methods, and imputing missing data in several high-dimensional time-series, such as measurements of pollutants across London. | ['Mark F. J. Steel', 'Ömer Deniz Akyildiz', 'Theodoros Damoulas', 'Gerrit J. J. van den Burg'] | 2019-10-09 | null | null | null | null | ['multivariate-time-series-imputation'] | ['time-series'] | [-4.19580452e-02 -3.88380975e-01 -9.84715968e-02 -2.29402632e-01
-7.13549256e-01 -6.31169140e-01 6.74204051e-01 -2.15246513e-01
-2.30624646e-01 7.16971397e-01 3.89697939e-01 -5.05926251e-01
-6.69114828e-01 -4.78595525e-01 -8.13277721e-01 -8.34305108e-01
-2.64200032e-01 3.82114708e-01 -2.26621702e-01 2.08441481e-01
-1.76686957e-01 8.21217179e-01 -1.36264145e+00 -1.71020076e-01
8.04610848e-01 6.05311632e-01 2.42101073e-01 7.99796700e-01
1.41163230e-01 1.78987607e-01 -2.54516900e-01 -9.73219275e-02
2.67375886e-01 2.73504090e-02 -1.75050110e-01 7.39049688e-02
4.21591587e-02 -2.10292429e-01 -5.37875175e-01 8.78655434e-01
2.04951480e-01 6.88088000e-01 7.57619381e-01 -1.27303493e+00
-3.35256517e-01 2.95776755e-01 -8.58382806e-02 1.31536201e-01
1.19637929e-01 -1.09571181e-01 5.56788385e-01 -8.52209985e-01
2.10470200e-01 1.54433072e+00 8.58016014e-01 1.97198138e-01
-1.59675825e+00 -4.66660619e-01 4.86076653e-01 1.12767465e-01
-1.47842288e+00 -4.51616466e-01 5.87909162e-01 -8.43144894e-01
9.06206787e-01 4.27993566e-01 4.53686208e-01 1.13425815e+00
4.58437830e-01 6.76415741e-01 1.06615293e+00 -6.40774965e-02
4.03614134e-01 -1.27643183e-01 3.10945451e-01 1.56904012e-01
1.77977279e-01 4.84481543e-01 -1.99678212e-01 -7.45097280e-01
7.69661903e-01 5.80364048e-01 -2.65874028e-01 -3.14405680e-01
-1.29494500e+00 6.73134804e-01 -1.72534630e-01 1.22363694e-01
-6.78027511e-01 8.96462202e-02 2.09568158e-01 2.58525759e-01
6.79199398e-01 4.97608259e-02 -8.15861523e-01 -2.60646552e-01
-1.00972843e+00 3.56396288e-01 7.96688497e-01 7.26339698e-01
8.19169760e-01 3.17966640e-01 -2.20882341e-01 4.59028333e-01
3.93609852e-01 1.22215426e+00 3.52983981e-01 -1.13685215e+00
2.55915076e-01 1.25314966e-02 6.90059483e-01 -8.56929958e-01
-6.39707267e-01 -4.08863872e-01 -1.00451410e+00 -3.17995489e-01
2.97669142e-01 -2.85128862e-01 -8.31413567e-01 2.07106018e+00
5.42368889e-01 9.34221327e-01 1.25459865e-01 5.83138049e-01
-1.57841295e-02 8.27155232e-01 -1.77672923e-01 -7.61030138e-01
1.04940701e+00 -2.69956142e-01 -9.85460043e-01 1.88379392e-01
3.53605449e-01 -5.36516845e-01 5.56931973e-01 4.84017938e-01
-7.60431409e-01 -5.42080343e-01 -3.98284078e-01 2.00348422e-01
-2.34189972e-01 2.53655910e-01 3.74885976e-01 5.12633085e-01
-8.46201837e-01 7.85565317e-01 -1.59226930e+00 -8.59207213e-02
-2.90331781e-01 2.95321494e-01 -4.20029163e-01 -1.09110020e-01
-1.25068641e+00 7.33696401e-01 2.00907275e-01 4.72083986e-01
-9.95328724e-01 -9.00272727e-01 -7.82532454e-01 -5.69984876e-02
1.91015288e-01 -8.74763191e-01 1.08624470e+00 -3.14302742e-01
-1.32515454e+00 -1.44692034e-01 -8.55556309e-01 -5.60070097e-01
3.61283749e-01 -4.24573630e-01 -1.01202524e+00 -9.94479060e-02
6.93425983e-02 -3.25416654e-01 1.51929986e+00 -7.02533126e-01
-3.99928659e-01 -4.87693995e-01 -2.64003187e-01 -5.02033718e-02
-1.68670237e-01 -1.62664175e-01 -2.77488887e-01 -8.88908088e-01
2.28810444e-01 -1.33416307e+00 -5.55210888e-01 -3.64180088e-01
-2.02207863e-01 -4.66107614e-02 7.62013257e-01 -1.01341248e+00
1.48528779e+00 -2.28347492e+00 4.96467769e-01 3.82328123e-01
-1.15210779e-01 -8.84558484e-02 -3.90270911e-02 6.81254566e-01
-2.15157285e-01 -3.83377284e-01 -4.15268153e-01 -5.30616879e-01
1.51271254e-01 7.78595686e-01 -6.89498842e-01 8.84586394e-01
9.95390490e-02 6.59176886e-01 -1.00171983e+00 2.68876582e-01
4.54649895e-01 6.89215243e-01 -4.37137395e-01 8.37943181e-02
3.09582520e-02 6.98199451e-01 -3.74212712e-01 2.00968549e-01
7.62047172e-01 1.05882749e-01 -4.59117591e-02 -2.41087958e-01
-4.03106868e-01 7.02768238e-03 -1.78919101e+00 1.61939359e+00
-6.40181541e-01 2.90702105e-01 2.04861104e-01 -1.10367620e+00
4.27139223e-01 5.86835623e-01 6.72994733e-01 -2.25610472e-02
-1.91560417e-01 4.91204858e-02 -3.87218952e-01 -5.33218443e-01
4.27464575e-01 -2.19313949e-01 -4.75494228e-02 3.17508191e-01
1.56995669e-01 1.03849500e-01 -4.04952057e-02 1.73797294e-01
9.80101705e-01 1.05063021e-01 1.42288357e-01 -3.19539428e-01
4.21079665e-01 -4.11735356e-01 8.16475630e-01 8.10536027e-01
5.85267991e-02 1.10662125e-01 8.04901645e-02 -1.64101109e-01
-8.30073357e-01 -1.26955700e+00 -4.29889888e-01 7.14603782e-01
-5.05754948e-01 -3.72916043e-01 -1.43897623e-01 -1.80096120e-01
3.65440190e-01 9.33042586e-01 -6.69757307e-01 -2.54073143e-01
-2.90027738e-01 -8.80852699e-01 8.27373862e-02 5.51636457e-01
-3.13982189e-01 -1.91733286e-01 5.10899723e-02 4.99256998e-01
-1.39887527e-01 -8.49359572e-01 -5.32259047e-01 4.53770995e-01
-1.11664486e+00 -7.76554108e-01 -6.32687807e-01 -8.42302293e-02
6.38436794e-01 2.47132868e-01 6.03086293e-01 -9.42834079e-01
6.01680614e-02 1.05198145e+00 1.12286117e-02 -3.17397088e-01
-3.56642842e-01 -5.45974135e-01 8.62652183e-01 5.12929559e-01
1.81488141e-01 -5.97666085e-01 -4.05855000e-01 2.54931062e-01
-1.02811050e+00 -3.00980866e-01 3.08950245e-01 1.06233120e+00
5.86874545e-01 3.65124524e-01 3.98090571e-01 -5.91782451e-01
4.46676701e-01 -7.69793630e-01 -7.96663761e-01 1.84125617e-01
-6.70421243e-01 2.61106640e-01 5.70146263e-01 -9.03030872e-01
-1.13947916e+00 2.29788437e-01 8.13123509e-02 -8.99823606e-01
-1.47691756e-01 7.37375021e-01 -8.09430182e-02 2.51759112e-01
3.72765630e-01 4.24308866e-01 1.27820894e-01 -8.81915331e-01
7.81589448e-01 4.65306133e-01 7.31659412e-01 -6.33911133e-01
1.01905346e+00 6.75496995e-01 2.41392523e-01 -9.03678536e-01
-5.54348171e-01 -9.54554737e-01 -7.69522071e-01 2.53445730e-02
4.90130991e-01 -1.40267277e+00 -5.30674040e-01 5.26049674e-01
-8.95146847e-01 -2.60069400e-01 -3.62110913e-01 1.05732787e+00
-6.31201565e-01 3.32006723e-01 -6.61615551e-01 -1.19873822e+00
2.21242473e-01 -7.92793274e-01 1.17811692e+00 -2.58025467e-01
-2.00041488e-01 -1.49379289e+00 4.07222539e-01 -2.60704249e-01
1.81671754e-01 1.25747025e-01 6.42638862e-01 -4.49100643e-01
-1.61292270e-01 -4.14182901e-01 2.23524064e-01 5.55421531e-01
3.97449315e-01 7.88621679e-02 -8.02934289e-01 -6.23891056e-01
3.17788690e-01 4.96485144e-01 7.18000710e-01 8.61043394e-01
7.56954789e-01 -4.81601089e-01 -4.84278321e-01 5.57562351e-01
1.07083750e+00 9.20423716e-02 5.75404987e-02 -2.26191401e-01
8.31106424e-01 4.25429851e-01 5.60691297e-01 8.25489521e-01
5.20529449e-01 6.38512909e-01 1.47927627e-01 2.78942078e-01
4.35864836e-01 -3.27568561e-01 5.90669334e-01 1.01561344e+00
2.00815722e-01 1.60756007e-01 -8.28635037e-01 7.87683964e-01
-2.12493992e+00 -1.09944236e+00 -4.75265563e-01 2.52803683e+00
5.22868097e-01 -3.93703729e-01 1.95309192e-01 1.01748608e-01
5.50377190e-01 -1.64828420e-01 -8.12260091e-01 1.05924057e-02
-3.83333713e-02 9.88380760e-02 7.29898095e-01 7.58561552e-01
-1.17915785e+00 3.14674199e-01 6.93688154e+00 5.31439364e-01
-8.75672698e-01 3.12198013e-01 -1.60727069e-01 8.19734007e-04
-3.41174066e-01 1.15067050e-01 -7.45687008e-01 5.36309958e-01
1.58801746e+00 -4.44875419e-01 6.33156717e-01 5.46044827e-01
8.43844354e-01 2.58355141e-01 -9.47362661e-01 1.02486408e+00
-3.06198418e-01 -8.46562862e-01 -8.99840370e-02 1.06404603e-01
9.24308062e-01 2.09226519e-01 2.12689310e-01 4.16654974e-01
4.97952312e-01 -5.71178377e-01 5.10084689e-01 1.07406509e+00
5.21524549e-01 -7.58126199e-01 3.22196186e-01 5.98481894e-01
-1.17968631e+00 -4.15952235e-01 -2.94307142e-01 -2.01031476e-01
5.60442924e-01 1.06353545e+00 -4.99142140e-01 6.81079447e-01
5.93825579e-01 1.12679124e+00 -6.37588575e-02 1.05941355e+00
4.78757545e-02 8.69130313e-01 -7.77647734e-01 5.70215762e-01
-1.44721353e-02 -5.77648580e-01 8.88470292e-01 9.82021332e-01
7.56638706e-01 1.71966270e-01 3.55359018e-01 5.78622997e-01
6.13487303e-01 -2.74255127e-01 -4.19946760e-01 -1.63162410e-01
3.70138675e-01 9.86838996e-01 -2.63135612e-01 -4.59255517e-01
-7.11890042e-01 9.74737048e-01 -2.23180801e-01 9.46507275e-01
-6.93193316e-01 1.13224061e-02 1.18205035e+00 -9.99715552e-02
5.64300716e-01 -8.10065925e-01 2.95038819e-01 -1.69790149e+00
-1.19300582e-01 -7.25847840e-01 3.49444091e-01 -4.49064940e-01
-1.41414750e+00 1.45528778e-01 2.54111528e-01 -1.25063777e+00
-5.92457592e-01 -5.28890014e-01 -3.60416681e-01 1.22251821e+00
-1.22168136e+00 -1.04438853e+00 2.86196083e-01 1.07267940e+00
3.01999867e-01 1.73141971e-01 8.68206203e-01 3.00249726e-01
-6.80864394e-01 -1.12800859e-02 8.87483656e-01 -3.67618322e-01
7.40728199e-01 -1.38250709e+00 5.76327980e-01 1.19395387e+00
1.66764110e-01 1.16566956e+00 1.09517384e+00 -8.05974603e-01
-1.69962633e+00 -1.49958885e+00 6.61342144e-01 -6.52956069e-01
1.18452024e+00 -4.09805119e-01 -1.10914052e+00 9.27591801e-01
-5.91627657e-01 6.45238608e-02 8.54713619e-01 2.43919134e-01
-3.05964202e-01 3.14382650e-02 -8.52164805e-01 3.84349972e-01
6.88097954e-01 -9.44779694e-01 -6.01177812e-01 4.74837780e-01
6.90125942e-01 -3.90191525e-01 -1.16062474e+00 2.89418131e-01
5.31698465e-01 -2.03717902e-01 1.01265454e+00 -9.97861266e-01
-5.32890201e-01 -7.32277632e-01 -3.50665033e-01 -1.59736609e+00
-8.55611920e-01 -9.96436715e-01 -7.89859593e-01 9.79211867e-01
1.56035513e-01 -8.45199943e-01 2.22706035e-01 9.93859887e-01
-3.27427872e-02 -6.45290241e-02 -1.16375399e+00 -1.13445783e+00
-7.52536133e-02 -1.04644632e+00 6.01303577e-01 8.49308968e-01
-2.00823829e-01 -1.36278272e-01 -8.66049528e-01 8.43144238e-01
8.28380942e-01 1.18479207e-01 6.99753642e-01 -1.42289066e+00
-6.37491107e-01 1.30262688e-01 -3.32126915e-01 -1.20569634e+00
4.97685343e-01 -5.72970748e-01 5.52924946e-02 -9.80043888e-01
-1.64651528e-01 -1.85122445e-01 -5.52679777e-01 2.69355774e-01
-2.88897276e-01 -2.86433458e-01 -2.36485638e-02 1.90690160e-01
-1.32708296e-01 6.87492490e-01 6.38771832e-01 -1.89683940e-02
-4.38631743e-01 6.31008029e-01 -3.68643314e-01 4.23533827e-01
3.37224513e-01 -4.13963079e-01 -6.52468860e-01 -5.37509695e-02
1.66994736e-01 3.69622320e-01 4.77338195e-01 -6.81400418e-01
1.80550829e-01 -5.33924282e-01 3.06557536e-01 -7.68855929e-01
3.94158453e-01 -1.08504438e+00 8.82615924e-01 4.23049778e-01
1.59887262e-02 1.15135051e-01 3.20744604e-01 1.30571866e+00
-6.43341169e-02 1.61856890e-01 2.70829231e-01 2.13299632e-01
-5.22617161e-01 5.05641341e-01 -7.41895199e-01 -4.71957326e-01
6.46892965e-01 2.45614275e-01 1.94588780e-01 -5.60827434e-01
-1.41573048e+00 3.15759569e-01 1.56020075e-01 4.72762197e-01
3.65811646e-01 -1.45195806e+00 -5.41872621e-01 3.52348953e-01
-4.56563681e-02 -3.57511580e-01 7.92017698e-01 1.18948507e+00
3.71794432e-01 6.02405250e-01 3.94238085e-01 -8.28740418e-01
-9.00724113e-01 9.52639818e-01 8.12280923e-02 -1.55064072e-02
-4.93876994e-01 4.22514051e-01 1.09966218e-01 -4.66009498e-01
-9.93437916e-02 -8.13797712e-01 1.17671117e-01 2.31356800e-01
7.28653669e-01 6.05993629e-01 1.76344961e-02 -8.29844117e-01
-3.52257997e-01 3.40687245e-01 2.89732307e-01 -3.15944225e-01
1.27478349e+00 -6.97948396e-01 -1.40462920e-01 1.12851882e+00
1.28181088e+00 -7.24239722e-02 -1.48523879e+00 -5.42991757e-01
-1.21813655e-01 -3.06870401e-01 3.03454638e-01 -4.19439554e-01
-4.90460277e-01 5.83984733e-01 5.92073321e-01 1.04376011e-01
1.02245808e+00 -3.85893166e-01 4.66641247e-01 2.83675134e-01
4.75543767e-01 -6.89571440e-01 -6.76402688e-01 5.44313252e-01
7.57733941e-01 -9.51463997e-01 -1.99395284e-01 -2.43892185e-02
-1.67116195e-01 1.00954759e+00 -2.54352599e-01 -1.37265131e-01
1.31049526e+00 1.50030836e-01 -1.67227790e-01 2.81578362e-01
-9.85166788e-01 -1.64388448e-01 5.99437237e-01 7.17867911e-01
-1.10472394e-02 5.06302416e-01 -1.61664873e-01 6.67302787e-01
-3.12289763e-02 -4.77627814e-02 4.08816308e-01 5.32944977e-01
-3.13919894e-02 -1.01717055e+00 -7.33969688e-01 4.75833535e-01
-2.67892689e-01 -3.81410085e-02 4.11761194e-01 3.65237296e-01
-1.47873834e-01 9.96028006e-01 9.70378518e-02 -1.44109920e-01
3.84217590e-01 3.97826046e-01 8.46720859e-02 -4.14702952e-01
6.26261681e-02 4.58994508e-01 -1.70861542e-01 -8.79427969e-01
-3.95742983e-01 -1.35575342e+00 -7.66165972e-01 -1.31082296e-01
-2.78601378e-01 4.26056564e-01 7.83021808e-01 1.11052954e+00
5.27676642e-01 4.29822356e-01 6.89448535e-01 -9.84906614e-01
-8.36942196e-01 -8.65736783e-01 -9.16649997e-01 3.56976628e-01
9.19263959e-01 -7.77094543e-01 -7.53940761e-01 1.95723757e-01] | [6.799228668212891, 3.6792781352996826] |
f130e550-9989-4aee-afc3-fb7c5126a896 | self-attention-fusion-for-audiovisual-emotion | 2201.11095 | null | https://arxiv.org/abs/2201.11095v1 | https://arxiv.org/pdf/2201.11095v1.pdf | Self-attention fusion for audiovisual emotion recognition with incomplete data | In this paper, we consider the problem of multimodal data analysis with a use case of audiovisual emotion recognition. We propose an architecture capable of learning from raw data and describe three variants of it with distinct modality fusion mechanisms. While most of the previous works consider the ideal scenario of presence of both modalities at all times during inference, we evaluate the robustness of the model in the unconstrained settings where one modality is absent or noisy, and propose a method to mitigate these limitations in a form of modality dropout. Most importantly, we find that following this approach not only improves performance drastically under the absence/noisy representations of one modality, but also improves the performance in a standard ideal setting, outperforming the competing methods. | ['Moncef Gabbouj', 'Alexandros Iosifidis', 'Kateryna Chumachenko'] | 2022-01-26 | null | null | null | null | ['facial-emotion-recognition'] | ['computer-vision'] | [ 5.88476360e-01 6.76364526e-02 2.19195075e-02 -3.11648250e-01
-1.00906134e+00 -5.65885127e-01 8.90579522e-01 4.66963416e-03
-7.44474888e-01 8.49331558e-01 3.95725995e-01 -8.89262091e-03
-2.03113854e-01 -3.02788466e-01 -7.11755633e-01 -8.50897372e-01
1.19957805e-01 2.83807844e-01 -1.46864235e-01 -2.46478394e-01
-1.03472024e-01 2.37874612e-01 -1.88210332e+00 4.05810356e-01
5.32153726e-01 1.01635790e+00 -3.01454067e-01 8.67175460e-01
2.20060512e-01 8.70603681e-01 -4.63641852e-01 -5.96388280e-01
2.17295393e-01 -3.07207465e-01 -8.31128597e-01 2.13050097e-01
5.42728484e-01 -2.36335143e-01 -4.39512163e-01 9.80497122e-01
8.31431389e-01 3.42931032e-01 5.54313838e-01 -1.47843468e+00
-2.32819140e-01 4.88946289e-01 -3.00852120e-01 -2.35622022e-02
7.38341630e-01 4.41744439e-02 1.02605832e+00 -7.76443064e-01
4.96644676e-01 1.38232565e+00 5.49086332e-01 7.56770909e-01
-1.59106553e+00 -3.01754981e-01 3.67543668e-01 1.09154150e-01
-1.33439851e+00 -9.44390655e-01 8.35980296e-01 -2.25917473e-01
8.48044276e-01 2.86949307e-01 1.48183510e-01 1.72935057e+00
-2.92109460e-01 1.13093436e+00 1.09916735e+00 -5.11333406e-01
3.04349035e-01 1.35190234e-01 2.26559356e-01 2.29225695e-01
-3.06466699e-01 8.17731693e-02 -9.36047971e-01 -4.55373585e-01
1.83500737e-01 -2.17050925e-01 -3.15560788e-01 -2.77670234e-01
-1.11322713e+00 6.53676212e-01 -1.01453751e-01 2.30710074e-01
-3.81185681e-01 -3.48850130e-03 4.96092021e-01 6.30147040e-01
4.00084734e-01 1.06596813e-01 -4.14726734e-01 -3.00410300e-01
-1.10210693e+00 2.51188248e-01 9.40019608e-01 7.13679552e-01
4.24709797e-01 1.64775148e-01 -1.71750993e-01 6.84212327e-01
1.65340558e-01 3.42616290e-01 2.48019800e-01 -1.10890698e+00
5.40054381e-01 1.43215016e-01 2.56550461e-01 -6.17524266e-01
-5.22296786e-01 -1.98338762e-01 -7.24270940e-01 1.55939072e-01
6.95450127e-01 -3.38737607e-01 -1.05660141e+00 2.34429789e+00
1.44500390e-01 4.54828709e-01 3.55350018e-01 9.11715209e-01
9.40652966e-01 3.65598828e-01 2.33406276e-01 -4.34929252e-01
1.17679191e+00 -5.26443064e-01 -1.09205699e+00 -1.39743537e-01
2.38240719e-01 -6.36734247e-01 9.04971242e-01 7.35571980e-01
-1.24309444e+00 -3.48297417e-01 -9.50933516e-01 -1.32441178e-01
-4.62621450e-01 -1.35237768e-01 6.30007565e-01 7.25096703e-01
-1.21892357e+00 3.72922093e-01 -7.45301485e-01 -4.83298033e-01
5.46114817e-02 5.35082757e-01 -6.05500221e-01 -8.34353939e-02
-1.33246779e+00 8.66579413e-01 2.31363937e-01 2.74570167e-01
-9.23367023e-01 -3.71166497e-01 -9.59658980e-01 1.46313012e-02
5.67921281e-01 -7.52857387e-01 1.30361223e+00 -1.40344274e+00
-1.52069950e+00 6.16213799e-01 -4.49461937e-01 -2.54541159e-01
5.59842288e-01 -3.72385591e-01 -5.37383914e-01 1.22266717e-01
-4.91097689e-01 5.55635273e-01 1.03890240e+00 -1.47292745e+00
-5.01310825e-01 -2.83529758e-01 2.78203130e-01 2.43299991e-01
-2.58786976e-01 1.00045353e-01 -3.91831785e-01 -4.00989532e-01
-1.32546619e-01 -7.13805139e-01 -3.60115469e-02 -2.56723106e-01
-4.81911778e-01 -2.39818357e-02 6.62995160e-01 -4.52655196e-01
9.17048693e-01 -2.41092134e+00 5.33736825e-01 2.68055111e-01
6.82631209e-02 2.82790400e-02 -3.39198381e-01 5.93413770e-01
-1.38799533e-01 7.68426433e-02 -3.51426721e-01 -9.72631991e-01
2.07064420e-01 4.90917146e-01 -3.54335696e-01 3.63943189e-01
4.91944849e-01 6.56672537e-01 -7.09864378e-01 -4.85800952e-01
1.37071773e-01 8.86137724e-01 -4.76483941e-01 2.58605957e-01
-2.08181422e-02 7.04896390e-01 -1.94334447e-01 7.57279932e-01
5.96326947e-01 7.99097773e-03 3.02865207e-01 -2.02185050e-01
1.19357847e-01 1.15768537e-01 -1.62584591e+00 1.87581408e+00
-3.27764630e-01 5.00329792e-01 6.01783633e-01 -9.77587759e-01
2.95075506e-01 8.87214720e-01 4.64764774e-01 -6.44830704e-01
1.83262512e-01 4.03216258e-02 -2.00134795e-02 -6.51259720e-01
4.76134181e-01 -3.63476276e-01 -1.68122113e-01 2.16320366e-01
4.80533123e-01 1.47078261e-01 4.02798839e-02 2.42182910e-01
1.00114501e+00 1.88804224e-01 1.19734108e-01 2.73797572e-01
3.32265109e-01 -4.54780698e-01 5.38119972e-01 1.20144081e+00
-2.82753944e-01 8.44313264e-01 5.70392430e-01 2.02702507e-01
-7.41713345e-01 -1.07975101e+00 -1.85073480e-01 1.35183227e+00
1.03130499e-02 -3.57727349e-01 -3.45431030e-01 -5.11605203e-01
-7.90542066e-02 6.65906250e-01 -6.16481721e-01 -1.66445240e-01
-2.18779907e-01 -9.47198987e-01 8.26665938e-01 4.79854733e-01
9.39297974e-02 -8.82119715e-01 -3.84396195e-01 -7.94116557e-02
-3.70847195e-01 -1.24789214e+00 5.28727546e-02 4.01077271e-01
-5.74309051e-01 -7.77048111e-01 -5.31105518e-01 -3.64918232e-01
1.07341416e-01 -2.48306200e-01 1.00571907e+00 -1.32449672e-01
8.15870389e-02 9.08210337e-01 -3.37061703e-01 -2.96809345e-01
-1.90865099e-01 -1.16765812e-01 1.44442812e-01 5.33941448e-01
2.16412202e-01 -5.80635369e-01 -3.07586193e-01 -2.15789661e-01
-1.30147350e+00 -1.26299873e-01 3.51780593e-01 9.09839272e-01
2.51269162e-01 4.72310148e-02 6.01565182e-01 -7.31954634e-01
6.30851746e-01 -6.37274802e-01 -9.26449373e-02 1.82629496e-01
-3.05031151e-01 2.28951331e-02 4.50869590e-01 -5.88713229e-01
-1.28221774e+00 2.70703435e-01 -2.63474375e-01 -4.83315915e-01
-6.14055932e-01 6.16118193e-01 -5.11681199e-01 8.51683095e-02
2.73634851e-01 2.84526311e-02 -4.62987982e-02 -6.24231040e-01
4.39621985e-01 5.21544814e-01 5.73317766e-01 -6.58298135e-01
5.24875998e-01 5.73990405e-01 1.63593590e-02 -8.04506421e-01
-5.59854507e-01 -3.48074645e-01 -5.44646442e-01 -2.31933221e-01
6.44632339e-01 -9.49865401e-01 -8.87151539e-01 3.59886169e-01
-1.01367855e+00 -5.73783107e-02 -2.85310119e-01 5.92650950e-01
-5.51922023e-01 5.12727618e-01 -7.50893950e-01 -1.35398328e+00
3.02363306e-01 -1.22639894e+00 1.23463452e+00 1.97555542e-01
-1.10613622e-01 -1.13795090e+00 6.46722168e-02 2.64333878e-02
3.58117074e-01 3.31645042e-01 7.01040387e-01 -8.78283560e-01
-1.07482761e-01 -2.23474413e-01 1.35000557e-01 4.09485012e-01
-9.87503752e-02 -9.24984813e-02 -1.69008136e+00 -3.08124959e-01
3.53933275e-02 -6.43832445e-01 1.06332588e+00 2.23993570e-01
7.89553344e-01 -3.88863795e-02 5.30836135e-02 3.49561304e-01
1.33348942e+00 -8.09603408e-02 4.26753044e-01 5.34103587e-02
5.98501921e-01 7.34931827e-01 2.16667026e-01 4.82239157e-01
3.76359999e-01 6.51233971e-01 6.08995795e-01 -2.24342942e-01
1.67602435e-01 2.61786357e-02 5.03060758e-01 5.85961998e-01
-9.15316194e-02 -6.81681871e-01 -6.06284320e-01 5.47553182e-01
-2.22718382e+00 -1.02245677e+00 1.31836072e-01 2.24401712e+00
7.35489547e-01 -7.79056773e-02 3.18881094e-01 3.23327333e-01
4.16239113e-01 2.21630782e-01 -2.94842631e-01 -4.22901750e-01
-6.72447622e-01 2.72733927e-01 9.42312628e-02 6.57878697e-01
-1.34077287e+00 4.18830186e-01 7.83749533e+00 4.27121222e-01
-9.28701997e-01 1.39927000e-01 2.39353254e-01 -5.21819830e-01
-2.90148735e-01 2.52112597e-02 -2.66885221e-01 3.11630458e-01
1.14496350e+00 4.19820100e-01 5.13090312e-01 2.39714146e-01
1.31983370e-01 -3.32432121e-01 -1.47282612e+00 1.02916491e+00
2.76465982e-01 -4.74308550e-01 -4.45291810e-02 -1.67190179e-01
2.86404610e-01 -1.47409767e-01 1.46267548e-01 2.32181713e-01
7.06473887e-02 -1.07902634e+00 8.06707561e-01 9.57619667e-01
4.27333236e-01 -6.89006388e-01 8.73245597e-01 4.17692959e-01
-7.82362223e-01 -1.68157309e-01 2.88420260e-01 -2.46798962e-01
3.88528168e-01 3.89610231e-01 -2.04498485e-01 9.25608695e-01
5.22790372e-01 4.58823830e-01 -5.90137661e-01 1.05965066e+00
7.73919523e-02 8.55997980e-01 -6.32484436e-01 5.37053764e-01
9.28045139e-02 7.38717467e-02 7.56484687e-01 1.41210783e+00
1.07656598e-01 -9.88425240e-02 2.76945829e-01 3.68262291e-01
-1.64830103e-03 -1.35467365e-01 -7.38064110e-01 -2.86860596e-02
3.59226525e-01 9.85736251e-01 -3.01648736e-01 -2.83778459e-01
-7.06919789e-01 9.69382167e-01 2.79048979e-01 8.80375504e-01
-5.46014607e-01 -1.75237566e-01 3.63857508e-01 -3.50067794e-01
1.91040799e-01 -6.98923692e-02 -1.58329144e-01 -1.38708425e+00
1.42216891e-01 -9.87800896e-01 7.74259388e-01 -7.58612394e-01
-1.47201800e+00 3.53660583e-01 1.25363037e-01 -1.00755584e+00
-4.45903778e-01 -6.42613292e-01 -4.35779035e-01 8.33653927e-01
-1.52073967e+00 -1.26566470e+00 8.50193799e-02 7.96596587e-01
1.88933030e-01 1.17106810e-01 9.20310318e-01 6.47809803e-01
-6.42635286e-01 7.07068205e-01 1.11523882e-01 -1.22025706e-01
9.93522346e-01 -1.24441588e+00 -4.98439848e-01 8.18706334e-01
2.12864533e-01 7.04780877e-01 1.17890930e+00 -2.59887367e-01
-1.61108947e+00 -4.45218474e-01 7.70473182e-01 -5.21611631e-01
5.05258024e-01 -6.26391768e-01 -9.67407465e-01 7.25773454e-01
5.99595070e-01 -1.99391857e-01 9.00558174e-01 3.84411573e-01
-5.14180064e-01 1.78850248e-01 -1.17660141e+00 4.79814023e-01
7.24494338e-01 -9.12858129e-01 -7.35113800e-01 6.63843676e-02
4.39518899e-01 -1.99500009e-01 -8.80739570e-01 6.03219748e-01
6.25384092e-01 -8.97020161e-01 9.04609799e-01 -9.48077977e-01
1.58320308e-01 -2.26575077e-01 -6.31789505e-01 -1.09831166e+00
1.28315836e-01 -7.55359113e-01 -3.81106496e-01 1.47532058e+00
3.58581215e-01 -4.49040920e-01 4.14249480e-01 9.10353959e-01
1.04439542e-01 -1.99361697e-01 -1.41442990e+00 -4.67063040e-01
4.06195922e-03 -8.52071166e-01 2.75505751e-01 9.21307862e-01
1.23657465e-01 3.67160589e-01 -8.76326144e-01 1.81260765e-01
4.25493747e-01 -1.68559253e-01 5.15238225e-01 -1.09103489e+00
-4.85773444e-01 -4.08362091e-01 -4.60598052e-01 -7.43785799e-01
4.05172765e-01 -5.41791797e-01 1.10746995e-01 -1.21984553e+00
3.24083567e-01 2.84718394e-01 -6.79925919e-01 4.56630647e-01
-1.91214859e-01 3.90066385e-01 2.60403156e-01 -4.51441796e-04
-8.55494261e-01 5.79403281e-01 6.58327639e-01 -9.73760411e-02
-7.90494755e-02 2.15548463e-02 -6.79797769e-01 5.87202609e-01
4.15153861e-01 -3.54786843e-01 -3.82324457e-01 -2.98442155e-01
3.41491163e-01 6.75145239e-02 7.17644155e-01 -6.80240333e-01
1.63879290e-01 -6.81076869e-02 3.41210008e-01 -3.04101527e-01
9.32004690e-01 -1.05140901e+00 2.02091157e-01 -1.31041974e-01
-6.15758359e-01 -8.46985802e-02 3.11108708e-01 6.84474230e-01
-3.77824306e-01 -1.61976337e-01 4.18762237e-01 1.39727801e-01
-6.44399107e-01 -2.05716193e-01 -6.16976321e-01 -9.61538777e-02
6.00828767e-01 -1.13333337e-01 -3.06810170e-01 -6.90709114e-01
-1.28959703e+00 2.40769461e-01 4.30412382e-01 3.61983806e-01
4.30715859e-01 -1.39467049e+00 -6.46930814e-01 1.42228037e-01
1.23210609e-01 -5.09129524e-01 3.34949195e-01 1.30931497e+00
4.60759550e-01 8.17529485e-02 6.00547669e-03 -5.35191834e-01
-1.19947195e+00 5.83817065e-01 2.80883193e-01 -2.07207464e-02
-2.11552456e-01 3.87653500e-01 5.96203059e-02 -3.77301633e-01
6.22159243e-01 9.80277434e-02 -3.47982258e-01 3.32737833e-01
4.18939561e-01 1.53989017e-01 2.33901516e-01 -9.51399565e-01
-4.22599196e-01 2.02648222e-01 8.00860673e-02 -6.18251562e-01
1.07330453e+00 -3.41109812e-01 1.36455640e-01 1.14686632e+00
1.03946936e+00 1.85417628e-03 -1.12801027e+00 -3.58720541e-01
3.13287005e-02 -2.98921824e-01 3.88145372e-02 -9.14362490e-01
-8.51360917e-01 9.04171050e-01 8.16317737e-01 4.15428221e-01
1.47041273e+00 5.85457832e-02 2.42302239e-01 4.19874907e-01
5.83165046e-03 -1.12454271e+00 -2.62325883e-01 5.31333208e-01
5.61654866e-01 -1.43982983e+00 -2.10900486e-01 -7.53900260e-02
-9.43069875e-01 8.74828339e-01 2.82377005e-01 1.53918013e-01
5.83867490e-01 4.13487136e-01 1.14451207e-01 -1.72907948e-01
-1.03312969e+00 -6.58042789e-01 4.31993335e-01 7.18326747e-01
5.64765692e-01 -2.22038671e-01 -1.70027688e-01 7.38737822e-01
3.40628624e-01 -8.58184136e-03 3.83742422e-01 1.31022835e+00
6.21592216e-02 -1.01422870e+00 -5.18779337e-01 3.20849985e-01
-7.30216980e-01 -4.42019291e-02 -7.52291501e-01 1.01951146e+00
3.34416777e-02 1.35748506e+00 -8.29272643e-02 -3.18838120e-01
5.05427659e-01 6.53670430e-01 5.63474953e-01 -1.72354758e-01
-7.57544935e-01 4.46116477e-01 3.33737552e-01 -5.78814507e-01
-9.99183416e-01 -7.78704762e-01 -8.37391496e-01 -1.90935452e-02
-2.92720586e-01 2.54253894e-02 4.12535667e-01 1.16351950e+00
3.21890831e-01 4.78269339e-01 3.82963032e-01 -9.56250906e-01
-4.97388601e-01 -9.78419363e-01 -7.22531736e-01 7.72000432e-01
8.87638032e-01 -8.61903548e-01 -6.08995378e-01 1.06583223e-01] | [13.22264575958252, 5.012643337249756] |
294e7d27-c633-4ad9-879b-9868c67d9cfe | when-do-graph-neural-networks-help-with-node | 2304.14274 | null | https://arxiv.org/abs/2304.14274v2 | https://arxiv.org/pdf/2304.14274v2.pdf | When Do Graph Neural Networks Help with Node Classification: Investigating the Homophily Principle on Node Distinguishability | Homophily principle, i.e. nodes with the same labels are more likely to be connected, has been believed to be the main reason for the performance superiority of Graph Neural Networks (GNNs) over node-based Neural Networks on Node Classification tasks. Recent research suggests that, even in the absence of homophily, the advantage of GNNs still exists as long as nodes from the same class share similar neighborhood patterns. However, this argument only considers intra-class Node Distinguishability (ND) and neglects inter-class ND, which provides incomplete understanding of homophily. In this paper, we first demonstrate the aforementioned insufficiency with examples and argue that an ideal situation for ND is to have smaller intra-class ND than inter-class ND. To formulate this idea, we propose Contextual Stochastic Block Model for Homophily (CSBM-H) and define two metrics, Probabilistic Bayes Error (PBE) and negative generalized Jeffreys divergence, to quantify ND, through which we can find how intra- and inter-class ND influence ND together. We visualize the results and give detailed analysis. Through experiments, we verified that the superiority of GNNs is indeed closely related to both intra- and inter-class ND regardless of homophily levels, based on which we propose a new performance metric beyond homophily, which is non-linear and feature-based. Experiments indicate it significantly more effective than the existing homophily metrics on revealing the advantage and disadvantage of GNNs on both synthetic and benchmark real-world datasets. | ['Doina Precup', 'Jure Leskovec', 'Jie Fu', 'Xiao-Wen Chang', 'Jiaqi Zhu', 'Qincheng Lu', 'Minkai Xu', 'Chenqing Hua', 'Sitao Luan'] | 2023-04-25 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [-3.50767285e-01 9.61254165e-02 -4.04665291e-01 -3.36380243e-01
5.35652041e-01 -5.30714571e-01 7.84678996e-01 3.38374972e-01
-1.35328710e-01 7.45929539e-01 -8.02916139e-02 -4.41092163e-01
-6.34405792e-01 -1.38107145e+00 -6.88093781e-01 -8.40774536e-01
-3.60071898e-01 2.14197606e-01 2.45773017e-01 -2.15355083e-01
4.88469303e-02 6.26499474e-01 -1.48933256e+00 -1.67933956e-01
9.05883372e-01 8.48976910e-01 -1.21250510e-01 1.60853356e-01
-1.37315914e-01 6.18902862e-01 -5.04478633e-01 -4.96850520e-01
3.17675263e-01 -6.20315194e-01 -6.33181691e-01 -6.33098483e-01
1.37411997e-01 1.30563855e-01 -5.10399461e-01 1.31654823e+00
2.93216586e-01 4.46067452e-02 8.70053828e-01 -1.75392854e+00
-7.58585811e-01 8.99969637e-01 -4.79635626e-01 1.46519303e-01
-8.47592857e-03 -1.52429849e-01 1.49459898e+00 -5.13370931e-01
8.28947902e-01 1.32289648e+00 9.88069177e-01 2.48986900e-01
-1.06743753e+00 -8.31089139e-01 1.64295703e-01 1.55034602e-01
-1.64815676e+00 1.20605275e-01 9.23449993e-01 -1.24483302e-01
4.70192641e-01 3.97268116e-01 8.61759841e-01 1.18048418e+00
4.93173361e-01 5.82967937e-01 1.26953423e+00 -2.22133592e-01
3.66567999e-01 3.01406264e-01 4.71800447e-01 5.33273518e-01
8.19612920e-01 1.11071967e-01 -3.90005022e-01 -1.46458626e-01
6.66108549e-01 3.34463060e-01 -3.57106090e-01 -6.81656361e-01
-9.30481732e-01 9.30329561e-01 9.57687974e-01 6.63507402e-01
-2.69349337e-01 -3.12908292e-02 3.42860579e-01 5.04078209e-01
4.35822845e-01 2.61579335e-01 -2.74129715e-02 1.48898661e-01
-7.30853796e-01 -2.29073018e-01 1.00673759e+00 5.13799906e-01
9.09474194e-01 -1.44881725e-01 -5.27288839e-02 5.15481293e-01
4.58445817e-01 3.61662179e-01 4.47800428e-01 -4.73030388e-01
3.43173742e-02 1.07968009e+00 -4.41521376e-01 -1.77896881e+00
-4.73704606e-01 -9.66949761e-01 -1.32507968e+00 8.24603811e-02
4.40478504e-01 1.45659253e-01 -6.24860406e-01 2.16031361e+00
3.38687062e-01 2.42734328e-01 -5.83321005e-02 6.82120800e-01
1.06837094e+00 4.81764168e-01 -7.93948844e-02 -1.86366200e-01
9.77672100e-01 -6.81223869e-01 -6.14044309e-01 2.24375084e-01
7.55493522e-01 -1.11122184e-01 9.76838410e-01 -1.12968117e-01
-5.74047267e-01 -1.65096521e-01 -1.22012448e+00 5.65172672e-01
-6.65358305e-01 -3.79649907e-01 7.61093974e-01 1.00192976e+00
-1.19241250e+00 9.66841459e-01 -6.93979621e-01 -5.95322251e-01
2.60638356e-01 1.39286011e-01 -3.58085871e-01 1.62479773e-01
-1.57391977e+00 5.78040361e-01 3.34712356e-01 -1.80580802e-02
-8.11408877e-01 -5.63073456e-01 -4.73013520e-01 4.06765312e-01
1.44405752e-01 -5.66553414e-01 3.58458191e-01 -1.00750029e+00
-1.04417002e+00 5.69710433e-01 1.13298930e-01 -4.60861802e-01
4.78419423e-01 4.24406797e-01 -5.18611014e-01 2.27999315e-01
-1.94280185e-02 4.27709639e-01 3.74907434e-01 -1.53661978e+00
-2.38493994e-01 -4.10836548e-01 1.22390188e-01 1.32032812e-01
-7.73065448e-01 -4.62556988e-01 -1.31723940e-01 -2.80075043e-01
3.33957285e-01 -9.25210238e-01 2.78475553e-01 2.11510807e-02
-6.06187999e-01 -6.83177769e-01 1.01755154e+00 5.15240468e-02
1.38783503e+00 -1.98010552e+00 -3.25796068e-01 7.95375884e-01
8.51971567e-01 2.44334415e-01 -8.63890797e-02 6.26840651e-01
-8.47252458e-02 3.94567728e-01 -4.13370319e-02 3.16755995e-02
2.67133228e-02 3.02306384e-01 1.36078000e-02 7.56940901e-01
-1.86241180e-01 9.59837973e-01 -9.19203162e-01 -4.39271986e-01
-1.00407908e-02 5.70056736e-01 -4.78930213e-02 -2.59790182e-01
2.74441540e-01 9.04263481e-02 -3.05984169e-01 6.20371401e-01
7.40159750e-01 -7.20124424e-01 3.77716511e-01 -7.48505741e-02
3.02669525e-01 6.13465309e-02 -1.12079155e+00 9.70573008e-01
-1.61064178e-01 7.81828880e-01 -7.42045641e-02 -1.07625175e+00
1.15088379e+00 4.40671183e-02 3.17168981e-01 -5.93406498e-01
3.06139737e-01 2.63836533e-01 3.35790783e-01 -1.22605497e-02
1.98011249e-01 5.00894934e-02 3.44393373e-01 6.56105042e-01
-1.49351746e-01 5.56534767e-01 1.81676984e-01 5.92180908e-01
1.19004655e+00 -5.49496472e-01 3.97774667e-01 -7.63450444e-01
1.88478276e-01 -5.18737078e-01 4.83939737e-01 1.02851689e+00
-5.02307236e-01 3.42904389e-01 8.93685162e-01 -3.43520045e-01
-5.87482810e-01 -1.10668802e+00 -2.95656413e-01 7.22774625e-01
8.36870432e-01 -4.57956642e-01 -5.64052045e-01 -8.81654501e-01
2.82136500e-01 5.00800133e-01 -1.12406397e+00 -3.86762202e-01
-1.48229659e-01 -9.66249049e-01 7.79321909e-01 3.72367501e-01
5.76210797e-01 -8.11245799e-01 -4.94413435e-01 -4.87346172e-01
2.08918247e-02 -6.34199679e-01 2.28597105e-01 1.32943064e-01
-8.56033742e-01 -1.18158817e+00 -7.06790507e-01 -5.34328222e-01
5.97907722e-01 6.68695271e-01 1.28794301e+00 6.66227341e-01
9.42130107e-04 1.47741273e-01 -4.69695747e-01 -4.95276041e-02
-3.11468780e-01 1.68206796e-01 2.48847514e-01 5.99225797e-02
4.26932752e-01 -1.05546486e+00 -9.89968717e-01 7.19854355e-01
-7.45919228e-01 -2.30865777e-01 6.63519263e-01 6.73696280e-01
1.90857276e-01 3.28915864e-01 5.70139408e-01 -1.00983989e+00
7.63765275e-01 -8.11408341e-01 -6.34163097e-02 4.32761133e-01
-1.35188544e+00 -2.29546085e-01 5.94287753e-01 -2.04956964e-01
-6.82273269e-01 -6.51225388e-01 2.78647393e-01 -2.97602892e-01
2.27949843e-02 6.81123912e-01 -1.15770876e-01 -2.98605025e-01
6.64007783e-01 -1.79689506e-03 7.43302181e-02 -6.43703490e-02
1.87131479e-01 6.05685949e-01 9.85232182e-03 -3.64167780e-01
6.89388156e-01 8.20973694e-01 3.10636848e-01 -7.28726506e-01
-3.36310536e-01 -3.25921804e-01 -4.12445009e-01 -3.19461554e-01
4.72243041e-01 -7.45448470e-01 -9.27063406e-01 3.49135756e-01
-7.79509783e-01 -1.22546420e-01 -1.83352187e-01 4.58879083e-01
1.22575343e-01 4.37463313e-01 -5.36038399e-01 -8.89614582e-01
-2.80141383e-01 -7.64576554e-01 4.42981809e-01 3.24675858e-01
-1.81145802e-01 -1.51234043e+00 -8.66547674e-02 -1.20684542e-01
5.37652016e-01 3.38151067e-01 1.10064065e+00 -1.07052612e+00
-5.00828981e-01 -3.41421574e-01 -5.26741505e-01 3.03875357e-02
8.95481706e-02 3.05229038e-01 -7.17711508e-01 -4.47027653e-01
-2.57408381e-01 1.72688160e-02 9.27235544e-01 4.66731966e-01
7.40751445e-01 -1.25904784e-01 -7.39618361e-01 5.08678257e-01
1.51450288e+00 1.08802933e-02 6.97139323e-01 3.23097378e-01
7.58566439e-01 7.86180794e-01 1.25995606e-01 1.71617866e-01
4.70355511e-01 3.43493581e-01 7.75834084e-01 -2.34189451e-01
-8.50847512e-02 -5.38517654e-01 8.39844868e-02 8.24147165e-01
-6.76999008e-03 -7.82022536e-01 -1.08846462e+00 4.68793720e-01
-1.84047401e+00 -8.56391430e-01 -3.88929605e-01 2.14332700e+00
3.88428122e-01 8.99721310e-02 2.49605373e-01 1.77423790e-01
1.10865331e+00 2.26407468e-01 -5.52819669e-01 1.28169982e-02
-4.84127343e-01 -3.92291903e-01 4.31392282e-01 2.70624369e-01
-6.42517328e-01 6.86879873e-01 5.84688139e+00 1.18261981e+00
-1.11345208e+00 4.78748651e-03 9.69706118e-01 2.43311226e-01
-6.95274234e-01 8.82609338e-02 -6.78394496e-01 5.66892564e-01
7.59570181e-01 -1.15932733e-01 1.34608010e-02 8.13324869e-01
-8.86774734e-02 -4.83249575e-02 -1.09755373e+00 7.65312433e-01
6.91870898e-02 -1.18214929e+00 1.44556120e-01 2.05154344e-01
7.95554399e-01 1.51271939e-01 -1.90759171e-02 5.67876697e-02
5.02503395e-01 -1.11999273e+00 3.45053315e-01 3.66947912e-02
4.04299080e-01 -6.33548617e-01 8.72353375e-01 3.90334189e-01
-1.29127610e+00 -4.43318747e-02 -3.67031991e-01 -1.60695508e-01
-1.15968175e-01 8.67797077e-01 -4.72739697e-01 7.52043784e-01
9.17493761e-01 9.19838846e-01 -6.34328187e-01 9.83986378e-01
-2.14388192e-01 5.89572251e-01 -4.09595788e-01 -5.93450189e-01
3.12887818e-01 -3.89528751e-01 6.56686902e-01 9.80126500e-01
3.42848301e-01 -3.32876563e-01 -2.06731632e-01 1.09870362e+00
-1.60540447e-01 1.83164373e-01 -7.31499255e-01 -2.21549277e-03
9.00712252e-01 1.45816481e+00 -1.46635246e+00 -2.66057521e-01
-1.74877778e-01 5.79398453e-01 3.04450512e-01 4.40719694e-01
-8.03860843e-01 -6.21995091e-01 3.96248937e-01 1.26313895e-01
-8.21080729e-02 7.56823942e-02 -5.67511141e-01 -9.02109563e-01
-3.20137776e-02 -5.69573760e-01 4.26083207e-01 -3.75615567e-01
-1.66391635e+00 6.51219606e-01 -6.29300699e-02 -1.18126523e+00
3.08728158e-01 -4.71152097e-01 -9.36358035e-01 5.13010681e-01
-1.45657253e+00 -1.01103365e+00 -5.86111784e-01 6.17172658e-01
-2.48935580e-01 -9.90158841e-02 4.54272330e-01 2.56033123e-01
-6.46408260e-01 8.61832857e-01 2.81679779e-01 3.39085162e-01
4.10289615e-01 -1.18616033e+00 1.89172238e-01 7.11582422e-01
3.21230233e-01 1.09652257e+00 6.70445144e-01 -8.90938580e-01
-1.11735964e+00 -8.06164265e-01 8.44546795e-01 -2.39654198e-01
8.53579879e-01 -3.09565485e-01 -8.80595267e-01 2.89477825e-01
-1.34816356e-02 7.35923722e-02 7.06407607e-01 4.01309758e-01
-6.24577463e-01 -2.49773607e-01 -1.01363182e+00 6.72101915e-01
1.43208873e+00 -5.19504428e-01 -1.65700108e-01 3.90522629e-01
6.94660068e-01 3.08716446e-01 -9.36960936e-01 7.63930023e-01
6.71769142e-01 -1.55878043e+00 8.64174664e-01 -3.88435662e-01
4.01863933e-01 -2.07415193e-01 -1.53123274e-01 -1.29740083e+00
-2.90090770e-01 -3.74856621e-01 -5.45036048e-02 1.19589603e+00
5.16023517e-01 -1.24956334e+00 1.02457643e+00 2.99020380e-01
1.68304443e-01 -9.86858070e-01 -9.94006693e-01 -1.26090062e+00
2.59775102e-01 -1.04005262e-02 7.41147339e-01 1.36037076e+00
7.97113255e-02 3.50233763e-01 -1.95043996e-01 -2.76889428e-02
4.64222550e-01 3.30297410e-01 6.23757958e-01 -1.83811796e+00
-7.69346952e-02 -7.58103967e-01 -6.76495075e-01 -8.44294250e-01
3.38822246e-01 -1.10824406e+00 -2.90935606e-01 -1.48415959e+00
4.94367748e-01 -8.07473421e-01 -7.00805664e-01 3.13622683e-01
-5.44688441e-02 2.87687689e-01 2.51190364e-02 7.71739364e-01
-5.81021726e-01 6.54798269e-01 1.22882879e+00 -7.76495934e-02
-2.00665936e-01 -4.94159050e-02 -6.62679255e-01 6.54743791e-01
7.28124857e-01 -4.61230338e-01 -6.80277169e-01 6.37865886e-02
6.51681900e-01 -3.04291844e-01 4.89057630e-01 -1.04551721e+00
2.90089786e-01 1.33626476e-01 2.04091042e-01 -2.79823691e-01
1.08575076e-01 -8.72669876e-01 2.15818852e-01 7.48400450e-01
-3.64777356e-01 2.01681226e-01 -4.76697981e-01 1.07578361e+00
-2.25661352e-01 5.33988550e-02 4.86848146e-01 4.47176956e-02
-4.25487757e-01 3.69063109e-01 7.28071248e-03 -6.20774925e-02
1.15497184e+00 -5.60379684e-01 -8.81410658e-01 -6.95532799e-01
-2.08912849e-01 1.56136408e-01 5.47417641e-01 3.08502495e-01
3.47510517e-01 -1.37570310e+00 -5.16495168e-01 3.66038978e-02
1.81420296e-01 -6.45434082e-01 1.58912078e-01 1.27133441e+00
-4.56500232e-01 2.73991227e-01 -2.99776137e-01 -9.12895858e-01
-1.25750911e+00 3.68934304e-01 3.34145427e-01 -1.40627369e-01
-5.38457155e-01 9.82580841e-01 3.96664053e-01 -5.60297966e-01
3.44117820e-01 1.03361718e-01 -3.28717947e-01 2.02488780e-01
-4.33377288e-02 6.02489054e-01 -7.73003930e-03 -6.38866067e-01
-4.96675491e-01 4.69028771e-01 5.54677509e-02 2.00401306e-01
1.03187835e+00 -1.13498248e-01 -4.38598394e-01 5.83763123e-01
1.18719542e+00 -1.97431892e-01 -8.18972468e-01 -1.03343979e-01
-7.95342699e-02 -3.86802495e-01 -2.09707215e-01 -5.59001088e-01
-1.13504338e+00 9.98142242e-01 5.41112840e-01 9.05079186e-01
7.67866254e-01 2.68605426e-02 7.10948944e-01 4.14476395e-01
3.49366367e-01 -1.01342952e+00 1.53426453e-01 4.06491458e-01
3.78779441e-01 -1.02741504e+00 2.25396901e-02 -6.60527527e-01
-3.43712360e-01 9.07429338e-01 6.36234820e-01 -1.38227955e-01
9.93047357e-01 -2.05935508e-01 -4.88355421e-02 -4.59562898e-01
-4.97032762e-01 7.54921287e-02 1.28105268e-01 3.28352302e-01
3.56917948e-01 2.83055097e-01 -4.99470830e-01 2.10493281e-01
-2.24887878e-01 -5.26686072e-01 2.57288516e-01 5.71356833e-01
-3.57220143e-01 -7.62754083e-01 8.55219290e-02 4.86913472e-01
-1.34574667e-01 -1.54150769e-01 -7.96207309e-01 1.05903530e+00
3.21996748e-03 8.70334744e-01 1.17902964e-01 -8.75124216e-01
-3.36070329e-01 -2.07578659e-01 -2.69079171e-02 -1.12472586e-01
-5.77589452e-01 -3.46446306e-01 -1.14861891e-01 -4.16842341e-01
-5.50891638e-01 -1.30973175e-01 -1.11822271e+00 -6.98924720e-01
-5.99001646e-01 4.58011001e-01 5.45133173e-01 7.96341717e-01
5.51992655e-01 2.54436702e-01 7.13900328e-01 -3.23913962e-01
-5.08034170e-01 -7.79511690e-01 -8.28193188e-01 6.56615913e-01
-3.92495096e-02 -9.12550688e-01 -1.05033243e+00 -9.08376276e-01] | [7.0693559646606445, 6.0787577629089355] |
cb362a8f-84d4-4b5c-8f82-250159111bad | semantic-image-manipulation-using-scene | 2004.03677 | null | https://arxiv.org/abs/2004.03677v1 | https://arxiv.org/pdf/2004.03677v1.pdf | Semantic Image Manipulation Using Scene Graphs | Image manipulation can be considered a special case of image generation where the image to be produced is a modification of an existing image. Image generation and manipulation have been, for the most part, tasks that operate on raw pixels. However, the remarkable progress in learning rich image and object representations has opened the way for tasks such as text-to-image or layout-to-image generation that are mainly driven by semantics. In our work, we address the novel problem of image manipulation from scene graphs, in which a user can edit images by merely applying changes in the nodes or edges of a semantic graph that is generated from the image. Our goal is to encode image information in a given constellation and from there on generate new constellations, such as replacing objects or even changing relationships between objects, while respecting the semantics and style from the original image. We introduce a spatio-semantic scene graph network that does not require direct supervision for constellation changes or image edits. This makes it possible to train the system from existing real-world datasets with no additional annotation effort. | ['Gregory D. Hager', 'Azade Farshad', 'Federico Tombari', 'Christian Rupprecht', 'Nassir Navab', 'Helisa Dhamo', 'Iro Laina'] | 2020-04-07 | semantic-image-manipulation-using-scene-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Dhamo_Semantic_Image_Manipulation_Using_Scene_Graphs_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Dhamo_Semantic_Image_Manipulation_Using_Scene_Graphs_CVPR_2020_paper.pdf | cvpr-2020-6 | ['layout-to-image-generation'] | ['computer-vision'] | [ 1.03305149e+00 4.90684479e-01 3.73504996e-01 -6.80195749e-01
-1.22167654e-02 -5.88435531e-01 9.02779937e-01 4.59212482e-01
-3.37174326e-01 5.53558469e-01 -4.57938202e-02 -1.27742916e-01
7.44725913e-02 -1.12130237e+00 -1.19519901e+00 -4.77853507e-01
2.51867980e-01 4.58854258e-01 3.63117188e-01 -4.30910915e-01
2.62324929e-01 7.80145526e-01 -1.92427671e+00 2.21311390e-01
7.57955372e-01 5.92885733e-01 6.35822535e-01 7.72468507e-01
-3.73482347e-01 4.97652918e-01 -7.51687884e-01 -4.73443121e-01
1.93990678e-01 -5.18314600e-01 -7.83773005e-01 6.84425235e-01
7.44093120e-01 -9.21801552e-02 -3.21104527e-01 1.09784532e+00
2.28795618e-01 1.54846832e-01 5.64799130e-01 -1.42846620e+00
-9.69467640e-01 6.57761574e-01 -1.77502677e-01 -3.60249341e-01
3.19248885e-01 2.51025945e-01 6.17081285e-01 -5.49338043e-01
1.10774457e+00 1.36367881e+00 -2.13718675e-02 5.13207912e-01
-1.38304222e+00 -5.39854281e-02 2.47217909e-01 1.05810128e-01
-1.15558994e+00 -3.62327993e-01 7.20494807e-01 -4.02629077e-01
7.17187107e-01 3.10644984e-01 6.34147763e-01 7.37441599e-01
-1.52131304e-01 5.69399953e-01 6.81461811e-01 -6.71089768e-01
3.57598096e-01 1.63381413e-01 -3.30636144e-01 7.20908523e-01
1.34026706e-01 -2.55355775e-01 -3.89203429e-01 2.52093881e-01
8.84521186e-01 3.90574755e-03 -2.10537985e-01 -7.24571526e-01
-1.46236908e+00 6.26848400e-01 8.52763593e-01 2.20308870e-01
-2.93264091e-01 4.58402008e-01 2.01314628e-01 3.83144408e-01
1.36474758e-01 7.80049980e-01 -2.75622994e-01 2.40856051e-01
-6.94488645e-01 4.23502028e-01 5.40360153e-01 1.31020391e+00
9.37752128e-01 4.43801805e-02 -2.33661845e-01 4.97655123e-01
-2.34861031e-01 5.60006857e-01 1.88035250e-01 -7.91260839e-01
4.67314810e-01 8.17503512e-01 1.15798093e-01 -1.05759525e+00
-1.17460132e-01 -1.22636460e-01 -9.28056538e-01 3.39750051e-01
1.32326290e-01 2.62860842e-02 -1.23029578e+00 1.66482031e+00
2.96607792e-01 -1.27446026e-01 -1.73631199e-02 5.95536411e-01
7.67412841e-01 7.39866376e-01 -4.64170948e-02 5.39715365e-02
1.09428060e+00 -8.57666135e-01 -6.38008833e-01 -3.04066449e-01
5.67154765e-01 -5.94767392e-01 1.21485460e+00 1.55460462e-01
-1.16189301e+00 -7.38652289e-01 -9.08835173e-01 -3.25782984e-01
-8.31831455e-01 -1.12149514e-01 5.17511845e-01 2.82723159e-01
-1.12330401e+00 6.04185104e-01 -3.40155333e-01 -5.22957623e-01
3.41547042e-01 2.49795571e-01 -5.95995247e-01 3.87550816e-02
-9.91253734e-01 7.51553476e-01 9.67943728e-01 -5.87522890e-03
-7.31138349e-01 -6.70004725e-01 -1.16391420e+00 4.39468063e-02
6.01820230e-01 -9.13842440e-01 1.14314473e+00 -1.46527660e+00
-1.24849403e+00 1.03226542e+00 6.19819053e-02 -4.21698838e-01
4.44597632e-01 1.89313546e-01 -1.98385820e-01 -1.29961744e-02
2.23914653e-01 1.31983221e+00 1.09840703e+00 -1.43456161e+00
-6.87264442e-01 -2.55236715e-01 4.75576967e-01 2.36205980e-01
4.48624901e-02 -1.28679588e-01 -7.04458773e-01 -7.91706860e-01
-1.50542171e-03 -1.01896238e+00 -3.06682438e-01 4.56800193e-01
-6.59050107e-01 4.06593680e-02 1.06398106e+00 -4.88790721e-01
8.30140591e-01 -2.17508984e+00 6.02571011e-01 2.96130270e-01
-1.22464634e-01 2.81962276e-01 -3.98038626e-01 5.91833770e-01
-1.82209268e-01 2.25379676e-01 -6.08052313e-01 -2.95776516e-01
-8.72285962e-02 3.10007215e-01 -3.25120449e-01 -1.41893193e-01
4.16998118e-01 9.63026226e-01 -1.12525260e+00 -3.55059087e-01
4.08516347e-01 3.13848406e-01 -6.76795363e-01 3.52677137e-01
-8.61240447e-01 6.82863891e-01 -1.69335574e-01 6.74295351e-02
4.91035819e-01 -1.58561140e-01 9.63330641e-02 -3.13539535e-01
1.38980849e-02 -1.23963371e-01 -1.18976176e+00 2.12260127e+00
-5.27749002e-01 6.49822116e-01 -3.76296997e-01 -9.38086867e-01
7.99379170e-01 1.53303370e-02 1.49224266e-01 -7.16659904e-01
-2.08557308e-01 -5.35882860e-02 -1.40372589e-01 -5.89549065e-01
7.93639421e-01 1.74200341e-01 -3.01862866e-01 4.96923059e-01
-1.31814688e-01 -7.98793018e-01 5.98890126e-01 4.97903615e-01
7.40965605e-01 2.41150394e-01 2.68771052e-01 1.39645353e-01
3.40086073e-01 1.82921261e-01 2.86189560e-02 8.31383109e-01
6.65694237e-01 6.87540054e-01 5.87738752e-01 -4.27516699e-01
-1.34516370e+00 -1.19148624e+00 1.48617864e-01 8.61565232e-01
1.85769200e-01 -4.23570007e-01 -8.04228008e-01 -5.72415531e-01
-1.12411782e-01 8.62173676e-01 -5.06804407e-01 -2.68791020e-01
-6.69898212e-01 -2.86258817e-01 2.18659520e-01 2.63561934e-01
6.13451958e-01 -1.40990603e+00 -7.99776495e-01 1.15960397e-01
-6.23995923e-02 -1.18255353e+00 -5.70726156e-01 -1.19772136e-01
-6.83310390e-01 -1.00170159e+00 -5.84825099e-01 -1.00694597e+00
1.30331588e+00 1.72144353e-01 1.07676935e+00 2.37148285e-01
-6.19037688e-01 3.30882132e-01 -5.18939376e-01 -5.40328145e-01
-7.13495255e-01 1.40878364e-01 -2.58193731e-01 2.59129047e-01
-4.72796917e-01 -5.00840664e-01 -2.83853471e-01 -6.06786124e-02
-1.72330415e+00 8.56575072e-01 4.80239540e-01 5.63998401e-01
6.85644329e-01 3.43627512e-01 4.18249041e-01 -1.20515800e+00
2.69965053e-01 -4.95771058e-02 -6.23024762e-01 3.70659381e-01
-4.84910868e-02 5.30375957e-01 6.44677162e-01 -2.52564728e-01
-1.23646104e+00 4.20343190e-01 1.73677027e-01 -1.89321786e-01
-3.08318198e-01 4.47001874e-01 -4.23620850e-01 1.62924021e-01
6.46041453e-01 2.29658589e-01 -1.04112700e-01 -3.27804506e-01
1.05917788e+00 5.19157410e-01 8.22569847e-01 -3.04698765e-01
9.24404979e-01 6.00344300e-01 9.91940573e-02 -9.70569015e-01
-6.32820487e-01 -1.09046707e-02 -1.14639568e+00 1.12574128e-03
1.00067377e+00 -4.57135409e-01 -2.02990934e-01 4.98703569e-01
-1.27672994e+00 -5.29618621e-01 -6.35542095e-01 -3.09522539e-01
-7.65766323e-01 2.38461256e-01 4.73945849e-02 -2.90533364e-01
-2.88516339e-02 -1.07242095e+00 1.31108189e+00 2.24631801e-01
-1.53426573e-01 -9.76169169e-01 -4.56602067e-01 -2.57973373e-02
2.51766205e-01 6.41651750e-01 1.33116615e+00 4.83054034e-02
-9.55954432e-01 -2.93080080e-02 -2.29887813e-01 2.39319026e-01
5.94680488e-01 1.36760131e-01 -4.89963204e-01 -1.94386318e-01
-5.47484994e-01 -5.20736016e-02 5.85351229e-01 -3.39913331e-02
1.55268645e+00 -3.77189755e-01 -3.40058446e-01 5.89814782e-01
1.44981384e+00 1.74181238e-01 8.25138390e-01 2.03627571e-01
9.74467039e-01 7.41713345e-01 3.72656345e-01 3.63109231e-01
1.80250362e-01 9.69617069e-01 6.67349100e-01 -2.86557734e-01
-3.80471855e-01 -6.05247438e-01 8.85390770e-03 2.37557843e-01
3.02949160e-01 -5.37159562e-01 -6.74230158e-01 6.55582726e-01
-1.90113449e+00 -8.11201155e-01 -2.01262519e-01 2.26987600e+00
8.66234899e-01 -3.40442844e-02 -2.29903564e-01 -3.92983779e-02
8.10751498e-01 1.50415614e-01 -4.56241071e-01 -5.01868010e-01
-1.34951815e-01 2.31636152e-01 4.46348250e-01 4.42506373e-01
-9.87030447e-01 1.18602180e+00 5.60704994e+00 4.80316281e-01
-1.12425399e+00 -3.32035393e-01 4.37912375e-01 -2.98661776e-02
-4.38000679e-01 1.71193406e-01 -3.85748923e-01 3.67776424e-01
4.26902890e-01 -1.54658452e-01 6.11204624e-01 4.72819209e-01
4.21675324e-01 -3.27134758e-01 -1.43810725e+00 1.04228544e+00
3.53384703e-01 -1.48468208e+00 8.48237932e-01 -1.56931892e-01
9.23467040e-01 -5.07642150e-01 -6.55424595e-02 -9.47186500e-02
3.06884766e-01 -9.78650451e-01 9.82501328e-01 8.73427927e-01
1.01281130e+00 -5.48226535e-01 2.51631945e-01 2.70782590e-01
-9.87519324e-01 2.12475374e-01 -2.87247896e-01 1.23539016e-01
4.66078639e-01 4.78475630e-01 -8.37972224e-01 5.63606203e-01
3.34894031e-01 7.56471634e-01 -9.02832627e-01 9.26117599e-01
-4.11755502e-01 -4.79705669e-02 -1.31014496e-01 1.63011998e-01
2.21364066e-01 -1.90612763e-01 4.23785776e-01 9.68348980e-01
4.10919040e-01 -7.42589608e-02 1.47886932e-01 1.10057712e+00
-3.09154898e-01 2.73914449e-02 -1.06168985e+00 -2.37458110e-01
2.53289789e-01 1.17450607e+00 -8.55059922e-01 -6.74658954e-01
-9.57328975e-02 1.48780811e+00 1.60926819e-01 2.62744486e-01
-8.42671335e-01 -6.94756746e-01 4.06131238e-01 3.46737772e-01
3.86053950e-01 -2.27848873e-01 1.61478166e-02 -1.03730214e+00
1.80172712e-01 -7.28760064e-01 4.28593643e-02 -1.34800053e+00
-8.05050850e-01 4.49957818e-01 2.64993548e-01 -9.05559719e-01
-1.83521986e-01 -4.70953554e-01 -5.47435522e-01 5.42757928e-01
-1.19817603e+00 -1.29281175e+00 -6.37467742e-01 5.76106906e-01
6.76889420e-01 3.84802781e-02 6.89519167e-01 1.10405773e-01
-1.36193097e-01 1.74364641e-01 -7.40581602e-02 1.61239237e-01
4.13742900e-01 -1.25279880e+00 7.34249592e-01 9.01949465e-01
4.03423131e-01 3.29955280e-01 6.97134972e-01 -6.48273647e-01
-1.11120391e+00 -1.57337320e+00 8.85573328e-01 -3.38018179e-01
3.53463829e-01 -6.48998678e-01 -6.95189297e-01 7.86724687e-01
2.77662009e-01 -1.88494906e-01 -8.20695236e-02 -5.49099267e-01
-2.26669207e-01 -7.04958811e-02 -9.14306879e-01 1.01698172e+00
1.44245303e+00 -4.16709960e-01 -3.90307039e-01 6.48939073e-01
8.18162739e-01 -5.65788388e-01 -5.02183437e-01 2.08910286e-01
4.66714129e-02 -6.55931830e-01 9.99193430e-01 -7.76403487e-01
4.24363613e-01 -7.39309132e-01 6.40768856e-02 -1.65874600e+00
-1.19301915e-01 -5.87023497e-01 4.58708137e-01 1.16121757e+00
3.09973925e-01 -2.60430723e-01 4.60420370e-01 5.92371881e-01
-1.61969587e-01 -2.86847204e-02 -5.00820398e-01 -6.83360457e-01
-3.79153103e-01 -2.01409921e-01 1.03230631e+00 6.99002266e-01
-3.69143277e-01 2.76887774e-01 -2.86196738e-01 1.21688113e-01
4.72226650e-01 2.21321002e-01 1.15728426e+00 -1.03785396e+00
-2.56030858e-01 -3.20793867e-01 -6.35294199e-01 -8.58346045e-01
1.40706599e-01 -1.12629795e+00 1.56146303e-01 -1.86024737e+00
1.11057572e-02 -5.77130795e-01 3.92291844e-01 6.16564095e-01
-1.23377562e-01 2.21236467e-01 5.50300658e-01 4.36130390e-02
-5.24273574e-01 5.67755461e-01 1.47696126e+00 -4.34131831e-01
-1.50380746e-01 -2.19965920e-01 -6.85081422e-01 6.14890754e-01
7.98767090e-01 -4.43228334e-01 -7.41603732e-01 -7.88282752e-01
5.15816391e-01 -3.97337191e-02 5.84676504e-01 -9.50765014e-01
4.31859903e-02 -3.02015066e-01 1.84307948e-01 -3.05027008e-01
2.59203464e-01 -8.84804487e-01 6.17160380e-01 2.99717307e-01
-5.51794887e-01 1.54951930e-01 5.19022495e-02 5.16918659e-01
-1.47814423e-01 -4.03592259e-01 7.63547540e-01 -4.43455428e-01
-1.07094681e+00 2.64946401e-01 -1.53491333e-01 -1.73773468e-01
1.26662552e+00 -3.45615417e-01 -2.65886456e-01 -4.84154165e-01
-8.67502630e-01 9.73082334e-03 7.83871531e-01 7.69496739e-01
5.46890020e-01 -1.35666132e+00 -6.19067252e-01 2.92301685e-01
5.25310576e-01 4.54852730e-01 1.84813648e-01 2.48036683e-01
-8.90870094e-01 2.46171534e-01 -2.93676436e-01 -5.59101582e-01
-1.33690798e+00 7.45085537e-01 3.20887476e-01 8.99552852e-02
-6.91042483e-01 5.37597716e-01 5.84614813e-01 -4.37091976e-01
8.09578318e-03 -6.01788700e-01 6.17966466e-02 -1.72620803e-01
4.11196262e-01 -2.23968960e-02 6.50451481e-02 -6.77075803e-01
9.49917212e-02 4.34218496e-01 6.51515797e-02 -1.24649972e-01
1.24919462e+00 -1.87053494e-02 -4.06718463e-01 3.94203901e-01
9.69297647e-01 -2.84441799e-01 -1.18353963e+00 -2.02983186e-01
-4.03051972e-02 -6.47220075e-01 -2.77796954e-01 -6.87027991e-01
-1.06209898e+00 7.52375662e-01 2.93776602e-01 1.92283064e-01
1.06456947e+00 1.55301705e-01 5.57958961e-01 5.65226138e-01
4.89360601e-01 -1.12315810e+00 4.43280786e-01 3.51999581e-01
1.30640316e+00 -1.03476107e+00 -1.73946485e-01 -5.22160351e-01
-6.64199412e-01 1.06371224e+00 5.93856871e-01 -3.34574245e-02
4.46791768e-01 -1.36215314e-01 -2.79242933e-01 -2.58536130e-01
-4.29848641e-01 -3.99577141e-01 2.31517598e-01 8.98916066e-01
4.21956219e-02 4.98619787e-02 5.04964180e-02 -3.96128684e-01
-3.09654385e-01 5.45233078e-02 7.24309921e-01 9.27006364e-01
-4.06001449e-01 -1.34272325e+00 -1.33458555e-01 4.72618401e-01
1.34498328e-01 -1.30302012e-01 -5.85688293e-01 7.12135613e-01
4.66503173e-01 6.52739346e-01 4.09141243e-01 1.47655457e-01
5.93439639e-01 -1.40712619e-01 6.34429216e-01 -1.10130119e+00
-2.39571765e-01 -4.77028251e-01 -5.48000447e-03 -3.56164783e-01
-5.03724098e-01 -6.13153875e-01 -1.37721717e+00 -4.72430959e-02
-3.34199965e-02 -1.73138827e-01 9.08155560e-01 7.40052462e-01
5.35011530e-01 7.39221632e-01 3.98144662e-01 -1.00549948e+00
7.42937997e-02 -5.99869490e-01 -6.65102720e-01 1.00561857e+00
1.99585319e-01 -3.45326006e-01 1.21030562e-01 8.10524762e-01] | [11.218180656433105, -0.24024303257465363] |
3c7e9b69-08bc-4cbb-8eaa-c68b2ee0baf7 | compressed-sensing-a-discrete-optimization | 2306.04647 | null | https://arxiv.org/abs/2306.04647v1 | https://arxiv.org/pdf/2306.04647v1.pdf | Compressed Sensing: A Discrete Optimization Approach | We study the Compressed Sensing (CS) problem, which is the problem of finding the most sparse vector that satisfies a set of linear measurements up to some numerical tolerance. CS is a central problem in Statistics, Operations Research and Machine Learning which arises in applications such as signal processing, data compression and image reconstruction. We introduce an $\ell_2$ regularized formulation of CS which we reformulate as a mixed integer second order cone program. We derive a second order cone relaxation of this problem and show that under mild conditions on the regularization parameter, the resulting relaxation is equivalent to the well studied basis pursuit denoising problem. We present a semidefinite relaxation that strengthens the second order cone relaxation and develop a custom branch-and-bound algorithm that leverages our second order cone relaxation to solve instances of CS to certifiable optimality. Our numerical results show that our approach produces solutions that are on average $6.22\%$ more sparse than solutions returned by state of the art benchmark methods on synthetic data in minutes. On real world ECG data, for a given $\ell_2$ reconstruction error our approach produces solutions that are on average $9.95\%$ more sparse than benchmark methods, while for a given sparsity level our approach produces solutions that have on average $10.77\%$ lower reconstruction error than benchmark methods in minutes. | ['Nicholas Johnson', 'Dimitris Bertsimas'] | 2023-06-05 | null | null | null | null | ['image-reconstruction', 'data-compression'] | ['computer-vision', 'time-series'] | [ 6.13124728e-01 1.88958973e-01 -1.24699540e-01 -1.69983581e-01
-1.34678292e+00 -5.47757864e-01 -4.60200369e-01 4.19986725e-01
-5.17873645e-01 6.55185938e-01 1.03120379e-01 -3.17353785e-01
-4.21969950e-01 -5.69849133e-01 -8.43679190e-01 -6.33152664e-01
-5.40507138e-01 2.57849723e-01 -3.10701847e-01 -1.08032867e-01
2.14585438e-01 9.42688063e-02 -6.96362019e-01 -1.00354627e-01
9.44561899e-01 1.48299801e+00 -1.27059996e-01 4.80379522e-01
4.05240148e-01 4.50134188e-01 -4.92986217e-02 -1.63719088e-01
7.01099813e-01 -5.43023825e-01 -6.53036773e-01 5.46531081e-01
3.23863536e-01 9.29974765e-02 -3.18630971e-02 1.42546380e+00
5.48205554e-01 2.22017094e-01 1.83123752e-01 -8.99693429e-01
-8.96425694e-02 5.41771710e-01 -9.66599345e-01 3.96572143e-01
4.45309877e-01 -2.71417558e-01 9.47968960e-01 -8.51893365e-01
6.09658062e-01 7.91753292e-01 9.00145173e-01 1.86339766e-01
-1.58926022e+00 -4.77204591e-01 -8.22308939e-03 -2.99003124e-01
-1.44780338e+00 -5.46182513e-01 6.53896987e-01 -3.28786582e-01
4.53432441e-01 5.92844188e-01 5.57735920e-01 1.64971486e-01
1.79232042e-02 6.62371159e-01 1.30259275e+00 -2.87818223e-01
5.15113175e-01 -1.89169437e-01 2.61717349e-01 7.49466777e-01
5.13558805e-01 -9.85737070e-02 -4.79655832e-01 -4.71755475e-01
7.03532934e-01 -4.88858223e-02 -5.83601177e-01 -2.46351317e-01
-1.26950037e+00 1.13683009e+00 1.80137724e-01 -2.69406270e-02
-6.07156456e-01 2.10201249e-01 4.25523520e-01 2.80319989e-01
4.17631865e-01 4.74790990e-01 -2.11095303e-01 -7.99030289e-02
-1.24099898e+00 2.99959570e-01 1.02070940e+00 7.51392186e-01
5.26152849e-01 2.93016404e-01 -1.61166042e-01 8.11323643e-01
5.06452918e-02 8.77355099e-01 -1.37009472e-01 -1.46475720e+00
5.64756632e-01 1.61124438e-01 5.73871024e-02 -1.58374977e+00
-3.11915964e-01 -8.76410306e-01 -1.14966941e+00 -2.89909571e-01
7.69717321e-02 -1.59078151e-01 -4.49840814e-01 1.67946684e+00
1.57986760e-01 4.10479784e-01 -1.06634356e-01 1.08490789e+00
4.31718439e-01 6.26464307e-01 -5.32673895e-01 -1.07699502e+00
1.20027506e+00 -5.09755611e-01 -7.98781514e-01 -3.88295919e-01
5.97004533e-01 -7.79074430e-01 5.15060842e-01 7.36732483e-01
-1.60739505e+00 3.46275754e-02 -1.07608557e+00 3.30932617e-01
6.61880672e-01 -1.01906434e-01 6.88495576e-01 6.66199505e-01
-8.96372497e-01 5.55921972e-01 -9.02578592e-01 1.43345088e-01
3.99034262e-01 4.34095085e-01 -3.32297087e-01 -5.42407393e-01
-5.24186015e-01 3.31194639e-01 -1.55066296e-01 1.30309239e-01
-7.11579382e-01 -9.97645795e-01 -1.11926377e+00 -1.00598320e-01
6.58358574e-01 -3.88104439e-01 9.45704222e-01 -5.67002416e-01
-7.93296456e-01 9.34053540e-01 -4.07529056e-01 -7.11044908e-01
4.26079959e-01 6.01304062e-02 -1.71374545e-01 3.04196209e-01
4.79065716e-01 -2.52195224e-02 7.07345665e-01 -1.13755476e+00
-2.97461122e-01 -4.96920049e-01 -1.10173665e-01 1.22828903e-02
-1.27014160e-01 -2.12731823e-01 -6.69019103e-01 -9.43597853e-01
7.89378524e-01 -1.24171388e+00 -9.61128950e-01 3.07106227e-02
-3.75880688e-01 4.42627132e-01 1.56374350e-01 -8.52311611e-01
1.46183741e+00 -2.12029719e+00 3.86214405e-01 7.82981157e-01
5.39499164e-01 -2.32332200e-01 -5.00187837e-02 2.66969204e-01
-3.79953496e-02 -1.86160594e-01 -8.21583569e-01 -5.12790859e-01
-2.40531221e-01 3.08732808e-01 -1.54106840e-01 1.01343739e+00
-4.73399550e-01 4.14113492e-01 -8.29090059e-01 -2.88566381e-01
-2.30416834e-01 1.96461678e-01 -1.15949416e+00 -1.27879396e-01
9.87796560e-02 3.86736006e-01 -3.20240349e-01 6.80476546e-01
9.74736333e-01 -4.81176078e-01 4.49422121e-01 -2.39885256e-01
2.11005762e-01 -2.13388428e-01 -1.79177022e+00 2.02724504e+00
-3.78372520e-01 4.83446777e-01 7.38109350e-01 -1.59241390e+00
7.52388716e-01 1.70144320e-01 1.20184290e+00 -5.08314967e-01
1.54558361e-01 4.52224404e-01 -2.07165226e-01 -3.12356651e-01
1.13242649e-01 -4.55193579e-01 -1.53595567e-01 4.88822252e-01
-5.54413021e-01 -2.46474221e-01 3.24108720e-01 4.22076374e-01
1.63392377e+00 -6.75990820e-01 1.57511324e-01 -6.25935078e-01
3.37821722e-01 7.40478700e-03 1.03895283e+00 7.78931141e-01
5.99540696e-02 6.83329284e-01 5.97341537e-01 -3.77929598e-01
-7.21411407e-01 -8.55596602e-01 -2.90491462e-01 5.30855834e-01
-5.74192330e-02 -5.72463572e-01 -5.60288906e-01 -1.34825662e-01
4.79827775e-03 1.35190055e-01 -4.59392905e-01 1.84204623e-01
-4.95213747e-01 -9.85518336e-01 2.02299058e-01 4.73043889e-01
4.41685796e-01 -4.00116265e-01 -6.66740656e-01 4.09956843e-01
-4.28363711e-01 -1.43811202e+00 -7.35866129e-01 1.50773376e-01
-1.25799227e+00 -1.07555079e+00 -8.50868464e-01 -6.33489192e-01
9.15086389e-01 3.03705722e-01 1.24953246e+00 2.33638689e-01
-5.00724494e-01 4.87419605e-01 -3.05380911e-01 -1.93833798e-01
-1.84906293e-02 -3.48777890e-01 2.78360605e-01 1.94189131e-01
-3.36998343e-01 -7.63133526e-01 -7.42129445e-01 -2.80924495e-02
-1.16510141e+00 -3.49492460e-01 3.38754147e-01 8.62785876e-01
1.24053502e+00 2.81959265e-01 2.78017342e-01 -1.17363822e+00
7.02161551e-01 -5.23743212e-01 -8.84697914e-01 -8.10797587e-02
-7.83009291e-01 2.31445402e-01 3.20252717e-01 -8.58006850e-02
-1.11529872e-01 5.17183125e-01 1.27901301e-01 -7.10607350e-01
7.45494425e-01 1.10132480e+00 2.83315480e-01 -3.12200695e-01
5.96974432e-01 4.56465840e-01 8.49112943e-02 -3.31647575e-01
-5.04395291e-02 2.47233391e-01 5.84537327e-01 -7.26520002e-01
6.84932113e-01 8.35287631e-01 4.00868267e-01 -8.29310834e-01
-1.03302622e+00 -7.16657043e-01 -3.25304046e-02 1.34769231e-01
4.99437928e-01 -8.58506441e-01 -9.04283464e-01 -2.38782078e-01
-7.31905997e-01 -1.05298139e-01 -3.77280176e-01 5.49102366e-01
-8.09269130e-01 5.62801719e-01 -4.04920012e-01 -8.32504451e-01
-5.44474661e-01 -1.12450016e+00 9.50319409e-01 -4.16470230e-01
4.84672971e-02 -7.71617651e-01 2.05254376e-01 6.27274752e-01
3.79220039e-01 6.28133714e-01 6.63674533e-01 -3.41608971e-02
-4.01552916e-01 -1.90056145e-01 -2.04733774e-01 4.33371067e-01
-1.36077091e-01 -6.88105822e-01 -2.01964661e-01 -7.45599926e-01
4.10879642e-01 -4.42291088e-02 8.99634898e-01 8.09442937e-01
1.24196911e+00 -7.77221262e-01 -2.60315776e-01 1.05656004e+00
1.77383566e+00 -3.17050666e-02 5.23344576e-01 -3.13935466e-02
2.57824600e-01 2.84810692e-01 6.21180415e-01 1.09757364e+00
9.82709751e-02 5.19725978e-01 5.00004947e-01 -8.22494775e-02
3.68616521e-01 3.69261891e-01 -1.53987156e-02 9.61945474e-01
-1.60467893e-01 3.04809004e-01 -7.52806425e-01 7.00881660e-01
-1.81036401e+00 -9.09693062e-01 -2.88211972e-01 2.42108941e+00
9.43490565e-01 5.71391582e-02 2.41658941e-01 6.04498088e-01
4.01860446e-01 9.17925537e-02 -5.81454098e-01 -3.74572366e-01
-4.61056568e-02 7.57370412e-01 8.33502948e-01 6.46437109e-01
-9.36320603e-01 2.61418909e-01 5.99609661e+00 6.25257015e-01
-1.05380356e+00 5.84019385e-02 5.48662305e-01 -3.52838427e-01
-4.50457811e-01 -6.92841038e-02 -2.94684142e-01 3.52664411e-01
5.92825055e-01 -3.32163304e-01 8.89365613e-01 7.79568315e-01
4.31788504e-01 -2.54627138e-01 -1.00876951e+00 1.60174024e+00
1.98751196e-01 -1.74472547e+00 -7.13064849e-01 4.30411130e-01
1.08220196e+00 -3.18571962e-02 1.45635098e-01 -2.89340436e-01
3.79982479e-02 -1.19617510e+00 4.57461506e-01 1.09096847e-01
1.00633883e+00 -7.07707167e-01 4.91056293e-01 3.33379447e-01
-1.13388824e+00 -1.33910328e-01 -3.75856489e-01 -4.85285968e-02
4.24931467e-01 1.33589816e+00 -4.50076848e-01 3.98720384e-01
7.49293506e-01 7.57500052e-01 3.73028181e-02 1.14790297e+00
2.26030871e-01 7.07516670e-01 -6.08495235e-01 3.71027619e-01
2.93539818e-02 -4.95207340e-01 8.75739038e-01 1.08326793e+00
4.14731443e-01 7.10594952e-01 5.47292352e-01 5.95450342e-01
-2.07299471e-01 2.18154222e-01 -4.68009591e-01 1.03252484e-02
4.95736927e-01 9.93319750e-01 -7.57557988e-01 -6.05587065e-02
-4.46074903e-02 9.91168559e-01 -7.98554253e-03 2.28328139e-01
-7.33238399e-01 -1.52786076e-01 5.26769161e-01 2.50225604e-01
3.71855944e-01 -4.00250971e-01 -6.00558639e-01 -1.25810206e+00
4.28507805e-01 -1.16764987e+00 6.58269167e-01 -1.03609763e-01
-1.12371016e+00 7.03764617e-01 -2.77151316e-01 -1.15984488e+00
-9.67588425e-02 -3.06299895e-01 -2.64964432e-01 6.18280888e-01
-1.19388497e+00 -3.11243564e-01 -3.39540184e-01 7.35577166e-01
2.07815990e-01 2.07524642e-01 8.60465765e-01 5.15633345e-01
-5.46621203e-01 5.95928967e-01 2.75540709e-01 -3.89101682e-03
4.35732044e-02 -9.30219054e-01 -5.46403050e-01 9.87397015e-01
1.08920999e-01 5.57040036e-01 9.68161762e-01 -4.43559855e-01
-1.87412417e+00 -8.00514400e-01 6.28819227e-01 1.78468436e-01
4.30737048e-01 -3.39745209e-02 -4.77095515e-01 6.41430140e-01
-1.95520639e-01 7.69819200e-01 7.74467230e-01 -5.99099249e-02
-2.12180242e-01 -3.27309787e-01 -1.39818609e+00 1.60145670e-01
1.11695600e+00 -2.93090433e-01 -4.54608016e-02 7.60947526e-01
3.57458204e-01 -6.42545462e-01 -1.20038259e+00 6.31420314e-01
3.13840896e-01 -8.12286615e-01 1.10395730e+00 -4.46361363e-01
4.03393209e-01 -8.79239440e-02 -8.09120595e-01 -9.79943931e-01
-1.74260750e-01 -1.05300212e+00 -1.18194439e-01 4.20187682e-01
3.72233272e-01 -4.68555689e-01 9.97809172e-01 5.98013103e-01
-1.83885530e-01 -1.22638667e+00 -1.20968640e+00 -8.95463943e-01
-1.34735465e-01 -8.25113416e-01 -1.03690021e-01 1.00675666e+00
4.89604957e-02 7.81432018e-02 -3.47875446e-01 3.06058854e-01
1.15947199e+00 3.80738705e-01 2.76518911e-01 -9.68435049e-01
-8.20072353e-01 -1.70022380e-02 -2.98863858e-01 -1.03306806e+00
-6.11263700e-02 -9.71100867e-01 5.01784310e-02 -1.45287871e+00
3.09948355e-01 -1.15342438e+00 -1.90023273e-01 2.87361622e-01
8.64489079e-02 6.36032820e-01 2.94862121e-01 3.60539518e-02
-5.55064917e-01 2.09763542e-01 1.04209805e+00 -1.89163208e-01
-2.80373812e-01 1.79212496e-01 -1.02618933e+00 5.80437660e-01
4.25490916e-01 -6.31313503e-01 -2.95382529e-01 -5.79490304e-01
6.79578066e-01 8.26412678e-01 1.64967209e-01 -1.06474030e+00
1.41997039e-01 -1.69468522e-01 -1.49926126e-01 -3.64317834e-01
5.29275775e-01 -8.94716561e-01 1.74974084e-01 8.13238263e-01
-3.06141138e-01 2.82929182e-01 7.32062757e-02 6.17979348e-01
-1.40082076e-01 -2.58394778e-01 8.65460694e-01 -6.03862777e-02
-1.31563872e-01 5.11491001e-01 -1.47858653e-02 4.54854906e-01
9.56030726e-01 -3.21621001e-02 3.05387676e-01 -8.56256962e-01
-9.52397346e-01 2.00107783e-01 2.19084248e-01 -3.95769596e-01
7.90539801e-01 -1.19296229e+00 -1.08992255e+00 1.00555733e-01
-1.84229836e-01 4.41461354e-02 2.64518380e-01 1.43281996e+00
-6.15958631e-01 4.31152910e-01 3.05346042e-01 -7.42358744e-01
-1.16485977e+00 4.96036261e-01 1.16007149e-01 -4.32926923e-01
-5.24172306e-01 1.04967475e+00 -1.29631832e-01 -6.79498464e-02
1.54056743e-01 -2.63833493e-01 3.94422233e-01 -2.29551092e-01
4.40349102e-01 8.49722147e-01 2.61449128e-01 -4.40079361e-01
-6.24934316e-01 5.30336380e-01 3.60618025e-01 -4.57314610e-01
1.78278995e+00 1.11670129e-01 -4.38720912e-01 -1.51852757e-01
1.40521479e+00 4.10058320e-01 -9.28430140e-01 -3.41888875e-01
-2.89339513e-01 -5.97850502e-01 3.07068825e-01 -5.19468546e-01
-1.59180951e+00 4.63444591e-01 4.39589471e-01 1.91047400e-01
1.59744346e+00 1.80101022e-02 9.71468866e-01 3.21552068e-01
6.03568137e-01 -8.91881943e-01 -8.62374008e-02 2.53744632e-01
9.27766383e-01 -1.05722380e+00 4.28402275e-01 -8.49294186e-01
-4.51689184e-01 7.38192022e-01 -1.44500971e-01 -5.68402410e-01
8.82621586e-01 5.44713616e-01 -2.93388069e-01 -2.38891184e-01
-4.11842823e-01 1.19669579e-01 3.58422309e-01 2.27349430e-01
3.20756316e-01 2.44174197e-01 -7.36842692e-01 7.13271677e-01
-2.35502481e-01 1.72867142e-02 4.13445115e-01 1.00912178e+00
-2.65527546e-01 -9.48296428e-01 -5.55095911e-01 7.68474102e-01
-7.44620085e-01 -3.34884137e-01 2.50838518e-01 4.92037758e-02
-4.62496057e-02 1.18186796e+00 -1.86342493e-01 -1.53433070e-01
2.12053061e-01 -6.19578421e-01 6.89112544e-01 -6.63967848e-01
-2.28149027e-01 2.77715415e-01 8.15599859e-02 -9.66337800e-01
-4.82293367e-01 -9.12181556e-01 -1.53236926e+00 -3.18083376e-01
-1.10546775e-01 3.44215810e-01 6.86491311e-01 8.87406707e-01
3.36323172e-01 2.38923401e-01 7.78579533e-01 -4.30854380e-01
-8.67920101e-01 -4.85761166e-01 -7.78372884e-01 3.57630879e-01
3.88398647e-01 -1.87467679e-01 -2.91585416e-01 7.25308433e-02] | [6.867467403411865, 4.477419376373291] |
62a83586-8d82-4f07-9840-857e7a22606f | workload-forecasting-of-a-logistic-node-using | 2211.04976 | null | https://arxiv.org/abs/2211.04976v1 | https://arxiv.org/pdf/2211.04976v1.pdf | Workload Forecasting of a Logistic Node Using Bayesian Neural Networks | Purpose: Traffic volume in empty container depots has been highly volatile due to external factors. Forecasting the expected container truck traffic along with having a dynamic module to foresee the future workload plays a critical role in improving the work efficiency. This paper studies the relevant literature and designs a forecasting model addressing the aforementioned issues. Methodology: The paper develops a forecasting model to predict hourly work and traffic volume of container trucks in an empty container depot using a Bayesian Neural Network based model. Furthermore, the paper experiments with datasets with different characteristics to assess the model's forecasting range for various data sources. Findings: The real data of an empty container depot is utilized to develop a forecasting model and to later verify the capabilities of the model. The findings show the performance validity of the model and provide the groundwork to build an effective traffic and workload planning system for the empty container depot in question. Originality: This paper proposes a Bayesian deep learning-based forecasting model for traffic and workload of an empty container depot using real-world data. This designed and implemented forecasting model offers a solution with which every actor in the container truck transportation benefits from the optimized workload. | ['Anisa Rizvanolli und Olaf Rendel', 'Emin Nakilcioglu'] | 2022-11-09 | null | null | null | null | ['probabilistic-deep-learning', 'probabilistic-time-series-forecasting'] | ['computer-vision', 'time-series'] | [-5.71157694e-01 -1.76571608e-01 -1.84818372e-01 -8.03384364e-01
-3.69957113e-03 8.29158127e-02 2.91106075e-01 -3.24195236e-01
-1.00770377e-01 6.12358391e-01 2.10840464e-01 -4.36211973e-01
-7.89150417e-01 -1.12714720e+00 -4.08994824e-01 -1.01207101e+00
1.10851526e-01 1.32333684e+00 -2.55355597e-01 -4.46453720e-01
3.24706256e-01 8.75314832e-01 -1.83460128e+00 2.49961123e-01
4.07322884e-01 1.01460421e+00 6.04929447e-01 3.65973324e-01
-2.97768116e-01 7.24992514e-01 -6.56621456e-01 9.23741162e-02
6.31249309e-01 5.06408572e-01 -3.93146515e-01 2.56417364e-01
-5.17273009e-01 -4.36606824e-01 -6.94535613e-01 2.28865147e-01
4.25450087e-01 4.91252542e-01 7.16826975e-01 -1.80920494e+00
-3.72031778e-01 7.15844691e-01 -1.64032772e-01 6.30912185e-01
-3.27972978e-01 3.11020434e-01 3.59491378e-01 -2.12890849e-01
1.12828016e-01 1.03242111e+00 5.89197457e-01 1.06973886e-01
-1.00781369e+00 -6.01156950e-01 2.08193749e-01 4.21022594e-01
-1.38367033e+00 -1.75477415e-02 7.36889243e-01 -6.83402538e-01
1.30431020e+00 7.39919627e-03 6.56234026e-01 9.41959798e-01
7.86353469e-01 6.16877258e-01 1.00549090e+00 -1.93677917e-01
1.61053941e-01 2.75655687e-01 2.28959158e-01 -3.46851237e-02
1.14269219e-01 5.11628032e-01 -2.28887573e-01 4.55135375e-01
5.20089045e-02 3.80221009e-01 5.81888318e-01 3.28511924e-01
-5.44509113e-01 8.21695387e-01 1.01707026e-01 3.70665371e-01
-1.15403557e+00 1.70603141e-01 6.42888248e-01 2.72093713e-02
8.39039743e-01 -7.02755898e-02 -5.24385333e-01 -2.59639084e-01
-1.19452500e+00 2.12624982e-01 7.90787935e-01 1.14066124e+00
3.47707510e-01 7.27371991e-01 -4.22986686e-01 2.88391441e-01
3.87486130e-01 9.17375624e-01 5.44696562e-02 -8.68729830e-01
4.05012280e-01 3.27420801e-01 3.03895801e-01 -1.16031051e+00
-1.01672769e+00 -2.78194129e-01 -5.87615788e-01 -2.35542372e-01
-1.46752447e-02 -2.51443058e-01 -1.00174558e+00 1.15274906e+00
-1.44344913e-02 -5.29721240e-03 -1.92757204e-01 7.93009877e-01
5.53498030e-01 1.27102399e+00 3.94091666e-01 6.12044521e-02
1.19079852e+00 -6.67985082e-01 -1.10109401e+00 -3.42491828e-02
2.88796842e-01 -5.99708855e-01 4.20978934e-01 2.79920727e-01
-8.25921893e-01 -8.33235204e-01 -4.38018888e-01 4.06163782e-01
-7.70584345e-01 2.80552447e-01 8.09840858e-01 6.31883383e-01
-8.50605488e-01 2.59401679e-01 -8.51380050e-01 -2.31632978e-01
-3.67452716e-03 4.78150606e-01 2.76687026e-01 -1.76160783e-01
-1.27394092e+00 1.55694771e+00 7.31892228e-01 8.17687988e-01
-9.56164539e-01 -5.97884357e-01 -7.18189836e-01 3.39037329e-01
2.01951340e-01 -5.70277572e-01 1.15382612e+00 -1.61702901e-01
-1.06162059e+00 1.28674045e-01 -9.81593058e-02 -3.92986983e-01
3.11274171e-01 4.47205603e-01 -7.53767014e-01 -4.49850142e-01
1.16034567e-01 2.16845229e-01 3.46960604e-01 -1.53373837e+00
-9.56748366e-01 -4.23409253e-01 -1.21358931e-01 -9.27232876e-02
7.92443529e-02 -1.68127820e-01 2.90618353e-02 -3.88647094e-02
-1.79923072e-01 -1.02146471e+00 -8.53144675e-02 -1.41833746e+00
-3.18629295e-01 -6.31403089e-01 6.84216559e-01 -7.44263470e-01
1.48840964e+00 -1.76268137e+00 -6.43782973e-01 4.63864535e-01
-2.73769736e-01 7.28923455e-02 1.34321243e-01 9.09832060e-01
2.11281031e-02 -3.80077451e-01 5.67207813e-01 -1.46359697e-01
5.38382053e-01 6.38354421e-01 -3.61618042e-01 3.19064677e-01
1.67188991e-03 7.95748591e-01 -4.63230908e-01 -2.53456742e-01
6.90887213e-01 5.21062374e-01 7.04595000e-02 1.19977951e-01
-1.73055440e-01 4.13154542e-01 -3.68686378e-01 6.95736170e-01
1.06345010e+00 3.96526277e-01 1.93023279e-01 -2.74595708e-01
-4.56744015e-01 -1.28113732e-01 -9.23877180e-01 6.95672691e-01
-9.37845409e-01 8.64772856e-01 -2.31709808e-01 -1.05507851e+00
1.50055754e+00 4.32890914e-02 8.66402447e-01 -1.03359556e+00
6.65338278e-01 -8.44919607e-02 -6.74643070e-02 -1.08062315e+00
1.00275922e+00 -2.38711789e-01 -1.79416463e-01 5.91288269e-01
-2.68362641e-01 9.43678468e-02 4.62643027e-01 -3.26710016e-01
7.91475296e-01 1.96845815e-01 -7.34625816e-01 -2.86210716e-01
1.43969253e-01 2.53132701e-01 5.45294583e-01 3.90038460e-01
-3.68861914e-01 -1.85306206e-01 3.15898269e-01 -1.15853751e+00
-1.27689981e+00 -6.71740174e-01 -1.96417257e-01 1.40484178e+00
-7.32056350e-02 4.94618535e-01 -3.91670674e-01 1.30148809e-02
3.19847882e-01 1.49890530e+00 -7.95415938e-01 5.14640883e-02
-5.16610861e-01 -9.37927425e-01 9.64724496e-02 7.79960811e-01
2.05521286e-01 -9.36839938e-01 -5.88853180e-01 4.15809363e-01
-3.45323622e-01 -1.23227036e+00 -2.77375188e-02 3.86182815e-01
-6.15522742e-01 -6.65641785e-01 -1.42479807e-01 -5.09929955e-01
6.01145267e-01 4.72400427e-01 1.08781290e+00 -2.19933957e-01
-8.21256563e-02 2.51111865e-01 1.73928186e-01 -1.12549269e+00
-3.62657458e-01 3.05445790e-01 1.78003967e-01 -1.69500515e-01
1.38230371e+00 -1.70601055e-01 -5.37470162e-01 7.08823562e-01
-8.36912215e-01 -1.49686471e-01 2.51489818e-01 4.99047965e-01
2.19864443e-01 1.01032364e+00 7.53803492e-01 -3.92128199e-01
8.71825516e-01 -1.05940950e+00 -9.96371329e-01 9.70072821e-02
-1.19332051e+00 -1.61726892e-01 2.89102286e-01 7.08624497e-02
-1.36490786e+00 -6.13880344e-02 2.26435270e-02 -3.18649977e-01
-3.95790756e-01 5.66770613e-01 2.55717456e-01 7.97670305e-01
-1.46597829e-02 3.63523424e-01 -9.84884053e-02 -3.84690493e-01
-1.53733209e-01 9.93488669e-01 2.22915098e-01 -4.94358838e-01
4.68444854e-01 3.77684861e-01 3.48993272e-01 -3.94094586e-01
-6.14950299e-01 -6.60800815e-01 -9.30055499e-01 -1.01919770e+00
7.22530901e-01 -8.41734171e-01 -1.69541764e+00 1.42077416e-01
-1.11982453e+00 -3.50893438e-01 -1.07574493e-01 5.80003798e-01
-5.44306219e-01 -6.16822064e-01 -3.91387284e-01 -1.34036613e+00
-6.35960773e-02 -1.41332603e+00 9.91722822e-01 1.28984824e-01
3.72078009e-02 -1.25476015e+00 2.34141387e-02 7.56562591e-01
1.04093754e+00 1.51473507e-01 9.44680393e-01 -6.06887281e-01
-5.36578894e-01 -4.26339805e-01 -2.02793643e-01 3.28896344e-01
-2.77121097e-01 2.43239775e-01 -8.79192650e-01 3.44930254e-02
-1.40946686e-01 3.34010780e-01 5.17067075e-01 9.51855540e-01
1.00947452e+00 1.43187210e-01 -2.70191610e-01 1.25575200e-01
1.42232358e+00 6.99374735e-01 5.34973860e-01 7.32819676e-01
1.19820945e-01 1.31076133e+00 9.38359261e-01 9.06087875e-01
1.02732432e+00 3.08881789e-01 5.73648691e-01 2.82974690e-01
2.59865701e-01 -4.85236160e-02 6.93268999e-02 1.09419894e+00
-2.63866723e-01 -5.66341281e-01 -1.35065067e+00 7.09445715e-01
-1.84009707e+00 -1.25190282e+00 -5.27814746e-01 1.47734582e+00
-1.99110508e-01 1.20419495e-01 1.83159426e-01 7.01450631e-02
8.15345645e-01 -2.12270860e-02 -4.66637582e-01 -1.17928553e+00
2.48904184e-01 -2.88353264e-01 9.64543819e-01 1.46706149e-01
-7.56522477e-01 6.46331072e-01 6.61822557e+00 7.28264689e-01
-7.73583949e-01 1.54400006e-01 7.51375318e-01 -2.19891801e-01
-1.28728420e-01 -3.33293915e-01 -1.21509838e+00 9.78363276e-01
1.77668643e+00 -2.49730736e-01 5.76587141e-01 8.49137723e-01
1.07741535e+00 -5.40716410e-01 -7.45249629e-01 5.84611058e-01
1.09131120e-01 -1.37523317e+00 -3.31993163e-01 2.42285177e-01
6.97742164e-01 2.93508619e-01 1.04574189e-01 8.49199474e-01
3.48627120e-01 -9.28790033e-01 6.37440205e-01 1.12678254e+00
9.65220109e-02 -1.29141366e+00 1.68330133e+00 4.93588835e-01
-8.75534296e-01 -6.68415844e-01 -5.47540128e-01 -8.66026953e-02
7.03894436e-01 5.33890009e-01 -1.29580867e+00 5.90087473e-01
5.90008080e-01 -1.50173083e-01 2.74265688e-02 7.56320179e-01
3.44616294e-01 5.03348529e-01 -3.24171513e-01 -7.12032467e-02
5.74951112e-01 -3.12899590e-01 2.63443473e-03 1.13174105e+00
5.70083022e-01 -1.88825727e-01 2.89705515e-01 8.45917583e-01
3.41563880e-01 -3.80237341e-01 -8.52566421e-01 1.26784042e-01
6.91624999e-01 1.09340119e+00 -6.46396577e-01 4.95119160e-03
-2.13718250e-01 -8.33771378e-03 -3.80625784e-01 3.12315315e-01
-1.18118799e+00 -2.54597038e-01 3.18949938e-01 4.15456861e-01
3.16567898e-01 -4.07605827e-01 -5.42576015e-01 -1.53837070e-01
-2.16335490e-01 -7.75246546e-02 6.60878718e-02 -1.09177303e+00
-1.40900338e+00 7.42404282e-01 5.75817943e-01 -1.10337842e+00
-2.64207274e-01 -8.00946593e-01 -6.04118466e-01 1.01274145e+00
-1.83538043e+00 -1.42808688e+00 -5.56984544e-01 2.99107671e-01
6.85871959e-01 -6.38105810e-01 4.37682480e-01 4.41040665e-01
-9.13336515e-01 -1.61060989e-01 4.17742193e-01 -3.33380222e-01
1.65579006e-01 -8.67591560e-01 1.96306676e-01 4.75361526e-01
-7.78960645e-01 4.21312422e-01 9.62736130e-01 -9.03970599e-01
-1.63383043e+00 -1.12738431e+00 1.21984971e+00 -5.54640055e-01
7.25712299e-01 -1.47949055e-01 -4.45511013e-01 8.07254553e-01
4.27480727e-01 -5.46690762e-01 9.01051223e-01 1.12444557e-01
4.05051619e-01 -6.09786391e-01 -1.26127243e+00 2.66183373e-02
2.26289377e-01 -1.39762923e-01 -5.20010412e-01 6.39613271e-01
3.31104696e-01 -1.85635716e-01 -9.16163802e-01 2.83612996e-01
6.18015051e-01 -8.34745347e-01 4.53772664e-01 -7.25459039e-01
1.45105824e-01 1.20642185e-02 -2.96511978e-01 -1.52974105e+00
-6.47127450e-01 -3.78834069e-01 -2.76806504e-02 1.30134749e+00
1.55862272e-01 -5.89076638e-01 7.72010207e-01 1.41983604e+00
-5.14332592e-01 -7.25139380e-01 -1.20108187e+00 -1.03865349e+00
-1.00242183e-01 -1.07372296e+00 1.43457484e+00 4.88888502e-01
-5.28651536e-01 -2.27404207e-01 -4.41728115e-01 3.59509140e-01
3.73327971e-01 8.36873800e-02 7.27350593e-01 -1.22279644e+00
5.92252672e-01 -2.25802585e-01 -4.13769126e-01 -3.81249726e-01
4.74261105e-01 -8.79625976e-01 1.91077471e-01 -1.84423614e+00
2.47194692e-01 -5.31068683e-01 -4.37736839e-01 3.22465807e-01
6.77117944e-01 -3.60953480e-01 2.06304356e-01 -2.08540093e-02
-1.50017008e-01 5.30700326e-01 9.68750179e-01 -1.49795711e-01
-3.93228009e-02 4.57818240e-01 -3.48379731e-01 5.16052127e-01
1.07415736e+00 -6.95991695e-01 -4.47165668e-01 -7.48425424e-01
3.38867635e-01 6.37006853e-03 9.76687893e-02 -6.50955975e-01
4.81919587e-01 -4.65770751e-01 3.89584482e-01 -1.41228080e+00
3.03797722e-01 -1.45510483e+00 5.41260540e-01 3.68841678e-01
-3.51508036e-02 7.37965822e-01 4.16129321e-01 4.99951214e-01
-5.78506403e-02 -3.56722832e-01 2.18573064e-01 1.88253939e-01
-8.91154826e-01 2.74265707e-01 -9.19548452e-01 -6.94340885e-01
1.51711822e+00 -3.49386185e-01 -4.09948289e-01 -3.73731047e-01
-7.98492849e-01 7.36127138e-01 -4.27364111e-01 6.14640236e-01
3.08916181e-01 -1.33714831e+00 -6.75241172e-01 3.30712736e-01
-2.38078728e-01 -3.68431032e-01 8.97373438e-01 8.80197644e-01
-7.01068223e-01 1.14534795e+00 -4.92692530e-01 -5.54831564e-01
-4.78506595e-01 6.76426053e-01 -3.08535453e-02 -2.99450159e-01
-1.72839284e-01 2.98179597e-01 -6.43099427e-01 -6.77422464e-01
2.01702833e-01 -1.41996861e-01 -2.74305552e-01 4.44331855e-01
2.40556061e-01 1.00921869e+00 4.43176270e-01 -9.57029164e-01
-1.48157328e-01 8.65182281e-02 3.10822248e-01 3.12413424e-01
1.88633084e+00 -6.70649052e-01 4.68104109e-02 6.32468700e-01
8.35421622e-01 -8.01411569e-01 -1.04870665e+00 3.32055166e-02
1.49851009e-01 -7.88170576e-01 7.17933655e-01 -8.94012213e-01
-1.31698287e+00 7.87879050e-01 7.33070135e-01 5.64840436e-01
1.06593704e+00 -3.25871915e-01 1.09137762e+00 2.99111187e-01
4.28564578e-01 -1.78131104e+00 -5.93439698e-01 8.81341994e-01
7.55786836e-01 -1.49842978e+00 -1.46245643e-01 3.36882651e-01
-1.03217208e+00 1.38018382e+00 5.37193835e-01 1.70451984e-01
1.00534046e+00 6.12786651e-01 1.33244380e-01 -5.50739348e-01
-9.66249466e-01 -3.03113490e-01 2.79960427e-02 6.89628482e-01
1.56178728e-01 6.14502728e-01 9.06129479e-02 6.68289602e-01
-2.93118775e-01 3.21687102e-01 4.81383026e-01 8.36822271e-01
-2.53576249e-01 -6.81437671e-01 -4.44290370e-01 5.99223256e-01
-2.05908164e-01 3.65234166e-01 6.25641286e-01 9.32415426e-01
4.12911922e-01 1.40510881e+00 5.39645851e-01 -6.12606287e-01
6.37931585e-01 1.61189318e-01 -6.95035011e-02 -2.41346687e-01
-6.51256621e-01 -5.11690378e-02 2.60979593e-01 -4.10683513e-01
-2.66552776e-01 -5.13826966e-01 -1.07162941e+00 -1.09025288e+00
-3.01431239e-01 4.88958210e-02 1.57659316e+00 9.68683302e-01
5.04516840e-01 1.12375617e+00 1.12582994e+00 -1.17923450e+00
-4.66895968e-01 -1.19672549e+00 -9.90662396e-01 1.36494180e-02
-1.26564965e-01 -1.01255476e+00 -2.03916699e-01 -2.88398743e-01] | [6.295473098754883, 2.6732888221740723] |
0c1ab2ff-4447-4d7f-ac72-41c4c0d00e35 | vehicle-detection-and-classification-without | 2305.08265 | null | https://arxiv.org/abs/2305.08265v2 | https://arxiv.org/pdf/2305.08265v2.pdf | Vehicle Detection and Classification without Residual Calculation: Accelerating HEVC Image Decoding with Random Perturbation Injection | In the field of video analytics, particularly traffic surveillance, there is a growing need for efficient and effective methods for processing and understanding video data. Traditional full video decoding techniques can be computationally intensive and time-consuming, leading researchers to explore alternative approaches in the compressed domain. This study introduces a novel random perturbation-based compressed domain method for reconstructing images from High Efficiency Video Coding (HEVC) bitstreams, specifically designed for traffic surveillance applications. To the best of our knowledge, our method is the first to propose substituting random perturbations for residual values, creating a condensed representation of the original image while retaining information relevant to video understanding tasks, particularly focusing on vehicle detection and classification as key use cases. By not using residual data, our proposed method significantly reduces the data needed in the image reconstruction process, allowing for more efficient storage and transmission of information. This is particularly important when considering the vast amount of video data involved in surveillance applications. Applied to the public BIT-Vehicle dataset, we demonstrate a significant increase in the reconstruction speed compared to the traditional full decoding approach, with our proposed method being approximately 56% faster than the pixel domain method. Additionally, we achieve a detection accuracy of 99.9%, on par with the pixel domain method, and a classification accuracy of 96.84%, only 0.98% lower than the pixel domain method. Furthermore, we showcase the significant reduction in data size, leading to more efficient storage and transmission. Our research establishes the potential of compressed domain methods in traffic surveillance applications, where speed and data size are critical factors. | ['Behçet Uğur Töreyin', 'Muhammet Sebul Beratoğlu'] | 2023-05-14 | null | null | null | null | ['image-reconstruction', 'video-understanding'] | ['computer-vision', 'computer-vision'] | [ 9.22905982e-01 -1.21762194e-01 -2.92415768e-01 -8.40281919e-02
-5.15981317e-01 -3.60348791e-01 2.23940313e-01 9.42355096e-02
-4.67560917e-01 4.79795814e-01 -1.49114773e-01 -7.08930910e-01
1.74314663e-01 -8.73811066e-01 -8.37494552e-01 -6.94575787e-01
-2.05053955e-01 -2.15902582e-01 4.89700854e-01 4.86707985e-02
2.52064526e-01 4.15720373e-01 -1.90362430e+00 2.50162989e-01
7.89598882e-01 1.45492542e+00 3.48539948e-01 9.07661974e-01
2.22124174e-01 7.74844110e-01 -5.64824402e-01 -4.39048946e-01
4.66491789e-01 -2.56717145e-01 -2.35276818e-01 2.84837514e-01
4.92954224e-01 -9.73318994e-01 -4.66784179e-01 1.08992767e+00
2.75980741e-01 -2.57153884e-02 3.11921865e-01 -1.12928081e+00
-1.01176903e-01 -7.45675266e-02 -5.51979423e-01 3.74968380e-01
4.19151098e-01 1.75571382e-01 4.18997347e-01 -3.83608669e-01
5.47338784e-01 9.33099747e-01 5.98228991e-01 4.28662837e-01
-9.82070565e-01 -8.27616155e-01 -2.61725187e-01 7.01553106e-01
-1.41883183e+00 -5.85204124e-01 7.75390029e-01 -2.77405083e-01
8.32993090e-01 4.98491168e-01 6.77964211e-01 7.07517564e-01
1.45829126e-01 5.63896537e-01 7.00492501e-01 -3.87367845e-01
2.00497091e-01 1.08039893e-01 -5.06329425e-02 5.35112083e-01
6.77084088e-01 2.05758274e-01 -2.74124920e-01 -8.94310176e-02
4.66048717e-01 2.43356854e-01 -5.24806798e-01 -1.39953658e-01
-9.65686142e-01 7.36077011e-01 -1.22950420e-01 -2.01155227e-02
-6.10209823e-01 1.92265540e-01 6.05334461e-01 3.88549984e-01
4.25453752e-01 -1.81369752e-01 -5.26903793e-02 -4.81588691e-01
-1.21621072e+00 1.08319350e-01 6.62934303e-01 9.53457892e-01
5.43962061e-01 3.51392254e-02 -6.26927242e-02 5.05220711e-01
1.16906993e-01 9.41349268e-01 1.23547778e-01 -1.31307232e+00
6.22105360e-01 3.89351249e-01 3.95717332e-04 -1.48849142e+00
8.89088288e-02 -7.20356032e-02 -9.42268133e-01 8.29968974e-02
1.41877875e-01 -1.43308491e-02 -5.72716057e-01 1.49165404e+00
2.65693963e-01 5.82603216e-01 4.07563090e-01 7.49428272e-01
5.64245284e-01 1.14019752e+00 -1.10269338e-01 -7.24875808e-01
1.53414166e+00 -5.13727725e-01 -8.06182265e-01 3.60302394e-03
5.80559254e-01 -7.29972541e-01 4.47124660e-01 5.31833291e-01
-9.10070240e-01 -5.60707092e-01 -1.09907019e+00 2.60637075e-01
-4.05323952e-02 5.16392961e-02 1.83905035e-01 9.18298125e-01
-9.60063636e-01 4.58742268e-02 -7.58542240e-01 -4.03860241e-01
5.01944840e-01 2.97317475e-01 -3.13577324e-01 -5.50077021e-01
-1.07572496e+00 5.78531504e-01 2.96357900e-01 -2.04953268e-01
-7.04395235e-01 -8.64144564e-01 -9.04945791e-01 9.41014215e-02
5.35420656e-01 -1.41576901e-01 1.01715553e+00 -7.36213803e-01
-1.20591378e+00 5.02028167e-01 -4.61640477e-01 -1.02981889e+00
4.29181874e-01 1.06939096e-02 -5.98065972e-01 7.77939320e-01
-1.44258469e-01 5.56899369e-01 1.06794071e+00 -8.11050594e-01
-7.74408340e-01 -3.51226516e-02 -8.94865915e-02 -1.71091124e-01
-5.54412186e-01 8.37373286e-02 -7.47640252e-01 -6.66222513e-01
-2.73439825e-01 -9.26667035e-01 -9.42826867e-02 4.06157464e-01
2.16450140e-01 9.45342034e-02 1.35633743e+00 -1.04666579e+00
1.47092330e+00 -2.37917280e+00 -2.81643569e-01 2.26386324e-01
2.71217585e-01 9.20411110e-01 -6.27104566e-02 3.55309367e-01
1.28679872e-01 -7.10903713e-03 -3.71924192e-01 6.36129826e-02
-5.76425970e-01 1.43252447e-01 -3.51489365e-01 5.65860033e-01
1.81343228e-01 5.37119746e-01 -7.13206291e-01 -5.57432890e-01
4.25916702e-01 6.97003901e-01 -7.23682880e-01 1.74777925e-01
-5.21402210e-02 1.59051567e-01 -4.94407237e-01 6.15103781e-01
1.05565906e+00 -6.63103387e-02 2.44011775e-01 -3.59659821e-01
-1.08444266e-01 -6.73005953e-02 -9.85893309e-01 9.57167685e-01
-3.35725367e-01 1.19673002e+00 2.43614808e-01 -1.29355645e+00
9.08127069e-01 4.80195463e-01 7.69693255e-01 -9.63825524e-01
9.04374048e-02 3.81594263e-02 -3.50465715e-01 -7.36256838e-01
5.93234360e-01 1.72995642e-01 2.44172543e-01 1.77404001e-01
-5.25427759e-01 1.18964419e-01 3.97937298e-01 1.61316350e-01
1.25758016e+00 -3.90623391e-01 2.91627735e-01 7.71967322e-02
5.89842677e-01 1.70682684e-01 4.85663027e-01 4.08916086e-01
-3.66820246e-01 2.54083812e-01 4.10925388e-01 -4.16048616e-01
-1.19289279e+00 -6.43662155e-01 -1.91662252e-01 3.22024733e-01
4.03730810e-01 -5.18235028e-01 -8.45899045e-01 -2.15206817e-01
-9.82708633e-02 5.17407775e-01 -1.61167234e-01 -1.96473107e-01
-7.44074285e-01 -5.00343621e-01 5.97452819e-01 1.92388386e-01
9.64981914e-01 -5.26580393e-01 -1.15084052e+00 1.60141289e-01
-4.66151506e-01 -1.76795053e+00 -2.88876534e-01 -6.72854304e-01
-1.06295431e+00 -1.33340728e+00 -6.19497120e-01 -5.40304720e-01
7.36365974e-01 8.04971755e-01 7.32046843e-01 3.13972801e-01
-3.29446554e-01 4.69610512e-01 -6.31077349e-01 -2.19740272e-01
-7.00231194e-01 -4.57486689e-01 -1.02699496e-01 2.45357797e-01
4.20978040e-01 -5.33341728e-02 -6.59551322e-01 4.03461933e-01
-1.17565215e+00 1.49911478e-01 5.33222497e-01 6.07524753e-01
4.72156346e-01 3.90109062e-01 1.65519357e-01 -5.23709059e-01
3.24502349e-01 -5.47730327e-01 -9.46370423e-01 6.18912093e-02
-4.82640862e-01 -2.36121580e-01 7.64285147e-01 -3.77349705e-01
-8.56408000e-01 -6.54445440e-02 -5.87735267e-04 -4.75578755e-01
-7.07186833e-02 2.09025398e-01 1.19222105e-01 -2.52033085e-01
3.46402377e-01 4.97157216e-01 6.22260213e-01 -2.23061025e-01
-1.13009125e-01 8.75259519e-01 5.40574670e-01 5.92691340e-02
6.22006476e-01 6.31610990e-01 2.14235455e-01 -1.39862144e+00
4.52987254e-02 -4.66630280e-01 -1.49865560e-02 -3.84228617e-01
8.25920880e-01 -8.93334389e-01 -6.96616888e-01 4.42386061e-01
-1.15783107e+00 8.93819556e-02 -2.45847423e-02 7.40908206e-01
-4.22957987e-01 9.83523607e-01 -2.99994528e-01 -7.93143511e-01
-2.07226217e-01 -1.41386676e+00 1.06917512e+00 -1.99159965e-01
2.30004445e-01 -6.09142661e-01 -3.73770297e-01 6.25720620e-01
5.78520715e-01 4.34642643e-01 4.96683985e-01 -1.78346902e-01
-9.05675232e-01 -4.84141231e-01 -4.85869735e-01 5.01407862e-01
-2.19412278e-02 -9.73512977e-02 -6.74255729e-01 -5.49392343e-01
6.83594197e-02 1.66305318e-01 8.34384322e-01 3.92551333e-01
1.21559036e+00 -4.62833017e-01 -3.66126567e-01 6.42624080e-01
1.39057195e+00 4.15769726e-01 8.41170609e-01 1.53923839e-01
1.94711342e-01 5.27468801e-01 9.77927089e-01 7.59496510e-01
2.22082570e-01 8.77972484e-01 5.81944525e-01 1.08435494e-03
-2.86608845e-01 1.04776910e-02 5.53524613e-01 4.98126805e-01
-1.00675583e-01 -5.29551208e-01 -8.01804245e-01 4.67886150e-01
-1.59405756e+00 -1.27454281e+00 -2.04610065e-01 2.40380597e+00
4.67764765e-01 -1.15539357e-01 -3.14186625e-02 4.48372781e-01
8.92193496e-01 1.92596659e-01 -3.60326171e-01 -4.84977394e-01
1.61851376e-01 -2.66945455e-02 9.25459802e-01 2.52031952e-01
-1.08550680e+00 4.86749887e-01 6.04248571e+00 9.80125666e-01
-1.12370038e+00 -2.72937715e-02 5.11465490e-01 -7.12153465e-02
8.84645954e-02 -2.14126050e-01 -6.45951867e-01 8.37091148e-01
1.41629839e+00 -2.06050813e-01 4.06241268e-01 6.77206218e-01
5.50815642e-01 -3.48061681e-01 -7.51117349e-01 1.37493956e+00
3.12886953e-01 -1.51116228e+00 1.39147580e-01 3.75976801e-01
4.78878349e-01 -2.57340491e-01 -8.94310698e-03 -2.27336824e-01
-4.46817875e-01 -6.78110182e-01 5.19548297e-01 2.52745956e-01
1.07335603e+00 -7.53733814e-01 8.43081713e-01 3.27695042e-01
-1.38881135e+00 -8.67204443e-02 -4.17022318e-01 7.87409209e-03
3.52676064e-01 6.63168371e-01 -6.48189306e-01 2.83023715e-01
8.18543673e-01 6.78085268e-01 -3.07671607e-01 8.74721169e-01
4.59779054e-01 9.34086680e-01 -4.45032030e-01 7.57920891e-02
-1.91290546e-02 -3.65818702e-02 7.68006742e-01 1.26569152e+00
6.41646445e-01 2.86931515e-01 2.45833807e-02 2.15870619e-01
4.02609147e-02 -7.29309320e-02 -9.53735530e-01 8.23503137e-02
4.96763974e-01 6.06313944e-01 -4.64286447e-01 -5.68897843e-01
-6.90772235e-01 7.00292468e-01 -4.22766298e-01 3.27583283e-01
-1.16277719e+00 -4.22489315e-01 7.59398520e-01 2.56962746e-01
7.20017970e-01 -4.71035480e-01 -9.18777427e-04 -9.47338402e-01
3.59726548e-01 -1.01195920e+00 1.08595930e-01 -2.84086049e-01
-4.35659349e-01 4.37157780e-01 4.21270698e-01 -1.75702584e+00
-3.95129383e-01 -5.93107402e-01 -2.13219076e-01 3.70560646e-01
-1.77354431e+00 -6.86005890e-01 -5.32299876e-01 5.54279804e-01
6.48304582e-01 -2.75292933e-01 3.28826815e-01 5.96051872e-01
-4.55690324e-01 6.27725184e-01 3.44954848e-01 -4.85516079e-02
3.11594188e-01 -1.85934842e-01 2.20139235e-01 1.24378216e+00
-3.52292895e-01 2.83957452e-01 7.99574912e-01 -5.30253887e-01
-1.95344698e+00 -1.24331605e+00 6.29048347e-01 1.59537300e-01
2.57703751e-01 -1.41187415e-01 -7.96867847e-01 1.10565938e-01
1.90879982e-02 -1.12174675e-02 7.11700439e-01 -9.28903222e-01
-1.16950288e-01 -4.28087980e-01 -1.48699403e+00 4.75648522e-01
7.79497206e-01 -4.18765217e-01 -2.79514670e-01 2.25143284e-01
6.76895082e-01 -1.26308590e-01 -6.51484728e-01 5.45079112e-01
5.62680840e-01 -9.95853901e-01 1.01394200e+00 1.50459915e-01
3.25964987e-01 -4.11458135e-01 -3.86220872e-01 -6.60121620e-01
7.31395632e-02 -5.20804167e-01 -4.43551451e-01 9.17863131e-01
-1.11897290e-02 -7.16955245e-01 8.07661474e-01 5.55901587e-01
1.59591809e-01 -5.69567800e-01 -1.05209267e+00 -8.24183106e-01
-6.64828241e-01 -1.01835418e+00 4.50345665e-01 3.74218017e-01
-3.73040468e-01 -4.36267853e-01 -6.68669820e-01 4.14288998e-01
8.65723312e-01 -6.92505762e-02 8.55128586e-01 -8.90537977e-01
-1.75264746e-01 -7.14516193e-02 -9.72678542e-01 -1.41423512e+00
-1.09935626e-01 -3.85399222e-01 -1.43890128e-01 -1.09668553e+00
1.78591684e-01 -2.42094398e-01 1.40505254e-01 9.43549722e-02
1.24115452e-01 7.23362088e-01 3.49805981e-01 3.17864895e-01
-4.02164906e-01 2.24799424e-01 9.79771376e-01 -1.99844524e-01
1.00561030e-01 2.54398286e-02 -4.25393432e-01 4.01773155e-01
9.22579885e-01 -2.70918608e-01 -5.67938268e-01 -4.16167527e-01
-1.46004215e-01 2.16767952e-01 5.99382937e-01 -1.21598661e+00
3.21996838e-01 -1.60073712e-01 -1.32461060e-02 -6.20858610e-01
4.49718475e-01 -1.26760375e+00 2.59476393e-01 8.77563477e-01
1.10240109e-01 -2.33006943e-03 4.29811954e-01 8.38359714e-01
-4.31728303e-01 -1.42943934e-01 7.14967966e-01 2.91553885e-01
-1.10989153e+00 1.22419097e-01 -7.50524342e-01 -2.27389038e-01
1.61608171e+00 -8.04133236e-01 -2.33944103e-01 -5.33864796e-01
-1.43151134e-01 4.58681211e-02 5.74362040e-01 3.07548106e-01
9.43312109e-01 -1.04728615e+00 -7.12528408e-01 7.05595255e-01
-3.19005586e-02 -2.89763570e-01 4.73699272e-01 7.44358718e-01
-1.08314681e+00 6.96949184e-01 -1.01810507e-01 -8.36209238e-01
-1.73053384e+00 7.06226289e-01 -2.11682230e-01 -3.00144348e-02
-6.93481028e-01 2.38634869e-01 1.01484373e-01 4.94154185e-01
1.97157502e-01 -4.93956268e-01 -5.49459495e-02 -2.58598775e-01
9.77553606e-01 8.11502099e-01 1.93038527e-02 -8.54856253e-01
-3.56375575e-01 8.61609161e-01 1.86202765e-01 1.60933241e-01
1.12012362e+00 -1.71284720e-01 8.17314237e-02 -4.02107984e-01
1.42761946e+00 -6.68302476e-02 -1.36373615e+00 -4.76211384e-02
-3.54774922e-01 -8.53017569e-01 3.31125647e-01 -1.83403686e-01
-1.27340639e+00 8.01321626e-01 9.52205420e-01 2.67181724e-01
1.73784137e+00 -2.53388941e-01 1.19290364e+00 3.57233584e-01
5.11501014e-01 -7.98554540e-01 -3.04853648e-01 5.87097891e-02
6.19502723e-01 -1.27475297e+00 1.63813800e-01 -7.11780250e-01
-4.30083752e-01 1.06978023e+00 1.36927620e-01 1.28928646e-01
5.26019752e-01 2.73569167e-01 -2.21770898e-01 1.03406176e-01
-7.84911454e-01 7.34359846e-02 1.62415817e-01 7.91071892e-01
1.13793850e-01 -4.49994095e-02 -5.47572196e-01 -1.38212875e-01
1.94779217e-01 2.37050295e-01 5.68155169e-01 1.06298459e+00
-5.50576925e-01 -1.07224739e+00 -7.28975713e-01 4.87541527e-01
-5.08548617e-01 -1.80610865e-02 2.10171416e-01 5.57449043e-01
7.18270838e-02 1.38379240e+00 3.44247967e-01 -4.29087460e-01
1.68373600e-01 -4.06348586e-01 7.45507404e-02 -8.82084593e-02
2.39999406e-02 -2.07941011e-01 6.57020658e-02 -7.66751647e-01
-7.07977414e-01 -5.02254248e-01 -9.41070139e-01 -7.64745116e-01
-2.02838719e-01 9.85375568e-02 7.89976776e-01 5.61751187e-01
6.37895107e-01 2.64364839e-01 7.45650530e-01 -7.55944252e-01
-3.07041556e-01 -3.95199567e-01 -1.79269060e-01 4.02938426e-01
6.77544296e-01 -5.61879218e-01 -1.95033535e-01 3.98870707e-01] | [10.323991775512695, -0.8518844842910767] |
668fb220-01ac-478e-a1b7-d4eca502be05 | improving-human-annotation-in-single-object | 1911.02807 | null | https://arxiv.org/abs/1911.02807v1 | https://arxiv.org/pdf/1911.02807v1.pdf | Improving Human Annotation in Single Object Tracking | Human annotation is always considered as ground truth in video object tracking tasks. It is used in both training and evaluation purposes. Thus, ensuring its high quality is an important task for the success of trackers and evaluations between them. In this paper, we give a qualitative and quantitative analysis of the existing human annotations. We show that human annotation tends to be non-smooth and is prone to partial visibility and deformation. We propose a smoothing trajectory strategy with the ability to handle moving scenes. We use a two-step adaptive image alignment algorithm to find the canonical view of the video sequence. We then use different techniques to smooth the trajectories at certain degree. Once we convert back to the original image coordination, we can compare with the human annotation. With the experimental results, we can get more consistent trajectories. At a certain degree, it can also slightly improve the trained model. If go beyond a certain threshold, the smoothing error will start eating up the benefit. Overall, our method could help extrapolate the missing annotation frames or identify and correct human annotation outliers as well as help improve the training data quality. | ['Yu Pang', 'Haibin Ling', 'Lin Yuan', 'Xinyi Li'] | 2019-11-07 | null | null | null | null | ['video-object-tracking'] | ['computer-vision'] | [ 6.77382424e-02 -1.14592962e-01 -1.18363857e-01 -2.98638731e-01
-3.86208743e-01 -5.21577299e-01 3.88879269e-01 -2.05810275e-02
-4.05337095e-01 5.39937973e-01 -9.71046239e-02 5.60640581e-02
2.02385888e-01 -2.90324420e-01 -7.17685699e-01 -7.29706347e-01
6.89505786e-02 2.88452238e-01 9.47016597e-01 4.08469550e-02
2.18142852e-01 4.69733238e-01 -1.61775053e+00 1.68381184e-01
7.48731077e-01 8.09808612e-01 2.11103559e-01 6.04206562e-01
1.17247917e-01 8.05285037e-01 -6.22417033e-01 -5.44145942e-01
3.72354746e-01 -2.38190740e-01 -6.64017677e-01 3.97197872e-01
7.22482145e-01 -3.04495186e-01 5.75010106e-03 1.40716922e+00
2.99291044e-01 1.53601468e-01 2.76499033e-01 -1.36442053e+00
-4.63138968e-02 1.69355914e-01 -6.77100182e-01 4.33988608e-02
4.91944879e-01 2.03993544e-01 4.81709749e-01 -9.29180145e-01
8.64816368e-01 1.02775764e+00 1.00767064e+00 5.78423917e-01
-8.71149242e-01 -5.89006722e-01 1.42298892e-01 2.15758428e-01
-1.35411847e+00 -4.95614171e-01 6.47175968e-01 -7.70853221e-01
1.41491950e-01 3.41278285e-01 6.98373914e-01 8.45218241e-01
-7.97013864e-02 7.53240168e-01 7.14374542e-01 -4.66579258e-01
9.60150361e-02 1.58047348e-01 1.22145191e-02 6.65801108e-01
3.10671926e-01 -1.39663126e-02 -3.55798513e-01 -2.33787987e-02
7.15180933e-01 4.79779392e-02 -5.16552448e-01 -6.52289212e-01
-1.28271997e+00 3.22249681e-01 2.28772372e-01 4.20541018e-01
-2.67321736e-01 1.46051571e-02 5.34984827e-01 -5.80147691e-02
3.77692908e-01 1.16679117e-01 -2.96662778e-01 -2.34376058e-01
-1.22186303e+00 2.56157905e-01 4.59499508e-01 9.86478627e-01
6.09786093e-01 -5.43361604e-02 -3.75554353e-01 4.70235050e-01
1.52790308e-01 4.27207947e-01 3.11823368e-01 -1.19916785e+00
1.66629434e-01 5.54148555e-01 5.75781763e-01 -1.12846410e+00
-3.11417431e-01 -8.90201405e-02 -6.29797459e-01 5.33957660e-01
1.06086802e+00 5.90743758e-02 -7.54810870e-01 1.38132000e+00
5.45869172e-01 4.78349686e-01 -3.89495343e-01 1.12069428e+00
3.78353596e-01 3.54346931e-01 1.65266633e-01 -4.54783380e-01
1.35658121e+00 -9.70734179e-01 -1.00040650e+00 -5.70180975e-02
8.01743984e-01 -1.02142107e+00 9.79829252e-01 3.86215836e-01
-1.11485469e+00 -8.53142500e-01 -8.51903439e-01 2.02414349e-01
1.34503869e-02 5.42325199e-01 1.54062241e-01 6.96393073e-01
-9.09765601e-01 8.03131819e-01 -1.16561055e+00 -4.29306000e-01
2.07339302e-01 1.71093345e-01 -3.98531616e-01 4.25282568e-02
-8.25172544e-01 9.04668450e-01 4.88805920e-01 1.12245657e-01
-4.27195817e-01 -4.35083359e-01 -7.08350301e-01 -2.64696449e-01
6.06343150e-01 -4.50318038e-01 1.27628493e+00 -1.09008002e+00
-1.23257291e+00 7.16502309e-01 -3.71841609e-01 -3.73035729e-01
9.93095160e-01 -4.35755312e-01 -2.47241393e-01 -3.09249898e-03
5.07846177e-02 5.76305330e-01 8.53869200e-01 -1.29802144e+00
-9.53021288e-01 -2.11470827e-01 -2.03270048e-01 8.35344493e-02
-1.88505337e-01 2.39922851e-01 -8.30658734e-01 -7.11031973e-01
4.65306453e-02 -1.25402653e+00 -1.24334216e-01 2.27139190e-01
-2.07912982e-01 -1.24447964e-01 9.86292243e-01 -9.26777720e-01
1.35669887e+00 -2.10993862e+00 -3.16249207e-02 1.80510700e-01
2.10181668e-01 5.20711482e-01 3.10675949e-01 -6.46804422e-02
3.41369957e-02 -4.87249419e-02 -8.29566270e-02 -4.49605435e-01
-3.89741570e-01 7.15889409e-02 5.36386147e-02 7.05985665e-01
8.48487541e-02 5.37708640e-01 -9.81861115e-01 -6.86379671e-01
3.36112797e-01 3.44979376e-01 -3.95899087e-01 2.39056826e-01
-6.12427816e-02 7.67711818e-01 -2.92330712e-01 4.94955182e-01
6.50003552e-01 -8.87068808e-02 4.47235070e-02 -4.79578555e-01
-2.48145550e-01 -2.09040910e-01 -1.50766194e+00 1.49815071e+00
5.88730313e-02 9.77994800e-01 2.29255706e-02 -7.69580245e-01
7.89118469e-01 3.99674922e-01 7.27142036e-01 -3.90082419e-01
1.87216088e-01 7.40554631e-02 9.30453390e-02 -5.67749321e-01
7.39265144e-01 3.30582500e-01 4.74133909e-01 3.92296165e-02
-2.35429585e-01 3.67066801e-01 3.39118034e-01 4.05574031e-02
7.56890833e-01 4.98022705e-01 3.60969514e-01 -1.66914582e-01
6.90707803e-01 1.13611542e-01 7.67980099e-01 5.20020366e-01
-4.58879352e-01 6.61190808e-01 2.72878945e-01 -5.65751970e-01
-1.35900331e+00 -5.45024157e-01 3.44097912e-02 8.52110565e-01
3.18192273e-01 -4.26481217e-01 -9.79582250e-01 -7.82897830e-01
-3.57459277e-01 4.10073549e-01 -4.16134119e-01 -3.97973917e-02
-8.41068625e-01 -2.89961308e-01 3.72417271e-01 5.75219631e-01
4.68978614e-01 -8.77266586e-01 -7.44503856e-01 6.35642484e-02
-4.25835460e-01 -1.31688762e+00 -6.87278926e-01 -2.85990387e-01
-7.54851878e-01 -1.24525857e+00 -8.53101611e-01 -6.83546960e-01
8.17828238e-01 4.25711095e-01 8.77335489e-01 5.66233158e-01
1.20054208e-01 1.80660725e-01 -4.25969005e-01 -2.48675182e-01
-6.53863609e-01 -2.34596014e-01 2.91297436e-01 6.09691776e-02
3.54357481e-01 2.66139652e-03 -6.35860264e-01 9.12849069e-01
-7.11646855e-01 5.21649569e-02 1.10751785e-01 5.32469213e-01
4.85780180e-01 3.97245921e-02 1.01839952e-01 -6.16390944e-01
1.60694644e-01 1.51547313e-01 -8.83156061e-01 3.44298244e-01
-5.67469954e-01 -2.87997909e-02 4.25708830e-01 -7.19832599e-01
-9.60002899e-01 5.57615697e-01 -6.13121204e-02 -8.76734555e-01
-2.35109121e-01 -3.77541757e-03 -1.02858275e-01 -2.86397070e-01
6.79696918e-01 -2.30099484e-01 -2.38959137e-02 -3.90980810e-01
2.43664131e-01 4.28193361e-01 6.40415251e-01 -3.58596116e-01
8.38382781e-01 5.02055168e-01 -2.08342612e-01 -7.59765804e-01
-8.59106064e-01 -7.46100605e-01 -8.74534249e-01 -7.72373736e-01
9.66382325e-01 -6.69702530e-01 -6.79387689e-01 2.92701751e-01
-1.32165825e+00 -3.00704658e-01 -9.33931619e-02 5.92745841e-01
-4.21759725e-01 7.32127964e-01 -3.20073575e-01 -9.59858477e-01
-4.79977652e-02 -1.24119115e+00 1.13516355e+00 2.54144371e-01
-1.74780726e-01 -9.55369413e-01 8.34710151e-02 5.86483628e-02
1.83126628e-02 2.77854204e-01 -9.89926700e-03 -5.94636261e-01
-5.75830102e-01 -2.33047158e-01 -4.21654806e-02 3.15032244e-01
-3.73469181e-02 4.21196818e-01 -8.57650638e-01 -3.66966456e-01
6.75604343e-02 2.63238728e-01 5.37763357e-01 4.83075798e-01
9.31227446e-01 -1.65569663e-01 -4.68411058e-01 3.76017243e-01
1.06454015e+00 1.99574634e-01 8.39833677e-01 5.57081342e-01
8.86436403e-01 6.88301206e-01 1.29858923e+00 2.12309405e-01
4.05891985e-02 1.14147174e+00 2.56910652e-01 -7.82018602e-02
-1.75378025e-01 -9.81757641e-02 5.17240882e-01 6.04629695e-01
-5.29087961e-01 6.79923594e-03 -1.02207839e+00 4.37619537e-01
-2.23772502e+00 -1.26591539e+00 -6.16657913e-01 2.67297745e+00
5.76327920e-01 2.84946263e-01 4.22217518e-01 2.65848905e-01
1.11715937e+00 -2.40777016e-01 -4.06503230e-02 1.20459370e-01
2.11822987e-01 -4.82764274e-01 7.21421957e-01 3.85644704e-01
-1.07170928e+00 9.10918474e-01 6.00603294e+00 7.23267138e-01
-9.96029675e-01 1.26002476e-01 2.61338443e-01 2.70234615e-01
3.73805523e-01 8.66706967e-02 -9.28963900e-01 6.17636561e-01
5.98089874e-01 -1.78931002e-02 9.09968391e-02 8.71043205e-01
5.45429647e-01 -2.07948923e-01 -1.03771377e+00 1.04020071e+00
-2.30406299e-02 -1.04830551e+00 -3.56421769e-01 -5.28598810e-03
7.27143466e-01 -3.30669999e-01 -4.28466171e-01 -4.88146953e-02
-5.60167357e-02 -6.15660250e-01 1.05583179e+00 6.79831088e-01
5.08539975e-01 -5.63415468e-01 9.66682136e-01 6.13278329e-01
-1.37951660e+00 1.58168092e-01 -3.88047636e-01 1.28731385e-01
5.49121082e-01 4.28423852e-01 -8.59105289e-01 6.12446189e-01
7.08475649e-01 6.82990193e-01 -8.27434003e-01 1.53803504e+00
-1.79378629e-01 4.65954244e-01 -2.71903068e-01 2.08741680e-01
-1.82147071e-01 -2.80413002e-01 6.98591709e-01 1.13145936e+00
4.80988443e-01 -1.96668077e-02 6.26731932e-01 3.74726802e-01
2.11447284e-01 1.77126348e-01 -5.00834882e-01 2.79493928e-01
3.64172876e-01 1.18102074e+00 -9.65272427e-01 -4.88281101e-01
-4.57478791e-01 9.63985801e-01 1.58663720e-01 2.95485288e-01
-1.15465200e+00 4.50848043e-03 3.02773237e-01 4.37966198e-01
1.44710779e-01 -2.48410016e-01 -1.52373731e-01 -1.21816337e+00
1.68633923e-01 -7.43059039e-01 2.87692517e-01 -9.30370808e-01
-7.17797399e-01 6.10742748e-01 -1.20517053e-02 -1.66826904e+00
-2.21840248e-01 -3.91884238e-01 -5.15763164e-01 3.90438050e-01
-1.04437792e+00 -9.78277862e-01 -6.19605839e-01 5.04303515e-01
5.31414747e-01 1.73153713e-01 2.68737197e-01 6.54081881e-01
-3.91019374e-01 4.19973314e-01 -1.05429530e-01 3.33289862e-01
1.08878887e+00 -1.02087080e+00 1.58870995e-01 1.15064704e+00
2.06795737e-01 4.45977360e-01 1.21512878e+00 -8.50378573e-01
-6.76570296e-01 -1.12812114e+00 6.88222706e-01 -7.49653816e-01
4.52254355e-01 3.57645787e-02 -1.31262898e+00 7.98883140e-01
3.17143947e-02 5.85705005e-02 1.54362842e-01 -1.25210822e-01
1.76183939e-01 6.09828383e-02 -9.35191751e-01 5.49486697e-01
1.00852239e+00 -4.32882383e-02 -4.89528507e-01 2.21300691e-01
4.66554046e-01 -7.62443662e-01 -8.15860927e-01 4.86360908e-01
6.55287206e-01 -1.09419572e+00 8.67218435e-01 -3.32572192e-01
-1.11233490e-03 -7.35145748e-01 2.83669233e-01 -1.07751048e+00
-2.56707251e-01 -5.32834709e-01 -1.38416037e-01 1.37640536e+00
1.24070458e-02 -2.57363264e-03 9.85538125e-01 7.76997745e-01
3.44116380e-03 -2.76241392e-01 -7.67634749e-01 -9.69350934e-01
-4.95547056e-01 -4.34990764e-01 2.07928658e-01 9.64735448e-01
-1.28332943e-01 -2.91344617e-02 -7.48055339e-01 3.56622547e-01
7.63157308e-01 -3.59104544e-01 1.12828779e+00 -1.28850365e+00
1.16276667e-01 -2.62230217e-01 -5.79464197e-01 -1.04451358e+00
-1.29191265e-01 -4.00467396e-01 2.71939993e-01 -1.27661967e+00
1.18960463e-01 -3.53400797e-01 1.61146410e-02 2.62944102e-01
-3.79636526e-01 2.85966963e-01 2.44257748e-01 4.81369495e-01
-8.06044221e-01 1.54811203e-01 1.13598323e+00 1.96237803e-01
-3.07695866e-01 3.10259104e-01 8.26951023e-03 1.04313838e+00
8.15883577e-01 -7.13037848e-01 -4.49212268e-02 -1.56974018e-01
2.11134227e-03 -1.35458723e-01 5.32580614e-01 -1.16460013e+00
3.84309649e-01 -1.31799832e-01 3.04743648e-01 -7.53230631e-01
-2.59750546e-03 -1.13748515e+00 4.19512451e-01 5.07828534e-01
-1.65526375e-01 1.89406589e-01 3.45795043e-02 4.87285107e-01
-1.48861542e-01 -4.54824120e-01 9.58671391e-01 4.14279178e-02
-7.48293936e-01 1.40245885e-01 -6.46164939e-02 -1.82564080e-01
1.18669009e+00 -6.21080279e-01 -5.22774383e-02 -3.89079720e-01
-7.62386382e-01 3.32437992e-01 1.01094067e+00 4.27664220e-01
2.94625878e-01 -1.47024643e+00 -4.61579740e-01 -4.74372357e-02
1.43098459e-01 -8.51555094e-02 6.24836935e-03 1.25000799e+00
-5.19616842e-01 6.36442080e-02 -1.18458949e-01 -1.13716197e+00
-1.58294261e+00 7.56424308e-01 2.62060463e-01 1.33228581e-02
-6.51251614e-01 2.41182134e-01 -1.65123530e-02 1.54441074e-01
5.04537582e-01 -3.76619369e-01 -4.13722306e-01 -1.78123582e-02
7.82648683e-01 6.08364582e-01 -2.97167972e-02 -9.68505025e-01
-4.03544813e-01 8.29965174e-01 1.81776732e-01 4.15967852e-02
8.16252172e-01 -3.73272985e-01 2.48930186e-01 2.66510695e-01
8.54577780e-01 4.26339418e-01 -1.57166016e+00 -7.38301203e-02
3.21083933e-01 -8.80376339e-01 -2.50155509e-01 -2.78740346e-01
-1.11299419e+00 6.88009143e-01 8.67594957e-01 4.32085007e-01
8.78109634e-01 -7.58645907e-02 5.32306492e-01 3.41953337e-02
2.98378378e-01 -1.18951094e+00 1.73065718e-02 1.83113679e-01
4.93428230e-01 -1.40308893e+00 1.53306127e-01 -6.22738242e-01
-8.86151016e-01 1.19247591e+00 7.06995249e-01 6.80435821e-02
2.02972069e-01 2.83230841e-01 6.03294037e-02 -3.94990072e-02
-2.71294951e-01 -3.62715006e-01 5.49213231e-01 6.11555159e-01
4.59055930e-01 -3.07904482e-01 -3.55813295e-01 2.16669008e-01
1.38649985e-01 1.56643927e-01 5.21037519e-01 6.95391893e-01
-8.48439634e-01 -1.02450037e+00 -7.13621676e-01 9.32360366e-02
-4.71129894e-01 3.65676612e-01 -1.86991483e-01 8.64471555e-01
2.01437622e-01 7.98487544e-01 -7.79491216e-02 -2.44570553e-01
5.29410303e-01 1.57042176e-01 3.59342784e-01 -2.10316926e-01
-3.00853342e-01 3.93174380e-01 5.70375361e-02 -7.40190446e-01
-9.12561476e-01 -7.90972114e-01 -1.27207041e+00 -3.54352921e-01
-6.94386065e-01 9.19693038e-02 4.93907243e-01 8.36590827e-01
9.24012214e-02 4.17960674e-01 2.57258207e-01 -9.84398067e-01
-3.00142705e-01 -7.57532239e-01 -1.17592581e-01 8.22960615e-01
1.89538181e-01 -8.73360634e-01 -2.80149788e-01 6.49419844e-01] | [8.76412296295166, -0.5624153017997742] |
a3084d0a-d77f-4527-b899-2773c0db1e53 | multi-target-range-and-angle-detection-for | 2302.14327 | null | https://arxiv.org/abs/2302.14327v1 | https://arxiv.org/pdf/2302.14327v1.pdf | Multi-target Range and Angle detection for MIMO-FMCW radar with limited antennas | Multiple-input multiple-output (MIMO) radar has several advantages with respect to the traditional radar array systems in terms of performance and flexibility. However, in order to achieve high angular resolution, a MIMO radar requires a large number of transmit and receive antennas, which increases hardware design and computational complexities. Although spatial compressive sensing (CS) has been recently considered for a pulsed-waveform MIMO radar with sparse random arrays, such methods for the frequency-modulated continuous wave (FMCW) radar remain largely unexplored. In this context, we propose a novel multi-target localization algorithm in the range-angle domain for a MIMO FMCW radar with a sparse array of randomly placed transmit and receive elements. In particular, we first obtain the targets' range-delays using a discrete Fourier transform (DFT)-based focusing operation. The target angles are then recovered at each detected range using CS-based techniques exploiting the sparsity of the target scene. Our simulation results demonstrate the effectiveness of the proposed algorithm over the classical methods in detecting multiple targets with a sparse array. | ['Arpan Chattopadhyay', 'Himali Singh'] | 2023-02-28 | null | null | null | null | ['compressive-sensing'] | ['computer-vision'] | [ 9.85554636e-01 -6.00888252e-01 6.12830698e-01 -3.84812266e-01
-7.87985206e-01 -6.41647577e-01 4.30673599e-01 -3.05780560e-01
-2.98331797e-01 6.63525701e-01 -1.95811868e-01 -1.36445761e-01
-8.98028910e-01 -7.23395765e-01 -1.33088484e-01 -1.14955628e+00
-4.40476030e-01 2.52696455e-01 -7.57835731e-02 -1.49241285e-02
1.63887262e-01 9.82894957e-01 -1.23532081e+00 -1.48552462e-01
7.53655195e-01 1.10514688e+00 2.29570583e-01 6.25524163e-01
5.17199874e-01 5.16952038e-01 -8.20826232e-01 1.41460538e-01
5.84421039e-01 -4.61853921e-01 7.85688579e-01 -3.80135626e-02
6.20871902e-01 -7.37366527e-02 -5.29277563e-01 1.16112924e+00
4.76684928e-01 -6.89136144e-03 5.87039292e-01 -8.28221321e-01
-2.91159242e-01 2.22448274e-01 -8.09890985e-01 2.21989051e-01
3.60107347e-02 -4.43960547e-01 4.37710971e-01 -1.16493499e+00
-1.47771820e-01 8.59029531e-01 4.99810010e-01 -2.82693446e-01
-5.92663109e-01 -8.61581147e-01 -3.27026576e-01 2.75132805e-01
-1.73311377e+00 -5.95703006e-01 7.30625510e-01 -2.27928415e-01
2.75100321e-01 4.21783805e-01 5.10335803e-01 3.64231646e-01
4.78602380e-01 -6.79114461e-02 1.27760351e+00 -5.56770563e-01
1.57597527e-01 -5.09247303e-01 -1.94604263e-01 2.51795888e-01
1.18443251e+00 6.89419270e-01 -3.83279949e-01 -2.26361841e-01
6.79896355e-01 4.12858218e-01 -5.73244810e-01 -3.71842265e-01
-1.16904354e+00 9.55313444e-01 3.73975784e-01 6.75191581e-01
-6.30087495e-01 1.27424046e-01 -4.50703889e-01 2.96683520e-01
2.09101215e-01 5.71605325e-01 2.05866009e-01 6.49674118e-01
-1.31314516e+00 2.19891071e-01 7.55166590e-01 7.32806742e-01
4.92446542e-01 8.69574189e-01 3.76412012e-02 3.58771980e-01
3.97225320e-01 1.65651906e+00 -2.52793431e-01 -4.64845628e-01
2.92974114e-01 -1.54234260e-01 2.98857898e-01 -1.39331460e+00
-2.99459398e-01 -1.44844759e+00 -1.26159680e+00 3.58425409e-01
9.28618833e-02 -4.75760758e-01 -6.79055452e-01 1.29913509e+00
2.33561605e-01 6.00442469e-01 5.79834223e-01 1.07946157e+00
3.19838405e-01 7.04397142e-01 -5.67149282e-01 -7.64047205e-01
1.19321704e+00 -3.13023210e-01 -9.40330327e-01 -9.43864286e-01
-8.01585615e-02 -1.01578999e+00 -4.38413620e-01 5.84277391e-01
-6.36453569e-01 -3.06796044e-01 -1.35128450e+00 1.07595575e+00
2.99057752e-01 1.31220534e-01 3.70874763e-01 7.00599492e-01
-3.88119817e-01 -8.08004215e-02 -3.53361458e-01 2.96155244e-01
7.96869844e-02 -1.99558824e-01 -1.23466194e-01 -8.40456247e-01
-9.39048946e-01 9.24736619e-01 2.25603860e-02 5.35891175e-01
-5.40062845e-01 -5.98069966e-01 -8.21978509e-01 2.25930512e-02
3.76521409e-01 -2.92368025e-01 6.11877561e-01 -3.35893899e-01
-8.89232516e-01 -1.20273642e-01 -2.46632785e-01 -5.14277577e-01
-1.39389947e-01 -3.27419043e-01 -9.64994609e-01 4.62991536e-01
1.51480198e-01 -4.70414191e-01 1.18455756e+00 -1.08623123e+00
-6.00474775e-01 -5.27825713e-01 -6.85483456e-01 1.31968260e-01
2.12197438e-01 -3.88605475e-01 5.56638539e-01 -8.27251494e-01
6.56912684e-01 -7.70763099e-01 -4.56580788e-01 -3.66376907e-01
1.18769184e-02 7.02480376e-01 9.34159636e-01 -3.09039414e-01
1.00963938e+00 -2.46495414e+00 8.86564031e-02 3.93911451e-01
1.16538405e-01 2.35063747e-01 -1.53716356e-01 7.62659311e-01
8.73024911e-02 -8.70857179e-01 -3.38991076e-01 3.39060187e-01
-5.71973085e-01 -1.57540023e-01 -5.51900089e-01 9.59863722e-01
-8.30532610e-02 4.60184544e-01 -4.79311258e-01 -6.45146668e-02
2.12845847e-01 4.74392802e-01 -1.70044109e-01 -1.31872781e-02
1.90402731e-01 5.94860554e-01 -7.50422060e-01 8.45836401e-01
1.28915560e+00 6.38920888e-02 1.11289516e-01 -4.10343200e-01
-7.15998232e-01 -7.04025745e-01 -1.43802845e+00 1.07184565e+00
-2.76842594e-01 6.11550152e-01 6.76961303e-01 -1.22274685e+00
1.36573219e+00 3.70773762e-01 4.52788174e-01 -7.85338223e-01
-1.74509436e-02 4.23252404e-01 3.21246147e-01 -2.14815110e-01
1.81990117e-01 -5.05924225e-01 -3.24791670e-01 1.45246536e-01
-4.89112645e-01 -2.52780706e-01 -1.12005293e-01 -1.43926769e-01
1.49850273e+00 -5.43523967e-01 7.71488070e-01 -1.98312500e-03
4.99684989e-01 2.33605951e-01 5.54984450e-01 7.80988097e-01
3.44515681e-01 1.84712633e-01 -6.87723875e-01 -1.84529096e-01
-5.97285092e-01 -1.12684619e+00 -1.78202495e-01 2.98680127e-01
3.34648758e-01 3.27659070e-01 1.87591836e-01 3.00422519e-01
2.54503310e-01 7.72264600e-01 -5.80178723e-02 1.33932993e-01
-8.44468653e-01 -6.15467072e-01 3.86103690e-01 -3.09078302e-02
4.25979674e-01 -4.41170216e-01 -1.21114659e+00 4.98124182e-01
1.70324683e-01 -1.35543895e+00 1.78050205e-01 1.83561310e-01
-7.71434128e-01 -1.08680987e+00 -5.32519639e-01 -6.53081656e-01
5.59360385e-01 1.24693966e+00 3.75465393e-01 -5.35558999e-01
-6.65252626e-01 4.25833404e-01 -5.94035387e-01 -5.95843613e-01
3.53725493e-01 -8.61456573e-01 1.90731421e-01 3.69353682e-01
1.55432774e-02 -7.46311784e-01 -4.38891947e-01 2.60313421e-01
-8.35970998e-01 -2.06378520e-01 1.56087160e+00 8.05871844e-01
3.20957065e-01 3.65216285e-01 7.39252269e-01 -4.11290109e-01
3.57395649e-01 -3.40948671e-01 -9.62725282e-01 2.20552415e-01
-1.21781439e-01 -2.23184958e-01 6.08380675e-01 -3.78292501e-01
-8.38586211e-01 1.15418039e-01 3.51849675e-01 -4.20389265e-01
4.96425442e-02 9.29142118e-01 2.24259049e-02 -8.23975980e-01
5.80567718e-01 7.47987092e-01 -1.17138796e-01 -2.34311178e-01
2.25067034e-01 4.58593696e-01 7.29741037e-01 1.65794060e-01
1.80108356e+00 5.07243931e-01 6.91317797e-01 -1.36275029e+00
-9.97095704e-01 -5.60942709e-01 -3.78030926e-01 -1.51656240e-01
3.24070215e-01 -1.09215629e+00 -4.93958384e-01 7.68678039e-02
-1.04391074e+00 5.60363054e-01 -6.22083014e-03 1.36268795e+00
-3.46839540e-02 4.03656036e-01 1.14876717e-01 -1.12488186e+00
-4.12846893e-01 -3.47552776e-01 7.33077526e-01 2.36472771e-01
2.17132747e-01 -5.50298035e-01 2.00064063e-01 -3.12289558e-02
8.83886755e-01 6.33368969e-01 6.00949883e-01 -5.52110374e-01
-9.74149525e-01 -7.32287228e-01 -1.47199333e-02 -1.50736392e-01
1.43054008e-01 -9.87361968e-01 -1.76023453e-01 -8.33684623e-01
5.87807298e-01 1.30973175e-01 5.00110805e-01 8.53642881e-01
2.56386131e-01 -1.60030246e-01 -7.05357969e-01 7.60540664e-01
1.96715355e+00 6.70617223e-01 2.70947039e-01 -1.02989733e-01
1.84252769e-01 1.98679164e-01 1.08850384e+00 8.27663898e-01
-3.89577121e-01 3.60296786e-01 4.69447017e-01 2.52040058e-01
3.23115140e-01 4.88490611e-01 1.26282300e-03 5.84474564e-01
1.02206856e-01 -4.19954628e-01 -2.88883060e-01 3.56280148e-01
-1.41545618e+00 -1.43257236e+00 -2.12343782e-01 2.31899643e+00
-5.07261939e-02 -2.25030154e-01 -8.03361356e-01 2.65081733e-01
7.55966127e-01 3.31292331e-01 -4.91352528e-01 2.92956442e-01
-2.88209230e-01 7.78756678e-01 8.08924735e-01 4.99702245e-01
-6.82660997e-01 1.37671947e-01 5.18360901e+00 6.61665618e-01
-1.06636441e+00 -7.85864890e-03 -5.70924640e-01 -7.34473243e-02
-2.68260926e-01 5.97099140e-02 -8.53054583e-01 -1.14614107e-02
5.56346714e-01 -4.60223734e-01 2.87332356e-01 3.76464784e-01
-5.82481436e-02 -3.64966720e-01 -6.18937135e-01 1.18056810e+00
4.25603807e-01 -1.18639243e+00 -2.49247313e-01 1.54962420e-01
4.13529307e-01 -3.41744334e-01 9.02861655e-02 1.63164780e-01
-1.00240357e-01 -9.84964669e-01 3.56372118e-01 5.84227979e-01
7.96247721e-01 -7.31452107e-01 6.06967807e-01 7.35733390e-01
-1.42291403e+00 -2.41908401e-01 -4.02661800e-01 -2.81533122e-01
6.39998317e-01 1.43514740e+00 -8.01561713e-01 9.25505280e-01
1.75762191e-01 2.79191196e-01 2.70134304e-02 1.38485980e+00
-6.72866777e-02 5.37649274e-01 -4.18032318e-01 -6.99180886e-02
3.50789160e-01 -4.44541693e-01 1.08305264e+00 8.77503216e-01
1.43737006e+00 7.98646450e-01 5.17265320e-01 3.99630010e-01
5.05265236e-01 -4.16281462e-01 -9.29438293e-01 -1.02188140e-01
1.03919971e+00 1.37603378e+00 -1.96089149e-01 6.33777305e-02
-3.53174448e-01 1.89527139e-01 -5.73547304e-01 3.78381550e-01
-6.38848782e-01 -8.81123185e-01 3.83841783e-01 9.32307467e-02
8.20811510e-01 -7.93969333e-01 4.53217402e-02 -6.58621252e-01
-1.48026571e-01 -6.97719038e-01 7.17564076e-02 -5.55544436e-01
-1.09088326e+00 7.64376819e-01 8.29535425e-02 -1.78483903e+00
-3.40130776e-01 -1.62602454e-01 -3.75890881e-01 1.14046371e+00
-1.66228247e+00 -1.07986259e+00 -6.52345836e-01 5.48020542e-01
2.43022069e-01 -4.61814493e-01 9.14642930e-01 1.15888476e-01
2.54703552e-01 -2.49885842e-02 2.51298994e-01 2.77777966e-02
5.67147732e-01 -3.92245471e-01 -1.69576377e-01 1.13786888e+00
9.95576307e-02 3.61729890e-01 1.06766319e+00 -8.82692099e-01
-1.96429121e+00 -1.08526182e+00 4.66187894e-01 3.90595496e-01
4.24056292e-01 -9.06905979e-02 -3.52504343e-01 5.76637328e-01
2.12206542e-01 2.38198310e-01 8.94798994e-01 -4.78570759e-01
-3.28367770e-01 -1.32673293e-01 -1.10035741e+00 1.17951304e-01
7.59001374e-01 2.48565495e-01 -7.47699916e-01 3.29657197e-01
1.72361434e-01 -3.20368171e-01 -6.42545521e-01 1.06053162e+00
6.09655023e-01 -7.20186353e-01 1.26976264e+00 2.28944480e-01
-1.99199945e-01 -7.49534130e-01 -8.45919013e-01 -1.57952762e+00
-8.79374266e-01 -3.23315561e-01 2.25314312e-02 4.31925774e-01
-1.22972950e-01 -9.49134707e-01 4.57252651e-01 -9.86219227e-01
-2.39284337e-01 -4.30920750e-01 -1.11771905e+00 -8.77476990e-01
-4.84903008e-01 -1.20683752e-01 7.51625225e-02 5.93384922e-01
-3.64159822e-01 5.18134654e-01 -6.56538963e-01 1.24094248e+00
1.56026602e+00 8.92354429e-01 4.76771027e-01 -1.72646952e+00
-5.90944111e-01 1.68011218e-01 -4.85256970e-01 -1.02721679e+00
-2.85070866e-01 -6.61429167e-01 2.04401925e-01 -1.50552928e+00
-4.27234083e-01 -7.01266766e-01 2.04695135e-01 -2.86598086e-01
2.15967163e-01 3.61072987e-01 1.64320543e-01 2.54967213e-01
-1.81822494e-01 4.98397261e-01 1.10096705e+00 -8.83879662e-02
3.33598047e-01 5.41611671e-01 -5.90381980e-01 4.05485034e-01
5.14539599e-01 -6.46442235e-01 -1.87996641e-01 -5.21924913e-01
-2.89470665e-02 9.63064492e-01 3.18801075e-01 -1.69042480e+00
7.09623814e-01 -3.76312137e-01 8.37372780e-01 -1.15186608e+00
8.19868326e-01 -1.05614734e+00 6.35292172e-01 9.67744887e-01
3.31562817e-01 -8.65403637e-02 -4.43738662e-02 9.82870400e-01
-5.70180893e-01 -3.55667651e-01 9.27361667e-01 4.42061499e-02
-6.31715894e-01 1.98177949e-01 -4.38790768e-01 -1.85837895e-01
1.38223624e+00 -2.49838650e-01 -4.16019827e-01 -3.82454038e-01
-4.73478228e-01 -5.22190472e-03 -2.44280592e-01 -1.83426127e-01
1.09906328e+00 -1.18155444e+00 -1.15150023e+00 5.47678173e-01
7.83104077e-02 -3.07287931e-01 5.80673575e-01 8.52765501e-01
-2.93098599e-01 8.62308145e-01 -3.25245708e-01 -5.56961060e-01
-1.27388930e+00 1.63726464e-01 -3.22457701e-02 -7.44148642e-02
-4.83369797e-01 5.77400506e-01 -5.54004833e-02 8.26250389e-02
-1.49674460e-01 3.65528017e-01 -1.30561784e-01 -1.63834944e-01
9.57377791e-01 1.58855528e-01 -6.48975521e-02 -4.57749277e-01
-5.37597775e-01 1.08349836e+00 3.57922882e-01 -4.07083809e-01
1.53380287e+00 2.10129037e-01 -1.23725690e-01 8.44365731e-02
5.75669765e-01 8.20475519e-01 -6.98131919e-01 -5.42665780e-01
-2.16052741e-01 -8.98032188e-01 9.80020761e-02 -5.82713068e-01
-6.17813230e-01 7.05992937e-01 5.17158806e-01 9.06880125e-02
1.15507293e+00 -2.77828693e-01 3.65382284e-01 6.96995795e-01
9.51111197e-01 -1.50505796e-01 7.25070164e-02 5.37670612e-01
7.82699823e-01 -6.42183363e-01 4.26855147e-01 -5.88627756e-01
-2.04746783e-01 1.23628092e+00 2.46181954e-02 -4.98244226e-01
6.45940542e-01 7.97332644e-01 -4.02331911e-02 -2.66197920e-01
-4.50310647e-01 -2.25880861e-01 4.84088659e-02 6.86594307e-01
4.66667376e-02 2.73763299e-01 -2.67340958e-01 1.69662073e-01
-1.04619734e-01 -2.26821870e-01 4.95621890e-01 1.27117109e+00
-1.31248057e+00 -8.58292878e-01 -1.18804777e+00 6.46203041e-01
-6.69480786e-02 -1.43151253e-01 1.76824078e-01 5.84327877e-01
-2.18475118e-01 1.24661100e+00 -9.11944546e-03 -2.42015898e-01
5.87655902e-01 -2.01524511e-01 8.43904376e-01 -7.69118905e-01
5.37322536e-02 3.58730018e-01 -1.96023628e-01 7.25079849e-02
-2.11915001e-01 -7.40906715e-01 -7.20758379e-01 1.26058966e-01
-4.77655917e-01 7.19215572e-01 7.11349249e-01 8.13243687e-01
2.12707490e-01 6.20704234e-01 9.24414456e-01 -7.66179860e-01
-1.01366258e+00 -1.06065679e+00 -9.37137306e-01 -5.05189478e-01
3.84035736e-01 -7.90060759e-01 -7.23442614e-01 -4.70738798e-01] | [6.631717205047607, 1.144851565361023] |
d1a579a5-9d49-4268-a324-7c0db8bc2bf2 | 3d-human-mesh-regression-with-dense-1 | 2006.05734 | null | https://arxiv.org/abs/2006.05734v2 | https://arxiv.org/pdf/2006.05734v2.pdf | 3D Human Mesh Regression with Dense Correspondence | Estimating 3D mesh of the human body from a single 2D image is an important task with many applications such as augmented reality and Human-Robot interaction. However, prior works reconstructed 3D mesh from global image feature extracted by using convolutional neural network (CNN), where the dense correspondences between the mesh surface and the image pixels are missing, leading to suboptimal solution. This paper proposes a model-free 3D human mesh estimation framework, named DecoMR, which explicitly establishes the dense correspondence between the mesh and the local image features in the UV space (i.e. a 2D space used for texture mapping of 3D mesh). DecoMR first predicts pixel-to-surface dense correspondence map (i.e., IUV image), with which we transfer local features from the image space to the UV space. Then the transferred local image features are processed in the UV space to regress a location map, which is well aligned with transferred features. Finally we reconstruct 3D human mesh from the regressed location map with a predefined mapping function. We also observe that the existing discontinuous UV map are unfriendly to the learning of network. Therefore, we propose a novel UV map that maintains most of the neighboring relations on the original mesh surface. Experiments demonstrate that our proposed local feature alignment and continuous UV map outperforms existing 3D mesh based methods on multiple public benchmarks. Code will be made available at https://github.com/zengwang430521/DecoMR | ['Wang Zeng', 'Wentao Liu', 'Ping Luo', 'Wanli Ouyang', 'Xiaogang Wang'] | 2020-06-10 | 3d-human-mesh-regression-with-dense | http://openaccess.thecvf.com/content_CVPR_2020/html/Zeng_3D_Human_Mesh_Regression_With_Dense_Correspondence_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Zeng_3D_Human_Mesh_Regression_With_Dense_Correspondence_CVPR_2020_paper.pdf | cvpr-2020-6 | ['3d-human-reconstruction'] | ['computer-vision'] | [ 1.90817267e-01 4.64657359e-02 -1.65407181e-01 -3.54597688e-01
-5.29747844e-01 3.15944105e-02 2.29831472e-01 -2.67715245e-01
-1.62247658e-01 4.66089100e-01 -3.80299725e-02 2.84355134e-01
1.24843009e-02 -1.20019031e+00 -1.26609552e+00 -2.92555600e-01
3.06854606e-01 7.00505614e-01 2.60272920e-01 -2.52138346e-01
-1.77523997e-02 5.38683534e-01 -1.67178631e+00 1.93154112e-01
4.38425720e-01 1.29054797e+00 2.41314054e-01 2.98130602e-01
-1.94938064e-01 2.63692528e-01 -7.16596544e-02 1.24029271e-01
5.03669024e-01 -4.32400823e-01 -8.48123133e-01 2.13752389e-01
8.01174045e-01 -3.75511765e-01 -3.59265894e-01 1.08360088e+00
3.25642705e-01 -1.19767614e-01 7.26075649e-01 -9.63670015e-01
-5.99425197e-01 -6.18611574e-02 -8.01692128e-01 -3.94327015e-01
4.43512261e-01 -1.99006706e-01 6.78960383e-01 -1.21041811e+00
1.09148204e+00 1.41462004e+00 9.03176010e-01 4.16549265e-01
-1.01680422e+00 -6.59291387e-01 -1.06846452e-01 -1.20222926e-01
-1.70110333e+00 -9.00578778e-03 1.05846334e+00 -5.56087971e-01
6.37462616e-01 2.24928796e-01 9.03483212e-01 7.57084906e-01
5.23947597e-01 4.73179251e-01 9.67846990e-01 -4.28224236e-01
-1.71581537e-01 -2.28883103e-01 -2.38076061e-01 1.29692304e+00
-2.40060501e-02 1.15017414e-01 -5.85322618e-01 -7.96413701e-03
1.75187182e+00 1.20961867e-01 -3.88315797e-01 -7.16927826e-01
-1.44754982e+00 5.45971811e-01 6.90904379e-01 7.32581271e-03
-5.03916562e-01 1.87366754e-01 1.97343051e-01 3.31312150e-01
8.01171660e-01 -8.94863680e-02 -6.10135138e-01 1.00704513e-01
-4.22915608e-01 2.50912398e-01 4.86474961e-01 9.62462783e-01
1.28210390e+00 -2.96528965e-01 1.73766688e-01 9.14667368e-01
3.81471813e-01 7.64750481e-01 1.04109414e-01 -1.11863136e+00
3.71961743e-01 8.90651882e-01 -4.09236774e-02 -1.31195092e+00
-4.50194508e-01 -1.04107156e-01 -1.07857478e+00 4.07765985e-01
4.09377307e-01 1.57497197e-01 -8.98076534e-01 1.46576238e+00
7.92590916e-01 3.09604019e-01 -4.35990542e-01 1.26108205e+00
1.00296569e+00 5.09811223e-01 -3.75029117e-01 2.64578551e-01
1.32422876e+00 -9.37784135e-01 -4.45738018e-01 -1.21294044e-01
2.89437771e-01 -6.22974873e-01 1.13851428e+00 1.25597373e-01
-1.17963493e+00 -7.85920084e-01 -1.01591861e+00 -4.03672427e-01
-1.14939608e-01 1.44718319e-01 2.46645346e-01 6.02668151e-02
-9.44934547e-01 7.25434363e-01 -9.61507440e-01 -3.64769399e-01
3.77733737e-01 4.60375458e-01 -7.19462931e-01 2.55008787e-02
-1.09060585e+00 6.36398733e-01 1.61490306e-01 3.78961653e-01
-5.00944436e-01 -7.45447040e-01 -1.26524782e+00 -3.72416914e-01
2.79691637e-01 -1.07590365e+00 9.52990711e-01 -8.41592371e-01
-1.60294604e+00 1.27343988e+00 -1.73440382e-01 9.99824181e-02
7.64879942e-01 -2.04470351e-01 1.25999525e-01 1.33346692e-01
2.22805247e-01 7.04590321e-01 9.15812314e-01 -1.39656413e+00
-5.62763870e-01 -4.70200092e-01 -1.04992114e-01 2.33472914e-01
6.10804521e-02 -3.77249449e-01 -7.55812764e-01 -6.13579452e-01
6.22814655e-01 -8.80293071e-01 8.03295001e-02 6.16715014e-01
-4.37902600e-01 -2.73972064e-01 6.95870996e-01 -7.43089259e-01
7.33935535e-01 -1.86856842e+00 3.50028247e-01 5.66095948e-01
2.96442956e-01 -3.80228251e-01 -1.28475457e-01 -2.28791535e-02
-1.40114903e-01 -1.81492805e-01 -3.51467043e-01 -5.42933106e-01
-1.41632691e-01 1.20939672e-01 2.54224449e-01 8.51585925e-01
9.01961178e-02 1.13820529e+00 -7.31441081e-01 -7.93835163e-01
5.58484137e-01 7.75736809e-01 -5.19617975e-01 2.68111676e-01
-9.72872004e-02 7.67872274e-01 -4.74747419e-01 8.27635586e-01
8.31191003e-01 -3.89206290e-01 -1.16936274e-01 -6.37527943e-01
-1.16341589e-02 -1.33185223e-01 -1.17098641e+00 2.19728827e+00
-6.57776594e-01 1.75423011e-01 1.30438522e-01 -8.85890305e-01
1.11642277e+00 2.23139852e-01 8.67640495e-01 -7.77455091e-01
4.45611298e-01 2.53500968e-01 -6.67779207e-01 -3.53468716e-01
1.34045765e-01 -1.27424905e-02 7.21707642e-02 2.61811227e-01
-8.14465433e-02 -4.79177758e-02 -6.22124314e-01 -3.64604533e-01
7.70289123e-01 5.71465790e-01 1.25168636e-01 -2.74830103e-01
5.03298402e-01 4.27340791e-02 5.51860154e-01 3.17509770e-01
1.52561665e-01 1.06880784e+00 1.05849527e-01 -7.53867149e-01
-1.15638399e+00 -1.22998798e+00 -3.00461829e-01 3.77422869e-01
6.39139414e-01 -3.04555684e-01 -7.97955215e-01 -5.58663249e-01
1.93216577e-01 -2.49369636e-01 -9.85256195e-01 -4.35160426e-03
-9.38833058e-01 8.19234401e-02 9.13379714e-02 4.77665216e-01
7.96539187e-01 -1.16508698e+00 -5.33821523e-01 -1.81973982e-03
-3.07980895e-01 -9.46942985e-01 -7.12249577e-01 -1.87044576e-01
-8.82155180e-01 -1.14227331e+00 -8.80131185e-01 -1.11118388e+00
1.04119861e+00 1.52461499e-01 1.13400054e+00 5.54195464e-01
-3.78386855e-01 4.30609286e-01 -5.40753789e-02 -9.79346130e-03
-1.96549729e-01 3.00862901e-02 8.59491602e-02 5.95557727e-02
9.47768539e-02 -5.97078264e-01 -7.67907619e-01 5.86215615e-01
-5.69020391e-01 5.16095102e-01 3.70643467e-01 9.83847082e-01
1.28786302e+00 1.79839600e-02 1.59011379e-01 -8.88803542e-01
7.66829923e-02 -3.57435018e-01 -5.37986934e-01 4.15032320e-02
-1.63630888e-01 -2.22766444e-01 9.52550843e-02 -4.36382383e-01
-8.81013274e-01 4.52032983e-01 -3.26500416e-01 -8.73018086e-01
-2.28012815e-01 3.79075557e-01 -2.52387285e-01 -2.34040931e-01
5.30689061e-01 5.65193743e-02 4.69013840e-01 -5.89796841e-01
2.81779747e-02 5.99295437e-01 4.76572692e-01 -7.04295874e-01
8.41648817e-01 6.98076189e-01 1.74675137e-01 -8.42325866e-01
-8.28933299e-01 -3.27509850e-01 -1.00590289e+00 -3.73731285e-01
1.07410729e+00 -9.97426748e-01 -6.12559378e-01 5.65692723e-01
-1.27199960e+00 -5.61185122e-01 -2.49239281e-01 4.40305859e-01
-8.56248260e-01 3.42639357e-01 -8.82574856e-01 -2.68802881e-01
-4.12530571e-01 -1.16912901e+00 1.68754268e+00 -3.62653919e-02
-3.80793989e-01 -9.16859210e-01 -3.04115526e-02 4.03314561e-01
-7.79675767e-02 6.11221492e-01 8.22230875e-01 3.89972270e-01
-6.18062079e-01 -1.21096633e-01 -2.47693703e-01 3.29332650e-01
4.37703609e-01 -3.04625154e-01 -7.37792969e-01 -2.57089771e-02
-3.00636142e-02 -1.76429018e-01 4.43125278e-01 6.09710574e-01
1.22051466e+00 6.38092384e-02 -3.19327623e-01 8.71192753e-01
1.40826321e+00 -2.61724979e-01 5.95009685e-01 4.77061391e-01
1.12890279e+00 5.74723005e-01 8.21696699e-01 2.54025906e-01
4.57857549e-01 8.76815915e-01 4.60675478e-01 -4.74618018e-01
-3.58505964e-01 -5.16977012e-01 5.43726124e-02 8.79882991e-01
-3.61151338e-01 3.26002449e-01 -8.40180278e-01 2.49508232e-01
-1.93500316e+00 -2.85172433e-01 -2.02573508e-01 2.25808811e+00
8.46990526e-01 7.12340996e-02 -1.11488216e-01 -2.60155171e-01
8.26139510e-01 -1.27107918e-01 -4.88744408e-01 2.77149379e-02
9.49383825e-02 4.39904988e-01 3.80904347e-01 6.51905239e-01
-1.05181623e+00 9.41375434e-01 5.08893347e+00 7.42155671e-01
-1.04913473e+00 1.66885585e-01 4.06672448e-01 2.16613978e-01
-6.64754212e-02 -3.52820605e-01 -4.84978378e-01 3.67530733e-01
2.59956360e-01 2.62264669e-01 2.17381254e-01 6.91197038e-01
9.23458859e-02 1.25253722e-01 -1.23414803e+00 1.42814422e+00
3.89362536e-02 -1.38793910e+00 1.59218878e-01 3.54070604e-01
7.96899021e-01 -1.56798869e-01 -1.13759778e-01 -1.24220639e-01
-1.40832558e-01 -9.97439563e-01 9.10966456e-01 8.79322588e-01
1.19881046e+00 -6.74001336e-01 7.60138273e-01 3.18612069e-01
-1.54887021e+00 5.56110740e-01 -5.58692455e-01 -3.53703834e-02
1.19413354e-01 5.52363634e-01 -4.82215375e-01 6.11393631e-01
9.77732062e-01 9.04963791e-01 -2.87296027e-01 6.55227244e-01
-5.33743091e-02 5.07317670e-02 -3.67062807e-01 4.34643686e-01
-2.20985293e-01 -4.29723233e-01 3.42689931e-01 6.57007396e-01
4.61420029e-01 9.68621895e-02 3.68586004e-01 1.14965093e+00
-2.83347785e-01 3.13371152e-01 -7.45355546e-01 5.66914618e-01
3.38782400e-01 1.07262778e+00 -7.52379954e-01 -2.84532100e-01
-3.53458941e-01 1.15013683e+00 3.33972782e-01 1.64696500e-01
-7.47135937e-01 -8.20923150e-02 6.50039852e-01 5.29447258e-01
1.21412249e-02 -2.58592814e-01 -4.21490312e-01 -1.04591799e+00
3.15079898e-01 -5.03720403e-01 7.03327954e-02 -8.78530145e-01
-1.32075644e+00 5.12641668e-01 -1.10700622e-01 -1.55755913e+00
1.35075375e-01 -5.13818145e-01 -3.29661906e-01 9.76661146e-01
-1.29621267e+00 -1.35865211e+00 -6.97964430e-01 7.12096691e-01
5.18287182e-01 1.79554433e-01 7.06067204e-01 3.33639681e-01
-1.75964862e-01 4.83178854e-01 -3.52662802e-01 3.54497343e-01
7.19621956e-01 -9.64902580e-01 5.09981453e-01 9.46831331e-02
-1.34115934e-01 4.03598905e-01 1.52282640e-01 -1.13631594e+00
-1.54135394e+00 -1.26538754e+00 5.93274891e-01 -5.62374711e-01
1.70558721e-01 -3.78168851e-01 -9.06179667e-01 7.30000377e-01
-2.17801839e-01 3.41073543e-01 4.36263531e-02 -1.92086607e-01
-7.16755986e-02 1.18598517e-03 -1.14518309e+00 5.28680027e-01
1.41116118e+00 -5.02499342e-01 -4.19622511e-01 2.79830307e-01
6.53145611e-01 -1.14759791e+00 -1.33674264e+00 5.98923266e-01
7.69782245e-01 -7.86603391e-01 1.29106712e+00 -1.30352274e-01
6.88471735e-01 -3.75013292e-01 -2.30630830e-01 -1.00023544e+00
-3.13269347e-01 -8.01408663e-03 -8.69746599e-03 8.21748555e-01
7.98119530e-02 -5.43282568e-01 1.03829992e+00 3.45804483e-01
-3.02591592e-01 -1.06906998e+00 -1.08496988e+00 -5.50921142e-01
9.74583402e-02 -2.09246516e-01 6.94242179e-01 9.95020986e-01
-5.10459960e-01 5.49425557e-02 -3.42476130e-01 2.90773779e-01
7.47034848e-01 3.46571475e-01 9.46982265e-01 -1.37863326e+00
1.46955833e-01 -1.83959022e-01 -6.41983151e-01 -1.27220428e+00
2.48407990e-01 -1.04966164e+00 2.30947956e-01 -1.58137333e+00
5.56827709e-02 -6.12513542e-01 -1.65380597e-01 4.40967888e-01
9.20754820e-02 7.11653173e-01 -1.86322495e-01 2.77810335e-01
-2.58328229e-01 7.43281364e-01 1.81328821e+00 -2.09313601e-01
-2.09150031e-01 -2.50588089e-01 -5.64325973e-02 9.59003150e-01
7.32343972e-01 -3.05875003e-01 -8.30154568e-02 -5.60078800e-01
8.72659907e-02 7.39523023e-02 7.10005879e-01 -1.00127530e+00
1.30331770e-01 -2.16599882e-01 7.17173696e-01 -7.11718440e-01
6.07584417e-01 -1.01822877e+00 4.71437871e-01 2.58039773e-01
-7.85547122e-03 -6.59785643e-02 -1.42620787e-01 4.33298886e-01
-8.27214718e-02 1.21148571e-01 8.08015347e-01 -3.91624689e-01
-6.40934229e-01 8.57987940e-01 4.01449591e-01 -1.58214033e-01
8.92722130e-01 -4.92360741e-01 1.33379951e-01 -7.76247829e-02
-7.67705679e-01 1.20385490e-01 9.48655784e-01 5.42966902e-01
1.10794365e+00 -1.76979077e+00 -6.81293249e-01 5.27474761e-01
-1.19115869e-02 6.96505249e-01 4.57002550e-01 8.59887302e-01
-9.56040561e-01 -1.47728324e-02 -3.96400601e-01 -1.12950385e+00
-1.32280219e+00 2.04232708e-01 6.63197696e-01 2.25387022e-01
-1.40964007e+00 6.05830491e-01 4.71930057e-01 -8.91673923e-01
2.30808586e-01 -4.70673621e-01 1.57338947e-01 -4.23112810e-01
2.08971381e-01 1.49656862e-01 1.40845522e-01 -1.00995696e+00
-2.39490762e-01 1.47125077e+00 2.96924323e-01 1.88017353e-01
1.27612007e+00 -2.16402337e-01 -4.94972140e-01 4.48712885e-01
1.64177990e+00 -1.52215540e-01 -1.24851871e+00 -4.78394300e-01
-4.68806893e-01 -6.81157112e-01 6.16622120e-02 -1.75798684e-01
-1.41148591e+00 8.11607420e-01 8.79300952e-01 -3.47517073e-01
8.62721741e-01 3.00215364e-01 1.13646686e+00 1.49257213e-01
7.49099731e-01 -9.17339206e-01 1.75752062e-02 4.13452089e-01
1.07777917e+00 -1.24194145e+00 1.47516593e-01 -8.05115640e-01
-1.62375420e-01 1.11512804e+00 8.76351535e-01 -3.72984082e-01
1.02920711e+00 3.26565616e-02 1.18972048e-01 -6.97537482e-01
-1.38804987e-01 8.76136422e-02 5.09140670e-01 5.92889845e-01
1.38471022e-01 1.07742198e-01 3.82353924e-02 3.58805925e-01
-3.00634682e-01 6.92751184e-02 -5.67774437e-02 9.49419618e-01
-2.75243998e-01 -9.27212119e-01 -5.74710727e-01 4.56374615e-01
4.55182418e-03 2.71226496e-01 -1.47113889e-01 8.35437775e-01
5.06963372e-01 3.21066648e-01 3.75729412e-01 -5.50478637e-01
6.62443399e-01 -2.73076028e-01 6.41085088e-01 -6.61939383e-01
-1.63631931e-01 1.75322846e-01 -1.63047493e-01 -8.62806916e-01
-5.20520270e-01 -3.85137886e-01 -1.50141978e+00 -2.28385553e-01
-1.30873784e-01 -2.24101901e-01 5.22698283e-01 7.80009031e-01
2.44264513e-01 4.22846705e-01 5.09824336e-01 -1.34276795e+00
1.25206765e-02 -7.13407934e-01 -7.91093230e-01 5.19966662e-01
2.66691029e-01 -1.19076812e+00 -1.08192876e-01 1.74352407e-01] | [7.124678134918213, -1.329338788986206] |
52a3eb2e-3bc6-42d1-9c7a-da2dd5bc58bf | towards-unsupervised-text-classification | null | null | https://aclanthology.org/P19-1036 | https://aclanthology.org/P19-1036.pdf | Towards Unsupervised Text Classification Leveraging Experts and Word Embeddings | Text classification aims at mapping documents into a set of predefined categories. Supervised machine learning models have shown great success in this area but they require a large number of labeled documents to reach adequate accuracy. This is particularly true when the number of target categories is in the tens or the hundreds. In this work, we explore an unsupervised approach to classify documents into categories simply described by a label. The proposed method is inspired by the way a human proceeds in this situation: It draws on textual similarity between the most relevant words in each document and a dictionary of keywords for each category reflecting its semantics and lexical field. The novelty of our method hinges on the enrichment of the category labels through a combination of human expertise and language models, both generic and domain specific. Our experiments on 5 standard corpora show that the proposed method increases F1-score over relying solely on human expertise and can also be on par with simple supervised approaches. It thus provides a practical alternative to situations where low cost text categorization is needed, as we illustrate with our application to operational risk incidents classification. | ["L{\\'e}a A. Deleris", 'Zied Haj-Yahia', 'Adrien Sieg'] | 2019-07-01 | null | null | null | acl-2019-7 | ['unsupervised-text-classification'] | ['natural-language-processing'] | [ 2.13626236e-01 2.59909838e-01 -2.47772276e-01 -5.34451962e-01
-5.03857076e-01 -7.52354145e-01 1.13984597e+00 9.89501178e-01
-6.39240921e-01 4.83213633e-01 2.81367242e-01 -4.20999855e-01
-4.39951926e-01 -7.97708333e-01 -3.39945522e-03 -4.23764825e-01
2.62857676e-01 8.68738115e-01 2.75960952e-01 -4.47759598e-01
6.34874225e-01 4.04764146e-01 -1.79458821e+00 2.57074535e-01
8.10445547e-01 1.02882946e+00 1.51343897e-01 1.02979980e-01
-7.12426364e-01 8.69147837e-01 -5.21142006e-01 -3.68831962e-01
-1.22734569e-02 -2.91745961e-01 -1.16539824e+00 1.77445993e-01
6.40622228e-02 2.10473940e-01 3.59047294e-01 9.59674716e-01
2.32112020e-01 1.45205483e-01 1.21214378e+00 -1.01448500e+00
-8.90821069e-02 7.24241436e-01 -2.71957576e-01 1.24021694e-01
4.53842461e-01 -5.60891807e-01 1.07220197e+00 -8.33555460e-01
5.43173790e-01 9.47660744e-01 7.54606843e-01 2.57089555e-01
-1.21299613e+00 -3.84096473e-01 2.84070075e-01 1.29069343e-01
-1.28371227e+00 -1.20833233e-01 6.04324996e-01 -9.59131479e-01
9.56847727e-01 4.57364582e-02 1.56133205e-01 6.98712766e-01
-3.44042294e-02 -3.18811126e-02 1.18954837e+00 -9.35838223e-01
4.91026908e-01 8.44083726e-01 7.22102046e-01 3.09077859e-01
3.51573527e-01 -3.38567108e-01 -1.98441237e-01 -2.33671367e-01
3.11898720e-02 8.24404731e-02 -4.11494300e-02 -4.59207773e-01
-1.01938820e+00 1.09749389e+00 2.74163514e-01 1.13006234e+00
-3.82922500e-01 -3.13340306e-01 7.46485412e-01 2.11606741e-01
6.38391018e-01 6.71287239e-01 -1.91367269e-01 1.41304180e-01
-1.06570244e+00 6.50155693e-02 1.00527883e+00 7.59841919e-01
8.05306196e-01 -4.99090075e-01 -4.12925519e-02 1.01210308e+00
1.95428520e-01 1.59061179e-01 6.80455744e-01 -4.59463120e-01
2.93704480e-01 9.38965917e-01 2.14231461e-01 -1.28980100e+00
-5.95389366e-01 -3.74521464e-01 -5.70322096e-01 3.93966921e-02
4.08417165e-01 1.13239214e-01 -6.80158675e-01 1.48524356e+00
2.20930070e-01 -5.11158049e-01 1.35100678e-01 5.25712729e-01
7.67681599e-01 5.34572184e-01 4.51020002e-01 -4.05720919e-01
1.53477979e+00 -7.09042668e-01 -7.29368091e-01 -2.37488344e-01
7.59960771e-01 -7.83053219e-01 9.28310752e-01 4.25763696e-01
-6.47825122e-01 -5.73919356e-01 -7.62581468e-01 1.87272131e-01
-9.81291711e-01 1.79717913e-01 4.39308226e-01 8.26846242e-01
-8.64689350e-01 5.10042429e-01 -1.78022712e-01 -1.06308961e+00
-4.91132541e-03 3.97064269e-01 -3.45971674e-01 2.78720230e-01
-1.23015118e+00 1.14124644e+00 7.62712777e-01 -1.88094169e-01
-1.52689874e-01 -2.07650632e-01 -6.15776837e-01 1.55825064e-01
3.86323392e-01 -3.28830868e-01 1.17072356e+00 -1.06110811e+00
-1.03722918e+00 1.24118960e+00 1.26710400e-01 -3.59385729e-01
3.99748951e-01 7.41179734e-02 -5.62142193e-01 3.59484047e-01
3.73191833e-01 3.14587712e-01 6.39097273e-01 -1.32536197e+00
-1.02527547e+00 -3.76240134e-01 1.33871347e-01 -2.38346728e-03
-8.14889014e-01 3.45605493e-01 -5.58703160e-03 -6.72813118e-01
3.46412174e-02 -7.61923432e-01 -1.94524065e-01 -4.29441124e-01
-1.59519717e-01 -7.32942522e-01 4.11973327e-01 -3.49502712e-01
1.49494672e+00 -1.88852215e+00 -8.07705987e-03 6.24103785e-01
3.21054637e-01 1.50019348e-01 5.22294104e-01 1.00952876e+00
-1.25874713e-01 3.40344310e-02 -2.03804001e-01 -3.34673189e-02
2.06309810e-01 4.35174555e-02 -5.29605448e-01 2.81324178e-01
-1.87500179e-01 2.70617872e-01 -1.01907754e+00 -7.48046517e-01
5.91180995e-02 1.41034022e-01 -3.05724382e-01 4.04084772e-02
2.87768394e-02 1.28366023e-01 -5.49119592e-01 1.41391993e-01
1.89769730e-01 -1.56854987e-01 4.19562191e-01 -5.14505506e-02
-2.70159036e-01 4.30228174e-01 -1.13191664e+00 1.25281513e+00
-6.81946635e-01 3.00224751e-01 -3.59731674e-01 -1.58515847e+00
1.13564193e+00 4.80077326e-01 3.43348801e-01 -3.35916549e-01
3.48993838e-01 5.05561948e-01 -1.38901949e-01 -4.67676729e-01
4.69241560e-01 -5.18287599e-01 -3.63115698e-01 6.62140131e-01
1.14937119e-01 -6.06489368e-02 4.82862592e-01 1.67860821e-01
7.15874553e-01 -1.87335357e-01 8.86161864e-01 -7.74453998e-01
8.28579009e-01 2.48690620e-01 5.69451079e-02 7.69706428e-01
2.18568295e-01 1.07227847e-01 4.99834448e-01 -4.47613835e-01
-8.72443557e-01 -6.91307843e-01 -3.93606663e-01 1.06243110e+00
-3.16846184e-02 -5.59162021e-01 -9.24801528e-01 -9.08971965e-01
-1.35261536e-01 9.10209894e-01 -7.24772453e-01 -1.24761350e-01
-2.31572464e-01 -3.97090673e-01 2.78614283e-01 3.82955670e-01
1.52463019e-01 -1.09116411e+00 -6.94977403e-01 3.76125872e-01
-1.49919257e-01 -9.49802041e-01 1.91669650e-02 4.47410882e-01
-6.82087719e-01 -9.66950417e-01 -5.56055725e-01 -8.90177071e-01
8.22423279e-01 1.14305198e-01 1.14460278e+00 3.00318561e-02
-4.81072590e-02 5.27144611e-01 -8.56079698e-01 -6.66742146e-01
-7.82591820e-01 3.82071376e-01 1.72434658e-01 1.77940369e-01
9.20938432e-01 -4.08810407e-01 -3.01880211e-01 3.87033790e-01
-1.01116037e+00 -2.69278258e-01 2.73686320e-01 7.28388011e-01
8.29871520e-02 4.02352780e-01 8.09470534e-01 -1.28549683e+00
7.19349742e-01 -6.64940119e-01 -2.75465190e-01 3.21953923e-01
-9.92681623e-01 -1.71590073e-03 8.42948616e-01 -2.77716517e-01
-1.04026461e+00 -5.54465093e-02 -1.20374128e-01 4.81270462e-01
-4.87709105e-01 6.37858212e-01 8.03060159e-02 8.21806863e-02
1.00699544e+00 -1.39998952e-02 -2.11928457e-01 -5.59996068e-01
1.38470441e-01 1.22036517e+00 2.49751285e-01 -4.97760236e-01
8.18958759e-01 3.30690861e-01 -1.34016708e-01 -6.35301113e-01
-9.36681449e-01 -9.93507981e-01 -9.50188518e-01 -3.53923291e-01
7.98767686e-01 -4.30806994e-01 -4.73896325e-01 4.19079587e-02
-9.51736927e-01 4.19382825e-02 -3.28080475e-01 4.89347756e-01
-4.62757468e-01 5.40128112e-01 -2.27826282e-01 -6.74864113e-01
-3.15557301e-01 -7.08167791e-01 8.13169003e-01 -1.07117973e-01
-6.88754261e-01 -1.29063582e+00 -6.50461093e-02 2.38508940e-01
3.45636576e-01 -1.78556107e-02 1.25523698e+00 -1.39248037e+00
3.98904681e-01 -6.16843939e-01 -3.89882140e-02 3.55680346e-01
3.30849409e-01 -4.79072690e-01 -8.58348191e-01 -1.74935102e-01
7.07148463e-02 -2.21948072e-01 7.53909230e-01 -1.44681841e-01
8.60553801e-01 -1.12958260e-01 -5.88168979e-01 -3.16947520e-01
1.49807811e+00 4.29036051e-01 2.74765998e-01 7.21691012e-01
4.23704892e-01 1.26558840e+00 8.73729408e-01 3.37639153e-01
5.91794997e-02 7.73499668e-01 -5.16118295e-02 -1.35426328e-01
3.22927952e-01 -7.46386722e-02 -1.48635820e-01 6.53846443e-01
-8.72707963e-02 -2.26509616e-01 -1.29809022e+00 6.21276021e-01
-1.71646464e+00 -1.04150891e+00 -4.87484708e-02 2.37121248e+00
7.34259248e-01 4.30481106e-01 4.15888995e-01 7.73707211e-01
8.31085265e-01 -2.12682933e-01 1.60868749e-01 -6.79151297e-01
2.96743810e-01 1.36765957e-01 2.24206984e-01 3.97394359e-01
-1.05041254e+00 6.76434040e-01 5.78519869e+00 9.32489276e-01
-8.43909621e-01 1.11791931e-01 4.19765115e-01 5.79666734e-01
-1.10453479e-01 -6.28952682e-02 -9.39125657e-01 4.62065786e-01
1.02524555e+00 -3.59446377e-01 -8.23444575e-02 8.03612709e-01
2.94197462e-02 -1.78628594e-01 -1.09532988e+00 7.02691197e-01
3.10602665e-01 -9.90200698e-01 1.67663425e-01 1.06431112e-01
3.55042487e-01 -6.42852902e-01 -1.68882549e-01 2.44058967e-01
-1.03529254e-02 -8.67121875e-01 8.28725576e-01 2.55998015e-01
6.51485801e-01 -7.47372627e-01 1.02017260e+00 4.93047416e-01
-9.59079683e-01 -3.98975730e-01 -3.30678612e-01 -1.53726399e-01
-7.13849952e-03 4.39386606e-01 -9.15963948e-01 5.53365469e-01
5.86461425e-01 4.40866232e-01 -4.37424570e-01 8.19159031e-01
-3.05575896e-02 5.98957419e-01 -1.29285097e-01 -2.02293247e-01
5.67839921e-01 -1.69533312e-01 2.25306526e-01 1.72992635e+00
3.42888743e-01 1.25400364e-01 2.77866662e-01 3.19622099e-01
3.45673651e-01 8.84241879e-01 -7.02533841e-01 3.97698209e-02
3.12037170e-01 1.32331419e+00 -1.47313356e+00 -5.50677538e-01
-4.64542747e-01 6.14145517e-01 1.41283303e-01 -6.98908120e-02
-4.05276060e-01 -8.04758132e-01 -7.11088628e-02 4.69953597e-01
1.27031028e-01 1.98540136e-01 -9.79589075e-02 -7.64133930e-01
-8.00475329e-02 -7.63712883e-01 6.85959995e-01 -3.80930364e-01
-1.37599814e+00 1.01935279e+00 3.95660460e-01 -1.51283836e+00
-4.67599988e-01 -7.75701284e-01 -2.58385628e-01 6.35411799e-01
-1.26088440e+00 -1.02332151e+00 -3.32137376e-01 3.96528959e-01
4.85075742e-01 -3.24528545e-01 1.08857536e+00 4.70297903e-01
-2.57613640e-02 4.06080782e-01 9.12850946e-02 -2.01459341e-02
8.06521952e-01 -1.40091145e+00 -4.27099615e-02 5.61831713e-01
7.65551776e-02 7.59795547e-01 9.82621849e-01 -4.59015191e-01
-3.33671957e-01 -7.92045236e-01 1.66169190e+00 -3.95261824e-01
8.80593956e-01 -3.82403761e-01 -8.24632943e-01 3.35669667e-01
2.39375919e-01 -5.08050442e-01 1.03811014e+00 1.84418663e-01
-4.65251237e-01 -2.03867823e-01 -1.16598797e+00 3.12702686e-01
7.89746940e-01 -4.57659304e-01 -1.05754483e+00 6.87304199e-01
1.74106181e-01 1.94288447e-01 -7.00982749e-01 1.11867182e-01
2.18626544e-01 -8.95886242e-01 7.03420699e-01 -6.00385427e-01
3.90689433e-01 -3.68843973e-01 -5.67136072e-02 -1.25908256e+00
-2.82658547e-01 -2.00087309e-01 6.05592430e-01 1.54783237e+00
4.31652009e-01 -6.65913999e-01 3.51837367e-01 4.36914921e-01
3.02229822e-02 -3.95225108e-01 -5.87231338e-01 -8.50061119e-01
1.29829094e-01 -3.38558406e-01 3.44843060e-01 1.11807001e+00
5.17546773e-01 5.21388531e-01 -2.41323505e-02 -1.50706381e-01
4.69768733e-01 1.99590608e-01 4.26159471e-01 -1.86606729e+00
-1.43855676e-01 -6.28355682e-01 -7.19828188e-01 -4.47040170e-01
3.48572493e-01 -1.08255398e+00 6.18655942e-02 -1.57919192e+00
2.75902271e-01 -6.67869091e-01 -2.65663117e-01 4.55091953e-01
2.91750077e-02 4.06677932e-01 3.53913605e-02 2.66302347e-01
-5.84329903e-01 -1.26367494e-01 3.27417493e-01 -5.94672635e-02
-2.21322298e-01 1.55480534e-01 -8.93231392e-01 9.41587031e-01
8.97174954e-01 -6.23082459e-01 -3.88591617e-01 3.44176106e-02
3.49600554e-01 -4.01919276e-01 3.60027738e-02 -1.00170588e+00
3.40081632e-01 -6.91765696e-02 -1.16231956e-01 -1.26617417e-01
-7.17293173e-02 -1.26119506e+00 1.46974891e-01 4.67522144e-01
-6.38937116e-01 -6.18983582e-02 -3.10043134e-02 5.02952814e-01
-3.51564437e-01 -8.45749736e-01 7.98471570e-01 -2.09409550e-01
-8.79933655e-01 -3.83935392e-01 -7.03593552e-01 -1.32629603e-01
1.07153678e+00 -4.24310714e-01 1.68967489e-02 -3.06440204e-01
-9.36893106e-01 -5.82479425e-02 4.32658136e-01 4.32477325e-01
1.03391662e-01 -1.07173777e+00 -4.87880290e-01 -1.70879886e-01
6.35810137e-01 -6.22868359e-01 -1.03762247e-01 6.83021486e-01
-5.62171638e-01 7.64407158e-01 -1.27565011e-01 -3.32519054e-01
-1.29339767e+00 8.95650268e-01 1.86925530e-01 -3.14808041e-01
-5.54306090e-01 3.66987586e-01 2.84346104e-01 -3.28172624e-01
4.18937892e-01 -5.20161353e-02 -1.03453374e+00 6.68524861e-01
6.29708767e-01 2.80942023e-01 3.07556510e-01 -1.01565027e+00
-3.59665990e-01 7.88820684e-01 -1.19547680e-01 -5.67591190e-02
1.21130288e+00 -2.22110510e-01 -3.39903116e-01 7.24753618e-01
1.03682244e+00 1.53082788e-01 -3.45727026e-01 -3.38428676e-01
6.26860440e-01 -1.36130258e-01 1.02359215e-02 -7.65041292e-01
-5.39216638e-01 8.04744482e-01 5.31556964e-01 7.40958929e-01
1.14786386e+00 1.95152491e-01 2.70554304e-01 5.73315084e-01
5.29066980e-01 -1.28811491e+00 -7.30508864e-02 2.45785773e-01
7.26680100e-01 -1.08181679e+00 7.31275300e-04 -6.44112110e-01
-5.53041577e-01 1.33460426e+00 1.68354213e-01 8.59466195e-02
8.01243782e-01 -1.42838120e-01 2.47405499e-01 -3.59757841e-01
-2.67739236e-01 -4.63710129e-01 3.11917007e-01 4.04961199e-01
5.78887761e-01 -1.31957605e-01 -1.04838729e+00 6.48484051e-01
-1.27751514e-01 -1.65699467e-01 4.20248419e-01 9.68750358e-01
-7.15181410e-01 -1.36721683e+00 -3.67566437e-01 5.15276253e-01
-7.87164509e-01 3.36945131e-02 -4.42165673e-01 8.32121968e-01
3.18235874e-01 1.28641307e+00 5.16946800e-03 -2.15085149e-01
4.22357529e-01 2.22286224e-01 8.93470943e-02 -1.04072273e+00
-9.13611948e-01 5.68077974e-02 2.39994153e-01 -2.86795534e-02
-7.42983997e-01 -7.12888241e-01 -1.14450920e+00 3.39797921e-02
-3.81908923e-01 8.81217241e-01 6.48719490e-01 1.06368172e+00
-6.46757483e-02 1.81330234e-01 7.63144970e-01 -6.62812889e-01
-5.67555308e-01 -9.66850996e-01 -8.41186941e-01 6.63389087e-01
3.11861373e-02 -8.52148712e-01 -6.32788122e-01 3.71957034e-01] | [10.16830825805664, 8.667800903320312] |
3e49dbd5-c3a3-48fb-9113-2100b779e9ef | deep-learning-for-entity-matching-a-design | null | null | https://doi.org/10.1145/3183713.3196926 | http://pages.cs.wisc.edu/~anhai/papers1/deepmatcher-sigmod18.pdf | Deep Learning for Entity Matching: A Design Space Exploration | Entity matching (EM) finds data instances that refer to the same real-world entity. In this paper we examine applying deep learning (DL) to EM, to understand DL's benefits and limitations. We review many DL solutions that have been developed for related matching tasks in text processing (e.g., entity linking, textual entailment, etc.). We categorize these solutions and define a space of DL solutions for EM, as embodied by four solutions with varying representational power: SIF, RNN, Attention, and Hybrid. Next, we investigate the types of EM problems for which DL can be helpful. We consider three such problem types, which match structured data instances, textual instances, and dirty instances, respectively. We empirically compare the above four DL solutions with Magellan, a state-of-the-art learning-based EM solution. The results show that DL does not outperform current solutions on structured EM, but it can significantly outperform them on textual and dirty EM. For practitioners, this suggests that they should seriously consider using DL for textual and dirty EM problems. Finally, we analyze DL's performance and discuss future research directions. | ['Vijay Raghavendra', 'Esteban Arcaute', 'Rohit Deep', 'Ganesh Krishnan', 'Youngchoon Park', 'AnHai Doan', 'Theodoros Rekatsinas', 'Han Li', 'Sidharth Mudgal'] | 2018-05-01 | null | null | null | sigmod-international-conference-on-management | ['entity-resolution'] | ['natural-language-processing'] | [-1.95183381e-01 1.83100268e-01 -1.31737694e-01 -3.46445978e-01
-7.95691729e-01 -3.77845436e-01 6.36090577e-01 3.06326717e-01
-8.09085310e-01 6.66743577e-01 3.49835455e-01 -4.82091129e-01
-2.39432901e-01 -8.73349071e-01 -7.23815858e-01 -1.59569606e-01
2.75837362e-01 1.09789789e+00 -1.71995163e-01 -3.20950329e-01
1.61532611e-01 2.43970066e-01 -1.45254457e+00 6.09197497e-01
1.02988040e+00 6.35849535e-01 2.63424873e-01 3.07985693e-01
-8.32804441e-01 1.24032414e+00 -8.25827420e-01 -1.26537895e+00
-4.25264128e-02 3.42157446e-02 -1.26437509e+00 -5.62005639e-01
8.00753891e-01 -3.91359329e-02 -6.70637131e-01 1.10214591e+00
8.32391143e-01 4.03670281e-01 7.86502302e-01 -1.54334486e+00
-9.85901535e-01 1.03384912e+00 -5.49714029e-01 2.90750146e-01
5.54380715e-01 -3.03049684e-01 1.15110195e+00 -1.13749456e+00
1.00064266e+00 1.46786475e+00 1.06789815e+00 5.45142710e-01
-7.75900900e-01 -7.86213160e-01 2.72163719e-01 6.82860970e-01
-1.34743166e+00 -4.01521623e-01 1.44815251e-01 -2.05520794e-01
1.70115888e+00 3.44874024e-01 -3.18589807e-02 1.15475917e+00
1.07856154e-01 1.29718804e+00 5.00407159e-01 -4.25639451e-01
-1.28421858e-01 1.76853850e-01 3.85518968e-01 2.51134545e-01
4.45017725e-01 -1.31000429e-01 -1.10720672e-01 -2.64317811e-01
8.00274592e-03 -1.99394718e-01 -1.52196631e-01 9.56771746e-02
-1.08080363e+00 8.20738792e-01 3.57281238e-01 5.01422107e-01
-3.11034530e-01 -3.54618914e-02 5.20979702e-01 5.29009998e-01
5.19443393e-01 7.04635501e-01 -6.37371659e-01 1.21965155e-01
-1.03235960e+00 4.51956242e-01 1.22149301e+00 1.48721969e+00
5.50195813e-01 -3.00354391e-01 -1.96331218e-01 9.80095983e-01
3.99358898e-01 4.10257071e-01 4.51074660e-01 -6.06925070e-01
1.03626704e+00 6.17026150e-01 -3.56456032e-03 -1.12072170e+00
-7.15672314e-01 -1.77484289e-01 -1.02481139e+00 -5.68729162e-01
-9.22693387e-02 -3.18283945e-01 -4.81621236e-01 1.81669152e+00
2.05175400e-01 3.19109857e-01 2.08355486e-01 4.84395951e-01
1.67143583e+00 7.20134199e-01 4.04853970e-01 3.13553996e-02
1.30591702e+00 -1.11849892e+00 -9.52811837e-01 -5.11847794e-01
1.08442199e+00 -1.00289774e+00 4.73062903e-01 -1.68660864e-01
-1.39182770e+00 -6.07108176e-01 -7.61286557e-01 -3.54948431e-01
-9.39322472e-01 6.54973686e-02 5.14624417e-01 1.90540791e-01
-1.07958448e+00 6.56630456e-01 -4.85061944e-01 -6.83175743e-01
-1.39397860e-01 4.00723070e-01 -3.06332052e-01 2.38881465e-02
-1.88988602e+00 1.76884711e+00 5.93051910e-01 6.96344227e-02
-6.48788959e-02 -6.53850794e-01 -1.18304396e+00 8.65366012e-02
1.45119220e-01 -1.22560000e+00 1.53879440e+00 -5.33218920e-01
-6.20046616e-01 1.31653726e+00 -4.20726776e-01 -4.43570226e-01
4.08948213e-01 -3.22603554e-01 -9.99632120e-01 -3.92048657e-01
2.41596833e-01 6.63956225e-01 2.59678245e-01 -1.16755426e+00
-8.51020038e-01 -1.55553982e-01 1.32724553e-01 8.33975151e-02
-2.85090595e-01 4.34433818e-01 -4.79804337e-01 -7.12421179e-01
-2.61816978e-01 -7.92390823e-01 -1.87420055e-01 -4.74688888e-01
-5.86071372e-01 -6.64201140e-01 4.15796250e-01 -7.05724180e-01
1.93660760e+00 -1.77377284e+00 3.19766998e-01 -6.34352118e-02
2.83957899e-01 4.81636643e-01 -4.48466510e-01 8.76080215e-01
-2.51819044e-01 1.81702808e-01 -7.23480508e-02 -6.91288233e-01
6.65052533e-01 5.27873375e-02 -8.55810195e-02 1.68799788e-01
1.02517389e-01 1.47090888e+00 -7.60612488e-01 -7.91011691e-01
5.67282215e-02 4.04814243e-01 -3.67518574e-01 4.95801151e-01
-1.60346925e-01 -2.91427284e-01 -1.61275834e-01 5.39778113e-01
9.32117224e-01 -4.58319396e-01 3.27697963e-01 -5.08942962e-01
8.58257636e-02 6.06788337e-01 -1.44157183e+00 1.46036041e+00
-6.27585351e-01 9.26278949e-01 7.69870728e-02 -1.06031585e+00
7.69235730e-01 3.54475975e-01 2.79604077e-01 -9.38734591e-01
1.10989504e-01 2.33402520e-01 -2.69239902e-01 -8.37798297e-01
1.03315270e+00 1.73973367e-01 -1.06872700e-01 6.75970614e-01
1.65133163e-01 4.96885389e-01 4.39061582e-01 3.47876102e-01
1.14220226e+00 -2.16318339e-01 3.29654843e-01 -1.07326306e-01
4.51423794e-01 -1.80723101e-01 4.00911152e-01 1.05492377e+00
1.41485795e-01 2.30567738e-01 -7.39742070e-02 -3.34732056e-01
-1.05199897e+00 -8.40853631e-01 -3.42125148e-01 1.34556639e+00
4.06086355e-01 -5.04528165e-01 -6.76971078e-01 -6.63319588e-01
2.63897806e-01 9.17215645e-01 -4.75870639e-01 -3.74519557e-01
-8.56162369e-01 -1.04456890e+00 6.65434182e-01 9.20406222e-01
5.03316760e-01 -1.48836792e+00 -2.09150717e-01 3.94537061e-01
-7.67676413e-01 -1.15546608e+00 -3.27774882e-01 2.31348932e-01
-6.09608471e-01 -9.16170835e-01 -5.95402777e-01 -1.16076088e+00
4.78141278e-01 1.50983796e-01 2.02872014e+00 2.21347347e-01
-2.11483389e-01 5.02066076e-01 -4.27640349e-01 -2.17832625e-01
-4.95102435e-01 4.93219823e-01 -1.21065937e-02 -5.43317258e-01
1.19744539e+00 -2.18806639e-01 -2.83604175e-01 2.41830617e-01
-9.52664673e-01 -2.03951746e-01 6.91748738e-01 8.63060415e-01
9.72642601e-02 -6.45375848e-02 6.23931766e-01 -1.13966787e+00
8.68573487e-01 -1.16116023e+00 -9.58681554e-02 7.89826095e-01
-5.76531172e-01 1.36884347e-01 2.56398797e-01 -5.15826643e-01
-1.11266398e+00 -5.94772339e-01 -5.74238718e-01 -1.67466313e-01
-1.67574078e-01 8.16998363e-01 -2.76281297e-01 2.14010805e-01
5.42674661e-01 -1.49538681e-01 -6.24343514e-01 -5.72940588e-01
5.42115629e-01 1.03727245e+00 6.17979407e-01 -7.52831995e-01
5.18622756e-01 1.42906429e-02 -5.35946429e-01 -2.98257828e-01
-9.79759455e-01 -6.98735833e-01 -5.77536404e-01 9.78502855e-02
5.94476163e-01 -8.81613672e-01 -6.69249415e-01 2.09372625e-01
-1.57291949e+00 -1.98088065e-02 7.75925741e-02 3.27107608e-01
-4.12131876e-01 4.26833659e-01 -1.16087508e+00 -4.29484099e-01
-7.10676193e-01 -7.79282451e-01 1.11785328e+00 1.26179531e-01
-7.16765225e-01 -1.42656589e+00 2.11001396e-01 3.04256022e-01
6.38368726e-01 1.23703005e-02 9.69154298e-01 -1.14239347e+00
-1.17502317e-01 -1.62965938e-01 -3.50792557e-01 -2.13961408e-01
-3.01479250e-01 3.52718122e-02 -7.72837698e-01 -3.82351756e-01
-2.60418236e-01 -1.45810291e-01 8.46158683e-01 3.24688792e-01
9.99552488e-01 -3.26172948e-01 -8.03487420e-01 6.47085309e-01
1.30664229e+00 1.57088086e-01 8.11767995e-01 9.49936271e-01
5.55410624e-01 6.39476955e-01 7.19166756e-01 3.79677624e-01
7.96038270e-01 6.26844645e-01 1.60269007e-01 -3.98361474e-01
-7.90439174e-03 -2.73694813e-01 1.65192828e-01 1.08728898e+00
3.17053378e-01 -8.28289270e-01 -8.01543951e-01 5.29333055e-01
-2.37522030e+00 -1.16226399e+00 -3.32608372e-01 1.63182414e+00
7.97217786e-01 -1.35769710e-01 -7.96055645e-02 -3.50889951e-01
1.13984191e+00 -1.60433486e-01 -5.38603902e-01 -5.28516710e-01
-4.82663661e-01 2.80576319e-01 2.60950536e-01 2.37125576e-01
-1.38158596e+00 9.07923102e-01 6.55237675e+00 9.03652489e-01
-4.63032067e-01 8.34943429e-02 3.17457736e-01 8.70418102e-02
-5.18436253e-01 -2.49431685e-01 -1.11800992e+00 6.05759084e-01
1.05041611e+00 -3.67587745e-01 1.15856625e-01 7.60018647e-01
-5.04660845e-01 3.39539856e-01 -1.58602190e+00 1.03709674e+00
1.56582519e-01 -1.29046822e+00 1.57299653e-01 -2.09797859e-01
6.93368018e-01 1.60820320e-01 -1.83181673e-01 1.00149632e+00
5.43837667e-01 -1.18865073e+00 4.04609710e-01 5.75837553e-01
3.38817418e-01 -9.05082941e-01 1.32198787e+00 3.29184026e-01
-1.33417487e+00 4.60278504e-02 -7.80487716e-01 3.82186443e-01
4.06757265e-01 5.73700964e-01 -2.17975616e-01 8.01155806e-01
8.05953085e-01 8.88319373e-01 -5.25104463e-01 1.09851921e+00
5.60527965e-02 5.53771574e-03 -1.81743443e-01 -8.17738920e-02
2.38049716e-01 4.37489115e-02 4.15886968e-01 2.03627086e+00
3.93551916e-01 -1.05531290e-01 -3.91085595e-02 1.18877363e+00
-4.24400896e-01 3.43265422e-02 -7.61017501e-01 2.63037950e-01
1.16047919e+00 1.26435423e+00 -8.79958123e-02 -5.69565415e-01
-7.66214907e-01 9.79635060e-01 7.89059341e-01 3.09561372e-01
-1.09796512e+00 -6.28503978e-01 5.96913278e-01 -3.16685468e-01
1.79757759e-01 2.24133328e-01 9.67996567e-02 -1.21790946e+00
1.61951065e-01 -1.08974206e+00 9.72776949e-01 -7.62604773e-01
-1.99830818e+00 8.24695587e-01 8.32518414e-02 -1.09219098e+00
-3.67000014e-01 -5.92801929e-01 -6.23093307e-01 5.95566869e-01
-1.59723353e+00 -1.03035605e+00 -3.62022370e-01 4.84443933e-01
3.39105994e-01 -3.01405460e-01 7.67733872e-01 1.13277888e+00
-7.82711327e-01 1.08853638e+00 3.25095356e-01 4.57140297e-01
9.88171160e-01 -1.21858275e+00 1.05202413e+00 4.44419324e-01
2.15514943e-01 9.04758215e-01 5.27363181e-01 -5.21468937e-01
-1.35906446e+00 -1.26998472e+00 1.66923296e+00 -7.41397083e-01
6.67289615e-01 7.65636237e-03 -1.06921792e+00 1.10499954e+00
5.31566560e-01 -4.73465800e-01 7.01884985e-01 3.28890711e-01
-2.88551271e-01 4.02932435e-01 -1.18026674e+00 6.53070152e-01
1.23848414e+00 -4.82834935e-01 -1.14743054e+00 2.70315111e-01
6.40423179e-01 -4.88166273e-01 -1.18950200e+00 7.63668299e-01
1.75165594e-01 -6.19241774e-01 1.24494040e+00 -1.18940055e+00
4.68146592e-01 -4.70620617e-02 -1.26070708e-01 -1.36824918e+00
-6.44738615e-01 -2.45202854e-01 -5.87551892e-01 1.69537091e+00
6.36463523e-01 -4.90143448e-01 4.44052130e-01 9.82147753e-01
-1.72281459e-01 -5.87685049e-01 -6.87717676e-01 -9.60084140e-01
6.14409328e-01 -3.49202901e-01 9.06225920e-01 1.57568240e+00
3.62685025e-01 4.71800148e-01 -2.18912795e-01 -3.81024480e-02
4.27830219e-01 2.94101149e-01 6.59716904e-01 -1.27232289e+00
-1.01981744e-01 -7.16081500e-01 -2.65956998e-01 -1.29872870e+00
9.06995118e-01 -1.33826065e+00 -2.82944232e-01 -1.74764407e+00
6.01706922e-01 -4.29745495e-01 -2.04969332e-01 2.69207358e-01
-5.76603472e-01 -2.39553109e-01 1.51987866e-01 6.57380223e-02
-9.67895806e-01 3.39789391e-01 6.53398275e-01 -4.80213493e-01
2.48454854e-01 -2.21590176e-01 -6.41457379e-01 4.82615978e-01
7.37367392e-01 -5.51283062e-01 1.19411878e-01 -8.83296192e-01
7.44638205e-01 -1.81619957e-01 -3.89140025e-02 -6.13578439e-01
7.83904910e-01 1.22235052e-01 2.51500160e-01 -8.40608895e-01
-2.38955803e-02 -8.36711764e-01 3.66189748e-01 5.16316444e-02
-5.62002420e-01 7.30503738e-01 3.83896500e-01 1.26099810e-01
-5.40583670e-01 -6.92826867e-01 5.88569283e-01 -2.63760805e-01
-9.25753355e-01 2.57444263e-01 -2.92515814e-01 7.11298108e-01
7.29903519e-01 2.88789291e-02 -6.55980527e-01 -3.78548563e-01
-6.55820608e-01 7.58133471e-01 4.32983115e-02 7.16073453e-01
4.83776361e-01 -1.63692021e+00 -1.07885313e+00 -2.03010589e-01
2.32699007e-01 -1.82139352e-01 1.14172541e-01 7.97103822e-01
-1.65116653e-01 5.81910968e-01 9.28584412e-02 -2.59209067e-01
-1.21235764e+00 8.80754232e-01 5.55619061e-01 -5.04060447e-01
-6.07348740e-01 8.16006362e-01 5.55297621e-02 -1.13543427e+00
6.68365717e-01 -6.42014444e-02 -1.85946479e-01 2.41735309e-01
5.08254766e-01 6.86668634e-01 3.80908012e-01 -5.02351463e-01
-5.05797923e-01 5.35982430e-01 -3.23263168e-01 4.17240769e-01
1.16146219e+00 -3.28474939e-01 -3.69305998e-01 7.14448914e-02
1.40255547e+00 -3.64752144e-01 -1.46022335e-01 -6.07304990e-01
6.87603891e-01 -5.33687025e-02 -1.83180824e-01 -7.77864993e-01
-1.14072740e+00 8.13998938e-01 1.99847788e-01 1.33239806e-01
8.62838745e-01 2.45281249e-01 1.27124786e+00 7.44677007e-01
2.66185790e-01 -1.38700974e+00 -4.36562359e-01 8.50815952e-01
5.71458042e-01 -1.34756970e+00 -1.47787958e-01 2.68396977e-02
-5.25368929e-01 1.06838024e+00 1.10302091e+00 2.74137318e-01
6.22882843e-01 6.44103169e-01 -9.44083706e-02 -4.41607684e-01
-1.05059779e+00 -4.44173932e-01 5.31565368e-01 4.67248738e-01
9.13940489e-01 -1.11271530e-01 -2.38573804e-01 7.03238249e-01
-3.01686287e-01 -3.95377785e-01 -2.12433413e-02 8.67895901e-01
-2.53901839e-01 -1.01257885e+00 -1.84180006e-01 5.84488273e-01
-4.72161919e-01 -4.68552321e-01 -8.11925292e-01 1.07452857e+00
6.13897331e-02 9.53789473e-01 4.21426445e-01 -4.19489175e-01
6.18145287e-01 1.39563695e-01 3.68581802e-01 -2.46928051e-01
-1.02085328e+00 -6.15862012e-01 5.48095047e-01 -2.73197860e-01
-3.56037825e-01 -4.45851833e-01 -1.18815863e+00 -1.06148267e+00
-6.42980039e-01 1.57371148e-01 2.77013570e-01 9.33230758e-01
4.65656757e-01 3.04199904e-01 2.64854252e-01 -4.09918517e-01
-3.26879293e-01 -1.16928256e+00 -1.23582937e-01 7.20054984e-01
1.40435725e-01 -5.99914372e-01 -4.56360698e-01 -4.58855122e-01] | [9.478570938110352, 8.581327438354492] |
9d54d36c-2ff2-41cd-90b8-c7089a7aba94 | mocapdeform-monocular-3d-human-motion-capture | 2208.08439 | null | https://arxiv.org/abs/2208.08439v1 | https://arxiv.org/pdf/2208.08439v1.pdf | MoCapDeform: Monocular 3D Human Motion Capture in Deformable Scenes | 3D human motion capture from monocular RGB images respecting interactions of a subject with complex and possibly deformable environments is a very challenging, ill-posed and under-explored problem. Existing methods address it only weakly and do not model possible surface deformations often occurring when humans interact with scene surfaces. In contrast, this paper proposes MoCapDeform, i.e., a new framework for monocular 3D human motion capture that is the first to explicitly model non-rigid deformations of a 3D scene for improved 3D human pose estimation and deformable environment reconstruction. MoCapDeform accepts a monocular RGB video and a 3D scene mesh aligned in the camera space. It first localises a subject in the input monocular video along with dense contact labels using a new raycasting based strategy. Next, our human-environment interaction constraints are leveraged to jointly optimise global 3D human poses and non-rigid surface deformations. MoCapDeform achieves superior accuracy than competing methods on several datasets, including our newly recorded one with deforming background scenes. | ['Vladislav Golyanik', 'Christian Theobalt', 'Bernt Schiele', 'Soshi Shimada', 'Zhi Li'] | 2022-08-17 | null | null | null | null | ['3d-human-pose-estimation'] | ['computer-vision'] | [ 2.41383985e-01 -4.44236435e-02 2.79382408e-01 -1.85670495e-01
-3.95662010e-01 -6.42897666e-01 5.60387969e-01 -8.19044292e-01
-6.29504025e-01 4.41169977e-01 4.20579404e-01 3.59938323e-01
2.45708764e-01 -2.47560024e-01 -7.41412461e-01 -6.28267884e-01
3.92493606e-01 9.36286747e-01 3.69281530e-01 -1.06564380e-01
-1.42493486e-01 6.41612530e-01 -1.57446444e+00 -3.86116877e-02
9.31427106e-02 4.82364625e-01 1.12420186e-01 1.12845647e+00
4.73305762e-01 4.51368064e-01 -8.82247910e-02 -2.04483822e-01
7.03074515e-01 -3.23238254e-01 -9.41435695e-01 6.07906401e-01
7.93804884e-01 -3.60198051e-01 -5.92658937e-01 6.01554811e-01
7.32071996e-01 4.20607030e-01 4.28467989e-01 -1.05057502e+00
7.12907016e-02 -3.68598282e-01 -4.19147402e-01 -5.40353246e-02
1.28500462e+00 3.97016615e-01 5.06949246e-01 -1.01406860e+00
1.40301609e+00 1.40784454e+00 6.91328228e-01 9.33058858e-01
-1.08479249e+00 -1.10435426e-01 1.91299662e-01 -2.97563188e-02
-1.17215431e+00 -4.09109414e-01 8.87140095e-01 -6.87802076e-01
9.56812203e-01 7.44569063e-01 1.13057911e+00 1.41431260e+00
1.48683950e-01 6.61628246e-01 1.01090312e+00 -2.15526074e-01
1.43534943e-01 -2.04703420e-01 -2.01000899e-01 6.47836685e-01
1.00829350e-02 2.67223299e-01 -6.66519523e-01 -3.30472499e-01
1.23889065e+00 1.62642717e-01 -4.77957457e-01 -1.18982780e+00
-1.56259859e+00 1.80595711e-01 1.65782914e-01 4.33337837e-02
-3.95611733e-01 2.21054673e-01 -1.29343802e-02 -7.75714964e-02
3.69392723e-01 1.94688886e-01 -5.59815288e-01 -3.13426793e-01
-7.24011540e-01 7.70755351e-01 6.56352997e-01 9.30864155e-01
5.39095104e-01 -1.05950229e-01 9.17209163e-02 2.77688175e-01
4.47397083e-01 7.18165815e-01 6.10903911e-02 -1.20161998e+00
3.84960413e-01 5.56003213e-01 4.75112945e-01 -9.12065923e-01
-7.04001307e-01 2.11909227e-02 -4.51750875e-01 3.05039257e-01
4.54734981e-01 -2.23753490e-02 -9.24567997e-01 1.46246028e+00
1.12223554e+00 2.26288155e-01 -1.41160950e-01 1.54712117e+00
7.62611091e-01 2.55207956e-01 -2.08449692e-01 9.60623622e-02
1.14920580e+00 -8.31462801e-01 -4.54783469e-01 -4.43454802e-01
3.61482531e-01 -7.03545094e-01 8.95053923e-01 4.25670296e-01
-1.46949935e+00 -4.81512040e-01 -4.59512949e-01 -3.35317045e-01
2.67211962e-02 -2.16907769e-01 1.98962212e-01 4.94916886e-01
-8.94887745e-01 3.47987801e-01 -1.15661180e+00 -5.64069986e-01
-1.49295926e-01 5.03770351e-01 -9.10781384e-01 -2.23724723e-01
-8.50024402e-01 1.01076400e+00 1.81070585e-02 3.42358470e-01
-6.88074231e-01 -3.99958104e-01 -1.14968371e+00 -8.02337646e-01
7.75679469e-01 -1.29374790e+00 1.02763939e+00 -6.62736416e-01
-1.72575700e+00 1.43916142e+00 -4.09377925e-02 1.81401465e-02
1.12328815e+00 -7.40990818e-01 2.70375907e-01 1.44511774e-01
-2.98758298e-01 4.46876347e-01 7.44017124e-01 -1.44202447e+00
-6.24897368e-02 -7.28414059e-01 4.04412970e-02 8.97173285e-01
4.75183278e-01 -6.82664141e-02 -8.77057135e-01 -6.10967696e-01
4.57991064e-01 -1.49592853e+00 -4.37207282e-01 3.19053501e-01
-5.06789923e-01 2.66935438e-01 7.33814716e-01 -8.21252763e-01
6.26118422e-01 -1.87046039e+00 1.23544109e+00 2.23561704e-01
7.41536915e-02 6.99088201e-02 1.51949853e-01 4.38453220e-02
2.20389694e-01 -3.54787529e-01 -4.32013869e-01 -6.86001062e-01
2.63281405e-01 5.57367682e-01 2.25327134e-01 9.89907801e-01
-2.19409391e-01 1.07131791e+00 -8.92509580e-01 -4.31343257e-01
5.98167360e-01 8.95897269e-01 -7.71381557e-01 6.55950427e-01
-4.35789488e-02 1.14433539e+00 -3.49013656e-01 8.12591791e-01
5.48402250e-01 -3.12047242e-03 -1.12068579e-01 -1.27787232e-01
-5.77580705e-02 -3.30284894e-01 -1.62159288e+00 2.22138429e+00
4.60265018e-02 8.46490785e-02 4.59225953e-01 -4.51778233e-01
4.63436782e-01 3.92172337e-01 8.42617571e-01 -1.66282341e-01
2.48857066e-01 1.01626672e-01 -5.80772460e-01 -7.77443409e-01
4.03140962e-01 -2.43409112e-01 -1.20295823e-01 7.27921501e-02
-1.95163950e-01 -5.06703556e-01 -7.88241625e-01 -2.30321154e-01
1.11677420e+00 1.00656056e+00 2.76773185e-01 -7.86695629e-02
5.28737724e-01 -7.61970133e-02 4.52730447e-01 2.46038705e-01
-3.02744508e-01 1.09759903e+00 -1.62298769e-01 -7.62713373e-01
-1.05370641e+00 -1.19455504e+00 9.28314254e-02 6.84784353e-01
4.56255555e-01 -5.23309316e-03 -9.02631640e-01 -4.16838199e-01
2.94872802e-02 8.92648026e-02 -6.73349142e-01 1.26417249e-01
-1.01596642e+00 -4.30066377e-01 1.34068862e-01 4.37325537e-01
3.09410274e-01 -1.05325937e+00 -1.48979485e+00 3.50328013e-02
-4.70569164e-01 -1.49694622e+00 -6.58865392e-01 -1.65410131e-01
-6.37625933e-01 -1.24154580e+00 -1.07009506e+00 -4.94202077e-01
7.12259829e-01 7.80362636e-02 1.08333242e+00 -1.14136398e-01
-7.09933102e-01 1.08525157e+00 -2.29610458e-01 1.60342768e-01
-4.25540507e-02 -3.55153084e-01 3.46539348e-01 1.87513337e-01
5.27738854e-02 -3.23406577e-01 -7.80302644e-01 5.28027236e-01
-6.74223602e-01 3.04411530e-01 1.30923064e-02 4.74662811e-01
8.17039132e-01 -4.32770550e-01 -6.04472518e-01 -7.88567245e-01
-2.46373221e-01 -2.66497076e-01 -3.73280317e-01 -1.72472019e-02
3.23764235e-01 -2.71087825e-01 -7.05795586e-02 -5.72815359e-01
-1.02057815e+00 8.89489949e-01 -8.58955383e-02 -7.99003363e-01
-5.52628100e-01 -1.06242076e-01 -4.32298869e-01 -2.45311141e-01
6.33001745e-01 -1.79343286e-03 -2.00620055e-01 -4.57269460e-01
2.57660151e-01 7.13351294e-02 8.83704066e-01 -4.92516160e-01
1.15390038e+00 1.02011514e+00 2.51183420e-01 -8.58691573e-01
-4.53248084e-01 -7.90556192e-01 -1.52866232e+00 -5.56145549e-01
1.50591481e+00 -9.24974442e-01 -6.77424788e-01 7.41477072e-01
-1.08167934e+00 -7.80348837e-01 -2.52083480e-01 5.77363789e-01
-1.04038286e+00 3.63577843e-01 -2.85808176e-01 -8.90624344e-01
-6.26914855e-03 -1.05528021e+00 1.75260866e+00 -3.54365826e-01
-5.00358820e-01 -1.04015362e+00 3.75291556e-01 6.99907303e-01
-1.30881324e-01 1.35759687e+00 1.53803781e-01 2.24402115e-01
-6.33297741e-01 -1.99715927e-01 5.00860691e-01 -1.77969024e-01
3.78242321e-02 -3.50526303e-01 -9.47154224e-01 -3.17579836e-01
1.86869413e-01 -1.93196520e-01 2.38561228e-01 3.46996158e-01
5.61006129e-01 -2.24407911e-01 -3.22833687e-01 9.60327268e-01
1.12924337e+00 -1.84339851e-01 5.46210647e-01 4.01462317e-01
1.31443918e+00 9.79258895e-01 6.78307593e-01 5.87211370e-01
5.40614307e-01 1.28181136e+00 5.37069976e-01 -2.50676692e-01
4.13096547e-02 -3.83683629e-02 4.59301531e-01 6.77982092e-01
-4.72286850e-01 1.11661013e-02 -1.12098110e+00 3.61787289e-01
-1.80727816e+00 -7.41387844e-01 -3.57026130e-01 2.26860833e+00
5.71904540e-01 -2.14457855e-01 4.22524929e-01 8.96135494e-02
3.52234244e-01 -8.94248560e-02 -7.08360016e-01 1.60350606e-01
-1.90516144e-01 8.52999091e-02 3.54161412e-01 8.73955488e-01
-9.75360334e-01 9.37319994e-01 5.56167412e+00 -3.20327193e-01
-8.56287301e-01 1.52378514e-01 -2.00373251e-02 -5.44198811e-01
-2.20816031e-01 -1.31204903e-01 -5.77129543e-01 9.13705602e-02
3.88791859e-01 4.30874735e-01 3.49424452e-01 5.01865625e-01
6.80615678e-02 -1.83660030e-01 -1.29260278e+00 1.34658599e+00
3.33294213e-01 -7.24890649e-01 -2.48203501e-01 3.21791679e-01
7.73888230e-01 -1.27254799e-01 -6.43527806e-02 -2.13597193e-01
6.28131106e-02 -9.08402383e-01 1.09400034e+00 9.85779107e-01
5.86714864e-01 -3.82493317e-01 3.86332065e-01 6.26249373e-01
-1.22573948e+00 2.70194978e-01 4.84596007e-02 -2.63803322e-02
6.10915422e-01 4.12889719e-02 -3.49767834e-01 5.44748783e-01
9.54446256e-01 6.92923248e-01 -4.17235404e-01 8.20808768e-01
1.18359312e-01 -2.29023527e-02 -3.86589795e-01 4.95018661e-01
2.68784408e-02 -1.48986608e-01 9.91828799e-01 1.08608866e+00
1.08654104e-01 6.26030505e-01 4.75643605e-01 5.99269271e-01
4.38510954e-01 -9.60208550e-02 -7.20622540e-01 6.37737215e-01
-1.84374496e-01 9.65983152e-01 -6.22613132e-01 -1.34033561e-01
-1.82424903e-01 1.67194390e+00 6.68573454e-02 4.90327418e-01
-7.87675679e-01 5.25160432e-01 6.90333128e-01 6.32681429e-01
9.11316574e-02 -6.43430591e-01 1.40751034e-01 -1.65173423e+00
2.20637277e-01 -8.15258205e-01 2.43562341e-01 -1.07571542e+00
-7.72465110e-01 6.17157638e-01 4.43067849e-01 -1.13798952e+00
-4.17008638e-01 -6.89386845e-01 4.30367105e-02 7.48548031e-01
-1.14634538e+00 -1.30184793e+00 -7.97916353e-01 1.09747016e+00
6.19672716e-01 3.89221400e-01 6.69574201e-01 2.35337894e-02
-1.20727226e-01 8.89272541e-02 -4.80954766e-01 -1.43567577e-01
6.41359925e-01 -1.38860118e+00 7.31918991e-01 7.20245242e-01
3.74063104e-02 3.54558855e-01 8.08433712e-01 -8.47347975e-01
-2.05918694e+00 -1.04026115e+00 5.89621067e-01 -1.47425413e+00
4.30579334e-02 -7.59132981e-01 -8.79669249e-01 8.96493256e-01
-1.83180034e-01 4.86493409e-01 3.37619483e-01 -5.70857286e-01
6.54248223e-02 4.39736634e-01 -1.29084373e+00 5.91293633e-01
1.78922760e+00 -4.83386338e-01 -6.22363508e-01 3.25001240e-01
3.99318933e-01 -1.32787001e+00 -8.57696414e-01 4.55085218e-01
6.90277874e-01 -8.06076944e-01 1.38689399e+00 -3.90236616e-01
-1.42166108e-01 -4.40667421e-01 -4.57907379e-01 -9.84744012e-01
-1.87914014e-01 -1.14418507e+00 -3.91692400e-01 4.40719247e-01
-3.96623731e-01 -2.01436162e-01 1.10367429e+00 1.04082763e+00
3.94084230e-02 -4.92148638e-01 -1.18415821e+00 -6.02947295e-01
-1.64999470e-01 -3.56622547e-01 2.54913449e-01 8.47048402e-01
-4.35809374e-01 -2.81017065e-01 -7.03397214e-01 3.55294824e-01
9.88148868e-01 -2.19560578e-01 1.39659762e+00 -1.38405406e+00
-3.12236488e-01 1.18442483e-01 -6.09051049e-01 -1.04458702e+00
2.76453137e-01 -5.67308009e-01 2.80455828e-01 -1.32775831e+00
7.22451583e-02 -1.49964690e-01 4.40539449e-01 2.14170963e-01
-1.64772257e-01 4.83065844e-01 1.73329711e-01 3.25398177e-01
-5.08941352e-01 5.02446830e-01 1.39013588e+00 4.25918311e-01
-2.45268792e-01 -1.30886301e-01 1.50172725e-01 1.00313091e+00
1.64809510e-01 -2.07741410e-01 -2.69548982e-01 -4.92846370e-01
2.65002251e-01 3.47031772e-01 1.01415157e+00 -1.04967761e+00
1.69035286e-01 -4.64665771e-01 4.08676356e-01 -5.56588948e-01
7.93851674e-01 -1.12262928e+00 9.88702476e-01 5.25783479e-01
-3.35962176e-02 2.92620391e-01 2.78125014e-02 6.00913823e-01
4.26076114e-01 3.30441415e-01 6.48423910e-01 -6.29310727e-01
-7.04015374e-01 5.46921194e-01 -6.08865693e-02 2.32331753e-01
1.14553058e+00 -6.35249496e-01 5.59827983e-01 -1.31051481e-01
-1.16630280e+00 -1.39236718e-01 1.12192774e+00 5.86118281e-01
8.82136345e-01 -1.44340384e+00 -6.87788308e-01 4.54314739e-01
-1.66149978e-02 5.01357734e-01 2.34684810e-01 6.89959288e-01
-7.84660399e-01 3.18592250e-01 -1.98258996e-01 -1.05724251e+00
-1.70176256e+00 3.71943474e-01 6.95266724e-01 1.74118131e-02
-1.04831648e+00 8.16303194e-01 4.55575198e-01 -9.58659112e-01
1.81035787e-01 -4.89186883e-01 1.36833236e-01 -4.95678514e-01
2.39028737e-01 6.30580902e-01 -3.54623944e-02 -1.56013215e+00
-5.57178855e-01 1.42760038e+00 5.52410841e-01 -5.07890761e-01
1.25700998e+00 -4.11577851e-01 1.34710848e-01 5.92889488e-01
1.05598176e+00 -3.87619659e-02 -1.86586177e+00 -3.93336207e-01
-2.33312175e-01 -7.92458057e-01 -2.81002611e-01 -5.23526609e-01
-7.90204108e-01 6.94881320e-01 5.25022566e-01 -5.33582270e-01
9.29568589e-01 1.82243660e-01 7.20344603e-01 9.34999064e-02
8.43889534e-01 -8.94472182e-01 3.67601752e-01 3.55563968e-01
1.19362962e+00 -1.04878271e+00 1.48013756e-01 -5.76470256e-01
-6.75649464e-01 7.71552742e-01 7.06639528e-01 -2.34440818e-01
5.39296746e-01 1.09640412e-01 8.91277045e-02 -4.17324364e-01
-3.04533601e-01 -3.46026989e-03 7.22783506e-01 6.38838530e-01
3.83154638e-02 1.02840699e-02 3.21083665e-01 -2.26867036e-03
-3.70843112e-01 -1.68978050e-01 2.69757032e-01 1.27848065e+00
1.94642823e-02 -8.70121360e-01 -9.16248143e-01 -3.48169416e-01
-2.07761243e-01 4.84197706e-01 -7.20235288e-01 8.47905338e-01
2.41822734e-01 5.63404620e-01 -9.31939334e-02 -3.21492493e-01
8.92328858e-01 -5.57219125e-02 8.97111654e-01 -7.85704553e-01
-8.11884582e-01 3.18289340e-01 -2.30546445e-01 -1.16845775e+00
-8.49756777e-01 -1.26192081e+00 -1.24407899e+00 -2.21251510e-02
-6.02257028e-02 -3.67339551e-01 4.83683497e-01 9.09001648e-01
1.38067305e-01 1.36338949e-01 2.97522545e-01 -1.67059183e+00
-1.52188063e-01 -3.41143131e-01 -3.30253363e-01 8.47833514e-01
7.70797491e-01 -9.29471076e-01 -3.16961676e-01 3.70865196e-01] | [7.091970443725586, -1.0710183382034302] |
46d1b4fe-819c-4a51-afcd-3d39e201f6ce | dear-xai-community-we-need-to-talk | 2306.04292 | null | https://arxiv.org/abs/2306.04292v1 | https://arxiv.org/pdf/2306.04292v1.pdf | Dear XAI Community, We Need to Talk! Fundamental Misconceptions in Current XAI Research | Despite progress in the field, significant parts of current XAI research are still not on solid conceptual, ethical, or methodological grounds. Unfortunately, these unfounded parts are not on the decline but continue to grow. Many explanation techniques are still proposed without clarifying their purpose. Instead, they are advertised with ever more fancy-looking heatmaps or only seemingly relevant benchmarks. Moreover, explanation techniques are motivated with questionable goals, such as building trust, or rely on strong assumptions about the 'concepts' that deep learning algorithms learn. In this paper, we highlight and discuss these and other misconceptions in current XAI research. We also suggest steps to make XAI a more substantive area of research. | ['Gunnar König', 'Timo Freiesleben'] | 2023-06-07 | null | null | null | null | ['misconceptions'] | ['miscellaneous'] | [-2.66316403e-02 6.42070830e-01 -8.42429936e-01 -7.67665923e-01
-1.95572317e-01 -4.46324378e-01 6.49997652e-01 2.33126804e-01
2.36518704e-03 9.41123426e-01 3.57707858e-01 -9.51799154e-01
-3.81871611e-01 -3.80128950e-01 -8.00956964e-01 -2.57230997e-01
1.14260413e-01 3.23686481e-01 -2.32499287e-01 8.27752426e-03
6.74899101e-01 1.71599999e-01 -1.36208665e+00 2.71703959e-01
8.07738841e-01 6.34393990e-01 -7.18512475e-01 2.32006595e-01
-2.90244848e-01 1.14027119e+00 -6.65895462e-01 -7.16186404e-01
-4.81354445e-02 -4.03288007e-01 -1.08410680e+00 -5.07966161e-01
5.21396756e-01 -5.33181846e-01 2.16670260e-02 8.90033066e-01
-1.53938383e-01 -3.37815464e-01 3.33914548e-01 -1.92894340e+00
-8.49964321e-01 9.78645027e-01 -6.09722078e-01 -7.58676976e-03
2.86203437e-02 2.65728354e-01 1.25075161e+00 -4.34142590e-01
5.41201353e-01 1.07235694e+00 5.86823821e-01 6.83863163e-01
-8.88559222e-01 -1.06730831e+00 4.56473142e-01 3.84867847e-01
-7.79425025e-01 -4.69668984e-01 7.95347452e-01 -2.95820326e-01
8.67471159e-01 6.09627008e-01 7.03079641e-01 1.40589166e+00
3.61290336e-01 9.38303590e-01 1.12511981e+00 -3.84415627e-01
3.43425483e-01 3.45111102e-01 6.75549626e-01 6.12143993e-01
1.20282090e+00 3.34232002e-01 -7.13858187e-01 -2.77137220e-01
7.00280726e-01 -1.46156587e-02 -1.62717909e-01 -4.60066289e-01
-9.88001049e-01 1.04431951e+00 2.66346484e-01 5.14584661e-01
-3.23030800e-01 5.30437589e-01 3.29848081e-01 1.79003984e-01
1.67764008e-01 8.00741136e-01 -6.36447787e-01 -5.43477118e-01
-8.67144942e-01 3.87602031e-01 6.99081659e-01 6.80370033e-01
7.33358383e-01 6.32486641e-02 3.93878043e-01 5.95208667e-02
6.63130581e-01 -6.80293441e-02 2.56883740e-01 -1.16086924e+00
-4.84740175e-02 6.39192581e-01 2.22470060e-01 -9.62359488e-01
-4.23724353e-01 -5.53610623e-01 -4.41023618e-01 6.95710599e-01
5.31009495e-01 -1.78899378e-01 -8.37918043e-01 1.77787316e+00
-1.94838583e-01 -5.35027124e-02 -2.17586800e-01 8.84165287e-01
5.59159279e-01 2.07865313e-01 3.46530110e-01 6.66552782e-02
1.11941493e+00 -8.63448024e-01 -7.78651536e-01 -5.90039670e-01
5.08435667e-01 -5.58439374e-01 1.28374624e+00 6.23779714e-01
-1.33184862e+00 -4.75379312e-03 -1.23939335e+00 -2.84134205e-02
-2.82681555e-01 -2.93965340e-01 1.35097647e+00 1.16501737e+00
-7.93597102e-01 4.97740716e-01 -8.65966558e-01 -2.37999797e-01
6.40832067e-01 2.31211707e-01 -1.37179894e-02 1.12610050e-01
-8.84266376e-01 9.83680964e-01 1.12923063e-01 8.05296749e-02
-6.95972562e-01 -8.86983812e-01 -5.34494400e-01 3.19581866e-01
5.17881095e-01 -7.92086184e-01 1.23920023e+00 -1.13755405e+00
-1.05719602e+00 5.04726887e-01 -1.65749993e-02 -5.34962356e-01
4.86213356e-01 -4.89893287e-01 -3.32924128e-01 -1.68457031e-01
7.49446377e-02 6.95692301e-01 4.49280620e-01 -1.64109993e+00
-6.74368560e-01 -3.20612013e-01 4.82021540e-01 -1.07847318e-01
-1.70713753e-01 1.38167948e-01 6.48728311e-02 -3.29644412e-01
1.76625848e-01 -9.44705069e-01 -1.22288741e-01 2.18997151e-01
-3.65882128e-01 -1.79331005e-01 6.90946996e-01 -2.14316949e-01
1.10239649e+00 -2.07563043e+00 -4.04568046e-01 1.54082179e-01
5.80039144e-01 -5.12942672e-02 1.79596886e-01 3.71767789e-01
-2.70201534e-01 9.45753634e-01 9.02937576e-02 1.59897998e-01
2.66250730e-01 -1.85368601e-02 -4.64469582e-01 4.04130429e-01
2.49276427e-03 1.02290332e+00 -9.99702036e-01 -2.19483837e-01
1.43048942e-01 3.12490076e-01 -4.91117716e-01 -2.04419315e-01
-2.78194994e-01 9.70990360e-02 -5.71768880e-01 6.62530124e-01
4.02558953e-01 -5.36634147e-01 3.30171824e-01 -8.18859339e-02
-2.89368570e-01 8.26648951e-01 -5.99456489e-01 1.57727361e+00
-1.12801068e-01 9.46390867e-01 -9.31857526e-02 -8.67732346e-01
4.64052707e-01 2.42623940e-01 4.46656734e-01 -4.20241475e-01
2.58352637e-01 2.39412442e-01 5.45088172e-01 -3.08027148e-01
3.54867488e-01 -3.44655037e-01 2.49263287e-01 1.12013924e+00
-4.82304513e-01 1.33893341e-01 -3.46276373e-01 2.57827014e-01
1.11088264e+00 2.86681443e-01 1.80342421e-01 -2.50515670e-01
-8.94685090e-02 3.59978616e-01 5.19066572e-01 7.42475927e-01
-3.99886101e-01 4.09962803e-01 7.75064886e-01 -7.60706902e-01
-9.50884044e-01 -8.31337512e-01 -1.91556200e-01 7.57418156e-01
9.33342874e-02 -6.85986638e-01 -6.25089884e-01 -1.19559288e+00
-1.26878098e-01 1.42788339e+00 -9.03526604e-01 -3.13463122e-01
-8.80567580e-02 -5.86909473e-01 4.41430718e-01 6.63845658e-01
4.23534125e-01 -9.88549829e-01 -1.27481294e+00 -1.18563741e-01
4.83349897e-02 -3.91576529e-01 2.82050103e-01 2.98839003e-01
-1.07812440e+00 -1.12896264e+00 -2.99648028e-02 -9.15637985e-02
5.80480933e-01 6.63635612e-01 1.20442307e+00 6.63251996e-01
5.29337116e-02 2.67472833e-01 -3.33039224e-01 -8.23881984e-01
-2.62667030e-01 1.14413975e-02 -1.55316725e-01 -7.47153461e-01
7.96589553e-01 -5.62906861e-01 -7.09813595e-01 1.68992862e-01
-8.05405200e-01 1.34421792e-02 8.23804796e-01 7.03703165e-01
-1.74410015e-01 -2.42075534e-03 4.57089841e-01 -1.26868308e+00
6.71818852e-01 -5.81985056e-01 -8.89273509e-02 4.38064665e-01
-1.42477250e+00 2.39716783e-01 3.83437686e-02 -2.25647047e-01
-1.13934994e+00 -7.68792868e-01 1.55603081e-01 -9.44137350e-02
-3.79225761e-01 6.70582056e-01 -4.64885123e-02 1.14039093e-01
9.33150411e-01 -5.51709831e-01 3.62132639e-02 -3.02428037e-01
3.67550522e-01 3.29598635e-01 1.75106347e-01 -8.37024212e-01
7.87797809e-01 4.70318586e-01 -4.90104347e-01 -3.54018807e-01
-8.63942206e-01 3.42165530e-01 -5.02841696e-02 -1.60299674e-01
5.12862802e-01 -3.65132272e-01 -8.63410115e-01 -2.24240974e-01
-1.02097869e+00 -4.09628034e-01 -3.42167258e-01 5.30062020e-01
-2.35181913e-01 2.13886619e-01 -4.64309007e-01 -7.39534676e-01
-1.79405242e-01 -1.16690445e+00 4.06436354e-01 4.52634305e-01
-8.90212893e-01 -8.88593078e-01 -7.29290694e-02 6.01925969e-01
7.23204911e-01 1.31863728e-01 1.34397829e+00 -7.46325135e-01
-7.49596596e-01 -2.30042547e-01 -4.17544574e-01 -3.39874290e-02
1.75763309e-01 4.07255918e-01 -1.07126451e+00 3.48602608e-02
1.05972424e-01 -4.77770507e-01 3.16843748e-01 4.53859776e-01
1.16001952e+00 -5.33050478e-01 -4.29406881e-01 3.31141859e-01
1.20868874e+00 5.64803243e-01 6.11787617e-01 8.18614900e-01
2.92226970e-01 8.31222594e-01 6.43337190e-01 3.21331322e-01
3.16851974e-01 1.85556233e-01 7.34111965e-01 -9.08490941e-02
6.19496368e-02 -3.85293931e-01 1.55099526e-01 7.68482015e-02
-3.62815976e-01 -1.18521258e-01 -1.35491979e+00 4.17861819e-01
-1.83581245e+00 -1.07017601e+00 -3.32870573e-01 2.09381366e+00
4.79411811e-01 7.94431090e-01 1.47644887e-02 4.21296097e-02
3.49422663e-01 6.15305714e-02 -8.52645755e-01 -5.93137264e-01
8.42087567e-02 7.43115470e-02 2.10753694e-01 4.12358701e-01
-4.46853518e-01 8.51967394e-01 7.59460783e+00 5.51674590e-02
-1.28217769e+00 1.04568992e-02 8.78117919e-01 -1.28218587e-02
-1.03627706e+00 6.20082140e-01 -2.55366653e-01 1.47157699e-01
9.25052822e-01 -3.33092391e-01 2.51996696e-01 1.12873411e+00
3.06888279e-02 -1.92039207e-01 -1.16301405e+00 4.38507855e-01
-1.99636400e-01 -1.62664509e+00 -1.94076598e-01 4.26266015e-01
3.06738555e-01 -5.08880392e-02 4.47241545e-01 1.73745513e-01
7.02977955e-01 -1.30235851e+00 1.00284934e+00 1.87763631e-01
3.48903716e-01 -7.70029485e-01 6.52391970e-01 -5.26375324e-02
-1.54778391e-01 -7.94054940e-02 -2.51108497e-01 -5.70318639e-01
-3.10587853e-01 2.59357125e-01 -6.97996497e-01 2.68763870e-01
8.47330511e-01 4.27347511e-01 -5.97673833e-01 9.36336577e-01
-6.32722437e-01 1.14102125e+00 -3.86901125e-02 -2.58735627e-01
4.59531903e-01 -8.66111591e-02 5.87145723e-02 6.27148569e-01
-7.23516718e-02 4.66062985e-02 -5.40449798e-01 1.06855869e+00
1.78502455e-01 -2.52186775e-01 -6.66019380e-01 -4.01348323e-01
4.99694556e-01 1.13958395e+00 -7.48791635e-01 -2.39836395e-01
-7.76903033e-01 4.83968645e-01 8.56856108e-02 3.90247196e-01
-1.02582502e+00 8.90163630e-02 1.06292117e+00 1.25670165e-01
-4.26940769e-01 -1.64364547e-01 -1.03793657e+00 -1.09876060e+00
-2.03908280e-01 -1.17006481e+00 2.75345832e-01 -9.18238997e-01
-9.78972554e-01 3.08452129e-01 1.61746740e-01 -6.46387100e-01
-2.30423346e-01 -4.77444053e-01 -8.92117023e-01 3.32733870e-01
-1.45380056e+00 -1.14638996e+00 -3.25074077e-01 2.10255221e-03
2.20212236e-01 6.24812581e-02 8.00400615e-01 -8.64569470e-02
-6.80833757e-01 5.80671310e-01 -1.16367877e-01 -3.21220338e-01
7.43955195e-01 -1.07203782e+00 5.68588555e-01 4.88123864e-01
3.69094133e-01 1.28793645e+00 1.12044895e+00 -5.40787041e-01
-1.22066665e+00 -2.58864582e-01 7.57706106e-01 -7.44054914e-01
7.49976814e-01 -1.83914110e-01 -9.55360532e-01 1.34150589e+00
7.91869879e-01 -6.71175480e-01 1.05671215e+00 1.04112816e+00
-5.12918115e-01 1.40817195e-01 -9.54510510e-01 7.10577667e-01
9.01027739e-01 -2.93369919e-01 -5.95901489e-01 2.86290377e-01
5.07034481e-01 1.87481265e-03 -4.85929310e-01 3.26068878e-01
8.54538083e-01 -1.31230724e+00 8.95202637e-01 -1.16184127e+00
7.55611837e-01 1.15213498e-01 1.08921476e-01 -8.96727026e-01
-3.75619709e-01 -5.96679926e-01 1.36789352e-01 1.20256090e+00
7.23076761e-01 -5.63886225e-01 1.42645812e+00 1.56897688e+00
-2.20468685e-01 -7.05233872e-01 -3.88567507e-01 -4.92693067e-01
3.29224586e-01 -7.40886152e-01 7.82045007e-01 1.42427027e+00
5.63300431e-01 5.45134008e-01 -2.83663750e-01 -7.29105473e-02
6.49662912e-01 2.22648889e-01 9.21272278e-01 -1.22917008e+00
6.88298941e-02 -7.74247706e-01 -1.26355678e-01 -4.94412452e-01
-1.70738921e-02 -5.23747504e-01 -3.11967671e-01 -1.61199939e+00
4.14473444e-01 -5.15240073e-01 -4.95220959e-01 9.30596054e-01
-1.75619975e-01 -6.78150877e-02 1.89766492e-04 3.51139992e-01
-3.33930820e-01 2.43387356e-01 5.45294285e-01 -1.60330102e-01
-5.66721931e-02 -3.37520272e-01 -1.52413893e+00 1.02127349e+00
1.23235273e+00 -3.01704943e-01 -8.75626504e-01 -6.38842702e-01
7.13812470e-01 -1.26974046e-01 5.06057978e-01 -9.18780088e-01
4.54621725e-02 -6.66956007e-01 2.99094588e-01 -3.15651923e-01
9.53579023e-02 -8.07079613e-01 4.32442665e-01 5.54395616e-01
-6.26730025e-01 2.72773683e-01 3.88711274e-01 1.63439140e-01
1.15841776e-01 -4.08271372e-01 4.80666578e-01 4.36544754e-02
-4.54409033e-01 -7.08832070e-02 -2.38273352e-01 -2.10556000e-01
9.98184323e-01 -2.22189814e-01 -6.95288479e-01 -5.76933086e-01
-3.07964385e-01 1.70764271e-02 8.91375422e-01 4.53242004e-01
5.59293270e-01 -8.23511422e-01 -3.33283007e-01 -1.79725930e-01
3.72103378e-02 -3.05616766e-01 6.13689274e-02 6.06804550e-01
-3.96957934e-01 6.02036893e-01 -6.18362427e-01 2.33345479e-02
-8.62988532e-01 7.05063522e-01 3.61679979e-02 1.54376090e-01
-7.63857961e-01 5.55234313e-01 3.70922804e-01 -5.09681925e-03
3.45538467e-01 -2.95283228e-01 1.17031194e-03 -2.91260928e-01
4.03056055e-01 2.04013035e-01 -3.23620379e-01 1.03293248e-01
-4.44394141e-01 -1.09551296e-01 -4.34102863e-01 -3.49217296e-01
1.54639339e+00 1.69773489e-01 -2.65339632e-02 4.66125011e-01
5.52482724e-01 9.66144353e-02 -1.06975734e+00 3.69794905e-01
2.13702217e-01 -7.79345989e-01 1.34382531e-01 -1.29149389e+00
-1.03931165e+00 1.08672619e+00 3.12362641e-01 3.28708112e-01
6.69211924e-01 7.05844583e-03 3.00421000e-01 1.55842215e-01
3.37575525e-01 -8.80846083e-01 -3.42608206e-02 -2.13209391e-01
9.05006945e-01 -1.09624493e+00 4.49482739e-01 -1.89330325e-01
-6.57186866e-01 1.10028863e+00 8.21315527e-01 3.25262308e-01
6.33278549e-01 2.00050220e-01 1.47087947e-01 -6.75620854e-01
-1.02212310e+00 3.14886838e-01 -2.88487881e-01 5.54462194e-01
9.85688508e-01 -1.25961572e-01 -8.50418925e-01 5.92692435e-01
-3.54920477e-01 5.97134754e-02 7.23927200e-01 9.35007572e-01
-3.91190678e-01 -1.22443604e+00 -3.74842763e-01 4.94009346e-01
-4.72333729e-01 -6.10792860e-02 -6.92470789e-01 1.27521074e+00
-1.30277470e-01 9.51557934e-01 -2.85944313e-01 -5.22891641e-01
-5.24574816e-02 8.48623216e-02 2.76371121e-01 -2.41469502e-01
-4.87299979e-01 -2.09907800e-01 2.96753049e-01 -5.66496551e-01
-2.15794832e-01 -9.30301189e-01 -1.26729763e+00 -8.75266969e-01
-3.21959287e-01 4.16037321e-01 1.19529283e+00 9.45346177e-01
3.71681482e-01 2.63899803e-01 -5.21154027e-04 -1.83020860e-01
-4.07616615e-01 -6.40580595e-01 -1.37314320e-01 6.03841022e-02
-9.43191648e-02 -6.90283656e-01 -2.38717586e-01 -4.00233954e-01] | [8.86583137512207, 6.009974002838135] |
57a87d9d-514b-4510-9193-237ed82908ca | blind-image-deblurring-via-reweighted-graph | 1712.08877 | null | http://arxiv.org/abs/1712.08877v1 | http://arxiv.org/pdf/1712.08877v1.pdf | Blind Image Deblurring via Reweighted Graph Total Variation | Blind image deblurring, i.e., deblurring without knowledge of the blur
kernel, is a highly ill-posed problem. The problem can be solved in two parts:
i) estimate a blur kernel from the blurry image, and ii) given estimated blur
kernel, de-convolve blurry input to restore the target image. In this paper, by
interpreting an image patch as a signal on a weighted graph, we first argue
that a skeleton image---a proxy that retains the strong gradients of the target
but smooths out the details---can be used to accurately estimate the blur
kernel and has a unique bi-modal edge weight distribution. We then design a
reweighted graph total variation (RGTV) prior that can efficiently promote
bi-modal edge weight distribution given a blurry patch. However, minimizing a
blind image deblurring objective with RGTV results in a non-convex
non-differentiable optimization problem. We propose a fast algorithm that
solves for the skeleton image and the blur kernel alternately. Finally with the
computed blur kernel, recent non-blind image deblurring algorithms can be
applied to restore the target image. Experimental results show that our
algorithm can robustly estimate the blur kernel with large kernel size, and the
reconstructed sharp image is competitive against the state-of-the-art methods. | ['Xian-Ming Liu', 'Yuanchao Bai', 'Wen Gao', 'Gene Cheung'] | 2017-12-24 | null | null | null | null | ['blind-image-deblurring'] | ['computer-vision'] | [ 3.09181184e-01 -3.40762079e-01 2.04046875e-01 -7.34170452e-02
-6.51804566e-01 -4.21219289e-01 5.76268993e-02 -9.64210689e-01
5.03489515e-03 7.30338991e-01 6.24482274e-01 4.53006700e-02
-3.74625534e-01 -1.92011431e-01 -6.96065545e-01 -9.09435213e-01
1.22503318e-01 -2.24853173e-01 -2.84553207e-02 2.33467713e-01
4.09647346e-01 1.52994975e-01 -1.03288078e+00 -8.66184309e-02
1.42842734e+00 8.74953270e-01 5.89975178e-01 9.62182462e-01
3.02523553e-01 1.10084975e+00 -3.53021055e-01 -1.39689058e-01
2.68619090e-01 -8.39722753e-01 -8.85988414e-01 6.28605068e-01
7.59880304e-01 -6.33679152e-01 -8.31895590e-01 1.79600310e+00
3.04723799e-01 2.07926571e-01 6.67362154e-01 -6.45110548e-01
-1.49723446e+00 3.24177653e-01 -1.17699718e+00 3.18213016e-01
1.20505214e-01 1.42772749e-01 4.90547359e-01 -1.14196491e+00
3.15578848e-01 1.04961991e+00 6.87071443e-01 4.19884980e-01
-1.36844301e+00 -2.63760775e-01 -2.15357333e-01 4.06893671e-01
-1.54399931e+00 -7.40846157e-01 9.61892009e-01 -3.69372100e-01
1.11881547e-01 6.19948864e-01 3.58880967e-01 6.89593196e-01
3.03445041e-01 5.13218105e-01 1.39585435e+00 -1.92689463e-01
1.64448380e-01 -3.55710953e-01 2.55680501e-01 7.40056276e-01
4.73935992e-01 9.68691409e-02 -3.59132677e-01 -1.98203698e-01
1.18931651e+00 1.50447905e-01 -1.43612540e+00 -2.83553690e-01
-1.33110332e+00 3.14442992e-01 6.35289133e-01 2.48501271e-01
-4.61039454e-01 3.20198208e-01 -2.20467269e-01 2.57051557e-01
6.52433157e-01 4.21074450e-01 -2.62677699e-01 4.84772474e-02
-1.23788261e+00 -5.26722930e-02 6.68149114e-01 6.80167437e-01
1.00643384e+00 -3.29428725e-02 -4.00342852e-01 1.12695444e+00
4.81693029e-01 6.53110862e-01 4.42394376e-01 -1.09578812e+00
1.27237886e-01 5.79816699e-02 5.46115100e-01 -8.54731381e-01
2.60736138e-01 -2.17752337e-01 -1.32661152e+00 2.31434956e-01
6.16988063e-01 -1.92333356e-01 -1.12621415e+00 1.53600478e+00
1.02407821e-01 9.65209424e-01 -9.75046158e-02 1.64112175e+00
4.89040732e-01 5.64900994e-01 -5.93442559e-01 -4.89295334e-01
1.42086649e+00 -1.16855276e+00 -1.04378688e+00 -5.83772182e-01
-3.97375733e-01 -1.02872217e+00 8.11171532e-01 1.98827371e-01
-1.33115613e+00 -5.41527867e-01 -9.35795546e-01 -2.77626127e-01
3.63199890e-01 1.72240600e-01 1.60360336e-01 4.90783691e-01
-1.30845010e+00 5.25895655e-01 -6.56632662e-01 -1.09644234e-01
2.69638568e-01 -1.06769623e-02 -1.54094875e-01 -6.89371407e-01
-1.00544810e+00 1.13292050e+00 -7.08357170e-02 3.80987316e-01
-9.62195873e-01 -8.20853889e-01 -8.82751882e-01 4.32908796e-02
2.22410738e-01 -1.01644540e+00 9.49155509e-01 -1.05914056e+00
-1.45855343e+00 6.33038044e-01 -4.36599910e-01 -4.07664105e-02
5.10029614e-01 -4.69459116e-01 -3.91497403e-01 2.30842754e-01
5.02360016e-02 -5.86225614e-02 1.82963538e+00 -1.59096909e+00
-2.11642489e-01 -3.01676840e-01 -3.87723655e-01 3.40003163e-01
6.23479486e-02 -8.41201991e-02 -7.65946805e-01 -1.11331749e+00
2.69437373e-01 -6.97698832e-01 -1.11239888e-01 3.19729857e-02
-3.56344342e-01 5.76391876e-01 8.76894414e-01 -1.27518857e+00
1.37114108e+00 -2.26894784e+00 5.88647485e-01 -1.05411261e-01
6.57595873e-01 1.25409603e-01 -9.76518989e-02 -2.27984339e-02
-2.35388890e-01 -3.48144561e-01 -8.23484719e-01 -9.53771546e-02
-4.21006143e-01 -2.40543336e-02 -3.65593195e-01 1.07190847e+00
-1.45321012e-01 1.05288577e+00 -1.11274028e+00 -1.11293040e-01
3.18141401e-01 6.38966799e-01 -2.33962521e-01 6.37394428e-01
3.59042615e-01 4.27105606e-01 -1.05790548e-01 4.45989072e-01
1.41486168e+00 -5.16629577e-01 -1.85137317e-01 -7.52184689e-01
-5.33579625e-02 -6.16832733e-01 -1.18278396e+00 1.84484208e+00
-3.11382800e-01 7.54823089e-01 7.70696402e-01 -6.93769157e-01
5.05928934e-01 5.41956127e-02 1.91462398e-01 8.69683698e-02
-2.88323425e-02 2.61781514e-01 -3.35753500e-01 -8.67038667e-01
4.19596940e-01 -1.05247922e-01 4.19478863e-01 2.84502327e-01
-8.74643326e-02 -2.79284000e-01 -3.41634691e-01 2.01379005e-02
1.14637625e+00 -1.25837028e-01 9.48652402e-02 -5.28858423e-01
6.05493069e-01 -4.42158639e-01 2.25480065e-01 8.38663697e-01
-2.86709666e-01 1.26008034e+00 -1.39914960e-01 -1.16337307e-01
-8.18676054e-01 -1.14304793e+00 -1.06569484e-01 5.44596314e-01
9.14134800e-01 2.34573811e-01 -1.20558679e+00 -4.96140301e-01
-7.31105879e-02 4.94404495e-01 -5.56464970e-01 -2.70549059e-01
-4.30877060e-01 -9.60866749e-01 4.75934744e-02 -4.40724902e-02
8.62885594e-01 -4.19303030e-01 -1.16251923e-01 4.13698331e-03
-5.44369161e-01 -8.93128216e-01 -1.60906327e+00 -2.46340811e-01
-6.90699697e-01 -1.26079988e+00 -1.62716627e+00 -1.01885498e+00
1.08488071e+00 1.12659824e+00 8.53011906e-01 3.66658457e-02
-1.67042807e-01 5.49705029e-01 1.45095692e-03 4.78739053e-01
-2.59043217e-01 -7.58248329e-01 -7.62657300e-02 7.85416067e-01
-1.67520016e-01 -5.13471365e-01 -1.04830778e+00 5.00384033e-01
-1.05716121e+00 2.73585767e-01 7.07132220e-01 1.02853322e+00
1.96936607e-01 3.52780819e-01 2.48827472e-01 -5.62865019e-01
9.73033130e-01 -3.47406000e-01 -3.71894479e-01 4.72476780e-01
-7.00081348e-01 1.63506806e-01 2.63058573e-01 -7.00236142e-01
-1.43377590e+00 -1.37424603e-01 5.56427300e-01 -9.96413827e-01
7.83702284e-02 4.93133545e-01 -3.95344123e-02 -5.01107574e-01
7.49881685e-01 4.48763102e-01 2.74669558e-01 -7.39186466e-01
8.17266047e-01 8.13615859e-01 1.01594734e+00 -3.05290580e-01
1.17800009e+00 5.09213865e-01 -1.65867940e-01 -9.32666540e-01
-7.93721378e-01 -6.95481062e-01 -2.33252674e-01 -3.21024507e-01
7.97356665e-01 -1.04468310e+00 -4.17161524e-01 1.08894515e+00
-1.24214053e+00 -3.43529224e-01 -8.23670328e-02 5.74235141e-01
-3.61600071e-01 1.02151263e+00 -9.77609932e-01 -6.81487918e-01
-2.35084563e-01 -1.07829177e+00 8.01560402e-01 4.57806855e-01
2.60722786e-01 -1.01409578e+00 7.45799392e-03 4.46659684e-01
7.69035399e-01 -1.73163921e-01 3.57524604e-01 4.88474280e-01
-8.17702949e-01 5.95791116e-02 -9.27449048e-01 5.41237831e-01
7.43167937e-01 -5.49303889e-01 -1.11283958e+00 -4.74574894e-01
7.31623113e-01 1.96566015e-01 1.11092186e+00 9.51587021e-01
1.03770530e+00 -7.40537226e-01 -1.59925237e-01 9.48609054e-01
1.43531001e+00 -3.97244781e-01 7.98824608e-01 -5.46179488e-02
1.09353161e+00 1.76220730e-01 2.14710742e-01 1.31527156e-01
2.03593120e-01 5.19892395e-01 3.31168622e-01 -1.61889717e-01
-7.71551013e-01 7.80041562e-03 4.26740676e-01 8.41970921e-01
-1.58296585e-01 7.88155869e-02 -4.20635551e-01 7.07848072e-01
-2.11094570e+00 -9.16515589e-01 -4.29078639e-01 2.21232414e+00
1.24138856e+00 -5.75907171e-01 -4.08671707e-01 -3.16270232e-01
1.22237527e+00 3.58161122e-01 -6.97783232e-01 4.65705514e-01
-1.88357607e-01 -8.11581984e-02 7.87447214e-01 1.08835399e+00
-9.72001374e-01 6.75236046e-01 6.18134451e+00 9.23074305e-01
-9.37449098e-01 1.80125818e-01 5.69886804e-01 3.21499944e-01
-1.87529325e-01 6.03808239e-02 1.26112893e-01 9.11131442e-01
4.90936399e-01 -4.07767653e-01 1.45559645e+00 5.15386581e-01
3.78441989e-01 -2.35727578e-01 -6.91926479e-01 1.49974835e+00
2.69898623e-01 -1.11208212e+00 -2.78240114e-01 2.13181097e-02
9.10866499e-01 -1.16627581e-01 1.13474935e-01 -3.75235856e-01
3.30336422e-01 -9.25662935e-01 6.91485107e-01 9.81519580e-01
1.02026033e+00 -1.90602377e-01 4.94049460e-01 2.07224622e-01
-9.59053755e-01 1.43906325e-01 -4.38303649e-01 5.98336644e-02
1.84519693e-01 1.15081346e+00 -1.21794909e-01 6.14600420e-01
7.93967545e-01 9.72840369e-01 -3.83926570e-01 1.45541811e+00
-3.03575248e-01 5.04268706e-01 3.48139256e-01 6.32834673e-01
-2.90864915e-01 -7.72871614e-01 9.40456927e-01 1.16816592e+00
6.08944833e-01 3.22844535e-01 -2.71756440e-01 1.16803396e+00
-1.63893476e-01 -5.83953202e-01 -1.87004223e-01 3.96126568e-01
1.81577474e-01 1.25209010e+00 -2.58503497e-01 -3.37164789e-01
-3.82595807e-01 1.88746619e+00 8.20150003e-02 1.13131964e+00
-7.87736595e-01 -4.17216599e-01 9.68871832e-01 -1.25844225e-01
3.14449936e-01 -9.79089960e-02 -3.32290143e-01 -1.80156600e+00
-6.25656471e-02 -8.16689134e-01 1.06086610e-02 -1.39936996e+00
-1.70887947e+00 5.14004469e-01 -3.17244500e-01 -1.20210874e+00
3.65175784e-01 -3.36132556e-01 -6.99767232e-01 1.37261999e+00
-1.76520038e+00 -1.07881403e+00 -7.69703746e-01 7.20483422e-01
5.62054396e-01 2.80327111e-01 1.95642650e-01 2.05027178e-01
-4.99151021e-01 1.17167272e-02 4.06799406e-01 -7.67372325e-02
9.85309720e-01 -1.43658006e+00 1.87466815e-01 1.33502042e+00
-2.96478599e-01 6.54937863e-01 9.12080705e-01 -7.75242329e-01
-1.68935752e+00 -1.19532895e+00 3.53737026e-01 -2.58729458e-01
8.22481990e-01 2.30245382e-01 -1.29005110e+00 4.78157610e-01
6.76765859e-01 3.75914484e-01 -4.88450825e-02 -5.78091800e-01
-2.02317446e-01 -9.03182849e-02 -1.16930687e+00 5.17052412e-01
9.58598912e-01 -6.94381356e-01 -7.47885585e-01 1.67154327e-01
6.63601577e-01 -7.09312737e-01 -5.75767517e-01 1.63064286e-01
1.64004743e-01 -9.38655496e-01 1.26075184e+00 -1.19517356e-01
3.69936079e-01 -8.65283728e-01 1.48328438e-01 -1.84634590e+00
-8.77772391e-01 -1.20974362e+00 -6.96359813e-01 1.06729472e+00
-1.55273914e-01 -6.14624441e-01 2.68525332e-01 7.11767733e-01
-1.73862010e-01 -2.46489510e-01 -6.81246102e-01 -6.12231851e-01
-5.56334376e-01 7.13435039e-02 2.76215911e-01 1.10991883e+00
-1.62303969e-01 2.12472290e-01 -9.71783698e-01 5.86109281e-01
1.38105392e+00 1.28306225e-01 3.31393510e-01 -7.14902818e-01
-2.42156208e-01 -3.59159470e-01 2.13836376e-02 -1.63631546e+00
-5.21493852e-02 -6.06420457e-01 4.47359741e-01 -1.57232010e+00
5.07246137e-01 7.82915670e-03 -1.95760131e-01 -9.25169140e-03
-8.59458625e-01 2.40380570e-01 -1.71143755e-01 7.32293606e-01
-2.53486037e-01 4.70389694e-01 1.69439757e+00 -3.55375886e-01
-3.98167111e-02 -8.78225192e-02 -8.80802155e-01 6.10899270e-01
2.50664055e-01 -2.14960471e-01 -2.80961394e-01 -6.14390790e-01
-2.44009823e-01 2.25598142e-01 6.11670792e-01 -6.53606296e-01
4.74195689e-01 -3.55315655e-01 4.14002240e-01 2.59507243e-02
1.50264993e-01 -8.77341807e-01 4.09033060e-01 1.80997580e-01
-7.01353922e-02 -5.88492692e-01 -1.74074575e-01 9.62332368e-01
-1.99532405e-01 -3.91910732e-01 1.07235515e+00 -5.98234870e-02
-5.29658377e-01 3.56235623e-01 -2.25380167e-01 7.91147873e-02
6.46367013e-01 -1.88787520e-01 -6.35887384e-01 -5.47192037e-01
-4.83057350e-01 -1.30210832e-01 8.21873426e-01 2.61207223e-01
8.38888466e-01 -1.42234123e+00 -8.14521551e-01 1.14721365e-01
-2.96818972e-01 -1.43362001e-01 5.72684646e-01 1.05487287e+00
-6.39022768e-01 -2.52602756e-01 7.92816654e-02 -2.96936065e-01
-1.02199697e+00 8.20966601e-01 5.75979531e-01 1.68406114e-01
-5.96506894e-01 1.02527070e+00 6.34636521e-01 2.98917890e-01
-3.14654559e-02 -2.97808528e-01 2.18168527e-01 -3.95021021e-01
8.78884673e-01 6.82024777e-01 -2.62311429e-01 -7.41788805e-01
-1.20990187e-01 9.48347509e-01 1.53398260e-01 -6.36722222e-02
1.18726337e+00 -8.70713770e-01 -7.24164963e-01 -1.66892290e-01
1.31851447e+00 3.22176784e-01 -1.73765075e+00 -3.76234740e-01
-2.37111881e-01 -1.13815939e+00 7.97624052e-01 -7.10015953e-01
-1.41003788e+00 3.75082016e-01 6.38408124e-01 4.44213212e-01
1.50217438e+00 4.93403450e-02 7.74711847e-01 -4.26259190e-01
4.04193811e-02 -5.69626570e-01 1.87623784e-01 1.47593126e-01
1.28966141e+00 -1.04541147e+00 7.77486190e-02 -5.34625709e-01
-5.26953936e-01 1.07454705e+00 2.92544544e-01 -2.30986625e-01
6.92295969e-01 6.86866697e-03 1.06170155e-01 -2.14418605e-01
-6.40038028e-02 -1.50538802e-01 8.59811127e-01 5.14133394e-01
7.41410777e-02 -1.01309463e-01 3.86354630e-03 3.31902087e-01
4.46348876e-01 3.23090971e-01 7.20788538e-01 4.62362677e-01
-6.11929417e-01 -4.65096653e-01 -1.03371501e+00 2.86048710e-01
-3.09558690e-01 -3.89270425e-01 -1.70100525e-01 -1.96104720e-01
-1.39275461e-01 1.21643376e+00 -3.25225592e-01 -1.98928803e-01
1.22411214e-01 -4.19478714e-01 6.03665709e-01 -2.20485806e-01
5.01046441e-02 4.13464755e-01 -3.51859897e-01 -4.81345505e-01
-4.15990502e-01 -3.08098465e-01 -5.63127339e-01 -4.20935273e-01
-6.19398594e-01 1.11351974e-01 2.19437107e-01 5.88081896e-01
4.05485809e-01 3.54954869e-01 7.05855191e-01 -1.21308720e+00
-5.34651220e-01 -1.03486407e+00 -1.00062990e+00 4.85471040e-01
1.07522988e+00 -3.02192658e-01 -1.02741969e+00 5.81341028e-01] | [11.610472679138184, -2.7526583671569824] |
074b9845-82ae-4047-b101-3ebd0a7358da | identitydp-differential-private | 2103.01745 | null | https://arxiv.org/abs/2103.01745v1 | https://arxiv.org/pdf/2103.01745v1.pdf | IdentityDP: Differential Private Identification Protection for Face Images | Because of the explosive growth of face photos as well as their widespread dissemination and easy accessibility in social media, the security and privacy of personal identity information becomes an unprecedented challenge. Meanwhile, the convenience brought by advanced identity-agnostic computer vision technologies is attractive. Therefore, it is important to use face images while taking careful consideration in protecting people's identities. Given a face image, face de-identification, also known as face anonymization, refers to generating another image with similar appearance and the same background, while the real identity is hidden. Although extensive efforts have been made, existing face de-identification techniques are either insufficient in photo-reality or incapable of well-balancing privacy and utility. In this paper, we focus on tackling these challenges to improve face de-identification. We propose IdentityDP, a face anonymization framework that combines a data-driven deep neural network with a differential privacy (DP) mechanism. This framework encompasses three stages: facial representations disentanglement, $\epsilon$-IdentityDP perturbation and image reconstruction. Our model can effectively obfuscate the identity-related information of faces, preserve significant visual similarity, and generate high-quality images that can be used for identity-agnostic computer vision tasks, such as detection, tracking, etc. Different from the previous methods, we can adjust the balance of privacy and utility through the privacy budget according to pratical demands and provide a diversity of results without pre-annotations. Extensive experiments demonstrate the effectiveness and generalization ability of our proposed anonymization framework. | ['Rong Xie', 'Ming Ding', 'Bo Liu', 'Li Song', 'Yunqian Wen'] | 2021-03-02 | null | null | null | null | ['face-anonymization'] | ['computer-vision'] | [ 2.12135896e-01 -6.31005317e-02 4.22696993e-02 -6.02950037e-01
-2.95094132e-01 -6.61074877e-01 4.32732701e-01 -3.46396148e-01
-3.38628769e-01 5.97525239e-01 1.06135145e-01 -3.15538682e-02
-1.18798241e-01 -7.78071404e-01 -4.25707191e-01 -8.94743145e-01
2.47596443e-01 6.98871836e-02 -5.12186527e-01 -9.09169670e-03
7.27495849e-02 8.22175324e-01 -1.57303560e+00 3.99397574e-02
8.53488803e-01 1.12425208e+00 -3.93827558e-01 1.04233310e-01
2.20310409e-02 1.68764904e-01 -4.14427489e-01 -1.00851893e+00
7.07517207e-01 -3.54120284e-01 -5.26322722e-01 2.51555145e-02
5.52746236e-01 -6.30637705e-01 -6.72161877e-01 1.42512560e+00
7.56968915e-01 -1.73176929e-01 2.17516169e-01 -1.63001037e+00
-1.18995559e+00 1.14917442e-01 -9.66504455e-01 3.30916680e-02
3.76362592e-01 1.75570101e-01 4.56864715e-01 -6.76929355e-01
4.30075973e-01 1.34091091e+00 7.05609083e-01 1.01059353e+00
-1.27600646e+00 -1.23443425e+00 4.61495854e-02 1.69056430e-01
-1.85111868e+00 -9.38489676e-01 7.37584710e-01 -2.69852489e-01
7.68607184e-02 4.17104185e-01 4.75805044e-01 1.17493665e+00
-1.78658962e-01 4.86565858e-01 1.01737773e+00 -7.69634694e-02
-6.39182776e-02 3.21839213e-01 -3.44921023e-01 5.05231917e-01
4.53505307e-01 1.93418548e-01 -3.76477867e-01 -2.80181170e-01
7.08920062e-01 3.62781793e-01 -5.20710707e-01 -4.13183033e-01
-8.24675918e-01 5.28109193e-01 3.54962111e-01 -1.34447381e-01
-3.70224476e-01 -1.01260811e-01 5.10450602e-01 2.31258899e-01
3.57039541e-01 1.52481928e-01 1.34997591e-01 2.53583670e-01
-7.62871742e-01 3.43603015e-01 4.90189463e-01 9.60538924e-01
8.54946554e-01 4.88195308e-02 -1.66095436e-01 7.09215760e-01
1.17218651e-01 5.37308931e-01 6.07332170e-01 -9.07701015e-01
2.92091846e-01 5.91876745e-01 1.41689450e-01 -1.60923350e+00
7.04366937e-02 -1.15091406e-01 -1.25286436e+00 2.50869215e-01
2.40683153e-01 -2.35854939e-01 -6.19282663e-01 2.09296417e+00
5.34215331e-01 2.59367585e-01 9.52373520e-02 9.02565897e-01
5.37526608e-01 4.05100405e-01 2.35937476e-01 -3.13112617e-01
1.54140913e+00 -3.57137144e-01 -8.12793791e-01 3.46522126e-03
2.05466617e-02 -6.17596269e-01 9.02377069e-01 3.37886102e-02
-8.20271492e-01 -4.32498962e-01 -1.03609765e+00 -9.10811722e-02
-1.99937522e-01 8.68811309e-02 7.63760924e-01 1.15382707e+00
-1.06215262e+00 4.26759720e-01 -5.61992407e-01 -3.30400109e-01
1.06319630e+00 8.49286318e-01 -8.87957454e-01 -2.88032722e-02
-1.00873506e+00 2.64486909e-01 3.34528685e-01 3.47701281e-01
-6.38794422e-01 -6.55033767e-01 -7.23762572e-01 9.65289921e-02
2.85235643e-01 -6.76163375e-01 8.02498460e-01 -1.33281147e+00
-1.33102143e+00 1.33464801e+00 -7.87593052e-02 -3.78390759e-01
8.12400401e-01 6.73332363e-02 -7.62963474e-01 -5.54162785e-02
6.51267692e-02 5.95863998e-01 1.07180905e+00 -1.11061633e+00
-6.13964081e-01 -9.12056029e-01 2.90678572e-02 1.79675356e-01
-8.02114964e-01 1.53380424e-01 -3.10985833e-01 -5.63323438e-01
-4.57249954e-03 -8.84781837e-01 -2.48449319e-03 4.78279173e-01
-4.29659009e-01 2.41512418e-01 1.15546024e+00 -7.75790036e-01
1.14842546e+00 -2.44641161e+00 -9.12587419e-02 1.02638528e-01
4.70376074e-01 6.22128129e-01 -1.62866130e-01 5.67733385e-02
-2.26862252e-01 4.43868250e-01 -2.07314029e-01 -3.70049119e-01
-8.24032202e-02 -8.08844119e-02 -4.12173808e-01 6.80312753e-01
1.45739496e-01 7.57294536e-01 -7.62787223e-01 -4.09983635e-01
4.79262248e-02 6.09889746e-01 -5.09081542e-01 2.33337745e-01
1.58804238e-01 5.18855512e-01 -5.64460099e-01 8.00323188e-01
1.38525736e+00 1.36343002e-01 2.28980094e-01 -3.04871529e-01
1.03971168e-01 -3.93022537e-01 -1.19922328e+00 1.32349396e+00
-2.88027227e-01 3.81589264e-01 4.25182372e-01 -7.56837964e-01
8.62627149e-01 3.60603780e-01 4.24876302e-01 -6.45572424e-01
3.21772039e-01 4.25042249e-02 -1.53382227e-01 -4.34098065e-01
3.73849958e-01 -6.68049678e-02 1.03315912e-01 5.52113295e-01
-2.81353354e-01 5.32915652e-01 -3.12742203e-01 -6.48625121e-02
5.97607374e-01 -9.88471135e-03 3.51495296e-01 -2.16916203e-01
7.85154223e-01 -6.19866729e-01 8.57343316e-01 2.04608873e-01
-5.74323595e-01 6.40801847e-01 3.29322815e-01 -4.34141159e-01
-9.96921778e-01 -9.94267344e-01 -8.43973830e-02 6.60356164e-01
2.54960746e-01 -7.10866302e-02 -1.00656605e+00 -7.43380725e-01
2.60483325e-01 1.70471728e-01 -6.41628742e-01 -5.21090388e-01
-5.08894444e-01 -7.48117864e-01 9.23995495e-01 1.11305915e-01
9.79966283e-01 -8.37172210e-01 4.02841680e-02 -1.91891178e-01
-9.56650227e-02 -9.66488659e-01 -7.57720292e-01 -8.33130836e-01
-4.11431193e-01 -1.00020719e+00 -6.41522944e-01 -6.07964754e-01
9.29686606e-01 6.54399037e-01 5.01469433e-01 1.30226910e-01
-3.36419702e-01 2.66826212e-01 3.07487771e-02 -4.09967273e-01
-3.60297441e-01 -9.39064622e-02 4.97283459e-01 9.20570493e-01
5.88521898e-01 -8.97002697e-01 -9.60460782e-01 2.32728988e-01
-1.04250789e+00 -2.68434763e-01 3.50221843e-01 6.90187693e-01
4.60294455e-01 1.36983737e-01 6.75733566e-01 -9.33326602e-01
7.44742632e-01 -4.98185754e-01 -5.38973749e-01 3.40499401e-01
-6.65716767e-01 -2.34352648e-01 7.74731636e-01 -4.91328686e-01
-1.18454754e+00 1.62895128e-01 -7.14255497e-02 -7.70710170e-01
-9.28439423e-02 -1.88827559e-01 -9.05430317e-01 -5.11747718e-01
4.73228514e-01 3.05286765e-01 4.25626338e-01 -2.73133695e-01
4.94621605e-01 9.62251008e-01 7.15844989e-01 -4.66970891e-01
1.00436902e+00 7.89997280e-01 -7.31792301e-02 -7.07549632e-01
-3.63238543e-01 -8.10518041e-02 -3.93412083e-01 -2.53016930e-02
6.26303792e-01 -9.63487566e-01 -1.27757502e+00 7.30319738e-01
-1.05165100e+00 6.37524366e-01 -1.67484045e-01 1.12912267e-01
-2.88229644e-01 7.92391539e-01 -2.09401444e-01 -8.57106209e-01
-5.20833075e-01 -1.01339006e+00 7.45218992e-01 5.82729340e-01
2.03115325e-02 -5.41181505e-01 -2.51825988e-01 3.60986829e-01
3.77349555e-01 3.75740528e-01 6.34404480e-01 -4.08983767e-01
-7.70321965e-01 -4.43482995e-01 -5.57011425e-01 4.51423258e-01
6.11342788e-01 -1.58250362e-01 -1.27948999e+00 -5.58612108e-01
1.08244196e-01 -2.04885200e-01 5.66862345e-01 -5.72953257e-04
1.56134737e+00 -8.35021079e-01 -3.76751542e-01 1.17968833e+00
1.22969401e+00 2.26511896e-01 8.29369426e-01 8.42766017e-02
9.47272718e-01 8.46937835e-01 2.07077563e-01 6.36320949e-01
3.21901202e-01 5.94073832e-01 5.89364588e-01 -6.24450259e-02
1.54738158e-01 -5.37890017e-01 8.40378273e-03 1.92276314e-01
-8.51133764e-02 -3.01823199e-01 -4.68481451e-01 4.31510985e-01
-1.71051955e+00 -1.23195612e+00 3.56952995e-01 2.49087071e+00
7.11028397e-01 -5.18594027e-01 1.50032163e-01 -1.29771262e-01
1.08385050e+00 2.37853110e-01 -8.31079483e-01 -2.59788454e-01
-1.81677997e-01 -1.28010958e-01 5.01006901e-01 1.38273910e-01
-1.09528327e+00 9.01944578e-01 5.37736702e+00 7.95742869e-01
-1.25188982e+00 8.75761583e-02 9.64092493e-01 -1.49197161e-01
-4.01246965e-01 -1.25208035e-01 -7.00436234e-01 7.78239131e-01
4.92036223e-01 -6.48724854e-01 7.91974485e-01 8.06447387e-01
1.13387987e-01 5.15082657e-01 -1.09917402e+00 1.44075525e+00
2.18517795e-01 -1.27990699e+00 3.00676137e-01 3.31174046e-01
4.66104597e-01 -6.48085177e-01 4.64485168e-01 -6.96759624e-03
1.58563048e-01 -1.09396124e+00 4.32329118e-01 4.67990458e-01
1.18452108e+00 -9.69644129e-01 4.09854233e-01 7.66848773e-02
-1.14878976e+00 -1.87167630e-01 -4.82132554e-01 1.92763045e-01
-2.49228105e-02 2.79909372e-01 -4.49124962e-01 6.48061156e-01
7.00754523e-01 5.69461584e-01 -2.91800827e-01 6.19508684e-01
-6.08289335e-03 5.77063337e-02 -2.36696601e-01 2.98881710e-01
-2.28946149e-01 -4.83663678e-01 4.82252270e-01 6.78792894e-01
3.77348334e-01 3.36210012e-01 -1.03343345e-01 1.10566175e+00
-5.67023873e-01 1.95356086e-01 -8.53740156e-01 -1.84609652e-01
8.43020797e-01 1.25421798e+00 -2.16809586e-01 6.15188926e-02
-2.37883985e-01 1.32896686e+00 3.26455414e-01 3.02069992e-01
-7.07900465e-01 -1.92574292e-01 1.31373250e+00 1.42679304e-01
-7.48566091e-02 1.69048607e-01 -1.58308327e-01 -1.26114285e+00
3.81782532e-01 -1.01410043e+00 3.52162123e-01 -4.06216860e-01
-1.45581484e+00 6.15460634e-01 -3.33200723e-01 -1.24688029e+00
-7.65509605e-02 -2.32325748e-01 -5.13273418e-01 1.07762194e+00
-1.31236243e+00 -1.41212165e+00 -3.86306792e-01 6.93299234e-01
-4.09305543e-02 -3.72407556e-01 8.43291044e-01 5.66651762e-01
-8.88539910e-01 1.20709264e+00 1.95227504e-01 2.38448083e-01
6.03540599e-01 -6.45528138e-01 5.47589540e-01 8.50069463e-01
-1.96478963e-01 8.50434899e-01 3.04385245e-01 -4.84407514e-01
-1.49591804e+00 -1.22924650e+00 5.67114890e-01 -1.86008334e-01
2.73604512e-01 -4.71546352e-01 -1.00515902e+00 6.94532931e-01
-6.53048232e-02 9.56959277e-02 7.45409906e-01 -2.84237713e-01
-5.59161365e-01 -5.88973820e-01 -1.64887547e+00 8.25123668e-01
1.30688131e+00 -7.95872271e-01 -4.76408266e-02 1.94685742e-01
6.06147408e-01 -8.36773589e-02 -5.63215554e-01 2.57196873e-01
8.06535959e-01 -1.14362383e+00 1.17340958e+00 -6.04079247e-01
1.84953082e-02 -2.41827875e-01 -5.82496971e-02 -8.45984578e-01
-2.96397001e-01 -9.37071562e-01 1.96597368e-01 1.76493919e+00
-1.65758044e-01 -8.96094084e-01 1.16174054e+00 1.00561643e+00
5.18476248e-01 -3.66421312e-01 -1.03477573e+00 -6.24766231e-01
-2.89818883e-01 1.15584940e-01 1.28004229e+00 1.31878960e+00
-4.45222497e-01 -1.40317529e-01 -7.44954407e-01 4.70786035e-01
8.33181977e-01 -3.29091474e-02 9.35804009e-01 -1.23886967e+00
1.66887969e-01 -3.57114673e-01 -6.82599485e-01 -6.49101973e-01
5.30982673e-01 -8.32752228e-01 -5.18138945e-01 -7.99257159e-01
2.89521396e-01 -4.18324649e-01 -2.12177694e-01 4.59080845e-01
-1.57034174e-02 4.40994650e-01 9.98621341e-03 3.47570717e-01
-1.22629879e-02 7.33586729e-01 1.03747737e+00 -1.38424516e-01
-5.26369922e-02 1.21978186e-01 -1.23731017e+00 5.89081347e-01
7.22645402e-01 -2.57182628e-01 -5.82264483e-01 -4.90058064e-01
-8.06949288e-02 -9.07384753e-02 5.41544795e-01 -7.74345279e-01
2.41626590e-01 -2.06958383e-01 3.80040288e-01 -2.00454239e-02
3.31869245e-01 -1.04665589e+00 4.18857545e-01 2.25454316e-01
-6.26764372e-02 1.04769327e-01 1.65888190e-01 7.30272412e-01
-1.73676103e-01 1.35537803e-01 9.59756672e-01 -1.48527011e-01
-6.94517851e-01 9.61264074e-01 1.60663366e-01 -1.20388858e-01
1.35438931e+00 -5.85533977e-01 -2.43542567e-01 -3.79572481e-01
-4.27800864e-01 1.23187996e-01 8.23335171e-01 5.69137394e-01
6.36598945e-01 -1.51313043e+00 -8.02738905e-01 7.29110301e-01
2.10751235e-01 -2.30607435e-01 6.93954170e-01 2.62723356e-01
-4.16986406e-01 -3.92207131e-02 -5.72283328e-01 -2.79976666e-01
-1.38766563e+00 8.22462320e-01 4.70167309e-01 3.89139831e-01
-4.20289159e-01 7.31399179e-01 5.10486186e-01 -3.76603633e-01
1.39960557e-01 4.94300008e-01 -6.97345957e-02 1.19365538e-02
8.33728373e-01 3.61848474e-01 -1.73302695e-01 -1.01217425e+00
-3.69108975e-01 5.55041134e-01 -2.87285507e-01 2.52113044e-01
9.39741075e-01 -4.39872116e-01 -3.41940045e-01 -3.78613114e-01
1.30646586e+00 7.75514692e-02 -1.27243984e+00 -1.74120128e-01
-3.57330471e-01 -1.14420164e+00 -4.56138179e-02 -3.55638593e-01
-1.42961776e+00 7.79866636e-01 9.23192143e-01 7.91537315e-02
1.33113706e+00 -4.86551166e-01 7.94241548e-01 7.58892596e-02
3.84018272e-01 -7.93889821e-01 -2.87437230e-01 -2.05005810e-01
7.39924490e-01 -1.30122054e+00 7.62876049e-02 -6.17407620e-01
-5.30519724e-01 6.97740793e-01 7.07659125e-01 8.33220631e-02
5.70325434e-01 -4.78870384e-02 4.60784622e-02 1.59052074e-01
-1.05564043e-01 3.16279173e-01 4.65593766e-03 9.10235584e-01
-7.43510351e-02 8.37304816e-02 -1.51890263e-01 6.08115196e-01
-2.21541241e-01 9.53055080e-03 2.68309563e-01 6.56323373e-01
1.17634200e-01 -1.19299078e+00 -3.19575101e-01 2.89723635e-01
-5.29741526e-01 8.56607333e-02 -4.31243360e-01 6.29693031e-01
2.42419854e-01 6.79531157e-01 5.08154891e-02 -5.47971249e-01
2.53740221e-01 -1.31690830e-01 2.65681565e-01 -2.70826012e-01
-3.40258360e-01 -5.27066767e-01 -3.04552495e-01 -5.11950374e-01
-2.77642250e-01 -5.17360866e-01 -8.72445881e-01 -1.06455386e+00
-9.94448140e-02 -1.23751229e-02 5.22000134e-01 6.02415621e-01
9.02879059e-01 -1.13229230e-01 1.07910621e+00 -6.48494422e-01
-7.04350114e-01 -3.42674196e-01 -7.99948931e-01 5.85167646e-01
4.02660400e-01 -4.38492596e-01 -2.78147519e-01 5.20068109e-02] | [12.76205062866211, 0.778588056564331] |
24def6d9-afb3-4080-972c-bc3601602010 | tensor-network-kalman-filtering-for-large | 2110.13501 | null | https://arxiv.org/abs/2110.13501v1 | https://arxiv.org/pdf/2110.13501v1.pdf | Tensor Network Kalman Filtering for Large-Scale LS-SVMs | Least squares support vector machines are a commonly used supervised learning method for nonlinear regression and classification. They can be implemented in either their primal or dual form. The latter requires solving a linear system, which can be advantageous as an explicit mapping of the data to a possibly infinite-dimensional feature space is avoided. However, for large-scale applications, current low-rank approximation methods can perform inadequately. For example, current methods are probabilistic due to their sampling procedures, and/or suffer from a poor trade-off between the ranks and approximation power. In this paper, a recursive Bayesian filtering framework based on tensor networks and the Kalman filter is presented to alleviate the demanding memory and computational complexities associated with solving large-scale dual problems. The proposed method is iterative, does not require explicit storage of the kernel matrix, and allows the formulation of early stopping conditions. Additionally, the framework yields confidence estimates of obtained models, unlike alternative methods. The performance is tested on two regression and three classification experiments, and compared to the Nystr\"om and fixed size LS-SVM methods. Results show that our method can achieve high performance and is particularly useful when alternative methods are computationally infeasible due to a slowly decaying kernel matrix spectrum. | ['Kim Batselier', 'Johan A. K. Suykens', 'Maximilian Lucassen'] | 2021-10-26 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [ 3.96731421e-02 -3.85337144e-01 -2.05807969e-01 -2.29006469e-01
-7.46535361e-01 -3.28828692e-01 5.41015387e-01 -2.02977121e-01
-4.31013703e-01 9.80234981e-01 -4.26771015e-01 -4.49929148e-01
-5.44931531e-01 -4.20635581e-01 -3.35661918e-01 -1.02412629e+00
2.69904695e-02 3.58464271e-01 1.54567346e-01 2.20107082e-02
4.03816909e-01 5.93254805e-01 -1.72492433e+00 1.93430260e-02
9.27207172e-01 1.34367168e+00 8.63135755e-02 4.90909785e-01
-8.90264958e-02 4.72260356e-01 -1.70742020e-01 -1.85653940e-01
2.68997371e-01 -3.22749704e-01 -3.41446906e-01 1.53480351e-01
2.65890419e-01 -3.05518478e-01 -2.24997252e-01 1.11808527e+00
1.76020056e-01 3.81226838e-01 7.51076639e-01 -1.16553128e+00
-2.19583973e-01 1.38465673e-01 -5.21609545e-01 7.62812868e-02
3.08646411e-01 -3.38430285e-01 8.45202446e-01 -1.23703158e+00
-1.17803633e-01 9.41823244e-01 9.38581109e-01 1.13594413e-01
-1.60792649e+00 -5.19003510e-01 -1.42686933e-01 3.07889909e-01
-1.57773077e+00 -6.45538807e-01 7.43136048e-01 -6.77628934e-01
6.61408186e-01 4.94533002e-01 4.48102742e-01 8.75335991e-01
1.09565377e-01 6.91407144e-01 1.48296297e+00 -5.11719227e-01
3.95176917e-01 5.53180873e-01 3.31049442e-01 8.25010836e-01
3.83828759e-01 1.57320142e-01 -4.87792343e-01 -7.48902857e-01
7.85475492e-01 -8.77735838e-02 -2.29786918e-01 -6.74199700e-01
-1.06326175e+00 1.14422965e+00 -8.42942670e-02 2.45672971e-01
-3.59051764e-01 3.34645249e-02 4.18411046e-01 8.69414508e-02
7.05421031e-01 2.01895222e-01 -3.15788239e-01 -6.14307728e-03
-1.32897103e+00 2.23083094e-01 1.02330494e+00 4.59797740e-01
5.22155285e-01 2.44391054e-01 2.38241509e-01 9.94929254e-01
6.03955805e-01 5.19835949e-01 4.80458438e-01 -8.16225886e-01
1.68561965e-01 1.70810163e-01 3.95473659e-01 -1.32464230e+00
-4.77727681e-01 -4.53873217e-01 -9.44495916e-01 3.69365752e-01
8.08834374e-01 1.35507494e-01 -5.65793753e-01 1.17838120e+00
3.84546310e-01 3.08967113e-01 7.79836848e-02 9.49826479e-01
3.64798248e-01 7.40541756e-01 -4.64772284e-01 -6.23631716e-01
1.14745891e+00 -6.77715421e-01 -8.62327874e-01 -2.39191815e-01
5.32858133e-01 -9.25414026e-01 7.47101724e-01 8.42580855e-01
-9.07125533e-01 -3.67843181e-01 -9.96758223e-01 2.58130133e-01
-7.45999888e-02 7.87554860e-01 8.43904614e-01 8.84806454e-01
-6.51948452e-01 8.52798820e-01 -1.03698564e+00 9.43470560e-03
-5.54056466e-02 3.84754747e-01 -2.25519896e-01 1.22479595e-01
-1.07409084e+00 9.79233682e-01 3.01142424e-01 7.07791924e-01
-9.16665122e-02 -4.34617192e-01 -7.96437562e-01 -2.35393662e-02
1.40691206e-01 -2.27914423e-01 1.02211952e+00 -7.52673388e-01
-1.71108639e+00 2.11734980e-01 -1.54499054e-01 -3.65596384e-01
6.23420954e-01 -3.60582441e-01 -2.40845695e-01 1.70532852e-01
-2.81201061e-02 -1.40858456e-01 1.41238368e+00 -7.75040030e-01
-4.84488338e-01 -2.15152234e-01 -2.54833579e-01 -1.97739787e-02
-4.44314241e-01 -1.58222333e-01 -2.46347889e-01 -6.84797943e-01
7.07054257e-01 -1.04677474e+00 -4.18734312e-01 -1.13985002e-01
-4.90763150e-02 -2.67054606e-02 6.24366164e-01 -7.89861321e-01
1.38445342e+00 -2.29630017e+00 1.02394842e-01 5.08646548e-01
-2.55309254e-01 2.03415602e-01 3.10250759e-01 3.72890413e-01
-2.03847662e-01 -3.78582627e-01 -2.21710980e-01 -6.66561872e-02
-1.69246316e-01 2.83055097e-01 -4.29627538e-01 9.97930706e-01
-4.57042195e-02 1.22122496e-01 -7.09289491e-01 -3.77296895e-01
4.25035059e-01 4.58295763e-01 -3.78937364e-01 -1.72857285e-01
1.87203556e-01 1.54306561e-01 -3.33828360e-01 3.66563618e-01
4.81503218e-01 -2.81844646e-01 1.68945372e-01 -5.89477539e-01
-2.13941947e-01 2.67448705e-02 -1.97419167e+00 1.25865293e+00
-4.59962785e-01 5.93334556e-01 2.14987010e-01 -1.57686985e+00
9.03268158e-01 3.99856508e-01 3.67813408e-01 -1.99194595e-01
-1.21516697e-02 5.43665409e-01 -2.54163086e-01 -3.70037585e-01
4.65743631e-01 -2.67945111e-01 1.60899520e-01 1.46678984e-01
-1.76950410e-01 -4.85732779e-02 3.16677213e-01 -6.41303211e-02
6.09963000e-01 2.38979250e-01 4.96778399e-01 -5.40820539e-01
8.44035268e-01 -5.66880032e-02 6.49809420e-01 6.64484143e-01
2.34542102e-01 2.68230051e-01 5.27425826e-01 -3.86405408e-01
-6.79328024e-01 -7.69810319e-01 -6.33669436e-01 7.61173308e-01
-5.89544736e-02 -1.08532965e-01 -3.87216926e-01 -4.86750126e-01
1.95970610e-01 6.12782300e-01 -2.76623517e-01 1.83794387e-02
-4.56629544e-01 -8.82396758e-01 3.47580224e-01 4.34441924e-01
1.74408972e-01 -2.22000867e-01 -3.41197699e-01 3.81070256e-01
-7.31037632e-02 -1.01863480e+00 -1.58937171e-01 2.29029015e-01
-1.24844968e+00 -1.08616781e+00 -6.64648473e-01 -4.72634852e-01
7.37737596e-01 2.80709296e-01 4.12326157e-01 -2.44354755e-01
-1.66466385e-01 2.77294219e-01 -1.90496877e-01 1.23990171e-01
-2.55292773e-01 -2.10585102e-01 6.01435781e-01 3.98025423e-01
2.06646964e-01 -4.09018725e-01 -3.17050636e-01 6.11927330e-01
-6.79444492e-01 3.78063694e-02 5.96111059e-01 1.51433086e+00
4.47641492e-01 3.33530307e-01 5.03805697e-01 -6.95690334e-01
5.46028912e-01 -3.25024039e-01 -1.19819832e+00 2.09268361e-01
-9.00036275e-01 2.27103636e-01 6.54044569e-01 -6.50608480e-01
-1.01215255e+00 3.55830491e-01 1.27937093e-01 -5.13280690e-01
2.56156862e-01 9.22321558e-01 2.68969148e-01 -3.92176390e-01
7.52717197e-01 3.52622598e-01 3.19431007e-01 -6.17454350e-01
1.15220197e-01 7.63339043e-01 3.04164112e-01 -7.00488448e-01
8.09814811e-01 4.13011819e-01 3.33138019e-01 -1.20734990e+00
-8.74093294e-01 -5.43263912e-01 -5.85388243e-01 -1.41406819e-01
2.97172815e-01 -7.31057763e-01 -6.29274547e-01 4.16270018e-01
-8.88040364e-01 1.33445755e-01 5.56962984e-03 1.11245811e+00
-6.36663735e-01 7.93668866e-01 -6.30824864e-01 -1.22143555e+00
1.53066879e-02 -1.06832981e+00 5.89332163e-01 -5.79800531e-02
-1.88626334e-01 -1.04614365e+00 -2.93068826e-01 3.09056759e-01
3.16750675e-01 6.85041100e-02 6.81802928e-01 -4.55218703e-01
-2.25025684e-01 -5.54685771e-01 -3.45177323e-01 6.66384697e-01
4.38933857e-02 7.73279965e-02 -8.26848030e-01 -6.01334095e-01
3.70562673e-01 -9.98101160e-02 6.25447392e-01 4.90119487e-01
9.87343311e-01 -4.33328331e-01 -2.52113283e-01 3.32353950e-01
1.21384573e+00 4.85079922e-02 9.06935036e-02 1.26527578e-01
3.28673303e-01 6.82517529e-01 9.07495320e-01 6.85638726e-01
-8.06304738e-02 7.04323828e-01 9.05616656e-02 5.98317198e-02
4.27121192e-01 1.68689758e-01 4.18300182e-01 1.03457391e+00
-7.66518563e-02 4.57814962e-01 -8.06903064e-01 2.93484896e-01
-2.17240977e+00 -9.85442877e-01 -5.51681399e-01 2.56745744e+00
5.54005623e-01 7.66173750e-02 3.94929461e-02 6.03290200e-01
5.68762898e-01 -1.85630679e-01 -4.85440165e-01 -2.32434005e-01
1.30345866e-01 9.34262499e-02 7.18112886e-01 5.47560871e-01
-1.17807031e+00 4.42184359e-01 6.52215910e+00 1.03821588e+00
-1.08190644e+00 1.01408616e-01 2.71694899e-01 -6.73821345e-02
2.05695257e-01 4.56899740e-02 -9.71121371e-01 4.79473919e-01
8.73252988e-01 2.15171296e-02 5.14869750e-01 1.01832616e+00
2.92303741e-01 -4.54591721e-01 -1.08693016e+00 1.25991821e+00
-4.21329029e-02 -1.11845458e+00 -4.64129955e-01 1.55438215e-01
3.73509556e-01 -2.38415986e-01 3.86721306e-02 2.72946030e-01
-1.73265770e-01 -7.31688321e-01 6.64873898e-01 5.17403603e-01
5.89529693e-01 -7.44687557e-01 5.99573076e-01 5.78946710e-01
-9.94070470e-01 -2.21976340e-01 -4.26570892e-01 -3.57288271e-01
5.16659953e-02 1.12709093e+00 -4.89859849e-01 4.36398596e-01
5.47719657e-01 4.00012106e-01 -1.57358974e-01 1.08615160e+00
1.08831756e-01 6.42697811e-01 -7.61400640e-01 -2.16444030e-01
1.66586921e-01 -7.66324103e-01 5.97457409e-01 1.06446791e+00
5.39804578e-01 5.51753044e-02 4.02660429e-01 4.92621630e-01
8.07563305e-01 2.30905682e-01 -4.60630685e-01 -1.03770539e-01
1.87008619e-01 1.16113400e+00 -8.79485369e-01 -1.71482742e-01
-6.50409639e-01 7.86295831e-01 1.11733124e-01 4.62119251e-01
-7.44729400e-01 -3.32493126e-01 5.21574855e-01 2.55690720e-02
4.07258362e-01 -4.96053219e-01 -2.49451816e-01 -1.43919766e+00
9.79169607e-02 -8.48245084e-01 4.01821494e-01 -3.78335297e-01
-1.22418237e+00 3.50603700e-01 1.16818577e-01 -1.36126363e+00
-5.42824686e-01 -8.78779531e-01 -2.24138573e-02 8.72690320e-01
-1.21223903e+00 -7.72153795e-01 1.62112623e-01 5.16450047e-01
3.51038337e-01 -2.23294809e-01 8.50983918e-01 5.23942709e-01
-7.34544933e-01 4.31690454e-01 7.58894861e-01 -2.71484613e-01
4.78869557e-01 -1.05370092e+00 -4.15550023e-01 8.44991565e-01
1.47969633e-01 6.86729372e-01 1.02064252e+00 -5.16781271e-01
-1.53217196e+00 -5.65399706e-01 6.81892753e-01 1.37417745e-02
9.68011081e-01 -2.50064760e-01 -9.93640363e-01 2.23040223e-01
-6.11001313e-01 3.15488040e-01 8.84782910e-01 3.75235915e-01
-1.66835725e-01 -1.59365639e-01 -1.06627762e+00 3.82858455e-01
3.57634246e-01 -5.28047919e-01 -5.19197881e-01 5.30838549e-01
-1.08054712e-01 -3.89393717e-01 -9.58212674e-01 3.68740171e-01
8.26843560e-01 -9.55613315e-01 9.56272662e-01 -4.97245431e-01
-1.07024759e-01 -4.72369015e-01 -3.00930589e-01 -1.08179057e+00
-3.92483413e-01 -6.38327122e-01 -3.93876821e-01 9.47700322e-01
3.53220820e-01 -7.58711398e-01 8.74094367e-01 6.09439492e-01
2.51898646e-01 -6.98211908e-01 -1.24865150e+00 -1.11697650e+00
-4.14085180e-01 -6.50339961e-01 -3.30313519e-02 1.00437641e+00
1.19845502e-01 2.63783127e-01 -6.58540368e-01 3.41175795e-01
9.20771360e-01 2.64095873e-01 5.82343221e-01 -1.49789500e+00
-6.73300028e-01 -2.68373728e-01 -3.51314902e-01 -1.05195212e+00
2.48549193e-01 -6.75348759e-01 2.91398242e-02 -1.01248789e+00
-2.11717695e-01 -7.88084984e-01 -3.21066380e-01 2.96987265e-01
9.66507941e-02 2.20266640e-01 -2.25059301e-01 3.65158439e-01
2.07095034e-02 4.60978836e-01 5.56691408e-01 2.05104560e-01
-3.57335269e-01 4.77107435e-01 -2.05711797e-01 1.14445090e+00
5.96262217e-01 -4.01856244e-01 -3.85522753e-01 5.84839620e-02
3.17299545e-01 4.89967048e-01 3.69949818e-01 -8.50272655e-01
3.31556469e-01 -1.53260976e-01 3.97467047e-01 -6.03314161e-01
6.07262075e-01 -8.75435710e-01 3.49991590e-01 6.04438901e-01
3.02697197e-02 -2.37125471e-01 2.28665173e-02 7.77918100e-01
-2.13350862e-01 -6.06317461e-01 9.24800336e-01 2.40582421e-01
-4.30664957e-01 -2.16009885e-01 -5.65353751e-01 -5.03936708e-01
8.79892170e-01 -4.24028546e-01 4.06512856e-01 -2.93809891e-01
-9.04062867e-01 -1.11245230e-01 3.58297378e-02 1.03182107e-01
6.33176088e-01 -1.35667789e+00 -4.90196347e-01 3.79717022e-01
-1.92641586e-01 -4.17655170e-01 5.29864058e-02 1.30118823e+00
-4.68343765e-01 5.42368174e-01 1.36450663e-01 -6.90887868e-01
-1.25669444e+00 5.68887234e-01 2.28460878e-02 -2.03958124e-01
-5.41555583e-01 5.83699763e-01 -3.66556533e-02 -2.29412019e-01
2.20779330e-01 -1.42267972e-01 -1.82362005e-01 3.57581526e-01
5.54834962e-01 7.87526011e-01 1.73688889e-01 -6.34982884e-01
-2.67629981e-01 4.13353711e-01 -3.75428200e-02 -2.02597409e-01
1.19018149e+00 -6.20730147e-02 -2.83068061e-01 5.65882921e-01
1.07082117e+00 -8.57589692e-02 -9.94539797e-01 -4.76670951e-01
1.91341057e-01 -7.15655863e-01 5.58190227e-01 -1.21711820e-01
-7.49491572e-01 5.97960532e-01 5.37019551e-01 3.62078249e-01
9.43868399e-01 -4.77732897e-01 4.97978508e-01 6.82789028e-01
3.90458077e-01 -1.27147472e+00 -2.17840165e-01 4.32786971e-01
7.71289051e-01 -1.25346100e+00 4.26061153e-01 -6.70367658e-01
-1.90338075e-01 1.53446007e+00 2.29136035e-01 -2.38278538e-01
8.10107350e-01 6.72829449e-02 -1.91096589e-01 1.40351370e-01
-4.27591592e-01 6.23903610e-02 5.73497951e-01 2.53895104e-01
4.09220040e-01 -1.18438706e-01 -8.69032681e-01 3.94833714e-01
2.52536461e-02 -2.00865753e-02 4.59640324e-01 7.03053117e-01
-3.16717654e-01 -1.03139102e+00 -8.28591466e-01 7.06442893e-01
-5.75436115e-01 3.37707512e-02 2.95747727e-01 6.70853794e-01
-3.76022905e-01 9.73365188e-01 -2.00492114e-01 -1.19600266e-01
1.10546298e-01 1.82059392e-01 4.48464453e-01 -3.26649219e-01
-4.15775068e-02 4.38285142e-01 2.18936890e-01 -5.29136479e-01
-3.42231363e-01 -1.00528932e+00 -8.53866816e-01 -1.92740798e-01
-9.27065194e-01 4.85430539e-01 9.71740365e-01 1.04635513e+00
3.96224223e-02 -3.85771715e-03 7.19989955e-01 -9.18494761e-01
-1.19365549e+00 -7.11546719e-01 -8.23728263e-01 -1.43077567e-01
2.59183615e-01 -1.05987275e+00 -6.32388234e-01 -1.86102569e-01] | [7.605876445770264, 4.155284404754639] |
41f66091-d220-4cff-b64e-a7a93d2bc77b | a-review-of-chatgpt-applications-in-education | 2305.00237 | null | https://arxiv.org/abs/2305.00237v1 | https://arxiv.org/pdf/2305.00237v1.pdf | A Review of ChatGPT Applications in Education, Marketing, Software Engineering, and Healthcare: Benefits, Drawbacks, and Research Directions | ChatGPT is a type of artificial intelligence language model that uses deep learning algorithms to generate human-like responses to text-based prompts. The introduction of the latest ChatGPT version in November of 2022 has caused shockwaves in the industrial and academic communities for its powerful capabilities, plethora of possible applications, and the great possibility for abuse. At the time of writing this work, several other language models (e.g., Google Bard and Meta LLaMA) just came out in an attempt to get a foothold in the vast possible market. These models have the ability to revolutionize the way we interact with computers and have potential applications in many fields, including education, software engineering, healthcare, and marketing. In this paper, we will discuss the possible applications, drawbacks, and research directions using advanced language Chatbots (e.g., ChatGPT) in each of these fields. We first start with a brief introduction and the development timeline of artificial intelligence based language models, then we go through possible applications of such models, after that we discuss the limitations and drawbacks of the current technological state of the art, and finally we point out future possible research directions. | ['Natheer Khasawneh', 'Mohammad Fraiwan'] | 2023-04-29 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [-3.66188616e-01 3.32227111e-01 -9.47054848e-02 -1.41930670e-01
-3.24720353e-01 -5.20759225e-01 5.98345995e-01 -2.14448005e-01
-2.41380945e-01 9.04307365e-01 4.01971489e-02 -6.43241644e-01
-1.38022508e-02 -8.32542777e-01 -1.50684610e-01 -3.39195430e-01
1.40265018e-01 6.50095463e-01 1.66193187e-01 -8.20878506e-01
3.67359787e-01 1.49363443e-01 -1.24222028e+00 4.90635782e-01
6.96930170e-01 6.73343599e-01 4.30076838e-01 5.41786969e-01
-8.34731936e-01 1.41208816e+00 -1.16851783e+00 -5.41519642e-01
-1.17614880e-01 -4.54904974e-01 -1.18972301e+00 -4.72161412e-01
-3.64883810e-01 -2.77893215e-01 -3.42799902e-01 7.10517108e-01
5.43362916e-01 8.79388452e-02 1.35225639e-01 -1.61829567e+00
-6.89747036e-01 8.15150738e-01 -2.05921695e-01 -1.55648500e-01
7.34476566e-01 3.42930704e-01 6.97600842e-01 -4.69823122e-01
7.12877989e-01 1.28136873e+00 6.55308664e-01 8.69360089e-01
-5.19130707e-01 -7.63870835e-01 8.07109177e-02 2.30798751e-01
-9.91789877e-01 3.07649956e-03 8.22419226e-01 -2.90497422e-01
1.22019839e+00 3.22833478e-01 4.81035352e-01 1.35643327e+00
4.03694451e-01 1.02269673e+00 1.02548361e+00 -6.28031015e-01
2.79022902e-01 5.16411066e-01 1.06029928e-01 5.26541770e-01
-2.75478661e-01 -3.66085410e-01 -4.81622636e-01 -3.38940591e-01
6.68889523e-01 -4.08021867e-01 2.60407239e-01 3.25823933e-01
-1.11066175e+00 1.13825583e+00 2.20357195e-01 8.84460449e-01
-3.69697958e-01 -1.15948550e-01 5.16320348e-01 3.77237022e-01
3.66122395e-01 6.92144692e-01 -3.45947504e-01 -6.60976589e-01
-2.06293702e-01 4.17795062e-01 1.22513437e+00 9.02606249e-01
3.53424460e-01 2.90263426e-02 2.10232232e-02 1.01530743e+00
2.04387605e-01 -7.27511197e-02 6.97036862e-01 -1.24092901e+00
3.46226871e-01 4.81985897e-01 1.66772306e-01 -9.20726240e-01
-4.65643644e-01 1.96083531e-01 -6.42172694e-01 8.80663171e-02
3.94608140e-01 -7.15271711e-01 -1.68278635e-01 1.22463644e+00
9.79098380e-02 -1.85071841e-01 -7.79366046e-02 7.47899771e-01
1.17415226e+00 9.76222754e-01 2.52960294e-01 4.92773112e-03
1.24009955e+00 -1.04855335e+00 -1.09672391e+00 -2.16681659e-01
7.93235362e-01 -7.72074699e-01 1.37604773e+00 4.39530551e-01
-9.26454127e-01 -3.58106822e-01 -6.93521857e-01 -8.85800645e-02
-6.24291539e-01 -4.24449772e-01 1.04584897e+00 8.38267922e-01
-1.01821685e+00 2.50641555e-01 -5.94926298e-01 -6.69520557e-01
-2.07282692e-01 2.79445589e-01 -4.77812300e-03 2.44968385e-01
-1.60170949e+00 9.76945221e-01 3.50624099e-02 -1.86071515e-01
-9.03562307e-02 -2.63096452e-01 -5.11540949e-01 -5.86504377e-02
2.67788649e-01 -4.84614462e-01 1.85040116e+00 -8.59480202e-01
-2.20079160e+00 6.97239757e-01 1.39308080e-01 -4.06220615e-01
4.51450229e-01 -4.38667648e-02 -2.62875825e-01 -1.24680407e-01
6.53591007e-02 6.58710718e-01 -6.18234538e-02 -7.22750843e-01
-5.59030771e-01 1.33673633e-02 5.89106083e-01 7.98723325e-02
-4.24057513e-01 6.54917717e-01 -1.47095978e-01 -3.05523425e-01
-5.44144213e-01 -8.26925218e-01 -3.91213268e-01 -6.70481771e-02
-2.07191050e-01 -6.55594289e-01 8.82797837e-01 -4.27268773e-01
1.09523213e+00 -1.75065398e+00 -4.10294503e-01 -3.58550102e-01
2.77919054e-01 5.40058851e-01 -1.57731116e-01 1.24105406e+00
3.46160263e-01 4.51341510e-01 3.17456156e-01 -1.00566680e-02
5.63604869e-02 1.70115232e-01 -1.22721866e-01 -1.92410916e-01
-6.21859506e-02 9.79225576e-01 -1.02941978e+00 -3.95774782e-01
4.43320066e-01 3.54758203e-01 -4.05769974e-01 3.22527707e-01
-6.24967694e-01 5.23837924e-01 -7.85213888e-01 2.40846783e-01
2.06416026e-01 -3.32659453e-01 2.65589535e-01 5.73158026e-01
-4.89524543e-01 7.60348916e-01 -4.84628141e-01 1.30426943e+00
-8.54272544e-01 8.42262328e-01 2.15877026e-01 -1.12969065e+00
1.06924140e+00 7.09506154e-01 4.52534735e-01 -7.29147792e-01
3.10155958e-01 1.90150112e-01 1.92340314e-01 -8.78157735e-01
6.17727578e-01 -1.52068630e-01 -3.28771859e-01 9.12935615e-01
-3.26192565e-02 -3.50143224e-01 6.39855638e-02 2.22402453e-01
1.03902411e+00 -8.83108526e-02 3.57989997e-01 1.45717591e-01
4.30794150e-01 6.43786415e-02 2.37219222e-02 8.78062248e-01
-4.37101930e-01 2.20937714e-01 6.11500740e-01 -7.97928333e-01
-7.83539414e-01 -3.70582461e-01 4.96823788e-01 1.40539801e+00
-3.72900516e-02 -3.88667196e-01 -9.21651483e-01 -5.94073713e-01
-4.66084599e-01 9.39285874e-01 -1.66905954e-01 4.66550365e-02
-4.33502406e-01 -3.29084128e-01 7.75378227e-01 3.04964572e-01
8.96620572e-01 -1.68126094e+00 -6.24214172e-01 3.75464827e-01
-6.65994763e-01 -1.28312111e+00 1.04849087e-02 -5.28404899e-02
-6.18316650e-01 -5.26115656e-01 -6.59558535e-01 -9.71892536e-01
-1.24989852e-01 2.85346180e-01 9.87098396e-01 3.03034961e-01
6.76273555e-02 4.53657508e-01 -5.95785201e-01 -7.93278694e-01
-9.03720558e-01 3.30787867e-01 2.36270502e-02 -3.84778529e-01
4.77704465e-01 -5.80724239e-01 -2.16701761e-01 1.94468766e-01
-6.25749111e-01 2.02089086e-01 2.65672654e-01 7.35390067e-01
-5.27639985e-01 -1.91276610e-01 8.42962980e-01 -8.85152340e-01
1.21396518e+00 -6.38834953e-01 -2.78884023e-01 1.69136614e-01
-1.64750040e-01 -2.30209127e-01 6.44885719e-01 -5.00982523e-01
-1.17614996e+00 -4.00512487e-01 -6.84071958e-01 1.21547289e-01
-4.35435623e-01 7.76411712e-01 1.02718964e-01 -1.44499287e-01
8.03547084e-01 1.32364511e-01 2.30057254e-01 -5.00816047e-01
1.76813066e-01 1.39625001e+00 9.29697156e-02 -6.40437007e-01
3.07021320e-01 -2.14991216e-02 -5.81413388e-01 -1.17495787e+00
-3.41442913e-01 -1.67103454e-01 -8.25349763e-02 -5.47221422e-01
6.97067261e-01 -4.16683942e-01 -1.16753852e+00 5.79900622e-01
-1.67481565e+00 -6.62813008e-01 1.82260256e-02 2.83317357e-01
-4.92685676e-01 2.50573784e-01 -1.13278234e+00 -1.12060285e+00
-4.33686435e-01 -9.23314571e-01 6.65563107e-01 5.31530440e-01
-6.95065737e-01 -1.20210910e+00 -2.55745351e-02 7.11459398e-01
8.14548492e-01 -1.19669870e-01 1.02032459e+00 -8.57998729e-01
-3.89319658e-01 -3.36389393e-01 -1.02764107e-02 2.01105714e-01
5.62992953e-02 -6.52669668e-02 -7.64853537e-01 2.52249092e-01
1.05699666e-01 -4.21985805e-01 -2.39999279e-01 8.79214425e-03
8.88888478e-01 -6.44427538e-01 -3.24146956e-01 -2.43453026e-01
9.03011382e-01 8.77877414e-01 6.18744731e-01 4.73425537e-01
1.61221951e-01 8.27932239e-01 5.09668827e-01 3.70941907e-01
3.76906663e-01 7.24721074e-01 1.45538360e-01 1.20247528e-01
1.13602780e-01 -2.95365691e-01 3.79678696e-01 1.30478704e+00
-7.78146833e-02 -5.68618476e-01 -1.17187774e+00 2.96853542e-01
-1.85754693e+00 -8.24349046e-01 -1.20172612e-01 1.80625224e+00
7.08581984e-01 1.99487776e-01 2.52841026e-01 -4.22288701e-02
7.28150010e-01 -3.01315002e-02 -2.50684947e-01 -1.04062009e+00
1.88098684e-01 2.52417922e-01 -2.47099146e-01 4.54446822e-01
-5.68248630e-01 1.11752439e+00 7.05036783e+00 5.79379916e-01
-1.44744360e+00 2.46548504e-01 7.16003835e-01 3.87873143e-01
1.36124939e-01 -2.80638814e-01 -3.00917238e-01 4.92006719e-01
1.15354943e+00 -5.65308630e-01 6.66354895e-01 9.67664480e-01
4.30158883e-01 -6.78141266e-02 -6.00469232e-01 7.92846084e-01
-9.90522653e-02 -1.49052238e+00 -1.71054617e-01 -1.80439815e-01
3.74483138e-01 1.98121995e-01 -7.34708980e-02 6.69678211e-01
6.18544698e-01 -8.42806458e-01 5.21531582e-01 4.96846661e-02
5.09628415e-01 -4.77830976e-01 9.20866728e-01 8.77782583e-01
-5.69374382e-01 -1.90490425e-01 -5.00217155e-02 -1.06202698e+00
1.93855062e-01 1.08583309e-01 -1.11315989e+00 3.84965748e-01
6.93333507e-01 1.11760125e-01 6.75127804e-02 9.99476671e-01
-2.21486613e-01 7.71849453e-01 -3.41221243e-01 -9.66274142e-01
3.50580692e-01 -8.78281295e-02 4.75110471e-01 1.07047951e+00
1.31091058e-01 4.26478297e-01 1.50235504e-01 8.50978792e-01
9.09095258e-02 8.98158252e-02 -8.32412481e-01 -3.02954763e-01
6.31774127e-01 1.05043888e+00 -7.54399061e-01 -3.47559571e-01
-5.71075916e-01 8.14612091e-01 2.47832593e-02 3.66709173e-01
-8.03306282e-01 -7.57084608e-01 7.58770585e-01 1.80096552e-01
-4.81709331e-01 -3.76696348e-01 -5.20888209e-01 -7.41489232e-01
-2.54317194e-01 -1.21286833e+00 -2.68236324e-02 -1.12545025e+00
-1.27219784e+00 7.75171518e-01 -1.02935815e-02 -1.00510526e+00
-6.20545805e-01 -5.35982728e-01 -8.58044088e-01 7.85537541e-01
-7.42376268e-01 -8.62726331e-01 2.53595170e-02 2.21929654e-01
8.79235506e-01 -1.95914090e-01 1.09237742e+00 1.89501628e-01
-2.70115733e-01 4.73334134e-01 1.15655875e-02 3.85415167e-01
5.62301040e-01 -6.93340242e-01 7.39804626e-01 1.21853212e-02
4.26543318e-02 7.96981454e-01 8.48285377e-01 -3.36448848e-01
-1.35355842e+00 -5.98208427e-01 1.40498447e+00 -4.41747755e-01
7.54379272e-01 -5.07764459e-01 -6.48488402e-01 7.91090190e-01
6.32918775e-01 -7.19749808e-01 4.63461131e-01 -1.15380408e-02
5.19252606e-02 1.00944474e-01 -1.12715030e+00 9.51327324e-01
5.44237256e-01 -4.94312674e-01 -4.48037893e-01 7.16404915e-01
9.93115664e-01 -3.96839380e-01 -3.57598037e-01 -1.29531637e-01
6.24714613e-01 -1.00089681e+00 4.24403697e-01 -6.52594626e-01
3.95871341e-01 2.20949590e-01 4.10571396e-01 -1.08038890e+00
-2.02205017e-01 -1.32120204e+00 2.22476766e-01 1.16839945e+00
2.39968479e-01 -9.47092831e-01 8.88438463e-01 9.49217260e-01
-1.27185464e-01 -6.45450354e-01 -9.20318067e-01 -6.76788628e-01
3.55928779e-01 -5.79255462e-01 5.61220527e-01 9.55090284e-01
9.06097472e-01 7.61850715e-01 -4.64545131e-01 -5.11656582e-01
-7.51922876e-02 -2.72776484e-01 8.50319386e-01 -1.15728414e+00
-1.33874699e-01 -4.62752759e-01 -1.77126110e-01 -1.33464599e+00
-2.03301781e-03 -5.29996812e-01 -1.35132596e-02 -1.68835330e+00
-1.11122981e-01 -3.08170885e-01 1.19559005e-01 7.48197973e-01
2.45228618e-01 6.84232563e-02 4.92493927e-01 -5.68883270e-02
-6.24157786e-01 3.25044870e-01 1.14885867e+00 -1.79930121e-01
-3.45959276e-01 3.22154433e-01 -8.38842809e-01 6.99214637e-01
1.18695307e+00 -2.71422148e-01 -2.15691835e-01 -4.67209786e-01
3.26867789e-01 3.92939836e-01 -4.85355593e-02 -9.31031168e-01
2.66743332e-01 -4.04481351e-01 -2.26274714e-01 8.86273533e-02
4.03621554e-01 -6.31367445e-01 -8.55033770e-02 3.52651983e-01
-5.77720821e-01 1.02596149e-01 2.36940756e-01 -3.93931754e-02
-3.02585989e-01 -2.86583573e-01 4.96079743e-01 -4.42106336e-01
-4.66445982e-01 -2.43886545e-01 -1.46292996e+00 -2.18361273e-01
9.01374280e-01 -1.28794998e-01 -3.11743677e-01 -1.18692493e+00
-3.16788614e-01 2.12934777e-01 3.28083374e-02 1.06026006e+00
2.46650904e-01 -9.21283185e-01 -3.27024370e-01 -1.11831687e-01
-3.08250606e-01 -4.93970305e-01 1.10415153e-01 6.19666696e-01
-6.75433695e-01 7.09918499e-01 -3.72449517e-01 -2.22920433e-01
-1.24330711e+00 4.71755296e-01 2.20054924e-01 -4.21205729e-01
-4.60108042e-01 6.63459957e-01 2.05048826e-02 -7.34411418e-01
3.55550408e-01 -7.37321898e-02 -3.59626412e-01 -5.03579855e-01
6.80644274e-01 2.56730676e-01 7.79011399e-02 -2.82874703e-01
-2.72842646e-01 -6.64554611e-02 -1.03747966e-02 -3.72208148e-01
1.07320666e+00 -7.97662139e-02 -2.98639208e-01 7.30859697e-01
7.85270631e-01 -2.86457986e-01 -3.07349503e-01 -1.03885196e-02
2.01554254e-01 1.48858950e-02 -4.38275218e-01 -1.11617792e+00
-7.83970833e-01 1.21569943e+00 1.72284737e-01 7.06329584e-01
7.40446389e-01 -2.28878902e-03 1.13910937e+00 6.74372435e-01
7.91606367e-01 -1.07680094e+00 2.87440360e-01 1.09273565e+00
8.27235460e-01 -8.22169900e-01 -4.96261984e-01 -2.60787308e-01
-7.80944705e-01 1.24911571e+00 7.83735931e-01 2.23835677e-01
3.72236520e-01 5.63252449e-01 5.98709285e-01 -4.14843336e-02
-9.85585570e-01 2.67952800e-01 -5.93499660e-01 1.02563715e+00
9.87888873e-01 -1.47736929e-02 -6.72870576e-01 8.12031209e-01
-4.31032062e-01 3.62938493e-01 7.63558924e-01 1.16997600e+00
-4.39637035e-01 -1.43969071e+00 -4.60249752e-01 9.85944122e-02
-6.22800648e-01 1.76287115e-01 -9.94213402e-01 7.92313397e-01
-1.23868711e-01 1.47564280e+00 -2.62832135e-01 -7.54426301e-01
2.01709390e-01 4.34608430e-01 3.38580683e-02 -6.21079743e-01
-1.16630793e+00 -4.82697099e-01 4.48309779e-01 -1.86611921e-01
-7.19169229e-02 -1.21724911e-01 -1.20772767e+00 -6.43069148e-01
-2.65287936e-01 6.58984303e-01 7.24773467e-01 8.97169828e-01
4.87862885e-01 1.55407295e-01 3.88898760e-01 -9.64815557e-01
-2.96603233e-01 -1.23182321e+00 -1.43141061e-01 -9.19652879e-02
-1.02423921e-01 -2.89644390e-01 4.26341072e-02 -2.39496455e-01] | [12.692681312561035, 7.949760913848877] |
60c3e38a-de67-448f-bd51-67adad339410 | ideals-idiomatic-expressions-for-advancement | 2305.13637 | null | https://arxiv.org/abs/2305.13637v2 | https://arxiv.org/pdf/2305.13637v2.pdf | IdEALS: Idiomatic Expressions for Advancement of Language Skills | Although significant progress has been made in developing methods for Grammatical Error Correction (GEC), addressing word choice improvements has been notably lacking and enhancing sentence expressivity by replacing phrases with advanced expressions is an understudied aspect. In this paper, we focus on this area and present our investigation into the task of incorporating the usage of idiomatic expressions in student writing. To facilitate our study, we curate extensive training sets and expert-annotated testing sets using real-world data and evaluate various approaches and compare their performance against human experts. | ['Zhou Yu', 'Sam Davidson', 'Bill Sun', 'Narutatsu Ri'] | 2023-05-23 | null | null | null | null | ['grammatical-error-correction'] | ['natural-language-processing'] | [ 3.96131724e-01 3.09110045e-01 2.21560393e-02 -6.81630611e-01
-8.58633339e-01 -5.84509671e-01 3.34313035e-01 5.79642415e-01
-6.56015754e-01 1.20268607e+00 3.34001958e-01 -5.90172350e-01
8.32029954e-02 -3.96486014e-01 -3.51310045e-01 9.72842500e-02
3.27697337e-01 4.44330484e-01 -4.12492231e-02 -6.87977552e-01
8.80684435e-01 4.63026203e-02 -1.15809655e+00 4.23721522e-01
1.44506550e+00 9.88138020e-02 -6.93305582e-02 6.52408898e-01
-3.00843030e-01 7.94414580e-01 -1.07311046e+00 -1.02672565e+00
-2.71069735e-01 -4.37291145e-01 -1.24643731e+00 -1.54211864e-01
5.96202433e-01 2.02233389e-01 4.15601104e-01 1.15788984e+00
6.60584629e-01 1.84851989e-01 4.66751158e-01 -6.01537287e-01
-9.32385743e-01 8.11413229e-01 -1.68163478e-01 5.46739697e-01
7.41059601e-01 -2.47063071e-01 1.06193793e+00 -8.99564207e-01
8.98131907e-01 9.63297963e-01 8.21665883e-01 9.89611030e-01
-1.17614162e+00 -3.81099492e-01 1.30012572e-01 3.36236030e-01
-1.35477674e+00 -4.45131749e-01 8.78758550e-01 -8.93697441e-02
1.45022893e+00 4.46363091e-01 4.21191275e-01 9.94545758e-01
-2.08406940e-01 8.41679335e-01 1.34364939e+00 -1.17737269e+00
-2.60960340e-01 4.85183507e-01 3.83924633e-01 7.74449468e-01
1.60243243e-01 -3.48308146e-01 -4.92634237e-01 1.79449975e-01
6.53637573e-02 -7.13732123e-01 -5.28759539e-01 4.98077720e-01
-5.54347277e-01 8.59535098e-01 -1.61129013e-02 5.92202067e-01
-1.69718377e-02 -1.35498242e-02 4.64569718e-01 5.86488187e-01
6.95407987e-01 1.07593799e+00 -6.32982612e-01 -7.86768198e-01
-7.87815213e-01 6.24679685e-01 1.09407353e+00 1.00154042e+00
2.29487449e-01 7.12866411e-02 1.89258039e-01 1.16705739e+00
-1.57466978e-02 3.26734036e-01 5.12060165e-01 -8.17765415e-01
7.21580207e-01 7.98329175e-01 1.66604921e-01 -1.13487816e+00
-2.61354625e-01 -4.50999618e-01 -2.41081715e-01 -3.03028524e-01
4.58569497e-01 -1.60204008e-01 -2.99575299e-01 1.59468615e+00
2.63891071e-02 -5.44301212e-01 3.28887515e-02 5.54290891e-01
8.49925339e-01 3.78989339e-01 4.23816532e-01 -1.45703360e-01
1.21273375e+00 -7.85789907e-01 -1.03672016e+00 -3.15333903e-01
1.24986470e+00 -1.13519394e+00 1.23741174e+00 4.89076793e-01
-1.35262382e+00 -2.50051886e-01 -9.07806754e-01 -4.13457066e-01
-5.17180264e-01 2.83998162e-01 5.16059577e-01 9.57252443e-01
-6.63439035e-01 6.40787959e-01 -5.42719901e-01 -3.66360307e-01
1.12035625e-01 8.57353881e-02 -4.77802545e-01 -7.37779811e-02
-1.40170872e+00 1.63003159e+00 2.35319704e-01 -7.61037394e-02
9.62276310e-02 -7.05945909e-01 -8.66580606e-01 -2.39937201e-01
2.84396321e-01 -3.95231664e-01 1.49286544e+00 -7.83344328e-01
-1.31646979e+00 1.14984274e+00 -3.03103834e-01 -2.80919135e-01
1.89114660e-01 -3.57821345e-01 -4.69343156e-01 -4.29749310e-01
2.19662413e-02 4.53639179e-01 1.33603722e-01 -1.03797019e+00
-5.23093998e-01 -2.76211202e-01 1.15550026e-01 2.75972992e-01
-3.54811281e-01 7.69403815e-01 1.85012043e-01 -1.09785676e+00
3.58411856e-02 -8.57335389e-01 1.17822781e-01 -4.88207221e-01
-7.07799718e-02 -6.84860766e-01 4.48428273e-01 -1.02363288e+00
2.05050755e+00 -1.59325659e+00 2.55132228e-01 1.42971009e-01
-7.34677985e-02 5.08745968e-01 1.79791614e-01 5.38179934e-01
-1.12872742e-01 6.42782629e-01 -4.00972009e-01 -6.64439619e-01
5.98612763e-02 4.15126413e-01 -1.33317977e-01 -1.21317104e-01
2.90451229e-01 8.47985029e-01 -1.06274438e+00 -6.15736127e-01
-1.08652025e-01 1.43435344e-01 -7.21339464e-01 -2.45166104e-02
-1.04258075e-01 1.33230284e-01 -2.11720780e-01 8.97624731e-01
2.30213344e-01 3.12457886e-02 2.51171142e-01 2.86347091e-01
-1.55099869e-01 9.65696394e-01 -9.62232292e-01 1.27613807e+00
-6.16666436e-01 6.80813074e-01 -2.84966826e-01 -8.82236958e-01
8.83834422e-01 3.24591756e-01 -2.27531552e-01 -7.03096449e-01
7.34027103e-02 6.45370901e-01 1.13058604e-01 -6.95308745e-01
1.13717055e+00 -3.86585683e-01 -3.78460348e-01 5.59720099e-01
1.37484178e-01 -3.48721117e-01 3.44882816e-01 1.59530222e-01
1.01845539e+00 1.18287504e-01 6.30411685e-01 -3.58115226e-01
6.54503405e-01 4.07199234e-01 3.51498336e-01 5.99387705e-01
-2.69771069e-01 3.11049700e-01 3.32323581e-01 -1.75405189e-01
-7.82262027e-01 -5.31174362e-01 -2.94112593e-01 1.10866272e+00
-4.34822023e-01 -9.33420122e-01 -9.59937334e-01 -9.96136248e-01
-2.88861722e-01 1.25217640e+00 -5.54114103e-01 6.07577004e-02
-9.73161697e-01 -8.40277255e-01 9.32594001e-01 5.55349410e-01
2.76249558e-01 -1.06764293e+00 -4.87944216e-01 4.69092816e-01
-5.31486571e-01 -9.35814321e-01 -1.83514416e-01 1.15639053e-01
-8.33766997e-01 -9.94071841e-01 -1.14723578e-01 -8.59104693e-01
7.65313029e-01 -2.17956752e-01 1.52125609e+00 7.28748083e-01
-1.25034153e-01 2.21599400e-01 -7.78754532e-01 -8.22598636e-01
-5.47923803e-01 2.87510395e-01 -2.93499500e-01 -9.60389614e-01
8.84725630e-01 -1.91311151e-01 1.63179502e-01 -9.01213959e-02
-6.36369824e-01 -9.50095057e-02 5.26299700e-02 9.49952841e-01
-7.99012110e-02 -2.55210251e-01 6.67754412e-01 -1.62486315e+00
1.20443010e+00 -2.40622044e-01 -1.08696423e-01 3.86861771e-01
-8.92691672e-01 1.38498724e-01 3.85568678e-01 5.76187856e-02
-1.21711636e+00 -4.31155235e-01 -7.79669464e-01 6.38767302e-01
-6.01147488e-03 8.19776237e-01 1.47433877e-01 -4.38379407e-01
8.52558970e-01 -2.12032214e-01 -1.78920954e-01 -4.69168395e-01
2.16978952e-01 6.97890103e-01 4.38139588e-01 -1.23832595e+00
3.56522679e-01 -3.19784343e-01 -3.72306943e-01 -6.46924496e-01
-1.10829663e+00 -2.08347246e-01 -5.27406931e-01 -2.42995396e-01
3.95216227e-01 -5.99722385e-01 -1.78007737e-01 1.69290558e-01
-1.61251819e+00 -2.80633550e-02 -2.05119193e-01 1.67429522e-01
-1.90544471e-01 5.11094570e-01 -7.97045648e-01 -7.10605681e-01
-3.43197137e-01 -9.45686042e-01 7.55610287e-01 1.54374287e-01
-1.00576091e+00 -1.44630492e+00 2.55895078e-01 5.82007647e-01
6.23473465e-01 -1.66785959e-02 1.16034842e+00 -8.51979017e-01
-4.48366329e-02 -3.30413193e-01 3.70152742e-02 4.14625704e-01
-1.61011979e-01 -2.40911637e-02 -7.73755789e-01 9.00754035e-02
-2.50227690e-01 -3.85775745e-01 4.79228199e-01 -2.89386988e-01
1.20747650e+00 -4.08869207e-01 2.29971446e-02 1.22756466e-01
1.22618711e+00 7.72824883e-03 5.86779177e-01 5.95068216e-01
5.03066957e-01 9.26361620e-01 7.49440074e-01 2.18205467e-01
7.17023313e-01 5.67281306e-01 2.23520510e-02 3.28493148e-01
-1.56910732e-01 -3.13090265e-01 1.15750365e-01 1.22922599e+00
-3.87325168e-01 -5.48902214e-01 -1.13300109e+00 6.33945048e-01
-1.43548453e+00 -8.44256341e-01 -4.38912213e-01 1.62765443e+00
1.42925692e+00 1.51246384e-01 -2.85852462e-01 3.58043522e-01
3.99644405e-01 -6.15718104e-02 3.35840672e-01 -9.61793542e-01
-2.00146392e-01 8.45183671e-01 1.42782062e-01 9.09103632e-01
-7.55676746e-01 1.27367735e+00 6.99847174e+00 6.71236336e-01
-8.43979180e-01 1.00003861e-01 3.44254076e-01 3.71925473e-01
-6.99726403e-01 1.12363407e-02 -1.03470504e+00 3.60593975e-01
8.34808052e-01 2.59529222e-02 1.83034033e-01 7.84945011e-01
-2.85052396e-02 -1.62302852e-01 -8.19537818e-01 6.17973864e-01
4.37278390e-01 -1.01317656e+00 -1.75569609e-01 -4.45386678e-01
8.45766962e-01 -5.19950569e-01 -2.65713334e-01 7.45289385e-01
3.49997759e-01 -1.07953000e+00 5.13055384e-01 5.05848825e-01
3.78708780e-01 -7.06485450e-01 1.17273009e+00 2.91090995e-01
-4.22594994e-01 2.14986339e-01 -1.12077810e-01 -5.78246891e-01
1.97281018e-01 2.43838474e-01 -7.77297974e-01 3.55529368e-01
6.13065362e-01 5.37114084e-01 -1.05587828e+00 6.44003570e-01
-8.96263063e-01 9.52852786e-01 -1.54016897e-01 -6.37879670e-01
1.37292251e-01 -3.52326483e-02 5.72121561e-01 1.63263524e+00
2.43506148e-01 3.31476182e-01 -7.58022964e-02 6.35644436e-01
-1.48488224e-01 6.14459991e-01 -4.25849944e-01 -1.59924164e-01
6.27169907e-01 1.09104812e+00 -3.41253698e-01 -4.42587227e-01
-3.67310286e-01 9.96418536e-01 8.61975789e-01 7.42197931e-02
-5.20500481e-01 -6.37321591e-01 4.64248717e-01 3.35310772e-02
-1.82468489e-01 -2.58221030e-01 -1.04014385e+00 -1.10165536e+00
9.23051462e-02 -1.27906060e+00 3.90450805e-01 -7.43416846e-01
-1.29387844e+00 3.76578271e-01 8.55459925e-03 -7.28813589e-01
-2.44385749e-01 -8.91349435e-01 -7.09845841e-01 1.08556533e+00
-1.43269265e+00 -7.44654894e-01 -6.91872761e-02 -4.60597128e-02
5.18526077e-01 -4.70863096e-02 1.13732493e+00 5.58498263e-01
-6.77259505e-01 1.03984559e+00 -3.13060433e-01 6.34487495e-02
9.29688871e-01 -1.49272609e+00 2.63234377e-01 1.03036726e+00
-7.05189956e-03 1.10888493e+00 1.10541737e+00 -7.14251041e-01
-6.85858190e-01 -7.51699924e-01 1.90694821e+00 -8.85213315e-01
8.04371893e-01 2.93227304e-02 -1.23225510e+00 7.85731971e-01
4.32088792e-01 -4.51128572e-01 9.52929437e-01 4.91503328e-01
-1.50887147e-01 4.82785344e-01 -1.07037127e+00 6.08555734e-01
1.04118121e+00 -6.17704153e-01 -1.14349413e+00 2.76779771e-01
2.88812757e-01 -7.95275033e-01 -9.01628852e-01 3.58671129e-01
1.28503352e-01 -3.95059824e-01 5.94070077e-01 -1.04811013e+00
7.06858695e-01 -2.29141906e-01 -4.15755808e-02 -1.76222563e+00
-5.43355793e-02 -3.12454015e-01 1.02505796e-01 1.49943876e+00
6.97540998e-01 -2.84577191e-01 6.97773457e-01 9.19380605e-01
-6.42578602e-01 -9.62920070e-01 -6.68431163e-01 -3.78931642e-01
3.44530940e-01 -6.28386021e-01 3.20366859e-01 1.29056656e+00
6.38391554e-01 5.02327740e-01 -1.79189399e-01 -1.08298033e-01
1.76660344e-01 -1.96706221e-01 4.59526062e-01 -1.01218355e+00
-1.76219240e-01 -4.11477268e-01 -2.57082522e-01 -7.74870694e-01
5.08834124e-01 -1.14943385e+00 7.83808902e-02 -1.30936480e+00
-9.51097310e-02 -4.15980846e-01 2.39743516e-01 6.39000416e-01
-7.07471490e-01 4.52589720e-01 2.28452925e-02 -3.57546091e-01
-4.72504228e-01 3.20166409e-01 8.10873926e-01 1.32327154e-01
1.72623068e-01 -3.82624805e-01 -8.70596588e-01 7.48444021e-01
1.13736105e+00 -5.32682359e-01 4.81570810e-02 -6.24567807e-01
5.98875523e-01 -3.23526114e-01 2.01632217e-01 -7.26878703e-01
5.58193512e-02 -1.46033093e-01 1.36487648e-01 -2.03376710e-01
-1.94199950e-01 -4.09281492e-01 -3.26621085e-01 1.79667339e-01
-5.70901752e-01 7.95192719e-01 3.04313332e-01 -1.21125095e-01
-2.73380786e-01 -1.03239679e+00 5.79879522e-01 -3.77183110e-01
-9.29247797e-01 -5.06759465e-01 -4.88227218e-01 4.91913825e-01
6.44545078e-01 -1.45214036e-01 -4.71898407e-01 -1.98348731e-01
-5.89102209e-01 2.15313166e-01 4.22425807e-01 3.94133896e-01
6.38435364e-01 -1.08898175e+00 -6.93645358e-01 5.17662056e-02
4.18852270e-01 -3.65018070e-01 -3.39353420e-02 7.15158224e-01
-8.15250397e-01 4.09770429e-01 -4.11889479e-02 -6.94003375e-03
-1.75636125e+00 6.19120942e-03 2.37513229e-01 -3.97321016e-01
-2.32343465e-01 1.13890290e+00 -9.90292132e-01 -8.86438012e-01
2.91068237e-02 -8.51219818e-02 -5.25586665e-01 -1.35765344e-01
4.37757820e-01 4.62456077e-01 4.21796530e-01 -5.05723059e-01
-1.50464028e-01 2.27909163e-01 -2.83227801e-01 7.03566149e-02
1.08096457e+00 -1.56413928e-01 -3.99111301e-01 4.58070785e-01
8.93621445e-01 4.58532929e-01 -1.86097145e-01 -1.53182313e-01
6.78540111e-01 -4.86910790e-01 -8.85784924e-02 -1.22258818e+00
-2.98174143e-01 7.57142305e-01 1.02099478e-01 2.28766967e-02
8.54608178e-01 -3.17488790e-01 7.53282189e-01 6.07079327e-01
4.09845233e-01 -1.57147157e+00 -1.84802353e-01 1.18351626e+00
7.25453854e-01 -1.30462480e+00 1.12540968e-01 -6.41629279e-01
-6.42484307e-01 1.17410290e+00 8.36189389e-01 -1.75390989e-02
3.50385815e-01 4.70426567e-02 2.52134442e-01 -3.69574308e-01
-7.40881741e-01 7.28305206e-02 3.09201241e-01 2.60715544e-01
1.51831043e+00 3.54647338e-02 -1.24594963e+00 6.16618037e-01
-7.08093941e-01 -8.23074505e-02 7.36961246e-01 1.09730411e+00
-3.68258715e-01 -1.57151306e+00 -1.86022833e-01 4.00661051e-01
-8.12236369e-01 -5.37377059e-01 -5.62573969e-01 9.21029568e-01
8.55160132e-02 8.28211725e-01 -2.21247926e-01 -7.61789009e-02
5.73884666e-01 5.55502057e-01 8.41501772e-01 -8.37225020e-01
-9.05031860e-01 -8.04299593e-01 9.09334123e-01 -1.33216873e-01
-5.04076421e-01 -8.03040028e-01 -1.14558268e+00 -5.71157753e-01
-4.01641577e-01 4.92891043e-01 7.23864019e-01 1.15578294e+00
4.26674224e-02 5.15176237e-01 -9.37505811e-03 -1.72924429e-01
-8.36653233e-01 -1.27121043e+00 -9.41504985e-02 5.27301133e-01
-2.46669948e-01 -5.08661032e-01 -3.67985338e-01 -4.19176854e-02] | [11.071439743041992, 10.670594215393066] |
5271dfc5-ff6d-4e5e-bc49-517a659853f6 | skin-lesion-segmentation-using-u-net-and-good | 1811.11314 | null | http://arxiv.org/abs/1811.11314v1 | http://arxiv.org/pdf/1811.11314v1.pdf | Skin lesion segmentation using U-Net and good training strategies | In this paper we approach the problem of skin lesion segmentation using a
convolutional neural network based on the U-Net architecture. We present a set
of training strategies that had a significant impact on the performance of this
model. We evaluated this method on the ISIC Challenge 2018 - Skin Lesion
Analysis Towards Melanoma Detection, obtaining threshold Jaccard index of
77.5%. | ['Teofilo E. deCampos', 'Fred Guth'] | 2018-11-27 | null | null | null | null | ['skin-lesion-segmentation'] | ['medical'] | [ 7.50680923e-01 1.69413149e-01 -3.06354225e-01 -1.62752822e-01
-6.18804932e-01 -4.74614829e-01 4.43914503e-01 8.89519826e-02
-8.40542257e-01 6.14043534e-01 -1.50851265e-01 -5.91712236e-01
-2.58167088e-01 -8.27946246e-01 -3.06779772e-01 -5.72696507e-01
-3.07262331e-01 -3.14259291e-01 2.94014812e-01 -1.78415731e-01
4.27177131e-01 4.36259866e-01 -1.21132922e+00 8.83041084e-01
9.12064552e-01 1.22087109e+00 -3.45024139e-01 1.34878552e+00
9.79747400e-02 5.94333529e-01 -6.28296077e-01 -5.87915480e-01
2.28811115e-01 -2.56865114e-01 -1.31472397e+00 -1.26475587e-01
5.48107982e-01 -6.18816018e-02 -7.25661516e-02 1.03052497e+00
4.60659742e-01 -3.91407520e-01 1.13567579e+00 -5.43098629e-01
-2.12349087e-01 5.59864044e-01 -3.97696286e-01 4.42104965e-01
1.44094497e-01 7.95291066e-02 7.56349742e-01 -8.13907012e-02
1.00221586e+00 7.25179970e-01 1.31049168e+00 7.27582932e-01
-6.08799398e-01 -1.08862862e-01 -4.27801639e-01 2.25963622e-01
-1.24335539e+00 5.24894111e-02 2.52559781e-01 -5.30877471e-01
9.23900306e-01 7.43985891e-01 7.47216821e-01 9.27288175e-01
2.95182198e-01 6.95017099e-01 1.41456079e+00 -7.85665154e-01
2.88685054e-01 -1.89606234e-01 3.37277770e-01 7.89274275e-01
2.32388884e-01 -4.85297441e-02 9.04727206e-02 -7.29791000e-02
7.86680460e-01 -3.17354918e-01 1.36695191e-01 3.37599218e-01
-6.72411144e-01 7.84358680e-01 9.74342525e-01 4.43829715e-01
-4.57539320e-01 3.77628803e-01 6.92963898e-01 9.79524404e-02
5.10348737e-01 5.06638050e-01 -1.57828983e-02 2.58732289e-01
-1.01044297e+00 -1.80728540e-01 6.36722565e-01 9.68014896e-02
1.10261284e-01 -5.87810457e-01 -5.65648079e-01 6.20245934e-01
1.12350024e-01 2.33660061e-02 2.62114018e-01 -7.44070292e-01
-8.77824873e-02 8.76970649e-01 -2.43946999e-01 -5.50311446e-01
-5.55300951e-01 -4.51415658e-01 -1.02913129e+00 2.10214645e-01
6.23970509e-01 -6.78335011e-01 -1.64566696e+00 8.70704770e-01
1.59372613e-01 2.12392479e-01 -4.30710651e-02 7.41257787e-01
7.19138086e-01 -1.29373893e-01 3.99679691e-01 3.67141783e-01
1.28532636e+00 -8.94535601e-01 -6.75283611e-01 3.96754563e-01
8.27676535e-01 -9.33404446e-01 3.02097797e-01 3.61945361e-01
-8.62545490e-01 -1.55290470e-01 -1.04517674e+00 3.02926719e-01
-8.15718472e-01 3.90741229e-01 7.28673697e-01 1.25107884e+00
-1.44268918e+00 7.84028947e-01 -6.03059947e-01 -9.19881880e-01
9.15581346e-01 3.91915023e-01 -4.40473050e-01 -5.06373681e-03
-1.20325291e+00 8.90035272e-01 5.67016423e-01 2.72648335e-01
-3.86429042e-01 -4.61103857e-01 -5.17928481e-01 -3.51025492e-01
1.65204909e-02 -5.01941264e-01 1.15072978e+00 -1.10409153e+00
-1.07172596e+00 9.85089302e-01 7.86502659e-02 -1.12978280e+00
8.34156990e-01 1.48593830e-02 -4.60346937e-01 5.73216379e-01
-2.84981519e-01 8.21605086e-01 4.68929917e-01 -8.39425206e-01
-8.80359352e-01 -1.53855368e-01 7.22452402e-02 -1.49680570e-01
-3.93958896e-01 8.26729387e-02 -4.41997617e-01 -3.72678816e-01
-2.99084753e-01 -7.23239064e-01 -6.15079403e-01 -1.81298822e-01
-5.04699886e-01 -2.54624456e-01 3.78359258e-01 -1.02128148e+00
1.31400609e+00 -1.60414565e+00 -2.28775561e-01 8.74935329e-01
2.41850674e-01 7.98393846e-01 -5.85258827e-02 2.68328846e-01
-1.54891923e-01 6.88001454e-01 -3.52667451e-01 1.56531081e-01
-3.13388169e-01 -1.21354297e-01 3.41586143e-01 3.05210114e-01
4.59735453e-01 9.48395729e-01 -6.65287852e-01 -6.55692577e-01
4.25352037e-01 5.87245107e-01 -3.44917387e-01 -1.11094683e-01
-1.81253791e-01 -2.55403161e-01 -1.07597284e-01 1.08525991e+00
6.98408782e-01 7.40008652e-02 2.01212168e-01 -2.65541375e-01
2.41528153e-02 -4.53448683e-01 -8.61554325e-01 1.39515650e+00
-1.48416027e-01 7.54985988e-01 -8.30897614e-02 -5.53036153e-01
4.89138693e-01 2.34262705e-01 6.38230026e-01 -5.54102242e-01
4.79593486e-01 9.72952098e-02 3.36769342e-01 -7.22364604e-01
3.09795111e-01 3.12709689e-01 4.21525061e-01 1.41644925e-02
1.91253290e-01 3.55270863e-01 5.61627984e-01 1.43889070e-01
1.25978231e+00 -1.43475637e-01 4.65308428e-01 -4.27248836e-01
6.74464345e-01 3.49140972e-01 1.30454749e-01 8.50476503e-01
-8.60118508e-01 7.75110424e-01 7.52916455e-01 -5.76112330e-01
-8.08880031e-01 -8.07446957e-01 -5.33962131e-01 5.76434791e-01
-3.72073501e-01 -1.71516836e-01 -1.54192770e+00 -9.11595702e-01
-1.43959850e-01 2.09677473e-01 -1.48230970e+00 2.17710808e-01
-2.95463800e-01 -1.18842578e+00 1.01047170e+00 5.59653759e-01
4.85621929e-01 -1.04500496e+00 -5.26416600e-01 -5.26210144e-02
2.89091259e-01 -9.06444430e-01 -1.45100623e-01 2.49139354e-01
-5.33301711e-01 -1.96111214e+00 -1.02840674e+00 -8.03417027e-01
9.57242072e-01 -2.04627037e-01 8.44419003e-01 2.09887594e-01
-1.20782673e+00 3.37182552e-01 -5.08290589e-01 -5.35067201e-01
-4.16790098e-01 4.91966516e-01 -5.78564525e-01 5.42782955e-02
6.90460026e-01 3.06209475e-01 -7.56427348e-01 -9.80230514e-03
-1.30761349e+00 -2.31096864e-01 7.97328353e-01 7.11163223e-01
5.76287568e-01 5.92357572e-03 4.37138170e-01 -1.26316273e+00
9.17523444e-01 -2.97538012e-01 -1.56530917e-01 6.24105692e-01
-2.94725209e-01 -4.51256275e-01 2.10745007e-01 -1.70308072e-02
-7.58159637e-01 4.84589040e-01 -5.47074318e-01 -4.85977307e-02
-4.80692387e-01 4.54014957e-01 5.36951005e-01 -6.81227028e-01
1.01903689e+00 -2.36834422e-01 2.01740503e-01 -6.96000382e-02
1.77934542e-01 9.36669230e-01 5.16192496e-01 9.58704650e-02
2.75185496e-01 3.20906192e-01 4.06139553e-01 -7.96352565e-01
-8.26862931e-01 -7.91563809e-01 -9.34007406e-01 -5.27780890e-01
1.33489656e+00 -5.30644298e-01 -4.68213052e-01 9.44407403e-01
-8.79535556e-01 -4.06476557e-01 6.19658269e-02 -2.22944990e-02
-7.30053186e-02 2.54212618e-01 -7.01864064e-01 -7.24613369e-01
-7.48600721e-01 -9.51590896e-01 6.11385047e-01 6.80000663e-01
-1.95661068e-01 -1.53979075e+00 2.09612027e-01 4.48960215e-01
5.99441648e-01 8.50939929e-01 5.73854387e-01 -6.62203312e-01
2.30522677e-01 -6.72215641e-01 -4.75466043e-01 5.72027326e-01
3.45255509e-02 4.94622529e-01 -1.20361471e+00 6.55739605e-02
-8.34378004e-01 -1.42777011e-01 1.44808507e+00 7.20137656e-01
1.42480385e+00 3.16106021e-01 -3.44831169e-01 7.22726583e-01
1.77591562e+00 4.14181836e-02 8.72270107e-01 3.61205220e-01
4.82621491e-01 5.23268044e-01 3.08937341e-01 -1.39249235e-01
-5.70985042e-02 -4.34706397e-02 7.33866632e-01 -5.89996636e-01
-5.56510091e-01 6.56402484e-02 -3.36351365e-01 1.77277878e-01
-3.86260569e-01 3.01526245e-02 -1.42403018e+00 6.83933914e-01
-1.49181306e+00 -6.02897465e-01 -3.42434287e-01 1.78845775e+00
6.04865849e-01 3.05601478e-01 2.30476588e-01 1.11333206e-01
9.18275595e-01 -1.70052558e-01 -2.42940128e-01 -7.63457179e-01
-1.62303403e-01 5.30824840e-01 7.99983084e-01 5.09893179e-01
-1.76916385e+00 9.29258823e-01 7.87786865e+00 9.43009853e-01
-1.23302794e+00 -2.82656580e-01 1.01316297e+00 3.58112127e-01
1.06196783e-01 -5.97302079e-01 -4.68847066e-01 2.44601250e-01
1.01358759e+00 3.88584614e-01 -1.17928013e-01 4.94378686e-01
-2.32161209e-01 -4.12894636e-01 -6.10735297e-01 4.83532101e-01
1.52576014e-01 -1.72108722e+00 -8.02999958e-02 1.11035749e-01
9.18014705e-01 1.18631892e-01 2.25590989e-01 3.94461909e-03
1.32898048e-01 -1.53808331e+00 -3.10854524e-01 8.23495924e-01
1.07135236e+00 -8.25419366e-01 1.43044007e+00 -4.38034773e-01
-6.11593843e-01 1.30218476e-01 -2.18494624e-01 1.96087211e-01
-3.14059228e-01 5.24948239e-01 -1.26509011e+00 5.68512142e-01
4.24717724e-01 4.58950251e-01 -1.20990038e+00 1.71034050e+00
9.28025767e-02 1.04102039e+00 -2.22118303e-01 -3.39797109e-01
6.28029883e-01 1.16952576e-01 1.96164936e-01 1.66444373e+00
-6.72228932e-02 -3.75997394e-01 -2.77076572e-01 4.38817412e-01
7.84499347e-02 3.70394409e-01 -4.37450051e-01 -6.20551556e-02
-1.05573095e-01 1.59351039e+00 -1.11632562e+00 -5.79244383e-02
-1.83537796e-01 7.27644026e-01 -1.80117879e-02 2.63628900e-01
-5.33105910e-01 -6.24116421e-01 3.41290325e-01 -7.87846185e-03
1.33068889e-01 4.42586124e-01 -6.31527245e-01 -6.02277339e-01
-2.97229797e-01 -5.56966960e-01 7.86664009e-01 -2.68600196e-01
-1.39663231e+00 6.25987172e-01 -4.80025619e-01 -7.96085954e-01
1.75391268e-02 -1.21388316e+00 -1.31875038e+00 1.01574600e+00
-1.73281181e+00 -1.29938948e+00 -3.39717358e-01 3.53768080e-01
5.80485240e-02 -4.02384967e-01 9.94676948e-01 -5.99735752e-02
-8.40103149e-01 6.62489891e-01 1.94493875e-01 4.56284225e-01
7.34120667e-01 -1.71817744e+00 5.83100855e-01 8.93517554e-01
-3.34593117e-01 3.71657670e-01 1.01236559e-01 -5.99370718e-01
-8.12493205e-01 -1.38323033e+00 6.65917575e-01 -2.72021264e-01
7.94242680e-01 -3.36053744e-02 -4.29616660e-01 2.71885097e-01
6.00551128e-01 -5.27262017e-02 1.23183131e+00 2.25071281e-01
-1.59434557e-01 3.29741508e-01 -1.49005198e+00 4.58805859e-01
4.10483837e-01 -3.23876500e-01 -1.50885925e-01 4.96063709e-01
1.36304274e-01 -2.66327262e-01 -1.14865947e+00 6.14843905e-01
5.23253262e-01 -7.60246873e-01 9.17256713e-01 -9.56898034e-01
7.19593227e-01 2.48706356e-01 1.15215071e-01 -1.25396466e+00
-2.89384633e-01 -3.15396965e-01 1.67528018e-02 6.63723171e-01
5.95708191e-01 -4.10650373e-01 1.19703329e+00 4.12529618e-01
1.16089590e-01 -1.23501718e+00 -7.52261162e-01 -2.13045254e-01
6.20166898e-01 -2.01292306e-01 -7.52065331e-03 6.18641913e-01
-2.37806454e-01 -3.31929386e-01 7.60160685e-02 -1.66005775e-01
6.34005666e-01 -5.67647755e-01 6.64267465e-02 -1.16830742e+00
3.58690977e-01 -6.73428833e-01 -5.90805233e-01 1.35419697e-01
-1.89375550e-01 -9.21951890e-01 -2.92024553e-01 -1.64053023e+00
2.14400709e-01 -5.55911586e-02 -8.55971038e-01 4.32453781e-01
-4.18179214e-01 7.41055548e-01 -7.90849924e-02 -2.03506619e-01
-6.37511015e-01 -6.27810299e-01 1.06455874e+00 -2.82030970e-01
5.73665053e-02 1.25621021e-01 -6.10026300e-01 6.80699110e-01
1.43597150e+00 -1.91995464e-02 6.03074133e-02 -3.18494290e-01
1.31542772e-01 -4.37049776e-01 3.58109146e-01 -1.28996730e+00
4.62188810e-01 -1.69926342e-02 8.41013432e-01 -5.22798896e-01
-2.02169508e-01 -5.17609000e-01 -4.53944743e-01 1.07085073e+00
-5.63930809e-01 -4.54914182e-01 3.43789399e-01 3.34382147e-01
-2.08295390e-01 -5.93404412e-01 9.14486229e-01 -1.66721761e-01
-8.37557375e-01 8.11604932e-02 -4.60975736e-01 -3.98608178e-01
1.43655360e+00 -2.27654696e-01 -4.69390661e-01 2.57744610e-01
-8.74375641e-01 2.84996450e-01 2.28486478e-01 1.09942026e-01
3.42376113e-01 -1.03633904e+00 -8.55385959e-01 -2.48317793e-02
1.71994746e-01 -4.52878237e-01 4.57462013e-01 9.32699800e-01
-1.25107265e+00 7.22004890e-01 -7.26927042e-01 -5.02596915e-01
-1.59623909e+00 -2.64432169e-02 9.28478062e-01 -4.54186916e-01
-1.47495046e-01 1.06811190e+00 -7.14541674e-01 -5.70438564e-01
3.95413309e-01 -6.13814965e-02 -8.94133270e-01 2.00199753e-01
5.52671909e-01 5.57323158e-01 2.27079809e-01 -2.54314005e-01
-3.38761449e-01 3.65931392e-01 -4.40989316e-01 2.18407765e-01
1.10722315e+00 4.86910343e-01 -3.61135572e-01 -1.68374628e-01
1.05314708e+00 -6.41262472e-01 -8.58860195e-01 1.55500665e-01
6.50522560e-02 -1.67811126e-01 3.90402973e-01 -1.51152802e+00
-1.00616145e+00 7.71581292e-01 1.30909228e+00 5.38036346e-01
1.15455651e+00 -6.42134666e-01 7.36324906e-01 4.11237687e-01
-3.45213234e-01 -1.47280920e+00 -3.44051868e-01 3.60472232e-01
2.87265211e-01 -1.56995893e+00 -2.86599454e-02 -7.17195332e-01
-4.49024916e-01 1.51233959e+00 5.48359573e-01 -4.35024053e-01
7.59539664e-01 2.82893151e-01 4.41878647e-01 -2.09636360e-01
-4.27641630e-01 -5.64731181e-01 6.76425457e-01 8.60559642e-01
7.25765526e-01 4.76013601e-01 -6.58958137e-01 1.95071638e-01
3.12219024e-01 2.82177180e-01 5.08766472e-01 7.82025635e-01
-4.22949106e-01 -1.08226633e+00 -2.46692479e-01 9.15996850e-01
-9.36436117e-01 -5.73696569e-02 -1.15937924e+00 8.20502222e-01
4.42543089e-01 6.07349813e-01 2.47597508e-02 -3.60429138e-01
-1.71618164e-01 4.53177579e-02 4.32144582e-01 -2.45814964e-01
-1.08336520e+00 -9.10619497e-02 2.97059357e-01 -3.47434521e-01
-5.26285946e-01 -3.74349892e-01 -8.00478339e-01 2.74683684e-02
-2.79519781e-02 -1.18969865e-01 1.07148767e+00 6.04650676e-01
9.80079025e-02 9.12509203e-01 2.62597769e-01 -1.53598040e-01
-3.15885335e-01 -1.18126142e+00 -4.36279386e-01 4.33022261e-01
4.90356088e-02 -1.05195947e-01 5.00905141e-02 2.01335605e-02] | [15.707708358764648, -3.003040313720703] |
d8d388bd-9bce-4309-abc0-1d5d5c3c3d1b | learning-word-representations-with-1 | null | null | https://aclanthology.org/K17-1016 | https://aclanthology.org/K17-1016.pdf | Learning Word Representations with Regularization from Prior Knowledge | Conventional word embeddings are trained with specific criteria (e.g., based on language modeling or co-occurrence) inside a single information source, disregarding the opportunity for further calibration using external knowledge. This paper presents a unified framework that leverages pre-learned or external priors, in the form of a regularizer, for enhancing conventional language model-based embedding learning. We consider two types of regularizers. The first type is derived from topic distribution by running LDA on unlabeled data. The second type is based on dictionaries that are created with human annotation efforts. To effectively learn with the regularizers, we propose a novel data structure, trajectory softmax, in this paper. The resulting embeddings are evaluated by word similarity and sentiment classification. Experimental results show that our learning framework with regularization from prior knowledge improves embedding quality across multiple datasets, compared to a diverse collection of baseline methods. | ['Chia-Jung Lee', 'Yan Song', 'Fei Xia'] | 2017-08-01 | null | null | null | conll-2017-8 | ['learning-word-embeddings'] | ['methodology'] | [-3.67293842e-02 2.40984023e-01 -7.78132796e-01 -4.50052232e-01
-6.80349708e-01 -5.10497928e-01 7.71663666e-01 1.77015543e-01
-7.56148696e-01 3.85790169e-01 5.39854109e-01 -1.86330140e-01
1.97568014e-01 -8.01432610e-01 -5.67808807e-01 -6.95159256e-01
2.09550977e-01 2.13279933e-01 -1.20861128e-01 -3.84562351e-02
8.50458592e-02 -1.56794950e-01 -1.06756759e+00 -1.15835607e-01
9.38615620e-01 9.42014575e-01 1.50251254e-01 9.45026428e-02
-3.89656574e-01 3.53675961e-01 -3.30940962e-01 -6.52264357e-01
2.03741193e-01 -1.94925591e-02 -3.80863637e-01 2.85387207e-02
3.77967618e-02 -3.59951675e-01 -3.51596922e-01 8.32081199e-01
3.48737508e-01 7.91845024e-02 8.24368298e-01 -1.06776261e+00
-1.15913713e+00 6.47117972e-01 -2.87885457e-01 -1.18625313e-01
-1.67255938e-01 -1.41628847e-01 1.41705477e+00 -1.43998969e+00
3.83279502e-01 1.24398935e+00 7.65624642e-01 3.96873862e-01
-1.43939495e+00 -5.27430296e-01 5.70375741e-01 9.55405552e-03
-1.30388713e+00 -2.52781987e-01 1.01128900e+00 -6.65632129e-01
8.93525720e-01 -1.98129579e-01 2.99057424e-01 1.31290948e+00
-5.48226722e-02 9.27725673e-01 8.31350505e-01 -7.83611953e-01
3.90812367e-01 7.31203973e-01 4.85806823e-01 4.11174506e-01
3.85587901e-01 -9.41985697e-02 -3.30019146e-01 -3.55391771e-01
3.38820010e-01 1.36134937e-01 -1.35505885e-01 -7.28703678e-01
-8.54861021e-01 1.31922841e+00 2.32022807e-01 1.29224941e-01
-1.28978103e-01 3.98742855e-02 4.16455567e-01 -3.38493288e-02
9.12603557e-01 4.12992239e-01 -7.40036368e-01 3.15724947e-02
-8.34684730e-01 -2.48724312e-01 6.61733091e-01 8.77384663e-01
1.19555497e+00 1.77132159e-01 -9.59582403e-02 1.07719016e+00
9.11422670e-01 3.98494959e-01 9.76378858e-01 -2.43770063e-01
3.92795444e-01 5.79881251e-01 8.50875601e-02 -8.96927834e-01
-9.05896202e-02 -3.36607575e-01 -2.34831095e-01 -1.78498223e-01
8.13165680e-02 -4.07139301e-01 -9.87351000e-01 1.90173745e+00
2.95212835e-01 4.85853046e-01 1.09055951e-01 5.76140165e-01
6.84443891e-01 6.84596419e-01 2.46885508e-01 7.26680504e-03
1.34791386e+00 -1.32138085e+00 -9.78979170e-01 -4.28326994e-01
8.11738908e-01 -5.83185732e-01 1.45514560e+00 2.43243814e-01
-4.59374160e-01 -5.85261226e-01 -1.16995656e+00 -6.94005564e-02
-7.79917598e-01 3.62340361e-01 5.64022124e-01 9.44957197e-01
-7.71980941e-01 2.63806105e-01 -9.97025669e-01 -2.49928758e-01
8.53816196e-02 1.84378296e-01 -1.03013396e-01 1.57523856e-01
-1.42111802e+00 8.66869032e-01 5.19246519e-01 -3.11641514e-01
-7.44205773e-01 -8.02619636e-01 -1.22051144e+00 -3.53363566e-02
2.65449047e-01 -3.42921883e-01 1.02527165e+00 -9.38552856e-01
-1.75567698e+00 7.18893111e-01 -1.32656083e-01 -4.46652085e-01
7.47196898e-02 -6.79622114e-01 -3.15203398e-01 -1.87094912e-01
4.11919244e-02 5.22269785e-01 1.03954768e+00 -1.40975881e+00
-1.88369542e-01 -1.08143441e-01 4.33834940e-02 -4.37810435e-04
-1.38801277e+00 -1.35790139e-01 -8.64091992e-01 -1.05803347e+00
-3.94165844e-01 -9.76145148e-01 -2.53801733e-01 2.21580286e-02
-3.09959114e-01 -4.83732164e-01 8.58572245e-01 -6.13648295e-01
1.70678115e+00 -2.15945935e+00 5.00105927e-03 2.83226460e-01
-3.95216839e-03 3.09196442e-01 -2.68251628e-01 5.60302079e-01
-1.02021731e-01 2.59535253e-01 -2.22791761e-01 -7.91181862e-01
4.19611841e-01 5.43377995e-01 -6.58653200e-01 3.38935822e-01
4.70969856e-01 8.10873866e-01 -8.79710078e-01 -3.26345086e-01
1.11123025e-01 7.12608099e-01 -7.32286870e-01 4.97990608e-01
-1.90440640e-01 1.30855888e-02 -3.41329485e-01 2.88868427e-01
5.09978652e-01 -3.92533541e-01 2.60436416e-01 -3.24008167e-01
7.68418610e-02 5.29066265e-01 -1.12218761e+00 1.89387763e+00
-8.22464943e-01 3.32886010e-01 -1.11563675e-01 -1.16482413e+00
1.03000820e+00 4.18438882e-01 3.76307011e-01 -8.57097134e-02
2.73396313e-01 -6.31140620e-02 -3.00718069e-01 -4.32906032e-01
4.45172489e-01 5.22252619e-02 -1.48188978e-01 6.58442914e-01
5.40793896e-01 1.08773671e-01 2.01269656e-01 2.59796172e-01
7.42026985e-01 2.28736445e-01 2.95236588e-01 -1.94867224e-01
3.30511600e-01 -3.37093472e-01 6.91145003e-01 4.71927494e-01
8.94482061e-03 5.47662735e-01 2.71180093e-01 -1.45560771e-01
-1.00283265e+00 -8.99847090e-01 -6.24924958e-01 1.28997505e+00
-8.34389925e-02 -9.85596895e-01 -5.65153718e-01 -9.57886279e-01
2.06499919e-01 8.27759027e-01 -7.73969829e-01 -4.40877557e-01
-2.08439916e-01 -9.62735653e-01 2.57404715e-01 7.96671629e-01
-1.49042886e-02 -5.77217579e-01 -9.96256992e-02 2.05947980e-01
1.66998237e-01 -1.16303372e+00 -6.28248215e-01 4.42077905e-01
-8.04028988e-01 -6.44638062e-01 -5.83576143e-01 -5.97579896e-01
8.36876214e-01 1.68694124e-01 9.23259914e-01 -2.10450366e-01
1.42349273e-01 8.68223071e-01 -5.41135669e-01 -4.60655302e-01
1.45601695e-02 1.59251511e-01 3.76741469e-01 2.58083314e-01
7.89133251e-01 -5.73879063e-01 -4.08650756e-01 3.17766480e-02
-1.07390893e+00 -3.98766786e-01 7.62143970e-01 1.12655950e+00
5.24525285e-01 -2.44772717e-01 5.46031296e-01 -1.07110679e+00
7.71708071e-01 -8.44962835e-01 -4.56290931e-01 1.92748338e-01
-1.02476132e+00 3.44897181e-01 4.89050835e-01 -8.23190272e-01
-1.11191332e+00 -3.83866206e-02 4.63193795e-03 -5.88467658e-01
1.83142070e-02 8.70082378e-01 -3.08897704e-01 9.44564641e-02
4.80958939e-01 5.45021929e-02 -1.96144864e-01 -7.08978951e-01
8.04787755e-01 7.27450669e-01 8.64708573e-02 -8.86036277e-01
1.01502407e+00 3.92242342e-01 -8.42610478e-01 -6.56852424e-01
-9.88922715e-01 -5.68057954e-01 -7.14830041e-01 1.99709699e-01
8.04648280e-01 -1.19913983e+00 1.20673880e-01 1.31292164e-01
-1.03907800e+00 -2.53050178e-01 -3.57350916e-01 9.00649667e-01
-1.23329952e-01 3.97344679e-01 -5.58224976e-01 -6.92022562e-01
-2.39923403e-01 -9.54953849e-01 9.90400255e-01 1.51141249e-02
-2.25801975e-01 -1.64108825e+00 5.75006604e-01 5.15113287e-02
3.96489322e-01 -4.31019776e-02 8.16854656e-01 -1.06127048e+00
-1.92100987e-01 -3.97678226e-01 -3.72430831e-02 8.14091682e-01
4.49134499e-01 -8.87351297e-03 -1.22233558e+00 -4.68642175e-01
1.10778725e-02 -4.53926384e-01 8.83025467e-01 1.24022692e-01
1.09834504e+00 -3.48207444e-01 -5.10662377e-01 6.16758108e-01
1.60083354e+00 -1.97354957e-01 3.28431517e-01 3.71911854e-01
6.40498400e-01 3.57981563e-01 3.68702203e-01 6.01402044e-01
5.49801171e-01 7.03101993e-01 1.20718986e-01 -1.72523335e-01
-2.57134810e-02 -4.45751250e-01 5.65953851e-01 1.35734189e+00
1.82742268e-01 -1.39975920e-01 -9.47646499e-01 7.12108910e-01
-1.73477888e+00 -3.24233502e-01 3.03262353e-01 2.01465344e+00
1.15472245e+00 2.14516133e-01 -6.60184696e-02 -1.26957387e-01
5.13355315e-01 1.59347847e-01 -5.72028160e-01 -3.00976068e-01
7.09961355e-02 3.65432113e-01 4.37705725e-01 5.92353404e-01
-1.22683489e+00 9.71140265e-01 6.04327965e+00 7.86671758e-01
-1.09432125e+00 4.48134363e-01 2.52720624e-01 1.07401736e-01
-5.57307363e-01 1.72875360e-01 -1.07457757e+00 5.23023129e-01
9.21612263e-01 -1.83544964e-01 4.43084463e-02 1.05060518e+00
8.31928849e-03 2.99260497e-01 -1.15170288e+00 7.29547143e-01
4.00754362e-01 -8.53932500e-01 -2.61886753e-02 4.73955609e-02
7.95068502e-01 6.06192797e-02 3.48836243e-01 6.59389138e-01
7.22930074e-01 -6.28177226e-01 5.87354600e-01 3.10067207e-01
5.39677680e-01 -4.77954119e-01 7.06228018e-01 1.59894213e-01
-1.05795288e+00 3.01961359e-02 -4.04421091e-01 1.39958128e-01
2.90584803e-01 7.82605827e-01 -6.36061430e-01 4.25473183e-01
4.34982300e-01 9.40731704e-01 -4.86898184e-01 6.74442291e-01
-5.97963154e-01 1.15792441e+00 -3.04450274e-01 1.38799459e-01
2.46137530e-01 -2.23430991e-01 3.12047422e-01 1.46549976e+00
2.35026702e-01 -3.56101781e-01 3.57257128e-01 7.50312090e-01
-1.71375498e-01 4.38127488e-01 -5.68984747e-01 -4.51609232e-02
5.24520278e-01 1.43204606e+00 -3.39163929e-01 -3.44515502e-01
-9.84365106e-01 6.60732210e-01 5.26353955e-01 5.72212160e-01
-8.26750875e-01 -4.69786078e-01 6.65627480e-01 -8.87930915e-02
6.17608964e-01 -3.87134135e-01 -1.70369491e-01 -1.52487195e+00
1.70551408e-02 -5.70688367e-01 3.53816658e-01 -3.22992772e-01
-1.79713750e+00 5.30796528e-01 1.82346880e-01 -1.12092268e+00
-1.19850352e-01 -9.85358536e-01 -6.81170523e-01 8.04400980e-01
-2.13675117e+00 -1.21154439e+00 3.96845303e-02 4.21104640e-01
3.52837592e-01 -2.51771808e-01 9.83431339e-01 4.69969690e-01
-8.95269156e-01 9.17568028e-01 3.69886845e-01 3.14649493e-01
1.16754889e+00 -1.33829188e+00 1.16168723e-01 7.02521741e-01
3.92193407e-01 9.47317183e-01 3.97022486e-01 -4.08764631e-01
-1.18892920e+00 -1.20954430e+00 7.15308845e-01 -6.56223893e-01
1.04118705e+00 -7.48101771e-01 -1.08803844e+00 8.58821511e-01
4.94403660e-01 1.86643168e-01 1.45660532e+00 5.32875299e-01
-8.22545469e-01 -1.54095784e-01 -6.97939038e-01 2.97917187e-01
5.41465998e-01 -7.21917868e-01 -8.23654115e-01 2.41994590e-01
9.89630520e-01 6.92656115e-02 -9.48725343e-01 3.76230001e-01
3.18897992e-01 -1.33538321e-01 1.02712607e+00 -9.81055319e-01
2.16094583e-01 -1.95583656e-01 -3.16880465e-01 -1.45626795e+00
-2.21200064e-01 -4.20440584e-01 -5.31162083e-01 1.61095548e+00
5.51838398e-01 -7.08223164e-01 6.44777000e-01 8.77692044e-01
-1.24073319e-01 -7.37921417e-01 -5.80633938e-01 -9.14724171e-01
4.38341975e-01 -5.77497542e-01 3.59616578e-01 1.25970447e+00
1.98555425e-01 5.54235041e-01 -4.68091667e-01 2.33186632e-01
4.48064625e-01 -2.23614380e-01 8.31868112e-01 -1.17447412e+00
-3.67674172e-01 -1.78727239e-01 -1.31435588e-01 -1.28811049e+00
6.15914345e-01 -1.07507467e+00 -3.56197320e-02 -1.30481148e+00
4.29413021e-02 -5.27557313e-01 -7.06878424e-01 7.02237666e-01
-4.54685450e-01 1.65826365e-01 -9.88235250e-02 2.52428532e-01
-5.52121103e-01 1.04129112e+00 6.66401803e-01 -1.40093327e-01
-2.37002835e-01 -3.66738677e-01 -8.66306782e-01 8.89863253e-01
7.35296011e-01 -6.04303122e-01 -5.68793058e-01 -6.36769176e-01
3.27348411e-01 -7.92977691e-01 -6.16007531e-03 -6.84515834e-01
1.09286591e-01 6.09510615e-02 5.06603308e-02 -1.99949920e-01
4.13903594e-01 -8.18445444e-01 -5.53686082e-01 1.92047823e-02
-5.15778303e-01 -1.08991742e-01 2.01087266e-01 8.95160615e-01
-3.13942075e-01 -4.70748127e-01 4.61966485e-01 2.48579264e-01
-6.48411036e-01 1.79980397e-01 -2.69249082e-01 2.10263625e-01
8.99307311e-01 2.01603785e-01 -1.55691057e-02 -3.52178574e-01
-7.41448283e-01 2.40287423e-01 1.71647638e-01 6.45723164e-01
3.44073385e-01 -1.76218569e+00 -4.62990016e-01 1.37770012e-01
5.18169999e-01 -3.13347727e-01 -3.63590181e-01 5.22889018e-01
4.22886670e-01 4.14284021e-01 3.40520859e-01 -4.21868056e-01
-6.75302267e-01 8.19448233e-01 -1.50677770e-01 -5.40150046e-01
-2.26302415e-01 7.53009915e-01 3.81866187e-01 -7.35998929e-01
4.20851171e-01 -4.45436329e-01 -3.85305554e-01 2.85967559e-01
3.30162376e-01 -1.49134606e-01 7.51507096e-03 -5.35726130e-01
-1.60776496e-01 6.39381468e-01 -2.48777375e-01 -9.53542516e-02
1.48693335e+00 -1.97133794e-01 8.12495276e-02 6.90531909e-01
1.37574315e+00 2.26154014e-01 -1.33176637e+00 -9.04383659e-01
2.87438065e-01 -4.47073132e-01 3.10807019e-01 -5.02398551e-01
-8.76214743e-01 9.99188304e-01 4.66424733e-01 1.12472355e-01
7.47352481e-01 -5.94679974e-02 7.44065344e-01 4.94775563e-01
-5.37078269e-02 -1.32537723e+00 4.75097060e-01 5.61082721e-01
5.46480417e-01 -1.38269842e+00 -1.33249313e-01 -1.71525672e-01
-6.41058743e-01 1.08430707e+00 7.00201750e-01 -3.00071865e-01
1.24396825e+00 3.00285965e-01 2.22995520e-01 1.60438687e-01
-7.43696153e-01 -2.10218295e-01 5.92631698e-01 4.53041762e-01
5.66134930e-01 -2.41299510e-01 -4.84489620e-01 1.20373631e+00
1.12627424e-01 -2.06475005e-01 5.25641888e-02 9.01345551e-01
-2.92231202e-01 -1.61414123e+00 -1.13465026e-01 2.73352444e-01
-3.93378615e-01 -3.60776305e-01 -5.53637892e-02 5.68643689e-01
2.75464684e-01 7.88393736e-01 1.53428450e-01 -2.43440717e-01
1.01719029e-01 3.91315132e-01 -2.77970619e-02 -9.69159663e-01
-2.26595461e-01 2.20841363e-01 2.66037658e-02 -1.05065823e-01
-5.52358985e-01 -5.07881463e-01 -7.91418672e-01 4.11919326e-01
-7.97808409e-01 2.77701020e-01 6.61599040e-01 9.51304317e-01
4.30575401e-01 3.23875993e-01 6.69416487e-01 -6.78941071e-01
-6.87701702e-01 -1.12015367e+00 -3.81163657e-01 4.55229223e-01
2.13775471e-01 -8.66712689e-01 -5.42207658e-01 2.95047313e-01] | [10.521241188049316, 8.573678016662598] |
ab677083-a07e-478b-8e80-7eabd1693a8f | analyzing-the-impact-of-feature-selection-on | 2206.03239 | null | https://arxiv.org/abs/2206.03239v1 | https://arxiv.org/pdf/2206.03239v1.pdf | Analyzing the impact of feature selection on the accuracy of heart disease prediction | Heart Disease has become one of the most serious diseases that has a significant impact on human life. It has emerged as one of the leading causes of mortality among the people across the globe during the last decade. In order to prevent patients from further damage, an accurate diagnosis of heart disease on time is an essential factor. Recently we have seen the usage of non-invasive medical procedures, such as artificial intelligence-based techniques in the field of medical. Specially machine learning employs several algorithms and techniques that are widely used and are highly useful in accurately diagnosing the heart disease with less amount of time. However, the prediction of heart disease is not an easy task. The increasing size of medical datasets has made it a complicated task for practitioners to understand the complex feature relations and make disease predictions. Accordingly, the aim of this research is to identify the most important risk-factors from a highly dimensional dataset which helps in the accurate classification of heart disease with less complications. For a broader analysis, we have used two heart disease datasets with various medical features. The classification results of the benchmarked models proved that there is a high impact of relevant features on the classification accuracy. Even with a reduced number of features, the performance of the classification models improved significantly with a reduced training time as compared with models trained on full feature set. | ['Soumyabrata Dev', 'Muhammad Mohisn Pathan', 'Avishek Nag', 'Muhammad Salman Pathan'] | 2022-06-07 | null | null | null | null | ['disease-prediction'] | ['medical'] | [-5.16010858e-02 -2.47755826e-01 -1.79714903e-01 -1.96748659e-01
-3.19127738e-02 -8.89794156e-02 1.81836084e-01 6.39271855e-01
-3.38995129e-01 7.39174902e-01 -7.74354637e-02 -2.23809272e-01
-6.58502698e-01 -8.88460755e-01 2.40440175e-01 -6.86260283e-01
-2.08175167e-01 6.36726916e-01 1.26282513e-01 -1.13181926e-01
2.70728737e-01 7.20667481e-01 -1.59318101e+00 -1.43457189e-01
9.98323619e-01 9.71181512e-01 1.84914589e-01 4.47129875e-01
-1.03913449e-01 4.55179453e-01 -4.42112178e-01 -2.24798828e-01
8.50091577e-02 -7.46313572e-01 -5.17215848e-01 -6.63966462e-02
-3.78019363e-01 -2.60737427e-02 2.55296007e-02 5.83124161e-01
6.10343099e-01 -9.34458002e-02 6.46581888e-01 -6.73818290e-01
-1.42904818e-02 3.99967609e-03 -2.15221524e-01 2.64960736e-01
1.08941175e-01 -1.88643754e-01 5.58039188e-01 -5.28804064e-01
2.64750242e-01 9.53513324e-01 4.55267847e-01 3.55162442e-01
-9.25188184e-01 -4.61710215e-01 -5.28379381e-01 5.16929448e-01
-1.22123981e+00 1.99245177e-02 7.83720136e-01 -5.93119740e-01
2.77587801e-01 5.60394287e-01 8.25227976e-01 3.82612437e-01
3.74527752e-01 8.98368508e-02 1.00716400e+00 -5.74596763e-01
1.35164149e-02 4.34209704e-01 2.53055602e-01 6.49786115e-01
5.02126873e-01 6.16935566e-02 1.10910788e-01 -2.22776860e-01
6.75746918e-01 5.78356802e-01 -2.44223922e-01 -2.18679681e-01
-1.29188716e+00 9.46145654e-01 2.96967298e-01 9.26433206e-01
-3.95350009e-01 -5.45688808e-01 4.38614696e-01 8.40253606e-02
2.97996134e-01 5.15955806e-01 -6.30920291e-01 -2.40395233e-01
-5.15677512e-01 -9.67034837e-04 7.71068454e-01 -4.77752499e-02
2.95105278e-01 -2.63402969e-01 4.83788662e-02 8.32038224e-01
1.03391387e-01 5.22567928e-01 5.93169928e-01 -4.74575073e-01
1.03892803e-01 1.28777027e+00 -1.51308894e-01 -1.40970504e+00
-6.29436135e-01 -8.79001975e-01 -1.50467944e+00 1.14063755e-01
6.20631099e-01 -1.05162293e-01 -5.34759045e-01 1.35063779e+00
4.29195136e-01 -2.42035035e-02 3.37242670e-02 7.50623703e-01
5.56519687e-01 5.14199555e-01 2.73963243e-01 -4.41119134e-01
1.54489911e+00 -2.24595070e-01 -7.07015276e-01 1.18213177e-01
6.53610229e-01 -7.45458186e-01 7.05759108e-01 5.05334973e-01
-3.89527470e-01 -5.87228775e-01 -7.37138927e-01 2.82707274e-01
-3.82003307e-01 3.74546081e-01 7.53301799e-01 6.91038787e-01
-4.15122241e-01 7.53197253e-01 -6.00441277e-01 -7.08114028e-01
3.06646973e-01 1.82212248e-01 -4.94468391e-01 -1.55225068e-01
-1.31040728e+00 1.12305105e+00 3.80258590e-01 -2.34171338e-02
-1.69571266e-01 -6.03160203e-01 -3.45134884e-01 5.42303845e-02
2.08384901e-01 -6.63911998e-01 4.08350110e-01 -6.42370462e-01
-8.79681647e-01 7.87770867e-01 6.17462071e-03 -1.81273088e-01
3.91411245e-01 -1.54797778e-01 -4.44794565e-01 9.88188088e-02
-2.76753083e-02 -3.13289575e-02 4.54658180e-01 -5.98393440e-01
-6.49899662e-01 -6.39027357e-01 -5.00306785e-01 -5.27158976e-02
-5.37272334e-01 5.17514423e-02 -2.29467284e-02 -5.40484428e-01
2.15347290e-01 -1.04074168e+00 -3.16902459e-01 -1.65351816e-02
-2.47405186e-01 -2.96089560e-01 8.31089735e-01 -8.65826428e-01
1.39102566e+00 -2.14801645e+00 3.46228510e-01 2.37687290e-01
3.73445928e-01 6.50628030e-01 6.16564929e-01 4.40670252e-01
-7.57443979e-02 2.88797468e-01 -8.16269219e-02 2.88689107e-01
-5.81493616e-01 2.80897260e-01 1.77307650e-01 1.72866732e-01
1.24009021e-01 3.55928719e-01 -6.67707920e-01 -5.87631881e-01
5.28251708e-01 6.76010787e-01 -3.04491580e-01 3.61295730e-01
1.19919270e-01 8.48814964e-01 -7.59552836e-01 3.97778332e-01
1.92635104e-01 -3.93844128e-01 1.66706935e-01 -1.47734553e-01
4.60779993e-03 -3.16243231e-01 -9.31225836e-01 1.23757327e+00
-9.28752199e-02 2.61460692e-01 -6.09054029e-01 -1.30003262e+00
1.18909597e+00 7.40376711e-01 7.82361925e-01 -6.31716132e-01
2.29657739e-01 3.31458181e-01 4.37141120e-01 -9.14937496e-01
-4.45790291e-01 -2.19989717e-01 9.52021033e-02 1.48368731e-01
-5.83974004e-01 2.77332336e-01 3.40137005e-01 -1.64589509e-01
7.72757530e-01 -4.72657233e-01 7.84206986e-01 -1.84927300e-01
8.65701795e-01 -3.05358414e-03 8.71399522e-01 2.78192222e-01
-9.23788771e-02 3.51026356e-01 2.90030241e-01 -1.00842142e+00
-8.12460959e-01 -6.93859875e-01 -5.40743232e-01 1.72244325e-01
-8.08404982e-02 -1.38778865e-01 -3.25369090e-01 -4.44678217e-01
-6.31163316e-03 2.61695594e-01 -4.44995224e-01 -1.05589099e-01
-3.73374581e-01 -1.16635633e+00 2.65237719e-01 1.63012549e-01
7.64672339e-01 -9.15852547e-01 -9.45518494e-01 2.56900370e-01
-1.09305680e-01 -6.60658419e-01 2.31543943e-01 -1.44966528e-01
-1.26474023e+00 -1.35039449e+00 -8.11965406e-01 -6.21484220e-01
5.50303698e-01 -4.14490663e-02 9.18621182e-01 6.69336379e-01
-9.72584486e-01 -2.07254484e-01 -5.55651009e-01 -6.33695424e-01
-6.31938219e-01 2.77697265e-01 -5.23581505e-02 1.01545490e-01
3.26925844e-01 -5.83792746e-01 -8.10746729e-01 2.57090837e-01
-5.57662606e-01 1.58485174e-01 9.32182550e-01 5.80079973e-01
2.64071465e-01 5.34122527e-01 8.85516465e-01 -8.35766852e-01
4.06052381e-01 -6.12043977e-01 -2.74442822e-01 1.65470928e-01
-8.80337536e-01 6.26767725e-02 6.04896188e-01 -1.61508113e-01
-6.18162572e-01 -1.20910645e-01 -9.71705317e-02 3.93648222e-02
-4.12096560e-01 5.22729218e-01 -9.39084217e-02 2.24816874e-01
5.49014807e-01 1.70474395e-01 3.11906874e-01 -7.73262680e-01
-4.35115099e-01 6.53914928e-01 -1.26917601e-01 -2.00793236e-01
5.99981070e-01 8.43467116e-02 7.44097054e-01 -1.07655466e+00
-8.25678706e-01 -5.64171612e-01 -5.86173236e-01 -3.13445300e-01
1.14203739e+00 -3.91014338e-01 -6.42641246e-01 4.39520061e-01
-7.89120376e-01 5.37476182e-01 1.69164203e-02 7.89568067e-01
1.60191342e-01 4.07032818e-01 -1.75698370e-01 -7.37281740e-01
-5.49739420e-01 -1.05482328e+00 3.62432688e-01 3.29724461e-01
-3.46427649e-01 -1.10444379e+00 2.23927692e-01 3.21708262e-01
6.63565159e-01 6.22616947e-01 1.54866803e+00 -6.86306775e-01
-4.41587478e-01 -4.11579251e-01 -1.75339788e-01 3.39407384e-01
6.90600514e-01 -4.49976213e-02 -5.79170108e-01 -6.24622107e-02
1.19831435e-01 1.58514276e-01 5.01102507e-01 3.55388850e-01
9.51843739e-01 -8.88725296e-02 -3.99342149e-01 2.91495532e-01
1.39313138e+00 7.20737457e-01 6.07883096e-01 3.15078795e-01
5.19053936e-01 5.49326301e-01 7.16704071e-01 4.73880231e-01
1.80691421e-01 5.72231770e-01 4.65353996e-01 -3.56373638e-01
1.54443100e-01 1.72936052e-01 -3.86956781e-01 8.20943236e-01
-4.36778009e-01 1.52881384e-01 -1.02270675e+00 3.26034546e-01
-1.60470307e+00 -1.10706151e+00 -4.18277472e-01 2.44984055e+00
7.03119874e-01 -8.46003741e-02 1.11110859e-01 9.64434803e-01
5.66010773e-01 -2.84282058e-01 -2.85832345e-01 -2.39889383e-01
2.34438300e-01 2.11976871e-01 -6.32852241e-02 3.07317972e-02
-1.01043224e+00 6.38695806e-02 5.71134853e+00 1.82322994e-01
-1.34827590e+00 -3.81009042e-01 9.04843450e-01 3.68481845e-01
3.12594801e-01 -1.42212346e-01 -5.31490207e-01 7.50292003e-01
7.68605828e-01 -8.23853686e-02 8.35243464e-02 7.67034292e-01
5.33889532e-01 -2.24235952e-01 -7.68132687e-01 9.86897230e-01
-1.81311652e-01 -9.69054222e-01 2.22774944e-03 1.40270710e-01
3.14906687e-01 -5.80071032e-01 -2.28362426e-01 -3.71098556e-02
-5.08108854e-01 -9.18506801e-01 -1.92223877e-01 7.71687746e-01
6.15597785e-01 -7.59974062e-01 1.05012167e+00 5.98812878e-01
-8.00820053e-01 -2.75737196e-01 -7.15790763e-02 -3.81330490e-01
1.31803885e-01 1.18705571e+00 -9.21545148e-01 5.93972564e-01
7.00708985e-01 7.05240965e-01 -5.61254740e-01 1.28395498e+00
3.85198109e-02 6.24734282e-01 -1.96452603e-01 -1.18806124e-01
-4.76698399e-01 -3.82309914e-01 4.35204864e-01 6.45667076e-01
5.63990712e-01 1.92947552e-01 2.33651131e-01 3.29439163e-01
4.00358081e-01 6.24532342e-01 -5.65805674e-01 1.53865886e-03
1.49174675e-01 1.26829576e+00 -8.09631407e-01 -2.99525768e-01
-3.01427245e-01 5.63427031e-01 -1.74499542e-01 -2.71128207e-01
-6.33404076e-01 -3.29502404e-01 4.26309079e-01 5.38365781e-01
-2.76376843e-01 5.61724678e-02 -2.36554742e-01 -9.75497246e-01
-2.66856849e-01 -7.38987148e-01 6.17484152e-01 -1.64329469e-01
-1.10295761e+00 5.13565063e-01 -1.48349300e-01 -1.27024829e+00
-2.45556027e-01 -3.07673007e-01 -4.59612548e-01 1.01692140e+00
-1.43646896e+00 -7.08205104e-01 -4.98282403e-01 4.21320975e-01
3.83274764e-01 -1.20429546e-01 1.33178735e+00 6.33295298e-01
-8.24965537e-01 -1.90047994e-02 -4.43839803e-02 2.69681334e-01
6.67803526e-01 -1.06551504e+00 -4.42512512e-01 4.99255925e-01
-3.05813968e-01 4.89496142e-01 6.15124404e-01 -5.97540736e-01
-9.71289933e-01 -6.69786036e-01 1.22159171e+00 -3.32631804e-02
7.59990811e-02 3.12795877e-01 -8.43764365e-01 4.65530269e-02
-2.66669124e-01 -4.16264981e-02 9.44924831e-01 1.06618404e-02
2.50080764e-01 -4.87086087e-01 -1.18760717e+00 1.73012733e-01
3.98135215e-01 -5.98594993e-02 -6.20021999e-01 1.64956450e-01
7.50560313e-02 1.07447401e-01 -1.19488740e+00 7.35836804e-01
6.71920538e-01 -1.12099814e+00 9.83870208e-01 -5.28198957e-01
2.93564081e-01 -2.83624053e-01 3.07124019e-01 -1.06552660e+00
-3.70828778e-01 -2.98818443e-02 2.14499548e-01 1.11400926e+00
2.52282143e-01 -8.39413404e-01 6.48279488e-01 7.35809743e-01
4.40856516e-01 -1.00347602e+00 -7.70794451e-01 -3.33538294e-01
-2.87283152e-01 1.79558262e-01 2.04999983e-01 1.01316249e+00
-1.03840671e-01 3.45890999e-01 -2.84696579e-01 -9.51952040e-02
7.24100471e-01 2.43399799e-01 6.26790643e-01 -2.12147021e+00
-2.17359945e-01 -2.15371057e-01 -8.09568703e-01 -4.63928096e-02
-5.38676620e-01 -6.95188880e-01 -7.36730337e-01 -1.70936012e+00
4.33337152e-01 -8.54427993e-01 -4.00164574e-01 4.04321492e-01
-4.92608577e-01 5.18316329e-02 8.60095490e-03 3.83274049e-01
1.44754782e-01 6.61289543e-02 1.40765619e+00 1.32869273e-01
-2.46999666e-01 4.55707699e-01 -5.65124810e-01 6.19830310e-01
1.09216988e+00 -5.79323471e-01 -5.83955109e-01 1.00870039e-02
1.86597124e-01 4.02921498e-01 1.47970274e-01 -1.18544424e+00
-1.93260550e-01 -2.00428218e-01 6.02923512e-01 -3.18718791e-01
1.39772087e-01 -1.19321334e+00 5.89163065e-01 1.07909536e+00
-1.31317154e-02 4.60379981e-02 -1.02616183e-01 2.38516942e-01
-2.83557117e-01 -6.76933676e-03 8.12434196e-01 -1.53843686e-01
-5.31355560e-01 1.52462408e-01 -1.58262163e-01 -9.48342457e-02
1.45123339e+00 -6.77730292e-02 1.30401952e-02 -1.57546669e-01
-8.07742894e-01 -1.28428042e-01 1.98165715e-01 3.64560962e-01
4.55378413e-01 -8.75823617e-01 -7.63797641e-01 1.78914830e-01
2.54060607e-02 -2.90850941e-02 3.70086372e-01 1.07317531e+00
-8.45977664e-01 5.73565662e-01 -3.50679696e-01 -6.18870735e-01
-1.65525353e+00 5.93244970e-01 2.09724948e-01 -5.36302030e-01
-7.67213762e-01 1.08999401e-01 -1.39315307e-01 2.36456633e-01
8.90403464e-02 -1.95677489e-01 -7.80323982e-01 7.17894956e-02
6.50813341e-01 6.99496686e-01 -1.52833918e-02 -4.82489705e-01
-3.18928957e-01 9.84209836e-01 -2.06746534e-02 4.34370130e-01
1.44256151e+00 1.32712439e-01 -3.16593349e-01 4.50027376e-01
7.86856711e-01 1.06178178e-02 -4.47417855e-01 9.60344151e-02
5.30820228e-02 -4.67929989e-01 6.81478158e-02 -1.08233082e+00
-1.16005373e+00 1.06806612e+00 9.47199106e-01 4.74923939e-01
1.29962373e+00 -2.55252123e-01 6.39293790e-01 1.24111786e-01
3.48138690e-01 -5.21492422e-01 -1.65541232e-01 -9.68405902e-02
5.61361790e-01 -1.39821434e+00 8.34613368e-02 -5.66975951e-01
-4.15367246e-01 1.26184332e+00 8.68110061e-02 1.33774713e-01
9.19191301e-01 -7.09140301e-02 2.58150220e-01 -6.49238825e-02
-3.11414659e-01 -6.04589768e-02 1.55584335e-01 3.11274916e-01
4.93256778e-01 9.68453959e-02 -8.58939171e-01 4.32862610e-01
2.85072416e-01 3.28410685e-01 4.42431755e-02 7.38055170e-01
-6.31286085e-01 -1.56469333e+00 -4.22223330e-01 7.52869606e-01
-8.70431483e-01 2.36640602e-01 -1.53367311e-01 7.63005197e-01
2.60336906e-01 9.16459918e-01 -4.34208244e-01 -2.39231139e-01
2.80263990e-01 3.12296867e-01 1.54830933e-01 -4.12404120e-01
-4.49292421e-01 -3.30953628e-01 -1.84289306e-01 -2.35155329e-01
-5.44061303e-01 -4.35718775e-01 -1.28896904e+00 -2.30125561e-01
-1.88626379e-01 4.29208547e-01 8.32071543e-01 9.31199849e-01
5.35394073e-01 5.93588531e-01 8.51578057e-01 -3.28921944e-01
-5.47012925e-01 -9.44981694e-01 -5.99873185e-01 8.06081772e-01
4.12684567e-02 -7.21036017e-01 -4.35356021e-01 1.82774942e-02] | [8.462912559509277, 4.854751110076904] |
3538b81c-ad40-4fbd-9222-3660a268cc0c | topic-aware-neural-keyphrase-generation-for | 1906.03889 | null | https://arxiv.org/abs/1906.03889v1 | https://arxiv.org/pdf/1906.03889v1.pdf | Topic-Aware Neural Keyphrase Generation for Social Media Language | A huge volume of user-generated content is daily produced on social media. To facilitate automatic language understanding, we study keyphrase prediction, distilling salient information from massive posts. While most existing methods extract words from source posts to form keyphrases, we propose a sequence-to-sequence (seq2seq) based neural keyphrase generation framework, enabling absent keyphrases to be created. Moreover, our model, being topic-aware, allows joint modeling of corpus-level latent topic representations, which helps alleviate the data sparsity that widely exhibited in social media language. Experiments on three datasets collected from English and Chinese social media platforms show that our model significantly outperforms both extraction and generation models that do not exploit latent topics. Further discussions show that our model learns meaningful topics, which interprets its superiority in social media keyphrase generation. | ['Hou Pong Chan', 'Michael R. Lyu', 'Jing Li', 'Yue Wang', 'Irwin King', 'Shuming Shi'] | 2019-06-10 | topic-aware-neural-keyphrase-generation-for-1 | https://aclanthology.org/P19-1240 | https://aclanthology.org/P19-1240.pdf | acl-2019-7 | ['keyphrase-generation'] | ['natural-language-processing'] | [ 1.66675195e-01 3.57077837e-01 -5.60063303e-01 3.49016011e-01
-9.97025073e-01 -5.99503517e-01 1.06866002e+00 4.96343523e-01
-4.53400612e-01 1.07060528e+00 1.23248982e+00 -2.88616151e-01
3.11779857e-01 -9.89590704e-01 -6.96093142e-01 -2.52080441e-01
-7.27858245e-02 6.89645335e-02 7.81306848e-02 -4.25832152e-01
4.94474322e-01 -2.99441069e-01 -1.35444820e+00 4.26948994e-01
1.06397378e+00 5.23134410e-01 7.31088042e-01 6.59064710e-01
-7.58498967e-01 1.02354431e+00 -6.24501288e-01 -2.30952963e-01
1.14232257e-01 -2.64968127e-01 -6.95441186e-01 2.69759409e-02
3.00661206e-01 -5.70730984e-01 -4.24770087e-01 9.02648330e-01
4.07414794e-01 1.10996529e-01 5.60944021e-01 -1.09334958e+00
-9.36507463e-01 1.50123060e+00 -5.58593452e-01 4.65564728e-02
4.43368345e-01 -2.89236635e-01 1.65134084e+00 -9.76332426e-01
7.39860117e-01 1.25140631e+00 2.82768041e-01 2.95348287e-01
-9.02650118e-01 -9.31265891e-01 3.44402909e-01 -2.85332888e-01
-1.21641576e+00 -1.88767631e-02 9.21802163e-01 -2.82181442e-01
7.98585951e-01 9.16042104e-02 8.79410326e-01 1.40636408e+00
4.03323442e-01 1.57994282e+00 8.17736089e-01 -2.59271502e-01
5.33716641e-02 9.37030613e-02 2.06617936e-02 4.29837018e-01
3.10270876e-01 -5.75066209e-01 -1.13742435e+00 -7.60546267e-01
6.89098895e-01 2.68661618e-01 -8.72104466e-02 1.98611483e-01
-1.76628840e+00 1.16736031e+00 -9.51771066e-02 2.22295582e-01
-8.96934867e-01 1.31481886e-01 4.75541592e-01 2.26048008e-01
1.01824772e+00 8.63109946e-01 -6.17043257e-01 -2.13159174e-01
-1.09605801e+00 7.85127640e-01 9.52494204e-01 1.07503843e+00
8.62242460e-01 -7.66835585e-02 -5.51709771e-01 6.71151400e-01
1.26464620e-01 5.23359120e-01 9.74020422e-01 -3.41984838e-01
5.23679495e-01 7.07059085e-01 2.45354086e-01 -1.27555084e+00
-7.19350800e-02 -1.97451934e-01 -8.23044002e-01 -1.13729799e+00
-9.84208360e-02 -6.79459929e-01 -7.14539886e-01 1.39992082e+00
7.03194961e-02 3.74874026e-01 3.72960158e-02 3.15743923e-01
8.71011734e-01 1.21269798e+00 2.33399048e-01 -3.89748871e-01
1.43032467e+00 -8.97769749e-01 -1.06499076e+00 -1.62961498e-01
5.00180066e-01 -8.84160280e-01 1.11483932e+00 4.26825553e-01
-9.63264108e-01 -3.36817056e-01 -6.87500417e-01 -2.75065958e-01
-5.01327872e-01 1.22158602e-01 9.15784121e-01 1.13235720e-01
-8.46744776e-01 2.14611337e-01 -5.13014555e-01 -1.26277223e-01
3.70185256e-01 -2.05249608e-01 7.95224980e-02 4.34532195e-01
-1.74029851e+00 6.14553690e-02 7.59135664e-01 -5.73202908e-01
-8.99730444e-01 -1.28766525e+00 -7.28352308e-01 4.86180931e-02
9.12831843e-01 -7.45832443e-01 1.37440336e+00 -5.45105219e-01
-1.34667087e+00 4.72986311e-01 -5.09159982e-01 -7.06076682e-01
1.61147013e-01 -8.42757285e-01 -3.32624078e-01 3.22915912e-01
4.63059545e-01 9.63010192e-01 1.19465709e+00 -9.28377569e-01
-9.37844992e-01 4.62304711e-01 9.08695161e-02 1.33581519e-01
-1.02470958e+00 1.81057557e-01 -3.46031755e-01 -1.36998868e+00
-1.77383170e-01 -6.55584753e-01 -5.05185843e-01 -5.36137342e-01
-9.05775964e-01 -6.55100822e-01 7.05876946e-01 -8.30298424e-01
1.69486892e+00 -1.65782130e+00 -6.65092245e-02 3.28662135e-02
5.71687341e-01 -2.15541899e-01 -1.85974553e-01 1.21568429e+00
3.28732371e-01 4.97890979e-01 8.44229087e-02 -1.70046955e-01
1.44204706e-01 -3.07554275e-01 -1.37284267e+00 -3.57802182e-01
1.21284137e-02 1.20507765e+00 -1.23473215e+00 -5.48498690e-01
-3.76546025e-01 1.70801044e-01 -6.47922575e-01 3.21877867e-01
-8.24705184e-01 1.15530789e-02 -1.08322108e+00 2.42488801e-01
2.13811442e-01 -6.12729788e-01 1.23538198e-02 2.75797918e-02
-1.47458032e-01 7.43911386e-01 -8.37911069e-01 1.88552165e+00
-5.95043719e-01 4.27599549e-01 -3.45291018e-01 -2.83636063e-01
9.28097963e-01 4.74485099e-01 8.02923381e-01 -3.07232082e-01
-1.45196721e-01 -1.80890650e-01 -6.66670263e-01 -2.62583519e-04
1.56398773e+00 2.56227762e-01 -4.19075012e-01 1.12214947e+00
-1.37410527e-02 -2.86894649e-01 5.10469794e-01 1.24029851e+00
7.43712842e-01 -3.10333937e-01 4.81224358e-01 -4.26722169e-01
1.44991830e-01 -5.08949123e-02 3.51246111e-02 1.07504690e+00
4.04550105e-01 3.72947991e-01 7.86601961e-01 -2.53393859e-01
-1.03243458e+00 -7.73325205e-01 3.61595809e-01 1.42659187e+00
2.74498332e-02 -1.25519383e+00 -4.55930442e-01 -6.73877239e-01
1.29106462e-01 6.90692246e-01 -5.84396064e-01 4.36199158e-02
-3.65941316e-01 -7.89282203e-01 3.45787674e-01 1.76450357e-01
9.66333225e-02 -1.06172299e+00 -3.30501236e-02 5.80058515e-01
-6.02866232e-01 -1.11222637e+00 -9.05600488e-01 -4.45735723e-01
-5.44108927e-01 -6.24821126e-01 -1.22158659e+00 -5.27142167e-01
5.66706777e-01 6.57139182e-01 1.30995011e+00 -1.76317036e-01
5.07281721e-03 5.37993848e-01 -6.60028219e-01 -8.36370111e-01
-3.12196642e-01 7.05657601e-01 1.87478825e-01 6.39410242e-02
4.17765111e-01 -6.08497560e-01 -5.45314074e-01 -3.85547251e-01
-1.21435857e+00 5.36803007e-01 5.04072130e-01 7.03392446e-01
2.66169101e-01 8.44005495e-02 8.48597467e-01 -1.07748652e+00
1.44413233e+00 -9.28326249e-01 -3.28321427e-01 3.41688953e-02
-7.61589408e-01 1.35673910e-01 5.31874120e-01 -7.18511522e-01
-1.09646523e+00 -5.09653091e-01 2.34304607e-01 1.35842919e-01
1.87460557e-01 8.15948486e-01 3.70303839e-01 8.10742974e-01
4.18173581e-01 8.06694746e-01 -3.75700682e-01 -5.36962867e-01
7.74223864e-01 5.88630617e-01 8.19509383e-03 -6.70277238e-01
9.32378054e-01 3.88226599e-01 -6.19265616e-01 -1.12441206e+00
-1.16134131e+00 -8.48868310e-01 -1.90188497e-01 8.24358761e-02
6.32883310e-01 -1.49012113e+00 -4.09089863e-01 4.34779942e-01
-1.29270434e+00 -5.65038435e-02 -3.20021719e-01 3.69618922e-01
-1.82591140e-01 4.56167817e-01 -7.97703922e-01 -6.98298395e-01
-9.97059405e-01 -5.55411398e-01 1.41063535e+00 1.19342737e-01
-5.63993037e-01 -1.04962349e+00 1.54470667e-01 1.05025381e-01
3.75248998e-01 -1.27938196e-01 7.74069369e-01 -9.55371499e-01
-8.19873035e-01 -1.46649837e-01 -2.59623885e-01 -4.99116890e-02
4.22600567e-01 -2.17899650e-01 -6.12030923e-01 -2.14286268e-01
-5.33134162e-01 -5.98718643e-01 1.18770647e+00 1.81955159e-01
1.24576271e+00 -9.93901014e-01 -2.90216237e-01 -1.03097849e-01
8.49535882e-01 -3.37188452e-01 2.86977023e-01 2.61779994e-01
8.33388567e-01 5.65507889e-01 3.94493371e-01 9.03676629e-01
8.58512163e-01 1.19444914e-01 -1.99450061e-01 -8.55183229e-02
3.72447193e-01 -1.08370996e+00 6.18128061e-01 1.34177792e+00
2.70410091e-01 -4.24224854e-01 -6.35416508e-01 8.04602683e-01
-1.70049870e+00 -8.96438062e-01 1.50486767e-01 1.56503487e+00
1.46391881e+00 1.70383453e-01 2.22623646e-01 -2.20981494e-01
4.39328790e-01 7.72869647e-01 -1.98323593e-01 1.21789455e-01
-1.49125591e-01 2.23634526e-01 4.13287371e-01 3.78324151e-01
-9.97185886e-01 1.38894463e+00 5.98880434e+00 1.12871242e+00
-1.00484180e+00 -4.80437018e-02 5.63266933e-01 2.84622442e-02
-9.90113199e-01 -2.01394618e-01 -1.27102280e+00 5.68916738e-01
7.00588226e-01 -8.67420256e-01 -8.83047581e-02 1.02384830e+00
4.39553559e-01 -4.37378883e-02 -6.60138190e-01 6.93566442e-01
1.86160028e-01 -1.69170415e+00 8.16522956e-01 1.23095468e-01
1.25237513e+00 -1.09227009e-01 2.32197810e-02 3.55521560e-01
8.11849177e-01 -4.81985360e-01 5.44284642e-01 5.35087407e-01
3.28347653e-01 -7.26429403e-01 2.80979276e-01 5.20847261e-01
-1.16447568e+00 2.03303769e-01 -3.71418566e-01 -4.67306264e-02
4.26100135e-01 1.04543614e+00 -1.35152316e+00 3.55617523e-01
2.94452548e-01 7.60436118e-01 -4.22588706e-01 4.82706845e-01
-4.37486738e-01 1.05533600e+00 -1.59342721e-01 -5.67112505e-01
5.63013077e-01 2.58194264e-02 5.81340551e-01 1.40907431e+00
3.49388391e-01 -7.22378716e-02 6.41032755e-01 8.47809553e-01
-4.08588618e-01 6.48432076e-01 -4.64837760e-01 -8.69731486e-01
5.91652095e-01 1.45611179e+00 -8.45105112e-01 -7.89049923e-01
-2.91307569e-01 9.77154136e-01 1.02794394e-02 4.08524007e-01
-2.28687897e-01 -3.73669803e-01 5.85563064e-01 8.20496529e-02
7.55316485e-03 -5.00264108e-01 1.72756717e-01 -1.48943913e+00
-1.56103000e-01 -1.01970291e+00 2.46826515e-01 -6.59619927e-01
-1.38510120e+00 3.53378057e-01 2.49185845e-01 -8.57473195e-01
-6.00269318e-01 -6.53926060e-02 -6.57668829e-01 6.92913771e-01
-1.51768148e+00 -1.42411053e+00 2.17294738e-01 3.92764360e-01
1.03796244e+00 -2.77860910e-01 5.55370450e-01 -1.44644603e-01
-3.19590531e-02 3.86545867e-01 -1.15526006e-01 2.16825858e-01
7.97193706e-01 -1.40739691e+00 1.07615280e+00 7.76965201e-01
4.27791208e-01 1.21536493e+00 5.73952556e-01 -1.28842819e+00
-1.50067568e+00 -1.12259817e+00 1.34189975e+00 -3.76925290e-01
1.26112819e+00 -6.01741195e-01 -6.95131600e-01 4.48451489e-01
5.67129552e-01 -7.19079077e-01 9.27727699e-01 2.16711223e-01
-3.30388278e-01 3.81250113e-01 -1.78661481e-01 1.15446508e+00
7.51206577e-01 -8.04906428e-01 -9.18693721e-01 6.22823775e-01
1.55603290e+00 -5.13040088e-02 -5.66966295e-01 6.37856200e-02
5.12620151e-01 -1.25713497e-01 9.74286795e-01 -7.68816233e-01
8.59806597e-01 9.45348963e-02 2.12987378e-01 -1.46175456e+00
-3.22593078e-02 -1.43216896e+00 -6.90039039e-01 1.25266504e+00
5.63143492e-01 -4.24213856e-01 7.72189200e-01 2.07662761e-01
3.09539050e-01 -5.41839182e-01 -2.23881498e-01 -3.31128418e-01
-4.58251219e-03 -5.16888618e-01 6.79308951e-01 1.10128093e+00
2.61382669e-01 6.68593049e-01 -9.30957258e-01 -2.94695467e-01
2.62005836e-01 1.45482048e-01 1.03898728e+00 -1.08845782e+00
-1.36716902e-01 -2.68461794e-01 4.24077600e-01 -1.67935646e+00
1.05448082e-01 -7.55738497e-01 -2.64385670e-01 -1.53469777e+00
4.17182833e-01 -2.05284253e-01 -5.85408211e-02 3.82287204e-01
-7.79079616e-01 4.26100986e-03 1.76033184e-01 2.49064192e-01
-8.57821703e-01 8.13926339e-01 1.61261928e+00 -1.22683197e-01
-4.00716782e-01 -1.02918856e-01 -1.23741639e+00 6.09557569e-01
5.63909769e-01 -4.65528220e-01 -5.73766172e-01 -9.22399759e-03
1.10245895e+00 -8.97503719e-02 -4.36472557e-02 -3.46055835e-01
2.46226490e-01 -3.52839738e-01 -2.14608610e-02 -9.21662748e-01
1.40310332e-01 -1.96077153e-01 -5.18174171e-01 2.09214836e-01
-1.06781447e+00 -4.49551158e-02 4.85176183e-02 7.69563317e-01
-2.31866926e-01 9.32355896e-02 -1.63853899e-01 -4.14618939e-01
-6.91721141e-01 6.63744569e-01 -9.14111376e-01 2.48557135e-01
4.18368906e-01 3.78874093e-01 -4.04813364e-02 -9.27150190e-01
-2.27231488e-01 2.31785387e-01 -1.57637581e-01 8.65152955e-01
5.93068540e-01 -1.24097979e+00 -1.17177057e+00 4.77899574e-02
3.30477059e-01 9.09702107e-02 1.14143848e-01 3.07255715e-01
8.79859477e-02 7.74227381e-01 3.92874718e-01 -1.59978047e-02
-7.64695406e-01 4.03912157e-01 -7.37045288e-01 -6.44261777e-01
-5.93843997e-01 8.98950756e-01 4.82896894e-01 -3.60722899e-01
-5.78558594e-02 -6.38086498e-01 -3.71186465e-01 6.29014730e-01
1.08967483e+00 1.18955351e-01 -2.65817553e-01 -1.36878297e-01
5.36126852e-01 -1.68051183e-01 -6.70891047e-01 -2.58364588e-01
1.13051176e+00 -3.17026705e-01 -2.47192428e-01 5.11050940e-01
1.05203414e+00 6.08049154e-01 -8.87531281e-01 -7.33461559e-01
2.20615909e-01 2.48399302e-02 2.03883395e-01 -4.53722954e-01
-4.92167741e-01 5.35094321e-01 -3.32787871e-01 4.15649146e-01
7.62202263e-01 6.90987185e-02 1.65040624e+00 8.72058153e-01
2.25161374e-01 -1.27981651e+00 4.78332877e-01 7.95816064e-01
1.00662506e+00 -1.11171472e+00 2.12812856e-01 -2.79322505e-01
-7.69986928e-01 8.75808895e-01 3.66693020e-01 4.25564088e-02
7.95222223e-01 1.75274357e-01 -1.71015054e-01 2.09947000e-03
-1.04908001e+00 -1.12974554e-01 6.06101573e-01 4.22849059e-02
5.10105789e-01 -1.25303864e-02 -4.71837729e-01 8.72957945e-01
-6.77977622e-01 -7.99855739e-02 5.74652135e-01 8.15711081e-01
-7.27186024e-01 -1.07375526e+00 -9.51650664e-02 6.67001069e-01
-1.02131283e+00 -8.92219961e-01 -5.96907198e-01 2.68689185e-01
-3.32680464e-01 6.38213158e-01 3.32744233e-02 -2.93507278e-01
-5.44155836e-01 9.94831026e-02 -3.35528731e-01 -9.40065324e-01
-4.79253769e-01 3.73949736e-01 5.80476150e-02 -2.52102822e-01
-1.89663187e-01 -3.68095040e-01 -9.71768737e-01 -2.07319960e-01
-2.48832792e-01 7.48946428e-01 5.23751736e-01 8.92936766e-01
6.06910944e-01 4.77674782e-01 7.87259042e-01 -6.77718222e-01
-3.36157441e-01 -1.36399269e+00 -3.99891675e-01 1.51554242e-01
3.61746281e-01 -9.67643782e-03 -9.48157534e-02 4.73162204e-01] | [12.315166473388672, 8.893037796020508] |
5ec5d058-b682-462e-94c1-c988ece5d2f2 | in-n-out-generative-learning-for-dense | 2203.15312 | null | https://arxiv.org/abs/2203.15312v4 | https://arxiv.org/pdf/2203.15312v4.pdf | In-N-Out Generative Learning for Dense Unsupervised Video Segmentation | In this paper, we focus on unsupervised learning for Video Object Segmentation (VOS) which learns visual correspondence (i.e., the similarity between pixel-level features) from unlabeled videos. Previous methods are mainly based on the contrastive learning paradigm, which optimize either in image level or pixel level. Image-level optimization (e.g., the spatially pooled feature of ResNet) learns robust high-level semantics but is sub-optimal since the pixel-level features are optimized implicitly. By contrast, pixel-level optimization is more explicit, however, it is sensitive to the visual quality of training data and is not robust to object deformation. To complementarily perform these two levels of optimization in a unified framework, we propose the In-aNd-Out (INO) generative learning from a purely generative perspective with the help of naturally designed class tokens and patch tokens in Vision Transformer (ViT). Specifically, for image-level optimization, we force the out-view imagination from local to global views on class tokens, which helps capture high-level semantics, and we name it as out-generative learning. As to pixel-level optimization, we perform in-view masked image modeling on patch tokens, which recovers the corrupted parts of an image via inferring its fine-grained structure, and we term it as in-generative learning. To discover the temporal information better, we additionally force the inter-frame consistency from both feature and affinity matrix levels. Extensive experiments on DAVIS-2017 val and YouTube-VOS 2018 val show that our INO outperforms previous state-of-the-art methods by significant margins. Code is available: https://github.com/pansanity666/INO_VOS | ['Yi Yang', 'Jingren Zhou', 'Hongxia Yang', 'Chang Zhou', 'Huiling Zhou', 'Zongxin Yang', 'Peike Li', 'Xiao Pan'] | 2022-03-29 | null | null | null | null | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [ 6.48824275e-02 3.05003095e-02 -3.83303285e-01 -2.84475178e-01
-7.91073263e-01 -4.93294597e-01 4.74721432e-01 -4.92345482e-01
-1.84013382e-01 4.76283133e-01 2.27760419e-01 1.00618079e-01
1.28257975e-01 -8.18295598e-01 -1.27229369e+00 -9.24321771e-01
3.98493230e-01 1.12936180e-02 3.33690166e-01 -7.04186484e-02
-1.31837530e-02 8.19298699e-02 -1.40503645e+00 3.41649383e-01
9.75936413e-01 8.65878284e-01 3.76030385e-01 3.83351326e-01
1.08721465e-01 8.13930631e-01 -2.29600087e-01 -2.33334512e-01
4.51195300e-01 -5.96088111e-01 -8.50735426e-01 5.86329103e-01
7.43347883e-01 -3.76903057e-01 -4.25214976e-01 1.29207480e+00
1.88905060e-01 1.48518132e-02 5.36326408e-01 -1.24496603e+00
-9.03338015e-01 4.21424508e-01 -6.86331928e-01 9.38157663e-02
1.79096669e-01 5.89322984e-01 1.10897410e+00 -9.97625053e-01
9.01912451e-01 1.21090484e+00 2.27754891e-01 5.09451628e-01
-1.48195541e+00 -4.54088062e-01 4.77721334e-01 1.88191220e-01
-1.21072519e+00 -2.99431831e-01 1.17830253e+00 -7.08859801e-01
3.39523822e-01 5.78676611e-02 8.80217373e-01 1.15788615e+00
4.76176031e-02 1.17098701e+00 1.32099843e+00 -1.37996316e-01
1.30624205e-01 -9.11438614e-02 -5.42289689e-02 9.24552500e-01
-4.00222242e-02 4.26789969e-01 -6.21241987e-01 3.56730521e-01
1.19089997e+00 2.21666887e-01 -5.14882863e-01 -5.89917660e-01
-1.33563054e+00 8.02491605e-01 7.58796930e-01 2.71458000e-01
-2.64453351e-01 2.45542362e-01 -7.56594986e-02 1.94285691e-01
4.91345793e-01 2.06470698e-01 -5.10536730e-01 1.59558654e-01
-9.70481873e-01 1.77076072e-01 3.65376413e-01 8.41597974e-01
1.18870413e+00 4.41812098e-01 -3.74422014e-01 5.75931728e-01
4.54978108e-01 3.69028181e-01 3.38375866e-01 -1.18396330e+00
2.76113629e-01 5.54692149e-01 -9.81737077e-02 -8.37914467e-01
-3.40122469e-02 -3.87284636e-01 -8.03388476e-01 3.11517298e-01
4.05454546e-01 -6.41492605e-02 -1.37008679e+00 2.01064944e+00
4.32425231e-01 5.71034968e-01 -9.91179198e-02 1.21770823e+00
1.14549124e+00 5.83530247e-01 -1.44145086e-01 -2.75811315e-01
1.30917954e+00 -1.30309975e+00 -5.65659761e-01 -3.55578154e-01
2.01636210e-01 -5.28108180e-01 1.32768393e+00 1.43387364e-02
-1.25137341e+00 -7.67166615e-01 -9.15488958e-01 -1.68481603e-01
-1.85585901e-01 5.21931462e-02 3.18086833e-01 2.72208720e-01
-1.05669332e+00 5.30122578e-01 -8.93884838e-01 -1.09803282e-01
6.74660683e-01 7.15297386e-02 -3.06077659e-01 -8.81177112e-02
-1.05530107e+00 4.07130182e-01 2.51198888e-01 9.74197835e-02
-1.23269498e+00 -8.72102797e-01 -9.70062196e-01 -9.74283963e-02
6.63576663e-01 -9.35622275e-01 7.21932530e-01 -1.13078940e+00
-1.57946038e+00 9.95094121e-01 -2.58181900e-01 -1.64074466e-01
5.55422962e-01 -4.78774123e-02 8.37151557e-02 2.84840643e-01
2.49784216e-01 9.87797320e-01 1.33190906e+00 -1.58124459e+00
-3.34398985e-01 -3.18672597e-01 2.26202086e-01 1.82345152e-01
-1.67726398e-01 -3.41127276e-01 -8.23604226e-01 -9.08424079e-01
2.36394927e-01 -8.69681239e-01 -8.06513205e-02 2.65774038e-02
-5.05553007e-01 1.07488390e-02 8.36993575e-01 -7.75796592e-01
9.35154736e-01 -2.21437478e+00 4.71308708e-01 -6.65975884e-02
3.53964299e-01 1.33533299e-01 -1.81810960e-01 -8.72403011e-02
-1.30181238e-01 1.70991763e-01 -4.47591692e-01 -3.13925713e-01
-1.82983384e-01 3.34528387e-01 -2.25369930e-01 5.07027924e-01
4.81856227e-01 1.52673078e+00 -9.30605531e-01 -5.50226927e-01
3.78559917e-01 3.46096486e-01 -8.46819222e-01 2.79069483e-01
-4.09709126e-01 9.93675053e-01 -5.47366142e-01 9.16005433e-01
5.94475865e-01 -4.80328798e-01 -1.06489033e-01 -5.43436050e-01
6.31202385e-02 -2.16691755e-02 -8.78288269e-01 1.95350361e+00
-8.20695683e-02 3.59814584e-01 7.94253573e-02 -1.19558024e+00
5.01491666e-01 1.10643476e-01 6.14735305e-01 -7.14728951e-01
8.61452743e-02 9.18147992e-03 -1.07431471e-01 -5.58877051e-01
1.47653788e-01 4.81268428e-02 2.86317449e-02 1.17154986e-01
2.79754430e-01 -1.95621565e-01 1.39078826e-01 2.92755067e-01
7.09734917e-01 6.41034782e-01 4.07640636e-03 -2.74921685e-01
4.16181952e-01 -2.59018451e-01 7.97291338e-01 6.02360427e-01
-2.68088579e-01 9.68655229e-01 4.25090969e-01 -1.90401331e-01
-8.89972210e-01 -1.24246013e+00 -1.04057953e-01 8.05734456e-01
5.97036481e-01 -3.80538195e-01 -9.75946248e-01 -7.91572034e-01
-1.07064836e-01 3.06808591e-01 -7.56311178e-01 -1.74774364e-01
-6.01036727e-01 -5.48777997e-01 1.10091805e-01 4.97157186e-01
7.76123762e-01 -1.19711614e+00 -3.30673426e-01 -3.54041792e-02
-4.08852816e-01 -1.29583824e+00 -7.76289403e-01 -5.32459430e-02
-8.41427982e-01 -9.73673284e-01 -7.39300251e-01 -7.97676384e-01
7.64483631e-01 1.77885100e-01 1.19002509e+00 1.26765013e-01
-2.50591815e-01 4.42137033e-01 -3.85063291e-01 2.93008592e-02
-2.07628366e-02 -7.96259493e-02 -1.75931618e-01 3.55495483e-01
-1.61249042e-01 -7.86168337e-01 -7.69225180e-01 2.94127375e-01
-1.09938598e+00 2.77403295e-01 6.81365788e-01 1.16237533e+00
9.39157546e-01 -1.68281108e-01 2.61743337e-01 -8.76170039e-01
1.02924734e-01 -2.97031254e-01 -4.69131559e-01 1.92635551e-01
-5.56683660e-01 1.51498094e-02 3.65280181e-01 -4.84678239e-01
-8.66769493e-01 1.65068880e-01 -4.42737453e-02 -9.63034987e-01
-7.04199821e-02 3.25937241e-01 -4.82627213e-01 -1.19864576e-01
3.22804660e-01 5.03234386e-01 -9.07820761e-02 -2.43048906e-01
6.06149018e-01 3.52315381e-02 5.75821638e-01 -7.21331656e-01
1.05687058e+00 6.81060135e-01 -3.47500652e-01 -6.25548124e-01
-1.15025675e+00 -2.90342540e-01 -6.50641143e-01 -3.79769027e-01
1.30593610e+00 -9.99208868e-01 -5.20564675e-01 6.63124382e-01
-9.40695286e-01 -6.79232359e-01 -6.25126541e-01 3.36828023e-01
-8.03847373e-01 2.72515178e-01 -6.50746167e-01 -3.07344496e-01
4.96268682e-02 -1.40699458e+00 1.33106685e+00 3.45007628e-01
2.12430984e-01 -9.83041763e-01 -2.06530526e-01 6.80705309e-01
8.85944162e-03 3.51236880e-01 7.36431897e-01 -5.31711914e-02
-1.14969432e+00 4.14566129e-01 -3.39647263e-01 6.04768634e-01
1.99375488e-02 -8.46102014e-02 -9.81211364e-01 -3.01557988e-01
2.07892746e-01 -3.56287420e-01 1.11259198e+00 7.89981246e-01
1.35338354e+00 -3.98912698e-01 -8.59363656e-03 1.11566246e+00
1.36382425e+00 -1.68682560e-01 8.54720414e-01 2.13520512e-01
1.28678393e+00 5.28963506e-01 5.98932564e-01 1.26313895e-01
4.35404837e-01 7.47232258e-01 5.80886781e-01 -3.56996328e-01
-5.04160821e-01 -4.62201864e-01 5.37600279e-01 7.16261625e-01
-2.47719690e-01 -1.26434818e-01 -4.12064850e-01 4.97031748e-01
-2.05433893e+00 -9.82006669e-01 9.76591259e-02 1.96064150e+00
9.56678331e-01 -3.65519635e-02 1.32958800e-01 -2.63834536e-01
6.18334532e-01 5.55447519e-01 -7.61594236e-01 2.53687739e-01
-3.09474647e-01 -3.23219672e-02 3.79447609e-01 5.72309613e-01
-1.03538704e+00 1.20101500e+00 5.26785707e+00 1.07448947e+00
-1.17167413e+00 2.88543254e-01 9.59005117e-01 -6.73318058e-02
-5.92893004e-01 1.93390712e-01 -5.15881002e-01 7.31414199e-01
7.44702145e-02 2.43203565e-01 5.03001511e-01 6.96938217e-01
1.63624510e-01 5.66643737e-02 -1.07330585e+00 1.17705309e+00
1.55166894e-01 -1.59530509e+00 8.10354948e-02 3.09269637e-01
1.19356215e+00 -3.19586252e-04 3.03673893e-01 1.41029447e-01
1.15309030e-01 -9.78473067e-01 1.09375846e+00 6.85556173e-01
6.98460102e-01 -3.93219680e-01 3.99726242e-01 2.20911220e-01
-1.20863438e+00 9.46679488e-02 -5.07649742e-02 2.94105500e-01
2.88755327e-01 5.31000078e-01 -3.85247618e-02 5.43828845e-01
9.64186728e-01 1.17462623e+00 -5.64133704e-01 5.33557951e-01
-4.90481287e-01 6.72486365e-01 -6.84796199e-02 4.93801385e-01
2.40943536e-01 -5.88661611e-01 6.58126593e-01 7.05895662e-01
3.68723609e-02 1.17558636e-01 4.64042515e-01 1.41965771e+00
-2.71509737e-01 -2.63493657e-01 -2.49224693e-01 9.39439703e-03
5.65936379e-02 1.27440131e+00 -6.42355144e-01 -3.98213029e-01
-3.12529385e-01 1.10843992e+00 2.42935076e-01 7.43719280e-01
-9.30148482e-01 2.05964819e-01 5.74446559e-01 3.20451975e-01
6.19699597e-01 -4.30718698e-02 -2.78685451e-01 -1.69664145e+00
2.37620369e-01 -8.37115586e-01 1.61787063e-01 -8.28624845e-01
-1.29731524e+00 3.24364722e-01 7.64629385e-03 -1.34618700e+00
-2.16714352e-01 -4.49711263e-01 -6.90966785e-01 5.91467559e-01
-1.62703133e+00 -1.46122587e+00 -4.01019096e-01 9.37258482e-01
6.96618855e-01 4.56838310e-02 1.37647316e-01 1.92735732e-01
-5.29881299e-01 5.40414691e-01 -4.59634587e-02 3.62042397e-01
6.05365098e-01 -1.12407458e+00 1.41110927e-01 1.05036128e+00
5.42549908e-01 4.42140043e-01 4.65689778e-01 -5.80568850e-01
-1.41684675e+00 -1.17511213e+00 4.07248259e-01 -4.84910399e-01
6.19118929e-01 -4.12950635e-01 -9.92553532e-01 7.11344242e-01
6.73986673e-02 5.00693083e-01 1.39331311e-01 -3.03826094e-01
-3.26269656e-01 -1.27335116e-01 -1.02856839e+00 6.58510625e-01
1.26798260e+00 -6.56104088e-01 -4.08406645e-01 2.76032805e-01
9.75800931e-01 -4.37139034e-01 -8.80032361e-01 5.64691186e-01
3.87730360e-01 -1.12409914e+00 1.16538358e+00 -4.19283688e-01
6.70062721e-01 -5.42979121e-01 -1.66750461e-01 -1.13478756e+00
-3.31692100e-01 -6.50450647e-01 -1.81958511e-01 1.36803854e+00
1.75748974e-01 -4.80843335e-01 8.50019097e-01 2.46929735e-01
-2.20399678e-01 -1.03767800e+00 -6.77471995e-01 -6.96538448e-01
3.68657634e-02 -4.08318937e-01 3.06401849e-01 1.01640117e+00
-5.98091125e-01 2.79962897e-01 -5.85009992e-01 9.43348184e-02
7.48570085e-01 3.74578297e-01 7.32467115e-01 -8.18143070e-01
-6.14758551e-01 -5.06436944e-01 -3.73718441e-01 -1.23508143e+00
2.67676324e-01 -8.18573773e-01 9.73565206e-02 -1.26604044e+00
3.43706250e-01 -1.63244948e-01 -2.46977672e-01 4.74126458e-01
-4.21231627e-01 6.28426552e-01 3.79569650e-01 4.65150267e-01
-5.19598067e-01 8.54490101e-01 1.81436396e+00 -3.38995367e-01
-1.64011836e-01 -2.56510943e-01 -6.61217272e-01 8.14174056e-01
4.90163833e-01 -3.31138998e-01 -3.90925556e-01 -4.21896607e-01
-9.02691409e-02 -6.54686540e-02 8.32281947e-01 -6.09287500e-01
4.48237024e-02 -3.77048314e-01 4.19330359e-01 -4.14319783e-01
3.56246769e-01 -4.09112155e-01 3.30711007e-02 3.50337833e-01
-2.52834976e-01 -3.37813735e-01 -2.62577266e-01 6.52552247e-01
-4.85046566e-01 3.04741766e-02 9.13421571e-01 -2.76120424e-01
-8.86562765e-01 8.36187005e-01 9.56867561e-02 4.03504282e-01
9.67823148e-01 -3.77581507e-01 -2.41572961e-01 -3.19007546e-01
-8.17149043e-01 2.22750008e-01 8.05712581e-01 4.31069493e-01
7.09249616e-01 -1.44459128e+00 -6.84112489e-01 3.47745180e-01
1.66810751e-01 4.26552266e-01 5.28028011e-01 1.06096804e+00
-2.21606836e-01 -5.19782975e-02 -1.91229850e-01 -9.94022906e-01
-1.00212944e+00 5.95436633e-01 2.17064962e-01 -1.26645967e-01
-6.75395906e-01 9.66461718e-01 9.65703487e-01 -2.99813300e-01
5.68548515e-02 -2.74107158e-01 1.65143143e-02 -9.10673514e-02
2.25873470e-01 -7.37795010e-02 -3.70207191e-01 -7.90474951e-01
-2.35603094e-01 9.01743174e-01 6.44470528e-02 -1.96496353e-01
1.21675324e+00 -2.36526489e-01 -1.94397360e-01 4.81889427e-01
1.33634543e+00 -7.67656416e-02 -1.86483085e+00 -3.29657495e-01
-4.72781092e-01 -6.79533899e-01 3.22300084e-02 -4.49043900e-01
-1.58544660e+00 8.57733965e-01 4.83168334e-01 -1.08175524e-01
1.19403303e+00 4.48632866e-01 7.70058930e-01 -1.14413381e-01
1.12261884e-01 -9.78640974e-01 5.88301599e-01 2.92711943e-01
7.56662071e-01 -1.50282598e+00 -7.37034604e-02 -5.45261204e-01
-6.97129965e-01 7.42940128e-01 7.72145212e-01 -4.96407479e-01
6.69545293e-01 4.41546924e-02 -3.00134402e-02 -2.84610361e-01
-5.03675163e-01 -4.44970399e-01 6.82026744e-01 5.87153375e-01
1.28478184e-01 -2.48228069e-02 -8.24605003e-02 4.19619560e-01
-6.34588301e-02 -2.26738036e-01 1.90937504e-01 6.70036972e-01
-1.07065022e-01 -9.84328687e-01 -1.45035654e-01 2.62562573e-01
-3.28799248e-01 -1.85873926e-01 -1.64412171e-01 6.92677855e-01
4.60109204e-01 6.12642169e-01 -4.48595583e-02 -4.96131778e-01
7.21456856e-02 -3.23675692e-01 6.53188884e-01 -5.81704140e-01
-3.68743002e-01 4.42195356e-01 -3.77448052e-01 -8.49275351e-01
-7.95759320e-01 -6.27667427e-01 -1.17208004e+00 1.61739103e-02
-1.80665076e-01 -2.61852741e-01 2.96206802e-01 1.13406873e+00
2.28084266e-01 6.97298169e-01 7.69678712e-01 -1.13537741e+00
-3.08505952e-01 -6.58362865e-01 -5.57314456e-01 7.16809809e-01
3.47728670e-01 -6.62413120e-01 -4.56269860e-01 4.66794103e-01] | [9.39154052734375, -0.1960684061050415] |
580a5cd0-46cd-4f2c-93ad-e7f22161213d | transfernet-an-effective-and-transparent | 2104.07302 | null | https://arxiv.org/abs/2104.07302v2 | https://arxiv.org/pdf/2104.07302v2.pdf | TransferNet: An Effective and Transparent Framework for Multi-hop Question Answering over Relation Graph | Multi-hop Question Answering (QA) is a challenging task because it requires precise reasoning with entity relations at every step towards the answer. The relations can be represented in terms of labels in knowledge graph (e.g., \textit{spouse}) or text in text corpus (e.g., \textit{they have been married for 26 years}). Existing models usually infer the answer by predicting the sequential relation path or aggregating the hidden graph features. The former is hard to optimize, and the latter lacks interpretability. In this paper, we propose TransferNet, an effective and transparent model for multi-hop QA, which supports both label and text relations in a unified framework. TransferNet jumps across entities at multiple steps. At each step, it attends to different parts of the question, computes activated scores for relations, and then transfer the previous entity scores along activated relations in a differentiable way. We carry out extensive experiments on three datasets and demonstrate that TransferNet surpasses the state-of-the-art models by a large margin. In particular, on MetaQA, it achieves 100\% accuracy in 2-hop and 3-hop questions. By qualitative analysis, we show that TransferNet has transparent and interpretable intermediate results. | ['Hanwang Zhang', 'Juanzi Li', 'Lei Hou', 'Shulin Cao', 'Jiaxin Shi'] | 2021-04-15 | null | https://aclanthology.org/2021.emnlp-main.341 | https://aclanthology.org/2021.emnlp-main.341.pdf | emnlp-2021-11 | ['multi-hop-question-answering'] | ['knowledge-base'] | [-3.02106589e-02 8.75335872e-01 -5.00356734e-01 -5.74220300e-01
-9.43735600e-01 -6.87179208e-01 3.85794848e-01 5.02425253e-01
-1.09057985e-01 1.10034156e+00 4.12618279e-01 -6.29080296e-01
-5.64950585e-01 -1.20763266e+00 -7.98243821e-01 -2.07545087e-01
1.52415574e-01 1.14584196e+00 4.23157483e-01 -5.58247328e-01
-1.93183601e-01 -1.19613660e-02 -8.38911235e-01 4.34862465e-01
1.20987606e+00 9.47231114e-01 -3.97908986e-01 4.01532829e-01
-4.03360784e-01 1.26675618e+00 -4.60917205e-01 -1.40578806e+00
-3.84960920e-01 -4.81418580e-01 -1.58897269e+00 -3.32841098e-01
2.42532268e-01 -1.54593691e-01 -3.81909162e-01 9.02843654e-01
1.55259937e-01 -6.02641292e-02 7.70380914e-01 -1.28764176e+00
-9.77328360e-01 8.80563855e-01 -3.01051766e-01 1.65405050e-01
8.14877927e-01 -1.40988559e-01 1.75309086e+00 -8.38259995e-01
6.77987337e-01 1.44251573e+00 6.24076724e-01 3.20543975e-01
-1.05943203e+00 -3.95686567e-01 2.12300882e-01 5.54872215e-01
-1.14971304e+00 -1.57952234e-01 6.41964853e-01 -2.54221648e-01
8.63219023e-01 5.17612576e-01 2.84911513e-01 8.56012762e-01
1.01708561e-01 6.43441677e-01 8.78662229e-01 -3.14972341e-01
-6.97906222e-03 -7.24721923e-02 5.02217889e-01 1.17076695e+00
2.18015984e-01 -5.04371643e-01 -5.54105461e-01 -2.56752998e-01
3.34851831e-01 -2.95736700e-01 -2.97736168e-01 3.87047231e-02
-1.00775802e+00 9.44813609e-01 8.31721067e-01 2.49136671e-01
-2.61293441e-01 5.87727614e-02 -4.22332250e-02 4.42342192e-01
4.00416315e-01 4.88095254e-01 -7.47051597e-01 7.78046697e-02
-2.57560879e-01 1.11529842e-01 1.21956992e+00 9.36699569e-01
1.16767752e+00 -8.06261301e-01 -5.20374179e-01 5.79479337e-01
3.99997771e-01 1.53687939e-01 -1.16661899e-01 -1.16496766e+00
1.08287048e+00 1.21890604e+00 9.12249088e-02 -1.25104070e+00
-3.87558103e-01 -3.27557504e-01 -8.51998568e-01 -6.14486754e-01
5.98224223e-01 -3.13716173e-01 -7.58840799e-01 1.65584385e+00
5.33739507e-01 5.36001921e-02 1.85131416e-01 7.22458959e-01
1.31971312e+00 6.62185371e-01 3.30710202e-01 -9.70466807e-02
1.78993714e+00 -1.30735946e+00 -1.06274223e+00 -3.58896852e-01
8.27282727e-01 -2.52522707e-01 9.69667494e-01 -1.54426051e-02
-1.00507057e+00 -1.81480169e-01 -7.23986745e-01 -6.15257204e-01
-4.59019214e-01 3.26805860e-02 7.95784414e-01 4.69549656e-01
-1.04881847e+00 4.30272311e-01 -4.46642786e-01 -2.82449186e-01
3.88961792e-01 4.14779902e-01 -5.14013350e-01 -2.52695769e-01
-1.79783642e+00 9.98056293e-01 4.23423320e-01 2.18806386e-01
-3.91878158e-01 -4.38293725e-01 -1.13706315e+00 3.86237234e-01
7.88331211e-01 -1.24267960e+00 1.20532990e+00 -3.94104987e-01
-1.39037669e+00 7.89419651e-01 -5.61371684e-01 -2.29033247e-01
3.38991284e-01 -1.45267442e-01 -5.69479585e-01 3.45779836e-01
3.32349092e-01 4.31114286e-01 4.33654308e-01 -1.26607776e+00
-5.23815989e-01 -5.16944468e-01 7.31225967e-01 2.17040986e-01
-2.98179895e-01 -1.68243200e-01 -7.37357080e-01 -8.22977945e-02
2.50472128e-01 -6.83897257e-01 1.04841284e-01 -2.35655800e-01
-6.46061122e-01 -7.94113696e-01 4.96421188e-01 -8.63042653e-01
1.44125187e+00 -1.46800303e+00 2.66191691e-01 1.64683804e-01
5.61411858e-01 5.11964932e-02 2.19652429e-01 5.35536110e-01
1.44586623e-01 3.13114792e-01 -3.60516459e-01 -2.00688690e-01
1.83372274e-01 6.50095820e-01 -2.22305328e-01 1.74161140e-02
4.53669369e-01 1.47290301e+00 -1.07429302e+00 -9.79505837e-01
-4.70902592e-01 9.76950005e-02 -5.78535736e-01 1.15324415e-01
-5.46499133e-01 4.18954313e-01 -8.66618812e-01 7.48494387e-01
3.44093204e-01 -9.07215834e-01 5.04831195e-01 -3.17125946e-01
4.06229168e-01 5.72193801e-01 -7.44833827e-01 1.53964174e+00
-3.77480924e-01 4.02251482e-01 -8.49693045e-02 -9.81445789e-01
8.11474025e-01 4.50019568e-01 6.01470172e-02 -6.92079186e-01
-9.66758505e-02 2.23337740e-01 -1.27563402e-02 -8.74337792e-01
4.89389122e-01 -8.61947797e-03 -2.67673343e-01 3.07333052e-01
1.26242042e-01 9.07434449e-02 3.10323328e-01 4.29792106e-01
1.21671081e+00 -4.37389873e-02 2.90681243e-01 5.20141572e-02
7.20915139e-01 1.16026476e-01 5.20557523e-01 5.73948681e-01
1.39054582e-01 2.75100201e-01 1.03335059e+00 -2.38749862e-01
-3.29132080e-01 -1.08840418e+00 1.70182139e-01 1.15338516e+00
3.73526514e-01 -6.57093704e-01 -5.42467296e-01 -1.27173281e+00
-3.55302170e-02 7.09417880e-01 -8.77212584e-01 -1.94907814e-01
-8.07307065e-01 -3.88610899e-01 6.32082045e-01 4.92217600e-01
7.09653080e-01 -1.08753884e+00 2.31906697e-01 1.81939930e-01
-8.62321854e-01 -1.32646644e+00 -9.00953859e-02 -1.87096179e-01
-7.29777277e-01 -1.17385030e+00 -1.76845700e-01 -8.75708580e-01
7.97128439e-01 -2.64959991e-01 1.65667474e+00 4.50910717e-01
3.11830312e-01 2.02936351e-01 -4.04945165e-01 -5.45166135e-02
2.98394673e-02 4.50151175e-01 -5.61825335e-01 1.17117882e-01
3.11183244e-01 -5.24516821e-01 -7.05776751e-01 3.69921237e-01
-6.74282491e-01 -7.30297863e-02 5.79444110e-01 7.74498045e-01
4.01800156e-01 -5.51084206e-02 7.06052184e-01 -1.32553041e+00
8.17408025e-01 -8.55053127e-01 -1.59737676e-01 9.13389504e-01
-6.10464811e-01 2.73951769e-01 4.51566964e-01 4.08029296e-02
-1.22063184e+00 -3.82220596e-01 -4.01820004e-01 2.80739635e-01
1.15140542e-01 9.34444070e-01 -3.44494402e-01 2.25437298e-01
6.15212798e-01 -2.22196162e-01 -4.89044726e-01 -3.11292529e-01
5.16633451e-01 4.90348846e-01 4.44710761e-01 -9.78075266e-01
8.40106845e-01 3.76840174e-01 1.22040114e-03 -1.81019485e-01
-1.50424361e+00 -2.47010723e-01 -5.28301179e-01 -1.32955909e-01
1.13082588e+00 -6.33983612e-01 -1.27510035e+00 -7.97837749e-02
-1.30559492e+00 -2.81172007e-01 -5.76850660e-02 1.39135942e-01
-1.24498047e-01 1.69856921e-01 -8.63319457e-01 -5.07894874e-01
-4.12644684e-01 -7.72442341e-01 1.06335759e+00 4.14757431e-01
-2.56743699e-01 -1.33647919e+00 -1.02182291e-01 1.10213375e+00
5.08330017e-02 1.79280028e-01 1.43633235e+00 -6.73882067e-01
-6.74509048e-01 8.06826539e-03 -5.10647476e-01 -1.78384736e-01
1.27802223e-01 -2.92017043e-01 -6.75164878e-01 1.76356375e-01
-3.65172446e-01 -5.01874268e-01 8.96126628e-01 -1.58396289e-01
1.10890985e+00 -5.94866812e-01 -4.25713092e-01 6.04903772e-02
1.01665437e+00 -2.04152480e-01 6.98474407e-01 7.90731087e-02
7.01274633e-01 8.21921527e-01 6.99286044e-01 -7.75830001e-02
1.36947608e+00 5.26111841e-01 5.29593706e-01 -1.09411813e-02
1.59540549e-02 -4.66245562e-01 1.08178137e-02 8.78080070e-01
-2.53226072e-01 -4.91727978e-01 -1.02475882e+00 5.58867574e-01
-2.07131243e+00 -6.92581296e-01 -5.51518857e-01 1.58740032e+00
1.05795574e+00 7.30005056e-02 -2.82368511e-01 1.05372164e-02
5.90909362e-01 1.85604095e-02 -4.33849156e-01 -2.03233749e-01
-1.13844410e-01 2.92596400e-01 -8.35952759e-02 9.05125558e-01
-7.32627153e-01 1.08031726e+00 5.29239655e+00 7.34984100e-01
-3.47318381e-01 3.69668573e-01 7.19155192e-01 6.13976955e-01
-8.11889946e-01 4.03832048e-01 -7.46543944e-01 1.55367106e-01
6.07476890e-01 -5.50850146e-02 2.09041074e-01 1.93422735e-01
-4.26813245e-01 -1.20849781e-01 -1.28559351e+00 5.23613334e-01
-8.24895874e-03 -1.43341255e+00 2.24210024e-01 -1.17670670e-01
7.28218019e-01 -4.63562965e-01 -2.14054570e-01 7.61680007e-01
5.01762390e-01 -1.30391002e+00 3.58186960e-01 6.04077697e-01
6.14383578e-01 -4.20527965e-01 9.20704722e-01 4.94330376e-01
-1.26909900e+00 -6.09859563e-02 -7.36799538e-02 -1.22214429e-01
2.72099257e-01 5.80538034e-01 -8.40892315e-01 1.32450879e+00
6.09838665e-01 3.87478143e-01 -6.40349388e-01 4.75961626e-01
-9.15532649e-01 7.87137210e-01 -7.53087252e-02 -1.65842116e-01
1.95944488e-01 -3.50502312e-01 7.62092993e-02 8.66931677e-01
3.44818413e-01 6.21092141e-01 -3.16130295e-02 7.99695849e-01
-7.05056548e-01 2.89109759e-02 -2.53482699e-01 1.80350933e-02
4.69168484e-01 1.35093153e+00 -3.22662175e-01 -3.93126726e-01
-3.97574335e-01 8.79430234e-01 9.66165423e-01 6.69087470e-01
-8.35976541e-01 -3.69078636e-01 2.41675392e-01 2.95926444e-02
1.30482584e-01 3.51677276e-02 -2.45122254e-01 -1.19206667e+00
2.77314246e-01 -4.78041351e-01 1.13757598e+00 -9.56780732e-01
-1.41160595e+00 5.76513469e-01 -2.01930180e-01 -6.02286816e-01
-3.48941326e-01 -4.57105398e-01 -4.51355934e-01 8.61915886e-01
-1.73852456e+00 -1.31303442e+00 -4.34701025e-01 7.27027535e-01
6.54279888e-02 2.58005738e-01 9.44204926e-01 4.27426159e-01
-4.86955583e-01 6.32241488e-01 -5.50090194e-01 4.84601021e-01
5.81726730e-01 -1.45005226e+00 1.82742849e-01 3.17500353e-01
1.98579684e-01 5.59124827e-01 5.08176863e-01 -5.97974598e-01
-1.27513051e+00 -1.03188074e+00 1.65631318e+00 -7.77318656e-01
8.02093804e-01 -1.60492375e-01 -1.00349748e+00 1.16488779e+00
4.01221335e-01 1.08018974e-02 7.28236258e-01 5.75473666e-01
-4.43000406e-01 -1.92724451e-01 -1.10798562e+00 5.89838445e-01
1.18233442e+00 -5.50529778e-01 -9.21157122e-01 4.77056026e-01
1.09096801e+00 -5.35262764e-01 -1.21656430e+00 6.20616257e-01
9.64631960e-02 -8.64437282e-01 9.05882776e-01 -1.02712941e+00
7.25333154e-01 -2.46566668e-01 1.10360779e-01 -1.14840174e+00
-3.19827586e-01 -5.85058272e-01 -6.33232117e-01 1.38256550e+00
1.13147438e+00 -9.00356352e-01 7.36650765e-01 7.50474632e-01
-3.15023735e-02 -1.09838212e+00 -9.42550719e-01 -3.12388480e-01
-2.49106735e-02 -1.44055352e-01 8.50399077e-01 1.19068384e+00
1.44322380e-01 9.04755890e-01 -1.07689537e-01 5.04966080e-01
3.38383704e-01 5.68666101e-01 5.06739318e-01 -1.35182464e+00
-3.05691123e-01 -2.38369167e-01 3.45857479e-02 -1.29154086e+00
4.23352689e-01 -1.03196788e+00 -3.24902028e-01 -2.34012461e+00
1.51632592e-01 -5.39434254e-01 -1.01765700e-01 6.62898600e-01
-6.64815426e-01 4.77043279e-02 -1.35878325e-01 -6.28759339e-02
-1.02107978e+00 6.03659511e-01 1.59962046e+00 -3.12066793e-01
1.63958132e-01 5.82514964e-02 -9.85440850e-01 6.51749671e-01
7.77866304e-01 -4.86629128e-01 -4.60591048e-01 -7.04701424e-01
9.36143398e-01 5.85615575e-01 3.85071546e-01 -3.57870162e-01
5.13394475e-01 -7.83817172e-02 1.27989233e-01 -2.73074090e-01
6.02269590e-01 -6.29003227e-01 -9.83200669e-02 2.49018788e-01
-3.72789323e-01 -1.78935770e-02 -2.50086695e-01 6.05718791e-01
-5.67557931e-01 -2.17386067e-01 -8.74426737e-02 -5.30059114e-02
-3.81024927e-01 4.65401143e-01 2.71975070e-01 5.82289219e-01
7.18217552e-01 2.39936337e-01 -8.66520345e-01 -7.17468619e-01
-9.07885432e-01 8.07836890e-01 -2.19844624e-01 3.54291916e-01
4.72388297e-01 -1.41861427e+00 -8.04975569e-01 -4.29060489e-01
1.89111516e-01 3.21770370e-01 1.97284862e-01 1.03164244e+00
-3.49063069e-01 4.96614546e-01 3.02334994e-01 -3.21947247e-01
-1.02373588e+00 1.99491501e-01 3.39100301e-01 -8.59436750e-01
-1.40460223e-01 9.26414788e-01 4.01379168e-02 -7.43869483e-01
1.13546560e-02 -2.10901916e-01 -6.77284062e-01 3.02190185e-01
3.45908888e-02 2.04389915e-01 3.87764536e-02 -4.16256875e-01
-3.18772227e-01 5.68791449e-01 9.64895915e-03 5.15827015e-02
1.21742868e+00 -1.09587178e-01 -6.90111995e-01 2.54833966e-01
1.00180638e+00 1.12224855e-01 -7.73041725e-01 -4.86622423e-01
2.21314713e-01 -2.04242215e-01 -5.00156939e-01 -9.64814305e-01
-9.17372525e-01 6.71232760e-01 -3.59571368e-01 6.17610633e-01
9.16456878e-01 6.50556564e-01 9.02494669e-01 7.12777138e-01
2.70223439e-01 -7.53717601e-01 2.28138342e-01 6.64888740e-01
9.37028408e-01 -1.29675531e+00 -1.19760811e-01 -9.74078417e-01
-6.10348403e-01 8.58913600e-01 8.61539721e-01 4.51636583e-01
4.78192359e-01 -2.55961984e-01 -4.56388034e-02 -7.04875708e-01
-8.83147478e-01 -3.31678420e-01 5.03292263e-01 1.89862564e-01
3.88060123e-01 1.82178512e-01 -4.23544228e-01 7.64958441e-01
-3.60270262e-01 -2.74346232e-01 3.19366008e-02 6.49452567e-01
-3.28731984e-01 -1.09951270e+00 -1.01957478e-01 5.29599011e-01
-4.34775144e-01 -6.74207136e-02 -7.09420145e-01 7.81248033e-01
1.68145001e-01 1.40266609e+00 -2.65740305e-01 -3.18088770e-01
4.49338585e-01 1.62904859e-01 3.91589731e-01 -4.74469721e-01
-6.86924338e-01 -7.73526251e-01 7.60747015e-01 -3.38209659e-01
-3.98293972e-01 -3.50376785e-01 -1.58384621e+00 -2.84622967e-01
-3.23905081e-01 5.89622021e-01 1.63496062e-01 1.39358151e+00
4.50446516e-01 6.73186660e-01 3.37566614e-01 2.25848660e-01
-2.62638003e-01 -9.61962163e-01 -1.67581707e-01 4.20528501e-01
1.69140935e-01 -5.09310544e-01 -1.21877380e-01 -5.20922653e-02] | [10.52070140838623, 7.892444610595703] |
10e56e2e-cab8-42aa-aef0-ee7b1df8cc30 | knowledge-base-question-answering-through | null | null | https://aclanthology.org/2021.eacl-main.35 | https://aclanthology.org/2021.eacl-main.35.pdf | Knowledge Base Question Answering through Recursive Hypergraphs | Knowledge Base Question Answering (KBQA) is the problem of predicting an answer for a factoid question over a given knowledge base (KB). Answering questions typically requires reasoning over multiple links in the given KB. Humans tend to answer questions by grouping different objects to perform reasoning over acquired knowledge. Hypergraphs provide a natural tool to model group relationships. In this work, inspired by typical human intelligence, we propose a new method for KBQA based on hypergraphs. Existing methods for KBQA, though effective, do not explicitly incorporate the recursive relational group structure in the given KB. Our method, which we name RecHyperNet (Recursive Hypergraph Network), exploits a new way of modelling KBs through recursive hypergraphs to organise such group relationships in KBs. Experiments on multiple KBQA benchmarks demonstrate the effectiveness of the proposed RecHyperNet. We have released the code. | ['Srinidhi G', 'Indira K M', 'Vaishnavi S', 'Dayanidhi R S', 'Naganand Yadati'] | 2021-04-01 | null | null | null | eacl-2021-2 | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-1.81637555e-01 1.01068926e+00 -8.30266923e-02 -3.80848795e-01
-4.99501288e-01 -6.38113678e-01 4.84824836e-01 5.42905986e-01
3.77263799e-02 1.08443654e+00 3.95507127e-01 -5.53681731e-01
-7.60636330e-01 -1.61635745e+00 -8.64320219e-01 -7.46150315e-02
9.25005600e-02 9.12940502e-01 1.06866419e+00 -6.82750225e-01
1.10241957e-01 1.69569299e-01 -1.14977682e+00 6.81248069e-01
9.87305880e-01 7.65713215e-01 -8.42310712e-02 5.25206983e-01
-7.80381799e-01 1.71164918e+00 -6.72446966e-01 -1.09649253e+00
-2.82962229e-02 -4.99336720e-01 -1.88682210e+00 -2.45204374e-01
4.19275135e-01 -1.58741269e-02 -4.05485064e-01 8.69572043e-01
-8.00124109e-02 2.58677214e-01 6.12317145e-01 -1.39152980e+00
-5.47158182e-01 1.06846452e+00 8.02268216e-04 4.64765400e-01
9.09982324e-01 -4.32959288e-01 1.69453251e+00 -4.35188681e-01
7.72224307e-01 1.35246181e+00 5.72123706e-01 4.39863533e-01
-9.94975567e-01 -1.44445568e-01 -2.20384012e-04 1.02272034e+00
-1.28424311e+00 1.43844754e-01 7.57410169e-01 -2.41394207e-01
9.84536946e-01 5.74870765e-01 7.63180435e-01 1.88636139e-01
-3.03149759e-03 6.38419926e-01 1.04580247e+00 -5.48749208e-01
1.83718607e-01 1.77853569e-01 9.89573240e-01 1.01118135e+00
4.37787265e-01 -6.79970622e-01 -5.94030142e-01 -6.67164981e-01
6.22865975e-01 -3.55382919e-01 -3.98413926e-01 -6.06085896e-01
-9.30812538e-01 1.04233027e+00 8.66462469e-01 2.78260827e-01
-2.87288934e-01 2.62029856e-01 1.67525873e-01 5.53599358e-01
-1.54575158e-03 7.51794040e-01 -4.53862876e-01 3.84513378e-01
-1.66053087e-01 4.39300925e-01 1.52437413e+00 9.82791841e-01
1.15900660e+00 -7.01930046e-01 -1.82811126e-01 6.68708205e-01
3.29180211e-01 1.47468954e-01 2.24867582e-01 -9.48257208e-01
5.45403004e-01 1.26937461e+00 4.03733887e-02 -1.37702811e+00
-4.84220445e-01 4.29318286e-02 -3.61261815e-01 -6.52014256e-01
3.42078745e-01 1.98550090e-01 -6.56045258e-01 1.27096105e+00
7.54502118e-01 1.37679368e-01 4.56579804e-01 5.99835575e-01
1.47462368e+00 5.65166175e-01 -5.61532080e-02 2.98053883e-02
1.52840519e+00 -9.67140913e-01 -7.97652006e-01 1.52689397e-01
7.52583086e-01 1.40258344e-02 9.25760806e-01 2.29253694e-01
-9.05731380e-01 -1.67112708e-01 -7.13495076e-01 -3.18342119e-01
-6.63828433e-01 -7.12911308e-01 7.12834299e-01 6.35306656e-01
-1.24013281e+00 6.08805194e-03 -2.51525164e-01 -3.73098671e-01
2.01689422e-01 4.68562156e-01 -2.26843163e-01 -3.85032624e-01
-1.81188178e+00 1.01837623e+00 1.01834357e+00 -1.76086742e-02
-5.37073374e-01 -6.32177174e-01 -8.20390642e-01 2.03831315e-01
1.27062523e+00 -1.24474633e+00 1.24082875e+00 -3.69604230e-01
-1.19731474e+00 6.49840772e-01 4.46982197e-02 -6.90056980e-01
-1.57680124e-01 -1.34243920e-01 -4.90317702e-01 5.09601891e-01
-1.23154923e-01 3.29751223e-01 4.26766008e-01 -1.52918589e+00
-5.45889735e-01 -3.27758491e-01 9.55038309e-01 2.44696021e-01
-2.41237700e-01 -2.13821843e-01 -4.76027757e-01 -6.72474876e-02
1.68577433e-01 -5.32165766e-01 -2.31882185e-01 -6.56359732e-01
-6.14115477e-01 -7.19332576e-01 6.54568732e-01 -6.28807425e-01
1.44007981e+00 -1.30764163e+00 2.19950244e-01 7.01122880e-01
8.06463599e-01 1.79235071e-01 1.85682014e-01 7.08357513e-01
3.28841746e-01 2.22474247e-01 -2.21453473e-01 7.50934303e-01
2.05932960e-01 7.88732588e-01 -5.49191296e-01 -2.42690086e-01
-8.55998024e-02 1.37722874e+00 -1.16652775e+00 -8.07947159e-01
-3.64541769e-01 -1.65795073e-01 -5.56682289e-01 1.59280181e-01
-9.50299501e-01 -1.18030079e-01 -8.07826519e-01 4.64891136e-01
3.78802299e-01 -6.18481457e-01 4.47923660e-01 -2.62738019e-01
6.65965855e-01 5.21944880e-01 -1.10393882e+00 1.31718612e+00
-9.71567184e-02 1.77428648e-01 -3.30953062e-01 -8.70684683e-01
9.47270155e-01 2.80697227e-01 1.42450809e-01 -4.14612144e-01
-2.05329895e-01 1.36462256e-01 1.01946399e-01 -8.29341948e-01
6.82300925e-01 -1.76150069e-01 -9.03080776e-02 4.63596076e-01
8.82775709e-02 -6.24567747e-01 4.37577993e-01 9.18475330e-01
1.55496693e+00 -2.89066762e-01 7.40391731e-01 -3.00826222e-01
7.78853238e-01 5.12879848e-01 1.23475254e-01 9.19980168e-01
3.70910943e-01 -8.15586448e-02 9.15445864e-01 -5.98086238e-01
-5.25745153e-01 -1.19240355e+00 2.92172968e-01 8.36630046e-01
2.62281418e-01 -7.89806247e-01 -4.94933814e-01 -1.14716339e+00
1.93638653e-01 6.26571834e-01 -4.93199944e-01 -5.36251962e-02
-5.80402315e-01 -3.12114358e-01 6.22853518e-01 3.13918263e-01
4.82148677e-01 -1.35514653e+00 -4.25711811e-01 2.04369053e-01
-5.33334374e-01 -1.19121325e+00 1.79913566e-01 -2.44688496e-01
-8.50419223e-01 -1.66508460e+00 -1.20302431e-01 -6.56253874e-01
5.09416461e-01 1.95653379e-01 1.73445380e+00 6.52333856e-01
1.06898881e-01 1.04776955e+00 -7.74638593e-01 -4.69691008e-01
-4.44043219e-01 2.66090065e-01 -4.02581066e-01 -1.75776139e-01
4.48785692e-01 -5.19630373e-01 -4.31287020e-01 3.03426623e-01
-1.29145610e+00 -5.20267077e-02 4.82592016e-01 6.60644114e-01
4.91203010e-01 3.67731631e-01 8.99448514e-01 -1.64938843e+00
9.80750740e-01 -8.50777566e-01 -4.82034683e-01 9.05698657e-01
-4.69181597e-01 2.81417459e-01 3.58157843e-01 2.56397799e-02
-1.13228250e+00 -4.40833867e-01 4.75041270e-02 1.54934272e-01
8.87869969e-02 1.25244069e+00 -2.96293586e-01 -1.78200647e-01
7.65361249e-01 -1.92239791e-01 -3.70586634e-01 -1.19635575e-01
8.40040267e-01 1.80946812e-01 4.93991107e-01 -8.67004573e-01
8.64147663e-01 3.90275747e-01 2.75010109e-01 -5.09021938e-01
-1.15466392e+00 -8.32290053e-01 -6.84101403e-01 -1.97768956e-01
6.82620525e-01 -4.45854992e-01 -1.05970371e+00 -2.39822522e-01
-1.18823934e+00 -2.66913235e-01 -3.79441202e-01 9.49473977e-02
-4.77667779e-01 4.77048814e-01 -4.03050125e-01 -7.02415228e-01
-1.65131167e-01 -2.74891555e-01 5.62999368e-01 9.34765562e-02
-2.15252072e-01 -1.30993354e+00 3.85668069e-01 9.68333840e-01
1.07608370e-01 2.58585572e-01 1.54195440e+00 -1.02625191e+00
-1.01471627e+00 3.66932377e-02 -3.37325543e-01 -3.07800192e-02
1.85265273e-01 -5.23164690e-01 -5.19146323e-01 2.16790885e-01
-2.06695765e-01 -4.74474519e-01 7.60668635e-01 -2.37520441e-01
9.91297543e-01 -7.95453131e-01 -2.95701414e-01 -2.20710650e-01
1.41040051e+00 -3.05834934e-02 8.95170569e-01 4.55321640e-01
8.74845982e-01 8.87780786e-01 5.16930521e-01 1.85126271e-02
1.00708985e+00 5.08417547e-01 5.16272604e-01 2.90788412e-01
3.02296616e-02 -2.23820567e-01 -1.47156671e-01 1.09715891e+00
-2.57329196e-01 -2.80293733e-01 -1.35027015e+00 7.08778918e-01
-1.99443948e+00 -8.24762881e-01 -5.70897043e-01 1.63850820e+00
1.20849407e+00 1.28254797e-02 1.34593382e-01 2.14214757e-01
4.72151190e-01 -1.71691403e-01 -3.93467069e-01 -2.75472730e-01
-1.58234492e-01 3.59714270e-01 1.78103328e-01 7.45647132e-01
-4.15560246e-01 1.06687522e+00 5.85844088e+00 6.24280393e-01
-4.84244861e-02 -7.64371380e-02 -3.71611444e-03 5.81889749e-01
-7.72190869e-01 3.65729660e-01 -8.70276451e-01 -2.09918141e-01
8.57608199e-01 -4.23137486e-01 3.79638165e-01 5.08525193e-01
-5.66142261e-01 -3.43354642e-01 -1.04231310e+00 4.75118726e-01
1.80023000e-01 -1.54701257e+00 5.92268646e-01 -2.21138611e-01
7.19980419e-01 -5.65068066e-01 -4.39570516e-01 4.78849411e-01
9.40435767e-01 -9.07808423e-01 1.56913698e-01 9.67092872e-01
1.79824755e-02 -7.12543488e-01 7.93903351e-01 4.03729647e-01
-1.08587170e+00 -1.70152545e-01 -5.89424491e-01 5.95062934e-02
6.43500239e-02 5.49652874e-01 -1.45900512e+00 1.38955212e+00
7.21040010e-01 2.56102443e-01 -9.28782523e-01 1.09879971e+00
-6.45422220e-01 8.14359546e-01 -5.02072508e-03 -2.85979241e-01
1.15604825e-01 -1.55346198e-02 3.87070388e-01 9.38272655e-01
-3.89434174e-02 6.92401290e-01 -3.97187248e-02 6.59962654e-01
-3.47478718e-01 1.78726420e-01 -5.47623754e-01 -2.24357937e-02
2.79389560e-01 1.10612917e+00 -6.24941111e-01 -5.46199083e-01
-4.19054598e-01 4.88453150e-01 6.40576422e-01 3.09728563e-01
-4.76267636e-01 -4.20641690e-01 -7.72473961e-02 1.96705177e-01
2.12640300e-01 -1.25945616e-03 9.68925357e-02 -8.98872435e-01
1.11342847e-01 -9.75831628e-01 1.26498222e+00 -1.16509151e+00
-1.55487692e+00 5.56892157e-01 4.50519383e-01 -4.36219156e-01
-3.48130971e-01 -4.72033143e-01 -1.39152393e-01 5.90192676e-01
-1.62533581e+00 -1.10772002e+00 -5.31943142e-01 1.00137997e+00
7.08266953e-03 1.94540843e-01 8.90375137e-01 -2.37943292e-01
2.10532159e-01 -1.67314569e-03 -4.84313041e-01 5.09895757e-02
3.00539464e-01 -1.69553196e+00 1.14078008e-01 2.94984818e-01
5.72363675e-01 8.26470137e-01 6.35821104e-01 -9.05553699e-01
-1.33485878e+00 -9.17471409e-01 9.22964752e-01 -8.91499639e-01
1.06158042e+00 -1.37644112e-01 -1.54085469e+00 9.12350297e-01
2.80086964e-01 -1.49455652e-01 8.13027143e-01 3.56607795e-01
-7.37063169e-01 -2.33523324e-01 -1.00536704e+00 4.30576742e-01
1.10660589e+00 -6.06464982e-01 -1.41349900e+00 5.74869514e-01
1.30094266e+00 -2.91904837e-01 -1.14883053e+00 5.50043404e-01
1.81704521e-01 -8.30025971e-01 9.44918931e-01 -9.87858295e-01
2.58775741e-01 -5.78071356e-01 -9.45360214e-02 -1.31630874e+00
-1.88752159e-01 -3.83949459e-01 -7.51326859e-01 9.82430816e-01
6.27077341e-01 -7.32442498e-01 1.01086724e+00 7.27809787e-01
2.19992474e-02 -5.97955048e-01 -7.88693070e-01 -7.75791347e-01
-2.39108458e-01 -1.84314132e-01 9.17027891e-01 1.11382699e+00
4.33756411e-01 7.33703554e-01 1.30257737e-02 5.63711345e-01
3.96999598e-01 2.17409953e-01 9.21036422e-01 -1.68477249e+00
-3.86519521e-01 -1.52662992e-01 -5.44545352e-01 -9.04142261e-01
2.38084182e-01 -9.61920142e-01 -1.94365367e-01 -2.35989070e+00
2.79964894e-01 -4.42717403e-01 -9.53623205e-02 5.58610737e-01
-3.83343816e-01 7.97155872e-03 7.92449992e-03 9.33306739e-02
-1.16250122e+00 2.95097083e-01 1.45591795e+00 -2.52311766e-01
-3.15168798e-02 -3.15413550e-02 -7.16545641e-01 5.90966105e-01
6.99193239e-01 -4.89098489e-01 -9.21093285e-01 -1.78243920e-01
1.15974677e+00 2.58264720e-01 5.57627201e-01 -6.05038762e-01
8.45520020e-01 -2.01822460e-01 -4.00407970e-01 -5.12837052e-01
3.55443001e-01 -9.26467478e-01 2.94938415e-01 2.51598716e-01
-2.86813229e-01 -1.61964167e-02 9.18935314e-02 8.04031312e-01
-6.12897813e-01 -4.63286191e-01 -1.56612936e-02 -2.28366405e-01
-9.47919309e-01 7.45128766e-02 -6.31495789e-02 3.71034503e-01
7.91559339e-01 2.79347179e-03 -8.10681641e-01 -5.78185022e-01
-8.32211792e-01 5.90568006e-01 -8.18759874e-02 7.87863880e-02
8.25250387e-01 -1.17070270e+00 -5.33135593e-01 -5.42348683e-01
3.15027505e-01 3.06230038e-01 1.56355515e-01 4.99120802e-01
-6.56769335e-01 7.04964399e-01 3.38560641e-02 -1.20600671e-01
-1.27836514e+00 6.70955896e-01 3.28264266e-01 -7.52809644e-01
-4.17283982e-01 6.89831555e-01 9.84859988e-02 -6.74527407e-01
-4.96719517e-02 -4.09863859e-01 -7.64762521e-01 -7.21225068e-02
3.70542228e-01 4.76760089e-01 2.09337711e-01 -3.01996619e-01
-9.71533731e-02 1.85182542e-01 -1.66720465e-01 -4.14472222e-02
1.13742507e+00 -1.09747745e-01 -8.86917531e-01 5.09946167e-01
4.75744933e-01 1.57592088e-01 -3.03490102e-01 -6.99325562e-01
5.91182768e-01 -3.19542408e-01 -4.97934908e-01 -9.52548683e-01
-6.06720626e-01 2.67739356e-01 -4.62790370e-01 1.08020377e+00
9.30456996e-01 6.45723760e-01 7.02763617e-01 1.25768638e+00
5.73237419e-01 -6.74481511e-01 3.47559810e-01 7.14652181e-01
1.09729159e+00 -7.75286794e-01 1.27123266e-01 -1.12955344e+00
-5.56130946e-01 9.78001773e-01 7.60844827e-01 2.05315068e-01
6.62389576e-01 -3.26830059e-01 -6.77467138e-02 -1.00954282e+00
-8.02801192e-01 -3.97693217e-01 5.16570449e-01 6.65241420e-01
-1.17270432e-01 -3.54185142e-02 -3.12930703e-01 6.96022093e-01
-5.26481330e-01 -6.58652335e-02 7.42657125e-01 9.68081176e-01
-7.31408656e-01 -1.15833473e+00 -4.79570627e-01 6.92256153e-01
-2.58194268e-01 -1.73318431e-01 -9.93446052e-01 9.37558234e-01
-1.16559602e-01 1.16950762e+00 -3.75095516e-01 -3.27475667e-01
5.12325764e-01 2.36838743e-01 6.31245553e-01 -8.75850677e-01
-6.66649163e-01 -8.92250955e-01 4.34220463e-01 -2.52715558e-01
-8.15810144e-01 -1.07162870e-01 -1.44000375e+00 -2.24231109e-01
-4.63956267e-01 8.33988726e-01 9.29213166e-02 1.09048045e+00
1.33424520e-01 4.85547036e-01 1.15444563e-01 5.70102870e-01
-1.44867346e-01 -7.64467061e-01 -7.59848058e-01 3.28609347e-01
5.55419028e-02 -5.78887701e-01 -4.00580093e-02 1.35201856e-01] | [10.433769226074219, 7.892094135284424] |
5dc10f91-8da9-4e4d-a3f4-e753061436fa | performance-evaluation-of-vanilla-residual | 2211.16570 | null | https://arxiv.org/abs/2211.16570v2 | https://arxiv.org/pdf/2211.16570v2.pdf | Performance Evaluation of Vanilla, Residual, and Dense 2D U-Net Architectures for Skull Stripping of Augmented 3D T1-weighted MRI Head Scans | Skull Stripping is a requisite preliminary step in most diagnostic neuroimaging applications. Manual Skull Stripping methods define the gold standard for the domain but are time-consuming and challenging to integrate into processing pipelines with a high number of data samples. Automated methods are an active area of research for head MRI segmentation, especially deep learning methods such as U-Net architecture implementations. This study compares Vanilla, Residual, and Dense 2D U-Net architectures for Skull Stripping. The Dense 2D U-Net architecture outperforms the Vanilla and Residual counterparts by achieving an accuracy of 99.75% on a test dataset. It is observed that dense interconnections in a U-Net encourage feature reuse across layers of the architecture and allow for shallower models with the strengths of a deeper network. | ['Mahesh H. Shindikar', 'Ketaki D. Kamble', 'Rashmika K. Patole', 'Anway S. Pimpalkar'] | 2022-11-29 | null | null | null | null | ['skull-stripping'] | ['medical'] | [ 3.61518785e-02 5.27681410e-01 1.14006557e-01 -5.42920768e-01
-2.37991616e-01 -2.32012570e-01 2.79469937e-01 2.27664202e-01
-5.13521314e-01 5.17191589e-01 2.01087333e-02 -2.83776134e-01
-7.40376487e-02 -8.36921751e-01 -5.68871140e-01 -4.89330560e-01
-3.37499470e-01 6.76378250e-01 6.05114460e-01 -1.23014942e-01
-2.13920280e-01 8.32227886e-01 -9.35600817e-01 1.13265947e-01
5.07776678e-01 8.40927005e-01 -1.19700581e-02 1.63024783e-01
-1.48792803e-01 2.60652661e-01 -3.77207667e-01 -4.32318717e-01
3.49730700e-01 -3.89195755e-02 -7.11973369e-01 -2.30358556e-01
5.58164537e-01 -3.83590251e-01 -4.39430177e-01 1.06741142e+00
6.65959895e-01 2.80078441e-01 6.79698050e-01 -8.82193327e-01
-3.36498290e-01 1.00984979e+00 -7.01191127e-01 7.81592846e-01
-4.43339385e-02 1.56587884e-01 4.22622681e-01 -6.97425544e-01
7.15813994e-01 8.83894444e-01 1.00505841e+00 7.12779105e-01
-1.36769021e+00 -7.98934340e-01 -7.04649016e-02 -1.77809343e-01
-1.37608230e+00 -5.09622455e-01 5.86511910e-01 -8.15694749e-01
1.15826976e+00 -1.99168641e-02 9.33940589e-01 8.20030332e-01
3.93007785e-01 8.42643142e-01 1.21634293e+00 2.38279309e-02
4.33668405e-01 -2.38181755e-01 8.75252008e-01 6.45683229e-01
5.63381195e-01 -1.66233182e-01 -2.37742051e-01 1.04255890e-02
1.29021907e+00 1.57270253e-01 -2.17428729e-01 -2.10296601e-01
-6.64392591e-01 7.81050324e-01 9.87195969e-01 4.35477883e-01
-6.09170556e-01 3.31885032e-02 5.64350963e-01 1.41633762e-04
5.26092887e-01 4.80710685e-01 -1.62375569e-01 3.20548624e-01
-1.66511524e+00 2.01706409e-01 2.97361404e-01 8.53727162e-01
4.55410570e-01 2.90235639e-01 -2.28146002e-01 7.22333491e-01
1.36658117e-01 -1.45282000e-01 7.44388938e-01 -4.49866623e-01
6.35579973e-02 7.35196173e-01 -5.98060191e-01 -3.89573663e-01
-1.05169713e+00 -7.71833420e-01 -1.06809640e+00 4.37310517e-01
3.44677567e-01 -4.96346712e-01 -1.44859409e+00 1.26592255e+00
4.11878228e-01 1.10063612e-01 -6.12622797e-01 7.89938688e-01
1.00340128e+00 -3.04344948e-02 2.14288577e-01 1.25676524e-02
1.25529468e+00 -1.02782047e+00 -6.26190007e-01 -5.49369574e-01
5.18800378e-01 -3.70999068e-01 5.87484121e-01 7.43215531e-02
-1.38835430e+00 -1.48678422e-01 -9.73010778e-01 -2.97893941e-01
-3.38340670e-01 -2.92889923e-01 9.62362826e-01 1.01762700e+00
-1.54946685e+00 7.79648483e-01 -1.17513573e+00 -2.22545788e-01
1.31161380e+00 7.75404751e-01 -6.83030546e-01 7.46151209e-02
-7.10147679e-01 1.11531973e+00 2.75815964e-01 1.11044228e-01
-8.07497680e-01 -1.00699389e+00 -7.46746302e-01 -9.10381845e-04
-7.96178505e-02 -8.61057878e-01 1.28709829e+00 -4.58152860e-01
-1.16730595e+00 1.04732215e+00 -2.99943946e-02 -9.56918597e-01
8.63793135e-01 -2.16336548e-03 1.43156946e-01 3.09278220e-01
-8.24102908e-02 8.33447516e-01 7.26149499e-01 -8.24306726e-01
-1.84295475e-01 -8.51740897e-01 -2.58281142e-01 -3.16004992e-01
1.74043536e-01 6.56500012e-02 -2.13824674e-01 -7.29328275e-01
5.43080389e-01 -5.81383884e-01 -4.59764987e-01 6.56185299e-02
-3.34546685e-01 -9.38684121e-02 7.20536470e-01 -9.56214786e-01
9.37840104e-01 -1.67746377e+00 -2.23332152e-01 5.24121046e-01
8.43421042e-01 4.16236401e-01 2.96597689e-01 -3.55911940e-01
-3.88945520e-01 3.32756527e-02 -5.99630177e-01 -4.21665996e-01
-3.59726816e-01 2.96679419e-02 4.07453179e-01 6.39014065e-01
-5.92723824e-02 9.82752979e-01 -8.34974945e-01 -2.39004657e-01
2.33226776e-01 5.52410185e-01 -8.26903343e-01 -4.43072896e-03
2.23655537e-01 4.82958913e-01 -5.10483682e-02 5.26334882e-01
7.65630066e-01 1.42183453e-01 1.02536343e-01 -2.30741844e-01
-7.61859044e-02 1.47174433e-01 -9.05559301e-01 1.77149129e+00
-1.51337743e-01 6.09999895e-01 4.26982939e-01 -8.24021816e-01
8.20212722e-01 3.90643090e-01 5.36270022e-01 -4.77543563e-01
8.74376953e-01 2.68768877e-01 6.07390046e-01 -2.18415678e-01
1.43423855e-01 -4.27159250e-01 5.01727939e-01 4.27741200e-01
5.55918455e-01 -3.31176192e-01 1.06277011e-01 1.29239604e-01
1.23466110e+00 -1.23315200e-01 2.30761409e-01 -6.85552180e-01
-9.37288776e-02 -3.71752471e-01 4.23864782e-01 6.83832824e-01
-5.60367346e-01 8.35362077e-01 4.97023493e-01 -5.32429636e-01
-1.01779044e+00 -1.07967234e+00 -5.92623651e-01 5.28420150e-01
-6.50516748e-01 -1.00568846e-01 -1.35466611e+00 -5.46049774e-01
-1.32387370e-01 6.61943853e-01 -1.04066408e+00 -3.26919705e-02
-5.70763350e-01 -8.98918867e-01 6.83006704e-01 7.83504784e-01
5.16963959e-01 -9.28047419e-01 -8.51916313e-01 3.91426265e-01
3.49844337e-01 -1.00972378e+00 -2.23192707e-01 6.55265570e-01
-1.27106690e+00 -1.02720261e+00 -8.84423494e-01 -7.92411506e-01
9.32243466e-01 -8.25273171e-02 8.88631642e-01 1.69021413e-01
-4.67131019e-01 -5.73293082e-02 2.16430258e-02 -5.27302682e-01
1.42724216e-01 3.24863464e-01 1.18408404e-01 -2.77447045e-01
5.04073262e-01 -9.14233923e-01 -8.02147746e-01 -1.90217756e-02
-8.98199618e-01 -2.30187327e-02 4.25237417e-01 5.33190668e-01
7.69044042e-01 -2.40483224e-01 5.74102819e-01 -1.20034099e+00
7.80956328e-01 -7.20631182e-01 -3.83421451e-01 -3.33057135e-01
-2.75186539e-01 -2.81508476e-01 1.71471670e-01 -1.91878021e-01
-7.83181071e-01 4.10903990e-03 -6.85903788e-01 -3.61418933e-01
-3.10028881e-01 2.62220234e-01 5.17835431e-02 -3.72509181e-01
7.03555107e-01 -2.91244268e-01 9.29999053e-02 -5.92887163e-01
1.14215776e-01 2.24120736e-01 5.70345104e-01 -2.64018953e-01
5.07106781e-01 5.53104520e-01 -3.08098681e-02 -1.04255033e+00
-7.14477301e-01 -3.46309870e-01 -1.31415856e+00 -1.72619015e-01
1.13088596e+00 -4.40819710e-01 -1.20555125e-01 5.29931664e-01
-1.10324728e+00 -6.83471501e-01 -3.32218379e-01 3.51319075e-01
-5.16483858e-02 -5.76873161e-02 -7.69579887e-01 -4.29191291e-01
-8.20198476e-01 -1.45396125e+00 7.07570493e-01 4.00215924e-01
-6.55199647e-01 -1.00241804e+00 -2.17258379e-01 3.11230510e-01
4.75902200e-01 2.65322059e-01 9.87007022e-01 -8.68538916e-01
-4.93274815e-02 -3.67614686e-01 -3.15307826e-01 2.45235130e-01
-1.06294388e-02 -1.01992957e-01 -1.36470294e+00 3.10161989e-02
9.51425061e-02 1.30198538e-01 9.51566517e-01 8.88229847e-01
1.36174941e+00 1.49995124e-03 -2.47877344e-01 8.00110519e-01
1.04439521e+00 9.39981937e-02 6.79342330e-01 2.22889885e-01
6.09086752e-01 5.34459472e-01 -4.62060452e-01 7.45750070e-02
2.32267231e-01 1.81704685e-01 4.02000844e-01 -1.89050525e-01
-5.20817876e-01 1.38403341e-01 -3.31615686e-01 8.48672271e-01
-1.27115399e-01 7.11662829e-01 -1.30859256e+00 7.11143196e-01
-1.45286572e+00 -6.81985319e-01 -3.67398292e-01 2.06516099e+00
8.51873815e-01 3.98235261e-01 1.45262167e-01 2.45494246e-01
4.36448574e-01 -2.81613559e-01 -5.12777627e-01 -3.01268280e-01
9.15314909e-03 7.93271720e-01 4.18327391e-01 4.23131108e-01
-1.07115662e+00 1.04053771e+00 7.12503242e+00 3.86683285e-01
-1.13253200e+00 4.88507241e-01 7.54083812e-01 -3.30995381e-01
3.78123112e-02 -5.07249296e-01 -7.48086154e-01 3.29692274e-01
8.51072311e-01 2.09169090e-01 1.16299264e-01 7.03036249e-01
2.88010836e-01 -2.54195392e-01 -1.10990560e+00 9.27916527e-01
-1.21455632e-01 -1.72697473e+00 -1.42347753e-01 -5.10592982e-02
8.44513714e-01 5.50252736e-01 -2.03381360e-01 1.18532717e-01
3.10812801e-01 -1.10840821e+00 7.54053056e-01 1.19698815e-01
7.15312839e-01 -5.18255353e-01 8.15109849e-01 1.29296765e-01
-8.37204635e-01 1.32085398e-01 -3.19991350e-01 -1.09766439e-01
3.99583459e-01 7.35108852e-01 -7.35853493e-01 1.30486801e-01
9.79288757e-01 2.30250672e-01 -4.15132821e-01 1.80134809e+00
-2.20583797e-01 5.90158939e-01 -5.93861341e-01 4.75464165e-01
4.51555699e-01 -3.11270624e-01 3.50074440e-01 1.33828199e+00
-3.53439972e-02 2.96890348e-01 -1.83851947e-03 7.18001664e-01
-4.43920881e-01 1.75833002e-01 -6.99557304e-01 5.09422719e-01
2.26573437e-01 1.30766666e+00 -1.42559099e+00 -3.08162153e-01
-2.98100352e-01 5.55270672e-01 4.22085077e-01 6.35083914e-02
-4.18791711e-01 -9.35232565e-02 7.68724024e-01 5.70527494e-01
-1.22653253e-01 -4.19249326e-01 -1.09323311e+00 -8.27007532e-01
-3.99784863e-01 -4.53807414e-01 2.42244318e-01 -5.69045365e-01
-1.05533397e+00 8.69935215e-01 -9.43883508e-02 -4.32217091e-01
1.46506041e-01 -6.27966166e-01 -9.42569554e-01 8.63956869e-01
-1.43968368e+00 -9.51416969e-01 -2.92129338e-01 7.30246007e-01
4.54194158e-01 -4.84957844e-02 8.17023516e-01 5.36308587e-01
-8.44043493e-01 3.22906882e-01 -2.56768107e-01 5.36088288e-01
1.60293072e-01 -9.70938444e-01 4.75212097e-01 1.05073392e+00
-1.34119749e-01 1.05299103e+00 5.34466743e-01 -9.96648908e-01
-8.67449224e-01 -1.16888344e+00 4.18156028e-01 -1.34722367e-01
7.85687745e-01 -2.79323399e-01 -1.10473275e+00 1.06691849e+00
2.61111945e-01 1.62019089e-01 9.32085335e-01 -3.64798419e-02
-2.18680739e-01 2.83714741e-01 -1.65298712e+00 4.95476395e-01
1.06542540e+00 -3.52624923e-01 -9.60823953e-01 2.74888963e-01
3.18574399e-01 -4.50310677e-01 -9.22926247e-01 1.02402240e-01
4.35728878e-01 -8.36313009e-01 8.75827432e-01 -5.67670643e-01
2.30136603e-01 1.06048547e-01 2.95791209e-01 -1.50285542e+00
-4.44813281e-01 -5.20924091e-01 -5.11111990e-02 9.53329504e-01
5.26841223e-01 -5.71334422e-01 1.01532400e+00 1.03887284e+00
-4.77338821e-01 -8.93342912e-01 -1.12030840e+00 -4.91197586e-01
6.11781180e-01 -5.97001553e-01 5.45280099e-01 8.94327044e-01
-2.13201299e-01 2.70039976e-01 4.08536404e-01 -1.17513046e-01
1.07086802e+00 -5.63261211e-01 2.16122016e-01 -1.52420700e+00
3.59526783e-01 -9.00015593e-01 -4.96821284e-01 -5.34603119e-01
1.62731737e-01 -1.40303469e+00 -4.63330448e-01 -2.02964878e+00
-9.19930786e-02 -4.95075047e-01 -9.68023110e-03 7.02841222e-01
1.79443732e-01 6.17649972e-01 -7.02656433e-02 4.74145114e-02
3.64237517e-01 2.68055439e-01 1.23957014e+00 -2.31258079e-01
-2.31197879e-01 4.65195030e-02 -4.90440518e-01 1.19452357e+00
9.49024498e-01 -3.75521600e-01 -4.04403627e-01 -9.22131598e-01
-3.48087817e-01 -3.63896102e-01 1.62360191e-01 -1.04087937e+00
3.55452329e-01 4.86330420e-01 7.50487149e-01 -4.62654412e-01
2.06208736e-01 -9.25649941e-01 -1.91861734e-01 3.68592441e-01
1.61421999e-01 8.74672458e-02 4.00673479e-01 -1.57899126e-01
1.02850907e-01 -2.95868486e-01 1.37726057e+00 -4.88674343e-01
-3.35247546e-01 7.50099301e-01 -4.76888657e-01 -1.56455353e-01
1.09130168e+00 -6.63410664e-01 -4.61419560e-02 1.73877239e-01
-1.26331973e+00 1.72628656e-01 8.57626572e-02 6.90517202e-02
7.52613068e-01 -8.79246354e-01 -5.11364877e-01 4.32209671e-01
-6.66360438e-01 4.65202153e-01 4.32852030e-01 1.17781341e+00
-8.93707454e-01 2.28395864e-01 -7.45705009e-01 -5.22356987e-01
-1.08651543e+00 -1.24431923e-02 7.57255912e-01 -1.20020524e-01
-1.27725816e+00 1.25612819e+00 2.27685437e-01 -4.35402930e-01
1.86532512e-01 -4.32704568e-01 -3.52590382e-01 4.48800147e-01
7.48062015e-01 5.24500191e-01 6.78423226e-01 -5.65270066e-01
-3.77196223e-01 1.86508745e-02 -3.28681618e-01 4.77667712e-02
1.87899244e+00 1.91396490e-01 -3.41063827e-01 -4.16699424e-02
9.45140362e-01 -6.26799464e-01 -1.04663122e+00 -1.75810501e-01
1.89961195e-01 -6.40587732e-02 6.54996753e-01 -5.87363422e-01
-1.71186399e+00 1.01420283e+00 6.23364747e-01 -1.84157625e-01
6.29504025e-01 -3.19102183e-02 7.53532350e-01 8.05561617e-03
4.52306360e-01 -9.97338951e-01 -5.24213374e-01 5.01875758e-01
8.85611653e-01 -8.14265132e-01 -1.74997494e-01 -2.55206406e-01
-2.08120048e-01 1.01982319e+00 7.64727533e-01 -4.71212864e-01
1.24240446e+00 7.42010832e-01 4.83952463e-02 -5.88119626e-01
-7.87848011e-02 -9.00266767e-02 1.21278442e-01 8.16835403e-01
6.87757373e-01 2.40878850e-01 -1.81026772e-01 9.32427764e-01
-5.43840110e-01 -3.72716486e-02 6.82533681e-01 8.82235944e-01
-3.43462497e-01 -6.98697031e-01 -3.05395126e-01 1.19175494e+00
-6.50892496e-01 -2.97417730e-01 -2.87705153e-01 6.69249713e-01
-1.03915082e-02 4.97331291e-01 1.68219626e-01 8.38410854e-02
3.57848316e-01 3.33032101e-01 7.76009858e-01 -9.55948949e-01
-1.25247884e+00 1.63808361e-01 -1.88294902e-01 -4.03321505e-01
9.28360671e-02 -8.86923552e-01 -1.50044537e+00 -2.72566825e-01
-1.27810299e-01 -2.19948292e-01 7.16493130e-01 1.00655150e+00
9.06836465e-02 9.11538005e-01 -1.96380585e-01 -1.35814857e+00
-5.54546081e-02 -1.12248600e+00 -6.95040524e-01 1.54373139e-01
1.25027120e-01 -9.27574098e-01 1.52559534e-01 -1.44061387e-01] | [14.272536277770996, -2.440572500228882] |
23b5d113-74c0-4490-b6a2-093884d6a6d9 | 190412602 | 1904.12602 | null | https://arxiv.org/abs/1904.12602v2 | https://arxiv.org/pdf/1904.12602v2.pdf | EV-Action: Electromyography-Vision Multi-Modal Action Dataset | Multi-modal human action analysis is a critical and attractive research topic. However, the majority of the existing datasets only provide visual modalities (i.e., RGB, depth and skeleton). To make up this, we introduce a new, large-scale EV-Action dataset in this work, which consists of RGB, depth, electromyography (EMG), and two skeleton modalities. Compared with the conventional datasets, EV-Action dataset has two major improvements: (1) we deploy a motion capturing system to obtain high quality skeleton modality, which provides more comprehensive motion information including skeleton, trajectory, acceleration with higher accuracy, sampling frequency, and more skeleton markers. (2) we introduce an EMG modality which is usually used as an effective indicator in the biomechanics area, also it has yet to be well explored in motion related research. To the best of our knowledge, this is the first action dataset with EMG modality. The details of EV-Action dataset are clarified, meanwhile, a simple yet effective framework for EMG-based action recognition is proposed. Moreover, state-of-the-art baselines are applied to evaluate the effectiveness of all the modalities. The obtained result clearly shows the validity of EMG modality in human action analysis tasks. We hope this dataset can make significant contributions to human motion analysis, computer vision, machine learning, biomechanics, and other interdisciplinary fields. | ['Lichen Wang', 'Yun Fu', 'Joseph Robinson', 'Taotao Jing', 'Bin Sun'] | 2019-04-20 | null | null | null | null | ['action-analysis', 'multimodal-activity-recognition', 'electromyography-emg'] | ['computer-vision', 'computer-vision', 'medical'] | [ 1.65490925e-01 -4.75833654e-01 -5.67295492e-01 1.67869702e-01
-6.03435814e-01 6.36168793e-02 2.95432538e-01 -4.55230325e-01
-4.66818690e-01 5.89773715e-01 5.86153924e-01 2.93186218e-01
-1.40074074e-01 -5.41908562e-01 -3.15151781e-01 -8.06478143e-01
5.60588278e-02 -1.18448824e-01 5.62537611e-01 -2.75772482e-01
2.80376554e-01 3.51814896e-01 -1.62371671e+00 7.84808397e-02
6.09892488e-01 9.97220397e-01 2.21345022e-01 3.60999703e-01
3.97531301e-01 4.59932059e-01 -2.34863982e-01 -2.92715877e-02
6.10141829e-02 -6.99425340e-01 -5.92923641e-01 1.10821940e-01
1.98291376e-01 -4.70000565e-01 -5.90438545e-01 7.92527080e-01
9.21976566e-01 3.64299893e-01 5.00788867e-01 -1.20652866e+00
-5.29944897e-01 1.62006676e-01 -7.96829820e-01 2.62763768e-01
6.15815759e-01 6.56970143e-01 5.91451645e-01 -7.02714145e-01
8.45144272e-01 1.06683087e+00 4.36517447e-01 6.46620989e-01
-5.20837963e-01 -5.57096958e-01 -1.18350074e-01 6.77489042e-01
-1.02374744e+00 -2.39040926e-01 1.03014553e+00 -3.54744047e-01
7.84762979e-01 2.89865881e-01 1.14982665e+00 1.42696023e+00
4.99986202e-01 1.15249753e+00 1.03557563e+00 -1.14896670e-01
3.96743976e-02 -6.78281188e-01 -5.12567833e-02 4.97892767e-01
6.38033003e-02 -1.09654129e-03 -7.91255116e-01 3.64947975e-01
9.38939214e-01 2.52071142e-01 -5.40773749e-01 -3.67105067e-01
-1.70353627e+00 2.59080976e-01 3.28211963e-01 4.12141979e-01
-6.01756871e-01 3.65476459e-01 6.04644477e-01 -1.01521879e-01
1.44704208e-01 -1.90091938e-01 -2.39511296e-01 -1.03111029e+00
-5.60367882e-01 2.12480187e-01 1.55051485e-01 6.01777017e-01
3.30222666e-01 1.99710295e-01 -2.48113610e-02 7.61510134e-01
3.52290094e-01 7.33857453e-01 9.07329440e-01 -1.00911438e+00
6.65249944e-01 7.54224658e-01 -2.20628291e-01 -1.10308897e+00
-6.92921400e-01 2.60419965e-01 -7.92644739e-01 1.54797480e-01
2.90482491e-01 9.80912372e-02 -5.85600913e-01 1.42730772e+00
4.70536262e-01 3.04233581e-01 -1.73583999e-01 1.48101783e+00
8.48574579e-01 1.99748293e-01 3.48368213e-02 -2.75329500e-01
1.35865211e+00 -8.99960876e-01 -1.01984680e+00 -9.55638364e-02
4.56065595e-01 -5.65459549e-01 1.32374692e+00 5.39519966e-01
-1.04356813e+00 -7.77945578e-01 -1.11495221e+00 -7.32933953e-02
-7.94841424e-02 4.77790534e-01 8.05868328e-01 3.82185996e-01
-4.51493651e-01 6.00157797e-01 -1.39467752e+00 -6.48761451e-01
3.55043173e-01 6.46254420e-02 -7.50098169e-01 1.38057945e-02
-1.24113452e+00 9.80987251e-01 1.69335112e-01 3.02343309e-01
-7.79801309e-01 -2.88851619e-01 -8.65253627e-01 -6.26211584e-01
4.60645467e-01 -7.41837442e-01 9.33195412e-01 -4.70133811e-01
-1.74552619e+00 6.43905997e-01 -6.37331158e-02 5.75913023e-03
5.23473680e-01 -5.26152372e-01 -4.87909436e-01 5.55411994e-01
-1.58868302e-02 2.27756768e-01 6.85268641e-01 -7.83744574e-01
-3.84965003e-01 -7.94098675e-01 -1.34624690e-01 2.94367015e-01
-5.85063577e-01 -1.25916764e-01 -5.99584818e-01 -8.53584290e-01
3.28237176e-01 -9.69327986e-01 9.46871266e-02 2.40708157e-01
-2.78644174e-01 -1.15018338e-01 8.30638289e-01 -8.56656969e-01
1.41341388e+00 -2.13435864e+00 6.88188195e-01 -1.46592438e-01
1.24232687e-01 3.36681068e-01 1.44144669e-01 3.89396578e-01
1.35070637e-01 -1.01834215e-01 -1.83485985e-01 -8.36032443e-03
-2.17526659e-01 2.42338330e-01 2.64392465e-01 8.46086085e-01
-1.39443249e-01 1.11779737e+00 -8.71342003e-01 -6.35196984e-01
7.56477237e-01 3.88060391e-01 -1.42094806e-01 -5.23201898e-02
2.75973678e-01 7.06109762e-01 -7.61442125e-01 1.07251573e+00
3.01626444e-01 1.07003041e-01 -1.50322497e-01 -6.31224930e-01
-9.01598111e-02 -2.81660974e-01 -1.39072073e+00 2.24947143e+00
2.22498532e-02 4.55219507e-01 -1.14565134e-01 -9.83704567e-01
7.23781884e-01 2.87393034e-01 1.04929507e+00 -8.04593205e-01
3.57982576e-01 2.33212322e-01 -1.05386324e-01 -1.10370469e+00
3.66710544e-01 -5.82065508e-02 2.07181826e-01 2.17432499e-01
-1.56131610e-01 2.72803634e-01 1.61651686e-01 -2.14935541e-01
1.06420243e+00 7.27103710e-01 3.70374471e-01 3.06128383e-01
5.53519666e-01 1.43818958e-02 6.27269566e-01 -2.12116074e-02
-7.11329699e-01 7.79651821e-01 5.38104661e-02 -8.56741667e-02
-8.24183464e-01 -1.07925820e+00 1.06898779e-02 5.00086427e-01
5.79184532e-01 -5.03997803e-01 -5.96320331e-01 -2.56363928e-01
1.63851753e-01 -3.96320783e-02 -5.40698409e-01 -3.79569858e-01
-9.13263083e-01 -6.39685988e-01 6.02142096e-01 1.04597950e+00
8.94967377e-01 -1.23556638e+00 -8.56527507e-01 7.81410858e-02
-4.98231292e-01 -1.16518593e+00 -3.45527887e-01 -6.22707188e-01
-1.30532730e+00 -1.44740725e+00 -1.06283581e+00 -4.90192026e-01
2.11046204e-01 4.99644637e-01 3.66670728e-01 -7.08609372e-02
-4.08129156e-01 6.21877015e-01 -6.34530187e-01 3.03944964e-02
1.48511216e-01 -1.90182447e-01 2.44042888e-01 -4.92234528e-02
2.47579664e-01 -6.64971828e-01 -1.05136693e+00 5.87006152e-01
-8.01445365e-01 8.36569741e-02 8.13823164e-01 6.02777243e-01
7.85663188e-01 -2.12676376e-01 2.32357368e-01 -4.42782231e-02
5.16543746e-01 -3.05629462e-01 1.93643883e-01 6.09836094e-02
-2.94408530e-01 -4.46047485e-01 3.37869495e-01 -5.79430699e-01
-1.05043268e+00 -1.02992453e-01 -2.14031532e-01 -4.38464344e-01
-2.46056065e-01 5.96898079e-01 -3.16793323e-01 -4.48640995e-02
5.29823899e-01 3.96925718e-01 3.25979322e-01 -6.33540571e-01
2.79362291e-01 7.33979166e-01 8.31874073e-01 -5.22642076e-01
3.33539218e-01 7.76038110e-01 3.02147567e-01 -8.59088957e-01
-2.03573808e-01 -7.97792792e-01 -8.19225609e-01 -7.98631191e-01
9.93177950e-01 -6.76644385e-01 -9.09689128e-01 8.83481741e-01
-9.69543993e-01 -1.20503329e-01 -4.97726351e-03 1.20839870e+00
-8.26591969e-01 1.03475142e+00 -7.59037018e-01 -8.57199907e-01
-2.81767428e-01 -1.06031156e+00 1.29103816e+00 1.35134116e-01
-1.94899499e-01 -5.82087576e-01 1.47294283e-01 8.56674254e-01
7.00738281e-02 7.73479342e-01 2.45489508e-01 2.15348899e-01
-3.14828992e-01 -1.62182406e-01 1.33313954e-01 2.58013308e-01
2.27193475e-01 6.98365867e-02 -4.30476308e-01 -1.54069573e-01
-4.82658148e-02 -4.65470165e-01 7.15645373e-01 4.41352427e-01
1.09864366e+00 1.86646834e-01 -3.36492091e-01 5.02039552e-01
1.09352303e+00 2.12799862e-01 1.13034117e+00 7.65286088e-01
9.08203363e-01 5.19709408e-01 1.17545915e+00 5.99013448e-01
3.08281362e-01 9.80114281e-01 5.23395896e-01 3.00816093e-02
-2.82062471e-01 -1.69303343e-01 4.88049775e-01 1.09603715e+00
-9.63988841e-01 3.83703187e-02 -8.87953937e-01 4.12575185e-01
-2.11282778e+00 -1.13101339e+00 -4.83756870e-01 1.92068338e+00
5.86376011e-01 -3.33408639e-02 4.29716974e-01 5.46601415e-01
4.41809863e-01 1.95367530e-01 -7.32582152e-01 2.92899013e-01
-1.70503244e-01 1.96322769e-01 1.71407789e-01 -1.82751536e-01
-1.06135714e+00 6.18425012e-01 5.60967636e+00 8.95591617e-01
-1.09050858e+00 5.71327060e-02 -3.29039365e-01 -2.08795384e-01
1.09518744e-01 -1.67560041e-01 -4.68482524e-01 6.98954761e-01
5.43685734e-01 3.78418155e-02 7.20469579e-02 8.18580806e-01
6.46253467e-01 -3.39821160e-01 -7.88732469e-01 1.20604444e+00
1.96713686e-01 -9.16299939e-01 -1.90274656e-01 1.46600097e-01
2.81125039e-01 -2.08973765e-01 -3.50217730e-01 5.93231395e-02
-6.55772865e-01 -7.35247672e-01 4.22209233e-01 1.01140881e+00
6.84017718e-01 -5.93988478e-01 6.86794937e-01 3.78555357e-01
-1.50387752e+00 1.44283967e-02 -1.43923566e-01 -2.09680825e-01
6.35763884e-01 3.36920500e-01 3.64203393e-01 1.02918136e+00
7.74731696e-01 1.47183776e+00 -4.73204434e-01 1.06133795e+00
-2.94710100e-01 4.90111530e-01 -1.32074356e-01 -3.01562436e-02
-3.64325717e-02 -3.08319628e-01 5.92799664e-01 9.12264228e-01
3.22193503e-01 2.56763518e-01 2.66799599e-01 4.50477153e-01
5.47946334e-01 2.22732350e-01 -5.46555877e-01 -1.55144736e-01
1.79742724e-02 1.24891400e+00 -5.47882617e-01 -1.20227464e-01
-4.63308752e-01 1.11104262e+00 -1.44956276e-01 2.70049691e-01
-9.48294878e-01 -3.19824904e-01 6.19314492e-01 1.15509525e-01
-1.47545040e-01 -6.25012934e-01 -1.68414295e-01 -1.43835437e+00
4.44131374e-01 -8.40444326e-01 4.84131396e-01 -8.96934509e-01
-1.09359777e+00 -1.81940570e-01 2.56628215e-01 -1.80722904e+00
-2.52274930e-01 -8.82138968e-01 -3.90504926e-01 5.33315122e-01
-1.14371991e+00 -1.13014710e+00 -7.34417081e-01 7.48945892e-01
5.44759929e-01 -4.23544757e-02 5.48898935e-01 4.68174607e-01
-8.42051744e-01 8.96358266e-02 -9.27444696e-02 1.75425217e-01
6.41537786e-01 -8.39056432e-01 3.13513279e-02 7.68102109e-01
-1.30828664e-01 7.48104751e-01 2.83225268e-01 -7.66943216e-01
-1.94115186e+00 -4.33453381e-01 2.12412879e-01 -3.74828488e-01
6.79838479e-01 3.04023296e-01 -6.93565428e-01 4.37776953e-01
-1.85471937e-01 4.37772796e-02 6.50849342e-01 -3.93852383e-01
3.81484717e-01 -1.70382813e-01 -8.13657165e-01 6.92122817e-01
1.35577226e+00 -2.84804910e-01 -7.19787121e-01 9.21695307e-02
1.64962173e-01 -6.60908759e-01 -1.29146779e+00 7.63208508e-01
1.14231241e+00 -9.02769327e-01 1.15105271e+00 -3.99829149e-01
7.07525849e-01 -5.37293136e-01 -6.43642843e-02 -8.97570193e-01
-8.66033584e-02 -2.08691999e-01 -6.65617406e-01 1.11373866e+00
-5.04857719e-01 -3.43036354e-01 8.62342894e-01 1.82587028e-01
-2.95361310e-01 -1.15634060e+00 -1.09861422e+00 -8.97321045e-01
-3.54629666e-01 -7.28011310e-01 2.79671997e-01 7.66728342e-01
3.17255139e-01 -7.68171251e-02 -8.46563518e-01 -3.39390725e-01
5.76987505e-01 1.04112431e-01 1.08891177e+00 -7.56720483e-01
-1.69575423e-01 -4.09322500e-01 -9.60336268e-01 -1.23743546e+00
-1.93002850e-01 -4.96651709e-01 -2.23456219e-01 -1.82471943e+00
1.18248068e-01 1.86945871e-01 -3.20674896e-01 3.31302017e-01
-2.50672430e-01 3.83641571e-01 1.10284738e-01 4.92370278e-01
-4.01937127e-01 8.61288786e-01 1.86422157e+00 -1.85724464e-03
-1.49661675e-02 -6.40703142e-02 -1.82205170e-01 8.32152903e-01
8.36683512e-01 -5.26187010e-02 -4.24969941e-01 -1.08779497e-01
-2.36908346e-01 1.77439943e-01 5.84840596e-01 -1.14639425e+00
1.97274864e-01 -5.47238946e-01 4.63908195e-01 -7.57290602e-01
4.72701102e-01 -6.88594699e-01 3.58476400e-01 7.35645115e-01
1.68485001e-01 3.95421451e-03 -1.16201991e-03 5.72323263e-01
-3.57048601e-01 1.69359356e-01 4.61167902e-01 -2.15026915e-01
-1.29606056e+00 4.40410972e-01 -2.08477393e-01 8.32259879e-02
1.28767383e+00 -8.42438459e-01 -2.88440913e-01 -1.06729530e-01
-5.41412890e-01 2.19377190e-01 4.11263466e-01 6.51783884e-01
9.34227884e-01 -1.77633047e+00 -3.78393084e-01 -8.21839944e-02
2.42163450e-01 -2.67550975e-01 5.87074995e-01 1.56263566e+00
-5.04506826e-01 2.63828427e-01 -7.79161870e-01 -7.88179278e-01
-1.34459901e+00 2.47289956e-01 -1.49480172e-03 1.75218940e-01
-9.60549474e-01 7.19580352e-02 -3.02540839e-01 1.68394449e-03
1.85609132e-01 -2.47882009e-01 -5.44950545e-01 6.63368218e-03
4.37251657e-01 1.09708798e+00 -2.64725327e-01 -8.66763830e-01
-6.70920253e-01 1.14850509e+00 6.14320278e-01 -1.84264451e-01
1.11826909e+00 -1.98690280e-01 2.60418244e-02 6.82643056e-01
1.04618955e+00 -1.99214786e-01 -1.17521882e+00 2.36038268e-01
-2.74151951e-01 -7.73104548e-01 -1.96072400e-01 -4.58691686e-01
-1.14737844e+00 1.08675969e+00 7.17040718e-01 -3.01976383e-01
1.37119102e+00 -2.06112400e-01 1.33685374e+00 -3.51094566e-02
6.45125031e-01 -1.42011786e+00 5.52446187e-01 -4.50204685e-02
1.03487539e+00 -1.13953829e+00 3.45043033e-01 -3.54807198e-01
-7.15111434e-01 1.27819359e+00 9.63390410e-01 1.40463367e-01
4.06658649e-01 9.04177222e-03 2.21911864e-03 -1.84990853e-01
-3.28794643e-02 -4.02075052e-01 5.43497622e-01 6.31649554e-01
4.40576166e-01 -9.47544202e-02 -1.06876493e+00 6.76254570e-01
8.87235478e-02 5.54434180e-01 3.30829173e-02 1.25379002e+00
-4.44220543e-01 -1.02520442e+00 -2.71366000e-01 2.61772573e-01
-4.32804734e-01 4.18191344e-01 -2.33116880e-01 1.08937633e+00
1.25850096e-01 8.34082127e-01 -5.12452126e-01 -7.73706615e-01
7.07984209e-01 -2.47890487e-01 6.91169977e-01 -1.59608036e-01
-2.69122183e-01 1.65016167e-02 -9.79613699e-03 -1.10867977e+00
-9.97046411e-01 -7.27636814e-01 -1.45520127e+00 -3.54304194e-01
-1.86309546e-01 -4.18779850e-01 5.76229692e-01 8.83289278e-01
1.75181463e-01 4.77806240e-01 2.36709133e-01 -1.07453048e+00
-3.62602413e-01 -1.06261945e+00 -7.64782906e-01 7.79864728e-01
-6.60270229e-02 -1.25195944e+00 -2.62145519e-01 3.58547419e-02] | [7.826160430908203, 0.38819077610969543] |
f17c42fd-f53e-4665-b5cb-993c77f15c9f | isoex-an-explainable-unsupervised-approach-to | 2306.09260 | null | https://arxiv.org/abs/2306.09260v1 | https://arxiv.org/pdf/2306.09260v1.pdf | IsoEx: an explainable unsupervised approach to process event logs cyber investigation | 39 seconds. That is the timelapse between two consecutive cyber attacks as of 2023. Meaning that by the time you are done reading this abstract, about 1 or 2 additional cyber attacks would have occurred somewhere in the world. In this context of highly increased frequency of cyber threats, Security Operation Centers (SOC) and Computer Emergency Response Teams (CERT) can be overwhelmed. In order to relieve the cybersecurity teams in their investigative effort and help them focus on more added-value tasks, machine learning approaches and methods started to emerge. This paper introduces a novel method, IsoEx, for detecting anomalous and potentially problematic command lines during the investigation of contaminated devices. IsoEx is built around a set of features that leverages the log structure of the command line, as well as its parent/child relationship, to achieve a greater accuracy than traditional methods. To detect anomalies, IsoEx resorts to an unsupervised anomaly detection technique that is both highly sensitive and lightweight. A key contribution of the paper is its emphasis on interpretability, achieved through the features themselves and the application of eXplainable Artificial Intelligence (XAI) techniques and visualizations. This is critical to ensure the adoption of the method by SOC and CERT teams, as the paper argues that the current literature on machine learning for log investigation has not adequately addressed the issue of explainability. This method was proven efficient in a real-life environment as it was built to support a company\'s SOC and CERT | ['Ismail Alaoui Hassani Atlas', 'Pierre Lavieille'] | 2023-06-07 | null | null | null | null | ['anomaly-detection', 'unsupervised-anomaly-detection'] | ['methodology', 'methodology'] | [ 1.63630679e-01 6.29946473e-04 2.15484768e-01 -1.48066714e-01
-1.85788035e-01 -7.67500818e-01 2.99888670e-01 7.62710273e-01
-4.37374003e-02 5.11078000e-01 -1.47746176e-01 -1.10442364e+00
-7.51207471e-01 -5.30636847e-01 -1.51211664e-01 -2.71016151e-01
-4.09493893e-01 2.52345085e-01 -1.84092999e-01 -1.86122939e-01
6.81144834e-01 9.34555650e-01 -1.37882483e+00 3.97740901e-01
4.46956515e-01 9.98847783e-01 -5.27339697e-01 7.65771389e-01
-1.34691730e-01 8.11363518e-01 -1.04650676e+00 1.35081410e-01
3.29278678e-01 -1.74014926e-01 -6.93913460e-01 -1.94314495e-01
-1.18453056e-01 -5.63395023e-02 1.40110955e-01 7.12318599e-01
3.92342079e-03 7.09035853e-03 5.02448261e-01 -1.76298451e+00
1.06058307e-01 2.91324079e-01 -6.18145883e-01 7.95758784e-01
6.76118076e-01 4.19640183e-01 5.70317388e-01 -5.14544785e-01
2.74887353e-01 7.68393099e-01 6.72876418e-01 -1.48656638e-02
-1.07065916e+00 -6.77872300e-01 2.88291514e-01 9.97866541e-02
-8.74773026e-01 1.73968151e-01 7.22220004e-01 -5.30691504e-01
1.53614759e+00 6.62894309e-01 5.80641150e-01 7.64562666e-01
5.05010247e-01 -2.45423913e-02 1.12188160e+00 -6.61315084e-01
3.03615689e-01 2.41883188e-01 5.47944009e-01 3.32171649e-01
6.72936201e-01 1.83450282e-01 -4.34715122e-01 -3.93103898e-01
3.12271684e-01 3.33070904e-01 -9.03942883e-02 6.25490323e-02
-9.28423226e-01 6.65544093e-01 -6.93963915e-02 6.92915797e-01
-4.82323855e-01 -2.35974640e-01 7.99815774e-01 5.41491389e-01
1.97571665e-01 9.68170762e-01 -5.14957309e-01 -6.29407465e-01
-6.05294764e-01 -1.41487330e-01 9.19559896e-01 4.60035473e-01
2.42137983e-01 4.93439019e-01 4.58747506e-01 -8.63555595e-02
2.68495232e-01 -1.00846536e-01 5.34975529e-02 -2.68867284e-01
2.74628103e-01 1.19321156e+00 7.34354556e-03 -1.14275694e+00
-4.94913667e-01 -3.06193471e-01 -5.18614054e-01 8.77632320e-01
3.01474661e-01 -1.84453756e-01 -8.37171674e-01 9.33371007e-01
1.74135879e-01 -1.22085445e-01 -1.43181324e-01 4.87077326e-01
-7.68178888e-03 5.60596824e-01 9.11986604e-02 -9.95813683e-02
1.13741100e+00 -3.35880592e-02 -8.41709495e-01 -2.93943891e-03
8.85148704e-01 -7.10283101e-01 9.46219027e-01 1.14561594e+00
-7.01494753e-01 -2.64414191e-01 -1.53392315e+00 8.76263976e-01
-9.36425805e-01 -4.55814242e-01 6.47311389e-01 8.98282468e-01
-4.16713476e-01 6.67360008e-01 -8.76088142e-01 -4.59069550e-01
9.61864591e-02 2.92879909e-01 -3.93991888e-01 2.17709050e-01
-8.44872952e-01 1.16835737e+00 6.32005394e-01 1.15038283e-01
-4.26286817e-01 -6.69285834e-01 -6.56287014e-01 8.57966617e-02
3.81312788e-01 7.84271806e-02 7.37694323e-01 -4.68483716e-01
-5.54078102e-01 3.98096621e-01 3.91116887e-01 -6.17665112e-01
2.79048324e-01 -3.07015419e-01 -9.52476025e-01 1.08212806e-01
-1.65873736e-01 -3.80324304e-01 6.06306076e-01 -1.07510376e+00
-8.36979151e-01 -5.01755595e-01 9.63308960e-02 -5.05814075e-01
-1.97135225e-01 4.84770775e-01 5.21575868e-01 -4.58894908e-01
2.91704349e-02 -7.30796993e-01 3.11315898e-02 -5.56608498e-01
-6.05803967e-01 1.28916487e-01 1.52774894e+00 -7.80538380e-01
1.60535228e+00 -2.24581409e+00 -6.32366300e-01 7.95442402e-01
2.46664643e-01 5.01997590e-01 7.17217863e-01 1.00552368e+00
-8.08207154e-01 4.02541608e-01 -2.66227633e-01 2.98537493e-01
1.43386927e-02 2.44699210e-01 -6.48236275e-01 4.54403341e-01
3.27497661e-01 1.31563574e-01 -6.76251709e-01 6.58672005e-02
5.38249969e-01 1.07522547e-01 -1.41832829e-02 1.81847438e-01
1.75462320e-01 5.56619406e-01 -2.26553604e-01 8.40890229e-01
3.87754500e-01 5.69875017e-02 2.99328994e-02 1.28394142e-01
-5.36672771e-01 1.15822271e-01 -1.20344234e+00 9.24105406e-01
-2.01153383e-01 8.32098305e-01 -1.19825885e-01 -1.03234804e+00
1.05295146e+00 6.36224985e-01 6.49169445e-01 -4.69602913e-01
5.21044508e-02 1.88569516e-01 2.19904944e-01 -6.27682030e-01
1.56109244e-01 1.35731861e-01 -8.46384913e-02 9.60874438e-01
-4.11184967e-01 1.39936596e-01 3.71472016e-02 4.15198386e-01
1.56441641e+00 -1.33748636e-01 5.26020527e-01 -1.72912300e-01
5.39267004e-01 3.90367985e-01 4.01453584e-01 4.52949166e-01
-1.52260363e-01 -2.77442206e-02 7.35480189e-01 -8.94267678e-01
-9.55733955e-01 -8.60259175e-01 -2.96929311e-02 5.31805813e-01
-2.76518106e-01 -4.56666619e-01 -6.41160846e-01 -9.45591986e-01
-1.07603110e-01 1.37113988e+00 -6.37581646e-01 -3.49030048e-01
-5.62761128e-01 -4.38623190e-01 5.21255374e-01 6.14850402e-01
4.51488979e-02 -1.02322459e+00 -1.36971962e+00 2.91421205e-01
3.62958550e-01 -8.22750151e-01 1.61576658e-01 6.88410282e-01
-9.23850179e-01 -1.52158117e+00 4.96636391e-01 1.36789620e-01
7.75385141e-01 -3.78276408e-02 9.45456386e-01 3.64933789e-01
-9.47393179e-01 7.49638736e-01 -6.36457622e-01 -1.13298523e+00
-5.75741827e-01 -4.44324315e-01 1.88842133e-01 -2.35255569e-01
7.35299170e-01 -6.33616686e-01 -3.29450279e-01 1.79176867e-01
-1.15598142e+00 -4.86788511e-01 4.54371423e-01 5.85389376e-01
6.85464367e-02 6.13651812e-01 6.73941910e-01 -7.37859547e-01
7.84753203e-01 -6.64704978e-01 -4.79091108e-01 1.67754725e-01
-1.22299194e+00 -1.57130733e-01 5.13322413e-01 -1.59700677e-01
-9.08244133e-01 -3.26260656e-01 2.43573040e-01 -1.81341574e-01
-6.26019478e-01 6.11852050e-01 1.78388357e-02 -6.63212612e-02
7.78345048e-01 -1.23211183e-01 9.19174701e-02 -4.50086474e-01
-3.30644667e-01 6.08106077e-01 5.20469606e-01 -2.40923852e-01
1.05575395e+00 2.85664588e-01 1.48126185e-01 -6.73207581e-01
-2.62438029e-01 -5.66586494e-01 -7.16636360e-01 -4.83766675e-01
6.23799801e-01 -1.29014835e-01 -9.25226867e-01 2.67938748e-02
-1.08029580e+00 1.60595924e-01 -2.53094286e-01 4.43517119e-01
-6.77064210e-02 2.71683186e-01 -1.50514811e-01 -1.33796811e+00
-3.49111229e-01 -6.72898710e-01 4.04726028e-01 -9.03862994e-04
-8.04097176e-01 -1.14288723e+00 -4.06034440e-02 1.60349250e-01
4.40657556e-01 1.06061733e+00 1.28694403e+00 -1.39526308e+00
-3.30814451e-01 -8.50264370e-01 -7.14636445e-02 4.31463540e-01
3.64254981e-01 3.92951608e-01 -8.37340891e-01 -4.85814333e-01
2.54627466e-01 1.41702265e-01 -1.21882007e-01 -1.54604167e-01
9.21448946e-01 -3.32187384e-01 -2.38208547e-01 2.57026609e-02
1.51176035e+00 1.02461004e+00 4.62999851e-01 7.47313023e-01
3.04022610e-01 7.97479212e-01 1.01079023e+00 7.17124701e-01
-4.26761150e-01 3.92683119e-01 7.10561931e-01 -1.87575057e-01
4.68017995e-01 -1.03196660e-02 2.29880780e-01 3.37863654e-01
-1.73182338e-01 8.29173997e-02 -1.40297127e+00 2.24861190e-01
-1.54435563e+00 -1.01306760e+00 -3.36950272e-01 2.37992787e+00
2.86574177e-02 7.53921092e-01 8.94769430e-02 1.00836611e+00
4.61491972e-01 -3.89245898e-01 -4.24100369e-01 -1.20713270e+00
3.81605595e-01 7.39656314e-02 2.56255955e-01 1.15801878e-01
-8.00083101e-01 -1.53752347e-03 6.10514641e+00 1.91812575e-01
-9.64019954e-01 -2.93438971e-01 5.77880025e-01 1.45894766e-01
1.10956788e-01 1.32180139e-01 -3.96210760e-01 4.83784109e-01
1.27215397e+00 -2.20961690e-01 1.11073725e-01 8.80453169e-01
5.03169239e-01 -3.02494079e-01 -1.12691498e+00 6.41249478e-01
-4.79162075e-02 -1.11428738e+00 -3.51598680e-01 2.81722307e-01
2.77084205e-02 -5.83927810e-01 -3.18393894e-02 3.37084457e-02
-9.14358646e-02 -1.16020286e+00 4.51525897e-01 6.13548756e-01
3.24394494e-01 -1.10844350e+00 1.00353825e+00 2.66583055e-01
-9.43410277e-01 -6.23039603e-01 3.65929186e-01 -5.14505148e-01
2.79851913e-01 3.78775895e-01 -1.42489433e+00 7.05858767e-01
9.48731065e-01 5.31658530e-02 -4.19046611e-01 9.36632752e-01
4.28921878e-02 8.32320035e-01 -2.88476944e-01 1.88244388e-01
4.00380373e-01 9.14112478e-03 7.93417156e-01 1.17464018e+00
2.19773352e-01 2.48689707e-02 -4.31148969e-02 5.66024899e-01
1.00853384e+00 -2.16844872e-01 -9.30626869e-01 -3.12622994e-01
4.73539174e-01 1.19930100e+00 -9.86696959e-01 2.16829067e-04
-3.77773046e-01 3.90483826e-01 -4.43613410e-01 1.62547663e-01
-7.23185301e-01 -8.05682480e-01 5.85657775e-01 5.35380840e-01
-1.24856003e-01 -2.99931794e-01 -6.90146208e-01 -4.04155821e-01
1.35971680e-01 -1.21721280e+00 6.76645696e-01 -5.56047976e-01
-1.21984112e+00 7.92566895e-01 3.50283742e-01 -1.42281055e+00
-5.86049557e-01 -8.72628033e-01 -1.09422910e+00 8.30333889e-01
-6.91004038e-01 -9.56053555e-01 -1.85395703e-01 5.35590351e-01
3.75436574e-01 -3.67256969e-01 1.02738321e+00 7.76074678e-02
-5.91742873e-01 2.19023749e-01 -2.75116324e-01 1.02620527e-01
2.72427380e-01 -1.46249688e+00 4.21816587e-01 1.13055027e+00
4.39440273e-02 8.03401828e-01 9.94336486e-01 -8.41126323e-01
-1.23214114e+00 -4.52733517e-01 6.62683725e-01 -7.24190176e-01
9.65316772e-01 -3.20899755e-01 -1.01889968e+00 7.42921829e-01
3.01140249e-01 -3.80239993e-01 1.12810576e+00 6.50904179e-02
-9.07983333e-02 1.04151964e-02 -1.26192129e+00 3.13365579e-01
4.08950657e-01 -3.74656707e-01 -7.92514086e-01 2.83418357e-01
2.32781738e-01 -5.76306395e-02 -9.10763681e-01 4.10006195e-01
2.20863700e-01 -1.28690016e+00 6.82165504e-01 -9.16846395e-01
-3.71312141e-01 -4.10295218e-01 2.06041291e-01 -9.68974531e-01
1.16436899e-01 -8.87712300e-01 -1.74397469e-01 1.12067199e+00
4.63026226e-01 -1.01008296e+00 5.91737807e-01 1.03806663e+00
-3.30346525e-01 -8.34452152e-01 -9.64072347e-01 -8.74398530e-01
-4.97901827e-01 -8.61031950e-01 5.97910702e-01 1.04460800e+00
4.36955631e-01 -2.24443704e-01 -1.86825499e-01 3.93031329e-01
4.53881890e-01 -2.05696538e-01 7.03198731e-01 -1.45754373e+00
-1.32136002e-01 -1.62915289e-01 -8.74137819e-01 3.25551599e-01
-5.26901126e-01 -1.51557162e-01 -4.89566118e-01 -1.12438440e+00
-2.40309656e-01 -3.35848451e-01 -4.45074916e-01 7.42926300e-01
2.49012679e-01 -1.73704028e-01 1.90250382e-01 1.25849411e-01
-7.43997246e-02 -3.72516006e-01 4.25315291e-01 7.45112449e-02
-2.56957591e-01 1.70146570e-01 -5.55625260e-01 8.01608086e-01
1.08140552e+00 -6.96291089e-01 -5.29155135e-01 2.84499913e-01
3.91114473e-01 -1.40560698e-03 4.62570906e-01 -1.16582501e+00
2.91472822e-01 -1.08584285e-01 6.33654773e-01 -8.70741487e-01
9.63555127e-02 -1.33491874e+00 4.57259119e-01 7.97015488e-01
-1.40454946e-02 9.30985808e-01 6.03326976e-01 4.68339324e-01
-4.06195253e-01 -3.36009443e-01 3.67756277e-01 1.39685929e-01
-5.23353815e-01 -1.50660515e-01 -6.49971485e-01 -4.62160736e-01
1.59615254e+00 -7.68598557e-01 -2.93499291e-01 -3.82526308e-01
-7.85632432e-01 -9.02751461e-02 3.15782279e-01 3.06369752e-01
6.96797848e-01 -7.62742400e-01 -1.85986355e-01 4.13859755e-01
7.35881403e-02 -5.19284189e-01 1.67828590e-01 1.01335883e+00
-7.34975338e-01 5.29375911e-01 -5.36587417e-01 -2.62589246e-01
-1.45803654e+00 7.78923213e-01 1.81242768e-02 -3.17738831e-01
-8.22122037e-01 2.73070276e-01 -5.31097233e-01 2.49870326e-02
3.04871351e-01 -1.01848386e-01 -3.14564586e-01 -3.19235735e-02
7.44360745e-01 7.06845701e-01 2.97089696e-01 -2.29498148e-01
-4.74715024e-01 8.75323862e-02 -3.77330691e-01 -4.25845459e-02
1.27761555e+00 1.97027341e-01 -9.80644599e-02 5.74969172e-01
6.49632156e-01 -4.13186662e-03 -8.80724788e-01 4.54302162e-01
6.56907141e-01 -6.91125989e-01 -2.10405827e-01 -1.22525930e+00
-6.03712261e-01 9.38102782e-01 7.84693062e-01 1.06429505e+00
1.31887877e+00 -3.60642612e-01 3.35131466e-01 3.77117068e-01
2.90366560e-01 -1.20938897e+00 1.36084586e-01 2.32117862e-01
7.31213629e-01 -9.90545034e-01 2.00747818e-01 -2.01425567e-01
-6.40107632e-01 1.50522053e+00 5.01111269e-01 1.72646284e-01
5.48020244e-01 6.13940537e-01 1.06471017e-01 -6.13829374e-01
-5.86904168e-01 4.59603220e-01 -3.48633081e-02 8.32348347e-01
3.88215274e-01 -8.24580267e-02 -1.28690735e-01 4.59622830e-01
6.99074492e-02 -3.38966578e-01 4.84864146e-01 1.45754254e+00
-4.49693948e-01 -1.01171410e+00 -8.99457932e-01 5.93383908e-01
-5.83838463e-01 1.37247786e-01 -6.03010774e-01 1.33890772e+00
1.62102118e-01 1.26075351e+00 6.88926205e-02 -5.67845345e-01
4.10604030e-01 2.33453706e-01 -2.01733246e-01 -4.25359219e-01
-8.65508437e-01 -3.51310104e-01 2.13839605e-01 -7.05471396e-01
2.06379861e-01 -6.39693081e-01 -1.41817594e+00 -4.67339128e-01
-1.60655320e-01 4.98795569e-01 1.12097859e+00 8.88517976e-01
3.19731951e-01 7.96591401e-01 6.76832199e-01 -4.26756799e-01
-4.72875774e-01 -8.34683716e-01 -6.27875805e-01 3.58763397e-01
6.79147691e-02 -7.06262708e-01 -7.26761043e-01 -1.61795765e-01] | [5.440705299377441, 7.161055088043213] |
a66d5c80-decf-4bd6-be96-e47bc85bb892 | object-based-slam-utilizing-unambiguous-pose | 2303.07872 | null | https://arxiv.org/abs/2303.07872v1 | https://arxiv.org/pdf/2303.07872v1.pdf | Object-based SLAM utilizing unambiguous pose parameters considering general symmetry types | Existence of symmetric objects, whose observation at different viewpoints can be identical, can deteriorate the performance of simultaneous localization and mapping(SLAM). This work proposes a system for robustly optimizing the pose of cameras and objects even in the presence of symmetric objects. We classify objects into three categories depending on their symmetry characteristics, which is efficient and effective in that it allows to deal with general objects and the objects in the same category can be associated with the same type of ambiguity. Then we extract only the unambiguous parameters corresponding to each category and use them in data association and joint optimization of the camera and object pose. The proposed approach provides significant robustness to the SLAM performance by removing the ambiguous parameters and utilizing as much useful geometric information as possible. Comparison with baseline algorithms confirms the superior performance of the proposed system in terms of object tracking and pose estimation, even in challenging scenarios where the baseline fails. | ['H. Jin Kim', 'Youngseok Jang', 'Taekbeom Lee'] | 2023-03-13 | null | null | null | null | ['simultaneous-localization-and-mapping'] | ['computer-vision'] | [-2.89814845e-02 -3.38651776e-01 1.36121854e-01 -2.75870740e-01
-4.51784551e-01 -9.22398210e-01 6.03898346e-01 8.56166631e-02
-5.68341136e-01 6.37550354e-01 -2.77728230e-01 9.43583772e-02
-4.31320816e-01 -3.22166979e-01 -6.58121645e-01 -9.77749944e-01
1.29838631e-01 8.73153448e-01 5.24618208e-01 8.62288699e-02
4.21959609e-01 9.24278796e-01 -1.70281458e+00 -4.84713495e-01
7.75660694e-01 8.23321462e-01 8.38749468e-01 3.46838832e-01
2.02136906e-03 1.75381869e-01 -5.89620411e-01 -1.05048172e-01
5.69974244e-01 1.59711823e-01 -3.12677503e-01 1.66725174e-01
8.30630004e-01 -1.79397747e-01 5.44593250e-03 1.21437287e+00
3.58239293e-01 4.49789204e-02 4.70143199e-01 -1.32261348e+00
7.06743151e-02 -1.23363152e-01 -5.95169723e-01 -9.60699841e-02
3.29074740e-01 -2.85055965e-01 7.17341602e-01 -9.19661522e-01
5.13738275e-01 1.07855475e+00 7.17263520e-01 6.85719401e-02
-1.11830795e+00 -6.59691393e-01 1.00856461e-01 1.79862216e-01
-1.95504606e+00 -5.64894080e-01 7.04474866e-01 -5.15075207e-01
3.01040351e-01 2.92411029e-01 3.06250751e-01 5.63630342e-01
2.19968945e-01 6.27612025e-02 7.15348125e-01 -3.82441550e-01
-2.12927572e-02 3.59673947e-01 4.93842959e-02 5.68410397e-01
9.27574158e-01 -1.81189999e-01 -4.03898865e-01 -1.53712556e-01
5.20761967e-01 1.42125666e-01 -1.47170156e-01 -1.19912791e+00
-1.58781767e+00 5.85071146e-01 4.38792080e-01 7.59397671e-02
-3.18443000e-01 -5.92737980e-02 -4.11657654e-02 2.57980078e-02
-1.51763961e-01 6.20004535e-01 -2.79928893e-01 3.01255047e-01
-5.45219541e-01 3.52898300e-01 4.67498094e-01 1.47407603e+00
8.49162102e-01 6.74426556e-04 2.82465518e-01 5.94777822e-01
5.06980777e-01 9.86767411e-01 -1.68739818e-02 -7.55489111e-01
5.95987678e-01 6.34991348e-01 7.22698569e-01 -1.26182079e+00
-6.38537943e-01 -4.44960177e-01 -5.35762608e-01 1.43826723e-01
3.46442491e-01 5.64729497e-02 -6.22011304e-01 1.69156361e+00
5.71351290e-01 -3.18192333e-01 1.50682911e-01 1.07530856e+00
6.25068665e-01 3.04559976e-01 -4.04872566e-01 -2.13270232e-01
1.35411835e+00 -7.45134413e-01 -7.86304116e-01 -3.93388122e-01
2.87727386e-01 -1.23955321e+00 3.51295531e-01 2.15313718e-01
-5.71403980e-01 -5.59672952e-01 -1.14296067e+00 2.35871255e-01
-2.37928778e-01 4.63315874e-01 5.32538831e-01 4.16983157e-01
-8.17594171e-01 -1.57889333e-02 -7.12071300e-01 -6.56438112e-01
-3.03218812e-01 8.27994525e-01 -5.65465510e-01 3.06457523e-02
-5.85721135e-01 1.27028322e+00 5.94553649e-01 3.87937158e-01
-6.71376109e-01 -1.54358447e-01 -7.91102588e-01 -2.91310042e-01
6.73047483e-01 -4.76853788e-01 7.69901872e-01 -4.89244580e-01
-1.03889418e+00 6.21782124e-01 -4.03339207e-01 -1.40796110e-01
6.32590950e-01 -3.61801356e-01 -3.11695129e-01 -7.22189471e-02
3.00118178e-01 5.54803610e-01 6.62024200e-01 -1.44365585e+00
-8.70117188e-01 -6.18389964e-01 -4.79647629e-02 5.96760869e-01
1.07958511e-01 -1.73345461e-01 -7.83076823e-01 -9.20525268e-02
1.05389082e+00 -1.27793956e+00 -1.10623367e-01 8.84159580e-02
-3.03163737e-01 1.11876063e-01 1.11353195e+00 -3.03880066e-01
6.41215026e-01 -1.94499040e+00 2.30025977e-01 2.30954066e-01
-1.97167203e-01 -3.90999913e-02 7.20411539e-02 4.42000061e-01
3.11920643e-01 -3.82147640e-01 2.93798894e-01 -3.82063240e-01
-2.38206610e-01 4.02886748e-01 -6.09168150e-02 1.00737751e+00
-1.93564475e-01 3.89835030e-01 -6.49086833e-01 -5.55937529e-01
4.75967944e-01 3.58157903e-01 -1.39470845e-01 2.51833022e-01
1.98224306e-01 7.35459208e-01 -4.58025396e-01 7.37600684e-01
1.05574727e+00 9.32588354e-02 2.60368884e-01 -5.65187454e-01
-3.87543827e-01 1.50472403e-01 -1.88776183e+00 1.31920934e+00
-3.36691976e-01 5.54290712e-01 2.59500623e-01 -5.40600359e-01
1.11726451e+00 1.73219547e-01 5.85777462e-01 -3.02193433e-01
4.44766991e-02 2.96849221e-01 2.38105983e-01 -3.78843337e-01
8.79148424e-01 1.50039256e-01 -3.82669456e-02 3.93571071e-02
5.07337078e-02 -1.65878236e-01 8.33058953e-02 -1.96848974e-01
5.18409431e-01 3.99879664e-02 5.39680243e-01 -4.99412239e-01
7.04057693e-01 -1.46485016e-01 6.12654626e-01 8.25816512e-01
-4.72003743e-02 3.52234721e-01 -1.75364986e-01 -2.67636836e-01
-1.13519943e+00 -1.23010027e+00 -4.49226707e-01 6.68779433e-01
1.15189910e+00 -4.28435812e-03 9.23051871e-03 -4.42357212e-01
3.77966404e-01 2.67404348e-01 -2.45843098e-01 5.86542971e-02
-5.03420115e-01 -5.89461923e-01 3.39400657e-02 1.07960150e-01
3.42638671e-01 -3.95173818e-01 -7.60713339e-01 -8.69689733e-02
-3.51440251e-01 -1.35399759e+00 -2.60553509e-01 1.69102505e-01
-7.68759072e-01 -1.27100182e+00 -2.64997959e-01 -7.00886428e-01
1.15041184e+00 9.82631624e-01 5.83917677e-01 1.12158902e-01
-1.73082054e-02 2.78099716e-01 -1.95126429e-01 -3.82871002e-01
-2.59643495e-01 -6.24036379e-02 5.81503093e-01 2.78292596e-01
8.21020529e-02 -2.19824687e-01 -3.60708773e-01 9.22502637e-01
-3.39388490e-01 -9.21081007e-02 4.86495197e-01 4.63709563e-01
3.72786164e-01 1.25298440e-01 8.13042521e-02 -3.81420046e-01
-6.00084849e-02 -1.53773576e-01 -1.14106703e+00 2.34029084e-01
-5.89836240e-01 7.85535052e-02 3.64472568e-01 -3.02039683e-01
-9.21302140e-01 4.96798933e-01 4.64044780e-01 -2.88720310e-01
-1.12548657e-01 5.97212799e-02 -3.90675187e-01 -8.40929568e-01
3.91727448e-01 1.18005477e-01 -5.56773543e-02 -6.81285024e-01
1.45985894e-02 5.07509053e-01 5.51554203e-01 -3.94080013e-01
1.20921004e+00 7.47499764e-01 6.13404155e-01 -9.31105733e-01
-6.27633989e-01 -1.05297160e+00 -1.07935250e+00 -2.45780081e-01
6.90133929e-01 -9.25277174e-01 -6.84055328e-01 4.57383484e-01
-1.31953740e+00 7.50600278e-01 2.64933050e-01 9.16792572e-01
-2.64402062e-01 6.75553083e-01 6.25139996e-02 -1.07531154e+00
1.99865073e-01 -1.51146042e+00 1.14807212e+00 1.67732731e-01
6.94546402e-02 -6.41985893e-01 -2.38974363e-01 1.60495773e-01
1.01735942e-01 2.21273109e-01 4.01128352e-01 -5.83768487e-01
-1.02881050e+00 -4.23628479e-01 -1.15107603e-01 -1.10196918e-01
2.70862937e-01 -4.38483804e-02 -7.04836369e-01 -7.94287622e-01
3.05815637e-02 1.89319611e-01 3.82498115e-01 7.80917183e-02
3.58044147e-01 -7.20770806e-02 -6.39371991e-01 6.96761727e-01
1.71164393e+00 4.10837382e-01 2.12099731e-01 5.80915332e-01
8.02739799e-01 6.57836616e-01 1.02728152e+00 2.97839433e-01
2.72081554e-01 1.25446486e+00 9.08578992e-01 9.65585709e-02
3.81473333e-01 6.10376708e-03 1.73084214e-01 5.61035812e-01
-4.76568341e-02 -1.47635922e-01 -8.31586897e-01 5.09949088e-01
-1.93228590e+00 -6.59979224e-01 -4.16662812e-01 2.64942598e+00
9.81825218e-03 -6.33555502e-02 -2.06927568e-01 -2.01848984e-01
9.75225925e-01 3.82658802e-02 -2.32825086e-01 2.50584155e-01
-1.92799628e-01 -5.69858968e-01 1.06325483e+00 6.16390824e-01
-1.09887624e+00 7.30659485e-01 6.11770678e+00 3.25799674e-01
-1.05575275e+00 -1.62968412e-02 -6.45170987e-01 9.82335731e-02
2.18349710e-01 3.82693648e-01 -1.37363613e+00 4.53402251e-01
2.39809185e-01 5.85821681e-02 9.69246775e-02 7.82961428e-01
-4.76576388e-02 -5.96618652e-01 -1.05640244e+00 1.08914304e+00
3.45954061e-01 -9.87014234e-01 1.95720121e-02 1.66348457e-01
5.79876006e-01 1.08837457e-02 -8.27758387e-02 -2.20993429e-01
-9.39758718e-02 -4.21573311e-01 9.56834435e-01 3.84825170e-01
2.21434444e-01 -6.29285693e-01 1.16458619e+00 6.09763741e-01
-1.29879916e+00 -1.16594642e-01 -5.42610586e-01 5.96431550e-03
2.21582368e-01 1.22378923e-01 -1.28531802e+00 9.94109750e-01
6.10842705e-01 2.99699903e-01 -7.65250146e-01 1.43246627e+00
-1.95545852e-02 -2.27624491e-01 -6.81475580e-01 1.72193870e-01
-2.11183229e-04 -5.16086757e-01 9.13294673e-01 7.45094836e-01
6.46230340e-01 -2.16678068e-01 4.98932987e-01 5.59338927e-01
3.90638053e-01 8.87663662e-02 -6.27236962e-01 5.58228970e-01
8.03277731e-01 1.17480981e+00 -8.27347815e-01 -1.04772493e-01
-5.34289658e-01 8.02926183e-01 1.26015455e-01 2.08965957e-01
-8.39289546e-01 -7.45607466e-02 5.24145246e-01 4.87278104e-02
4.39214230e-01 -6.07228935e-01 -9.32419449e-02 -1.14474678e+00
3.83322030e-01 -4.88543451e-01 2.61429012e-01 -7.60377765e-01
-7.62681723e-01 5.30317366e-01 2.40759984e-01 -1.62499189e+00
-9.06434879e-02 -6.24750257e-01 -5.14394455e-02 7.44274080e-01
-1.28027308e+00 -1.34927833e+00 -5.85585892e-01 5.16355395e-01
2.23304763e-01 -1.04320422e-01 4.86522496e-01 3.34451377e-01
-2.08930701e-01 3.53104562e-01 4.07804608e-01 -2.33226091e-01
9.52317834e-01 -1.00009930e+00 -3.80261898e-01 1.09497666e+00
2.69359559e-01 7.83320785e-01 1.08566177e+00 -7.59522080e-01
-1.58821952e+00 -8.62439632e-01 8.10144365e-01 -6.44837618e-01
3.71220738e-01 -6.18611753e-01 -5.20709813e-01 6.66537404e-01
-3.35745364e-01 -2.30442807e-01 1.94381580e-01 1.52844489e-01
-7.99891874e-02 -3.34926695e-01 -1.07520485e+00 4.06443447e-01
7.12588310e-01 -3.08651567e-01 -6.24927461e-01 5.06163716e-01
4.55249369e-01 -6.58757687e-01 -5.83978117e-01 5.95030487e-01
5.32907009e-01 -9.34851587e-01 1.05629802e+00 -4.43092585e-02
-6.94573462e-01 -1.03949380e+00 -4.97258663e-01 -8.33336532e-01
-2.85656691e-01 -3.86346221e-01 2.63848662e-01 1.23254478e+00
6.16700053e-02 -7.48250306e-01 6.31970227e-01 1.84858918e-01
-8.17712396e-02 5.98614328e-02 -1.07879829e+00 -1.12104142e+00
-7.97914445e-01 -2.40572374e-02 6.06509924e-01 8.02470386e-01
-6.48790002e-01 2.13193595e-01 -7.68704057e-01 1.10644448e+00
8.90613139e-01 5.04465163e-01 1.35097075e+00 -1.62469149e+00
-1.15079299e-01 2.19015181e-02 -8.26949239e-01 -1.06230009e+00
-2.35346165e-02 -5.31534314e-01 3.34293365e-01 -1.20589364e+00
2.19907656e-01 -6.47665739e-01 -5.80364652e-02 1.48702428e-01
4.00296785e-02 2.99654037e-01 3.85614902e-01 6.15653515e-01
-6.07256949e-01 3.05028588e-01 8.38035762e-01 1.59896985e-01
-4.27603684e-02 2.94787824e-01 -1.35277137e-01 8.41203392e-01
3.00072849e-01 -5.41050732e-01 -1.23361535e-01 -6.26215816e-01
4.84510623e-02 -4.00653109e-02 4.65846837e-01 -1.20559835e+00
4.42248970e-01 -2.31166750e-01 3.46371412e-01 -1.09523451e+00
6.40140593e-01 -1.63642848e+00 6.76626861e-01 5.14332771e-01
2.51585841e-01 1.79474816e-01 1.06905058e-01 5.35106719e-01
-1.63329974e-01 -5.76499283e-01 8.64306152e-01 5.06089106e-02
-1.00853097e+00 1.42055348e-01 2.22220439e-02 -6.73944592e-01
1.27308977e+00 -4.71454203e-01 -2.05693662e-01 -4.32725698e-01
-5.27887404e-01 2.87518144e-01 9.15111959e-01 5.30606568e-01
1.91556275e-01 -1.36969078e+00 -3.34880352e-01 3.51678938e-01
4.27803218e-01 1.05938591e-01 1.27473667e-01 1.09865248e+00
-6.45227492e-01 6.60619318e-01 -3.20526630e-01 -1.13965583e+00
-1.60535586e+00 6.34438157e-01 2.83602387e-01 9.43515599e-02
-2.18636110e-01 3.95789593e-01 4.24897045e-01 -4.73596692e-01
2.41908610e-01 3.09333969e-02 -1.42102316e-01 4.87540103e-03
1.27205580e-01 4.38790649e-01 1.01972051e-01 -1.21455371e+00
-7.53956497e-01 1.13832545e+00 4.88905422e-02 1.70290723e-01
1.05153108e+00 -6.56226158e-01 -2.50225842e-01 4.51549232e-01
9.21471715e-01 5.19135773e-01 -1.09492600e+00 -3.26930434e-01
1.16321310e-01 -8.14629614e-01 -2.34570667e-01 -3.95760298e-01
-6.64654493e-01 6.15329564e-01 7.75977433e-01 -1.24419488e-01
7.01172829e-01 2.48299967e-02 9.63417068e-02 6.56522274e-01
8.80653560e-01 -8.67593527e-01 -2.46038899e-01 5.08805096e-01
7.32305884e-01 -1.34123600e+00 4.44718003e-01 -6.11185670e-01
-3.56299281e-01 1.10941601e+00 7.92359591e-01 -1.36747703e-01
3.52166109e-02 5.21437079e-02 1.14722386e-01 -8.24427456e-02
-1.02910921e-01 -2.05235347e-01 4.65153813e-01 6.63444877e-01
-5.16484901e-02 -7.47008100e-02 -2.84562886e-01 -1.16931006e-01
-1.40126437e-01 -7.15127409e-01 4.52325523e-01 9.55194235e-01
-6.94594383e-01 -9.61261332e-01 -9.08082902e-01 -4.10248637e-02
-1.56272277e-01 2.68471181e-01 -2.14831665e-01 1.25567818e+00
2.99694151e-01 7.12679207e-01 1.46359056e-01 -1.46945015e-01
4.24382776e-01 -7.84112886e-02 6.45267904e-01 -3.67117256e-01
-1.26370668e-01 3.00374061e-01 -7.28298165e-03 -5.18167138e-01
-5.77522695e-01 -8.30169022e-01 -9.47098613e-01 -4.23374847e-02
-8.06886196e-01 2.84579962e-01 9.51411247e-01 8.16146195e-01
2.14699551e-01 7.00886780e-03 6.00436509e-01 -9.15355086e-01
-6.74299777e-01 -6.89990163e-01 -7.41187632e-01 3.93489242e-01
6.49576545e-01 -1.22284544e+00 -4.36792403e-01 -2.08915666e-01] | [7.38260555267334, -2.258282423019409] |
60da9243-368e-441c-8aa9-d61bbbb3db67 | select-and-attend-towards-controllable | 1909.04453 | null | https://arxiv.org/abs/1909.04453v1 | https://arxiv.org/pdf/1909.04453v1.pdf | Select and Attend: Towards Controllable Content Selection in Text Generation | Many text generation tasks naturally contain two steps: content selection and surface realization. Current neural encoder-decoder models conflate both steps into a black-box architecture. As a result, the content to be described in the text cannot be explicitly controlled. This paper tackles this problem by decoupling content selection from the decoder. The decoupled content selection is human interpretable, whose value can be manually manipulated to control the content of generated text. The model can be trained end-to-end without human annotations by maximizing a lower bound of the marginal likelihood. We further propose an effective way to trade-off between performance and controllability with a single adjustable hyperparameter. In both data-to-text and headline generation tasks, our model achieves promising results, paving the way for controllable content selection in text generation. | ['Satoshi Sekine', 'Jun Suzuki', 'Xiaoyu Shen', 'Kentaro Inui', 'Dietrich Klakow', 'Hui Su'] | 2019-09-10 | select-and-attend-towards-controllable-1 | https://aclanthology.org/D19-1054 | https://aclanthology.org/D19-1054.pdf | ijcnlp-2019-11 | ['headline-generation'] | ['natural-language-processing'] | [ 4.39189643e-01 6.92464411e-01 -3.92309785e-01 -3.89636308e-01
-1.02991068e+00 -7.14558125e-01 9.24457729e-01 -2.59198789e-02
-2.95994967e-01 8.66229355e-01 5.69171369e-01 -2.72098213e-01
3.87219727e-01 -8.28220248e-01 -6.24607980e-01 -5.32043576e-01
4.78229105e-01 6.23361290e-01 -1.75991692e-02 -2.37670019e-01
1.40379101e-01 -1.04929116e-02 -1.30049622e+00 2.46178389e-01
1.05025434e+00 6.59933507e-01 6.40118778e-01 8.05599689e-01
-1.95791170e-01 4.67899323e-01 -7.60793805e-01 -4.43885714e-01
2.91849338e-02 -8.86865497e-01 -6.08393490e-01 3.37196350e-01
1.40432164e-01 -3.87133032e-01 -2.70573329e-02 9.92247403e-01
6.65804029e-01 -1.06016614e-01 8.03482354e-01 -1.01778090e+00
-6.74809217e-01 1.07563376e+00 -1.70376599e-01 -3.77430528e-01
2.68974781e-01 2.50642300e-01 1.37059808e+00 -7.46291935e-01
7.75770903e-01 1.27810061e+00 -6.56114444e-02 7.16747820e-01
-1.57513404e+00 -3.61896932e-01 3.22077215e-01 -3.89181674e-01
-1.09495056e+00 -6.42223656e-01 6.61452770e-01 -5.67574382e-01
8.46223235e-01 1.82170197e-01 6.82454646e-01 1.24924707e+00
1.66134521e-01 9.67320681e-01 5.69942117e-01 -6.37742877e-01
3.77322286e-01 2.88619578e-01 -3.75458181e-01 5.79976261e-01
2.41178885e-01 -9.80916172e-02 -5.77539742e-01 2.07938757e-02
7.94362724e-01 -5.75763941e-01 -3.42872739e-01 -1.89436615e-01
-1.37883854e+00 9.27119851e-01 2.17931028e-02 -1.88576076e-02
-1.45878002e-01 3.69681388e-01 3.72448236e-01 2.83937395e-01
4.51754779e-01 8.45718920e-01 -3.61195773e-01 -2.45620802e-01
-1.01410103e+00 4.30980265e-01 8.19544733e-01 1.35906410e+00
4.43711817e-01 1.79432526e-01 -9.12148237e-01 8.71735334e-01
4.66921359e-01 5.93333483e-01 5.13607502e-01 -6.56398773e-01
7.79588997e-01 3.23581427e-01 3.90898168e-01 -3.96474242e-01
-6.37455732e-02 -4.21248227e-01 -5.89529991e-01 7.91647583e-02
3.52210104e-01 -6.72380984e-01 -1.11509180e+00 1.98741746e+00
2.10393280e-01 -4.23541099e-01 -9.71793085e-02 8.60434949e-01
4.24455136e-01 9.13079441e-01 1.36440411e-01 -6.50250390e-02
1.42839229e+00 -1.15008020e+00 -1.02799296e+00 -3.17889661e-01
6.99071944e-01 -7.65274584e-01 1.45766187e+00 2.20586032e-01
-1.37206137e+00 -3.19462329e-01 -1.08598959e+00 -2.27297157e-01
3.40482011e-03 5.40310502e-01 3.36747855e-01 4.49841946e-01
-1.00352526e+00 3.71095926e-01 -6.23147786e-01 1.70002148e-01
1.66231021e-01 4.09385115e-01 -3.50352935e-02 4.77150023e-01
-1.23434591e+00 6.71119392e-01 5.31444430e-01 -5.30353859e-02
-9.28427041e-01 -4.62472439e-01 -7.64344513e-01 2.13882193e-01
3.61826301e-01 -1.25283146e+00 1.68565106e+00 -1.06313765e+00
-2.07450247e+00 6.78685129e-01 -2.89080650e-01 -3.20961058e-01
6.96319938e-01 -3.28418076e-01 -2.53050160e-02 -1.18961379e-01
2.58872390e-01 9.07449305e-01 1.21338391e+00 -1.14886045e+00
-6.20613813e-01 2.50226408e-01 4.11458760e-02 3.50118309e-01
-3.70760471e-01 -2.03299358e-01 -7.10967839e-01 -1.09166777e+00
-3.74277443e-01 -9.70317543e-01 -2.72846758e-01 2.71429718e-02
-7.51317441e-01 -1.43903390e-01 3.90316278e-01 -3.89139831e-01
1.36490297e+00 -1.87464714e+00 4.58653659e-01 -1.56605691e-01
5.49830822e-03 1.30315304e-01 -3.84645104e-01 4.14184749e-01
2.77464241e-01 3.74635339e-01 -3.07597965e-02 -5.41996002e-01
3.19153905e-01 -2.52659678e-01 -5.72147369e-01 -1.51450345e-02
3.96781325e-01 9.44001734e-01 -9.54920530e-01 -4.68930513e-01
5.80366189e-03 3.97087008e-01 -9.09686208e-01 5.92231214e-01
-9.48181152e-01 3.90867054e-01 -6.91183031e-01 2.54937112e-01
2.35792071e-01 -3.62951159e-01 2.53026098e-01 1.06408447e-01
2.37708408e-02 7.68563330e-01 -9.44776475e-01 1.52082968e+00
-7.98055708e-01 7.21446812e-01 7.84003288e-02 -2.25955591e-01
8.45414996e-01 4.74199623e-01 -3.00343800e-02 -5.00441551e-01
2.55345106e-01 3.33489999e-02 -4.41258177e-02 -2.58599550e-01
8.01940739e-01 -3.57449204e-02 -2.28295162e-01 5.89709580e-01
-6.24175295e-02 -5.17725050e-01 3.08232933e-01 9.04665664e-02
6.37100101e-01 2.71218687e-01 1.70733497e-01 -2.92477727e-01
2.16044381e-01 -6.18382096e-02 3.39630872e-01 8.16095769e-01
3.34114105e-01 8.40950966e-01 8.90143037e-01 1.46792576e-01
-1.30752838e+00 -8.56830299e-01 -9.38967802e-03 1.12933850e+00
-1.04910955e-01 -5.87093890e-01 -1.18218124e+00 -6.44360542e-01
-2.62186646e-01 1.16271365e+00 -4.88231987e-01 -1.04559116e-01
-6.00903273e-01 -5.28268754e-01 4.08214569e-01 3.85495991e-01
2.16899872e-01 -1.10296822e+00 -4.89207298e-01 4.16757166e-01
-3.12240571e-01 -1.11283898e+00 -9.16424155e-01 2.16178894e-01
-7.35235333e-01 -3.61317128e-01 -7.84571290e-01 -6.93793714e-01
8.88205469e-01 -6.85134828e-02 1.24236262e+00 -3.77871394e-02
1.42041981e-01 -1.08142018e-01 -2.56984949e-01 -4.53895748e-01
-7.45165110e-01 6.15931511e-01 -4.28247452e-01 -2.09217072e-02
-1.86065167e-01 -3.83355319e-01 -5.82850337e-01 2.13024423e-01
-9.52606916e-01 7.43137062e-01 5.82987547e-01 1.00543594e+00
5.85504413e-01 -3.11816543e-01 7.21254468e-01 -9.04120445e-01
9.87293422e-01 -2.93874741e-01 -8.76253068e-01 3.24784428e-01
-6.85694277e-01 5.95746934e-01 8.66390109e-01 -4.99429941e-01
-1.24920225e+00 2.34390438e-01 -1.30979329e-01 -4.18507047e-02
6.28576279e-02 4.09974724e-01 -5.53709090e-01 6.15386963e-01
3.53948504e-01 1.49487585e-01 -2.31779054e-01 -2.77506888e-01
8.23312044e-01 7.17333794e-01 1.81545392e-01 -5.63788950e-01
6.43131435e-01 3.22737060e-02 -3.00315201e-01 -6.60341978e-01
-8.59617949e-01 1.78860530e-01 -4.56223756e-01 -2.67249700e-02
9.37761724e-01 -1.08591986e+00 -3.47023904e-01 2.14709073e-01
-1.33895910e+00 -7.41610825e-01 -3.99870425e-01 1.18326336e-01
-7.32712448e-01 -3.62505428e-02 -4.93976504e-01 -6.78168237e-01
-5.36942661e-01 -1.29257894e+00 1.59961426e+00 3.23693976e-02
-6.63996816e-01 -9.72527206e-01 -1.64526388e-01 1.88002959e-01
3.95381838e-01 -5.16241305e-02 1.07201529e+00 -4.42155242e-01
-7.60947883e-01 -2.92492270e-01 -1.60988718e-01 -2.05467157e-02
2.73969900e-02 4.85814102e-02 -8.98886323e-01 -8.57152939e-02
-1.87520310e-01 -2.61006713e-01 8.25937033e-01 3.47928971e-01
1.10316658e+00 -6.70900941e-01 -2.38675833e-01 6.41366601e-01
1.04663420e+00 1.82116944e-02 5.70414305e-01 1.70079991e-01
7.70546317e-01 4.78420824e-01 4.64913994e-01 6.39596760e-01
3.21424067e-01 9.34495389e-01 2.99722165e-01 4.39694375e-02
-2.60323077e-01 -7.85978258e-01 4.27065313e-01 6.38793409e-01
3.79514188e-01 -9.71444309e-01 -6.48345172e-01 5.07053375e-01
-1.89479232e+00 -7.38869548e-01 -6.18981139e-04 2.09685349e+00
1.32420886e+00 4.09673035e-01 -3.66327912e-02 -1.91502675e-01
7.31376171e-01 2.21726254e-01 -3.99162024e-01 -3.04360986e-01
1.17216028e-01 -1.31243065e-01 3.44901294e-01 9.25795794e-01
-8.58387113e-01 1.26799798e+00 6.67090607e+00 9.25174832e-01
-1.06660140e+00 -1.10849008e-01 7.66065836e-01 -3.50477576e-01
-8.75919044e-01 1.85907811e-01 -1.26325321e+00 6.81436300e-01
9.68644500e-01 -3.00030440e-01 2.98175156e-01 8.12588573e-01
7.11946607e-01 1.23129576e-01 -1.34443760e+00 6.59918368e-01
-2.28907615e-01 -1.38102889e+00 4.42820013e-01 4.38939855e-02
7.74385273e-01 -4.29365575e-01 2.89795231e-02 2.19054058e-01
3.96670133e-01 -1.05120993e+00 1.17956805e+00 3.58931005e-01
1.08377075e+00 -4.82568383e-01 2.33227193e-01 3.42487603e-01
-8.64552557e-01 1.67031348e-01 -9.01924521e-02 1.12545609e-01
6.87359333e-01 7.80169845e-01 -1.08490896e+00 -1.25127569e-01
-1.94902673e-01 2.89026529e-01 -2.80416578e-01 6.73655748e-01
-7.19732404e-01 5.37350416e-01 1.09168449e-02 -4.08961743e-01
2.01079071e-01 -2.42669508e-01 6.19291067e-01 1.30373681e+00
3.15665632e-01 -2.20688894e-01 1.98901802e-01 1.23765469e+00
-3.71470928e-01 2.52910238e-02 -3.74196887e-01 -3.55573207e-01
5.75876296e-01 1.06143796e+00 -5.66602290e-01 -4.92613643e-01
-4.16537076e-02 1.09967494e+00 3.29303652e-01 4.54025090e-01
-8.48742068e-01 -5.62068760e-01 5.65516829e-01 3.47416192e-01
3.99288744e-01 -2.15830088e-01 -6.34671271e-01 -1.27252340e+00
5.20279221e-02 -9.86255169e-01 -2.58274496e-01 -7.08959877e-01
-8.65242004e-01 6.79991663e-01 9.01938509e-03 -1.09669423e+00
-6.65575147e-01 -3.62187952e-01 -5.07388532e-01 1.02954662e+00
-1.26555693e+00 -9.82331276e-01 2.70128306e-02 1.54673412e-01
7.96626031e-01 -1.26783848e-01 7.74960637e-01 6.46856651e-02
-6.33495033e-01 8.20962012e-01 -4.21547554e-02 1.00096352e-02
6.28537536e-01 -1.39376938e+00 8.70810330e-01 7.60971606e-01
-1.87531382e-01 4.70745355e-01 8.79744351e-01 -7.85052061e-01
-1.23772025e+00 -1.18008149e+00 1.21993387e+00 -3.11112195e-01
6.00438774e-01 -8.39872897e-01 -5.09876430e-01 4.59086716e-01
4.36731607e-01 -6.18072987e-01 4.62588519e-01 2.57493947e-02
-2.10576266e-01 2.24901915e-01 -7.20315814e-01 1.25249791e+00
8.60278249e-01 -4.31474626e-01 -2.31101334e-01 4.80172038e-01
1.16629755e+00 -5.12654722e-01 -5.44707477e-01 -2.11103130e-02
3.95670086e-01 -5.43223500e-01 5.43505669e-01 -4.50041503e-01
8.00464272e-01 -1.30148962e-01 1.14508465e-01 -1.52253425e+00
-2.07390785e-01 -1.21213984e+00 -2.42007464e-01 1.20970154e+00
9.62168634e-01 -2.28645325e-01 7.53284037e-01 9.02965903e-01
-2.32699066e-01 -7.35723138e-01 -4.46162432e-01 -4.38256472e-01
3.65208760e-02 -1.93268165e-01 7.72637546e-01 4.24976438e-01
1.31136864e-01 9.07742560e-01 -5.64012647e-01 -3.37846190e-01
2.98086822e-01 4.10075784e-02 7.69133985e-01 -8.16439569e-01
-5.14599383e-01 -7.96827853e-01 3.19625854e-01 -1.71724129e+00
1.76944926e-01 -8.16516221e-01 3.35567147e-01 -1.57776964e+00
1.37675097e-02 -4.55017269e-01 4.66086328e-01 3.80116433e-01
-4.26125884e-01 -3.05259049e-01 3.03069562e-01 8.45559239e-02
-3.12417477e-01 8.97443354e-01 1.64703333e+00 -4.34233956e-02
-4.44312781e-01 6.50296267e-03 -8.72671187e-01 4.03601199e-01
1.03497875e+00 -5.03006458e-01 -7.83471167e-01 -7.47608542e-01
4.20389920e-01 4.29865479e-01 -1.33446138e-02 -4.22231197e-01
1.91888139e-02 -4.36477482e-01 1.27665654e-01 -3.88682008e-01
3.51193130e-01 -4.51305479e-01 -1.73836738e-01 9.44209471e-02
-1.06985879e+00 -2.49047391e-03 -4.25420403e-02 5.13730407e-01
-9.18486342e-02 -4.18631673e-01 7.73689270e-01 -2.19727848e-02
-3.07892188e-02 2.02874258e-01 -6.75536454e-01 3.23862821e-01
7.59467125e-01 1.66365579e-02 -1.55164868e-01 -7.32081831e-01
-5.51624537e-01 3.40776265e-01 3.61601770e-01 4.81719702e-01
4.79940265e-01 -1.27709937e+00 -8.35149825e-01 3.31536084e-01
-4.15162854e-02 8.10078532e-02 -2.10825250e-01 2.43636340e-01
-3.91663164e-01 4.75645959e-01 2.54482001e-01 -3.25334638e-01
-8.83319974e-01 2.92340755e-01 3.00142318e-01 -3.99026066e-01
-5.22130430e-01 6.99101210e-01 2.22330675e-01 -8.78029019e-02
4.48024094e-01 -2.49459758e-01 1.03150301e-01 1.16778284e-01
6.23886347e-01 -9.70417261e-02 -1.72650874e-01 -2.64977455e-01
2.84396678e-01 1.78769007e-01 -2.00709745e-01 -6.73180580e-01
1.01642537e+00 -3.13765496e-01 1.88983176e-02 2.57228017e-01
1.17065156e+00 -3.23538594e-02 -1.49470603e+00 1.21848816e-02
-4.69299592e-02 -2.75653243e-01 2.19858259e-01 -8.54483724e-01
-8.08502316e-01 8.12099755e-01 1.27565973e-02 2.45425180e-01
8.58495951e-01 -2.51849502e-01 9.24508095e-01 3.52636904e-01
-2.69244937e-03 -1.29657066e+00 2.56015033e-01 6.50295913e-01
9.66282725e-01 -1.12214005e+00 -2.73030579e-01 -3.13301027e-01
-9.90639269e-01 1.00148344e+00 6.58693016e-01 1.82570741e-02
4.66575295e-01 5.73241889e-01 -1.67071447e-02 9.17752236e-02
-1.30522895e+00 6.82627363e-03 4.55036998e-01 4.29049432e-01
8.82198811e-01 1.73899561e-01 -3.90913606e-01 4.68513131e-01
-5.10777235e-01 -2.26716578e-01 5.98539948e-01 5.74760556e-01
-6.69866145e-01 -1.40216768e+00 -1.40677452e-01 3.74854743e-01
-5.42661130e-01 -3.63790274e-01 -4.83393818e-01 3.98975283e-01
-1.90217525e-01 9.88880157e-01 -2.78310832e-02 -1.05436906e-01
1.11816175e-01 1.04376584e-01 1.73901275e-01 -9.66583729e-01
-6.46134913e-01 4.94384110e-01 4.81221825e-01 -1.84190989e-01
2.44811103e-01 -3.65503043e-01 -1.11358106e+00 -6.12961538e-02
-6.54888034e-01 2.22014457e-01 5.55068612e-01 8.15580130e-01
4.43390965e-01 7.30990648e-01 5.73493242e-01 -7.78559864e-01
-7.51603007e-01 -9.07743037e-01 -3.82220000e-01 2.17014149e-01
4.58713621e-01 -1.79677740e-01 -1.32404581e-01 3.66315782e-01] | [11.84296989440918, 9.088920593261719] |
5c2d904a-4a9a-4b68-b36b-af7b3bad8ce7 | a-general-framework-for-uncertainty-1 | 2306.01189 | null | https://arxiv.org/abs/2306.01189v1 | https://arxiv.org/pdf/2306.01189v1.pdf | A General Framework for Uncertainty Quantification via Neural SDE-RNN | Uncertainty quantification is a critical yet unsolved challenge for deep learning, especially for the time series imputation with irregularly sampled measurements. To tackle this problem, we propose a novel framework based on the principles of recurrent neural networks and neural stochastic differential equations for reconciling irregularly sampled measurements. We impute measurements at any arbitrary timescale and quantify the uncertainty in the imputations in a principled manner. Specifically, we derive analytical expressions for quantifying and propagating the epistemic and aleatoric uncertainty across time instants. Our experiments on the IEEE 37 bus test distribution system reveal that our framework can outperform state-of-the-art uncertainty quantification approaches for time-series data imputations. | ['Balasubramaniam Natarajan', 'Sai Munikoti', 'Shweta Dahale'] | 2023-06-01 | null | null | null | null | ['imputation', 'imputation', 'imputation'] | ['computer-vision', 'miscellaneous', 'time-series'] | [-1.50730446e-01 -7.11416528e-02 2.25151047e-01 -4.96612877e-01
-1.41896594e+00 -3.49713385e-01 3.01474690e-01 6.26328774e-03
-8.14032853e-02 1.44684315e+00 3.72741520e-01 -3.78729135e-01
-8.36023331e-01 -9.73013699e-01 -9.81694818e-01 -9.74714816e-01
-2.96645105e-01 5.74642479e-01 -6.89353108e-01 1.27718642e-01
-2.65066009e-02 3.81994098e-01 -1.03319526e+00 -5.64943440e-02
1.00229621e+00 1.24190259e+00 -7.95881152e-01 2.72658616e-01
3.17535065e-02 9.95138705e-01 -9.05958414e-01 -4.49448287e-01
9.28108990e-02 -2.69910574e-01 -4.18303698e-01 -6.27903461e-01
-4.75438297e-01 -4.67791229e-01 -6.14611685e-01 1.16890514e+00
5.79995811e-01 2.03853548e-01 8.85868847e-01 -1.50947928e+00
-4.92117554e-01 1.21369731e+00 -5.18942535e-01 1.83902904e-01
-3.61733064e-02 1.17739860e-03 3.53107125e-01 -3.79029959e-01
-3.17784883e-02 7.82679260e-01 1.27007365e+00 3.08434069e-01
-1.36216640e+00 -6.91646397e-01 -2.82996744e-01 1.38380080e-01
-1.39834177e+00 -2.97102064e-01 1.02986848e+00 -4.21221912e-01
6.26174152e-01 1.14147916e-01 1.86044902e-01 1.40108109e+00
8.83851469e-01 6.53084397e-01 1.05923748e+00 1.33317605e-01
7.34549999e-01 -6.41511142e-01 4.11055923e-01 -1.39509231e-01
1.71863168e-01 5.01514196e-01 -4.08370703e-01 -4.26125586e-01
4.71316904e-01 2.03731209e-01 -4.88915481e-02 5.58024310e-02
-1.10811329e+00 6.71589375e-01 1.37558207e-01 -2.55788535e-01
-7.19293594e-01 7.35490024e-01 5.79205990e-01 3.56727630e-01
8.12077105e-01 9.12884399e-02 -4.59907383e-01 -4.80379105e-01
-1.03638256e+00 2.94228286e-01 8.26611102e-01 6.68769240e-01
3.28785151e-01 5.80193460e-01 -6.16485953e-01 3.33016008e-01
3.46887469e-01 7.82244921e-01 -1.07576750e-01 -1.30580103e+00
4.60868835e-01 -1.12971403e-01 5.14225006e-01 -4.60447937e-01
-4.93863553e-01 -7.30192840e-01 -1.56442463e+00 1.47043422e-01
5.96799195e-01 -8.43323231e-01 -7.71416426e-01 1.96778774e+00
5.17386124e-02 7.03140438e-01 2.14108333e-01 5.87845743e-01
2.53643215e-01 5.78198791e-01 -6.66746870e-02 -4.11059648e-01
8.38806272e-01 -9.18468162e-02 -1.13410795e+00 2.87683934e-01
1.24370098e-01 -2.70294487e-01 1.19310349e-01 3.42730254e-01
-1.10453093e+00 -1.11371405e-01 -1.11553085e+00 3.42931747e-01
-1.29636854e-01 -3.33345592e-01 4.78629678e-01 7.46178210e-01
-5.09339809e-01 1.00080466e+00 -1.05310476e+00 6.37818754e-01
2.95641571e-01 1.15651898e-01 6.17003925e-02 4.02332366e-01
-1.65776348e+00 9.56566334e-01 1.65505633e-01 9.59611773e-01
-1.05201209e+00 -1.23143995e+00 -8.46172392e-01 5.11107855e-02
1.27691954e-01 -7.06149161e-01 1.39043319e+00 -1.53957605e-01
-1.61957645e+00 -4.71285060e-02 -4.66853604e-02 -1.10682881e+00
8.05638731e-01 -8.84841457e-02 -7.16622531e-01 -4.79509026e-01
3.36911939e-02 -4.28644329e-01 7.13479877e-01 -8.74592066e-01
5.71969151e-03 -3.39925528e-01 -2.96133757e-01 -4.41312283e-01
3.98988485e-01 -3.78765345e-01 5.50976515e-01 -6.62439704e-01
2.55970657e-01 -5.41616261e-01 -5.00705242e-01 -4.29603308e-01
-6.91231072e-01 1.20713646e-02 2.58349299e-01 -9.20281708e-01
9.80519474e-01 -1.78196633e+00 1.87773868e-01 4.70874548e-01
1.92035422e-01 -3.89051497e-01 3.33830595e-01 6.23526633e-01
-1.13196947e-01 -4.85154465e-02 -6.52874231e-01 -5.04394412e-01
6.46262646e-01 2.21624255e-01 -8.82713854e-01 7.20829785e-01
1.51241139e-01 9.14121628e-01 -8.24533165e-01 7.14163780e-02
4.20398951e-01 6.88683033e-01 1.01968646e-01 5.96382134e-02
-2.94165045e-01 8.11297894e-01 -3.79202902e-01 5.14979362e-01
9.73087847e-01 1.80505365e-02 -1.59340799e-01 -1.97640732e-01
1.20159257e-02 1.60415560e-01 -1.31514287e+00 1.78313828e+00
-5.37465155e-01 3.93719792e-01 -1.49943888e-01 -9.04394388e-01
8.40268433e-01 3.45483333e-01 7.13595033e-01 -5.93976557e-01
4.41725791e-01 1.72446311e-01 -3.66292685e-01 -1.22695431e-01
4.62791115e-01 -3.47155392e-01 -7.62890756e-01 8.77505422e-01
3.42708863e-02 -4.04580273e-02 -1.01670906e-01 -3.72234434e-02
1.18422031e+00 1.90816492e-01 -1.29929423e-01 -1.85917854e-01
2.16970354e-01 -5.47703922e-01 8.59459162e-01 9.86645222e-01
-1.59810528e-01 5.25046706e-01 6.84557080e-01 -4.33338881e-01
-1.11830747e+00 -1.65042520e+00 -5.47662318e-01 1.99328303e-01
-2.63472348e-01 1.37028337e-01 -8.07055950e-01 -1.99653342e-01
3.05685550e-01 1.25412214e+00 -8.97042274e-01 -2.73502618e-01
-2.55253404e-01 -1.29335773e+00 7.90801346e-01 7.79166937e-01
4.00097787e-01 -7.43547916e-01 -2.00178966e-01 6.49963677e-01
-3.74654591e-01 -7.06405759e-01 -3.19443680e-02 5.36038518e-01
-6.70083880e-01 -7.16151953e-01 -6.02085292e-01 2.77358234e-01
3.35253537e-01 -5.59808791e-01 1.03237092e+00 -6.79395199e-01
1.69642687e-01 3.49717766e-01 2.01483920e-01 -6.40069485e-01
-5.21263242e-01 -2.91275680e-01 1.85202315e-01 -1.23138055e-02
4.03467715e-01 -9.63644564e-01 -2.52157688e-01 1.17738739e-01
-7.77084112e-01 -4.19313371e-01 1.30480856e-01 9.50440466e-01
8.38389874e-01 3.80002499e-01 1.10679913e+00 -4.83171225e-01
1.07579517e+00 -9.43493128e-01 -1.04457784e+00 2.82587171e-01
-4.40651149e-01 4.88577902e-01 5.12292504e-01 -1.37899250e-01
-1.25441968e+00 -2.45001346e-01 -2.62068242e-01 -3.08174968e-01
5.71801178e-02 8.76516342e-01 -8.70585442e-02 1.83862701e-01
4.05287981e-01 2.38959789e-01 -2.32663259e-01 -2.85359830e-01
3.76388818e-01 7.27033734e-01 1.00476301e+00 -1.25530982e+00
6.04249716e-01 5.60714722e-01 4.15452629e-01 -2.81901449e-01
-8.85875821e-01 2.28577018e-01 -4.51964468e-01 -4.87017363e-01
4.40460414e-01 -9.47866678e-01 -1.37773871e+00 7.84557223e-01
-1.38178051e+00 -2.26550192e-01 -7.91540682e-01 5.46794772e-01
-1.01373267e+00 2.38378361e-01 -8.28269541e-01 -1.41447115e+00
-5.19660413e-01 -9.12491083e-01 9.31169868e-01 1.47688225e-01
-4.87281531e-02 -9.37392712e-01 3.78192902e-01 -1.49089098e-01
6.05062962e-01 9.27732587e-01 7.10627854e-01 -2.40015537e-01
-5.94566822e-01 -3.71777415e-01 -1.30926356e-01 3.04148972e-01
-3.35592985e-01 -1.50237918e-01 -8.57653916e-01 4.60505038e-02
5.28635800e-01 -1.72950365e-02 5.44024169e-01 7.68638730e-01
1.34510684e+00 -4.12321314e-02 7.53373429e-02 7.55294025e-01
1.43072152e+00 1.60057217e-01 9.86381769e-01 1.76700115e-01
1.11259066e-01 2.07041711e-01 -1.96778998e-02 9.44606900e-01
3.37005377e-01 1.91801310e-01 5.63678145e-01 6.81173205e-01
4.82893437e-01 -2.41267398e-01 2.30835035e-01 9.17196631e-01
5.79453632e-02 -2.39666402e-01 -9.05385256e-01 7.45712280e-01
-2.06411195e+00 -1.16383636e+00 -4.30013090e-01 2.40118194e+00
7.99079776e-01 1.12924933e-01 -3.51460338e-01 2.59752989e-01
5.97953677e-01 -3.23360935e-02 -1.03121650e+00 -2.97812641e-01
-2.79473394e-01 2.71752447e-01 6.92101061e-01 4.00442183e-01
-8.96665573e-01 -3.08869928e-02 7.21531391e+00 6.99758410e-01
-4.89282846e-01 3.20940703e-01 6.32745266e-01 -1.10983878e-01
-4.85265493e-01 -2.63492763e-01 -4.81628299e-01 9.39250112e-01
1.65810490e+00 -6.77308381e-01 5.61117887e-01 2.57786572e-01
2.31068701e-01 1.16794379e-02 -1.03314841e+00 8.56015861e-01
-4.29594517e-01 -1.54734623e+00 -5.33554614e-01 -1.02618307e-01
9.33608770e-01 3.37489843e-01 2.37589747e-01 4.19090152e-01
8.86667848e-01 -1.12861061e+00 6.74195230e-01 1.43324101e+00
4.03550416e-01 -1.18059611e+00 1.14742482e+00 2.86291629e-01
-6.96186662e-01 -4.74711731e-02 -2.80903995e-01 -2.55972892e-01
7.57518291e-01 1.42425704e+00 6.26933528e-03 8.55814457e-01
5.97718656e-01 6.04164004e-01 1.84799954e-01 1.26901412e+00
-3.77199620e-01 6.82531953e-01 -6.68083131e-01 1.71610638e-01
-6.97303191e-03 -4.77167189e-01 5.50667048e-01 6.25655651e-01
6.87861443e-01 -3.71179357e-02 -5.86984515e-01 1.47569883e+00
-2.44453713e-01 -8.50645602e-01 -3.53235930e-01 2.38438994e-01
6.40632749e-01 6.42164767e-01 -9.97690037e-02 2.11866181e-02
3.65692712e-02 4.83418643e-01 5.83587065e-02 7.55547106e-01
-1.35148180e+00 -4.41361010e-01 7.30110168e-01 -5.70117116e-01
-1.28782600e-01 -2.48714104e-01 -7.70130932e-01 -1.30702770e+00
2.72306323e-01 -3.54838610e-01 3.56878221e-01 -9.52537894e-01
-2.01168513e+00 4.18128252e-01 1.20902956e-01 -1.27991509e+00
-8.94874811e-01 -2.12321028e-01 -7.35261977e-01 1.23789907e+00
-1.34812272e+00 -8.17203760e-01 1.34807214e-01 5.99483013e-01
-1.70650661e-01 5.16513735e-02 8.76323938e-01 1.38800502e-01
-7.35919476e-01 3.78426939e-01 1.04636657e+00 2.22523227e-01
6.53605163e-02 -1.20057905e+00 4.67377424e-01 9.70923901e-01
-2.36660033e-01 3.93380612e-01 1.14800060e+00 -6.66339159e-01
-1.53091753e+00 -1.11440372e+00 3.43339115e-01 -5.29952526e-01
1.20968056e+00 -2.32329264e-01 -9.95593548e-01 8.34132671e-01
2.48083845e-01 4.62675802e-02 6.82893634e-01 2.25595161e-01
-2.38355413e-01 -7.50700831e-02 -1.49560070e+00 2.92250603e-01
5.03386319e-01 -7.73407936e-01 -8.81234705e-01 3.65001291e-01
7.53410876e-01 -6.70293987e-01 -1.42636323e+00 5.32652259e-01
5.63964128e-01 -7.12711334e-01 1.07328069e+00 -5.68068385e-01
3.62751275e-01 -4.14742500e-01 -3.74771357e-01 -1.59066784e+00
1.34554982e-01 -1.03122079e+00 -7.07854271e-01 1.22469783e+00
2.78766245e-01 -9.51192856e-01 7.39297211e-01 9.40232217e-01
-2.77060010e-02 -2.98875570e-01 -1.70969963e+00 -1.02799451e+00
6.15225852e-01 -1.06235480e+00 1.20559061e+00 6.74634576e-01
6.11106446e-03 -3.82315814e-01 -7.34788418e-01 5.16320884e-01
1.60052633e+00 1.36613473e-02 2.75691897e-01 -1.23706532e+00
-1.49241731e-01 -2.29749903e-01 -2.89670199e-01 -4.29347306e-01
5.33881128e-01 -5.95986545e-01 1.80564716e-01 -1.23738968e+00
-9.61302072e-02 -3.29647511e-01 -7.03180075e-01 6.96619600e-02
4.00375634e-01 -1.46253899e-01 -2.01307565e-01 -1.41394526e-01
-4.57062751e-01 1.07590127e+00 4.27548736e-01 -3.93098027e-01
1.84578493e-01 2.80066937e-01 -3.76646399e-01 4.54463691e-01
8.98420036e-01 -7.67782748e-01 -1.68272957e-01 -5.01038074e-01
6.39544666e-01 7.85717666e-01 6.97427630e-01 -1.12826121e+00
4.54867810e-01 -7.03921914e-02 4.42031085e-01 -1.21147060e+00
2.70961821e-01 -8.56628418e-01 8.79686832e-01 2.39535496e-01
-3.63846689e-01 -5.48999012e-02 5.18537045e-01 1.01093304e+00
-1.90462738e-01 -1.95556834e-01 3.64325911e-01 3.84529799e-01
-1.16260104e-01 2.29312226e-01 -4.36329186e-01 1.55604720e-01
7.67179608e-01 4.62554634e-01 -6.67037904e-01 -5.73199332e-01
-1.06368923e+00 4.58803415e-01 -1.23452820e-01 -3.71265188e-02
5.91126382e-01 -1.62976551e+00 -9.52471435e-01 1.36513442e-01
-3.32479566e-01 -8.82271752e-02 8.55646193e-01 8.35547268e-01
9.77143124e-02 1.74870625e-01 -1.75448611e-01 -6.42071307e-01
1.17956482e-01 4.10452574e-01 7.35786498e-01 -3.02038968e-01
-3.63050342e-01 2.32749149e-01 -7.40266144e-01 -7.69998550e-01
1.89056039e-01 -4.64474440e-01 3.33931923e-01 -3.73481810e-02
6.11619413e-01 8.45478296e-01 1.85799539e-01 -3.28061730e-01
-8.73538405e-02 2.65538096e-01 4.18779492e-01 -2.61506438e-01
1.52740932e+00 -2.44287431e-01 -4.30803746e-01 8.93968701e-01
1.04080343e+00 -4.66469735e-01 -1.52427435e+00 -1.95678875e-01
4.50723656e-02 4.19336632e-02 1.17772043e-01 -1.12205374e+00
-9.64584470e-01 8.30510437e-01 4.59633589e-01 3.26202631e-01
9.25595999e-01 -4.63646203e-01 7.72532165e-01 4.53326702e-01
8.43826234e-01 -9.70277429e-01 -9.23116326e-01 6.54323339e-01
8.53397906e-01 -8.34705472e-01 -8.25970769e-02 2.66833812e-01
-7.18632713e-02 1.07889652e+00 -1.35565177e-01 -2.56261647e-01
1.03664744e+00 8.90116274e-01 -3.14807087e-01 1.72026992e-01
-6.85552955e-01 3.05767864e-01 -2.24550087e-02 6.53395832e-01
9.09465626e-02 3.79434526e-01 -7.24546239e-02 1.19942880e+00
-1.20666198e-01 3.54503185e-01 9.02254224e-01 5.31963527e-01
2.47440726e-01 -6.83975279e-01 -5.89017093e-01 5.16210496e-01
-4.20960248e-01 8.22196603e-02 4.24119055e-01 4.24474895e-01
-3.76426160e-01 9.86164331e-01 1.37286723e-01 -1.88640654e-01
3.10966849e-01 2.29604155e-01 2.39062428e-01 2.78178990e-01
-3.29298675e-01 -2.28839815e-01 -7.63282478e-02 -5.25922000e-01
-2.39758104e-01 -9.18926001e-01 -1.24902058e+00 -7.67928779e-01
-1.75422784e-02 3.21309090e-01 9.41018760e-01 1.15133798e+00
3.46669525e-01 1.01015139e+00 4.60122555e-01 -6.87183559e-01
-1.26885211e+00 -9.18987811e-01 -9.49206829e-01 -3.34419869e-02
5.16305029e-01 -5.35160005e-01 -7.49930680e-01 -5.82944334e-01] | [6.90901517868042, 3.4161715507507324] |
321123e7-5267-4505-9fb1-776bfcace541 | differentiating-objects-by-motion-joint | 1709.04666 | null | http://arxiv.org/abs/1709.04666v3 | http://arxiv.org/pdf/1709.04666v3.pdf | Differentiating Objects by Motion: Joint Detection and Tracking of Small Flying Objects | While generic object detection has achieved large improvements with rich
feature hierarchies from deep nets, detecting small objects with poor visual
cues remains challenging. Motion cues from multiple frames may be more
informative for detecting such hard-to-distinguish objects in each frame.
However, how to encode discriminative motion patterns, such as deformations and
pose changes that characterize objects, has remained an open question. To learn
them and thereby realize small object detection, we present a neural model
called the Recurrent Correlational Network, where detection and tracking are
jointly performed over a multi-frame representation learned through a single,
trainable, and end-to-end network. A convolutional long short-term memory
network is utilized for learning informative appearance change for detection,
while learned representation is shared in tracking for enhancing its
performance. In experiments with datasets containing images of scenes with
small flying objects, such as birds and unmanned aerial vehicles, the proposed
method yielded consistent improvements in detection performance over deep
single-frame detectors and existing motion-based detectors. Furthermore, our
network performs as well as state-of-the-art generic object trackers when it
was evaluated as a tracker on the bird dataset. | ['ShaoDi You', 'Rei Kawakami', 'Makoto Iida', 'Tu Tuan Trinh', 'Takeshi Naemura', 'Ryota Yoshihashi'] | 2017-09-14 | null | null | null | null | ['small-object-detection'] | ['computer-vision'] | [-1.08031360e-02 -4.93994415e-01 -1.83241948e-01 -3.94784182e-01
-4.54707772e-01 -7.36590743e-01 4.06126648e-01 -2.69631177e-01
-6.99965179e-01 3.38938355e-01 -5.75291403e-02 3.86861354e-01
1.48743927e-01 -2.67634004e-01 -9.08257067e-01 -7.34691858e-01
-5.15003145e-01 1.33712143e-01 7.88740098e-01 6.99312612e-02
-1.62592083e-01 4.94241029e-01 -1.63973749e+00 3.26808214e-01
3.02230477e-01 1.19316518e+00 5.65692425e-01 8.40639770e-01
4.93356317e-01 9.16642547e-01 -5.17340600e-01 1.12720683e-01
5.51284432e-01 -8.32159445e-02 -2.98232466e-01 1.31367058e-01
1.16078472e+00 -7.68351495e-01 -5.29370248e-01 1.08911991e+00
4.23344851e-01 2.21823797e-01 3.65031809e-01 -1.07916629e+00
-7.33963192e-01 5.08970916e-01 -7.08381236e-01 7.11286306e-01
2.63462335e-01 6.49675310e-01 8.81191373e-01 -8.94562960e-01
4.68424171e-01 1.57197309e+00 6.78643048e-01 8.07204306e-01
-1.07175589e+00 -7.87104905e-01 5.08077264e-01 3.15168574e-02
-1.03721881e+00 -3.63538742e-01 4.58966553e-01 -7.66934097e-01
9.70989466e-01 -7.74506405e-02 8.14759433e-01 1.06537282e+00
4.19270873e-01 1.13895309e+00 5.03923655e-01 1.35184139e-01
-2.49566704e-01 -2.50940740e-01 3.50011051e-01 9.16013479e-01
6.94152713e-01 6.85508966e-01 -4.12085444e-01 1.51105169e-02
8.61321986e-01 3.23751181e-01 -3.62732530e-01 -5.84373355e-01
-1.43119049e+00 6.03410780e-01 9.24044788e-01 6.57826960e-02
-3.51521194e-01 6.35596693e-01 2.17860833e-01 1.21214323e-01
2.45834514e-01 2.21210346e-01 -4.57058281e-01 1.76787674e-01
-9.41351652e-01 2.88941294e-01 4.83257085e-01 9.10932899e-01
5.40065527e-01 5.01177967e-01 -5.92372537e-01 4.07283396e-01
6.95421576e-01 9.97715771e-01 4.42108154e-01 -6.99455738e-01
1.48241729e-01 4.88898486e-01 4.70292002e-01 -1.05715716e+00
-6.99565411e-01 -7.74960339e-01 -4.81904954e-01 2.67076612e-01
4.66948748e-01 -1.22736715e-01 -1.20464563e+00 1.89674199e+00
4.60966408e-01 4.39635158e-01 -8.20482820e-02 1.46519685e+00
1.05528641e+00 4.93474245e-01 2.06406206e-01 3.39889526e-02
1.57926524e+00 -1.17540705e+00 -4.01198745e-01 -5.68270326e-01
3.94128442e-01 -7.10133672e-01 3.55250269e-01 6.13482744e-02
-9.23128664e-01 -1.07046151e+00 -9.82993484e-01 1.04044504e-01
-6.33105114e-02 6.43497467e-01 7.29262888e-01 4.28987056e-01
-9.87872303e-01 5.00519216e-01 -1.37515783e+00 -5.28145611e-01
5.45033514e-01 4.84225839e-01 -1.67133272e-01 8.20068642e-02
-7.76977003e-01 7.27597475e-01 5.54889619e-01 4.49230075e-01
-1.77419055e+00 -5.29650271e-01 -9.33915734e-01 -1.41310409e-01
2.96960801e-01 -1.01234591e+00 1.21790719e+00 -9.89435673e-01
-1.17627180e+00 8.46778274e-01 6.72084000e-03 -9.17642534e-01
3.03858012e-01 -5.46637714e-01 -4.85350013e-01 1.28628224e-01
1.91664502e-01 1.17274272e+00 1.25533557e+00 -9.96438384e-01
-1.06262863e+00 -3.24091256e-01 2.05277964e-01 -4.58256789e-02
-4.79427278e-02 3.70776802e-01 -4.41188306e-01 -6.63400412e-01
-6.50599226e-02 -1.21207380e+00 -2.80097455e-01 4.78775024e-01
4.53058211e-03 -3.89261305e-01 1.34223807e+00 -5.47974169e-01
8.14196944e-01 -2.10413599e+00 2.52346769e-02 -4.52085048e-01
2.22184941e-01 6.04521513e-01 -3.22725326e-01 -2.66606688e-01
2.62567014e-01 -4.08124238e-01 9.22758058e-02 -1.33464992e-01
-9.69702527e-02 3.76492999e-02 -2.99898803e-01 8.98229659e-01
4.64571446e-01 1.02576816e+00 -1.14820611e+00 -2.67371356e-01
3.78883392e-01 5.54451823e-01 -4.06523257e-01 3.03966433e-01
-3.43191028e-01 4.84661698e-01 -5.70241332e-01 9.79973257e-01
5.28177440e-01 -3.28303128e-01 -3.08611184e-01 -3.72259051e-01
-1.95936725e-01 -1.39016926e-01 -1.08309066e+00 1.77864206e+00
3.19854468e-02 8.77246261e-01 1.65147349e-01 -6.17404699e-01
6.95113778e-01 8.15842897e-02 2.75230348e-01 -3.60258013e-01
2.22181216e-01 -6.13285676e-02 3.73041809e-01 -4.93930876e-01
6.09915376e-01 3.15310866e-01 1.17387488e-01 -1.68844350e-02
2.63062507e-01 4.40438569e-01 3.64198983e-01 -5.24753891e-03
1.05879414e+00 3.75061542e-01 1.36635769e-02 -2.57604450e-01
3.72253507e-01 -3.15561742e-02 9.26788867e-01 9.50075388e-01
-5.70099413e-01 5.40287018e-01 -3.56403023e-01 -8.74191105e-01
-5.83187640e-01 -9.76703405e-01 -1.40228048e-01 1.32080889e+00
6.40894055e-01 -2.22643688e-01 -3.05343747e-01 -5.20571589e-01
2.15369582e-01 1.02029093e-01 -6.05531156e-01 -2.31047779e-01
-7.19440818e-01 -8.43292832e-01 4.44389552e-01 9.20905828e-01
5.89954615e-01 -9.68524933e-01 -1.25178707e+00 3.20271134e-01
1.71949744e-01 -1.30814946e+00 -7.56868422e-01 8.74126703e-02
-8.20949376e-01 -1.19401932e+00 -6.05305254e-01 -8.02563727e-01
5.49492121e-01 8.83600414e-01 9.00849223e-01 -1.36798337e-01
-7.48381376e-01 5.00766397e-01 -7.88590908e-02 -3.39596003e-01
1.55693159e-01 -3.17871392e-01 2.87733465e-01 7.39804432e-02
2.87727296e-01 -5.13019711e-02 -1.10184813e+00 5.87599814e-01
-7.42889285e-01 -3.45785797e-01 7.99227834e-01 8.86376858e-01
3.84290934e-01 -3.98421943e-01 1.01256073e-01 -2.62538821e-01
-3.54048721e-02 -2.62631118e-01 -9.43791270e-01 1.77147761e-01
-1.59275159e-01 1.83265030e-01 3.11962724e-01 -6.75071597e-01
-8.08307290e-01 6.15660846e-01 2.83417016e-01 -8.29644799e-01
-1.03385471e-01 -3.10845096e-02 1.12327822e-01 -5.22839010e-01
6.55433118e-01 6.13717772e-02 -3.37802529e-01 -4.48085040e-01
3.74109656e-01 9.56291035e-02 9.03913379e-01 -2.93658048e-01
1.03626156e+00 6.23389542e-01 -1.91954598e-01 -8.20491374e-01
-9.49684560e-01 -8.34285915e-01 -6.94822490e-01 -4.31632400e-01
1.10382760e+00 -1.57924283e+00 -7.63275266e-01 5.15500307e-01
-1.22146475e+00 -3.57907489e-02 -4.46816161e-03 7.94095278e-01
-1.21352591e-01 4.87542301e-02 -6.30732298e-01 -8.42894733e-01
-4.28938627e-01 -1.11578131e+00 1.43139410e+00 5.72228253e-01
1.52606115e-01 -8.19679737e-01 -2.31743138e-02 7.86246583e-02
4.42178398e-01 4.46733594e-01 -3.77803296e-02 -3.45255882e-01
-1.16864932e+00 -4.11251664e-01 -2.29099289e-01 1.62267223e-01
4.56771180e-02 2.20086917e-01 -9.97359574e-01 -9.41351414e-01
-2.48901427e-01 -4.76131380e-01 1.54644489e+00 6.58408344e-01
8.08971643e-01 -1.20661423e-01 -8.08793247e-01 9.90240574e-01
1.22074866e+00 1.53041273e-01 -2.21819822e-02 1.70063525e-01
9.72211480e-01 -2.50165146e-02 6.92672968e-01 4.06177968e-01
2.06099093e-01 8.64901304e-01 6.34268403e-01 -2.37156916e-03
-3.98531824e-01 -9.23822373e-02 8.30776155e-01 3.06011587e-01
-1.58542365e-01 3.85790542e-02 -5.33026516e-01 4.98608023e-01
-2.01233983e+00 -1.17933869e+00 -3.32009122e-02 2.04495311e+00
2.96559721e-01 2.59422213e-01 4.17941630e-01 -6.09697700e-01
8.19146991e-01 3.54950517e-01 -8.51716518e-01 4.97953176e-01
-3.73890013e-01 -4.89411473e-01 7.36213446e-01 5.94923124e-02
-1.72848475e+00 1.17024457e+00 6.08706808e+00 3.23210031e-01
-1.36810422e+00 9.48566049e-02 8.11016560e-03 -2.74190724e-01
3.59422654e-01 -8.35492164e-02 -1.35528743e+00 2.72514075e-01
7.23407328e-01 7.74647892e-02 -8.04444496e-03 1.13719380e+00
1.37810498e-01 1.55608281e-01 -1.09770310e+00 9.86526847e-01
1.41468346e-01 -1.06390250e+00 -1.19626001e-01 -3.26113671e-01
8.06086600e-01 5.27353823e-01 2.50000596e-01 4.23696846e-01
4.72695976e-01 -7.07038760e-01 9.20604050e-01 5.29580176e-01
2.40754604e-01 -3.58846366e-01 5.27055562e-01 2.32100636e-01
-1.87338376e+00 -2.99094707e-01 -8.17906499e-01 -1.58661813e-01
-2.89538279e-02 1.98402137e-01 -6.13446414e-01 3.67640942e-01
9.10125077e-01 1.27926230e+00 -9.79268909e-01 1.59423745e+00
-7.25676417e-02 5.51226318e-01 -3.32763940e-01 1.27176538e-01
5.39644301e-01 2.83109635e-01 9.14462268e-01 1.37217164e+00
1.50462523e-01 -1.26357928e-01 8.40597868e-01 8.88106287e-01
-3.43415476e-02 -5.44533908e-01 -5.85398555e-01 -5.96931539e-02
1.79057077e-01 1.70512342e+00 -7.97146201e-01 -2.55887330e-01
-4.81877744e-01 7.05210865e-01 2.33357042e-01 3.54060471e-01
-1.04408526e+00 6.76046908e-02 9.59250510e-01 -1.54264495e-01
7.57641494e-01 -3.90276283e-01 6.60502076e-01 -1.37259495e+00
-6.90911785e-02 -4.92402107e-01 5.13982475e-01 -4.97116148e-01
-1.26154828e+00 7.13477671e-01 -1.58153191e-01 -1.60830545e+00
-1.58399910e-01 -9.30830896e-01 -5.45464814e-01 4.21914756e-01
-1.39562750e+00 -1.40921569e+00 -5.93989670e-01 5.02850294e-01
5.45736194e-01 -2.11400717e-01 3.65550727e-01 3.33337188e-01
-8.20795536e-01 5.38149357e-01 4.31402437e-02 5.09119332e-01
8.16731572e-01 -1.05902231e+00 4.50538188e-01 1.33161533e+00
4.74165618e-01 6.05624259e-01 6.10701442e-01 -6.33470058e-01
-1.85560524e+00 -1.69662452e+00 -2.67978553e-02 -5.73269963e-01
6.01108074e-01 -2.76656121e-01 -8.23910952e-01 8.07313561e-01
2.00274795e-01 8.97801638e-01 7.31529221e-02 -2.92132556e-01
-3.20823401e-01 -2.30963141e-01 -6.65517092e-01 1.49590388e-01
1.23997378e+00 -3.45957845e-01 -6.60744488e-01 3.37881476e-01
7.67419875e-01 -7.48165250e-01 -5.27997732e-01 5.80593765e-01
6.41783953e-01 -6.23430848e-01 1.20244491e+00 -7.85704374e-01
-2.03613669e-01 -1.00046289e+00 3.81579399e-02 -1.02841306e+00
-8.14236879e-01 -5.81510186e-01 -3.95668149e-01 8.93171966e-01
1.24000810e-01 -2.36318260e-01 6.36709690e-01 2.45922759e-01
-3.20050806e-01 -2.94184327e-01 -6.77134812e-01 -1.09550989e+00
-5.03670871e-01 -1.61643684e-01 7.01113418e-02 5.61702073e-01
-6.29950404e-01 3.99574578e-01 -7.80401409e-01 6.10908031e-01
9.24493134e-01 5.69382489e-01 7.97596514e-01 -1.28728747e+00
-3.14350456e-01 -2.92112470e-01 -9.74216938e-01 -1.56111801e+00
1.62673131e-01 -9.07611012e-01 4.30676162e-01 -1.31658459e+00
3.44561547e-01 1.43557176e-01 -5.66414356e-01 4.70855385e-01
-4.49188620e-01 2.77202874e-01 4.25274223e-01 3.32648844e-01
-1.21855676e+00 6.05053723e-01 1.14402366e+00 -4.99932915e-01
-2.18934398e-02 1.73487321e-01 -2.67521292e-01 8.30786347e-01
1.65802926e-01 -6.00179255e-01 -1.17116421e-01 -5.32280564e-01
-3.85870755e-01 -8.48739035e-03 8.92507911e-01 -1.51018417e+00
7.31354177e-01 1.71588063e-01 8.92924607e-01 -9.95972335e-01
3.77749860e-01 -7.76394308e-01 7.53094926e-02 8.66866708e-01
-2.94243157e-01 1.00915834e-01 4.75373149e-01 1.00251138e+00
-1.87183276e-01 2.25963488e-01 9.66860533e-01 -3.33673432e-02
-1.35104108e+00 6.41080797e-01 -1.03900492e-01 7.51895979e-02
1.01662838e+00 -1.85235679e-01 -4.10829097e-01 1.01357467e-01
-4.98262018e-01 5.07326841e-01 1.89635277e-01 8.55992496e-01
7.53101587e-01 -1.32171762e+00 -7.53101230e-01 8.42949823e-02
1.70262143e-01 2.61460077e-02 9.84143242e-02 8.35175037e-01
-2.33556181e-01 5.97449601e-01 -4.01763052e-01 -1.29382503e+00
-1.28388727e+00 7.06965625e-01 4.87342566e-01 2.58726895e-01
-7.87852108e-01 1.18502796e+00 6.98006213e-01 1.06823221e-01
4.56354350e-01 -9.13277388e-01 -1.05984949e-01 -6.90207407e-02
8.06053042e-01 8.80498216e-02 -2.89033115e-01 -8.14531207e-01
-6.44744456e-01 8.98373306e-01 -1.71190992e-01 5.52217126e-01
1.21821988e+00 -1.67769909e-01 3.22366983e-01 3.59645605e-01
8.38473439e-01 -5.71694732e-01 -1.98341739e+00 -4.00167197e-01
7.61471614e-02 -4.60782737e-01 8.54241028e-02 -5.13856530e-01
-1.45614314e+00 6.62901640e-01 1.02815902e+00 -1.93559080e-01
8.60373557e-01 -7.09452927e-02 7.19983459e-01 7.71409392e-01
6.09342992e-01 -6.97203517e-01 2.28306115e-01 5.45848966e-01
6.62198842e-01 -1.73927999e+00 -7.50229433e-02 -4.25723083e-02
-5.05872548e-01 1.16493833e+00 9.67160165e-01 -3.51830393e-01
4.71888036e-01 1.71635658e-01 3.13907489e-02 -1.45918250e-01
-9.62265551e-01 -6.25326931e-01 8.38801801e-01 4.38610017e-01
4.29743141e-01 -1.00883925e-02 3.08499396e-01 3.66898447e-01
2.48480305e-01 -6.10642023e-02 5.07716984e-02 8.06458831e-01
-7.95995772e-01 -5.16692758e-01 -3.90487641e-01 2.25575715e-01
-3.26429039e-01 1.46978661e-01 -3.11102778e-01 9.06824052e-01
2.58623570e-01 8.19914877e-01 1.65861063e-02 -4.35014218e-01
2.58141339e-01 -3.33967000e-01 5.04955530e-01 -6.11974359e-01
-7.14298069e-01 1.38118207e-01 -1.07038908e-01 -8.94595921e-01
-8.04229498e-01 -7.25317597e-01 -1.12148738e+00 1.97437346e-01
-4.80880827e-01 -1.31892517e-01 2.42634118e-01 9.45916295e-01
2.03204319e-01 7.35662878e-01 2.96886325e-01 -1.24984431e+00
-6.36054456e-01 -8.67755175e-01 -5.00341654e-01 4.17162806e-01
7.55351841e-01 -1.05315471e+00 -2.04684913e-01 2.34169275e-01] | [6.312832355499268, -2.055046796798706] |
3cd1a7db-265f-4388-8cf1-74f6f97a9d9e | ranking-causal-anomalies-via-temporal-and | null | null | https://www.kdd.org/kdd2016/papers/files/rfp0445-chengAemb.pdf | https://www.kdd.org/kdd2016/papers/files/rfp0445-chengAemb.pdf | Ranking Causal Anomalies via Temporal and Dynamical Analysis on Vanishing Correlations | Modern world has witnessed a dramatic increase in our ability to collect, transmit and distribute real-time monitoring
and surveillance data from large-scale information systems and cyber-physical systems. Detecting system anomalies
thus attracts significant amount of interest in many fields such as security, fault management, and industrial optimization. Recently, invariant network has shown to be a powerful way in characterizing complex system behaviours.
In the invariant network, a node represents a system component and an edge indicates a stable, significant interaction
between two components. Structures and evolutions of the
invariance network, in particular the vanishing correlations, can shed important light on locating causal anomalies
and performing diagnosis. However, existing approaches to
detect causal anomalies with the invariant network often
use the percentage of vanishing correlations to rank possible casual components, which have several limitations: 1)
fault propagation in the network is ignored; 2) the root casual anomalies may not always be the nodes with a highpercentage of vanishing correlations; 3) temporal patterns
of vanishing correlations are not exploited for robust detection. To address these limitations, in this paper we propose
a network diffusion based framework to identify significant causal anomalies and rank them. Our approach can effectively model fault propagation over the entire invariant
network, and can perform joint inference on both the structural, and the time-evolving broken invariance patterns. As
a result, it can locate high-confidence anomalies that are truly responsible for the vanishing correlations, and can compensate for unstructured measurement noise in the system.
Extensive experiments on synthetic datasets, bank information system datasets, and coal plant cyber-physical system
datasets demonstrate the effectiveness of our approach. | ['Wei Wang', 'Zhengzhang Chen', 'Guofei Jiang', 'Haifeng Chen', 'Kai Zhang', 'Wei Cheng'] | 2016-07-19 | null | null | null | acm-sigkdd-international-conference-on-2 | ['root-cause-ranking'] | ['graphs'] | [ 1.76755726e-01 -3.07516664e-01 -1.80106059e-01 2.81410098e-01
8.71420428e-02 -6.63513541e-01 5.22716403e-01 4.67823893e-01
5.98561883e-01 5.87049246e-01 -2.90262312e-01 -4.85227764e-01
-9.35096502e-01 -9.28089797e-01 -4.61300045e-01 -7.93591917e-01
-8.69425952e-01 1.59749582e-01 5.92266262e-01 -1.26034498e-01
1.53355449e-01 6.77813053e-01 -1.21801054e+00 -2.21781135e-01
7.24115312e-01 1.13453805e+00 -2.08680496e-01 4.60064918e-01
2.67199278e-01 8.03209186e-01 -9.58155513e-01 3.44343692e-01
1.27493948e-01 -5.49856782e-01 -4.81033564e-01 2.38826767e-01
-3.05441529e-01 -1.60054080e-02 -4.51431990e-01 1.26278234e+00
-4.65060025e-02 -1.34214208e-01 4.69837934e-01 -1.86825454e+00
-2.04566717e-01 4.19382244e-01 -9.38002408e-01 4.96748388e-01
3.14901203e-01 1.44924313e-01 8.37702036e-01 -3.39991897e-01
3.07201058e-01 1.23746204e+00 4.68478948e-01 -1.58432171e-01
-1.08706892e+00 -7.52531111e-01 4.32651162e-01 2.75714248e-01
-1.36322522e+00 5.29707111e-02 1.03819168e+00 -5.01453519e-01
6.78968668e-01 4.06016916e-01 6.46082282e-01 7.57510722e-01
7.71011770e-01 4.02918160e-01 7.35819519e-01 -8.28684121e-02
2.80369610e-01 -4.97452796e-01 3.76690537e-01 7.81064749e-01
4.65463489e-01 5.24225794e-02 -4.18367684e-01 -6.95727587e-01
9.07580435e-01 4.65951055e-01 -3.59380245e-01 8.49213172e-03
-1.13730621e+00 5.47712266e-01 4.76894617e-01 4.53894377e-01
-5.82802951e-01 1.85506865e-01 4.05426502e-01 6.80562794e-01
4.40440387e-01 4.67171401e-01 -5.74171185e-01 -8.14783499e-02
-3.72629076e-01 -2.27882609e-01 6.68828607e-01 5.73071778e-01
5.97113609e-01 2.24159271e-01 1.18490636e-01 6.24892652e-01
1.35386035e-01 6.25878811e-01 2.27672011e-01 -6.42271578e-01
2.38353118e-01 1.05225730e+00 -2.32375428e-01 -1.52747548e+00
-4.63393629e-01 -4.08021122e-01 -1.35518885e+00 -1.42488703e-02
1.75841749e-01 -1.03091143e-01 -7.15513468e-01 1.58058667e+00
3.27564657e-01 6.75952494e-01 -3.10120702e-01 5.07039905e-01
-1.94630548e-01 5.63316047e-01 -6.31029665e-01 -4.97064829e-01
9.88830924e-01 -1.57376483e-01 -7.62443364e-01 1.97807744e-01
3.38643461e-01 -5.84262669e-01 5.38521826e-01 4.53359514e-01
-6.08191192e-01 -1.40487194e-01 -9.70127404e-01 1.22494221e+00
-1.64082959e-01 -3.91807646e-01 4.10460353e-01 2.33658865e-01
-6.21925831e-01 7.92482257e-01 -1.11512232e+00 -3.14849734e-01
1.88436687e-01 8.50162804e-02 -1.93190962e-01 -3.00339282e-01
-1.16366696e+00 3.97918850e-01 2.72077173e-01 3.10128689e-01
-9.13818479e-01 -4.71507937e-01 -5.49783170e-01 -8.48657638e-03
9.87535954e-01 -1.56109005e-01 6.11445665e-01 -5.87723136e-01
-8.82961988e-01 -2.46129438e-01 -3.84203568e-02 -2.12361038e-01
1.69718415e-01 6.70538843e-02 -1.02660382e+00 1.16153128e-01
1.28650874e-01 -5.56861460e-01 8.97807539e-01 -9.88197386e-01
-5.43100655e-01 -3.88729364e-01 -1.38929427e-01 -5.82369685e-01
-5.25517583e-01 -8.57343227e-02 -1.48770690e-01 -6.79365456e-01
6.39080644e-01 -8.97672355e-01 -3.03781360e-01 -2.72407264e-01
-9.34139788e-01 -2.26472795e-01 1.38867927e+00 -5.13092935e-01
1.52920282e+00 -2.05905676e+00 -2.80153900e-02 9.37239289e-01
4.09045011e-01 1.78468321e-02 1.30459294e-01 8.00637424e-01
-2.29775533e-01 2.05775604e-01 -4.78782088e-01 4.75454092e-01
-4.43409711e-01 5.60738266e-01 -5.50388932e-01 7.85141349e-01
5.65750599e-01 3.55277658e-01 -1.04814124e+00 5.04215658e-02
2.30420008e-01 3.90833914e-02 -1.63048267e-01 4.79512177e-02
1.04347929e-01 3.68368596e-01 -7.99190640e-01 8.52192581e-01
4.02753413e-01 -5.21555305e-01 1.18619591e-01 -6.66655675e-02
-4.27121557e-02 -1.87820271e-01 -1.37710202e+00 6.69196784e-01
-7.99018592e-02 5.94378710e-01 4.80769016e-02 -1.25883818e+00
8.44507277e-01 4.20404434e-01 1.00410163e+00 -4.99269605e-01
-1.13120914e-01 4.10213023e-02 3.24377596e-01 -4.20302540e-01
-3.17949317e-02 3.33365321e-01 -1.33781716e-01 5.57259679e-01
-1.83800668e-01 1.91207677e-01 3.45540822e-01 2.66104579e-01
2.01056242e+00 -8.70187283e-01 2.18280271e-01 -3.61959368e-01
5.95051467e-01 -1.45113811e-01 9.21944618e-01 5.42178810e-01
-8.30706432e-02 6.28791898e-02 1.09330142e+00 -2.71682829e-01
-6.13151550e-01 -1.13921320e+00 -1.07798025e-01 8.93644914e-02
3.56670022e-01 -4.18031901e-01 -1.74660295e-01 -8.11819673e-01
2.98272729e-01 4.44807172e-01 -5.48310101e-01 -6.11304224e-01
-4.25910920e-01 -8.12807083e-01 4.73742932e-01 3.66394728e-01
4.03349340e-01 -7.67351925e-01 -1.99798435e-01 3.80422741e-01
4.67612548e-03 -8.51632595e-01 -7.09693134e-02 8.83098692e-02
-8.55064809e-01 -1.73890376e+00 -1.57396749e-01 -1.85426190e-01
1.06486905e+00 2.93753654e-01 7.92452931e-01 3.46947521e-01
-6.46616340e-01 3.45119834e-01 -2.09771261e-01 -1.99885249e-01
-5.68899035e-01 -5.16345203e-01 3.84801716e-01 2.44976252e-01
-2.40219742e-01 -6.50564790e-01 -3.07074577e-01 8.01082373e-01
-9.62823689e-01 -6.36114478e-01 4.45391446e-01 8.78917217e-01
4.30460155e-01 1.27272713e+00 7.06442654e-01 -6.40638471e-01
7.54478693e-01 -9.32536542e-01 -8.14928889e-01 3.08988959e-01
-9.45108593e-01 -1.36770546e-01 8.65730345e-01 -5.07734358e-01
-4.39314723e-01 -2.66911566e-01 5.15601635e-01 -6.58082068e-01
-7.71718100e-02 9.57879961e-01 -3.28391306e-02 1.25732839e-01
3.18185836e-01 -1.49563246e-03 3.40700038e-02 -2.53993571e-01
-9.56325606e-02 2.98689306e-01 3.37519079e-01 -3.38918656e-01
1.23141730e+00 4.84825075e-01 6.40581787e-01 -1.02415240e+00
-4.16154742e-01 -6.30492032e-01 -5.33133328e-01 -5.87685645e-01
2.52465606e-01 -4.58014399e-01 -7.07226098e-01 6.16472721e-01
-8.94836068e-01 4.41509709e-02 -1.35180265e-01 3.83881211e-01
6.39872029e-02 4.16561574e-01 -7.77297854e-01 -1.01835871e+00
1.36429071e-01 -8.47838402e-01 7.65467048e-01 -2.87873875e-02
-2.83603936e-01 -1.18471801e+00 8.09758157e-02 -3.71156663e-01
3.85230005e-01 7.72525251e-01 1.16715848e+00 -4.09169674e-01
-5.50220549e-01 -7.82957852e-01 -1.38457388e-01 3.14003795e-01
8.60878944e-01 5.38422763e-01 -3.74855846e-01 -5.19590616e-01
3.83715928e-02 2.69127339e-01 4.14854139e-01 2.68869549e-01
1.07860315e+00 -3.92641276e-01 -6.34604752e-01 -7.86042064e-02
1.13928425e+00 3.40710610e-01 1.76040113e-01 -2.43752271e-01
7.48310149e-01 6.21533632e-01 5.85040629e-01 6.37367427e-01
-3.17248523e-01 3.59470367e-01 7.10325539e-01 -5.37169836e-02
3.60100597e-01 4.31923848e-03 6.37019694e-01 1.10054302e+00
-1.99693837e-03 -3.52575928e-01 -1.11499512e+00 5.04891217e-01
-1.94999874e+00 -1.01154959e+00 -4.93826985e-01 2.24026132e+00
2.85491705e-01 4.73987192e-01 -3.65673378e-02 6.17432237e-01
9.42587972e-01 1.63508591e-03 -8.85424674e-01 1.32861987e-01
-8.91658068e-02 -2.63481319e-01 4.31287080e-01 1.95275530e-01
-8.45268250e-01 1.66800708e-01 5.79404449e+00 5.16999781e-01
-1.02041829e+00 -2.47012630e-01 3.60592157e-01 2.05988601e-01
1.13796882e-01 4.03708033e-02 -2.85802633e-01 4.14758235e-01
9.87892330e-01 -4.82406855e-01 3.11084867e-01 5.51459610e-01
2.96411544e-01 -1.82486083e-02 -8.74624908e-01 4.68206137e-01
-2.02402547e-01 -8.96229327e-01 -2.16912404e-01 4.28924024e-01
7.85406351e-01 -3.90115567e-02 -9.98327658e-02 -2.83922613e-01
6.04140043e-01 -6.17016613e-01 2.36975461e-01 5.73969305e-01
3.01717132e-01 -9.26818252e-01 9.04398561e-01 3.70510548e-01
-1.50621092e+00 -3.71995032e-01 -1.02639511e-01 -9.00162086e-02
3.54248673e-01 1.34368432e+00 -7.34186113e-01 9.20415938e-01
6.88046455e-01 1.22591388e+00 -4.50855911e-01 1.01247132e+00
-2.76844084e-01 1.01684725e+00 -4.86545742e-01 2.17734665e-01
-4.27256152e-02 -2.70913601e-01 9.03280258e-01 5.85464537e-01
4.41324234e-01 -1.45720586e-01 5.77303469e-01 8.11117291e-01
3.93620521e-01 -5.16916752e-01 -8.64935756e-01 -4.28762883e-01
5.55812120e-01 1.03325593e+00 -1.39375710e+00 -2.24444456e-02
-3.35645139e-01 6.09777808e-01 -2.99481809e-01 4.55760568e-01
-7.41382182e-01 -4.28198695e-01 8.49029899e-01 1.22300461e-01
-2.42482126e-02 -4.72955078e-01 -7.83119276e-02 -9.79696095e-01
5.49993888e-02 -7.69550085e-01 5.60615301e-01 -3.53210717e-01
-1.65297413e+00 4.19240624e-01 1.07869893e-01 -1.42519760e+00
-4.44476783e-01 -4.46337581e-01 -8.96276355e-01 6.00372255e-01
-9.69291031e-01 -5.58037519e-01 -3.83574247e-01 9.21888471e-01
3.36834669e-01 -7.79272243e-02 6.39023364e-01 2.21169487e-01
-1.19391930e+00 -4.70604897e-02 4.41205092e-02 2.43240550e-01
3.01450074e-01 -1.07033896e+00 3.44760686e-01 1.27770305e+00
1.03072785e-01 6.24771893e-01 7.37101793e-01 -1.12824023e+00
-1.46549404e+00 -1.19975996e+00 1.63639277e-01 -2.31241032e-01
1.40609217e+00 -2.37860858e-01 -1.05418098e+00 4.30802345e-01
-1.27943039e-01 1.95907816e-01 2.61059433e-01 1.84199572e-01
-1.67179495e-01 -1.16226003e-01 -9.49574590e-01 6.10252023e-01
8.63876760e-01 -3.65391374e-01 -6.02331683e-02 6.24293327e-01
6.31042600e-01 2.89784223e-01 -9.61221099e-01 6.59642577e-01
2.35486180e-02 -7.40927100e-01 7.16356218e-01 -4.15922493e-01
2.53732532e-01 -5.52860320e-01 3.64110246e-02 -1.45917428e+00
-4.09474313e-01 -6.29495740e-01 -4.92737889e-01 1.30204725e+00
2.38918915e-01 -1.15471363e+00 2.88389415e-01 2.23809376e-01
-1.53166756e-01 -6.89899325e-01 -1.02639294e+00 -1.21896720e+00
-5.08683622e-01 -4.50579822e-01 6.93426847e-01 1.32923532e+00
1.31961599e-01 1.52418584e-01 -1.64141521e-01 8.48231733e-01
6.91760898e-01 1.08461179e-01 5.08245766e-01 -1.70576084e+00
-3.46006751e-01 -5.11274636e-01 -7.03381002e-01 -3.95356238e-01
-2.94636227e-02 -4.06834602e-01 2.42793839e-02 -1.12129223e+00
-1.14201643e-01 -5.25245011e-01 -7.28107631e-01 6.39187098e-01
-1.78455506e-02 -1.34240568e-01 -3.68078440e-01 4.98574853e-01
-2.83647507e-01 4.42730933e-01 7.63927400e-01 -9.50346962e-02
-3.42217721e-02 3.33355278e-01 -8.68221372e-02 8.21115375e-01
7.34448433e-01 -5.26061535e-01 -5.47277570e-01 2.78285801e-01
2.74076313e-01 3.88495654e-01 4.90530491e-01 -1.04604268e+00
2.92062581e-01 -4.16055858e-01 1.33538932e-01 -6.17386281e-01
-6.91514835e-02 -1.12826526e+00 3.82732570e-01 9.88876760e-01
2.06139699e-01 6.23610139e-01 1.87471315e-01 1.12452245e+00
-4.69358146e-01 2.53647029e-01 3.29668462e-01 2.31466413e-01
-5.48916399e-01 3.82278919e-01 -5.30070186e-01 -2.95583606e-01
1.17658997e+00 2.31668845e-01 -5.86187541e-01 -4.85825777e-01
-3.83813798e-01 5.02583921e-01 2.44588539e-01 7.53311574e-01
6.80915296e-01 -1.11122477e+00 -4.46035832e-01 4.34498876e-01
1.40874768e-02 -2.35846445e-01 1.32297009e-01 1.08699727e+00
-1.54977456e-01 2.11160034e-01 -8.92769254e-04 -9.59073246e-01
-1.20702076e+00 5.65981507e-01 1.17193274e-01 -3.00605088e-01
-7.04284549e-01 3.59893054e-01 3.32377516e-02 -2.86971796e-02
-1.10572599e-01 -3.29127163e-01 -4.48772982e-02 -2.64097899e-02
3.40197772e-01 7.16025949e-01 6.87261522e-02 -4.27716285e-01
-4.68771458e-01 4.20688570e-01 9.39189345e-02 2.46471122e-01
1.22479975e+00 -4.71667275e-02 -5.64410448e-01 8.91287982e-01
8.52814198e-01 -1.92351013e-01 -1.10608268e+00 -1.55211613e-01
2.78741747e-01 -4.64727253e-01 1.46168739e-01 -5.13074934e-01
-1.45063102e+00 4.66376454e-01 3.37479800e-01 1.18974078e+00
1.36199737e+00 5.41274855e-03 5.22001147e-01 3.58486563e-01
5.88567555e-01 -6.80252552e-01 3.46664131e-01 4.90512162e-01
6.72000051e-01 -7.75760531e-01 1.79512277e-02 -6.51824057e-01
-1.53111473e-01 1.14082074e+00 5.10667264e-01 -2.35325396e-01
1.13300884e+00 4.11546588e-01 -3.18705559e-01 -5.98573983e-01
-9.45800424e-01 1.45986900e-01 2.03005210e-01 2.14799687e-01
-2.30802912e-02 1.44325957e-01 1.27116755e-01 7.87949339e-02
1.09628983e-01 -6.68073416e-01 5.97128510e-01 1.13890791e+00
-3.14477086e-01 -8.90379488e-01 -5.45311749e-01 7.86499679e-01
-2.12428480e-01 3.50785524e-01 -6.11267567e-01 7.18590558e-01
-3.38959545e-01 1.21807230e+00 1.13410480e-01 -6.89793348e-01
6.12752140e-01 -1.39474049e-01 -1.44979015e-01 -3.69007736e-01
-1.16697825e-01 -1.68505069e-02 -2.38732040e-01 -8.57767940e-01
-6.38155714e-02 -7.94529796e-01 -1.19927323e+00 -4.33438987e-01
-8.65209997e-01 9.89552066e-02 3.89632106e-01 1.02115691e+00
4.51443553e-01 1.11664033e+00 1.13607800e+00 -3.68675441e-01
-4.62222427e-01 -8.83582652e-01 -9.64946091e-01 2.72396713e-01
5.50154924e-01 -9.29342270e-01 -7.89876342e-01 -3.10193658e-01] | [7.293174743652344, 2.9400551319122314] |
eef06fec-ff8c-4548-a616-fc0a79b594ee | lidar-sensor-modeling-and-data-augmentation | 1905.07290 | null | https://arxiv.org/abs/1905.07290v1 | https://arxiv.org/pdf/1905.07290v1.pdf | LiDAR Sensor modeling and Data augmentation with GANs for Autonomous driving | In the autonomous driving domain, data collection and annotation from real vehicles are expensive and sometimes unsafe. Simulators are often used for data augmentation, which requires realistic sensor models that are hard to formulate and model in closed forms. Instead, sensors models can be learned from real data. The main challenge is the absence of paired data set, which makes traditional supervised learning techniques not suitable. In this work, we formulate the problem as image translation from unpaired data and employ CycleGANs to solve the sensor modeling problem for LiDAR, to produce realistic LiDAR from simulated LiDAR (sim2real). Further, we generate high-resolution, realistic LiDAR from lower resolution one (real2real). The LiDAR 3D point cloud is processed in Bird-eye View and Polar 2D representations. The experimental results show a high potential of the proposed approach. | ['Mohamed Zahran', 'Nader Essam', 'Ibrahim Sobh', 'Ahmad El Sallab'] | 2019-05-17 | null | null | null | null | ['sensor-modeling', 'point-cloud-generation'] | ['computer-vision', 'computer-vision'] | [ 5.42105854e-01 3.05751204e-01 -2.78825670e-01 -6.49376810e-01
-5.52161396e-01 -4.63252425e-01 7.04418659e-01 4.03061276e-03
-6.55794322e-01 1.02741623e+00 -4.56841469e-01 -2.79972523e-01
8.87424946e-02 -1.09450841e+00 -1.09818637e+00 -5.26279330e-01
1.37240916e-01 7.56540835e-01 6.47580251e-02 -3.26119423e-01
-5.32841161e-02 7.08189547e-01 -2.17753077e+00 -3.92304182e-01
1.02965558e+00 7.63663411e-01 5.92059553e-01 3.99261326e-01
-2.22205803e-01 2.18574405e-01 -1.75325066e-01 -8.09258372e-02
6.65302157e-01 -6.57854006e-02 -9.03307498e-02 2.06878573e-01
5.02442300e-01 -3.88178527e-01 -1.22561149e-01 1.25036788e+00
2.73983598e-01 -2.12183923e-01 5.36263585e-01 -1.83468986e+00
-2.58208439e-02 7.55571350e-02 -6.95976019e-01 -3.94665569e-01
1.54062033e-01 1.32137075e-01 4.10979807e-01 -6.93625152e-01
5.51687777e-01 1.30201662e+00 5.80012679e-01 5.06741226e-01
-1.33920133e+00 -1.07935464e+00 -8.21108893e-02 4.90333401e-02
-1.47329617e+00 -4.16959167e-01 8.43989491e-01 -4.08338040e-01
1.91700876e-01 9.06514153e-02 7.68175125e-01 1.00278974e+00
-6.10508509e-02 4.44060475e-01 1.46164548e+00 -2.26152256e-01
1.63216844e-01 5.84459543e-01 -1.65669233e-01 4.63089705e-01
6.16358995e-01 7.22452402e-01 -4.85601634e-01 2.90603787e-01
5.73365986e-01 1.77162096e-01 -7.67033175e-02 -5.98543167e-01
-1.11732209e+00 8.73676777e-01 5.31879544e-01 -4.23895746e-01
-2.75128633e-01 4.29604501e-02 1.28979027e-01 2.04125687e-01
1.22327156e-01 1.50526345e-01 -2.52781719e-01 1.36775747e-01
-7.81435251e-01 4.98494893e-01 3.00053686e-01 1.48127425e+00
1.33485389e+00 3.40335488e-01 6.78235412e-01 2.54843444e-01
5.12628317e-01 1.22596204e+00 3.14920098e-01 -8.91017437e-01
4.12653923e-01 4.80512649e-01 1.20774962e-01 -7.80351341e-01
-1.87430799e-01 7.36438781e-02 -9.44731832e-01 6.98783994e-01
9.70644876e-02 -2.26038203e-01 -1.00349677e+00 1.52821696e+00
3.58451635e-01 5.19494116e-01 5.42612672e-01 8.06866527e-01
9.06425655e-01 6.91594362e-01 -2.39307716e-01 -3.23105335e-01
1.04561996e+00 -6.17041647e-01 -7.90464878e-01 -5.21304667e-01
6.78267598e-01 -5.87374508e-01 9.24847603e-01 2.75286734e-01
-5.28925955e-01 -8.64309072e-01 -1.30242860e+00 1.58775281e-02
-6.41581118e-01 -6.97409287e-02 3.80713612e-01 4.94467467e-01
-7.23730624e-01 2.19880849e-01 -5.98093927e-01 -3.83739144e-01
2.53436357e-01 4.25582796e-01 -6.39063597e-01 -2.87978262e-01
-1.10916948e+00 1.07893407e+00 7.58039951e-01 1.71052232e-01
-9.49599504e-01 -5.51113963e-01 -1.20523345e+00 -6.49516404e-01
4.50426519e-01 -4.08695817e-01 9.46503699e-01 -4.67838287e-01
-1.28001904e+00 8.96488488e-01 -1.90148517e-01 -7.63505578e-01
6.99006379e-01 -8.05736408e-02 -3.77173126e-01 -3.66177857e-01
1.04568221e-01 1.15092957e+00 8.39403927e-01 -1.83199799e+00
-7.60956168e-01 -4.63322192e-01 -1.49123490e-01 1.99359313e-01
1.62285000e-01 -6.07334971e-01 -3.96710075e-02 3.47904488e-02
1.98857009e-01 -9.91188347e-01 -4.47146744e-01 1.54117703e-01
-1.53487280e-01 2.16695428e-01 1.35520792e+00 -1.43434212e-01
2.94208288e-01 -2.07605267e+00 -3.95060807e-01 1.96402878e-01
-4.38639037e-02 3.73936415e-01 -1.97207823e-01 4.65406179e-02
3.07142884e-02 2.76775248e-02 -3.05951208e-01 -4.40069437e-01
-2.29641750e-01 9.36687529e-01 -5.15718699e-01 4.05461490e-01
4.31166202e-01 7.52212048e-01 -7.88552523e-01 -7.75897920e-01
6.86795592e-01 4.71239418e-01 -3.98004115e-01 3.84265900e-01
-3.33378613e-01 9.23503458e-01 -3.95098776e-01 3.92977804e-01
1.24642718e+00 6.16976857e-01 -2.25349694e-01 -2.75192320e-01
-5.63407362e-01 -9.40094516e-03 -1.38538420e+00 1.54435706e+00
-5.83368778e-01 5.89093328e-01 1.16929255e-01 -9.73583639e-01
1.43768752e+00 -6.20913580e-02 3.46399873e-01 -7.64937282e-01
1.88917831e-01 2.36543626e-01 -3.28666680e-02 -5.88127613e-01
7.39230990e-01 -3.16298723e-01 -1.40814438e-01 -1.48732159e-02
-3.00465584e-01 -1.10274339e+00 5.71153080e-03 -2.26054043e-01
3.79877687e-01 4.33660179e-01 9.35901776e-02 6.64361194e-02
5.40979087e-01 3.00224930e-01 6.93032742e-01 4.18116927e-01
1.17948018e-01 6.81197643e-01 1.87001005e-02 -1.75934434e-01
-1.34092152e+00 -9.50145245e-01 -2.83199310e-01 5.58764748e-02
4.25556451e-01 -1.04774386e-01 -4.37623292e-01 -3.75033438e-01
2.19378665e-01 7.83207238e-01 -1.19091675e-01 -9.85388160e-02
-5.65735459e-01 -3.76270890e-01 5.31106889e-01 3.12263280e-01
7.01989591e-01 -6.46005094e-01 -8.97347152e-01 -4.71103154e-02
-1.10197933e-02 -1.57426357e+00 2.73724049e-01 2.21291259e-01
-8.77657890e-01 -8.63246799e-01 -5.22249341e-02 -6.09755635e-01
5.56783140e-01 5.17117202e-01 7.00133204e-01 -7.27073923e-02
-1.54272407e-01 -1.52680874e-01 -9.52025056e-02 -9.71956074e-01
-4.29324150e-01 -3.77317637e-01 4.35457170e-01 -1.23195127e-02
4.17695045e-01 -5.97083449e-01 3.28317136e-02 2.92197376e-01
-9.53735530e-01 2.55717725e-01 7.64094353e-01 7.38556802e-01
1.04214275e+00 7.49837756e-02 3.56499851e-01 -7.83872902e-01
9.15576518e-02 -2.04436734e-01 -1.27368236e+00 -3.36049736e-01
-7.92709053e-01 5.24128973e-02 6.56100392e-01 -2.10967481e-01
-8.82426023e-01 6.79383576e-01 -5.39961867e-02 -6.96667552e-01
-5.58146536e-01 4.00782853e-01 -5.18377185e-01 -1.85063705e-01
6.17569566e-01 2.28782654e-01 4.16684419e-01 -3.29280943e-01
4.61242676e-01 8.62535536e-01 8.65929365e-01 -3.70771557e-01
1.35298550e+00 7.13753045e-01 4.78331000e-01 -1.25351143e+00
-5.30513644e-01 -2.66526878e-01 -9.26746249e-01 -1.39454007e-01
7.53363371e-01 -1.18827689e+00 -5.13343930e-01 1.85251191e-01
-1.11671519e+00 -1.06491148e-02 -6.75257802e-01 8.00616443e-01
-7.27520645e-01 2.93116391e-01 4.11333144e-01 -1.05197847e+00
1.57173857e-01 -1.26619065e+00 1.25833189e+00 3.53573740e-01
1.53550670e-01 -5.55579484e-01 -5.43741211e-02 3.52060795e-01
6.89237239e-03 5.86027503e-01 5.44852495e-01 -3.45214248e-01
-9.61515427e-01 -3.57060581e-01 -3.32859188e-01 4.78263080e-01
-2.30792724e-02 1.28402770e-01 -1.04346848e+00 -1.08542532e-01
-6.11046180e-02 -4.62779462e-01 5.01668930e-01 -2.72005815e-02
9.90396678e-01 7.44132623e-02 -5.05653858e-01 7.20546603e-01
1.35315073e+00 1.26684234e-01 5.37947714e-01 3.18845138e-02
7.27874935e-01 7.42226839e-01 1.39173079e+00 1.72402278e-01
6.29245579e-01 5.76487660e-01 9.33984399e-01 -1.69683203e-01
1.38377892e-02 -7.02451169e-01 2.42868721e-01 6.61131382e-01
1.20075911e-01 -7.60530829e-02 -1.00828290e+00 4.86912668e-01
-1.73533571e+00 -6.80752158e-01 -7.51167655e-01 2.40279889e+00
4.65770125e-01 3.32249016e-01 -2.19758138e-01 2.85492569e-01
5.23844302e-01 -1.19490430e-01 -4.64283139e-01 -1.83929071e-01
-2.68041164e-01 1.57324955e-01 9.15056825e-01 6.88395619e-01
-8.66252840e-01 1.03014135e+00 5.81141615e+00 4.64974314e-01
-1.26147497e+00 4.18927744e-02 1.66069083e-02 2.74207056e-01
-4.22289312e-01 3.01457316e-01 -1.07637894e+00 4.82872605e-01
1.07859969e+00 -2.04299614e-01 5.03869839e-02 8.31166089e-01
5.68653226e-01 -3.77943367e-01 -1.28056145e+00 1.27935767e+00
-4.15497907e-02 -1.06877649e+00 -4.18822169e-02 4.54533577e-01
5.98041475e-01 2.03546792e-01 5.77698881e-03 2.39971042e-01
4.09289181e-01 -1.01670814e+00 5.90844929e-01 5.21079838e-01
9.82934058e-01 -7.21181691e-01 8.09360266e-01 1.01585186e+00
-9.90004063e-01 6.58041909e-02 -5.34075022e-01 -1.96249977e-01
3.39961767e-01 3.99567276e-01 -1.13382137e+00 6.83564246e-01
5.70952713e-01 7.95864344e-01 -3.55757058e-01 9.38046813e-01
-2.14314297e-01 2.60146141e-01 -7.81092227e-01 1.55372411e-01
1.38766062e-03 -5.92015624e-01 4.80073810e-01 7.21567810e-01
5.02065659e-01 -1.25914603e-01 3.90784532e-01 7.96064496e-01
7.25955069e-02 -1.59432575e-01 -1.59443724e+00 8.03845003e-02
5.99321246e-01 1.30053103e+00 -2.22388908e-01 -2.37273544e-01
-1.74456671e-01 4.28454369e-01 -1.94849581e-01 4.94559109e-02
-8.24433267e-01 -9.14897919e-02 6.33839846e-01 3.26769441e-01
-1.65957417e-02 -4.34661865e-01 -2.92279512e-01 -9.18444932e-01
-9.55908839e-03 -7.02688873e-01 -2.11371928e-01 -1.01424670e+00
-9.08221364e-01 4.39923733e-01 3.67748648e-01 -1.77466357e+00
-6.24449968e-01 -5.94015241e-01 -3.24354023e-01 1.00610185e+00
-1.86472344e+00 -1.51312792e+00 -8.82460713e-01 5.59816539e-01
2.62572229e-01 -1.60607457e-01 7.35790014e-01 4.12392884e-01
-1.93699852e-01 2.18318045e-01 -8.39420259e-02 -2.57139146e-01
6.00560188e-01 -8.55775893e-01 2.53339380e-01 7.90467024e-01
2.45174579e-02 -2.43866025e-03 1.08354890e+00 -6.72176301e-01
-1.52289450e+00 -1.51349878e+00 7.34210730e-01 -2.27373004e-01
4.87321287e-01 -5.40648818e-01 -6.90821409e-01 7.13598847e-01
7.00481981e-02 2.81708807e-01 1.67008072e-01 -5.73825240e-01
-1.03602692e-01 -4.98202771e-01 -1.42752886e+00 4.44230288e-01
9.33404386e-01 -3.44638646e-01 -4.67048109e-01 2.87697881e-01
6.83386147e-01 -3.62105846e-01 -4.58256006e-01 7.02071846e-01
1.95584640e-01 -7.84623086e-01 8.04140270e-01 -3.63239527e-01
1.43052384e-01 -7.86872983e-01 -4.44926083e-01 -1.30349755e+00
4.42565113e-01 -2.99179494e-01 2.55241096e-01 1.31660461e+00
2.61638969e-01 -7.90652394e-01 1.05928230e+00 1.12370320e-01
-1.21173680e-01 -8.42123628e-02 -1.02381158e+00 -9.02728438e-01
-1.81542139e-03 -7.04745591e-01 7.50802815e-01 8.24519515e-01
-6.88281715e-01 5.53826630e-01 -4.72144991e-01 3.62571090e-01
8.83370459e-01 8.65354091e-02 1.45078623e+00 -1.65729690e+00
1.55842185e-01 2.10802868e-01 -8.29007506e-01 -1.01515758e+00
3.98308247e-01 -7.00220823e-01 4.14221376e-01 -1.15994084e+00
-3.56730551e-01 -7.74573445e-01 4.37692523e-01 1.66824147e-01
3.48842978e-01 3.86379570e-01 1.39206499e-01 1.11799026e-02
1.64108247e-01 8.03362608e-01 1.00203598e+00 -1.37851149e-01
8.50364491e-02 1.61633864e-01 -3.44614059e-01 9.63021696e-01
8.45182002e-01 -4.34418261e-01 -7.39598334e-01 -3.91046226e-01
8.85962620e-02 -8.25781934e-03 5.15364826e-01 -1.19355249e+00
8.87290612e-02 -3.49659055e-01 7.46477097e-02 -1.28726041e+00
7.96547234e-01 -1.44687557e+00 3.17960858e-01 2.67357111e-01
2.07554668e-01 2.50383973e-01 1.45571306e-01 4.78789240e-01
-4.37900454e-01 -3.93126011e-01 8.61833572e-01 -4.88777943e-02
-7.51792192e-01 5.65000176e-01 -1.60255134e-01 -2.60571748e-01
1.32612717e+00 -4.62584823e-01 8.04431662e-02 -3.77714008e-01
-3.54461581e-01 4.86121267e-01 5.84617615e-01 4.21423346e-01
7.94547558e-01 -1.52393472e+00 -6.74090028e-01 7.03961432e-01
5.08287728e-01 8.92578006e-01 -8.29545110e-02 6.92772746e-01
-4.92058426e-01 2.01026648e-01 -3.73002499e-01 -1.03120041e+00
-1.32537246e+00 6.64010346e-01 6.90537170e-02 2.75509030e-01
-4.58036989e-01 3.26023996e-01 -1.37780691e-04 -1.08237278e+00
-1.79073960e-02 -3.34731549e-01 -2.42454484e-01 1.21492848e-01
1.74766228e-01 4.16694442e-03 -3.95272207e-03 -1.21579766e+00
-1.22167118e-01 7.53809333e-01 3.80717546e-01 -2.67964393e-01
1.16977060e+00 -3.07713896e-01 7.32355490e-02 5.60192049e-01
1.01069677e+00 3.55238244e-02 -1.07876384e+00 -2.82948524e-01
-1.49222612e-01 -6.32465243e-01 2.44365074e-02 -2.13375241e-02
-9.08729434e-01 1.27944982e+00 9.89644289e-01 -7.03472691e-03
7.83229172e-01 -3.05107355e-01 5.72645545e-01 5.87343156e-01
4.97352779e-01 -1.00026941e+00 -3.67993385e-01 4.83859956e-01
6.88869417e-01 -1.63571274e+00 -1.32370247e-02 -6.34694993e-01
-4.93324608e-01 8.38661313e-01 8.97523046e-01 -7.06859306e-02
6.03028953e-01 5.95443845e-01 2.27427602e-01 1.31931677e-01
-4.80494738e-01 -5.54326177e-01 -3.02086771e-01 1.04919887e+00
-9.94921327e-02 7.51577467e-02 -1.69686917e-02 3.21945958e-02
-7.06698000e-01 1.48817152e-01 7.80591011e-01 9.49735284e-01
-4.92011160e-01 -1.20771003e+00 -6.14859104e-01 1.96525872e-01
3.18852514e-01 2.85728216e-01 -1.37852892e-01 9.92100596e-01
7.92747796e-01 9.00755048e-01 3.91913205e-01 -6.02591932e-01
5.26314497e-01 -9.56899077e-02 3.11963528e-01 -4.75900620e-01
2.25852877e-01 -8.64359513e-02 -1.14574069e-02 -2.42084906e-01
-4.72816020e-01 -7.49176323e-01 -1.36455512e+00 -3.11742723e-01
-3.18188310e-01 1.66912094e-01 1.11579466e+00 7.68567085e-01
1.88146755e-01 2.64276654e-01 8.01373482e-01 -9.95030642e-01
-5.27625144e-01 -7.60094523e-01 -3.77586305e-01 2.51242667e-01
5.90090513e-01 -9.65929389e-01 -1.04281843e-01 1.52188942e-01] | [7.9899001121521, -2.519354820251465] |
c4ff3a68-6da1-4a46-8164-79b85ee9a213 | 3d-face-parsing-via-surface-parameterization | 2206.09221 | null | https://arxiv.org/abs/2206.09221v1 | https://arxiv.org/pdf/2206.09221v1.pdf | 3D Face Parsing via Surface Parameterization and 2D Semantic Segmentation Network | Face parsing assigns pixel-wise semantic labels as the face representation for computers, which is the fundamental part of many advanced face technologies. Compared with 2D face parsing, 3D face parsing shows more potential to achieve better performance and further application, but it is still challenging due to 3D mesh data computation. Recent works introduced different methods for 3D surface segmentation, while the performance is still limited. In this paper, we propose a method based on the "3D-2D-3D" strategy to accomplish 3D face parsing. The topological disk-like 2D face image containing spatial and textural information is transformed from the sampled 3D face data through the face parameterization algorithm, and a specific 2D network called CPFNet is proposed to achieve the semantic segmentation of the 2D parameterized face data with multi-scale technologies and feature aggregation. The 2D semantic result is then inversely re-mapped to 3D face data, which finally achieves the 3D face parsing. Experimental results show that both CPFNet and the "3D-2D-3D" strategy accomplish high-quality 3D face parsing and outperform state-of-the-art 2D networks as well as 3D methods in both qualitative and quantitative comparisons. | ['Guangquan Zhou', 'Jing Jin', 'Zongpu Yu', 'Yangang Wang', 'Ping Zhou', 'Wenyuan Sun'] | 2022-06-18 | null | null | null | null | ['face-parsing', '2d-semantic-segmentation'] | ['computer-vision', 'computer-vision'] | [-4.61700782e-02 3.54819983e-01 -5.69239557e-02 -6.79643273e-01
-4.10890490e-01 -3.23296756e-01 2.46021211e-01 -5.98226368e-01
1.89404130e-01 1.21451311e-01 -1.30383193e-01 -3.12789269e-02
-8.63444954e-02 -1.10745931e+00 -4.94518787e-01 -4.91891056e-01
2.86681298e-02 9.43633556e-01 3.32491666e-01 -1.54622048e-01
1.26209870e-01 1.36744630e+00 -1.82272172e+00 4.01377641e-02
4.09869462e-01 1.43412578e+00 -1.07980020e-01 2.66571313e-01
-1.13688648e+00 -6.13008440e-01 -3.65164757e-01 -4.26214010e-01
5.74415565e-01 -3.85636032e-01 -6.80756032e-01 3.17508757e-01
6.19921684e-01 -1.38640985e-01 2.12116331e-01 1.15505445e+00
5.72369576e-01 -2.26389110e-01 7.08315551e-01 -1.24096918e+00
-4.68195856e-01 -2.61009410e-02 -9.55052674e-01 -3.99832428e-01
5.41790247e-01 -3.88819635e-01 2.09907457e-01 -1.04830456e+00
6.44450068e-01 2.16254568e+00 8.04939449e-01 1.12169433e+00
-8.25947285e-01 -8.82626891e-01 1.59551784e-01 -4.99739677e-01
-1.29114866e+00 -3.07022393e-01 1.42613125e+00 -3.40965867e-01
7.01189995e-01 -4.76333080e-03 7.10206151e-01 4.51402515e-01
-1.01917312e-02 4.05667067e-01 1.21624815e+00 -3.41507137e-01
6.67600334e-02 -3.12088847e-01 6.73827827e-02 1.35109949e+00
-4.84590828e-02 -2.01830655e-01 -4.51351911e-01 -4.77221934e-03
1.47281945e+00 -9.18254405e-02 2.05002367e-01 -2.48171851e-01
-3.78820926e-01 6.34391725e-01 2.43521273e-01 1.08824700e-01
-1.66713327e-01 -9.74040478e-02 1.58090711e-01 1.14235058e-01
1.26129889e+00 -1.89997256e-01 -5.67300975e-01 2.79051006e-01
-9.57625985e-01 1.79908693e-01 6.76969528e-01 9.34664190e-01
1.06899214e+00 -2.60124188e-02 2.18709350e-01 1.02861035e+00
5.87770939e-01 7.47606695e-01 2.24708654e-02 -1.26142681e+00
2.58871228e-01 1.26473379e+00 -4.69111115e-01 -1.03771663e+00
-7.23069429e-01 2.19997153e-01 -7.62796819e-01 4.80981588e-01
1.70517296e-01 6.75627515e-02 -1.32756078e+00 1.37123919e+00
1.03944099e+00 1.10575415e-01 -1.21276267e-01 8.20114136e-01
1.35992146e+00 5.04249811e-01 -1.83334090e-02 -1.95888445e-01
1.57177567e+00 -5.05545795e-01 -4.92318630e-01 2.72665918e-02
1.18505746e-01 -6.83469415e-01 9.69827771e-01 6.38066456e-02
-1.36603916e+00 -6.00526869e-01 -5.78605115e-01 -1.18922740e-01
-3.15945983e-01 5.33155426e-02 6.26292408e-01 1.01361990e+00
-1.32370234e+00 3.58641356e-01 -6.23453557e-01 -4.77814347e-01
1.11519885e+00 8.18001926e-01 -6.76241517e-01 2.44866945e-02
-8.52990091e-01 3.92250538e-01 1.40647203e-01 1.44336775e-01
-5.40757418e-01 -7.55494356e-01 -8.95343900e-01 -2.13759407e-01
2.79227376e-01 -6.11526728e-01 9.16941524e-01 -4.50487822e-01
-1.75333941e+00 1.56137741e+00 -4.16074485e-01 3.27310920e-01
3.51402342e-01 2.07206532e-01 -1.10791437e-01 4.96973544e-01
1.55979723e-01 1.03045976e+00 9.08723354e-01 -1.35638893e+00
-5.30545235e-01 -1.13337719e+00 -1.52531946e-02 2.55998343e-01
-1.49372578e-01 1.13003582e-01 -9.98120666e-01 -4.00811255e-01
9.10166204e-01 -5.10668635e-01 6.60954341e-02 3.91564429e-01
-2.53980398e-01 -7.62154281e-01 1.58363330e+00 -5.92839658e-01
5.78062832e-01 -1.94201851e+00 1.46216035e-01 4.59064424e-01
2.18219250e-01 2.45817304e-01 -7.32198730e-02 -1.25312194e-01
1.38343722e-01 4.73977178e-01 -4.62357670e-01 -6.35477543e-01
-1.45181417e-01 1.62202075e-01 3.38966578e-01 3.41145843e-01
5.05169034e-01 6.82850003e-01 -4.18254346e-01 -9.61400211e-01
2.53301144e-01 5.66659212e-01 -6.33298934e-01 1.50241315e-01
-4.51666594e-01 5.21417439e-01 -9.17650938e-01 1.17590451e+00
1.32790601e+00 8.64791721e-02 1.01525955e-01 -3.80563766e-01
-1.27770531e-03 -2.73474813e-01 -1.10414684e+00 1.73345768e+00
-3.17289263e-01 3.77410511e-03 7.48006642e-01 -1.02084804e+00
1.42634225e+00 1.80273816e-01 7.18230844e-01 -8.14193010e-01
3.01620126e-01 2.19520450e-01 -7.36877203e-01 -4.82352972e-01
-1.76287711e-01 -3.60186160e-01 -1.62111357e-01 1.18315421e-01
1.00035714e-02 -7.37029076e-01 -2.56032526e-01 -2.09025890e-01
3.57825249e-01 5.12379944e-01 -2.69854456e-01 -6.31361604e-01
8.39544177e-01 -9.19020027e-02 7.74651706e-01 -7.58373737e-02
2.81137936e-02 5.20645738e-01 9.31860268e-01 -5.35465658e-01
-8.17415953e-01 -1.05627823e+00 -5.03111422e-01 4.31158960e-01
4.26675409e-01 -1.17526138e-02 -1.53725135e+00 -1.04672527e+00
2.75760502e-01 -1.46811586e-02 -5.50158978e-01 3.06634724e-01
-9.41134334e-01 -6.39364898e-01 4.43815947e-01 4.40760046e-01
1.01303554e+00 -9.13247228e-01 -2.24377856e-01 -8.06039199e-02
1.74501657e-01 -9.94561136e-01 -4.43516403e-01 -2.31692284e-01
-1.04695237e+00 -1.17732680e+00 -6.46189570e-01 -1.24312997e+00
1.01483929e+00 1.58260107e-01 9.09169257e-01 3.70004445e-01
-3.12385499e-01 4.89048839e-01 -2.09685862e-01 -3.01972091e-01
-3.03080827e-01 -1.55113637e-01 2.11966202e-01 -5.45611344e-02
3.57595623e-01 -5.98889589e-01 -5.21850705e-01 5.13326585e-01
-7.39423692e-01 -6.18645363e-02 2.90820390e-01 4.79485929e-01
1.04343486e+00 3.44466686e-01 6.66281521e-01 -8.63247752e-01
3.73111784e-01 -5.31518757e-02 -6.58973813e-01 1.29285693e-01
-3.81417066e-01 -1.90212160e-01 4.60022360e-01 -1.44295558e-01
-1.38416374e+00 2.56208092e-01 -5.24053097e-01 -6.22082055e-01
-3.82201850e-01 -7.36626834e-02 -8.78050745e-01 -3.93837035e-01
8.02623853e-02 -5.61680458e-02 5.83298802e-01 -8.44517827e-01
1.85672432e-01 7.47287214e-01 2.40827173e-01 -9.36809063e-01
8.14732492e-01 7.51920700e-01 6.01724088e-01 -1.09313476e+00
-7.66720772e-01 1.68937985e-02 -1.10677993e+00 -4.57970470e-01
1.31379187e+00 -5.39611578e-01 -7.91541874e-01 9.02102768e-01
-1.44777846e+00 -2.66535550e-01 -1.15355246e-01 -1.91992093e-02
-5.51777065e-01 4.16702867e-01 -5.44088781e-01 -8.21166337e-01
-4.39273119e-01 -1.22490036e+00 1.71676540e+00 4.82058793e-01
2.90464073e-01 -9.36570644e-01 -5.12315869e-01 5.07284582e-01
7.58501282e-03 3.51060629e-01 1.42983651e+00 -2.00994983e-01
-3.38777423e-01 -5.83622195e-02 -4.32063371e-01 3.99327606e-01
1.81798160e-01 1.38201430e-01 -9.23128188e-01 1.63869932e-01
1.75613075e-01 -2.86667403e-02 3.62888485e-01 5.02358198e-01
1.39418662e+00 2.55236235e-02 -5.50314128e-01 7.65352726e-01
1.18125117e+00 2.73910433e-01 4.72907215e-01 -2.66456962e-01
8.27524006e-01 1.17462826e+00 5.17455697e-01 5.04111387e-02
4.61573124e-01 6.57202423e-01 6.19899929e-01 -2.37563714e-01
-6.15346491e-01 -3.82179171e-01 -2.85518970e-02 6.42083943e-01
-1.31746707e-02 2.00695589e-01 -9.12291229e-01 1.50587782e-02
-1.20499015e+00 -4.79762346e-01 -1.83315739e-01 1.89768970e+00
4.11426634e-01 3.72915342e-02 1.50537625e-01 8.41436815e-03
1.08345234e+00 -5.63730001e-02 -6.41593754e-01 -4.45701778e-01
-9.90734100e-02 6.53475761e-01 1.12351100e-03 4.31790680e-01
-8.83903205e-01 1.34817886e+00 5.82504463e+00 1.02339256e+00
-9.99230325e-01 1.07365958e-01 8.41742814e-01 2.61956334e-01
-3.04257661e-01 -2.94842333e-01 -1.32861710e+00 3.74628752e-01
3.97109896e-01 3.54255408e-01 9.49428752e-02 7.00190723e-01
1.29016832e-01 8.36297031e-03 -8.56250823e-01 1.28614569e+00
1.63070396e-01 -1.39485455e+00 4.35351700e-01 2.53488898e-01
4.68000591e-01 -6.10861719e-01 -1.26457766e-01 -4.51283045e-02
-2.91644663e-01 -1.11100817e+00 6.19548917e-01 4.35390741e-01
1.47192371e+00 -8.10446024e-01 3.51627260e-01 3.21247071e-01
-1.74774599e+00 1.44016415e-01 -3.25228035e-01 4.02991116e-01
1.64035290e-01 6.74899578e-01 -3.14583957e-01 5.21884799e-01
1.04468310e+00 5.33551097e-01 -2.27773800e-01 4.33340311e-01
6.54704571e-02 1.66674718e-01 -5.42707086e-01 7.75466412e-02
-1.07280957e-02 -6.26546800e-01 3.68164957e-01 6.89996541e-01
5.49916506e-01 5.14461637e-01 1.86211750e-01 8.13760102e-01
-4.87842470e-01 1.67070091e-01 -6.25863850e-01 2.18228251e-01
6.17699683e-01 1.33859396e+00 -1.41232824e+00 -1.79001302e-01
-3.48470002e-01 8.22240591e-01 8.64466205e-02 -1.45887390e-01
-5.26047349e-01 -1.97898209e-01 6.92850590e-01 4.39982206e-01
1.16968147e-01 -2.73209095e-01 -2.60978937e-01 -6.24490678e-01
1.41006231e-01 -2.58041739e-01 1.68095604e-01 -6.40417159e-01
-1.30856395e+00 6.06628001e-01 2.52988040e-01 -7.66852319e-01
3.00338030e-01 -1.02485120e+00 -5.44905484e-01 9.75169897e-01
-1.44110644e+00 -1.29180694e+00 -2.34132290e-01 5.91408193e-01
6.81940734e-01 -2.09727958e-01 7.44441807e-01 3.13849390e-01
-5.85112572e-01 4.61863905e-01 -4.26650017e-01 1.63274124e-01
1.08691812e-01 -8.11459184e-01 4.87658054e-01 3.13026816e-01
-4.07119483e-01 2.14521036e-01 -1.31069124e-01 -1.06061482e+00
-1.71608126e+00 -1.05946314e+00 5.36731064e-01 -3.56724709e-01
-1.28292188e-01 -5.39518595e-01 -7.71523297e-01 1.16385214e-01
-5.68454385e-01 1.93084553e-01 3.11064184e-01 -2.57536441e-01
-1.57354221e-01 -2.87896991e-01 -2.09222126e+00 3.79328966e-01
1.79826427e+00 -4.25286829e-01 -4.25017148e-01 2.38965496e-01
8.22896302e-01 -5.44974685e-01 -1.14409935e+00 6.28193617e-01
5.32028913e-01 -9.54032719e-01 1.06637764e+00 -3.33743691e-01
6.67717457e-02 -1.96222723e-01 -2.17358693e-01 -8.07756245e-01
4.99841385e-02 -4.24586475e-01 1.32583410e-01 1.50485039e+00
1.98128335e-02 -6.85452521e-01 1.49789691e+00 3.20848405e-01
-3.37064713e-01 -9.41938639e-01 -1.41677368e+00 -6.94786251e-01
3.39965582e-01 -2.38788098e-01 1.12657428e+00 6.77835107e-01
-8.26025784e-01 1.99203342e-02 4.06473666e-01 2.86182433e-01
9.66792524e-01 3.37036073e-01 5.19288182e-01 -1.91922748e+00
7.00792193e-01 -6.66754484e-01 -5.04628122e-01 -1.27223861e+00
6.92172050e-01 -1.14202571e+00 -2.73071021e-01 -1.54970753e+00
-1.83904961e-01 -6.49159789e-01 5.59834003e-01 2.82826036e-01
3.97058576e-01 3.70229125e-01 -1.98051885e-01 4.89243157e-02
6.69540539e-02 4.51929122e-01 1.69876432e+00 8.34084153e-02
-2.36469358e-01 -1.23086065e-01 -5.11640728e-01 9.75912750e-01
5.59388041e-01 -1.14154793e-01 -3.20569873e-01 -3.96511763e-01
-3.50395381e-01 3.46826583e-01 1.60396218e-01 -7.54102051e-01
-6.17092662e-02 -8.97298977e-02 7.10536301e-01 -9.68072474e-01
7.33454764e-01 -9.22934175e-01 -1.08753003e-01 1.20978132e-01
3.99005324e-01 -1.66952629e-02 1.73203140e-01 2.72690654e-01
-2.59956885e-02 -7.15975538e-02 1.00752401e+00 -4.19972509e-01
-6.47313654e-01 9.06812668e-01 3.15420270e-01 2.27577565e-03
1.11180067e+00 -7.73256660e-01 -2.67280005e-02 1.93210989e-01
-7.47202814e-01 2.07709342e-01 6.37945652e-01 3.07208747e-01
9.67337191e-01 -1.36042237e+00 -4.71498728e-01 8.13878536e-01
-5.31101406e-01 7.51399219e-01 4.64868665e-01 2.45347232e-01
-7.20419526e-01 3.06548536e-01 -2.85526067e-01 -9.35141385e-01
-1.31848216e+00 6.75696433e-02 6.52193308e-01 3.64756167e-01
-7.16666579e-01 9.36768651e-01 2.28009984e-01 -1.04648626e+00
7.57498518e-02 9.42118019e-02 -2.07189485e-01 2.29729697e-01
1.53914869e-01 4.25279140e-01 9.23125967e-02 -1.03187907e+00
-5.08488238e-01 1.70270193e+00 2.98082173e-01 1.36664465e-01
1.28824961e+00 -1.07126839e-01 -5.36307514e-01 -2.13700905e-01
1.35527706e+00 -1.78795189e-01 -1.40496612e+00 4.95924018e-02
-3.04531902e-01 -4.43620324e-01 -1.01560175e-01 -3.78411233e-01
-1.62055528e+00 1.13357818e+00 6.43046021e-01 2.24771336e-01
1.23098183e+00 4.68092859e-01 9.62544620e-01 -1.24883540e-01
7.04033315e-01 -1.11587000e+00 7.86108598e-02 4.69242305e-01
7.31162727e-01 -9.34481323e-01 -1.31273612e-01 -1.13477695e+00
1.41103342e-02 1.40380549e+00 8.94426465e-01 -6.36753067e-02
1.20026207e+00 3.17299902e-01 1.51245408e-02 -6.76613331e-01
1.60840720e-01 3.01978644e-02 2.47351035e-01 7.18101919e-01
4.21326719e-02 -5.43896742e-02 -8.52559879e-02 6.67718709e-01
-2.32243136e-01 -3.15481812e-01 -1.46459386e-01 7.21781731e-01
-4.46098387e-01 -1.35185087e+00 -5.87632775e-01 4.67444539e-01
-2.40941092e-01 4.63811636e-01 -3.69701773e-01 9.42618728e-01
4.40815538e-01 6.32564962e-01 4.24569339e-01 -2.43974820e-01
5.54578424e-01 3.24386120e-01 8.24185371e-01 -8.38379741e-01
-2.49495313e-01 1.97098374e-01 -1.86988845e-01 -6.03478968e-01
-5.44408381e-01 -4.74680573e-01 -1.84816933e+00 -2.84434825e-01
-2.36971378e-02 4.24165800e-02 1.12916124e+00 6.94095194e-01
5.93720019e-01 1.67962424e-02 1.02599406e+00 -1.37845457e+00
-3.50472494e-03 -4.65279609e-01 -6.36078775e-01 2.64374197e-01
-2.13024318e-01 -1.04637957e+00 -3.33760887e-01 -1.65831834e-01] | [13.250783920288086, 0.14374059438705444] |
54076760-1627-438c-9ea7-9a78b2841f13 | mitigating-label-noise-through-data | 2305.13764 | null | https://arxiv.org/abs/2305.13764v1 | https://arxiv.org/pdf/2305.13764v1.pdf | Mitigating Label Noise through Data Ambiguation | Label noise poses an important challenge in machine learning, especially in deep learning, in which large models with high expressive power dominate the field. Models of that kind are prone to memorizing incorrect labels, thereby harming generalization performance. Many methods have been proposed to address this problem, including robust loss functions and more complex label correction approaches. Robust loss functions are appealing due to their simplicity, but typically lack flexibility, while label correction usually adds substantial complexity to the training setup. In this paper, we suggest to address the shortcomings of both methodologies by "ambiguating" the target information, adding additional, complementary candidate labels in case the learner is not sufficiently convinced of the observed training label. More precisely, we leverage the framework of so-called superset learning to construct set-valued targets based on a confidence threshold, which deliver imprecise yet more reliable beliefs about the ground-truth, effectively helping the learner to suppress the memorization effect. In an extensive empirical evaluation, our method demonstrates favorable learning behavior on synthetic and real-world noise, confirming the effectiveness in detecting and correcting erroneous training labels. | ['Eyke Hüllermeier', 'Julian Lienen'] | 2023-05-23 | null | null | null | null | ['memorization'] | ['natural-language-processing'] | [ 5.03661573e-01 2.76328057e-01 -1.63294941e-01 -5.55844665e-01
-9.38764036e-01 -5.37834167e-01 4.21392500e-01 6.86116457e-01
-6.77103937e-01 1.00444865e+00 -1.17325433e-01 -7.72006139e-02
-7.35844374e-02 -7.24392951e-01 -8.24974179e-01 -7.88150728e-01
3.05183351e-01 3.45277101e-01 1.40744433e-01 1.24979287e-01
2.29787976e-01 3.23789567e-02 -1.58317375e+00 7.47915059e-02
1.20756698e+00 1.14665210e+00 -6.85580373e-02 5.31552620e-02
-1.44746769e-02 1.11093235e+00 -7.31707633e-01 -8.66846681e-01
-7.06499349e-03 -2.83651769e-01 -6.29446149e-01 9.33604985e-02
6.49063289e-01 -2.38206107e-02 1.78216979e-01 1.49282873e+00
4.35059696e-01 1.91181436e-01 6.84397340e-01 -1.12332630e+00
-4.69557941e-01 8.29738200e-01 -5.33934951e-01 -1.48508074e-02
1.40676633e-01 8.90482566e-04 1.13646519e+00 -8.34916115e-01
1.55954912e-01 1.12539482e+00 1.04721928e+00 7.87724674e-01
-1.53391957e+00 -8.99387777e-01 4.17189926e-01 2.17520911e-02
-1.31527567e+00 -4.97493505e-01 7.43888140e-01 -4.28183883e-01
2.24058539e-01 2.62108326e-01 1.40494019e-01 1.25499821e+00
9.78763215e-03 6.90156162e-01 1.25347924e+00 -5.51780879e-01
5.08491039e-01 5.55867493e-01 2.51884460e-01 6.87887967e-01
4.96735930e-01 1.19629484e-02 -4.59200293e-01 -3.08645695e-01
2.68159449e-01 -2.85444688e-02 -3.36230248e-01 -3.22563946e-01
-8.49985480e-01 9.28797901e-01 4.80511785e-01 1.19774893e-01
-2.73391783e-01 1.41520232e-01 3.69054556e-01 3.35553885e-01
7.97023594e-01 6.59480453e-01 -2.75099635e-01 1.15917295e-01
-7.84861743e-01 2.46255785e-01 5.76100528e-01 5.73027372e-01
8.92628014e-01 -2.07345575e-01 -4.49760675e-01 9.88395274e-01
2.22764686e-01 2.14214355e-01 4.65542942e-01 -8.80168140e-01
4.80976999e-01 5.31994641e-01 2.43398547e-01 -1.01161468e+00
-4.22060937e-01 -9.54086006e-01 -9.99223053e-01 1.89334005e-01
5.90708137e-01 5.77854067e-02 -9.78330851e-01 2.24765635e+00
3.25338304e-01 3.00521702e-01 3.46244243e-03 9.24017847e-01
4.21835035e-01 2.56762892e-01 5.69947064e-01 -2.00930238e-01
1.13413584e+00 -7.60306358e-01 -6.31586313e-01 -4.10755724e-01
8.62184703e-01 -4.49152589e-01 1.29345894e+00 5.38585365e-01
-9.45127726e-01 -4.64598060e-01 -9.40738082e-01 1.86727926e-01
-2.13075891e-01 3.16851437e-02 5.34538090e-01 6.75815642e-01
-6.91101849e-01 7.30601966e-01 -4.10612494e-01 5.33784255e-02
5.68980455e-01 2.06505656e-01 -2.13871852e-01 -1.98966175e-01
-1.31018758e+00 9.55411196e-01 5.46484530e-01 9.75681543e-02
-8.14697385e-01 -6.89727485e-01 -8.35519254e-01 1.58037603e-01
6.59421623e-01 -5.09023905e-01 1.49253273e+00 -1.10983157e+00
-1.23711646e+00 9.55995440e-01 7.60676488e-02 -5.86804450e-01
9.64297950e-01 -2.78575748e-01 -1.74449503e-01 -2.99443066e-01
3.13709199e-01 5.84438741e-01 9.88760293e-01 -1.62637019e+00
-7.12412000e-01 -3.82988036e-01 2.02491164e-01 1.36412382e-01
-4.05867070e-01 -4.02279139e-01 -1.46228317e-02 -8.47911477e-01
3.29025179e-01 -8.90343249e-01 -3.82548183e-01 -1.07670523e-01
-5.29225826e-01 -3.85530502e-01 4.35946852e-01 -2.21248105e-01
1.28441203e+00 -2.07611203e+00 -1.53706059e-01 3.04082513e-01
4.96878058e-01 4.39380586e-01 -1.56570189e-02 -6.20988123e-02
3.66859101e-02 2.88015574e-01 -4.67857003e-01 -6.54568255e-01
3.17431316e-02 2.58985519e-01 -4.17823225e-01 3.56223375e-01
2.01334104e-01 6.40528262e-01 -1.08492398e+00 -2.89280981e-01
-1.03637971e-01 1.12148821e-01 -4.00075614e-01 2.30643809e-01
-4.14553493e-01 5.81634402e-01 -3.82909864e-01 3.14576596e-01
6.32623553e-01 -4.24956799e-01 1.14925234e-02 3.34563851e-02
2.97745436e-01 3.20840687e-01 -1.35194850e+00 1.24294448e+00
-6.03854358e-01 2.40024284e-01 -5.70533760e-02 -1.03110480e+00
9.04204488e-01 1.59345835e-01 7.00423568e-02 -6.67385578e-01
2.29781553e-01 2.36861005e-01 -2.13108182e-01 -3.87281120e-01
2.73729473e-01 -4.60143387e-01 -4.15976122e-02 4.55029666e-01
-1.30651399e-01 2.93367803e-02 3.04400492e-02 8.56369734e-02
8.44355285e-01 -2.25886256e-01 1.83912486e-01 -5.62343784e-02
5.22631109e-01 -3.24834555e-01 7.74476826e-01 1.20900953e+00
-2.50636667e-01 4.98993307e-01 3.98084134e-01 -2.34482512e-01
-7.43710816e-01 -8.85770440e-01 -2.01888800e-01 1.28006589e+00
2.22289234e-01 -2.82175690e-01 -7.65336514e-01 -1.06915545e+00
1.14164427e-01 8.77739489e-01 -7.19965875e-01 -5.16438246e-01
-3.30447733e-01 -8.07016730e-01 6.66829050e-01 5.11362791e-01
4.28692967e-01 -8.80238116e-01 -4.00981098e-01 3.06213558e-01
-3.58459204e-01 -8.36127102e-01 -2.68636107e-01 5.44683099e-01
-7.99770951e-01 -1.06609941e+00 -4.19080108e-01 -6.70176625e-01
9.20082927e-01 1.02430277e-01 1.20415187e+00 3.27218950e-01
1.05045572e-01 4.59984243e-02 -2.80799150e-01 -5.36276221e-01
-6.63267255e-01 4.42748666e-02 1.73750445e-01 1.29136279e-01
2.72411704e-01 -2.77288079e-01 -2.32232422e-01 3.11449468e-01
-1.01321507e+00 -9.82556939e-02 5.27107775e-01 1.20064306e+00
4.95399743e-01 2.42830351e-01 9.94411528e-01 -1.49256504e+00
7.80323625e-01 -5.20197213e-01 -5.84584475e-01 2.65958011e-01
-1.00041509e+00 2.12578103e-01 8.72278929e-01 -5.90996623e-01
-1.09105241e+00 -1.51767954e-01 -1.63950875e-01 -3.02765042e-01
-1.17599748e-01 5.41393101e-01 -1.45803675e-01 -1.38031840e-01
8.80607188e-01 -4.15486796e-03 -1.67737126e-01 -5.94248116e-01
1.90588325e-01 4.44100916e-01 4.94018584e-01 -7.83436418e-01
7.29726315e-01 2.84828186e-01 -1.10294074e-01 -1.89048529e-01
-1.70078731e+00 -3.15700084e-01 -3.42267007e-01 -1.11259304e-01
3.27595979e-01 -7.64761984e-01 -6.26037061e-01 6.04066491e-01
-1.06625438e+00 -2.59562939e-01 -3.60977739e-01 3.89256239e-01
-2.67071515e-01 2.37382054e-01 -5.62806487e-01 -8.77632976e-01
-1.63646802e-01 -1.07699299e+00 9.28019166e-01 2.94329613e-01
-2.88848758e-01 -1.07220805e+00 -6.91965669e-02 3.29051495e-01
3.55842978e-01 1.16574772e-01 1.10684979e+00 -9.59968090e-01
-1.40383855e-01 -2.61907339e-01 -2.46271849e-01 7.01829255e-01
-2.10110638e-02 -4.18044597e-01 -1.33313596e+00 -3.81031394e-01
7.61055499e-02 -7.71155834e-01 1.17873919e+00 1.33803353e-01
1.38642359e+00 -3.24239641e-01 -2.82811850e-01 2.52571583e-01
1.18519425e+00 -1.04620330e-01 2.84612209e-01 3.53617221e-01
6.09286308e-01 7.53584445e-01 6.37837589e-01 3.95048052e-01
2.94258267e-01 5.25678515e-01 4.66655463e-01 4.70913667e-03
2.27515046e-02 -4.49037552e-01 -4.88734208e-02 5.63903868e-01
2.84006983e-01 -3.92709166e-01 -7.61887133e-01 2.44111180e-01
-1.84344161e+00 -7.43606150e-01 -2.81180106e-02 2.43570733e+00
1.30384970e+00 4.69575286e-01 -2.06845477e-01 4.51152831e-01
8.16950560e-01 -7.55021768e-03 -6.52998626e-01 1.15763895e-01
-7.80425519e-02 9.92060825e-03 3.89096349e-01 6.05085313e-01
-1.28550971e+00 8.62535417e-01 5.75196981e+00 9.49986517e-01
-1.11457992e+00 2.86603987e-01 9.63890254e-01 2.12180465e-01
-3.67608964e-01 -1.24868087e-01 -7.93644547e-01 6.45416796e-01
5.92671990e-01 1.05182126e-01 9.12625715e-02 8.85765612e-01
1.05979882e-01 -2.61996955e-01 -1.19960439e+00 9.00669813e-01
-5.50848246e-03 -1.06607032e+00 -1.16157830e-01 -2.48895869e-01
8.05996001e-01 -3.28320146e-01 1.29226387e-01 6.48794234e-01
3.03238064e-01 -1.05648780e+00 9.91116405e-01 4.19552326e-01
6.65316701e-01 -7.89022148e-01 9.39212084e-01 6.17411435e-01
-6.68001473e-01 -2.48241708e-01 -4.96091127e-01 -1.84035376e-01
-1.29270449e-01 1.06913531e+00 -7.78854012e-01 1.22494802e-01
4.92914617e-01 3.89755398e-01 -8.21084440e-01 1.04243588e+00
-5.92091799e-01 8.89663875e-01 -2.70284176e-01 3.35966982e-02
1.57607093e-01 4.78049815e-02 2.81702846e-01 1.07793820e+00
7.91072994e-02 -1.40350938e-01 4.21356022e-01 8.69104147e-01
-5.87093294e-01 1.96749922e-02 -4.62452561e-01 3.57767463e-01
7.24180162e-01 1.08793843e+00 -5.91039598e-01 -3.48856121e-01
-1.23653896e-01 7.70809472e-01 7.74198890e-01 2.59851694e-01
-8.34305227e-01 -1.41116962e-01 2.67846465e-01 8.88134688e-02
-1.60751835e-01 2.22113967e-01 -6.19015694e-01 -1.05656993e+00
2.60361642e-01 -8.53432000e-01 4.38453734e-01 -3.53743792e-01
-1.53228104e+00 4.46471334e-01 -3.26591909e-01 -1.05744791e+00
-5.30001614e-03 -3.87769938e-01 -4.19267684e-01 6.58742785e-01
-1.63199222e+00 -8.94043982e-01 -3.84889990e-01 3.43012780e-01
3.32526058e-01 1.51516154e-01 7.07127094e-01 5.17251849e-01
-6.48990154e-01 1.00672948e+00 1.15123764e-01 -8.13371390e-02
9.30429339e-01 -1.36898625e+00 -3.98958288e-02 6.32985950e-01
9.68318060e-02 4.92561758e-01 8.84655297e-01 -5.75593770e-01
-6.79386616e-01 -1.32750225e+00 8.81149709e-01 -4.85602230e-01
5.00538647e-01 -4.92034703e-01 -1.35852742e+00 5.37829518e-01
-4.47080463e-01 2.51491554e-02 6.52266204e-01 2.93303162e-01
-7.00633943e-01 -1.50654987e-01 -1.37976134e+00 5.11889637e-01
9.13026571e-01 -3.90999198e-01 -4.91066664e-01 4.17783707e-01
6.14956260e-01 -3.31835866e-01 -5.24028063e-01 5.28125167e-01
2.36049056e-01 -1.00422668e+00 6.33434653e-01 -6.00427091e-01
1.89389035e-01 -2.28005484e-01 5.61646977e-03 -1.54859531e+00
-3.99168730e-01 -2.84121662e-01 -2.22596042e-02 1.44844306e+00
5.14455259e-01 -6.82963431e-01 9.35692012e-01 6.49336100e-01
-9.89920795e-02 -7.80195296e-01 -9.28586245e-01 -7.64593482e-01
-2.13183612e-02 -5.59436440e-01 3.70185852e-01 1.17223263e+00
-2.29858860e-01 3.79996002e-01 -4.83423412e-01 1.88394442e-01
8.05923998e-01 -2.60938853e-01 4.73832011e-01 -1.60318387e+00
-2.58792400e-01 -4.48499650e-01 -2.67401755e-01 -8.50335419e-01
5.38479567e-01 -8.76664460e-01 4.47704941e-01 -1.04670811e+00
2.60025889e-01 -1.05525827e+00 -5.71812093e-01 7.02124000e-01
-5.61495602e-01 5.00724971e-01 -6.07132800e-02 2.91476637e-01
-8.04127336e-01 6.22768998e-01 1.03159404e+00 -2.30666131e-01
7.58577362e-02 3.41475099e-01 -8.67902100e-01 9.02846873e-01
7.36931145e-01 -8.63024831e-01 -5.01104414e-01 -4.32308674e-01
3.61129403e-01 -3.95364612e-01 4.93478626e-01 -1.06419575e+00
2.22869799e-01 -1.55147361e-02 2.64798254e-01 6.04754910e-02
1.62853330e-01 -8.75104129e-01 -1.84186116e-01 4.14159000e-01
-8.85928094e-01 -2.06080571e-01 -5.86989298e-02 8.46198320e-01
-2.58282602e-01 -5.87632596e-01 1.23688078e+00 -1.55001124e-02
-4.44260687e-01 1.01664543e-01 -7.79019669e-02 3.62788320e-01
9.16634798e-01 3.52399834e-02 -3.09309989e-01 -4.17156696e-01
-6.76685154e-01 2.13201448e-01 4.35163409e-01 2.37734243e-01
3.59887272e-01 -1.24352026e+00 -6.99739337e-01 1.87915981e-01
3.96623790e-01 1.07037090e-01 9.01917741e-02 6.69861734e-01
1.31954104e-02 1.25177875e-01 2.56038874e-01 -5.74738920e-01
-8.69107902e-01 5.59568107e-01 4.08719867e-01 -4.00979072e-01
-3.31541777e-01 1.15319371e+00 2.59569138e-01 -5.29065847e-01
7.79390812e-01 -2.88775057e-01 -1.18253134e-01 2.06539541e-01
5.59859812e-01 3.37176293e-01 3.91634494e-01 -3.39004904e-01
-1.25539705e-01 3.24186683e-01 -4.10108775e-01 1.96336091e-01
9.24198687e-01 -1.61936954e-01 5.56645170e-02 3.98964822e-01
8.81258190e-01 6.70983121e-02 -1.32523310e+00 -6.76063955e-01
4.70163167e-01 -4.26867336e-01 2.71537807e-02 -1.07598865e+00
-8.05106103e-01 8.90944839e-01 5.25564432e-01 3.89259577e-01
9.91814613e-01 -8.04439709e-02 5.81063271e-01 4.43529963e-01
3.76460820e-01 -1.09995949e+00 2.71655738e-01 5.27624726e-01
4.83660966e-01 -1.67060709e+00 -2.51737118e-01 -4.25217777e-01
-5.65628648e-01 8.29097569e-01 8.99430275e-01 1.05570547e-01
3.09839278e-01 1.31702125e-01 2.78558910e-01 -5.85030438e-03
-6.24481082e-01 7.93391913e-02 1.35929510e-01 4.49689776e-01
3.65123957e-01 2.05717701e-02 -3.51655304e-01 7.83724844e-01
2.88798790e-02 -2.33163506e-01 3.76057237e-01 7.66450644e-01
-6.10428154e-01 -9.18181002e-01 -3.93852502e-01 5.87012708e-01
-6.12155020e-01 -1.69538692e-01 -1.75176620e-01 4.18638825e-01
4.00824308e-01 1.05785680e+00 -1.63119853e-01 -2.98654109e-01
3.87033761e-01 1.63825735e-01 2.92555243e-01 -7.72580087e-01
-5.28148353e-01 -1.92483872e-01 2.75321230e-02 -2.61680931e-01
-2.32611343e-01 -2.46950760e-01 -1.00173533e+00 -1.25364095e-01
-7.25035250e-01 4.11476731e-01 4.54131663e-01 1.11228800e+00
1.26236960e-01 4.07660514e-01 6.19299829e-01 -6.26462221e-01
-1.22161329e+00 -1.01186001e+00 -5.78793764e-01 7.74849951e-01
4.31161106e-01 -1.02755344e+00 -5.26630282e-01 -1.87916711e-01] | [9.282781600952148, 3.885890483856201] |
b1af881a-6944-44dc-8514-4ac0de861ee0 | totto-a-controlled-table-to-text-generation | 2004.14373 | null | https://arxiv.org/abs/2004.14373v3 | https://arxiv.org/pdf/2004.14373v3.pdf | ToTTo: A Controlled Table-To-Text Generation Dataset | We present ToTTo, an open-domain English table-to-text dataset with over 120,000 training examples that proposes a controlled generation task: given a Wikipedia table and a set of highlighted table cells, produce a one-sentence description. To obtain generated targets that are natural but also faithful to the source table, we introduce a dataset construction process where annotators directly revise existing candidate sentences from Wikipedia. We present systematic analyses of our dataset and annotation process as well as results achieved by several state-of-the-art baselines. While usually fluent, existing methods often hallucinate phrases that are not supported by the table, suggesting that this dataset can serve as a useful research benchmark for high-precision conditional text generation. | ['Manaal Faruqui', 'Xuezhi Wang', 'Dipanjan Das', 'Ankur P. Parikh', 'Sebastian Gehrmann', 'Diyi Yang', 'Bhuwan Dhingra'] | 2020-04-29 | null | https://aclanthology.org/2020.emnlp-main.89 | https://aclanthology.org/2020.emnlp-main.89.pdf | emnlp-2020-11 | ['conditional-text-generation', 'table-to-text-generation'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.65029365e-01 1.01218295e+00 -3.17152053e-01 -1.47380367e-01
-1.46947718e+00 -8.51069212e-01 9.62325871e-01 3.16887408e-01
-1.37608841e-01 1.61817682e+00 8.13275874e-01 -1.35105610e-01
5.84238529e-01 -1.10742271e+00 -9.89341617e-01 6.86794594e-02
3.18153530e-01 9.54462588e-01 1.74811542e-01 -6.77278459e-01
3.85038108e-01 -3.39358449e-01 -1.19174457e+00 9.26460147e-01
1.18255508e+00 3.19914848e-01 5.68298772e-02 4.60786343e-01
-3.32089245e-01 8.86937678e-01 -1.17779768e+00 -1.30536413e+00
3.69816460e-02 -5.29504359e-01 -1.11426413e+00 -1.04668774e-01
7.59056568e-01 -8.33115205e-02 -1.91296905e-01 1.00306988e+00
6.23971939e-01 -3.53868003e-03 9.05135691e-01 -1.00298524e+00
-1.18063545e+00 1.72529829e+00 -6.38311217e-03 8.48263726e-02
8.89546514e-01 1.42134294e-01 1.27092791e+00 -1.25257897e+00
1.29823458e+00 1.18113887e+00 5.94107389e-01 8.85463357e-01
-1.62995112e+00 -6.65644646e-01 -2.87209630e-01 -2.45392606e-01
-1.33868217e+00 -7.80055106e-01 2.60656685e-01 -2.45341688e-01
1.22934914e+00 2.03120545e-01 3.94591510e-01 1.83159542e+00
2.58550197e-01 6.75044715e-01 6.79165840e-01 -7.03941703e-01
-3.20249461e-02 6.48792148e-01 -2.80548304e-01 4.23579037e-01
7.36505985e-01 -3.51650089e-01 -7.81839371e-01 -1.19960785e-01
2.73527533e-01 -1.07017279e+00 -3.73666286e-01 4.68636639e-02
-1.81644261e+00 8.66012156e-01 2.73204535e-01 -2.89137624e-02
-1.89257026e-01 -1.12500355e-01 5.41382670e-01 2.53355503e-02
4.68717754e-01 1.10437882e+00 -4.63412672e-01 -2.82448083e-01
-8.14288557e-01 8.27551365e-01 1.12309611e+00 1.82983804e+00
5.89698970e-01 -3.47536653e-02 -9.35754597e-01 6.45144522e-01
-1.43421069e-01 7.06119478e-01 3.97202998e-01 -7.51431346e-01
1.25897872e+00 5.11963844e-01 4.49898064e-01 -8.02071333e-01
1.70758784e-01 -1.13965787e-01 -6.55482531e-01 -3.78757119e-01
2.89752662e-01 -2.62038738e-01 -5.73888421e-01 1.50641727e+00
2.95364261e-02 -7.09637046e-01 5.82125008e-01 1.51612833e-01
1.22952068e+00 7.75747478e-01 5.85098378e-02 -2.59865940e-01
1.14924192e+00 -1.07832277e+00 -9.75408554e-01 -6.02884233e-01
9.25962627e-01 -7.48187363e-01 1.55411816e+00 1.56498045e-01
-1.09588194e+00 -4.92663562e-01 -9.30430174e-01 -3.37996423e-01
-8.52671385e-01 3.21012437e-01 3.38599026e-01 4.64482635e-01
-9.61537480e-01 4.03595418e-01 -1.44320711e-01 -3.92578542e-01
2.28183568e-01 -2.86128342e-01 -4.94309187e-01 -8.00594166e-02
-1.58368075e+00 1.16749489e+00 9.49021578e-01 -3.77922833e-01
-7.50782549e-01 -7.32643127e-01 -1.13345408e+00 -2.75094032e-01
4.66039777e-01 -7.91136801e-01 1.52652013e+00 -2.98354954e-01
-9.26470339e-01 1.00776756e+00 -2.58748591e-01 -6.76519871e-01
4.82506573e-01 -4.10813183e-01 -5.56252658e-01 8.05164222e-03
9.17521656e-01 1.09706950e+00 4.64165509e-01 -1.33697391e+00
-4.17810053e-01 1.04117990e-01 5.45133837e-02 2.72291660e-01
-2.19804123e-01 -3.29415239e-02 -2.56593615e-01 -1.10405481e+00
-4.28063720e-01 -6.69760227e-01 -1.03444070e-01 -5.02851486e-01
-1.49971187e+00 -1.92337587e-01 1.37105569e-01 -8.43673646e-01
1.79849935e+00 -1.44616044e+00 -6.86831772e-02 -3.10553700e-01
-3.91875431e-02 -1.37228444e-01 -1.38587236e-01 9.22292888e-01
3.09294388e-02 8.66851807e-01 -4.03514564e-01 -2.43222609e-01
1.67035341e-01 -3.41766626e-01 -8.87242019e-01 -4.19514298e-01
4.65177089e-01 1.05986667e+00 -1.10596967e+00 -8.69231343e-01
-3.33671212e-01 -2.19079927e-02 -4.71543670e-01 1.19168982e-01
-7.47161329e-01 7.92953968e-02 -2.47022152e-01 5.73050857e-01
3.05859268e-01 -8.55222717e-02 -1.09650165e-01 -2.92647749e-01
-4.10763919e-02 1.03038168e+00 -9.26396906e-01 1.61119032e+00
-3.47519934e-01 6.38955355e-01 -6.29865289e-01 1.80456694e-02
9.74396169e-01 3.67220968e-01 -3.67877573e-01 -3.66382301e-01
-2.97835231e-01 3.31584990e-01 -2.19654366e-01 -5.39512098e-01
1.18370318e+00 -8.95334259e-02 -6.83692038e-01 5.40888190e-01
1.36640474e-01 -7.84065068e-01 1.10543799e+00 9.14013863e-01
1.10116398e+00 2.77147591e-01 6.17785513e-01 -2.58669257e-01
3.30400199e-01 6.47307813e-01 1.93475217e-01 1.13250756e+00
3.14413339e-01 7.50954807e-01 5.37490547e-01 -2.28991345e-01
-1.33915412e+00 -1.10863769e+00 -2.04555601e-01 7.14363933e-01
-2.03917027e-01 -1.07663941e+00 -8.45526040e-01 -9.53566372e-01
-7.54701868e-02 1.60364950e+00 -9.99733865e-01 -1.24690779e-01
-3.35295141e-01 -4.59690183e-01 1.08458209e+00 5.63225448e-01
4.04028624e-01 -1.49633384e+00 -2.25327298e-01 3.55154663e-01
-8.72867644e-01 -1.20429122e+00 -5.83995044e-01 1.50211319e-01
-5.34947038e-01 -8.75970066e-01 -3.50415438e-01 -8.18839312e-01
6.94271743e-01 -2.98825771e-01 2.01042962e+00 -2.66062886e-01
1.02855608e-01 -6.26214966e-02 -4.88085985e-01 -5.87003827e-01
-1.10402155e+00 4.44279671e-01 -1.63151473e-01 -6.79968476e-01
4.92559522e-01 2.69598644e-02 2.02450767e-01 -1.82670832e-01
-8.39934289e-01 5.38371682e-01 6.02611363e-01 8.22623909e-01
5.61101496e-01 -2.55670995e-01 8.23854089e-01 -1.41713393e+00
1.20386267e+00 -4.78057772e-01 -2.08975345e-01 6.36507332e-01
-4.21299011e-01 2.42558956e-01 8.06537151e-01 -1.53375402e-01
-1.29573631e+00 3.81174870e-02 1.34952709e-01 3.06376040e-01
-2.10175198e-02 7.07744658e-01 -3.04398745e-01 7.43691683e-01
1.36829758e+00 4.51646537e-01 -3.57236266e-01 -7.82580152e-02
7.01776445e-01 7.24972844e-01 5.85012436e-01 -6.75979733e-01
9.91949439e-01 -5.00941575e-02 -5.55986404e-01 -4.70492661e-01
-1.30136347e+00 3.33283275e-01 -8.11602473e-01 7.58458599e-02
7.58025408e-01 -1.22148108e+00 2.48040676e-01 2.79568508e-02
-1.62145054e+00 -4.87742245e-01 -4.37569022e-01 -9.59070846e-02
-4.57781076e-01 -6.79896772e-02 -5.87791145e-01 -5.52610040e-01
-6.64407253e-01 -7.16737032e-01 1.25144696e+00 -1.92980897e-02
-8.83398056e-01 -9.57620144e-01 2.32818350e-01 2.81599849e-01
2.14876249e-01 2.01599598e-01 1.03471661e+00 -9.25924659e-01
-4.08392638e-01 -1.79230720e-01 -2.29783561e-02 4.94291596e-02
1.09801240e-01 4.11471188e-01 -7.14199543e-01 1.21129237e-01
-5.94619513e-01 -1.10237956e+00 7.37006247e-01 -1.60719752e-01
7.90337920e-01 -9.30931211e-01 -5.75771511e-01 3.86555940e-02
1.08852625e+00 -5.98951355e-02 7.94776678e-01 4.43137407e-01
4.74017650e-01 7.14941740e-01 7.67364383e-01 3.77847075e-01
6.32772624e-01 5.40686965e-01 -8.53644907e-02 4.06489253e-01
-3.14988852e-01 -1.07976711e+00 5.45665205e-01 7.12225974e-01
6.40350640e-01 -7.83427835e-01 -8.66038799e-01 8.54288042e-01
-1.32867265e+00 -1.19236207e+00 -3.68986398e-01 2.12746978e+00
1.78830707e+00 5.45390427e-01 -2.10105315e-01 -1.30185604e-01
8.44440579e-01 6.96919411e-02 -1.79325283e-01 -2.40518361e-01
-5.60491741e-01 1.25555560e-01 3.94578725e-02 5.12147427e-01
-1.08438492e+00 1.30406046e+00 7.41347837e+00 9.13281679e-01
-4.72254187e-01 -1.64153650e-01 5.91918111e-01 4.84543554e-02
-7.47777104e-01 2.48854030e-02 -1.32934690e+00 4.42131311e-01
1.02238679e+00 -8.08652878e-01 2.48409882e-01 9.26727891e-01
-2.17230972e-02 -8.75509903e-02 -1.40396392e+00 5.61896801e-01
4.79111940e-01 -1.64073145e+00 7.91642129e-01 -3.16927433e-01
9.29105222e-01 -3.04604352e-01 -2.32416078e-01 8.21526587e-01
5.37759781e-01 -1.17276585e+00 1.09958041e+00 2.56408423e-01
1.15447748e+00 -4.31658924e-01 5.92406809e-01 3.20256412e-01
-8.64325941e-01 3.84679377e-01 -4.88175720e-01 -2.34413818e-02
3.85142833e-01 7.28708625e-01 -1.22223413e+00 2.62916356e-01
4.77214515e-01 6.87831819e-01 -1.20702779e+00 6.57401443e-01
-7.64790773e-01 4.01316583e-01 -6.22952916e-02 -5.36415279e-01
-1.14466503e-01 2.04840615e-01 5.86176455e-01 1.55520523e+00
3.12656164e-01 -1.88806742e-01 1.26150586e-02 1.43432188e+00
-7.12224841e-01 3.28360766e-01 -1.27355087e+00 -3.77364337e-01
1.06057310e+00 1.15284348e+00 -5.67976892e-01 -7.48824716e-01
-2.05103114e-01 9.91623044e-01 5.99087536e-01 8.71406719e-02
-8.33099365e-01 -9.66122210e-01 8.33863541e-02 3.41062732e-02
1.42070368e-01 1.92003146e-01 -5.93893528e-01 -1.39219856e+00
3.21367949e-01 -1.13967121e+00 1.91165671e-01 -1.17930710e+00
-1.38638568e+00 1.01250207e+00 1.36362255e-01 -1.29160547e+00
-5.61878860e-01 -1.77558437e-01 -5.55005133e-01 8.23098063e-01
-9.32758987e-01 -1.07607162e+00 -1.26739979e-01 3.12669009e-01
9.50346887e-01 -2.10683376e-01 1.14222682e+00 -6.39849901e-02
-4.93968040e-01 5.87681890e-01 -1.89509347e-01 5.05410790e-01
1.05672681e+00 -1.60766625e+00 1.01915169e+00 1.16480708e+00
3.70954067e-01 1.04493940e+00 9.46044624e-01 -1.38551557e+00
-8.45865726e-01 -1.45992482e+00 1.76544857e+00 -1.25662911e+00
7.09473550e-01 -7.96002209e-01 -6.64498627e-01 1.03252923e+00
7.88525999e-01 -2.98815966e-01 6.46332264e-01 -2.16710735e-02
-3.51976842e-01 1.69095770e-01 -9.97454226e-01 1.00881815e+00
1.03277695e+00 -5.36566436e-01 -1.09904587e+00 7.61899531e-01
9.89921689e-01 -8.17834854e-01 -5.16170561e-01 8.04304481e-02
9.74534601e-02 -6.71403885e-01 4.66312915e-01 -7.70654380e-01
1.38013196e+00 -2.41578981e-01 -3.16678882e-02 -1.80861771e+00
-1.41477302e-01 -7.34008193e-01 -1.81880042e-01 1.68310225e+00
1.42849004e+00 -6.64690435e-02 4.47382361e-01 7.83949733e-01
-3.28162044e-01 -4.11449164e-01 -6.22446895e-01 -6.68778300e-01
3.74973297e-01 -2.48296604e-01 3.70914549e-01 7.78662205e-01
5.61009645e-01 1.16969621e+00 -2.72184819e-01 -4.02286857e-01
4.80412483e-01 -2.67745882e-01 8.92436922e-01 -8.70786130e-01
3.78885306e-02 -1.05734132e-01 4.36579585e-01 -9.11823094e-01
2.67951012e-01 -1.01037145e+00 3.92421663e-01 -1.88815224e+00
3.84089082e-01 -2.60983080e-01 6.95947289e-01 5.83421648e-01
-4.61128235e-01 3.75056356e-01 9.17313062e-03 -6.10397663e-04
-6.02850080e-01 6.84103608e-01 1.16833830e+00 -2.54622400e-01
-1.11792453e-01 -3.80743921e-01 -1.29342353e+00 2.36986771e-01
5.11698186e-01 -6.87334001e-01 -4.32424515e-01 -6.14592493e-01
6.08200848e-01 -2.58873310e-02 -1.54372156e-02 -1.09465635e+00
9.32573751e-02 -1.19343735e-01 4.32762593e-01 -6.04009092e-01
9.77792591e-02 -1.44977823e-01 1.41787343e-02 1.92669943e-01
-1.08685875e+00 3.13987881e-01 3.06169420e-01 2.76044399e-01
-2.64085174e-01 -4.77090150e-01 4.59127516e-01 -4.35104787e-01
-3.22629482e-01 -2.09634125e-01 -6.43843412e-01 9.42333162e-01
8.66523206e-01 2.05532983e-02 -9.21997964e-01 -5.05886555e-01
-2.56722301e-01 6.21973164e-02 4.61145848e-01 6.64211333e-01
5.89036047e-01 -1.39498889e+00 -1.15703940e+00 -1.70569152e-01
6.86160564e-01 2.87990887e-02 -3.80067945e-01 1.01360492e-01
-4.89273578e-01 7.65456021e-01 -5.54049797e-02 4.36320454e-02
-5.84556580e-01 5.26863992e-01 -8.02561343e-02 -4.31221157e-01
-4.57400769e-01 8.59635711e-01 -2.42300510e-01 -7.55650640e-01
3.86165157e-02 -3.69623721e-01 -1.58245370e-01 1.08829424e-01
7.69957781e-01 9.63746831e-02 1.76924944e-01 -3.94835979e-01
-1.97731316e-01 -3.15181643e-01 -1.85667545e-01 -4.66078788e-01
8.60969841e-01 -1.34577304e-01 -1.28701881e-01 6.11185074e-01
6.84698462e-01 7.10028052e-01 -6.38426840e-01 6.22859113e-02
1.57841310e-01 -2.53208756e-01 -4.82867241e-01 -1.39711511e+00
-3.64714503e-01 4.96434391e-01 -4.90286201e-01 2.60380894e-01
4.80340064e-01 1.03599705e-01 6.76279068e-01 8.36991489e-01
3.93370271e-01 -1.06103551e+00 2.90603936e-01 9.65988755e-01
1.45215464e+00 -1.35134888e+00 -6.43734261e-02 -7.50155330e-01
-1.19642460e+00 1.09749496e+00 1.22984612e+00 3.57201666e-01
1.25740781e-01 2.87237674e-01 -6.26282953e-03 -2.04609707e-02
-1.27176476e+00 -2.57914234e-02 9.77069438e-02 9.45910871e-01
9.34096873e-01 -1.56127468e-01 -2.90771842e-01 1.01001632e+00
-9.06605721e-01 -1.80194214e-01 1.21123755e+00 6.21366918e-01
-2.06642628e-01 -9.68285143e-01 -7.48785511e-02 7.35554516e-01
-3.53271395e-01 -8.91808569e-01 -9.69353616e-01 9.28612649e-01
-1.62622064e-01 9.88987327e-01 -1.10719934e-01 -2.54997194e-01
3.93392086e-01 3.04549962e-01 2.57022709e-01 -1.33900082e+00
-7.54405320e-01 -5.74898303e-01 9.49004292e-01 -1.57740474e-01
-6.59159422e-02 -5.28335869e-01 -1.11035120e+00 -8.68478045e-02
-2.73309588e-01 5.10242403e-01 1.16524016e-02 6.49242461e-01
4.77139682e-01 1.38341650e-01 9.96533111e-02 -4.51368421e-01
-6.33604944e-01 -1.62736309e+00 -2.54108459e-01 5.39443493e-01
-1.37828574e-01 -3.27436358e-01 -2.30677083e-01 5.70746720e-01] | [11.546813011169434, 8.863469123840332] |
97b8fb35-05c7-44a1-b803-107af3294295 | reevaluating-data-partitioning-for-emotion | 2303.13364 | null | https://arxiv.org/abs/2303.13364v1 | https://arxiv.org/pdf/2303.13364v1.pdf | Reevaluating Data Partitioning for Emotion Detection in EmoWOZ | This paper focuses on the EmoWoz dataset, an extension of MultiWOZ that provides emotion labels for the dialogues. MultiWOZ was partitioned initially for another purpose, resulting in a distributional shift when considering the new purpose of emotion recognition. The emotion tags in EmoWoz are highly imbalanced and unevenly distributed across the partitions, which causes sub-optimal performance and poor comparison of models. We propose a stratified sampling scheme based on emotion tags to address this issue, improve the dataset's distribution, and reduce dataset shift. We also introduce a special technique to handle conversation (sequential) data with many emotional tags. Using our proposed sampling method, models built upon EmoWoz can perform better, making it a more reliable resource for training conversational agents with emotional intelligence. We recommend that future researchers use this new partitioning to ensure consistent and accurate performance evaluations. | ['Michael D. Porter', 'Moeen Mostafavi'] | 2023-03-15 | null | null | null | null | ['emotional-intelligence'] | ['natural-language-processing'] | [-2.52457321e-01 3.27452384e-02 1.29191261e-02 -6.33402169e-01
-2.36789644e-01 -4.21864241e-01 2.19410181e-01 1.86057091e-01
-3.12378705e-01 6.85435534e-01 3.72696280e-01 -7.64223235e-03
1.11202467e-02 -8.60325754e-01 2.58351147e-01 -6.57400310e-01
8.36950168e-02 5.77440381e-01 4.81333360e-02 -3.28449041e-01
3.06101203e-01 2.58377511e-02 -1.95193672e+00 5.45492887e-01
9.57664430e-01 8.83225083e-01 1.69791669e-01 4.39634353e-01
-5.82593203e-01 7.67200410e-01 -9.40771282e-01 -7.05075264e-01
1.14864789e-01 -4.59287643e-01 -9.90351558e-01 8.46691057e-02
-2.68038273e-01 -9.85179991e-02 3.16402644e-01 8.03082466e-01
7.86677957e-01 6.72493100e-01 7.91268826e-01 -1.59775019e+00
-3.98135595e-02 5.11334956e-01 -4.94740158e-01 -2.02029720e-01
4.21939164e-01 -2.11523741e-01 8.53290558e-01 -3.53540421e-01
4.95146394e-01 1.22943485e+00 8.80330265e-01 8.58478487e-01
-7.78477669e-01 -7.57980525e-01 1.83842227e-01 3.15755397e-01
-1.05031395e+00 -4.45498735e-01 9.35651004e-01 -3.73695940e-01
8.00353587e-01 5.39889514e-01 7.22096860e-01 1.08765519e+00
-2.04960629e-01 7.47291148e-01 1.51522744e+00 -4.46588367e-01
4.26547855e-01 6.23594463e-01 3.72171491e-01 4.57423925e-01
-1.42792270e-01 -5.45211256e-01 -3.65125418e-01 -5.73636889e-01
5.93093736e-03 -7.41634741e-02 -1.33393019e-01 -3.45005840e-01
-7.02114105e-01 1.04897165e+00 -2.58565009e-01 5.12105525e-01
-5.03023744e-01 -5.96405923e-01 8.54040861e-01 3.71837080e-01
8.98614943e-01 6.77614868e-01 -8.27896357e-01 -9.66707826e-01
-4.62178260e-01 1.84321746e-01 1.14262092e+00 5.19966662e-01
7.24387884e-01 -2.52714485e-01 -1.83635101e-01 1.65925431e+00
1.92618847e-01 9.11074057e-02 7.84633279e-01 -8.62685382e-01
3.20418477e-01 7.86549449e-01 1.56886801e-01 -1.20854390e+00
-5.94580770e-01 -4.51425277e-03 -6.31202817e-01 1.97818223e-02
4.85685945e-01 -5.15316308e-01 -3.36144596e-01 1.80897379e+00
4.73849922e-01 -3.20627064e-01 3.01967561e-01 6.73481703e-01
1.03772891e+00 6.44583046e-01 3.56175721e-01 -2.93514937e-01
1.58519113e+00 -8.06600153e-01 -1.16153562e+00 9.32367519e-02
9.78270888e-01 -7.26568162e-01 1.33376276e+00 3.18975389e-01
-7.03210115e-01 -2.22236246e-01 -8.27428758e-01 2.89351135e-01
-7.31193125e-01 -2.54452508e-02 1.03444755e+00 1.11791039e+00
-8.15860927e-01 2.08474472e-01 -3.65462214e-01 -4.80731994e-01
-7.07830710e-04 2.42028832e-01 -2.66392827e-01 2.55173951e-01
-1.50721526e+00 1.16532910e+00 3.04598272e-01 -1.48256645e-01
-9.63771809e-03 -2.68426985e-01 -8.68288219e-01 1.02134511e-01
1.31729618e-01 -2.95834929e-01 1.34520912e+00 -1.08722985e+00
-2.01432586e+00 9.12454069e-01 -2.90906429e-01 2.93739121e-02
3.96747619e-01 3.94624144e-01 -4.28004205e-01 -1.89934194e-01
-2.30068311e-01 4.04311091e-01 3.17885667e-01 -1.32952130e+00
-5.76850832e-01 -5.13065994e-01 -9.19268373e-03 4.23624009e-01
-6.73856437e-01 1.93852894e-02 -4.07476455e-01 -3.99528295e-01
-4.43034731e-02 -9.66665149e-01 -1.03813410e-01 -1.09274697e+00
-8.70359167e-02 -7.78694451e-01 7.76803911e-01 -5.35693824e-01
1.38179779e+00 -1.97996199e+00 -2.20615149e-01 1.87831864e-01
2.64675289e-01 2.55165368e-01 -8.64208862e-02 5.63753128e-01
-7.20372126e-02 2.60923892e-01 2.16492310e-01 -7.28028595e-01
1.64567515e-01 2.13420928e-01 2.00313374e-01 5.46048395e-02
-1.38856396e-01 4.24931198e-01 -8.04881513e-01 -4.73795563e-01
3.06125171e-02 1.16296045e-01 -5.58313251e-01 5.85587740e-01
1.38229072e-01 9.15022790e-02 -4.82066661e-01 4.73923594e-01
7.74275601e-01 2.13200733e-01 2.22560376e-01 -2.08680078e-01
-1.04503576e-02 2.72062629e-01 -8.58009994e-01 1.18275726e+00
-8.23791683e-01 5.22121906e-01 1.19433645e-02 -7.93976009e-01
1.37731123e+00 4.71055478e-01 7.97746301e-01 -4.27890360e-01
7.00991750e-01 -1.58307001e-01 8.91609639e-02 -8.09318423e-01
8.44152570e-01 -2.92333275e-01 -2.88849652e-01 6.96595430e-01
-2.60543525e-01 -1.80368409e-01 1.28384352e-01 -2.57857069e-02
8.20576668e-01 -2.99988747e-01 1.68750823e-01 -1.87220022e-01
2.33780578e-01 -7.72704184e-02 8.42866123e-01 4.50285614e-01
-6.72023773e-01 2.27460891e-01 6.16587162e-01 -4.12609905e-01
-7.48940885e-01 -3.27662915e-01 -1.55934170e-01 1.38218808e+00
1.03951044e-01 -6.09506428e-01 -9.41259563e-01 -9.89376307e-01
-3.35424989e-01 7.30587304e-01 -7.43291616e-01 -1.11782894e-01
2.94423044e-01 -1.14034486e+00 5.33753812e-01 1.05060339e-01
7.12021112e-01 -1.13727665e+00 -6.76663101e-01 1.13269307e-01
-7.71561444e-01 -7.93959439e-01 -4.67006629e-03 3.45346779e-01
-5.47103524e-01 -1.07750130e+00 -4.92745757e-01 -5.45279622e-01
4.18842643e-01 2.68172860e-01 1.18664503e+00 -3.18155587e-02
2.36742198e-01 3.81744474e-01 -9.48033988e-01 -6.00794852e-01
-3.51911455e-01 3.75717759e-01 -1.06310379e-02 7.56570995e-02
8.96176159e-01 -3.67998779e-01 -2.79588103e-01 5.39824069e-01
-7.83096552e-01 1.56884059e-01 -9.09647644e-02 8.69697094e-01
-2.55416334e-01 4.70300168e-01 8.96257520e-01 -1.05142820e+00
1.26742554e+00 -6.18945658e-01 1.52744383e-01 1.68630823e-01
-3.12251449e-01 -2.67584443e-01 4.49177146e-01 -6.07757390e-01
-1.43937290e+00 -4.30173695e-01 -3.16175580e-01 -3.13491859e-02
-3.69596004e-01 3.16749930e-01 -3.69100422e-01 1.22221388e-01
3.10713857e-01 -4.08181071e-01 3.00808191e-01 -3.24358672e-01
-1.40010193e-01 1.44647753e+00 -3.00228477e-01 -6.48620307e-01
4.18846458e-02 -3.21369506e-02 -7.06695497e-01 -7.08666444e-01
-5.12135029e-01 -5.49327314e-01 -2.04366893e-01 -4.97796267e-01
6.98523760e-01 -6.14913225e-01 -1.16183603e+00 7.40561247e-01
-1.03086829e+00 -5.82264781e-01 5.54038733e-02 5.19075334e-01
-2.78368890e-01 2.60340750e-01 -7.35211194e-01 -1.28856409e+00
-2.59735882e-01 -9.77088094e-01 7.52261281e-01 3.66427094e-01
-7.03254640e-01 -1.10808027e+00 2.14694396e-01 5.69662631e-01
5.11206925e-01 9.29068848e-02 9.00574267e-01 -1.01670301e+00
7.09819436e-01 -2.37195238e-01 -1.01448536e-01 3.39903682e-01
3.24900538e-01 5.75103424e-02 -1.24853599e+00 -4.80353385e-02
1.74039736e-01 -7.87939787e-01 4.27508026e-01 -7.70610198e-02
1.15764940e+00 -7.74096027e-02 -9.99216139e-02 5.24671823e-02
8.43142092e-01 7.98098862e-01 6.84073508e-01 4.71718490e-01
3.04557204e-01 1.22047341e+00 8.58014762e-01 7.24217117e-01
8.83085966e-01 5.98639429e-01 1.77932411e-01 -2.13137835e-01
5.23650527e-01 1.98586777e-01 3.45674127e-01 1.08984363e+00
-1.59777831e-02 -3.31535757e-01 -7.56149292e-01 4.33494449e-01
-2.01923800e+00 -9.23955619e-01 9.66172218e-02 1.88515508e+00
8.86725128e-01 -3.12948972e-01 2.65854001e-01 2.25910246e-01
8.64682853e-01 2.19225466e-01 -2.47416407e-01 -1.01445854e+00
1.24172419e-01 -1.21624328e-01 -4.46811877e-02 4.38938588e-01
-9.69441414e-01 9.00510073e-01 6.40585089e+00 8.72581899e-01
-9.71488535e-01 1.68071032e-01 1.13886487e+00 8.61884952e-02
-3.20587248e-01 -4.35070902e-01 -5.73095560e-01 8.01651776e-01
7.02232182e-01 7.44533986e-02 2.51477659e-01 8.72850001e-01
2.75508970e-01 -4.19774443e-01 -5.54759145e-01 1.18927014e+00
9.99885648e-02 -5.67650557e-01 -4.33158696e-01 -3.82927223e-03
5.48544765e-01 -1.73466146e-01 -3.50648522e-01 8.21794450e-01
5.07924259e-01 -6.64853275e-01 2.61222392e-01 2.19583482e-01
3.59120458e-01 -9.64960933e-01 1.33184361e+00 3.32834125e-01
-7.63767898e-01 2.04761446e-01 -3.27722877e-01 -4.49808866e-01
-2.86768883e-01 5.25077879e-01 -9.67602551e-01 3.15665275e-01
8.80235672e-01 4.00288343e-01 -4.50444609e-01 7.67805517e-01
3.77388209e-01 4.13602024e-01 -1.92397639e-01 -5.69023788e-01
6.74401745e-02 -5.16585290e-01 2.45052963e-01 1.33724582e+00
2.45866433e-01 1.78674355e-01 2.99957544e-01 1.22083068e-01
1.05751865e-01 5.71692526e-01 -7.19987094e-01 2.43141040e-01
6.76114082e-01 1.53921103e+00 -8.80102873e-01 -4.41809356e-01
-2.20647693e-01 8.04375350e-01 3.56632203e-01 2.37324640e-01
-8.05523336e-01 -5.76699197e-01 9.40159142e-01 -3.71212363e-01
-2.38349244e-01 8.76136124e-02 -3.21096629e-01 -1.19730830e+00
-2.64715135e-01 -9.46700454e-01 4.57381666e-01 -6.25368059e-01
-1.59065127e+00 9.32084620e-01 -4.96000350e-02 -1.03260589e+00
-3.58326316e-01 -3.94979239e-01 -5.21962345e-01 7.24607289e-01
-1.03802860e+00 -8.16572905e-01 -5.26935697e-01 4.50391233e-01
4.81404364e-01 -1.88386887e-01 1.19015729e+00 2.20387951e-01
-8.16544771e-01 6.36548162e-01 -1.14817306e-01 1.84315536e-02
9.67588961e-01 -1.26451659e+00 -3.81438673e-01 1.22620359e-01
-4.46627110e-01 5.22264481e-01 7.49971688e-01 -4.73810852e-01
-9.04513299e-01 -7.54647195e-01 8.70946944e-01 -3.13681781e-01
2.83923894e-01 -4.92339253e-01 -7.67178774e-01 2.73496538e-01
6.34601414e-01 -8.24633241e-01 1.40797853e+00 8.62140596e-01
-1.87095448e-01 -3.70453484e-02 -1.59960699e+00 6.81594610e-01
6.46883786e-01 -3.27763110e-01 -7.23461390e-01 -1.27197048e-02
2.96656936e-01 -2.53717959e-01 -9.06711757e-01 4.45777416e-01
5.94143569e-01 -1.36102903e+00 2.46033907e-01 -4.19329673e-01
2.73672104e-01 4.80116606e-02 1.41622052e-02 -2.02587676e+00
-1.79460183e-01 -4.79439437e-01 3.50514621e-01 1.72112536e+00
2.46086687e-01 -1.00011075e+00 8.79480839e-01 9.51509476e-01
1.37169540e-01 -7.65423954e-01 -5.98397076e-01 -3.08622032e-01
-3.09450001e-01 -5.32789350e-01 8.68239403e-01 1.51878798e+00
7.11278498e-01 3.69411260e-01 -5.11738300e-01 -4.24472451e-01
-6.52542412e-02 4.84327711e-02 9.18161094e-01 -1.14005220e+00
1.63299292e-02 -6.23662829e-01 -2.55355388e-01 -6.74601376e-01
3.87625694e-01 -3.81305724e-01 2.54812717e-01 -1.17729020e+00
2.37811387e-01 -8.50767493e-01 -1.04377260e-02 6.75251365e-01
-3.73101711e-01 5.19261479e-01 1.98708981e-01 -1.31396368e-01
-7.02615321e-01 8.49143922e-01 1.04353809e+00 7.11618364e-03
-4.30498660e-01 1.22210778e-01 -8.49226236e-01 7.55849957e-01
1.14943206e+00 -1.34475052e-01 -5.51030517e-01 -4.05286364e-02
1.05698243e-01 -2.18316317e-01 -3.95254135e-01 -8.17179978e-01
-8.71705934e-02 -1.58720806e-01 1.56750992e-01 -3.34297776e-01
4.19305861e-01 -1.02030671e+00 1.20696193e-02 -1.12099044e-01
-2.64294356e-01 9.98306945e-02 3.57884169e-01 -1.90825015e-01
-3.88886184e-01 -4.04331833e-01 5.84410131e-01 -2.47643748e-03
-3.66169691e-01 3.48538458e-02 -9.54549849e-01 7.83717260e-03
1.09806085e+00 -2.63689935e-01 -1.91106021e-01 -8.04778457e-01
-5.90692043e-01 3.69118989e-01 5.98978281e-01 6.01145566e-01
2.11685613e-01 -1.26225984e+00 -3.21021378e-01 2.41352245e-02
2.73898661e-01 -2.74306506e-01 5.39859116e-01 6.26095355e-01
-1.19485803e-01 5.57047091e-02 -4.29821730e-01 -1.06818698e-01
-1.75784898e+00 3.57053131e-02 2.28005394e-01 -4.82787251e-01
-6.26158118e-02 4.89072710e-01 -3.18543091e-02 -9.08517897e-01
3.87279212e-01 -2.17062995e-01 -8.00348401e-01 7.67415941e-01
5.14210820e-01 4.36136216e-01 1.81190059e-01 -6.16663754e-01
-1.59612343e-01 1.77193999e-01 4.13718708e-02 -5.60937710e-02
1.25795329e+00 -4.17460948e-01 -3.68701518e-01 8.37425888e-01
1.09364116e+00 3.17316294e-01 -8.46561790e-01 1.34259969e-01
-1.73953101e-02 -4.49894011e-01 -1.27736613e-01 -8.25715184e-01
-9.38525319e-01 5.96859753e-01 5.44468760e-01 9.57678497e-01
1.13720024e+00 -2.74606198e-01 7.29899406e-01 1.82800859e-01
4.25734699e-01 -1.26065266e+00 -3.52726728e-02 8.86386037e-01
3.99241388e-01 -1.25254369e+00 -5.26588857e-01 -4.71152306e-01
-1.17807639e+00 9.11899745e-01 9.57590878e-01 5.26513577e-01
5.00510573e-01 2.64153689e-01 5.64941347e-01 -2.34317258e-01
-9.11924958e-01 -2.49859691e-01 -1.41377494e-01 7.40616620e-01
6.00084126e-01 2.81184912e-01 -5.27294636e-01 1.04889226e+00
-3.28246027e-01 -2.91810721e-01 3.16366524e-01 7.51183271e-01
-2.25612625e-01 -1.48521972e+00 -4.31516647e-01 5.55537999e-01
-4.20538515e-01 1.08717255e-01 -6.23796463e-01 6.40676558e-01
2.39653632e-01 1.48949111e+00 1.21090196e-01 -9.23628449e-01
4.28763688e-01 5.97122669e-01 6.32705465e-02 -3.12902510e-01
-9.37098026e-01 -2.16210976e-01 7.25337029e-01 -3.18612814e-01
-6.16062939e-01 -4.82323647e-01 -9.86550093e-01 -6.08040333e-01
-4.67333287e-01 9.89513814e-01 6.43698156e-01 8.79409969e-01
5.20834982e-01 5.11300147e-01 1.09424233e+00 -8.50185096e-01
-3.36917117e-02 -1.46920478e+00 -6.80567384e-01 9.80553508e-01
-3.34486008e-01 -9.31042671e-01 -4.55372781e-01 -3.32844496e-01] | [12.969099044799805, 6.114431858062744] |
a6cc46b8-15a3-4f29-bd71-9d858524d655 | accurate-detection-of-mediastinal-lesions | 2303.11214 | null | https://arxiv.org/abs/2303.11214v1 | https://arxiv.org/pdf/2303.11214v1.pdf | Accurate Detection of Mediastinal Lesions with nnDetection | The accurate detection of mediastinal lesions is one of the rarely explored medical object detection problems. In this work, we applied a modified version of the self-configuring method nnDetection to the Mediastinal Lesion Analysis (MELA) Challenge 2022. By incorporating automatically generated pseudo masks, training high capacity models with large patch sizes in a multi GPU setup and an adapted augmentation scheme to reduce localization errors caused by rotations, our method achieved an excellent FROC score of 0.9922 at IoU 0.10 and 0.9880 at IoU 0.3 in our cross-validation experiments. The submitted ensemble ranked third in the competition with a FROC score of 0.9897 on the MELA challenge leaderboard. | ['Klaus H. Maier-Hein', 'Peter M. Full', 'Michael Baumgartner'] | 2023-03-20 | null | null | null | null | ['medical-object-detection'] | ['computer-vision'] | [ 1.00092115e-02 2.46879030e-02 -2.91337296e-02 1.22812271e-01
-1.35427487e+00 -4.82134640e-01 5.78994215e-01 4.14751798e-01
-7.06964731e-01 5.62788963e-01 5.46317287e-02 -4.32486236e-01
7.54446760e-02 -2.94322312e-01 -5.57772636e-01 -6.93957448e-01
-8.76675323e-02 6.18682683e-01 8.54830265e-01 -1.74314436e-02
-9.76333302e-03 6.15143001e-01 -1.05067801e+00 6.37780905e-01
5.02834082e-01 1.00923347e+00 1.49843737e-01 1.41807306e+00
5.17881453e-01 7.13646650e-01 -6.02230787e-01 -5.84704094e-02
2.51084298e-01 -1.30767569e-01 -8.57469738e-01 -2.47915871e-02
4.91834909e-01 -4.02766615e-02 -2.63374686e-01 4.67368603e-01
6.57867610e-01 2.81292293e-02 6.35740638e-01 -6.63545012e-01
3.14522505e-01 1.92894995e-01 -7.38602936e-01 8.92934620e-01
-1.18650600e-01 1.50424421e-01 5.88294685e-01 -9.27446723e-01
6.38277829e-01 6.05885446e-01 9.92689371e-01 3.78030181e-01
-1.14129031e+00 -1.81063801e-01 -4.12383944e-01 1.89340096e-02
-1.33618581e+00 -9.14920121e-02 -6.67632744e-02 -1.65112182e-01
1.13314402e+00 6.99476242e-01 3.61487001e-01 8.40882361e-01
4.99046266e-01 4.14524168e-01 1.13076746e+00 -5.18783987e-01
-2.60885013e-03 1.97764933e-01 6.29052445e-02 1.01053798e+00
2.66623527e-01 -3.54016945e-02 -3.34292293e-01 -5.70977509e-01
6.14271462e-01 -1.19783297e-01 -6.68849722e-02 -1.57589674e-01
-1.83256614e+00 8.24537516e-01 6.53176188e-01 2.77429789e-01
-4.15205687e-01 1.98277935e-01 4.75080639e-01 -1.28044993e-01
4.08714414e-01 3.61554652e-01 -1.40279487e-01 1.51797794e-02
-7.76606500e-01 -1.22141682e-01 5.69594622e-01 6.33441269e-01
-1.60109803e-01 -2.17558429e-01 -5.41047871e-01 7.75229037e-01
8.83534551e-02 2.58644015e-01 7.19637811e-01 -6.29808307e-01
1.84737965e-02 2.51157790e-01 -1.87732160e-01 -3.88906002e-01
-1.08068061e+00 -9.61612105e-01 -1.03344977e+00 1.52550355e-01
5.63036025e-01 -2.58057773e-01 -1.26688862e+00 1.29856503e+00
6.45284176e-01 1.57360658e-01 -3.91278192e-02 9.65021729e-01
7.61381328e-01 5.21518111e-01 4.27819230e-02 4.00394052e-02
1.77164650e+00 -1.29723597e+00 -3.39393690e-02 -1.40128762e-01
1.18601263e+00 -1.11652851e+00 6.88704014e-01 5.11478782e-01
-9.21274006e-01 -3.90297741e-01 -1.15320122e+00 2.91914135e-01
7.67586976e-02 5.62346280e-01 4.18249428e-01 7.84033298e-01
-1.27448523e+00 4.79334921e-01 -1.08627594e+00 -5.68202078e-01
4.04835135e-01 4.14035887e-01 -2.81737655e-01 2.59407703e-02
-5.97766221e-01 1.01299655e+00 4.87387985e-01 -1.93199515e-01
-7.57983744e-01 -6.40952528e-01 -2.14300334e-01 -3.85267258e-01
5.11803508e-01 -8.98384929e-01 1.23900652e+00 -4.87820506e-01
-1.11368239e+00 1.17912996e+00 2.25159392e-01 -9.91307616e-01
8.41651917e-01 1.78633958e-01 -3.98645848e-01 4.07834172e-01
3.37451696e-02 7.50843585e-01 6.58029079e-01 -6.41561806e-01
-9.20913517e-01 -2.38701433e-01 -2.84467340e-01 2.14603722e-01
-2.60811038e-02 -7.99918361e-03 -6.04631305e-01 -8.17359567e-01
3.13715488e-01 -1.28616142e+00 -5.19279301e-01 1.25628030e-02
-2.55526155e-01 7.22147971e-02 4.91880327e-01 -6.93955958e-01
7.66071498e-01 -1.86483395e+00 -2.72989839e-01 5.72815478e-01
5.68633854e-01 2.27622420e-01 1.30401149e-01 -6.83333427e-02
-5.46471477e-02 -1.61597028e-01 3.23319845e-02 -2.52220094e-01
-5.20814717e-01 1.50704399e-01 2.04215631e-01 6.93630338e-01
2.34412821e-03 8.07595432e-01 -8.05767477e-01 -7.83767402e-01
1.79594770e-01 5.30911446e-01 -5.00345170e-01 2.33992785e-02
1.67642355e-01 2.62121350e-01 -1.69894069e-01 4.83554184e-01
3.38599294e-01 -6.28211141e-01 7.85461888e-02 -1.70153618e-01
9.17234868e-02 7.81570561e-03 -1.00154197e+00 1.71569490e+00
-4.19122905e-01 5.11421204e-01 7.95635357e-02 -3.70152354e-01
5.31556904e-01 3.92279148e-01 6.07566953e-01 -3.72524440e-01
2.43439153e-01 2.84294695e-01 3.17438960e-01 -1.76990747e-01
3.90651852e-01 -2.01225460e-01 2.88668573e-02 4.37370002e-01
5.39985187e-02 4.65362296e-02 8.72419849e-02 4.75007027e-01
1.62911820e+00 -4.48683619e-01 3.87111008e-01 -5.68695247e-01
4.53408718e-01 2.85624087e-01 9.69093814e-02 1.20604765e+00
-3.85829568e-01 1.21240270e+00 2.34253958e-01 -4.76733029e-01
-9.10307109e-01 -9.89471734e-01 -4.52139378e-01 9.57104683e-01
-1.81395113e-01 -5.09545505e-01 -6.71676636e-01 -1.06994522e+00
-3.31883907e-01 3.84046078e-01 -7.12699294e-01 -8.41179416e-02
-8.39791059e-01 -1.35253870e+00 5.24303555e-01 8.04046094e-01
1.91684902e-01 -8.64106417e-01 -8.80317807e-01 2.61186123e-01
-2.77304441e-01 -1.16959107e+00 -5.30572116e-01 3.67313564e-01
-8.51488709e-01 -1.17081380e+00 -1.08655071e+00 -8.12395334e-01
6.93342507e-01 2.82224029e-01 1.20800161e+00 2.98163265e-01
-9.38684821e-01 1.51555553e-01 -3.05439144e-01 -3.71642202e-01
-7.27124929e-01 3.47473085e-01 -1.87256306e-01 -3.10076982e-01
-2.69792795e-01 8.42100754e-02 -6.64496958e-01 5.24892330e-01
-7.85841882e-01 3.30177367e-01 6.97755277e-01 1.14637840e+00
6.48090661e-01 -4.87226337e-01 2.63487935e-01 -9.35737789e-01
1.81955665e-01 -1.50832206e-01 -5.28164446e-01 1.18046813e-01
-5.28196096e-01 -1.36071891e-01 4.37705696e-01 -3.92113060e-01
-5.07569611e-01 1.50561258e-01 -3.53884727e-01 -2.43876487e-01
-1.01871438e-01 9.12793428e-02 5.56590438e-01 -3.54661107e-01
1.11751962e+00 -2.04209890e-02 -6.67334571e-02 -2.19417423e-01
-5.92232905e-02 3.30823183e-01 5.62067628e-01 -2.01418132e-01
4.86458033e-01 5.85955143e-01 3.38773429e-01 -7.25030303e-01
-7.87329316e-01 -7.24695086e-01 -5.19438565e-01 -2.23584130e-01
7.54241765e-01 -7.81709075e-01 -2.49305978e-01 4.28493589e-01
-9.05075669e-01 -3.55043888e-01 -3.66578549e-01 5.84366500e-01
-1.74432322e-01 4.33292329e-01 -5.69087207e-01 -4.10883099e-01
-7.58647084e-01 -1.13802183e+00 1.13173330e+00 3.37255858e-02
-3.62750918e-01 -7.21242309e-01 2.63769567e-01 4.65721220e-01
7.36670136e-01 4.72004592e-01 4.08706635e-01 -8.24260175e-01
-2.63176233e-01 -7.28711903e-01 -3.18415612e-01 6.36600181e-02
-1.33820847e-01 -2.29070559e-01 -9.53175426e-01 -5.58511078e-01
-9.51108187e-02 -1.97481573e-01 1.09281039e+00 3.29491675e-01
1.33435619e+00 1.75322354e-01 -5.82194149e-01 5.86420476e-01
1.05443358e+00 -1.44707158e-01 2.96386778e-01 3.93383324e-01
4.78987098e-01 1.01166116e-02 5.14569044e-01 3.23680967e-01
3.81883122e-02 7.73923695e-01 5.37751257e-01 -2.61104941e-01
-4.15202290e-01 1.72532469e-01 -2.34090850e-01 3.98759991e-01
-5.02489656e-02 -3.06181848e-01 -1.29711485e+00 4.47690874e-01
-1.56621432e+00 -4.71939743e-01 -5.17977715e-01 2.16903305e+00
4.98414338e-01 5.65481186e-01 3.20786744e-01 -1.20142959e-01
7.07907557e-01 5.03694229e-02 -8.52383301e-02 -3.93040776e-02
-8.18107128e-02 3.83846641e-01 7.58829892e-01 3.91906559e-01
-1.42922473e+00 4.41232681e-01 6.40808773e+00 9.72987711e-01
-9.20145690e-01 5.77377379e-01 9.60394800e-01 -4.53205943e-01
5.27012587e-01 -1.67099640e-01 -7.52814949e-01 3.67007852e-01
1.14400220e+00 2.75768459e-01 -1.45658597e-01 7.05790162e-01
-2.92317212e-01 -4.96961236e-01 -7.37703919e-01 9.40769315e-01
3.38988066e-01 -1.47217822e+00 -3.72755170e-01 1.17443681e-01
7.90165484e-01 6.95062280e-01 -6.93180859e-02 1.67688787e-01
-1.80938337e-02 -1.05630493e+00 3.22590172e-01 1.95208669e-01
8.32079053e-01 -4.65544522e-01 1.05721855e+00 4.13151205e-01
-7.64666557e-01 3.39459121e-01 -1.65837705e-01 4.14468259e-01
6.11136071e-02 6.26357138e-01 -1.72940600e+00 3.52085918e-01
6.44435287e-01 6.13325872e-02 -8.34113538e-01 1.35630763e+00
2.95083392e-02 8.09394658e-01 -9.17944849e-01 -8.39278921e-02
3.02912205e-01 6.51877344e-01 6.94845974e-01 1.44419181e+00
1.78399637e-01 1.12270549e-01 -5.34032881e-02 -1.98931880e-02
-1.25251204e-01 9.98355225e-02 1.51394103e-02 7.35180557e-01
-1.40873520e-02 1.70472693e+00 -1.18110681e+00 -4.66397047e-01
5.43640107e-02 9.00306106e-01 -3.96295562e-02 -2.50912756e-01
-1.16305912e+00 1.83437802e-02 -3.57389778e-01 4.85614628e-01
3.83363575e-01 1.01263702e-01 -1.89067826e-01 -1.08329940e+00
-9.06672254e-02 -7.04923809e-01 9.53575432e-01 -6.72151208e-01
-9.20046449e-01 9.91815329e-01 -3.30775142e-01 -1.22327197e+00
-5.68149537e-02 -7.71191061e-01 -7.83995330e-01 6.76320732e-01
-1.09698343e+00 -9.45652962e-01 -4.68637884e-01 2.58500904e-01
2.64323294e-01 -6.68903813e-02 9.81056929e-01 8.20993036e-02
-4.71319556e-01 9.61879849e-01 1.15834504e-01 -7.97230527e-02
7.88720012e-01 -1.30475819e+00 3.32619220e-01 6.80462003e-01
1.95681676e-01 8.34927186e-02 4.99909282e-01 -4.81884629e-01
-1.03049886e+00 -1.23265481e+00 6.99446738e-01 -6.93753123e-01
5.86646974e-01 -1.78034157e-01 -8.46748233e-01 4.96653676e-01
5.27669350e-03 5.69569647e-01 6.10471487e-01 -8.39510411e-02
-5.92142381e-02 3.94113474e-02 -1.28834641e+00 2.60034680e-01
5.77120900e-01 -4.42611836e-02 -1.84549153e-01 8.93594921e-01
6.40449822e-02 -1.04982316e+00 -9.29048061e-01 6.04441285e-01
4.09469336e-01 -8.15514326e-01 1.07778132e+00 -3.78810436e-01
-9.53092277e-02 -1.77936062e-01 1.36862863e-02 -8.64940107e-01
-3.72303009e-01 -4.63820577e-01 -1.08182929e-01 4.50815916e-01
5.69393754e-01 -6.38268054e-01 1.02022314e+00 9.52384546e-02
-2.35981613e-01 -1.05583000e+00 -1.21233940e+00 -5.20357788e-01
-6.91800341e-02 -3.09715271e-01 -2.30335563e-01 4.54578400e-01
-1.66059092e-01 2.95449317e-01 -7.47369777e-04 2.13504165e-01
5.18337846e-01 -3.68956551e-02 6.74814820e-01 -9.30779755e-01
-8.68409216e-01 -4.86751080e-01 -4.32064921e-01 -5.94224155e-01
-5.87335646e-01 -1.04974782e+00 -1.98605374e-01 -1.34295285e+00
6.13007724e-01 -3.33669424e-01 -6.02986872e-01 5.22420228e-01
-3.80566061e-01 8.58066261e-01 1.21986233e-01 4.13020283e-01
-8.35731506e-01 -7.25117475e-02 8.62860799e-01 9.84117389e-02
2.76284337e-01 2.96345830e-01 -4.68118459e-01 7.77610898e-01
1.16765630e+00 -7.46502340e-01 2.69265264e-01 -1.90048367e-01
-1.48258030e-01 2.27769539e-01 6.84304595e-01 -1.26644075e+00
4.49168645e-02 3.14115226e-01 6.47414863e-01 -9.81712043e-01
5.60039639e-01 -5.39271176e-01 -2.78649554e-02 1.07509971e+00
-1.23880982e-01 2.67524987e-01 5.14273584e-01 4.97841299e-01
1.94918707e-01 -9.27767605e-02 1.13941085e+00 -3.00753992e-02
-3.75960082e-01 2.63417326e-02 -3.77946377e-01 8.86407197e-02
1.35718226e+00 -1.11564137e-01 -3.02639395e-01 4.20125313e-02
-8.70538056e-01 2.95567393e-01 3.47870171e-01 2.55865812e-01
3.02770138e-01 -9.64034438e-01 -9.73797560e-01 3.04098219e-01
1.35682255e-01 5.63091338e-02 3.31689745e-01 1.43966448e+00
-9.24560130e-01 4.93384808e-01 2.06859663e-01 -1.07978082e+00
-1.58299398e+00 1.90400615e-01 7.01662421e-01 -8.95579696e-01
-8.09506774e-01 1.23680675e+00 8.29108283e-02 -3.32194299e-01
1.46910623e-01 -1.67171016e-01 1.84420258e-01 -3.39970917e-01
5.34202754e-01 5.54802954e-01 7.53764093e-01 -4.80296880e-01
-7.17258453e-01 2.07114011e-01 -3.80680859e-01 -1.76447973e-01
9.85756755e-01 3.93979609e-01 3.14462662e-01 -7.59864002e-02
1.02883720e+00 -3.45004983e-02 -7.99040616e-01 -2.37323508e-01
-1.86870605e-01 -1.35908559e-01 4.16435659e-01 -1.25482070e+00
-8.20058048e-01 6.07147038e-01 9.65481341e-01 3.90262678e-02
1.05368638e+00 1.49450168e-01 6.41685009e-01 2.48517811e-01
7.60870203e-02 -8.28747272e-01 1.55552119e-01 4.34772700e-01
6.46175027e-01 -1.32220757e+00 1.44479081e-01 -3.46857548e-01
-6.99509263e-01 9.73198771e-01 5.17767906e-01 -3.15629542e-01
5.32313287e-01 5.03458142e-01 1.96383342e-01 -1.25888020e-01
-9.40594673e-01 1.91332906e-01 4.33827996e-01 2.35066414e-01
3.47139984e-01 1.28238112e-01 -4.21024501e-01 1.88941136e-01
-8.20416398e-03 -3.24132413e-01 4.07530338e-01 1.06880832e+00
-5.33269286e-01 -8.43506694e-01 -7.47949004e-01 6.77335262e-01
-7.94589818e-01 -6.65341225e-03 -1.91046640e-01 9.33543205e-01
-2.74461120e-01 5.72964668e-01 -1.50196508e-01 -1.39143765e-01
4.21504259e-01 -9.91327986e-02 5.18683732e-01 -6.07976198e-01
-1.03054225e+00 5.16544521e-01 2.54493412e-02 -4.84085172e-01
7.21356692e-03 -7.35786617e-01 -1.36488402e+00 -4.80579920e-02
-5.06078899e-01 3.75419706e-02 7.46766090e-01 6.02701247e-01
2.96499759e-01 8.87495399e-01 5.32872677e-01 -6.55431569e-01
-7.47917950e-01 -1.05878186e+00 -2.31530890e-01 1.43985495e-01
1.75815582e-01 -2.21872300e-01 -3.25863302e-01 -3.68695259e-01] | [15.181581497192383, -2.2189218997955322] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.