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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
91e1f26d-2bb7-476b-ae21-13a21c32e7e3 | rethinking-visual-geo-localization-for-large | 2204.02287 | null | https://arxiv.org/abs/2204.02287v2 | https://arxiv.org/pdf/2204.02287v2.pdf | Rethinking Visual Geo-localization for Large-Scale Applications | Visual Geo-localization (VG) is the task of estimating the position where a given photo was taken by comparing it with a large database of images of known locations. To investigate how existing techniques would perform on a real-world city-wide VG application, we build San Francisco eXtra Large, a new dataset covering a whole city and providing a wide range of challenging cases, with a size 30x bigger than the previous largest dataset for visual geo-localization. We find that current methods fail to scale to such large datasets, therefore we design a new highly scalable training technique, called CosPlace, which casts the training as a classification problem avoiding the expensive mining needed by the commonly used contrastive learning. We achieve state-of-the-art performance on a wide range of datasets and find that CosPlace is robust to heavy domain changes. Moreover, we show that, compared to the previous state-of-the-art, CosPlace requires roughly 80% less GPU memory at train time, and it achieves better results with 8x smaller descriptors, paving the way for city-wide real-world visual geo-localization. Dataset, code and trained models are available for research purposes at https://github.com/gmberton/CosPlace. | ['Barbara Caputo', 'Carlo Masone', 'Gabriele Berton'] | 2022-04-05 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Berton_Rethinking_Visual_Geo-Localization_for_Large-Scale_Applications_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Berton_Rethinking_Visual_Geo-Localization_for_Large-Scale_Applications_CVPR_2022_paper.pdf | cvpr-2022-1 | ['visual-place-recognition'] | ['computer-vision'] | [-4.82587159e-01 -5.36176920e-01 -2.87510991e-01 -5.97382858e-02
-1.13873303e+00 -6.41645908e-01 7.26999819e-01 3.27197671e-01
-3.94417793e-01 5.87265074e-01 1.85012713e-01 -3.23567092e-01
1.77923188e-01 -9.29537416e-01 -8.21792006e-01 -4.46449786e-01
-2.20997944e-01 4.71700400e-01 5.69048703e-01 1.67955589e-02
2.29768455e-01 5.30451775e-01 -1.69398069e+00 1.49432197e-01
3.99620593e-01 1.07674348e+00 1.15529016e-01 6.44119918e-01
2.85916418e-01 7.47797012e-01 -5.49712658e-01 -1.84855804e-01
4.31641310e-01 8.72148722e-02 -7.47540414e-01 -5.71798757e-02
1.07726836e+00 -2.73445457e-01 -5.66907823e-01 7.98550725e-01
7.18180060e-01 2.10657775e-01 4.28135961e-01 -1.58235681e+00
-8.04309845e-01 -6.78521544e-02 -8.06437314e-01 4.31375951e-01
4.54073697e-01 7.69782737e-02 8.38918090e-01 -9.70707178e-01
7.83797741e-01 8.73806834e-01 1.14729345e+00 -6.43886328e-02
-9.01719689e-01 -5.67258954e-01 4.70453091e-02 3.19417506e-01
-2.03550911e+00 -4.56026077e-01 2.97800452e-01 -3.81583303e-01
1.06235754e+00 1.98311895e-01 5.34053922e-01 9.44570899e-01
-2.22790256e-01 6.05976820e-01 9.57591176e-01 -4.47611064e-01
1.99457422e-01 6.87801139e-03 -2.77570367e-01 9.18134391e-01
2.83430398e-01 -3.19570787e-02 -3.90657425e-01 -4.53334928e-01
5.30059457e-01 1.81121886e-01 -2.51553237e-01 -6.09126151e-01
-1.33222628e+00 7.36898541e-01 8.03153098e-01 2.57118762e-01
-5.48726246e-02 5.15939891e-01 3.71566594e-01 5.40512316e-02
6.28640771e-01 -2.67348494e-02 -2.34527051e-01 -1.65007025e-01
-1.08697629e+00 2.42104322e-01 5.79024076e-01 1.19577086e+00
1.05203176e+00 -1.65851995e-01 1.14989184e-01 6.14218533e-01
7.96480179e-02 7.78871834e-01 3.95793855e-01 -9.82072055e-01
7.09941864e-01 4.87694472e-01 2.06764758e-01 -1.18265402e+00
-3.24721277e-01 -2.38000974e-01 -5.83010018e-01 1.73089147e-01
7.80589163e-01 4.80502024e-02 -8.29823196e-01 1.44924486e+00
5.44191837e-01 5.98301888e-01 -3.30348641e-01 7.23772347e-01
7.67539561e-01 6.42412484e-01 6.84626028e-02 4.71621692e-01
1.32059693e+00 -1.04188704e+00 2.79058386e-02 -5.65339625e-01
8.88547361e-01 -6.67424560e-01 1.09679711e+00 4.27399166e-02
-6.22334003e-01 -4.42312986e-01 -9.77959514e-01 -2.03895450e-01
-9.97312903e-01 2.65966028e-01 8.45382571e-01 5.86203277e-01
-1.41243804e+00 2.67729402e-01 -7.61591375e-01 -1.00824857e+00
5.90947866e-01 1.34934738e-01 -7.13970006e-01 -3.64061117e-01
-7.91689992e-01 8.17028821e-01 1.32216886e-01 -4.11366373e-01
-6.82490349e-01 -7.74241030e-01 -1.02579391e+00 -1.50239751e-01
1.37461200e-01 -6.15968943e-01 1.09149694e+00 -4.76758331e-01
-6.50969088e-01 1.20228171e+00 -3.12435597e-01 -3.09344828e-01
6.07598364e-01 1.33967772e-01 -5.93880594e-01 6.48592487e-02
7.29902208e-01 7.76429892e-01 5.75144827e-01 -1.09501612e+00
-1.06118619e+00 -4.26665306e-01 -1.82038546e-02 -5.72619364e-02
-2.69947022e-01 -1.23273291e-01 -8.44109297e-01 -4.48454350e-01
-2.38259193e-02 -1.06429100e+00 -1.69643730e-01 2.33289137e-01
-1.88653767e-02 -1.17832109e-01 8.97271574e-01 -5.32185733e-01
1.04702437e+00 -2.35516381e+00 -7.05262303e-01 2.66531736e-01
3.42953771e-01 2.86822200e-01 -1.01928614e-01 7.34798074e-01
6.90162927e-02 4.26701754e-02 1.81431025e-01 -5.46202183e-01
5.23128398e-02 -3.43263336e-02 -2.65410215e-01 9.54524338e-01
-3.54142547e-01 9.79466379e-01 -1.05885100e+00 -5.20548880e-01
3.83915663e-01 4.61799920e-01 -2.85078734e-01 -2.06150874e-01
1.98301747e-01 2.02963918e-01 -3.62956375e-01 9.89475608e-01
9.51237202e-01 -7.15966344e-01 -5.70421331e-02 1.20191209e-01
-1.35385364e-01 -1.03017367e-01 -1.24207568e+00 1.75052702e+00
-5.23119867e-01 9.61736679e-01 -2.37158671e-01 -8.85788143e-01
7.30287313e-01 -8.38280320e-02 4.12687302e-01 -9.76678014e-01
-1.68622598e-01 2.73933977e-01 -8.39804232e-01 -2.07113564e-01
6.22542918e-01 5.60262263e-01 -2.47137785e-01 2.70921469e-01
-1.02129973e-01 1.66506097e-01 2.27822214e-01 2.28747889e-01
1.33563852e+00 -6.98258504e-02 5.79595804e-01 -8.75909105e-02
1.53854802e-01 5.49368560e-01 1.86579376e-01 9.38730299e-01
-4.50144142e-01 8.32721055e-01 4.30059657e-02 -6.92535043e-01
-1.07882619e+00 -1.06482708e+00 -1.00925140e-01 1.11863971e+00
3.21441770e-01 -5.67601323e-01 -5.67523658e-01 -7.04338372e-01
3.19708198e-01 3.24498385e-01 -6.10343456e-01 2.74755388e-01
-4.71662313e-01 -6.63033724e-01 8.20454180e-01 6.79561257e-01
7.95310259e-01 -5.54205954e-01 -6.17513001e-01 -9.64359045e-02
-1.46735460e-01 -1.35676301e+00 -4.99104708e-01 -2.39291072e-01
-4.83439088e-01 -1.17162681e+00 -8.92045200e-01 -8.64164412e-01
5.28401077e-01 7.89812863e-01 1.41029215e+00 2.02232987e-01
-3.78562808e-01 5.41793525e-01 -2.14549556e-01 -1.33713722e-01
8.70510191e-02 2.15582043e-01 -2.00016484e-01 -2.55633742e-01
5.78626633e-01 -5.63368261e-01 -8.54276299e-01 3.15408140e-01
-4.16043729e-01 -2.00679168e-01 2.82327861e-01 5.13894498e-01
7.12682843e-01 -4.06746604e-02 2.66073138e-01 -6.15550697e-01
2.02089325e-01 -8.38846266e-01 -9.03512657e-01 2.65389472e-01
-4.83245403e-01 -3.22272152e-01 5.04823565e-01 -3.13244373e-01
-5.49047410e-01 2.92075187e-01 2.44102869e-02 -3.75114799e-01
-3.59094858e-01 2.76059449e-01 1.21743001e-01 -5.36380529e-01
9.49042797e-01 2.28358284e-01 -4.19512391e-01 -3.59503329e-01
4.90483552e-01 6.53852820e-01 6.87575936e-01 -2.89712936e-01
9.22284901e-01 9.47913826e-01 4.93162349e-02 -7.74250925e-01
-7.37091541e-01 -1.08398414e+00 -4.98434573e-01 -8.09318051e-02
6.59547746e-01 -1.39793026e+00 -6.72889292e-01 3.57800603e-01
-8.71532381e-01 -5.58871925e-01 -8.10635313e-02 3.12956393e-01
-4.96159285e-01 2.53648102e-01 -1.11952163e-01 -3.22554708e-01
-1.52176432e-02 -7.78970540e-01 1.47020030e+00 9.52073187e-02
5.44351786e-02 -1.18523622e+00 3.62304479e-01 2.55051970e-01
4.73424286e-01 4.95777518e-01 2.85273701e-01 -3.48120004e-01
-8.64066005e-01 -3.00900936e-01 -6.60792530e-01 -1.91420734e-01
-8.88557434e-02 -2.26578206e-01 -1.00228631e+00 -4.80604529e-01
-6.55347884e-01 -2.27884591e-01 9.50418234e-01 2.19721690e-01
1.09104323e+00 -4.81184721e-02 -8.32053483e-01 9.61869597e-01
1.91982710e+00 -1.17944397e-01 6.88383758e-01 8.54178011e-01
8.89415681e-01 2.30490882e-02 8.18882108e-01 6.06511950e-01
9.73578691e-01 9.27393377e-01 5.30183077e-01 -4.08752680e-01
-4.13852960e-01 -5.74561715e-01 9.76670813e-03 5.79835773e-02
-1.41792685e-01 -4.51997995e-01 -1.30152202e+00 1.07686496e+00
-2.08970070e+00 -1.15193152e+00 -2.73180246e-01 2.29745650e+00
2.31777325e-01 -2.66005158e-01 2.71395147e-01 -4.92727309e-02
5.74375212e-01 4.40646470e-01 -2.37409800e-01 1.85063437e-01
-6.36752546e-02 7.23307356e-02 1.29877329e+00 2.69473135e-01
-1.53527677e+00 1.12726772e+00 6.55050421e+00 1.09199071e+00
-1.35076761e+00 5.05990684e-01 5.27900338e-01 -3.87180075e-02
2.07357854e-01 3.84497531e-02 -8.89218330e-01 5.44032514e-01
9.23908234e-01 -6.03641830e-02 3.74896437e-01 1.41634429e+00
1.65832683e-01 -4.52294588e-01 -6.99719608e-01 1.38904583e+00
1.19717360e-01 -1.80823338e+00 -4.09138590e-01 1.49641812e-01
9.98973310e-01 7.53740966e-01 -2.03111582e-03 2.09574312e-01
4.00586307e-01 -8.82352948e-01 8.76677752e-01 1.39941022e-01
1.03891170e+00 -5.14265716e-01 6.09196424e-01 2.15001807e-01
-1.68634737e+00 -1.15520671e-01 -5.30316651e-01 -3.67594697e-02
1.38741732e-01 4.17667240e-01 -9.02394950e-01 4.13296163e-01
1.08485925e+00 9.20041442e-01 -1.23516011e+00 1.47030175e+00
-9.51863378e-02 3.66363972e-01 -4.65009600e-01 2.03993186e-01
3.84871662e-01 3.46678734e-01 1.17450133e-01 1.36979473e+00
7.55042374e-01 -3.78837705e-01 4.19653565e-01 3.02299410e-01
-1.15551583e-01 4.69443621e-03 -1.05266786e+00 3.96300405e-01
7.36141324e-01 1.12881005e+00 -7.81273007e-01 -4.84680355e-01
-5.69548726e-01 1.04836214e+00 6.91732466e-01 3.60580385e-01
-1.11785662e+00 -3.38992715e-01 5.62808275e-01 5.83518744e-01
5.58827937e-01 -4.57038641e-01 -8.50335062e-02 -1.24756277e+00
7.42194057e-02 -5.42328715e-01 6.06946647e-01 -1.02012563e+00
-1.05961704e+00 3.85255009e-01 7.85594359e-02 -1.55405962e+00
-1.87726274e-01 -5.65594614e-01 -5.60283065e-01 5.46402395e-01
-1.69199717e+00 -1.56611216e+00 -9.30156767e-01 9.58942533e-01
2.67513692e-01 -1.03442177e-01 7.47863173e-01 7.49845505e-01
-2.85052478e-01 5.79765737e-01 4.20151860e-01 2.04020485e-01
9.75531578e-01 -1.09531188e+00 7.66680360e-01 1.02250910e+00
5.06568789e-01 2.54319221e-01 3.96049559e-01 -4.60641593e-01
-1.21842813e+00 -1.42379653e+00 1.14016914e+00 -7.42957234e-01
8.14407110e-01 -4.83186573e-01 -3.98807555e-01 1.02447152e+00
2.01510992e-02 7.66797662e-01 2.64194489e-01 -5.27764633e-02
-5.63134730e-01 -1.86289400e-01 -1.23271561e+00 4.63076264e-01
1.49189711e+00 -6.12468004e-01 -6.70525208e-02 6.84905767e-01
3.57028633e-01 -6.86549544e-01 -7.11463571e-01 3.51426825e-02
2.73948371e-01 -1.02257812e+00 1.30545914e+00 -1.43680826e-01
-5.12793809e-02 -6.18373036e-01 -2.86469817e-01 -1.18033087e+00
-4.38830733e-01 -2.76650250e-01 1.41256303e-02 1.14791322e+00
1.95601866e-01 -9.46440935e-01 8.89179170e-01 1.89967334e-01
1.04135983e-02 -4.59373504e-01 -1.04824555e+00 -1.13377130e+00
-2.57603794e-01 -5.14800429e-01 8.58582914e-01 1.11587000e+00
-1.92744389e-01 -1.72647834e-01 -3.20417702e-01 5.23265183e-01
5.89096606e-01 8.30877125e-02 1.09544802e+00 -8.70533347e-01
-2.42995840e-04 -9.70555246e-02 -9.99581516e-01 -1.04880250e+00
7.79641941e-02 -7.85205841e-01 -2.87621945e-01 -1.80749512e+00
3.16885076e-02 -8.64393294e-01 -6.01219460e-02 8.18977416e-01
7.17415437e-02 9.99349654e-01 8.42271075e-02 3.87836546e-01
-1.01794612e+00 9.57190320e-02 7.52265215e-01 -3.27047646e-01
1.43465087e-01 -1.03351489e-01 -5.73698044e-01 5.92510045e-01
7.19550312e-01 -5.53335547e-01 -4.27382439e-01 -6.16844356e-01
7.20030516e-02 -3.72119963e-01 7.83455789e-01 -1.44308579e+00
5.06516218e-01 -9.34681371e-02 5.16161084e-01 -5.54257095e-01
4.06805307e-01 -7.49350369e-01 3.16461802e-01 1.75126746e-01
2.93264657e-01 4.89366293e-01 3.05219144e-01 5.12527764e-01
-2.86472470e-01 1.64412916e-01 6.37965798e-01 -2.84855098e-01
-1.60048544e+00 5.48005998e-01 -3.00968755e-02 4.05530423e-01
1.29051685e+00 -4.42327052e-01 -8.49594891e-01 -3.62533689e-01
-4.80345786e-01 7.64560774e-02 9.48890150e-01 2.90233105e-01
3.45100462e-01 -1.56297874e+00 -5.28116763e-01 1.56469420e-02
6.68829739e-01 -3.08339447e-01 1.69669986e-01 7.71752238e-01
-8.80744815e-01 3.69988948e-01 -3.46675739e-02 -7.44811177e-01
-1.25824463e+00 6.69144928e-01 2.96847075e-01 -2.41825819e-01
-7.22426951e-01 6.49474382e-01 -7.68878087e-02 -3.09084415e-01
4.95216437e-03 -1.13474645e-01 1.70518890e-01 2.15073228e-02
6.42082751e-01 5.00452995e-01 4.02293444e-01 -8.83824229e-01
-7.19720006e-01 7.25857973e-01 4.48561579e-01 4.71590832e-02
1.32757163e+00 -2.23912194e-01 2.57942349e-01 6.31129136e-03
1.36780906e+00 2.81039059e-01 -1.24538660e+00 -1.29853278e-01
-6.78509697e-02 -1.00410867e+00 5.26413135e-02 -5.00392437e-01
-1.10915482e+00 5.33211529e-01 9.28582251e-01 1.03189200e-01
9.82130647e-01 2.81552911e-01 6.95028841e-01 2.99922347e-01
8.91659617e-01 -1.00059903e+00 -9.26169753e-02 4.00106996e-01
6.14928007e-01 -1.55454171e+00 1.72160938e-01 -3.09590846e-01
-5.12123466e-01 6.86915278e-01 3.72847199e-01 -2.46823803e-01
7.77351797e-01 1.79418683e-01 2.00354844e-01 -3.59123856e-01
-3.71577889e-01 -5.53833544e-01 8.12479407e-02 1.07542455e+00
-5.62667847e-02 6.71835393e-02 2.87823617e-01 -1.49293870e-01
-2.35450774e-01 2.09371820e-02 2.21952796e-01 9.99394894e-01
-2.73249328e-01 -9.57841396e-01 -4.74574924e-01 1.15825176e-01
-8.69459659e-02 -1.54707342e-01 -9.19915140e-02 1.23233163e+00
4.30581570e-01 8.28622878e-01 1.63800046e-01 -1.22418553e-01
3.43751907e-01 -3.78704071e-01 3.09155256e-01 -4.46425408e-01
-2.86811858e-01 -5.19972444e-01 2.43850544e-01 -9.44450080e-01
-3.79403949e-01 -6.94410026e-01 -9.64856446e-01 -8.40259850e-01
1.12922527e-02 -2.14754716e-01 6.20018125e-01 6.42076373e-01
7.77379751e-01 -4.92096767e-02 4.14582938e-01 -1.16105342e+00
-1.09235056e-01 -4.94422257e-01 -5.78918874e-01 5.79416335e-01
3.04251641e-01 -7.48951733e-01 -3.19528997e-01 -5.46216927e-02] | [7.683736324310303, -1.8760429620742798] |
3583a8df-743a-441c-93de-dd752f1e8cc9 | gpu-accelerated-guided-source-separation-for | 2212.05271 | null | https://arxiv.org/abs/2212.05271v1 | https://arxiv.org/pdf/2212.05271v1.pdf | GPU-accelerated Guided Source Separation for Meeting Transcription | Guided source separation (GSS) is a type of target-speaker extraction method that relies on pre-computed speaker activities and blind source separation to perform front-end enhancement of overlapped speech signals. It was first proposed during the CHiME-5 challenge and provided significant improvements over the delay-and-sum beamforming baseline. Despite its strengths, however, the method has seen limited adoption for meeting transcription benchmarks primarily due to its high computation time. In this paper, we describe our improved implementation of GSS that leverages the power of modern GPU-based pipelines, including batched processing of frequencies and segments, to provide 300x speed-up over CPU-based inference. The improved inference time allows us to perform detailed ablation studies over several parameters of the GSS algorithm -- such as context duration, number of channels, and noise class, to name a few. We provide end-to-end reproducible pipelines for speaker-attributed transcription of popular meeting benchmarks: LibriCSS, AMI, and AliMeeting. Our code and recipes are publicly available: https://github.com/desh2608/gss. | ['Sanjeev Khudanpur', 'Daniel Povey', 'Desh Raj'] | 2022-12-10 | null | null | null | null | ['target-speaker-extraction'] | ['audio'] | [ 2.98685312e-01 -2.90046841e-01 2.17930764e-01 -6.09760106e-01
-1.93972492e+00 -8.43899727e-01 5.16573608e-01 5.21252528e-02
-8.13731626e-02 3.12445283e-01 8.48643124e-01 -3.57269347e-01
7.26719573e-02 1.61255404e-01 -3.72665197e-01 -6.52503014e-01
-3.23500097e-01 1.64062798e-01 2.56784596e-02 -1.09132752e-01
1.26831546e-01 5.96449934e-02 -1.39851713e+00 4.34941769e-01
5.59397399e-01 7.12806463e-01 -9.42554772e-02 1.09213626e+00
2.93992400e-01 4.83229905e-01 -6.49853110e-01 -3.88447084e-02
5.90265393e-02 -6.47026241e-01 -4.84959096e-01 -3.31478268e-01
4.95175600e-01 -2.90582627e-01 -3.93687963e-01 6.81106985e-01
1.23524618e+00 1.20292872e-01 4.03333694e-01 -1.23004305e+00
-3.13611003e-03 1.02310920e+00 -5.84774196e-01 6.47385597e-01
5.38339853e-01 2.53130943e-01 1.10239756e+00 -1.08222938e+00
7.84579292e-02 1.13322806e+00 9.31400478e-01 2.30102882e-01
-1.41397238e+00 -9.10643935e-01 -7.82907382e-02 2.67870096e-03
-1.57904279e+00 -1.48967361e+00 4.87974256e-01 -2.77894258e-01
1.20453608e+00 6.13781631e-01 3.14956069e-01 1.36454058e+00
-5.66561580e-01 8.25611591e-01 9.42637265e-01 -5.18042743e-01
3.69112462e-01 -2.56424606e-01 2.81043708e-01 2.64228880e-01
-2.45958924e-01 9.14181098e-02 -1.30706263e+00 -5.54810226e-01
2.08614364e-01 -5.83768785e-01 -5.53779066e-01 4.14390594e-01
-1.48276699e+00 4.76229012e-01 -2.83096526e-02 1.63401127e-01
-1.59344181e-01 1.88582361e-01 4.59872931e-01 4.00118940e-02
7.08212852e-01 2.79849261e-01 -5.43796718e-01 -5.89762688e-01
-1.60777581e+00 4.95613366e-01 1.00193596e+00 9.27357376e-01
3.28444511e-01 1.31026700e-01 -6.83741510e-01 9.69741046e-01
2.58100063e-01 6.63746536e-01 2.92191595e-01 -8.81126702e-01
4.61888373e-01 -1.89857349e-01 1.07728153e-01 -6.32054687e-01
-5.90417206e-01 -6.75484657e-01 -5.77356339e-01 -3.24670225e-01
5.03778219e-01 -4.96318758e-01 -9.24506545e-01 1.83859015e+00
2.92590410e-01 7.30951428e-01 -1.13460086e-01 1.15968645e+00
8.43405247e-01 5.59721410e-01 -1.77294016e-01 -1.75509498e-01
1.48927331e+00 -1.09922731e+00 -7.18578637e-01 -3.68095517e-01
2.47198179e-01 -1.16441286e+00 8.93382549e-01 4.26021993e-01
-1.29331124e+00 -1.89622760e-01 -1.00602388e+00 -4.61058766e-02
9.71996486e-02 -3.12087946e-02 5.42369723e-01 9.02694464e-01
-1.33393800e+00 3.30944270e-01 -1.07162654e+00 -2.86525249e-01
2.85386801e-01 3.41473967e-01 1.14931785e-01 2.48082772e-01
-1.01186001e+00 2.52799183e-01 -2.45465040e-01 6.44808635e-02
-8.66127431e-01 -1.11142433e+00 -7.44760394e-01 1.92518950e-01
3.35640341e-01 -5.47979236e-01 1.89822781e+00 -6.69374108e-01
-1.68161690e+00 5.97166300e-01 -7.63885260e-01 -5.52748859e-01
1.35323599e-01 -2.16802031e-01 -6.35345578e-01 4.18066941e-02
8.36775377e-02 3.28398675e-01 9.22780871e-01 -6.90768301e-01
-3.67154568e-01 -2.19902545e-01 -5.60961187e-01 3.54360282e-01
-1.85094520e-01 5.83323956e-01 -7.91960239e-01 -8.66791964e-01
4.07674819e-01 -8.39938939e-01 -1.40272453e-01 -6.02896512e-01
-7.39491284e-01 1.95282698e-01 1.54180497e-01 -1.14163518e+00
1.19499671e+00 -2.54338408e+00 3.61509472e-02 3.52657884e-01
2.28642613e-01 -1.74666718e-02 -1.93194449e-01 5.65756261e-01
-1.79051608e-01 -1.73140407e-01 -4.88440067e-01 -8.69568408e-01
2.66820520e-01 -6.21489942e-01 -3.74760091e-01 4.95667070e-01
2.44862288e-02 6.34072304e-01 -7.77868927e-01 -2.37359554e-01
-1.77289993e-02 7.84321606e-01 -7.29339600e-01 3.64778668e-01
1.29847780e-01 6.24751389e-01 6.93838522e-02 6.93438172e-01
7.67370164e-01 9.38371867e-02 1.45902365e-01 -1.74938500e-01
-1.70593113e-01 1.01993656e+00 -1.29639423e+00 2.12673450e+00
-3.92654061e-01 7.57790148e-01 7.26839721e-01 -4.79045093e-01
4.43632960e-01 6.21712327e-01 3.34980637e-01 -3.26295316e-01
3.17078643e-02 4.04329032e-01 -5.39800450e-02 -2.92627722e-01
3.39493006e-01 1.31186277e-01 -7.06333444e-02 4.81664628e-01
2.30270118e-01 -2.98887998e-01 7.81689361e-02 4.83365238e-01
1.30051112e+00 -8.50521773e-02 1.19472958e-01 -3.22347879e-01
1.41105503e-01 -2.35250130e-01 3.73200417e-01 7.44119227e-01
-1.05195269e-01 9.61357117e-01 2.73848146e-01 3.37693959e-01
-6.98708951e-01 -1.15463388e+00 -2.78442763e-02 1.54050124e+00
-4.37087297e-01 -8.28949571e-01 -1.01572227e+00 -2.21632645e-02
-1.58496886e-01 7.98683226e-01 -1.19870856e-01 1.85445026e-01
-4.25191581e-01 -9.45003271e-01 1.12038136e+00 4.01689500e-01
6.40523210e-02 -7.24633336e-01 -3.74613643e-01 3.10083508e-01
-5.34838557e-01 -1.15996480e+00 -1.00492632e+00 3.62808287e-01
-4.04744804e-01 -5.49813509e-01 -7.40314424e-01 -5.34761310e-01
1.72306970e-01 4.00685370e-01 1.16891241e+00 -2.31986001e-01
-1.98289976e-01 4.43067282e-01 -2.07020089e-01 -5.38402021e-01
-2.89910197e-01 2.96193063e-01 1.65822387e-01 8.08493569e-02
2.91426331e-01 -9.01234150e-01 -7.58923769e-01 1.05558477e-01
-3.61105561e-01 8.99305940e-02 2.85816878e-01 5.92690170e-01
2.12116897e-01 -3.67369533e-01 6.28523111e-01 -4.74725813e-01
6.89875782e-01 -5.02462506e-01 -5.09064317e-01 -2.08575517e-01
-2.56528825e-01 -1.91947311e-01 1.51666462e-01 -8.52755085e-02
-1.02577722e+00 2.80166082e-02 -5.53043485e-01 -1.93364881e-02
-8.10961276e-02 4.70487744e-01 -1.37521029e-01 2.59237826e-01
6.43346310e-01 2.13175461e-01 -1.45167187e-01 -6.96425259e-01
2.86244243e-01 1.10178185e+00 8.48925352e-01 -3.79603416e-01
5.29366553e-01 1.77983567e-01 -5.75215220e-01 -8.62337708e-01
-6.80464327e-01 -8.07118714e-01 -1.43447429e-01 1.96216092e-01
4.98974979e-01 -1.42630553e+00 -6.92525744e-01 6.46544635e-01
-1.09013999e+00 -4.73716468e-01 1.48193628e-01 5.61612904e-01
-3.22163701e-01 1.04907401e-01 -6.78650379e-01 -1.13038540e+00
-6.84670448e-01 -1.01586759e+00 1.31948352e+00 2.26410734e-03
-5.47802985e-01 -5.00588417e-01 5.67351356e-02 4.94607270e-01
6.74603164e-01 -3.46751064e-02 2.07651421e-01 -7.46990979e-01
-2.18158454e-01 8.43806192e-02 5.05343489e-02 1.02884442e-01
-1.08558625e-01 -1.67352796e-01 -1.70686615e+00 -4.87042785e-01
-6.45573111e-03 -3.03351823e-02 9.89277840e-01 7.05700994e-01
9.68167663e-01 -3.30609381e-01 -2.83718795e-01 9.49689269e-01
8.86865854e-01 -2.03933388e-01 2.43824035e-01 -1.47231594e-01
4.60770309e-01 3.05779576e-01 1.77729055e-01 6.30321145e-01
4.02822047e-01 9.01057780e-01 3.71441394e-02 -2.08410159e-01
-3.66051316e-01 4.14957888e-02 6.46484256e-01 1.20557535e+00
1.91438049e-01 -1.76609099e-01 -1.01854587e+00 6.66980863e-01
-1.52550411e+00 -8.94311190e-01 -2.80257523e-01 2.37093067e+00
1.17234719e+00 -5.34696458e-03 4.92977470e-01 3.21232200e-01
6.57888472e-01 1.52323708e-01 -4.83083993e-01 -1.90368071e-01
-2.20604599e-01 4.10511404e-01 4.91313457e-01 9.24208641e-01
-1.05523074e+00 6.44903839e-01 6.17463160e+00 8.56839538e-01
-9.96115565e-01 4.97597575e-01 4.37535226e-01 -8.20958674e-01
-1.41904101e-01 -1.78983241e-01 -8.53903055e-01 4.31982517e-01
1.56226301e+00 -6.15264960e-02 9.64062035e-01 3.19915086e-01
4.59361374e-01 -1.48045957e-01 -1.19779718e+00 1.13287723e+00
7.98689723e-02 -1.10848105e+00 -7.72915065e-01 -2.07694188e-01
4.18729812e-01 6.14483297e-01 6.92425892e-02 1.60046726e-01
2.24056214e-01 -8.63965988e-01 1.01180172e+00 -9.60003287e-02
8.33696008e-01 -5.63631892e-01 3.36052030e-01 1.11246509e-02
-1.19996333e+00 -1.17681921e-02 2.41612718e-01 -3.47693153e-02
2.09396094e-01 8.90763998e-01 -1.15004814e+00 4.92322326e-01
7.55698264e-01 2.12250382e-01 -3.99001628e-01 9.78395402e-01
-1.49824649e-01 1.57331109e+00 -7.71153629e-01 2.56224841e-01
7.84713551e-02 3.08700919e-01 9.03339505e-01 1.85843325e+00
4.34237450e-01 1.41055927e-01 -4.21401300e-02 6.10098124e-01
4.59274314e-02 -8.56730249e-03 -5.20014092e-02 1.70534924e-01
9.83295262e-01 1.35906196e+00 -5.42723000e-01 -2.05136433e-01
-2.97722608e-01 8.88010383e-01 1.16047999e-02 5.96073031e-01
-1.10573924e+00 -4.23921496e-01 8.47445607e-01 1.53066322e-01
3.17540824e-01 -2.64487892e-01 -4.21010107e-01 -1.01519549e+00
-8.86150897e-02 -1.25257409e+00 2.88055748e-01 -5.32274306e-01
-8.56535971e-01 5.40641904e-01 -1.15660869e-01 -8.50230575e-01
-4.46755290e-01 -1.46920653e-02 -6.90479755e-01 1.21474016e+00
-1.14393294e+00 -9.16083574e-01 -1.82961330e-01 6.45497680e-01
5.98548412e-01 -1.39571220e-01 9.96948123e-01 4.79051650e-01
-7.42486596e-01 9.41977978e-01 1.56256139e-01 -3.48377898e-02
1.04911542e+00 -1.14288509e+00 9.58435297e-01 1.06584358e+00
3.40079188e-01 7.06417620e-01 1.13370824e+00 -1.64179683e-01
-1.44754207e+00 -9.16202247e-01 8.96673501e-01 -4.82880414e-01
5.82217157e-01 -8.92338216e-01 -5.49782753e-01 6.96085215e-01
6.06255710e-01 -1.62337184e-01 9.24867153e-01 5.56328654e-01
-4.47567940e-01 -1.57214791e-01 -6.64481223e-01 5.46184063e-01
1.06947005e+00 -5.67992032e-01 -2.93549567e-01 3.29744935e-01
6.73897266e-01 -7.28160620e-01 -5.39230287e-01 -2.57600192e-02
4.93576616e-01 -1.12627709e+00 1.01934326e+00 3.30955051e-02
-7.75620863e-02 -3.78844857e-01 -3.27166826e-01 -1.45569968e+00
-3.45980644e-01 -1.36509311e+00 -9.97971892e-02 1.60523701e+00
6.82041109e-01 -5.65519571e-01 2.69623160e-01 3.55731726e-01
-4.13072169e-01 -3.93418461e-01 -1.01942050e+00 -5.56112587e-01
-4.96090531e-01 -8.20397496e-01 6.61084950e-01 6.97714508e-01
1.14239715e-01 5.73203862e-01 -3.29580843e-01 5.11345088e-01
7.33676136e-01 -3.59589905e-02 7.48585820e-01 -7.64323175e-01
-7.33296454e-01 -3.19669068e-01 1.38895884e-01 -9.54893053e-01
-7.48304799e-02 -9.16543663e-01 3.63479763e-01 -1.25904238e+00
5.28219752e-02 -1.51150376e-01 -3.99022341e-01 5.95408857e-01
-2.68840820e-01 5.36136925e-01 1.12356298e-01 6.36727065e-02
-5.36625922e-01 2.53964782e-01 3.48920971e-01 2.80675050e-02
-5.29067755e-01 7.52734616e-02 -9.80280578e-01 6.02202296e-01
9.15842474e-01 -5.33670127e-01 -2.64947504e-01 -6.18251204e-01
-7.31567293e-02 1.04729585e-01 4.32658792e-01 -1.20253420e+00
3.58341187e-01 3.56455803e-01 1.64647833e-01 -5.99549651e-01
6.50597513e-01 -2.33541816e-01 8.78061652e-02 1.30793288e-01
-4.60479975e-01 -2.11885303e-01 4.94202673e-01 3.10828179e-01
-1.00561842e-01 2.52378404e-01 5.15539527e-01 3.60038429e-01
-2.57671297e-01 -3.05977482e-02 -4.80651975e-01 3.21475416e-01
3.59206110e-01 2.56104231e-01 2.20176601e-03 -6.49362028e-01
-5.61372042e-01 9.77720469e-02 -9.34105590e-02 2.78584540e-01
2.35805273e-01 -1.07476008e+00 -1.27717817e+00 2.23046035e-01
-5.82869351e-02 -5.77640533e-02 4.11537170e-01 1.31232762e+00
-1.26299664e-01 3.78538102e-01 4.26841140e-01 -7.07682133e-01
-1.38580036e+00 1.14860140e-01 2.06511080e-01 1.73427053e-02
-5.11310160e-01 1.29847443e+00 1.04420856e-01 -2.71582276e-01
5.43828666e-01 -4.49996501e-01 3.69661510e-01 -8.57373103e-02
9.04419541e-01 5.31882048e-01 4.89902854e-01 -4.51398611e-01
-8.96468461e-01 9.02229846e-02 2.27751955e-01 -8.78501356e-01
1.24211836e+00 -2.82387495e-01 3.42286639e-02 3.65881205e-01
1.07230926e+00 5.66858351e-01 -1.25842524e+00 -2.07849473e-01
-2.53859963e-02 -1.46614641e-01 4.85002697e-01 -1.14520526e+00
-7.82794297e-01 7.26790428e-01 4.91504222e-01 1.10043269e-02
1.39974511e+00 -2.68700756e-02 7.66955256e-01 -6.50893077e-02
4.51152846e-02 -8.47460270e-01 -3.87176365e-01 4.25322175e-01
9.62909639e-01 -1.00536287e+00 -1.80173934e-01 -4.74728823e-01
-4.41664189e-01 7.68572807e-01 9.80145037e-02 3.25099856e-01
5.92277408e-01 9.00658786e-01 2.30441824e-01 1.06083587e-01
-7.38339067e-01 -1.21059828e-01 2.76771009e-01 4.69974577e-01
8.07527959e-01 1.83810204e-01 1.12412795e-01 1.01930320e+00
-5.95218956e-01 -2.01057822e-01 1.84127182e-01 6.08835995e-01
-1.74765229e-01 -7.88119912e-01 -8.34986091e-01 2.30694294e-01
-6.70701683e-01 -7.68629670e-01 -2.23644719e-01 4.78096418e-02
-2.59053230e-01 1.56565809e+00 -3.61666270e-02 -2.77177483e-01
2.77309328e-01 2.61636674e-01 4.26889271e-01 -5.55969119e-01
-9.25731301e-01 7.38026679e-01 3.82785857e-01 -6.64705694e-01
-2.31397584e-01 -9.87165391e-01 -1.16866457e+00 -4.14041936e-01
-1.85380802e-01 2.76396304e-01 9.03147221e-01 6.73580945e-01
7.91936517e-01 7.92274654e-01 5.01367271e-01 -1.24108624e+00
-6.37669384e-01 -1.23378599e+00 -4.66928005e-01 2.33581942e-02
6.08936012e-01 -1.94269910e-01 -5.50874054e-01 1.37247160e-01] | [14.804398536682129, 6.054569244384766] |
955763d2-7d1c-4fa3-89cf-88b7b26986d6 | multimodal-pathology-image-search-between-h-e | 2306.06780 | null | https://arxiv.org/abs/2306.06780v1 | https://arxiv.org/pdf/2306.06780v1.pdf | Multimodal Pathology Image Search Between H&E Slides and Multiplexed Immunofluorescent Images | We present an approach for multimodal pathology image search, using dynamic time warping (DTW) on Variational Autoencoder (VAE) latent space that is fed into a ranked choice voting scheme to retrieve multiplexed immunofluorescent imaging (mIF) that is most similar to a query H&E slide. Through training the VAE and applying DTW, we align and compare mIF and H&E slides. Our method improves differential diagnosis and therapeutic decisions by integrating morphological H&E data with immunophenotyping from mIF, providing clinicians a rich perspective of disease states. This facilitates an understanding of the spatial relationships in tissue samples and could revolutionize the diagnostic process, enhancing precision and enabling personalized therapy selection. Our technique demonstrates feasibility using colorectal cancer and healthy tonsil samples. An exhaustive ablation study was conducted on a search engine designed to explore the correlation between multiplexed Immunofluorescence (mIF) and Hematoxylin and Eosin (H&E) staining, in order to validate its ability to map these distinct modalities into a unified vector space. Despite extreme class imbalance, the system demonstrated robustness and utility by returning similar results across various data features, which suggests potential for future use in multimodal histopathology data analysis. | ['Jacob M Luber', 'Helen H Shang', 'Michael Robben', 'Parisa Boodaghi Malidarreh', 'Aarti Darji', 'Jai Prakash Veerla', 'Mohammad S Nasr', 'MD Jillur Rahman Saurav', 'Amir Hajighasemi'] | 2023-06-11 | null | null | null | null | ['dynamic-time-warping'] | ['time-series'] | [ 3.04604590e-01 -4.37137961e-01 -3.94755036e-01 -1.06180541e-01
-1.10021079e+00 -9.24886107e-01 4.42112088e-01 3.67505074e-01
-7.14565873e-01 5.81886351e-01 2.79671758e-01 -4.21859324e-01
-4.50684607e-01 -6.45189404e-01 -5.86791709e-02 -1.44138348e+00
1.46613829e-02 6.64127827e-01 -3.25838745e-01 1.27938762e-01
-7.49421716e-02 8.38893533e-01 -1.28244889e+00 3.95929813e-01
1.37590557e-01 6.59177542e-01 4.30951267e-01 9.35917377e-01
-2.74040010e-02 6.28112435e-01 -4.23431545e-01 -3.80051732e-01
-1.42798126e-01 -4.32683378e-02 -9.03073370e-01 -3.52695771e-02
4.60015804e-01 -3.85021806e-01 -3.33584249e-01 8.22867155e-01
7.02856898e-01 -2.69750282e-02 9.35124278e-01 -1.12911904e+00
-4.95478332e-01 1.90234147e-02 -3.91393930e-01 6.22184157e-01
7.96070695e-02 2.43034899e-01 1.14366281e+00 -8.58872294e-01
1.51310158e+00 6.69892192e-01 5.41244626e-01 5.54068208e-01
-1.30551684e+00 -3.02039415e-01 -5.56080401e-01 4.22104448e-01
-1.26602995e+00 -3.20928097e-02 3.83972764e-01 -7.76781976e-01
1.00105691e+00 6.21264517e-01 9.26887751e-01 9.87802565e-01
9.30465817e-01 6.27556860e-01 1.04204035e+00 -3.15676033e-01
8.54916051e-02 9.02125910e-02 1.14526361e-01 7.59790838e-01
-7.94939026e-02 1.12230562e-01 -4.19875354e-01 -4.75109965e-01
7.04155684e-01 4.25074369e-01 -3.89969766e-01 -3.07543606e-01
-1.51175356e+00 9.03594553e-01 4.86383677e-01 7.38938689e-01
-3.62591684e-01 -7.31876539e-03 4.04630989e-01 2.15926960e-01
2.29793355e-01 4.66208786e-01 1.67008881e-02 3.50637466e-01
-1.16669428e+00 -1.09883457e-01 2.58783430e-01 -1.56307787e-01
3.22534859e-01 -3.87601584e-01 -2.48121500e-01 6.84882045e-01
3.44086945e-01 4.62233245e-01 9.44332898e-01 -9.82756853e-01
-4.24599320e-01 5.85909247e-01 -2.26255298e-01 -8.29798818e-01
-5.92644930e-01 -3.41540724e-01 -7.43821383e-01 3.18700343e-01
1.36165366e-01 2.28864089e-01 -9.45277810e-01 1.41806149e+00
4.18452621e-01 -1.36739150e-01 -9.30311829e-02 1.00199044e+00
6.24840736e-01 2.41971537e-01 3.54629517e-01 -1.65031195e-01
1.71320534e+00 -2.71280080e-01 -5.62068701e-01 5.18727481e-01
8.30063879e-01 -8.06715786e-01 6.27847970e-01 1.41853020e-01
-6.71199083e-01 -9.00051277e-03 -1.04749620e+00 -2.83796132e-01
-5.35754859e-01 1.41388327e-01 7.60580778e-01 2.91100532e-01
-1.48347545e+00 5.61540067e-01 -1.21838248e+00 -6.63845718e-01
3.32055539e-01 6.38428509e-01 -9.25808012e-01 2.06445679e-02
-9.72685695e-01 1.09282541e+00 -2.44800057e-02 2.98221633e-02
-5.29937327e-01 -7.43900061e-01 -8.02050233e-01 -8.43794420e-02
-2.80076653e-01 -9.47567701e-01 8.84286284e-01 -4.61379260e-01
-1.31055248e+00 1.19446528e+00 -2.99483240e-01 -3.19802910e-01
2.30053440e-01 8.40654910e-01 -5.05177438e-01 5.27104676e-01
1.94551167e-03 9.33133841e-01 5.17144024e-01 -7.55632699e-01
-5.82675815e-01 -5.64051926e-01 -3.53095114e-01 1.53874293e-01
-2.60412931e-01 -3.00515473e-01 -1.01583481e-01 -4.08221781e-01
-1.90077070e-02 -1.00052392e+00 -1.92857206e-01 3.09964329e-01
-1.85684711e-01 -2.45558962e-01 7.80528307e-01 -7.55594730e-01
9.59724724e-01 -2.05113602e+00 3.51192623e-01 5.68489432e-01
6.58790827e-01 -1.44913539e-01 -2.07797792e-02 3.48367423e-01
-2.59712905e-01 3.44590873e-01 2.35153779e-01 -1.75864160e-01
1.00549608e-02 1.70255631e-01 -9.42908321e-03 1.11259449e+00
1.76603049e-01 1.37340355e+00 -9.24090922e-01 -9.71564949e-01
5.07483840e-01 6.83890462e-01 -2.74980426e-01 -1.97446048e-01
2.26704016e-01 2.53426582e-01 -1.19977057e-01 1.11269808e+00
2.89164215e-01 -7.35707581e-01 6.27938449e-01 -3.43684673e-01
7.64487982e-02 -6.44744396e-01 -5.04390299e-01 1.37001860e+00
-1.73955590e-01 1.06400037e+00 1.14512086e-01 -4.76759464e-01
3.85308057e-01 2.50149995e-01 8.74457896e-01 -6.88514352e-01
3.15355897e-01 3.92610691e-02 -1.11399323e-01 -6.62969351e-01
2.87673920e-01 -5.54784119e-01 1.48483783e-01 5.11573434e-01
3.36927056e-01 9.36776474e-02 2.16118768e-01 9.74174291e-02
1.23957109e+00 -2.76373357e-01 2.70359725e-01 -1.48118377e-01
5.79260349e-01 2.22097367e-01 1.52824610e-01 2.66197383e-01
-6.03798270e-01 2.83234507e-01 3.50397557e-01 -4.63185698e-01
-1.00814700e+00 -1.34507620e+00 -5.84682882e-01 8.17755580e-01
-6.85321093e-02 3.13616097e-01 -2.77114689e-01 -5.28911293e-01
2.67130733e-01 3.29530776e-01 -1.34848177e+00 2.03851983e-01
-2.44559005e-01 -8.84023428e-01 4.79173005e-01 4.36158627e-01
-3.58293086e-01 -6.82869136e-01 -7.41969764e-01 2.24283293e-01
-5.55014014e-02 -4.37791795e-01 -3.56987566e-01 3.40187907e-01
-7.75246739e-01 -1.41846347e+00 -1.12917614e+00 -8.86471272e-01
8.57900202e-01 2.33891942e-02 8.37682903e-01 -3.13016400e-02
-9.08453822e-01 6.77117586e-01 -6.49354467e-03 -1.06178880e-01
-4.74972188e-01 -1.91061124e-01 5.97904325e-02 -1.61101624e-01
3.63096029e-01 -2.31661305e-01 -9.24879611e-01 -3.05971829e-04
-1.19935322e+00 -3.18456829e-01 8.44743609e-01 1.44306052e+00
1.17538726e+00 -2.78871328e-01 -3.22960205e-02 -5.36269963e-01
5.13371944e-01 -5.48737764e-01 -4.02850658e-01 5.77541769e-01
-6.32880390e-01 -8.91564265e-02 2.82687694e-01 -2.33070821e-01
-7.67104745e-01 -3.06851715e-01 -3.24553847e-02 -7.85994470e-01
5.59890755e-02 7.54718065e-01 7.98117042e-01 -2.76966751e-01
5.56251943e-01 1.70756102e-01 4.54263091e-01 1.72752008e-01
4.19718139e-02 5.73099673e-01 7.32483089e-01 2.45157570e-01
3.94391596e-01 1.07712746e+00 3.23953420e-01 -7.28885233e-01
4.02559489e-02 -8.17476213e-01 -4.29000169e-01 -3.16924602e-01
9.56220746e-01 -5.38300335e-01 -9.22621787e-01 7.27768317e-02
-4.98614401e-01 -1.34636283e-01 -1.05713524e-01 6.74142420e-01
-3.25187743e-01 2.79094160e-01 -1.01221204e+00 -2.07899675e-01
-6.17201090e-01 -1.28989661e+00 1.28599524e+00 9.58826393e-02
-5.99485040e-01 -1.39722180e+00 7.74066448e-01 3.35792363e-01
3.33963752e-01 5.58638871e-01 1.34007180e+00 -5.66150010e-01
-3.91004950e-01 -5.72145522e-01 5.96223138e-02 -1.98060960e-01
7.03962371e-02 6.35045707e-01 -7.46099293e-01 -3.86246473e-01
-2.03408048e-01 -5.84946051e-02 8.51295352e-01 5.98487675e-01
7.55197167e-01 1.25987053e-01 -7.52355516e-01 7.19183922e-01
1.70375180e+00 2.66579330e-01 3.10751051e-01 3.94316941e-01
1.82738766e-01 6.74186945e-01 2.33519644e-01 1.34397224e-01
2.57881522e-01 4.10695046e-01 2.90649533e-01 -3.35208744e-01
3.53819095e-02 8.19378197e-02 -7.21241310e-02 6.05625391e-01
2.29553327e-01 -3.39404345e-01 -1.15295374e+00 8.01931441e-01
-1.20513821e+00 -1.06347144e+00 4.00142819e-01 1.70724416e+00
5.64596355e-01 -4.60738569e-01 -3.20100307e-01 -1.46059811e-01
7.88471699e-01 5.38148880e-02 -6.03067040e-01 -2.65355766e-01
-2.86689967e-01 4.07052338e-01 5.90201616e-01 4.18808043e-01
-9.12136853e-01 3.19966972e-01 6.90599632e+00 7.01043725e-01
-1.86648321e+00 8.29535276e-02 7.81037271e-01 -4.53198999e-01
-6.23782575e-01 -5.67034960e-01 -1.49817944e-01 1.28457189e-01
6.72726572e-01 -4.06267285e-01 2.25460172e-01 3.07388842e-01
7.09624514e-02 -1.58415169e-01 -1.12399650e+00 9.33805346e-01
-2.15739030e-02 -1.94614434e+00 -1.13696374e-01 3.63318563e-01
5.02100050e-01 3.22291672e-01 6.42687440e-01 1.44042626e-01
2.79314309e-01 -1.16463757e+00 -1.38923526e-01 6.81548059e-01
1.38828731e+00 -5.77534914e-01 1.24335504e+00 -2.84810454e-01
-7.07199335e-01 6.17729537e-02 1.51954340e-02 7.54596949e-01
-2.53817409e-01 2.91373551e-01 -1.58664286e+00 3.49753499e-01
5.07006347e-01 3.30154985e-01 -4.41992015e-01 5.86317837e-01
3.03237230e-01 1.33354500e-01 -2.38053918e-01 -2.66610742e-01
2.90751934e-01 2.81044066e-01 4.92206484e-01 1.23822045e+00
3.36204469e-01 -7.51561895e-02 -2.96512455e-01 5.72926819e-01
1.30772278e-01 3.37821633e-01 -2.14782998e-01 -3.74431700e-01
3.19593519e-01 1.61881888e+00 -1.04434788e+00 -2.56600887e-01
-1.33731335e-01 7.71718323e-01 1.50069722e-03 4.10957456e-01
-6.07167423e-01 -1.53780878e-01 6.30259931e-01 -1.25974372e-01
1.05489567e-01 3.23656172e-01 -1.60092618e-02 -1.05338788e+00
-4.76888567e-01 -5.46880484e-01 7.58480728e-01 -7.99244642e-01
-1.30417812e+00 5.19099951e-01 -3.53890508e-01 -1.37613153e+00
-5.29510260e-01 -8.57134342e-01 -4.98433143e-01 8.43740880e-01
-1.39283407e+00 -1.14824915e+00 -1.70278065e-02 5.29781878e-01
6.08057939e-02 -1.43678010e-01 1.17792809e+00 -8.02393723e-03
-4.85907763e-01 6.63077950e-01 5.85316837e-01 1.40141007e-02
9.22901869e-01 -1.48002267e+00 -4.57199872e-01 3.93976569e-01
-3.76591422e-02 8.73877168e-01 5.15417337e-01 -5.24721920e-01
-1.43434370e+00 -9.25013244e-01 7.73861110e-01 -5.35210431e-01
8.74568880e-01 3.94376427e-01 -5.42350769e-01 5.92735648e-01
3.43386024e-01 2.67350152e-02 1.63338566e+00 -2.23362133e-01
-9.06763226e-02 3.23198527e-01 -1.47024870e+00 7.26742566e-01
1.35720789e-01 -9.15295482e-01 -4.38144982e-01 2.68232375e-01
9.54397023e-02 -4.77183729e-01 -1.45232439e+00 1.84239030e-01
1.06255066e+00 -6.08046591e-01 7.45033562e-01 -5.43606102e-01
5.34799218e-01 -2.52976894e-01 -2.44799569e-01 -1.20583534e+00
-6.78572953e-01 -5.79552278e-02 1.13973126e-01 3.86147618e-01
2.80440420e-01 -6.27054393e-01 9.87847686e-01 4.96530890e-01
2.70037055e-01 -1.19453347e+00 -1.32053936e+00 -4.74979132e-02
1.08883202e-01 -9.84029379e-04 2.96465665e-01 1.07822776e+00
2.61981249e-01 -3.92915070e-01 3.45352769e-01 8.29721093e-02
5.36201596e-01 2.48355895e-01 1.48294806e-01 -9.01699305e-01
-2.41057411e-01 -8.74151886e-01 -8.88546646e-01 1.22180670e-01
2.14360684e-01 -1.40494633e+00 -4.31973606e-01 -1.41498888e+00
3.88556689e-01 2.11839974e-01 -9.23762560e-01 5.21868408e-01
-7.16418028e-02 6.22131348e-01 -6.41965419e-02 3.53801638e-01
-2.45970041e-01 5.82276955e-02 1.13268459e+00 -6.16486609e-01
-3.37253399e-02 -5.98221838e-01 -7.42593944e-01 3.24961185e-01
3.82513374e-01 -3.33936006e-01 -2.75712982e-02 -8.59029964e-02
3.13686915e-02 4.64467764e-01 5.07537186e-01 -7.07449079e-01
5.46823561e-01 -1.81866974e-01 9.92681503e-01 -7.19146788e-01
1.58408210e-01 -6.50544524e-01 4.03196037e-01 7.33767271e-01
-3.48215669e-01 1.28311351e-01 1.38537556e-01 5.68107963e-01
-4.83086735e-01 1.75724626e-01 6.45603657e-01 2.42851079e-02
-8.49235475e-01 4.82339025e-01 -7.38808393e-01 -7.47786999e-01
1.14556623e+00 -4.38591331e-01 -5.20837247e-01 -1.54049858e-01
-1.05074549e+00 4.19962913e-01 6.79979563e-01 1.05759695e-01
6.52557850e-01 -1.26749933e+00 -6.95917547e-01 1.89497709e-01
2.68836617e-01 -4.95136052e-01 9.32549477e-01 1.15453529e+00
-8.95247281e-01 6.38515651e-01 -4.34152365e-01 -9.04289305e-01
-1.51283908e+00 4.32817787e-01 5.46884358e-01 -7.97639608e-01
-1.89029232e-01 1.18779552e+00 6.56262320e-03 -3.33461344e-01
-1.77377626e-01 4.01807278e-02 -2.86991328e-01 5.44871449e-01
4.57551867e-01 2.15755641e-01 1.72617108e-01 -7.18294740e-01
-5.57682276e-01 3.40425432e-01 -3.03671837e-01 -3.91838193e-01
1.24329066e+00 7.59914611e-03 -2.44442493e-01 3.16924363e-01
1.74320567e+00 6.24733865e-02 -7.37845302e-01 2.75528654e-02
-2.74645478e-01 -1.35178596e-01 5.70357442e-01 -8.32122147e-01
-1.06762111e+00 6.18851900e-01 1.12917662e+00 -1.03743248e-01
1.01041627e+00 -4.62402739e-02 6.45893455e-01 2.58982033e-01
1.36471927e-01 -8.40766907e-01 -3.07279974e-01 -4.74411286e-02
4.99391913e-01 -1.07089710e+00 1.71375528e-01 2.16703892e-01
-5.36798537e-01 1.53962946e+00 5.32187298e-02 4.30070236e-02
4.35057074e-01 3.89163554e-01 6.46595597e-01 -4.95740026e-01
-1.08615482e+00 1.01623766e-01 1.86154887e-01 4.94862586e-01
4.35704172e-01 2.23689467e-01 -1.49980664e-01 8.58308747e-02
-9.41802338e-02 1.63563311e-01 2.93679595e-01 9.47278976e-01
1.03928037e-02 -1.07683098e+00 -2.75613338e-01 9.09776449e-01
-6.75291955e-01 1.71594933e-01 -3.08775991e-01 1.05356419e+00
3.65597457e-02 2.82312214e-01 3.48468781e-01 -4.06621218e-01
-1.75567836e-01 3.16642851e-01 5.23051023e-01 -1.46493360e-01
-4.40299302e-01 2.74359941e-01 -4.56141144e-01 -5.15886188e-01
-4.05260295e-01 -9.46189821e-01 -9.55667913e-01 -1.85869232e-01
-2.72007376e-01 9.91852861e-03 6.53509736e-01 7.81490624e-01
2.35634416e-01 5.84097147e-01 5.51682770e-01 -6.37515485e-01
-3.58259380e-01 -6.02548778e-01 -8.03760648e-01 2.71437585e-01
8.09408247e-01 -5.92641056e-01 -4.41705972e-01 1.22778654e-01] | [15.090066909790039, -2.9848921298980713] |
529f8820-75b9-4340-81a7-d59db2b7733d | scalable-lossless-coding-of-dynamic-medical | 2302.01589 | null | https://arxiv.org/abs/2302.01589v1 | https://arxiv.org/pdf/2302.01589v1.pdf | Scalable Lossless Coding of Dynamic Medical CT Data Using Motion Compensated Wavelet Lifting with Denoised Prediction and Update | Professional applications like telemedicine often require scalable lossless coding of sensitive data. 3-D subband coding has turned out to offer good compression results for dynamic CT data and additionally provides a scalable representation in terms of low- and highpass subbands. To improve the visual quality of the lowpass subband, motion compensation can be incorporated into the lifting structure, but leads to inferior compression results at the same time. Prior work has shown that a denoising filter in the update step can improve the compression ratio. In this paper, we present a new processing order of motion compensation and denoising in the update step and additionally introduce a second denoising filter in the prediction step. This allows for reducing the overall file size by up to 4.4%, while the visual quality of the lowpass subband is kept nearly constant. | ['André Kaup', 'Franz Schilling', 'Daniela Lanz'] | 2023-02-03 | null | null | null | null | ['motion-compensation'] | ['computer-vision'] | [ 5.86275578e-01 1.55824333e-01 -1.81521982e-01 -1.61985293e-01
-8.47008705e-01 -9.76481810e-02 8.55832621e-02 5.81001401e-01
-4.57294345e-01 5.16644955e-01 6.84251904e-01 -2.08537936e-01
-7.60331899e-02 -7.89893866e-01 -3.70278925e-01 -7.75496483e-01
-1.82204664e-01 -2.56713182e-01 7.53140211e-01 2.22057793e-02
8.21803883e-03 3.37482184e-01 -1.31489336e+00 8.60961735e-01
7.10522711e-01 1.22287321e+00 4.17680025e-01 8.38870823e-01
1.58754036e-01 9.92728770e-01 -3.73537421e-01 -2.39703164e-01
3.57890487e-01 -5.11445642e-01 -7.98687160e-01 1.72347963e-01
3.13484281e-01 -1.08215702e+00 -6.72388554e-01 7.38654017e-01
7.58143306e-01 4.13028523e-02 4.21738140e-02 -6.60644889e-01
1.84064373e-01 3.25445533e-01 -6.06986165e-01 1.95237249e-01
4.13055062e-01 -9.95635763e-02 6.16964698e-01 -3.98986757e-01
8.01695287e-01 9.18545187e-01 8.65205884e-01 2.73570657e-01
-1.40541768e+00 -2.40851939e-01 -6.35823965e-01 4.65115637e-01
-1.06493425e+00 -4.50619727e-01 8.79860699e-01 4.66565490e-02
7.90566206e-01 4.11110252e-01 9.35096323e-01 2.41445854e-01
7.27425456e-01 3.66151094e-01 1.02353883e+00 -4.59121138e-01
1.26444176e-01 -4.22170550e-01 -2.31283411e-01 6.06200218e-01
1.56180337e-01 9.26853120e-02 -5.76042414e-01 -1.21788509e-01
9.39836144e-01 6.78616539e-02 -6.01889133e-01 -2.18448609e-01
-9.75982487e-01 7.07776368e-01 4.53058779e-01 4.92074579e-01
-5.92457414e-01 5.19486964e-01 8.08615208e-01 3.50433499e-01
5.26321948e-01 -1.88642770e-01 -1.03749350e-01 -1.46813750e-01
-1.29510665e+00 1.23281397e-01 6.21543348e-01 5.94013870e-01
4.64222342e-01 -4.12460156e-02 -3.99776012e-01 9.12837565e-01
5.89824282e-02 2.42667366e-02 4.42402422e-01 -1.48760557e+00
5.36240578e-01 1.72458783e-01 -1.34907857e-01 -9.71983492e-01
-3.13234240e-01 -4.89904106e-01 -9.87041652e-01 4.11942005e-01
2.59459257e-01 1.87775835e-01 -9.02197003e-01 1.38737798e+00
3.34532589e-01 4.38559689e-02 -4.94564511e-02 8.82787704e-01
6.71465397e-01 6.77392781e-01 4.81243953e-02 -6.33002818e-01
1.39777339e+00 -6.35054708e-01 -1.18116760e+00 2.41785143e-02
6.51071608e-01 -1.19145119e+00 7.74309456e-01 5.56400836e-01
-1.70898604e+00 -4.79594380e-01 -1.17018437e+00 -4.61202711e-01
4.49562669e-01 -1.86920509e-01 2.24367157e-01 6.17197096e-01
-1.00591946e+00 8.62004519e-01 -1.07065785e+00 7.13830162e-03
4.45419043e-01 3.11940670e-01 -3.99768621e-01 -6.50199592e-01
-9.74206150e-01 8.76527429e-01 1.74701288e-01 -4.82540160e-01
-4.12202448e-01 -9.54085767e-01 -7.74999082e-01 3.18720698e-01
7.42988959e-02 -5.83280861e-01 1.06692815e+00 -5.85219622e-01
-1.24057984e+00 6.86325788e-01 -1.87323987e-01 -7.09829867e-01
6.53260469e-01 1.28981154e-02 -2.60854542e-01 9.85627174e-01
-2.68607795e-01 3.89897346e-01 1.00226998e+00 -8.86919975e-01
-4.13977325e-01 -1.93334058e-01 2.01922525e-02 3.15858364e-01
-4.02269125e-01 -3.43559951e-01 -3.70174974e-01 -1.20410299e+00
7.05420136e-01 -6.84130073e-01 -2.52504230e-01 6.53480411e-01
2.70227134e-01 7.34934270e-01 8.35998535e-01 -1.17798901e+00
1.46506596e+00 -2.47127914e+00 -1.56088263e-01 2.39530027e-01
1.84568256e-01 -1.08923651e-02 1.13992125e-01 4.14120585e-01
-2.96278223e-02 -2.51540154e-01 -3.78888011e-01 -2.38044038e-01
-6.84223294e-01 6.05972111e-01 -1.47758638e-02 6.83384001e-01
-3.79345745e-01 4.72736061e-01 -5.46449482e-01 -6.19437754e-01
3.46019298e-01 9.37229991e-01 -1.05878532e+00 -1.38421878e-01
4.91265446e-01 3.60062391e-01 2.37358790e-02 4.27708745e-01
1.00596058e+00 -2.68116985e-02 2.50474006e-01 -6.88358366e-01
1.05952799e-01 4.48279917e-01 -1.11575508e+00 1.89805758e+00
-5.13286173e-01 5.17113626e-01 5.46656251e-01 -8.89166653e-01
5.34699261e-01 6.79518521e-01 1.04425657e+00 -8.33663106e-01
4.31812853e-02 3.56709391e-01 -1.97935849e-02 -5.17859459e-01
6.76411033e-01 -8.74367297e-01 3.21572602e-01 8.57619345e-02
-2.43481398e-01 -4.64020133e-01 5.10947146e-02 7.17729777e-02
1.18529856e+00 -2.66315430e-01 2.58507609e-01 -1.79402664e-01
5.28742731e-01 -1.18303910e-01 5.11269391e-01 3.44123840e-01
-1.42872453e-01 1.01147211e+00 9.47476849e-02 -4.47985470e-01
-1.31210589e+00 -7.89924741e-01 -3.35069448e-01 5.99388003e-01
2.20869884e-01 -6.65899873e-01 -4.41773564e-01 -6.47683442e-02
-6.27943128e-02 4.28427815e-01 -1.10413082e-01 -3.24740082e-01
-8.13304842e-01 -5.68038583e-01 2.18480632e-01 4.47329253e-01
6.44300878e-01 -5.19562900e-01 -1.05879056e+00 7.29953945e-01
-7.57546246e-01 -1.14168918e+00 -3.91499549e-01 3.09953749e-01
-1.65388143e+00 -7.09918797e-01 -8.94899786e-01 -5.34736037e-01
4.70474601e-01 2.60746568e-01 8.57762396e-01 5.28286040e-01
-4.89814848e-01 5.71524240e-02 -6.61016643e-01 -3.54974978e-02
-5.64939678e-01 -3.28940719e-01 -3.85472000e-01 -3.92359227e-01
-3.75494152e-01 -6.78035796e-01 -9.89144087e-01 5.81122190e-02
-1.29895079e+00 -7.86654186e-03 2.49174818e-01 9.36999440e-01
7.47520387e-01 4.62327659e-01 3.48679163e-02 -5.26043892e-01
2.53495127e-01 4.59328666e-02 -9.83949602e-02 -2.88759589e-01
-5.23657441e-01 3.11798956e-02 5.61937749e-01 -2.87107170e-01
-9.73530412e-01 2.04046220e-02 -7.06286788e-01 -2.97864377e-01
3.47376972e-01 4.34899151e-01 3.69712919e-01 -3.53596747e-01
5.75539052e-01 2.44660050e-01 3.24053913e-01 -5.83924890e-01
-1.72602721e-02 3.70965004e-01 5.03068507e-01 -6.10847361e-02
4.84489202e-01 9.22194242e-01 4.18199241e-01 -7.86676586e-01
-4.12922621e-01 -4.20788199e-01 -3.82565111e-01 -2.54711926e-01
5.90792239e-01 -9.40422297e-01 -4.79062825e-01 1.87754631e-01
-7.80139506e-01 -2.12889045e-01 -7.49550343e-01 8.00163507e-01
-8.21556866e-01 7.14318395e-01 -1.13951993e+00 -4.57803339e-01
-5.31555295e-01 -1.06603289e+00 7.61637211e-01 -3.64496559e-01
-2.58512080e-01 -6.35955691e-01 -2.24454060e-01 2.71967769e-01
6.56399429e-01 3.50169241e-01 1.01292098e+00 3.73000920e-01
-4.10738856e-01 -2.43299127e-01 1.51858985e-01 4.88538027e-01
7.85893202e-02 -6.39375329e-01 -7.68373668e-01 -5.81537008e-01
5.48978209e-01 -6.87130988e-02 1.00763643e+00 7.65028775e-01
1.21012688e+00 -3.58342677e-01 -1.02715930e-02 8.54020178e-01
1.59347367e+00 1.86820909e-01 1.14870620e+00 1.60435930e-01
7.67588615e-02 3.98491830e-01 5.79931378e-01 6.26953483e-01
-1.95268139e-01 6.02413118e-01 3.12421173e-01 -2.39445955e-01
-7.48488069e-01 -1.01039864e-01 4.77742925e-02 9.35195804e-01
-1.59283094e-02 -4.05788273e-02 -4.60307866e-01 4.31158721e-01
-1.42010868e+00 -1.06726837e+00 -1.36794373e-01 2.30926108e+00
8.32164943e-01 1.40472010e-01 -1.46864176e-01 9.61804271e-01
3.57678324e-01 2.04972789e-01 -1.78665370e-01 -4.80196714e-01
7.89074525e-02 4.78180915e-01 8.85520995e-01 7.56791413e-01
-6.86497152e-01 9.77683589e-02 6.96051741e+00 9.77606177e-01
-1.13000834e+00 3.14300597e-01 5.24413049e-01 -3.82058173e-01
-2.96611637e-01 6.33869972e-03 -1.24654710e-01 4.83396769e-01
7.63115525e-01 -5.70887141e-02 1.45290285e-01 4.89067942e-01
3.84547979e-01 -6.04407310e-01 -5.36585212e-01 9.84139621e-01
-1.92481086e-01 -1.29144180e+00 -9.18225870e-02 2.54760802e-01
2.70733356e-01 -3.85337353e-01 -7.39467740e-02 -3.92202079e-01
-5.15582860e-01 -5.98272264e-01 6.41975343e-01 2.65502989e-01
1.11165917e+00 -1.05953753e+00 7.31092632e-01 2.28614718e-01
-1.26255035e+00 -1.74106881e-01 -3.65564138e-01 -4.57493439e-02
5.35117984e-01 1.03113317e+00 -4.72666502e-01 5.49326897e-01
7.92423308e-01 4.28291649e-01 2.33022776e-02 1.30052125e+00
1.14135683e-01 5.78485787e-01 -4.03841317e-01 8.43558729e-01
1.83591638e-02 2.80774385e-02 5.77341735e-01 1.01745772e+00
5.99935353e-01 6.42865062e-01 1.49931544e-02 -3.55758741e-02
6.64380267e-02 6.21750280e-02 -1.58614516e-01 4.74503040e-01
-3.05176917e-02 7.34868407e-01 -6.91800117e-01 -4.20113385e-01
-4.49428409e-01 1.27903700e+00 -3.90747249e-01 -3.04491390e-02
-5.34289896e-01 -3.74066144e-01 4.67333287e-01 8.50202918e-01
5.44575572e-01 -2.33593240e-01 -4.11097765e-01 -9.18149292e-01
1.12434685e-01 -7.35232890e-01 5.54954529e-01 -4.71776336e-01
-6.83304846e-01 1.85550153e-01 -1.74677864e-01 -1.47384441e+00
-2.12415755e-01 3.04863099e-02 -7.41776004e-02 7.97729671e-01
-1.64256263e+00 -7.00014472e-01 -3.18508863e-01 7.17690587e-01
3.98738652e-01 4.74806488e-01 7.65657067e-01 7.38470376e-01
3.39610040e-01 5.13423622e-01 2.68635511e-01 -3.72755766e-01
5.99660516e-01 -6.23536229e-01 -2.52919644e-01 7.53729284e-01
-5.34950137e-01 2.89553732e-01 9.95043695e-01 -6.97920799e-01
-1.34251821e+00 -6.17212057e-01 9.12477612e-01 4.32488829e-01
-6.30700290e-02 5.12822121e-02 -1.06694007e+00 1.67751908e-01
4.80188057e-02 1.80071056e-01 7.11953521e-01 -6.82837605e-01
-7.76549708e-03 -3.34867686e-01 -1.90472007e+00 1.50213346e-01
5.68229079e-01 -6.33382261e-01 -3.28923166e-01 1.33746803e-01
5.35144389e-01 -7.13500619e-01 -1.26823997e+00 4.13601041e-01
7.97985852e-01 -1.22679937e+00 1.40155828e+00 3.47323895e-01
7.41794169e-01 -1.57274783e-01 -2.97035098e-01 -8.15465927e-01
-5.12637854e-01 -4.89028990e-01 -3.61756861e-01 6.61615849e-01
-2.10750148e-01 -2.43069157e-01 9.37116802e-01 4.96701330e-01
-6.33350834e-02 -7.38397241e-01 -1.40870440e+00 -6.44521534e-01
-3.19883943e-01 -5.12332737e-01 -3.39321531e-02 5.32172024e-01
3.31355602e-01 -1.62798166e-01 -5.83457530e-01 -1.43959880e-01
6.49664104e-01 -3.40521596e-02 4.59467694e-02 -9.03171360e-01
-5.23000479e-01 9.48225334e-02 -5.30678213e-01 -1.28441501e+00
-6.20293140e-01 -7.24257052e-01 -6.96700141e-02 -1.71851802e+00
1.38455302e-01 -1.79786593e-01 -1.17063396e-01 2.67865360e-01
4.79058512e-02 8.86353672e-01 3.80271405e-01 2.50264376e-01
2.76356369e-01 2.62607872e-01 1.60360479e+00 -1.62321143e-02
-1.81205615e-01 -1.17305964e-01 -3.91960204e-01 5.92263818e-01
6.15385115e-01 -6.18510783e-01 -5.32122791e-01 -4.23706323e-01
-2.25015003e-02 5.01597047e-01 1.19683690e-01 -1.15359032e+00
8.80175978e-02 3.24123502e-01 5.35432935e-01 -6.61260903e-01
7.31546223e-01 -1.35893083e+00 3.55167270e-01 1.25661683e+00
-2.14539364e-01 -1.76510096e-01 2.14445770e-01 3.87407392e-01
-4.01331902e-01 -3.51952314e-01 1.29790246e+00 -1.75097570e-01
-2.53370285e-01 6.32093009e-03 -6.11243308e-01 -3.26970339e-01
7.83992171e-01 -5.60875058e-01 2.57969618e-01 -5.89983881e-01
-9.20668542e-01 -2.47857809e-01 6.29144192e-01 -2.40910530e-01
9.19476688e-01 -1.19468510e+00 -6.17973387e-01 3.20289463e-01
-3.63458306e-01 -2.56479472e-01 6.85845554e-01 1.13401854e+00
-1.04039764e+00 4.86978367e-02 -4.52594459e-01 -5.21552801e-01
-1.69809425e+00 4.03990299e-01 1.53125122e-01 -5.91856956e-01
-1.19324577e+00 5.94249308e-01 -3.64330858e-01 5.70601583e-01
2.51567036e-01 -3.97611111e-01 1.03016637e-01 1.26285940e-01
8.47993135e-01 6.21406555e-01 3.93493652e-01 -6.70426071e-01
-2.31314600e-01 5.91005266e-01 1.51400715e-02 -2.28969991e-01
1.43267763e+00 -4.24890310e-01 -4.74324822e-02 -2.90276669e-02
1.57975149e+00 -4.29830886e-02 -1.11577356e+00 -2.85903998e-02
-4.77786064e-01 -9.61076200e-01 5.72700441e-01 -4.38885808e-01
-1.25060022e+00 7.85989404e-01 9.21680868e-01 1.25525191e-01
1.92080760e+00 -2.98242182e-01 1.44426322e+00 -3.11775386e-01
3.97655040e-01 -1.06329799e+00 1.24796309e-01 4.19962592e-02
7.78635025e-01 -7.01482713e-01 6.67723060e-01 -7.81527996e-01
-1.99710250e-01 1.25722504e+00 -4.06804204e-01 -5.32764047e-02
6.90918803e-01 7.07544148e-01 -2.28014812e-01 4.08361405e-02
-4.55031604e-01 9.85032022e-02 7.84892440e-02 3.32861036e-01
6.57788932e-01 -2.13047147e-01 -1.05102205e+00 -8.88102129e-02
-1.06754802e-01 3.11216682e-01 5.29860795e-01 1.32044184e+00
-6.33654356e-01 -1.21207547e+00 -7.04947114e-01 6.69258237e-01
-1.11215472e+00 -1.62895367e-01 5.36794305e-01 3.24608415e-01
1.41945332e-01 9.53329384e-01 1.21597640e-01 -1.60383180e-01
4.10574883e-01 -1.32636070e-01 6.58065796e-01 -1.87699527e-01
-8.10032487e-01 7.12461710e-01 7.38374963e-02 -9.77560222e-01
-5.37465036e-01 -3.13662082e-01 -1.30439746e+00 -5.70528090e-01
-1.27047524e-01 -1.48983479e-01 5.49712777e-01 3.69003654e-01
5.19731157e-02 6.53108239e-01 4.35931206e-01 -6.94104493e-01
-3.18808436e-01 -5.79152584e-01 -8.16556275e-01 5.94830215e-01
7.37084031e-01 -2.01681525e-01 -1.99905992e-01 3.77276093e-01] | [11.498762130737305, -2.30005145072937] |
fcfa82f8-c0d3-43bd-8834-a75118498063 | building-hvac-scheduling-using-reinforcement | 1910.05313 | null | https://arxiv.org/abs/1910.05313v2 | https://arxiv.org/pdf/1910.05313v2.pdf | Building HVAC Scheduling Using Reinforcement Learning via Neural Network Based Model Approximation | Buildings sector is one of the major consumers of energy in the United States. The buildings HVAC (Heating, Ventilation, and Air Conditioning) systems, whose functionality is to maintain thermal comfort and indoor air quality (IAQ), account for almost half of the energy consumed by the buildings. Thus, intelligent scheduling of the building HVAC system has the potential for tremendous energy and cost savings while ensuring that the control objectives (thermal comfort, air quality) are satisfied. Recently, several works have focused on model-free deep reinforcement learning based techniques such as Deep Q-Network (DQN). Such methods require extensive interactions with the environment. Thus, they are impractical to implement in real systems due to low sample efficiency. Safety-aware exploration is another challenge in real systems since certain actions at particular states may result in catastrophic outcomes. To address these issues and challenges, we propose a model-based reinforcement learning approach that learns the system dynamics using a neural network. Then, we adopt Model Predictive Control (MPC) using the learned system dynamics to perform control with random-sampling shooting method. To ensure safe exploration, we limit the actions within safe range and the maximum absolute change of actions according to prior knowledge. We evaluate our ideas through simulation using widely adopted EnergyPlus tool on a case study consisting of a two zone data-center. Experiments show that the average deviation of the trajectories sampled from the learned dynamics and the ground truth is below $20\%$. Compared with baseline approaches, we reduce the total energy consumption by $17.1\% \sim 21.8\%$. Compared with model-free reinforcement learning approach, we reduce the required number of training steps to converge by 10x. | ['Viktor K. Prasanna', 'Rajgopal Kannan', 'Sanmukh R. Kuppannagari', 'Chi Zhang'] | 2019-10-11 | null | null | null | null | ['safe-exploration'] | ['robots'] | [ 1.11372687e-01 2.94317663e-01 1.78745482e-02 1.61047876e-01
-7.53218055e-01 -6.24879360e-01 3.51732224e-01 2.35604122e-01
-7.13606104e-02 1.20401406e+00 -2.12657899e-01 -4.71101433e-01
-4.58544642e-01 -1.08628607e+00 -7.76851058e-01 -9.76402223e-01
-1.46838501e-01 9.25861150e-02 -1.88562393e-01 -1.83076903e-01
-1.21633068e-01 1.75770372e-01 -1.37967145e+00 -4.64518309e-01
9.56911981e-01 1.06824577e+00 1.67965904e-01 6.59316182e-01
5.34500182e-01 5.52733779e-01 -3.16890121e-01 6.49457037e-01
4.33240801e-01 -5.84382534e-01 -6.02752984e-01 -1.22493058e-01
-4.24757659e-01 -5.62587500e-01 -2.54873000e-02 9.80378449e-01
5.97281277e-01 7.49970615e-01 4.94074017e-01 -1.44812429e+00
6.93860725e-02 9.26049575e-02 -2.46555433e-01 -2.82128211e-02
-6.23259880e-03 6.34260535e-01 5.25509298e-01 2.79291067e-02
-2.15418451e-02 7.11768389e-01 3.80375594e-01 5.16621470e-01
-1.12741923e+00 -5.35294950e-01 9.75356922e-02 6.20911308e-02
-1.35961354e+00 -7.49416277e-02 6.58262849e-01 -2.28264987e-01
1.28299582e+00 5.78438640e-01 1.17901623e+00 4.96122420e-01
4.00162756e-01 3.63298267e-01 1.27365506e+00 -3.19429129e-01
8.93535912e-01 1.59003675e-01 -3.55845928e-01 5.97934306e-01
1.42676188e-02 6.57516122e-01 4.67404515e-01 -1.17574446e-01
4.95671004e-01 1.41643077e-01 -2.21590430e-01 -2.52930850e-01
-5.90972722e-01 7.37027466e-01 6.56259060e-01 9.98670608e-02
-4.14951146e-01 5.12344062e-01 2.84276456e-01 4.73976135e-02
-2.43113413e-02 5.73164642e-01 -5.60974181e-01 -2.04185218e-01
-7.43318796e-01 4.07751828e-01 5.16946435e-01 8.09232593e-01
5.17077982e-01 2.69661546e-01 -1.47935525e-01 4.62947935e-01
4.61856484e-01 5.29423535e-01 2.60811806e-01 -1.16657257e+00
1.34365514e-01 7.23061323e-01 6.24102890e-01 -4.31174368e-01
-3.47252220e-01 -1.94815218e-01 -9.25716341e-01 4.35505569e-01
-8.63148831e-03 -6.90884113e-01 -9.80067253e-01 1.65074372e+00
6.66857421e-01 -1.47669753e-02 -9.09324288e-02 7.53860593e-01
-2.68930554e-01 1.23495948e+00 1.80199206e-01 -6.26329541e-01
1.09093678e+00 -7.54098713e-01 -6.74228549e-01 1.31381089e-02
2.19829619e-01 -3.34423900e-01 9.22454417e-01 1.89021319e-01
-1.26522565e+00 -5.11369348e-01 -1.30635965e+00 6.74346924e-01
-5.71799040e-01 -2.45665148e-01 -6.70629442e-02 7.45252013e-01
-9.05534208e-01 6.66818917e-01 -1.06275213e+00 -2.72358865e-01
1.11910537e-01 5.34425974e-01 5.36757112e-01 3.54170203e-01
-1.25113440e+00 9.13565099e-01 5.23478329e-01 1.62577644e-01
-1.32465959e+00 -9.02194977e-01 -5.50116837e-01 2.26789474e-01
5.75752139e-01 -5.42583525e-01 1.51098347e+00 -3.88069540e-01
-1.79596567e+00 -2.30278611e-01 3.52896929e-01 -4.78559852e-01
4.52192038e-01 -2.48104975e-01 -2.83239573e-01 -6.54218346e-02
-4.19880122e-01 3.84441257e-01 4.79112446e-01 -1.30824888e+00
-8.89764309e-01 1.63585469e-02 1.41399488e-01 2.31789947e-01
-2.65460253e-01 -6.34228706e-01 3.28068465e-01 -1.54759884e-01
-3.83434266e-01 -1.28149593e+00 -6.69339120e-01 -3.06962490e-01
-3.03371817e-01 -2.53181428e-01 1.02140367e+00 -7.00871408e-01
1.33189428e+00 -1.74986839e+00 -2.02183481e-02 5.59916317e-01
-4.82721388e-01 1.85845196e-01 5.89276373e-01 6.73659325e-01
2.58336775e-02 1.13496751e-01 -5.38812697e-01 5.63524216e-02
1.33803546e-01 4.14308041e-01 -1.32615507e-01 3.20347697e-01
1.04711642e-02 6.13898456e-01 -8.31942856e-01 -1.87083915e-01
7.95609534e-01 4.22903746e-01 -4.02433962e-01 4.82023239e-01
-5.74771047e-01 3.99718702e-01 -5.91208816e-01 2.70760894e-01
4.84249294e-01 8.16945359e-02 4.25709426e-01 -7.49579296e-02
-3.43632042e-01 -1.20388322e-01 -1.38721645e+00 1.35608578e+00
-9.84703422e-01 1.38944328e-01 3.66979569e-01 -1.13437879e+00
5.58451176e-01 5.39133787e-01 6.30168796e-01 -8.61960590e-01
2.66573012e-01 4.16760780e-02 -2.31181532e-01 -4.08722460e-01
1.85619935e-01 -3.95551950e-01 1.01223891e-03 3.28231335e-01
-3.28723401e-01 -5.82928598e-01 -2.00567633e-01 -2.79074401e-01
1.15133476e+00 1.98268011e-01 4.12119359e-01 -5.62022746e-01
7.37027287e-01 1.07104540e-01 5.79732597e-01 3.11553366e-02
-3.33726138e-01 -2.48942763e-01 7.38701075e-02 -1.73128277e-01
-1.19911027e+00 -9.94107723e-01 1.33625558e-02 6.17888808e-01
1.76012754e-01 8.22601616e-02 -9.98618305e-01 -3.53623450e-01
-1.50402009e-01 1.31341326e+00 -4.04034168e-01 -4.67134327e-01
-7.71088421e-01 -6.88265741e-01 -1.54153839e-01 5.27211845e-01
6.79412603e-01 -8.21804047e-01 -1.53670990e+00 3.30129772e-01
1.85497120e-01 -7.06056714e-01 -1.54015705e-01 5.16165376e-01
-7.59799421e-01 -1.09339774e+00 -2.98911452e-01 -4.60489631e-01
7.41845667e-01 -2.56403536e-01 7.46944547e-01 5.65768853e-02
-4.89809781e-01 4.86858785e-01 1.56211615e-01 -2.96828508e-01
-4.82206851e-01 -8.44046697e-02 7.37110525e-02 -5.76252699e-01
-1.96122542e-01 -3.64783823e-01 -1.12874877e+00 2.99558431e-01
-8.32255781e-01 1.13112926e-01 2.66744643e-01 7.50823736e-01
7.09363103e-01 1.00904894e+00 6.89491749e-01 -3.83868486e-01
6.04256868e-01 -4.23474550e-01 -1.21327233e+00 2.67855316e-01
-9.68040586e-01 1.68278754e-01 1.14721048e+00 -1.73732519e-01
-1.15824783e+00 2.80548602e-01 1.51608005e-01 -2.24592701e-01
-1.73142016e-01 -9.63152200e-02 -1.80127442e-01 1.84709892e-01
4.03022170e-02 2.88436264e-01 -1.69312254e-01 -1.57287598e-01
9.74170938e-02 2.56371230e-01 1.60607561e-01 -6.59010887e-01
1.07248437e+00 -6.33430842e-04 3.55486065e-01 -4.13750559e-01
-3.77701998e-01 -9.11715478e-02 -1.29140332e-01 -4.98058230e-01
9.34821248e-01 -7.82223880e-01 -1.25908959e+00 -2.01081410e-02
-4.80516076e-01 -7.05092192e-01 -4.48837072e-01 1.41071603e-01
-8.67712498e-01 -7.70388693e-02 -1.76197946e-01 -1.54686141e+00
-6.72546625e-01 -1.09568858e+00 5.68679571e-01 6.00556374e-01
-2.03976527e-01 -8.74386013e-01 4.13240016e-01 3.83802280e-02
6.15158141e-01 9.87550259e-01 1.14015567e+00 1.25952482e-01
-6.49542749e-01 3.66962254e-02 3.51307750e-01 5.58741808e-01
2.75605500e-01 -4.51826602e-02 -8.34528208e-01 -7.59701312e-01
1.50520310e-01 -3.29961240e-01 1.51752219e-01 5.09476840e-01
1.26515377e+00 -7.40210474e-01 -5.40456593e-01 6.77601025e-02
2.15991068e+00 1.08379626e+00 2.79081911e-01 2.97956795e-01
2.68543333e-01 4.40950692e-01 8.99765074e-01 8.38570356e-01
6.81072325e-02 4.65868860e-01 8.91846061e-01 -1.15326298e-02
5.05227864e-01 -3.59524965e-01 2.18686029e-01 4.06621069e-01
5.71133159e-02 -4.64945957e-02 -8.42200875e-01 6.56393468e-01
-1.70289850e+00 -7.15107620e-01 3.28090072e-01 2.47105336e+00
7.91120529e-01 2.23755971e-01 1.16496429e-01 3.84472162e-01
4.03512627e-01 -1.17103346e-01 -8.18194568e-01 -7.23929822e-01
7.45051920e-01 3.80766451e-01 7.56973386e-01 4.56807375e-01
-8.54071379e-01 1.91134229e-01 5.54183197e+00 5.57030618e-01
-8.77587736e-01 -4.13619950e-02 8.85664761e-01 -4.28085119e-01
2.09450740e-02 5.70542030e-02 -3.61868411e-01 6.58109963e-01
1.71791494e+00 -4.46559519e-01 9.55157995e-01 9.71782982e-01
8.52400184e-01 -5.55096030e-01 -1.15995455e+00 5.06577611e-01
-8.41868281e-01 -1.04618597e+00 -7.51725435e-01 1.08383335e-01
1.05141747e+00 -3.44900280e-01 -9.22071710e-02 5.96613348e-01
5.47743320e-01 -9.82439756e-01 5.09048641e-01 3.73777896e-01
4.67273295e-01 -1.43472886e+00 6.81148171e-01 6.14498436e-01
-1.35261559e+00 -5.65939128e-01 -1.00376472e-01 4.31023538e-03
4.41763073e-01 2.56603420e-01 -8.61633837e-01 5.31117201e-01
8.26387167e-01 -5.08555099e-02 -2.96924710e-02 7.21398890e-01
6.30559698e-02 6.37532711e-01 -5.96293986e-01 -4.70259815e-01
3.70789230e-01 -4.14066702e-01 1.02197796e-01 6.35523319e-01
3.79920095e-01 4.01266307e-01 3.39472979e-01 9.88268673e-01
3.09457481e-01 -2.43701354e-01 -5.45908570e-01 1.51362270e-01
4.62314516e-01 1.24588609e+00 -5.55300534e-01 -3.36359978e-01
-5.09613426e-03 5.82396150e-01 -3.60303074e-01 2.79701531e-01
-1.30180573e+00 -3.52978110e-01 5.74888468e-01 1.57825798e-01
1.07282288e-01 -1.58244476e-01 -6.09057546e-02 -4.71903272e-02
-1.60058171e-01 -5.63682973e-01 1.10676743e-01 -4.65782464e-01
-7.38914073e-01 1.54307991e-01 3.62174928e-01 -1.04581428e+00
-5.63136697e-01 -2.98066348e-01 -6.84783995e-01 7.94827998e-01
-1.35681081e+00 -4.83698070e-01 -4.40753341e-01 4.26714033e-01
7.17999041e-01 2.43029028e-01 8.04061294e-01 2.14953303e-01
-9.25355971e-01 1.06671251e-01 5.42427421e-01 -5.05980849e-01
-6.37650266e-02 -1.37304759e+00 -8.04528743e-02 6.65901184e-01
-9.40043151e-01 1.95564598e-01 1.01195276e+00 -3.83183032e-01
-1.66078568e+00 -1.28878319e+00 -8.99941549e-02 3.98170799e-02
4.02735889e-01 -2.55991697e-01 -6.05315387e-01 2.51259834e-01
7.70834148e-01 -1.25889033e-01 2.53876716e-01 -6.70876622e-01
7.11472690e-01 -3.80126297e-01 -1.61696243e+00 5.06730497e-01
4.63606685e-01 -1.88496247e-01 -4.88164611e-02 2.39592180e-01
7.62209177e-01 -1.61178097e-01 -1.03388286e+00 4.60671961e-01
3.19355279e-01 -6.88090026e-01 6.93182826e-01 -2.29216158e-01
2.80232280e-02 -4.80725110e-01 -2.95917958e-01 -1.49875700e+00
-7.76038542e-02 -7.01570809e-01 -6.51284099e-01 1.28275430e+00
1.82435378e-01 -4.55855221e-01 7.31682539e-01 9.86537039e-01
-1.90729275e-01 -1.14579463e+00 -1.19288421e+00 -7.83911467e-01
9.19764414e-02 -5.93580045e-02 7.86960065e-01 5.95490992e-01
-6.89112097e-02 -7.79924095e-02 -7.31200278e-02 3.84396493e-01
7.22819388e-01 -3.95310931e-02 2.91776717e-01 -4.28998590e-01
-3.72667402e-01 -1.32279709e-01 3.07566971e-01 -2.58219182e-01
-2.22977474e-01 -1.47029772e-01 3.89076471e-01 -1.80498326e+00
-3.59008871e-02 -3.27613562e-01 -6.53005064e-01 3.59725237e-01
3.26252058e-02 -6.28970861e-01 -3.27647291e-02 -5.13895035e-01
-1.46705732e-01 1.06231403e+00 1.03226316e+00 -1.88529462e-01
-4.58702117e-01 2.50359982e-01 -9.45088938e-02 3.55740726e-01
1.18772900e+00 -3.13645422e-01 -1.02205420e+00 1.50823608e-01
4.08809222e-02 4.10066724e-01 2.92154491e-01 -1.38315594e+00
-9.60716605e-02 -7.49445260e-01 3.05490673e-01 -6.92781091e-01
2.37776011e-01 -1.51652730e+00 7.14135408e-01 1.25641572e+00
-1.90008461e-01 2.67654598e-01 4.26882505e-01 7.65673816e-01
6.83990121e-02 1.28926679e-01 1.12354767e+00 -3.03910285e-01
-5.05979657e-01 1.91236492e-02 -5.71712434e-01 -3.31402451e-01
1.71983469e+00 -6.26010373e-02 -1.00010946e-01 -2.15031967e-01
-3.05950165e-01 7.09493935e-01 2.79854387e-01 2.49524236e-01
2.27073655e-01 -1.29242587e+00 -5.06629162e-02 -9.84759703e-02
-3.93317610e-01 1.76864207e-01 4.46544349e-01 3.66445035e-01
-4.59964067e-01 5.56302786e-01 -1.79469854e-01 -3.49998593e-01
-8.75244856e-01 8.39598656e-01 9.23403680e-01 -4.07424629e-01
-3.87061000e-01 2.82817245e-01 -1.10446274e-01 -2.93950409e-01
4.48840708e-02 -7.67385483e-01 2.93018609e-01 -2.72324264e-01
7.96188042e-02 8.22838128e-01 2.82409675e-02 1.75661117e-01
-5.05876422e-01 3.42845649e-01 5.25263608e-01 -1.57052562e-01
1.41004586e+00 1.08054103e-02 2.68407971e-01 2.36677289e-01
1.21151817e+00 -5.55877626e-01 -1.67481852e+00 4.45177794e-01
-1.45776868e-01 -1.14661142e-01 4.15120929e-01 -1.10691822e+00
-9.61784899e-01 4.81734157e-01 1.21557474e+00 5.31230688e-01
1.74807477e+00 -6.74567163e-01 7.79759884e-01 1.15495808e-01
3.97946566e-01 -1.73593628e+00 -5.27432151e-02 -3.63805853e-02
6.06767058e-01 -9.93879855e-01 4.21411544e-02 2.63604373e-01
-3.04227531e-01 7.77657390e-01 1.04681683e+00 -3.49064201e-01
8.13947618e-01 1.95443466e-01 -4.57169503e-01 2.15246052e-01
-9.34683800e-01 3.17255884e-01 -1.58197135e-01 2.58996308e-01
1.13449492e-01 3.60789418e-01 -1.38733268e-01 1.83837652e-01
1.67381912e-01 1.50266672e-02 1.88586697e-01 1.36798906e+00
-6.02751136e-01 -7.98709571e-01 -5.46091914e-01 2.67139763e-01
-1.11578323e-01 4.20816272e-01 4.31049645e-01 9.86711323e-01
2.47054592e-01 9.66559112e-01 -1.47825539e-01 -1.45871744e-01
6.05733693e-01 1.35977656e-01 2.47844085e-01 -1.55483991e-01
-6.61392331e-01 9.15257111e-02 3.71657610e-02 -6.01088166e-01
-1.11696169e-01 -4.42007482e-01 -1.60451305e+00 -4.03545201e-01
-2.39254028e-01 2.34056950e-01 7.96261907e-01 9.23957288e-01
2.51042366e-01 1.15765762e+00 1.07820737e+00 -4.40110087e-01
-8.24496865e-01 -7.27745295e-01 -5.55629849e-01 -9.49722826e-02
5.36720693e-01 -7.71007657e-01 -3.08739275e-01 -2.98066318e-01] | [5.531165599822998, 2.4461286067962646] |
a10ca705-f6c4-4445-8f88-eab0a6583e16 | convformer-parameter-reduction-in-transformer | 2304.02147 | null | https://arxiv.org/abs/2304.02147v1 | https://arxiv.org/pdf/2304.02147v1.pdf | ConvFormer: Parameter Reduction in Transformer Models for 3D Human Pose Estimation by Leveraging Dynamic Multi-Headed Convolutional Attention | Recently, fully-transformer architectures have replaced the defacto convolutional architecture for the 3D human pose estimation task. In this paper we propose \textbf{\textit{ConvFormer}}, a novel convolutional transformer that leverages a new \textbf{\textit{dynamic multi-headed convolutional self-attention}} mechanism for monocular 3D human pose estimation. We designed a spatial and temporal convolutional transformer to comprehensively model human joint relations within individual frames and globally across the motion sequence. Moreover, we introduce a novel notion of \textbf{\textit{temporal joints profile}} for our temporal ConvFormer that fuses complete temporal information immediately for a local neighborhood of joint features. We have quantitatively and qualitatively validated our method on three common benchmark datasets: Human3.6M, MPI-INF-3DHP, and HumanEva. Extensive experiments have been conducted to identify the optimal hyper-parameter set. These experiments demonstrated that we achieved a \textbf{significant parameter reduction relative to prior transformer models} while attaining State-of-the-Art (SOTA) or near SOTA on all three datasets. Additionally, we achieved SOTA for Protocol III on H36M for both GT and CPN detection inputs. Finally, we obtained SOTA on all three metrics for the MPI-INF-3DHP dataset and for all three subjects on HumanEva under Protocol II. | ['Dmitriy Shin', 'Alec Diaz-Arias'] | 2023-04-04 | null | null | null | null | ['3d-human-pose-estimation', 'monocular-3d-human-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-3.15032691e-01 6.53847307e-02 2.29861531e-02 -2.50250667e-01
-9.12861049e-01 -3.79025936e-01 3.22321028e-01 -6.45794868e-01
-6.17508709e-01 4.19295698e-01 4.59143966e-01 1.69956863e-01
-7.59135485e-02 -1.96760982e-01 -8.30181360e-01 -4.45472300e-01
-4.12573576e-01 5.63922107e-01 2.85092980e-01 -2.27517635e-01
-4.91057366e-01 3.33677799e-01 -1.13054025e+00 2.61184275e-01
2.45460093e-01 1.14277112e+00 3.65245603e-02 9.01204348e-01
9.21221614e-01 7.32460439e-01 -4.95345533e-01 -5.32286763e-01
4.64677274e-01 -8.74044299e-02 -1.00022936e+00 -6.22728430e-02
6.64338410e-01 -7.07631588e-01 -9.90809321e-01 4.65218842e-01
9.76594925e-01 1.98756784e-01 5.27171016e-01 -1.40398288e+00
-2.01353833e-01 1.88613653e-01 -5.81693351e-01 4.52124089e-01
5.60749471e-01 6.15746200e-01 8.56024325e-01 -9.48987365e-01
7.63815284e-01 1.37973940e+00 9.22025502e-01 4.33171898e-01
-9.56272304e-01 -6.53004766e-01 9.34281647e-02 3.21548790e-01
-1.39463973e+00 -1.91205293e-01 4.71808970e-01 -4.02921855e-01
1.26937068e+00 1.59177244e-01 1.01960504e+00 1.63496852e+00
5.75675547e-01 1.15755010e+00 7.86801279e-01 1.22168129e-02
-2.22477436e-01 -7.39874423e-01 3.65376584e-02 9.80356216e-01
-1.16243675e-01 4.42800760e-01 -1.05465698e+00 4.82882634e-02
1.02445149e+00 -3.83979023e-01 -1.43487960e-01 -2.47162029e-01
-1.59032798e+00 5.37881315e-01 5.43679535e-01 2.05925107e-02
-4.64292407e-01 7.45080709e-01 5.75647473e-01 7.57916346e-02
3.36632311e-01 5.28958580e-03 -6.81676269e-01 -4.92356032e-01
-7.16485083e-01 5.79806566e-01 5.24988770e-01 1.15540910e+00
1.51424244e-01 -2.98282225e-02 -5.64113140e-01 5.48458695e-01
2.62363493e-01 8.47988367e-01 8.28954205e-02 -1.42419600e+00
6.93908453e-01 1.40336826e-01 1.91220254e-01 -9.70989466e-01
-9.88380551e-01 -7.52535999e-01 -8.31160367e-01 -8.05391967e-02
4.63350594e-01 -2.53249943e-01 -1.16458511e+00 1.90502465e+00
4.48795199e-01 -2.28556599e-02 -3.63702416e-01 1.34072864e+00
7.90473878e-01 3.42093587e-01 4.98667508e-02 2.84897834e-01
1.61273432e+00 -1.00824738e+00 -3.98902208e-01 -2.24665493e-01
6.22596502e-01 -5.43083429e-01 9.79868531e-01 5.28680563e-01
-1.04404354e+00 -8.20016325e-01 -8.92272115e-01 -3.09898943e-01
2.11876899e-01 3.79744798e-01 5.28210282e-01 3.96934181e-01
-1.09794319e+00 5.06504297e-01 -1.27626479e+00 -3.82536650e-01
1.38706341e-01 5.84730983e-01 -5.99493802e-01 1.52954891e-01
-1.25277269e+00 8.97804022e-01 8.76861811e-02 6.03495002e-01
-1.15929377e+00 -5.29754221e-01 -8.12220514e-01 -1.30533323e-01
4.34335679e-01 -1.22285509e+00 1.25941706e+00 -1.07095927e-01
-1.53062105e+00 8.18631232e-01 8.62635672e-03 -6.95445418e-01
7.30470538e-01 -9.78024960e-01 -1.61215276e-01 4.85972881e-01
2.44387686e-01 1.17092919e+00 6.52340412e-01 -7.17656791e-01
-4.13157910e-01 -4.82878178e-01 4.13819253e-02 3.80536079e-01
1.11384718e-02 6.97759688e-02 -9.82809603e-01 -1.01673305e+00
4.72717285e-02 -1.51558971e+00 -8.39554146e-02 1.11967348e-01
-6.57080770e-01 -2.12960914e-01 5.98388433e-01 -9.28384423e-01
9.40864980e-01 -1.79538357e+00 7.30548620e-01 2.40602046e-01
5.07356644e-01 1.34457499e-01 -6.96773231e-02 1.23077050e-01
1.17159843e-01 -2.71138400e-01 2.02344611e-01 -6.49332941e-01
2.30954841e-01 2.49075636e-01 1.46557659e-01 7.54962027e-01
-1.89101741e-01 1.17608619e+00 -5.01566291e-01 -6.06483042e-01
3.91516805e-01 8.53843033e-01 -9.21779275e-01 1.36828393e-01
4.19970788e-02 6.74202502e-01 -4.77571249e-01 6.70759737e-01
2.73595780e-01 -3.45575035e-01 -7.49612674e-02 -5.80697477e-01
1.79291978e-01 -8.35942701e-02 -1.00899863e+00 1.92448425e+00
-4.54350151e-02 5.32941282e-01 -1.79939102e-02 -5.35972059e-01
4.23910797e-01 4.88752335e-01 1.09119070e+00 -6.27923012e-01
6.14697635e-01 -8.01783577e-02 -1.56312242e-01 -3.70013684e-01
5.05114853e-01 4.16489840e-01 -3.93568307e-01 1.60219803e-01
4.25619483e-01 3.51194799e-01 5.26280433e-04 6.40340671e-02
1.19469750e+00 5.48868895e-01 4.49630544e-02 -4.09636497e-01
1.81982309e-01 -2.36616999e-01 6.19532645e-01 6.95320368e-01
-6.18674517e-01 7.87422299e-01 3.65281433e-01 -5.67639887e-01
-9.79225993e-01 -1.31517696e+00 2.86119998e-01 1.15455651e+00
4.26746607e-02 -6.68740034e-01 -8.93054903e-01 -6.36191428e-01
-6.72535151e-02 5.63358404e-02 -8.02685559e-01 -1.59759685e-01
-9.31350827e-01 -5.42060196e-01 1.15503991e+00 9.98824656e-01
8.10316920e-01 -8.53054345e-01 -1.15705645e+00 1.40242741e-01
-7.60090649e-01 -1.46862948e+00 -9.92192805e-01 6.46075839e-03
-3.92910421e-01 -1.11440885e+00 -1.01140118e+00 -4.49149877e-01
1.02825668e-02 -4.21225354e-02 9.55293953e-01 -3.86682898e-01
-4.41854775e-01 4.87741143e-01 -2.87710369e-01 -4.53897230e-02
3.62276733e-01 1.66368589e-01 3.48903358e-01 -2.47267604e-01
-3.94354953e-04 -4.96855289e-01 -8.97982240e-01 7.29527414e-01
-1.55157328e-01 1.05122939e-01 4.98179406e-01 6.90481842e-01
5.85510612e-01 -4.85259682e-01 1.17903262e-01 -2.10972831e-01
8.51816833e-02 -3.48088890e-02 -2.57543951e-01 -6.22852519e-02
2.43158769e-02 8.76558349e-02 1.20018892e-01 -4.05187994e-01
-8.61405373e-01 1.53724641e-01 -3.96241039e-01 -8.53052855e-01
7.43926466e-02 2.69388080e-01 -8.80327076e-02 6.49296939e-02
6.03506744e-01 6.62025288e-02 -2.43866481e-02 -4.31173861e-01
2.99276710e-01 1.65105015e-01 1.07526100e+00 -7.30276942e-01
6.71928644e-01 7.03285396e-01 1.18082881e-01 -7.32149303e-01
-8.72280955e-01 -4.95945960e-01 -8.69612157e-01 -4.60778236e-01
1.30261374e+00 -1.19744492e+00 -1.27345693e+00 8.19863141e-01
-1.13818169e+00 -6.41733050e-01 2.54842755e-03 7.52733707e-01
-9.65894699e-01 4.99825269e-01 -8.62125158e-01 -3.75772804e-01
-6.28535569e-01 -1.20597064e+00 1.51684952e+00 -1.88775435e-01
-5.48415542e-01 -6.16444290e-01 -1.30050126e-02 6.90819502e-01
1.66620091e-01 5.49164474e-01 3.49665821e-01 -1.69232011e-01
-3.67138445e-01 -7.99418315e-02 -8.55950788e-02 1.32497624e-01
-2.45288715e-01 -4.72482979e-01 -7.53903925e-01 -5.97198784e-01
-3.11753511e-01 -3.80418330e-01 6.98123097e-01 5.94847023e-01
9.39990342e-01 -1.19835660e-01 -4.25446302e-01 8.65405500e-01
5.80423415e-01 -1.49452791e-01 5.85934758e-01 2.13854671e-01
9.94956136e-01 2.07817048e-01 6.03880346e-01 6.00548625e-01
5.93800724e-01 1.27464962e+00 3.36795479e-01 -1.32689014e-01
-3.26036066e-01 -1.67737588e-01 5.00292420e-01 5.52458704e-01
-4.94152755e-01 -3.24605644e-01 -8.77175033e-01 4.79445308e-01
-1.91193128e+00 -8.10992897e-01 1.08970083e-01 1.80972040e+00
5.10604024e-01 2.44196951e-01 5.45628965e-01 -5.64094968e-02
3.98215413e-01 1.17859930e-01 -4.49326605e-01 1.82492256e-01
5.34251034e-02 3.18669528e-01 4.57154840e-01 3.84457260e-01
-1.44674778e+00 9.42878425e-01 5.86243963e+00 6.06275737e-01
-8.48635793e-01 2.25420341e-01 2.72011757e-01 -5.17908990e-01
3.78794014e-01 -3.74921203e-01 -9.26553607e-01 2.76569486e-01
7.83318758e-01 2.73112983e-01 1.85073525e-01 6.57948971e-01
2.65110433e-01 -2.14911764e-03 -1.20683122e+00 1.06232572e+00
-9.42056701e-02 -8.77751172e-01 -1.39017612e-01 3.35617751e-01
3.47054303e-01 2.66002834e-01 1.61705866e-01 3.89900208e-01
2.59981602e-01 -9.93745267e-01 1.09619331e+00 4.11471099e-01
7.40910172e-01 -7.62760341e-01 5.86410999e-01 2.39506468e-01
-1.69104898e+00 2.85620764e-02 9.47172642e-02 1.67806670e-01
4.42338645e-01 2.77301759e-01 -6.31155074e-01 7.11665630e-01
1.20641696e+00 6.81963086e-01 -5.49018621e-01 7.80579329e-01
-2.16142833e-01 3.48725438e-01 -6.68550849e-01 3.50314528e-01
2.70110399e-01 5.46374202e-01 8.25480759e-01 1.19192040e+00
1.41518027e-01 1.80179432e-01 4.08136398e-01 4.09811497e-01
1.39699489e-01 -3.20227087e-01 -1.70085832e-01 4.16941822e-01
1.82056800e-01 9.29498553e-01 -5.58254302e-01 -9.92041752e-02
-7.17704818e-02 1.23555100e+00 2.21011043e-01 3.14788699e-01
-1.43010139e+00 6.87404862e-03 8.47361326e-01 1.22177176e-01
4.81240779e-01 -6.25627577e-01 1.10031649e-01 -1.18985260e+00
1.28741547e-01 -8.39552164e-01 7.02534735e-01 -8.08547437e-01
-9.14171636e-01 6.60415232e-01 5.55414855e-01 -1.09193289e+00
-5.18417239e-01 -6.16415322e-01 -5.99726252e-02 4.62828964e-01
-6.88577175e-01 -1.56064045e+00 -4.57911283e-01 1.08517957e+00
3.55302960e-01 5.83026186e-02 5.57704866e-01 5.87797701e-01
-5.11223614e-01 9.84864771e-01 -5.00758052e-01 3.91962349e-01
6.66885734e-01 -9.93087649e-01 7.52359807e-01 8.71272624e-01
-1.24152843e-02 5.61369300e-01 7.62242436e-01 -7.28153408e-01
-1.48579764e+00 -1.08520365e+00 5.59546709e-01 -6.47709489e-01
3.04916948e-01 -4.94125009e-01 -2.80628622e-01 1.20901597e+00
1.14992388e-01 1.53301209e-01 1.99737370e-01 1.42378449e-01
-2.50058293e-01 9.59324837e-02 -9.71885562e-01 5.89238644e-01
1.57950449e+00 -4.22620267e-01 -5.10337770e-01 3.53468835e-01
9.17411089e-01 -9.32518542e-01 -1.18119562e+00 7.27993846e-01
9.10691559e-01 -8.14388931e-01 1.22764707e+00 -3.92530769e-01
1.08745381e-01 -2.82346457e-01 -3.72485757e-01 -1.05287170e+00
-5.07938325e-01 -9.16191936e-01 -4.53014016e-01 5.28916299e-01
-5.15261367e-02 -1.61164418e-01 1.11529315e+00 1.92304641e-01
-2.90549785e-01 -8.11910212e-01 -1.35224044e+00 -8.12474787e-01
-7.48717189e-02 -6.70401812e-01 6.78166449e-02 3.40398222e-01
-2.37619951e-01 2.24704504e-01 -9.08453941e-01 2.19417155e-01
9.09177780e-01 -3.16047788e-01 1.06261063e+00 -6.51421964e-01
-5.80550313e-01 -6.40910193e-02 -6.06948733e-01 -1.50172126e+00
-5.72884921e-03 -6.34841800e-01 8.39077309e-03 -1.23930085e+00
2.96296835e-01 1.48405833e-02 -4.84546348e-02 7.64599442e-01
6.29502460e-02 4.55889970e-01 3.46918583e-01 3.50028127e-02
-6.96295917e-01 6.30732119e-01 1.33251345e+00 3.53784449e-02
1.12137899e-01 -1.76131725e-01 -1.41579553e-01 6.71566248e-01
3.87098014e-01 -2.29928061e-01 -3.99091393e-01 -5.73507309e-01
-4.60537598e-02 2.87293732e-01 1.04189551e+00 -1.42719221e+00
1.96450964e-01 2.07061008e-01 6.53202951e-01 -1.06381953e+00
8.37681830e-01 -5.37817836e-01 3.34650159e-01 6.02227867e-01
-2.92314659e-03 3.02227169e-01 2.58216649e-01 4.47946429e-01
1.94732845e-01 8.43970597e-01 7.67790437e-01 -4.13886048e-02
-7.96903729e-01 5.59712052e-01 -3.33374351e-01 3.24085265e-01
8.49256396e-01 -1.40523911e-01 -1.32164866e-01 -4.40496981e-01
-1.21415210e+00 5.60284734e-01 1.23927752e-02 5.85044742e-01
6.46246731e-01 -1.46452379e+00 -7.46346712e-01 7.65290335e-02
4.55623493e-03 -6.62769005e-02 4.94771034e-01 1.09139204e+00
-5.89563489e-01 7.78870702e-01 -4.61771965e-01 -9.87328947e-01
-1.30833387e+00 2.58635998e-01 7.02620029e-01 -5.20148814e-01
-9.56289828e-01 9.59809005e-01 3.08530062e-01 -4.57643598e-01
4.67041492e-01 -5.32561719e-01 2.51391411e-01 -2.37599626e-01
1.50251180e-01 6.48345649e-01 -7.38459527e-02 -9.86887872e-01
-6.93781078e-01 6.60483062e-01 7.68286586e-02 -3.38087559e-01
1.17853415e+00 -2.67014951e-01 4.81915683e-01 -1.38629124e-01
1.42098534e+00 -4.11520302e-01 -1.68538857e+00 -1.09079674e-01
-4.48874444e-01 -2.51156807e-01 -2.01427013e-01 -8.78034830e-01
-1.23680949e+00 6.97231054e-01 7.26818979e-01 -6.89532816e-01
9.61855948e-01 1.28409266e-01 1.09774685e+00 4.65873599e-01
7.90832102e-01 -1.05464208e+00 5.46058178e-01 7.07699060e-01
1.02571237e+00 -7.47031033e-01 5.44519462e-02 -4.75709051e-01
-6.64587557e-01 6.67806566e-01 7.93471038e-01 -1.21323705e-01
6.69591010e-01 2.63717830e-01 -1.57558605e-01 -4.91465837e-01
-6.29035413e-01 -1.76871926e-01 5.55746913e-01 4.19647634e-01
3.29349786e-01 2.56753080e-02 7.39573129e-03 6.73157454e-01
-5.64401388e-01 2.08813593e-01 -4.85128686e-02 1.02421546e+00
-1.36396959e-01 -6.76912606e-01 -2.46369511e-01 3.32845338e-02
-5.05084753e-01 2.41471618e-01 -2.59313971e-01 1.18508112e+00
1.26165986e-01 7.29195058e-01 -2.46064335e-01 -8.52835536e-01
5.71845472e-01 -1.99981049e-01 9.00606453e-01 -1.54657066e-01
-7.42593348e-01 2.97199667e-01 3.58615786e-01 -1.29619503e+00
-2.73096681e-01 -8.40942919e-01 -1.40666306e+00 -5.64676404e-01
5.58207519e-02 -3.01193774e-01 4.46941294e-02 1.05542767e+00
3.84158164e-01 7.05365419e-01 1.20003581e-01 -1.28252673e+00
-4.15788949e-01 -1.02619183e+00 -2.92336345e-01 2.55319566e-01
1.39645517e-01 -1.08932745e+00 1.44583136e-01 -7.13744313e-02] | [7.110671043395996, -0.6745200157165527] |
b73b5d75-8592-4b51-8153-ee579a93126b | spectral-response-function-guided-deep | 2011.09701 | null | https://arxiv.org/abs/2011.09701v2 | https://arxiv.org/pdf/2011.09701v2.pdf | Spectral Response Function Guided Deep Optimization-driven Network for Spectral Super-resolution | Hyperspectral images are crucial for many research works. Spectral super-resolution (SSR) is a method used to obtain high spatial resolution (HR) hyperspectral images from HR multispectral images. Traditional SSR methods include model-driven algorithms and deep learning. By unfolding a variational method, this paper proposes an optimization-driven convolutional neural network (CNN) with a deep spatial-spectral prior, resulting in physically interpretable networks. Unlike the fully data-driven CNN, auxiliary spectral response function (SRF) is utilized to guide CNNs to group the bands with spectral relevance. In addition, the channel attention module (CAM) and reformulated spectral angle mapper loss function are applied to achieve an effective reconstruction model. Finally, experiments on two types of datasets, including natural and remote sensing images, demonstrate the spectral enhancement effect of the proposed method. And the classification results on the remote sensing dataset also verified the validity of the information enhanced by the proposed method. | ['Liangpei Zhang', 'Huanfeng Shen', 'Qiangqiang Yuan', 'Jie Li', 'Jiang He'] | 2020-11-19 | null | null | null | null | ['spectral-super-resolution'] | ['computer-vision'] | [ 8.58675182e-01 -3.06147784e-01 1.88303308e-03 -4.16028887e-01
-8.32641065e-01 -4.54697907e-02 3.54543358e-01 -6.34990394e-01
-1.16199113e-01 9.64931011e-01 1.61391303e-01 -2.62980819e-01
-4.64199513e-01 -1.00338089e+00 -6.08071268e-01 -1.10198307e+00
1.92947581e-01 -3.47768784e-01 -4.38778877e-01 -4.46670145e-01
1.04959287e-01 7.22088158e-01 -1.53925383e+00 2.34476402e-01
1.44580710e+00 1.21782768e+00 7.35028267e-01 2.58543640e-01
1.67131484e-01 5.23703516e-01 -8.20147097e-02 4.54245716e-01
5.72097778e-01 -3.71105283e-01 -7.12016046e-01 4.45541054e-01
2.92753309e-01 -4.32254076e-01 -5.86587310e-01 1.61122024e+00
5.86689949e-01 4.86700922e-01 3.34083468e-01 -5.82287788e-01
-1.12189209e+00 7.80887604e-01 -9.77685750e-01 1.04220472e-01
-2.06819355e-01 2.15238735e-01 9.72529471e-01 -9.67328072e-01
1.60613000e-01 1.18211949e+00 4.42265481e-01 1.91573367e-01
-1.12657034e+00 -7.53168583e-01 -2.04532500e-02 3.09697866e-01
-1.68570769e+00 -1.42519072e-01 1.03146362e+00 -2.29748607e-01
5.36517084e-01 3.75004441e-01 6.04848206e-01 5.39280593e-01
-2.36981034e-01 5.11853278e-01 1.31977379e+00 -1.66199729e-01
-1.84625983e-01 -1.14547409e-01 5.56709208e-02 2.59450018e-01
1.40930831e-01 5.50201595e-01 -2.25044772e-01 2.54369825e-01
1.03020298e+00 2.96307921e-01 -8.00273895e-01 1.74643949e-01
-8.76118422e-01 8.40177059e-01 1.15479887e+00 1.90872718e-02
-6.98267043e-01 -1.90907493e-01 -9.88234580e-02 -1.13869220e-01
7.89363801e-01 3.47743869e-01 -3.16690534e-01 7.07913220e-01
-9.23505843e-01 -1.16354324e-01 -7.12284520e-02 6.64848149e-01
1.06390107e+00 5.89784205e-01 -8.82060751e-02 1.26280069e+00
4.41002995e-01 6.05305851e-01 1.93516374e-01 -9.27488029e-01
6.26901165e-02 4.05973703e-01 1.81797311e-01 -7.77466595e-01
-4.10945624e-01 -9.84589994e-01 -1.40393221e+00 7.67570361e-02
-4.92689461e-01 -1.14874542e-01 -1.08332491e+00 1.42543066e+00
6.86845407e-02 1.27737507e-01 1.77553371e-01 1.60023487e+00
9.85981941e-01 1.00594866e+00 6.81854486e-02 -3.91070217e-01
9.33496714e-01 -9.78700340e-01 -7.57477105e-01 -9.68733057e-02
-1.85183119e-02 -3.89374286e-01 1.05654943e+00 3.47740740e-01
-7.63155162e-01 -7.00907469e-01 -1.20797729e+00 -6.51270598e-02
-3.03661734e-01 3.97712231e-01 8.20567369e-01 2.32106328e-01
-7.82304466e-01 8.22940528e-01 -5.06990194e-01 3.18955928e-02
9.14954841e-01 3.80881913e-02 -1.50509784e-02 4.61212657e-02
-1.34272385e+00 6.00570381e-01 7.73003161e-01 7.96346486e-01
-8.99756134e-01 -7.42459714e-01 -7.33113468e-01 3.10904860e-01
7.45015070e-02 -3.28828186e-01 8.68019760e-01 -1.34526479e+00
-1.54042530e+00 6.74831212e-01 -8.29477832e-02 -2.49242410e-01
5.03145270e-02 -1.17364503e-01 -7.48469293e-01 4.82543379e-01
-6.40788674e-02 4.48124051e-01 7.44853556e-01 -1.44185352e+00
-7.12714970e-01 -4.49723214e-01 -6.87956139e-02 4.51603085e-01
-2.85537273e-01 -1.81847334e-01 2.00245693e-01 -4.48665291e-01
5.12729406e-01 -5.08156776e-01 -3.61422330e-01 -2.12283090e-01
-5.98810911e-01 2.18718365e-01 7.72169113e-01 -9.75451171e-01
7.29524314e-01 -2.14516330e+00 1.86079130e-01 2.55829334e-01
1.73372388e-01 3.66088480e-01 -2.20153138e-01 4.60747518e-02
-6.78559482e-01 3.99282843e-01 -6.55424356e-01 3.08744550e-01
-2.37703636e-01 -1.88105509e-01 -4.26875353e-01 6.31030619e-01
3.32781076e-01 8.25054884e-01 -5.90862513e-01 5.34332581e-02
5.07043958e-01 8.57613266e-01 9.39222649e-02 1.84586167e-01
-2.25897953e-01 7.08468854e-01 -4.84013647e-01 6.85673892e-01
1.22286236e+00 -5.46714783e-01 -1.25581905e-01 -7.22302735e-01
-4.16156769e-01 -1.05086826e-01 -1.01805794e+00 1.53027689e+00
-3.00094098e-01 3.63349199e-01 3.74415755e-01 -8.62796307e-01
1.14049983e+00 6.58393130e-02 4.34728146e-01 -9.88889277e-01
-6.53321370e-02 1.83042333e-01 -3.13340276e-01 -5.52029788e-01
5.95409334e-01 -2.97531188e-01 8.31951141e-01 -2.73948591e-02
-4.78762180e-01 1.83297694e-02 -4.92470890e-01 -3.48401219e-01
-2.97251623e-02 4.13126081e-01 2.83051934e-02 -2.95087546e-01
7.58711815e-01 1.64592251e-01 5.65110862e-01 4.10739034e-01
5.35788871e-02 7.29457855e-01 -2.43123099e-01 -5.33013284e-01
-1.27231216e+00 -5.56095779e-01 -5.96524715e-01 8.08539391e-01
4.52032357e-01 5.35415113e-01 -3.92562389e-01 8.45895410e-02
-2.27970794e-01 6.23940051e-01 -5.17238796e-01 -1.04644932e-01
-9.09814611e-03 -1.37416017e+00 6.72219098e-02 4.74247605e-01
1.16781890e+00 -1.05252087e+00 -2.01540947e-01 1.21786356e-01
-3.05939406e-01 -8.88228893e-01 9.56440195e-02 -1.29520655e-01
-8.93303573e-01 -8.94402504e-01 -6.57848060e-01 -4.18710470e-01
3.93486887e-01 8.72281551e-01 3.77385288e-01 -1.77462995e-01
-3.32720250e-01 -2.73062021e-01 -4.10095900e-01 -2.78428346e-01
2.61393543e-02 2.52565350e-02 -4.53506142e-01 3.77134234e-01
1.19195491e-01 -5.63328445e-01 -8.17024648e-01 -2.33412385e-02
-1.13056374e+00 3.92099857e-01 7.73326695e-01 1.01569831e+00
9.17713940e-01 4.55572039e-01 6.30442798e-01 -7.26608753e-01
4.01898235e-01 -4.84547615e-01 -7.87116945e-01 1.00386828e-01
-7.43189394e-01 -3.12075168e-01 6.67771637e-01 -2.35520899e-02
-1.58998001e+00 9.20683742e-02 -8.99074674e-02 -4.97154206e-01
-3.13805014e-01 1.05278695e+00 -2.20954195e-01 -2.73731858e-01
8.08854401e-01 6.65768862e-01 3.12583055e-03 -5.36210597e-01
2.78009921e-01 1.01808035e+00 5.30102313e-01 -7.05085769e-02
7.84845293e-01 7.86422729e-01 2.53712889e-02 -1.11050260e+00
-1.28902495e+00 -4.12972331e-01 -3.38455528e-01 -1.23574853e-01
9.81982172e-01 -1.40377200e+00 -6.41422927e-01 6.32409513e-01
-8.80774021e-01 -2.84670234e-01 1.34895490e-02 6.42313778e-01
-4.19075400e-01 4.60416317e-01 -5.07818997e-01 -1.10214901e+00
-6.51678026e-01 -9.42796886e-01 9.73510563e-01 6.16983473e-01
9.17230785e-01 -8.00809085e-01 -4.82730120e-01 5.98702192e-01
5.89342415e-01 3.16030771e-01 8.11588943e-01 -2.87508965e-03
-8.12780142e-01 1.74893484e-01 -1.04325688e+00 5.33607185e-01
2.37733006e-01 1.03815854e-01 -1.40650237e+00 -3.10558975e-01
2.69079506e-02 -3.52435917e-01 1.28290093e+00 8.48870218e-01
1.87050831e+00 -1.35609016e-01 9.43608582e-02 1.34246361e+00
2.07822680e+00 2.86181599e-01 9.78803635e-01 5.20740807e-01
1.00705063e+00 5.19441903e-01 7.01091290e-01 4.53943878e-01
2.58876979e-02 7.88401216e-02 1.03132772e+00 -7.34462202e-01
1.32792547e-01 -1.06229214e-02 -3.78029831e-02 1.79162011e-01
-5.16871035e-01 4.41213809e-02 -6.57716036e-01 2.90723801e-01
-1.80495882e+00 -1.24424374e+00 -4.95678693e-01 1.96548593e+00
6.79032147e-01 -5.61113834e-01 -3.74602824e-01 -8.94871578e-02
9.89809811e-01 6.28094494e-01 -9.08593655e-01 3.64339978e-01
-8.03105056e-01 2.17922464e-01 8.65783036e-01 5.44037104e-01
-1.37149012e+00 1.04210150e+00 5.72058249e+00 7.57277012e-01
-1.59184885e+00 3.08725480e-02 7.15701878e-01 2.65054256e-01
-5.60681343e-01 -1.49139479e-01 -3.88125181e-01 7.12432563e-02
5.70718586e-01 -1.15688778e-01 9.90783513e-01 6.24936044e-01
8.73554528e-01 9.62919965e-02 -2.60689706e-01 1.07194543e+00
-2.71052212e-01 -1.41316271e+00 2.41595700e-01 1.10980906e-01
8.37110817e-01 2.95388311e-01 3.61909062e-01 -1.44329414e-01
1.73374310e-01 -1.41204596e+00 4.59292531e-01 8.25123608e-01
1.14859915e+00 -9.19968843e-01 6.65686965e-01 1.39285505e-01
-1.27379775e+00 -4.31977987e-01 -9.32905197e-01 5.83356172e-02
-1.79959431e-01 6.64813519e-01 -3.27284098e-01 1.14467371e+00
7.38599539e-01 1.21650946e+00 -2.98658401e-01 9.35374439e-01
-2.98931271e-01 6.10400975e-01 1.35315314e-01 4.57745910e-01
3.09814453e-01 -7.56160557e-01 5.08194387e-01 8.94017398e-01
3.44830126e-01 5.84256649e-01 3.98090743e-02 1.48926830e+00
5.41282594e-02 3.59928198e-02 -4.07529563e-01 -2.03379110e-01
4.16113019e-01 1.64628887e+00 -7.04348460e-02 -5.07314205e-02
-2.16687679e-01 7.49215782e-01 -1.31528422e-01 9.65608776e-01
-7.44538784e-01 -4.01834339e-01 4.98006582e-01 -1.67489842e-01
2.91063517e-01 -4.82346080e-02 -4.70110565e-01 -1.20630562e+00
-2.50172138e-01 -8.80245984e-01 1.29100457e-01 -1.53968799e+00
-1.18695533e+00 5.27880430e-01 -3.45220417e-01 -1.33187783e+00
5.38113534e-01 -5.81593633e-01 -4.15079206e-01 1.65776062e+00
-2.59546065e+00 -1.29474366e+00 -9.98713195e-01 6.11223161e-01
3.92486781e-01 -3.47236134e-02 6.21405423e-01 1.80532351e-01
-8.48609984e-01 -1.63101479e-01 3.90780926e-01 -3.49016935e-02
2.72851706e-01 -1.04778397e+00 -2.55575895e-01 1.01748252e+00
-6.50163531e-01 4.19702560e-01 4.48659092e-01 -4.19649988e-01
-1.30345976e+00 -1.71869624e+00 -3.12012341e-02 5.77698708e-01
5.46527207e-01 4.63117719e-01 -1.20167291e+00 5.06594181e-01
2.46989846e-01 -2.15749498e-02 6.15342081e-01 -2.23322690e-01
-3.13491434e-01 -4.94720697e-01 -9.92043555e-01 3.05878341e-01
8.39647353e-01 -6.18184268e-01 -2.30479479e-01 7.06537724e-01
8.54867458e-01 -1.16360776e-01 -9.95997846e-01 7.43027866e-01
2.31473729e-01 -9.20064628e-01 1.00429380e+00 -5.19407332e-01
8.60797822e-01 -4.59894031e-01 -3.36643845e-01 -1.49957716e+00
-7.98018456e-01 -1.67050466e-01 3.33110899e-01 6.77129447e-01
4.36918408e-01 -7.15855658e-01 4.80421335e-01 2.06006050e-01
-5.43295562e-01 -3.17000002e-01 -3.59984159e-01 -6.01312459e-01
9.98797547e-03 -7.09952638e-02 9.14392173e-01 1.32747364e+00
-5.62160969e-01 1.78727254e-01 -4.14930373e-01 9.85672474e-01
8.73270035e-01 4.66558367e-01 2.16913268e-01 -1.35970962e+00
8.41676816e-02 -6.78415298e-01 2.77325422e-01 -8.37398350e-01
2.20424548e-01 -9.38631058e-01 1.01365566e-01 -1.79731333e+00
4.35867339e-01 -1.63885131e-01 -5.33370912e-01 4.56505328e-01
-4.25517797e-01 9.18880701e-02 -2.03659937e-01 4.34648126e-01
-1.45530440e-02 9.44749534e-01 1.62235868e+00 -3.89952898e-01
-2.58396119e-01 -1.97853804e-01 -8.41816604e-01 3.81388813e-01
1.04706061e+00 4.32563759e-02 -3.86168629e-01 -5.22307456e-01
2.16667414e-01 3.44564289e-01 7.65371501e-01 -6.57958210e-01
3.51011641e-02 -6.04710162e-01 5.32066643e-01 -8.16118240e-01
3.90889764e-01 -7.97995269e-01 2.75550753e-01 2.16343880e-01
-4.07017469e-01 -8.39839160e-01 1.36230260e-01 3.34462196e-01
-3.43329787e-01 4.99406010e-02 1.22426391e+00 -1.88204631e-01
-1.03239548e+00 9.59219933e-01 1.22956835e-01 -6.32581472e-01
7.49601662e-01 -3.97006035e-01 -5.42450368e-01 -2.76441455e-01
-6.55813634e-01 3.27136785e-01 1.96235657e-01 4.27159667e-02
9.80630577e-01 -1.23119366e+00 -1.15503097e+00 3.29619765e-01
1.99673191e-01 2.89234370e-01 8.42629135e-01 6.86277807e-01
-7.95580804e-01 1.83672950e-01 -3.95879537e-01 -5.39608061e-01
-8.27394962e-01 3.40192080e-01 1.02941430e+00 2.97573596e-01
-4.97185379e-01 7.96625137e-01 1.20102145e-01 -5.44330835e-01
-2.37602070e-01 7.22243711e-02 -6.16192758e-01 -5.20498306e-02
5.89694560e-01 4.12557155e-01 -1.00664698e-01 -6.62484825e-01
7.79486895e-02 4.28147227e-01 3.33260775e-01 -8.38975087e-02
1.88403261e+00 -3.29879016e-01 -5.39094746e-01 1.95840195e-01
1.01774156e+00 -4.51288313e-01 -1.60284340e+00 -5.00544012e-01
-5.33086240e-01 -6.09226763e-01 7.77432621e-01 -8.38267565e-01
-1.45389879e+00 1.15954530e+00 7.45554745e-01 3.16417426e-01
1.47398388e+00 -4.76278216e-01 5.38103282e-01 3.36292595e-01
3.35697271e-02 -1.16142547e+00 -4.41299289e-01 4.40680891e-01
9.96922016e-01 -1.51090658e+00 6.86412230e-02 -5.38448513e-01
-4.85299826e-01 1.31087399e+00 6.22581720e-01 1.23451434e-01
5.37817776e-01 -4.08664256e-01 -1.33186756e-02 -2.44414017e-01
-1.42900273e-01 -4.58293110e-01 2.33617663e-01 6.14082992e-01
3.68062109e-01 2.01282203e-01 -1.74349949e-01 4.13288265e-01
4.29409631e-02 6.03559315e-02 5.38060486e-01 3.26741546e-01
-8.52772653e-01 -3.12266141e-01 -3.48321050e-01 5.44892967e-01
-1.23847581e-01 -5.09143353e-01 -9.57344100e-02 3.29661459e-01
-1.48617759e-01 1.05830181e+00 -7.98013285e-02 -4.16759104e-01
8.30305815e-02 -3.29682380e-01 -3.01613915e-03 -6.59952700e-01
-2.63698865e-02 3.32015783e-01 -2.92703807e-01 -5.03321648e-01
-8.89187992e-01 -2.85795480e-01 -1.28641570e+00 -2.28129685e-01
-5.55974662e-01 2.79121171e-03 7.50809968e-01 8.58511031e-01
2.23440528e-01 7.52198875e-01 1.17054784e+00 -9.13644969e-01
-6.39728189e-01 -1.10320818e+00 -1.33631516e+00 1.50911719e-01
5.90722859e-01 -2.74756432e-01 -4.20107812e-01 -3.22328545e-02] | [10.137588500976562, -1.9806994199752808] |
373cad38-02e7-4e10-af24-c202c02b7e1e | gibbonr-an-r-package-for-the-detection-and | 1906.02572 | null | https://arxiv.org/abs/1906.02572v2 | https://arxiv.org/pdf/1906.02572v2.pdf | GIBBONFINDR: An R package for the detection and classification of acoustic signals | The recent improvements in recording technology, data storage and battery life have led to an increased interest in the use of passive acoustic monitoring for a variety of research questions. One of the main obstacles in implementing wide scale acoustic monitoring programs in terrestrial environments is the lack of user-friendly, open source programs for processing large sound archives. Here we describe the new, open-source R package GIBBONFINDR which has functions for detection, classification and visualization of acoustic signals using a variety of readily available machine learning algorithms in the R programming environment. We provide a case study showing how GIBBONFINDR functions can be used in a workflow to detect and classify Bornean gibbon (Hylobates muelleri) calls in long-term acoustic data sets recorded in Danum Valley Conservation Area, Sabah, Malaysia. Machine learning is currently one of the most rapidly growing fields-- with applications across many disciplines-- and our goal is to make commonly used signal processing techniques and machine learning algorithms readily available for ecologists who are interested in incorporating bioacoustics techniques into their research. | ['Dena J. Clink', 'Holger Klinck'] | 2019-06-06 | null | null | null | null | ['acoustic-modelling'] | ['speech'] | [ 1.17398977e-01 -6.65214896e-01 9.46489573e-01 -4.47321504e-01
-8.52043986e-01 -6.58111989e-01 -1.59486830e-01 4.89211857e-01
-7.79547811e-01 3.66699189e-01 6.33910745e-02 -4.69584525e-01
-2.99508423e-01 -5.38065016e-01 -2.36752689e-01 -9.60978448e-01
-9.21008408e-01 1.93670556e-01 3.68144065e-01 -1.12029657e-01
3.47326905e-01 7.85044074e-01 -1.86715925e+00 -9.09943730e-02
1.28002197e-01 5.57346880e-01 6.55961871e-01 1.14399171e+00
1.00049622e-01 3.39142047e-03 -7.77662933e-01 5.43080084e-03
8.57566744e-02 -8.74423459e-02 -4.34459329e-01 -9.60524082e-01
1.18696606e-02 -1.02299407e-01 2.98377275e-01 5.93003035e-01
1.18603539e+00 2.89862126e-01 3.53257239e-01 -8.26239407e-01
1.23857960e-01 6.58400059e-01 -4.22014683e-01 5.56034982e-01
3.21596950e-01 2.01362610e-01 8.94681931e-01 -3.83318573e-01
-1.16975591e-01 1.31102526e+00 1.03993058e+00 3.18424672e-01
-1.12405384e+00 -1.06752229e+00 -4.48509067e-01 1.33560583e-01
-1.14970803e+00 -7.78579175e-01 4.47012782e-01 -5.62764049e-01
1.30054045e+00 7.58766294e-01 7.25754619e-01 4.16914374e-01
9.65429023e-02 1.04025252e-01 1.29752600e+00 -2.07807153e-01
2.98222780e-01 6.33831369e-03 4.64428030e-02 3.64002287e-01
1.35977522e-01 1.87733963e-01 -9.12466764e-01 -5.76633513e-01
3.88754666e-01 -2.62652516e-01 -1.66019529e-01 5.42018414e-01
-8.51634979e-01 7.28831470e-01 1.63389012e-01 2.45816916e-01
-1.29231900e-01 1.10118285e-01 4.67129558e-01 2.86479056e-01
6.70679748e-01 4.68045980e-01 -6.54933214e-01 -7.44704127e-01
-9.57989275e-01 -4.25731391e-03 9.11888003e-01 2.57989049e-01
4.02660102e-01 3.51463825e-01 8.15969229e-01 1.53383946e+00
1.00937486e+00 7.89467573e-01 4.84843075e-01 -6.38805926e-01
-1.65611476e-01 3.05668842e-02 -2.09499717e-01 -9.29275393e-01
-6.33204281e-01 -1.46855950e-01 -3.44983071e-01 2.31829301e-01
2.19111100e-01 -2.91317403e-01 -7.17460692e-01 1.29642177e+00
5.64604521e-01 1.58392027e-01 3.52688543e-02 5.74388444e-01
1.01798344e+00 1.02663338e+00 1.51564091e-01 -2.46770456e-01
1.39077306e+00 -1.21595509e-01 -3.56073946e-01 -2.62599736e-01
2.57902682e-01 -8.28953683e-01 8.89198005e-01 5.19674540e-01
-5.90745091e-01 -1.41949788e-01 -9.68457401e-01 4.18764532e-01
-7.46802032e-01 -3.07299286e-01 6.70627534e-01 1.09531999e+00
-1.01697993e+00 2.74733871e-01 -1.37890017e+00 -6.88407063e-01
2.75169164e-01 5.03528655e-01 -5.41599333e-01 3.75210017e-01
-6.72496498e-01 7.01180637e-01 -2.55955637e-01 4.16592062e-01
-1.04564548e+00 -6.62182927e-01 -6.63782597e-01 -2.31072456e-01
-3.59038740e-01 2.55739927e-01 1.32306564e+00 -1.27497539e-01
-1.48417556e+00 8.59563291e-01 9.10635293e-02 -6.05021939e-02
2.89820116e-02 -1.99256226e-01 -4.81258899e-01 -2.33116910e-01
9.57160667e-02 3.45345050e-01 2.56474525e-01 -9.56925988e-01
-9.11304712e-01 -6.39542639e-01 -6.88720822e-01 7.60713518e-02
-1.99177206e-01 7.39285409e-01 2.97595263e-01 -5.93139350e-01
2.03806937e-01 -6.43288434e-01 -1.99056312e-01 9.38137546e-02
2.08096504e-01 -1.86368055e-03 7.33616292e-01 -8.68756175e-01
1.08225620e+00 -2.24052000e+00 -1.34897247e-01 1.48691177e-01
-3.90821874e-01 2.44951308e-01 -3.07384459e-03 6.08967543e-01
2.09092513e-01 1.43043429e-01 -5.47854662e-01 -1.37526244e-01
-2.96870619e-01 5.21463513e-01 8.06377754e-02 7.42535889e-01
-3.66021752e-01 -2.06696212e-01 -1.03082883e+00 -2.47830853e-01
2.44523346e-01 5.66485465e-01 -2.21379027e-01 3.97593290e-01
2.84298778e-01 5.88243544e-01 -1.12442493e-01 8.68498921e-01
8.55974972e-01 8.23022187e-01 -2.32381999e-01 3.12729836e-01
-8.40463877e-01 4.89768744e-01 -1.28067386e+00 1.32658720e+00
-6.85281515e-01 9.18114781e-01 9.28373277e-01 -8.10081780e-01
1.11615527e+00 3.60425800e-01 4.05472159e-01 -1.70564398e-01
-1.17862239e-01 5.41802824e-01 2.33486563e-01 -8.77819479e-01
1.58244312e-01 -3.51662576e-01 2.46450320e-01 5.99181414e-01
4.39602584e-02 -3.61118585e-01 -1.11090504e-01 -3.40515345e-01
1.00599599e+00 -5.43570258e-02 1.98797360e-01 -3.64694387e-01
1.72687665e-01 -1.54250339e-01 4.26272064e-01 6.46596193e-01
-1.92552924e-01 1.73932001e-01 -1.15531355e-01 -2.93518871e-01
-4.33945477e-01 -7.77889192e-01 -7.09866703e-01 1.71834111e+00
-7.66379952e-01 -1.79745361e-01 -3.26285392e-01 3.98723960e-01
5.06561659e-02 4.17695493e-01 -3.75946879e-01 3.24256688e-01
-4.45598453e-01 -1.18921793e+00 1.23115969e+00 1.64430410e-01
2.41109192e-01 -1.50934148e+00 -1.17157173e+00 5.36677599e-01
5.92041090e-02 -3.88766617e-01 2.26037800e-01 8.50312114e-01
-1.03394794e+00 -5.15919089e-01 -5.84626377e-01 -8.41026783e-01
2.45043725e-01 2.63277173e-01 6.11337960e-01 4.10115793e-02
-7.73066461e-01 4.99248177e-01 -3.95411134e-01 -1.07272851e+00
-4.11327571e-01 -9.47458819e-02 2.42194101e-01 -2.91635871e-01
4.00849760e-01 -9.51003909e-01 -3.15114558e-01 3.84484887e-01
-7.54266500e-01 -5.13379633e-01 2.55374074e-01 4.79103863e-01
3.94373715e-01 -1.60556391e-01 8.97499919e-01 -3.59039068e-01
4.79487032e-01 -6.77078664e-01 -8.03960800e-01 1.21109840e-02
-1.13986492e-01 -3.25262487e-01 5.90882823e-02 -2.23416716e-01
-6.23301387e-01 3.12256128e-01 -9.33407485e-01 6.45090163e-01
-5.06194651e-01 8.61786246e-01 -1.32890165e-01 -3.03766817e-01
6.00179791e-01 3.86384398e-01 1.78641155e-02 -9.28226471e-01
-1.86966538e-01 1.46902096e+00 4.20442373e-01 -3.75663728e-01
3.34114134e-01 3.48468572e-01 -6.75610006e-02 -1.83144999e+00
-2.43685152e-02 -8.36339235e-01 -4.48985368e-01 -6.25713646e-01
6.64447010e-01 -6.64154768e-01 -6.85341299e-01 9.82966661e-01
-7.94503391e-01 -2.54489005e-01 2.51023531e-01 5.91214597e-01
7.97100440e-02 1.63644955e-01 -9.99632850e-02 -1.54431832e+00
-6.38806522e-01 -9.13915694e-01 8.28801692e-01 2.75545061e-01
-4.63071279e-02 -8.04483533e-01 9.67361569e-01 2.10916579e-01
7.09929049e-01 2.46160720e-02 4.01262850e-01 -7.54498303e-01
1.43222734e-01 -2.95012087e-01 3.95934135e-01 3.12178999e-01
3.84219021e-01 3.21781993e-01 -1.63566279e+00 -1.30002722e-01
2.78326534e-02 -2.34064922e-01 6.65330112e-01 3.53834480e-01
8.63214254e-01 -2.63781041e-01 -1.45962924e-01 6.00563347e-01
1.13893533e+00 5.76408327e-01 1.87421098e-01 2.73982853e-01
1.50130346e-01 8.85257840e-01 3.96675318e-01 6.81390703e-01
2.01965913e-01 3.78956556e-01 3.41663748e-01 2.54110545e-01
3.95791620e-01 -8.40874314e-02 5.51557899e-01 8.82100821e-01
-1.31309450e-01 5.62198088e-02 -1.24758363e+00 7.74750888e-01
-1.25286257e+00 -1.01100409e+00 -3.40261906e-01 2.54037356e+00
8.74489069e-01 -5.51108539e-01 2.57848889e-01 4.04575318e-01
3.30191433e-01 -1.24447808e-01 -1.51258558e-01 -6.12810075e-01
1.53102517e-01 3.72720659e-01 4.35528636e-01 5.11768579e-01
-1.04706299e+00 5.26295304e-01 6.66911125e+00 3.75615537e-01
-1.30529702e+00 1.73873484e-01 -3.72243337e-02 -1.07536383e-01
1.43956244e-01 -2.85261482e-01 -6.80383027e-01 1.57379717e-01
1.49825907e+00 2.06918120e-01 3.72238666e-01 6.05150580e-01
7.80799866e-01 -4.62155551e-01 -7.47461200e-01 7.81175852e-01
-1.49391383e-01 -7.41329908e-01 -6.42721832e-01 -8.04478023e-03
-1.05638966e-01 5.93784034e-01 -2.86222041e-01 -9.04622078e-02
2.12050378e-01 -8.94663095e-01 4.86358017e-01 1.52369991e-01
5.85292578e-01 -5.93726873e-01 7.09194481e-01 2.28110850e-01
-1.28901911e+00 -1.03312721e-02 -6.38634145e-01 -3.78714025e-01
1.57112211e-01 4.26014096e-01 -8.05894494e-01 6.58100545e-02
1.45731378e+00 3.45478266e-01 -5.65931737e-01 1.79679346e+00
1.71166211e-01 1.36525345e+00 -1.11161411e+00 -5.35553753e-01
-2.09955052e-02 1.64503887e-01 6.59189999e-01 1.73541629e+00
4.65901047e-01 1.84667975e-01 -4.06397194e-01 2.01796532e-01
6.83962405e-01 3.44075233e-01 -4.78251874e-01 -1.51997358e-01
6.45146549e-01 1.40129542e+00 -8.35464358e-01 3.76479834e-01
-3.88138801e-01 8.83769840e-02 -2.24572822e-01 -2.04756737e-01
-3.45406502e-01 -6.72875464e-01 7.16903269e-01 -3.21372790e-04
-6.17221445e-02 -3.86424661e-01 -2.18545467e-01 -3.27256024e-01
-4.38028097e-01 -6.81489289e-01 6.58803880e-01 -5.83682775e-01
-1.18701482e+00 4.94386911e-01 3.75041544e-01 -9.81533349e-01
1.04240604e-01 -6.59257233e-01 -8.29327762e-01 8.99811685e-01
-1.23259735e+00 -1.05575144e+00 -2.54783869e-01 3.12187701e-01
5.20140886e-01 -1.19873583e-01 1.31901944e+00 5.29097497e-01
-3.84162754e-01 3.04552794e-01 5.72034180e-01 -6.28023073e-02
4.59475338e-01 -1.10829365e+00 1.44535825e-01 4.46454078e-01
1.71956986e-01 6.76806152e-01 9.67777133e-01 -6.91409945e-01
-1.28934324e+00 -7.06641912e-01 5.07508099e-01 6.08669966e-02
9.62537527e-01 -4.95066017e-01 -9.40479577e-01 3.92017424e-01
6.49307594e-02 -4.06777412e-01 1.69895077e+00 -4.05232683e-02
1.02187954e-01 -5.06260574e-01 -1.16910946e+00 5.44555187e-02
2.91574210e-01 -1.70354635e-01 -3.65682602e-01 2.18396083e-01
-2.31007580e-02 -3.27311270e-02 -7.64412045e-01 -7.59966075e-02
1.15526307e+00 -3.77164334e-01 7.11480200e-01 -2.57266939e-01
-3.74760419e-01 -4.71598059e-01 -2.06446141e-01 -1.36149657e+00
2.68725365e-01 -6.45561635e-01 7.54359424e-01 1.58716893e+00
6.02631152e-01 -8.85856092e-01 1.52406916e-01 2.18068734e-01
-3.15553725e-01 -3.44816782e-02 -1.63181317e+00 -4.73891288e-01
-3.74714471e-02 -8.55007291e-01 2.82632262e-01 2.62839586e-01
1.16717018e-01 -4.84773964e-02 -5.52267492e-01 5.52133977e-01
6.88441336e-01 -2.76898235e-01 7.64774978e-01 -1.48193669e+00
-4.10236299e-01 -5.60717136e-02 -6.46673918e-01 -6.15876734e-01
-5.30812979e-01 -5.65872490e-01 6.78146303e-01 -1.47381949e+00
-2.38267407e-01 -6.66841209e-01 -1.47176474e-01 9.05944169e-01
2.51151353e-01 2.92856276e-01 -2.86955953e-01 8.35704282e-02
2.42533654e-01 1.09702550e-01 5.62040389e-01 1.67110577e-01
-3.66464466e-01 5.38548708e-01 -4.59522605e-01 6.22264206e-01
7.38945127e-01 -1.05452549e+00 7.36220554e-02 -6.58096433e-01
1.80926427e-01 -1.03258669e-01 2.51081139e-01 -1.17406034e+00
4.04388845e-01 -1.94365427e-01 -5.06747253e-02 -6.65176809e-01
3.79773945e-01 -9.71156001e-01 4.31601644e-01 6.08890474e-01
-1.70199469e-01 1.70758635e-01 6.95081770e-01 3.68683845e-01
1.22167043e-01 -5.81092536e-01 9.58923876e-01 -2.60455728e-01
-4.50650930e-01 -2.93862551e-01 -1.16361237e+00 -4.93329167e-01
6.30225420e-01 -1.61315322e-01 -2.23409340e-01 -3.11483651e-01
-5.91690123e-01 2.09201172e-01 -1.68925792e-01 3.50895435e-01
6.91816807e-01 -4.60049868e-01 -9.54803109e-01 4.66304064e-01
2.31405590e-02 -2.07786530e-01 2.08752751e-01 7.02706575e-01
-1.32656848e+00 1.65162697e-01 -4.94674534e-01 -6.64103210e-01
-1.79712653e+00 -3.60195547e-01 4.53981817e-01 3.44832957e-01
-4.31232870e-01 1.31244326e+00 -2.97767788e-01 -6.21747851e-01
5.25377989e-01 -1.94568098e-01 -4.02746618e-01 1.64434642e-01
6.81740880e-01 6.60715520e-01 1.79054603e-01 -7.71531761e-01
-7.63418972e-01 3.55392873e-01 4.44957584e-01 -5.53206682e-01
2.06621432e+00 -3.05971894e-02 -3.66309285e-01 1.05353212e+00
8.84593427e-01 1.65680751e-01 -7.82408655e-01 3.29472333e-01
9.59968008e-03 -4.42953527e-01 2.91257977e-01 -1.06307316e+00
-8.72524917e-01 1.23412919e+00 1.26600492e+00 4.08248484e-01
1.09773386e+00 8.51944536e-02 1.67201459e-01 5.02677262e-01
3.65587026e-01 -8.97374809e-01 -5.55940628e-01 3.34017724e-01
1.10830748e+00 -1.00264931e+00 6.44970611e-02 3.60752106e-01
-3.32807183e-01 1.08696878e+00 2.82099128e-01 3.85730416e-01
1.04768419e+00 7.44301260e-01 6.50068402e-01 -2.38306105e-01
-5.75137973e-01 -7.87199363e-02 -2.43945792e-01 1.20376420e+00
7.40754187e-01 9.26428512e-02 -3.43976736e-01 6.76937401e-01
-5.83715856e-01 -4.27338928e-01 4.74846095e-01 1.28976417e+00
-9.45885837e-01 -1.10216248e+00 -8.49545360e-01 6.63105011e-01
-8.93037021e-01 -3.87212238e-03 -3.71249199e-01 2.53002644e-01
-9.11451429e-02 1.39037836e+00 -2.13435948e-01 -3.46525550e-01
1.13476485e-01 -4.95352931e-02 -1.66031629e-01 -7.92215347e-01
-9.77870286e-01 6.24848247e-01 1.12012617e-01 9.38725471e-02
-6.02426112e-01 -1.10821080e+00 -1.11625433e+00 -8.27438310e-02
-7.23955572e-01 7.64474332e-01 1.52394783e+00 4.82888371e-01
-7.17554241e-02 1.06017344e-01 5.02423942e-01 -1.06321418e+00
-1.83876917e-01 -1.43589485e+00 -9.34816003e-01 -5.83143890e-01
3.30264449e-01 -4.71187949e-01 -7.23438263e-01 -7.70758241e-02] | [8.59467601776123, -1.1181389093399048] |
d3e0de8f-aa75-43cf-aaf2-2b58473f7ce9 | custom-structure-preservation-in-face-aging | 2207.11025 | null | https://arxiv.org/abs/2207.11025v1 | https://arxiv.org/pdf/2207.11025v1.pdf | Custom Structure Preservation in Face Aging | In this work, we propose a novel architecture for face age editing that can produce structural modifications while maintaining relevant details present in the original image. We disentangle the style and content of the input image and propose a new decoder network that adopts a style-based strategy to combine the style and content representations of the input image while conditioning the output on the target age. We go beyond existing aging methods allowing users to adjust the degree of structure preservation in the input image during inference. To this purpose, we introduce a masking mechanism, the CUstom Structure Preservation module, that distinguishes relevant regions in the input image from those that should be discarded. CUSP requires no additional supervision. Finally, our quantitative and qualitative analysis which include a user study, show that our method outperforms prior art and demonstrates the effectiveness of our strategy regarding image editing and adjustable structure preservation. Code and pretrained models are available at https://github.com/guillermogotre/CUSP. | ['Óscar Cordón', 'Pablo Mesejo', 'Stéphane Lathuilière', 'Guillermo Gomez-Trenado'] | 2022-07-22 | null | null | null | null | ['face-age-editing'] | ['computer-vision'] | [ 3.40897232e-01 3.68354023e-01 -3.22244465e-02 -5.53140044e-01
-5.80160990e-02 -5.37484527e-01 5.95715702e-01 -1.49685696e-01
-5.04653037e-01 4.79423732e-01 3.49689066e-01 -1.06782377e-01
3.61150980e-01 -6.33674502e-01 -7.53673196e-01 -5.00953794e-01
3.84473562e-01 4.07512859e-03 4.02424671e-02 -9.31617469e-02
2.90599227e-01 4.95785117e-01 -1.66576517e+00 3.11996490e-01
9.16512907e-01 8.26333404e-01 2.49677420e-01 6.14249647e-01
2.15973571e-01 5.43213964e-01 -5.16439438e-01 -7.61698067e-01
2.81557858e-01 -4.86675709e-01 -5.99408805e-01 1.32161617e-01
1.17603457e+00 -5.70573986e-01 -4.65826899e-01 1.06695950e+00
7.29362667e-01 3.04521173e-02 5.60767293e-01 -1.00389028e+00
-1.11050832e+00 4.76908118e-01 -5.60696959e-01 2.16028735e-01
8.18409324e-02 2.97729552e-01 6.66787028e-01 -1.04868793e+00
7.11666346e-01 1.32383716e+00 4.69075412e-01 9.34120715e-01
-1.51362026e+00 -7.41854787e-01 3.98972571e-01 2.15132087e-01
-1.25147820e+00 -8.56504619e-01 8.24190676e-01 -3.70151430e-01
5.40811598e-01 1.84569791e-01 7.10689604e-01 1.36794806e+00
3.97009283e-01 4.24897820e-01 1.17263830e+00 -4.93396074e-01
1.31504729e-01 1.00028150e-01 -1.26747802e-01 9.55533087e-01
2.09635735e-01 2.19975442e-01 -6.83081865e-01 5.25477193e-02
7.96820521e-01 -2.37228587e-01 -3.00372452e-01 -3.57897401e-01
-8.05984557e-01 5.87842286e-01 3.28247577e-01 1.79629669e-01
-7.55960792e-02 1.22997560e-01 1.70403346e-01 3.92319322e-01
6.53732359e-01 3.29721689e-01 -2.93225199e-01 2.39389405e-01
-1.12912941e+00 1.36124045e-01 4.29612696e-01 8.96249950e-01
7.04070866e-01 5.68786636e-02 -6.33204520e-01 8.27879608e-01
2.67130852e-01 3.78351897e-01 2.06669778e-01 -1.27724266e+00
1.28351435e-01 5.90180755e-01 -2.21382275e-01 -8.86222959e-01
-6.33667856e-02 -2.47261733e-01 -7.84317076e-01 7.66476333e-01
3.41714710e-01 -4.73943166e-02 -1.28519094e+00 2.26536846e+00
2.49522835e-01 9.15982574e-02 -3.35750014e-01 7.27601826e-01
7.40069211e-01 1.66797310e-01 2.96937376e-01 7.88877718e-03
1.46418440e+00 -9.31331635e-01 -6.96390629e-01 -3.72352958e-01
1.01865917e-01 -7.42124736e-01 1.42331886e+00 2.71587372e-01
-1.49170554e+00 -6.09016836e-01 -1.22985828e+00 -5.56715429e-01
-2.81394243e-01 6.05750918e-01 3.84893119e-01 7.28488028e-01
-1.51721406e+00 8.04761231e-01 -7.52339900e-01 -3.58946174e-01
6.87944174e-01 4.72884268e-01 -4.37127829e-01 1.47422940e-01
-9.32228148e-01 9.28184628e-01 1.83066219e-01 5.80971800e-02
-9.07068670e-01 -8.44985127e-01 -8.85584593e-01 6.11889958e-02
3.11269790e-01 -1.03538370e+00 1.24931049e+00 -1.43456876e+00
-1.57166064e+00 1.26017773e+00 -3.53724599e-01 -1.42449766e-01
6.30878806e-01 -2.35408768e-01 -2.67357469e-01 2.07599595e-01
-1.85987771e-01 1.06224477e+00 1.42360115e+00 -1.25565386e+00
-2.91583657e-01 -4.36248183e-01 2.74305135e-01 1.63285196e-01
-4.70200539e-01 3.88991684e-02 -7.75027335e-01 -1.23869181e+00
-2.60039270e-01 -9.39263523e-01 5.40974829e-03 6.46294713e-01
-2.46252924e-01 3.61804277e-01 5.97145617e-01 -9.90529060e-01
1.42087197e+00 -2.28842568e+00 4.73899245e-01 1.63512453e-01
5.24676383e-01 1.32859200e-01 -3.11776489e-01 1.25689581e-01
-1.37660220e-01 2.55829483e-01 -3.93623769e-01 -8.18126917e-01
-1.94722533e-01 1.18892007e-02 -8.61796066e-02 2.77882427e-01
3.23385119e-01 7.77735770e-01 -5.68892896e-01 -5.24898350e-01
-7.41656497e-02 7.40648627e-01 -8.69976163e-01 1.62212104e-01
-1.60140708e-01 5.59681654e-01 7.61117637e-02 5.92860281e-01
8.50184679e-01 -4.11616452e-02 2.38250300e-01 -4.80407387e-01
4.47651483e-02 1.43117175e-01 -9.29606616e-01 1.86178970e+00
-3.17358553e-01 5.84017694e-01 3.49119246e-01 -2.12268710e-01
5.84504545e-01 1.99602172e-01 -1.86288301e-02 -7.55791903e-01
2.50547349e-01 -1.41868398e-01 -2.95417290e-02 -1.77633822e-01
6.17291987e-01 -9.75846034e-03 2.17432484e-01 5.64245164e-01
1.65624887e-01 6.97276965e-02 2.46447831e-01 4.44206268e-01
7.57977128e-01 2.75041938e-01 2.60560717e-02 -3.98080379e-01
5.29322982e-01 -7.24565625e-01 5.83123446e-01 4.78311509e-01
-2.02601582e-01 8.77074003e-01 6.40171409e-01 -1.80983797e-01
-1.23983371e+00 -1.12564826e+00 -8.68669059e-03 1.27179444e+00
-1.36256024e-01 -5.92337906e-01 -1.13281989e+00 -6.90906882e-01
-8.11956525e-02 5.74325502e-01 -1.16867685e+00 -4.95506555e-01
-5.87010145e-01 -3.25908124e-01 5.96720040e-01 4.93841946e-01
4.10909027e-01 -1.03796947e+00 -3.76681119e-01 -3.02988261e-01
-2.57174559e-02 -8.35943162e-01 -1.06191707e+00 -4.62258518e-01
-9.13306117e-01 -8.63357365e-01 -5.84015608e-01 -7.44324565e-01
1.23556006e+00 -1.58553004e-01 1.07240224e+00 4.63550508e-01
-2.85024583e-01 2.81946003e-01 -3.62994485e-02 -2.19803676e-01
-5.00301421e-01 2.10066006e-01 -1.44034121e-02 1.24894276e-01
-1.72491610e-01 -8.47761571e-01 -8.01381886e-01 1.11040451e-01
-1.00234878e+00 4.44928259e-01 5.75613499e-01 5.78195572e-01
3.89069736e-01 -2.87531585e-01 3.11317354e-01 -1.06366777e+00
5.84766448e-01 5.05762827e-03 -3.59709650e-01 2.87785202e-01
-8.84088457e-01 2.58844376e-01 3.49651605e-01 -5.43608606e-01
-1.30355835e+00 2.24130675e-02 -7.89509267e-02 -3.79377544e-01
1.04108684e-01 1.21481933e-01 -4.50759500e-01 -7.67662376e-02
6.26766324e-01 -5.17655984e-02 2.37131506e-01 -5.91690183e-01
5.46214461e-01 2.61209726e-01 5.97578824e-01 -6.31491303e-01
7.79362440e-01 6.48315966e-01 -3.08766514e-01 -3.83599162e-01
-7.09642708e-01 3.70089531e-01 -8.60036850e-01 -2.81956941e-01
6.72299445e-01 -9.73689914e-01 -4.84430254e-01 6.95789337e-01
-1.04119349e+00 -4.81288224e-01 -2.35113800e-01 -4.82471958e-02
-3.26519340e-01 4.00807351e-01 -7.36561298e-01 -4.29818034e-01
-4.63372111e-01 -1.00761080e+00 9.36452687e-01 3.26568067e-01
-4.41415966e-01 -8.62883687e-01 -6.91200793e-02 2.89336175e-01
4.58666444e-01 8.82348642e-02 1.00702417e+00 6.53952211e-02
-4.70593423e-01 2.18214884e-01 -2.53266543e-01 3.68429631e-01
1.88396648e-01 2.94786662e-01 -1.15178943e+00 -6.74886405e-01
-1.97376743e-01 -8.60143974e-02 1.09476745e+00 2.11024255e-01
1.39043081e+00 -5.29086471e-01 -1.31134242e-02 8.58605981e-01
1.11809254e+00 -1.40318379e-01 1.02171457e+00 1.06176071e-01
7.88045406e-01 5.48859835e-01 1.21742427e-01 3.66330177e-01
2.41784349e-01 5.33401906e-01 1.86341196e-01 -3.33503455e-01
-5.63842475e-01 -4.36436892e-01 5.70121050e-01 4.48545218e-01
-1.67617127e-01 -6.23034388e-02 -3.72529417e-01 4.58034605e-01
-1.50118315e+00 -9.74965572e-01 3.37899625e-01 2.18370056e+00
1.32268977e+00 1.65226400e-01 3.63262333e-02 -1.14313543e-01
7.13810861e-01 2.64813662e-01 -6.91589415e-01 -5.46942770e-01
-1.50364503e-01 5.36580622e-01 3.55331808e-01 7.98069119e-01
-8.70458186e-01 1.04246616e+00 6.46540737e+00 5.76204300e-01
-1.18176019e+00 1.10452443e-01 7.57702768e-01 -3.99702191e-01
-5.29539764e-01 -7.31709376e-02 -6.87159836e-01 4.95565951e-01
7.36540794e-01 -7.26963207e-02 6.39178693e-01 4.42638248e-01
4.34980899e-01 -6.41503334e-02 -1.17493081e+00 6.54155314e-01
2.88675010e-01 -1.17387617e+00 3.17481369e-01 6.14105724e-02
5.37030518e-01 -5.82246959e-01 5.23451924e-01 3.79271992e-03
7.24525675e-02 -1.06171608e+00 1.05464733e+00 6.79230690e-01
1.20907199e+00 -5.36189437e-01 7.15639591e-02 -2.21959040e-01
-9.40322578e-01 -7.18079284e-02 -1.20509766e-01 3.92149314e-02
-1.48610458e-01 5.19150496e-01 -2.98286557e-01 1.69588163e-01
7.11032033e-01 6.16983891e-01 -1.25063825e+00 6.15621150e-01
-5.97253978e-01 3.97906721e-01 4.49120626e-02 6.85451090e-01
-4.63972867e-01 -1.73905939e-01 3.78748119e-01 1.04675686e+00
3.06767613e-01 -2.77909860e-02 -3.78745258e-01 1.06292295e+00
-4.58118767e-01 -2.98827142e-02 -4.27298963e-01 2.54347380e-02
5.03151774e-01 1.26054943e+00 -6.00359797e-01 -3.03940237e-01
-2.74473637e-01 1.43975866e+00 6.14585519e-01 4.06045645e-01
-8.75816286e-01 -2.27305114e-01 6.97275043e-01 2.95078605e-01
4.51931477e-01 -1.96266338e-01 -5.91961920e-01 -1.22158670e+00
1.86814189e-01 -1.13269627e+00 3.12838644e-01 -8.75845551e-01
-9.78420079e-01 5.99037886e-01 -3.85820642e-02 -8.95302713e-01
1.78837121e-01 -5.10894418e-01 -7.58813918e-01 8.59115303e-01
-1.13522136e+00 -1.40887296e+00 -2.91232675e-01 4.45964396e-01
3.24667811e-01 -5.52787781e-02 6.11468256e-01 4.15053368e-01
-8.39984596e-01 1.11073220e+00 -3.38793993e-01 -1.71987852e-03
1.08763349e+00 -1.13024890e+00 4.99648422e-01 1.09493577e+00
-1.55150846e-01 9.09290075e-01 7.25728691e-01 -8.49697948e-01
-9.38635230e-01 -9.86028016e-01 9.10910130e-01 -4.84759241e-01
3.43922079e-01 -5.99840462e-01 -1.06092536e+00 7.83495009e-01
5.96973121e-01 -3.92802835e-01 5.60022891e-01 -6.52925149e-02
-7.30169356e-01 -1.76866114e-01 -1.23536408e+00 1.07212949e+00
1.27967048e+00 -6.38144374e-01 -4.53320146e-01 -2.80533105e-01
6.55410588e-01 -4.30288225e-01 -6.63275957e-01 3.51821572e-01
7.89148510e-01 -1.00810874e+00 8.98401320e-01 -3.77182007e-01
5.34158051e-01 -4.11620945e-01 2.17267394e-01 -1.28269291e+00
-5.71782887e-01 -6.21171594e-01 -2.67535567e-01 1.37966967e+00
5.64466298e-01 -4.47954297e-01 7.79033780e-01 8.21203172e-01
-7.79130170e-03 -5.83442211e-01 -7.34639168e-01 -3.74856651e-01
2.17049524e-01 -2.42888369e-02 5.55265605e-01 6.90830171e-01
-3.18229556e-01 3.12245309e-01 -4.63482201e-01 1.65224999e-01
5.88809788e-01 -1.93471387e-01 4.43101257e-01 -9.98015344e-01
-3.30501527e-01 -5.28578639e-01 -6.65648580e-02 -5.94572663e-01
3.36747229e-01 -7.57907987e-01 -2.13866130e-01 -1.17287159e+00
3.18224460e-01 -9.04444605e-02 -2.27093935e-01 8.05751860e-01
-3.54276925e-01 5.13954103e-01 3.32327574e-01 1.05102874e-01
-1.17086753e-01 6.13559842e-01 1.39593351e+00 -1.38298348e-01
-1.16313070e-01 -2.65254974e-01 -1.07384312e+00 5.65515459e-01
1.08046293e+00 -2.99304694e-01 -5.54327428e-01 -7.22992659e-01
1.88339651e-01 -5.57118952e-01 4.24215317e-01 -9.00246024e-01
-2.91415043e-02 -6.57471865e-02 6.95265412e-01 -1.06314048e-01
4.27380770e-01 -6.03017867e-01 1.97400123e-01 4.70686823e-01
-5.62163055e-01 2.83773422e-01 4.55070347e-01 3.42201084e-01
2.76603878e-01 -1.43188968e-01 1.16798341e+00 9.24625695e-02
-4.78545636e-01 4.42386091e-01 -3.79560351e-01 -9.58831385e-02
7.18996882e-01 -2.20705181e-01 -3.46901536e-01 -4.13010597e-01
-9.76902604e-01 7.54991844e-02 1.10076725e+00 5.35639703e-01
6.37253463e-01 -1.38864696e+00 -6.44162059e-01 4.63117629e-01
-1.20466743e-02 -4.99094367e-01 3.62103462e-01 7.07529962e-01
-3.41975451e-01 -2.71046668e-01 -4.62812990e-01 -6.15005195e-02
-1.64174628e+00 5.89552283e-01 3.88570100e-01 8.29553977e-02
-5.23955762e-01 7.38491714e-01 3.53425354e-01 -1.96628124e-01
2.33530462e-01 -3.50370347e-01 -6.97639957e-02 8.04181024e-02
6.43272519e-01 1.63109943e-01 2.90563721e-02 -5.43548703e-01
-9.63515639e-02 5.08023620e-01 -4.19866174e-01 -3.52566600e-01
1.18040442e+00 -3.71299773e-01 -3.46627653e-01 2.35463768e-01
8.83831322e-01 2.34264269e-01 -1.62546110e+00 -9.63752270e-02
-2.87864059e-01 -6.25665188e-01 1.10089935e-01 -9.76055622e-01
-1.39523780e+00 7.75413156e-01 8.53987098e-01 -3.68319303e-01
1.43547869e+00 -1.10626355e-01 5.11987388e-01 -1.01221919e-01
-1.18861251e-01 -1.16240990e+00 1.86836913e-01 3.59384537e-01
1.18058598e+00 -8.70677769e-01 2.10677028e-01 -4.21878457e-01
-5.61282635e-01 7.74586499e-01 9.38036740e-01 1.20987501e-02
6.11654937e-01 3.17788213e-01 1.04446441e-01 -1.30811617e-01
-8.87774825e-01 -5.97373582e-02 4.33143348e-01 4.99712825e-01
5.27960658e-01 -2.53829062e-01 -4.69666302e-01 4.02454823e-01
-3.42012376e-01 9.18738320e-02 4.67191130e-01 8.94301832e-01
-1.84216931e-01 -1.42700160e+00 -2.97434449e-01 3.75401199e-01
-5.13464749e-01 -3.51503342e-01 -6.51152849e-01 3.46250564e-01
3.24618816e-01 5.74989378e-01 7.12535530e-02 -3.61700207e-01
4.04631227e-01 2.12177575e-01 9.65713441e-01 -6.27364397e-01
-7.30233192e-01 -4.21219766e-02 9.65860114e-02 -5.31233430e-01
-3.14398170e-01 -4.91358370e-01 -9.15586114e-01 -5.20576417e-01
-1.90443899e-02 -2.89239556e-01 3.90651911e-01 5.49367547e-01
7.19773650e-01 4.42469001e-01 5.09224474e-01 -9.41928625e-01
-6.97720945e-02 -8.73291552e-01 -3.87710780e-01 4.99026984e-01
3.97386760e-01 -6.07684314e-01 -3.52070183e-01 6.02232575e-01] | [12.462071418762207, -0.27507761120796204] |
514b8faf-d8ef-488f-b0c1-68ced2c46043 | marked-personas-using-natural-language | 2305.18189 | null | https://arxiv.org/abs/2305.18189v1 | https://arxiv.org/pdf/2305.18189v1.pdf | Marked Personas: Using Natural Language Prompts to Measure Stereotypes in Language Models | To recognize and mitigate harms from large language models (LLMs), we need to understand the prevalence and nuances of stereotypes in LLM outputs. Toward this end, we present Marked Personas, a prompt-based method to measure stereotypes in LLMs for intersectional demographic groups without any lexicon or data labeling. Grounded in the sociolinguistic concept of markedness (which characterizes explicitly linguistically marked categories versus unmarked defaults), our proposed method is twofold: 1) prompting an LLM to generate personas, i.e., natural language descriptions, of the target demographic group alongside personas of unmarked, default groups; 2) identifying the words that significantly distinguish personas of the target group from corresponding unmarked ones. We find that the portrayals generated by GPT-3.5 and GPT-4 contain higher rates of racial stereotypes than human-written portrayals using the same prompts. The words distinguishing personas of marked (non-white, non-male) groups reflect patterns of othering and exoticizing these demographics. An intersectional lens further reveals tropes that dominate portrayals of marginalized groups, such as tropicalism and the hypersexualization of minoritized women. These representational harms have concerning implications for downstream applications like story generation. | ['Dan Jurafsky', 'Esin Durmus', 'Myra Cheng'] | 2023-05-29 | null | null | null | null | ['story-generation'] | ['natural-language-processing'] | [ 2.94950843e-01 7.20315516e-01 -3.30214262e-01 -4.13216025e-01
-3.57033581e-01 -8.21089864e-01 1.39208937e+00 6.15734458e-01
-2.24469885e-01 4.27160233e-01 1.40874100e+00 -2.95062274e-01
1.46510586e-01 -9.79928732e-01 -3.55837971e-01 -3.53393912e-01
1.23821728e-01 5.22851110e-01 -5.85615754e-01 -5.68554759e-01
5.38910806e-01 2.50887066e-01 -1.37011039e+00 6.57883227e-01
7.03066349e-01 4.56118017e-01 -4.85768944e-01 2.06142321e-01
-6.13090694e-01 9.32356954e-01 -8.68862510e-01 -8.38800132e-01
1.01009741e-01 -4.22916770e-01 -5.60197771e-01 -2.11053953e-01
1.22279441e+00 -1.22872762e-01 -1.55144647e-01 1.17427444e+00
6.68347061e-01 -4.24285904e-02 1.09640729e+00 -1.14982891e+00
-1.07636511e+00 1.41386282e+00 -7.28145897e-01 1.07191630e-01
6.81314647e-01 2.18826771e-01 7.02147543e-01 -7.72957504e-01
8.25532734e-01 2.42098999e+00 8.88581812e-01 7.03801155e-01
-1.77274036e+00 -1.06722569e+00 1.34520650e-01 -7.35425770e-01
-1.67181373e+00 -7.72616088e-01 6.06211960e-01 -1.06308603e+00
4.84210730e-01 6.67427480e-01 5.04029274e-01 1.70020449e+00
2.98983883e-02 7.53052756e-02 1.18143320e+00 -1.75514668e-01
1.00967169e-01 4.07296717e-01 1.91927716e-01 5.94888091e-01
4.29035306e-01 1.94320202e-01 -7.83835769e-01 -7.34884262e-01
4.92978126e-01 -3.01483989e-01 2.52436191e-01 2.47187197e-01
-1.13115883e+00 1.26896238e+00 1.10555641e-01 3.72260273e-01
-2.59205937e-01 1.36455774e-01 6.23703897e-01 -2.29599047e-02
1.07403278e+00 6.34499311e-01 1.52314588e-01 -7.58997872e-02
-9.53218937e-01 6.63537025e-01 8.04166734e-01 8.54728699e-01
5.73585927e-01 1.17377669e-01 -7.35111475e-01 1.01821911e+00
9.43360329e-02 7.03196347e-01 2.16974363e-01 -8.77332389e-01
4.73952174e-01 6.36482477e-01 1.45306408e-01 -1.73796415e+00
-5.29992104e-01 1.49150249e-02 -8.45938444e-01 -6.41632974e-02
5.26760638e-01 -3.86982799e-01 -5.94207764e-01 2.24528980e+00
1.61771566e-01 -3.77806872e-01 -4.58747633e-02 6.23578846e-01
1.01202500e+00 6.98976219e-01 9.92680073e-01 -5.59948348e-02
1.45426071e+00 3.29458714e-02 -6.52713954e-01 -4.70099479e-01
7.77168989e-01 -7.96262145e-01 1.27958858e+00 -3.20268571e-01
-1.00860941e+00 -4.05884087e-01 -5.45984685e-01 -1.23163246e-01
-5.42910457e-01 -4.50346380e-01 5.73155582e-01 1.14599037e+00
-9.77697670e-01 5.00930369e-01 5.70182018e-02 -8.39458168e-01
5.75772643e-01 -3.58262379e-03 -1.36552691e-01 3.87362242e-01
-1.22720671e+00 9.76698756e-01 1.17162116e-01 -2.71016926e-01
-6.58733845e-01 -1.10699260e+00 -1.18208551e+00 -2.96325713e-01
2.40371644e-01 -4.90076452e-01 8.07099640e-01 -1.27283382e+00
-7.59551823e-01 1.65171528e+00 -3.66550386e-02 -2.34592408e-01
5.00292063e-01 -2.55728751e-01 -7.14685500e-01 -3.14510942e-01
5.91617048e-01 9.71606314e-01 5.90402007e-01 -1.50048220e+00
-3.91329437e-01 -3.00214469e-01 2.69073933e-01 2.14238033e-01
-4.91490930e-01 7.71444261e-01 7.54481375e-01 -8.60332072e-01
-1.81139544e-01 -5.30270100e-01 -1.77528441e-01 -1.43780604e-01
-1.07254303e+00 -1.74142811e-02 4.96333480e-01 -7.00055063e-01
1.56566095e+00 -2.22791576e+00 -2.54055828e-01 2.99002200e-01
6.37857914e-01 4.71620895e-02 -3.32151819e-03 9.18709040e-01
-1.17896542e-01 8.35938096e-01 1.70024350e-01 -1.26115188e-01
4.04708326e-01 -1.80884406e-01 -6.06285810e-01 5.13888061e-01
5.19019701e-02 6.71081901e-01 -1.06162560e+00 -6.32452428e-01
1.25776559e-01 2.07388744e-01 -7.48047411e-01 -2.39544719e-01
-9.80387554e-02 3.83103311e-01 3.42855789e-03 3.34393412e-01
6.24306023e-01 5.46672583e-01 4.26539965e-02 -2.24723190e-01
-6.65080786e-01 4.34997588e-01 -7.74805903e-01 9.82465982e-01
-1.69240877e-01 5.93855500e-01 6.61855340e-02 -1.94076300e-01
1.21622169e+00 -6.56007752e-02 -1.69230439e-02 -5.36579967e-01
2.38706768e-01 3.86319458e-02 1.23338066e-01 -4.17020470e-01
9.33386564e-01 -7.03230917e-01 -8.73508751e-01 6.10653579e-01
-3.93207937e-01 -2.97696412e-01 -2.53802128e-02 3.73552561e-01
6.06491506e-01 -2.30347618e-01 4.63019162e-01 -1.12830269e+00
1.35026544e-01 -2.22610645e-02 5.99344671e-01 9.53531623e-01
-1.05676889e-01 3.51638347e-01 8.72943521e-01 -6.11794949e-01
-1.15091503e+00 -1.38202286e+00 -7.16682598e-02 1.49253869e+00
-2.03521196e-02 -6.06054664e-01 -7.91110575e-01 -3.14134181e-01
2.55077839e-01 1.40842819e+00 -7.79594898e-01 -3.85795742e-01
-5.02504051e-01 -5.12206674e-01 1.16330647e+00 -6.69846609e-02
7.98671842e-02 -9.72019374e-01 -5.81915081e-01 -1.12690777e-01
-1.22749850e-01 -7.46439397e-01 -5.07451117e-01 -5.17137051e-01
-2.27743939e-01 -6.73760712e-01 -5.08215666e-01 -6.80698574e-01
7.26254046e-01 -2.40537420e-01 1.31746340e+00 -5.07359067e-03
1.15085982e-01 8.43587965e-02 -6.47386536e-02 -6.40672386e-01
-1.04194880e+00 -2.06466496e-01 2.99961895e-01 1.11335240e-01
4.84373182e-01 -4.87948179e-01 -1.94149181e-01 -6.70123706e-03
-8.76759946e-01 5.04665494e-01 -1.15207292e-01 2.72544976e-02
-1.84754148e-01 -5.87336421e-01 4.02334064e-01 -1.58801079e+00
9.77788866e-01 -8.00000131e-01 5.97735271e-02 -1.00345619e-01
-1.59188703e-01 -1.49359897e-01 6.10561252e-01 -7.90522754e-01
-1.04959142e+00 -6.53845727e-01 6.91595227e-02 3.09220016e-01
-3.66727322e-01 3.00728142e-01 -2.07930982e-01 5.73384881e-01
1.11539125e+00 -3.10564935e-01 -1.79083854e-01 -4.33613241e-01
6.90868437e-01 8.42207193e-01 6.94534123e-01 -9.49241102e-01
9.15220439e-01 5.52389383e-01 -3.94757688e-01 -1.06338513e+00
-6.67733550e-01 5.55376448e-02 -3.29039007e-01 -4.54406291e-01
8.50201130e-01 -8.56942534e-01 -7.09077775e-01 5.05374849e-01
-1.16310239e+00 -4.63610470e-01 -4.00023967e-01 -3.59350517e-02
-3.10995042e-01 9.85061079e-02 -6.20480001e-01 -9.61897731e-01
-3.95687521e-01 -4.81885940e-01 1.04943752e+00 1.51828766e-01
-1.44457781e+00 -1.13818288e+00 -8.62281471e-02 -3.58087867e-02
4.18782979e-01 8.87063324e-01 1.58647633e+00 -9.30526316e-01
5.03186464e-01 1.67596206e-01 -1.85045183e-01 -3.12963963e-01
2.46556789e-01 1.49402440e-01 -8.38604331e-01 7.66076073e-02
-2.34169841e-01 -3.86181474e-01 2.26513892e-01 1.68727264e-01
5.70726275e-01 -1.04169261e+00 -4.00071651e-01 3.65985781e-01
1.12651968e+00 -1.93684306e-02 4.50767666e-01 1.09155335e-01
7.85661876e-01 1.25512195e+00 2.56584197e-01 6.22619629e-01
4.56004024e-01 2.75241852e-01 -1.26595959e-01 -2.00251222e-01
-1.19291566e-01 -1.05314708e+00 7.33666539e-01 3.72612804e-01
4.05506968e-01 -1.19638517e-01 -1.23813117e+00 5.25955439e-01
-1.37170255e+00 -1.21436048e+00 -4.96265739e-01 1.97162449e+00
9.25622523e-01 1.55425966e-01 2.02191904e-01 -2.66709000e-01
1.19053757e+00 8.56060743e-01 -1.60015061e-01 -1.12096918e+00
-4.03771579e-01 -5.47769107e-02 3.75470757e-01 8.10082853e-01
-9.08555329e-01 1.36311257e+00 6.61009645e+00 9.50580895e-01
-8.65392625e-01 5.86396009e-02 9.56810474e-01 -1.89601690e-01
-9.58122730e-01 3.52507643e-02 -7.04786539e-01 5.17563462e-01
9.23385203e-01 -6.52766287e-01 3.38600516e-01 3.44316155e-01
6.68583930e-01 -1.04717679e-01 -1.36512959e+00 9.18620646e-01
2.32763693e-01 -1.27794516e+00 4.29552555e-01 5.75260110e-02
8.21261942e-01 -6.10197186e-01 3.46268654e-01 3.50014716e-01
7.09185541e-01 -1.22060943e+00 1.52417564e+00 3.92024606e-01
1.15256369e+00 -6.34703934e-01 1.92811444e-01 2.38750905e-01
-6.72131777e-01 -1.19547963e-01 -3.81992728e-01 -7.54864573e-01
2.67217666e-01 5.21903932e-01 -5.66236019e-01 -1.91651896e-01
1.85843453e-01 3.88669938e-01 -6.91169560e-01 1.80283099e-01
1.55629739e-02 6.17716134e-01 -1.68778956e-01 1.28504440e-01
2.81473905e-01 -2.42436126e-01 7.44642556e-01 1.75277448e+00
2.26754606e-01 4.08400036e-03 8.89209881e-02 1.31162763e+00
4.85361218e-02 1.78345636e-01 -1.17091966e+00 -3.59529734e-01
8.47447515e-01 1.09734046e+00 -5.96710265e-01 -4.24949139e-01
-1.77892804e-01 5.79307675e-01 1.66579098e-01 3.49173009e-01
-6.91372871e-01 1.06540754e-01 1.00316238e+00 6.01884782e-01
-8.25406075e-01 5.69322295e-02 -6.87334716e-01 -6.01107836e-01
-8.27097654e-01 -1.17945850e+00 3.36418301e-01 -6.87239885e-01
-1.59386027e+00 1.67053953e-01 4.41524059e-01 -5.03131509e-01
-4.18408364e-01 -1.86486781e-01 -7.04731643e-01 1.01312530e+00
-2.19032645e-01 -1.29777265e+00 -7.62741342e-02 1.54165566e-01
1.78052008e-01 2.53040553e-03 9.50698435e-01 1.23890772e-01
-4.49976295e-01 7.44154871e-01 -4.38600391e-01 3.30541968e-01
7.06461847e-01 -9.50478494e-01 8.05776596e-01 6.22797608e-01
-2.87595958e-01 8.14402044e-01 1.18444967e+00 -1.13369083e+00
-7.42293715e-01 -1.00352550e+00 1.50118160e+00 -5.57018638e-01
9.02031064e-01 -8.78484607e-01 -4.37636316e-01 8.30900311e-01
1.43994406e-01 -8.42043340e-01 1.03877318e+00 3.84342641e-01
-6.59578204e-01 3.95883977e-01 -1.43324089e+00 1.36695790e+00
1.67357969e+00 -8.14921916e-01 -7.09657550e-01 2.89779156e-01
7.33307481e-01 -1.30757228e-01 -6.33529365e-01 -5.21896631e-02
5.96922338e-01 -8.68891656e-01 8.54002178e-01 -8.06586802e-01
8.21420670e-01 3.66567969e-01 -2.56658047e-01 -1.37541151e+00
-3.74956995e-01 -8.93031895e-01 5.34390569e-01 2.12405276e+00
3.08402687e-01 -6.41007006e-01 2.80893534e-01 1.15382361e+00
-9.58695039e-02 -4.93894033e-02 -5.85114777e-01 -3.86166304e-01
6.91665947e-01 -2.88607925e-01 7.65428603e-01 1.46134818e+00
1.35063931e-01 4.48579013e-01 -4.27354902e-01 -1.48699597e-01
6.16144180e-01 -2.02345937e-01 6.79634392e-01 -1.30326343e+00
2.96671242e-01 -7.68757463e-01 -2.47773439e-01 -3.18706065e-01
2.79054582e-01 -1.09274495e+00 -3.02743703e-01 -1.18753970e+00
5.06509006e-01 -4.75652218e-01 5.11893630e-01 2.45773569e-01
4.28077728e-02 3.02871197e-01 5.40214479e-01 -1.80163942e-02
-7.21100792e-02 1.20667115e-01 9.46230054e-01 -2.91084088e-02
-2.75060266e-01 -6.61085963e-01 -1.51129174e+00 1.29370379e+00
7.12914824e-01 -6.77478909e-01 -9.64085683e-02 -3.76607090e-01
6.01849973e-01 -3.22459430e-01 5.54397643e-01 -6.96646869e-01
-2.42460117e-01 -6.79826021e-01 2.81481147e-01 -1.33447245e-01
2.67867237e-01 -3.47224236e-01 5.21192491e-01 5.82815766e-01
-8.94528329e-01 3.55660766e-02 1.91712186e-01 -1.16046205e-01
2.28998408e-01 -1.83270827e-01 7.10026503e-01 -1.97936952e-01
-3.94970745e-01 -8.30170214e-02 -8.70679438e-01 7.60402858e-01
7.17065752e-01 -2.29880676e-01 -8.76162410e-01 -6.52131915e-01
-4.38141733e-01 -1.44108400e-01 8.05518210e-01 6.12700284e-01
2.26990238e-01 -1.50337553e+00 -1.14535999e+00 4.37796190e-02
2.53500819e-01 -5.67796588e-01 3.79025005e-03 3.85468394e-01
-3.55585933e-01 -5.75101934e-02 -3.24289471e-01 5.33946790e-02
-9.23883140e-01 4.86010969e-01 1.32749021e-01 2.50516266e-01
-3.60137433e-01 7.28647768e-01 9.01347637e-01 -6.08774900e-01
-9.77321118e-02 -4.66720723e-02 -2.84775913e-01 6.16764665e-01
5.34499705e-01 5.96571684e-01 -8.12095761e-01 -1.38143885e+00
-2.81677812e-01 2.12076515e-01 1.90499857e-01 -3.29581618e-01
9.50081766e-01 -2.58569181e-01 -3.18505347e-01 7.39721060e-01
1.16235030e+00 6.41104043e-01 -6.89235508e-01 9.54928026e-02
3.44385684e-01 -6.02097511e-01 -4.66055065e-01 -8.75242770e-01
-3.91729414e-01 5.86806297e-01 1.21113151e-01 3.40673953e-01
3.51357847e-01 2.85045624e-01 5.41980386e-01 -3.53718638e-01
3.40370715e-01 -1.12382472e+00 -2.65582383e-01 4.61589783e-01
1.31646979e+00 -6.60537541e-01 -2.38760054e-01 -5.62723100e-01
-8.47081721e-01 5.39891720e-01 6.26599491e-01 8.95128772e-02
2.24488437e-01 2.70711482e-01 4.27180976e-01 -4.65110183e-01
-4.34354246e-01 -2.67892648e-02 5.19723333e-02 9.14526105e-01
7.25324333e-01 6.33956254e-01 -8.22162688e-01 9.13759112e-01
-1.00633693e+00 -7.18072057e-01 5.20218313e-01 3.16680402e-01
-2.68496990e-01 -4.94940490e-01 -8.19676280e-01 5.64584970e-01
-7.54026353e-01 -5.22924125e-01 -1.10260057e+00 9.64612424e-01
6.53148770e-01 1.19693315e+00 5.45497835e-01 -6.15706801e-01
1.75754204e-01 1.58546448e-01 1.36426255e-01 -9.65533197e-01
-1.24376523e+00 -2.69011021e-01 8.59522462e-01 -3.75096262e-01
-6.91679642e-02 -6.36057198e-01 -1.08643162e+00 -1.12539208e+00
4.63391513e-01 -1.82403505e-01 1.44600987e-01 6.20236754e-01
7.37816617e-02 -1.93304852e-01 2.32114136e-01 -8.66454899e-01
-3.64302993e-01 -1.08915949e+00 -6.22628391e-01 1.04782939e+00
-4.00393344e-02 -5.29329896e-01 -4.17351454e-01 -2.12002099e-01] | [9.252728462219238, 10.214604377746582] |
156472c7-c642-4ce3-97ec-e1dde0df2465 | unis-mmc-multimodal-classification-via | 2305.09299 | null | https://arxiv.org/abs/2305.09299v1 | https://arxiv.org/pdf/2305.09299v1.pdf | UniS-MMC: Multimodal Classification via Unimodality-supervised Multimodal Contrastive Learning | Multimodal learning aims to imitate human beings to acquire complementary information from multiple modalities for various downstream tasks. However, traditional aggregation-based multimodal fusion methods ignore the inter-modality relationship, treat each modality equally, suffer sensor noise, and thus reduce multimodal learning performance. In this work, we propose a novel multimodal contrastive method to explore more reliable multimodal representations under the weak supervision of unimodal predicting. Specifically, we first capture task-related unimodal representations and the unimodal predictions from the introduced unimodal predicting task. Then the unimodal representations are aligned with the more effective one by the designed multimodal contrastive method under the supervision of the unimodal predictions. Experimental results with fused features on two image-text classification benchmarks UPMC-Food-101 and N24News show that our proposed Unimodality-Supervised MultiModal Contrastive UniS-MMC learning method outperforms current state-of-the-art multimodal methods. The detailed ablation study and analysis further demonstrate the advantage of our proposed method. | ['Eng Siong Chng', 'Deepu Rajan', 'Yuchen Hu', 'Chen Chen', 'Meng Shen', 'Heqing Zou'] | 2023-05-16 | null | null | null | null | ['image-text-classification'] | ['miscellaneous'] | [ 5.57159901e-01 1.02695443e-01 -3.87629569e-01 -3.60601097e-01
-1.29779780e+00 -3.23484659e-01 8.76437366e-01 1.59281462e-01
-1.96513444e-01 8.03485453e-01 4.63359773e-01 1.73643142e-01
-9.26521420e-02 -2.31143862e-01 -7.80071676e-01 -1.12481570e+00
3.33473176e-01 3.28265935e-01 -3.29534858e-01 -2.55794197e-01
1.40644968e-01 -2.73438215e-01 -1.66748106e+00 8.60794961e-01
8.98561001e-01 1.24872983e+00 2.13571101e-01 4.93809044e-01
-4.80133072e-02 8.07372034e-01 -1.99209854e-01 -4.53965753e-01
-1.81693941e-01 -4.68674123e-01 -6.84071541e-01 1.07458219e-01
5.02010167e-01 -4.34022509e-02 -2.11313501e-01 9.45990860e-01
7.80811548e-01 1.55257851e-01 9.47970092e-01 -1.50280035e+00
-5.85929394e-01 8.54714394e-01 -8.16694438e-01 -3.92009437e-01
6.77534282e-01 -4.13958244e-02 9.57684517e-01 -8.36623549e-01
4.38905597e-01 1.42567921e+00 1.60468906e-01 6.72628760e-01
-1.27728271e+00 -6.30508661e-01 4.28183913e-01 7.01392069e-02
-1.02942979e+00 -4.71068770e-01 8.99262130e-01 -1.61018670e-01
6.35599673e-01 2.13051319e-01 7.28348345e-02 1.54883540e+00
1.39434904e-01 1.39062381e+00 1.19966042e+00 -4.07178283e-01
-8.93640444e-02 2.29659915e-01 2.02602252e-01 7.56002724e-01
-1.30777135e-01 -1.52667567e-01 -9.65080142e-01 -1.39287800e-01
1.22883618e-01 2.95479536e-01 -4.30194616e-01 -2.48715892e-01
-1.87227786e+00 5.12082636e-01 3.23830903e-01 2.16685608e-01
-4.29157346e-01 1.28297389e-01 4.74628747e-01 2.77254254e-01
7.39057958e-02 3.95762883e-02 -4.14808512e-01 4.67030443e-02
-5.11185586e-01 -1.88980296e-01 4.36099231e-01 9.41900909e-01
9.54279661e-01 -2.36902922e-01 -2.58551389e-01 9.12848532e-01
7.51557529e-01 9.32932079e-01 6.91682398e-01 -8.59030962e-01
9.43337381e-01 7.52946794e-01 -2.19952330e-01 -8.90062034e-01
-5.24888277e-01 1.30303532e-01 -1.09392238e+00 -2.19332874e-01
1.43405825e-01 -2.93460339e-01 -8.58696759e-01 1.73967350e+00
3.79305743e-02 -1.19973883e-01 6.22909188e-01 9.46017802e-01
1.24568415e+00 7.88119972e-01 4.71593320e-01 -1.58818230e-01
1.16069126e+00 -1.15510976e+00 -8.14736068e-01 -1.90206468e-01
6.04776621e-01 -8.41819525e-01 8.02128553e-01 4.32559073e-01
-9.17321861e-01 -6.14345312e-01 -1.05938315e+00 1.80455074e-01
-5.60351670e-01 4.42313582e-01 5.75124919e-01 4.12205607e-01
-4.86369133e-01 1.73765495e-01 -3.76837879e-01 -2.77441651e-01
3.10532123e-01 4.57079470e-01 -8.56742680e-01 -3.20243388e-01
-1.19881439e+00 7.78456151e-01 6.76309705e-01 1.62484959e-01
-8.92118931e-01 -2.38542199e-01 -1.14427662e+00 -1.75049290e-01
2.07983494e-01 -8.31228375e-01 9.99675453e-01 -1.24204576e+00
-1.44102550e+00 5.91597617e-01 -2.23731577e-01 7.67850727e-02
1.09043464e-01 -1.14703275e-01 -4.35959816e-01 2.97991663e-01
-1.26840457e-01 1.41002047e+00 1.01992440e+00 -1.84929705e+00
-8.25477839e-01 -4.04327691e-01 -2.59715140e-01 7.14540601e-01
-5.33097625e-01 -6.24859452e-01 -3.06663424e-01 -3.71041149e-01
2.10710883e-01 -6.80431783e-01 8.46869648e-02 -5.22909760e-01
-6.60909653e-01 -3.15375388e-01 7.68204272e-01 -4.53895062e-01
9.23807561e-01 -2.08722615e+00 8.85717750e-01 1.61933854e-01
1.30332962e-01 -3.45784366e-01 -6.22845650e-01 4.54206139e-01
-1.44835085e-01 -1.69910505e-01 -2.44307429e-01 -7.84317374e-01
1.52581617e-01 2.53529787e-01 -1.18071303e-01 3.41629475e-01
2.58035332e-01 9.28458989e-01 -8.21736515e-01 -8.12662840e-01
4.06370580e-01 5.27992249e-01 7.58701414e-02 2.43770808e-01
3.67437825e-02 5.80803990e-01 -3.53114992e-01 1.32503045e+00
6.95914626e-01 -1.31047592e-01 3.15766096e-01 -8.27157676e-01
2.26079360e-01 -3.97421956e-01 -7.93962419e-01 2.30825043e+00
-3.61083537e-01 3.96057278e-01 -2.40246188e-02 -1.23619878e+00
6.71226382e-01 3.08773816e-01 4.64347184e-01 -9.92126286e-01
3.67416978e-01 1.41532183e-01 -4.45705354e-01 -7.39375353e-01
4.54594463e-01 -2.32939348e-01 -3.59299242e-01 2.30007723e-01
5.85510075e-01 2.23132342e-01 7.94581417e-03 3.06027502e-01
5.34070730e-01 3.44286591e-01 1.47974595e-01 2.33184338e-01
6.59406960e-01 -1.68570384e-01 1.58328459e-01 6.18345022e-01
-2.41900653e-01 5.75151980e-01 3.30019742e-01 1.26292050e-01
-3.96040350e-01 -1.14334095e+00 7.79478550e-02 1.63528943e+00
5.60385942e-01 -2.14353368e-01 -2.97022432e-01 -1.08972657e+00
-9.18340907e-02 5.53049862e-01 -6.93963945e-01 -1.48379609e-01
-9.33132619e-02 -8.59326422e-01 5.87784827e-01 6.46917403e-01
4.98425454e-01 -6.73293293e-01 -1.10492609e-01 -1.83375493e-01
-7.70388782e-01 -8.87367010e-01 -2.08337009e-01 2.92969137e-01
-7.58763909e-01 -8.90038490e-01 -1.21467042e+00 -9.34897363e-01
7.47614384e-01 3.37415338e-01 7.91085780e-01 -4.00242239e-01
1.83910415e-01 1.11359406e+00 -3.88666511e-01 -2.37039551e-01
-2.21368670e-01 -1.47362798e-02 -1.20903388e-01 3.97391826e-01
2.41782635e-01 -1.56381831e-01 -3.97154748e-01 1.78505316e-01
-9.30132806e-01 3.64156246e-01 1.09460068e+00 1.21101749e+00
5.81695795e-01 -2.41139278e-01 8.25088680e-01 -3.96518886e-01
4.89294678e-01 -7.48887956e-01 -8.13497696e-03 7.56804347e-01
-2.57513463e-01 2.84598559e-01 2.52689719e-01 -7.81842589e-01
-1.52324629e+00 5.91012955e-01 3.69416535e-01 -4.07983720e-01
-4.15618211e-01 8.68255019e-01 -5.81994534e-01 8.32251622e-04
5.09560943e-01 3.93965393e-01 1.34492353e-01 -1.55479178e-01
7.47294009e-01 8.40973973e-01 7.23321438e-01 -8.24778974e-01
2.50943363e-01 3.11973751e-01 1.84480712e-01 -5.41890502e-01
-4.20583397e-01 -5.38632512e-01 -4.72595602e-01 -6.46462381e-01
9.42214966e-01 -1.29118109e+00 -9.44343209e-01 4.70587432e-01
-1.22968352e+00 2.64375180e-01 2.61008471e-01 6.87911630e-01
-5.84052920e-01 5.91412723e-01 -4.71059084e-01 -9.54671383e-01
-2.86503881e-01 -1.31499481e+00 1.68886900e+00 5.08100927e-01
-5.29354671e-03 -9.86496627e-01 2.03546900e-02 9.25271869e-01
5.63709065e-02 9.59582180e-02 9.73325729e-01 -6.81996465e-01
-3.61668587e-01 -2.83646435e-01 -4.47822064e-01 3.41396868e-01
1.03563696e-01 -2.84790993e-01 -1.29581296e+00 6.50629103e-02
-6.64990306e-01 -9.02485788e-01 1.34950399e+00 2.10566089e-01
6.83632612e-01 1.28851950e-01 -4.56830025e-01 -5.26082516e-02
1.21546841e+00 -7.00083077e-02 5.15498459e-01 1.76274162e-02
8.46725523e-01 9.57379997e-01 6.89877510e-01 3.08386117e-01
5.21988750e-01 1.64577916e-01 7.97686696e-01 -1.88095704e-01
1.64551497e-01 -1.21243916e-01 7.57080078e-01 8.84452701e-01
-1.04012728e-01 -4.49022084e-01 -6.94650352e-01 2.98791766e-01
-2.40580535e+00 -7.83363283e-01 -2.28530280e-02 1.98105633e+00
6.64382160e-01 -3.68742406e-01 8.07363614e-02 -1.49783552e-01
7.33173192e-01 -7.76206478e-02 -2.89061785e-01 1.10103875e-01
-6.32603168e-01 -4.62424934e-01 2.02377185e-01 1.68592617e-01
-1.46912742e+00 4.46832418e-01 5.56976652e+00 1.03795099e+00
-8.88938725e-01 1.35943331e-02 6.24750197e-01 5.92335425e-02
-5.12782216e-01 -3.52069229e-01 -5.72994649e-01 2.84623325e-01
7.33281851e-01 2.41091803e-01 3.39984834e-01 4.74640727e-01
-3.22113633e-01 -4.36189562e-01 -1.40509367e+00 1.41369295e+00
5.78728914e-01 -9.14077818e-01 3.61211151e-01 -2.21755415e-01
8.38170767e-01 -1.91527054e-01 4.81301039e-01 4.52475339e-01
-2.08603621e-01 -1.04161441e+00 5.23716092e-01 1.16592693e+00
3.37299168e-01 -5.72000086e-01 1.10330653e+00 3.08348060e-01
-1.11000276e+00 -1.90227494e-01 -2.88644042e-02 4.14824307e-01
1.22925788e-01 2.12860167e-01 -1.51350960e-01 1.18947327e+00
4.51780260e-01 1.06018031e+00 -7.86816359e-01 5.66425323e-01
2.54035354e-01 3.07719976e-01 -1.28414303e-01 -1.96599104e-02
2.02377126e-01 2.78799096e-03 3.98234844e-01 1.31715858e+00
4.21186298e-01 -2.01736018e-01 2.72946637e-02 6.04023695e-01
-1.83224604e-01 2.69335777e-01 -6.80502653e-01 -3.82174969e-01
2.89397445e-02 1.34814632e+00 -1.53824627e-01 -4.90133286e-01
-7.89116383e-01 9.68187928e-01 3.30170393e-02 5.07188261e-01
-9.28075433e-01 -9.08825174e-02 1.79823060e-02 -7.49994397e-01
-1.99799538e-01 2.64101326e-01 -3.74378443e-01 -1.28454220e+00
-1.52811617e-01 -9.55393076e-01 7.09036171e-01 -1.00309837e+00
-1.82986343e+00 4.87677991e-01 1.73464641e-01 -1.56896341e+00
-2.03260124e-01 -7.49686599e-01 -3.90580803e-01 7.52206445e-01
-1.51065850e+00 -1.75541091e+00 -3.12346995e-01 8.44364583e-01
4.85533506e-01 -6.39348447e-01 8.87451768e-01 2.84026116e-01
-6.81174040e-01 5.13411641e-01 3.29998881e-02 -3.02010566e-01
1.24636066e+00 -1.08300066e+00 -1.03446949e+00 2.87825674e-01
-2.58727014e-01 4.52572316e-01 4.77551103e-01 -5.99212706e-01
-1.70217395e+00 -8.26188624e-01 5.42408228e-01 -2.79361188e-01
5.29443860e-01 9.90375429e-02 -6.12744749e-01 4.63508278e-01
9.73014891e-01 -4.06772673e-01 1.12400973e+00 7.20305443e-02
-5.79312742e-01 -2.12299287e-01 -8.94598484e-01 5.65029442e-01
4.67684239e-01 -4.34978575e-01 -7.00495303e-01 3.07947010e-01
6.40532374e-01 -7.15919025e-03 -9.98785973e-01 7.28940547e-01
8.07718158e-01 -6.69054449e-01 9.09151435e-01 -6.21644735e-01
8.03533375e-01 -3.10816288e-01 -7.98694611e-01 -1.36070681e+00
2.52472591e-02 -1.97710097e-01 -2.07079649e-01 1.36923385e+00
6.88136220e-01 -4.39512461e-01 3.34760070e-01 4.12817955e-01
-1.01824671e-01 -4.73612964e-01 -8.89383316e-01 -3.52869987e-01
5.15143611e-02 -2.28689849e-01 2.78756320e-01 1.06259453e+00
6.75058544e-01 6.00499570e-01 -6.66917801e-01 6.53571188e-02
7.54232585e-01 9.38564315e-02 5.15602171e-01 -9.83636796e-01
-8.31275582e-02 -5.16354442e-01 -2.61297375e-01 -1.04825461e+00
4.40425515e-01 -8.85235846e-01 1.27115712e-01 -1.23491728e+00
6.86742246e-01 2.66908735e-01 -6.82493150e-01 8.12155426e-01
-2.82174021e-01 1.14619508e-01 3.40349048e-01 1.01371370e-01
-1.21414483e+00 8.26052070e-01 1.19382977e+00 -8.49034846e-01
-6.04714528e-02 -3.20494324e-01 -7.97727883e-01 4.60204959e-01
5.80736399e-01 -9.27288830e-02 -4.30639416e-01 -2.35929728e-01
2.08713487e-01 1.31432787e-01 4.92079109e-01 -7.19456553e-01
3.91965657e-01 -2.18829617e-01 6.71718299e-01 -8.32169235e-01
6.52460039e-01 -1.07460129e+00 -8.92489925e-02 2.34764423e-02
-6.84841156e-01 -2.20858067e-01 2.61155665e-01 9.14218366e-01
-5.19067764e-01 4.54372577e-02 4.43474770e-01 2.26521924e-01
-8.44850540e-01 -3.02610874e-01 -3.64223450e-01 -5.55807114e-01
9.06696856e-01 2.16315508e-01 -9.77584302e-01 -5.53995788e-01
-8.52264047e-01 5.74442029e-01 -5.42740524e-02 5.55792272e-01
9.03656363e-01 -1.72054279e+00 -5.70919037e-01 -2.37262323e-02
6.13480866e-01 -4.70247149e-01 6.70772254e-01 1.38664770e+00
2.81521767e-01 3.63427997e-01 -4.14512753e-01 -9.48622465e-01
-1.35760891e+00 6.03539288e-01 2.49700367e-01 -8.55798200e-02
1.90317184e-01 5.41393399e-01 4.12321359e-01 -6.53886199e-01
4.87905711e-01 4.10923474e-02 -4.57730830e-01 4.08285201e-01
2.73460925e-01 4.75451082e-01 -2.46608227e-01 -9.83708024e-01
-3.85812491e-01 6.32863343e-01 3.82145159e-02 -2.38922074e-01
8.36808980e-01 -5.39190888e-01 -1.37724906e-01 7.31512964e-01
1.32156968e+00 -5.53012311e-01 -7.93023229e-01 -3.07362974e-01
-4.54095721e-01 1.11989394e-01 -4.57003303e-02 -1.13375413e+00
-1.06666100e+00 1.08617151e+00 9.39245045e-01 -1.41179085e-01
1.47588873e+00 1.42714471e-01 6.35556400e-01 6.33326828e-01
-4.00210805e-02 -1.34017110e+00 3.86052787e-01 1.44709811e-01
8.87214065e-01 -1.84996319e+00 -1.78435147e-01 -2.53258795e-01
-1.31650221e+00 1.25534546e+00 8.00975144e-01 6.52824163e-01
5.13884246e-01 -1.54204279e-01 6.30818233e-02 -8.10329020e-02
-8.24312985e-01 -3.47855300e-01 7.13558495e-01 6.53998017e-01
4.72030073e-01 1.73379466e-01 -1.00481972e-01 1.02230620e+00
7.80226052e-01 -2.21699342e-01 -3.60130854e-02 1.08642197e+00
-3.87951374e-01 -9.50918317e-01 -7.10313916e-01 2.25644380e-01
1.22726135e-01 -6.65265173e-02 -5.65090656e-01 5.38955092e-01
1.44727230e-01 1.32090974e+00 -3.20087612e-01 -8.08811724e-01
2.17979282e-01 3.60118419e-01 6.30213976e-01 3.24418396e-02
-2.74174184e-01 5.53556979e-01 5.74636161e-02 -4.55247879e-01
-1.12832355e+00 -6.26916885e-01 -1.28973126e+00 4.97347414e-02
-3.22406471e-01 -2.34375462e-01 6.79127812e-01 1.18225694e+00
4.21148688e-01 5.01046240e-01 4.73639637e-01 -1.39999723e+00
-2.60812104e-01 -1.00984979e+00 -4.21502382e-01 5.99574864e-01
2.70433992e-01 -7.92311728e-01 -4.30134743e-01 3.08980316e-01] | [13.124393463134766, 5.020175933837891] |
d3d7af0c-fb5f-4c96-89cf-41b10cd7a747 | mutual-information-maximization-for-effective | 2003.06439 | null | https://arxiv.org/abs/2003.06439v1 | https://arxiv.org/pdf/2003.06439v1.pdf | Mutual Information Maximization for Effective Lip Reading | Lip reading has received an increasing research interest in recent years due to the rapid development of deep learning and its widespread potential applications. One key point to obtain good performance for the lip reading task depends heavily on how effective the representation can be to capture the lip movement information and meanwhile to resist the noises resulted from the change of pose, lighting conditions, speaker's appearance and so on. Towards this target, we propose to introduce the mutual information constraints on both the local feature's level and the global sequence's level to enhance the relations of the features with the speech content. On the one hand, we constraint the features generated at each time step to enable them carry a strong relation with the speech content by imposing the local mutual information maximization constraint (LMIM), leading to improvements over the model's ability to discover fine-grained lip movements and the fine-grained differences among words with similar pronunciation, such as ``spend'' and ``spending''. On the other hand, we introduce the mutual information maximization constraint on the global sequence's level (GMIM), to make the model be able to pay more attention to discriminate key frames related with the speech content, and less to various noises appeared in the speaking process. By combining these two advantages together, the proposed method is expected to be both discriminative and robust for effective lip reading. To verify this method, we evaluate on two large-scale benchmark. We perform a detailed analysis and comparison on several aspects, including the comparison of the LMIM and GMIM with the baseline, the visualization of the learned representation and so on. The results not only prove the effectiveness of the proposed method but also report new state-of-the-art performance on both the two benchmarks. | ['Shiguang Shan', 'Xilin Chen', 'Xing Zhao', 'Shuang Yang'] | 2020-03-13 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [ 1.49930799e-02 -3.66959721e-01 -2.23410144e-01 -1.79996029e-01
-6.14737093e-01 -1.31562948e-01 5.32630324e-01 -1.10870279e-01
-3.42736334e-01 4.75061297e-01 5.26326239e-01 1.54046670e-01
-1.36680469e-01 -3.23453516e-01 -4.89054114e-01 -9.64943171e-01
1.62444741e-01 -3.54845136e-01 4.17126387e-01 -8.78602788e-02
4.07974809e-01 3.38383466e-01 -2.11128736e+00 4.85753976e-02
1.00014544e+00 1.18542314e+00 6.43927217e-01 2.34288365e-01
-1.84388787e-01 5.22589445e-01 -4.69497770e-01 -6.81981966e-02
-7.70065188e-02 -3.69383991e-01 -4.05602366e-01 5.71164154e-02
1.80235088e-01 -1.01852179e-01 -1.44497693e-01 1.27985454e+00
1.09011579e+00 2.55479693e-01 5.90253055e-01 -1.09987569e+00
-3.29543203e-01 2.31370896e-01 -7.67451346e-01 1.21281281e-01
3.47702920e-01 4.48532134e-01 9.35793638e-01 -7.68573582e-01
3.89310956e-01 1.27875865e+00 4.56083745e-01 5.07223248e-01
-8.43561769e-01 -6.80529833e-01 2.80545235e-01 6.53003752e-01
-1.25002825e+00 -8.44343960e-01 1.17278862e+00 -3.17140907e-01
3.96593451e-01 3.75695407e-01 4.98114914e-01 9.72669899e-01
5.25481664e-02 1.05636358e+00 1.14032757e+00 -5.19698560e-01
-9.79838613e-03 1.09431490e-01 -3.03946044e-02 4.47205275e-01
-2.13776320e-01 6.10364005e-02 -5.62926114e-01 3.99788827e-01
3.90050888e-01 -3.45614821e-01 -7.42229879e-01 -2.04606339e-01
-1.16405058e+00 5.28942287e-01 2.42380187e-01 5.59400022e-01
-2.75836855e-01 -1.52601853e-01 4.43925947e-01 -1.77195042e-01
3.64375949e-01 1.58188734e-02 -4.11280453e-01 -2.73976386e-01
-7.98757374e-01 -2.04208925e-01 4.22194034e-01 6.10330760e-01
6.34950519e-01 -1.25521019e-01 -4.77095872e-01 1.21800745e+00
5.24521589e-01 4.59907085e-01 9.51967359e-01 -4.90546912e-01
5.69393635e-01 3.03570569e-01 -1.81163177e-01 -1.21276271e+00
-3.16754431e-01 -4.04080331e-01 -9.09538746e-01 1.49028659e-01
3.96491766e-01 -6.56995699e-02 -6.59297884e-01 2.18179131e+00
3.35733443e-01 1.83573678e-01 -6.45305291e-02 8.53099346e-01
9.50924337e-01 5.18778205e-01 -8.62200037e-02 -6.58028126e-01
1.52610290e+00 -7.85251796e-01 -1.11679173e+00 -1.26544923e-01
2.14480743e-01 -1.14862740e+00 1.13939261e+00 4.03774679e-02
-9.37750638e-01 -9.76784766e-01 -8.18712234e-01 1.13645412e-01
-1.85311481e-01 3.96836907e-01 3.00775617e-01 3.73866409e-01
-9.41179872e-01 3.59193146e-01 -4.85452384e-01 -1.80077538e-01
2.24558383e-01 2.17708230e-01 -2.75489956e-01 1.63645476e-01
-1.34426463e+00 8.11810315e-01 1.86755136e-01 2.19627723e-01
-3.57474536e-01 -4.64667648e-01 -7.80286610e-01 1.52774677e-01
2.44078994e-01 -2.93105602e-01 9.26659703e-01 -9.35443699e-01
-1.85159600e+00 6.60474002e-01 -5.00266314e-01 5.81463501e-02
7.64076412e-01 -1.41639467e-02 -4.23224777e-01 7.20448717e-02
-5.91410361e-02 6.81793869e-01 8.96896303e-01 -1.00955224e+00
-6.65424764e-01 -3.47167939e-01 -2.90653646e-01 4.59240228e-01
-3.74636948e-01 7.81045556e-02 -7.27911770e-01 -7.97079504e-01
1.13256663e-01 -7.90517926e-01 4.64233071e-01 5.36604337e-02
-3.75444174e-01 -6.27447128e-01 9.48511958e-01 -8.72038066e-01
1.16621304e+00 -2.49000812e+00 9.24726874e-02 -1.10103115e-01
-1.12847164e-01 5.20296991e-01 -2.02664226e-01 2.24388227e-01
-5.99068180e-02 6.46220371e-02 7.15930611e-02 -4.63033646e-01
-7.47961402e-02 -9.29515883e-02 1.33430541e-01 5.28676212e-01
6.48189634e-02 7.70079195e-01 -6.65943325e-01 -6.32386982e-01
3.33995789e-01 7.13253498e-01 -3.53428036e-01 3.05769295e-01
-8.14212486e-02 5.33203423e-01 -3.53365898e-01 2.30899200e-01
9.21077967e-01 7.74433985e-02 -1.41783133e-01 -5.44282138e-01
-2.15204582e-01 3.59959364e-01 -1.22824752e+00 1.41699338e+00
-3.93940419e-01 8.24735761e-01 1.94938526e-01 -8.24616611e-01
9.93422210e-01 4.70048934e-01 4.78700966e-01 -9.87464011e-01
1.72609776e-01 1.54723197e-01 1.52781650e-01 -8.42580914e-01
-4.94875014e-02 6.56665191e-02 4.55608547e-01 3.32191795e-01
-2.03150943e-01 4.39147800e-02 -7.89377838e-02 -2.01678559e-01
3.25095594e-01 -8.31209049e-02 2.19766319e-01 -3.56399208e-01
1.09936392e+00 -9.81809735e-01 4.94376481e-01 8.91708434e-02
-4.20551986e-01 3.30770880e-01 3.00510019e-01 9.99251753e-02
-6.40075684e-01 -6.75038159e-01 -2.82546192e-01 1.02899301e+00
4.96824414e-01 -1.38851134e-02 -9.00100946e-01 -5.48954844e-01
-2.06172138e-01 5.30211568e-01 -5.56790411e-01 -2.96553344e-01
-3.15310925e-01 -4.98471290e-01 3.82109821e-01 4.05538559e-01
1.04083228e+00 -1.26469696e+00 -5.16623259e-01 -7.54789114e-02
-5.46049953e-01 -9.97757614e-01 -8.38141263e-01 -1.84196040e-01
-2.26659998e-01 -9.03866231e-01 -9.79452193e-01 -8.74065995e-01
3.72682661e-01 2.89701909e-01 5.90804696e-01 4.73246649e-02
-1.70212388e-01 1.07506633e-01 -3.66194516e-01 -3.88526738e-01
-2.93864071e-01 1.48288277e-03 6.88807294e-03 4.64096636e-01
1.57722861e-01 -5.10196209e-01 -6.73441768e-01 6.73263907e-01
-8.91555309e-01 1.51003897e-01 5.24922967e-01 7.76045144e-01
4.33878183e-01 3.23370486e-01 6.92966223e-01 -4.43326775e-03
7.76836514e-01 -6.78168833e-02 -3.24479938e-01 2.18928963e-01
-4.42356646e-01 7.82246515e-02 3.78685355e-01 -8.28054905e-01
-1.08729672e+00 -3.35986495e-01 -6.27091169e-01 -2.61854976e-01
-2.07522780e-01 2.28074044e-01 -8.63994360e-01 8.24488848e-02
4.84407991e-02 5.90597272e-01 3.94176632e-01 -6.10967398e-01
1.22917838e-01 1.02946973e+00 3.92807811e-01 -3.64670515e-01
5.16185880e-01 2.38186613e-01 -8.29080418e-02 -1.03354502e+00
-5.03777444e-01 -5.46932876e-01 -4.92562950e-01 -3.16070557e-01
9.62310255e-01 -5.30563891e-01 -1.11456645e+00 1.08075213e+00
-1.15613067e+00 -1.57048032e-01 4.28628996e-02 6.55063033e-01
-6.20737374e-01 6.45460427e-01 -3.18381041e-01 -9.11474764e-01
-3.25818330e-01 -1.60254800e+00 9.61848080e-01 4.00589496e-01
1.05433233e-01 -9.71679866e-01 -6.68349490e-02 1.91712469e-01
5.28567493e-01 -1.41217694e-01 9.44576561e-01 -5.42517185e-01
-3.76099318e-01 1.20707452e-01 -4.01152313e-01 6.01239443e-01
5.97954571e-01 4.83886003e-02 -1.18357837e+00 -2.70835906e-01
1.43516555e-01 -4.25138287e-02 8.13667119e-01 5.55889130e-01
1.00595224e+00 -3.94319922e-01 -2.43377343e-01 4.76207912e-01
1.07048178e+00 2.91768789e-01 8.85168612e-01 1.58008501e-01
4.94060189e-01 9.12903011e-01 5.60570359e-01 2.40200698e-01
1.46935821e-01 1.19569492e+00 5.38948059e-01 -2.67933398e-01
-5.28640449e-01 -4.62492019e-01 2.97609776e-01 1.00232828e+00
-9.45309103e-02 -2.15980127e-01 -4.30010617e-01 4.56477791e-01
-1.67271125e+00 -1.09523368e+00 3.22200917e-02 2.39058900e+00
8.77840459e-01 -6.10379241e-02 1.21113904e-01 4.81655031e-01
1.08302391e+00 4.41103905e-01 -5.37709057e-01 -1.67799711e-01
-1.75713331e-01 -3.92997414e-02 -4.88303276e-03 4.96577322e-01
-9.35862482e-01 7.98134685e-01 5.06558037e+00 1.29701626e+00
-1.57723737e+00 -1.31241202e-01 5.70209980e-01 2.54247546e-01
-1.92290068e-01 -4.79142010e-01 -8.26250851e-01 9.00556803e-01
3.45346242e-01 -3.11176796e-02 3.71455342e-01 5.60273826e-01
5.42425811e-01 -1.20632149e-01 -7.78197229e-01 1.13007939e+00
1.27405301e-01 -9.13333952e-01 -1.45997956e-01 2.18532741e-01
3.96763474e-01 -2.88069278e-01 3.25219929e-01 1.93028338e-02
-4.19815630e-01 -9.08207059e-01 8.51165950e-01 6.44566119e-01
8.73460770e-01 -7.98854113e-01 7.43308008e-01 6.54794037e-01
-1.48033845e+00 4.60041761e-02 -2.43394583e-01 3.19462508e-01
-2.63672471e-02 4.54537362e-01 -7.26811945e-01 4.53697652e-01
5.17430246e-01 6.50776505e-01 -3.70594740e-01 1.08752310e+00
-4.52196211e-01 3.59251767e-01 -1.46556273e-01 -1.57795623e-01
1.44828642e-02 2.55170166e-02 6.51679754e-01 1.18858922e+00
2.12879881e-01 -2.19941229e-01 -6.22612014e-02 7.18315065e-01
8.12198296e-02 5.59061646e-01 -2.68100291e-01 1.58269122e-01
4.96580333e-01 1.00815523e+00 -2.81356812e-01 -4.75386195e-02
-2.67316759e-01 7.49751568e-01 4.61107772e-03 4.22401845e-01
-7.98997045e-01 -3.78630638e-01 8.92224252e-01 1.64743438e-01
3.18990320e-01 3.24937366e-02 -6.21891841e-02 -9.45205450e-01
2.23818198e-01 -1.10072863e+00 1.62454173e-02 -6.73364699e-01
-1.10143614e+00 5.62270224e-01 -3.17537516e-01 -1.19695485e+00
-2.25311980e-01 -4.27762240e-01 -6.50966167e-01 1.12695158e+00
-1.84504914e+00 -1.03977370e+00 -3.58443052e-01 9.06738877e-01
8.07269275e-01 -1.34803383e-02 5.73415399e-01 3.47956181e-01
-7.07367897e-01 1.04591370e+00 -1.21131530e-02 1.60186321e-01
8.12915862e-01 -7.30246305e-01 -1.51339263e-01 7.01161146e-01
-5.62105933e-03 4.93008941e-01 6.09950542e-01 -3.15394521e-01
-1.09853184e+00 -6.63539410e-01 8.58334482e-01 1.70139804e-01
3.44903141e-01 -1.60460234e-01 -9.70634043e-01 1.19156532e-01
1.17501579e-01 -9.47399586e-02 2.82297373e-01 -8.86036456e-02
-2.34775335e-01 -4.14676368e-01 -1.01520300e+00 4.29274350e-01
8.79869044e-01 -5.95454156e-01 -4.80127841e-01 7.15787336e-02
5.64795732e-01 -1.74493954e-01 -5.27437806e-01 8.02873313e-01
7.36419916e-01 -1.23807132e+00 7.92762756e-01 -1.89018652e-01
2.10392371e-01 -2.99792200e-01 -1.07283838e-01 -1.25721323e+00
-2.59300709e-01 -5.72185159e-01 1.36429340e-01 1.75505817e+00
1.99324712e-01 -8.14225376e-01 3.28809559e-01 2.39441752e-01
5.80859631e-02 -1.00337458e+00 -1.10249507e+00 -7.01118052e-01
-3.57085317e-02 -3.70447785e-01 6.13912106e-01 6.45815969e-01
8.93226862e-02 2.79679447e-01 -5.01477480e-01 2.84250733e-02
2.60945082e-01 -3.79159860e-02 7.28680253e-01 -1.21647584e+00
-2.45232463e-01 -7.04403281e-01 -4.62756753e-01 -1.43307149e+00
3.14683229e-01 -4.91704077e-01 2.35998288e-01 -1.31747127e+00
3.86406392e-01 -3.03555816e-01 -3.73567462e-01 2.07052380e-01
-5.14035583e-01 -5.34631824e-03 3.21309716e-01 4.01348382e-01
-2.64085025e-01 6.87011480e-01 1.53628910e+00 -1.43656611e-01
-2.52467275e-01 4.74801511e-01 -5.41509151e-01 7.48907208e-01
5.45700312e-01 7.46241733e-02 -4.24138993e-01 -2.52879649e-01
-2.58917421e-01 4.31436719e-03 7.23053887e-02 -9.29571092e-01
2.49562904e-01 -4.58575673e-02 1.55840531e-01 -4.96686399e-01
4.44650441e-01 -6.92446053e-01 -1.65084213e-01 2.78196275e-01
-3.87137711e-01 -2.66788512e-01 3.13533872e-01 4.15383637e-01
-3.77555788e-01 -1.11586437e-01 1.31592822e+00 1.33012533e-01
-6.57212973e-01 2.78938413e-01 -5.92735745e-02 2.63110306e-02
9.67028797e-01 -2.07754627e-01 -1.87834322e-01 -7.44396389e-01
-4.94610220e-01 2.04907474e-03 3.70074064e-01 6.01782799e-01
4.19130087e-01 -1.18731737e+00 -6.51760817e-01 5.00248730e-01
-4.40072082e-02 -3.32602143e-01 4.18579787e-01 1.08567095e+00
7.59920850e-02 4.38800365e-01 -1.67318329e-01 -7.23193347e-01
-1.52651131e+00 6.15268648e-01 4.01575506e-01 -2.07962289e-01
-4.08987820e-01 8.39651406e-01 5.04428208e-01 2.96968576e-02
5.72924852e-01 -1.99351415e-01 -5.95923781e-01 2.58555770e-01
6.53041601e-01 3.05886656e-01 -2.98278015e-02 -9.71820831e-01
-3.75328928e-01 9.91520584e-01 1.60328463e-01 7.87498727e-02
9.60822344e-01 -4.91417944e-01 1.04768343e-01 3.21819484e-01
1.38153851e+00 5.52031100e-01 -1.28161156e+00 -1.42450511e-01
-2.37413660e-01 -4.82637286e-01 5.00076078e-02 -8.36833537e-01
-1.10213852e+00 1.24596417e+00 7.88628817e-01 1.63637921e-01
1.29274035e+00 -5.56291901e-02 6.55834019e-01 -3.35954875e-01
1.38851851e-01 -9.98439014e-01 3.11913669e-01 2.08166048e-01
9.29701865e-01 -1.03344834e+00 -3.68834585e-01 -5.10012388e-01
-6.34066820e-01 1.05474722e+00 2.37933651e-01 4.77053076e-01
5.23197830e-01 8.81982446e-02 2.68088788e-01 3.28751266e-01
-3.52845728e-01 -4.56923276e-01 6.26263738e-01 6.35854185e-01
4.64197695e-01 -4.70884405e-02 -3.27678144e-01 3.42996538e-01
-1.86494187e-01 -7.84108266e-02 -2.43893623e-01 1.70377314e-01
-5.91034353e-01 -1.06731856e+00 -2.62400389e-01 -8.78291111e-03
-4.01835024e-01 -6.88787103e-02 -2.12427348e-01 7.47768342e-01
3.74497712e-01 1.16023159e+00 -8.35660920e-02 -3.24654281e-01
3.30683976e-01 1.67642802e-01 1.53204933e-01 -2.36187294e-01
2.24600844e-02 4.34553385e-01 -1.83042035e-01 -3.88069898e-01
-3.77058625e-01 -7.28909850e-01 -9.75898623e-01 -1.18778452e-01
-6.91676140e-01 2.09610276e-02 8.33466351e-01 1.11477363e+00
3.40825826e-01 5.38873494e-01 8.90589774e-01 -9.09537494e-01
-7.17661738e-01 -1.33256626e+00 -5.05737901e-01 6.54195547e-01
2.71005332e-01 -8.22566271e-01 -6.47358298e-01 -1.61747441e-01] | [14.278482437133789, 4.89946174621582] |
95671e9c-2f22-4cb9-bf00-2b134ff177f4 | mind-your-language-abuse-and-offense | 1809.08652 | null | http://arxiv.org/abs/1809.08652v1 | http://arxiv.org/pdf/1809.08652v1.pdf | Mind Your Language: Abuse and Offense Detection for Code-Switched Languages | In multilingual societies like the Indian subcontinent, use of code-switched
languages is much popular and convenient for the users. In this paper, we study
offense and abuse detection in the code-switched pair of Hindi and English
(i.e. Hinglish), the pair that is the most spoken. The task is made difficult
due to non-fixed grammar, vocabulary, semantics and spellings of Hinglish
language. We apply transfer learning and make a LSTM based model for hate
speech classification. This model surpasses the performance shown by the
current best models to establish itself as the state-of-the-art in the
unexplored domain of Hinglish offensive text classification.We also release our
model and the embeddings trained for research purposes | ['Rajiv Ratn Shah', 'Roger Zimmermann', 'Kshitij Rajput', 'Yaman Kumar', 'Raghav Kapoor', 'Ponnurangam Kumaraguru'] | 2018-09-23 | null | null | null | null | ['abuse-detection'] | ['natural-language-processing'] | [-6.06746137e-01 -2.33892903e-01 -2.39363447e-01 -1.31065786e-01
-6.46764934e-01 -7.55818129e-01 7.60088444e-01 -9.05725956e-02
-6.05283499e-01 7.04901278e-01 2.55430073e-01 -4.87597346e-01
2.85982102e-01 -2.99547613e-01 -4.33768660e-01 -5.87334156e-01
-2.21575171e-01 3.70070457e-01 -8.50413442e-02 -5.01956284e-01
1.76189452e-01 4.66533571e-01 -9.43484724e-01 2.32500732e-01
9.44426715e-01 3.86722684e-01 1.03841975e-01 4.44004685e-01
-1.60007905e-02 8.12371194e-01 -5.51129043e-01 -7.71884084e-01
-4.84359413e-02 -3.81580472e-01 -6.06775582e-01 -2.11446792e-01
5.59231460e-01 -7.92442977e-01 -9.85300004e-01 1.00310183e+00
4.60286766e-01 -3.11781839e-02 7.66531110e-01 -1.12180662e+00
-1.32201350e+00 8.14472854e-01 -4.92853433e-01 4.31964785e-01
1.76632583e-01 -4.32054251e-01 7.69734740e-01 -1.07286632e+00
8.04477453e-01 1.14401793e+00 6.24216855e-01 6.18908703e-01
-7.18664765e-01 -1.10608089e+00 -3.17752600e-01 4.26272035e-01
-1.58355784e+00 -5.14224410e-01 8.08737338e-01 -4.52454776e-01
1.06613421e+00 2.82340534e-02 4.64475214e-01 1.55897415e+00
3.85258228e-01 1.02633440e+00 1.08845615e+00 -4.82227564e-01
1.12705762e-02 5.48526347e-01 9.03891400e-02 6.82525039e-01
-2.63706967e-02 -1.44883126e-01 -7.53205478e-01 -4.75947469e-01
2.32340991e-01 8.74592438e-02 -2.72699803e-01 -3.08292843e-02
-5.18203855e-01 1.36349750e+00 1.93962336e-01 7.98193574e-01
2.29247697e-02 -1.21140718e-01 5.12815118e-01 5.95871270e-01
7.08880186e-01 2.76797920e-01 -3.14344645e-01 -5.15858054e-01
-1.20913768e+00 1.42353848e-01 9.90179300e-01 9.14951384e-01
1.78769007e-01 2.91275114e-01 3.13816011e-01 1.41110194e+00
3.63187760e-01 6.41173840e-01 6.18913472e-01 -2.54057854e-01
4.63226020e-01 -1.36055470e-01 -3.09795231e-01 -8.71513724e-01
-6.41130358e-02 -3.34608883e-01 -3.15201968e-01 -1.58108741e-01
4.24962491e-01 -4.40959394e-01 -9.08573389e-01 1.39192688e+00
-1.81772798e-01 -8.08515251e-02 -1.64634138e-01 4.07330334e-01
5.91757774e-01 9.21598494e-01 4.13139090e-02 -3.55814360e-02
1.06284213e+00 -8.27652156e-01 -8.35291922e-01 -3.79624292e-02
8.47194135e-01 -8.46935809e-01 7.59923935e-01 5.07266343e-01
-6.17176116e-01 1.79634504e-02 -1.13913643e+00 -2.58135200e-01
-7.90592134e-01 -1.76716402e-01 3.16416055e-01 9.65878844e-01
-8.48420143e-01 5.37660599e-01 -4.95869517e-01 -1.01350033e+00
4.37466681e-01 6.74481094e-02 -6.35414481e-01 -9.22741592e-02
-1.56436789e+00 1.25372303e+00 1.79546669e-01 -1.15307145e-01
-1.01565897e+00 -3.26158494e-01 -7.55784571e-01 -2.66312122e-01
-1.59982279e-01 6.47632718e-01 1.12195158e+00 -6.07036829e-01
-1.24462640e+00 1.11797011e+00 2.06873596e-01 -3.02431822e-01
2.24069670e-01 -5.94909489e-01 -7.88151145e-01 -7.58943334e-02
9.03544296e-03 1.01330563e-01 1.00350010e+00 -9.12766397e-01
-3.36299688e-01 -4.48846877e-01 -1.62634060e-01 -7.04518482e-02
-1.00004244e+00 7.92502999e-01 -1.30699515e-01 -7.88467765e-01
-4.65592980e-01 -1.16938901e+00 5.93809366e-01 -2.58149445e-01
-4.55444723e-01 -3.17999870e-01 1.35082340e+00 -1.43858147e+00
1.64237952e+00 -2.62646842e+00 1.65413454e-01 3.99088580e-03
1.41075000e-01 5.27417183e-01 1.19908698e-01 1.08392048e+00
1.80919230e-01 1.69064745e-01 -2.01724470e-01 -3.75708789e-01
1.87087851e-03 4.55877095e-01 -5.08842468e-01 9.28404629e-01
-1.72377393e-01 5.90207338e-01 -7.50301778e-01 -3.89983565e-01
-3.96753475e-02 4.28761482e-01 -3.63944441e-01 5.45349896e-01
5.75967312e-01 3.41053084e-02 -3.80666584e-01 8.05400670e-01
5.16824722e-01 6.06124878e-01 5.25435545e-02 4.47521895e-01
-4.71935511e-01 3.18557411e-01 -2.84041137e-01 1.57860994e+00
-4.29475099e-01 9.86846566e-01 2.58837700e-01 -4.08800244e-01
9.95110631e-01 3.74539196e-01 -2.01344118e-01 -2.36027807e-01
2.65503228e-01 5.78318417e-01 3.11020732e-01 -8.36508811e-01
3.77997994e-01 -9.63274166e-02 -2.96695679e-01 3.10044497e-01
3.00327420e-01 -2.20526680e-01 -2.20755711e-01 2.64025390e-01
8.88125181e-01 2.33903863e-02 4.74611104e-01 -7.05199540e-01
1.71628222e-01 -2.98392653e-01 2.88825274e-01 6.84405625e-01
-4.95795906e-01 4.12130922e-01 5.51096201e-01 -2.42596015e-01
-1.35071170e+00 -1.10706091e+00 -6.14014924e-01 1.63665843e+00
-5.97335458e-01 -2.01389402e-01 -9.28192735e-01 -8.43208492e-01
1.03460357e-01 1.12118435e+00 -6.27413332e-01 -5.03964238e-02
-6.56818628e-01 -4.92650092e-01 1.22041512e+00 2.50659257e-01
3.91234905e-01 -1.04212201e+00 -2.80730009e-01 2.38565400e-01
1.46949172e-01 -1.05263507e+00 -6.33204937e-01 4.39420104e-01
-2.16907069e-01 -6.18717074e-01 -9.00886655e-01 -1.08853650e+00
1.31142482e-01 -1.61172554e-01 4.97819394e-01 3.40354890e-02
1.71950217e-02 -1.21748596e-01 -7.22668946e-01 -6.49971962e-01
-4.62801397e-01 3.89559209e-01 8.90789852e-02 -1.64260164e-01
7.28617251e-01 -4.88240927e-01 1.61382362e-01 -1.64375767e-01
-8.67639959e-01 -5.75378418e-01 1.98358849e-01 7.81833291e-01
-6.29681289e-01 -3.72027040e-01 4.53316808e-01 -1.03605461e+00
6.09991431e-01 -9.43110704e-01 -1.00289457e-01 1.92751780e-01
-4.85685021e-02 -2.76123881e-01 6.86955988e-01 -3.71644586e-01
-8.01729321e-01 -2.41169319e-01 -3.49693120e-01 -1.83421865e-01
-1.68274745e-01 5.31801343e-01 -9.86268371e-02 -1.36559978e-01
3.46481085e-01 2.05382958e-01 -2.31828257e-01 -8.25457215e-01
2.38481209e-01 1.58712888e+00 2.65531898e-01 -3.11208755e-01
7.57809997e-01 1.02218390e-01 -7.13628113e-01 -1.74449134e+00
-4.79245096e-01 -4.17087615e-01 -8.86168301e-01 -1.29518658e-01
1.12696064e+00 -6.71769321e-01 1.01971544e-01 1.08509994e+00
-1.37387466e+00 -1.35701403e-01 5.79380095e-01 3.98738474e-01
-1.24322988e-01 5.66982090e-01 -1.30682373e+00 -1.05422139e+00
-1.55829668e-01 -7.58540690e-01 8.40061903e-01 -3.19819689e-01
-4.59242612e-01 -9.70616996e-01 3.56551915e-01 -4.20273952e-02
5.64424038e-01 1.48046970e-01 1.29108727e+00 -1.33634543e+00
2.17965379e-01 -4.68130670e-02 -1.23263560e-02 5.68935692e-01
9.63821188e-02 1.87630728e-01 -1.14052761e+00 -3.86170775e-01
2.02274725e-01 -8.16436112e-01 6.87176645e-01 -1.15657188e-02
6.73385441e-01 -3.65170479e-01 -1.19232990e-01 4.06416774e-01
1.03217494e+00 5.80272794e-01 4.95695561e-01 4.03632581e-01
6.87605739e-01 8.58650029e-01 8.59652385e-02 5.22988260e-01
2.65557736e-01 5.98708987e-01 -4.84956540e-02 1.64212257e-01
1.04653247e-01 -5.05461037e-01 6.85211360e-01 1.47538424e+00
2.50600010e-01 -3.42020333e-01 -1.37605500e+00 7.02222049e-01
-1.55237353e+00 -9.70998228e-01 -3.91593389e-02 2.00726366e+00
8.82499039e-01 -1.10710420e-01 9.82703045e-02 -5.83430715e-02
7.57327855e-01 3.96813601e-01 5.32766245e-03 -1.10731435e+00
1.68507081e-02 2.78708249e-01 5.34731686e-01 6.07736409e-01
-1.32150638e+00 1.22479606e+00 6.92982960e+00 8.92512560e-01
-1.13421321e+00 5.63186526e-01 2.98355311e-01 -1.53342932e-01
7.70810572e-03 -4.15445656e-01 -7.78502941e-01 6.88895047e-01
1.35163498e+00 -2.75193080e-02 6.55716300e-01 1.01249695e+00
-1.37498885e-01 2.60235131e-01 -9.97083843e-01 9.84095037e-01
7.32395589e-01 -5.68178773e-01 -3.62038970e-01 3.19707364e-01
3.60210836e-01 5.33815920e-01 1.95548877e-01 6.29847109e-01
3.25193763e-01 -1.08546567e+00 7.61383832e-01 -1.06141128e-01
7.50094056e-01 -9.61655140e-01 8.07373762e-01 3.35296243e-01
-5.15757799e-01 -2.93247253e-01 -4.36575443e-01 -3.46832797e-02
1.10273674e-01 1.55417740e-01 -5.27798057e-01 -1.62683681e-01
7.76706040e-01 7.92825758e-01 -5.27079463e-01 5.77920318e-01
-3.03566158e-01 1.03731596e+00 -1.86745733e-01 -3.67735118e-01
6.44196808e-01 -4.80364077e-02 6.12709641e-01 1.70155370e+00
3.92577112e-01 -1.99217528e-01 5.66968732e-02 3.25403720e-01
-1.90464839e-01 2.31032744e-01 -1.25224507e+00 -5.17529368e-01
5.28752387e-01 8.91766310e-01 -2.58148968e-01 -3.84050123e-02
-6.47640884e-01 1.32396460e+00 6.88377976e-01 3.87924165e-01
-8.58253598e-01 -8.47945690e-01 5.69712400e-01 9.53334123e-02
2.60452271e-01 -5.33805549e-01 2.51465231e-01 -1.28782833e+00
-3.10135305e-01 -1.02257609e+00 1.83817133e-01 -3.14704955e-01
-1.42760086e+00 1.09593129e+00 2.05943823e-01 -7.02307343e-01
-5.64632773e-01 -8.73562574e-01 -4.06078726e-01 5.15426278e-01
-1.29428804e+00 -1.33748400e+00 4.84184325e-01 4.34324771e-01
6.30553544e-01 -6.46669149e-01 8.85938227e-01 5.99887013e-01
-6.89526916e-01 7.45377421e-01 6.64691806e-01 7.31070518e-01
8.13222587e-01 -9.91183102e-01 3.99309158e-01 7.55842745e-01
-9.84001346e-03 7.09465444e-01 8.24733853e-01 -7.27600753e-01
-1.22058845e+00 -6.30948901e-01 1.48086834e+00 -6.77903354e-01
1.43283916e+00 -1.12808597e+00 -9.77046132e-01 8.15099597e-01
7.66313195e-01 -4.24106985e-01 1.07584393e+00 2.00904056e-01
-7.15956986e-01 2.50870675e-01 -1.24273300e+00 4.48163092e-01
7.91363895e-01 -1.17938364e+00 -1.03410900e+00 3.29017758e-01
4.05775249e-01 1.33339167e-01 -7.05211520e-01 -3.91451150e-01
8.69110405e-01 -7.12935507e-01 2.33366370e-01 -7.36679792e-01
5.60513437e-01 4.21284735e-01 -5.31325400e-01 -1.39021218e+00
-4.85043585e-01 -5.06260335e-01 -1.50425723e-02 1.58437085e+00
4.50039744e-01 -6.59662604e-01 4.68565732e-01 6.52399361e-01
-1.51043683e-01 -4.24639821e-01 -1.36299372e+00 -1.06114018e+00
7.97922432e-01 -1.16760179e-01 1.31263524e-01 1.43952978e+00
4.81975615e-01 2.61937559e-01 -9.53683376e-01 -3.20013076e-01
5.43032825e-01 -3.74850601e-01 3.08070302e-01 -8.88608158e-01
-2.83036828e-01 -1.04189344e-01 -3.95059168e-01 -6.02362037e-01
5.95434725e-01 -1.02636528e+00 1.45415878e-02 -9.13241684e-01
2.08910778e-01 2.35057697e-02 -1.85923323e-01 6.25716507e-01
3.96029651e-01 3.30695629e-01 5.54641843e-01 1.95630729e-01
-1.33788794e-01 5.29970706e-01 4.03796196e-01 -9.78430882e-02
-1.46095334e-02 -5.23446023e-01 -3.59248906e-01 6.17534637e-01
8.82769048e-01 -8.78560781e-01 -1.03578702e-01 -4.43517029e-01
-1.41325027e-01 -2.20896035e-01 -2.97482073e-01 -7.32061386e-01
-1.81651667e-01 3.86938304e-02 1.50047451e-01 -2.41342694e-01
3.46132517e-01 -7.96931267e-01 -3.85554075e-01 6.74481809e-01
-3.25571179e-01 1.86945230e-01 1.14469506e-01 -2.57700160e-02
-2.01866925e-01 -6.23063505e-01 8.02245736e-01 5.28144427e-02
-5.60186803e-01 1.73470825e-01 -9.32567060e-01 1.99647814e-01
9.39584076e-01 -6.44249916e-02 -2.89855450e-01 -6.25594974e-01
-2.75411487e-01 -1.31717637e-01 4.34050649e-01 7.39328146e-01
5.40609419e-01 -1.08776021e+00 -1.13065469e+00 3.00056010e-01
2.67915368e-01 -1.26104796e+00 -1.88058928e-01 4.68292624e-01
-6.30480826e-01 4.42756414e-01 -4.40910488e-01 2.81777054e-01
-1.19729567e+00 4.05711979e-01 -1.61748212e-02 1.45266250e-01
-3.60527337e-01 7.48727620e-01 4.92218174e-02 -5.44742227e-01
1.40526906e-01 2.02834934e-01 -2.83861190e-01 1.54107481e-01
6.36503696e-01 4.85106856e-01 -1.22981325e-01 -1.15894306e+00
-5.66000760e-01 9.62499157e-02 -4.60081279e-01 -5.10182381e-02
1.49151075e+00 4.12990823e-02 -4.24881071e-01 7.79452741e-01
1.75282466e+00 4.09607649e-01 -6.25380039e-01 2.12854609e-01
5.15174679e-02 -6.86009347e-01 1.39446571e-01 -5.34646690e-01
-6.78364813e-01 9.77881491e-01 3.41807425e-01 3.07258338e-01
4.03119206e-01 1.04971297e-01 1.24215293e+00 4.40889388e-01
3.42893004e-01 -1.33179486e+00 -4.69107330e-01 9.44895387e-01
1.06802988e+00 -9.01208401e-01 -2.31602639e-01 -5.19692451e-02
-7.34851122e-01 1.12703478e+00 4.13335234e-01 -2.76661247e-01
1.07884347e+00 3.84577394e-01 2.37481222e-01 5.36230244e-02
-3.85999233e-01 2.42490798e-01 1.43214062e-01 7.85349071e-01
7.57703185e-01 1.07409656e-01 -5.45417309e-01 2.77303964e-01
-4.89296436e-01 -3.61980259e-01 8.12919259e-01 1.04188359e+00
-4.80975538e-01 -1.07021713e+00 -1.58755317e-01 6.01563632e-01
-9.91311550e-01 -4.65631038e-01 -7.30659187e-01 8.81302953e-01
7.31421188e-02 1.10224533e+00 -1.10205166e-01 -5.40387809e-01
1.87830552e-02 3.42808396e-01 2.45650664e-01 -6.79101408e-01
-8.68118942e-01 -3.02094333e-02 2.78863311e-01 -3.66049558e-02
-7.76024116e-03 -6.98199868e-01 -8.65514457e-01 -7.21316755e-01
-2.02099830e-01 2.45935023e-01 6.88210428e-01 9.57150698e-01
-4.73761670e-02 -3.88975263e-01 5.23652256e-01 -6.12952411e-01
-6.09996438e-01 -1.09149897e+00 -1.28707135e+00 3.97403896e-01
4.63292032e-01 -4.31263983e-01 -4.51754361e-01 -2.78758943e-01] | [8.835644721984863, 10.571843147277832] |
d68beaef-e26d-475c-a9d7-955a5db87dbe | neural-field-conditioning-strategies-for-2d | 2304.14371 | null | https://arxiv.org/abs/2304.14371v1 | https://arxiv.org/pdf/2304.14371v1.pdf | Neural Field Conditioning Strategies for 2D Semantic Segmentation | Neural fields are neural networks which map coordinates to a desired signal. When a neural field should jointly model multiple signals, and not memorize only one, it needs to be conditioned on a latent code which describes the signal at hand. Despite being an important aspect, there has been little research on conditioning strategies for neural fields. In this work, we explore the use of neural fields as decoders for 2D semantic segmentation. For this task, we compare three conditioning methods, simple concatenation of the latent code, Feature Wise Linear Modulation (FiLM), and Cross-Attention, in conjunction with latent codes which either describe the full image or only a local region of the image. Our results show a considerable difference in performance between the examined conditioning strategies. Furthermore, we show that conditioning via Cross-Attention achieves the best results and is competitive with a CNN-based decoder for semantic segmentation. | ['Stefan Wermter', 'Sven Magg', 'Martin Gromniak'] | 2023-04-12 | null | null | null | null | ['2d-semantic-segmentation'] | ['computer-vision'] | [ 7.11406469e-01 2.59719610e-01 -5.02137616e-02 -4.51918364e-01
-7.31839299e-01 -3.28584909e-01 7.95494497e-01 -6.34001642e-02
-4.32846069e-01 6.62636280e-01 2.29743049e-01 -1.33202523e-01
1.46663800e-01 -7.28425384e-01 -9.47128475e-01 -7.43302882e-01
-1.26700914e-02 3.12572151e-01 2.58056432e-01 6.44972175e-02
3.81154329e-01 3.91253710e-01 -1.35427594e+00 4.48949695e-01
6.32094681e-01 1.01067328e+00 6.25393212e-01 6.40622854e-01
-4.80063818e-02 7.11489379e-01 -6.41412973e-01 -1.77360550e-01
-1.32335424e-01 -6.74354613e-01 -8.85262787e-01 2.53181886e-02
5.87753892e-01 1.23706259e-01 -2.76612997e-01 1.22454739e+00
3.61151725e-01 -5.07848822e-02 8.38517725e-01 -9.71193790e-01
-6.43749416e-01 5.28870463e-01 -4.21363533e-01 1.76717192e-01
3.75731736e-01 -1.13577537e-01 9.57996786e-01 -5.16714811e-01
7.56397486e-01 1.15097094e+00 4.96436864e-01 7.18750894e-01
-1.63253593e+00 -5.92531681e-01 3.02861542e-01 3.66841778e-02
-1.33066058e+00 -4.16753054e-01 6.60572350e-01 -3.91605616e-01
1.10110736e+00 7.50515833e-02 4.80359614e-01 1.01681101e+00
3.99815530e-01 8.33917320e-01 1.47075677e+00 -4.51484591e-01
2.31056079e-01 4.83598672e-02 -1.28334360e-02 5.88303626e-01
-1.17290370e-01 6.24726303e-02 -7.64241457e-01 1.40279695e-01
8.33845437e-01 -4.30918813e-01 -4.26854670e-01 1.84659474e-02
-1.24492168e+00 8.64781797e-01 6.04140282e-01 5.60054898e-01
-3.76571298e-01 6.07462764e-01 -5.61042270e-03 1.83055535e-01
4.35539126e-01 5.22193372e-01 -3.02239716e-01 5.27286418e-02
-1.31639588e+00 2.22844228e-01 7.02402234e-01 9.14272010e-01
8.90433192e-01 -4.66919951e-02 -2.91348636e-01 6.02262139e-01
3.82226288e-01 4.48988020e-01 2.45816633e-01 -9.44845259e-01
3.89592916e-01 1.77988127e-01 -2.78499722e-01 -1.03653610e+00
-3.32270592e-01 -5.93086779e-01 -7.36292601e-01 1.38108745e-01
3.69984180e-01 -2.73852330e-02 -1.19388676e+00 2.12950349e+00
-2.68629164e-01 3.57002765e-01 -6.12563863e-02 1.01072490e+00
5.57260215e-01 8.88814986e-01 1.24330454e-01 -1.19397871e-01
1.07828677e+00 -8.71863246e-01 -8.36209297e-01 -5.13522565e-01
3.94550085e-01 -8.28120589e-01 6.28658295e-01 3.68775457e-01
-1.20479333e+00 -6.61845207e-01 -1.17307317e+00 -9.82634798e-02
-4.99140948e-01 8.18099454e-02 7.19749689e-01 5.40017664e-01
-1.43645084e+00 5.76507509e-01 -7.89465249e-01 -1.94620788e-01
2.91517526e-01 7.08565235e-01 -3.43231231e-01 7.09424764e-02
-1.29807162e+00 9.46553707e-01 1.78756282e-01 -4.86060651e-03
-8.40476751e-01 -3.65367621e-01 -8.65897298e-01 1.48035556e-01
-1.33154422e-01 -5.06433189e-01 1.10105217e+00 -1.22207952e+00
-1.40954328e+00 1.04580367e+00 -3.71497452e-01 -6.71352863e-01
1.26404718e-01 -6.70633540e-02 -2.19228536e-01 2.75251538e-01
1.43695325e-01 1.41432834e+00 9.44885254e-01 -1.34926140e+00
-3.49377900e-01 -1.28255159e-01 1.24354862e-01 1.02851138e-01
6.58554062e-02 -8.90104324e-02 -7.62916327e-01 -6.82974935e-01
3.50839764e-01 -8.79001856e-01 -2.30797723e-01 -2.33103350e-01
-5.37704825e-01 4.01939526e-02 6.22012138e-01 -5.90891182e-01
1.02234769e+00 -2.41760349e+00 4.42171961e-01 3.38818312e-01
-2.20146347e-02 -1.14896297e-01 -1.07954212e-01 1.64272383e-01
-2.60394931e-01 1.46338820e-01 -5.57626247e-01 -5.42945981e-01
4.78159487e-02 3.54247272e-01 -4.24872369e-01 6.94537997e-01
5.62308609e-01 1.01421452e+00 -6.00631118e-01 -4.90759671e-01
2.90570613e-02 7.18966186e-01 -7.29501843e-01 -6.42128065e-02
-4.03352886e-01 6.03384316e-01 -2.48912275e-01 3.72494280e-01
6.27618551e-01 -4.02621090e-01 -1.09741958e-02 -9.92689952e-02
-2.32793003e-01 4.73234147e-01 -1.05831504e+00 2.10823298e+00
-3.79563987e-01 1.00247121e+00 1.59267709e-01 -1.18005514e+00
8.48426759e-01 4.12108332e-01 2.60989428e-01 -9.01432812e-01
1.95431516e-01 1.94416195e-01 -2.21023738e-01 -1.17415056e-01
2.90574580e-01 -1.85127586e-01 -1.49187803e-01 3.86131376e-01
5.43423057e-01 -5.99917620e-02 1.64329574e-01 1.17669575e-01
9.98023152e-01 1.70305565e-01 -4.28117141e-02 -3.32769156e-01
4.00108218e-01 -4.05351706e-02 1.94442645e-01 8.42889667e-01
1.51580811e-01 8.71943772e-01 6.76132798e-01 2.18477190e-01
-7.80969679e-01 -8.93258572e-01 -3.26545894e-01 8.76476884e-01
4.59622145e-01 -2.71046162e-01 -9.84876752e-01 -3.37837964e-01
-1.92994237e-01 5.54946184e-01 -8.65844548e-01 -1.29207253e-01
-5.30565083e-01 -6.27931416e-01 7.03466535e-01 4.31487352e-01
6.44607186e-01 -1.02655661e+00 -6.92562401e-01 2.41255417e-01
-2.07730144e-01 -1.00840306e+00 -5.59516370e-01 8.56264949e-01
-9.48545814e-01 -6.74859047e-01 -8.69616210e-01 -9.45993602e-01
8.36254418e-01 -3.55698168e-02 1.04779124e+00 -9.46661010e-02
5.05920574e-02 1.67612731e-01 -1.92442343e-01 -3.81058753e-02
-2.10491359e-01 2.65154272e-01 -4.92390513e-01 2.01248214e-01
2.23421797e-01 -5.63132226e-01 -3.00935298e-01 8.24559927e-02
-9.33636308e-01 3.41352701e-01 6.94688439e-01 8.12696993e-01
6.02036357e-01 -3.67619991e-01 2.12356612e-01 -9.64091063e-01
5.57578802e-01 -3.63392115e-01 -5.60729265e-01 3.40964310e-02
-3.22187871e-01 3.95914972e-01 3.04084361e-01 -4.05176014e-01
-8.81209195e-01 4.97636616e-01 -3.05155098e-01 -3.14742684e-01
-4.41493005e-01 7.07271338e-01 -3.96023467e-02 -1.59338132e-01
5.99264622e-01 2.31121510e-01 -1.61227688e-01 -4.16234851e-01
3.73611748e-01 4.72494513e-01 6.04093194e-01 -4.03043747e-01
3.42529923e-01 4.63737130e-01 7.22513050e-02 -7.11981297e-01
-6.94951594e-01 -3.80692303e-01 -8.19371819e-01 -2.36277044e-01
1.39259303e+00 -8.17350388e-01 -3.35690349e-01 4.23781574e-01
-1.50747716e+00 -3.54844451e-01 -1.34449840e-01 4.82732654e-01
-8.38951766e-01 -1.72167808e-01 -5.86559117e-01 -5.70585430e-01
2.70652413e-01 -1.33620656e+00 1.20957506e+00 1.92880407e-01
-3.12909156e-01 -9.89728689e-01 1.00941984e-02 -8.08519721e-02
3.77633393e-01 3.08814645e-01 6.60434783e-01 -3.26269031e-01
-9.09853399e-01 -7.84210265e-02 -2.27950409e-01 1.68795854e-01
-1.70836866e-01 -3.36224109e-01 -1.28873348e+00 -1.02042146e-01
1.83710568e-02 -2.86097109e-01 1.30341494e+00 7.41520584e-01
1.03595328e+00 8.33800733e-02 -4.25510466e-01 7.84384489e-01
1.54973388e+00 3.50329340e-01 8.38337362e-01 4.19441462e-02
3.80371004e-01 5.87933064e-01 1.09779619e-01 5.98620921e-02
1.55466855e-01 6.19401276e-01 4.61568028e-01 -3.97222608e-01
-2.52627850e-01 -1.94925174e-01 2.03606844e-01 7.87836969e-01
3.98083590e-04 -2.50224859e-01 -6.78269267e-01 5.52878320e-01
-1.59104860e+00 -8.44241619e-01 -2.60553300e-01 1.75141144e+00
8.19739878e-01 2.55322009e-01 -4.05046463e-01 2.63942271e-01
8.18823338e-01 2.17234552e-01 -3.34964216e-01 -5.05490303e-01
-4.31663245e-01 5.99863827e-01 7.78825581e-01 7.06317782e-01
-1.14477110e+00 1.02796447e+00 7.33700657e+00 8.94343317e-01
-1.60683692e+00 3.72319400e-01 6.10414743e-01 3.42538267e-01
-3.92597646e-01 1.00966461e-01 -6.62089407e-01 3.83187294e-01
8.83333564e-01 4.36031699e-01 5.08360207e-01 4.94229198e-01
-6.81704432e-02 -4.77739960e-01 -1.11684358e+00 9.25265729e-01
5.05637042e-02 -1.39232481e+00 -7.51039237e-02 2.23401040e-01
1.01685107e+00 5.14622815e-02 3.18691999e-01 -7.40369260e-02
-2.27282625e-02 -1.35622895e+00 1.06654310e+00 7.28867769e-01
8.23966324e-01 -5.73636830e-01 5.26637375e-01 3.10022295e-01
-9.89552617e-01 2.94740885e-01 -1.39218062e-01 -1.12607427e-01
2.31692925e-01 3.87242287e-01 -2.18332842e-01 1.74280956e-01
5.73328793e-01 6.55409932e-01 -3.96075606e-01 9.57695246e-01
-3.63051593e-01 8.96617413e-01 -2.77431220e-01 -7.70373344e-02
5.25721014e-01 1.11529052e-01 3.50545496e-01 1.48122823e+00
1.90881565e-01 -1.16500044e-02 -1.25415921e-01 1.23963475e+00
1.04985297e-01 -1.65751241e-02 -6.61634922e-01 -9.74008814e-02
8.65013339e-03 7.79166877e-01 -9.86126125e-01 -3.19662422e-01
-5.47528625e-01 1.29463172e+00 2.05453217e-01 6.30025983e-01
-8.71847391e-01 -3.42886448e-01 3.43979448e-01 -1.70176346e-02
6.67715788e-01 -4.62830931e-01 -7.49742210e-01 -9.43357885e-01
-3.13418120e-01 -3.87424618e-01 3.68318893e-02 -9.61895287e-01
-9.76957679e-01 6.40305400e-01 1.12569496e-01 -9.38046753e-01
-1.15508676e-01 -6.66055083e-01 -3.96036863e-01 1.16186917e+00
-1.57336628e+00 -9.59168255e-01 -1.26376763e-01 7.10777402e-01
2.43974552e-01 5.48913591e-02 9.25921619e-01 4.65810150e-01
1.42416442e-02 3.34992826e-01 -5.81549630e-02 4.67791222e-02
4.60642666e-01 -1.28567982e+00 2.92277277e-01 7.12608278e-01
4.81211245e-01 8.08519006e-01 7.32030690e-01 -5.29398024e-01
-9.88065362e-01 -8.34189355e-01 1.17558885e+00 -3.73582579e-02
2.25198999e-01 -4.97984946e-01 -9.00094390e-01 5.95048308e-01
5.57404578e-01 -2.90830255e-01 4.55061734e-01 -4.41194810e-02
-1.79090917e-01 1.42389700e-01 -7.97511399e-01 3.48854184e-01
8.36770296e-01 -7.10268795e-01 -5.85850060e-01 -4.37339097e-02
6.56917632e-01 -5.54367602e-01 -5.63898206e-01 2.69331992e-01
2.15483710e-01 -8.89954448e-01 7.95693517e-01 -9.94568393e-02
4.64971930e-01 -3.80649805e-01 -3.02889377e-01 -1.32212627e+00
-4.15852189e-01 -2.25970417e-01 2.38387495e-01 1.00107694e+00
6.93472445e-01 -4.57519710e-01 7.87336648e-01 2.19501942e-01
-3.60213757e-01 -5.09856164e-01 -1.09249651e+00 -3.46345067e-01
1.61833465e-01 -7.58288205e-01 3.41418147e-01 8.93118620e-01
-1.25652716e-01 3.61649305e-01 -4.43438351e-01 2.40643308e-01
4.18957293e-01 2.28151932e-01 2.41741329e-01 -9.27308261e-01
-3.04802328e-01 -5.93291342e-01 -4.94900942e-01 -1.69591522e+00
2.55621165e-01 -1.16621411e+00 3.05918545e-01 -1.57238626e+00
-1.17752561e-02 -1.58888176e-01 -3.60445946e-01 4.13661629e-01
1.77282006e-01 4.52419162e-01 1.50816724e-01 1.26880497e-01
-4.23957527e-01 3.47389877e-01 1.43686306e+00 -4.08568472e-01
-6.80712312e-02 -1.61305949e-01 -6.60542011e-01 3.50651711e-01
7.75755823e-01 -5.51748157e-01 -2.36737609e-01 -5.92001438e-01
2.17859253e-01 1.72514677e-01 4.90569979e-01 -1.19581306e+00
5.84522843e-01 1.54885948e-01 4.89592046e-01 -6.17595673e-01
4.95946944e-01 -7.44175255e-01 2.15539992e-01 4.73380238e-01
-6.38624728e-01 -1.42654866e-01 1.70582697e-01 3.99377435e-01
-6.61975861e-01 -3.31998348e-01 8.95498872e-01 -2.91770905e-01
-7.23004103e-01 3.82270776e-02 -7.30062902e-01 -9.71904546e-02
6.27652109e-01 -4.43934351e-01 8.58717877e-03 -4.39007014e-01
-1.15195012e+00 1.32219400e-02 1.76025525e-01 1.87824070e-01
6.41322792e-01 -1.34892249e+00 -4.94913965e-01 4.39761698e-01
-1.66598082e-01 -7.54018128e-02 1.72328146e-04 8.93406749e-01
-5.73839068e-01 7.66478479e-01 -3.77021134e-01 -8.45332146e-01
-9.13850963e-01 2.85090894e-01 3.73783737e-01 -3.49637978e-02
-2.03805462e-01 1.17085183e+00 3.40284973e-01 -1.64891794e-01
3.58467221e-01 -5.45942724e-01 -3.54301244e-01 9.48051959e-02
2.24401802e-01 -1.85523927e-01 -1.50528550e-01 -9.64501619e-01
-3.06759298e-01 8.38024259e-01 3.56142133e-01 -4.76554990e-01
1.16278267e+00 -5.38085960e-03 -2.03195736e-01 6.09377444e-01
1.32292223e+00 -1.29227534e-01 -1.49219525e+00 -1.05333619e-01
1.17363092e-02 -2.90338725e-01 3.28997284e-01 -6.88468695e-01
-1.21419322e+00 1.19386482e+00 5.91145039e-01 3.24715912e-01
1.34591091e+00 2.00787604e-01 3.89383078e-01 -1.54790029e-01
3.15777957e-01 -9.97603595e-01 3.54403108e-02 7.07553089e-01
8.41229260e-01 -8.42047811e-01 -3.33930463e-01 -3.92277002e-01
-5.25683582e-01 1.02149880e+00 5.18700957e-01 -4.21801776e-01
8.45105648e-01 3.51423562e-01 -1.40887946e-02 -2.90183634e-01
-8.25504959e-01 -5.50232708e-01 4.44171876e-01 5.19693077e-01
6.24995053e-01 -7.00808838e-02 -2.64544725e-01 3.72832894e-01
-1.97827592e-01 -4.87360209e-02 1.44329369e-01 8.16583514e-01
-4.26827282e-01 -9.92291272e-01 -4.80966330e-01 3.21427166e-01
-6.26889169e-01 -1.87497735e-01 -4.01969939e-01 5.41163743e-01
6.00879312e-01 8.12867761e-01 2.64699310e-01 -3.51396084e-01
1.55714914e-01 8.25271532e-02 7.02201784e-01 -7.27646470e-01
-5.50933421e-01 4.06374991e-01 -2.01698869e-01 -5.16036987e-01
-9.40434515e-01 -8.22211385e-01 -1.39513469e+00 -2.66684834e-02
-5.15001297e-01 4.59421724e-02 9.13147330e-01 1.13733709e+00
-7.86226690e-02 6.77501976e-01 2.29259968e-01 -9.94382203e-01
1.18970923e-01 -9.65273559e-01 -6.54623926e-01 2.82668144e-01
4.65229362e-01 -7.30682731e-01 -2.27691397e-01 4.47874457e-01] | [9.727343559265137, 0.7147372364997864] |
dd0d8a3c-5346-4066-9345-c15609766a83 | degenerative-adversarial-neuroimage-nets | 1907.02787 | null | https://arxiv.org/abs/1907.02787v2 | https://arxiv.org/pdf/1907.02787v2.pdf | Degenerative Adversarial NeuroImage Nets: Generating Images that Mimic Disease Progression | Simulating images representative of neurodegenerative diseases is important for predicting patient outcomes and for validation of computational models of disease progression. This capability is valuable for secondary prevention clinical trials where outcomes and screening criteria involve neuroimaging. Traditional computational methods are limited by imposing a parametric model for atrophy and are extremely resource-demanding. Recent advances in deep learning have yielded data-driven models for longitudinal studies (e.g., face ageing) that are capable of generating synthetic images in real-time. Similar solutions can be used to model trajectories of atrophy in the brain, although new challenges need to be addressed to ensure accurate disease progression modelling. Here we propose Degenerative Adversarial NeuroImage Net (DaniNet) --- a new deep learning approach that learns to emulate the effect of neurodegeneration on MRI by simulating atrophy as a function of ages, and disease progression. DaniNet uses an underlying set of Support Vector Regressors (SVRs) trained to capture the patterns of regional intensity changes that accompany disease progression. DaniNet produces whole output images, consisting of 2D-MRI slices that are constrained to match regional predictions from the SVRs. DaniNet is also able to maintain the unique brain morphology of individuals. Adversarial training ensures realistic brain images and smooth temporal progression. We train our model using 9652 T1-weighted (longitudinal) MRI extracted from the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. We perform quantitative and qualitative evaluations on a separate test set of 1283 images (also from ADNI) demonstrating the ability of DaniNet to produce accurate and convincing synthetic images that emulate disease progression. | ['Neil P. Oxtoby', 'Daniele Ravi', 'Daniel C. Alexander'] | 2019-07-05 | null | null | null | null | ['predicting-patient-outcomes'] | ['medical'] | [ 3.12543184e-01 -1.06853463e-01 3.31286043e-01 -5.43605864e-01
-6.52382076e-01 -2.71134406e-01 5.99664927e-01 -3.37719023e-01
-4.63308096e-01 8.33170474e-01 3.17603260e-01 -3.41026783e-01
-8.70771147e-03 -5.56686044e-01 -7.19064772e-01 -6.20327830e-01
-8.43688190e-01 7.81671941e-01 1.61103323e-01 -1.32007480e-01
-2.54389316e-01 5.38118184e-01 -9.66885567e-01 2.86147565e-01
9.23052132e-01 6.74416363e-01 1.31132990e-01 9.36587512e-01
5.20622849e-01 4.70263422e-01 -5.67208827e-01 -2.72397518e-01
2.74214506e-01 -3.95851612e-01 -6.97684944e-01 -1.31443702e-02
6.13247216e-01 -6.38532579e-01 -5.28986752e-01 7.83895910e-01
8.25014174e-01 -3.20257515e-01 1.01634407e+00 -1.08491218e+00
-7.55955696e-01 2.12940365e-01 -4.47786331e-01 7.43795455e-01
-2.84634084e-02 7.72387505e-01 3.92267168e-01 -4.97150362e-01
9.67259288e-01 1.16268635e+00 1.10162175e+00 1.01713574e+00
-1.85421824e+00 -5.46520710e-01 -2.04088509e-01 1.09193422e-01
-8.35501730e-01 -2.17693448e-01 3.70856583e-01 -8.50049675e-01
7.35337615e-01 2.07004339e-01 9.92676556e-01 1.64841437e+00
8.44760001e-01 2.15450808e-01 1.57403159e+00 5.00164106e-02
2.53236651e-01 -6.04379833e-01 2.45072469e-02 5.11244476e-01
-3.10303140e-02 3.33631992e-01 2.28250921e-01 -3.41820300e-01
1.12279820e+00 -6.32708296e-02 -3.24267238e-01 -2.96306014e-01
-1.32212269e+00 8.11481833e-01 6.17856503e-01 2.64444947e-02
-5.95933139e-01 3.59605551e-01 5.74164927e-01 3.75285536e-01
4.98544723e-01 1.45406500e-01 -3.61218423e-01 5.26288629e-01
-1.26646090e+00 8.10340762e-01 1.37836576e-01 1.65057287e-01
-1.90627635e-01 4.04479235e-01 -2.64668375e-01 9.79983032e-01
2.20988303e-01 6.06089711e-01 9.00739193e-01 -1.22504890e+00
1.51548252e-01 3.84649426e-01 -9.62170437e-02 -4.64779437e-01
-8.04621696e-01 -5.42576015e-01 -1.25335908e+00 9.52731669e-01
6.11446381e-01 -3.53865713e-01 -1.31078327e+00 1.79985476e+00
9.74060148e-02 5.32023385e-02 -2.46972024e-01 8.97769034e-01
3.14428627e-01 2.77570546e-01 5.61213672e-01 1.82340331e-02
1.26819003e+00 -2.53800184e-01 -2.35292614e-01 -2.04820052e-01
7.30826914e-01 -1.17385842e-01 1.07767725e+00 1.38990954e-01
-1.41647732e+00 -1.62767187e-01 -8.43644977e-01 4.94837314e-02
-1.33542299e-01 -8.56021866e-02 3.77329916e-01 5.28514624e-01
-1.50992310e+00 6.69507921e-01 -1.22376573e+00 -2.28450462e-01
9.58236873e-01 4.88697588e-01 -4.22999710e-01 -9.04375166e-02
-1.32129157e+00 1.22340035e+00 -1.14368916e-01 4.59060557e-02
-1.46389687e+00 -1.30574715e+00 -5.50505757e-01 -3.47561389e-01
-6.47780359e-01 -1.19258511e+00 1.09427524e+00 -1.09164548e+00
-5.64595819e-01 1.16816366e+00 7.57961050e-02 -1.00609910e+00
1.04847610e+00 5.76328859e-02 -5.27385473e-01 4.35658932e-01
1.65923268e-01 1.00165641e+00 7.36694574e-01 -1.23362434e+00
2.17717290e-01 -6.90499187e-01 -3.61032367e-01 -6.93948269e-02
6.46114051e-02 4.24187839e-01 4.42497671e-01 -1.03506768e+00
-2.32793599e-01 -9.92969096e-01 -6.99781358e-01 6.54892802e-01
-3.04482013e-01 5.06368697e-01 7.99093127e-01 -1.42034793e+00
7.53233135e-01 -1.49661577e+00 1.40724614e-01 7.79314786e-02
5.25410056e-01 2.85634100e-01 -2.57107288e-01 -2.41378337e-01
-5.73540807e-01 4.25670743e-01 -8.21212649e-01 -2.86993086e-01
-3.10375899e-01 1.24564297e-01 -4.39284369e-02 6.52765512e-01
4.51558143e-01 1.21216810e+00 -6.48608923e-01 -3.64973783e-01
-1.53842062e-01 8.22526753e-01 -5.08026779e-01 -2.55505852e-02
-2.72883952e-01 8.36856842e-01 -1.67927787e-01 4.51202482e-01
4.65184838e-01 -2.62194984e-02 1.26577914e-01 8.68031289e-03
3.05965751e-01 -2.07405299e-01 -3.13293397e-01 1.34887767e+00
-2.99708605e-01 3.21007669e-01 2.07239762e-01 -1.05403852e+00
6.00515902e-01 4.31712806e-01 6.51810527e-01 -8.36279809e-01
1.66968122e-01 2.38708347e-01 4.70142901e-01 -4.21449274e-01
-4.97912139e-01 -6.29938602e-01 1.54110506e-01 6.86810851e-01
-1.83278322e-01 -4.68216427e-02 1.07698321e-01 -7.52435625e-03
1.50015330e+00 6.25002459e-02 -3.48343551e-01 -2.14186668e-01
3.71319115e-01 2.21578941e-01 4.34533358e-01 3.71782154e-01
-5.99431217e-01 8.18935156e-01 5.71995735e-01 -7.77853370e-01
-1.90683877e+00 -1.47441840e+00 -4.08448368e-01 2.34152913e-01
-8.09880137e-01 3.74740392e-01 -1.00008857e+00 -6.54694021e-01
-6.17101043e-02 5.22218704e-01 -9.52295363e-01 -3.42891723e-01
-9.31345224e-01 -1.47800648e+00 9.39604402e-01 6.93361104e-01
2.73050427e-01 -1.17089534e+00 -6.15362883e-01 2.17422083e-01
6.17672391e-02 -6.82498336e-01 -4.16690379e-01 -2.64244229e-01
-1.35464096e+00 -1.31935942e+00 -1.39739823e+00 -9.43105936e-01
9.07412887e-01 -5.70383668e-01 1.17105675e+00 1.46862775e-01
-8.25463176e-01 2.79932082e-01 1.38605371e-01 -1.48538321e-01
-9.29367542e-01 -3.24289292e-01 3.46567273e-01 -3.46502602e-01
-1.95730329e-01 -1.15004456e+00 -1.14042640e+00 2.62695938e-01
-1.00337982e+00 -9.18051898e-02 5.99784434e-01 8.56763363e-01
8.29147220e-01 -2.00330585e-01 9.19212401e-01 -7.11029530e-01
8.57475638e-01 -6.61765218e-01 -8.09012502e-02 1.13368966e-01
-5.89746237e-01 -1.57745108e-01 6.73599422e-01 -6.67751908e-01
-8.27058375e-01 -4.27127816e-02 -1.94335654e-01 -4.20100003e-01
-1.34853527e-01 2.45082080e-01 1.73915327e-01 -8.76071304e-02
9.44472909e-01 2.55986810e-01 6.23347998e-01 -3.60043436e-01
1.55941620e-01 1.74269348e-01 7.36149073e-01 -4.94068742e-01
5.67771077e-01 5.82039952e-01 3.32604438e-01 -5.50451577e-01
-3.28999579e-01 3.70427042e-01 -7.99231291e-01 -3.38234901e-01
8.34615946e-01 -8.12906384e-01 -2.10702077e-01 8.14280987e-01
-7.68520296e-01 -9.20038044e-01 -3.05168241e-01 2.63963997e-01
-8.07697177e-01 1.20898984e-01 -7.69441843e-01 -2.76961237e-01
-8.15330386e-01 -1.22611284e+00 6.44512951e-01 -2.68262863e-01
-5.28737962e-01 -1.31950295e+00 3.31437290e-01 3.59610796e-01
6.64289773e-01 8.87299180e-01 1.49992037e+00 -4.70808715e-01
-5.38551025e-02 -1.18280306e-01 -1.93549469e-01 5.64266503e-01
-2.12207019e-01 -1.83335990e-01 -4.51808184e-01 -2.94717550e-01
4.98463493e-03 -4.01594371e-01 7.45599329e-01 8.21197033e-01
9.92008924e-01 -3.46411705e-01 -1.90356970e-01 5.38929462e-01
1.26367533e+00 1.23494565e-01 9.39507782e-01 4.11617875e-01
5.12936890e-01 6.11141443e-01 -1.14144176e-01 -1.57685488e-01
2.34503329e-01 5.76470256e-01 4.08107400e-01 -3.72515082e-01
-6.13409281e-01 3.31604183e-01 4.39145803e-01 2.55724877e-01
-2.83769965e-01 3.07721674e-01 -1.17522991e+00 5.78959167e-01
-1.24092376e+00 -1.13880110e+00 -4.94008601e-01 2.01211572e+00
1.10304928e+00 2.31655002e-01 5.70095778e-01 -1.52191110e-02
7.82645166e-01 -8.77040029e-02 -8.59634995e-01 -3.97852302e-01
-2.79637128e-01 5.38111925e-01 4.05989379e-01 2.67946005e-01
-8.73795033e-01 2.61804104e-01 7.34341812e+00 5.32936305e-03
-1.08410883e+00 4.77318287e-01 1.07052362e+00 -5.38890243e-01
-2.48669997e-01 -4.54704434e-01 -1.26768082e-01 6.99995816e-01
1.45312119e+00 1.01762237e-02 4.03002143e-01 2.74393350e-01
8.03089976e-01 2.56912619e-01 -7.67330289e-01 2.51933068e-01
-2.11585313e-01 -1.18953753e+00 6.70516640e-02 2.30067670e-02
7.72998393e-01 3.33167523e-01 2.85803467e-01 8.87437165e-02
4.69942182e-01 -1.41286004e+00 5.58588982e-01 1.07755291e+00
1.27530253e+00 -6.23129249e-01 5.61904252e-01 4.75633927e-02
-4.53806490e-01 -2.19300184e-02 5.51432744e-03 3.28577608e-01
4.21965688e-01 5.34336150e-01 -8.51513147e-01 -8.03252533e-02
6.35835052e-01 3.48931760e-01 -7.55923629e-01 1.11848259e+00
-4.16166969e-02 7.68806875e-01 -5.23324264e-03 7.62096167e-01
1.20772555e-01 -1.11716807e-01 6.00435555e-01 1.08042145e+00
3.79804581e-01 -2.21361056e-01 -1.51854917e-01 1.28598702e+00
-2.51535922e-02 -2.12537736e-01 -3.22614670e-01 3.26075777e-02
-2.49922145e-02 9.28888559e-01 -5.78209460e-01 -6.83518723e-02
-1.70592442e-01 9.15584207e-01 2.09534690e-01 3.39302897e-01
-6.22316658e-01 5.42354099e-02 7.53406882e-01 8.05572867e-01
-2.05023382e-02 -4.35999721e-01 -4.74258244e-01 -7.36185133e-01
-9.71148238e-02 -9.77541029e-01 1.32272124e-01 -1.04267716e+00
-1.42382324e+00 6.36502385e-01 -1.53502628e-01 -7.34446347e-01
-2.57568628e-01 -5.85964680e-01 -1.04984736e+00 1.30619645e+00
-1.16208363e+00 -1.33626103e+00 4.93287283e-04 5.27083874e-01
2.48750433e-01 -3.25744413e-02 8.96981835e-01 3.72546941e-01
-6.20517492e-01 3.06146324e-01 7.43048415e-02 1.74004495e-01
6.55045569e-01 -1.35750270e+00 8.01408768e-01 7.15699553e-01
-5.55427074e-01 5.68786740e-01 7.62693524e-01 -1.16242635e+00
-6.13154650e-01 -1.59404516e+00 6.76519454e-01 -6.14753306e-01
9.25446987e-01 6.21162914e-02 -1.06906605e+00 8.47667992e-01
-2.46002287e-01 3.26279879e-01 4.39695895e-01 -5.39157808e-01
-2.14059860e-01 9.68431234e-02 -1.54644799e+00 7.04002440e-01
1.02401900e+00 -3.25215042e-01 -6.03928208e-01 4.24791902e-01
3.06403130e-01 -2.05899969e-01 -1.22405541e+00 3.28462631e-01
7.73921847e-01 -7.37747967e-01 1.45055354e+00 -1.04370654e+00
8.20860326e-01 4.42174338e-02 3.40866834e-01 -1.70201612e+00
-1.73857689e-01 -1.39703169e-01 -2.06436872e-01 7.77269900e-01
2.44875610e-01 -5.46694815e-01 7.27112174e-01 8.36344481e-01
-8.10248107e-02 -1.14714670e+00 -9.61278796e-01 -7.77907908e-01
7.29637980e-01 -1.77475452e-01 2.97315329e-01 5.89691639e-01
-6.63384736e-01 -1.85338914e-01 -1.00268170e-01 8.13138261e-02
9.29323018e-01 -6.60706401e-01 1.79469418e-02 -1.42693257e+00
8.77214894e-02 -6.17813468e-01 -6.32381856e-01 2.30904296e-01
3.08217704e-01 -1.08935392e+00 -4.06917602e-01 -1.61474967e+00
2.78035611e-01 -7.04864323e-01 -3.13225895e-01 4.18352693e-01
1.40061053e-02 6.97780371e-01 -1.66164994e-01 4.68239158e-01
2.32315585e-01 2.96475440e-01 1.66106927e+00 -3.16730857e-01
5.28885610e-02 -7.91344866e-02 -5.09016454e-01 6.32103503e-01
1.03961504e+00 -4.78274226e-01 -4.20889527e-01 -2.09657356e-01
-1.31379455e-01 -1.07797839e-01 1.28869045e+00 -1.12490916e+00
-2.91127354e-01 2.12501436e-01 1.08034372e+00 2.64984928e-03
2.80587733e-01 -4.80119497e-01 4.10220802e-01 9.04548109e-01
-4.13978279e-01 3.38256687e-01 8.11462551e-02 3.34297568e-01
3.40946168e-01 -5.42661473e-02 9.96038020e-01 -4.64530468e-01
-5.34768961e-02 6.87309563e-01 -6.19816840e-01 2.96058893e-01
9.59457099e-01 -1.07139513e-01 -3.37044507e-01 -1.93754226e-01
-1.36066115e+00 2.06844136e-01 6.61686242e-01 3.26987579e-02
5.49855292e-01 -1.29541814e+00 -1.07993293e+00 -2.96464115e-02
-4.76978987e-01 -2.34302029e-01 4.73881751e-01 1.16206253e+00
-8.30214143e-01 1.56367019e-01 -8.74689519e-01 -2.95573086e-01
-1.11724818e+00 3.19953680e-01 8.32538843e-01 -6.03040993e-01
-9.80078161e-01 5.96089065e-01 7.15196356e-02 -3.00763100e-01
-1.79791331e-01 -1.48506597e-01 -8.61701369e-02 -8.28805268e-02
6.63984239e-01 3.40428025e-01 9.35594961e-02 -7.52294481e-01
-1.45942912e-01 1.51056185e-01 -2.90995717e-01 -1.68942317e-01
1.68901753e+00 -5.31501044e-03 -3.29145603e-02 7.12713599e-02
9.12177145e-01 -6.36557519e-01 -1.50324368e+00 2.03190327e-01
-2.24517778e-01 2.78711498e-01 2.57367462e-01 -1.37812269e+00
-1.34423065e+00 8.42535734e-01 1.17964709e+00 -3.02975439e-03
9.32274282e-01 -2.59048045e-01 1.14933479e+00 -3.28628629e-01
1.66775957e-01 -5.19774139e-01 -8.73420313e-02 -1.02211395e-02
1.30494761e+00 -7.60799348e-01 -1.05794892e-01 1.52039349e-01
-5.03299296e-01 1.02388263e+00 4.66549605e-01 -5.55129170e-01
4.99004424e-01 3.34416956e-01 2.07983404e-01 -3.96289587e-01
-4.87279326e-01 4.32349831e-01 2.01095223e-01 1.14585149e+00
2.45760873e-01 9.35309902e-02 -4.92168546e-01 6.47048831e-01
-2.00700149e-01 1.94870904e-01 4.94637251e-01 7.36293912e-01
-4.29780670e-02 -1.22082365e+00 -4.68795836e-01 1.00813746e+00
-6.88399017e-01 -5.36246225e-02 -2.17762023e-01 7.82869875e-01
2.05461398e-01 1.16146684e-01 3.73813398e-02 1.17103137e-01
1.79403037e-01 4.10958529e-01 7.19022810e-01 -3.45691085e-01
-5.72005093e-01 -4.33180630e-01 7.12172389e-02 -4.20116305e-01
-2.66323119e-01 -1.10597718e+00 -1.25493872e+00 -6.60950467e-02
4.62440640e-01 -6.12968504e-01 5.93699694e-01 7.64354467e-01
2.80336291e-01 9.43561256e-01 3.11722666e-01 -1.07696331e+00
-6.08140945e-01 -9.83530283e-01 -6.07731819e-01 5.15522659e-01
2.67979532e-01 -7.55632222e-01 -3.08286279e-01 4.81494039e-01] | [14.103503227233887, -1.9327499866485596] |
ce8acb4d-afc1-4664-a073-54565d2ca603 | ridcp-revitalizing-real-image-dehazing-via | 2304.03994 | null | https://arxiv.org/abs/2304.03994v1 | https://arxiv.org/pdf/2304.03994v1.pdf | RIDCP: Revitalizing Real Image Dehazing via High-Quality Codebook Priors | Existing dehazing approaches struggle to process real-world hazy images owing to the lack of paired real data and robust priors. In this work, we present a new paradigm for real image dehazing from the perspectives of synthesizing more realistic hazy data and introducing more robust priors into the network. Specifically, (1) instead of adopting the de facto physical scattering model, we rethink the degradation of real hazy images and propose a phenomenological pipeline considering diverse degradation types. (2) We propose a Real Image Dehazing network via high-quality Codebook Priors (RIDCP). Firstly, a VQGAN is pre-trained on a large-scale high-quality dataset to obtain the discrete codebook, encapsulating high-quality priors (HQPs). After replacing the negative effects brought by haze with HQPs, the decoder equipped with a novel normalized feature alignment module can effectively utilize high-quality features and produce clean results. However, although our degradation pipeline drastically mitigates the domain gap between synthetic and real data, it is still intractable to avoid it, which challenges HQPs matching in the wild. Thus, we re-calculate the distance when matching the features to the HQPs by a controllable matching operation, which facilitates finding better counterparts. We provide a recommendation to control the matching based on an explainable solution. Users can also flexibly adjust the enhancement degree as per their preference. Extensive experiments verify the effectiveness of our data synthesis pipeline and the superior performance of RIDCP in real image dehazing. | ['Chong-Yi Li', 'Zhi Chai', 'Chun-Le Guo', 'Zheng-Peng Duan', 'Rui-Qi Wu'] | 2023-04-08 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wu_RIDCP_Revitalizing_Real_Image_Dehazing_via_High-Quality_Codebook_Priors_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wu_RIDCP_Revitalizing_Real_Image_Dehazing_via_High-Quality_Codebook_Priors_CVPR_2023_paper.pdf | cvpr-2023-1 | ['image-dehazing'] | ['computer-vision'] | [ 3.12293202e-01 -4.55430383e-03 5.15433371e-01 -2.19898656e-01
-7.44030833e-01 -4.10035342e-01 5.99782646e-01 -3.14258724e-01
-2.33916551e-01 3.79236579e-01 3.54553550e-01 5.68291219e-03
-7.95606822e-02 -9.36260343e-01 -8.20823252e-01 -1.22688067e+00
2.17615306e-01 -7.61519149e-02 3.43529224e-01 -4.71204758e-01
1.35700777e-01 2.47674897e-01 -1.76828027e+00 8.38524774e-02
1.22493100e+00 8.37880790e-01 5.64802110e-01 6.13792300e-01
3.62228423e-01 4.81209010e-01 -7.60176182e-01 -4.41230297e-01
7.34150767e-01 -5.22961378e-01 -2.26256639e-01 2.94134319e-01
5.90249777e-01 -6.18746519e-01 -4.51890349e-01 1.29414701e+00
5.61136007e-01 5.33369370e-02 3.48944485e-01 -1.37572217e+00
-8.46187115e-01 1.43648684e-01 -3.79890233e-01 -9.30848345e-02
1.78440690e-01 6.30098581e-01 7.63392508e-01 -7.27235436e-01
4.71316427e-01 1.12799382e+00 4.27914143e-01 4.40530956e-01
-9.81319904e-01 -7.59186208e-01 4.08534296e-02 2.15865433e-01
-1.49658847e+00 -6.53319240e-01 8.12234342e-01 -3.14584762e-01
4.92633045e-01 3.02871764e-01 6.81320548e-01 1.10571599e+00
1.96815774e-01 3.53498042e-01 1.13466513e+00 -3.41943979e-01
2.88571626e-01 8.92324075e-02 -4.30837840e-01 3.89881074e-01
2.51589507e-01 2.70958513e-01 -5.18390357e-01 1.66450694e-01
5.91014802e-01 2.35195067e-02 -7.23518550e-01 -2.43055046e-01
-1.24863660e+00 5.84310770e-01 4.22818750e-01 2.02490408e-02
-4.08200413e-01 -9.87240765e-03 -1.70699835e-01 3.77255321e-01
2.92753130e-01 7.09475040e-01 -2.62775868e-01 3.22805315e-01
-1.02016723e+00 3.07728410e-01 3.20545703e-01 1.02583814e+00
1.09753954e+00 1.35876805e-01 -2.77079284e-01 6.62606776e-01
2.85430074e-01 7.67370582e-01 1.68223023e-01 -1.12723029e+00
2.64268041e-01 1.37636542e-01 3.91896486e-01 -1.05330741e+00
1.34761846e-02 -4.65646148e-01 -8.19780231e-01 4.35852528e-01
3.18126410e-01 8.07056800e-02 -1.12889898e+00 1.57318151e+00
2.52199113e-01 4.28149760e-01 2.14320511e-01 1.13741934e+00
4.63410646e-01 1.08902669e+00 -3.18422139e-01 -1.55464828e-01
1.28900528e+00 -1.00586522e+00 -9.12936985e-01 -1.48545489e-01
1.92364931e-01 -8.14859450e-01 1.27537668e+00 6.45636439e-01
-8.95522952e-01 -4.48803425e-01 -1.35973942e+00 -1.29119352e-01
-2.81727403e-01 -5.04209772e-02 1.75091729e-01 5.42729735e-01
-1.22291756e+00 4.72112685e-01 -6.98813498e-01 -2.90607363e-01
1.79288954e-01 1.46574005e-01 -1.57530665e-01 -4.84525979e-01
-1.42157280e+00 8.41806889e-01 3.91968310e-01 2.88077295e-01
-1.20444775e+00 -9.13463354e-01 -8.89563382e-01 -8.80829394e-02
3.91468257e-01 -7.85432875e-01 7.45380878e-01 -9.45531726e-01
-1.76767743e+00 4.85182971e-01 2.01220810e-01 -3.01047862e-01
4.23855692e-01 -1.80097878e-01 -6.68343008e-01 3.04263383e-01
-1.12220839e-01 8.71607721e-01 1.56213176e+00 -1.54423368e+00
-4.55273002e-01 1.76862687e-01 3.10205847e-01 2.30327696e-01
-3.73331875e-01 -8.03206563e-02 -6.65745854e-01 -8.96268904e-01
-4.20780741e-02 -7.52453923e-01 -1.52448624e-01 2.16327190e-01
-3.34353089e-01 5.03138483e-01 7.48400033e-01 -8.51785421e-01
1.11718452e+00 -2.49538803e+00 1.42832100e-01 1.44185901e-01
4.02505070e-01 3.24314773e-01 -4.07576740e-01 3.95326942e-01
3.21409628e-02 -6.05308451e-03 -5.43417692e-01 -3.71474266e-01
-2.47491486e-02 3.14020097e-01 -4.62861687e-01 5.72674692e-01
5.28775692e-01 6.76161706e-01 -1.01060498e+00 -3.28777462e-01
3.64657670e-01 6.78800821e-01 -6.86779380e-01 6.00079000e-01
-3.29781860e-01 5.50552070e-01 -8.75970796e-02 5.61609209e-01
1.24498272e+00 -2.26335227e-02 -1.78812742e-01 -5.78744948e-01
-2.45874703e-01 1.58651769e-01 -1.08159530e+00 1.64231980e+00
-5.23133397e-01 5.42868197e-01 3.28403622e-01 -5.52145243e-01
7.91347563e-01 4.03531194e-02 1.13240451e-01 -8.00703466e-01
7.50972703e-02 1.40249953e-01 -1.96713135e-01 -6.41372561e-01
5.44628263e-01 -2.12827086e-01 1.99914321e-01 1.13703586e-01
4.51744683e-02 -5.97324491e-01 -7.48272836e-02 1.83917478e-01
1.03381038e+00 1.42073914e-01 -3.07480358e-02 -2.51306802e-01
5.68685174e-01 -2.69197792e-01 6.10052586e-01 6.27841175e-01
-9.71178785e-02 1.35613108e+00 2.56927550e-01 -8.46886933e-02
-1.28736484e+00 -1.11986005e+00 -1.94536219e-03 6.86415434e-01
5.05610287e-01 -5.02986431e-01 -8.71496558e-01 -3.01882118e-01
-4.12409931e-01 7.71297038e-01 -6.08939290e-01 -2.88032264e-01
-4.88094717e-01 -8.84981692e-01 4.38593417e-01 -1.73497513e-01
8.77328753e-01 -5.90195417e-01 -3.93669099e-01 2.54624635e-02
-2.84211665e-01 -1.21275222e+00 -6.20595276e-01 -1.22728750e-01
-2.41108954e-01 -7.03823566e-01 -5.57009757e-01 -4.14720744e-01
6.06725097e-01 8.62546265e-01 8.02881420e-01 4.53891814e-01
1.00680791e-01 3.98415998e-02 -6.30752146e-01 -3.74442011e-01
-6.86682463e-01 -2.92836994e-01 -1.31454468e-02 3.57751042e-01
-2.59005487e-01 -8.12164605e-01 -9.20138538e-01 3.62110645e-01
-1.45762217e+00 2.69708306e-01 6.53947353e-01 7.50638008e-01
3.04218441e-01 5.81062913e-01 1.37825429e-01 -3.44354212e-01
2.92197287e-01 -4.70511526e-01 -8.04435194e-01 1.58873707e-01
-6.70484066e-01 4.16219756e-02 7.17962384e-01 -5.20525873e-01
-1.17632914e+00 -2.11407483e-01 -2.02353045e-01 -6.53106809e-01
-7.36489668e-02 1.72787711e-01 -6.98755980e-01 -2.09754378e-01
6.50229692e-01 3.12752396e-01 7.55051747e-02 -2.47891247e-01
5.74076772e-01 6.38791382e-01 7.68204868e-01 -4.37040657e-01
1.59618783e+00 7.04904020e-01 -3.76583189e-01 -7.16174841e-01
-7.70747840e-01 -1.62359774e-01 -3.45115602e-01 -1.05666019e-01
1.03722656e+00 -1.25574839e+00 -3.85227054e-01 7.25768924e-01
-1.10635352e+00 -3.90589744e-01 -1.70358628e-01 4.23313290e-01
-1.99935675e-01 5.46592593e-01 -4.81410682e-01 -4.74717557e-01
-1.72774956e-01 -1.28392029e+00 1.22504294e+00 1.76232412e-01
4.62363392e-01 -7.16816604e-01 -1.76995531e-01 2.88613558e-01
7.14680731e-01 1.29666999e-01 8.03061247e-01 6.38149679e-02
-1.01370656e+00 1.72781378e-01 -3.95092428e-01 6.49988234e-01
2.07710460e-01 1.77715227e-01 -1.20902622e+00 -4.87571061e-01
1.63671434e-01 8.48244429e-02 7.73415804e-01 2.37659179e-02
1.07643163e+00 -3.47468644e-01 1.39910087e-01 1.11986995e+00
1.46244419e+00 -1.16407342e-01 1.14101040e+00 6.04462683e-01
7.25409508e-01 6.65862620e-01 6.54052675e-01 3.25601220e-01
3.73097956e-01 7.25016296e-01 7.92620122e-01 -3.16875219e-01
-4.51191515e-01 -2.29042053e-01 6.85120225e-01 7.09592342e-01
1.66778877e-01 -6.54310703e-01 -5.71566343e-01 4.45528090e-01
-1.51169634e+00 -6.65705383e-01 -1.45096974e-02 2.19230771e+00
8.85446727e-01 6.03932738e-02 -2.88657576e-01 -8.38347822e-02
5.60523748e-01 4.98010188e-01 -2.21593991e-01 5.58069125e-02
-4.02962089e-01 9.73955393e-02 6.99826360e-01 6.13242805e-01
-8.94315660e-01 9.17076766e-01 5.56633186e+00 9.59914029e-01
-1.41788363e+00 1.86058640e-01 2.04996124e-01 -1.32173121e-01
-6.85070932e-01 1.84801176e-01 -5.69611967e-01 7.49875546e-01
9.24346805e-01 1.76183861e-02 7.96338737e-01 1.95433512e-01
4.05928791e-01 -2.55636014e-02 -7.54795909e-01 1.08158517e+00
1.83179483e-01 -1.33036876e+00 2.53890514e-01 1.07577428e-01
6.84116185e-01 -1.00246407e-02 1.85369328e-01 -1.53111843e-02
1.88640609e-01 -8.14561963e-01 1.10647929e+00 5.74595630e-01
7.96063483e-01 -4.03878361e-01 5.96834958e-01 2.55013764e-01
-7.76687920e-01 -9.73843783e-02 -5.20508409e-01 1.48175016e-01
1.47611618e-01 9.13851142e-01 -6.99795544e-01 8.91155779e-01
8.71741951e-01 6.73331618e-01 -7.20018268e-01 9.50918138e-01
-4.86875772e-01 6.42321527e-01 -2.97946215e-01 7.07296193e-01
1.16866358e-01 -4.08312798e-01 4.31320101e-01 1.05475378e+00
6.70405209e-01 2.37654999e-01 -1.75242037e-01 9.39072192e-01
5.56173176e-02 -2.39344746e-01 -3.90578330e-01 2.01669306e-01
4.92314279e-01 1.27259731e+00 -3.54286551e-01 -1.34507090e-01
-3.93641561e-01 1.08814013e+00 -1.59478948e-01 5.05601227e-01
-1.06209791e+00 -2.62489140e-01 8.90373528e-01 9.67580825e-02
3.93935382e-01 -2.46988177e-01 6.62669390e-02 -1.37745607e+00
1.53350189e-01 -1.17327738e+00 -9.22654048e-02 -1.24913907e+00
-1.32624376e+00 6.40418530e-01 8.14049393e-02 -1.56474316e+00
2.27138996e-01 -4.70817834e-01 -5.62687695e-01 9.03531671e-01
-2.05177808e+00 -1.31365490e+00 -8.60660851e-01 5.48283756e-01
3.65024298e-01 2.12939739e-01 3.65107775e-01 4.79495227e-01
-6.37710929e-01 4.92919922e-01 1.05270326e-01 -2.07108274e-01
1.09198308e+00 -1.01358902e+00 4.34279650e-01 1.36594760e+00
-2.41799116e-01 5.17878711e-01 1.19303226e+00 -4.38181549e-01
-1.40841889e+00 -1.36019969e+00 2.30630040e-01 -4.77897704e-01
6.71995819e-01 -6.18445575e-01 -1.37541807e+00 3.12348038e-01
4.75481123e-01 1.38851836e-01 2.76712120e-01 -8.02682281e-01
-4.70276564e-01 -3.74081552e-01 -1.02012908e+00 6.61255300e-01
1.02127826e+00 -5.66368937e-01 -4.20451820e-01 2.72962898e-01
1.25314522e+00 -3.92188311e-01 -8.41624558e-01 3.71282339e-01
2.93493032e-01 -1.10673094e+00 9.77525353e-01 7.25489110e-02
4.33467746e-01 -9.10000324e-01 -3.32629681e-01 -1.47192526e+00
-3.26614022e-01 -1.00612473e+00 1.56289846e-01 1.28802741e+00
1.22877531e-01 -6.10942781e-01 4.27204698e-01 3.48146588e-01
-3.10876042e-01 -3.50524336e-01 -5.22091389e-01 -7.78881371e-01
-7.05516860e-02 -4.49584901e-01 1.15785205e+00 1.03716803e+00
-6.95016444e-01 -2.12517932e-01 -7.86907375e-01 9.04676795e-01
7.24304497e-01 -6.85965940e-02 9.18193161e-01 -7.44931936e-01
-3.92515481e-01 -8.85476023e-02 -3.10115546e-01 -1.08727002e+00
-1.95155516e-02 -4.49060261e-01 4.49579269e-01 -1.22158229e+00
-1.09050490e-01 -4.63735014e-01 -2.07081467e-01 2.30443284e-01
-2.99184293e-01 3.82694513e-01 2.47145772e-01 4.06556994e-01
-3.77728134e-01 8.44506383e-01 1.37391925e+00 -1.54597104e-01
-8.95604342e-02 -3.92635196e-01 -7.72789359e-01 3.91512007e-01
7.00238585e-01 -5.68173230e-01 -5.72674036e-01 -7.30182111e-01
3.79391164e-01 -1.90677032e-01 5.64609408e-01 -1.07403374e+00
1.71962231e-01 -2.82941282e-01 9.73942280e-02 -2.68953025e-01
3.31392050e-01 -7.06524670e-01 3.20575148e-01 6.18176418e-04
4.39742729e-02 -2.57807255e-01 -6.93056434e-02 5.94079792e-01
-3.46235573e-01 -6.73431009e-02 9.38551307e-01 7.23354891e-02
-5.94199479e-01 3.26585740e-01 -2.37539947e-01 -2.12509647e-01
7.31227398e-01 -3.06370854e-01 -6.57526910e-01 -5.85441053e-01
-2.20212385e-01 2.07534492e-01 1.13772607e+00 3.90407383e-01
7.99658537e-01 -1.00730658e+00 -7.42902637e-01 5.13492763e-01
3.80855948e-01 2.41163343e-01 3.34776103e-01 7.91551948e-01
-7.14692414e-01 -2.07518503e-01 -1.36712670e-01 -4.59254950e-01
-8.05249155e-01 8.92086387e-01 5.11103213e-01 2.03034535e-01
-8.28770697e-01 6.39984846e-01 5.73802233e-01 -2.34150440e-01
-3.02853417e-02 -2.70827353e-01 8.75190720e-02 -1.76174194e-01
7.28009164e-01 3.76996286e-02 2.99793512e-01 -5.25038242e-01
-6.30012900e-02 4.31666434e-01 1.20292433e-01 -1.70024224e-02
1.40886176e+00 -5.70185304e-01 -2.43908241e-01 -9.23633799e-02
1.21773875e+00 2.66917586e-01 -1.85327792e+00 -1.62613451e-01
-4.21536714e-01 -9.17503059e-01 2.73649514e-01 -5.14491737e-01
-1.30012012e+00 8.27785790e-01 6.74863160e-01 9.37265251e-03
1.61822569e+00 -1.64627358e-01 9.70812261e-01 1.79373413e-01
2.10166335e-01 -7.83261061e-01 2.07302108e-01 1.66828454e-01
9.52353716e-01 -1.06080508e+00 4.22499813e-02 -5.73378086e-01
-5.20838261e-01 9.36657012e-01 4.56453294e-01 3.28491116e-03
5.38175285e-01 1.82130441e-01 3.67852479e-01 -1.91908583e-01
-6.90418482e-01 -1.26072854e-01 4.94511202e-02 7.07772195e-01
-2.63414741e-01 -3.03906471e-01 1.15663096e-01 2.46760428e-01
-4.14672256e-01 -3.52888584e-01 9.77832258e-01 6.07243955e-01
-4.19699609e-01 -9.63955522e-01 -6.83370531e-01 2.53047012e-02
-1.26234323e-01 -2.63763726e-01 1.92353278e-01 5.60912132e-01
5.36725342e-01 1.16013384e+00 -1.06596082e-01 -5.96675873e-01
3.19781989e-01 -4.39738423e-01 2.96030223e-01 -4.11992311e-01
-2.88233906e-01 1.59291416e-01 -1.62122846e-01 -6.10249400e-01
-4.91712183e-01 -4.27067935e-01 -9.34574425e-01 -2.83956677e-01
-3.62000316e-01 6.96886927e-02 5.74974597e-01 8.24566126e-01
4.62798983e-01 3.88949126e-01 9.49110389e-01 -9.96510208e-01
-4.01388049e-01 -7.10923016e-01 -6.41907692e-01 5.63162327e-01
7.07284153e-01 -7.43506253e-01 -7.48918951e-01 2.32636213e-01] | [10.936593055725098, -3.1161489486694336] |
4957b07b-245e-464b-a0ce-8bfaab7afaff | vader-video-alignment-differencing-and | 2303.13193 | null | https://arxiv.org/abs/2303.13193v2 | https://arxiv.org/pdf/2303.13193v2.pdf | VADER: Video Alignment Differencing and Retrieval | We propose VADER, a spatio-temporal matching, alignment, and change summarization method to help fight misinformation spread via manipulated videos. VADER matches and coarsely aligns partial video fragments to candidate videos using a robust visual descriptor and scalable search over adaptively chunked video content. A transformer-based alignment module then refines the temporal localization of the query fragment within the matched video. A space-time comparator module identifies regions of manipulation between aligned content, invariant to any changes due to any residual temporal misalignments or artifacts arising from non-editorial changes of the content. Robustly matching video to a trusted source enables conclusions to be drawn on video provenance, enabling informed trust decisions on content encountered. | ['John Collomosse', 'Viswanathan Swaminathan', 'Ritwik Sinha', 'Stefano Petrangeli', 'Md. Mehrab Tanjim', 'Tu Bui', 'Simon Jenni', 'Alexander Black'] | 2023-03-23 | null | null | null | null | ['video-alignment', 'misinformation'] | ['computer-vision', 'miscellaneous'] | [ 1.84391648e-01 -2.88838506e-01 -6.89942956e-01 -2.63412409e-02
-6.22666180e-01 -1.11871040e+00 6.36224508e-01 7.12285101e-01
-2.83061564e-01 3.31407398e-01 7.40462422e-01 1.12277701e-01
-1.89635858e-01 -3.03492010e-01 -7.68907607e-01 -1.68355227e-01
-7.69895732e-01 -2.02146154e-02 8.09397280e-01 2.38236383e-01
8.77145946e-01 6.53816223e-01 -1.35503221e+00 6.09687507e-01
6.40059039e-02 1.08914816e+00 -1.96758986e-01 8.13369811e-01
3.57643723e-01 1.24571192e+00 -7.94238389e-01 -2.82747120e-01
7.65901387e-01 -1.85325190e-01 -8.64519835e-01 6.24672659e-02
7.57341325e-01 -8.66739154e-01 -6.14538789e-01 9.66385543e-01
1.08995855e-01 7.88071156e-02 6.09299950e-02 -1.59301484e+00
-1.79845780e-01 5.02719641e-01 -4.58991587e-01 8.41527045e-01
1.24807036e+00 3.15259635e-01 5.38382113e-01 -6.46689177e-01
1.53307176e+00 9.27520096e-01 7.89291799e-01 4.44300361e-02
-9.26093280e-01 -5.96505582e-01 1.51409224e-01 7.43184924e-01
-1.61283481e+00 -9.46262181e-01 4.72813785e-01 -4.87362653e-01
9.35604930e-01 5.35409927e-01 6.25631213e-01 8.87804091e-01
6.52084351e-01 2.28211984e-01 4.83313054e-01 -3.11095119e-02
3.77021819e-01 -2.40021393e-01 -5.07224500e-01 7.24535644e-01
1.64841369e-01 1.64500162e-01 -1.19945729e+00 -7.38205075e-01
5.22450864e-01 -2.12831110e-01 -4.61090207e-01 -4.25491124e-01
-1.60314798e+00 3.29208046e-01 2.08531395e-01 1.94021817e-02
-8.34471345e-01 2.76611865e-01 9.77488756e-01 4.78815317e-01
1.55301958e-01 5.62048018e-01 -6.74942136e-02 -4.45415944e-01
-1.50036573e+00 3.56837124e-01 5.47282219e-01 1.28547072e+00
5.80779791e-01 -1.24197841e-01 -4.23326492e-01 -3.18041116e-01
1.43119320e-01 2.76663065e-01 4.46009822e-02 -1.43815017e+00
4.05456007e-01 1.87983185e-01 1.67892292e-01 -1.70027113e+00
3.13391924e-01 2.11023107e-01 -1.72762245e-01 2.26525262e-01
-6.01275414e-02 3.80630732e-01 -3.71721774e-01 1.14172685e+00
7.48547077e-01 3.73248100e-01 -2.96362698e-01 8.60622048e-01
3.86420846e-01 3.14879566e-01 -8.11125636e-02 -5.97322047e-01
1.25858295e+00 -3.82527828e-01 -7.68509448e-01 2.58554697e-01
1.08606540e-01 -1.02849817e+00 5.89265767e-03 4.25844342e-01
-1.26564300e+00 -1.95997551e-01 -1.18480086e+00 1.14377923e-01
-3.25571653e-03 -7.38374174e-01 -2.46487662e-01 2.23913878e-01
-1.04493666e+00 7.45129466e-01 -4.65761632e-01 -2.65442044e-01
3.53044420e-01 1.28715783e-01 -8.00097287e-01 -2.82325447e-02
-1.14972436e+00 7.38323510e-01 4.94418174e-01 -3.25239867e-01
-1.31261504e+00 -9.48910654e-01 -7.38111734e-01 -2.35733837e-01
8.56353223e-01 -4.84828681e-01 1.20568335e+00 -1.11072850e+00
-9.56374884e-01 7.36539185e-01 -3.01097006e-01 -7.29514122e-01
6.46266758e-01 -1.40012249e-01 -5.22379994e-01 9.89988744e-01
4.93966550e-01 3.45076561e-01 1.56153309e+00 -9.96602356e-01
-8.94086123e-01 -2.07986906e-01 -5.72147705e-02 1.94293648e-01
2.41667032e-01 4.50307757e-01 -5.28722465e-01 -9.98901069e-01
-9.62286890e-02 -6.34897411e-01 1.10729247e-01 3.68348241e-01
-1.81809813e-01 3.70103300e-01 1.23672616e+00 -1.25199139e+00
1.70797336e+00 -2.24941850e+00 1.54179081e-01 5.69532394e-01
3.11440438e-01 -1.64180640e-02 -1.02073558e-01 8.58855784e-01
7.92724937e-02 2.31581911e-01 3.30153465e-01 3.38385403e-01
-3.83340925e-01 -3.76256168e-01 -5.59813023e-01 9.00551081e-01
-2.13369802e-01 4.15318698e-01 -1.16826451e+00 -6.00838006e-01
-1.20123606e-02 4.78734449e-02 -4.61890429e-01 3.30921143e-01
-2.56208837e-01 3.20180833e-01 -2.17685997e-01 1.14577019e+00
5.07054090e-01 3.02907377e-01 1.30193695e-01 -6.36543155e-01
-2.49187708e-01 9.06010419e-02 -1.09783602e+00 1.71674514e+00
1.75557390e-01 8.12814415e-01 3.28420848e-01 -2.29586557e-01
4.10953432e-01 6.02313817e-01 7.61857390e-01 -7.02398896e-01
-9.16050524e-02 -1.20042317e-01 -7.19907939e-01 -8.91041636e-01
9.58631516e-01 5.46385527e-01 -1.74273610e-01 7.45803595e-01
-2.06459552e-01 1.50634661e-01 1.02990448e-01 7.39251792e-01
1.65874338e+00 1.90775260e-01 6.03772521e-01 1.70841590e-01
3.05068225e-01 4.13169682e-01 5.49770355e-01 6.92016125e-01
-6.34908259e-01 2.56291956e-01 3.39763165e-01 -6.39667869e-01
-1.30874610e+00 -8.86907876e-01 3.05898666e-01 1.11330485e+00
6.36730790e-01 -1.11254752e+00 -5.89113832e-01 -8.51457953e-01
6.79671615e-02 5.22834718e-01 -5.41373849e-01 -2.72436917e-01
-6.56519949e-01 5.13571680e-01 7.24988341e-01 2.31870458e-01
5.05579829e-01 -7.37522602e-01 -1.27089310e+00 3.00345570e-01
-4.09571737e-01 -1.08233738e+00 -1.07741988e+00 -4.67192680e-01
-6.44934952e-01 -1.55857646e+00 9.74592194e-02 -2.09337085e-01
4.38683301e-01 5.95816195e-01 8.19658875e-01 3.23384672e-01
-3.84853035e-01 9.37456667e-01 -5.30560017e-01 2.39778310e-01
-8.08138669e-01 -8.77475381e-01 3.14397752e-01 -1.38094738e-01
-6.26534373e-02 -1.17619142e-01 -5.96070766e-01 7.03153372e-01
-9.79982495e-01 -1.86778396e-01 -6.80566058e-02 4.65206414e-01
6.63135469e-01 6.65735826e-02 5.91899976e-02 -2.96624362e-01
4.82363522e-01 -7.84493864e-01 -4.67691571e-01 5.02040029e-01
-5.42869508e-01 -2.44102642e-01 2.96759665e-01 -5.14257491e-01
-7.84159541e-01 -1.95017740e-01 7.38907039e-01 -1.01757896e+00
4.01186533e-02 4.77460563e-01 2.89053712e-02 -1.13347955e-01
6.29569232e-01 4.69058752e-02 -1.08602189e-01 -4.69881529e-03
5.17442167e-01 4.89979982e-01 1.00360405e+00 -3.61834407e-01
8.55071366e-01 7.65984416e-01 -1.68840110e-01 -4.07887697e-01
-3.47232744e-02 -4.67767626e-01 -7.86314905e-01 -6.32744849e-01
7.41460741e-01 -9.50372517e-01 -6.71493769e-01 1.29188344e-01
-1.25354779e+00 1.46686539e-01 -3.47496003e-01 1.44861326e-01
-6.19906187e-01 8.87425184e-01 -4.14449543e-01 -4.14332867e-01
-2.90574729e-01 -8.78894329e-01 6.53470755e-01 -1.08374596e-01
-8.48372638e-01 -3.45388353e-01 1.77448377e-01 3.34944725e-01
3.57135981e-01 4.43523526e-01 4.36669350e-01 -5.68500102e-01
-1.08899832e+00 -4.43989754e-01 3.50277163e-02 -2.04846457e-01
1.99651465e-01 4.97736394e-01 -4.52350348e-01 -7.39920855e-01
-1.24691792e-01 2.68623624e-02 1.51059330e-01 9.50678140e-02
6.80025220e-01 -1.18025959e+00 -4.02573019e-01 4.97420788e-01
1.02933538e+00 5.21176815e-01 5.36096275e-01 6.79654300e-01
3.37478220e-01 4.11087453e-01 1.27699375e+00 8.30172002e-01
1.53202470e-02 7.39092827e-01 6.18185341e-01 5.38494468e-01
-1.70796141e-01 -5.62499285e-01 6.51544571e-01 2.23194793e-01
5.44204451e-02 -1.07856311e-01 -3.86037767e-01 5.98246872e-01
-1.79333568e+00 -1.80057800e+00 2.29922310e-01 2.55809569e+00
4.73712504e-01 -3.15935798e-02 2.08057597e-01 -1.93551689e-01
1.15922582e+00 3.49588573e-01 -6.06762946e-01 -3.87371182e-01
1.77671596e-01 -2.71352351e-01 8.89589787e-01 2.19069526e-01
-1.03329635e+00 8.29697251e-01 7.36902523e+00 7.89993346e-01
-7.89892972e-01 1.62311822e-01 6.76409081e-02 -4.98330891e-01
-3.44888330e-01 3.56710285e-01 -2.47415140e-01 5.88680506e-01
7.87325323e-01 -6.25397682e-01 6.65429115e-01 5.25106430e-01
5.70181727e-01 -4.85321164e-01 -1.18134701e+00 8.04298520e-01
5.07403255e-01 -1.91308331e+00 -7.99267367e-03 -7.35876486e-02
4.53272760e-01 -1.94692761e-01 -1.98488817e-01 -4.04509455e-01
1.86285391e-01 -6.69977129e-01 1.44394886e+00 5.35012901e-01
9.56130862e-01 -6.57768548e-01 2.98328102e-01 -7.13448524e-02
-1.35064805e+00 -1.24132477e-01 8.36693775e-03 2.93545961e-01
2.63927966e-01 5.43533415e-02 -1.11920118e+00 5.99374294e-01
9.33593333e-01 6.87156916e-01 -6.19543493e-01 1.25029576e+00
-7.74368122e-02 3.78549933e-01 -1.20016150e-02 6.22589886e-01
-1.72362193e-01 2.00000629e-01 1.46606183e+00 1.15088379e+00
3.69608432e-01 -2.34612953e-02 1.81201294e-01 4.24417049e-01
-2.19739624e-03 -1.54651240e-01 -8.61228108e-01 -2.90582143e-02
1.24581099e+00 8.93354416e-01 -6.35401845e-01 -4.84040707e-01
7.84957316e-03 1.18304491e+00 -2.42153063e-01 2.87059039e-01
-8.98798108e-01 -1.11925542e-01 7.58100927e-01 4.29126859e-01
5.25982618e-01 -2.13008404e-01 3.25246274e-01 -1.12156403e+00
1.05574183e-01 -1.27442420e+00 8.89858305e-01 -9.13710356e-01
-8.24049115e-01 4.88798916e-01 3.96146089e-01 -1.76457059e+00
-4.56268311e-01 4.82926637e-01 -4.83050436e-01 2.88357913e-01
-7.12568402e-01 -9.24643278e-01 -1.95743933e-01 8.92455459e-01
4.91178334e-01 -3.70375991e-01 4.34614033e-01 -1.47434577e-01
-1.97920978e-01 4.05967295e-01 -3.04581076e-01 -5.21931835e-02
1.09636676e+00 -4.21532273e-01 9.39190835e-02 1.38390768e+00
-6.88486472e-02 6.78349435e-01 1.03331828e+00 -1.45600498e+00
-1.70551848e+00 -1.01747859e+00 6.38671339e-01 -3.26609343e-01
1.00197220e+00 -1.43414503e-02 -9.56727266e-01 7.99674869e-01
6.14732146e-01 -4.89122681e-02 5.50454140e-01 -6.39037848e-01
-9.23660278e-01 -1.85733899e-01 -1.56030107e+00 4.10639316e-01
9.86541629e-01 -9.66748655e-01 -7.10578382e-01 1.28185645e-01
6.54804945e-01 -4.87026811e-01 -1.11600733e+00 1.88086450e-01
7.51236081e-01 -1.07318282e+00 8.90240073e-01 -6.13702774e-01
1.03438094e-01 -8.31173241e-01 -2.21371159e-01 -8.53037834e-01
-3.26276362e-01 -1.29171944e+00 -2.96256959e-01 1.14117241e+00
-4.83790874e-01 -2.39098165e-02 4.42879021e-01 7.92819798e-01
8.49039331e-02 6.37653247e-02 -1.42032850e+00 -7.88897514e-01
-9.96743202e-01 -2.59205014e-01 5.31566024e-01 1.25178790e+00
6.26740992e-01 -4.21606362e-01 -6.73769772e-01 5.03715158e-01
7.19675303e-01 -5.42458780e-02 7.08143830e-01 -4.98870850e-01
-1.00966439e-01 -1.72474861e-01 -8.26051772e-01 -4.96426225e-01
-7.51627833e-02 -5.26503563e-01 -5.72655573e-02 -1.03526139e+00
3.08559716e-01 -7.44079705e-03 1.42232999e-01 4.63577926e-01
2.62957990e-01 2.79326341e-03 3.45721781e-01 7.78753459e-01
-9.81348932e-01 5.60119301e-02 6.16174579e-01 -2.42644399e-01
-1.07471317e-01 -3.07540029e-01 -1.36266246e-01 3.39161068e-01
4.99743283e-01 -6.89131141e-01 -1.39284626e-01 -1.60097182e-01
3.05933893e-01 4.36232120e-01 6.70661271e-01 -1.07635534e+00
7.14278340e-01 -5.76674879e-01 2.80164659e-01 -6.84699655e-01
-1.37973696e-01 -1.02828348e+00 9.10387397e-01 5.20894647e-01
-5.74991941e-01 5.48204780e-01 1.80217206e-01 9.20717478e-01
-2.08954066e-01 -3.08651656e-01 4.75387782e-01 -1.69144198e-01
-1.01362884e+00 2.44095981e-01 -8.95325303e-01 -2.28250310e-01
1.62151027e+00 -7.07568765e-01 -5.47819972e-01 -5.89856625e-01
-4.66830552e-01 2.11052209e-01 1.12869644e+00 5.24203360e-01
9.18230593e-01 -1.42626250e+00 -5.34186482e-01 1.03372470e-01
1.68483570e-01 -3.77860218e-01 4.37546492e-01 7.63865173e-01
-6.49821639e-01 -3.00903320e-02 -4.01189119e-01 -3.83926123e-01
-1.91534555e+00 1.09159660e+00 3.39940712e-02 2.48859882e-01
-5.39210796e-01 4.59130406e-01 -5.77472985e-01 6.60632074e-01
3.76291990e-01 -4.03906927e-02 2.64640957e-01 2.12300375e-01
1.06491125e+00 8.81411016e-01 -3.64362150e-02 -9.87000048e-01
-6.37208164e-01 2.35716030e-01 -1.58366784e-01 -2.00534865e-01
1.02044761e+00 -6.62791073e-01 -1.06613599e-01 3.19514759e-02
1.08178687e+00 4.95260537e-01 -1.49153590e+00 -1.67827576e-01
7.09741861e-02 -1.13793147e+00 -3.48237008e-02 -7.06825376e-01
-7.77454853e-01 -7.53014758e-02 3.23851913e-01 3.67212705e-02
1.11524558e+00 -7.38111511e-03 8.05632174e-01 -8.19161069e-03
7.60817647e-01 -1.07238424e+00 2.10082412e-01 5.28623238e-02
1.06302857e+00 -5.83481610e-01 7.30827630e-01 -1.67167932e-01
-7.58450389e-01 1.27037644e+00 3.52223933e-01 8.94948095e-02
3.91313642e-01 3.87488067e-01 -1.72958165e-01 -3.91435355e-01
-8.80791247e-01 4.87444550e-01 1.38874665e-01 8.41891408e-01
-3.88056755e-01 -3.89086187e-01 -3.62033769e-02 -2.08361879e-01
3.00256252e-01 1.29899651e-01 8.13701570e-01 1.38911891e+00
-4.10520792e-01 -7.53413498e-01 -8.97449136e-01 4.78498042e-02
-2.55021662e-01 5.02875671e-02 -4.58198696e-01 5.40559053e-01
2.74332851e-01 9.27253187e-01 4.75839764e-01 -7.57955015e-01
1.14576057e-01 -3.58233362e-01 4.41920340e-01 -1.96367145e-01
-1.02610886e+00 3.21764976e-01 2.59361327e-01 -1.32886779e+00
-5.73447943e-01 -1.01290381e+00 -1.12343442e+00 -6.68238580e-01
-4.63454006e-03 9.60172340e-02 3.34619761e-01 4.68391478e-01
8.65235031e-01 -3.22237670e-01 8.39045644e-01 -8.68954718e-01
-3.05297643e-01 -2.53057569e-01 -1.61434695e-01 6.98294640e-01
5.04029095e-01 -4.19758260e-01 -1.19309843e-01 4.61172611e-01] | [12.305644035339355, 1.0001720190048218] |
52a46cf9-8e7d-4e9d-9080-c487bffeca97 | disassembling-the-dataset-a-camera-alignment | 2001.08680 | null | https://arxiv.org/abs/2001.08680v3 | https://arxiv.org/pdf/2001.08680v3.pdf | Rethinking the Distribution Gap of Person Re-identification with Camera-based Batch Normalization | The fundamental difficulty in person re-identification (ReID) lies in learning the correspondence among individual cameras. It strongly demands costly inter-camera annotations, yet the trained models are not guaranteed to transfer well to previously unseen cameras. These problems significantly limit the application of ReID. This paper rethinks the working mechanism of conventional ReID approaches and puts forward a new solution. With an effective operator named Camera-based Batch Normalization (CBN), we force the image data of all cameras to fall onto the same subspace, so that the distribution gap between any camera pair is largely shrunk. This alignment brings two benefits. First, the trained model enjoys better abilities to generalize across scenarios with unseen cameras as well as transfer across multiple training sets. Second, we can rely on intra-camera annotations, which have been undervalued before due to the lack of cross-camera information, to achieve competitive ReID performance. Experiments on a wide range of ReID tasks demonstrate the effectiveness of our approach. The code is available at https://github.com/automan000/Camera-based-Person-ReID. | ['Zijie Zhuang', 'Tianyu Zhang', 'Longhui Wei', 'Haizhou Ai', 'Qi Tian', 'Lingxi Xie', 'Hengheng Zhang', 'Haozhe Wu'] | 2020-01-23 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1493_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123570137.pdf | eccv-2020-8 | ['direct-transfer-person-re-identification'] | ['computer-vision'] | [-1.33217171e-01 -2.93281287e-01 4.39535007e-02 -6.06973171e-01
-6.14797890e-01 -8.98248851e-01 4.92500752e-01 -1.62063807e-01
-6.24233961e-01 5.56883872e-01 2.03165814e-01 1.47364259e-01
9.47068483e-02 -2.34204620e-01 -7.44598389e-01 -6.69842482e-01
2.86051035e-01 2.35036016e-01 -4.63436432e-02 6.89757690e-02
6.79814145e-02 3.80092263e-01 -1.26701975e+00 -1.10820457e-01
6.40058517e-01 4.13415879e-01 2.37424094e-02 6.09744608e-01
1.86849236e-01 4.58119303e-01 -5.30444622e-01 -1.12456989e+00
6.70845985e-01 -3.79270375e-01 -7.64634132e-01 1.32032484e-01
9.42800999e-01 -4.88345414e-01 -7.35142648e-01 1.24119604e+00
5.03667414e-01 2.17206672e-01 4.11555141e-01 -1.44180262e+00
-8.98667753e-01 2.85281271e-01 -8.34011674e-01 1.17300525e-01
4.67276424e-01 -1.86335355e-01 7.00579882e-01 -8.36601615e-01
4.47650135e-01 1.17435908e+00 8.47254932e-01 9.77255583e-01
-1.11903656e+00 -9.39443827e-01 2.35569581e-01 1.57504454e-01
-1.70836699e+00 -6.06205404e-01 5.05134284e-01 -4.18523997e-01
3.69924456e-01 2.26963788e-01 4.78581131e-01 1.26657343e+00
-2.41191700e-01 6.50023699e-01 8.64805520e-01 -4.06507045e-01
-2.33390301e-01 3.99510026e-01 1.29968241e-01 3.67176235e-01
5.04748344e-01 7.34171718e-02 -5.14566422e-01 1.25417998e-02
8.03387046e-01 3.55777383e-01 -4.36079115e-01 -5.50707102e-01
-1.27436554e+00 5.29955328e-01 6.55495763e-01 1.03308655e-01
2.33798936e-01 -1.22483075e-01 3.07766020e-01 8.39870647e-02
3.00929517e-01 2.91116685e-01 -1.61933556e-01 1.43554389e-01
-6.29208744e-01 2.31214732e-01 5.80142021e-01 1.45444846e+00
7.74750352e-01 -4.92108166e-01 1.65241569e-01 9.19151247e-01
-4.18920182e-02 6.08182371e-01 2.76711196e-01 -8.91399503e-01
6.66319191e-01 3.89931887e-01 2.28762001e-01 -1.02930963e+00
-1.12073749e-01 -2.90553302e-01 -9.86427128e-01 -1.27047956e-01
6.72165036e-01 -1.52678818e-01 -5.32661855e-01 1.91555274e+00
3.70122850e-01 3.79169643e-01 -9.51718241e-02 9.27666426e-01
5.12036383e-01 2.73575008e-01 1.06935836e-01 2.40913793e-01
1.26055443e+00 -1.04252326e+00 -4.90125239e-01 -4.11540538e-01
5.06219745e-01 -8.27804863e-01 6.92506492e-01 8.11661854e-02
-8.86721849e-01 -7.15513110e-01 -1.01939905e+00 -1.43652633e-01
-4.11237329e-01 2.15824634e-01 4.87438560e-01 6.55699134e-01
-1.15065289e+00 2.88246095e-01 -6.57208920e-01 -8.04937303e-01
4.31366295e-01 5.51275969e-01 -9.95983899e-01 -4.58739638e-01
-7.63472557e-01 8.76908302e-01 2.92401552e-01 2.89663076e-01
-4.69333619e-01 -6.32613301e-01 -9.64699268e-01 -1.73466340e-01
1.11959040e-01 -9.10434246e-01 1.29239595e+00 -1.02014005e+00
-1.12077951e+00 1.07051146e+00 -4.33950275e-01 -1.41176376e-02
8.98625314e-01 -3.75882924e-01 -3.39749098e-01 1.94613729e-02
3.81917626e-01 8.45152199e-01 6.43727779e-01 -1.45273209e+00
-7.77242839e-01 -4.32786465e-01 6.03369996e-02 3.82483989e-01
-6.58563912e-01 1.71171233e-01 -8.90781283e-01 -5.73027492e-01
-1.00229479e-01 -1.23457515e+00 9.55439508e-02 -3.85082178e-02
-5.16055048e-01 -3.99910919e-02 6.41083419e-01 -6.16786063e-01
8.18885207e-01 -2.35443163e+00 7.72874206e-02 8.98219794e-02
2.30079949e-01 2.51044750e-01 -2.53601730e-01 3.24805468e-01
-3.71877819e-01 7.49811251e-03 -6.10503480e-02 -9.11400855e-01
-2.14032248e-01 1.12075217e-01 -1.72785535e-01 8.09948683e-01
-3.54995802e-02 7.85338402e-01 -8.29356551e-01 -4.79143739e-01
1.68320879e-01 4.71758425e-01 -3.78921986e-01 4.35487181e-01
6.70766234e-01 7.68109143e-01 -1.69774994e-01 4.30055827e-01
1.09050930e+00 -1.51736334e-01 5.88326231e-02 -1.32214665e-01
6.70776740e-02 -1.75384536e-01 -1.27461636e+00 1.60065961e+00
-7.60667324e-02 5.73184550e-01 1.32281485e-03 -7.93337405e-01
7.23516226e-01 1.39816299e-01 2.93374121e-01 -4.59322095e-01
2.15632981e-03 -6.71953708e-02 -3.06090504e-01 -3.36583048e-01
4.71749693e-01 1.50065124e-02 -8.96095783e-02 3.20933223e-01
1.06862471e-01 3.54447424e-01 1.60429209e-01 2.89657116e-01
6.63578629e-01 -3.81582454e-02 1.27737552e-01 -1.70043290e-01
5.45684755e-01 -1.54227868e-01 7.58328974e-01 8.37812424e-01
-4.20434058e-01 9.41837549e-01 3.77188772e-02 -3.99267614e-01
-1.23790050e+00 -1.22183108e+00 -8.85488391e-02 9.69117045e-01
6.19295180e-01 -3.36236328e-01 -9.11904097e-01 -7.42134571e-01
1.65953711e-01 2.73337752e-01 -7.02216744e-01 -1.45006225e-01
-6.31731987e-01 -6.77995503e-01 5.10951817e-01 8.23660135e-01
6.07131898e-01 -2.94424355e-01 1.49619877e-01 -2.17918634e-01
-3.09058785e-01 -1.46114814e+00 -1.03147519e+00 -4.27975357e-01
-6.36372924e-01 -1.27057636e+00 -1.10788715e+00 -9.43257987e-01
1.16456866e+00 1.03673065e+00 6.24331474e-01 2.29625300e-01
-2.41026208e-01 7.54927933e-01 -3.43013048e-01 -3.20469201e-01
-1.01450793e-01 7.89046735e-02 4.73277986e-01 2.04801589e-01
6.32128417e-01 -3.07013184e-01 -6.52844429e-01 7.75355935e-01
-4.97621298e-01 -1.24484912e-01 4.15661126e-01 8.68554294e-01
2.82676011e-01 -2.99788266e-02 4.01699364e-01 -7.18186378e-01
3.56305033e-01 -2.77228862e-01 -6.30177200e-01 4.51305360e-01
-4.01384056e-01 -4.48983043e-01 3.40474159e-01 -5.94224215e-01
-1.21659231e+00 2.00725704e-01 3.40923578e-01 -4.13736582e-01
-2.52605468e-01 -1.59609392e-01 -3.39110523e-01 -2.83259898e-01
4.29221243e-01 1.15282573e-01 -1.49960201e-02 -5.60874283e-01
3.46040368e-01 7.01066971e-01 9.45080578e-01 -5.43704689e-01
1.16746891e+00 6.32217705e-01 -4.16459739e-01 -5.87213576e-01
-9.31256413e-01 -8.84220064e-01 -1.12091196e+00 -1.81355596e-01
9.46932256e-01 -1.35238671e+00 -7.31099486e-01 6.57074928e-01
-1.19250393e+00 -7.77280629e-02 3.23383212e-02 7.03026474e-01
-2.00714782e-01 7.29339600e-01 -6.40157163e-01 -5.97504497e-01
-1.02700613e-01 -9.08292353e-01 9.15692210e-01 6.35545194e-01
-1.57058030e-01 -9.37241435e-01 -1.49456551e-02 6.19504273e-01
1.18874088e-01 -1.50424466e-01 2.90821373e-01 -6.47521973e-01
-4.99889404e-01 -7.69065678e-01 -3.64902228e-01 4.34358716e-01
2.68771887e-01 -1.57660320e-01 -1.16531980e+00 -6.97086811e-01
-2.13429302e-01 -8.45638365e-02 7.10870206e-01 7.58749843e-02
1.05639386e+00 -1.96419239e-01 -6.21358573e-01 8.43826115e-01
1.12234557e+00 -1.52602300e-01 5.20113468e-01 4.26348776e-01
1.03530192e+00 7.37073660e-01 3.00423831e-01 1.57280296e-01
8.11321735e-01 7.26037383e-01 1.12211816e-01 -2.48518139e-01
-7.64761716e-02 -4.57134813e-01 2.82587171e-01 7.15468466e-01
-2.68592179e-01 -2.66987383e-01 -8.46232295e-01 5.46596229e-01
-1.84947085e+00 -9.58949149e-01 -1.25868738e-01 2.54645586e+00
5.08003592e-01 -2.73285776e-01 2.96751648e-01 -3.06521267e-01
1.37826872e+00 -2.51721174e-01 -4.39756423e-01 1.18023254e-01
-1.86094880e-01 -3.56998682e-01 7.82988966e-01 4.27772194e-01
-1.30986214e+00 7.86805749e-01 6.16105509e+00 3.71869802e-01
-6.62794292e-01 1.90791443e-01 4.58866656e-01 -5.73680513e-02
3.21968079e-01 2.29866058e-02 -1.16049254e+00 5.41780293e-01
5.19065917e-01 -3.09354126e-01 4.62555021e-01 8.85192335e-01
-1.94132179e-01 3.86342034e-02 -1.48780382e+00 1.41742969e+00
4.06010151e-01 -9.18461680e-01 -1.24803744e-01 2.26965278e-01
8.55963290e-01 -1.08034134e-01 4.88380492e-02 5.98285124e-02
2.21889064e-01 -7.75302768e-01 6.69769824e-01 3.88382554e-01
8.03496182e-01 -6.58556283e-01 1.09303045e+00 1.43248364e-01
-1.15894389e+00 -1.25306293e-01 -7.31667936e-01 5.81061691e-02
7.02212006e-02 2.24097818e-01 -5.72441757e-01 6.76771045e-01
9.11856890e-01 8.29719484e-01 -7.80239701e-01 1.32763696e+00
-2.38785088e-01 9.58839711e-03 -1.91629052e-01 4.56149966e-01
-2.43713662e-01 -2.53500581e-01 3.00423115e-01 1.31670332e+00
4.28058654e-01 1.20420039e-01 1.25343487e-01 5.58823526e-01
-3.53100121e-01 -1.03929520e-01 -6.07362032e-01 3.80307168e-01
7.21614897e-01 1.20956683e+00 -3.18942010e-01 -2.84316838e-01
-6.85052574e-01 1.38316631e+00 4.57538873e-01 5.55200756e-01
-7.56361604e-01 -2.06091046e-01 8.25561941e-01 2.70982292e-02
9.25439373e-02 -2.61650443e-01 -2.88801156e-02 -1.52212870e+00
3.05399507e-01 -6.64215505e-01 6.63154185e-01 -5.58306754e-01
-1.89899707e+00 4.39442724e-01 1.37234464e-01 -1.41781592e+00
7.63732269e-02 -5.93251169e-01 -6.61187887e-01 1.02066302e+00
-1.34913409e+00 -1.38257694e+00 -7.37325191e-01 8.64331901e-01
2.09032923e-01 -1.42559960e-01 6.85627222e-01 6.43977642e-01
-9.55397725e-01 1.19596744e+00 1.35165617e-01 7.20136106e-01
1.38730085e+00 -1.12562776e+00 4.69182849e-01 1.12527645e+00
-1.62398577e-01 9.51206148e-01 3.80311728e-01 -5.26429057e-01
-1.24272859e+00 -1.13099718e+00 8.41660142e-01 -8.66663575e-01
4.69066769e-01 -7.36952364e-01 -9.35297310e-01 1.10952854e+00
8.01682696e-02 1.12104535e-01 8.86866331e-01 2.61018842e-01
-7.58257866e-01 -3.60766053e-01 -9.99888599e-01 5.30750215e-01
1.21369135e+00 -6.71385884e-01 -4.90750134e-01 3.38300169e-01
2.25014776e-01 -4.62109268e-01 -8.16284895e-01 3.60359102e-02
5.69956660e-01 -9.35004592e-01 1.15178430e+00 -6.05430424e-01
-4.22708206e-02 -3.46167773e-01 -1.09643124e-01 -1.11392736e+00
-4.11220908e-01 -5.32469869e-01 2.82362461e-01 1.71375537e+00
1.27532721e-01 -9.57212746e-01 5.66059947e-01 1.20161402e+00
1.33644074e-01 -2.32615322e-02 -7.75861621e-01 -1.14164531e+00
1.43752083e-01 1.69372559e-03 7.48171508e-01 1.22524691e+00
4.24025953e-02 1.43892720e-01 -6.80201113e-01 5.73668480e-01
9.24877822e-01 -2.66520083e-01 1.24347782e+00 -1.17800581e+00
-1.04362540e-01 -2.88949549e-01 -6.42175198e-01 -1.15794587e+00
2.85670668e-01 -7.99173832e-01 -6.99567050e-02 -1.11685121e+00
8.72167408e-01 -5.25469899e-01 -2.32545868e-01 4.35529560e-01
-4.98608142e-01 3.80141526e-01 4.54022735e-01 6.10761821e-01
-6.10311210e-01 3.10674548e-01 1.00539064e+00 -3.62414457e-02
8.16804394e-02 4.89336550e-02 -9.48638082e-01 6.54857457e-01
8.69868040e-01 -5.35449505e-01 -2.12691098e-01 -7.85335720e-01
-2.05436498e-01 -4.48712707e-01 5.93265116e-01 -9.76204276e-01
6.20525956e-01 1.54342890e-01 8.77160132e-01 -2.04535261e-01
3.50331664e-01 -8.05954814e-01 2.52834648e-01 8.10108185e-02
-1.66248143e-01 3.65913033e-01 1.31274939e-01 6.42041206e-01
-1.26777604e-01 -1.33537665e-01 8.55900168e-01 -6.90742657e-02
-6.40571237e-01 5.14437735e-01 2.57505238e-01 -2.26912126e-02
1.07981730e+00 -3.74212295e-01 -4.97413218e-01 -4.02028382e-01
-4.38406199e-01 3.79970580e-01 9.46007848e-01 5.81928074e-01
3.01941395e-01 -1.37544596e+00 -6.77966654e-01 2.35085532e-01
4.01240319e-01 -5.45227043e-02 5.29945731e-01 7.52198935e-01
-2.93837488e-01 2.54700810e-01 -2.13449121e-01 -6.93719268e-01
-1.44807065e+00 7.85298884e-01 4.65177417e-01 2.91274846e-01
-7.17024863e-01 9.22260165e-01 6.03947461e-01 -5.79145432e-01
2.75736839e-01 3.65002990e-01 1.45513207e-01 -1.24930829e-01
9.59847569e-01 3.79665673e-01 -1.60167217e-01 -9.94933903e-01
-5.22042811e-01 9.73918200e-01 -4.04276490e-01 1.48149595e-01
1.12165010e+00 -5.29690087e-01 -9.24592465e-02 -6.76520616e-02
1.37388384e+00 2.99835335e-02 -1.44213963e+00 -2.29134202e-01
-1.74384996e-01 -9.43421304e-01 -4.03822571e-01 -4.85934645e-01
-9.17776644e-01 6.92815185e-01 6.91765368e-01 -1.52119428e-01
8.88834059e-01 1.05747417e-01 6.93476021e-01 3.84580791e-01
4.24701273e-01 -1.02963161e+00 -8.28461573e-02 2.20384449e-01
6.08144522e-01 -1.57288527e+00 6.43282980e-02 -5.42653263e-01
-6.49401426e-01 9.59946513e-01 7.48707652e-01 7.34348819e-02
3.00634325e-01 1.23019040e-01 2.88290858e-01 1.99682415e-01
-4.20593144e-03 -1.18873711e-03 1.75026193e-01 8.51317823e-01
2.20206574e-01 8.95488560e-02 2.30024949e-01 3.94612581e-01
-2.12510914e-01 -1.63765356e-01 5.98487914e-01 7.54875243e-01
6.92323819e-02 -1.30214095e+00 -6.17096901e-01 -9.46355052e-03
-2.52334088e-01 1.95207313e-01 -4.68209118e-01 9.87838209e-01
1.00461945e-01 1.05446613e+00 1.52105212e-01 -3.80231678e-01
4.20376062e-01 -1.29674658e-01 5.83777785e-01 -5.19436836e-01
-2.84234196e-01 -1.70103952e-01 -2.62766063e-01 -4.46664959e-01
-4.83675867e-01 -8.71766090e-01 -6.66287899e-01 -8.02368999e-01
-3.20836037e-01 1.26742259e-01 5.06730080e-01 6.96828187e-01
4.46844399e-01 -1.16801448e-01 6.69528127e-01 -9.48964894e-01
-4.87638831e-01 -7.36573875e-01 -5.34682155e-01 7.21316874e-01
4.53203917e-01 -6.37495458e-01 -3.95863414e-01 3.86932045e-01] | [14.698918342590332, 0.9948827624320984] |
00a307a7-5015-4cde-8145-b9a4383e3fc1 | inducing-early-neural-collapse-in-deep-neural | 2209.08378 | null | https://arxiv.org/abs/2209.08378v3 | https://arxiv.org/pdf/2209.08378v3.pdf | Linking Neural Collapse and L2 Normalization with Improved Out-of-Distribution Detection in Deep Neural Networks | We propose a simple modification to standard ResNet architectures--L2 normalization over feature space--that substantially improves out-of-distribution (OoD) performance on the previously proposed Deep Deterministic Uncertainty (DDU) benchmark. We show that this change also induces early Neural Collapse (NC), an effect linked to better OoD performance. Our method achieves comparable or superior OoD detection scores and classification accuracy in a small fraction of the training time of the benchmark. Additionally, it substantially improves worst case OoD performance over multiple, randomly initialized models. Though we do not suggest that NC is the sole mechanism or a comprehensive explanation for OoD behaviour in deep neural networks (DNN), we believe NC's simple mathematical and geometric structure can provide a framework for analysis of this complex phenomenon in future work. | ['Bernhard Rabus', 'William Yolland', 'Jarrod Haas'] | 2022-09-17 | null | null | null | null | ['l2-regularization'] | ['methodology'] | [-1.13445930e-01 2.56472945e-01 8.24549794e-02 -4.02697027e-01
-4.30343539e-01 -5.80197752e-01 1.07907832e+00 8.28704089e-02
-6.44712031e-01 7.37781703e-01 2.75967181e-01 -5.20080745e-01
-2.16498375e-01 -7.76794255e-01 -8.90457988e-01 -7.36494541e-01
-3.46933156e-01 2.74453491e-01 5.04342437e-01 7.44338194e-03
1.43728524e-01 7.66627252e-01 -1.57326007e+00 -1.05511129e-01
6.15346849e-01 1.18578899e+00 -1.89922284e-02 6.19339943e-01
-7.01356083e-02 7.82527089e-01 -1.07415259e+00 -3.73330146e-01
4.30004597e-01 -1.15177907e-01 -5.45580804e-01 -5.11742890e-01
7.16070116e-01 -4.61287528e-01 -6.46644652e-01 1.24042034e+00
8.33221078e-01 1.79657772e-01 1.08522987e+00 -1.20846701e+00
-6.95858777e-01 6.89837813e-01 -1.96076319e-01 6.96653426e-01
-4.10785317e-01 -1.87647417e-01 1.00642288e+00 -7.71819592e-01
6.71823978e-01 1.26736617e+00 7.28307843e-01 6.81731045e-01
-1.27891314e+00 -6.85184300e-01 3.74227809e-03 -1.53088599e-01
-1.29848909e+00 -3.57429564e-01 4.88325745e-01 -2.27199763e-01
1.21240902e+00 2.01155022e-02 4.43374157e-01 1.16673493e+00
6.98737144e-01 9.17766750e-01 9.59078491e-01 -3.16900134e-01
7.25686252e-01 -1.12507120e-01 4.28077787e-01 5.12862623e-01
6.88068032e-01 4.64894474e-01 -3.20278019e-01 5.05417511e-02
7.81711459e-01 -3.87129039e-01 -6.01997487e-02 -1.08717181e-01
-7.82506406e-01 6.50412142e-01 4.58667278e-01 4.00786996e-01
-9.54340026e-02 6.63720489e-01 3.67651790e-01 4.81679291e-01
6.62100792e-01 4.62934285e-01 -5.13677239e-01 -2.45924801e-01
-9.73950446e-01 5.51674306e-01 8.77752244e-01 9.17794883e-01
4.82309014e-01 3.49454194e-01 -9.18293595e-02 6.65306866e-01
3.60563934e-01 2.92039841e-01 5.75609565e-01 -1.17490494e+00
2.14244545e-01 1.37459159e-01 -1.26858309e-01 -7.59894073e-01
-6.67506039e-01 -7.20862150e-01 -6.64348960e-01 5.12496948e-01
7.21534014e-01 -1.89287871e-01 -1.10287416e+00 2.03113437e+00
-1.72666579e-01 1.03541262e-01 1.12135775e-01 4.79642838e-01
8.70339513e-01 3.24485511e-01 -5.47626661e-03 1.56601101e-01
1.04784632e+00 -5.34624577e-01 -5.68798244e-01 -1.77712694e-01
6.68714881e-01 -3.65804255e-01 7.25295424e-01 2.62019902e-01
-8.79208922e-01 -1.48588002e-01 -1.26848221e+00 -7.33089000e-02
-6.52925849e-01 -3.56159091e-01 6.98552549e-01 9.35948431e-01
-1.16123140e+00 1.00163615e+00 -8.84955227e-01 -1.83674246e-01
7.50513077e-01 3.99233788e-01 5.43543249e-02 7.77222365e-02
-1.38837159e+00 1.03131866e+00 3.57624948e-01 -1.28271639e-01
-1.01529026e+00 -8.07070851e-01 -7.24055648e-01 -6.36116788e-02
-1.53699979e-01 -4.87881362e-01 1.30668795e+00 -2.27528334e-01
-1.37425005e+00 6.95694685e-01 1.47782844e-02 -9.48667705e-01
7.70491302e-01 -4.20764863e-01 -1.85260564e-01 -2.48450607e-01
-1.92186281e-01 1.10468638e+00 6.73939168e-01 -1.26248813e+00
-2.53271192e-01 -2.69564033e-01 -1.11544177e-01 3.91913503e-02
-2.01479644e-01 -5.81139438e-02 -7.19673932e-02 -9.06265616e-01
1.43415153e-01 -8.83296728e-01 -9.86536294e-02 -1.43796131e-01
-5.51494122e-01 -6.01643682e-01 3.54120433e-01 -1.82757545e-02
1.01664448e+00 -1.93362522e+00 -3.21100324e-01 1.01429462e-01
5.01095414e-01 1.01928815e-01 -1.58056274e-01 3.81875932e-02
-2.40601048e-01 5.34159243e-01 -3.67080308e-02 -4.57249969e-01
4.70104426e-01 2.16793031e-01 -5.26828349e-01 6.09735847e-01
3.18855226e-01 9.73360717e-01 -7.54829168e-01 -2.62482554e-01
1.75377741e-01 3.87106568e-01 -4.58174497e-01 -2.61645287e-01
-2.90761173e-01 -2.64774382e-01 -1.29714683e-01 5.02558470e-01
8.61424327e-01 1.01171225e-01 -4.07673746e-01 8.49493220e-02
-4.72448170e-02 4.47146654e-01 -1.10038102e+00 1.31108510e+00
-3.02076310e-01 1.25329196e+00 -6.22148454e-01 -7.70427704e-01
8.45819712e-01 9.05456543e-02 2.40616933e-01 -6.61345720e-01
6.09544158e-01 2.50337899e-01 3.94946694e-01 -3.45342085e-02
4.12813723e-01 7.34507218e-02 5.69279753e-02 2.87512124e-01
4.22465503e-01 -1.54660687e-01 1.37079254e-01 4.11898606e-02
1.33349717e+00 -1.81488112e-01 1.37670726e-01 -8.50012600e-01
-2.87828058e-01 -3.01197022e-01 2.88203925e-01 1.28908360e+00
-4.55207974e-01 8.61474812e-01 8.09540153e-01 -3.55180532e-01
-1.10236406e+00 -1.42604351e+00 -8.47059965e-01 9.48536932e-01
8.40348527e-02 -2.92260289e-01 -9.61184978e-01 -6.25883043e-01
7.07258657e-02 1.12655616e+00 -8.02426875e-01 -2.62310266e-01
-1.57184318e-01 -1.19141805e+00 1.04470634e+00 5.21868229e-01
5.21134555e-01 -8.96291494e-01 -4.22413141e-01 2.05950975e-01
4.51749682e-01 -9.68956769e-01 -7.08850026e-02 8.95839989e-01
-1.01709926e+00 -6.82848752e-01 -9.26652014e-01 -5.56908429e-01
3.38906169e-01 -1.61122754e-01 1.04347289e+00 -2.65913218e-01
-2.06607312e-01 4.38379124e-02 9.69002321e-02 -9.37537909e-01
-5.95276892e-01 8.78606588e-02 3.97227705e-01 -6.84313715e-01
6.50844216e-01 -6.76016331e-01 -4.47443634e-01 1.99867502e-01
-9.81383145e-01 -5.29611826e-01 3.15127134e-01 6.55407667e-01
2.71364927e-01 5.25155067e-01 5.69612384e-01 -6.18468285e-01
8.53435636e-01 -4.54498470e-01 -6.76665962e-01 -1.55015633e-01
-8.04393589e-01 5.93060136e-01 6.13526464e-01 -4.95209247e-01
-9.24805343e-01 -3.87604088e-01 -3.47743392e-01 -5.10356784e-01
-3.77891064e-01 1.45373493e-01 1.21376179e-01 -1.93167269e-01
9.92136240e-01 -8.29858473e-04 -2.67361760e-01 -4.62575912e-01
1.63805142e-01 4.19488817e-01 4.61069643e-01 -5.22531390e-01
7.36721218e-01 3.05213660e-01 3.19466926e-02 -8.17341089e-01
-9.31195796e-01 4.46243808e-02 -3.51649582e-01 6.85662255e-02
7.98749149e-01 -9.30709183e-01 -5.04900038e-01 7.42990315e-01
-1.28628230e+00 -5.09580135e-01 -3.38934451e-01 5.92813194e-01
-2.47496724e-01 -1.06563523e-01 -6.63810372e-01 -8.34560394e-01
-1.92201100e-02 -1.09809816e+00 8.12796116e-01 4.43617076e-01
-2.03520477e-01 -1.21634209e+00 1.11398086e-01 -2.81714380e-01
4.92906988e-01 6.59282953e-02 8.23898673e-01 -1.19196856e+00
-3.71014684e-01 -2.24705458e-01 -4.09725457e-01 5.00842869e-01
-1.72765240e-01 1.62763581e-01 -1.38064635e+00 1.71950758e-02
4.53380421e-02 -3.64160240e-01 1.29469895e+00 7.64734328e-01
1.20717847e+00 5.92627451e-02 -1.75585955e-01 3.85942757e-01
1.34519339e+00 1.34266287e-01 7.37125516e-01 6.04242980e-01
3.18657398e-01 3.29802006e-01 1.19703613e-01 3.83437365e-01
2.23589942e-01 3.82078081e-01 6.59110725e-01 2.06487060e-01
-4.69274729e-01 -1.47384340e-02 2.85195500e-01 4.58212554e-01
-4.33360115e-02 -4.89141941e-01 -9.15060520e-01 4.53588814e-01
-1.41646397e+00 -9.18094158e-01 2.77715791e-02 2.09109926e+00
8.33728015e-01 8.26097727e-01 -1.06586099e-01 9.09575373e-02
6.18209720e-01 3.00826341e-01 -7.85301626e-01 -3.65518093e-01
-2.85842240e-01 2.49217361e-01 9.85667825e-01 3.93042237e-01
-9.61723030e-01 8.37966263e-01 7.73157310e+00 9.92520988e-01
-9.98260677e-01 1.46305814e-01 7.62572765e-01 -1.77033380e-01
-4.73733813e-01 -5.25292814e-01 -1.24389267e+00 4.80378598e-01
1.09866476e+00 -1.48058131e-01 3.49778116e-01 1.02218151e+00
-9.31723490e-02 -1.26178235e-01 -1.29326689e+00 7.46194065e-01
-5.05997054e-02 -1.54538357e+00 -5.13576493e-02 3.98040533e-01
9.62291718e-01 8.39381337e-01 1.90458089e-01 4.17593390e-01
7.80011892e-01 -1.08961606e+00 1.08235025e+00 2.58387566e-01
6.78226531e-01 -7.52152443e-01 8.81415725e-01 3.02070946e-01
-7.90433586e-01 5.09969369e-02 -8.27737153e-01 6.04909733e-02
-2.59534836e-01 8.93864870e-01 -6.54819250e-01 7.58715346e-02
7.25850523e-01 3.26999545e-01 -7.87223876e-01 1.10536754e+00
1.30236698e-02 6.91636324e-01 -7.47376025e-01 -3.70031923e-01
2.78084129e-01 4.12648886e-01 7.30083466e-01 1.11180294e+00
2.10847706e-01 -2.20833108e-01 -6.54392838e-01 1.18755388e+00
-3.73735368e-01 -4.42303360e-01 -7.71473527e-01 2.08511185e-02
8.11449111e-01 7.73972929e-01 -8.56764913e-01 -5.74036874e-02
-1.81627408e-01 5.21651030e-01 3.66493136e-01 1.62878484e-01
-9.49527800e-01 -7.65186131e-01 7.97717631e-01 -3.99558246e-01
1.32417187e-01 -2.11638108e-01 -6.72261834e-01 -1.16380382e+00
-1.63334385e-01 -3.37318689e-01 -1.53660169e-02 -4.94163752e-01
-1.28022289e+00 7.50387132e-01 1.13538779e-01 -9.40007627e-01
-2.18190327e-01 -1.01592255e+00 -6.62952363e-01 4.59574461e-01
-1.53966558e+00 -3.07007194e-01 -9.42683741e-02 2.42521256e-01
3.97722363e-01 -3.22537333e-01 7.14411020e-01 1.36366382e-01
-6.99901640e-01 8.63044024e-01 5.66436470e-01 2.12534726e-01
5.21009505e-01 -1.53407681e+00 8.48560035e-01 7.55588710e-01
3.96352500e-01 5.96137464e-01 1.19369614e+00 -3.63966823e-01
-7.38146782e-01 -1.07538724e+00 7.90981114e-01 -6.84254110e-01
8.05882037e-01 -3.88559610e-01 -6.51870847e-01 5.64236999e-01
3.57280895e-02 1.86237216e-01 3.13764274e-01 3.07924841e-02
-2.59431452e-01 4.77989279e-02 -1.12416029e+00 7.86519170e-01
1.17106390e+00 -4.29430008e-01 -1.05731201e+00 2.96583682e-01
8.82990360e-01 -4.65931654e-01 -1.01087022e+00 2.60631919e-01
5.17818868e-01 -1.17555714e+00 9.55832005e-01 -4.35904205e-01
4.21386898e-01 1.46744981e-01 -4.43099409e-01 -1.45342314e+00
-5.17673045e-02 -5.23933351e-01 -1.12342834e-01 1.23948240e+00
4.00423825e-01 -8.67521465e-01 8.26980472e-01 6.62714899e-01
-3.49265009e-01 -9.25758898e-01 -1.23560834e+00 -1.16543710e+00
5.81666768e-01 -1.09953296e+00 4.41710830e-01 5.23668468e-01
-3.35993826e-01 -4.67348807e-02 2.14310989e-01 9.99858752e-02
9.72528636e-01 -5.35833180e-01 2.16472879e-01 -1.43079472e+00
1.87091231e-01 -8.21472645e-01 -6.56584322e-01 -9.46279049e-01
3.03160995e-01 -8.82994056e-01 3.65872518e-03 -1.32866919e+00
-1.98719278e-01 -5.63529432e-01 -6.01333499e-01 2.07457751e-01
2.43143111e-01 4.07187253e-01 1.10892728e-01 1.10090755e-01
-5.43243408e-01 6.01214290e-01 9.21520233e-01 1.65306792e-01
-1.27439037e-01 2.26466823e-02 -6.52112663e-01 8.90282094e-01
8.46788049e-01 -9.07991529e-01 -3.35361332e-01 -5.38901329e-01
7.93001801e-02 -5.69891930e-01 4.38961923e-01 -1.36767149e+00
6.47021458e-02 1.90817371e-01 6.36431396e-01 -5.19203067e-01
9.54380929e-02 -5.84497273e-01 -4.61145163e-01 4.36713099e-01
-5.84587634e-01 -1.56324357e-01 4.76211727e-01 6.09958827e-01
1.22123696e-01 -6.26316607e-01 9.42868888e-01 -2.75672562e-02
-5.99128485e-01 2.97510326e-01 -5.84631920e-01 6.05677724e-01
8.89280558e-01 -1.49004877e-01 -7.70252466e-01 -2.20429257e-01
-6.19837105e-01 -2.55589578e-02 3.89375389e-01 5.24329185e-01
2.12264881e-01 -1.38184559e+00 -3.79442751e-01 1.16892569e-01
-7.52400979e-02 1.27224758e-01 -1.06448896e-01 3.80077183e-01
-6.67590499e-01 6.76335692e-01 -1.15591384e-01 -5.31219363e-01
-7.23018467e-01 -1.54151479e-02 6.90665424e-01 -4.24443111e-02
-4.13698822e-01 1.31888914e+00 1.51991382e-01 -4.54441875e-01
7.28842139e-01 -6.63951397e-01 -5.70314564e-03 1.37452945e-01
4.98850644e-01 3.94970834e-01 1.98393434e-01 -1.45257413e-01
-3.74878526e-01 1.67411998e-01 -2.06868276e-01 -1.48833171e-01
1.45211029e+00 -8.62989500e-02 2.46180475e-01 6.95061386e-01
1.27787352e+00 -1.17969118e-01 -1.57421875e+00 -1.24662153e-01
1.38833206e-02 -5.02572618e-02 5.21879017e-01 -6.06946528e-01
-9.11603630e-01 9.95338380e-01 7.27895260e-01 5.02178609e-01
5.40876687e-01 2.86739618e-01 4.56060797e-01 7.46944904e-01
2.32597440e-01 -9.93242085e-01 -8.99220072e-03 7.45136321e-01
6.90991402e-01 -1.07429421e+00 -6.51885942e-02 2.91321069e-01
-3.13471675e-01 1.24527633e+00 6.66426957e-01 -5.01593232e-01
1.04059339e+00 7.58679450e-01 -1.48005262e-01 -2.98295692e-02
-9.28624630e-01 -7.27890730e-02 1.76242709e-01 6.04432344e-01
2.43660107e-01 -3.50919455e-01 9.89582613e-02 6.01876736e-01
-5.67189097e-01 -1.23642221e-01 5.11124551e-01 6.12032890e-01
-6.06981337e-01 -6.73116505e-01 -8.28665216e-03 7.32992589e-01
-6.45533919e-01 -3.86926711e-01 -1.61275923e-01 9.43336427e-01
4.52856645e-02 5.78313529e-01 4.76585031e-01 -5.19361436e-01
3.68372761e-02 3.23198020e-01 5.36187589e-01 -4.15310293e-01
-1.51715234e-01 -1.80207819e-01 -1.35989045e-03 -4.85855371e-01
-1.44635543e-01 -5.87767720e-01 -1.36199784e+00 -5.25504947e-01
-5.77966511e-01 -2.87319183e-01 8.50839257e-01 1.13925374e+00
3.24184895e-01 7.13462532e-01 3.36043328e-01 -9.52704132e-01
-8.25249791e-01 -1.03479493e+00 -9.20024037e-01 3.80655304e-02
4.61027890e-01 -9.70472276e-01 -9.31549788e-01 -3.08367521e-01] | [7.567513942718506, 3.70713472366333] |
1f3b6f5b-5b39-4db6-8e9d-18ca25d507aa | learning-iterative-neural-optimizers-for | 2303.16206 | null | https://arxiv.org/abs/2303.16206v1 | https://arxiv.org/pdf/2303.16206v1.pdf | Learning Iterative Neural Optimizers for Image Steganography | Image steganography is the process of concealing secret information in images through imperceptible changes. Recent work has formulated this task as a classic constrained optimization problem. In this paper, we argue that image steganography is inherently performed on the (elusive) manifold of natural images, and propose an iterative neural network trained to perform the optimization steps. In contrast to classical optimization methods like L-BFGS or projected gradient descent, we train the neural network to also stay close to the manifold of natural images throughout the optimization. We show that our learned neural optimization is faster and more reliable than classical optimization approaches. In comparison to previous state-of-the-art encoder-decoder-based steganography methods, it reduces the recovery error rate by multiple orders of magnitude and achieves zero error up to 3 bits per pixel (bpp) without the need for error-correcting codes. | ['Kilian Q Weinberger', 'Varsha Kishore', 'Xiangyu Chen'] | 2023-03-27 | null | null | null | null | ['image-steganography'] | ['computer-vision'] | [ 1.02552485e+00 5.19402564e-01 4.59577031e-02 1.50778778e-02
-4.54778105e-01 -1.72583327e-01 5.37083030e-01 -4.08863187e-01
-5.30451596e-01 5.22889316e-01 6.64921254e-02 -6.47430480e-01
3.59532386e-01 -5.86826563e-01 -1.15161240e+00 -9.03239489e-01
-4.62037921e-01 -1.93031989e-02 -1.99414164e-01 -3.99525672e-01
5.58889568e-01 1.44687563e-01 -9.18596268e-01 -2.59595141e-02
6.51321411e-01 7.60087609e-01 -3.07972841e-02 1.07692587e+00
3.88864756e-01 1.05648029e+00 -3.92866164e-01 -3.58844638e-01
5.20276248e-01 -8.17698598e-01 -7.45838284e-01 4.14801091e-01
1.18245587e-01 -6.11293495e-01 -7.16628015e-01 1.45769763e+00
2.55958050e-01 -2.56521463e-01 2.59240001e-01 -8.85867417e-01
-1.00929856e+00 5.88157058e-01 -3.53571177e-01 -3.47015500e-01
1.13605507e-01 2.04979926e-01 6.08123243e-01 -7.25719512e-01
6.65105641e-01 1.08916140e+00 6.92134559e-01 6.73853695e-01
-1.14314628e+00 -4.58332360e-01 -2.83790231e-01 1.15359426e-01
-1.29628038e+00 -7.33052731e-01 8.86439443e-01 -9.48560014e-02
8.83295119e-01 4.88571465e-01 6.61613822e-01 7.33100832e-01
4.92666185e-01 6.31555974e-01 1.08616781e+00 -7.79107273e-01
-1.85567334e-01 1.11968204e-01 -6.04644775e-01 1.07696819e+00
2.67736167e-01 4.79222596e-01 -4.37865525e-01 1.14324540e-01
9.02136624e-01 -4.56735045e-02 -6.68423355e-01 -5.58593452e-01
-1.60838759e+00 1.15243375e+00 5.43321252e-01 2.41961360e-01
-2.71085620e-01 8.89274776e-01 8.34965482e-02 6.34365976e-01
3.95339578e-01 6.97396398e-01 -1.23955108e-01 4.93500046e-02
-1.01073325e+00 -3.46125662e-01 1.23282099e+00 5.49777448e-01
8.63311589e-01 5.27996719e-01 4.82005000e-01 1.23400241e-01
5.94657063e-01 5.39303541e-01 2.70982325e-01 -9.78895009e-01
3.77663374e-01 1.46235183e-01 -8.28106776e-02 -1.53107631e+00
1.27654478e-01 -1.92557111e-01 -1.27982926e+00 6.20482624e-01
1.44880235e-01 -2.28197590e-01 -9.44978595e-01 1.43407357e+00
-2.94488166e-02 1.44088849e-01 1.67095125e-01 7.19964385e-01
1.93790480e-01 8.27571988e-01 -6.70747399e-01 -3.01294655e-01
7.96846569e-01 -1.06841648e+00 -8.74918997e-01 -3.62052500e-01
5.43265462e-01 -7.24233449e-01 4.98271644e-01 2.06611827e-01
-1.15190721e+00 -1.14584483e-01 -1.48442614e+00 1.46898046e-01
-2.35126063e-01 -3.66820604e-01 3.52275968e-01 1.23539126e+00
-1.27976763e+00 9.35555816e-01 -7.98045814e-01 1.62466034e-01
5.49808562e-01 7.30176270e-01 -5.06892204e-01 1.14763724e-02
-1.08728528e+00 9.11904097e-01 5.39384067e-01 2.93155164e-01
-1.22727966e+00 -4.16953772e-01 -1.23471773e+00 -1.46741346e-01
2.86595851e-01 -4.16186899e-01 6.01425886e-01 -1.20150852e+00
-1.98286283e+00 1.03218639e+00 -1.14318162e-01 -8.47090721e-01
5.72350681e-01 -2.42650863e-02 -2.63856858e-01 1.76501483e-01
-5.37386179e-01 6.60499930e-01 1.68918538e+00 -1.28182232e+00
-1.40883505e-01 1.26311421e-01 -8.58805105e-02 -1.82046846e-01
-3.23321581e-01 4.59939502e-02 -3.70354682e-01 -6.23609543e-01
2.83476442e-01 -1.15697062e+00 -5.84377229e-01 3.26980680e-01
-5.19272804e-01 7.20517635e-01 1.04357564e+00 -9.58457708e-01
1.04043460e+00 -1.90466452e+00 4.24895167e-01 3.65476459e-01
5.95748663e-01 5.26918292e-01 -2.20276520e-01 3.42442960e-01
-1.26652241e-01 2.73833454e-01 -8.32304716e-01 -5.78452647e-01
7.90765509e-03 1.90549493e-01 -2.66602393e-02 1.21371043e+00
6.80268034e-02 1.23605978e+00 -7.95030892e-01 -2.68364996e-01
3.39825332e-01 8.39924395e-01 -5.70718825e-01 6.29598126e-02
6.91943467e-02 7.77010977e-01 2.37501338e-02 4.69660163e-01
8.96061480e-01 -5.80854893e-01 3.22459817e-01 9.37167928e-02
1.08138919e-01 2.18001574e-01 -8.07703078e-01 1.57155585e+00
-3.72310758e-01 1.19804275e+00 3.53750408e-01 -1.37737179e+00
7.18957424e-01 4.75493282e-01 4.61840928e-01 -4.11376148e-01
3.12436432e-01 3.57684284e-01 -8.48241523e-02 -4.17865783e-01
4.00009960e-01 -6.53802790e-03 1.95144355e-01 6.82920277e-01
-1.71908885e-01 -1.61805719e-01 -3.03423941e-01 -4.77639362e-02
9.97090220e-01 -9.09693614e-02 3.27650160e-01 -2.06168741e-01
6.84344649e-01 -2.97944903e-01 2.50580430e-01 7.63535917e-01
3.07444413e-03 4.88243014e-01 2.40436450e-01 -5.53111494e-01
-1.52095628e+00 -4.66048479e-01 4.96073477e-02 5.56121111e-01
4.38215703e-01 -3.41109186e-01 -8.97870541e-01 -5.23432672e-01
-3.16366374e-01 3.50191057e-01 -5.19753218e-01 -2.25831106e-01
-9.26150978e-01 -7.08267570e-01 7.38383234e-01 -3.75563890e-01
1.12604547e+00 -1.05755627e+00 -6.02183640e-01 3.68430316e-01
-1.69969320e-01 -1.17586792e+00 -6.62240624e-01 -7.89756402e-02
-8.26972425e-01 -9.62976098e-01 -8.83428693e-01 -9.83762264e-01
1.16903508e+00 1.16855957e-01 1.03550494e+00 6.31490231e-01
-1.30366802e-01 1.18705839e-01 -2.69385874e-01 -5.02170436e-02
-1.07463300e+00 -9.86350849e-02 -2.95225114e-01 2.45229915e-01
-7.37927929e-02 -5.87354183e-01 -6.11863792e-01 1.05246976e-01
-1.15990162e+00 1.59643501e-01 6.72661841e-01 1.10865736e+00
3.86685312e-01 3.10154140e-01 -3.12582344e-01 -9.18072820e-01
3.84495556e-01 -2.22143695e-01 -9.02746677e-01 9.26672146e-02
-8.95712018e-01 5.14807284e-01 3.91070932e-01 -5.25259793e-01
-2.94931978e-01 1.72339529e-01 -2.09213510e-01 -2.83160716e-01
3.57494861e-01 4.13596690e-01 1.92872137e-01 -1.28099954e+00
3.80504310e-01 6.82178736e-01 5.74578643e-01 -1.10617682e-01
2.50364095e-01 4.20120835e-01 3.55259031e-01 1.85774550e-01
1.37088215e+00 8.96911263e-01 2.10333228e-01 -7.90201664e-01
-2.12856963e-01 3.82795488e-03 -5.25623679e-01 -1.84311941e-01
8.10157001e-01 -6.12198591e-01 -9.24093544e-01 6.93479717e-01
-1.15126359e+00 -5.24335384e-01 -1.68650866e-01 3.73928547e-01
-7.04697132e-01 8.63526583e-01 -6.43911719e-01 -7.07126737e-01
-4.38294113e-01 -1.30114031e+00 7.55738139e-01 -3.78877223e-01
2.07439944e-01 -1.34031618e+00 -1.22207597e-01 1.35646001e-01
7.57990360e-01 7.26136804e-01 3.54253232e-01 6.06230553e-03
-1.11083615e+00 -3.74016672e-01 -9.41507891e-02 5.67935586e-01
1.31590083e-01 -5.21977961e-01 -5.63481629e-01 -7.99686551e-01
6.81012690e-01 -6.30141348e-02 1.16129851e+00 3.20133865e-01
9.78369236e-01 -1.11744416e+00 -1.10756218e-01 1.34596252e+00
1.74373794e+00 -3.23752426e-02 1.18480182e+00 5.00318229e-01
8.61725390e-01 3.79554242e-01 -1.92309171e-01 1.40741348e-01
1.29837587e-01 3.11631024e-01 8.97782743e-01 -2.18544185e-01
-3.07142977e-02 -1.83564261e-01 6.67251647e-01 7.94061065e-01
-3.85635272e-02 -3.87151480e-01 -5.13525724e-01 5.28044283e-01
-1.62438929e+00 -8.89105976e-01 1.11833297e-01 2.18198967e+00
6.88665092e-01 -1.89181592e-03 -3.93362463e-01 3.65816146e-01
8.13963711e-01 6.91693485e-01 -3.39051485e-01 -4.70757425e-01
-3.05059999e-01 2.00425833e-03 1.35753250e+00 9.17268753e-01
-1.24807227e+00 9.08285499e-01 7.51548576e+00 6.92492425e-01
-1.28733206e+00 1.70134589e-01 5.67845523e-01 1.59680724e-01
-2.85688818e-01 1.99780881e-01 -1.44382462e-01 4.07106638e-01
8.32883537e-01 1.83202595e-01 1.14695895e+00 3.91691953e-01
-2.70774662e-01 2.57111877e-01 -8.68632734e-01 1.12361014e+00
4.10940528e-01 -1.79294884e+00 -1.70278713e-01 6.03620052e-01
1.10417640e+00 -7.82120600e-03 4.59691763e-01 -4.51880634e-01
3.21007252e-01 -1.47010922e+00 4.19273943e-01 2.82964766e-01
9.30336952e-01 -5.43245614e-01 5.56319416e-01 3.66968900e-01
-8.33438039e-01 6.57746792e-02 -3.41873288e-01 8.89153033e-02
4.97213930e-01 4.31417614e-01 -6.40786171e-01 1.79446220e-01
3.16064775e-01 7.75647938e-01 -9.18784291e-02 6.32990539e-01
-5.10178983e-01 6.93916500e-01 -2.84746885e-01 1.14892628e-02
6.09436214e-01 -1.80095151e-01 1.11971891e+00 1.01028550e+00
4.00392026e-01 -2.03231558e-01 -1.75378144e-01 7.99117863e-01
-3.86340797e-01 -3.56916875e-01 -7.02875435e-01 -2.82927364e-01
-9.48605239e-02 6.51401222e-01 -5.09733796e-01 -2.21917406e-01
-8.96356329e-02 1.60337329e+00 -1.65049657e-01 4.00233567e-01
-4.62670267e-01 -5.54777145e-01 3.78732532e-01 -8.75846595e-02
7.86625803e-01 -5.35626888e-01 -2.52823383e-01 -1.30432630e+00
-3.70732024e-02 -1.11376488e+00 -3.49350631e-01 -2.40256742e-01
-6.28415644e-01 4.06232774e-01 -6.12064123e-01 -1.14586735e+00
-3.88367891e-01 -6.55319154e-01 -3.69812012e-01 6.31765842e-01
-1.86573017e+00 -1.06088901e+00 5.68581969e-02 6.25139534e-01
1.83894724e-01 -3.38269919e-01 9.22106981e-01 -1.53756794e-02
-2.76270419e-01 5.35945058e-01 6.06904149e-01 4.49969739e-01
1.35771126e-01 -1.02224219e+00 8.67816806e-01 1.24989629e+00
1.26586989e-01 4.75130588e-01 9.52046514e-01 -5.50970614e-01
-1.74087083e+00 -8.47800970e-01 9.70544875e-01 -1.02372617e-01
6.40103757e-01 -2.50226468e-01 -7.55432069e-01 7.81201065e-01
6.24254942e-01 -1.64843388e-02 2.68627048e-01 -7.02332377e-01
-3.90781224e-01 3.51686984e-01 -1.36556315e+00 5.37700713e-01
8.09475660e-01 -6.93011343e-01 -7.18461499e-02 5.18124342e-01
7.20646918e-01 -5.67350566e-01 -6.58456028e-01 5.24942316e-02
5.62811792e-01 -8.92947316e-01 1.20692599e+00 -1.71207234e-01
4.43076730e-01 -2.07115769e-01 -1.18374452e-01 -1.17010641e+00
-1.83799312e-01 -1.59275901e+00 -4.91685361e-01 2.62773454e-01
3.46982628e-01 -8.32996905e-01 1.02423322e+00 1.31689936e-01
1.69445813e-01 -5.29833138e-01 -1.12040079e+00 -6.54967427e-01
-6.97078109e-02 -1.39131531e-01 4.88109201e-01 9.81432796e-01
-1.45472780e-01 -2.90539324e-01 -1.33252215e+00 3.62696022e-01
1.23851490e+00 -2.23517492e-01 6.58050179e-01 -6.85157657e-01
-4.31164593e-01 -2.87420928e-01 -7.90817440e-01 -1.47246766e+00
1.57109350e-01 -8.25881958e-01 2.60981202e-01 -1.16714621e+00
-2.54683405e-01 -2.52125651e-01 1.66232176e-02 1.51772276e-01
5.96732832e-02 7.61653781e-01 2.72904992e-01 3.93200904e-01
-4.72104609e-01 5.02320647e-01 1.48388505e+00 -3.78880590e-01
5.03632315e-02 -1.45708695e-01 -5.61948180e-01 6.17017925e-01
8.66019964e-01 -8.75237286e-01 1.06958799e-01 -6.27470016e-01
3.86076689e-01 -1.44328386e-01 5.04583180e-01 -8.78108621e-01
4.19541478e-01 9.76716578e-02 9.33620259e-02 -7.72581398e-02
5.52556574e-01 -9.11195576e-01 1.77248374e-01 1.16949904e+00
-2.50922978e-01 -2.94689357e-01 -2.69324780e-01 5.25819004e-01
-1.79106086e-01 -3.57445717e-01 9.28533673e-01 -3.90989840e-01
-5.74031711e-01 2.74934411e-01 -5.79593122e-01 -3.25958818e-01
8.00679922e-01 -5.23962557e-01 -3.93981785e-02 -8.77810180e-01
-4.04293865e-01 -1.02044687e-01 6.74271822e-01 -5.39476285e-03
1.10503769e+00 -1.10585344e+00 -9.19031560e-01 5.52796960e-01
-3.39174509e-01 -2.50411958e-01 -1.01035811e-01 7.52230704e-01
-1.22251070e+00 3.77095759e-01 -6.59014955e-02 -4.83219713e-01
-1.29303932e+00 6.52078211e-01 6.58045888e-01 -4.70832705e-01
-7.44209051e-01 9.86580491e-01 -3.07568550e-01 -3.94299358e-01
1.06807545e-01 2.77998894e-01 1.14810169e-01 -6.91803992e-01
6.04024112e-01 3.38349938e-01 -4.07432675e-01 -9.17488515e-01
-8.14388394e-02 7.34463274e-01 1.10035062e-01 -2.60665148e-01
1.38632047e+00 -3.89237911e-01 -6.36390865e-01 -1.72079816e-01
1.81868446e+00 -7.02438056e-02 -1.34904051e+00 -3.21351767e-01
-2.00482771e-01 -6.99444652e-01 3.15313578e-01 -1.12913907e-01
-1.22126746e+00 7.68929124e-01 7.03413486e-01 4.24535483e-01
9.52471197e-01 -3.47684979e-01 1.05334949e+00 6.99372470e-01
2.41097316e-01 -8.65410566e-01 1.46848619e-01 4.26411539e-01
7.71587908e-01 -1.49587452e+00 9.16813165e-02 -3.07621956e-01
-2.25309998e-01 1.10765910e+00 -3.87183875e-01 -5.48304975e-01
8.85349214e-01 1.62012726e-01 -1.18208230e-02 -2.23070085e-01
-1.67280197e-01 1.26685202e-01 3.31176937e-01 7.12709248e-01
1.43136270e-02 -9.27567557e-02 -6.31693453e-02 -7.52383709e-01
-2.28187308e-01 -1.24470860e-01 5.67415655e-01 1.10795903e+00
-5.01196444e-01 -1.20081580e+00 -6.55010164e-01 2.84632295e-02
-8.62314284e-01 -3.70059073e-01 -1.12855650e-01 5.19022465e-01
-1.55671269e-01 1.06640029e+00 -1.26914740e-01 -4.63179946e-01
-4.70637023e-01 -3.62352788e-01 4.14408922e-01 -2.06379235e-01
-3.97789389e-01 -1.04012668e-01 -2.38118947e-01 -6.25624657e-01
-7.76594877e-01 -1.98601022e-01 -7.36007869e-01 -8.45828116e-01
-5.58901310e-01 -1.04320422e-02 1.02796376e+00 9.46031392e-01
1.50046408e-01 3.91193002e-01 1.05781662e+00 -1.05120158e+00
-4.74737227e-01 -5.72971880e-01 -3.99536997e-01 9.01128873e-02
1.16919482e+00 1.42485321e-01 -7.81292677e-01 3.06670010e-01] | [4.342766761779785, 8.038812637329102] |
2873f4d2-ed31-4d1d-8516-ec06dec24468 | exploiting-the-textual-potential-from-vision | 2303.04497 | null | https://arxiv.org/abs/2303.04497v1 | https://arxiv.org/pdf/2303.04497v1.pdf | Exploiting the Textual Potential from Vision-Language Pre-training for Text-based Person Search | Text-based Person Search (TPS), is targeted on retrieving pedestrians to match text descriptions instead of query images. Recent Vision-Language Pre-training (VLP) models can bring transferable knowledge to downstream TPS tasks, resulting in more efficient performance gains. However, existing TPS methods improved by VLP only utilize pre-trained visual encoders, neglecting the corresponding textual representation and breaking the significant modality alignment learned from large-scale pre-training. In this paper, we explore the full utilization of textual potential from VLP in TPS tasks. We build on the proposed VLP-TPS baseline model, which is the first TPS model with both pre-trained modalities. We propose the Multi-Integrity Description Constraints (MIDC) to enhance the robustness of the textual modality by incorporating different components of fine-grained corpus during training. Inspired by the prompt approach for zero-shot classification with VLP models, we propose the Dynamic Attribute Prompt (DAP) to provide a unified corpus of fine-grained attributes as language hints for the image modality. Extensive experiments show that our proposed TPS framework achieves state-of-the-art performance, exceeding the previous best method by a margin. | ['Shouhong Ding', 'Qiong Jia', 'Junjie Li', 'Fufu Yu', 'Guanshuo Wang'] | 2023-03-08 | null | null | null | null | ['person-search'] | ['computer-vision'] | [ 2.10053220e-01 -6.38467595e-02 -4.26560789e-01 -4.96953905e-01
-1.02465832e+00 -4.41461593e-01 1.03429997e+00 -1.18569963e-01
-7.54553795e-01 5.93245804e-01 4.52525020e-01 -3.39868404e-02
1.42486900e-01 -5.74783444e-01 -9.42158103e-01 -5.49290776e-01
4.32362676e-01 5.00499308e-01 4.41543370e-01 -7.81794116e-02
6.24225996e-02 5.10700457e-02 -1.82773650e+00 7.58338034e-01
8.22876155e-01 1.26075006e+00 4.01416391e-01 5.15596688e-01
-1.86684579e-01 8.88860643e-01 -1.78010330e-01 -1.02175379e+00
2.77509212e-01 -3.49729620e-02 -7.34373987e-01 2.28160605e-01
8.26790512e-01 -4.91188228e-01 -7.44025648e-01 8.18139672e-01
6.91456079e-01 3.82464737e-01 7.98760653e-01 -1.62331140e+00
-1.06038284e+00 2.92747498e-01 -5.46318412e-01 6.00725673e-02
5.01965106e-01 4.59574729e-01 1.14277375e+00 -1.29506099e+00
5.97258806e-01 1.41196692e+00 5.88604152e-01 7.24613547e-01
-9.01623666e-01 -3.62465739e-01 3.92483711e-01 8.34568262e-01
-1.55386555e+00 -7.17765510e-01 5.02073944e-01 -3.32155615e-01
1.17240036e+00 2.27553904e-01 6.01482630e-01 1.62058771e+00
-4.46754068e-01 1.31768239e+00 1.00553548e+00 -4.85492766e-01
-1.58104599e-01 3.20389479e-01 6.81977943e-02 9.19532120e-01
-4.53240313e-02 1.46223381e-01 -9.70733762e-01 1.81748383e-02
4.10337955e-01 -1.37776703e-01 -1.72320694e-01 -5.66937029e-01
-1.32785678e+00 4.71773088e-01 3.59356910e-01 -2.23474562e-01
-2.57515550e-01 6.44480214e-02 5.98354399e-01 -3.53494324e-02
1.91378161e-01 -2.83731446e-02 -3.03946763e-01 -4.02629189e-02
-8.63182604e-01 1.64766125e-02 4.96638864e-01 1.39715707e+00
6.17240548e-01 -2.55425245e-01 -1.10082543e+00 9.89518881e-01
4.05247301e-01 8.44953775e-01 1.60635382e-01 -1.00049984e+00
8.14279974e-01 3.85546595e-01 1.53701216e-01 -5.35640955e-01
2.53339093e-02 -4.27881807e-01 -6.54729187e-01 -3.76100928e-01
4.40495610e-01 2.92199343e-01 -1.03108048e+00 1.76634395e+00
3.42523515e-01 7.50318021e-02 1.02287836e-01 1.03408706e+00
9.84876215e-01 5.72474182e-01 5.47568262e-01 9.89679471e-02
1.66356051e+00 -1.50202835e+00 -3.76855731e-01 -3.95985514e-01
3.60654712e-01 -4.63643461e-01 1.48294795e+00 -9.98143777e-02
-9.20024335e-01 -7.35386252e-01 -7.43911505e-01 -5.42841375e-01
-4.99887586e-01 3.87785882e-01 4.16510701e-01 6.02514565e-01
-1.21878994e+00 2.53973659e-02 -4.56541210e-01 -8.12999368e-01
6.28178298e-01 3.84078249e-02 -4.85556424e-01 -4.45848972e-01
-1.27046514e+00 9.41803992e-01 3.26307029e-01 1.16371617e-01
-9.70617831e-01 -6.90201223e-01 -9.49040473e-01 5.59731349e-02
7.04271674e-01 -1.28076363e+00 1.03985977e+00 -6.84072673e-01
-1.23380709e+00 1.09719539e+00 -6.18771732e-01 -4.51700538e-01
7.65445352e-01 -3.04883629e-01 -1.39302090e-01 5.51519692e-01
4.36195672e-01 1.14246440e+00 9.94677722e-01 -1.31058633e+00
-7.47980356e-01 -2.26992473e-01 2.33883426e-01 5.32822251e-01
-5.74476123e-01 9.91198607e-03 -1.22630227e+00 -3.78116459e-01
-3.84859532e-01 -7.87833452e-01 3.00843120e-02 3.34809333e-01
-5.01586676e-01 -5.33178985e-01 4.36705112e-01 -6.02699935e-01
8.49674523e-01 -2.01084542e+00 -6.30089501e-03 -7.10335746e-02
7.70413429e-02 1.91164955e-01 -4.11287636e-01 4.15145457e-01
4.49664950e-01 -1.68458596e-01 1.34814188e-01 -8.20121527e-01
4.29748267e-01 4.21865165e-01 -2.60325044e-01 1.72084168e-01
4.02609557e-01 1.25492227e+00 -8.30087245e-01 -1.10643971e+00
2.69007504e-01 5.88682950e-01 -4.39573705e-01 1.45771831e-01
-1.88712895e-01 4.80429620e-01 -4.97782081e-01 9.35494244e-01
5.76428235e-01 -3.85329068e-01 -4.27326351e-01 -6.04477167e-01
-2.18504686e-02 -9.27515551e-02 -6.88328326e-01 1.90710449e+00
-2.26800054e-01 5.62470853e-01 -2.98287719e-01 -9.37092423e-01
5.33163548e-01 1.30489036e-01 3.46329898e-01 -1.11205602e+00
-5.40740229e-02 -4.25321832e-02 -5.77465475e-01 -6.49711072e-01
6.84878469e-01 2.12806612e-01 -1.75950438e-01 -4.96483035e-03
3.09289724e-01 5.61264575e-01 4.19317245e-01 5.41713774e-01
7.56333470e-01 3.44956577e-01 1.00838855e-01 8.03318173e-02
8.29086959e-01 -6.26095086e-02 3.69394511e-01 1.24376690e+00
-6.48841739e-01 6.56072617e-01 2.32140228e-01 -2.12228522e-01
-1.21864629e+00 -1.09892988e+00 -5.52759729e-02 1.59577942e+00
3.17512840e-01 -4.28433299e-01 -5.96811235e-01 -7.45825589e-01
-9.50141996e-02 6.04364932e-01 -5.98458588e-01 -8.75645876e-02
-2.86743909e-01 -4.55995888e-01 8.61023128e-01 7.24044502e-01
1.05557263e+00 -8.23211014e-01 -4.78738844e-01 -2.69061685e-01
-8.03012848e-01 -1.82119775e+00 -6.67819560e-01 -1.93024173e-01
-2.17920646e-01 -8.20121646e-01 -1.09625649e+00 -7.32608378e-01
5.57309151e-01 4.55447316e-01 1.09252024e+00 -6.01657778e-02
-1.43199578e-01 1.03759241e+00 -6.19553387e-01 -2.19233006e-01
1.38030142e-01 -6.01322912e-02 8.28424767e-02 2.68004209e-01
5.31487584e-01 -7.07037151e-02 -7.32522428e-01 2.81276822e-01
-4.22472268e-01 3.79949033e-01 8.65484774e-01 1.07908010e+00
7.26755857e-01 -4.29170817e-01 3.16507846e-01 -3.97238225e-01
1.91616461e-01 -2.03869671e-01 -1.03149869e-01 7.89485812e-01
-4.65424269e-01 2.06999287e-01 3.05067897e-01 -5.95440269e-01
-1.41269767e+00 1.52592719e-01 -9.99293849e-02 -6.62050486e-01
-2.58908361e-01 2.84315616e-01 -4.14706916e-01 -8.31431672e-02
4.09754872e-01 7.60575712e-01 -3.32752556e-01 -3.49553555e-01
7.43412137e-01 4.99192506e-01 9.10973370e-01 -9.14011896e-01
7.18556046e-01 5.44743419e-01 -7.54786283e-02 -6.40115619e-01
-1.15822637e+00 -7.97313333e-01 -7.81911194e-01 -3.99505496e-01
1.07386780e+00 -1.18840456e+00 -7.15555429e-01 3.65754753e-01
-1.15970290e+00 -1.68183804e-01 -1.40587181e-01 2.97707260e-01
-6.67580068e-01 6.25819504e-01 -4.73182052e-01 -8.11047196e-01
-2.47479692e-01 -9.97584403e-01 1.38296902e+00 1.76954195e-01
3.89716595e-01 -5.34095705e-01 -4.06650484e-01 8.15988004e-01
3.19523275e-01 -3.18013757e-01 6.91081166e-01 -5.92809141e-01
-7.06767380e-01 4.40681055e-02 -8.34427714e-01 1.85100976e-02
-4.54478741e-01 -5.00907183e-01 -1.21844435e+00 -2.16652140e-01
-3.37278277e-01 -6.05909586e-01 1.35352433e+00 2.44933397e-01
1.12870777e+00 -2.38759711e-01 -4.97863233e-01 7.14542747e-01
1.25809693e+00 -2.87189215e-01 7.45603919e-01 6.63356125e-01
9.61178601e-01 6.34730697e-01 8.34613264e-01 5.10204375e-01
1.02583778e+00 7.62626469e-01 2.46396974e-01 -1.46370083e-01
-5.79380572e-01 -5.25232375e-01 4.67334956e-01 2.89010733e-01
-3.13766152e-01 -3.79599959e-01 -7.64106870e-01 5.51979184e-01
-2.07109451e+00 -1.15816307e+00 4.12320308e-02 2.12779570e+00
7.04931140e-01 -9.60068256e-02 1.82380587e-01 -2.64625818e-01
6.61370695e-01 1.75399810e-01 -5.82980216e-01 1.89336717e-01
-3.61537814e-01 -2.98106641e-01 5.47139883e-01 3.07046264e-01
-1.41854453e+00 1.27011764e+00 5.47476625e+00 9.79960024e-01
-5.38461328e-01 3.82618397e-01 3.48642468e-01 -1.18581221e-01
-6.09969981e-02 -1.12252362e-01 -1.14339626e+00 3.75217110e-01
5.66931129e-01 -5.66845424e-02 1.86947525e-01 8.10563743e-01
-4.79732640e-02 1.49586834e-02 -1.35861385e+00 1.23663306e+00
4.75732833e-01 -1.10834563e+00 5.27721524e-01 -4.03848030e-02
4.20753866e-01 -2.82283337e-03 3.39546621e-01 7.51825094e-01
-9.28649232e-02 -7.53587067e-01 9.61645246e-01 6.99561298e-01
9.80701566e-01 -4.41677749e-01 6.14067733e-01 2.16815263e-01
-1.62166369e+00 -3.03492039e-01 -5.34388304e-01 3.19155753e-01
3.21467072e-01 9.52617824e-02 -6.80875957e-01 6.58655822e-01
8.24324250e-01 7.87675619e-01 -1.04000461e+00 1.07828116e+00
-8.86387527e-02 2.59010106e-01 -6.25044033e-02 7.40023106e-02
2.94042975e-01 2.43361607e-01 5.93143225e-01 1.38477886e+00
2.92172253e-01 -1.49539532e-02 3.78064752e-01 7.61219263e-01
3.06743849e-02 -3.72855142e-02 -4.35224205e-01 1.01910047e-01
6.54257655e-01 9.57141995e-01 -3.06035638e-01 -6.05289459e-01
-9.28868711e-01 1.36062002e+00 3.94396096e-01 5.57582259e-01
-9.32873607e-01 -2.28052847e-02 3.89154881e-01 -2.62737833e-02
5.28053880e-01 -9.56220925e-02 9.44891423e-02 -1.38054931e+00
2.74417490e-01 -7.01113701e-01 6.94204330e-01 -1.05432892e+00
-1.71694684e+00 5.16271770e-01 1.10932164e-01 -1.24281359e+00
-1.60324082e-01 -6.77233756e-01 -4.55139801e-02 7.98046470e-01
-2.00293827e+00 -2.19692492e+00 -5.10985196e-01 1.27062225e+00
8.31555605e-01 -2.97698081e-01 6.53946817e-01 4.61910903e-01
-4.49721843e-01 1.12212062e+00 -2.43059725e-01 1.63940921e-01
1.17139220e+00 -1.06457317e+00 9.35245976e-02 1.01710677e+00
5.68855256e-02 5.27977347e-01 5.95680118e-01 -5.95164061e-01
-1.51150846e+00 -1.16575897e+00 9.64493811e-01 -6.87154889e-01
5.60854852e-01 -2.93896168e-01 -6.63301408e-01 6.36699736e-01
2.11613789e-01 1.24257088e-01 4.77927387e-01 1.59471780e-01
-8.53732705e-01 -9.70404595e-02 -8.61893833e-01 5.89943230e-01
1.40582144e+00 -8.60506654e-01 -6.28220499e-01 3.49323362e-01
8.86635184e-01 -2.24537790e-01 -5.30062616e-01 3.09349597e-01
6.77125692e-01 -7.01529682e-01 1.59172976e+00 -4.98377681e-01
3.32402259e-01 -2.79018134e-01 -4.83011395e-01 -7.83035517e-01
-3.89732033e-01 -5.67007586e-02 -2.19835371e-01 1.47963905e+00
9.79740322e-02 -3.17635208e-01 6.02611780e-01 9.07465994e-01
-3.06717902e-01 -4.39346582e-01 -9.73479688e-01 -8.72248888e-01
-3.04334372e-01 -5.57467997e-01 2.66943306e-01 6.72655761e-01
-1.54806241e-01 4.02302474e-01 -6.99130237e-01 4.43037480e-01
1.08990657e+00 -1.63796261e-01 7.37109363e-01 -1.02577937e+00
-3.74580324e-01 -3.15703213e-01 -3.67687821e-01 -1.38227057e+00
2.11567014e-01 -8.65754426e-01 1.42889544e-01 -1.71986842e+00
8.45685065e-01 -2.73749858e-01 -4.94264334e-01 7.08554387e-01
-3.31134856e-01 2.54898757e-01 4.66974884e-01 4.33296531e-01
-1.27906621e+00 7.44404793e-01 1.21045351e+00 -4.10562098e-01
1.32605836e-01 -2.34994516e-01 -6.08167529e-01 5.43901086e-01
3.09532940e-01 -1.60090357e-01 -5.89705646e-01 -7.71337092e-01
-2.62453360e-03 -1.18917696e-01 9.24880087e-01 -9.47696686e-01
7.05678880e-01 -1.13556653e-01 5.67062736e-01 -7.88488269e-01
7.59111166e-01 -6.38660729e-01 -4.46529865e-01 -4.50700382e-03
-5.60479105e-01 -2.76823103e-01 7.56074116e-02 8.81850302e-01
-1.45471156e-01 -2.43162401e-02 4.02759284e-01 -1.33865669e-01
-1.36384749e+00 6.17021561e-01 -8.21294487e-02 3.99909774e-03
7.87785351e-01 -2.36006394e-01 -7.86217988e-01 -3.12796295e-01
-6.82023764e-01 6.02340758e-01 3.83855194e-01 4.89831030e-01
7.94671178e-01 -1.56272650e+00 -6.66914642e-01 -2.63429955e-02
7.18060791e-01 -4.02382493e-01 5.69791913e-01 1.05879843e+00
1.27731189e-01 5.53933144e-01 -4.02903289e-01 -7.44100332e-01
-1.29041898e+00 8.67004275e-01 1.29361033e-01 -1.73449114e-01
-7.63111770e-01 1.04190779e+00 4.33178961e-01 -1.74280271e-01
5.47430575e-01 2.61615425e-01 -2.85522521e-01 1.99383609e-02
6.57582581e-01 1.13014080e-01 -2.92547613e-01 -9.11092639e-01
-5.23347974e-01 5.91984034e-01 -8.63114819e-02 -4.10090387e-01
9.26591873e-01 -6.23546064e-01 1.84035271e-01 1.74778789e-01
1.00026727e+00 -2.98141807e-01 -1.44660437e+00 -6.39622629e-01
-1.44419283e-01 -6.33354425e-01 -1.23874746e-01 -1.00681341e+00
-7.80364931e-01 1.06511319e+00 6.65357172e-01 -3.74297589e-01
1.16973221e+00 3.65063936e-01 8.02903891e-01 7.31243730e-01
5.37868500e-01 -1.23946583e+00 5.00060380e-01 5.60240328e-01
7.63337553e-01 -1.52435839e+00 -2.12865129e-01 -4.62684155e-01
-1.25995588e+00 7.87958503e-01 8.40248168e-01 4.78515029e-01
1.06998853e-01 -2.50969619e-01 -2.54736334e-01 1.13785535e-01
-7.60954082e-01 -8.97853494e-01 7.17004359e-01 9.88896549e-01
-3.59850563e-02 -1.82541370e-01 2.46192310e-02 8.34209681e-01
2.05476120e-01 -2.19143145e-02 -5.54895923e-02 7.13009775e-01
-5.02765715e-01 -9.98923004e-01 -2.48799741e-01 2.13934824e-01
-3.67102548e-02 -4.62153554e-01 -3.57293934e-01 5.93297422e-01
2.06487402e-01 9.75787222e-01 -1.47753716e-01 -4.08389091e-01
2.70423114e-01 3.32012147e-01 6.14412665e-01 -1.66226611e-01
-8.40553418e-02 -7.55784959e-02 4.23658073e-01 -7.23137259e-01
-6.09441876e-01 -6.63793445e-01 -7.99667060e-01 -1.26188055e-01
1.93477049e-02 -2.85697043e-01 3.15065354e-01 1.13186014e+00
4.55695421e-01 2.74885774e-01 1.21427350e-01 -8.53280067e-01
-3.01858336e-01 -8.83746743e-01 -7.91689083e-02 5.98635495e-01
2.54512399e-01 -1.02707541e+00 -1.13972407e-02 2.81447738e-01] | [14.57864761352539, 0.8904938101768494] |
3374a5b8-e5ea-4221-b6de-0a9603ccdb51 | weight-poisoning-attacks-on-pretrained-models | null | null | https://aclanthology.org/2020.acl-main.249 | https://aclanthology.org/2020.acl-main.249.pdf | Weight Poisoning Attacks on Pretrained Models | Recently, NLP has seen a surge in the usage of large pre-trained models. Users download weights of models pre-trained on large datasets, then fine-tune the weights on a task of their choice. This raises the question of whether downloading untrusted pre-trained weights can pose a security threat. In this paper, we show that it is possible to construct {``}weight poisoning{''} attacks where pre-trained weights are injected with vulnerabilities that expose {``}backdoors{''} after fine-tuning, enabling the attacker to manipulate the model prediction simply by injecting an arbitrary keyword. We show that by applying a regularization method which we call RIPPLe and an initialization procedure we call Embedding Surgery, such attacks are possible even with limited knowledge of the dataset and fine-tuning procedure. Our experiments on sentiment classification, toxicity detection, and spam detection show that this attack is widely applicable and poses a serious threat. Finally, we outline practical defenses against such attacks. | ['Keita Kurita', 'Paul Michel', 'Graham Neubig'] | 2020-07-01 | null | null | null | acl-2020-6 | ['spam-detection'] | ['natural-language-processing'] | [ 4.27272290e-01 1.85889736e-01 2.32352167e-02 -8.00854117e-02
-7.71210074e-01 -1.44414127e+00 4.27674145e-01 1.83809698e-01
-7.49235868e-01 6.13587260e-01 -2.63315886e-01 -6.86111450e-01
1.97372884e-01 -7.85266161e-01 -1.15355182e+00 -8.30351293e-01
-1.97726816e-01 1.79345623e-01 2.95838505e-01 -8.10749307e-02
5.31659424e-01 6.95741653e-01 -9.38053489e-01 2.94271111e-01
3.87715459e-01 6.13873839e-01 -2.98398614e-01 8.26562822e-01
5.41140279e-03 1.08522654e-01 -1.03986931e+00 -1.05242085e+00
6.91651404e-01 2.88464159e-01 -8.72857451e-01 -4.74183232e-01
2.85532653e-01 -3.38223338e-01 -1.93897113e-01 1.41487706e+00
3.99080604e-01 -3.07359666e-01 4.36975688e-01 -1.38287795e+00
-5.07943630e-01 8.54299784e-01 -6.86159968e-01 2.67970860e-01
9.14339796e-02 3.27945650e-01 8.26286376e-01 -2.15619609e-01
4.91255164e-01 8.99842739e-01 5.64888835e-01 1.01714957e+00
-1.49450982e+00 -1.16940868e+00 -7.74290636e-02 -3.23080927e-01
-1.08890915e+00 -2.72411317e-01 5.37043214e-01 -4.69302267e-01
8.70362520e-01 6.94463074e-01 1.92253962e-01 1.69332755e+00
1.74919382e-01 3.94208014e-01 1.04226720e+00 -3.78808528e-01
3.07102710e-01 6.49859548e-01 2.30405599e-01 7.07320511e-01
7.14647233e-01 1.56127572e-01 -4.52158779e-01 -1.20793772e+00
2.52668530e-01 -9.07645822e-02 -4.03361261e-01 -3.17443371e-01
-9.18556452e-01 1.14549446e+00 2.02652305e-01 9.16052517e-03
2.76703574e-02 3.95220637e-01 5.55765033e-01 3.87127042e-01
4.06040311e-01 1.07795072e+00 -1.14360309e+00 1.24370664e-01
-5.88261902e-01 3.81130725e-02 1.02174389e+00 6.17124259e-01
5.40665030e-01 -3.55682448e-02 3.04406285e-01 2.81694353e-01
5.08563146e-02 4.44771320e-01 5.49454451e-01 -5.39164186e-01
5.32230139e-01 7.23247677e-02 2.99218059e-01 -8.01966846e-01
-2.64204115e-01 -7.30969310e-02 -2.92960465e-01 3.64737839e-01
7.13514090e-01 -7.04471350e-01 -7.65671432e-01 2.04371190e+00
5.04717648e-01 2.54984468e-01 3.43079828e-02 3.72117281e-01
1.00297250e-01 5.95741808e-01 4.01613474e-01 2.28311568e-02
1.46737063e+00 -5.21319270e-01 -2.00843111e-01 -7.34187961e-02
7.50892699e-01 -6.22960150e-01 1.19141638e+00 5.34894109e-01
-8.12789261e-01 2.03876197e-01 -1.22677410e+00 1.55236065e-01
-9.08659279e-01 -4.41807121e-01 7.58274555e-01 1.39079607e+00
-5.23757935e-01 8.16726148e-01 -7.35426128e-01 -1.81854278e-01
6.24477684e-01 7.50414133e-01 -5.32440484e-01 4.34242845e-01
-1.39253938e+00 8.72079849e-01 1.88427493e-01 -3.61414105e-01
-1.02856123e+00 -1.02641976e+00 -3.84166241e-01 1.89549983e-01
2.18973204e-01 -3.60827118e-01 9.17602122e-01 -6.03959978e-01
-1.21845412e+00 8.66409242e-01 2.31778860e-01 -6.44795477e-01
5.14742911e-01 -1.17568702e-01 -3.47637534e-01 1.42219931e-01
-2.38569677e-01 2.55535156e-01 1.36372364e+00 -1.12449360e+00
-2.98116595e-01 -4.07817304e-01 3.75262976e-01 -4.59755123e-01
-1.16304779e+00 4.40109760e-01 1.65595561e-01 -7.05811560e-01
-5.52183032e-01 -1.07473505e+00 -2.91779697e-01 2.80478727e-02
-8.79287660e-01 2.62875170e-01 8.32755208e-01 -4.73076046e-01
1.33625340e+00 -2.24138165e+00 -7.24166557e-02 7.60734975e-01
4.10466522e-01 7.95780897e-01 -2.54094392e-01 5.34534812e-01
-2.22269192e-01 1.00178730e+00 -2.20502794e-01 4.37482968e-02
1.91828951e-01 2.65876632e-02 -8.64791512e-01 5.66257894e-01
-3.82071361e-02 6.76898837e-01 -5.33422053e-01 4.52402830e-02
-3.60235006e-01 4.51950967e-01 -6.78539574e-01 3.90170068e-02
-3.83793354e-01 4.36103418e-02 -8.16106081e-01 2.62463212e-01
6.69785202e-01 -3.10295284e-01 1.07979164e-01 -1.78219661e-01
2.13724419e-01 2.52732217e-01 -8.70639920e-01 1.17377889e+00
-1.95657089e-01 1.65861398e-01 5.94725944e-02 -6.87580407e-01
5.03377020e-01 2.06740364e-01 1.75991029e-01 2.46974036e-01
2.08478361e-01 1.15626872e-01 -3.33815925e-02 -4.62952137e-01
4.65078577e-02 2.50951853e-02 -2.71571368e-01 9.81598854e-01
3.01054167e-03 -1.01170538e-03 -2.89623052e-01 4.38071638e-01
1.59183216e+00 -3.93552274e-01 -2.25169659e-01 -2.25622639e-01
4.44842219e-01 -1.51561752e-01 1.59216642e-01 8.42077494e-01
-7.20843226e-02 1.43158749e-01 8.56981575e-01 -4.90745842e-01
-1.07486403e+00 -9.80100036e-01 -1.23095520e-01 1.27004242e+00
-2.15432033e-01 -7.29586065e-01 -1.13852441e+00 -1.33582664e+00
3.09641629e-01 6.93500578e-01 -1.05861115e+00 -4.81705904e-01
-2.91438669e-01 -1.07473743e+00 1.13696122e+00 1.59314156e-01
2.83124685e-01 -8.53384078e-01 -4.40104336e-01 4.33284044e-02
1.77873343e-01 -7.79486477e-01 -6.95267320e-01 5.89941323e-01
-6.60379112e-01 -1.16682136e+00 -3.66239756e-01 -3.45010310e-01
9.30624366e-01 4.32818979e-02 6.47047222e-01 3.38416845e-01
-4.35395956e-01 2.22791120e-01 -2.11841948e-02 -7.36680329e-01
-5.10046959e-01 4.47805792e-01 7.21142590e-02 1.68013312e-02
6.46800220e-01 -6.99935257e-01 -3.65841419e-01 1.52861565e-01
-1.24871278e+00 -6.43482804e-01 2.47216836e-01 5.87297380e-01
6.63335323e-02 5.87634258e-02 5.25199652e-01 -1.56821334e+00
9.26343799e-01 -4.39784288e-01 -9.16885614e-01 4.02793705e-01
-5.69002807e-01 2.88851559e-01 6.76645994e-01 -1.06191194e+00
-5.37178993e-01 7.99949914e-02 -1.95778370e-01 -4.02847886e-01
-1.82125643e-02 1.43384948e-01 -3.33158314e-01 -6.05872929e-01
1.08461142e+00 -1.18342757e-01 -1.70218468e-01 -5.50922573e-01
4.93444115e-01 5.27772009e-01 -3.01742386e-02 -7.42232621e-01
1.43273318e+00 4.80449170e-01 -1.53843060e-01 -4.54578906e-01
-7.84171939e-01 3.71968634e-02 -8.64335001e-02 3.92023414e-01
8.50941777e-01 -5.47741413e-01 -9.14238036e-01 3.98387700e-01
-1.16750979e+00 -2.54136860e-01 -1.83406919e-01 -1.31645165e-02
-1.06566608e-01 5.11238575e-01 -7.33428895e-01 -4.40356880e-01
-5.65889716e-01 -1.15833139e+00 5.40771127e-01 -9.96221602e-02
-3.85519922e-01 -9.28839803e-01 3.26757468e-02 4.44753557e-01
6.16193771e-01 1.60960719e-01 1.20725954e+00 -1.33512616e+00
-3.03950459e-01 -7.27395058e-01 -6.39702380e-03 3.82871538e-01
-1.21281095e-01 1.33521348e-01 -1.32530582e+00 -3.93348873e-01
1.54327542e-01 -4.41746414e-01 5.97842157e-01 -2.96793312e-01
1.56418085e+00 -8.12246501e-01 -4.39331174e-01 6.51231587e-01
1.21615279e+00 -9.40582082e-02 5.85682690e-01 4.89812940e-01
6.46797121e-01 3.31517816e-01 -1.57386556e-01 4.43920761e-01
-3.29145193e-01 2.56435215e-01 3.35833341e-01 2.77813047e-01
8.16472650e-01 -3.30825239e-01 4.32600915e-01 7.98596144e-02
1.32657498e-01 -1.11338116e-01 -5.27950287e-01 7.61976168e-02
-1.32649612e+00 -8.65406930e-01 1.38073564e-01 2.36046410e+00
1.31203091e+00 5.71501315e-01 -1.03935741e-01 -1.63437277e-02
8.17921996e-01 1.20047450e-01 -6.75593853e-01 -8.62668395e-01
3.24528843e-01 8.19519758e-01 1.13747728e+00 4.40795094e-01
-1.24434626e+00 1.14112926e+00 6.61323977e+00 1.03879869e+00
-1.18308496e+00 4.23763484e-01 4.56408978e-01 -1.88291237e-01
-5.66793501e-01 -4.74831760e-02 -9.17398870e-01 6.18651450e-01
1.19023776e+00 -2.39340544e-01 5.88782728e-01 8.71845961e-01
-2.34442070e-01 3.95006031e-01 -8.58293176e-01 3.74448687e-01
-1.22762159e-01 -1.44056511e+00 5.85318543e-02 3.68297607e-01
3.96019548e-01 -7.03540221e-02 3.48686546e-01 1.74129635e-01
7.68714726e-01 -8.80753696e-01 1.88935757e-01 -3.84633876e-02
6.30930066e-01 -9.75687325e-01 1.44777611e-01 2.83041865e-01
-3.87712806e-01 -1.81759849e-01 -6.21621013e-01 3.90982926e-01
-7.66454041e-02 4.89908665e-01 -5.73805749e-01 -1.69496298e-01
5.62696159e-01 -8.38305876e-02 -6.56637430e-01 7.66317010e-01
-5.28815389e-01 1.00993252e+00 -6.02357686e-01 -3.33041213e-02
1.06161311e-01 7.62331709e-02 4.12863374e-01 1.12621701e+00
2.00435910e-02 -1.17857531e-01 -1.98644236e-01 7.00811744e-01
-5.55198729e-01 -3.87505852e-02 -7.24373221e-01 -3.35047126e-01
3.60302150e-01 1.34280765e+00 -4.83939976e-01 -1.44995168e-01
-2.70328045e-01 1.00990319e+00 3.57807249e-01 3.47784162e-01
-9.82798040e-01 -5.94941378e-01 1.03515542e+00 1.65646616e-02
2.52450734e-01 1.98436193e-02 -1.84458420e-01 -1.28214347e+00
-1.73581094e-01 -9.59806144e-01 4.24787819e-01 -5.73146343e-01
-1.36683559e+00 4.10152763e-01 -2.26374760e-01 -7.35038400e-01
3.39647889e-01 -6.17947996e-01 -6.62340462e-01 7.68039584e-01
-1.27316689e+00 -9.08122301e-01 5.91584504e-01 6.65848911e-01
2.23324783e-02 -2.04470560e-01 1.22274435e+00 8.28732252e-02
-6.77297473e-01 1.09096980e+00 1.06793806e-01 2.73942351e-01
7.14182198e-01 -1.02220750e+00 5.53966463e-01 7.74168134e-01
2.81488299e-01 1.06137300e+00 8.38924825e-01 -6.96408093e-01
-1.37998831e+00 -9.61857617e-01 8.49489808e-01 -8.46176863e-01
1.20351446e+00 -9.60733712e-01 -9.79497790e-01 8.03776503e-01
7.88591430e-02 -1.45974737e-02 1.17888498e+00 -1.28048480e-01
-1.10678744e+00 1.48307130e-01 -1.71141851e+00 6.08512580e-01
5.68697155e-01 -7.25164175e-01 -3.04228455e-01 9.63719904e-01
1.02957094e+00 -4.31537554e-02 -7.52098262e-01 5.82283586e-02
6.02318525e-01 -3.85068268e-01 1.18915284e+00 -1.48341715e+00
1.36510760e-01 4.02824879e-02 -4.32154015e-02 -1.02965629e+00
-2.18523350e-02 -1.10193002e+00 -1.66362256e-01 1.24812257e+00
9.95990872e-01 -1.06029737e+00 8.50061774e-01 1.19490421e+00
6.01184964e-01 -2.86753654e-01 -7.78041363e-01 -6.10123694e-01
6.40660584e-01 -3.14297944e-01 6.47931814e-01 1.17757022e+00
2.10969016e-01 -5.65022044e-02 -4.97440368e-01 6.29123390e-01
7.44551480e-01 -4.62507963e-01 6.76885962e-01 -1.05534267e+00
-5.05360067e-01 -5.23937382e-02 -2.12237954e-01 -4.78243560e-01
3.82549345e-01 -7.94537067e-01 -4.60877717e-01 -4.44228977e-01
7.78961629e-02 -3.96565288e-01 -4.51614678e-01 7.32834995e-01
2.62635052e-02 3.83004695e-01 1.88409865e-01 1.29701272e-01
-1.57965690e-01 -2.03753024e-01 7.43629873e-01 -2.48971075e-01
1.04655683e-01 1.88897952e-01 -1.02703238e+00 7.45686889e-01
1.20460105e+00 -1.22670031e+00 -3.84991825e-01 -1.49819449e-01
7.46295035e-01 -5.20556927e-01 2.37354144e-01 -5.56830049e-01
1.21486887e-01 -3.92813608e-02 2.81882584e-01 1.74679637e-01
2.78182268e-01 -1.05227637e+00 -1.58544928e-02 5.07441282e-01
-6.55145705e-01 -1.09864324e-02 3.39342147e-01 7.22855210e-01
5.82677245e-01 -6.95514381e-01 8.61118078e-01 -2.84899652e-01
1.54079601e-01 3.12092960e-01 -5.02924800e-01 3.20458375e-02
1.13910615e+00 2.69254714e-01 -5.60199857e-01 -6.51018247e-02
-9.75037158e-01 6.09011576e-02 4.71959412e-01 1.81689233e-01
3.10276121e-01 -7.99102128e-01 -2.22207487e-01 4.93091524e-01
-3.25758904e-02 -7.12228894e-01 -1.32591218e-01 2.76512086e-01
-3.82067561e-01 -1.13352522e-01 9.12155360e-02 4.98533696e-02
-1.42801750e+00 1.04562986e+00 1.40496820e-01 -2.71876186e-01
-2.70485461e-01 1.06274605e+00 5.57791181e-02 -5.03630042e-01
4.62043047e-01 -1.15493005e-02 1.14403330e-02 1.59035306e-02
7.05005169e-01 3.20308618e-02 2.75695287e-02 -1.10328861e-01
-2.50361770e-01 2.17540205e-01 -5.22871017e-01 -3.97851355e-02
1.25767136e+00 1.84450209e-01 -1.81522012e-01 -9.02888253e-02
1.42124069e+00 4.51852560e-01 -9.85451758e-01 1.41801396e-02
-1.88184977e-02 -5.88673711e-01 -3.08091909e-01 -1.07103550e+00
-1.21232808e+00 9.60924864e-01 4.91818905e-01 6.32197618e-01
6.29236162e-01 -2.01576054e-01 9.75585520e-01 7.53414035e-01
4.78647739e-01 -8.92401993e-01 -1.02562785e-01 1.19291179e-01
5.39776862e-01 -8.05649459e-01 -4.23800312e-02 -4.26557302e-01
-5.85305274e-01 1.01177120e+00 4.13032651e-01 -2.73881286e-01
1.06085348e+00 4.80442405e-01 7.91210979e-02 -1.13535032e-01
-6.49259269e-01 4.09820914e-01 -2.10365757e-01 5.60234129e-01
-2.55276382e-01 -2.21917313e-02 -1.44883037e-01 7.07066178e-01
-2.27687135e-01 -1.82543874e-01 7.63640046e-01 1.00028276e+00
-5.00157297e-01 -1.52284443e+00 -4.61183786e-01 6.38978243e-01
-1.15064633e+00 -3.31275821e-01 -5.55938244e-01 4.67421472e-01
-7.17230290e-02 8.35606515e-01 -5.46471179e-01 -5.52950621e-01
1.37588397e-01 5.48621118e-01 2.60318279e-01 -6.25578344e-01
-1.16628861e+00 -2.52146751e-01 1.06146798e-01 -4.75570023e-01
1.75888121e-01 -4.58516777e-01 -9.51882720e-01 -5.87428629e-01
-6.83716238e-01 3.48556042e-01 7.92258680e-01 6.97867811e-01
4.39719528e-01 -1.41982287e-01 7.54123569e-01 -4.49233681e-01
-1.26049685e+00 -6.19970202e-01 -6.76713943e-01 4.51349914e-01
1.80798531e-01 -2.60363966e-01 -1.02784371e+00 1.09106295e-01] | [5.861743450164795, 7.680727481842041] |
4624fe2a-653a-4975-9dab-3a29b8de5bca | generalizable-one-shot-neural-head-avatar | 2306.08768 | null | https://arxiv.org/abs/2306.08768v1 | https://arxiv.org/pdf/2306.08768v1.pdf | Generalizable One-shot Neural Head Avatar | We present a method that reconstructs and animates a 3D head avatar from a single-view portrait image. Existing methods either involve time-consuming optimization for a specific person with multiple images, or they struggle to synthesize intricate appearance details beyond the facial region. To address these limitations, we propose a framework that not only generalizes to unseen identities based on a single-view image without requiring person-specific optimization, but also captures characteristic details within and beyond the face area (e.g. hairstyle, accessories, etc.). At the core of our method are three branches that produce three tri-planes representing the coarse 3D geometry, detailed appearance of a source image, as well as the expression of a target image. By applying volumetric rendering to the combination of the three tri-planes followed by a super-resolution module, our method yields a high fidelity image of the desired identity, expression and pose. Once trained, our model enables efficient 3D head avatar reconstruction and animation via a single forward pass through a network. Experiments show that the proposed approach generalizes well to unseen validation datasets, surpassing SOTA baseline methods by a large margin on head avatar reconstruction and animation. | ['Jan Kautz', 'Umar Iqbal', 'Koki Nagano', 'Sifei Liu', 'Shalini De Mello', 'Xueting Li'] | 2023-06-14 | null | null | null | null | ['super-resolution'] | ['computer-vision'] | [ 2.50437587e-01 3.39246958e-01 3.36810797e-01 -4.30865794e-01
-5.45918107e-01 -5.73964179e-01 5.06674111e-01 -4.92620379e-01
1.03477277e-01 5.40753663e-01 2.11181313e-01 3.56781751e-01
5.36312640e-01 -6.76652014e-01 -7.22927511e-01 -5.57919383e-01
3.75222355e-01 7.18926132e-01 1.79631524e-02 -3.16559434e-01
-2.38817960e-01 9.34472263e-01 -1.69599557e+00 3.00222374e-02
4.00403023e-01 1.07752597e+00 -3.81993294e-01 6.36245906e-01
7.59166926e-02 3.00445080e-01 -6.12918198e-01 -7.66903818e-01
3.33462328e-01 -4.71525878e-01 -5.83585083e-01 7.23555684e-01
1.05103588e+00 -7.80743062e-01 -1.09493859e-01 7.70956099e-01
4.68165994e-01 5.69519075e-03 5.20417750e-01 -1.19519675e+00
-4.51680243e-01 9.21234041e-02 -7.86818087e-01 -5.36395907e-01
8.48959982e-01 1.98438510e-01 7.24384606e-01 -1.04434693e+00
1.07095373e+00 1.46974659e+00 8.17613006e-01 9.41954434e-01
-1.49173725e+00 -7.77528584e-01 2.56785005e-01 -2.69456536e-01
-1.54332912e+00 -6.21482074e-01 1.00950229e+00 -2.59120136e-01
3.20113271e-01 3.26067775e-01 1.21490598e+00 1.34894884e+00
-1.39829256e-02 6.20292127e-01 1.20803714e+00 -1.30720362e-01
4.65149060e-02 1.64889231e-01 -3.76710087e-01 9.99255478e-01
-2.53263414e-01 1.29879906e-03 -4.87452626e-01 -1.93347752e-01
1.15821087e+00 -9.84178185e-02 -4.17806655e-01 -6.00749195e-01
-8.49322975e-01 6.46681309e-01 2.51863837e-01 -1.37321308e-01
-3.57502431e-01 -3.86738069e-02 4.38682325e-02 -1.25573546e-01
5.02664983e-01 3.04933507e-02 -1.48770750e-01 6.17428049e-02
-9.94882524e-01 3.87478888e-01 6.60329103e-01 9.25610602e-01
7.89129734e-01 3.80681604e-01 2.00182259e-01 6.50451005e-01
1.72742307e-01 5.71437657e-01 -2.74806153e-02 -1.15067792e+00
7.52484947e-02 5.32877207e-01 -5.86654134e-02 -8.60737801e-01
-4.78030384e-01 -3.03167075e-01 -8.77408624e-01 5.82863033e-01
4.88728225e-01 -1.64241880e-01 -9.11854148e-01 1.79815495e+00
9.17949378e-01 1.41672948e-02 -2.96392709e-01 1.13756084e+00
1.05471814e+00 4.12511259e-01 -1.46599963e-01 -1.55811280e-01
1.45614576e+00 -8.41251194e-01 -6.57174945e-01 -1.49767712e-01
1.78840980e-02 -5.37927032e-01 1.08822048e+00 3.45997155e-01
-1.53832269e+00 -3.50143611e-01 -8.18767190e-01 -3.53965521e-01
5.35121597e-02 -9.15958807e-02 3.96741807e-01 6.71651602e-01
-1.24824119e+00 5.40823877e-01 -4.37801093e-01 -2.48874918e-01
4.03678596e-01 4.66406256e-01 -8.67536128e-01 1.90569043e-01
-7.34177113e-01 7.76946425e-01 -2.28247195e-01 1.33637890e-01
-7.73963809e-01 -8.68319690e-01 -1.08704185e+00 -1.02884091e-01
1.80945829e-01 -1.06573558e+00 9.62718904e-01 -1.34078860e+00
-1.91840351e+00 1.18128049e+00 -1.79111391e-01 8.43631625e-02
8.19646955e-01 1.67822819e-02 -1.96239948e-01 2.62377590e-01
-2.38744378e-01 7.83558249e-01 1.15495169e+00 -1.61269569e+00
-2.16335148e-01 -7.65377283e-01 1.31832123e-01 3.36133987e-01
-1.79460179e-02 -1.27094025e-02 -5.80011308e-01 -6.38647258e-01
1.57169580e-01 -1.02385736e+00 -1.64490554e-03 6.33386135e-01
-5.14086187e-01 3.15016776e-01 1.04451191e+00 -8.80607784e-01
6.84046090e-01 -2.11055303e+00 5.43449819e-01 2.82532781e-01
4.53935862e-01 -3.41667570e-02 -2.24413108e-02 -4.26982641e-02
-1.30089998e-01 -1.03108428e-01 -1.31735831e-01 -8.47575307e-01
-1.91351205e-01 3.33155282e-02 8.15830380e-02 7.09120333e-01
4.18199673e-02 8.27097654e-01 -6.50087893e-01 -6.51950181e-01
2.23756522e-01 9.99402165e-01 -6.22480571e-01 3.02678764e-01
-1.64696470e-01 8.93103242e-01 -3.02494198e-01 7.76462674e-01
7.31199682e-01 -2.32169762e-01 1.55155733e-01 -3.40383619e-01
1.92458689e-01 -2.76320815e-01 -1.25481522e+00 1.81645536e+00
-4.17196989e-01 3.82044733e-01 6.23270512e-01 -3.40346992e-01
9.78653193e-01 3.60610306e-01 5.07480085e-01 -3.27935398e-01
3.54096323e-01 2.72718314e-02 -5.48160791e-01 -2.60233223e-01
3.30130011e-01 -5.73503315e-01 -8.89316499e-02 5.20353019e-01
4.26996723e-02 -3.45797926e-01 -4.87556458e-01 7.55751738e-03
4.16382253e-01 4.91517663e-01 1.00526676e-01 1.83668092e-01
6.07417583e-01 -2.87344337e-01 3.86215717e-01 9.18220542e-03
3.22836079e-02 1.07319689e+00 3.43422920e-01 -7.13145733e-01
-1.32593083e+00 -1.12688398e+00 -5.38112074e-02 8.12014282e-01
7.98926726e-02 -2.46345326e-01 -1.06970799e+00 -5.96754670e-01
-5.52137829e-02 4.62629199e-01 -9.58109200e-01 3.89767438e-02
-7.69490123e-01 -3.05714726e-01 3.80256474e-01 3.58710319e-01
4.14977044e-01 -8.35352004e-01 -5.24963737e-01 -1.48594588e-01
-1.45743236e-01 -1.47907484e+00 -8.08348775e-01 -4.77929354e-01
-7.15479016e-01 -7.84107566e-01 -9.23598766e-01 -6.89241886e-01
8.71277809e-01 -6.59565702e-02 1.09015989e+00 1.28626898e-01
-2.67337888e-01 4.44227070e-01 4.05734107e-02 -2.41844673e-02
-5.26522815e-01 -2.81931520e-01 1.17829010e-01 6.19432509e-01
-4.39023882e-01 -9.63082552e-01 -5.82331896e-01 1.81637824e-01
-4.68783826e-01 5.05661547e-01 1.88567832e-01 6.85710669e-01
7.17756093e-01 -5.64873695e-01 5.94473109e-02 -7.16678083e-01
2.92984128e-01 -1.04005009e-01 -5.10500073e-01 1.97035670e-01
-2.17220813e-01 -1.37171060e-01 7.67429292e-01 -6.87570453e-01
-1.13994241e+00 5.43356955e-01 -2.46551484e-01 -8.02594602e-01
-2.69037306e-01 -3.25203836e-01 -4.52704161e-01 -3.08369517e-01
4.67739344e-01 2.05434054e-01 4.84274596e-01 -4.51892376e-01
4.37876940e-01 2.96094149e-01 7.32674420e-01 -4.86035913e-01
9.96586621e-01 9.00907397e-01 3.04982364e-01 -7.77012587e-01
-6.82815909e-01 1.36691004e-01 -1.01785576e+00 -5.89880884e-01
8.93286228e-01 -9.30226147e-01 -9.47009504e-01 3.86808723e-01
-1.11095750e+00 -1.88640431e-01 -3.25033128e-01 1.66576728e-01
-7.19960868e-01 2.99808323e-01 -5.52717030e-01 -6.98377550e-01
-5.38715720e-01 -1.18607259e+00 1.61382520e+00 1.73629299e-01
-4.55517948e-01 -9.06076372e-01 -1.21869661e-01 6.50842190e-01
8.56981277e-02 9.16097283e-01 7.11554110e-01 9.30577703e-03
-5.50865948e-01 -2.51414597e-01 8.82375911e-02 9.02015641e-02
-1.53384462e-01 1.78529382e-01 -1.12495399e+00 -1.47340938e-01
-2.75873452e-01 -3.92955452e-01 1.61076933e-01 2.63594121e-01
9.28800762e-01 -4.47512835e-01 -4.89249937e-02 1.15649915e+00
1.07914531e+00 -1.70419946e-01 4.68061656e-01 -2.65792347e-02
9.59241748e-01 8.07392836e-01 2.15256318e-01 5.35686016e-01
5.45383692e-01 1.14727759e+00 5.99370122e-01 -3.99149418e-01
-4.22674924e-01 -4.80330735e-01 3.82807374e-01 3.75524104e-01
-5.35509109e-01 2.39781052e-01 -3.93271387e-01 1.72584027e-01
-1.30297399e+00 -9.03471947e-01 -4.10362259e-02 2.05507112e+00
6.55520558e-01 -3.84162158e-01 6.07406974e-01 1.25430562e-02
4.83843565e-01 1.25326559e-01 -8.03911984e-01 -3.93062353e-01
-2.27320954e-01 2.04683855e-01 3.55543979e-02 6.62945569e-01
-7.07695365e-01 9.63829458e-01 6.42293406e+00 4.66671407e-01
-1.30354393e+00 2.52709482e-02 6.19533420e-01 -3.10544163e-01
-6.40177667e-01 -2.32725754e-01 -5.97661436e-01 -8.71202722e-02
4.34994817e-01 2.14616377e-02 5.81988394e-01 7.65276611e-01
7.76015967e-02 2.40701854e-01 -9.75970149e-01 1.15631044e+00
5.02899706e-01 -1.32149899e+00 2.76325285e-01 2.72703767e-01
5.70399284e-01 -6.37194276e-01 1.63872823e-01 -1.55729458e-01
6.73891306e-02 -1.20236015e+00 1.15450394e+00 4.05322820e-01
1.28998697e+00 -7.51396894e-01 2.43089899e-01 3.02673638e-01
-1.07228351e+00 1.87854484e-01 2.22125530e-01 2.65816063e-01
3.93465459e-01 -1.07657380e-01 -6.88746989e-01 3.17680955e-01
6.12434566e-01 4.35535342e-01 -4.28611279e-01 4.21194643e-01
-1.71903238e-01 -4.13137041e-02 -3.52201253e-01 2.64845282e-01
-9.08831134e-02 -3.75521123e-01 7.01592267e-01 8.56954634e-01
2.52480298e-01 2.83712268e-01 -7.21254721e-02 1.02749348e+00
-1.02292560e-01 2.54100859e-01 -5.66395998e-01 4.03854311e-01
8.32514465e-02 1.42853081e+00 -3.94045085e-01 -1.78217754e-01
-1.92184612e-01 1.26014161e+00 3.74710351e-01 4.09828365e-01
-8.63551378e-01 2.90690571e-01 6.94239855e-01 5.90068877e-01
1.42848879e-01 -1.16536714e-01 -2.67840445e-01 -1.30507255e+00
7.72259980e-02 -9.82802153e-01 5.39205223e-02 -1.03342652e+00
-7.55805850e-01 9.08548653e-01 -9.24696028e-02 -1.06891084e+00
-4.36607957e-01 -2.86507666e-01 -5.17861843e-01 7.89864659e-01
-1.14538920e+00 -1.68924415e+00 -6.63823366e-01 9.42761898e-01
4.58675444e-01 4.74649062e-03 1.07105470e+00 8.42480138e-02
-5.81607401e-01 8.25556219e-01 -5.31091452e-01 9.52667370e-02
4.97229338e-01 -9.65725839e-01 4.14065897e-01 3.58398438e-01
6.17992990e-02 2.27315024e-01 6.91855252e-01 -3.69012445e-01
-1.44593704e+00 -7.85133719e-01 5.28366148e-01 -4.54534113e-01
1.66238233e-01 -6.06005967e-01 -7.49619722e-01 7.59286106e-01
4.00425419e-02 1.57811239e-01 4.76070374e-01 -9.78544876e-02
-4.63785738e-01 -1.75731748e-01 -1.45909894e+00 7.41883278e-01
1.09935868e+00 -3.26102495e-01 -1.43943757e-01 1.22687668e-01
2.84156770e-01 -8.60336065e-01 -1.03483295e+00 1.58454344e-01
1.12982106e+00 -1.20068192e+00 1.15117037e+00 -5.10872722e-01
3.99630010e-01 -1.91394463e-01 1.13974763e-02 -1.04392719e+00
-9.37948897e-02 -9.75004315e-01 -7.03316703e-02 9.86566544e-01
2.88400263e-01 -2.27368608e-01 9.28696275e-01 8.50382626e-01
1.50254950e-01 -8.45687509e-01 -8.94850552e-01 -3.82720411e-01
5.95144788e-03 -1.43596008e-01 9.01901722e-01 7.63151526e-01
-2.96400100e-01 4.38298553e-01 -7.12267697e-01 1.74806714e-01
1.01283562e+00 3.22448194e-01 1.24069583e+00 -1.25491798e+00
-1.67832881e-01 -2.70865023e-01 -4.41282302e-01 -9.28791344e-01
4.10109997e-01 -6.44655406e-01 -2.98517019e-01 -1.18542802e+00
2.43618533e-01 -2.76961923e-01 7.24962234e-01 3.58130813e-01
1.07445791e-01 6.53362453e-01 3.64966929e-01 1.43633232e-01
-6.44825771e-02 5.81996441e-01 1.84597635e+00 1.48624375e-01
-1.62040532e-01 4.11576889e-02 -5.22940516e-01 9.98596668e-01
4.55020338e-01 -9.37871113e-02 -2.47914031e-01 -2.55307853e-01
-7.58004934e-02 5.30015469e-01 6.85449302e-01 -6.48452282e-01
-2.06302125e-02 -9.33387652e-02 7.96482325e-01 -3.74787539e-01
1.09870541e+00 -8.88040423e-01 7.91490376e-01 7.21908361e-02
-9.47362706e-02 1.49686068e-01 5.51274419e-02 3.78221750e-01
8.46214890e-02 2.08760619e-01 1.08934939e+00 -2.20759377e-01
-2.10554078e-01 6.41250134e-01 7.44653195e-02 -3.95134911e-02
1.00624359e+00 -6.50283277e-01 1.74351647e-01 -8.97839367e-01
-1.16940892e+00 -1.39218509e-01 1.07048297e+00 4.00716484e-01
8.24509442e-01 -1.43140197e+00 -8.44207048e-01 6.37305200e-01
-2.01551527e-01 1.47791073e-01 3.88418525e-01 7.27686048e-01
-5.94126523e-01 -2.42019162e-01 -4.38273281e-01 -7.09255397e-01
-1.72292173e+00 3.26292127e-01 6.91664219e-01 7.10281655e-02
-1.13526654e+00 7.19729841e-01 5.73481023e-01 -4.26483959e-01
2.39053696e-01 1.40606180e-01 -1.41886503e-01 -6.08408414e-02
5.14790416e-01 -6.83263503e-03 -3.25643092e-01 -1.53056741e+00
-1.83015108e-01 1.22890997e+00 1.99322447e-01 -2.47914821e-01
1.31420934e+00 -2.67508775e-01 5.20228259e-02 2.91212797e-01
1.30519223e+00 2.89324045e-01 -1.63118994e+00 -3.95244844e-02
-7.80581653e-01 -5.78905880e-01 -9.98508334e-02 -5.43938458e-01
-1.48505425e+00 8.52440894e-01 2.25849211e-01 -3.28278989e-01
1.23600018e+00 1.08425260e-01 9.33759630e-01 -2.82496899e-01
3.97149861e-01 -7.56683230e-01 1.26778066e-01 4.93868515e-02
1.24025595e+00 -9.38743532e-01 1.96344718e-01 -6.59391463e-01
-8.17437172e-01 1.28789747e+00 5.86093366e-01 -7.36437514e-02
5.17024696e-01 3.29794198e-01 1.66021481e-01 -2.90267259e-01
-4.57935601e-01 1.42221019e-01 5.21141410e-01 6.64020658e-01
4.83240373e-02 1.90248825e-02 3.32484484e-01 3.66931647e-01
-5.78592479e-01 -2.54737407e-01 5.42673171e-01 4.37770247e-01
4.11072373e-02 -8.66932273e-01 -6.56975985e-01 -9.05384123e-02
-4.14161980e-01 2.98628896e-01 -6.53832078e-01 1.03407550e+00
1.62322909e-01 5.23715794e-01 9.91265252e-02 -3.64270627e-01
6.68409765e-01 7.27624819e-03 8.92167985e-01 -4.06616092e-01
-7.00436831e-01 3.32560182e-01 1.10589884e-01 -6.94286823e-01
-3.41296017e-01 -7.81688869e-01 -1.04361236e+00 -6.59090936e-01
-1.49120599e-01 -3.65730882e-01 5.91545701e-01 6.77669942e-01
2.56925613e-01 7.16258734e-02 7.12749779e-01 -1.32195997e+00
-2.22064614e-01 -6.10628784e-01 -6.75293386e-01 6.77013814e-01
6.25228643e-01 -6.58084571e-01 -3.99655223e-01 2.14782074e-01] | [12.789714813232422, -0.3517579138278961] |
fefe8a75-7d62-4455-8141-ccf3ccc318a7 | deltar-depth-estimation-from-a-light-weight | 2209.13362 | null | https://arxiv.org/abs/2209.13362v1 | https://arxiv.org/pdf/2209.13362v1.pdf | DELTAR: Depth Estimation from a Light-weight ToF Sensor and RGB Image | Light-weight time-of-flight (ToF) depth sensors are small, cheap, low-energy and have been massively deployed on mobile devices for the purposes like autofocus, obstacle detection, etc. However, due to their specific measurements (depth distribution in a region instead of the depth value at a certain pixel) and extremely low resolution, they are insufficient for applications requiring high-fidelity depth such as 3D reconstruction. In this paper, we propose DELTAR, a novel method to empower light-weight ToF sensors with the capability of measuring high resolution and accurate depth by cooperating with a color image. As the core of DELTAR, a feature extractor customized for depth distribution and an attention-based neural architecture is proposed to fuse the information from the color and ToF domain efficiently. To evaluate our system in real-world scenarios, we design a data collection device and propose a new approach to calibrate the RGB camera and ToF sensor. Experiments show that our method produces more accurate depth than existing frameworks designed for depth completion and depth super-resolution and achieves on par performance with a commodity-level RGB-D sensor. Code and data are available at https://zju3dv.github.io/deltar/. | ['Zhaopeng Cui', 'yinda zhang', 'Guofeng Zhang', 'Hujun Bao', 'Han Zhou', 'Wenqi Dong', 'Xinyang Liu', 'Yijin Li'] | 2022-09-27 | null | null | null | null | ['depth-completion'] | ['computer-vision'] | [ 3.01711291e-01 -2.21593723e-01 -1.86401624e-02 -4.43072885e-01
-7.13046312e-01 -1.83924884e-01 1.69328526e-01 -3.54232490e-01
-4.53791827e-01 4.41503644e-01 6.46328256e-02 -1.52656138e-01
1.35721490e-01 -9.46041524e-01 -5.38085282e-01 -5.98996699e-01
4.12584841e-01 2.05438629e-01 4.60012048e-01 -1.11223578e-01
2.54102260e-01 5.40306091e-01 -1.97232771e+00 6.99365512e-02
9.35713351e-01 1.44249570e+00 5.94404101e-01 5.95835268e-01
1.12875402e-01 5.30920208e-01 -3.39461505e-01 2.83977650e-02
4.24677432e-01 -1.90155998e-01 -5.28954268e-02 -3.61781986e-03
5.62043190e-01 -1.05531347e+00 -6.11470699e-01 9.33596373e-01
6.54063404e-01 -5.24047427e-02 1.94606677e-01 -9.56929505e-01
-2.44708702e-01 -1.55752420e-01 -7.66225636e-01 5.32143824e-02
6.74914300e-01 2.21133783e-01 3.29709560e-01 -8.64141881e-01
4.70827043e-01 9.16858971e-01 4.30827528e-01 8.27081859e-01
-6.95803523e-01 -5.64066589e-01 -1.55559346e-01 1.23109877e-01
-1.16363454e+00 -5.63945353e-01 8.58844936e-01 -1.24484360e-01
8.96664321e-01 1.83792859e-01 8.23944509e-01 1.04592383e+00
3.14365029e-01 4.41480547e-01 1.01761174e+00 -9.75381210e-02
3.29409271e-01 -3.32505591e-02 -5.71709216e-01 6.38153613e-01
2.85056174e-01 1.99955106e-01 -7.89813161e-01 3.80605072e-01
1.22232080e+00 4.93171751e-01 -5.61574101e-01 -4.36534762e-01
-1.21655381e+00 4.35748667e-01 6.91867709e-01 2.03504160e-01
-2.88040638e-01 3.36104214e-01 -1.91228613e-01 -1.57565642e-02
4.16492492e-01 1.50542036e-01 -2.02759817e-01 -5.55554390e-01
-6.31564021e-01 -1.78030327e-01 1.70617133e-01 9.08358693e-01
9.61399257e-01 -1.24493316e-01 2.99636275e-01 3.18595111e-01
3.31517249e-01 9.03083742e-01 6.90763950e-01 -1.32234192e+00
5.08692682e-01 7.10544646e-01 2.91916966e-01 -7.27119684e-01
-4.78666574e-01 2.37467773e-02 -6.14965141e-01 5.74337840e-01
2.65556872e-01 -2.16928925e-02 -8.39755893e-01 1.17578220e+00
5.01689017e-01 4.40743752e-02 2.73990557e-02 1.32113159e+00
8.66579533e-01 4.46214885e-01 -6.16659105e-01 -1.55886048e-02
1.11296749e+00 -7.03232229e-01 -6.15132868e-01 -4.54440743e-01
3.53123933e-01 -3.15484196e-01 1.35969126e+00 6.60859704e-01
-1.00051463e+00 -4.46643651e-01 -1.26793230e+00 -6.96937501e-01
-1.44080862e-01 -2.03881711e-02 8.14953089e-01 5.80726326e-01
-1.01866806e+00 3.84156346e-01 -9.50987577e-01 -1.72284275e-01
2.31282204e-01 2.21820205e-01 -4.14226979e-01 -4.02171910e-01
-8.30929041e-01 6.23218298e-01 1.00097626e-01 -2.02809554e-03
-7.08347380e-01 -6.12702847e-01 -7.19357669e-01 -1.70817003e-01
2.27995947e-01 -7.09810436e-01 1.09628677e+00 -5.84032536e-01
-1.87833011e+00 7.64810920e-01 -2.49945477e-01 -1.76301718e-01
4.62883860e-01 -3.71203423e-01 -1.28387585e-01 5.89693666e-01
-1.27127627e-02 5.78350484e-01 7.82106042e-01 -1.05678749e+00
-8.65139902e-01 -8.24584603e-01 2.46753216e-01 3.68131340e-01
-4.06798780e-01 -5.37502229e-01 -4.95938390e-01 1.78938493e-01
6.20670021e-01 -5.63502133e-01 -2.33993053e-01 3.88019294e-01
3.08637153e-02 2.87355125e-01 9.50833797e-01 -2.56802112e-01
8.01091373e-01 -2.26482701e+00 -1.11719117e-01 -1.88194215e-01
4.35123116e-01 6.27944246e-02 3.71695071e-01 -2.79171066e-03
3.32301497e-01 -3.16864073e-01 -1.18516609e-01 -5.86518943e-01
-4.42073733e-01 1.89095303e-01 -2.57274173e-02 5.96157551e-01
-2.88287610e-01 6.71485603e-01 -1.08756340e+00 -2.72556990e-01
8.33883286e-01 9.25146639e-01 -3.91810238e-01 2.03636274e-01
-9.86674950e-02 7.91130602e-01 -4.17523533e-01 9.74035800e-01
8.39578331e-01 -8.09401721e-02 -1.13009863e-01 -3.30826759e-01
-4.20984834e-01 2.48847231e-01 -9.61484432e-01 2.22695708e+00
-6.62706852e-01 6.52080595e-01 1.08348154e-01 -2.96343893e-01
9.86778200e-01 4.67784144e-02 6.40178204e-01 -1.24999523e+00
4.37951386e-01 4.81750250e-01 -4.38296169e-01 -6.16349816e-01
7.43116081e-01 -5.45470184e-03 9.66389999e-02 1.81755126e-01
-2.78916359e-01 -5.15521646e-01 -1.61345690e-01 -5.55805378e-02
1.11790693e+00 3.38561773e-01 5.84422126e-02 3.05732012e-01
2.77245373e-01 -1.95572734e-01 4.55334455e-01 2.39685416e-01
-2.38250762e-01 9.03382063e-01 7.89103750e-03 -3.56391311e-01
-1.05365431e+00 -1.03138077e+00 -3.07982296e-01 4.69425648e-01
6.92065418e-01 -1.62601620e-01 -6.12134099e-01 -2.10491329e-01
6.97360709e-02 3.46211076e-01 -5.44327974e-01 -7.51418411e-04
-2.70671040e-01 -2.62307286e-01 5.66514470e-02 5.72526217e-01
9.17596042e-01 -4.86853361e-01 -1.50725889e+00 -1.70657951e-02
-2.28341028e-01 -1.25644743e+00 -2.55865067e-01 1.06741115e-01
-1.06427944e+00 -9.95748460e-01 -5.30052304e-01 -1.52857929e-01
5.60003102e-01 8.50640953e-01 7.35902905e-01 -2.83517599e-01
-2.58685261e-01 4.56026822e-01 -3.75712693e-01 -3.08649153e-01
4.25377786e-01 -2.76156336e-01 1.50152436e-02 1.22365053e-03
3.09604198e-01 -7.68000841e-01 -1.16781700e+00 1.85298592e-01
-9.09182668e-01 2.37595171e-01 5.29665112e-01 3.20964545e-01
7.85267234e-01 -1.22610725e-01 -1.58301696e-01 -5.23019552e-01
9.83624831e-02 -1.76862106e-01 -9.36494410e-01 -3.45111579e-01
-5.98596573e-01 -1.01244919e-01 3.06872159e-01 -3.23477536e-01
-9.82439876e-01 2.62508780e-01 -2.40322202e-01 -7.10648179e-01
2.87001263e-02 -4.98550385e-03 -3.54991704e-01 -2.82489389e-01
7.15378165e-01 1.13601521e-01 -1.91178527e-02 -4.13911939e-01
7.23844171e-02 8.69077086e-01 8.33905816e-01 -1.28056213e-01
6.38269484e-01 1.14343667e+00 1.83926493e-01 -7.00421453e-01
-6.40641928e-01 -5.51073790e-01 -4.69181001e-01 -4.25162315e-01
7.23143101e-01 -1.22734058e+00 -9.06028986e-01 5.37341475e-01
-9.45751309e-01 -3.59872341e-01 -2.37173274e-01 7.46530831e-01
-4.84309852e-01 2.37015292e-01 -3.33181471e-01 -8.12336326e-01
-3.78570706e-01 -1.02669108e+00 1.45191646e+00 6.32531524e-01
3.50750297e-01 -7.10915506e-01 -7.99159706e-02 5.46393514e-01
5.57422161e-01 4.52803791e-01 8.76558498e-02 6.08039558e-01
-9.57919061e-01 -2.65524179e-01 -2.60492027e-01 5.85980043e-02
2.60802299e-01 -1.19972184e-01 -1.35952485e+00 -1.38691068e-01
3.30256462e-01 -1.80240378e-01 8.53804350e-01 4.36300486e-01
1.23525620e+00 2.38679156e-01 -2.98655540e-01 1.34104443e+00
1.75623381e+00 2.81543404e-01 8.68746161e-01 4.40659523e-01
1.05018699e+00 3.34535807e-01 8.09256256e-01 7.15204597e-01
6.18609309e-01 7.74306536e-01 9.30135787e-01 5.72908968e-02
-1.43914133e-01 -3.42029899e-01 3.67351294e-01 4.42929536e-01
-3.42209578e-01 -1.00626625e-01 -6.79814637e-01 2.48081118e-01
-1.54139471e+00 -7.30393827e-01 -1.81351006e-01 2.44826531e+00
5.72845101e-01 1.11827612e-01 -8.91938061e-02 2.19864562e-01
2.70319939e-01 5.42537160e-02 -8.59260678e-01 -2.16736868e-01
-1.45072669e-01 6.98612928e-02 9.07573640e-01 6.07184231e-01
-6.34906173e-01 5.79778254e-01 5.23147392e+00 3.20174396e-01
-1.57957351e+00 2.57293642e-01 2.57817209e-01 -5.86688876e-01
-5.60327947e-01 -2.25487366e-01 -6.38604403e-01 5.20968199e-01
7.92997599e-01 9.91421044e-02 4.99700159e-01 9.90746498e-01
2.38663375e-01 -6.75836980e-01 -9.72885668e-01 1.56170321e+00
7.41981119e-02 -1.02325344e+00 -3.18058670e-01 1.93221420e-01
6.05893970e-01 2.97090858e-01 1.50666118e-01 -2.98256874e-01
-1.73067704e-01 -7.57358849e-01 6.44283652e-01 4.50808167e-01
1.31139457e+00 -7.72810757e-01 5.48312962e-01 3.78504306e-01
-1.00130749e+00 -1.05869517e-01 -5.55625439e-01 -2.83706367e-01
9.73484367e-02 9.61334467e-01 -5.85342526e-01 4.17579174e-01
1.00409746e+00 7.44047105e-01 -4.09660488e-01 8.15078974e-01
-3.92682314e-01 -2.49392554e-01 -4.74018961e-01 1.46901933e-03
-1.07063733e-01 -1.99627280e-01 2.36617416e-01 5.56398630e-01
7.45210648e-01 2.53883570e-01 -3.29798251e-01 7.41306484e-01
6.37756884e-02 -3.67968917e-01 -7.68032670e-01 3.01539361e-01
4.51162577e-01 1.40129387e+00 -4.70263571e-01 -1.41674384e-01
-5.93831480e-01 1.01600695e+00 8.32985714e-02 1.83604360e-01
-8.61428201e-01 -4.88666207e-01 8.18192005e-01 5.11216342e-01
1.10555969e-01 -4.94603187e-01 -5.56050777e-01 -1.40116131e+00
2.84572929e-01 -2.37609178e-01 -7.64717758e-02 -1.23420703e+00
-6.07649803e-01 6.42650247e-01 -3.84999782e-01 -1.63280046e+00
-1.94090113e-01 -8.13439906e-01 -2.46910602e-01 7.27372766e-01
-1.80541003e+00 -8.01674426e-01 -1.19799376e+00 8.94907653e-01
2.38726780e-01 3.44854027e-01 6.57677472e-01 4.66013968e-01
-3.96458179e-01 2.73942769e-01 6.99931607e-02 -1.41877800e-01
6.20055735e-01 -1.07451987e+00 2.22359210e-01 9.51802135e-01
-2.47824550e-01 2.78930008e-01 4.33360279e-01 -5.14614344e-01
-2.13202047e+00 -7.07006037e-01 3.05226415e-01 -4.30090189e-01
2.02439025e-01 -5.39001584e-01 -5.11479378e-01 3.60692352e-01
-3.05548519e-01 3.83879572e-01 1.31604016e-01 -4.27391678e-01
-1.22188166e-01 -5.70799589e-01 -1.48989451e+00 4.22581494e-01
1.25885820e+00 -7.14207590e-01 -8.67314711e-02 5.38437441e-02
7.29352415e-01 -1.01185310e+00 -7.19687343e-01 3.84595990e-01
8.17879796e-01 -1.70435524e+00 8.99482191e-01 6.59480989e-01
4.96986568e-01 -5.12237668e-01 -3.16805333e-01 -9.80963707e-01
-5.14687635e-02 -4.14082438e-01 -3.70218486e-01 6.95294857e-01
2.09746528e-02 -9.59845603e-01 1.08997953e+00 7.06495523e-01
-3.48278910e-01 -5.99199653e-01 -1.31555533e+00 -3.80082965e-01
-6.19102061e-01 -6.44401371e-01 6.86165631e-01 4.41931903e-01
7.19075277e-02 1.61819026e-01 -2.75454611e-01 3.43025744e-01
7.73028612e-01 2.43064597e-01 8.26846361e-01 -1.03799140e+00
-2.80436367e-01 -1.12332083e-01 -6.19987190e-01 -1.35738885e+00
-5.72701514e-01 -2.57316858e-01 2.58881617e-02 -1.71476769e+00
-2.53689528e-01 -3.71814102e-01 1.16424166e-01 1.25086874e-01
1.12893075e-01 5.62765062e-01 -1.36707947e-01 1.94915369e-01
-5.14911771e-01 5.26728094e-01 1.42300200e+00 1.61670864e-01
-2.04101324e-01 -1.58190042e-01 -5.87312698e-01 6.44982815e-01
6.04069531e-01 -1.20841444e-01 -5.48088968e-01 -8.60839486e-01
3.44853014e-01 3.30558568e-01 3.45218450e-01 -1.44818556e+00
3.60599995e-01 -5.71129434e-02 6.79957032e-01 -6.08013451e-01
8.73512328e-01 -1.07588935e+00 9.95570347e-02 4.52986002e-01
4.00735497e-01 -7.63838589e-02 -9.52428058e-02 2.44123951e-01
-1.16461232e-01 1.40974090e-01 7.66242325e-01 -1.98601365e-01
-1.01165736e+00 5.49915612e-01 6.30497038e-02 -2.79348135e-01
1.03898370e+00 -7.15594053e-01 -5.14773190e-01 -5.30003488e-01
9.82159935e-03 -2.45741215e-02 1.01693285e+00 2.82880276e-01
1.21741927e+00 -1.24054420e+00 -2.55259365e-01 6.27243578e-01
2.92908341e-01 5.62235773e-01 4.21616256e-01 8.06670666e-01
-7.95510232e-01 4.15614396e-01 -3.06643635e-01 -8.93914938e-01
-7.96371698e-01 4.84371424e-01 5.01231313e-01 3.75487447e-01
-8.65606070e-01 7.87569940e-01 1.02148339e-01 -3.20786284e-03
1.71914861e-01 -6.74205124e-01 1.29383564e-01 -2.19718784e-01
8.90421867e-01 3.34518075e-01 1.66825131e-01 -3.58229935e-01
-3.24337870e-01 1.05539536e+00 4.89027619e-01 -1.88299403e-01
1.19984329e+00 -5.77437937e-01 1.73542276e-01 5.59551299e-01
1.08244944e+00 8.16134885e-02 -1.80130994e+00 -6.22585788e-02
-6.78311229e-01 -1.12962162e+00 3.91614258e-01 -4.02045667e-01
-1.20295179e+00 9.47005987e-01 9.12598670e-01 -7.71167204e-02
1.54787266e+00 -3.38433571e-02 8.68908405e-01 8.61756429e-02
9.33629811e-01 -9.65254784e-01 3.61214578e-02 2.89752454e-01
5.48469841e-01 -1.40938365e+00 1.43048889e-03 -3.10754538e-01
-3.28380644e-01 1.14626384e+00 7.29616284e-01 1.13610886e-01
2.83261269e-01 5.58690310e-01 1.72120035e-01 -6.23879097e-02
-4.48806763e-01 -2.97896385e-01 -1.10434234e-01 8.91737223e-01
1.27226159e-01 -2.66341697e-02 1.73823386e-01 6.86245263e-02
-1.76906645e-01 1.91556975e-01 6.45443559e-01 9.27106798e-01
-6.77164495e-01 -7.74682820e-01 -3.62946689e-01 2.17429087e-01
-1.41220853e-01 9.69709530e-02 1.63464621e-01 5.00241697e-01
3.09111059e-01 9.25235033e-01 2.38228530e-01 -7.50099361e-01
4.64711934e-01 -5.10580599e-01 7.70724237e-01 -3.73344541e-01
2.92651746e-02 -5.23478165e-02 -1.36209175e-01 -1.24278843e+00
-4.45362061e-01 -5.26385248e-01 -1.35570717e+00 -5.33812046e-01
1.15057558e-01 -3.75513226e-01 1.12021422e+00 6.18908048e-01
4.37874883e-01 3.59946340e-01 8.61529708e-01 -1.22164547e+00
1.04437925e-01 -8.95339727e-01 -7.65154004e-01 1.76305786e-01
6.30241096e-01 -7.31555998e-01 -5.20530820e-01 -3.10000509e-01] | [9.043135643005371, -2.4657716751098633] |
ba97f202-94cf-4d77-849a-22f56a3dd65c | findings-of-the-vardial-evaluation-campaign-3 | 2305.20080 | null | https://arxiv.org/abs/2305.20080v1 | https://arxiv.org/pdf/2305.20080v1.pdf | Findings of the VarDial Evaluation Campaign 2023 | This report presents the results of the shared tasks organized as part of the VarDial Evaluation Campaign 2023. The campaign is part of the tenth workshop on Natural Language Processing (NLP) for Similar Languages, Varieties and Dialects (VarDial), co-located with EACL 2023. Three separate shared tasks were included this year: Slot and intent detection for low-resource language varieties (SID4LR), Discriminating Between Similar Languages -- True Labels (DSL-TL), and Discriminating Between Similar Languages -- Speech (DSL-S). All three tasks were organized for the first time this year. | ['Marcos Zampieri', 'Yves Scherrer', 'Barbara Plank', 'Kai North', 'Nikola Ljubešić', 'Mourhaf Kazzaz', 'Tommi Jauhiainen', 'Rob van der Goot', 'Çağrı Çöltekin', 'Noëmi Aepli'] | 2023-05-31 | null | null | null | null | ['intent-detection'] | ['natural-language-processing'] | [-3.87755513e-01 6.73766062e-02 -2.92122483e-01 -6.61924720e-01
-1.27923703e+00 -9.01062369e-01 8.03919554e-01 4.48857635e-01
-5.80367923e-01 4.83859777e-01 5.75139165e-01 -4.37292993e-01
2.99356043e-01 -2.16586590e-01 -2.93170232e-02 2.93175448e-02
-7.15209171e-02 7.54646897e-01 2.05774769e-01 -1.98685586e-01
-3.73975784e-02 5.53140819e-01 -9.96225059e-01 7.73433030e-01
5.81342280e-01 4.39559370e-01 3.22916776e-01 6.06382251e-01
-4.97781783e-01 5.79321980e-01 -3.28404397e-01 -2.94576377e-01
1.97743878e-01 -2.00164974e-01 -1.24153900e+00 1.05629399e-01
5.95208228e-01 9.75149591e-03 2.64396846e-01 1.03082085e+00
4.21761394e-01 8.37735645e-03 4.74622726e-01 -1.23940659e+00
-3.60728592e-01 9.82026577e-01 -3.44087481e-01 4.25787657e-01
8.75691593e-01 1.02662429e-01 1.29434526e+00 -1.19845188e+00
1.04486370e+00 1.82304513e+00 9.91379678e-01 2.91176349e-01
-1.52717960e+00 -6.08017802e-01 2.51655757e-01 -2.20629781e-01
-1.79625750e+00 -1.10678482e+00 4.14752990e-01 -6.56619549e-01
1.70240104e+00 3.82675052e-01 1.65267587e-01 9.86029387e-01
-2.51165092e-01 1.20748246e+00 1.37347090e+00 -7.12637603e-01
1.99450880e-01 6.64130390e-01 5.99346638e-01 4.51569140e-01
-1.02350853e-01 9.81944147e-03 -4.37607437e-01 -2.17419997e-01
1.33215174e-01 -1.14338434e+00 1.30794644e-01 2.03498572e-01
-9.42060947e-01 1.08678973e+00 -3.45642477e-01 7.98300564e-01
-3.11439455e-01 -6.32028043e-01 9.98292923e-01 4.50794905e-01
8.35134029e-01 4.88393217e-01 -8.43212426e-01 8.46591368e-02
-7.65743852e-01 4.67777997e-01 1.07511413e+00 1.09506083e+00
4.74305987e-01 3.15332323e-01 -3.87422442e-01 1.41234267e+00
6.47768795e-01 4.50748593e-01 6.10687613e-01 -4.88317490e-01
6.24054790e-01 3.05551916e-01 -5.06913625e-02 -6.31893873e-01
-7.20861375e-01 1.10974006e-01 -3.26955467e-01 -2.46175423e-01
4.42418814e-01 -4.19720173e-01 -7.54085302e-01 1.69378710e+00
1.89083740e-01 -5.84737718e-01 4.36364233e-01 4.47846174e-01
1.24484026e+00 1.03261483e+00 6.30347848e-01 -3.91149163e-01
1.70210421e+00 -6.25573695e-01 -7.55524993e-01 -4.89408940e-01
1.11383998e+00 -1.19715464e+00 9.60614204e-01 1.27180651e-01
-1.11374819e+00 -4.46386337e-01 -5.91659486e-01 -1.69179007e-01
-7.26714730e-01 2.52142429e-01 5.32696366e-01 1.06773913e+00
-1.31491768e+00 -4.02578712e-01 -2.11283222e-01 -1.05798745e+00
-1.84451908e-01 -2.40347851e-02 -4.65089947e-01 5.27113974e-02
-1.36474991e+00 8.43990266e-01 4.44481641e-01 -2.86023051e-01
-7.86850333e-01 -5.13555825e-01 -1.16263938e+00 -3.97789180e-01
1.38541073e-01 3.24741930e-01 1.45422888e+00 -9.81076896e-01
-1.37327564e+00 1.98372829e+00 -2.39229187e-01 -2.24374726e-01
2.36723006e-01 -1.69439301e-01 -1.26155639e+00 -5.90361476e-01
7.51518905e-01 7.43031204e-01 2.18772709e-01 -9.28480387e-01
-7.55521894e-01 -5.17638564e-01 -3.33117217e-01 1.92407563e-01
3.43758881e-01 1.28024209e+00 -2.32505918e-01 -6.58706903e-01
5.43776155e-02 -8.91986966e-01 -5.46036996e-02 -6.96557343e-01
-7.69605815e-01 -8.87697101e-01 6.38886929e-01 -1.31416619e+00
9.00094807e-01 -2.49806571e+00 -2.83870429e-01 -2.91937012e-02
-2.33257383e-01 5.82740724e-01 -3.58283669e-01 6.05920017e-01
-1.71560243e-01 2.88323045e-01 1.36296496e-01 -4.90980476e-01
3.17959219e-01 1.68273240e-01 -3.76008332e-01 3.71830076e-01
2.19124570e-01 6.26764774e-01 -6.26171589e-01 -3.11862707e-01
8.42553824e-02 -6.54015094e-02 -7.95196295e-02 -1.20143838e-01
-3.89526039e-01 -5.71623109e-02 -8.27116296e-02 7.02032864e-01
8.05923939e-01 6.41462266e-01 4.74464536e-01 7.28862584e-02
-9.69789922e-01 8.31032872e-01 -1.20233405e+00 1.29582620e+00
-4.39801872e-01 6.95426404e-01 5.66933692e-01 -6.34460211e-01
9.35722828e-01 5.99530339e-01 -8.32123458e-02 -6.91012025e-01
-7.07556978e-02 5.56395352e-01 -9.83996019e-02 -3.61323029e-01
9.41593647e-01 -1.65428698e-01 -6.85451984e-01 1.87075615e-01
3.73408407e-01 -1.16773114e-01 2.78341591e-01 -5.58571657e-03
6.77593172e-01 -1.51667416e-01 9.17591333e-01 -8.60741913e-01
7.41521418e-01 4.01985228e-01 9.20135736e-01 5.70075572e-01
-6.29062057e-01 8.73343199e-02 4.94030893e-01 -3.07158947e-01
-8.67018044e-01 -1.08998859e+00 -3.27086776e-01 1.42950439e+00
-3.99145871e-01 -6.39329314e-01 -4.45910245e-01 -6.82848096e-01
-1.29875124e-01 1.25621021e+00 -1.65003575e-02 4.84140486e-01
-5.49615502e-01 -4.10592079e-01 1.36157370e+00 2.62041003e-01
5.05081177e-01 -1.15943968e+00 8.53929520e-02 1.79541007e-01
-1.25977799e-01 -1.46913826e+00 -3.54204506e-01 4.75500882e-01
-1.11355431e-01 -6.05582178e-01 -4.74315614e-01 -1.63406396e+00
-1.49008632e-01 -1.35395110e-01 1.14113343e+00 -6.95023298e-01
-3.86253655e-01 1.98674962e-01 -2.45133728e-01 -3.89680892e-01
-8.36153388e-01 8.59020874e-02 3.93973365e-02 -2.69945025e-01
8.88685763e-01 2.57782340e-01 6.39995098e-01 1.07361846e-01
-2.42569208e-01 -1.45180538e-01 5.06600514e-02 4.57573175e-01
4.55887914e-01 -1.77347288e-01 4.80201244e-01 -1.16984022e+00
6.53309643e-01 -4.98809755e-01 -8.13520670e-01 3.85647506e-01
-2.17013419e-01 -3.01211715e-01 5.35349607e-01 -1.10354327e-01
-1.37196922e+00 3.23829144e-01 -7.14097261e-01 3.65890414e-01
-9.09707248e-01 4.64787275e-01 -6.81275964e-01 3.68329495e-01
4.42282766e-01 9.62444767e-02 -3.99838954e-01 -7.30718791e-01
4.80410188e-01 1.00291240e+00 4.91226673e-01 -3.58790666e-01
3.14742893e-01 -2.19993770e-01 -7.00155437e-01 -1.53730690e+00
-6.36746049e-01 -8.75877380e-01 -6.71657681e-01 9.05189570e-03
1.16020715e+00 -1.26497018e+00 4.98315841e-02 9.13786292e-01
-1.38627112e+00 -3.08652252e-01 -2.65428007e-01 3.54482561e-01
-1.01683281e-01 1.22489072e-01 -8.61474574e-01 -8.04558814e-01
-1.06251083e-01 -1.04859293e+00 1.08394361e+00 -2.74122894e-01
-7.73782909e-01 -1.26765573e+00 5.04310846e-01 2.99124401e-02
2.47284785e-01 -2.78437346e-01 1.11399710e+00 -1.48002613e+00
2.70576149e-01 1.33684531e-01 -3.10909361e-01 3.30308616e-01
-2.11448908e-01 -4.78348546e-02 -1.02812946e+00 -2.59577874e-02
-3.24669242e-01 -7.08655596e-01 4.95210290e-01 4.72163439e-01
2.14161873e-02 -2.07951531e-01 -3.06987077e-01 2.44939432e-01
1.34269929e+00 3.91358972e-01 2.42364630e-01 1.31113350e-01
3.07584673e-01 9.07360554e-01 5.60661852e-01 4.07142878e-01
6.49495959e-01 7.91161120e-01 -4.60086077e-01 1.30537972e-02
-3.86401325e-01 -1.96010277e-01 8.28636825e-01 9.72300053e-01
7.74482071e-01 -2.54916757e-01 -1.45849442e+00 9.07770813e-01
-1.41564369e+00 -7.36568153e-01 -6.40139222e-01 1.90303588e+00
6.61982179e-01 -1.60634443e-01 5.32532394e-01 -1.78474724e-01
9.57610309e-01 5.66895545e-01 -5.37061021e-02 -1.11872256e+00
-5.91713488e-01 -9.52672400e-03 2.62199312e-01 8.90612841e-01
-1.35796523e+00 1.69866371e+00 7.40310717e+00 8.98826897e-01
-9.58572328e-01 4.21389371e-01 6.31140172e-01 6.39940739e-01
-2.87837774e-01 1.10174812e-01 -1.53244042e+00 9.09571908e-03
1.19191647e+00 -3.28372210e-01 2.58138418e-01 9.14263785e-01
5.54412641e-02 -2.13194057e-01 -8.82609844e-01 5.89161336e-01
1.74403310e-01 -8.53794694e-01 -1.88406825e-01 -1.60757467e-01
3.62409323e-01 6.27451718e-01 -1.95815265e-01 7.58265495e-01
7.17224360e-01 -5.78671455e-01 9.03701305e-01 -9.71844569e-02
9.20031190e-01 -6.38106823e-01 6.94679320e-01 1.76829889e-01
-1.36151505e+00 2.08550498e-01 -4.67536420e-01 1.45228714e-01
4.30687785e-01 5.80167055e-01 -9.51329410e-01 2.63570160e-01
4.33593929e-01 5.70493996e-01 -4.93042827e-01 6.66962266e-01
-1.09106407e-01 8.34476471e-01 -2.81288981e-01 -1.99163303e-01
2.83983469e-01 6.93604872e-02 1.10058594e+00 1.98858166e+00
-1.50569201e-01 -1.94280460e-01 7.19134331e-01 6.26992404e-01
1.38520151e-01 5.91217458e-01 -6.34615779e-01 -3.99465382e-01
4.48487818e-01 1.14609921e+00 -7.19058275e-01 -3.74837488e-01
-4.43308085e-01 1.00326741e+00 2.79920250e-01 5.69884607e-04
-2.30389819e-01 -2.81456769e-01 8.12070489e-01 -1.26279518e-01
-4.59510423e-02 -2.91242331e-01 -1.85918882e-01 -7.32789397e-01
-3.51434827e-01 -9.04930651e-01 6.52597725e-01 -4.48682159e-01
-1.76351476e+00 9.62966681e-01 5.83394803e-02 -5.12889862e-01
-5.41758358e-01 -6.81719303e-01 -2.54586339e-01 1.14366853e+00
-1.01063979e+00 -1.30959284e+00 2.88987458e-01 7.01608121e-01
9.59937334e-01 -6.88715398e-01 1.31733096e+00 2.45022923e-01
-6.87305450e-01 6.79479480e-01 -3.43787968e-02 2.61201322e-01
7.88778067e-01 -1.12195468e+00 7.59391546e-01 8.47081006e-01
1.72755614e-01 3.47829647e-02 4.52617079e-01 -7.49333858e-01
-1.02063191e+00 -1.24264312e+00 1.95583475e+00 1.32351041e-01
1.06857681e+00 -7.27848113e-01 -5.44127464e-01 8.15267622e-01
4.22295481e-01 -4.39582050e-01 8.56043637e-01 4.89803284e-01
-6.13935590e-01 1.01941533e-01 -1.39077091e+00 3.22074175e-01
7.92250156e-01 -1.06196439e+00 -5.93932748e-01 7.93542325e-01
1.04003692e+00 -1.95275739e-01 -5.38681746e-01 -4.28499281e-02
1.82014808e-01 -5.41141093e-01 3.52448821e-01 -4.90163565e-01
-2.02482387e-01 -1.45312417e-02 -9.09298837e-01 -1.42875063e+00
-4.44110602e-01 -5.85990489e-01 9.98976707e-01 2.00928974e+00
6.31251037e-01 -7.61238277e-01 4.48017269e-02 3.77649307e-01
-2.77898043e-01 4.44206357e-01 -1.06629622e+00 -1.06379306e+00
1.09445944e-01 -8.86539876e-01 3.25733215e-01 1.26908469e+00
1.54974118e-01 7.90867388e-01 -2.65985206e-02 2.73281753e-01
3.37064952e-01 -3.07658345e-01 3.63359362e-01 -1.05345690e+00
-2.27718830e-01 -3.61946046e-01 -6.80803895e-01 -5.70887804e-01
6.92556202e-01 -1.27508616e+00 2.95912981e-01 -1.12811375e+00
-1.61049113e-01 -5.67092776e-01 1.81799859e-01 7.86299467e-01
3.78981233e-01 -3.77125107e-02 3.69425982e-01 -3.02101206e-03
-4.28761393e-01 4.75492254e-02 8.82460997e-02 1.30139152e-02
-4.51655358e-01 5.97264059e-02 -5.73611677e-01 6.49004459e-01
7.96525240e-01 -3.77415627e-01 -8.61361343e-03 -3.81397396e-01
-9.92559567e-02 5.85440136e-02 2.83952896e-02 -8.58011901e-01
-1.10794790e-01 -7.50748217e-02 1.22951576e-02 -7.88669586e-01
1.44863218e-01 -3.11071128e-01 1.36614917e-02 3.03667873e-01
-3.66378903e-01 3.41045827e-01 7.67905235e-01 -2.74689317e-01
-3.49867404e-01 -2.84697413e-01 8.05682421e-01 -2.16796249e-01
-1.36422586e+00 -1.04374982e-01 -1.24363053e+00 4.46492165e-01
1.21662343e+00 8.34763050e-02 -2.82420427e-01 -6.93509579e-02
-7.80582607e-01 1.54103711e-01 3.75981815e-02 6.23607278e-01
2.79017359e-01 -1.05119908e+00 -1.06022549e+00 5.49796641e-01
3.09372663e-01 -7.53752053e-01 1.32500604e-01 5.11128128e-01
-4.20223743e-01 9.56588805e-01 -2.19672024e-01 -3.67028683e-01
-1.47293186e+00 3.74624759e-01 1.68540180e-01 -4.39309120e-01
-2.61500299e-01 1.02558446e+00 -2.19769981e-02 -1.09659266e+00
2.22943068e-01 1.05479546e-01 -2.91475743e-01 3.33375871e-01
5.36830246e-01 2.54432857e-01 1.65223219e-02 -1.40120661e+00
-7.84346342e-01 -1.33933909e-02 -3.46322119e-01 -5.11805654e-01
1.04661906e+00 -1.65346086e-01 -3.75537366e-01 8.51333857e-01
1.10785580e+00 4.65564042e-01 -2.36586943e-01 -4.49724495e-01
6.93985760e-01 3.91386524e-02 1.75348848e-01 -1.07041669e+00
-3.47582608e-01 4.11058009e-01 6.33956254e-01 3.27008754e-01
7.02745676e-01 1.53471738e-01 7.07566559e-01 -9.64072300e-04
3.72524589e-01 -1.41587079e+00 -8.04556191e-01 1.07685339e+00
8.65995586e-01 -9.85479057e-01 -4.51163113e-01 -8.32509816e-01
-8.22336972e-01 7.81085610e-01 5.43009937e-01 1.82288617e-01
9.50909615e-01 4.14077431e-01 2.02327162e-01 -1.14163943e-01
-7.54197955e-01 -3.56853694e-01 1.89484671e-01 8.96422386e-01
8.81911695e-01 6.87831044e-01 -5.21498322e-01 6.04002893e-01
-1.75268576e-01 -6.55167460e-01 1.18540794e-01 7.81636655e-01
-3.64060342e-01 -1.25242305e+00 -2.68381387e-01 1.98373586e-01
-2.26110205e-01 -4.44005400e-01 -9.24344420e-01 1.09889817e+00
2.23238483e-01 1.12105763e+00 1.16813593e-01 -3.42625886e-01
3.27055931e-01 5.75361550e-01 4.21505943e-02 -1.02480483e+00
-9.84800458e-01 1.64071262e-01 1.10845077e+00 -4.92946655e-01
-1.34608760e-01 -1.25452077e+00 -1.02474225e+00 -3.05844158e-01
1.84864085e-02 8.05068016e-02 6.84713900e-01 8.40710580e-01
-5.05978800e-02 -1.13235041e-01 4.51142579e-01 -5.22011101e-01
-4.39042956e-01 -9.51241612e-01 -1.13227332e+00 -1.50813609e-01
-9.84533131e-02 -1.54003233e-01 -2.65599877e-01 -2.05112547e-01] | [10.235885620117188, 10.701041221618652] |
e58b18a7-ef4f-4de8-acfa-016caa66d29f | audio-visual-video-face-hallucination-with | 2211.10883 | null | https://arxiv.org/abs/2211.10883v1 | https://arxiv.org/pdf/2211.10883v1.pdf | Audio-visual video face hallucination with frequency supervision and cross modality support by speech based lip reading loss | Recently, there has been numerous breakthroughs in face hallucination tasks. However, the task remains rather challenging in videos in comparison to the images due to inherent consistency issues. The presence of extra temporal dimension in video face hallucination makes it non-trivial to learn the facial motion through out the sequence. In order to learn these fine spatio-temporal motion details, we propose a novel cross-modal audio-visual Video Face Hallucination Generative Adversarial Network (VFH-GAN). The architecture exploits the semantic correlation of between the movement of the facial structure and the associated speech signal. Another major issue in present video based approaches is the presence of blurriness around the key facial regions such as mouth and lips - where spatial displacement is much higher in comparison to other areas. The proposed approach explicitly defines a lip reading loss to learn the fine grain motion in these facial areas. During training, GANs have potential to fit frequencies from low to high, which leads to miss the hard to synthesize frequencies. Therefore, to add salient frequency features to the network we add a frequency based loss function. The visual and the quantitative comparison with state-of-the-art shows a significant improvement in performance and efficacy. | ['Vivek Singh Bawa', 'Vinay Kumar', 'Abhinav Dhall', 'Shailza Sharma'] | 2022-11-20 | null | null | null | null | ['face-hallucination'] | ['computer-vision'] | [ 1.00812286e-01 2.37495840e-01 8.02044570e-02 -6.95394203e-02
-6.01428926e-01 -1.26026839e-01 5.60708344e-01 -7.87332654e-01
1.45620942e-01 8.87559056e-01 6.46778166e-01 5.88138163e-01
-8.47803354e-02 -5.59239566e-01 -8.91074657e-01 -9.94237781e-01
5.73582985e-02 -1.55366853e-01 -2.33271588e-02 -2.98963815e-01
5.32330833e-02 5.27770579e-01 -1.76532352e+00 5.43091059e-01
5.47817349e-01 1.01080585e+00 2.38904923e-01 3.58613819e-01
-9.08736214e-02 7.28029311e-01 -5.19262314e-01 -2.82134205e-01
1.69225305e-01 -7.40559340e-01 -4.60435271e-01 1.44639418e-01
6.10050499e-01 -3.39016348e-01 -7.60096490e-01 1.21540594e+00
6.43981695e-01 1.01644203e-01 7.81822622e-01 -1.50668454e+00
-7.99209058e-01 -5.80842001e-03 -8.91000450e-01 1.99781731e-01
4.44313914e-01 2.45540664e-01 4.84845012e-01 -1.10096216e+00
6.95714355e-01 1.56636643e+00 8.07248592e-01 9.19426739e-01
-8.96462381e-01 -6.87239587e-01 -1.92338601e-01 4.50814426e-01
-1.48634267e+00 -7.12725282e-01 1.18931675e+00 -3.97645503e-01
4.75304574e-01 6.59695268e-02 5.34647465e-01 1.50046182e+00
3.31092715e-01 3.72314006e-01 9.62440252e-01 -2.57905155e-01
-4.15252894e-02 1.81799326e-02 -6.73068464e-01 6.25002742e-01
-3.96295905e-01 3.47768247e-01 -7.95542836e-01 1.30578965e-01
9.62991953e-01 1.23768404e-01 -5.58598757e-01 -1.80707470e-01
-7.61902750e-01 7.64359653e-01 4.39411730e-01 3.20130289e-01
-3.53353977e-01 2.76737958e-01 3.06341588e-01 1.12685814e-01
3.62257600e-01 7.26088509e-02 -1.27764165e-01 -2.54114605e-02
-1.22797561e+00 2.20028147e-01 2.58288115e-01 5.34650147e-01
5.84816873e-01 5.53151071e-01 -2.40286842e-01 8.71165931e-01
3.82332116e-01 3.44838858e-01 6.32703483e-01 -1.04874218e+00
2.47059718e-01 1.16971754e-01 -9.04732868e-02 -1.13640988e+00
-2.16221847e-02 -2.21831515e-01 -1.22728491e+00 3.46581370e-01
2.41039619e-01 -8.07579011e-02 -9.98571217e-01 1.96360493e+00
3.45163137e-01 6.42614424e-01 -9.42212436e-03 1.10166430e+00
8.34792972e-01 8.56829643e-01 8.29031970e-03 -5.59478939e-01
1.22124946e+00 -8.14518392e-01 -1.27206469e+00 -6.58192113e-02
-3.11219335e-01 -9.00220275e-01 1.05980349e+00 1.19925842e-01
-1.26969540e+00 -8.34522545e-01 -8.91085804e-01 -1.11797517e-02
-4.93924096e-02 -1.03550099e-01 4.26663041e-01 3.29150170e-01
-1.25622797e+00 4.81211692e-01 -5.58014512e-01 -2.59455442e-01
5.12014031e-01 2.65520781e-01 -5.82988739e-01 -1.34290665e-01
-1.34512293e+00 8.30035746e-01 1.30722046e-01 1.30743474e-01
-1.05125129e+00 -8.83371651e-01 -1.01585090e+00 5.96781448e-02
1.66110456e-01 -6.32365882e-01 8.79250467e-01 -1.35679328e+00
-1.58262885e+00 5.03399730e-01 -2.59394139e-01 -1.27237022e-01
6.47527039e-01 -4.80284132e-02 -6.45743728e-01 4.03114766e-01
-3.17322351e-02 1.03651512e+00 1.52927709e+00 -1.30446792e+00
-1.89887062e-01 -3.20427120e-01 -3.13108921e-01 1.72245681e-01
-3.64862978e-01 -1.06424429e-01 -3.41982603e-01 -1.16430593e+00
-2.01124653e-01 -7.73374438e-01 3.33530396e-01 3.60018492e-01
-8.49620625e-02 1.50035582e-02 1.41932178e+00 -8.98157597e-01
1.15238857e+00 -2.43507099e+00 1.51855260e-01 -3.14529181e-01
3.71500850e-02 2.43205249e-01 -1.73546061e-01 3.82742703e-01
-3.46414775e-01 -7.96452835e-02 -2.08128542e-01 -3.04003030e-01
-2.92895138e-01 1.39131084e-01 -4.45149302e-01 5.46427488e-01
4.28999037e-01 8.07266593e-01 -7.44520366e-01 -4.00598764e-01
2.14094728e-01 1.07868373e+00 -7.38844454e-01 2.71811545e-01
-9.57694724e-02 6.55996978e-01 -1.63951620e-01 6.49281621e-01
1.04107714e+00 5.18418178e-02 -2.96973377e-01 -4.95105714e-01
1.52608067e-01 -1.63002834e-01 -8.04948568e-01 1.89946234e+00
-4.27470356e-01 6.71707153e-01 2.19858021e-01 -9.18631494e-01
5.17848670e-01 7.11550951e-01 6.00095510e-01 -5.62927186e-01
-4.59304228e-02 5.85219525e-02 -5.87501377e-02 -8.70441973e-01
1.77254543e-01 -4.39400256e-01 4.51672405e-01 -1.70544624e-01
3.70729864e-01 -4.77536768e-02 -4.28030044e-01 -2.86644608e-01
6.79137886e-01 2.00067952e-01 6.27374500e-02 -1.35647550e-01
5.73765635e-01 -5.45876443e-01 5.11882961e-01 -1.18895300e-01
-1.33508503e-01 9.85032201e-01 3.23376924e-01 -1.91815913e-01
-1.06772649e+00 -1.10511208e+00 -1.40228614e-01 4.76555765e-01
8.45503882e-02 -1.97136492e-01 -1.01709592e+00 -4.77667302e-01
-1.31477907e-01 2.80292511e-01 -8.66877556e-01 -5.05726337e-01
-4.96593714e-01 -3.26439828e-01 5.38995504e-01 3.89439434e-01
8.37369442e-01 -1.19543481e+00 -2.61052608e-01 9.42502916e-02
-3.28628778e-01 -1.24454486e+00 -9.44496334e-01 -4.45915014e-01
-5.44584930e-01 -5.76373994e-01 -1.13728702e+00 -8.88802350e-01
4.16529298e-01 1.08902961e-01 6.06278300e-01 -3.58601332e-01
-6.36639059e-01 1.02721989e-01 -1.24608591e-01 -1.91956863e-01
-3.83323789e-01 -5.74525416e-01 1.11009926e-01 3.63666475e-01
1.01586029e-01 -8.13963711e-01 -8.77338946e-01 2.96275735e-01
-1.04962230e+00 -1.86625376e-01 3.71732086e-01 1.12347186e+00
3.13439488e-01 2.95945406e-01 7.08543599e-01 -2.46570528e-01
3.56200665e-01 -5.78659832e-01 -1.49980351e-01 -6.71041086e-02
-1.26911610e-01 -1.42937407e-01 6.23870850e-01 -6.69442475e-01
-1.03050160e+00 -5.45892604e-02 -2.86670834e-01 -1.03583944e+00
-1.01592213e-01 3.00389510e-02 -4.07909453e-01 -9.72416103e-02
4.15714324e-01 2.99835175e-01 4.14320081e-01 -2.64693677e-01
1.43807113e-01 4.00161117e-01 6.11001790e-01 -7.26361647e-02
7.46322930e-01 5.49576581e-01 1.11988902e-01 -1.26016176e+00
-5.60416102e-01 -1.28523827e-01 -1.28960565e-01 -3.72930676e-01
1.10111451e+00 -9.19199228e-01 -7.06736684e-01 5.03935575e-01
-1.30278945e+00 -6.64567053e-02 -2.89053649e-01 3.52891266e-01
-8.64143252e-01 4.41463172e-01 -6.35962307e-01 -6.74126923e-01
-1.74398676e-01 -1.26720631e+00 1.12739110e+00 2.57035792e-01
-8.45480114e-02 -8.43780816e-01 -1.65267885e-01 2.98504472e-01
5.79018354e-01 4.40372467e-01 8.65815461e-01 1.95491672e-01
-5.58752835e-01 2.66407251e-01 -5.84665202e-02 4.71022218e-01
3.34845304e-01 3.73559296e-02 -1.32860780e+00 -3.69209409e-01
2.76559621e-01 -2.59525418e-01 7.44380772e-01 6.62602961e-01
1.29906631e+00 -6.65254474e-01 -8.08335021e-02 7.51860559e-01
1.33265185e+00 2.17321441e-01 1.07450092e+00 -2.37957671e-01
6.40894055e-01 8.30133677e-01 4.65515524e-01 3.59927475e-01
-1.58076227e-01 1.06837273e+00 5.06379306e-01 -1.72101364e-01
-7.61362672e-01 -5.13312817e-01 7.42836893e-01 4.72727448e-01
-5.96141592e-02 -1.91493258e-01 -4.09269601e-01 5.35609365e-01
-1.54547071e+00 -1.32287240e+00 4.08767223e-01 1.97606015e+00
7.55990386e-01 -3.52013290e-01 4.00049658e-03 2.21886054e-01
7.84534514e-01 1.98240787e-01 -3.80164713e-01 -1.90013394e-01
-1.11662596e-01 2.88308859e-01 -2.63500493e-03 5.08710563e-01
-8.28392148e-01 8.69673014e-01 6.20168638e+00 1.26602983e+00
-1.57315290e+00 3.06838423e-01 5.65581858e-01 -9.68894511e-02
-2.53632516e-01 -4.66305196e-01 -5.48214853e-01 8.00937235e-01
8.44512045e-01 1.64937675e-01 5.55449545e-01 3.98967505e-01
4.40166593e-01 2.26925761e-01 -8.39121222e-01 1.39163482e+00
2.82128870e-01 -1.33315670e+00 3.24591309e-01 9.92418230e-02
8.36369812e-01 -4.66174066e-01 5.08252680e-01 -1.43804671e-02
-3.81886691e-01 -1.58683693e+00 6.01818681e-01 6.99377298e-01
1.35447216e+00 -1.00484407e+00 4.39084440e-01 1.20605998e-01
-1.10361481e+00 -9.95570645e-02 -3.49589229e-01 3.66089314e-01
1.53822094e-01 2.89065689e-01 -6.74159586e-01 3.26626718e-01
8.03979695e-01 5.85217357e-01 -6.74705580e-02 7.79863298e-01
1.45511823e-02 2.33933747e-01 -1.01231020e-02 5.64210832e-01
1.73628941e-01 -1.71155185e-02 7.23739386e-01 9.21438873e-01
7.65223563e-01 -7.43105859e-02 -1.46203175e-01 9.84505892e-01
-1.79993555e-01 -6.68003485e-02 -9.38327730e-01 -6.08140137e-03
2.28115395e-01 9.64676261e-01 -2.09508806e-01 7.54144713e-02
-3.89453113e-01 1.29341495e+00 -3.61027457e-02 4.65268821e-01
-9.91727352e-01 -2.35163104e-02 9.48640645e-01 6.38313174e-01
1.99600667e-01 7.25586712e-02 2.52344191e-01 -9.16296303e-01
9.11376253e-02 -9.16979432e-01 3.38558890e-02 -8.88993740e-01
-1.20486927e+00 6.91510260e-01 -1.66791543e-01 -1.34385145e+00
-4.72335100e-01 -2.96082795e-01 -5.46791553e-01 9.22641933e-01
-1.53975463e+00 -1.23447955e+00 -5.20280778e-01 1.06561422e+00
9.81986761e-01 -3.91800016e-01 7.30300009e-01 5.34422338e-01
-1.39132857e-01 9.33376431e-01 -2.52290249e-01 1.02363694e-02
9.00292397e-01 -6.78335488e-01 -1.11823231e-01 6.52767658e-01
-1.82571441e-01 3.69860739e-01 8.45942080e-01 -4.79860723e-01
-1.29055679e+00 -1.24926388e+00 5.59318900e-01 -4.67771627e-02
2.70855606e-01 -2.29995579e-01 -9.53215957e-01 2.76649028e-01
5.55623531e-01 3.82029921e-01 4.28293586e-01 -7.42415011e-01
-2.54352629e-01 -2.00654954e-01 -1.46717823e+00 5.60769081e-01
9.61285174e-01 -8.11046541e-01 -4.23836619e-01 1.26435056e-01
6.92849040e-01 -2.51142890e-03 -6.14981592e-01 5.90396404e-01
5.17008722e-01 -1.10260284e+00 1.08160472e+00 -3.24687272e-01
7.65174866e-01 -4.63253200e-01 -2.51183420e-01 -1.33960021e+00
-1.21212102e-01 -9.73385274e-01 -3.04227740e-01 1.34383309e+00
-3.20408233e-02 -3.77449274e-01 7.37470388e-01 -3.60875912e-02
-1.04546048e-01 -5.54267287e-01 -1.10423064e+00 -5.76860249e-01
-3.70816179e-02 1.88086435e-01 3.53458643e-01 1.06601644e+00
-1.50728881e-01 2.77886868e-01 -8.00245762e-01 1.34541184e-01
5.93122900e-01 -3.09018672e-01 4.10535216e-01 -9.56222236e-01
-3.42306614e-01 -2.32682928e-01 -6.57996893e-01 -8.15345645e-01
3.06895584e-01 -6.35573864e-01 -1.04601547e-01 -1.16072047e+00
7.29595637e-03 3.88005143e-03 -2.64296010e-02 2.22896218e-01
9.55055878e-02 5.70264518e-01 8.48765150e-02 6.65256456e-02
1.68469414e-01 1.01848662e+00 1.77680707e+00 -1.50823191e-01
-4.18314431e-03 -1.41984820e-01 -3.78064990e-01 7.79459476e-01
4.26231861e-01 -2.02619404e-01 -7.16944396e-01 -2.29425505e-01
-2.74250329e-01 7.27261901e-01 5.48808634e-01 -1.08536935e+00
1.02552608e-01 -5.96173257e-02 6.87249303e-01 -3.44508260e-01
9.26015198e-01 -9.74740088e-01 4.30653006e-01 3.25413346e-01
-1.64861053e-01 2.77347676e-02 2.80475140e-01 6.81555331e-01
-7.96757102e-01 1.55236110e-01 1.42180693e+00 -1.60622895e-01
-6.13605976e-01 5.66064119e-01 -1.85110733e-01 -1.86832324e-02
1.19632328e+00 -3.99257839e-01 -6.34970237e-03 -6.45840406e-01
-8.42358112e-01 -3.63691866e-01 4.27900851e-01 6.44460082e-01
7.67980754e-01 -1.49510050e+00 -6.57587171e-01 5.58542609e-01
-1.93261400e-01 -2.57548183e-01 7.55249381e-01 8.15906584e-01
-4.06879932e-01 2.00361222e-01 -7.07173586e-01 -5.42217672e-01
-1.12609363e+00 7.13658750e-01 4.80283022e-01 2.47992903e-01
-7.73269832e-01 7.98729658e-01 6.44145548e-01 2.77155906e-01
3.23116601e-01 6.62563443e-02 -3.91805559e-01 5.37201129e-02
6.35409713e-01 2.03247890e-01 -5.40264547e-02 -8.68600667e-01
-1.75853789e-01 8.07244837e-01 1.87284514e-01 -9.01286229e-02
1.17581463e+00 -1.73719272e-01 6.53384160e-03 1.52940497e-01
1.52232599e+00 1.31559312e-01 -1.61044812e+00 7.78102130e-02
-4.38866794e-01 -7.23727643e-01 3.51675190e-02 -5.60144305e-01
-1.40221167e+00 1.09572530e+00 8.91612709e-01 -1.34810984e-01
1.37409973e+00 -1.93187252e-01 8.00579786e-01 -4.81090546e-01
2.02089444e-01 -8.90533805e-01 5.51514566e-01 2.69651443e-01
1.29240334e+00 -1.05611742e+00 -4.52689737e-01 -6.48173451e-01
-6.41401947e-01 1.01859951e+00 6.25036657e-01 -1.42378733e-01
8.38410378e-01 3.39445293e-01 -1.25329152e-01 -3.57628576e-02
-5.90155959e-01 1.64152086e-01 3.65292847e-01 8.36626351e-01
3.96967292e-01 -3.32584381e-01 -1.06712669e-01 2.65282303e-01
-1.51937574e-01 -2.13457905e-02 4.03767496e-01 3.97970557e-01
-1.43768907e-01 -7.51452863e-01 -4.30814058e-01 4.65679914e-02
-7.49469280e-01 -8.27130023e-03 6.61065336e-03 7.87628651e-01
3.55732888e-01 8.35501134e-01 9.23375636e-02 -1.23403035e-01
1.34902015e-01 1.06671974e-01 6.68040276e-01 -2.86418676e-01
-1.24891564e-01 4.05063689e-01 -3.69844466e-01 -6.89120352e-01
-4.22225952e-01 -4.75263715e-01 -1.15468240e+00 -3.25343102e-01
5.81080019e-02 -5.23979813e-02 5.15847325e-01 6.22721851e-01
3.25587213e-01 6.84615612e-01 6.71826601e-01 -9.95144308e-01
-4.16723520e-01 -1.08574760e+00 -7.15498686e-01 5.06892562e-01
8.94665837e-01 -1.06728959e+00 -5.24137795e-01 1.76201150e-01] | [13.107150077819824, -0.31075215339660645] |
9b1020ff-3696-45c5-946a-f98dab312c5e | stochastic-multi-person-3d-motion-forecasting | 2306.05421 | null | https://arxiv.org/abs/2306.05421v1 | https://arxiv.org/pdf/2306.05421v1.pdf | Stochastic Multi-Person 3D Motion Forecasting | This paper aims to deal with the ignored real-world complexities in prior work on human motion forecasting, emphasizing the social properties of multi-person motion, the diversity of motion and social interactions, and the complexity of articulated motion. To this end, we introduce a novel task of stochastic multi-person 3D motion forecasting. We propose a dual-level generative modeling framework that separately models independent individual motion at the local level and social interactions at the global level. Notably, this dual-level modeling mechanism can be achieved within a shared generative model, through introducing learnable latent codes that represent intents of future motion and switching the codes' modes of operation at different levels. Our framework is general; we instantiate it with different generative models, including generative adversarial networks and diffusion models, and various multi-person forecasting models. Extensive experiments on CMU-Mocap, MuPoTS-3D, and SoMoF benchmarks show that our approach produces diverse and accurate multi-person predictions, significantly outperforming the state of the art. | ['Liang-Yan Gui', 'Yu-Xiong Wang', 'Sirui Xu'] | 2023-06-08 | null | null | null | null | ['motion-forecasting'] | ['computer-vision'] | [-2.79255271e-01 8.76841396e-02 4.69213650e-02 -1.70989960e-01
-3.53841037e-01 -4.53868926e-01 1.08447349e+00 -6.22462511e-01
4.75613438e-02 5.41007400e-01 9.37538445e-01 5.10545969e-02
1.90966100e-01 -8.04720283e-01 -5.90476573e-01 -8.10792625e-01
-9.39060450e-02 6.74573243e-01 3.25966328e-01 -3.66090328e-01
-2.58645564e-01 3.03644627e-01 -1.36203587e+00 3.46932709e-01
5.24254620e-01 3.82570237e-01 -1.40341505e-01 1.17117500e+00
2.09281608e-01 1.28324330e+00 -4.72166806e-01 -5.65813601e-01
1.95148624e-02 -5.93055069e-01 -7.09436834e-01 1.68231860e-01
2.48698741e-01 -4.93345529e-01 -8.35222304e-01 3.89189094e-01
6.71793818e-01 4.44287509e-01 1.00550187e+00 -1.56983292e+00
-9.14555669e-01 5.24302363e-01 -2.26557359e-01 -1.60210449e-02
5.62900901e-01 4.69749600e-01 6.47018731e-01 -6.22880220e-01
6.55699015e-01 1.77757919e+00 7.62282312e-01 1.05350280e+00
-9.54667807e-01 -2.28287950e-01 5.58883786e-01 8.81919477e-05
-1.11908746e+00 -3.49528104e-01 5.85456371e-01 -9.43167150e-01
8.95963550e-01 1.52611420e-01 7.40521848e-01 1.85498798e+00
4.57739532e-01 1.08952594e+00 4.92437661e-01 1.22080408e-01
2.05398008e-01 -5.26921630e-01 2.25777831e-02 7.26926923e-01
-6.77913129e-02 3.60221080e-02 -4.82342422e-01 -3.75166833e-01
8.49544764e-01 1.30432606e-01 6.22856766e-02 -2.85901964e-01
-1.45578074e+00 8.27516317e-01 1.45211771e-01 1.43210873e-01
-3.94048810e-01 6.42508268e-01 1.01640187e-01 -3.03137004e-01
5.89052081e-01 -4.27151710e-01 3.65354456e-02 -2.92172998e-01
-8.98731768e-01 8.63470793e-01 8.49458396e-01 1.00126708e+00
4.27857846e-01 2.46527299e-01 -5.15091181e-01 4.26068246e-01
5.22972703e-01 6.68908179e-01 5.46873927e-01 -1.03739369e+00
2.75279701e-01 1.75746843e-01 1.60674363e-01 -1.18570566e+00
-4.66346234e-01 -3.50118190e-01 -1.13001180e+00 -7.63588250e-02
2.93850124e-01 -3.55356961e-01 -1.03134465e+00 2.09087920e+00
3.80044341e-01 7.16434836e-01 5.32331988e-02 8.50925386e-01
9.65871930e-01 8.82976413e-01 3.80151302e-01 3.10870618e-01
9.67007101e-01 -1.55295861e+00 -4.89627898e-01 -3.19535047e-01
4.73591268e-01 -3.44790161e-01 5.81271768e-01 -1.62830755e-01
-1.14063978e+00 -8.74158740e-01 -6.39600873e-01 -1.08806245e-01
-1.20714299e-01 7.24883936e-03 7.93493807e-01 6.05990410e-01
-1.23734272e+00 4.64844018e-01 -1.28671265e+00 -3.54772508e-01
1.98431343e-01 6.33125135e-04 -1.01278745e-01 5.17534576e-02
-1.15313864e+00 6.17925763e-01 -7.06020296e-02 3.46636549e-02
-1.11927950e+00 -4.48234439e-01 -8.07625830e-01 -1.64088070e-01
-1.14916578e-01 -1.79494274e+00 1.06738782e+00 -4.57260340e-01
-1.51240277e+00 5.60218513e-01 -2.80780345e-01 -3.73488396e-01
1.02055550e+00 -4.34869975e-01 -4.26100522e-01 -1.74509600e-01
1.80953428e-01 8.81168962e-01 9.37407434e-01 -1.13323033e+00
-3.34840268e-01 -1.51726127e-01 -3.97007391e-02 3.46387923e-01
1.41600715e-02 -2.08541229e-01 -6.03706956e-01 -9.24744308e-01
-2.01299265e-01 -1.51595950e+00 -6.08884752e-01 -4.28658426e-01
-5.13452411e-01 -1.59231752e-01 6.29796088e-01 -5.22816718e-01
1.36616695e+00 -2.01007318e+00 7.28332162e-01 -9.68871936e-02
3.09316039e-01 -1.22391298e-01 -1.59803361e-01 5.01125634e-01
3.74775261e-01 6.71176910e-02 -5.63949756e-02 -1.00937867e+00
2.89915383e-01 3.18780243e-01 -2.94851482e-01 2.21831769e-01
-1.38876602e-01 1.43376470e+00 -9.89214540e-01 -2.23407209e-01
1.62562475e-01 7.35862434e-01 -8.19196820e-01 1.09430924e-01
-1.56037495e-01 8.65635157e-01 -6.52603030e-01 6.45104051e-01
2.57354707e-01 -5.74615717e-01 -2.48674378e-02 2.35593140e-01
2.97358632e-01 -1.12071700e-01 -1.12930310e+00 1.81764293e+00
-1.29597276e-01 3.49743694e-01 -4.62453902e-01 -3.00055236e-01
6.71483696e-01 1.61284119e-01 6.92210913e-01 4.61677685e-02
-1.34721369e-01 -1.85195386e-01 -3.25705379e-01 -4.26330060e-01
8.11459303e-01 7.12442845e-02 -4.75983977e-01 6.16763771e-01
3.51961739e-02 1.76735401e-01 -2.49413267e-01 2.38039479e-01
1.09381378e+00 4.86641437e-01 -8.86502713e-02 2.68816762e-02
2.54817188e-01 -2.68721521e-01 4.35774654e-01 9.75201666e-01
-3.26731890e-01 8.70546699e-01 1.12451591e-01 -7.04185784e-01
-9.22999799e-01 -1.17191041e+00 5.30817866e-01 1.07194781e+00
1.44066885e-01 -4.08665448e-01 -7.45476365e-01 -5.84134996e-01
1.68605477e-01 4.92057472e-01 -8.46418083e-01 -1.56108871e-01
-8.54445875e-01 -9.84424710e-01 6.36377454e-01 7.01382875e-01
2.38461152e-01 -9.15410101e-01 -5.28608859e-01 4.07086313e-01
-5.23187101e-01 -1.30766618e+00 -5.60102165e-01 -6.63019598e-01
-6.46279693e-01 -5.70991278e-01 -1.13036621e+00 -5.02337337e-01
3.16215456e-01 1.29953325e-01 1.26321137e+00 -1.49563238e-01
-9.58834216e-02 6.48206830e-01 -1.70979038e-01 -8.62886161e-02
-4.35719639e-01 2.28885755e-01 1.65037602e-01 2.22312123e-01
1.27226219e-01 -7.30918884e-01 -8.24063838e-01 3.25680077e-01
-7.40182340e-01 3.29479516e-01 2.07439795e-01 6.56413078e-01
2.17231438e-01 -2.56349623e-01 4.78907704e-01 -6.99268579e-01
6.48951173e-01 -9.45479631e-01 2.46659413e-01 1.99090853e-01
-4.10039686e-02 -1.05565459e-01 1.58373386e-01 -6.64236784e-01
-1.09500897e+00 8.01457614e-02 -2.20702648e-01 -6.01491511e-01
-3.71305138e-01 1.99124157e-01 -1.21497959e-01 1.63831636e-01
5.26245832e-01 4.05795246e-01 -2.24459186e-01 -1.94799244e-01
7.73194015e-01 1.46416053e-01 7.42004097e-01 -5.44729471e-01
8.53364348e-01 9.32520211e-01 5.87240458e-02 -6.18957281e-01
-5.10081947e-01 -1.69462711e-01 -7.00696945e-01 -6.45652771e-01
1.29461229e+00 -1.29192972e+00 -6.80380106e-01 1.06864727e+00
-1.33841145e+00 -5.65622270e-01 -1.34813100e-01 2.92423368e-01
-8.12040985e-01 3.87253463e-01 -9.69439149e-01 -8.64190817e-01
-2.02292800e-02 -1.12895286e+00 1.42933536e+00 8.54725316e-02
-6.42381251e-01 -1.39886415e+00 5.01070559e-01 4.17344302e-01
3.19247186e-01 7.65330553e-01 5.71081877e-01 -3.58980507e-01
-9.08100367e-01 -1.92320094e-01 5.21349072e-01 -1.16540238e-01
4.97439690e-02 7.44824484e-02 -8.27479005e-01 -7.72192627e-02
-3.79668176e-01 -1.38303116e-01 1.06106102e+00 6.91506922e-01
7.29779005e-01 -3.42582822e-01 -5.73248446e-01 6.59118772e-01
8.48285139e-01 -2.49955337e-02 8.16819847e-01 1.46666422e-01
1.16907811e+00 5.14895499e-01 1.24689370e-01 7.30239987e-01
1.09879994e+00 7.40805984e-01 2.33352527e-01 1.09834544e-01
-1.66734844e-01 -6.80959821e-01 5.33414602e-01 9.04113054e-01
-5.53226888e-01 -7.77037323e-01 -9.27601278e-01 5.13499975e-01
-2.51640439e+00 -1.52566206e+00 -2.20899075e-01 1.54139185e+00
8.10507536e-02 -4.90014255e-03 6.44597352e-01 -3.59048694e-01
5.70481956e-01 6.37302101e-01 -3.53516638e-01 3.50813218e-03
-2.85232455e-01 -4.11257207e-01 1.72967941e-01 6.29033387e-01
-1.26949930e+00 1.02226722e+00 7.34435701e+00 5.77325761e-01
-7.06807077e-01 1.86331570e-01 6.30589485e-01 -3.01613927e-01
-6.32817209e-01 -1.84250534e-01 -1.15834355e+00 7.00416148e-01
8.93485725e-01 2.86192577e-02 2.21110806e-01 5.96928298e-01
1.54599637e-01 3.75373840e-01 -1.12816370e+00 9.43646967e-01
2.62783080e-01 -1.41569304e+00 4.34128612e-01 2.15193048e-01
1.06297100e+00 2.57471241e-02 1.74260870e-01 3.37052286e-01
7.92636991e-01 -1.04943395e+00 1.15532494e+00 1.19622159e+00
1.10148542e-01 -6.20100319e-01 2.69800812e-01 6.82013094e-01
-1.30492699e+00 -2.35566683e-02 -7.77400881e-02 -1.20651193e-01
8.96192670e-01 3.05877000e-01 -1.49769858e-01 6.99187040e-01
6.11514211e-01 1.00948894e+00 -4.81611580e-01 5.02240598e-01
-8.54861736e-02 3.94760579e-01 -1.30518928e-01 -6.17674589e-02
3.95182550e-01 -9.94611830e-02 7.67157197e-01 1.35895705e+00
5.96479774e-01 9.10007954e-02 3.46978337e-01 8.42597246e-01
2.39033550e-01 -3.79907668e-01 -7.03275621e-01 8.78926069e-02
1.29451960e-01 8.80109072e-01 -5.29311895e-01 -4.73404348e-01
-3.73298168e-01 1.26099062e+00 1.46428049e-01 5.94390333e-01
-1.17019582e+00 5.85049391e-01 8.89620721e-01 1.71343297e-01
2.42133930e-01 -7.07536697e-01 -1.05839055e-02 -1.51413560e+00
-2.43780792e-01 -5.61808228e-01 3.43884647e-01 -7.28827775e-01
-1.33391547e+00 7.05262899e-01 4.59694602e-02 -1.20449078e+00
-9.27918971e-01 -2.87789375e-01 -7.00386643e-01 7.14026034e-01
-9.81527448e-01 -1.65672100e+00 -3.71474773e-01 7.05142438e-01
4.96003211e-01 -2.37393618e-01 7.26904929e-01 3.09808075e-01
-3.94069970e-01 4.07234699e-01 -1.18013345e-01 1.20981820e-01
4.39821303e-01 -9.28550959e-01 1.24288487e+00 8.36734831e-01
1.42366141e-01 3.78196388e-01 6.51034594e-01 -9.04445112e-01
-1.15741265e+00 -1.28878045e+00 1.10123086e+00 -1.11305809e+00
5.46978533e-01 -4.63843048e-01 -6.36026621e-01 1.01056278e+00
-5.71809970e-02 -3.21790241e-02 8.23993802e-01 -1.53925642e-01
-1.11535773e-01 6.40082300e-01 -7.06006587e-01 9.56021845e-01
1.65979660e+00 -4.49053466e-01 -3.05635244e-01 2.30553895e-01
6.90742910e-01 -5.33928216e-01 -6.90594435e-01 3.91323447e-01
8.00287247e-01 -1.03781939e+00 1.50649178e+00 -9.36173201e-01
7.38799453e-01 -1.73607051e-01 -2.06879288e-01 -1.18173003e+00
-8.51109445e-01 -7.23019063e-01 -7.29600072e-01 1.00025439e+00
1.34234488e-01 -2.62222707e-01 9.90723252e-01 8.78610909e-01
-1.92327663e-01 -6.12330437e-01 -7.78337359e-01 -6.88304901e-01
2.73248225e-01 -3.96694332e-01 7.42580950e-01 8.70890021e-01
-4.93225455e-01 1.62512198e-01 -1.20722532e+00 9.89779904e-02
7.27841437e-01 2.25116196e-03 1.27246475e+00 -9.96785164e-01
-8.80614936e-01 -4.18419391e-01 -6.19788408e-01 -1.53188896e+00
4.10471112e-01 -7.22075224e-01 -1.53523147e-01 -1.62685716e+00
2.42030755e-01 -3.06279324e-02 1.33472994e-01 2.67073780e-01
-2.92332828e-01 7.45447725e-02 5.10890245e-01 5.39493501e-01
-7.53928423e-01 8.71980786e-01 1.36774564e+00 -2.16951355e-01
-2.97698677e-01 1.54103845e-01 -4.31345493e-01 9.51360643e-01
3.39103401e-01 -2.70687550e-01 -5.10288477e-01 -7.05643892e-01
5.07767983e-02 3.12633455e-01 9.17490423e-01 -1.31492257e+00
5.30315042e-01 -3.19670677e-01 3.96388739e-01 -6.33922279e-01
6.36202872e-01 -2.23283991e-01 8.21136475e-01 4.74202603e-01
-2.51524776e-01 1.67335480e-01 -9.28798094e-02 9.85981882e-01
4.60707210e-02 5.10318995e-01 3.17302406e-01 -2.76201338e-01
-8.21412802e-01 6.59012794e-01 -5.64298213e-01 3.52402888e-02
9.88337874e-01 -2.86444545e-01 -2.46372953e-01 -8.91378760e-01
-1.36517656e+00 1.76517636e-01 4.77248013e-01 8.98019195e-01
4.89311069e-01 -1.73196161e+00 -8.65868032e-01 -1.99625408e-03
1.71554778e-02 -1.59894139e-01 7.57348299e-01 5.15151262e-01
-3.13912570e-01 1.68927774e-01 -1.51944265e-01 -7.94494092e-01
-8.79990995e-01 2.55499959e-01 3.62031460e-01 -3.85548651e-01
-6.35885656e-01 8.24987054e-01 4.25434232e-01 -4.56523687e-01
-2.89579034e-02 -1.42663851e-01 -1.35659859e-01 -2.12590545e-01
3.29853714e-01 6.83117807e-01 -5.73711276e-01 -1.25336838e+00
-2.89118588e-01 6.46008134e-01 4.46495563e-01 -2.23770887e-01
1.15135312e+00 -5.05908608e-01 3.39083016e-01 6.79927289e-01
9.19758737e-01 -1.23800240e-01 -1.54303527e+00 -5.19652739e-02
-2.79863238e-01 -2.19323874e-01 -5.16589105e-01 -4.14122373e-01
-8.45781386e-01 7.89600253e-01 1.46584213e-01 2.33985651e-02
5.89309752e-01 1.81590259e-01 1.07184017e+00 6.96467012e-02
6.28260553e-01 -7.04849958e-01 2.27611363e-01 6.93254232e-01
6.91634417e-01 -9.15933490e-01 -3.55122924e-01 -3.22261512e-01
-8.60281825e-01 6.91733241e-01 4.88776058e-01 -2.43030936e-01
8.91915858e-01 1.54087543e-01 -6.95910901e-02 -8.76131002e-03
-9.22413826e-01 -4.55338024e-02 5.62407255e-01 5.51386237e-01
2.42558509e-01 3.14044118e-01 2.06396282e-01 8.87437701e-01
-2.50826418e-01 1.04842111e-01 2.02057615e-01 8.25678885e-01
-1.44785300e-01 -9.93273795e-01 -2.62204975e-01 -1.30064100e-01
-2.80444741e-01 1.53692380e-01 -3.20516139e-01 5.89758635e-01
1.91927016e-01 7.26039290e-01 1.02189265e-01 -7.39782572e-01
1.07559629e-01 1.47188842e-01 2.69134074e-01 -2.69356221e-01
-3.93820971e-01 1.38274893e-01 -4.43136431e-02 -6.78771377e-01
-7.74332583e-01 -9.30895746e-01 -9.08241153e-01 -7.21047521e-01
4.81418341e-01 -2.66334355e-01 1.95859045e-01 9.42127526e-01
6.21872604e-01 6.49947226e-01 3.08754921e-01 -1.24086523e+00
-3.53776991e-01 -7.03584433e-01 -3.29272211e-01 7.42904067e-01
3.64256352e-01 -6.75719678e-01 -3.18561912e-01 4.45120156e-01] | [7.235209941864014, -0.08526485413312912] |
38eeb920-4f6e-49f6-be2f-08b9bf45f833 | camembert-bio-a-tasty-french-language-model | 2306.15550 | null | https://arxiv.org/abs/2306.15550v1 | https://arxiv.org/pdf/2306.15550v1.pdf | CamemBERT-bio: a Tasty French Language Model Better for your Health | Clinical data in hospitals are increasingly accessible for research through clinical data warehouses, however these documents are unstructured. It is therefore necessary to extract information from medical reports to conduct clinical studies. Transfer learning with BERT-like models such as CamemBERT has allowed major advances, especially for named entity recognition. However, these models are trained for plain language and are less efficient on biomedical data. This is why we propose a new French public biomedical dataset on which we have continued the pre-training of CamemBERT. Thus, we introduce a first version of CamemBERT-bio, a specialized public model for the French biomedical domain that shows 2.54 points of F1 score improvement on average on different biomedical named entity recognition tasks. | ['Eric de la Clergerie', 'Laurent Romary', 'Rian Touchent'] | 2023-06-27 | null | null | null | null | ['transfer-learning'] | ['miscellaneous'] | [-3.71689588e-01 3.31659973e-01 -4.32875872e-01 -5.57369232e-01
-1.02008247e+00 -2.33030066e-01 2.27722630e-01 8.43202889e-01
-8.76980305e-01 1.34196115e+00 1.99230447e-01 -4.27736342e-01
1.02411546e-01 -8.36205423e-01 -7.53031909e-01 -4.38097805e-01
-4.51933555e-02 8.53528082e-01 -9.07235146e-02 3.03090475e-02
-1.78329378e-01 5.61818004e-01 -5.94479322e-01 8.40223312e-01
7.58868873e-01 5.88676691e-01 -1.78420827e-01 5.34592688e-01
-2.71631539e-01 8.64178658e-01 -6.84732735e-01 -5.80651760e-01
-9.48318020e-02 -3.60436976e-01 -7.97112048e-01 -2.42311120e-01
-1.46728054e-01 -3.51691246e-02 -4.55289669e-02 6.83736503e-01
6.73133969e-01 -3.14816415e-01 6.49335146e-01 -9.57965076e-01
-7.55901396e-01 5.90041339e-01 -1.31411552e-01 -1.55641004e-01
2.56538868e-01 -1.37854800e-01 6.13935888e-01 -8.51477027e-01
1.19714236e+00 5.69433510e-01 8.66489589e-01 7.48858988e-01
-9.79495943e-01 -6.20827079e-01 -4.13058519e-01 1.77773640e-01
-1.38670421e+00 -4.51868862e-01 1.76185638e-01 -4.69677061e-01
1.04736042e+00 1.31631210e-01 4.90774393e-01 1.20145524e+00
7.15231180e-01 8.72784972e-01 1.03313541e+00 -2.88185745e-01
3.37457746e-01 5.73747516e-01 1.22839026e-01 7.37333775e-01
5.14648855e-01 -2.37335548e-01 -4.75962013e-01 -4.55987900e-01
4.00997907e-01 2.02043608e-01 -2.59449959e-01 -2.73968905e-01
-1.54479694e+00 9.34895933e-01 2.35860080e-01 6.85985208e-01
-6.24221265e-01 -3.20849061e-01 4.48544443e-01 4.72011864e-01
4.78751361e-01 6.14411414e-01 -1.04288447e+00 -1.00920670e-01
-8.74209046e-01 2.24488638e-02 1.14100361e+00 1.27008820e+00
2.97654510e-01 -6.35727406e-01 -1.39166847e-01 7.08626688e-01
2.83205032e-01 3.32148463e-01 7.09498405e-01 -2.31661871e-01
5.17544925e-01 7.66363263e-01 7.22050145e-02 -6.97870672e-01
-7.13144779e-01 -3.97817641e-01 -1.13107359e+00 -5.60555041e-01
3.02984715e-01 -4.81496066e-01 -1.04394615e+00 1.23232412e+00
3.16978991e-01 -7.12053701e-02 4.98225093e-01 3.67256433e-01
1.07722032e+00 5.99081039e-01 5.61822653e-01 -2.46342927e-01
1.57812750e+00 -6.99391782e-01 -1.12475646e+00 8.91014859e-02
1.18295062e+00 -5.82677186e-01 3.35867792e-01 3.83644074e-01
-8.01773071e-01 -1.21953547e-01 -7.18165398e-01 -5.21822907e-02
-1.00577068e+00 4.13924932e-01 5.76626003e-01 7.68053174e-01
-8.28935325e-01 5.02471328e-01 -1.11231935e+00 -5.21380961e-01
8.42249751e-01 4.27050620e-01 -9.52565908e-01 -2.05178410e-01
-1.18037689e+00 9.77350116e-01 5.11293113e-01 1.56250205e-02
-5.09642601e-01 -9.41411436e-01 -7.48811960e-01 -9.24841464e-02
1.74327508e-01 -8.40110302e-01 1.01546812e+00 -1.46284238e-01
-1.15639019e+00 1.05215704e+00 -2.58419067e-02 -4.50307012e-01
5.74344158e-01 -8.52961317e-02 -7.91449308e-01 3.92644405e-02
2.33202562e-01 4.37049717e-01 7.38345366e-03 -9.13728893e-01
-5.52642465e-01 -5.22431850e-01 -6.67608738e-01 -6.33949101e-01
-3.24632853e-01 2.03849748e-01 -3.41847062e-01 -6.39163673e-01
-3.03612590e-01 -1.01589012e+00 -5.17002881e-01 -2.50180930e-01
-3.79728705e-01 -1.87306076e-01 3.22745949e-01 -9.16646957e-01
1.13425446e+00 -1.88442266e+00 -1.93013698e-01 6.56127110e-02
3.61519694e-01 4.88019973e-01 -9.82356668e-02 4.84276801e-01
-4.01386350e-01 4.42616373e-01 -2.58517623e-01 4.89715897e-02
-3.50071043e-01 -6.04732409e-02 2.21186787e-01 3.85568112e-01
5.87394476e-01 1.11502945e+00 -8.33457589e-01 -7.83416450e-01
-2.79154897e-01 4.70848143e-01 -6.38636589e-01 3.94620419e-01
1.23680402e-02 6.67750955e-01 -9.23318148e-01 7.14560866e-01
5.97508132e-01 -5.68639576e-01 2.56820768e-01 -2.52410710e-01
4.27161790e-02 8.31709877e-02 -5.84720969e-01 1.85314572e+00
-3.15725952e-01 1.31351128e-01 -7.13091195e-02 -9.21723008e-01
8.88486683e-01 9.24893737e-01 1.08844340e+00 -2.72747844e-01
1.86870724e-01 3.47421378e-01 -1.54848501e-01 -8.10176015e-01
4.49890122e-02 -5.12630403e-01 -6.12715706e-02 9.69459265e-02
2.19041005e-01 1.01859003e-01 3.50037366e-01 3.41005879e-03
1.61500943e+00 -5.10786520e-03 7.56360650e-01 -7.98353702e-02
5.66357791e-01 4.52564478e-01 8.21007371e-01 2.74409294e-01
-3.20265621e-01 3.76615644e-01 4.40530211e-01 -5.88498175e-01
-6.74672127e-01 -8.33806872e-01 -5.12949109e-01 6.05000079e-01
-7.91754007e-01 -6.20442510e-01 -5.98991454e-01 -1.24064541e+00
1.13818787e-01 7.11005569e-01 -6.54391110e-01 -1.61098205e-02
-4.91354913e-01 -1.04760325e+00 6.99378729e-01 4.21428561e-01
2.24369705e-01 -1.19306850e+00 -4.99874443e-01 5.72080314e-01
-1.29183963e-01 -1.28031588e+00 -2.35884145e-01 3.62726837e-01
-1.11913979e+00 -1.24304497e+00 -1.15543914e+00 -7.62161553e-01
6.16109967e-01 -6.87883019e-01 1.27340806e+00 -4.72278625e-01
-4.72555101e-01 1.77896112e-01 -4.64873642e-01 -1.03238487e+00
-7.70906031e-01 5.36626399e-01 -2.76090771e-01 -1.63639665e-01
8.09666574e-01 -1.03202257e-02 -4.41385657e-01 1.29795283e-01
-8.87079120e-01 -1.28541186e-01 1.12159848e+00 9.27440524e-01
5.62265694e-01 -5.41438520e-01 1.21883929e+00 -1.46892869e+00
5.48857808e-01 -7.78063118e-01 -2.13619485e-01 5.99603474e-01
-6.40148938e-01 4.98863235e-02 5.07311881e-01 -4.23492566e-02
-1.00166357e+00 1.97338641e-01 -4.80236351e-01 -1.21983550e-01
-5.16477883e-01 9.06231940e-01 -7.75090978e-02 4.82134819e-01
5.88958025e-01 -1.65966645e-01 -1.34799510e-01 -5.63023269e-01
-5.40780909e-02 1.18719614e+00 1.50830681e-02 -1.96168050e-01
1.92089424e-01 1.05461791e-01 -2.23782212e-01 -7.79261410e-01
-6.75219238e-01 -6.33755326e-01 -7.66250610e-01 4.19819504e-01
1.24410868e+00 -1.01507831e+00 -6.96257949e-01 7.46856108e-02
-1.21745741e+00 1.22902943e-02 -2.77694702e-01 8.82677436e-01
-4.44968075e-01 -7.76302963e-02 -9.20623362e-01 -3.15189391e-01
-6.24313891e-01 -1.02619493e+00 9.65579510e-01 -1.20220654e-01
-2.87316352e-01 -1.03284609e+00 4.41016436e-01 3.18671137e-01
4.83392566e-01 3.55148226e-01 1.19192803e+00 -1.28172278e+00
-2.46879444e-01 -5.36087573e-01 -1.54826954e-01 1.59343526e-01
5.25039375e-01 -3.78057092e-01 -7.98423767e-01 -2.95790941e-01
2.55069765e-03 -2.15544224e-01 6.99168921e-01 1.48174673e-01
1.09988117e+00 -7.59410858e-02 -7.82388389e-01 3.54098737e-01
1.29022467e+00 5.21492064e-01 6.17759645e-01 3.34979475e-01
4.11715746e-01 5.72606087e-01 5.48036277e-01 3.22407484e-01
3.74661565e-01 1.27347887e-01 -2.81454086e-01 -5.09849489e-01
2.34468833e-01 -1.37274656e-02 4.55963612e-02 1.08056200e+00
-2.30166659e-01 -1.10559486e-01 -1.22173226e+00 7.31496930e-01
-1.58152115e+00 -4.43782985e-01 -9.06448141e-02 1.94546711e+00
1.48029017e+00 -3.03312689e-01 -1.83423102e-01 -4.93774682e-01
4.56169695e-01 -5.39610684e-01 -3.66060317e-01 -2.31761128e-01
1.78971849e-02 4.70802158e-01 5.43837965e-01 -3.06773245e-01
-1.18386424e+00 7.54997194e-01 6.58404255e+00 4.11799222e-01
-9.75528419e-01 2.90517062e-01 8.98683012e-01 1.64402634e-01
2.57244200e-01 -5.29879212e-01 -8.08939934e-01 2.10596800e-01
1.71552038e+00 -2.59298623e-01 -6.29300356e-01 9.02024269e-01
1.47489712e-01 2.05671057e-01 -1.30350065e+00 7.53026664e-01
-3.17340828e-02 -1.50399411e+00 -5.50779477e-02 2.24217787e-01
5.59757710e-01 3.03409576e-01 -3.40918779e-01 4.54894394e-01
1.94502518e-01 -1.03166807e+00 -2.08603367e-01 6.70755684e-01
8.88512433e-01 -5.78976989e-01 1.34468365e+00 1.46817014e-01
-4.96543527e-01 2.91589797e-01 -4.31878209e-01 9.31814909e-01
1.37738332e-01 8.93990874e-01 -1.63536894e+00 8.28307748e-01
5.94271064e-01 7.53735602e-01 -6.43154860e-01 1.14539862e+00
2.10229591e-01 6.58264458e-01 9.74818990e-02 -7.30415899e-03
4.56120037e-02 3.38262618e-02 1.39318099e-02 1.53204346e+00
5.78146636e-01 7.40078613e-02 9.30608213e-02 6.14127934e-01
-4.66958642e-01 7.79832304e-01 -7.39530027e-01 -5.04514098e-01
-1.77415475e-01 1.21796787e+00 -6.15924060e-01 -5.89785933e-01
-6.40058577e-01 7.45223820e-01 1.64387807e-01 1.00586712e-01
-8.44135106e-01 -5.55375218e-01 2.36073658e-01 9.82317924e-02
2.03358293e-01 9.41505358e-02 1.24700494e-01 -1.38467801e+00
-5.06399214e-01 -1.15423298e+00 6.29048645e-01 -3.81903678e-01
-1.54247367e+00 9.34094489e-01 -1.29267916e-01 -1.15518117e+00
-2.85574734e-01 -9.17252183e-01 2.08098024e-01 6.23642683e-01
-1.49094951e+00 -1.22224581e+00 1.92379177e-01 6.36637330e-01
2.47080952e-01 -4.45672899e-01 1.43076456e+00 6.38795197e-01
-5.85571766e-01 5.21823823e-01 2.50886351e-01 6.62365139e-01
1.41393304e+00 -1.15437627e+00 6.77425116e-02 -5.98852662e-03
4.82380986e-02 9.29311454e-01 2.39210069e-01 -8.69047642e-01
-1.27138329e+00 -1.47037864e+00 1.50507820e+00 -6.21697068e-01
6.03192151e-01 -1.28931835e-01 -9.73537087e-01 1.00224090e+00
3.30657452e-01 1.01077845e-02 1.54676390e+00 1.15542844e-01
-1.59786791e-01 -1.08371191e-02 -1.49964118e+00 2.00709045e-01
4.96771365e-01 -1.06496863e-01 -7.97064900e-01 6.84734344e-01
5.44513702e-01 -8.90249982e-02 -1.74252212e+00 4.43533599e-01
2.42644802e-01 -4.59919006e-01 6.69961631e-01 -1.25250232e+00
4.17633027e-01 -2.91916728e-02 1.16111472e-01 -1.37275112e+00
-2.71093529e-02 -3.53500575e-01 1.84431657e-01 1.08200276e+00
1.03437734e+00 -8.45782042e-01 8.29115152e-01 7.01788843e-01
1.02625713e-01 -6.54457688e-01 -8.51005256e-01 -4.60687846e-01
3.01259905e-01 -9.40239280e-02 5.97481847e-01 1.41151941e+00
2.84858018e-01 4.79430169e-01 1.43157467e-01 -2.07635731e-01
2.38565847e-01 4.77474518e-02 3.93122256e-01 -1.26080632e+00
6.04058206e-02 1.58862337e-01 -3.68404001e-01 -1.71160966e-01
3.62533107e-02 -1.11131001e+00 -7.42059201e-04 -1.68609965e+00
3.71666729e-01 -6.35019183e-01 -6.26122296e-01 9.67976034e-01
-9.93365198e-02 5.02707250e-02 -1.63048074e-01 1.86326914e-02
-4.75001067e-01 2.60652721e-01 1.02120292e+00 -3.28769505e-01
-3.09214920e-01 5.69241978e-02 -4.98198062e-01 4.81045097e-01
8.92420232e-01 -8.56287718e-01 2.67485678e-01 -1.53357089e-01
1.91723809e-01 2.39161223e-01 -2.30521813e-01 -9.01812732e-01
2.86815196e-01 -9.26261023e-02 6.13741696e-01 -6.76675737e-01
-1.50994673e-01 -9.36889768e-01 2.45986372e-01 6.23458326e-01
-3.68994892e-01 2.69773304e-01 4.14147496e-01 4.63024110e-01
-4.30582851e-01 -8.12113136e-02 5.14387786e-01 -2.53095895e-01
-2.84542829e-01 3.78957897e-01 -5.43148816e-01 -1.67352661e-01
1.20144081e+00 4.07925397e-01 -2.40834579e-01 2.43448794e-01
-1.19902718e+00 2.79846098e-02 3.48225757e-02 2.49882311e-01
5.21797657e-01 -1.03626204e+00 -1.21005476e+00 1.58324465e-01
5.74510217e-01 -1.74268261e-01 1.61410034e-01 1.01719630e+00
-6.46749675e-01 1.01626682e+00 -3.49439085e-01 -3.96332473e-01
-1.18217599e+00 8.17421436e-01 1.65946372e-02 -8.13649416e-01
-4.38285112e-01 4.98237461e-01 -6.31861240e-02 -8.48792672e-01
-4.09871459e-01 -5.67670763e-01 -3.57128710e-01 2.89079696e-01
4.95149672e-01 1.40836790e-01 7.14655042e-01 -4.25946444e-01
-5.94186306e-01 1.10743307e-01 -3.02110106e-01 1.59135729e-01
1.76329279e+00 3.50197405e-01 -3.97698849e-01 4.43340480e-01
1.43961489e+00 3.05200905e-01 -1.42587036e-01 1.50423586e-01
5.79532743e-01 3.60921733e-02 -5.15049770e-02 -1.19608068e+00
-1.17354321e+00 5.87257206e-01 5.91919959e-01 -1.63348451e-01
9.96203601e-01 1.34180516e-01 6.39646232e-01 8.44691038e-01
4.62997645e-01 -7.58172810e-01 -5.04042387e-01 2.98074782e-01
7.70969689e-01 -1.29866147e+00 -3.50351222e-02 -3.59535336e-01
-6.59697652e-01 9.85187113e-01 1.93491042e-01 3.80344480e-01
1.07007325e+00 5.73662817e-01 3.39438409e-01 -3.64887565e-01
-7.88575709e-01 1.05412856e-01 2.10290700e-01 5.90370774e-01
1.05348170e+00 6.57946244e-02 -6.31554961e-01 9.99390244e-01
1.24402367e-01 8.48872364e-01 2.97085494e-01 1.29222262e+00
1.49177179e-01 -1.49043012e+00 -3.19726378e-01 6.59711063e-01
-1.19002819e+00 -2.66323984e-01 -3.64697665e-01 9.73473370e-01
9.23875421e-02 8.22492719e-01 -4.02238607e-01 3.17248702e-02
6.55828178e-01 5.01298547e-01 1.56823561e-01 -8.47525895e-01
-9.09573019e-01 8.30064993e-03 2.21009180e-01 -4.34063852e-01
-5.33076584e-01 -3.58414203e-01 -1.34214485e+00 2.11485729e-01
-2.76220471e-01 7.48087883e-01 9.01833177e-01 5.36461651e-01
9.77002919e-01 7.09415257e-01 1.02552712e-01 5.04220426e-02
-2.42004648e-01 -1.11735797e+00 -4.87733185e-01 2.07593873e-01
4.14272323e-02 -2.18880162e-01 1.19620197e-01 4.95487243e-01] | [8.465446472167969, 8.73388957977295] |
e1069af5-131d-4be6-b840-cfe17923d187 | fine-grained-visual-entailment | 2203.15704 | null | https://arxiv.org/abs/2203.15704v1 | https://arxiv.org/pdf/2203.15704v1.pdf | Fine-Grained Visual Entailment | Visual entailment is a recently proposed multimodal reasoning task where the goal is to predict the logical relationship of a piece of text to an image. In this paper, we propose an extension of this task, where the goal is to predict the logical relationship of fine-grained knowledge elements within a piece of text to an image. Unlike prior work, our method is inherently explainable and makes logical predictions at different levels of granularity. Because we lack fine-grained labels to train our method, we propose a novel multi-instance learning approach which learns a fine-grained labeling using only sample-level supervision. We also impose novel semantic structural constraints which ensure that fine-grained predictions are internally semantically consistent. We evaluate our method on a new dataset of manually annotated knowledge elements and show that our method achieves 68.18\% accuracy at this challenging task while significantly outperforming several strong baselines. Finally, we present extensive qualitative results illustrating our method's predictions and the visual evidence our method relied on. Our code and annotated dataset can be found here: https://github.com/SkrighYZ/FGVE. | ['Shih-Fu Chang', 'YiPeng Zhang', 'Christopher Thomas'] | 2022-03-29 | null | null | null | null | ['visual-entailment'] | ['reasoning'] | [ 2.77751476e-01 5.43397605e-01 -5.00690401e-01 -7.18150973e-01
-9.75832999e-01 -6.62520945e-01 8.51496398e-01 1.06285304e-01
4.45271693e-02 6.58942759e-01 4.32080448e-01 -2.51275897e-01
3.90843004e-02 -5.29690325e-01 -1.30286288e+00 -2.71033328e-02
5.03566504e-01 5.95718920e-01 2.56424993e-01 2.50332747e-02
2.52698749e-01 2.25449622e-01 -1.54043317e+00 1.16734898e+00
4.95920241e-01 1.10975993e+00 -6.24389574e-02 6.10046923e-01
-4.66318801e-02 1.29910707e+00 -1.86116472e-01 -8.64647329e-01
7.59573132e-02 -1.71780378e-01 -1.42086661e+00 4.31551039e-01
1.21849930e+00 -2.59114444e-01 -9.28323567e-02 8.83113742e-01
-1.34365022e-01 -5.37785925e-02 7.85247028e-01 -1.54241443e+00
-8.39934886e-01 5.15067041e-01 -6.20897532e-01 -1.84690043e-01
5.52212119e-01 -1.70689058e-02 1.48511267e+00 -1.13639700e+00
8.86235714e-01 1.34196365e+00 6.07848525e-01 4.87514436e-01
-1.24061763e+00 -4.92495567e-01 3.92546982e-01 4.45126802e-01
-1.32038796e+00 -6.78899944e-01 5.63524187e-01 -5.57624102e-01
1.03923154e+00 1.40104577e-01 3.48892123e-01 1.10289681e+00
-4.24242690e-02 8.73324454e-01 1.20063019e+00 -5.63164413e-01
7.09383842e-03 -5.77690266e-03 1.64941519e-01 1.28829873e+00
5.56359626e-02 -3.88658106e-01 -8.56771827e-01 6.65098801e-02
6.50448740e-01 -1.50371775e-01 -1.03376620e-01 -4.42395717e-01
-1.36780095e+00 5.58625698e-01 5.94508708e-01 -2.47784052e-02
-4.96916994e-02 5.38026989e-01 1.48181856e-01 -1.01331331e-01
3.22399855e-01 2.36587971e-01 -4.18068290e-01 1.43667132e-01
-6.86075389e-01 1.15587957e-01 8.17614734e-01 1.10776019e+00
8.22115004e-01 -4.37558711e-01 -2.66836882e-01 5.73433995e-01
4.63022053e-01 2.93996036e-01 -5.64268604e-02 -1.41615260e+00
6.20351613e-01 7.89374650e-01 2.28138432e-01 -9.34307933e-01
-2.35390201e-01 1.38066756e-03 -4.15698946e-01 2.75362879e-01
5.42676210e-01 2.14701310e-01 -8.50371480e-01 1.44136500e+00
2.03955516e-01 7.14803264e-02 -1.38347745e-01 9.99085188e-01
9.24476624e-01 4.19240057e-01 3.32191020e-01 1.76018462e-01
1.49125516e+00 -1.24799716e+00 -5.76266110e-01 -3.53030205e-01
4.12340373e-01 -7.56921291e-01 1.40090144e+00 3.18707973e-01
-1.16779280e+00 -3.89038205e-01 -8.48138630e-01 -5.96813381e-01
-4.01442766e-01 2.96502948e-01 6.16961241e-01 1.13431029e-01
-1.13142276e+00 2.87648946e-01 -6.43257260e-01 -4.68353003e-01
7.20956922e-01 1.22761115e-01 -5.97922742e-01 -2.47322470e-01
-7.56111741e-01 9.78473008e-01 4.10043269e-01 -1.59328982e-01
-8.39942992e-01 -4.87942398e-01 -9.30114269e-01 5.19474111e-02
6.04635179e-01 -9.77774024e-01 1.47933793e+00 -9.86623228e-01
-9.13038075e-01 1.36253870e+00 -6.00901008e-01 -2.55555183e-01
4.93539065e-01 -1.57571763e-01 -9.83913541e-02 3.26114386e-01
3.61964792e-01 1.15033436e+00 8.07865560e-01 -1.77696621e+00
-6.96810782e-01 -1.46667063e-01 4.88732249e-01 2.42967770e-01
-5.34191579e-02 -1.44910291e-01 -7.55738616e-01 -5.92669666e-01
-1.11788601e-01 -7.34723687e-01 2.67800361e-01 3.13163102e-01
-6.85018420e-01 -4.40108985e-01 7.79976785e-01 -5.54666579e-01
9.16476011e-01 -1.92236507e+00 1.06053501e-01 4.88866940e-02
3.75983179e-01 -3.70397508e-01 -1.19014420e-01 3.94244403e-01
3.17193568e-02 3.11273783e-01 -3.24002653e-01 -6.24373913e-01
5.11179447e-01 2.76107907e-01 -5.97357988e-01 1.43865675e-01
3.38301420e-01 1.21493483e+00 -7.72193074e-01 -8.41329932e-01
1.20918028e-01 4.01875705e-01 -3.30298871e-01 1.01650812e-01
-8.51384521e-01 2.45390266e-01 -2.36919224e-01 9.86974239e-01
2.61962682e-01 -9.30801749e-01 2.87649423e-01 -7.78863966e-01
2.83449709e-01 3.34413513e-03 -8.21438611e-01 1.75526810e+00
-2.54501045e-01 6.17700577e-01 -1.91142693e-01 -7.87941694e-01
5.59278667e-01 1.93729520e-01 2.72090230e-02 -6.02522433e-01
-4.10681665e-02 -1.22978315e-01 -5.34812391e-01 -6.51149690e-01
4.06345367e-01 -3.56799424e-01 -2.77571857e-01 6.56597853e-01
-1.10936478e-01 -1.98273703e-01 -1.64397545e-02 5.86079299e-01
1.08135223e+00 6.63929880e-01 4.70077902e-01 5.20197637e-02
3.09973806e-01 4.02849197e-01 3.82051051e-01 7.45735168e-01
-2.03791618e-01 4.18778181e-01 5.80682933e-01 -4.89605486e-01
-1.15841627e+00 -1.15917218e+00 -6.23721397e-03 1.27179694e+00
4.25264865e-01 -6.63356066e-01 -6.04922593e-01 -1.01603937e+00
1.55409068e-01 6.36950672e-01 -8.34251881e-01 3.96101534e-01
-3.29670340e-01 -1.29895777e-01 6.35240376e-01 7.82349825e-01
5.37122011e-01 -9.83189166e-01 -3.70658666e-01 -4.51366633e-01
-6.00225508e-01 -1.57911897e+00 -2.73568988e-01 -1.48506448e-01
-5.82285106e-01 -1.35647190e+00 3.57384533e-02 -8.52928519e-01
9.30421472e-01 1.62574854e-02 1.43808115e+00 3.70365798e-01
-3.15506309e-01 8.38813424e-01 -3.04111660e-01 -2.78766632e-01
-2.17688560e-01 -2.96254992e-01 -3.50317597e-01 -1.59750730e-01
3.75510305e-01 -2.34037805e-02 -4.48358476e-01 3.89901131e-01
-8.51899385e-01 7.44516790e-01 4.94999617e-01 7.58398533e-01
9.97447729e-01 -4.20629643e-02 3.09926271e-01 -9.86589491e-01
3.12756807e-01 -2.99037486e-01 -2.67978251e-01 6.86404765e-01
-4.43593979e-01 1.95554867e-01 3.82386208e-01 1.02533571e-01
-1.39348876e+00 3.57187301e-01 2.45902658e-01 -3.16388488e-01
-5.00331342e-01 2.60984182e-01 -1.32759243e-01 3.73887755e-02
5.24283648e-01 -2.17658862e-01 -3.06631505e-01 -2.64586955e-01
6.94253981e-01 4.95982468e-01 7.00724006e-01 -9.70042288e-01
6.91941142e-01 8.29929590e-01 1.48461804e-01 -2.41301700e-01
-1.48120272e+00 -1.77709520e-01 -8.30819011e-01 -3.07070673e-01
9.96316373e-01 -1.12895381e+00 -9.80193317e-01 -1.74111910e-02
-1.26812971e+00 -7.26752341e-01 -5.09081408e-02 2.67576575e-02
-8.30376387e-01 4.02387947e-01 -6.59855604e-01 -4.64791864e-01
-1.16133764e-01 -8.25633764e-01 1.53121078e+00 -1.16739146e-01
-4.99418855e-01 -1.13763773e+00 -3.02270114e-01 1.15983999e+00
-4.43023778e-02 2.91228205e-01 9.78606522e-01 -3.04469973e-01
-9.13599670e-01 1.87023953e-01 -7.36025333e-01 -8.64850581e-02
-7.58633018e-02 1.23982459e-01 -9.53818381e-01 -3.34497467e-02
-5.23609042e-01 -9.39606190e-01 8.81773531e-01 9.11031067e-02
1.40591121e+00 -4.40502554e-01 -5.12719572e-01 2.98591286e-01
1.43419254e+00 -3.63672197e-01 3.82399976e-01 4.32477444e-01
9.56664264e-01 7.70242095e-01 8.73482525e-01 3.51023704e-01
8.80204082e-01 6.92028224e-01 6.71299696e-01 -2.23922916e-02
-4.08329397e-01 -4.47429538e-01 5.48144169e-02 1.22371100e-01
-1.99880987e-01 -4.08300459e-01 -1.09875250e+00 5.22641957e-01
-2.40241003e+00 -1.18458104e+00 -1.46409497e-01 1.61346829e+00
9.10126925e-01 2.84233224e-02 -4.39251289e-02 -2.46734768e-01
6.63080156e-01 -1.64149534e-02 -4.92968172e-01 -3.58254254e-01
-1.09382393e-02 -2.54292488e-02 2.61271924e-01 9.06899154e-01
-1.23778141e+00 1.18612814e+00 6.28973055e+00 6.18040204e-01
-6.29225373e-01 1.12892762e-01 6.61168873e-01 -2.83032119e-01
-5.61584413e-01 1.10971749e-01 -8.82269800e-01 5.54401502e-02
4.27684069e-01 3.16736966e-01 3.12132865e-01 4.20326501e-01
1.01229064e-01 -2.35541388e-01 -1.42618048e+00 8.34694266e-01
4.39630687e-01 -1.58507776e+00 1.98154762e-01 -1.94927782e-01
8.34327579e-01 -2.65528291e-01 -9.45411772e-02 3.65849435e-02
5.37659883e-01 -1.51417184e+00 9.08496857e-01 8.54759991e-01
9.73221362e-01 -4.25089300e-01 3.24641198e-01 2.51352072e-01
-1.25037766e+00 7.49999881e-02 -2.39869896e-02 -2.77797699e-01
1.66725591e-01 4.64448214e-01 -1.00976515e+00 3.90125364e-01
8.64134431e-01 9.86322403e-01 -7.15995252e-01 5.44138193e-01
-8.15919161e-01 3.66421878e-01 1.79457874e-03 2.11289838e-01
3.71651947e-02 4.23937052e-01 5.90038039e-02 1.40636754e+00
-1.89375260e-03 2.09326550e-01 2.26458609e-01 9.21963751e-01
-5.55147231e-01 -2.79262424e-01 -5.38967788e-01 1.79231450e-01
5.31715035e-01 1.36381447e+00 -5.98093688e-01 -4.87241864e-01
-5.16554892e-01 1.18259966e+00 8.62780094e-01 4.07781601e-01
-9.68545556e-01 -6.47347718e-02 3.17551047e-01 4.18185294e-02
4.62765723e-01 2.10622959e-02 -5.93758404e-01 -1.33629489e+00
2.53099144e-01 -6.24393284e-01 6.01516187e-01 -1.50665700e+00
-1.48253655e+00 2.70668536e-01 -5.49785979e-03 -1.00713825e+00
-9.63656977e-02 -9.26981032e-01 -3.55819434e-01 5.88776469e-01
-1.56780958e+00 -1.87179744e+00 -6.63831890e-01 7.45242953e-01
5.69316328e-01 1.84132278e-01 7.84359038e-01 -1.24250650e-01
-3.19281489e-01 4.23409849e-01 -4.88830000e-01 2.55565673e-01
1.03776324e+00 -1.38178098e+00 1.09729715e-01 7.46994019e-01
2.58370072e-01 4.91584510e-01 7.21043408e-01 -8.20055604e-01
-1.28371286e+00 -1.12210691e+00 1.12692606e+00 -9.93071496e-01
8.81839991e-01 -3.38611394e-01 -7.94811249e-01 1.39472723e+00
4.67384279e-01 2.55558670e-01 8.02978933e-01 2.18517736e-01
-1.07426023e+00 1.13962173e-01 -1.09346700e+00 5.23236215e-01
1.36035585e+00 -7.76785672e-01 -8.23063970e-01 4.59477127e-01
5.91238499e-01 -5.69237173e-01 -1.09423018e+00 4.41558063e-01
7.88752437e-01 -9.84207153e-01 9.96683776e-01 -8.22145224e-01
1.08111155e+00 -6.16178691e-01 -4.11151797e-01 -8.85208249e-01
-2.91496158e-01 7.95114636e-02 -3.72831404e-01 9.90297437e-01
5.31633556e-01 -1.82696775e-01 9.05942440e-01 8.82097125e-01
-1.05287865e-01 -8.61080587e-01 -6.44405603e-01 -4.90576327e-01
-1.38987035e-01 -4.71399635e-01 2.93452233e-01 1.16803288e+00
4.44540262e-01 5.34552097e-01 -3.36369753e-01 3.54511529e-01
8.57088745e-01 5.89179993e-01 6.93711877e-01 -9.37351048e-01
-3.93249571e-01 -3.40510875e-01 -2.61246979e-01 -8.78151119e-01
5.67359328e-01 -1.03118443e+00 1.93805043e-02 -2.06585526e+00
6.85204864e-01 -2.36930311e-01 -9.57829785e-03 1.21451926e+00
-3.35857384e-02 6.65306091e-01 3.14092368e-01 3.69694054e-01
-1.30670393e+00 2.03383192e-01 1.28473175e+00 -3.82524550e-01
4.95842248e-01 -6.22240007e-01 -8.29466641e-01 8.51382792e-01
7.03809679e-01 -1.11969963e-01 -3.73863786e-01 -6.86816514e-01
5.60005903e-01 -3.94383371e-02 1.00568426e+00 -5.33086181e-01
4.23114330e-01 -3.35135549e-01 7.14122295e-01 -4.17655289e-01
5.15031397e-01 -8.81484985e-01 1.89152285e-01 6.87257499e-02
-5.23991287e-01 -9.42080095e-02 2.73290277e-01 5.94869435e-01
-2.40321890e-01 8.90365243e-03 4.62002814e-01 -1.21485151e-01
-1.26530707e+00 1.05983973e-01 3.32091227e-02 8.97320285e-02
1.19759345e+00 -9.88121256e-02 -1.00549030e+00 -3.50976437e-01
-9.42677081e-01 4.06654239e-01 8.21031630e-01 2.99735606e-01
7.50885665e-01 -1.42574573e+00 -6.10484183e-01 -3.67373526e-01
5.51841497e-01 3.97754833e-02 1.48808137e-01 7.67228067e-01
-2.82782316e-01 5.24019420e-01 -1.44625142e-01 -5.39596319e-01
-1.42392778e+00 3.90160859e-01 2.21140012e-01 -1.42187811e-02
-5.19274175e-01 8.74580503e-01 3.39021891e-01 -4.11943853e-01
2.57546872e-01 -1.09327927e-01 -1.37296626e-02 -2.25858226e-01
2.82993764e-01 9.90650579e-02 -1.51640296e-01 -7.36137390e-01
-5.62205851e-01 8.10774267e-01 -6.92969770e-04 -2.73115039e-02
1.03724122e+00 -3.61328632e-01 -2.16451481e-01 5.58144450e-01
1.02010560e+00 6.09234199e-02 -1.47210336e+00 -3.29232126e-01
-4.75309156e-02 -4.93285358e-01 -2.61118919e-01 -1.38834083e+00
-5.80169022e-01 7.61709392e-01 -3.89968045e-02 -7.80660510e-02
1.07666028e+00 7.52568364e-01 4.22032177e-01 4.60545719e-01
2.70499468e-01 -8.85923743e-01 3.88719559e-01 4.87691671e-01
9.38890398e-01 -1.76747513e+00 1.58336192e-01 -6.32023752e-01
-9.37501013e-01 1.11591470e+00 8.95445764e-01 1.48191169e-01
2.30546758e-01 2.92614847e-01 2.93851607e-02 -5.34710050e-01
-1.10350025e+00 -1.73624426e-01 6.47881210e-01 3.99653077e-01
6.34630442e-01 1.45278156e-01 2.21542194e-01 4.30265844e-01
-7.99943954e-02 1.20885983e-01 3.53349090e-01 8.26358557e-01
-5.00352740e-01 -9.57769334e-01 -3.73791456e-01 4.41127181e-01
-2.57936716e-01 -1.17306262e-01 -7.11009979e-01 7.56714761e-01
1.71925813e-01 9.34925497e-01 1.13089196e-01 -1.93406075e-01
2.31621116e-01 1.95619941e-01 7.59468913e-01 -6.38149142e-01
-1.93021730e-01 -2.59569138e-01 4.63950902e-01 -8.16391349e-01
-8.03447545e-01 -5.93892038e-01 -1.61700022e+00 -3.34389359e-01
2.52479017e-01 -2.25345135e-01 3.76324266e-01 1.21450806e+00
4.27211761e-01 3.40822756e-01 -9.79407039e-03 -6.24118447e-01
7.43506774e-02 -4.47083384e-01 -2.03885436e-01 7.99796820e-01
3.04942906e-01 -6.02127016e-01 -1.06583066e-01 7.32171297e-01] | [10.653221130371094, 1.5937604904174805] |
562b073d-7bd6-4a76-af81-e4f6becc958d | lifted-message-passing-for-the-generalized | 1610.01525 | null | http://arxiv.org/abs/1610.01525v1 | http://arxiv.org/pdf/1610.01525v1.pdf | Lifted Message Passing for the Generalized Belief Propagation | We introduce the lifted Generalized Belief Propagation (GBP) message passing
algorithm, for the computation of sum-product queries in Probabilistic
Relational Models (e.g. Markov logic network). The algorithm forms a compact
region graph and establishes a modified version of message passing, which
mimics the GBP behavior in a corresponding ground model. The compact graph is
obtained by exploiting a graphical representation of clusters, which reduces
cluster symmetry detection to isomorphism tests on small local graphs. The
framework is thus capable of handling complex models, while remaining
domain-size independent. | ['Udi Apsel'] | 2016-10-05 | null | null | null | null | ['symmetry-detection'] | ['computer-vision'] | [-1.59155354e-01 6.22540951e-01 -2.71248490e-01 -3.69837433e-01
-7.42964745e-01 -6.39787376e-01 9.86135244e-01 6.12173021e-01
-4.46177460e-02 3.04806739e-01 -2.26207256e-01 -4.70569313e-01
-6.97168350e-01 -1.40799773e+00 -8.40984166e-01 -5.20978868e-01
-7.17124164e-01 1.23722124e+00 9.55446959e-01 -1.53949410e-01
1.35802254e-01 8.77687037e-01 -1.27530229e+00 3.75576407e-01
-2.92413887e-02 6.17966413e-01 -1.31524041e-01 8.68470848e-01
-6.52380884e-02 1.28808975e+00 -4.20615263e-02 -6.78574979e-01
-1.37145016e-02 1.18788935e-01 -1.36863029e+00 -7.58015215e-02
-3.65533046e-02 3.12639517e-03 -8.67901802e-01 1.29240978e+00
-3.23140919e-01 3.12775075e-01 5.52333295e-01 -1.74784315e+00
-2.97267288e-01 1.24768126e+00 -2.54458413e-02 -9.22513902e-02
7.14463532e-01 -4.56375480e-01 1.51276648e+00 -6.18364513e-01
7.90126085e-01 1.70320809e+00 2.74151355e-01 -1.24836229e-01
-1.92531443e+00 3.74260470e-02 -1.50377363e-01 2.90381968e-01
-2.00731301e+00 -7.01988488e-03 3.59497875e-01 -4.09020782e-01
1.05660832e+00 5.84797502e-01 2.45916560e-01 5.15551090e-01
2.33097300e-01 5.35313964e-01 1.08808053e+00 -6.11860931e-01
5.54957271e-01 2.43734390e-01 7.04644501e-01 1.10555565e+00
3.82436723e-01 1.82684928e-01 -5.23513138e-01 -8.75694752e-01
6.17091000e-01 1.29333511e-01 1.60881132e-01 -1.02469110e+00
-1.06778228e+00 9.61913645e-01 3.13689649e-01 2.11345568e-01
-5.39453961e-02 4.06661898e-01 8.48709978e-03 4.65848714e-01
-1.55474365e-01 -9.27929394e-03 -2.71672457e-01 3.45875829e-01
-7.04361856e-01 4.93561894e-01 1.67807877e+00 1.36990452e+00
1.23396695e+00 -6.67858958e-01 -1.31248325e-01 1.11246035e-01
1.05705774e+00 6.31610453e-01 -3.13403547e-01 -1.02943850e+00
9.78830904e-02 7.14276791e-01 6.68396577e-02 -1.10813785e+00
-4.05474007e-01 4.65210006e-02 -5.51666915e-01 -9.44198072e-02
2.60715902e-01 5.73488712e-01 -6.50764465e-01 1.60912585e+00
4.18006063e-01 -1.46778539e-01 1.34026751e-01 1.93689436e-01
3.41949314e-01 6.35774493e-01 -1.86978236e-01 3.05983350e-02
1.40195203e+00 -4.79376912e-01 -3.29349309e-01 2.31866568e-01
7.98494637e-01 -2.93637007e-01 3.18008959e-01 4.32273179e-01
-1.04114306e+00 -1.18148118e-01 -9.25244987e-01 -7.08524287e-02
-7.08351254e-01 -6.67606175e-01 1.00780308e+00 7.57377148e-01
-1.44907141e+00 6.12505913e-01 -1.17692769e+00 -2.84765035e-01
-2.00662792e-01 6.64177775e-01 -5.89740813e-01 -3.59904975e-01
-1.27225113e+00 8.09112668e-01 9.03240323e-01 5.42706698e-02
-1.33082402e+00 -9.35749710e-02 -1.08182430e+00 9.45240408e-02
5.83651483e-01 -3.48753095e-01 1.17903399e+00 -1.02122715e-02
-1.38360298e+00 9.43687022e-01 -1.63129374e-01 -5.69014490e-01
1.05447240e-01 4.11884159e-01 -4.57746357e-01 2.85362303e-01
-2.53812253e-01 3.81466985e-01 6.98491395e-01 -1.19921517e+00
-4.21326667e-01 -5.02477467e-01 5.15980363e-01 -3.21810782e-01
5.93548834e-01 3.12295526e-01 -5.57266951e-01 1.32846028e-01
5.91253757e-01 -1.11311185e+00 -5.74534416e-01 -6.08268797e-01
-8.99406612e-01 -2.98974574e-01 4.55074549e-01 -1.14881240e-01
1.30621552e+00 -2.06058955e+00 2.85957128e-01 1.38940573e+00
5.69284141e-01 -7.22083747e-01 1.49777323e-01 9.89553869e-01
-2.73665856e-03 1.31142974e-01 -1.70875072e-01 -3.53300795e-02
4.93655086e-01 6.05183363e-01 -3.89896691e-01 6.16380155e-01
2.71996632e-02 6.97677255e-01 -8.10445309e-01 -8.32357585e-01
1.98056437e-02 -4.12007459e-02 -6.97803259e-01 6.37674853e-02
-7.16755569e-01 -2.85343170e-01 -4.48773235e-01 6.23733521e-01
9.05107498e-01 -4.21626627e-01 9.39331055e-01 1.73235372e-01
1.17749266e-01 1.80554807e-01 -1.80097365e+00 1.46224642e+00
3.46623287e-02 1.37948647e-01 1.78140983e-01 -8.14927816e-01
7.96608806e-01 2.28470534e-01 1.22478984e-01 4.77774404e-02
5.46455383e-03 -2.26406723e-01 -3.10755730e-01 -1.96038157e-01
5.88706374e-01 -6.78455234e-02 -5.15162289e-01 6.57537460e-01
3.13010544e-01 -2.39964858e-01 2.47171357e-01 9.82575178e-01
1.30856657e+00 -2.18376979e-01 1.66751564e-01 -4.70771313e-01
4.16880369e-01 8.69004354e-02 2.30453178e-01 1.45919907e+00
4.74053882e-02 1.56953827e-01 1.10391474e+00 -9.10156444e-02
-6.10858798e-01 -1.72040439e+00 -1.20136224e-01 1.28373241e+00
2.19177455e-01 -1.03074002e+00 -5.22998035e-01 -3.72902095e-01
1.10656559e-01 5.02346158e-01 -3.64086777e-01 -2.02583402e-01
-1.10172927e-01 -5.91793776e-01 7.40118861e-01 2.40550369e-01
-1.35532115e-02 -6.72470212e-01 8.26334804e-02 1.67912394e-01
1.10069253e-01 -9.49702382e-01 2.61333793e-01 2.91619509e-01
-9.21285391e-01 -1.22464526e+00 2.81454772e-01 -7.08326101e-01
5.02341807e-01 -6.77625239e-02 1.22108555e+00 -1.40334427e-01
-4.26349323e-03 8.15654993e-01 6.98258653e-02 -6.44919053e-02
-6.61114514e-01 -1.65027305e-01 3.19380150e-03 5.91289178e-02
4.85033154e-01 -4.78154778e-01 1.48034349e-01 2.61012882e-01
-1.22604561e+00 -3.31394404e-01 3.96893919e-01 6.22439146e-01
8.18864405e-01 5.72126508e-01 -4.79280263e-01 -1.17271996e+00
3.83007109e-01 -6.14790082e-01 -1.23846531e+00 5.51744759e-01
-4.96235281e-01 6.30311728e-01 1.25404403e-01 6.91726208e-02
-9.66390252e-01 7.29888678e-02 4.34665501e-01 -4.11113948e-02
-2.15923101e-01 9.61479247e-01 -2.50937581e-01 -2.59133577e-01
6.59392059e-01 -4.61699441e-02 -3.95205095e-02 -2.99942940e-01
7.72707760e-01 3.05302501e-01 4.81353998e-01 -1.12945819e+00
7.77515590e-01 7.23950088e-01 7.86417842e-01 -5.78476489e-01
-2.73436069e-01 -4.56259578e-01 -7.46535897e-01 1.52301028e-01
4.68236774e-01 -7.49577403e-01 -1.13972688e+00 2.31680065e-01
-1.13463271e+00 -2.35470727e-01 -1.21565521e-01 7.17217445e-01
-7.17329562e-01 5.07761896e-01 -9.97393906e-01 -8.66898179e-01
5.68350911e-01 -9.84254718e-01 7.03257799e-01 -3.96084100e-01
-4.36784029e-02 -1.06337368e+00 7.16746926e-01 -3.98708284e-02
-7.33219981e-02 -5.30117191e-02 1.35753167e+00 -1.18443942e+00
-1.17280972e+00 -6.06750607e-01 -2.73300171e-01 4.25791740e-02
-8.56868148e-01 3.53795081e-01 -8.57863426e-01 -5.12447730e-02
-3.10589701e-01 1.38630634e-02 5.73299170e-01 -1.62604228e-02
5.66102028e-01 -2.02491686e-01 -6.33137822e-01 -2.23611798e-02
1.65422487e+00 -1.34477511e-01 7.58810997e-01 -1.02810532e-01
4.36916798e-01 4.02775109e-01 1.10346086e-01 2.74397761e-01
5.85915208e-01 5.00419021e-01 3.40974420e-01 4.86048102e-01
4.90813583e-01 -4.34801608e-01 2.01659501e-01 6.91918910e-01
-6.87652826e-02 1.80409968e-01 -1.29487491e+00 3.27340096e-01
-2.10623741e+00 -9.15268064e-01 -4.36357915e-01 2.20733571e+00
6.75345421e-01 2.58992285e-01 -4.86268848e-03 -3.86166945e-02
8.33675623e-01 -1.82731692e-02 3.62451911e-01 -4.62818623e-01
6.45047706e-03 5.54319143e-01 7.12703824e-01 9.47286367e-01
-8.43326449e-01 8.58381331e-01 7.81714964e+00 7.30077863e-01
-1.56757846e-01 3.87129366e-01 -5.55067249e-02 4.16118175e-01
-4.83132809e-01 8.74758124e-01 -6.89133525e-01 -6.62950054e-02
1.38035858e+00 -1.32583693e-01 7.28142321e-01 1.11718404e+00
-4.74725932e-01 -3.57355267e-01 -1.59859800e+00 6.01951241e-01
-4.29455191e-01 -1.24857461e+00 4.56749164e-02 3.22549082e-02
4.23204094e-01 1.62306547e-01 -2.45102987e-01 5.05350471e-01
1.28807497e+00 -7.82903612e-01 7.95278192e-01 7.58318543e-01
1.19539313e-01 -8.42578232e-01 6.15781128e-01 1.04245670e-01
-1.03247249e+00 7.75392354e-02 -5.03556728e-01 8.72346088e-02
-1.37551442e-01 6.17690325e-01 -8.55450332e-01 8.21992397e-01
4.57666963e-01 -6.53461069e-02 -4.84112263e-01 7.49687195e-01
-1.18681699e-01 4.79163706e-01 -8.23336005e-01 -8.05302430e-03
1.51781306e-01 -3.74968797e-01 5.87139189e-01 1.35514367e+00
-1.79255351e-01 -1.12279035e-01 3.28899562e-01 9.16407287e-01
-1.86545238e-01 -1.56493515e-01 -8.81686807e-01 3.95087749e-02
6.39998972e-01 9.61640120e-01 -8.25424135e-01 -5.06541431e-01
-4.43973154e-01 4.48467016e-01 4.49002653e-01 4.95963514e-01
-5.76128066e-01 -3.48418355e-01 1.87261268e-01 -3.44108753e-02
4.03102249e-01 -2.86461800e-01 2.81271636e-01 -9.71039355e-01
-2.57384092e-01 -5.78560710e-01 9.21759248e-01 -6.53230369e-01
-9.88656700e-01 3.54147315e-01 5.97611248e-01 -4.79285210e-01
-2.46040791e-01 -6.11970782e-01 -3.17297518e-01 5.98149955e-01
-9.36330020e-01 -1.24603331e+00 2.64904559e-01 1.08018887e+00
-7.64485538e-01 1.97456464e-01 1.21953535e+00 -1.91778719e-01
-3.97935450e-01 2.41279960e-01 6.57984838e-02 2.38973852e-02
1.35643378e-01 -1.49200106e+00 -1.25692189e-01 9.15112495e-01
3.82269442e-01 1.09153092e+00 6.57351434e-01 -5.98012567e-01
-1.92494631e+00 -1.07037234e+00 9.55214679e-01 -6.91813231e-01
1.07013941e+00 -4.36393797e-01 -7.55860209e-01 1.37320375e+00
-7.31928200e-02 -3.45558301e-02 5.73063970e-01 6.23781204e-01
-9.31247413e-01 -2.42440879e-01 -1.26963127e+00 4.99484152e-01
7.06476688e-01 -1.19454575e+00 -6.68552577e-01 6.32720172e-01
7.15776145e-01 -1.59552604e-01 -1.47651708e+00 1.67834327e-01
2.63918430e-01 -8.69682252e-01 8.38940322e-01 -7.64742732e-01
-2.63808906e-01 -7.66762257e-01 -6.57805860e-01 -4.46702898e-01
-6.67609692e-01 -1.00062323e+00 -4.74833280e-01 8.06760073e-01
7.08034694e-01 -7.87885785e-01 6.11168921e-01 9.39831793e-01
2.44644165e-01 -1.63337752e-01 -1.06726956e+00 -6.19012773e-01
-3.97959262e-01 -9.53702629e-01 6.61810398e-01 7.03169525e-01
6.70004785e-01 1.37653470e-01 5.83370812e-02 8.11419189e-01
9.41663682e-01 2.40681961e-01 5.27727425e-01 -1.32508457e+00
-8.19650888e-01 -2.02055991e-01 -1.08216441e+00 -7.30268300e-01
1.96150422e-01 -1.30135787e+00 -7.47996494e-02 -1.06731737e+00
3.93073946e-01 -4.37424988e-01 -4.13774222e-01 3.79327595e-01
5.71331739e-01 -7.32599050e-02 -3.43628377e-02 3.24211121e-01
-1.23096633e+00 -3.98819707e-02 4.41715300e-01 -1.54271558e-01
4.14367467e-02 1.88398585e-01 -3.82565528e-01 6.35628104e-01
5.13375521e-01 -7.79414177e-01 -1.36768356e-01 2.11516708e-01
9.52204704e-01 3.28633904e-01 6.78476930e-01 -6.55456066e-01
7.64146745e-01 -1.94639906e-01 -3.19704682e-01 -9.00204778e-01
1.44161642e-01 -6.73432171e-01 7.89127052e-01 4.73265857e-01
-4.88026112e-01 6.77694753e-02 -2.89384544e-01 1.03770447e+00
-3.65845442e-01 -4.61573899e-01 5.45233905e-01 -3.08213204e-01
-7.70104170e-01 2.81263322e-01 -4.40365285e-01 -3.18874508e-01
8.49967897e-01 2.27403626e-01 -1.45151675e-01 -2.12179780e-01
-1.29166281e+00 5.45444787e-02 4.99377698e-01 -3.77294153e-01
4.97626513e-01 -1.25792491e+00 -2.34322518e-01 1.17104910e-01
3.19739580e-01 1.05652347e-01 1.57681078e-01 1.00351465e+00
-8.44908655e-01 6.42714083e-01 3.82649392e-01 -5.80751836e-01
-9.98839915e-01 1.06763780e+00 1.86068028e-01 -5.55704534e-01
-4.03570086e-01 7.57384837e-01 -1.30180284e-01 -5.06211996e-01
3.02918792e-01 -5.76725364e-01 1.59250483e-01 -4.72578585e-01
5.05342662e-01 4.99654353e-01 1.55170187e-01 -3.27548265e-01
-3.94910932e-01 -3.85906249e-02 -2.10945100e-01 -5.53798914e-01
8.40151072e-01 -1.08364441e-01 -1.11553347e+00 5.99957287e-01
9.30000842e-01 8.86343867e-02 -3.77565712e-01 -5.47938108e-01
5.46335459e-01 -1.37062445e-01 -1.08384281e-01 -1.01870202e-01
-4.39530820e-01 3.53226423e-01 1.06029347e-01 9.07028139e-01
4.04416054e-01 7.11378753e-01 -1.26651213e-01 1.23711991e+00
7.61807978e-01 -9.54737186e-01 -5.33680797e-01 6.23459935e-01
3.95561755e-01 -6.51879311e-01 -4.42606732e-02 -5.69091737e-01
-1.07623331e-01 1.06469047e+00 -1.38092965e-01 -5.74317463e-02
1.28115344e+00 3.73459280e-01 -6.55951679e-01 -7.56564736e-01
-9.23542440e-01 -1.28549993e-01 -2.11929440e-01 6.87943339e-01
-2.78006524e-01 6.19584441e-01 1.89405888e-01 5.15605569e-01
-9.02533084e-02 -1.58019602e-01 6.89787865e-01 1.23346734e+00
-4.79330093e-01 -1.11605978e+00 -6.52426243e-01 2.88816571e-01
-2.62498856e-01 -1.83253035e-01 -3.01118672e-01 9.75868583e-01
-2.92913318e-01 1.04869950e+00 1.38567597e-01 -4.71022874e-01
3.77903320e-02 1.72176510e-01 7.74580896e-01 -6.62443876e-01
-8.49581498e-04 -4.13765311e-01 2.33192652e-01 -8.00454021e-01
-4.38700676e-01 -4.49384660e-01 -9.90435600e-01 -7.58529544e-01
-5.21632016e-01 5.12513518e-01 5.36288083e-01 7.39711940e-01
3.35992604e-01 -1.17223002e-01 5.61659873e-01 -1.73726708e-01
-6.82692587e-01 -7.67325342e-01 -1.33566558e+00 1.31391004e-01
-2.92035826e-02 -5.77322900e-01 -2.65003622e-01 -4.13162522e-02] | [8.400256156921387, 6.523624420166016] |
50d54c6d-d7b5-4400-be3a-d4804306e99b | a-real-time-fire-segmentation-method-based-on | null | null | https://www.sciencedirect.com/science/article/pii/S2405896322005055 | https://pdf.sciencedirectassets.com/313346/1-s2.0-S2405896322X00074/1-s2.0-S2405896322005055/main.pdf | A Real-time Fire Segmentation Method Based on A Deep Learning Approach | As a kind of the forest “fault”, fire is highly destructive and difficult to rescue. Fire segmentation is helpful for firefighters to understand the fire scale and formulate a reasonable fire-fighting plan. Therefore, this paper proposes a real-time fire segmentation method based on deep learning. This method is an improved version of deeplbav3+, which is an encoder-decoder structure network. Encoder network is composed of deep convolutional neural network and atrous spatial pyramid pooling. Different from deeplabv3+, in order to improve the segmentation speed, this paper uses the lightweight network mobilenetv3 to build a new deep convolutional neural network and does not use atrous convolution, but it will affect the segmentation accuracy. Therefore, in order to compensate for the loss of segmentation accuracy, on the basis of the original decoder network, this paper adds two different shallow features to make the network contain rich fire feature information. Experimental results show that the comprehensive performance of this method is better than the original deeplabv3+, especially the segmentation speed of the network is greatly improved, which is about 59 FPS. | ['Yi Yingmin', 'Guo Xie', 'Han Liu', 'Shangbin Jiao', 'Ziquan Yu', 'Jing Xin', 'Lingxia Mu', 'Youmin Zhang', 'Mengna Li'] | 2022-07-22 | null | null | null | ifac-papersonline-2022-7 | ['real-time-semantic-segmentation'] | ['computer-vision'] | [-1.06388830e-01 -4.67617571e-01 3.53874895e-03 -1.05931982e-01
2.51330525e-01 -8.21091607e-02 3.53507064e-02 -4.57520813e-01
-8.33639443e-01 6.29435718e-01 1.69248790e-01 -4.28598970e-01
3.93144153e-02 -1.44744861e+00 -4.10886824e-01 -5.97282588e-01
3.02134603e-01 9.11096390e-03 6.57663286e-01 -3.54415953e-01
2.70280931e-02 5.27649641e-01 -1.12749445e+00 2.64280498e-01
6.70764327e-01 1.08872688e+00 4.87361729e-01 4.58662927e-01
-4.69345093e-01 8.27065885e-01 -5.45739532e-01 5.34057505e-02
4.04586941e-01 -2.47078314e-01 -9.80159283e-01 1.08273728e-02
-4.29232627e-01 -9.46832061e-01 -8.20027530e-01 9.27397788e-01
7.29946733e-01 -8.36568251e-02 2.26258263e-01 -1.01374567e+00
-3.22194397e-01 6.35133207e-01 -4.54022259e-01 4.85430568e-01
-2.12306693e-01 3.51407558e-01 3.72129828e-01 -3.04257870e-01
2.01124743e-01 1.24635029e+00 1.02002645e+00 2.83841431e-01
-4.82641518e-01 -9.70036447e-01 2.51806319e-01 3.97611737e-01
-1.70694089e+00 -2.23623216e-02 5.46501875e-01 -3.55184942e-01
5.16561449e-01 1.09034225e-01 1.01643276e+00 7.71564424e-01
1.61873475e-01 7.41372824e-01 7.80860186e-01 7.60382563e-02
-5.23917377e-02 -4.31687355e-01 1.31632462e-01 9.43097293e-01
2.31537387e-01 2.02620685e-01 2.33545154e-01 4.10465747e-01
1.33477008e+00 6.02157474e-01 -5.97239256e-01 3.27430576e-01
-1.11820471e+00 6.76961601e-01 1.18103027e+00 5.08727551e-01
-3.16186339e-01 1.85506806e-01 5.88866293e-01 -1.43694496e-02
2.26546943e-01 -2.88004011e-01 -5.35932958e-01 -3.41820568e-01
-1.03285635e+00 2.55668312e-01 6.57210350e-01 6.73921287e-01
9.87253010e-01 2.17700750e-01 -2.01986164e-01 5.72118580e-01
2.67011404e-01 5.25610805e-01 5.53407729e-01 -6.35362327e-01
3.69977981e-01 8.79332840e-01 -2.25760341e-01 -9.93584335e-01
-7.02405870e-01 -4.52201396e-01 -1.33559144e+00 3.19775075e-01
9.75500494e-02 -5.90328097e-01 -1.11281419e+00 1.03909087e+00
5.16375564e-02 1.91251025e-01 -1.85262859e-01 1.29525161e+00
1.13659835e+00 8.45161319e-01 5.96332029e-02 1.76813543e-01
1.17179406e+00 -1.04420364e+00 -6.68083251e-01 -1.02629483e-01
3.71823698e-01 -4.88378197e-01 1.02088165e+00 1.74218670e-01
-5.08610487e-01 -8.66319478e-01 -1.03264546e+00 -4.86229695e-02
-4.46875662e-01 1.15234509e-01 7.97060251e-01 6.53059959e-01
-5.87682784e-01 6.48039222e-01 -7.53481984e-01 -3.53278577e-01
8.02228153e-01 1.39942877e-02 4.81408387e-02 -8.11075568e-02
-1.44084036e+00 7.85711229e-01 8.96857917e-01 6.45897806e-01
-7.66608834e-01 -4.27439094e-01 -5.59503496e-01 1.75661176e-01
3.26505601e-01 -5.52182138e-01 1.28284478e+00 -7.57553041e-01
-1.07723761e+00 1.92876965e-01 7.70584345e-02 -2.90165931e-01
7.09936023e-01 -7.46861994e-02 -3.75195026e-01 1.16243988e-01
6.86242282e-02 7.69594967e-01 2.93756515e-01 -8.31630468e-01
-8.45321953e-01 -2.63327479e-01 3.41070384e-01 1.71555832e-01
-1.18219241e-01 -1.34398162e-01 -4.01747227e-01 -5.66779077e-01
1.35141805e-01 -5.65071166e-01 -5.20298362e-01 7.55352378e-02
-1.38110265e-01 8.51575136e-02 1.13845098e+00 -8.63603294e-01
1.40228760e+00 -1.99071181e+00 -3.49120080e-01 1.96944475e-01
8.90813917e-02 5.49372256e-01 2.42533788e-01 7.95863569e-03
2.38466516e-01 4.24589068e-01 -7.51051486e-01 5.80475748e-01
-3.55067015e-01 5.58321297e-01 1.34307951e-01 1.05445251e-01
-1.21237300e-02 7.91987062e-01 -6.40769303e-01 -7.64030397e-01
3.30494136e-01 5.18649459e-01 -4.55398709e-01 6.35565668e-02
-6.51805699e-02 3.70635509e-01 -5.94342172e-01 6.57620430e-01
1.09059203e+00 2.75029540e-01 -4.51393723e-01 -3.61300856e-01
-5.03481150e-01 -2.80134529e-01 -1.23893213e+00 1.67069781e+00
-3.29020172e-01 3.77797335e-01 2.70704120e-01 -9.82423902e-01
1.07355165e+00 2.73539305e-01 6.67736232e-01 -6.95548475e-01
6.75280571e-01 2.14634445e-02 -1.51384473e-01 -8.47009182e-01
1.98285520e-01 -3.33111793e-01 1.40513584e-01 1.34479001e-01
-2.89248139e-01 1.53229646e-02 2.23355994e-01 -1.96052045e-01
1.06205475e+00 2.80569673e-01 -2.63651997e-01 -2.01706558e-01
4.56612706e-01 2.43377477e-01 9.00147617e-01 4.03829992e-01
-3.24342906e-01 6.44811988e-01 3.13023984e-01 -8.26257885e-01
-6.73492968e-01 -7.50933111e-01 -2.06180871e-01 3.36552590e-01
6.22689843e-01 -5.15526235e-01 -9.31928217e-01 -5.95739722e-01
-2.47060046e-01 1.31323621e-01 -3.29425097e-01 -2.47172460e-01
-6.18475318e-01 -8.26596797e-01 8.15930247e-01 9.65070903e-01
1.95175111e+00 -1.50084400e+00 -9.87943411e-01 4.13070828e-01
-5.94803393e-01 -8.38071704e-01 -2.27013379e-01 -1.86483741e-01
-9.01965678e-01 -1.26436687e+00 -9.30521190e-01 -9.66686666e-01
1.92554340e-01 6.13461792e-01 5.73967159e-01 4.80204165e-01
-5.48866451e-01 -2.64243096e-01 -7.55319238e-01 -4.98886347e-01
3.12582582e-01 4.00531024e-01 -4.73211706e-01 -3.08702022e-01
4.32482630e-01 -5.90753496e-01 -6.62234545e-01 6.49372041e-02
-1.13610566e+00 3.65101516e-01 7.50073075e-01 4.83012885e-01
2.67804414e-01 7.40802944e-01 2.31643789e-03 -6.40935600e-01
6.33714914e-01 -2.08808601e-01 -3.14377874e-01 3.53373177e-02
-5.09401143e-01 -2.14495122e-01 7.67143309e-01 1.47101864e-01
-1.08968437e+00 -1.82028040e-02 -6.08828008e-01 -2.25283429e-01
-2.71584541e-01 6.56205714e-01 -3.24056268e-01 -3.44602987e-02
2.54217118e-01 2.37685904e-01 -3.56805511e-02 -7.81378269e-01
1.22019857e-01 9.96179700e-01 5.22804677e-01 -1.77927420e-01
5.72711349e-01 1.46319464e-01 -1.75601825e-01 -6.54257178e-01
-6.00335360e-01 -2.23927557e-01 -7.57096529e-01 -4.04902995e-01
1.12561810e+00 -7.68204212e-01 -9.95395720e-01 6.71245217e-01
-1.45442724e+00 -5.33862710e-01 -1.18605316e-01 6.13097310e-01
-2.93259323e-01 3.34756106e-01 -7.69491196e-01 -3.83497387e-01
-5.43096244e-01 -1.16410232e+00 5.69392741e-01 6.01194978e-01
5.33306956e-01 -5.69434047e-01 -2.72047788e-01 1.48769483e-01
6.92156971e-01 3.79692227e-01 4.96407628e-01 2.76655912e-01
-6.12458229e-01 8.37760698e-03 -7.66508043e-01 4.63306427e-01
2.19558980e-02 1.52796194e-01 -6.54412925e-01 1.92527160e-01
1.21797575e-02 2.05687732e-01 1.37581396e+00 3.77659887e-01
1.70826018e+00 -3.23996037e-01 -2.35234693e-01 9.91028130e-01
1.51542330e+00 5.01622200e-01 1.15541577e+00 8.42569470e-01
8.48830104e-01 7.94209316e-02 5.16896546e-01 4.68391359e-01
4.56086904e-01 1.66566744e-01 7.31676877e-01 -6.63093269e-01
-1.39472619e-01 -1.61636069e-01 2.10831046e-01 4.79299217e-01
-5.85677803e-01 9.48294178e-02 -9.42456543e-01 1.76709905e-01
-1.78404212e+00 -1.25468111e+00 -4.32013512e-01 1.56994390e+00
5.12310922e-01 9.86948907e-02 2.81237159e-02 4.11244482e-01
7.29673624e-01 4.40753490e-01 -2.83967495e-01 -3.52783412e-01
1.58240106e-02 4.82262880e-01 8.65137696e-01 3.40428978e-01
-1.37120867e+00 9.94439483e-01 5.65349960e+00 9.66856003e-01
-1.31928015e+00 1.29935667e-01 2.79448420e-01 2.29509056e-01
8.54022428e-02 5.37775382e-02 -6.97749972e-01 6.34034693e-01
4.26644236e-01 2.26830781e-01 5.42602479e-01 6.63974643e-01
3.51175964e-01 1.29890442e-03 -4.55962121e-02 9.64278340e-01
-3.43275219e-01 -1.22746396e+00 -6.85359761e-02 -7.96410069e-02
7.76568726e-02 8.32081214e-02 -6.00371242e-01 5.16562581e-01
1.87063441e-01 -9.26161110e-01 6.12037599e-01 5.87926269e-01
7.58532584e-01 -1.14811313e+00 1.16613448e+00 5.86232305e-01
-1.64342344e+00 -9.90190655e-02 -8.94688487e-01 -4.50561732e-01
4.57533360e-01 7.27071345e-01 -3.15358192e-01 8.10118020e-01
1.05765605e+00 9.51892674e-01 -4.25855845e-01 1.28409588e+00
-3.56111884e-01 6.11599386e-01 -2.33844221e-01 1.80200353e-01
6.39230907e-01 -3.20197225e-01 7.72187114e-02 1.35782993e+00
2.79550821e-01 4.56735671e-01 7.27801204e-01 7.46367753e-01
5.62813766e-02 4.72402247e-03 -4.24668044e-01 1.39577746e-01
1.36070758e-01 1.44351768e+00 -7.80730367e-01 -3.14433545e-01
-1.37417302e-01 9.93824065e-01 -1.23615503e-01 2.55289912e-01
-1.20291507e+00 -9.65386450e-01 3.05505067e-01 2.58902073e-01
2.42828980e-01 -4.10661042e-01 -2.15414912e-01 -1.08459365e+00
-1.30886823e-01 -2.75742531e-01 1.33168042e-01 -8.46378684e-01
-6.66352391e-01 6.59178615e-01 -2.73453087e-01 -1.12841928e+00
5.81235886e-01 -5.00499189e-01 -8.52718294e-01 8.97576511e-01
-1.70414340e+00 -1.23238409e+00 -9.74283695e-01 6.83936596e-01
5.76964259e-01 6.57478645e-02 7.06172168e-01 5.57511270e-01
-8.31099510e-01 -1.90728251e-02 -1.88990176e-01 6.86808228e-01
2.67549276e-01 -7.73340225e-01 2.65090078e-01 9.10635293e-01
-3.32104087e-01 2.46518776e-02 -3.34084965e-02 -6.82705522e-01
-7.03273416e-01 -1.45326304e+00 5.90250731e-01 3.74293238e-01
2.42992546e-02 -2.11592093e-02 -6.91314638e-01 5.82855642e-01
6.19762652e-02 -1.07810669e-01 2.15928435e-01 -2.44691417e-01
1.36372387e-01 -5.35664082e-01 -1.17110801e+00 5.37097454e-01
1.21237910e+00 -1.19876608e-01 -4.40606862e-01 3.36197168e-01
9.02570605e-01 -3.72985274e-01 -5.84529042e-01 5.42461693e-01
5.90559304e-01 -1.14931905e+00 8.13823819e-01 -2.13816240e-01
5.12119114e-01 -6.37898266e-01 1.02668948e-01 -9.91014957e-01
-8.20747852e-01 4.19614725e-02 4.07787770e-01 1.14257336e+00
-7.56646916e-02 -5.38231969e-01 9.01628554e-01 2.32939407e-01
-3.78130704e-01 -6.91061258e-01 -6.91734791e-01 -7.98155367e-01
-1.73902869e-01 -4.15800124e-01 1.01503372e+00 6.62252247e-01
-3.00295979e-01 2.28990674e-01 -8.92114863e-02 -2.86406707e-02
4.78960723e-02 9.05772019e-03 6.00645483e-01 -1.24383199e+00
3.01190138e-01 -4.10033733e-01 -4.01990175e-01 -1.27507305e+00
-2.04360932e-01 -8.55913579e-01 1.25021845e-01 -1.99172425e+00
-2.49664746e-02 -5.76242387e-01 -1.88866273e-01 8.19638789e-01
5.57261743e-02 1.55721307e-01 2.74959266e-01 8.76963660e-02
-1.81543067e-01 6.92535818e-01 1.69331491e+00 -4.03248757e-01
-6.34599626e-02 8.93428177e-02 -6.59159362e-01 8.49660814e-01
1.01390958e+00 -3.06841075e-01 -1.66329607e-01 -8.94489765e-01
-3.53185833e-01 5.89825585e-02 6.17159426e-01 -1.55348492e+00
3.01262319e-01 -3.12358022e-01 7.25024581e-01 -7.71108329e-01
-2.89463834e-03 -1.22045124e+00 4.72560674e-02 1.13116229e+00
1.41263351e-01 7.58181214e-02 2.65829712e-01 2.24544287e-01
-2.64425248e-01 -3.16727489e-01 5.13266623e-01 -5.41916013e-01
-1.18402910e+00 8.30600679e-01 -4.51630145e-01 -1.50097355e-01
9.13993537e-01 -4.27179396e-01 -1.85249880e-01 -8.02640691e-02
-5.21122932e-01 3.83638561e-01 2.09165424e-01 5.05465828e-02
6.79407895e-01 -1.37666655e+00 -7.44391322e-01 1.75409555e-01
-5.31880319e-01 5.44405997e-01 5.34331203e-01 6.91151738e-01
-1.40813434e+00 3.02122593e-01 -6.80838645e-01 -3.33529413e-01
-8.43265951e-01 3.85427147e-01 4.97183204e-01 -1.10456320e-02
-8.74225318e-01 7.56939769e-01 -1.45073803e-02 -3.93009841e-01
-4.58780937e-02 -4.73839730e-01 -5.21833122e-01 -1.33083269e-01
5.50868332e-01 5.98573685e-01 -1.29729733e-01 -4.18710887e-01
-3.86197984e-01 6.05460763e-01 2.44593143e-01 -4.94582504e-02
1.30931723e+00 1.32223397e-01 -2.74857104e-01 -3.37492637e-02
9.98749614e-01 -5.12730956e-01 -1.27905440e+00 1.87847435e-01
-5.61364830e-01 -5.77144325e-01 4.23575670e-01 -6.91598415e-01
-1.74640310e+00 1.28763247e+00 7.29736805e-01 -9.47874598e-03
1.38439834e+00 -6.82066798e-01 1.47714221e+00 2.44093955e-01
4.27595049e-01 -1.06265950e+00 -4.41266835e-01 8.10022831e-01
6.57723188e-01 -8.74652445e-01 -3.87779996e-02 -3.92961413e-01
-5.54850876e-01 1.43454492e+00 8.02552283e-01 -2.23239437e-01
8.79109561e-01 3.12454373e-01 1.04850732e-01 -9.54966247e-02
1.53549731e-01 -5.99491358e-01 -3.86561811e-01 5.87072730e-01
1.27162129e-01 8.96073803e-02 -6.47822559e-01 9.91113126e-01
-2.51790941e-01 5.19190729e-01 2.63889372e-01 1.04387283e+00
-9.91439819e-01 -8.81643593e-01 -2.43017942e-01 5.73577046e-01
-1.95895970e-01 -1.10861041e-01 -1.03644673e-02 6.99994504e-01
8.89450490e-01 9.57951188e-01 1.48032546e-01 -8.76421571e-01
5.75322151e-01 -3.82128954e-01 1.09481983e-01 -2.20681816e-01
-6.55139208e-01 -1.53220370e-01 -2.93350160e-01 -3.78894150e-01
-5.38504481e-01 3.49405445e-02 -1.62527120e+00 -7.74595082e-01
-3.43449205e-01 1.58004224e-01 5.08043468e-01 1.02199066e+00
-2.59237420e-02 9.25470531e-01 5.55180669e-01 -7.27685452e-01
1.64017864e-02 -9.29432750e-01 -8.57076347e-01 7.23405834e-03
-1.40560597e-01 -6.37423277e-01 1.47382692e-01 -2.55422015e-02] | [9.348976135253906, -0.8326335549354553] |
a9a9e1f8-cbab-42e1-ad11-142974acd9b9 | wordrank-learning-word-embeddings-via-robust | 1506.02761 | null | http://arxiv.org/abs/1506.02761v4 | http://arxiv.org/pdf/1506.02761v4.pdf | WordRank: Learning Word Embeddings via Robust Ranking | Embedding words in a vector space has gained a lot of attention in recent
years. While state-of-the-art methods provide efficient computation of word
similarities via a low-dimensional matrix embedding, their motivation is often
left unclear. In this paper, we argue that word embedding can be naturally
viewed as a ranking problem due to the ranking nature of the evaluation
metrics. Then, based on this insight, we propose a novel framework WordRank
that efficiently estimates word representations via robust ranking, in which
the attention mechanism and robustness to noise are readily achieved via the
DCG-like ranking losses. The performance of WordRank is measured in word
similarity and word analogy benchmarks, and the results are compared to the
state-of-the-art word embedding techniques. Our algorithm is very competitive
to the state-of-the- arts on large corpora, while outperforms them by a
significant margin when the training set is limited (i.e., sparse and noisy).
With 17 million tokens, WordRank performs almost as well as existing methods
using 7.2 billion tokens on a popular word similarity benchmark. Our multi-node
distributed implementation of WordRank is publicly available for general usage. | ['S. V. N. Vishwanathan', 'Shin Matsushima', 'Hyokun Yun', 'Shihao Ji', 'Pinar Yanardag'] | 2015-06-09 | wordrank-learning-word-embeddings-via-robust-1 | https://aclanthology.org/D16-1063 | https://aclanthology.org/D16-1063.pdf | emnlp-2016-11 | ['learning-word-embeddings'] | ['methodology'] | [-1.72712162e-01 -1.94577530e-01 -5.72690189e-01 -1.36510357e-01
-9.80644524e-01 -5.30451417e-01 8.82511079e-01 6.46183074e-01
-9.10527468e-01 2.51932353e-01 7.92520523e-01 -2.74016887e-01
-1.51317954e-01 -7.79144645e-01 -3.20733219e-01 -4.95599300e-01
-1.64916828e-01 5.61398447e-01 1.20521933e-01 -5.93090534e-01
5.31003475e-01 7.71933421e-02 -1.45086777e+00 -5.18261082e-02
6.70622706e-01 6.60761178e-01 1.57568440e-01 5.25133967e-01
-3.87432188e-01 2.12871671e-01 -2.66728103e-01 -5.68613172e-01
1.37834698e-01 -1.49677368e-03 -8.80790472e-01 -6.36639833e-01
4.08753276e-01 -3.53706211e-01 -6.86824441e-01 1.17352259e+00
6.82387590e-01 3.66866291e-01 6.05028212e-01 -1.04041338e+00
-1.11109817e+00 7.59606838e-01 -2.17462316e-01 1.80693030e-01
4.14735436e-01 -2.75848329e-01 2.15349412e+00 -1.46114218e+00
3.70483279e-01 1.30809104e+00 5.32757401e-01 2.28174821e-01
-1.28905833e+00 -3.53136808e-01 2.55143285e-01 5.34353971e-01
-1.48635685e+00 -1.55790269e-01 7.02410579e-01 -2.39117354e-01
1.26864171e+00 2.36627474e-01 5.88836432e-01 9.34619784e-01
4.21334207e-02 9.38104928e-01 5.71409643e-01 -6.63498044e-01
2.80857116e-01 5.51959090e-02 6.44834101e-01 5.27210593e-01
5.37032843e-01 -2.99454425e-02 -5.42345285e-01 -6.15008235e-01
3.12626004e-01 3.00820619e-01 -4.91310917e-02 -3.98039401e-01
-1.35019815e+00 1.39544725e+00 4.18990403e-01 5.44235170e-01
-3.36106628e-01 5.10111749e-01 8.62613738e-01 3.25280935e-01
7.80004680e-01 5.90058148e-01 -2.51076072e-01 -3.76729310e-01
-7.19707489e-01 3.01528662e-01 6.54419839e-01 6.43067718e-01
7.22839117e-01 -1.34593815e-01 -1.33266032e-01 1.13935149e+00
5.90440750e-01 4.23171520e-01 9.69534755e-01 -5.17234802e-01
3.85424674e-01 3.80436987e-01 -2.21607685e-01 -1.22106421e+00
-9.46396738e-02 -2.70656854e-01 -6.86135769e-01 -2.41681516e-01
4.88241017e-03 3.17455024e-01 -4.64416504e-01 1.67852914e+00
1.72010690e-01 3.19056928e-01 8.50659143e-03 8.34866285e-01
7.35579610e-01 8.27067554e-01 -1.43498152e-01 1.01833329e-01
1.62807000e+00 -1.01461577e+00 -7.64406085e-01 -3.53432447e-01
8.87956977e-01 -7.44365335e-01 1.35414326e+00 8.03392529e-02
-8.56217384e-01 -3.11511546e-01 -1.03003073e+00 -2.99969375e-01
-6.27343178e-01 -4.32611614e-01 1.01552343e+00 5.55018663e-01
-1.00584280e+00 6.04570150e-01 -6.19350135e-01 -6.15991473e-01
3.88430641e-03 9.75070298e-02 -4.07612830e-01 -2.28336170e-01
-1.65709007e+00 9.96353269e-01 1.63961664e-01 -2.88989156e-01
-4.34253037e-01 -7.73501813e-01 -9.82962668e-01 1.00774035e-01
3.29558700e-02 -5.05220175e-01 1.04531229e+00 -3.71960253e-01
-1.23064339e+00 6.23028517e-01 -1.23662777e-01 -3.88787836e-01
4.68533672e-02 -5.13580501e-01 -3.76015842e-01 8.08412358e-02
1.79038942e-01 2.93022662e-01 6.29310727e-01 -7.96794236e-01
-3.36322516e-01 -2.00994983e-01 1.86562583e-01 1.00445583e-01
-1.21686506e+00 1.25799000e-01 -3.85727555e-01 -9.69092250e-01
-1.59413844e-01 -7.90902197e-01 -3.50276798e-01 1.70871943e-01
-8.47405195e-02 -7.33453274e-01 3.39758307e-01 -3.26624155e-01
1.60051715e+00 -2.09733462e+00 1.42396078e-01 2.82734841e-01
3.46350044e-01 3.61232579e-01 -6.69961035e-01 1.11860943e+00
-1.76996872e-01 1.86357245e-01 -9.31418240e-02 -4.16047424e-01
5.29628754e-01 2.52580374e-01 -6.57082081e-01 6.56947315e-01
-1.01472646e-01 1.02613473e+00 -1.32080746e+00 -4.89594966e-01
7.63465911e-02 6.38557076e-01 -7.89241314e-01 1.58753276e-01
2.49345422e-01 -7.11941898e-01 -6.08499587e-01 3.55494499e-01
3.09167147e-01 -3.05814266e-01 2.19868630e-01 -2.54025102e-01
6.58048466e-02 5.84656000e-01 -1.14853179e+00 1.99089336e+00
-6.83506370e-01 6.88926935e-01 -4.21197951e-01 -1.10289085e+00
7.59992063e-01 2.85386026e-01 4.33362097e-01 -5.19955099e-01
-1.29843086e-01 3.65810037e-01 -3.28603722e-02 -1.71596855e-01
1.04037952e+00 7.55531564e-02 -1.62094325e-01 8.44414413e-01
-1.14053309e-01 2.43919622e-02 4.20477390e-01 6.72939718e-01
1.30406654e+00 -5.67059219e-01 3.84394020e-01 -4.35537577e-01
3.04222584e-01 -3.09009820e-01 1.18394621e-01 5.94253898e-01
2.43007448e-02 4.31299955e-01 3.19136083e-01 -3.22980851e-01
-1.01080632e+00 -1.08696878e+00 -1.81533530e-01 1.41135895e+00
8.38750675e-02 -1.08100009e+00 -3.38733792e-01 -4.46937680e-01
4.11389023e-01 5.75676918e-01 -7.05658078e-01 -3.20590883e-01
-4.28657115e-01 -7.07361400e-01 5.02922237e-01 6.76991701e-01
-1.07124165e-01 -8.38478625e-01 -2.63579451e-02 2.92259306e-01
4.77862805e-02 -8.10484350e-01 -7.53847778e-01 1.00459836e-01
-7.91921377e-01 -7.78165102e-01 -9.13130999e-01 -8.55303407e-01
5.62697828e-01 6.19918585e-01 1.31044352e+00 2.34714627e-01
-4.06832993e-01 4.41772968e-01 -7.34224081e-01 -2.35922169e-02
-1.61981396e-02 3.45193058e-01 3.82152766e-01 -2.02620164e-01
7.17064857e-01 -6.57114863e-01 -7.87245214e-01 -5.62538318e-02
-1.11456501e+00 -5.67531824e-01 6.06400788e-01 1.20365238e+00
3.99369240e-01 -4.32250023e-01 6.38604164e-01 -7.34214485e-01
1.10789442e+00 -5.88684320e-01 -4.11586344e-01 1.25918537e-01
-1.12404311e+00 4.67523515e-01 5.33943653e-01 -5.30436754e-01
-1.76178619e-01 -3.65248829e-01 -2.07500458e-01 -1.37106240e-01
4.32137281e-01 7.69666314e-01 2.86909729e-01 1.52340129e-01
5.13748407e-01 2.13740543e-01 -1.83949113e-01 -6.02953970e-01
9.43408966e-01 8.33789110e-01 7.18208030e-02 -6.50707781e-01
9.95085895e-01 3.74187380e-01 -2.59091794e-01 -8.38686347e-01
-6.49629295e-01 -9.19329405e-01 -3.58159125e-01 2.75352478e-01
5.56205273e-01 -8.79456818e-01 -2.88778424e-01 -8.17994773e-02
-1.41878462e+00 1.51710287e-01 -4.21101332e-01 6.62147164e-01
-1.30766302e-01 7.38637626e-01 -7.40171552e-01 -6.17254496e-01
-5.87673008e-01 -9.32316244e-01 1.11423254e+00 -4.86492932e-01
-4.44069773e-01 -1.29924357e+00 6.67042136e-01 -4.76755500e-02
6.01614654e-01 -3.55621070e-01 1.17080295e+00 -9.52748716e-01
-1.82023376e-01 -6.14941299e-01 -2.74151981e-01 3.96549970e-01
9.11704153e-02 -2.12454706e-01 -7.75263608e-01 -4.08393174e-01
-4.96583968e-01 -2.87015289e-01 1.11610377e+00 1.81846209e-02
9.59455073e-01 -2.86363304e-01 -3.04544508e-01 2.64075577e-01
1.47798765e+00 -5.82581878e-01 4.65066880e-01 3.82059932e-01
6.29517496e-01 4.07947063e-01 5.28947234e-01 6.08004391e-01
4.95802701e-01 8.83827806e-01 4.40827519e-01 1.92261338e-02
-8.50327779e-03 -3.41509700e-01 3.51973534e-01 1.63421619e+00
2.18160510e-01 -1.65723741e-01 -7.58654356e-01 9.58497524e-01
-1.94924855e+00 -8.77335548e-01 -6.63821213e-03 2.17486358e+00
9.70000088e-01 -1.58089586e-02 9.54142883e-02 3.37390125e-01
6.37158632e-01 7.90107548e-01 -2.29629681e-01 -5.97968876e-01
-1.09604420e-02 3.30891728e-01 4.99817163e-01 6.88012242e-01
-7.63286471e-01 9.41124797e-01 6.54362345e+00 1.08355176e+00
-6.98120117e-01 2.24067077e-01 1.18185371e-01 -1.16235562e-01
-8.24193716e-01 -1.64194889e-02 -6.16015911e-01 4.98844326e-01
8.37044895e-01 -6.14237010e-01 4.34364974e-01 8.60733986e-01
-1.85637042e-01 4.48540658e-01 -1.32364571e+00 1.25903571e+00
3.69334847e-01 -1.23271465e+00 1.80793494e-01 1.59125924e-01
6.15149379e-01 3.33218217e-01 2.13473931e-01 3.00107360e-01
3.35932165e-01 -1.02346849e+00 3.96358788e-01 1.94561541e-01
7.87666380e-01 -7.92328000e-01 7.30260253e-01 1.71377361e-02
-1.49745405e+00 7.50039369e-02 -8.66126657e-01 -1.76470250e-01
1.77721709e-01 1.04884279e+00 -2.40327090e-01 2.98939466e-01
3.83866787e-01 8.71478260e-01 -4.46132153e-01 7.72410214e-01
-3.46922308e-01 6.38572335e-01 -2.56780893e-01 -5.06611943e-01
4.27920848e-01 -1.84597969e-01 4.28572983e-01 1.37197161e+00
2.46168658e-01 -1.01284452e-01 5.04317321e-02 5.11461556e-01
-3.37185651e-01 5.97741425e-01 -7.22661674e-01 -3.35098386e-01
6.30189896e-01 1.34962046e+00 -2.06394151e-01 -2.38361672e-01
-6.30608857e-01 9.88395452e-01 5.74283540e-01 2.23731384e-01
-5.92158496e-01 -9.31975186e-01 1.19750690e+00 -9.65582877e-02
3.66982758e-01 -3.70892107e-01 7.81748667e-02 -1.14414418e+00
1.94567576e-01 -5.75061500e-01 2.52800524e-01 -3.56865793e-01
-1.81470704e+00 6.18309438e-01 -2.99306940e-02 -1.18038797e+00
-4.25308853e-01 -8.37457478e-01 -6.16635859e-01 6.59485579e-01
-1.77075243e+00 -7.52724171e-01 1.21482417e-01 2.82779336e-01
4.39390093e-01 -2.93021917e-01 1.23277187e+00 5.71249425e-01
-3.08456808e-01 8.74816835e-01 6.73230767e-01 2.30194181e-01
7.88484097e-01 -1.25150764e+00 9.29956257e-01 5.44920504e-01
7.52082705e-01 9.11704123e-01 5.75336337e-01 -3.04295212e-01
-1.78052819e+00 -8.26765597e-01 1.49176490e+00 -3.06531698e-01
1.37002301e+00 -6.06762648e-01 -8.48819852e-01 2.54400134e-01
2.32759774e-01 4.47626442e-01 9.53720093e-01 4.23616618e-01
-9.01186645e-01 -6.56286851e-02 -6.38638198e-01 8.55821550e-01
1.08694673e+00 -1.00592685e+00 -9.12920356e-01 7.65343785e-01
1.02283466e+00 2.33355850e-01 -8.79239500e-01 -1.30072251e-01
7.72234976e-01 -4.03646588e-01 1.28586292e+00 -8.64580154e-01
3.88388216e-01 6.65336102e-03 -6.97683811e-01 -1.50889075e+00
-5.70902228e-01 -6.00186110e-01 -2.63466060e-01 1.10238707e+00
3.71270210e-01 -9.26364422e-01 5.58513403e-01 2.92206913e-01
1.51747912e-01 -8.70710731e-01 -9.78828907e-01 -1.13648236e+00
3.97377938e-01 -5.38398027e-01 5.84732354e-01 8.90306830e-01
4.98776257e-01 4.31554675e-01 -3.39028060e-01 -2.85096824e-01
5.59760809e-01 6.98566809e-02 4.64065045e-01 -1.28660142e+00
-3.12819988e-01 -5.20122886e-01 -7.03925550e-01 -1.36590207e+00
6.61033571e-01 -1.35086179e+00 -1.09758362e-01 -1.65704012e+00
2.60693341e-01 -2.68726349e-01 -6.98810041e-01 2.91755259e-01
-4.04353112e-01 3.90362829e-01 1.13710441e-01 2.58590966e-01
-7.20833361e-01 9.20423746e-01 6.95670843e-01 -2.45781332e-01
3.93571287e-01 -6.00644588e-01 -8.14839780e-01 3.76705885e-01
6.65727019e-01 -6.28146768e-01 -3.28352392e-01 -8.10504496e-01
6.41538918e-01 -6.84553981e-01 3.99073996e-02 -4.25262183e-01
3.52872103e-01 1.84189454e-01 -3.38597745e-01 -3.41740221e-01
5.04827976e-01 -6.65084243e-01 -5.02459288e-01 5.17813921e-01
-6.70966625e-01 5.52663267e-01 -2.33993456e-01 7.56733298e-01
-3.24739546e-01 -4.31646377e-01 3.18398029e-01 1.95435062e-01
-6.68569684e-01 3.35653871e-01 -2.65233904e-01 4.68603790e-01
4.99014318e-01 -1.26029039e-03 -2.73160964e-01 -4.30308849e-01
7.29286969e-02 -5.35644107e-02 2.87240297e-01 6.08399332e-01
7.99432755e-01 -1.90945625e+00 -1.02936852e+00 6.70660362e-02
6.17938578e-01 -5.95850408e-01 -1.91066861e-01 4.44294006e-01
-2.32873097e-01 5.90281665e-01 3.15617204e-01 -2.56058425e-01
-1.11514628e+00 5.58763206e-01 -3.73528421e-01 -4.96326894e-01
-6.28626108e-01 7.34447002e-01 5.63503653e-02 -3.03951770e-01
2.57198811e-01 -2.45722800e-01 -8.02202076e-02 2.30290443e-01
8.31982613e-01 4.79472846e-01 5.51185012e-02 -5.55580258e-01
-5.72760582e-01 8.21277618e-01 -2.76511878e-01 -2.05325201e-01
1.49053764e+00 5.75549304e-02 -1.82426915e-01 4.88910317e-01
1.76952255e+00 6.92975745e-02 -3.75117272e-01 -7.11924970e-01
1.45118266e-01 -6.67369187e-01 1.96781099e-01 -1.33503839e-01
-9.53947961e-01 1.03072476e+00 5.04974544e-01 3.34848255e-01
5.24299860e-01 1.13060661e-01 1.12751484e+00 8.43695104e-01
2.68858761e-01 -1.18996859e+00 3.65741968e-01 8.10105920e-01
8.17552805e-01 -1.20340550e+00 3.32143337e-01 -9.56813321e-02
-3.59522700e-01 9.75614548e-01 -4.33480963e-02 -5.73107004e-01
8.21552694e-01 -1.47117982e-02 -1.06829695e-01 -1.15954615e-01
-8.98067057e-01 -2.52103746e-01 5.83095014e-01 5.27784169e-01
7.51638532e-01 -1.04636385e-03 -6.61598742e-01 5.35825491e-01
-1.15116216e-01 -6.56491756e-01 1.75958738e-01 6.60874903e-01
-5.92201352e-01 -1.56271458e+00 1.06472475e-03 5.22307217e-01
-4.45392162e-01 -8.30323040e-01 -2.25192294e-01 5.77129841e-01
-5.98577976e-01 9.20228779e-01 6.26938194e-02 -3.33894819e-01
2.35141024e-01 -4.71066833e-02 9.24654081e-02 -8.58772397e-01
-5.89642763e-01 -4.12099838e-01 2.18231529e-02 -6.97547495e-01
-1.23885021e-01 -4.65134054e-01 -1.05681217e+00 -3.75382811e-01
-5.29929519e-01 4.62209195e-01 8.50315154e-01 6.91097856e-01
3.18140954e-01 3.08725476e-01 6.94443703e-01 -7.64280915e-01
-1.14119422e+00 -8.87766898e-01 -7.96611905e-01 5.86463690e-01
1.26680106e-01 -5.40657938e-01 -6.28434479e-01 -5.57889640e-01] | [10.581413269042969, 8.583974838256836] |
cb4eda21-6b00-4a11-b019-23e82eac5483 | terrain-analysis-in-starcraft-1-and-2-as | 2205.08683 | null | https://arxiv.org/abs/2205.08683v1 | https://arxiv.org/pdf/2205.08683v1.pdf | Terrain Analysis in StarCraft 1 and 2 as Combinatorial Optimization | Terrain analysis in Real-Time Strategy games is a necessary step to allow spacial reasoning. The goal of terrain analysis is to gather and process data about the map topology and properties to have a qualitative spatial representation. On StarCraft games, all previous works on terrain analysis propose a crisp analysis based on connected component detection, Voronoi diagram computation and pruning, and region merging. Those methods have been implemented as game-specific libraries, and they can only offer the same kind of analysis for all maps and all users. In this paper, we propose a way to consider terrain analysis as a combinatorial optimization problem. Our method allows different kinds of analysis by changing constraints or the objective function in the problem model. We also present a library, Taunt, implementing our method and able to handle both StarCraft 1 and StarCraft 2 maps. This makes our library a universal tool for StarCraft bots with different spatial representation needs. We believe our library unlocks the possibility to have real adaptive AIs playing StarCraft, and can be the starting point of a new wave of bots. | ['Florian Richoux'] | 2022-05-18 | null | null | null | null | ['real-time-strategy-games'] | ['playing-games'] | [-4.23661977e-01 1.77798107e-01 5.55344336e-02 3.33291851e-03
6.84194714e-02 -9.65749025e-01 6.82631552e-01 4.34250206e-01
-5.17367482e-01 8.08291137e-01 -3.28898966e-01 -6.43099964e-01
-4.87136692e-01 -1.71635115e+00 -2.44590208e-01 -3.83601874e-01
-4.97897029e-01 1.21415401e+00 1.26483095e+00 -1.22546875e+00
3.86876673e-01 7.67796934e-01 -1.89775920e+00 -1.93197466e-02
8.91681015e-01 7.39311516e-01 3.60276163e-01 6.82052255e-01
-4.29531634e-01 9.09254551e-01 -6.21060312e-01 -3.41705441e-01
4.08750534e-01 -1.76631525e-01 -9.62523997e-01 -3.24627757e-01
-3.84013742e-01 -2.81543154e-02 -7.36733750e-02 1.11192501e+00
2.33276114e-01 2.69020975e-01 7.40306556e-01 -1.53327072e+00
2.05492333e-01 7.16147661e-01 -1.01322211e-01 3.51403095e-02
3.35337758e-01 3.31508033e-02 6.38624251e-01 -1.57764405e-01
8.12109768e-01 9.72233534e-01 4.98495907e-01 1.59816355e-01
-8.26184213e-01 -1.72441274e-01 6.08730596e-03 6.53961480e-01
-1.47363997e+00 1.30132750e-01 5.83837748e-01 -5.51159799e-01
6.70305312e-01 8.89470518e-01 1.37576973e+00 3.83432418e-01
1.58637047e-01 4.05571222e-01 1.05959558e+00 -2.85273671e-01
8.07815671e-01 7.85009097e-03 -1.46595404e-01 6.28852546e-01
4.87712920e-01 -1.79556116e-01 9.78097972e-03 3.79228666e-02
1.14029443e+00 -1.82470292e-01 6.91851974e-02 -7.23326981e-01
-8.46334577e-01 9.27743971e-01 4.36249852e-01 4.81671065e-01
-2.61894554e-01 2.37404227e-01 2.55343348e-01 2.91262507e-01
1.27437294e-01 7.26285279e-01 -2.70996671e-02 -5.20557642e-01
-9.65355396e-01 9.22462225e-01 1.20208454e+00 7.35688031e-01
9.31793451e-01 -1.91913962e-01 1.85995564e-01 1.87123761e-01
-3.98878716e-02 2.86449671e-01 1.12224616e-01 -7.91129947e-01
1.55930504e-01 1.01617575e+00 2.09126011e-01 -1.25944579e+00
-5.69275796e-01 -1.38704479e-01 -3.78495783e-01 1.18178904e+00
6.41594470e-01 4.32944745e-02 -7.27640569e-01 1.00867963e+00
3.36469382e-01 -1.27188414e-01 -1.33599892e-01 1.01636839e+00
6.78220332e-01 4.69550580e-01 -3.74003559e-01 1.64837450e-01
1.31922460e+00 -4.79304850e-01 -6.18506730e-01 1.64751232e-01
6.37720346e-01 -4.29680228e-01 1.03092563e+00 7.73690104e-01
-1.06181347e+00 1.37113810e-01 -1.20259607e+00 8.34706277e-02
-1.12335837e+00 -2.79370815e-01 8.34029317e-01 8.09525967e-01
-1.18043852e+00 7.52026320e-01 -7.19258785e-01 -3.78667682e-01
3.26006189e-02 4.36597049e-01 -1.13031633e-01 3.91515911e-01
-1.13934040e+00 1.39734757e+00 6.30022824e-01 -6.49506673e-02
-6.01710379e-01 -1.37068229e-02 -6.25906348e-01 1.39572918e-01
7.82767534e-01 -4.13167179e-01 1.04326272e+00 -3.77210975e-01
-1.37305546e+00 8.89179707e-01 3.75335306e-01 -6.09166622e-01
9.75166440e-01 4.61725265e-01 -1.22448295e-01 -2.41199806e-01
2.24381700e-01 2.11030573e-01 1.74265608e-01 -1.10703385e+00
-9.08393562e-01 -2.06279218e-01 5.66586316e-01 2.74731308e-01
1.64947778e-01 -1.82498321e-01 -4.58530754e-01 -3.84269029e-01
7.06833526e-02 -4.64639246e-01 -7.47758627e-01 -2.89208025e-01
-9.15133506e-02 -1.01907998e-01 5.90074241e-01 -4.31330293e-01
1.33434081e+00 -1.73632598e+00 3.72064561e-01 8.14923942e-01
3.74571770e-01 9.01815593e-02 4.18637961e-01 6.07677341e-01
4.66031313e-01 2.94712931e-01 -2.85751969e-01 2.41709813e-01
3.71190995e-01 4.60138172e-01 -1.48550004e-01 1.52945355e-01
-1.81315169e-01 6.41133785e-01 -1.05475974e+00 -6.73015594e-01
6.01166010e-01 -1.21441774e-01 -4.80382204e-01 -2.92347461e-01
-5.78044653e-01 3.12594348e-03 -5.91975927e-01 5.37950575e-01
7.87927091e-01 4.09217000e-01 2.98367858e-01 3.16340446e-01
-7.88750708e-01 7.40950704e-02 -1.59027910e+00 1.54670036e+00
-1.86886117e-01 6.08152270e-01 1.45183489e-01 -8.46948683e-01
1.31246662e+00 -1.87628537e-01 5.27043223e-01 -7.42428541e-01
3.91421437e-01 3.96912307e-01 -5.18302470e-02 -3.00891310e-01
9.36432183e-01 1.13566704e-01 -3.38611424e-01 2.75568038e-01
-2.49836460e-01 -8.71542215e-01 6.25023901e-01 1.77754387e-02
1.16180074e+00 9.89654362e-02 4.37405348e-01 -6.32085025e-01
3.25031489e-01 8.25174093e-01 2.73738205e-01 6.72487319e-01
2.29960799e-01 1.12651154e-01 9.64863300e-01 -8.30288947e-01
-1.18293464e+00 -1.15738225e+00 1.04905695e-01 7.53335536e-01
6.69376791e-01 -6.54928446e-01 -8.73215258e-01 -3.52868468e-01
-1.72909752e-01 6.65663123e-01 -6.20119512e-01 3.06696385e-01
-5.48087597e-01 -6.82290614e-01 4.48050529e-01 3.24762985e-02
4.95366544e-01 -1.00213385e+00 -1.30954635e+00 3.32467675e-01
-1.39072791e-01 -5.91364563e-01 5.16282737e-01 3.72800201e-01
-7.03742981e-01 -1.08718204e+00 -2.30140820e-01 -4.81273979e-01
2.85055041e-01 -7.06813186e-02 9.79981303e-01 2.17894912e-01
-2.06313804e-01 -6.61309436e-02 -6.28477991e-01 -7.30758309e-01
-1.18366085e-01 3.87103200e-01 -3.08093339e-01 -4.93006945e-01
-2.88693234e-02 -9.34959948e-01 -1.08065657e-01 5.90959013e-01
-7.83812284e-01 1.48598388e-01 1.54373005e-01 1.15810215e-01
4.15315419e-01 6.81177974e-01 -6.64992854e-02 -5.59935093e-01
9.35932875e-01 -3.76788467e-01 -1.15321839e+00 2.00826347e-01
-3.32843781e-01 2.42696553e-01 5.84076166e-01 4.33242992e-02
-5.75354099e-01 1.21786609e-01 -2.34410569e-01 -7.44953007e-02
-8.05381462e-02 6.97403252e-01 -4.41640645e-01 -2.93736309e-01
8.43240380e-01 4.45347950e-02 -1.79755270e-01 -4.12301123e-01
5.24140060e-01 3.67163211e-01 6.34427369e-01 -5.90143025e-01
7.43871927e-01 6.21202588e-01 3.58001292e-01 -7.19253421e-01
3.03044856e-01 -5.05783796e-01 -7.80680478e-01 -6.89015329e-01
8.16694319e-01 -1.92170784e-01 -1.09404266e+00 3.12040269e-01
-1.13988900e+00 -8.48531246e-01 -6.13875329e-01 1.00633286e-01
-7.67402768e-01 1.60843089e-01 4.53995429e-02 -1.00946033e+00
2.72967964e-01 -1.14588630e+00 6.77535772e-01 2.65871912e-01
1.00138234e-02 -9.36214209e-01 5.22376478e-01 -2.35522956e-01
3.48674238e-01 7.27883041e-01 5.79492450e-01 -3.78449470e-01
-7.87643850e-01 -1.62032664e-01 1.52201969e-02 -5.34613192e-01
-2.17794273e-02 2.34679997e-01 -4.63885486e-01 3.23103517e-01
-1.57676786e-01 3.31209242e-01 6.55187905e-01 2.57705152e-01
8.69528711e-01 -2.50100732e-01 -4.04677421e-01 5.96382916e-01
1.65724158e+00 4.58443373e-01 1.21075439e+00 9.45739627e-01
2.66422212e-01 7.15466082e-01 1.06465280e+00 4.75240856e-01
4.41818446e-01 1.20003140e+00 1.12182868e+00 -2.31963992e-01
3.02355230e-01 -1.03515275e-01 8.65347981e-02 1.32474214e-01
-8.54116201e-01 -1.11368321e-01 -1.31272078e+00 5.88607669e-01
-2.44960308e+00 -1.27245700e+00 -7.19027698e-01 2.13657808e+00
3.33662271e-01 2.45769233e-01 6.94357991e-01 4.38030213e-01
5.24279118e-01 -1.23125874e-02 1.10451825e-01 -6.73475623e-01
-1.82822291e-02 4.23465908e-01 9.98473048e-01 8.94993842e-01
-1.02502108e+00 1.31015909e+00 6.25144625e+00 1.14554417e+00
-1.13359439e+00 8.50077122e-02 -1.65894389e-01 -1.43383760e-02
-4.24501240e-01 3.20835710e-01 -3.90426427e-01 3.02714497e-01
4.48772788e-01 -3.94993246e-01 7.32957125e-01 1.09819436e+00
2.57732332e-01 -5.13967991e-01 -2.86052972e-01 7.90192664e-01
-3.04482281e-01 -1.74465311e+00 -1.32346645e-01 3.88860911e-01
2.00468481e-01 -1.06116802e-01 -4.47415292e-01 7.42661357e-02
7.90312588e-01 -1.31749952e+00 1.06930101e+00 6.53660357e-01
6.00694299e-01 -9.09668446e-01 8.40625763e-01 5.03015518e-01
-1.49248874e+00 2.81772390e-02 -3.51638436e-01 -6.21379018e-01
3.33573878e-01 1.97783977e-01 -9.51449335e-01 9.46862400e-01
7.17439175e-01 2.15606943e-01 -6.98351681e-01 1.45532978e+00
-2.32905209e-01 -1.03090126e-02 -6.08282030e-01 -7.14045942e-01
4.16825742e-01 -6.35691881e-01 7.26019382e-01 9.82359290e-01
2.82109320e-01 1.48348317e-01 8.11799616e-02 9.48601663e-01
8.73337805e-01 4.08489674e-01 -8.21293712e-01 3.74485582e-01
2.60396689e-01 1.15553975e+00 -1.58614254e+00 -9.54565555e-02
2.70515651e-01 6.96798682e-01 5.78457341e-02 -1.46544233e-01
-8.99034560e-01 -7.68587947e-01 6.07743323e-01 6.19175732e-01
2.29520332e-02 -5.44779181e-01 -7.24599004e-01 -9.84726250e-01
-1.03270516e-01 -5.98965049e-01 1.77535400e-01 -8.85906875e-01
-4.16753620e-01 6.58648968e-01 4.41528052e-01 -1.08605409e+00
-4.31429237e-01 -9.13536608e-01 -8.84977221e-01 6.48098707e-01
-8.15236568e-01 -1.01627171e+00 -4.46323544e-01 7.53325403e-01
1.93931967e-01 -2.00056151e-01 7.56244004e-01 3.72418799e-02
-6.17513470e-02 -1.73021004e-01 -1.46757707e-01 -1.15111955e-02
-7.75262946e-03 -1.54197848e+00 5.40786088e-01 9.61975157e-01
6.80270568e-02 7.04074726e-02 1.06540132e+00 -7.80373514e-01
-9.94675398e-01 -5.40663540e-01 5.75387716e-01 -5.28555095e-01
7.11127937e-01 -7.06017911e-01 -6.11900628e-01 3.93586338e-01
-5.21380967e-03 -7.83572912e-01 -2.13085338e-02 1.15662046e-01
3.35340351e-01 -7.08419606e-02 -1.16167080e+00 9.30446684e-01
1.14551163e+00 -9.09288973e-03 -5.13609469e-01 7.42599368e-02
3.02442044e-01 -5.57561934e-01 -5.84193647e-01 2.02012151e-01
4.23583269e-01 -1.32986784e+00 8.13107371e-01 -2.81894743e-01
-1.16545409e-01 -8.51409495e-01 -1.01030096e-02 -1.21984494e+00
-1.17154233e-01 -6.34074867e-01 4.54681665e-01 7.97221005e-01
2.54700184e-01 -6.53332293e-01 9.49734867e-01 2.33044446e-01
-2.54950672e-01 -3.93300802e-01 -1.10299611e+00 -9.55027640e-01
-3.97723764e-02 -7.47234702e-01 1.13927639e+00 8.94664645e-01
6.38059914e-01 -2.26710707e-01 -7.75930882e-02 2.50992756e-02
4.00840908e-01 -3.39725763e-02 1.13410294e+00 -1.60606921e+00
-1.58366889e-01 -1.01088488e+00 -1.04541373e+00 -3.48880142e-01
-2.98794627e-01 -7.34318972e-01 -1.06454201e-01 -1.97080159e+00
-6.29554212e-01 -8.11139166e-01 4.90700871e-01 5.06532669e-01
5.88709712e-01 1.52662136e-02 1.77947998e-01 -1.17371809e-02
-3.73134762e-01 1.02693066e-01 1.01445484e+00 -1.27212601e-02
-5.19452393e-01 6.64379960e-03 -3.21840614e-01 7.87045062e-01
9.99642551e-01 -4.98590678e-01 -4.55140442e-01 7.66456034e-03
8.36425662e-01 -1.64590240e-01 5.96772134e-01 -1.40237987e+00
4.97976571e-01 -8.56354237e-01 -2.05787078e-01 -6.48300469e-01
4.31487530e-01 -9.11569118e-01 6.56124592e-01 6.75791204e-01
5.11518121e-01 4.42713052e-02 2.44881034e-01 -2.97555719e-02
-1.95872903e-01 -6.25630319e-01 4.18967634e-01 -4.03406650e-01
-1.28108954e+00 -3.43885680e-04 -8.24052572e-01 -2.05386162e-01
1.28857553e+00 -6.97560191e-01 -3.65285069e-01 -5.52376688e-01
-9.08464193e-01 4.35127407e-01 1.11369944e+00 -2.87531968e-02
4.47465181e-01 -1.00153673e+00 -5.29438436e-01 -3.82413566e-02
-1.90208131e-03 1.00864202e-01 6.52445778e-02 4.57961589e-01
-1.51183224e+00 5.74822202e-02 -9.92328405e-01 -4.52916175e-01
-1.05532181e+00 5.35851181e-01 5.52043378e-01 -3.62856686e-01
-4.41279799e-01 3.51062715e-01 -2.88937211e-01 -4.34395522e-01
-2.11284071e-01 -5.60192168e-01 -5.65731883e-01 1.45846352e-01
3.02922487e-01 5.79676569e-01 2.01891258e-01 -4.78984624e-01
-4.60331112e-01 7.17429698e-01 8.07267070e-01 -7.61597157e-01
1.36640954e+00 1.61020055e-01 -3.62326860e-01 3.84447306e-01
7.48017430e-02 3.37676018e-01 -7.58150339e-01 7.43682504e-01
1.09532125e-01 -4.62549835e-01 -4.88414839e-02 -6.79310560e-01
-5.62880576e-01 6.02420092e-01 3.15919518e-01 7.83488095e-01
1.00882220e+00 -5.49246147e-02 4.26977575e-02 3.89190257e-01
1.13355756e+00 -1.22374046e+00 -5.05753696e-01 6.88412726e-01
8.69692862e-01 -5.89888275e-01 -1.25628402e-02 -7.10475266e-01
-6.44290507e-01 1.22848344e+00 2.97757834e-01 -3.60780358e-01
5.07573247e-01 6.77417219e-01 -1.65659308e-01 -5.46902180e-01
-4.15376872e-01 -9.84194577e-01 -1.25868991e-01 1.11419964e+00
-3.22099268e-01 4.25790548e-01 -6.50021851e-01 6.75839007e-01
-9.23119605e-01 1.15013652e-01 9.49256063e-01 9.59957063e-01
-1.01311707e+00 -1.31765795e+00 -7.40113497e-01 1.72494397e-01
9.31166038e-02 1.78507194e-01 -8.79995644e-01 1.29138613e+00
5.42081416e-01 7.82769799e-01 1.95439458e-01 -6.53744102e-01
7.67953217e-01 -3.78876716e-01 6.25352800e-01 -5.29181004e-01
-6.26030803e-01 -3.59435767e-01 3.80628020e-01 -5.26104331e-01
8.07409268e-03 -4.96234566e-01 -1.43195176e+00 -9.05203760e-01
-3.64193581e-02 5.89666307e-01 9.56677258e-01 6.88608050e-01
-1.14298902e-01 2.99210757e-01 1.34856388e-01 -8.75704706e-01
2.93523639e-01 -4.89854604e-01 -9.26320374e-01 -5.44602945e-02
-2.59269983e-01 -1.09015036e+00 2.22616214e-02 -3.13781947e-01] | [3.4765446186065674, 1.4654655456542969] |
97fda81b-885c-4238-ab2c-c7c76d58e21b | weighted-histogram-equalization-using-entropy | 2111.08578 | null | https://arxiv.org/abs/2111.08578v3 | https://arxiv.org/pdf/2111.08578v3.pdf | Weighted Histogram Equalization Using Entropy of Probability Density Function | Low-contrast image enhancement is essential for high-quality image display and other visual applications. However, it is a challenging task as the enhancement is expected to increase the visibility of an image while maintaining its naturalness. In this paper, the weighted histogram equalization using the entropy of the probability density function is proposed. The computation of the local mapping functions utilizes the relationship between non-height bin and height bin distributions. Finally, the complete tone mapping function is produced by concatenating local mapping functions. Computer simulation results on the CSIQ dataset demonstrate that the proposed method produces images with higher visibility and visual quality, which outperforms traditional and recently proposed contrast enhancement algorithms methods in qualitative and quantitative metrics. | ['Sos Agaian', 'Thaweesak Trongtirakul'] | 2021-11-16 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 4.57154751e-01 -7.44831920e-01 3.24204803e-01 -2.75860995e-01
-5.09076536e-01 -2.14984179e-01 2.89922565e-01 3.00879866e-01
-4.47207183e-01 8.12942147e-01 1.05298236e-01 -2.34207794e-01
-4.02573287e-01 -9.33775544e-01 -2.50118285e-01 -1.05554783e+00
-3.92658621e-01 -6.48669422e-01 4.93153453e-01 -2.56508321e-01
5.68800867e-01 4.07653511e-01 -1.92756271e+00 -7.30306804e-02
1.32350874e+00 1.44179809e+00 5.11622429e-01 7.74323821e-01
2.94639438e-01 5.70937693e-01 -6.50968552e-01 -2.39868492e-01
3.69546235e-01 -5.17652571e-01 -2.44584098e-01 2.34118581e-01
2.91139066e-01 -4.62311745e-01 -6.04415759e-02 1.49085116e+00
6.08407438e-01 3.87335747e-01 5.32070339e-01 -9.78650749e-01
-6.50219917e-01 -2.54289865e-01 -7.63818800e-01 5.45533240e-01
4.16703224e-02 -1.13188080e-01 5.97214997e-01 -6.78254724e-01
3.02943349e-01 6.65768921e-01 2.87745357e-01 -4.50165421e-01
-1.01648569e+00 -6.64145410e-01 -4.87877935e-01 6.28466547e-01
-1.65261245e+00 -1.67017892e-01 6.80855572e-01 -1.70659631e-01
4.65574235e-01 4.73990411e-01 5.50770760e-01 -2.16050610e-01
6.47465825e-01 4.99563850e-02 1.70348918e+00 -6.50972545e-01
2.50814967e-02 1.77550897e-01 -3.33625674e-01 6.70688689e-01
1.41702652e-01 3.94506127e-01 -5.21972656e-01 1.41928628e-01
8.16433430e-01 -1.69003472e-01 -6.53568268e-01 -2.29023308e-01
-8.47643614e-01 5.75100124e-01 4.56217319e-01 3.44958484e-01
-5.42861879e-01 -2.45606184e-01 -4.41684313e-02 1.88288346e-01
3.67695063e-01 2.84109682e-01 2.51815498e-01 -1.34900108e-01
-1.07877314e+00 4.28813100e-02 1.22960284e-01 5.86376786e-01
7.54654586e-01 1.66196465e-01 -2.65494972e-01 7.81819522e-01
-1.98240038e-02 7.07328320e-01 7.04193041e-02 -1.00563014e+00
8.48752707e-02 1.49027586e-01 3.75376880e-01 -1.37449205e+00
-1.14392333e-01 -6.46252036e-01 -9.81092572e-01 8.63244176e-01
2.41499752e-01 1.49131536e-01 -7.07596242e-01 1.39553988e+00
1.27656490e-01 -9.21619087e-02 6.20871410e-02 9.99060810e-01
4.61879551e-01 9.72008288e-01 -1.82292033e-02 -5.19157827e-01
1.30919623e+00 -5.26347518e-01 -1.29254425e+00 1.50374338e-01
-3.23538959e-01 -1.21467507e+00 8.38799655e-01 5.61832070e-01
-1.46114564e+00 -8.00085723e-01 -1.41982520e+00 2.30523217e-02
-2.02685386e-01 2.44255394e-01 2.07584992e-01 7.78477788e-01
-1.07000172e+00 4.42539483e-01 -2.72912741e-01 6.56128451e-02
1.78005069e-01 1.00701787e-01 -9.42533240e-02 -1.35423899e-01
-1.10773396e+00 9.09860075e-01 3.19681585e-01 1.17763132e-01
-2.41693169e-01 -5.58219016e-01 -9.41684246e-01 3.75059992e-01
6.57507032e-02 -1.12297907e-02 8.47957492e-01 -4.53626812e-01
-1.41725791e+00 4.58700269e-01 -2.43352205e-01 -2.79939622e-01
1.93000615e-01 2.33736753e-01 -7.22795129e-01 5.59999347e-01
-1.89861983e-01 3.09010237e-01 9.53591526e-01 -1.30667949e+00
-9.57268119e-01 -7.93710053e-02 -2.28549093e-01 4.03187215e-01
-2.58190364e-01 1.48151955e-02 -3.59448314e-01 -5.63493967e-01
9.76131931e-02 -3.03346127e-01 1.76201448e-01 2.23054513e-01
-1.36968508e-01 4.86392379e-01 8.19820285e-01 -1.21091557e+00
1.43437839e+00 -2.43357873e+00 -4.17252928e-01 6.16035819e-01
2.59049445e-01 1.73087746e-01 1.56810984e-01 1.25095725e-01
1.85304165e-01 -2.76539832e-01 -3.65005702e-01 2.98182338e-01
-2.10462138e-01 -4.33328152e-01 9.71179605e-02 6.33327961e-01
-3.13384831e-02 4.00679052e-01 -5.91689706e-01 -6.67976081e-01
6.08760536e-01 7.17247427e-01 -2.51480490e-01 2.46859506e-01
5.51046073e-01 1.40097708e-01 1.52688429e-01 4.89224046e-01
1.26189351e+00 -8.50346498e-03 5.46661988e-02 -5.29985845e-01
-5.45504570e-01 -3.10789526e-01 -1.23053467e+00 8.17342460e-01
-4.48791981e-01 9.85866606e-01 1.21098503e-01 -5.32855451e-01
1.22134781e+00 1.35805234e-01 3.37730467e-01 -1.37011087e+00
3.69086832e-01 2.43773699e-01 4.38661575e-02 -2.69867003e-01
9.00383115e-01 -3.92664373e-01 2.04408422e-01 1.31975085e-01
-3.30845654e-01 -1.40342221e-01 3.08403105e-01 -1.59478188e-03
3.87202442e-01 -4.24913973e-01 7.27184236e-01 -2.88672328e-01
6.63177907e-01 -4.01977986e-01 3.27856809e-01 4.02364135e-01
-4.44656551e-01 6.48941755e-01 1.38471678e-01 2.73854196e-01
-1.28098333e+00 -1.51439834e+00 -5.42036295e-01 5.44377506e-01
7.47475684e-01 -4.03411724e-02 -6.11507356e-01 1.47483334e-01
-3.80124569e-01 6.23486757e-01 -3.07421833e-01 -2.51592129e-01
-2.53273010e-01 -5.56776166e-01 2.19264403e-02 2.71826416e-01
1.14409125e+00 -9.17995095e-01 -8.75114024e-01 3.11550479e-02
-4.43968683e-01 -1.16893828e+00 -5.11310875e-01 -6.74313009e-02
-4.02074635e-01 -8.24968994e-01 -9.32362556e-01 -9.85256672e-01
5.80732167e-01 6.57371342e-01 6.67864442e-01 2.28017226e-01
-4.34021503e-01 -3.85705680e-02 -3.39617789e-01 -8.33192021e-02
-2.78265059e-01 -6.86151683e-01 -3.93477917e-01 1.53344616e-01
-1.11718513e-01 -4.34314907e-01 -9.60709751e-01 4.24564630e-01
-1.00913405e+00 8.61357003e-02 7.38256931e-01 8.67119074e-01
8.32020402e-01 1.08289373e+00 3.60106289e-01 -1.84636995e-01
8.30305994e-01 4.59605567e-02 -9.11305845e-01 2.65084654e-01
-6.66103125e-01 -1.96134984e-01 5.35412312e-01 -1.82882115e-01
-1.69649100e+00 -4.31194872e-01 1.15006439e-01 1.13604546e-01
2.18551919e-01 4.02125984e-01 -1.88413724e-01 -4.79712069e-01
4.17935014e-01 6.37691438e-01 2.12685987e-01 -1.80979624e-01
8.41843486e-02 6.81734741e-01 9.47030425e-01 5.05069867e-02
8.63481879e-01 5.77933371e-01 2.64692485e-01 -1.13076186e+00
-1.14796765e-01 -5.12012780e-01 -1.21520996e-01 -5.73551774e-01
9.27989602e-01 -4.95990157e-01 -8.86612654e-01 5.84717691e-01
-7.69230843e-01 -1.30740386e-02 7.16111064e-02 6.55529678e-01
-3.33199680e-01 5.07288754e-01 -4.37874675e-01 -9.19128835e-01
-2.91571021e-01 -1.16703105e+00 6.29617453e-01 7.65712798e-01
4.64095801e-01 -8.79858375e-01 -2.15431362e-01 -4.05657031e-02
8.94723296e-01 2.12487951e-01 8.36414158e-01 5.18461466e-01
-6.15219235e-01 -5.58106713e-02 -7.36847103e-01 6.08478725e-01
2.83240050e-01 -1.59700394e-01 -7.58281469e-01 -3.04784864e-01
-4.91036810e-02 1.13246158e-01 4.87818778e-01 7.35562444e-01
1.24670255e+00 2.63822116e-02 1.64655298e-01 6.98643327e-01
1.90395367e+00 6.25665963e-01 1.23334384e+00 4.70521450e-01
-9.28128585e-02 5.37398279e-01 9.85727370e-01 6.25732005e-01
-9.63593647e-02 6.39081478e-01 3.29586297e-01 -5.62028944e-01
-2.21420035e-01 4.19372618e-02 -1.86141878e-02 5.41159332e-01
-5.54649457e-02 -3.19988251e-01 -3.15935284e-01 5.30699015e-01
-1.16524339e+00 -1.15763402e+00 -1.39435455e-01 2.42195940e+00
8.02168489e-01 1.00529492e-01 -3.56459498e-01 4.00209963e-01
9.70499873e-01 1.00121923e-01 -1.77989155e-01 -5.33608675e-01
-4.29700822e-01 4.99376386e-01 6.22707546e-01 8.19774687e-01
-9.72729683e-01 2.64323294e-01 6.40801716e+00 1.05493867e+00
-1.19654846e+00 -3.90171297e-02 5.18154442e-01 2.94629037e-01
-1.78405076e-01 -1.37242645e-01 -3.41558278e-01 7.07875907e-01
4.59202379e-01 -5.09460092e-01 4.87483650e-01 1.98581412e-01
5.30801177e-01 -7.90795088e-01 -7.78125180e-03 1.10889816e+00
1.59497187e-01 -9.92919147e-01 -1.37138277e-01 1.04832768e-01
6.68815911e-01 -7.12287068e-01 5.40264487e-01 -2.16407999e-01
-3.10612172e-01 -8.79827857e-01 5.66419423e-01 7.10406125e-01
9.83191013e-01 -1.01139569e+00 8.10891986e-01 -2.22627353e-02
-1.46994972e+00 -4.85707354e-03 -1.71630576e-01 1.39454097e-01
5.14066160e-01 6.55013680e-01 -5.45181930e-01 5.81890404e-01
8.46476972e-01 7.27677047e-02 -5.78850091e-01 1.65747607e+00
-8.84273201e-02 2.75584072e-01 -6.44479990e-02 2.20136285e-01
-6.57668188e-02 -5.81019759e-01 5.24944127e-01 1.11694753e+00
6.90356016e-01 1.86862215e-01 -2.23011941e-01 7.40727067e-01
1.49800867e-01 2.41259754e-01 -1.52297437e-01 2.74541587e-01
4.62611139e-01 1.28631783e+00 -6.15287423e-01 -1.79913446e-01
-3.18895191e-01 9.77140367e-01 -3.67363751e-01 4.92332578e-01
-9.00664926e-01 -1.11651123e+00 2.76885062e-01 1.34347752e-01
3.71941984e-01 -2.43690819e-01 -3.29019517e-01 -4.12938923e-01
1.19008064e-01 -7.45117605e-01 1.41671613e-01 -1.14497578e+00
-7.31506407e-01 6.36548519e-01 3.87683436e-02 -1.55926716e+00
1.10860057e-01 -4.45720285e-01 -5.93329728e-01 1.20057428e+00
-1.91119957e+00 -7.39056706e-01 -7.96339989e-01 5.64463019e-01
8.42667744e-02 7.61281475e-02 3.73590291e-01 6.80254400e-01
-1.61715392e-02 5.61889887e-01 3.88954937e-01 -2.53241211e-01
7.44301260e-01 -1.14632940e+00 -4.50861722e-01 1.28312707e+00
-5.09355426e-01 2.37636432e-01 1.00032699e+00 -5.23181677e-01
-5.89974344e-01 -6.63628638e-01 7.74119735e-01 7.47340739e-01
1.77990481e-01 6.36206791e-02 -1.04268992e+00 -9.11790356e-02
6.08523786e-01 -4.75714505e-02 5.05955458e-01 -6.56125963e-01
2.01169059e-01 -3.56818110e-01 -1.48923814e+00 4.26501662e-01
3.48301440e-01 -4.61688727e-01 -3.04809790e-02 -2.66932994e-01
3.56746107e-01 -3.60325783e-01 -9.82827663e-01 3.62916261e-01
6.34808540e-01 -1.26090884e+00 9.82804000e-01 3.71545136e-01
3.68145049e-01 -7.88834155e-01 -4.96872216e-01 -1.35230482e+00
-4.08381909e-01 -4.13949043e-01 2.81537533e-01 1.12288380e+00
2.42464006e-01 -5.97850144e-01 2.81249613e-01 1.02099061e-01
2.12678667e-02 -4.72983211e-01 -6.46306396e-01 -8.11677992e-01
-5.58777332e-01 1.19555015e-02 4.43356723e-01 5.03367662e-01
-1.41579896e-01 -1.77220151e-01 -5.49712718e-01 5.39392352e-01
1.09980309e+00 1.67032152e-01 3.19426924e-01 -7.38873541e-01
-2.24487692e-01 -3.91763866e-01 -6.91951454e-01 -8.46193373e-01
-5.49740076e-01 -3.72643083e-01 2.10839033e-01 -1.58220851e+00
3.35654110e-01 -2.92739987e-01 -5.03612041e-01 -1.77687570e-01
-2.77600914e-01 8.40990961e-01 8.99231210e-02 -5.03476188e-02
-3.24682713e-01 7.11772859e-01 1.62165904e+00 -8.07832554e-02
-1.78549737e-01 -7.95781612e-02 -5.48713088e-01 3.38008404e-01
8.77345681e-01 -8.07689503e-02 -4.33572739e-01 1.84955914e-02
3.80670018e-02 7.41884410e-02 4.09119546e-01 -1.17243850e+00
3.03768396e-01 -1.64991111e-01 5.26780069e-01 -7.64813483e-01
5.67488611e-01 -8.62226069e-01 -2.04967316e-02 3.15086365e-01
-2.35811789e-02 3.13865840e-02 1.66578650e-01 4.03863996e-01
-7.26059973e-01 -1.09670043e-01 1.30620873e+00 3.85372728e-01
-9.31610465e-01 -4.34452295e-02 -3.78158450e-01 -3.08680654e-01
1.11666846e+00 -4.91735101e-01 -4.25810009e-01 -7.60563195e-01
-3.89599264e-01 -1.93455711e-01 4.38331455e-01 -5.80566861e-02
1.06842041e+00 -1.36555064e+00 -6.26860082e-01 3.04574549e-01
1.22284099e-01 -6.81928456e-01 6.68981850e-01 1.10330200e+00
-7.95427382e-01 5.86934835e-02 -7.91885316e-01 -4.73043680e-01
-1.54310346e+00 3.06139588e-01 3.58238190e-01 -7.08289966e-02
-3.69217485e-01 3.08053851e-01 1.17113113e-01 2.89947540e-01
-3.20188212e-03 -1.38863802e-01 -4.76839215e-01 -4.17895317e-01
9.24694777e-01 7.99173295e-01 9.97104794e-02 -7.11879492e-01
-4.08982346e-03 6.27993047e-01 2.48320460e-01 -4.45723832e-01
1.05162561e+00 -6.21229827e-01 -1.43493250e-01 -1.49686724e-01
1.40264690e+00 5.08832932e-01 -1.15507734e+00 -9.11569446e-02
-5.18076718e-01 -1.17534101e+00 5.46002269e-01 -9.34000969e-01
-9.78244483e-01 8.75638962e-01 1.18703771e+00 3.44197601e-01
1.89708650e+00 -4.54916954e-01 7.80459523e-01 -3.46613705e-01
3.87556553e-02 -1.16874874e+00 1.49116352e-01 -6.71845600e-02
6.80937529e-01 -9.84474421e-01 3.08671623e-01 -6.76000118e-01
-5.36627948e-01 8.99640381e-01 3.05475265e-01 1.90095365e-01
6.57737017e-01 3.69760334e-01 1.39003783e-01 3.15476991e-02
-8.99936408e-02 -4.69598472e-01 4.59467858e-01 7.03129292e-01
2.19911709e-01 5.98608777e-02 -6.69084251e-01 1.78703919e-01
-1.27322406e-01 -2.57469773e-01 6.14693344e-01 9.21469867e-01
-8.79666150e-01 -7.43976712e-01 -7.02167809e-01 3.76431972e-01
-6.06897831e-01 -2.21152708e-01 2.40229070e-01 6.66966796e-01
2.89627969e-01 1.29426837e+00 1.94033355e-01 -3.35583836e-01
2.69291520e-01 -5.96097946e-01 5.42097211e-01 1.94317281e-01
-1.10829249e-01 2.83494145e-01 -6.18239418e-02 -2.66910940e-01
-4.52347517e-01 -2.90479839e-01 -1.29982173e+00 -4.74611819e-01
-5.00108898e-01 1.77451804e-01 8.25357854e-01 5.08527875e-01
5.76712377e-02 6.58679426e-01 9.44185317e-01 -5.33497572e-01
-1.32781059e-01 -6.90532744e-01 -1.26139092e+00 4.98842657e-01
4.24236149e-01 -7.61109233e-01 -5.46930850e-01 1.60697103e-01] | [11.041389465332031, -2.3362462520599365] |
eb184e6c-16c0-4e8a-906d-874e62aec354 | vihealthbert-pre-trained-language-models-for | null | null | http://www.lrec-conf.org/proceedings/lrec2022/pdf/2022.lrec-1.35.pdf | http://www.lrec-conf.org/proceedings/lrec2022/pdf/2022.lrec-1.35.pdf | ViHealthBERT: Pre-trained Language Models for Vietnamese in Health Text Mining | Pre-trained language models have become crucial to achieving competitive results across many Natural Language Processing (NLP) problems. For monolingual pre-trained models in low-resource languages, the quantity has been significantly increased. However, most of them relate to the general domain, and there are limited strong baseline language models for domain-specific. We introduce ViHealthBERT, the first domain-specific pre-trained language model for Vietnamese healthcare. The performance of our model shows strong results while outperforming the general domain language models in all health-related datasets. Moreover, we also present Vietnamese datasets for the healthcare domain for two tasks are Acronym Disambiguation (AD) and Frequently Asked Questions (FAQ) Summarization. We release our ViHealthBERT to facilitate future research and downstream application for Vietnamese NLP in domain-specific. | ['Steven Quoc Hung', 'Trung Huu and Truong', 'Huy Duc and Bui', 'Vu and Ta', 'Vu Hoang and Hoang', 'Nguyen and Tran', 'Minh'] | 2022-06-01 | null | null | null | lrec-2022-6 | ['word-sense-disambiguation', 'medical-named-entity-recognition', 'named-entity-recognition-in-vietnamese', 'vietnamese-datasets'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 1.32171556e-01 3.24008435e-01 -6.46611691e-01 -3.79180819e-01
-1.55956781e+00 -4.92849469e-01 4.49778914e-01 7.78914690e-01
-9.01726663e-01 1.11064208e+00 1.10058737e+00 -3.47107649e-01
-1.56771746e-02 -4.04766589e-01 -2.03520164e-01 -2.83213794e-01
1.78587154e-01 1.02332509e+00 -3.02014738e-01 -6.46819353e-01
-2.15241775e-01 1.73420697e-01 -2.26186454e-01 7.18362868e-01
1.20058417e+00 5.45070134e-02 2.92498648e-01 5.81106007e-01
-2.77814269e-01 9.61741805e-01 -6.31674647e-01 -5.19104004e-01
-2.67586440e-01 -4.58623320e-01 -1.25864840e+00 -4.16289151e-01
3.05150717e-01 -1.59313437e-02 -1.57478765e-01 6.98285520e-01
9.37676907e-01 -1.37257919e-01 7.98396945e-01 -7.57426381e-01
-7.26401925e-01 9.90139425e-01 -3.87644678e-01 2.98302650e-01
7.12293684e-01 4.56905663e-02 1.15605330e+00 -6.96454287e-01
1.25017917e+00 1.48168242e+00 6.57413185e-01 7.33668804e-01
-7.86365271e-01 -6.24671698e-01 1.32341221e-01 1.10536128e-01
-1.39281583e+00 -4.54620957e-01 3.75968695e-01 -2.83930242e-01
1.73791754e+00 7.81689361e-02 1.96910143e-01 1.49979675e+00
4.63180304e-01 1.09582973e+00 7.13090360e-01 -2.78569460e-01
-1.54632136e-01 5.44318259e-02 2.34113246e-01 6.45580351e-01
3.32925558e-01 -3.87264699e-01 -3.78192216e-01 -4.87965882e-01
1.81458220e-01 -2.78433412e-01 -2.86207706e-01 4.36889380e-01
-1.53981721e+00 1.15636241e+00 2.94961095e-01 4.88142937e-01
-7.32883811e-01 -3.14372212e-01 7.51081824e-01 3.55055273e-01
5.95356822e-01 1.00439012e+00 -1.03614855e+00 1.82927493e-02
-8.80792320e-01 3.46883297e-01 1.13693643e+00 1.11637056e+00
2.41618395e-01 -1.88658729e-01 -5.95161796e-01 1.22518837e+00
-7.36835375e-02 7.76398003e-01 4.43360686e-01 -8.44888687e-01
8.38270664e-01 4.02890652e-01 -2.78493762e-01 -9.71232355e-01
-9.72140968e-01 -3.62967134e-01 -1.08376110e+00 -1.16637838e+00
2.48834819e-01 -4.84888583e-01 -6.81237638e-01 1.70190251e+00
-6.08629435e-02 -5.17749906e-01 7.55898058e-01 5.12834549e-01
1.59057760e+00 9.34791684e-01 6.22029722e-01 -5.59535503e-01
1.90913773e+00 -1.00341725e+00 -1.15062404e+00 -6.21582031e-01
8.89562786e-01 -9.55213130e-01 9.70118165e-01 3.60910416e-01
-8.70893061e-01 -1.02070034e-01 -4.73026335e-01 -7.29656339e-01
-4.96143192e-01 1.85517997e-01 5.34095109e-01 2.64269441e-01
-8.66364062e-01 -2.03562900e-01 -5.50718844e-01 -9.50349510e-01
2.39840314e-01 -1.06462143e-01 -2.88040221e-01 -3.94345015e-01
-1.73714352e+00 1.15496886e+00 6.52668238e-01 -3.52170527e-01
-7.63337612e-01 -9.81260955e-01 -1.18829286e+00 -2.19113156e-01
4.22149599e-01 -1.23912108e+00 1.40882742e+00 -3.75229508e-01
-1.01836455e+00 1.26236951e+00 -3.97921175e-01 -6.70161009e-01
2.98747599e-01 -3.92397970e-01 -8.41117680e-01 3.59655052e-01
5.95897198e-01 7.88577437e-01 1.43820122e-01 -7.92766988e-01
-5.15177071e-01 -4.62527387e-02 -2.45554045e-01 2.79926062e-01
-1.27974764e-01 4.61394280e-01 -3.61248434e-01 -8.47933412e-01
-6.55442417e-01 -5.90442181e-01 -4.35880423e-01 -5.97945690e-01
-4.84317869e-01 -6.89913511e-01 1.51372209e-01 -1.24228787e+00
1.74108636e+00 -1.83631289e+00 -1.47257820e-01 -3.69014800e-01
5.72696589e-02 3.11090469e-01 -6.17233276e-01 1.06292379e+00
3.64889354e-02 2.32303008e-01 -5.22152066e-01 -5.67143559e-02
-1.85226530e-01 5.79247653e-01 -4.85246092e-01 1.06596254e-01
5.60734391e-01 1.21828437e+00 -1.28658736e+00 -8.89339328e-01
-1.97405249e-01 2.83203542e-01 -6.27732337e-01 8.47472250e-02
-3.44938636e-01 3.34174842e-01 -5.55691957e-01 6.32394969e-01
3.88810992e-01 -1.19995721e-01 4.89216626e-01 -1.03057303e-01
-6.49395362e-02 8.61265898e-01 -2.22587630e-01 1.99154389e+00
-6.50929093e-01 1.07106335e-01 7.07176775e-02 -7.93763578e-01
4.57805872e-01 4.63454694e-01 6.19324446e-01 -8.10061038e-01
-2.21964866e-02 3.12620938e-01 8.93718079e-02 -9.50998306e-01
6.30254030e-01 -1.37892768e-01 -6.37732983e-01 1.56742096e-01
9.44942236e-02 -3.90165448e-01 3.66071016e-01 6.47333801e-01
8.49782526e-01 -3.49029690e-01 1.25723791e+00 -6.11133099e-01
6.47712171e-01 7.16660976e-01 6.67099476e-01 7.31698155e-01
-2.08643749e-01 6.85535014e-01 4.54870254e-01 -2.58453965e-01
-5.84517777e-01 -8.85039449e-01 -3.96753460e-01 1.11369824e+00
-4.38284665e-01 -7.65073001e-01 -5.60236692e-01 -8.85047257e-01
-4.22885679e-02 1.09684479e+00 -1.63000181e-01 8.77767801e-02
-1.08948171e+00 -9.36506271e-01 1.01080263e+00 2.39006475e-01
4.54160810e-01 -1.23132920e+00 -1.41815647e-01 5.34008622e-01
-7.85371423e-01 -1.60579669e+00 -8.61961842e-01 -4.03974652e-02
-5.88184178e-01 -1.09407139e+00 -1.02935338e+00 -1.19805062e+00
1.87049702e-01 -1.57298192e-01 1.58951163e+00 -2.76199102e-01
-2.42611781e-01 5.90813160e-01 -3.14213842e-01 -8.59672844e-01
-6.44422889e-01 5.14604390e-01 -1.19167482e-02 -7.69419789e-01
6.72539055e-01 1.83895260e-01 -3.68005782e-01 -3.66841525e-01
-7.89654732e-01 -9.74285901e-02 6.77481949e-01 8.81278813e-01
6.01357341e-01 -3.98002952e-01 9.91426706e-01 -1.42673588e+00
1.18925130e+00 -8.46152067e-01 1.73196584e-01 4.98449624e-01
-3.78994673e-01 4.80717756e-02 7.29521096e-01 -3.07519406e-01
-1.01322269e+00 -2.67130971e-01 -6.28558993e-01 6.57973647e-01
-1.98878661e-01 1.07030189e+00 -6.11266568e-02 7.30085850e-01
7.86317527e-01 1.24403894e-01 -2.07098588e-01 -6.50513351e-01
5.54987013e-01 7.91973531e-01 5.64768374e-01 -4.56219047e-01
2.60198623e-01 8.75386149e-02 -5.07645130e-01 -1.34372365e+00
-1.09825695e+00 -7.99542665e-01 -3.90917242e-01 5.96518576e-01
1.34379351e+00 -1.19748855e+00 -4.35990810e-01 3.59398313e-02
-1.49963379e+00 -2.60484010e-01 -1.56691119e-01 4.57005769e-01
-8.83229729e-03 5.44625521e-01 -8.70292783e-01 -4.33863640e-01
-8.52634370e-01 -9.15347219e-01 1.30548477e+00 -1.17162988e-01
-8.09798896e-01 -1.32606840e+00 3.71936977e-01 4.95702535e-01
2.75720805e-01 1.23146169e-01 1.59542310e+00 -1.04653037e+00
3.34226370e-01 5.93956932e-02 -2.17263296e-01 1.75594334e-02
2.98548758e-01 -5.58014035e-01 -4.24360603e-01 -1.65381312e-01
1.73579026e-02 -5.22310972e-01 1.12916875e+00 5.75093091e-01
8.56488287e-01 -8.05284679e-01 -4.54188973e-01 2.71380067e-01
1.05619216e+00 -1.96379442e-02 1.44379139e-01 2.67560661e-01
5.29545307e-01 7.05780864e-01 5.73493600e-01 2.28707835e-01
1.17573774e+00 3.56247634e-01 -4.57822919e-01 -5.24286687e-01
-2.77029276e-01 -3.39878470e-01 3.67420614e-01 1.43420136e+00
4.48225260e-01 -5.27935863e-01 -1.37809896e+00 8.40010703e-01
-1.67971265e+00 -5.70365906e-01 -1.91569030e-01 1.52162409e+00
1.43070304e+00 -4.71913248e-01 1.89872429e-01 -7.71657050e-01
3.02505642e-01 2.46017411e-01 -3.48214775e-01 -7.95837224e-01
-4.74625349e-01 3.56685847e-01 2.89386570e-01 6.39857829e-01
-1.31659138e+00 1.27117240e+00 6.74515009e+00 9.33552444e-01
-7.74359763e-01 2.84818619e-01 4.89297003e-01 2.38387346e-01
-3.32128793e-01 -3.34055305e-01 -8.57509851e-01 5.45276813e-02
9.82290804e-01 -4.50802386e-01 2.97009908e-02 5.95702589e-01
2.76697516e-01 5.47143146e-02 -1.01527166e+00 1.04388380e+00
2.88548946e-01 -1.21695733e+00 4.13166046e-01 -1.95901603e-01
8.12514961e-01 3.44136149e-01 -1.76012188e-01 7.17339456e-01
3.67806584e-01 -1.13343620e+00 -7.88281485e-02 2.74705946e-01
8.87866557e-01 -7.45319247e-01 1.03036511e+00 5.83978772e-01
-8.62180710e-01 1.21788993e-01 -4.62911814e-01 3.47055107e-01
5.32025874e-01 5.88621974e-01 -1.24819052e+00 8.73672128e-01
3.59262317e-01 7.94757724e-01 -4.27460790e-01 5.76995373e-01
-3.50659758e-01 6.96421504e-01 -6.53394833e-02 -1.15207873e-01
5.42848766e-01 1.03329331e-01 5.50064921e-01 1.94741118e+00
1.53757418e-02 4.98395145e-01 5.71911275e-01 4.54641730e-01
-3.23679805e-01 7.44268775e-01 -6.60388470e-01 -5.06052196e-01
2.83724576e-01 1.00528383e+00 -2.42150545e-01 -6.22975469e-01
-5.58930457e-01 9.43244398e-01 3.37805152e-01 3.92210126e-01
-4.17461544e-01 -2.85038054e-01 6.09942913e-01 -1.74443215e-01
-4.30053353e-01 -2.66613692e-01 -1.55546501e-01 -1.40187979e+00
-4.95721161e-01 -1.52720511e+00 1.18198526e+00 -2.79483855e-01
-1.86233234e+00 7.07973063e-01 2.01861843e-01 -8.50077569e-01
-5.77530205e-01 -5.07620692e-01 -1.89615283e-02 7.33644187e-01
-1.79574859e+00 -1.20321846e+00 3.47683668e-01 5.99546432e-01
9.41359580e-01 -3.39150339e-01 1.20533192e+00 4.21167523e-01
-4.73802328e-01 6.66424215e-01 -1.15146421e-01 4.05622244e-01
1.31900644e+00 -1.13326919e+00 3.28129143e-01 5.89945018e-01
1.23062385e-02 9.42641497e-01 5.26101708e-01 -1.09889638e+00
-1.29831350e+00 -1.26564443e+00 2.08077097e+00 -7.60622919e-01
6.46333873e-01 -3.44680429e-01 -8.64676178e-01 6.49991989e-01
8.58565211e-01 -8.03159297e-01 1.01483643e+00 2.32240021e-01
-7.60619342e-02 1.91559643e-01 -1.35224628e+00 5.38107216e-01
9.98737037e-01 -6.04161322e-01 -1.13681352e+00 8.84606481e-01
1.08814371e+00 -4.00013298e-01 -9.59397435e-01 4.71736044e-01
1.39566064e-01 1.30750909e-01 1.16406071e+00 -1.13179255e+00
5.31285703e-01 8.59881565e-02 8.76006559e-02 -1.33210635e+00
-4.36279237e-01 -4.64765131e-01 2.17497900e-01 1.13079381e+00
7.99127758e-01 -7.59893417e-01 1.06420722e-02 3.26721460e-01
-1.90314725e-01 -4.85209376e-01 -6.79016531e-01 -6.37152195e-01
6.96731269e-01 -3.82246584e-01 3.62087518e-01 1.39274251e+00
1.72231898e-01 9.26743567e-01 -3.35198641e-01 -3.22375484e-02
1.75825194e-01 -3.86966318e-02 3.56922835e-01 -8.08030427e-01
-1.41038090e-01 -3.63188446e-01 3.00745159e-01 -9.09716249e-01
5.14526188e-01 -1.45881712e+00 -8.34318027e-02 -2.14458036e+00
5.16441643e-01 -3.17162834e-03 7.96622932e-02 9.00394738e-01
-4.67319131e-01 -2.10230947e-01 -6.97853044e-02 1.53154209e-02
-5.94015241e-01 3.38581532e-01 1.17597699e+00 -4.02392030e-01
-2.92299837e-01 -3.45405430e-01 -1.13224435e+00 5.17440915e-01
7.29799151e-01 -6.61122322e-01 -2.38990739e-01 -8.75211477e-01
1.68781236e-01 2.62317032e-01 -2.68803656e-01 -2.25574195e-01
2.25651905e-01 -3.95141780e-01 1.02884531e-01 -5.91829062e-01
-2.23197058e-01 -3.81271780e-01 -3.67512971e-01 8.48195314e-01
-6.63423777e-01 4.42661017e-01 3.41369182e-01 9.50645357e-02
-3.43875736e-01 -3.07895746e-02 4.91848528e-01 -4.65380669e-01
-6.98009253e-01 3.64298016e-01 -7.88058281e-01 9.40892994e-01
4.18396711e-01 5.61933398e-01 -5.47874272e-01 -4.13027823e-01
-4.67542410e-01 8.76635730e-01 -8.56968760e-02 5.79904199e-01
5.38119256e-01 -9.39476907e-01 -1.39046061e+00 -1.72059789e-01
3.77194315e-01 -5.13959602e-02 2.68559992e-01 8.66821468e-01
-7.31181264e-01 1.20196164e+00 2.84843773e-01 -1.95135921e-01
-1.01844227e+00 8.06661189e-01 -3.32750529e-02 -9.13398921e-01
-6.63038552e-01 7.30187237e-01 3.58229131e-01 -8.45437288e-01
1.91261023e-01 -5.34508646e-01 -4.83068347e-01 2.33809575e-01
7.77649641e-01 3.15476954e-01 2.24742889e-02 -6.48848951e-01
-8.35914135e-01 2.39215508e-01 -2.00495347e-01 1.51440520e-02
1.36477208e+00 -1.52198046e-01 -3.89082193e-01 2.17282102e-01
1.31808746e+00 2.25271747e-01 8.08935687e-02 -4.13454980e-01
4.47563350e-01 3.75315040e-01 -2.65814453e-01 -1.38182342e+00
-5.43177903e-01 7.66827106e-01 -2.50048023e-02 -3.48526359e-01
1.09959245e+00 2.63545036e-01 1.31277442e+00 8.06022942e-01
2.34893933e-01 -1.27264619e+00 -1.75498620e-01 9.96927619e-01
1.23306346e+00 -1.28206933e+00 1.01427741e-01 -4.06780958e-01
-1.22683156e+00 1.00018156e+00 3.03576201e-01 1.71038419e-01
4.92554784e-01 3.04988533e-01 6.27768099e-01 -4.09053355e-01
-7.50171781e-01 -2.54821569e-01 6.60320938e-01 6.12306654e-01
9.74541903e-01 2.67235845e-01 -1.09483302e+00 1.03882241e+00
-6.07676983e-01 -1.66572213e-01 2.39377677e-01 5.92542231e-01
6.47693276e-02 -1.31998038e+00 1.54684316e-02 3.98697346e-01
-9.13558424e-01 -7.24844992e-01 -6.90542579e-01 9.32587683e-01
-1.07550837e-01 1.13489628e+00 -4.20979232e-01 2.72988766e-01
5.53710639e-01 8.39560702e-02 2.60493994e-01 -1.01672292e+00
-9.69571471e-01 1.71613663e-01 6.91626787e-01 -4.86689568e-01
-3.54459256e-01 -5.02294540e-01 -1.43182862e+00 -1.21270437e-02
3.72039378e-01 2.76983291e-01 1.36116877e-01 7.49838769e-01
4.04095024e-01 4.21444178e-01 -3.58303301e-02 -1.80120189e-02
-4.04931545e-01 -9.61184978e-01 -1.69731289e-01 3.27847153e-01
4.42157209e-01 3.43654677e-02 4.95388806e-02 -2.02852219e-01] | [8.835389137268066, 8.899136543273926] |
2a3046bd-acd5-44f9-8d23-f7ac179d8285 | syntax-ignorant-n-gram-embeddings-for | null | null | https://aclanthology.org/W19-4604 | https://aclanthology.org/W19-4604.pdf | Syntax-Ignorant N-gram Embeddings for Sentiment Analysis of Arabic Dialects | Arabic sentiment analysis models have employed compositional embedding features to represent the Arabic dialectal content. These embeddings are usually composed via ordered, syntax-aware composition functions and learned within deep neural frameworks. With the free word order and the varying syntax nature across the different Arabic dialects, a sentiment analysis system developed for one dialect might not be efficient for the others. Here we present syntax-ignorant n-gram embeddings to be used in sentiment analysis of several Arabic dialects. The proposed embeddings were composed and learned using an unordered composition function and a shallow neural model. Five datasets of different dialects were used to evaluate the produced embeddings in the sentiment analysis task. The obtained results revealed that, our syntax-ignorant embeddings could outperform word2vec model and doc2vec both variant models in addition to hand-crafted system baselines, while a competent performance was noticed towards baseline systems that adopted more complicated neural architectures. | ['Ismail Babao{\\u{g}}lu', 'Mourad Gridach', 'Hatem Haddad', 'Hala Mulki'] | 2019-08-01 | null | null | null | ws-2019-8 | ['arabic-sentiment-analysis'] | ['natural-language-processing'] | [-4.34347719e-01 -1.51850179e-01 4.27480012e-01 -7.02786863e-01
-1.48977980e-01 -9.73019719e-01 9.38507736e-01 3.21154833e-01
-6.46897554e-01 2.06599295e-01 6.06442094e-01 -3.38494003e-01
2.14226097e-01 -8.76825929e-01 -2.68286735e-01 -7.45909572e-01
-1.23869166e-01 3.72835368e-01 -2.59854525e-01 -1.23606682e+00
3.68752897e-01 1.87922329e-01 -1.44065309e+00 3.54600310e-01
8.65224659e-01 8.03928971e-01 -4.69182283e-02 7.44680524e-01
-4.82563585e-01 5.39024234e-01 -6.27517223e-01 -8.72655809e-01
2.16224506e-01 -2.46095613e-01 -4.53440398e-01 -1.55453831e-01
3.30746025e-01 -1.42724365e-01 -1.12988591e-01 9.02242720e-01
4.18978751e-01 -5.93110435e-02 9.69996691e-01 -7.42463529e-01
-1.59867978e+00 9.30820704e-01 -1.18771382e-01 -6.54255897e-02
2.76062846e-01 -9.70274210e-02 1.32644033e+00 -1.18492627e+00
3.49435478e-01 1.29119849e+00 6.89432681e-01 3.61872196e-01
-8.69606435e-01 -2.99307644e-01 2.96862185e-01 1.44571707e-01
-1.04974627e+00 -8.77701715e-02 7.51432836e-01 -3.38058949e-01
1.10454750e+00 4.08173017e-02 5.60406327e-01 1.10513508e+00
3.60649109e-01 6.27438188e-01 1.15890336e+00 -6.08884871e-01
8.46752489e-04 6.45984948e-01 6.34806514e-01 7.85483479e-01
2.29308024e-01 -4.53389019e-01 -3.44287664e-01 2.40811203e-02
-1.74258780e-02 -8.48074034e-02 -7.10531473e-02 -2.93848634e-01
-9.95750904e-01 1.19584692e+00 6.67260349e-01 4.40988392e-01
-3.53124261e-01 -1.22956492e-01 8.73105645e-01 6.63393199e-01
1.64673984e-01 6.17078722e-01 -8.43273997e-01 1.17333964e-01
-4.05231446e-01 2.21725091e-01 7.74520040e-01 7.65093565e-01
6.39934242e-01 5.26823223e-01 1.26587078e-01 9.80586410e-01
5.99861801e-01 6.68773353e-01 1.13468301e+00 2.20451895e-02
2.83943087e-01 1.09452105e+00 -4.25337106e-02 -1.20591164e+00
-4.47483063e-01 -1.53546613e-02 -5.26161969e-01 2.37829700e-01
5.17169118e-01 -5.50136447e-01 -1.02407634e+00 1.39299381e+00
6.41845018e-02 -7.73387849e-01 4.09594268e-01 9.40305412e-01
7.24378884e-01 8.83997142e-01 -5.75249456e-02 4.92616445e-01
1.73452663e+00 -1.12311399e+00 -8.48559201e-01 -1.21199057e-01
7.55102098e-01 -1.10354352e+00 1.56376660e+00 6.07871652e-01
-7.71876693e-01 -6.17952943e-01 -1.50410688e+00 -1.86154321e-01
-1.19958305e+00 3.30654442e-01 7.02215612e-01 1.31886458e+00
-1.17702484e+00 2.55538911e-01 -5.08471131e-01 -4.98225063e-01
5.48279583e-02 2.86951393e-01 -6.27339125e-01 5.97285517e-02
-1.37128854e+00 1.26858044e+00 2.93524235e-01 5.60122252e-01
-8.36649656e-01 -1.94054276e-01 -1.17442322e+00 -9.39667318e-03
-3.61401141e-01 -5.36330007e-02 8.08596194e-01 -1.49949503e+00
-1.58685160e+00 7.45595217e-01 1.64740503e-01 -2.85799414e-01
-6.92553744e-02 -4.58125293e-01 -7.05554724e-01 -2.03323871e-01
-2.69974917e-01 1.85714111e-01 8.08908761e-01 -1.16520274e+00
-4.35559660e-01 -3.98895264e-01 2.63079286e-01 3.69236469e-01
-9.04125810e-01 7.98024908e-02 7.89224058e-02 -8.50073814e-01
2.38234065e-02 -9.97132301e-01 -1.22830614e-01 -4.05365974e-01
-1.01378947e-01 -1.94214210e-02 6.90708995e-01 -7.88782477e-01
1.29297614e+00 -1.97093797e+00 3.34508747e-01 2.10448831e-01
-1.53805301e-01 3.84331673e-01 -3.92398030e-01 8.13294351e-01
-4.59841862e-02 8.74511227e-02 -2.11130574e-01 -1.49986669e-01
4.40517306e-01 7.91980103e-02 -2.70027399e-01 4.93991703e-01
5.64420760e-01 6.60162091e-01 -7.20323145e-01 5.86071936e-03
1.21093281e-01 7.58300304e-01 -5.32202542e-01 2.86213249e-01
-2.30478123e-01 -3.72477829e-01 -8.37623700e-02 8.81312191e-01
6.94867730e-01 4.43492740e-01 5.59835732e-01 -4.83785599e-01
-1.33326324e-02 1.95609838e-01 -9.70206380e-01 1.51744020e+00
-9.33205307e-01 6.10715151e-01 -3.38614546e-02 -7.45855510e-01
1.22624731e+00 2.64226466e-01 -2.39188820e-01 -6.27882123e-01
6.35566294e-01 4.38579440e-01 3.43458593e-01 -4.38496858e-01
1.04455447e+00 -2.79579401e-01 -5.09204447e-01 6.21703029e-01
4.11619276e-01 4.58194241e-02 1.67826697e-01 5.78799956e-02
6.85723186e-01 1.39408067e-01 1.67004406e-01 -5.90469539e-01
1.02105296e+00 3.36212548e-03 -1.03045300e-01 2.88583547e-01
-2.06552461e-01 4.84728217e-01 5.93950689e-01 -8.27270985e-01
-9.69131172e-01 -9.78058517e-01 -3.21972251e-01 1.74241960e+00
-1.41660407e-01 -5.08424938e-01 -7.08490670e-01 -9.51716721e-01
-2.49431849e-01 5.52998602e-01 -9.32024777e-01 6.83729127e-02
-6.71546638e-01 -1.17775679e+00 7.85839498e-01 2.73585916e-01
1.60914451e-01 -1.18372953e+00 -2.20917761e-01 3.38835537e-01
2.84133673e-01 -6.26145482e-01 -2.78389454e-01 6.71724856e-01
-6.08469546e-01 -9.23427522e-01 -7.49011338e-01 -1.30129087e+00
6.72567725e-01 -1.90625295e-01 1.06796348e+00 -1.14478149e-01
2.34141529e-01 -3.10285240e-02 -1.02807164e+00 -6.34274602e-01
-2.96032608e-01 3.22454959e-01 1.26035273e-01 2.86372781e-01
1.01202297e+00 -3.26279938e-01 -4.55451220e-01 -3.32370549e-02
-1.20409298e+00 -7.58543611e-01 3.83919626e-01 9.66693461e-01
-1.48613170e-01 -2.89414376e-01 6.23697162e-01 -1.08363891e+00
1.16622770e+00 -6.10002935e-01 -2.15162352e-01 1.58881694e-01
-6.14971638e-01 4.04652119e-01 1.13911295e+00 -3.99322093e-01
-8.04195344e-01 -1.07069135e-01 -2.98271537e-01 3.98231745e-01
5.31058758e-02 7.38803267e-01 -2.74954319e-01 1.57955021e-01
8.11482787e-01 3.52694005e-01 8.95145312e-02 -2.86981046e-01
7.15072453e-01 9.83510435e-01 -5.36661893e-02 -6.39004946e-01
4.03603405e-01 1.71398044e-01 -5.81792712e-01 -8.40796471e-01
-3.65927905e-01 3.48206721e-02 -7.11555004e-01 1.34212866e-01
9.90001619e-01 -8.97024930e-01 -2.08836779e-01 8.71375740e-01
-1.09789670e+00 8.14019353e-04 2.75563508e-01 1.92583799e-01
3.15416940e-02 3.43508214e-01 -7.33943999e-01 -5.61178446e-01
-4.15782481e-01 -1.48186255e+00 7.72700369e-01 2.15525832e-03
-3.63217264e-01 -1.29238808e+00 3.23190182e-01 4.94658425e-02
6.92443252e-01 5.54024205e-02 1.49493635e+00 -1.06241429e+00
1.50624827e-01 -4.34919268e-01 -1.39776215e-01 7.01811910e-01
3.97325873e-01 3.20641577e-01 -1.20067179e+00 -1.10103950e-01
-1.03006974e-01 -4.83566731e-01 5.16787767e-01 -8.74069259e-02
4.18978214e-01 -4.17989939e-01 5.40029943e-01 4.68381137e-01
1.68800497e+00 1.07686095e-01 4.85372096e-01 8.22212398e-01
7.79122770e-01 8.59047890e-01 2.67043173e-01 2.08266079e-01
5.89690030e-01 2.45817468e-01 5.29905081e-01 2.17084706e-01
2.40213960e-01 8.75968114e-02 9.48517323e-01 1.48357689e+00
2.29301691e-01 -4.03555572e-01 -9.05498922e-01 8.45763981e-01
-1.34150147e+00 -3.49855542e-01 -2.91840702e-01 1.72579718e+00
7.66176283e-01 7.02131763e-02 -1.08094011e-02 2.84804612e-01
2.08932757e-01 4.09970880e-01 -6.06343709e-02 -1.56482553e+00
-5.44096351e-01 6.00690246e-01 2.78454840e-01 5.83563447e-01
-1.07663023e+00 1.25068724e+00 5.85028839e+00 3.76126379e-01
-1.33074224e+00 7.56099001e-02 1.41339406e-01 1.38927847e-01
-5.35484493e-01 -2.99865484e-01 -6.31294072e-01 3.57762069e-01
1.40888274e+00 1.43930227e-01 1.83514327e-01 7.71019161e-01
-1.78771690e-01 3.35779399e-01 -8.79938722e-01 5.66687465e-01
4.22434211e-01 -9.73866999e-01 4.97858167e-01 -2.38009065e-01
6.10940158e-01 1.70639411e-01 2.58682430e-01 5.71533263e-01
4.93930578e-01 -1.34644461e+00 6.57830417e-01 8.16526711e-02
3.20717782e-01 -1.06692815e+00 1.19495332e+00 -3.39991659e-01
-9.58970904e-01 -1.86741441e-01 -5.04057884e-01 -3.29672545e-01
3.81448157e-02 -7.15633668e-03 -6.01924241e-01 4.81251001e-01
5.45202672e-01 7.14300573e-01 -8.98917496e-01 1.34391263e-01
-2.14191884e-01 5.65390408e-01 -4.52852547e-02 -6.18979335e-01
9.60887313e-01 -6.09092653e-01 3.13036814e-02 1.60687804e+00
1.56186640e-01 -6.01128519e-01 -4.25687104e-01 1.74052536e-01
8.48663598e-02 7.64443338e-01 -7.14425087e-01 -2.29639038e-01
1.29291758e-01 1.42890179e+00 -4.71202195e-01 -1.89641267e-01
-5.18227816e-01 1.12639260e+00 3.63657355e-01 2.56721228e-01
-8.14977765e-01 -8.94613206e-01 9.75803077e-01 -2.41932034e-01
4.18535203e-01 -3.77812177e-01 -4.04786944e-01 -1.17654729e+00
-3.82975712e-02 -1.27557409e+00 1.80758744e-01 -7.32668281e-01
-1.57559931e+00 1.42446232e+00 -5.10396004e-01 -1.02387702e+00
-7.67352805e-02 -1.46886003e+00 -5.30782044e-01 1.16066062e+00
-1.42811096e+00 -1.54612672e+00 3.45889106e-02 4.44841027e-01
6.07512355e-01 -9.22479808e-01 1.38326931e+00 3.83200616e-01
-5.09921253e-01 7.55465984e-01 3.21690083e-01 4.46981281e-01
8.03078830e-01 -1.68419254e+00 6.58665597e-02 8.70480657e-01
1.81044742e-01 1.19308352e+00 6.22590721e-01 -1.56404436e-01
-1.68560791e+00 -8.36300075e-01 1.16250253e+00 -7.48250782e-01
1.07104719e+00 -6.10830784e-01 -8.18845928e-01 6.17055595e-01
9.45837379e-01 -3.74823928e-01 1.17651963e+00 2.39949331e-01
-7.86967516e-01 -1.86702564e-01 -9.16381538e-01 8.15106392e-01
2.15068549e-01 -5.83614945e-01 -1.01497853e+00 4.41166386e-02
5.79152584e-01 1.13064758e-01 -8.02074194e-01 -1.13128632e-01
8.37513030e-01 -8.87030900e-01 6.98904157e-01 -9.07932222e-01
7.76626527e-01 -3.02901000e-01 -7.11330116e-01 -1.55577993e+00
-8.71326849e-02 -7.24603534e-02 1.67751685e-01 1.27453709e+00
7.97595024e-01 -6.05589211e-01 4.52155530e-01 2.97124505e-01
-1.44378111e-01 -5.91859400e-01 -3.26222062e-01 -2.88713366e-01
7.83338904e-01 -2.88970947e-01 8.16369474e-01 1.08011520e+00
9.22243446e-02 6.51781082e-01 -1.84956476e-01 1.68145284e-01
9.65347793e-03 -1.65935263e-01 5.91274679e-01 -8.66909504e-01
1.98440030e-01 -5.37986100e-01 -5.90892613e-01 -5.90571463e-01
2.13970289e-01 -1.05325282e+00 -1.03160433e-01 -1.28718400e+00
-6.39998317e-01 -3.53873074e-01 -4.53092426e-01 2.70154476e-01
-1.46703452e-01 4.58494186e-01 1.45960033e-01 -3.07016224e-01
3.93254720e-02 6.61068380e-01 8.07129741e-01 -2.44659156e-01
-8.92399326e-02 -5.35142541e-01 -1.17249751e+00 7.80173600e-01
8.61983955e-01 -3.80949110e-01 -4.81760472e-01 -8.49019170e-01
7.37463117e-01 -7.11360335e-01 -3.50587070e-01 -6.27235472e-01
-9.25227031e-02 2.22278550e-01 2.35172108e-01 -2.82906383e-01
2.64729589e-01 -1.11819196e+00 -5.45288324e-01 2.43493855e-01
-1.62933350e-01 7.92438567e-01 3.07530820e-01 9.94027182e-02
-4.64738876e-01 -4.27184492e-01 5.02483189e-01 -7.04143122e-02
-8.76449168e-01 -1.13496110e-01 -7.72221804e-01 -2.77655572e-01
7.19573498e-01 -3.40813249e-01 -1.01449415e-01 -2.79825032e-02
-5.25103390e-01 -5.36961257e-02 3.70129973e-01 7.03474224e-01
5.63158572e-01 -1.26977074e+00 -8.59816194e-01 4.79135633e-01
4.29473460e-01 -3.29537034e-01 2.06991155e-02 3.56898278e-01
-1.18274736e+00 2.23846838e-01 -6.66160285e-01 -1.08503789e-01
-1.02680945e+00 3.55326265e-01 3.49374026e-01 -1.73367962e-01
-1.09887542e-02 8.47030342e-01 -7.35158324e-02 -1.33722568e+00
7.29554072e-02 -4.48422939e-01 -8.11495125e-01 6.86009824e-01
6.20745301e-01 -1.79535721e-03 4.40513462e-01 -1.05320704e+00
-3.79841179e-01 6.52025104e-01 -3.31450075e-01 -1.81811392e-01
1.50484657e+00 -2.18792528e-01 -3.36180270e-01 6.13807440e-01
1.27143443e+00 3.58433336e-01 -5.30086517e-01 3.35742794e-02
9.23951343e-02 -1.63298383e-01 -9.62729454e-02 -9.43153203e-01
-9.85175133e-01 1.10266483e+00 6.80311739e-01 3.95796418e-01
9.61555064e-01 -8.09347868e-01 7.02939212e-01 6.22940481e-01
1.48841396e-01 -9.65730846e-01 -1.06379069e-01 1.05390656e+00
7.52975285e-01 -1.33271682e+00 -3.09401393e-01 3.57988268e-01
-1.14131534e+00 1.60843062e+00 6.85804069e-01 -6.74688876e-01
9.63046432e-01 8.69230032e-02 8.60820174e-01 -2.70210505e-01
-4.64311808e-01 -6.99499696e-02 2.94507623e-01 7.71389604e-01
1.10449672e+00 -2.26535578e-03 -4.47221607e-01 9.20756221e-01
-6.14240706e-01 -5.23149550e-01 9.02982354e-01 7.91065633e-01
-2.30574340e-01 -1.42527413e+00 -2.79955804e-01 2.88165063e-01
-4.94702756e-01 -4.20112014e-01 -4.51967984e-01 7.97764897e-01
2.28057742e-01 8.71117294e-01 1.32006437e-01 -5.19339144e-01
4.90339577e-01 3.05407017e-01 1.95885479e-01 -5.93382120e-01
-1.48365593e+00 -3.34352404e-01 1.23241685e-01 -2.02885523e-01
-3.96308154e-01 -4.15933400e-01 -1.04609442e+00 -3.57209474e-01
-1.74316198e-01 1.10055387e-01 9.71582949e-01 6.74852252e-01
3.99732664e-02 4.02292460e-01 7.39004672e-01 -6.67261481e-01
-4.87877786e-01 -1.33662903e+00 -7.39680529e-01 5.05904853e-01
4.58693147e-01 -2.16710821e-01 -3.23722690e-01 2.64213812e-02] | [11.15727424621582, 7.212823867797852] |
075bf644-bdd6-4736-89a4-9d5659235c96 | more-for-less-compact-convolutional | 2307.00213 | null | https://arxiv.org/abs/2307.00213v1 | https://arxiv.org/pdf/2307.00213v1.pdf | More for Less: Compact Convolutional Transformers Enable Robust Medical Image Classification with Limited Data | Transformers are very powerful tools for a variety of tasks across domains, from text generation to image captioning. However, transformers require substantial amounts of training data, which is often a challenge in biomedical settings, where high quality labeled data can be challenging or expensive to obtain. This study investigates the efficacy of Compact Convolutional Transformers (CCT) for robust medical image classification with limited data, addressing a key issue faced by conventional Vision Transformers - their requirement for large datasets. A hybrid of transformers and convolutional layers, CCTs demonstrate high accuracy on modestly sized datasets. We employed a benchmark dataset of peripheral blood cell images of eight distinct cell types, each represented by approximately 2,000 low-resolution (28x28x3 pixel) samples. Despite the dataset size being smaller than those typically used with Vision Transformers, we achieved a commendable classification accuracy of 92.49% and a micro-average ROC AUC of 0.9935. The CCT also learned quickly, exceeding 80% validation accuracy after five epochs. Analysis of per-class precision, recall, F1, and ROC showed that performance was strong across cell types. Our findings underscore the robustness of CCTs, indicating their potential as a solution to data scarcity issues prevalent in biomedical imaging. We substantiate the applicability of CCTs in data-constrained areas and encourage further work on CCTs. | ['Andrew Kean Gao'] | 2023-07-01 | null | null | null | null | ['image-captioning', 'medical-image-classification', 'text-generation'] | ['computer-vision', 'medical', 'natural-language-processing'] | [ 4.97163445e-01 3.18862535e-02 -6.29124194e-02 -3.13344032e-01
-1.19150460e+00 -5.93438089e-01 4.28432494e-01 1.53664470e-01
-5.84718645e-01 8.74691844e-01 9.76397991e-02 -5.74368179e-01
8.96513090e-03 -4.81846750e-01 -6.14172995e-01 -7.31331706e-01
1.54151976e-01 3.73456687e-01 3.14152129e-02 2.63694495e-01
3.41687687e-02 4.54318076e-01 -1.16397655e+00 4.73901510e-01
6.81418657e-01 1.36538279e+00 1.61027431e-01 6.07000232e-01
-1.40181435e-02 6.92738831e-01 -7.94027746e-01 -4.70185965e-01
9.22010541e-02 -5.77013567e-02 -6.71551645e-01 4.89442423e-02
6.59090281e-01 -3.18470657e-01 -1.35698915e-01 5.97213745e-01
9.27314878e-01 -4.89142478e-01 7.39940763e-01 -1.05235767e+00
-7.07198560e-01 1.77597672e-01 -4.44749594e-01 6.94403827e-01
1.83684155e-01 6.18751645e-01 7.33502328e-01 -7.55600631e-01
4.81871873e-01 7.78002024e-01 1.00135481e+00 7.27926373e-01
-1.34118021e+00 -8.78063977e-01 -3.19701821e-01 -2.93247819e-01
-1.14458179e+00 -6.66501760e-01 1.36668324e-01 -7.26405978e-01
1.09102142e+00 1.89444169e-01 6.40072703e-01 1.22119772e+00
4.91715252e-01 4.25434172e-01 1.46035457e+00 -7.09673914e-04
1.31040856e-01 2.57073253e-01 1.18821338e-02 6.38418198e-01
4.30571318e-01 -1.96271271e-01 -5.52978933e-01 -1.97274804e-01
8.94475162e-01 -1.65214449e-01 -4.20868099e-01 1.38283938e-01
-1.51319897e+00 6.33589685e-01 3.46381903e-01 3.79429102e-01
-3.28579694e-01 4.09042649e-02 6.44748926e-01 3.43481570e-01
5.11653543e-01 6.62164927e-01 -5.60056150e-01 -2.29233012e-01
-8.23247612e-01 -6.99619576e-02 5.41961551e-01 9.42017972e-01
-1.07835695e-01 1.41101807e-01 -3.56178403e-01 8.47171664e-01
-1.37616917e-01 6.46382153e-01 5.87463915e-01 -7.20446289e-01
3.99447232e-01 5.96390069e-01 -3.58047634e-02 -6.51307166e-01
-5.86296856e-01 -6.12134039e-01 -1.05270994e+00 -5.41288033e-03
8.52959812e-01 -1.01206206e-01 -1.00954998e+00 1.64683509e+00
8.88790414e-02 -1.41794547e-01 -4.32188027e-02 7.78337717e-01
1.09432805e+00 3.08502793e-01 3.30070972e-01 -2.16031075e-01
1.62826049e+00 -3.65695536e-01 -5.91826260e-01 -1.36526868e-01
6.40178800e-01 -7.09948540e-01 1.35668862e+00 4.83081490e-01
-1.00277901e+00 -2.64308661e-01 -9.09977794e-01 -1.94308251e-01
-3.35293084e-01 3.33895624e-01 7.52056122e-01 7.70728350e-01
-1.26455688e+00 3.43555897e-01 -5.99425316e-01 -4.06147122e-01
1.21526515e+00 5.79101682e-01 -5.82040250e-01 -7.78653100e-02
-7.38745391e-01 9.38958228e-01 -2.75293022e-01 9.50235948e-02
-8.28486741e-01 -1.22893119e+00 -5.14210284e-01 -2.00060830e-01
-2.13621229e-01 -7.86723197e-01 1.19214332e+00 -6.84637547e-01
-1.14147413e+00 1.13517284e+00 8.52242261e-02 -5.88795960e-01
7.83089936e-01 6.71292767e-02 -2.08215013e-01 3.38318855e-01
2.80821025e-01 7.69814312e-01 5.64880848e-01 -7.15779543e-01
-3.94803137e-01 -2.45170310e-01 -1.61346465e-01 -2.30091019e-03
-5.23914814e-01 -4.81552109e-02 -1.62160024e-01 -4.36560035e-01
-2.23820537e-01 -7.42452979e-01 -5.60288988e-02 3.21773529e-01
-2.79547125e-01 -8.35821182e-02 5.63388348e-01 -5.41352153e-01
7.58332491e-01 -2.01635599e+00 -5.27926326e-01 -9.89159122e-02
6.00659311e-01 3.68642181e-01 -5.78828491e-02 1.54132443e-02
-5.66418990e-02 2.42833972e-01 -8.42742026e-02 -6.81453012e-03
-2.92984337e-01 -2.05020994e-01 -1.50935009e-01 5.76902926e-01
3.47871870e-01 1.31789017e+00 -7.20491409e-01 -5.85039258e-01
1.76404923e-01 6.62875533e-01 -1.08029492e-01 1.59307569e-02
8.65059346e-02 5.74090600e-01 -3.71648550e-01 8.94992650e-01
4.07608211e-01 -9.22945917e-01 -3.75769474e-02 -5.15405774e-01
4.55547683e-02 1.59295619e-01 -4.40217882e-01 1.41279745e+00
-3.08072150e-01 9.82682526e-01 -8.93553421e-02 -9.40299273e-01
7.37705290e-01 5.44457912e-01 8.23713183e-01 -9.61102664e-01
2.27514789e-01 2.50709742e-01 1.94168836e-02 -6.33841813e-01
-3.59468418e-03 -5.27917624e-01 9.28344131e-02 3.09872568e-01
-2.68083382e-02 -1.72451496e-01 -1.29786480e-04 -2.30845138e-02
1.40951312e+00 -2.55823165e-01 6.78846166e-02 -3.87127101e-01
3.20278890e-02 1.37550592e-01 2.31498331e-01 6.88437939e-01
-3.45785439e-01 9.30497944e-01 6.30181730e-01 -6.26017094e-01
-1.24709594e+00 -1.00799346e+00 -6.60114765e-01 5.81444502e-01
-4.36110258e-01 -1.01958245e-01 -6.91849887e-01 -4.82791215e-01
-3.23373452e-02 3.78633104e-02 -7.31848955e-01 -2.55039819e-02
-1.84613794e-01 -1.22579408e+00 9.41752732e-01 6.82467878e-01
4.90626156e-01 -9.07019019e-01 -9.43788171e-01 3.31294060e-01
-2.65978456e-01 -1.62231576e+00 -3.56619954e-01 1.73594609e-01
-9.75041807e-01 -1.22756577e+00 -1.09807694e+00 -8.46611977e-01
7.37025559e-01 4.00750563e-02 1.16747642e+00 3.35700996e-02
-6.19382203e-01 4.44788635e-01 5.21066748e-02 -6.86258912e-01
-2.01011330e-01 9.03842300e-02 -8.42970237e-03 -3.28530788e-01
2.89682955e-01 -3.71460408e-01 -9.06488836e-01 2.05620259e-01
-6.64330006e-01 1.73704386e-01 8.59724939e-01 1.02970481e+00
6.46785557e-01 -4.07817274e-01 9.41478372e-01 -9.57439423e-01
7.40086436e-01 -2.45194316e-01 -3.23700279e-01 8.57943371e-02
-6.99279547e-01 -3.47986281e-01 5.96463978e-01 -7.16111124e-01
-6.93268478e-01 -3.96374241e-02 1.54533312e-02 -2.30506822e-01
-3.20180431e-02 3.62706333e-01 3.48439395e-01 -1.54757336e-01
9.24342334e-01 1.09346062e-01 3.91811341e-01 -3.45239043e-02
-1.75954267e-01 7.91981220e-01 4.93794680e-01 -4.58826751e-01
2.53307879e-01 5.73890507e-01 -5.01215551e-03 -9.77824628e-01
-8.42842758e-01 -2.81193495e-01 -3.21626335e-01 -7.99653679e-02
8.07276428e-01 -1.03303790e+00 -8.73737097e-01 5.42365670e-01
-7.54350424e-01 -6.22635603e-01 -2.70024866e-01 4.47866052e-01
-2.88912594e-01 6.27978072e-02 -8.48895669e-01 -4.02957678e-01
-9.76171374e-01 -1.09577262e+00 1.18683052e+00 7.38548636e-02
-3.98058712e-01 -9.63933170e-01 -2.82792330e-01 7.11182833e-01
5.95140755e-01 5.04383206e-01 1.02341533e+00 -4.94465142e-01
-3.56216490e-01 -2.59315431e-01 -6.08854055e-01 2.01931953e-01
3.91851127e-01 -2.66758502e-01 -1.31361699e+00 -4.48360950e-01
-2.63553381e-01 -7.71889031e-01 5.71646869e-01 7.40878463e-01
1.31546021e+00 -9.91771072e-02 -4.79587227e-01 4.65156436e-01
1.22817504e+00 5.81475198e-02 6.12621248e-01 1.72018185e-01
5.59385180e-01 3.26383710e-01 2.62488604e-01 2.83637196e-01
2.95078158e-01 3.02895069e-01 9.23106149e-02 -5.61410248e-01
-2.95815408e-01 1.56344891e-01 7.05077266e-03 4.85894442e-01
1.81117505e-01 -1.90972283e-01 -1.00974786e+00 7.19651282e-01
-1.18557632e+00 -6.14692807e-01 -1.29837796e-01 2.19781232e+00
1.06949925e+00 2.67092139e-01 1.80914208e-01 1.37961254e-01
3.92877728e-01 -3.34886312e-01 -7.57709205e-01 -4.70351130e-02
-2.41688162e-01 3.61936986e-01 4.61388201e-01 -1.77284032e-02
-9.03319180e-01 4.85366702e-01 7.12822342e+00 6.17165744e-01
-1.50084937e+00 5.50442785e-02 1.23294497e+00 -2.75845230e-01
-6.19210564e-02 -5.64041972e-01 -4.83334422e-01 6.03287339e-01
1.18586755e+00 -2.66456418e-03 2.89180763e-02 3.97890538e-01
2.96007246e-01 -1.89926118e-01 -1.21801901e+00 1.14005184e+00
-1.00046329e-01 -1.57413387e+00 1.53691974e-03 1.89240053e-01
5.40844738e-01 2.52227277e-01 4.91744041e-01 8.96857381e-02
-8.59222468e-03 -1.61006951e+00 4.13214505e-01 2.43061736e-01
1.61054444e+00 -2.24848688e-01 8.22057486e-01 -1.18573112e-02
-8.61653149e-01 6.06518351e-02 -3.22441369e-01 1.10690661e-01
-1.64243340e-01 8.27861369e-01 -1.23783088e+00 -2.01197769e-02
8.91900182e-01 6.01994276e-01 -6.55987024e-01 9.98829365e-01
4.44260299e-01 5.88920474e-01 -3.68192673e-01 -1.51622236e-01
1.25669986e-01 2.70122409e-01 -2.94554234e-02 1.40386045e+00
3.63801062e-01 2.45270893e-01 -2.10212514e-01 6.83928311e-01
-2.36858845e-01 5.72543740e-02 -7.80941010e-01 -2.28585631e-01
5.42740464e-01 1.37140918e+00 -1.05005825e+00 -3.80405098e-01
-4.55436289e-01 5.50508201e-01 2.48170599e-01 2.38554001e-01
-7.42910326e-01 -1.15052268e-01 2.94055998e-01 3.17403138e-01
6.30623698e-02 2.02876166e-01 -7.36970723e-01 -7.57201672e-01
1.26083463e-01 -1.02114213e+00 3.34558159e-01 -8.55024099e-01
-1.41537952e+00 6.41438007e-01 -2.82984346e-01 -1.25384712e+00
1.87841386e-01 -9.26433504e-01 -2.22350791e-01 7.30739594e-01
-1.47961915e+00 -1.32156253e+00 -5.03205121e-01 5.54945529e-01
3.44016761e-01 -7.13738129e-02 9.56916809e-01 3.63559008e-01
-3.60905409e-01 8.11802804e-01 5.99818379e-02 2.44210184e-01
6.89680219e-01 -1.12773609e+00 1.66770354e-01 2.86203891e-01
-1.07899219e-01 6.70222461e-01 4.75306243e-01 -4.43325043e-01
-1.43279600e+00 -1.10330176e+00 7.21086740e-01 -6.74020767e-01
3.98727983e-01 -3.14671218e-01 -7.42607892e-01 5.44431567e-01
-1.78271178e-02 2.54369169e-01 8.62154365e-01 -1.21360555e-01
-3.08025002e-01 -2.40367919e-01 -1.41055369e+00 4.94016200e-01
8.02026749e-01 -5.98352969e-01 -1.07238546e-01 4.88014489e-01
3.24188918e-01 -5.88827908e-01 -1.27504635e+00 3.24732661e-01
7.48927891e-01 -6.50658786e-01 8.98005009e-01 -3.36058289e-01
5.18272400e-01 5.35549819e-02 1.36806577e-01 -1.12092233e+00
-2.54424989e-01 -3.28716308e-01 2.76498139e-01 1.04022264e+00
7.08646715e-01 -7.90863514e-01 8.31219494e-01 7.55494177e-01
-1.30081907e-01 -1.08201265e+00 -1.20716262e+00 -4.94641811e-01
3.37729037e-01 -3.89345914e-01 4.18620944e-01 9.67831790e-01
1.70811921e-01 4.69261289e-01 3.96432653e-02 -2.77171791e-01
5.59805274e-01 -1.25202343e-01 4.41346258e-01 -1.19917810e+00
1.51178613e-01 -4.24193710e-01 -4.29996341e-01 -5.40868223e-01
-1.34124056e-01 -8.82467270e-01 -1.96641669e-01 -1.58237970e+00
4.91411448e-01 -7.69543827e-01 -2.57301092e-01 7.06103027e-01
-1.70241445e-02 7.23469913e-01 -2.22500667e-01 1.41769469e-01
-3.98867309e-01 2.64920443e-02 1.50264525e+00 -4.11191165e-01
1.83739349e-01 -3.86196882e-01 -1.00318301e+00 3.10704857e-01
9.02540267e-01 -2.30013251e-01 -6.77282810e-01 -5.61746120e-01
-2.33591031e-02 6.64175376e-02 3.94017518e-01 -1.08521533e+00
1.28112018e-01 3.44075412e-02 8.47900331e-01 -2.24489436e-01
3.85397911e-01 -6.04732513e-01 3.05963755e-01 5.74001551e-01
-4.02185768e-01 5.47063798e-02 4.65928346e-01 2.54662305e-01
1.55094534e-01 2.92728186e-01 9.10077572e-01 -3.01992178e-01
-2.76566774e-01 4.33015704e-01 -4.44548279e-01 3.88028145e-01
9.35531080e-01 -5.65893769e-01 -5.96933305e-01 -1.33436084e-01
-3.97840053e-01 8.88267905e-02 3.27577353e-01 1.99526146e-01
8.07879269e-01 -1.15267050e+00 -7.49068975e-01 9.10185054e-02
3.19629610e-01 1.26651466e-01 1.89078242e-01 1.18320978e+00
-5.62377453e-01 5.96705019e-01 -3.97185117e-01 -9.81841147e-01
-1.27548909e+00 2.40338549e-01 5.27549088e-01 -2.16048837e-01
-8.72668207e-01 7.96742916e-01 2.81273127e-01 -1.18391708e-01
5.55666909e-02 -5.98075926e-01 -1.62445635e-01 4.99032438e-02
5.56533575e-01 1.07205614e-01 3.59868884e-01 -2.67516524e-01
-3.91234577e-01 4.40023750e-01 -2.56216556e-01 2.40756813e-02
1.38495028e+00 9.05249566e-02 1.99338228e-01 4.21913981e-01
1.14032817e+00 -4.27775443e-01 -1.29548728e+00 -6.56177327e-02
-2.15329558e-01 -1.92235187e-01 -6.17644191e-03 -1.01309943e+00
-1.12524080e+00 9.48312223e-01 9.06459808e-01 -3.10253017e-02
9.61048365e-01 1.05007201e-01 7.19187379e-01 1.15982309e-01
3.44921827e-01 -8.99680316e-01 2.77559370e-01 1.11422636e-01
7.88004279e-01 -1.30527747e+00 8.32146034e-02 -1.54871657e-01
-6.80036664e-01 9.00764942e-01 6.51384115e-01 2.99875319e-01
3.40268642e-01 6.07909381e-01 3.17148477e-01 -5.10151803e-01
-9.11252856e-01 2.81514734e-01 1.30449295e-01 9.43442643e-01
7.91647434e-01 7.53380731e-03 -1.09505259e-01 4.36175108e-01
-2.97989279e-01 3.93954337e-01 5.30080914e-01 8.78030658e-01
-9.86372307e-02 -4.64708716e-01 -1.20601624e-01 1.14619958e+00
-9.90316272e-01 -2.71940500e-01 -1.51000559e-01 6.55476511e-01
-1.14307351e-01 9.57813919e-01 5.97485304e-02 -2.22014964e-01
3.05974394e-01 -1.48581624e-01 5.70573270e-01 -4.15963233e-01
-6.41334474e-01 -9.19942260e-02 7.90147632e-02 -3.35023582e-01
-3.50720912e-01 -6.19403243e-01 -1.11543751e+00 -3.10264081e-01
-2.47510269e-01 -2.07097083e-01 4.03093457e-01 8.30948412e-01
3.80441815e-01 6.02568209e-01 2.49485075e-01 -2.74084419e-01
-4.29259270e-01 -1.07364988e+00 -3.98995280e-01 3.81482631e-01
5.61200619e-01 -4.91794139e-01 -8.97047073e-02 3.63058150e-01] | [14.871862411499023, -2.4614195823669434] |
1c851b13-72b7-4d66-be4f-30505cb59234 | research-on-multi-agent-communication-and | 2305.17141 | null | https://arxiv.org/abs/2305.17141v1 | https://arxiv.org/pdf/2305.17141v1.pdf | Research on Multi-Agent Communication and Collaborative Decision-Making Based on Deep Reinforcement Learning | In a multi-agent environment, In order to overcome and alleviate the non-stationarity of the multi-agent environment, the mainstream method is to adopt the framework of Centralized Training Decentralized Execution (CTDE). This thesis is based on the framework of CTDE, and studies the cooperative decision-making of multi-agent based on the Multi-Agent Proximal Policy Optimization (MAPPO) algorithm for multi-agent proximal policy optimization. In order to alleviate the non-stationarity of the multi-agent environment, a multi-agent communication mechanism based on weight scheduling and attention module is introduced. Different agents can alleviate the non-stationarity caused by local observations through information exchange between agents, assisting in the collaborative decision-making of agents. The specific method is to introduce a communication module in the policy network part. The communication module is composed of a weight generator, a weight scheduler, a message encoder, a message pool and an attention module. Among them, the weight generator and weight scheduler will generate weights as the selection basis for communication, the message encoder is used to compress and encode communication information, the message pool is used to store communication messages, and the attention module realizes the interactive processing of the agent's own information and communication information. This thesis proposes a Multi-Agent Communication and Global Information Optimization Proximal Policy Optimization(MCGOPPO)algorithm, and conducted experiments in the SMAC and the MPE. The experimental results show that the improvement has achieved certain effects, which can better alleviate the non-stationarity of the multi-agent environment, and improve the collaborative decision-making ability among the agents. | ['Zeng Da'] | 2023-05-23 | null | null | null | null | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [-5.71788192e-01 -6.48116767e-02 -3.38781834e-01 1.01203537e-02
-9.49087217e-02 2.74471462e-01 3.63993049e-01 1.28060296e-01
-6.74186885e-01 6.93673670e-01 3.81853729e-01 -1.20893858e-01
-2.98369557e-01 -8.29302669e-01 -1.64592922e-01 -1.15697944e+00
-3.47938985e-01 8.34474921e-01 2.01895431e-01 -4.86499369e-01
2.70717829e-01 1.19327158e-01 -1.44715369e+00 3.21374208e-01
9.35556412e-01 7.03156710e-01 9.39514816e-01 5.21437645e-01
-2.73113668e-01 1.04960895e+00 -9.31682646e-01 3.96538138e-01
8.58168229e-02 -2.85180986e-01 -6.06480598e-01 9.37452763e-02
-8.09796929e-01 -4.79836851e-01 3.71450447e-02 1.03418815e+00
1.00047231e+00 4.83649850e-01 3.20773512e-01 -1.64220524e+00
-3.45779121e-01 9.19080615e-01 -4.62490171e-01 3.31935167e-01
1.70277581e-01 1.91008747e-01 5.53250730e-01 -4.33523208e-01
2.83783197e-01 1.68850851e+00 2.19093114e-01 5.61137319e-01
-5.72566152e-01 -5.80729485e-01 6.67998552e-01 6.48204088e-01
-1.09296834e+00 -1.38129205e-01 6.16302013e-01 -1.39189020e-01
9.15434301e-01 1.96600959e-01 9.66452897e-01 7.05052197e-01
7.79174805e-01 9.74396408e-01 7.84989178e-01 -5.09616792e-01
4.19497192e-01 1.67040303e-01 -1.24035813e-01 7.04356551e-01
-1.71233341e-01 2.90711433e-01 -2.15730399e-01 -3.17712665e-01
7.27395356e-01 1.11671492e-01 2.98661571e-02 1.13584690e-01
-1.16602588e+00 9.53183055e-01 2.77624398e-01 3.86125118e-01
-8.04777682e-01 6.50020316e-03 5.10110259e-01 6.99980557e-01
5.84708512e-01 6.65003732e-02 -4.79318500e-01 -1.37148306e-01
-2.87844300e-01 2.75565624e-01 9.42618906e-01 6.07245564e-01
6.35321915e-01 3.23897809e-01 -3.76188636e-01 6.89546049e-01
8.79487097e-01 4.88098472e-01 1.20941687e+00 -1.02911198e+00
5.32354295e-01 6.77161634e-01 9.29645970e-02 -9.54404116e-01
-6.02821350e-01 -5.13095677e-01 -8.76241028e-01 6.65729702e-01
-1.81283012e-01 -8.18308711e-01 -2.90200412e-01 1.83433604e+00
7.35660672e-01 -2.23895553e-02 3.58469844e-01 1.16003621e+00
5.74895203e-01 9.84292686e-01 3.25562537e-01 -7.61438072e-01
1.27100575e+00 -1.37702143e+00 -7.47923791e-01 -8.32486600e-02
6.11152470e-01 -8.48821700e-01 5.60826182e-01 4.91811931e-02
-1.09659708e+00 -5.65878570e-01 -9.42449093e-01 4.87899601e-01
-2.46655524e-01 -6.67028725e-02 4.83529747e-01 -2.04947088e-02
-7.60153413e-01 3.94756585e-01 -7.35908985e-01 -4.83721010e-02
-8.82266238e-02 6.67525291e-01 2.16099918e-01 2.73613334e-01
-1.37970853e+00 1.10806859e+00 8.19011331e-01 -1.33184537e-01
-1.13666570e+00 -4.90892589e-01 -5.53975642e-01 3.10945928e-01
3.96812975e-01 -1.03356671e+00 1.25781024e+00 -1.44625473e+00
-2.02197576e+00 -4.93716784e-02 9.99156684e-02 -2.40437061e-01
3.70273322e-01 2.50863701e-01 -4.42013860e-01 -1.47238255e-01
1.54784396e-01 4.34708238e-01 5.92056632e-01 -1.15552318e+00
-1.32167315e+00 -3.30213308e-02 9.14131999e-02 1.07446826e+00
-1.91739932e-01 3.21133316e-01 -9.21427608e-02 -5.55465639e-01
-3.56332570e-01 -7.19593108e-01 -4.88958895e-01 -7.03521252e-01
6.90552220e-02 -6.82482302e-01 1.13348329e+00 -3.59120667e-01
1.32558274e+00 -2.30848312e+00 4.73253161e-01 2.66915321e-01
1.47081167e-01 2.20278561e-01 -2.97843784e-01 6.04640305e-01
7.69934244e-03 -2.60157555e-01 3.76872450e-01 -4.61156726e-01
2.86321253e-01 6.10076427e-01 1.29323244e-01 3.47445279e-01
-4.45424855e-01 4.58404958e-01 -1.02698600e+00 -7.82014251e-01
2.28243768e-01 4.42365408e-02 -2.83150941e-01 4.83393312e-01
-5.46348095e-01 3.80101949e-01 -1.07972586e+00 3.82868618e-01
4.13727283e-01 -1.64104387e-01 1.62555873e-01 2.74648458e-01
-3.76120061e-01 -1.38098896e-01 -1.62058079e+00 1.47014463e+00
-2.49637470e-01 -2.24329196e-02 8.75756502e-01 -1.03790009e+00
7.56247401e-01 7.65938282e-01 9.75152135e-01 -7.06233203e-01
4.49062765e-01 -1.01340069e-02 3.05471987e-01 -9.08666193e-01
4.33342755e-01 2.07046837e-01 1.40295282e-01 9.38063860e-01
-2.33599111e-01 4.70664114e-01 5.51523268e-01 2.86253244e-01
8.07434440e-01 -1.86431840e-01 5.92208877e-02 -4.89963681e-01
8.22684646e-01 7.95176849e-02 1.10172129e+00 3.72795492e-01
-2.35899314e-01 -4.49730307e-01 8.66189227e-02 -6.80874705e-01
-8.26940894e-01 -3.52283537e-01 5.31118155e-01 1.82241416e+00
5.89331746e-01 -4.15054739e-01 -5.23741782e-01 -5.09872556e-01
8.73094350e-02 4.47419703e-01 -1.91828579e-01 -2.53494948e-01
-6.33848727e-01 -1.08408558e+00 -1.40222833e-01 1.39676765e-01
9.48632002e-01 -1.43640327e+00 -5.30147851e-01 7.40225315e-01
-2.27650821e-01 -5.40878177e-01 -7.39451647e-01 -1.64758280e-01
-4.03409570e-01 -8.56082916e-01 -3.87509853e-01 -9.15877879e-01
7.81774998e-01 3.26077968e-01 8.50608706e-01 4.53448296e-01
1.89388096e-01 4.44364220e-01 -4.08486277e-01 -8.95476997e-01
-6.09320939e-01 1.94787517e-01 2.57125467e-01 1.47484347e-01
3.63303088e-02 -5.56655765e-01 -5.15509725e-01 5.16221821e-01
-7.36512721e-01 4.07543689e-01 7.33280838e-01 8.05589378e-01
1.21950135e-01 2.89897501e-01 7.27647901e-01 -4.20276612e-01
1.22326851e+00 -7.98280001e-01 -7.71831572e-01 3.46812397e-01
-7.94726849e-01 -1.01629972e-01 8.38487089e-01 -6.03775442e-01
-1.37252319e+00 -3.23820919e-01 9.47143286e-02 -3.36081758e-02
1.24808922e-02 5.32141745e-01 -1.33890346e-01 -4.41647321e-02
2.24142134e-01 1.30737394e-01 3.67706805e-01 -1.96326539e-01
8.05606470e-02 8.26018810e-01 -1.50768235e-01 -8.30943823e-01
3.70912522e-01 7.50197098e-02 -1.51667267e-01 -2.96090424e-01
-1.71008170e-01 -8.40599164e-02 1.20440863e-01 -5.46825886e-01
8.99780452e-01 -9.19650316e-01 -1.42489278e+00 3.61776799e-01
-1.24724686e+00 -5.97890973e-01 -8.04294869e-02 1.00179946e+00
-3.72680038e-01 2.68813252e-01 -9.42072988e-01 -7.26769209e-01
-6.31893456e-01 -1.53087473e+00 3.83472830e-01 7.73401499e-01
6.46717995e-02 -1.31005394e+00 5.78330569e-02 1.64321825e-01
8.07173729e-01 -1.84212729e-01 8.77399802e-01 -5.93583405e-01
-6.74206555e-01 4.21678543e-01 1.61427915e-01 5.47617339e-02
-2.00385541e-01 -1.21071443e-01 -3.34339797e-01 -6.54830158e-01
1.70672014e-02 -1.62401184e-01 1.91276625e-01 4.81859386e-01
6.90585256e-01 -5.36701083e-01 -3.77179593e-01 2.49943629e-01
1.38452244e+00 5.96341550e-01 5.20514734e-02 7.23447502e-01
2.32548922e-01 5.46788394e-01 6.96780384e-01 7.57149339e-01
9.84817743e-01 4.75402653e-01 7.97036052e-01 -2.39457160e-01
-5.13562262e-02 2.14010537e-01 7.92480648e-01 1.37096298e+00
-2.50515848e-01 -2.80349016e-01 -4.46281046e-01 2.16960475e-01
-2.63096905e+00 -1.05015159e+00 1.55927539e-01 1.71815550e+00
7.31362104e-01 -2.18467206e-01 -1.33472914e-03 -3.24760795e-01
9.24803555e-01 2.63384264e-02 -3.45355719e-01 -5.39016485e-01
1.74288042e-02 -7.04931617e-01 3.75444263e-01 7.47474670e-01
-8.35924923e-01 7.00092673e-01 5.63066244e+00 1.06026995e+00
-1.08241558e+00 3.78859282e-01 1.79613844e-01 4.17792052e-02
6.24031760e-02 -1.24743320e-01 -8.18104208e-01 1.06187904e+00
6.71588480e-01 -4.78659898e-01 8.65940094e-01 8.34831953e-01
6.89769089e-01 -3.72739732e-01 -7.28721142e-01 8.95016730e-01
-1.68783933e-01 -1.19665730e+00 -2.22957179e-01 1.33436665e-01
8.71300936e-01 2.74089128e-01 -2.14404166e-01 4.73393768e-01
9.65211570e-01 -3.70331407e-01 6.03566945e-01 7.02063262e-01
-1.57282680e-01 -7.95707285e-01 1.00419319e+00 9.53238666e-01
-1.36003053e+00 -6.50962174e-01 -5.41949570e-01 -5.60427308e-01
1.63540229e-01 2.97637522e-01 -5.11975467e-01 9.99868691e-01
4.43421572e-01 4.67991740e-01 1.81735698e-02 1.02668333e+00
-1.52376920e-01 5.03711626e-02 -3.37876707e-01 -2.16079906e-01
3.52906168e-01 -4.92247790e-01 7.23510444e-01 8.52355957e-01
8.22745413e-02 1.51940972e-01 1.23934257e+00 3.34514499e-01
1.91071719e-01 2.30998605e-01 -5.51898926e-02 6.00779831e-01
7.58252740e-01 1.30515695e+00 -3.66093963e-01 -8.60828459e-01
-3.48902345e-01 3.96639615e-01 4.48180586e-01 5.38202882e-01
-7.62503564e-01 -3.97325903e-01 5.77086866e-01 -6.69883430e-01
-1.21592976e-01 -1.68818116e-01 5.54759763e-02 -7.80200481e-01
-8.69008824e-02 -9.58390713e-01 6.28701687e-01 -5.75046420e-01
-1.13090038e+00 2.42544934e-01 3.87040935e-02 -1.03342092e+00
-4.42610681e-01 3.06478124e-02 -1.19490695e+00 9.60101664e-01
-1.67646146e+00 -7.88260162e-01 -1.86282068e-01 8.36440265e-01
7.49352038e-01 -7.94314146e-01 1.01407266e+00 6.86370492e-01
-1.09174001e+00 2.54954677e-02 3.05815578e-01 -3.81914489e-02
4.63409036e-01 -1.05692744e+00 -7.57127643e-01 6.96383059e-01
-5.98539889e-01 1.85467973e-01 5.75712264e-01 -4.95825619e-01
-1.56020594e+00 -1.05091512e+00 6.34715021e-01 5.19472659e-01
3.52746427e-01 2.94106990e-01 -6.61796153e-01 5.02245903e-01
7.25394964e-01 -4.33362573e-01 5.63762903e-01 -1.45380765e-01
6.42787099e-01 -4.89836305e-01 -1.02303052e+00 5.64976215e-01
3.82815868e-01 1.95938185e-01 -6.44917071e-01 7.56818831e-01
6.73747897e-01 -4.51396704e-01 -9.27507579e-01 -3.46610174e-02
1.06497109e-01 -5.08682132e-01 7.37077475e-01 -4.39814776e-01
8.56997371e-02 -6.57617986e-01 2.88265735e-01 -1.81910014e+00
-8.29531252e-01 -6.76638842e-01 2.55252328e-02 1.21352410e+00
2.53927022e-01 -9.93107677e-01 5.82261026e-01 6.85297430e-01
-5.35969079e-01 -6.87655091e-01 -1.07517624e+00 -4.96616572e-01
-3.68443400e-01 1.14341423e-01 9.85921383e-01 8.91824782e-01
3.55074704e-01 4.41160887e-01 -2.83060521e-01 2.57133007e-01
4.80816394e-01 1.95704073e-01 7.71498561e-01 -8.80059838e-01
-5.03977478e-01 -7.61257529e-01 2.81434178e-01 -1.11247432e+00
2.69384027e-01 -7.41592705e-01 -2.51994152e-02 -1.76356792e+00
-2.00044475e-02 -7.06315339e-01 -3.64822209e-01 5.01983762e-01
-1.20818555e-01 -6.45271480e-01 5.46472371e-01 5.28326809e-01
-1.13781726e+00 8.49095523e-01 1.60065746e+00 -2.00972348e-01
-4.11193103e-01 6.95803612e-02 -3.59054059e-01 8.12577486e-01
1.12886357e+00 -5.01463056e-01 -5.13449550e-01 -9.05321598e-01
3.22525054e-02 3.37489605e-01 -1.19545497e-01 -7.62055457e-01
8.59759390e-01 -6.38942778e-01 1.92664266e-01 -2.78178036e-01
2.42297724e-01 -1.09932804e+00 -4.75741215e-02 1.06881177e+00
-2.09969804e-01 6.99706376e-01 -1.70694038e-01 3.64633918e-01
-2.21629336e-01 -3.11773658e-01 4.84995157e-01 -4.40788239e-01
-7.87186086e-01 3.57588381e-01 -9.32478309e-01 -2.64295399e-01
1.49173510e+00 3.68111342e-01 -5.72929263e-01 -2.33279124e-01
-7.59633839e-01 1.15500891e+00 -1.79565638e-01 1.79855153e-01
5.02977014e-01 -1.47112465e+00 -9.92408514e-01 3.16092074e-01
-6.28184199e-01 5.94120957e-02 2.19787896e-01 9.38602090e-01
-4.09815609e-01 5.86909726e-02 -3.82052124e-01 -2.17289820e-01
-1.36581790e+00 3.52388203e-01 3.50396901e-01 -5.92663229e-01
-3.75887215e-01 4.70530987e-01 -2.64217049e-01 -1.66328460e-01
5.20299256e-01 -1.97174288e-02 -5.94064057e-01 -8.18194970e-02
5.48992336e-01 7.03578711e-01 -2.65673995e-01 -3.65918815e-01
-7.69656077e-02 3.11522126e-01 -1.33101538e-01 -7.71701932e-02
1.28113627e+00 -4.81536359e-01 -6.31298125e-01 6.89720586e-02
6.73284352e-01 -2.83029705e-01 -1.00476956e+00 -3.51043582e-01
-3.64686638e-01 -2.97426313e-01 1.50780320e-01 -7.14172423e-01
-1.11506081e+00 2.25107431e-01 4.28678125e-01 5.67588031e-01
1.01524591e+00 -4.63995695e-01 7.56660700e-01 2.44395494e-01
3.56508195e-01 -1.67676902e+00 -1.24042287e-01 7.29865611e-01
8.10111046e-01 -1.09437549e+00 1.39800578e-01 -8.87218788e-02
-7.43970096e-01 1.22520018e+00 8.62266958e-01 -1.23771384e-01
7.25022137e-01 2.55270690e-01 1.98749721e-01 -6.28302321e-02
-1.33927393e+00 -1.04864895e-01 -3.17324936e-01 6.00038290e-01
-6.09372929e-02 2.85723418e-01 -6.71711981e-01 4.53355640e-01
-1.81038883e-02 -7.07143396e-02 2.14437649e-01 1.03478181e+00
-8.35841179e-01 -1.36005402e+00 -6.71423376e-01 6.49386272e-02
-2.79117703e-01 2.73442864e-01 3.29639226e-01 4.51713175e-01
5.97147167e-01 1.32903779e+00 1.77067742e-01 -3.37548673e-01
8.97005796e-02 -2.79364198e-01 -1.79575235e-01 -3.03190529e-01
-1.10982859e+00 2.44764879e-01 1.84969872e-01 -4.23134208e-01
-6.04761064e-01 -2.46383876e-01 -1.57022154e+00 -7.87857473e-01
-8.23991075e-02 7.04972088e-01 6.35961533e-01 1.16424835e+00
6.39513016e-01 1.10078812e+00 9.18422759e-01 -8.43868971e-01
-7.65655935e-01 -1.05604362e+00 -3.38075042e-01 3.56037259e-01
8.95010903e-02 -5.14732122e-01 -5.86507097e-03 -3.36252898e-01] | [3.7771575450897217, 2.0258948802948] |
a3ba69af-3400-4f94-bbda-60d9b09cfaa5 | kqgc-knowledge-graph-embedding-with-smoothing | 2205.12102 | null | https://arxiv.org/abs/2205.12102v1 | https://arxiv.org/pdf/2205.12102v1.pdf | KQGC: Knowledge Graph Embedding with Smoothing Effects of Graph Convolutions for Recommendation | Leveraging graphs on recommender systems has gained popularity with the development of graph representation learning (GRL). In particular, knowledge graph embedding (KGE) and graph neural networks (GNNs) are representative GRL approaches, which have achieved the state-of-the-art performance on several recommendation tasks. Furthermore, combination of KGE and GNNs (KG-GNNs) has been explored and found effective in many academic literatures. One of the main characteristics of GNNs is their ability to retain structural properties among neighbors in the resulting dense representation, which is usually coined as smoothing. The smoothing is specially desired in the presence of homophilic graphs, such as the ones we find on recommender systems. In this paper, we propose a new model for recommender systems named Knowledge Query-based Graph Convolution (KQGC). In contrast to exisiting KG-GNNs, KQGC focuses on the smoothing, and leverages a simple linear graph convolution for smoothing KGE. A pre-trained KGE is fed into KQGC, and it is smoothed by aggregating neighbor knowledge queries, which allow entity-embeddings to be aligned on appropriate vector points for smoothing KGE effectively. We apply the proposed KQGC to a recommendation task that aims prospective users for specific products. Extensive experiments on a real E-commerce dataset demonstrate the effectiveness of KQGC. | ['Pablo Loyola', 'Takuma Ebisu', 'Manoj Kondapaka', 'Satyen Abrol', 'Yu Hirate', 'Md Mostafizur Rahman', 'Toyotaro Suzumura', 'Daisuke Kikuta'] | 2022-05-23 | null | null | null | null | ['entity-embeddings'] | ['methodology'] | [-3.27067733e-01 2.57973254e-01 -4.11249757e-01 -1.77338824e-01
4.86304201e-02 -3.48862350e-01 5.80139697e-01 2.71620035e-01
1.22536927e-01 2.26761162e-01 6.28818095e-01 -4.78591591e-01
-5.29858053e-01 -1.27331126e+00 -6.87013030e-01 -5.56426406e-01
-1.64917395e-01 -7.10958838e-02 1.34276196e-01 -5.70851386e-01
3.19133922e-02 2.83099890e-01 -1.02341342e+00 3.48800048e-02
1.02802992e+00 8.72900069e-01 -2.31797155e-02 2.51350015e-01
-3.79842907e-01 4.80035871e-01 -1.12792917e-01 -9.39091265e-01
2.52379864e-01 -7.74763227e-02 -3.51875782e-01 -1.78765669e-01
3.57201189e-01 -7.05455169e-02 -1.03606653e+00 1.08120406e+00
2.65977323e-01 5.55658937e-01 5.98566651e-01 -1.26809335e+00
-1.71004176e+00 9.23928022e-01 -4.06797528e-01 1.78267404e-01
1.62297547e-01 -1.41911089e-01 1.42905092e+00 -9.26450908e-01
3.93178552e-01 1.34232008e+00 7.17076063e-01 1.61512569e-01
-1.18955743e+00 -5.33354282e-01 5.49097240e-01 1.25174358e-01
-1.28167665e+00 1.25837281e-01 7.91256428e-01 -2.70082086e-01
9.61710334e-01 4.12183329e-02 8.24666977e-01 7.95062542e-01
3.01041394e-01 7.48310030e-01 4.86530900e-01 -1.38042703e-01
1.57193661e-01 6.40455037e-02 5.66816092e-01 7.67758429e-01
5.94606817e-01 1.44746661e-01 -3.08118969e-01 -2.62103051e-01
1.06186569e+00 6.73870385e-01 -3.84526163e-01 -5.29765010e-01
-8.92997444e-01 1.14593875e+00 1.15386701e+00 1.26724094e-01
-5.87482929e-01 1.71253830e-01 3.64378482e-01 4.08033758e-01
6.40713215e-01 4.26338643e-01 -1.83095977e-01 3.67258251e-01
-4.17514265e-01 1.15394255e-03 8.19757521e-01 1.01752353e+00
6.86328888e-01 2.02795491e-01 -2.36928210e-01 9.04282391e-01
4.78579253e-01 2.44464457e-01 5.11304021e-01 -2.05893531e-01
2.52147347e-01 1.03572381e+00 -3.58556688e-01 -1.60341549e+00
-2.08394960e-01 -7.55227149e-01 -9.61699545e-01 -4.17939514e-01
-2.45836340e-02 9.56737436e-03 -7.22723484e-01 1.44272923e+00
2.61651963e-01 7.02595413e-01 1.36252359e-01 8.27782154e-01
1.10236096e+00 7.75848746e-01 7.25669712e-02 1.80761114e-01
1.09672034e+00 -1.13650930e+00 -5.55759251e-01 1.52580947e-01
7.39872575e-01 -3.69827211e-01 1.10711789e+00 1.43691555e-01
-4.93348688e-01 -5.33710837e-01 -8.61352324e-01 2.57425494e-02
-7.52673686e-01 -1.23336844e-01 1.20948815e+00 5.86907744e-01
-1.11769521e+00 7.65476525e-01 -5.64444304e-01 -6.21330202e-01
5.39690256e-01 3.54683995e-01 -2.73656219e-01 -3.87111127e-01
-1.35871661e+00 4.99770790e-01 3.22360069e-01 1.66139051e-01
-3.51861477e-01 -8.44249308e-01 -9.19711709e-01 5.24405599e-01
6.11272931e-01 -6.34477496e-01 7.16306150e-01 -5.73132336e-01
-1.46943104e+00 9.61833298e-02 3.40659231e-01 -4.46384430e-01
-2.75995791e-01 -2.06451908e-01 -1.08424485e+00 -2.60535359e-01
-3.17320794e-01 1.01875901e-01 7.32292414e-01 -9.53383684e-01
-3.14440757e-01 -2.99046248e-01 4.91812557e-01 -2.00365903e-03
-7.73701727e-01 -3.32965583e-01 -4.69130009e-01 -9.33348536e-01
-5.30649014e-02 -8.49190652e-01 -4.28923607e-01 -4.61572737e-01
-3.40446621e-01 -6.04761660e-01 6.79503202e-01 -5.04811049e-01
1.77369034e+00 -2.24902964e+00 4.69031744e-02 5.77526569e-01
6.73624575e-01 6.00408256e-01 -5.16619325e-01 7.70351112e-01
-2.62213293e-02 2.06685606e-02 3.14782292e-01 1.47351250e-01
3.16226095e-01 3.18967611e-01 -4.40816641e-01 2.90697902e-01
1.19475112e-03 1.46625650e+00 -1.13085723e+00 4.60647419e-02
2.29905993e-01 6.97183967e-01 -8.37905288e-01 1.29024819e-01
-3.04333627e-01 -2.34194010e-01 -7.57263899e-01 4.17521268e-01
6.65675700e-01 -7.39566803e-01 3.63136411e-01 -5.54140747e-01
2.05147788e-01 2.26136282e-01 -1.06579494e+00 1.30775821e+00
-3.63547295e-01 -5.18336240e-03 -3.06689084e-01 -9.79294538e-01
1.15342867e+00 1.10311853e-02 2.80052841e-01 -6.49709702e-01
1.16242124e-02 -5.41636068e-03 3.16190757e-02 -1.12914346e-01
8.08634162e-01 2.55985826e-01 7.59758353e-02 4.85006690e-01
2.78946787e-01 3.25502664e-01 -9.73185897e-02 6.77966893e-01
1.05996525e+00 -1.16914332e-01 4.53128785e-01 -2.98095435e-01
5.21768570e-01 -4.81490225e-01 5.00070095e-01 6.39086187e-01
1.52936324e-01 1.93985254e-01 3.99707407e-01 -3.52073491e-01
-6.26409829e-01 -8.81830156e-01 3.05335850e-01 1.20354426e+00
1.86699778e-01 -9.33999479e-01 -3.71746570e-01 -1.02278483e+00
5.25525093e-01 6.15338087e-01 -6.06042087e-01 -4.58521187e-01
-3.15503061e-01 -4.77874428e-01 7.35849589e-02 6.09956443e-01
2.71526307e-01 -9.91792381e-01 4.62033212e-01 3.73448074e-01
5.71334541e-01 -6.72832668e-01 -8.62427771e-01 -1.43264771e-01
-7.88451910e-01 -1.01451659e+00 -6.18388534e-01 -7.07847178e-01
7.77244568e-01 9.56716716e-01 1.05535352e+00 2.40999162e-01
1.86585113e-01 7.31759727e-01 -7.44190931e-01 -6.99462071e-02
1.98062346e-03 2.29914322e-01 1.99540094e-01 3.63403350e-01
5.12520373e-01 -7.89894104e-01 -7.68556237e-01 2.05934793e-01
-1.02058280e+00 -2.47086823e-01 7.28430390e-01 8.49619925e-01
6.35823011e-01 1.06067166e-01 9.39475954e-01 -1.52068245e+00
1.13053751e+00 -9.61200774e-01 -5.10754466e-01 4.21785682e-01
-1.07560647e+00 -1.82270687e-02 9.35525715e-01 -5.01072884e-01
-7.32370198e-01 -5.15172184e-01 -5.80491498e-02 -7.89692104e-01
2.78662205e-01 1.15338624e+00 -9.43062361e-03 -3.32822412e-01
3.80067885e-01 6.70037642e-02 -1.83611050e-01 -5.85174799e-01
1.12294304e+00 4.83392805e-01 2.60631531e-01 -2.99296886e-01
8.43033612e-01 2.87598604e-03 -1.73905075e-01 -6.91322684e-01
-7.32769430e-01 -5.33235610e-01 -1.51315108e-01 4.13171276e-02
5.03016591e-01 -7.16455996e-01 -7.20657468e-01 -1.65738557e-02
-6.29961848e-01 -1.24735661e-01 -3.63900572e-01 7.07326353e-01
-1.50287496e-02 5.02336323e-01 -6.09796941e-01 -5.37408352e-01
-4.77584690e-01 -6.62842870e-01 5.34323633e-01 5.46059728e-01
2.28889197e-01 -1.42993963e+00 2.36199494e-03 -2.45640855e-02
7.23370314e-01 -3.11187804e-02 1.13748932e+00 -9.94111538e-01
-6.96649849e-01 -2.93124318e-01 -5.58398604e-01 3.23796839e-01
3.11296910e-01 -2.70453751e-01 -5.09212613e-01 -5.17808557e-01
-5.11366248e-01 1.09554864e-01 8.74132931e-01 4.17493612e-01
1.02145159e+00 -4.47913170e-01 -4.19746161e-01 7.28274524e-01
1.44077325e+00 -1.33163288e-01 5.38814723e-01 2.67437436e-02
1.16572917e+00 1.48710966e-01 4.74319607e-01 2.43996024e-01
6.25641108e-01 4.95781302e-01 3.91384989e-01 -1.44818723e-01
-1.43778816e-01 -7.17304707e-01 3.10694188e-01 1.40233397e+00
-1.79234982e-01 -3.04540873e-01 -3.60341251e-01 3.05185646e-01
-2.13002825e+00 -6.64046288e-01 -1.18119374e-01 1.98110926e+00
1.96889833e-01 -1.13329194e-01 1.34797424e-01 -3.23754042e-01
7.38062084e-01 4.39015567e-01 -6.15756989e-01 -4.05877978e-01
-1.91614889e-02 4.27144736e-01 5.94572961e-01 3.18854898e-01
-9.93541777e-01 1.08086252e+00 5.12026930e+00 7.89459884e-01
-9.70517874e-01 1.52525036e-02 1.78460464e-01 2.50091702e-01
-6.49660468e-01 7.54920170e-02 -6.58662379e-01 4.10483688e-01
8.75842214e-01 -5.98964691e-01 6.45598412e-01 1.03796768e+00
-1.61482871e-01 5.49739122e-01 -7.99548745e-01 1.04710853e+00
9.14582759e-02 -1.56026387e+00 3.85276526e-01 2.44863927e-01
8.79768789e-01 4.62798476e-02 2.66805947e-01 8.99127543e-01
7.82350481e-01 -9.58799541e-01 -5.42919785e-02 5.67572951e-01
4.96025324e-01 -8.19031179e-01 8.10145378e-01 -3.37108448e-02
-1.40390313e+00 -2.07100049e-01 -8.83706331e-01 1.86723009e-01
1.13887124e-01 8.38226080e-01 -6.59207463e-01 1.21515286e+00
4.03841764e-01 1.26062965e+00 -5.04021823e-01 1.05071890e+00
-2.51900733e-01 9.81611252e-01 1.79847553e-02 -1.73758835e-01
3.96414369e-01 -8.28255177e-01 2.73777038e-01 1.13832211e+00
4.23012793e-01 3.19301903e-01 2.98608363e-01 8.09986711e-01
-4.45279777e-01 5.19801080e-01 -5.62961459e-01 -5.78958988e-01
4.31847423e-01 1.57842696e+00 -4.77323145e-01 -2.13843092e-01
-9.41409290e-01 7.39209771e-01 5.85670471e-01 5.44600964e-01
-5.55603147e-01 -4.45917070e-01 7.38797903e-01 4.38565873e-02
7.16153026e-01 -4.98911627e-02 4.33224618e-01 -1.27006161e+00
-3.30622733e-01 -7.28665471e-01 7.13171482e-01 -4.07687187e-01
-2.02735424e+00 6.27726257e-01 -4.23648030e-01 -1.07287610e+00
3.40011537e-01 -5.79793990e-01 -7.19337404e-01 9.61367011e-01
-1.77043438e+00 -1.35637701e+00 -2.98323005e-01 7.66673148e-01
-8.61087367e-02 -3.85164142e-01 8.04693103e-01 4.48115021e-01
-6.52562916e-01 8.40750456e-01 3.68856341e-01 7.14339390e-02
4.67104167e-01 -1.27369106e+00 7.25537717e-01 5.03657937e-01
3.86629045e-01 1.14564764e+00 1.26235917e-01 -7.85190582e-01
-1.88891518e+00 -1.61244810e+00 5.73607147e-01 -1.71394587e-01
9.83072221e-01 -2.28882536e-01 -1.23793399e+00 8.81754339e-01
-4.66171978e-03 4.30978715e-01 8.58036995e-01 6.04253590e-01
-6.52023256e-01 -1.87789395e-01 -7.75440574e-01 5.61605632e-01
1.22460365e+00 -6.11927211e-01 -4.73736107e-01 3.97549123e-01
1.02076375e+00 -1.11457631e-01 -1.25962186e+00 1.14704549e-01
4.29649651e-01 -7.87708342e-01 1.05979574e+00 -9.49770570e-01
1.84719130e-01 -2.91738719e-01 -1.47108540e-01 -1.77269888e+00
-9.90781844e-01 -6.09542370e-01 -5.53525031e-01 1.21805751e+00
1.46215826e-01 -1.02602470e+00 8.05348873e-01 3.55115801e-01
-2.42510781e-01 -9.92837369e-01 -2.55820453e-01 -8.42643976e-01
-1.22074164e-01 -3.94185893e-02 9.98022556e-01 1.21088290e+00
1.36539742e-01 4.21169609e-01 -4.20255423e-01 3.32451522e-01
3.52201402e-01 3.20507348e-01 8.11837077e-01 -1.34289920e+00
-4.89827663e-01 -3.97169739e-01 -6.24403119e-01 -1.29988205e+00
1.03014596e-01 -1.63610935e+00 -6.51156664e-01 -1.79190028e+00
-6.85531795e-02 -5.96313715e-01 -8.31853211e-01 2.96368778e-01
-4.20985311e-01 -1.38809318e-02 1.96351990e-01 9.08867195e-02
-7.65154421e-01 7.66519725e-01 1.36616385e+00 5.86370565e-02
-4.14444894e-01 7.11491257e-02 -1.18846393e+00 3.69382501e-01
4.82647121e-01 -5.90482280e-02 -7.37521768e-01 -1.80503026e-01
4.54816282e-01 -1.37959689e-01 1.11106403e-01 -5.53128719e-01
3.18358332e-01 -4.65738513e-02 3.48723866e-03 -2.97997504e-01
3.22761051e-02 -7.13862419e-01 2.97614425e-01 2.68934807e-03
-1.98718846e-01 -3.90075967e-02 -1.54365540e-01 1.22230256e+00
-2.11944744e-01 2.57851958e-01 3.01780194e-01 1.60108075e-01
-9.00806606e-01 9.44268942e-01 3.64155561e-01 -3.04149330e-01
6.86471820e-01 -8.00419301e-02 -5.41749597e-01 -4.50833291e-01
-6.18947506e-01 2.96593189e-01 1.05798133e-01 6.05818033e-01
6.93323612e-01 -1.69055331e+00 -4.22935456e-01 2.44547486e-01
3.95491183e-01 -3.75871450e-01 3.50188822e-01 8.35244834e-01
7.59697258e-02 4.62490648e-01 1.83471918e-01 -7.99227133e-02
-8.72653365e-01 9.97549891e-01 -3.52002606e-02 -3.99618745e-01
-1.02469087e+00 8.70333731e-01 3.96903872e-01 -6.52382910e-01
1.02334835e-01 -5.29493988e-01 -4.75178182e-01 -8.74887109e-02
4.29898739e-01 3.31790686e-01 -2.77071036e-02 -4.34666485e-01
-1.14194907e-01 2.69064724e-01 -4.58479196e-01 7.54564226e-01
1.62202859e+00 -1.75075158e-02 -1.28328562e-01 7.49660358e-02
1.10050488e+00 1.65989161e-01 -8.74046445e-01 -8.19587231e-01
-1.22378640e-01 -5.61631560e-01 2.83413738e-01 -5.23554385e-01
-1.50593150e+00 5.66645920e-01 1.47855401e-01 5.80805540e-01
7.99393952e-01 -4.15634960e-02 1.13230383e+00 3.31847459e-01
4.33367312e-01 -7.03546345e-01 -3.69215012e-03 5.46487272e-01
6.50715113e-01 -8.70655417e-01 9.73608792e-02 -6.71890378e-01
-6.63344860e-01 1.06885827e+00 3.85871053e-01 -5.96982419e-01
1.20796800e+00 -2.69574970e-01 -3.09981585e-01 -4.71111357e-01
-4.37316746e-01 -3.30761880e-01 5.65344989e-01 5.57036519e-01
4.01203364e-01 4.00663584e-01 -3.59716505e-01 1.22302055e+00
-7.09546581e-02 1.10315338e-01 3.09764236e-01 5.65454721e-01
-2.12691039e-01 -1.14102268e+00 3.22286367e-01 1.09786069e+00
-1.24378696e-01 -3.69182527e-01 -2.28656754e-01 5.08386791e-01
-2.02145576e-01 8.41338336e-01 -2.59354532e-01 -8.39893997e-01
6.21044874e-01 -1.99401423e-01 1.25365943e-01 -8.12969685e-01
-7.86743224e-01 -1.74164280e-01 -7.82170668e-02 -6.56580269e-01
-1.03017695e-01 -1.14221849e-01 -1.09144247e+00 -5.46515226e-01
-6.72722101e-01 4.45532173e-01 2.52969980e-01 5.59752345e-01
9.03468132e-01 6.72288895e-01 7.35697567e-01 -5.48988640e-01
-6.23111188e-01 -7.45996118e-01 -1.26662111e+00 6.13526702e-01
-3.55439298e-02 -8.34996462e-01 -2.83579528e-01 -4.47071999e-01] | [10.203763008117676, 5.648401737213135] |
44b47356-38a4-4984-a9f3-6c0f06e08cd4 | conformalized-semi-supervised-random-forest | 2302.02237 | null | https://arxiv.org/abs/2302.02237v1 | https://arxiv.org/pdf/2302.02237v1.pdf | Conformalized semi-supervised random forest for classification and abnormality detection | Traditional classifiers infer labels under the premise that the training and test samples are generated from the same distribution. This assumption can be problematic for safety-critical applications such as medical diagnosis and network attack detection. In this paper, we consider the multi-class classification problem when the training data and the test data may have different distributions. We propose conformalized semi-supervised random forest (CSForest), which constructs set-valued predictions $C(x)$ to include the correct class label with desired probability while detecting outliers efficiently. We compare the proposed method to other state-of-art methods in both a synthetic example and a real data application to demonstrate the strength of our proposal. | ['Leying Guan', 'Mingwenchan Xu', 'Yujin Han'] | 2023-02-04 | null | null | null | null | ['medical-diagnosis'] | ['medical'] | [ 3.83179694e-01 4.45509590e-02 -3.28565717e-01 -6.95376754e-01
-6.23614907e-01 -3.77677083e-01 2.34833121e-01 6.01266563e-01
-1.31428704e-01 1.16255236e+00 -6.24299645e-01 -6.44856572e-01
-2.43557259e-01 -8.73648643e-01 -6.53131425e-01 -8.17518234e-01
-2.31374055e-01 6.45809531e-01 1.63501292e-01 4.40497190e-01
2.49551088e-01 6.02266431e-01 -1.43181801e+00 4.06937182e-01
1.07002211e+00 1.19533741e+00 -7.98974991e-01 4.52546537e-01
2.63356149e-01 9.19672072e-01 -9.77164030e-01 -3.70610178e-01
4.40814108e-01 -2.55358309e-01 -5.59848845e-01 3.20931226e-02
2.72323579e-01 6.89124735e-03 3.30142587e-01 1.22166204e+00
2.71248460e-01 -1.78140312e-01 1.10494792e+00 -2.09267306e+00
-4.85089347e-02 3.71400684e-01 -6.23554409e-01 -2.78645903e-02
5.64668216e-02 -1.93107769e-01 7.22108603e-01 -5.32769263e-01
2.23046169e-01 8.99298429e-01 8.51327777e-01 3.99035454e-01
-1.31687379e+00 -1.09995806e+00 -2.89384881e-03 3.72630030e-01
-1.47576571e+00 -8.51997733e-03 6.38309956e-01 -4.97705102e-01
4.42547262e-01 2.83162445e-01 2.17080161e-01 1.02122569e+00
6.30492747e-01 4.95511711e-01 1.52516103e+00 -3.05711597e-01
6.39226973e-01 5.18303633e-01 3.20737690e-01 4.69179690e-01
5.57702899e-01 3.92901033e-01 -1.44181326e-01 -8.56607735e-01
9.98742059e-02 4.69309539e-01 -6.70416504e-02 -3.37447912e-01
-8.49401295e-01 9.91349936e-01 -6.93815276e-02 -2.12180480e-01
-3.28249246e-01 -2.55621910e-01 3.85480344e-01 4.53044087e-01
4.31052625e-01 3.03484052e-02 -5.69157720e-01 2.67478377e-01
-9.91150439e-01 3.36745009e-02 8.81686449e-01 8.78093600e-01
4.99764264e-01 2.43406929e-02 -8.93749576e-03 4.22742128e-01
2.28443041e-01 4.36919034e-01 3.72632742e-01 -4.37030971e-01
2.62294143e-01 4.70311254e-01 2.08255351e-01 -6.92586601e-01
-2.41042703e-01 -6.16574466e-01 -9.74955380e-01 5.79731703e-01
4.80815887e-01 -3.37769806e-01 -9.08189952e-01 1.41060781e+00
4.50993747e-01 5.88774502e-01 2.13918075e-01 3.67048293e-01
2.66162515e-01 3.72683972e-01 8.67021754e-02 -5.13903141e-01
6.74689293e-01 -5.93352914e-01 -6.04683757e-01 9.46439356e-02
6.40887082e-01 -5.95045686e-01 7.48655796e-01 8.75899732e-01
-3.66217136e-01 -3.69152427e-01 -1.08741999e+00 8.89991820e-01
-1.27406985e-01 -2.09330209e-02 2.93751210e-01 9.10890877e-01
-2.95827568e-01 7.32721567e-01 -6.02638066e-01 -1.81913488e-02
5.55420578e-01 3.32016945e-01 -3.54166776e-01 -1.63648218e-01
-9.22764301e-01 7.59386897e-01 3.16798776e-01 -5.98946847e-02
-9.43103731e-01 -3.03165019e-01 -4.36355650e-01 -2.20019400e-01
3.37341487e-01 -2.30405241e-01 9.94590878e-01 -1.00125039e+00
-9.59371030e-01 5.57903230e-01 -1.33541137e-01 -6.46723509e-01
7.73665786e-01 -1.48681194e-01 -7.77601600e-01 -1.03806697e-01
1.41751304e-01 -7.33071864e-02 1.06876671e+00 -1.12454915e+00
-8.98660481e-01 -4.49810594e-01 -4.76959229e-01 -3.71397704e-01
1.04034208e-01 -1.39743060e-01 5.93805969e-01 -5.85802197e-01
1.07801102e-01 -8.98791790e-01 -3.40189904e-01 -1.65278509e-01
-1.02794838e+00 -7.50965625e-02 1.00860643e+00 -3.12812984e-01
1.09046316e+00 -2.09642434e+00 -5.16275167e-01 8.68769526e-01
-8.51622690e-03 -4.09487728e-03 4.37453181e-01 2.87957311e-01
-6.03820205e-01 -5.23864524e-03 -4.93791103e-01 -1.37667377e-02
-3.96744817e-01 1.52169138e-01 -6.50252521e-01 9.31887329e-01
1.85008571e-01 1.23411883e-02 -6.47679567e-01 -5.87866545e-01
-7.47499913e-02 -3.28683257e-02 -1.68547735e-01 3.71450692e-01
-5.40768877e-02 4.90182966e-01 -3.92492950e-01 8.65554154e-01
6.44879997e-01 -5.15675694e-02 4.30712141e-02 1.73975557e-01
3.81509453e-01 -1.79860562e-01 -1.43555307e+00 6.32427990e-01
-5.84298968e-02 2.15463683e-01 -5.00245273e-01 -1.38948011e+00
1.23769188e+00 4.34280813e-01 5.09718835e-01 -1.10622741e-01
1.47446588e-01 1.54745296e-01 2.40850195e-01 -2.46118054e-01
-8.78861696e-02 -6.11986697e-01 -1.61754802e-01 5.60418606e-01
-2.40834445e-01 8.45784619e-02 -2.30424866e-01 -1.33232921e-01
1.23761320e+00 -2.49825954e-01 7.14297652e-01 -2.66746342e-01
4.17541891e-01 9.03917418e-04 9.81258214e-01 8.15492988e-01
-3.46676469e-01 4.10234600e-01 6.81738853e-01 -3.73829842e-01
-6.89768910e-01 -1.36920488e+00 -5.24250984e-01 5.07483304e-01
-1.56311989e-01 3.63091193e-03 -5.07170200e-01 -1.38784552e+00
2.92272210e-01 1.07763982e+00 -5.33275247e-01 -1.73472598e-01
-2.27083176e-01 -9.49720502e-01 4.91539121e-01 3.33336741e-01
1.11419514e-01 -8.03570509e-01 -6.26506209e-01 1.01749144e-01
1.28254578e-01 -9.80451524e-01 -1.06180623e-01 6.73643708e-01
-8.56266797e-01 -1.60976374e+00 -6.46681001e-04 -6.23801768e-01
1.05768883e+00 -3.09391499e-01 8.50660682e-01 1.21961571e-02
-2.73700297e-01 -7.86218494e-02 -3.53080660e-01 -7.57158577e-01
-7.50549674e-01 -2.31388003e-01 3.35048884e-01 3.13042730e-01
4.16087598e-01 -5.42063534e-01 -2.62327433e-01 5.98451614e-01
-8.34068716e-01 -4.11382854e-01 4.13888156e-01 1.08329618e+00
6.70016825e-01 7.57488072e-01 1.09403849e+00 -1.62709796e+00
5.76138437e-01 -8.66395831e-01 -5.97629905e-01 3.52289766e-01
-1.03445280e+00 -5.43038957e-02 1.11562204e+00 -6.18156314e-01
-5.34499049e-01 1.22410320e-01 6.71656244e-03 -4.34746265e-01
-5.71148336e-01 2.43608564e-01 -2.72644013e-01 1.10193230e-01
7.75238514e-01 1.03995502e-02 -2.23958436e-02 -2.58949846e-01
-2.00701922e-01 1.10601449e+00 5.37621677e-01 -5.54039240e-01
8.72904480e-01 3.71586204e-01 3.64971340e-01 -3.93065661e-01
-8.05414319e-01 -3.00183296e-01 -5.47799528e-01 -1.11727394e-01
2.29511395e-01 -6.38126016e-01 -6.31518960e-01 3.19252908e-01
-7.24478006e-01 -1.18209578e-01 -3.68417501e-01 5.49871027e-01
-4.81054842e-01 3.30743641e-01 -2.68166125e-01 -1.25822115e+00
-3.56285810e-01 -9.82588053e-01 6.71718538e-01 -5.82439192e-02
-3.11844766e-01 -8.97113144e-01 1.89523343e-02 -3.91393825e-02
-3.83923240e-02 8.11813653e-01 1.10104156e+00 -1.46845675e+00
-4.40002121e-02 -8.72249424e-01 1.41779453e-01 6.90793812e-01
2.44859204e-01 1.39042899e-01 -9.24483836e-01 -5.44011652e-01
2.30541706e-01 -2.45418832e-01 4.40534174e-01 2.31805518e-01
1.44648921e+00 -4.37240034e-01 -4.71312255e-01 3.31300914e-01
1.36410677e+00 4.73694891e-01 4.37317818e-01 1.00115746e-01
2.87860692e-01 5.84075332e-01 1.09627199e+00 6.91730201e-01
-2.06702113e-01 3.43586534e-01 4.96974498e-01 3.31582576e-02
5.88827610e-01 -2.99907297e-01 2.41428703e-01 2.88151205e-01
4.26151901e-01 -1.94609657e-01 -1.01352501e+00 4.89068627e-01
-1.83540463e+00 -9.99924362e-01 -1.34708077e-01 2.56441808e+00
7.44405031e-01 6.18559301e-01 2.35126689e-01 8.43392849e-01
1.08447421e+00 -5.39024651e-01 -8.07923377e-01 -3.02403897e-01
1.60253704e-01 5.22407770e-01 5.89010954e-01 3.03384542e-01
-1.15885854e+00 2.15261921e-01 6.54047680e+00 8.80285978e-01
-1.27850962e+00 -3.26943360e-02 1.03143311e+00 7.99980164e-02
8.09899196e-02 3.36547121e-02 -7.64942944e-01 6.42579257e-01
1.15764320e+00 -1.35710478e-01 -1.88345999e-01 1.15231681e+00
-8.05358738e-02 -1.78129837e-01 -1.33047307e+00 7.95748591e-01
2.33431486e-03 -7.50602365e-01 -3.44319731e-01 -2.46924460e-02
4.72670615e-01 -3.20521533e-01 3.45955603e-02 1.46241099e-01
4.67307717e-01 -1.16614306e+00 6.56424165e-01 4.01524007e-01
1.01065147e+00 -9.11983669e-01 1.08552110e+00 7.16874599e-01
-6.76052272e-01 -3.01564813e-01 4.56351135e-03 -8.87323767e-02
-1.14938073e-01 1.10845113e+00 -1.23200214e+00 4.99360681e-01
5.66437483e-01 6.39822543e-01 -4.82668459e-01 1.23826230e+00
-2.13387564e-01 1.17557001e+00 -3.68140817e-01 2.44704317e-02
-2.27580473e-01 2.66387351e-02 3.79310161e-01 8.82994711e-01
3.46165299e-01 -1.90405935e-01 4.63739365e-01 4.14565086e-01
3.22039247e-01 1.56194523e-01 -6.39375508e-01 3.92212868e-01
6.35875344e-01 8.76873195e-01 -9.11246836e-01 -3.39839548e-01
-3.53307903e-01 5.19704938e-01 1.79607227e-01 1.79606199e-01
-1.01425469e+00 -5.35450697e-01 3.85517955e-01 2.58607298e-01
1.20228149e-01 2.13386297e-01 -5.72537005e-01 -9.51962650e-01
5.72760552e-02 -7.38137543e-01 6.41343594e-01 -9.54755321e-02
-1.68990338e+00 7.16880143e-01 8.61683860e-02 -1.94928813e+00
-2.75298923e-01 -5.21059453e-01 -8.02048385e-01 6.89564109e-01
-1.31739414e+00 -7.40417480e-01 -8.75709802e-02 7.48753369e-01
1.29690124e-02 -3.98061991e-01 1.06182003e+00 9.56189930e-02
-6.86510742e-01 8.47136080e-01 3.37833256e-01 1.87865704e-01
9.74157274e-01 -1.20616829e+00 2.53888723e-02 8.42880785e-01
-1.33455083e-01 2.28816941e-01 9.37692821e-01 -7.79562950e-01
-4.62276399e-01 -1.51273870e+00 4.62085128e-01 -2.64905959e-01
4.34958160e-01 -6.09266870e-02 -8.25582385e-01 6.89145744e-01
-3.46603721e-01 5.54276526e-01 1.25931048e+00 1.11738414e-01
-5.38396478e-01 -2.97644943e-01 -1.87010121e+00 7.74319023e-02
4.48846549e-01 -2.37311050e-01 -4.61673558e-01 4.21602398e-01
1.21584751e-01 -2.09741265e-01 -9.09190536e-01 5.76397359e-01
4.12805885e-01 -9.95025218e-01 5.96361399e-01 -7.21032739e-01
1.13399066e-01 -4.38929945e-01 -3.12875181e-01 -1.41910684e+00
4.82327826e-02 -3.42381120e-01 -6.01596907e-02 1.01587415e+00
5.00342786e-01 -9.68705118e-01 8.36111486e-01 4.28135097e-01
1.93039447e-01 -8.93064678e-01 -1.21016419e+00 -8.45840096e-01
-1.18061397e-02 -5.87267876e-01 6.56386793e-01 1.09993720e+00
2.54636914e-01 1.59689993e-01 -5.16854763e-01 4.04814720e-01
9.86633718e-01 2.11555466e-01 6.05484903e-01 -1.68284726e+00
-4.34061944e-01 1.78223699e-01 -7.64787972e-01 -1.08881623e-01
3.35172892e-01 -6.88699782e-01 1.37912989e-01 -6.58620477e-01
2.16201648e-01 -8.23897660e-01 -5.50134182e-01 6.35932028e-01
-2.93210328e-01 1.30673006e-01 -3.02914500e-01 1.13027103e-01
-3.67000610e-01 2.09526896e-01 5.89438617e-01 -8.90954677e-03
1.78079843e-03 8.80415559e-01 -6.54387295e-01 8.57812285e-01
1.01074564e+00 -1.02482915e+00 -4.63791668e-01 3.56310904e-01
-2.42393047e-01 3.76216829e-01 2.14819893e-01 -1.18938386e+00
5.19044371e-03 -5.14732003e-01 4.58184361e-01 -4.68514591e-01
-6.84367493e-02 -1.23626447e+00 2.30368271e-01 7.55685270e-01
-4.11674917e-01 1.66917250e-01 -1.11637630e-01 8.75535607e-01
-2.92257249e-01 -2.58783072e-01 1.04624486e+00 4.06480193e-01
-2.88884435e-02 3.27586532e-01 -1.11390680e-01 -1.29245922e-01
1.58229876e+00 -1.43190518e-01 -3.94575387e-01 -4.34737712e-01
-6.83420837e-01 6.10953346e-02 3.18896919e-01 -3.79951783e-02
8.89884412e-01 -1.13829732e+00 -7.55576968e-01 5.96684277e-01
2.59714127e-01 9.24534574e-02 -2.38357902e-01 6.93149209e-01
-2.36610547e-01 1.04118064e-02 -1.82245135e-01 -6.06551766e-01
-1.41714346e+00 6.54154241e-01 2.71297008e-01 -2.55046666e-01
-2.91970462e-01 3.50197226e-01 -2.17951864e-01 -4.40265954e-01
3.14891398e-01 -1.92156807e-01 -1.29342766e-03 -3.70814234e-01
4.39701587e-01 5.24427116e-01 1.50589451e-01 -4.45112795e-01
-3.92775059e-01 9.09325257e-02 -2.19739407e-01 1.11969136e-01
1.07185566e+00 3.88585120e-01 -3.00797448e-02 6.98157609e-01
1.03152990e+00 3.98975387e-02 -9.03504550e-01 -2.85559058e-01
1.54254943e-01 -6.09333634e-01 -3.80798787e-01 -8.26490760e-01
-9.05115008e-01 8.71192932e-01 7.95090139e-01 3.28274131e-01
1.17859077e+00 -2.96029866e-01 4.55940872e-01 1.90439314e-01
6.37335837e-01 -9.39036191e-01 -1.80305764e-01 -4.62892950e-02
2.14916199e-01 -1.19652069e+00 1.23686448e-01 -4.91772294e-01
-7.67505467e-01 1.06087112e+00 6.62982583e-01 -3.56420159e-01
1.08666456e+00 3.51720512e-01 3.39202173e-02 2.92885005e-01
-9.46377277e-01 2.52249807e-01 -9.56212450e-03 8.21608126e-01
2.67418236e-01 4.13756311e-01 -1.69018149e-01 7.64059782e-01
-1.53597847e-01 2.04386432e-02 5.74158013e-01 1.16823673e+00
-2.36755043e-01 -1.11939204e+00 -6.86273634e-01 1.06118834e+00
-6.59473658e-01 1.42874509e-01 -3.44533660e-02 4.94860888e-01
3.51844370e-01 1.07421458e+00 -7.75851980e-02 -6.01147354e-01
3.31896722e-01 3.70185137e-01 -2.65632961e-02 -6.12693846e-01
-4.05439556e-01 -1.21227250e-01 -4.03381102e-02 -1.42082691e-01
-2.09800214e-01 -7.74188638e-01 -1.24165595e+00 -1.74483836e-01
-5.67098916e-01 5.93129873e-01 4.59908634e-01 9.64607596e-01
-1.21777263e-02 4.22653139e-01 1.10916400e+00 -9.78067368e-02
-9.18159902e-01 -9.43079829e-01 -9.98003066e-01 4.89266664e-01
4.47812676e-01 -7.89300919e-01 -6.45009339e-01 -6.95511028e-02] | [8.653070449829102, 4.027576446533203] |
1452dc18-e27e-4e47-826b-9ebe69e5a3c6 | efficient-automation-of-neural-network-design | 2304.05405 | null | https://arxiv.org/abs/2304.05405v2 | https://arxiv.org/pdf/2304.05405v2.pdf | Efficient Automation of Neural Network Design: A Survey on Differentiable Neural Architecture Search | In the past few years, Differentiable Neural Architecture Search (DNAS) rapidly imposed itself as the trending approach to automate the discovery of deep neural network architectures. This rise is mainly due to the popularity of DARTS, one of the first major DNAS methods. In contrast with previous works based on Reinforcement Learning or Evolutionary Algorithms, DNAS is faster by several orders of magnitude and uses fewer computational resources. In this comprehensive survey, we focus specifically on DNAS and review recent approaches in this field. Furthermore, we propose a novel challenge-based taxonomy to classify DNAS methods. We also discuss the contributions brought to DNAS in the past few years and its impact on the global NAS field. Finally, we conclude by giving some insights into future research directions for the DNAS field. | ['Hedi Tabia', 'Hichem Arioui', 'Ahmad Nasser', 'Alexandre Heuillet'] | 2023-04-11 | null | null | null | null | ['architecture-search'] | ['methodology'] | [ 1.91793498e-02 4.47800830e-02 -2.15220407e-01 -2.76354074e-01
-1.41461343e-01 -6.16421223e-01 6.46900833e-01 -9.08275247e-02
-6.94583893e-01 7.55932927e-01 -5.65534607e-02 -4.30923641e-01
-2.59425104e-01 -7.83703625e-01 -5.94071031e-01 -7.44354069e-01
1.39550656e-01 4.39219475e-01 9.36953202e-02 -4.77738470e-01
5.14230430e-01 8.76239419e-01 -1.92813551e+00 7.43208081e-02
7.71418035e-01 1.19135821e+00 2.51770318e-01 3.42645615e-01
-3.18107575e-01 4.58260059e-01 -6.69973254e-01 -4.16651338e-01
3.11062604e-01 -2.31599182e-01 -6.56392574e-01 -5.73357403e-01
3.99306476e-01 -1.32036641e-01 -5.84681511e-01 1.12659049e+00
4.92318213e-01 2.51939535e-01 6.40497804e-01 -1.19480884e+00
-6.70363307e-01 5.95650971e-01 -1.06054328e-01 3.05315554e-01
-2.09941313e-01 2.05768108e-01 1.03694594e+00 -7.85281539e-01
6.84599280e-01 1.31350362e+00 6.19841397e-01 8.24501812e-01
-6.82357013e-01 -5.04148543e-01 3.41640502e-01 7.45050073e-01
-1.17741394e+00 -2.87049979e-01 7.54307389e-01 -2.80978590e-01
1.41654432e+00 3.40971857e-01 9.52664375e-01 1.36009240e+00
3.45477387e-02 1.28012204e+00 8.49379718e-01 -5.48819244e-01
7.80922353e-01 -2.96799481e-01 3.77991617e-01 7.58371830e-01
3.38427126e-01 4.15178865e-01 -5.44637322e-01 1.49576012e-02
5.66059768e-01 -4.34587866e-01 2.79715478e-01 -2.90747434e-01
-9.87059414e-01 1.12452102e+00 5.59081197e-01 5.66446722e-01
-5.03937006e-01 2.56379932e-01 5.03441453e-01 1.47718772e-01
1.31916657e-01 9.46138024e-01 -4.33786631e-01 -4.82113451e-01
-8.93231034e-01 4.50424671e-01 7.38409519e-01 4.06813920e-01
3.15490842e-01 7.47583091e-01 1.50525570e-01 7.14800835e-01
2.08875656e-01 9.75791663e-02 5.69879472e-01 -9.45683479e-01
-6.21535350e-03 6.39805913e-01 -1.75299212e-01 -6.77763581e-01
-4.21574295e-01 -7.74370909e-01 -9.36611474e-01 2.73770779e-01
2.55724132e-01 -2.07416266e-01 -7.85992861e-01 1.53735387e+00
1.34758323e-01 7.20736384e-02 2.07035188e-02 7.63541937e-01
6.67197704e-01 6.01457000e-01 -1.89421996e-01 1.37328893e-01
9.23934281e-01 -1.05548930e+00 -3.14141303e-01 5.82131557e-03
2.22397283e-01 -2.41760090e-01 7.33342111e-01 4.83873278e-01
-9.29494262e-01 -4.12798554e-01 -1.21308005e+00 2.72230417e-01
-8.11777115e-01 -2.71285623e-02 8.18000972e-01 1.03121412e+00
-1.33998311e+00 6.91958010e-01 -1.00175953e+00 -3.19385201e-01
4.57562685e-01 5.10211527e-01 2.95996487e-01 3.25327218e-01
-1.24166024e+00 1.14722359e+00 7.73367763e-01 4.94476482e-02
-7.05822825e-01 -5.55489838e-01 -4.50993776e-01 1.18375018e-01
3.06090027e-01 -8.04774284e-01 1.22631168e+00 -9.19898987e-01
-1.83408380e+00 5.76014400e-01 -7.27041364e-02 -9.32512283e-01
2.00335696e-01 -3.69240418e-02 -2.68710315e-01 -2.75900900e-01
-4.74567473e-01 1.01472318e+00 4.15030658e-01 -7.10497022e-01
-7.36674666e-01 -3.39281410e-01 4.54617850e-02 1.40460089e-01
-6.56565785e-01 1.08994722e-01 -2.34975293e-01 -6.27261937e-01
-1.82953387e-01 -8.51664543e-01 -2.67525613e-01 -1.65276915e-01
-2.44326591e-01 -7.80518353e-01 3.90976250e-01 -2.76425511e-01
1.15613627e+00 -2.00468636e+00 6.30689502e-01 8.42546225e-02
1.27151042e-01 6.09439850e-01 -4.29120690e-01 4.94294554e-01
-9.24226828e-03 -3.29162665e-02 -2.30308816e-01 -3.76703925e-02
2.52156824e-01 3.30270708e-01 -4.17621225e-01 2.49951929e-01
1.48630917e-01 1.02875435e+00 -7.96170413e-01 2.01359421e-01
6.77131563e-02 3.87544006e-01 -3.90875638e-01 1.59389704e-01
-4.07624245e-01 1.42691374e-01 -1.92155987e-01 8.03237736e-01
2.61897027e-01 -3.38035449e-02 1.41508430e-01 -5.31258993e-02
-5.95243573e-01 3.05701017e-01 -7.87769020e-01 1.53065944e+00
-7.12559670e-02 8.71969402e-01 -2.13969797e-01 -1.40127265e+00
1.08959329e+00 1.17648747e-02 3.23132902e-01 -8.21323216e-01
3.74978840e-01 6.23888135e-01 4.66734111e-01 -2.31771186e-01
4.11994934e-01 2.26646587e-01 2.08409309e-01 9.05593038e-02
1.55117869e-01 2.95205712e-01 4.30835158e-01 -1.48515597e-01
9.45747674e-01 8.33035260e-02 4.77261066e-01 -3.41868460e-01
6.51172042e-01 1.11534998e-01 4.78202909e-01 7.58028209e-01
-2.82392412e-01 7.61698112e-02 1.77959129e-01 -1.04420948e+00
-1.20334029e+00 -9.95329559e-01 9.78629105e-03 1.03943050e+00
-1.07637011e-01 -4.11334708e-02 -9.63248432e-01 -5.65755188e-01
4.59140465e-02 7.26831079e-01 -7.39379644e-01 7.22429436e-03
-6.93150640e-01 -9.00679946e-01 9.51837361e-01 6.64486468e-01
8.02036762e-01 -1.47480321e+00 -8.46817493e-01 2.31185108e-01
1.87867358e-01 -6.37758255e-01 1.49126291e-01 2.36259744e-01
-1.06594169e+00 -8.61009896e-01 -1.00725627e+00 -8.08064163e-01
1.25421226e-01 -1.25614315e-01 9.31785941e-01 2.23697871e-01
-4.51778591e-01 5.40084764e-02 -2.38465831e-01 -6.11767173e-01
-5.22583783e-01 7.69952595e-01 3.46880853e-01 -3.23390782e-01
5.55590153e-01 -5.13152659e-01 -5.30196905e-01 -1.46889895e-01
-7.57018268e-01 8.56157765e-02 6.94721341e-01 9.57543015e-01
2.08178554e-02 -1.49966359e-01 8.40234935e-01 -4.30543125e-01
8.28779995e-01 -5.75521588e-01 -8.81008446e-01 3.01482499e-01
-9.76985931e-01 3.50173771e-01 5.26914477e-01 -3.07315826e-01
-8.08043838e-01 -1.12840652e-01 -4.94301140e-01 -1.95750535e-01
-4.20553595e-01 6.49326861e-01 3.41614634e-01 -2.52742350e-01
7.20028043e-01 4.83643383e-01 9.53255892e-02 -6.66747212e-01
1.62287086e-01 5.75366735e-01 4.27950293e-01 -4.57639158e-01
5.87766469e-01 4.71609607e-02 -1.25032827e-01 -1.14461982e+00
-5.14054120e-01 -1.37283042e-01 -4.83806074e-01 -3.19098622e-01
5.96415162e-01 -3.61236244e-01 -1.00353658e+00 8.79431069e-01
-1.32346940e+00 -2.72278160e-01 -5.23169041e-02 1.25742733e-01
-3.79466146e-01 2.91872084e-01 -6.63794041e-01 -1.00581479e+00
-6.32509410e-01 -1.25562477e+00 5.58498502e-01 5.49106240e-01
-2.17321187e-01 -1.02802718e+00 2.31599540e-01 2.09187388e-01
8.54172349e-01 1.70920435e-02 1.12198341e+00 -6.81375384e-01
-4.73363101e-01 1.64864793e-01 -1.60898402e-01 2.85545439e-01
-8.88556838e-02 -1.04138419e-01 -8.65879536e-01 -2.00021863e-01
-1.05949000e-01 -1.52056903e-01 8.99314046e-01 3.09450924e-01
1.14896178e+00 -2.40556806e-01 -2.34283209e-01 6.33026838e-01
1.40295577e+00 9.09209847e-01 6.34806752e-01 1.10355771e+00
2.86667734e-01 5.54150939e-01 4.63930219e-01 1.48760661e-01
8.69093910e-02 4.96921837e-01 6.77745283e-01 8.73220339e-02
-2.42983386e-01 8.49845782e-02 2.75996029e-01 1.19491577e+00
-2.74710596e-01 -3.59111458e-01 -1.18164945e+00 4.36948359e-01
-1.80712700e+00 -6.79580450e-01 4.79611009e-01 1.72419775e+00
6.24314785e-01 -5.83799258e-02 3.33910137e-01 2.37201259e-01
6.60864651e-01 -8.53180811e-02 -1.00230348e+00 -5.26147962e-01
-4.01029170e-01 4.17730004e-01 3.10754389e-01 2.04168841e-01
-9.39642727e-01 1.17470682e+00 7.45511627e+00 8.74361992e-01
-1.36394429e+00 -2.84943908e-01 6.31256878e-01 -3.50686572e-02
-1.27377495e-01 -4.89366114e-01 -9.00585532e-01 5.91593385e-01
1.09648931e+00 -9.25785676e-02 8.42025757e-01 1.11522174e+00
-3.10949385e-01 2.45917410e-01 -7.88450956e-01 9.26016450e-01
1.53829154e-04 -1.84256136e+00 1.34860232e-01 -2.34946962e-02
7.07063317e-01 5.99848747e-01 3.89596373e-01 3.12336296e-01
2.93063700e-01 -1.13642907e+00 8.26339424e-01 1.77771240e-01
4.85651195e-01 -9.04599547e-01 7.33914495e-01 2.26568058e-01
-1.00558734e+00 -4.64001656e-01 -3.83274168e-01 -1.55862659e-01
-1.89600468e-01 8.35186020e-02 -1.05637050e+00 4.24222887e-01
6.88555419e-01 4.02739227e-01 -4.05429840e-01 1.40986609e+00
-2.51239184e-02 7.50377059e-01 -2.47844696e-01 -8.37503254e-01
8.01286221e-01 -2.72285134e-01 5.83321750e-01 1.22900116e+00
2.77528942e-01 -1.73242882e-01 -1.74631521e-01 7.87914157e-01
8.23243037e-02 -1.23903193e-01 -3.86240631e-01 -3.22514862e-01
6.31368160e-01 9.72201705e-01 -7.58965850e-01 -3.16169113e-01
-2.91670412e-01 8.62424016e-01 3.94335151e-01 3.57137263e-01
-7.36622810e-01 -5.85273147e-01 8.43259215e-01 -3.87010992e-01
2.32339770e-01 -4.38176721e-01 -3.57076287e-01 -7.77137697e-01
-4.22869503e-01 -1.15342271e+00 2.00871184e-01 -5.04489303e-01
-1.18589950e+00 8.63796592e-01 -3.03587288e-01 -6.36637092e-01
-3.90059948e-01 -1.08237731e+00 -2.81953186e-01 5.45601666e-01
-1.52187181e+00 -6.90116882e-01 3.03297341e-02 1.33755652e-03
8.46471965e-01 -9.30214584e-01 1.06342292e+00 2.16589779e-01
-6.89097404e-01 5.72285056e-01 4.90363419e-01 3.06279957e-02
6.73095137e-02 -1.08123326e+00 1.22452688e+00 5.81339538e-01
2.73001760e-01 7.85657346e-01 6.51650429e-01 -4.07185227e-01
-1.58452773e+00 -6.46235704e-01 4.79736239e-01 1.38479069e-01
7.60746121e-01 -2.85956591e-01 -8.43288541e-01 2.91025549e-01
4.08982277e-01 -6.31564796e-01 3.82091939e-01 1.84003592e-01
-4.36238378e-01 1.62171528e-01 -9.42221642e-01 8.27312648e-01
1.06851470e+00 -2.91780651e-01 -5.94195664e-01 -1.15542099e-01
5.80608547e-01 -4.18128967e-01 -5.12300372e-01 4.75596488e-01
7.70395577e-01 -9.14291978e-01 9.91391778e-01 -8.47229660e-01
-4.30018082e-02 -3.08785047e-02 -2.63014644e-01 -1.40010762e+00
-5.89728892e-01 -5.04658043e-01 -2.55409569e-01 5.63661873e-01
2.55082548e-01 -8.78878593e-01 1.31160891e+00 2.66093284e-01
-3.51747513e-01 -1.14548063e+00 -1.27764142e+00 -1.08676124e+00
3.61903906e-01 -2.47532040e-01 7.02729881e-01 8.05989087e-01
-1.62260905e-01 1.18396446e-01 -6.13981076e-02 -2.80230194e-01
5.31598687e-01 -3.69228981e-02 2.37159774e-01 -1.41665590e+00
-2.56883919e-01 -1.03046012e+00 -5.67239702e-01 -9.52912211e-01
2.31347233e-01 -9.72504616e-01 -1.37686700e-01 -1.28563249e+00
-3.05182040e-02 -2.34260276e-01 -4.98318434e-01 4.45363998e-01
2.62538880e-01 2.14407355e-01 1.27377868e-01 -7.32431561e-02
-3.29063565e-01 5.81830680e-01 7.18986332e-01 -2.29322746e-01
3.33748944e-02 -8.31774026e-02 -5.91740608e-01 7.49923706e-01
1.26967692e+00 -1.00462355e-01 -2.74749458e-01 -5.95107198e-01
4.52735811e-01 -5.63376069e-01 2.11462751e-03 -1.16267931e+00
4.37321961e-01 -2.95890540e-01 2.18251854e-01 -5.76558948e-01
4.73586947e-01 -4.25477982e-01 -9.90490541e-02 9.29767907e-01
-3.09044689e-01 4.33435768e-01 4.08594668e-01 2.66993463e-01
-6.14724979e-02 -4.52126324e-01 8.15172851e-01 -2.91968226e-01
-9.89532113e-01 1.15965322e-01 -8.63759398e-01 -2.78037757e-01
7.83716977e-01 -2.94175923e-01 -2.54696518e-01 7.85362050e-02
-5.00553250e-01 4.31240164e-02 5.68730831e-02 7.72742450e-01
6.95090532e-01 -1.08590961e+00 -4.36673015e-01 7.77775072e-04
-2.14421511e-01 -4.12370503e-01 1.36938035e-01 2.90387332e-01
-5.60348630e-01 9.18272972e-01 -5.67641079e-01 -4.21093643e-01
-1.29924703e+00 4.35655206e-01 5.49424350e-01 1.07376695e-01
-3.37723672e-01 1.02763462e+00 -7.97836035e-02 -4.42957073e-01
7.69824982e-01 -1.69376284e-02 -3.48487437e-01 -1.34612590e-01
5.58251262e-01 7.43822217e-01 2.34876782e-01 -1.44409910e-01
-4.67487425e-01 2.02339411e-01 -1.87209770e-01 -2.01579764e-01
1.67376065e+00 3.51898730e-01 -1.41169548e-01 5.02234340e-01
8.86121869e-01 -7.78778195e-01 -9.72939372e-01 1.04650997e-01
3.89783531e-01 9.64263231e-02 1.80203944e-01 -1.05531287e+00
-1.10820997e+00 8.27435613e-01 9.30638909e-01 1.39393777e-01
9.54924941e-01 -3.08732122e-01 7.35918820e-01 8.26660335e-01
4.86692995e-01 -1.21192110e+00 5.90897575e-02 9.70507622e-01
9.54535484e-01 -7.30558574e-01 -3.93584430e-01 9.75181386e-02
-2.39539906e-01 1.44882178e+00 6.39195621e-01 -2.97963560e-01
3.66196722e-01 3.37965667e-01 -2.01869041e-01 -1.14375137e-01
-8.01322162e-01 4.88251932e-02 1.73242375e-01 6.36879563e-01
5.19331574e-01 -1.67496741e-01 -1.06621668e-01 3.57403368e-01
-4.32854742e-01 2.57018916e-02 1.38860956e-01 8.18579018e-01
-7.39833713e-01 -1.51314807e+00 -2.17840597e-01 2.63289452e-01
-1.30332738e-01 -2.68055618e-01 -5.25587440e-01 5.42892098e-01
-3.50906476e-02 4.81951326e-01 1.04678884e-01 -5.47118962e-01
1.83166593e-01 2.11839631e-01 6.02881849e-01 -1.51754707e-01
-8.78173053e-01 -4.78197962e-01 -5.68128936e-02 -4.52882856e-01
-2.44632196e-02 -6.84344649e-01 -1.15239346e+00 -5.06393790e-01
-1.72314331e-01 1.73392102e-01 1.16245615e+00 8.81828189e-01
5.70549011e-01 5.71537018e-01 3.12385857e-01 -8.65620911e-01
-7.11532593e-01 -7.33175278e-01 -4.70852628e-02 -4.11292076e-01
9.98242348e-02 -7.33426929e-01 -1.42673869e-02 -5.00633240e-01] | [8.395282745361328, 3.237260341644287] |
313872fd-9fed-4397-b0b6-759d5b21f818 | exploring-unknown-states-with-action-balance | 2003.04518 | null | https://arxiv.org/abs/2003.04518v2 | https://arxiv.org/pdf/2003.04518v2.pdf | Exploring Unknown States with Action Balance | Exploration is a key problem in reinforcement learning. Recently bonus-based methods have achieved considerable successes in environments where exploration is difficult such as Montezuma's Revenge, which assign additional bonuses (e.g., intrinsic rewards) to guide the agent to rarely visited states. Since the bonus is calculated according to the novelty of the next state after performing an action, we call such methods as the next-state bonus methods. However, the next-state bonus methods force the agent to pay overmuch attention in exploring known states and ignore finding unknown states since the exploration is driven by the next state already visited, which may slow the pace of finding reward in some environments. In this paper, we focus on improving the effectiveness of finding unknown states and propose action balance exploration, which balances the frequency of selecting each action at a given state and can be treated as an extension of upper confidence bound (UCB) to deep reinforcement learning. Moreover, we propose action balance RND that combines the next-state bonus methods (e.g., random network distillation exploration, RND) and our action balance exploration to take advantage of both sides. The experiments on the grid world and Atari games demonstrate action balance exploration has a better capability in finding unknown states and can improve the performance of RND in some hard exploration environments respectively. | ['Yujing Hu', 'Yan Song', 'Changjie Fan', 'Yingfeng Chen'] | 2020-03-10 | null | null | null | null | ['montezumas-revenge'] | ['playing-games'] | [-3.56096715e-01 2.82245427e-01 -5.73441088e-01 8.87712985e-02
-2.62967229e-01 -3.62721682e-01 5.47899485e-01 -1.43732846e-01
-8.88892949e-01 1.38448870e+00 7.60638192e-02 -4.94100213e-01
-4.53393698e-01 -1.00989795e+00 -6.60953641e-01 -1.03364336e+00
-3.82292330e-01 3.70485246e-01 2.51599103e-01 -4.63442445e-01
4.35691983e-01 2.27980465e-01 -1.09561563e+00 -4.24393564e-01
1.11474848e+00 8.26462865e-01 4.08842325e-01 8.12596977e-02
-2.08087802e-01 6.95376694e-01 -5.54736078e-01 1.78175613e-01
5.33873975e-01 -6.01096451e-01 -8.80578399e-01 -2.01099634e-01
-5.63007355e-01 -6.45976543e-01 -3.79648447e-01 1.42418838e+00
3.43964547e-01 5.20256996e-01 9.12787020e-02 -1.34523869e+00
-2.23713130e-01 1.09630203e+00 -8.79632831e-01 6.06936753e-01
1.93995714e-01 5.02187729e-01 7.29908168e-01 2.77184919e-02
4.43289429e-01 1.33439124e+00 -1.83812946e-01 8.06958497e-01
-9.14019167e-01 -1.00511026e+00 6.37666047e-01 3.80934089e-01
-7.98579276e-01 -5.01849577e-02 6.68378592e-01 4.26851586e-02
7.02768624e-01 4.89839613e-02 8.84446084e-01 9.62955952e-01
3.22850704e-01 1.15608025e+00 1.49671292e+00 -2.11731553e-01
9.05084968e-01 -7.07316473e-02 -1.63186401e-01 4.77776200e-01
1.53522044e-01 8.43127072e-01 -3.43692094e-01 -1.26597852e-01
9.50165808e-01 8.54072049e-02 -2.55893737e-01 -4.31232572e-01
-1.11471021e+00 1.05615091e+00 8.51188242e-01 -6.19962215e-02
-7.64106512e-01 3.94695550e-01 3.49895805e-01 3.92804742e-01
-1.10821791e-01 6.63942575e-01 -2.54994929e-01 -6.12089217e-01
-4.85148013e-01 3.20697904e-01 4.32659715e-01 4.93820459e-01
7.97231376e-01 4.07961875e-01 -2.93965012e-01 3.68727744e-01
1.75440624e-01 2.51806110e-01 8.40537727e-01 -9.87114131e-01
4.76490855e-01 4.68078613e-01 3.91107798e-01 -5.09459317e-01
-3.45102847e-01 -6.22449100e-01 -6.59412324e-01 7.84835100e-01
1.29100725e-01 -6.30843937e-01 -1.07802594e+00 2.06146216e+00
3.13388109e-01 2.32098207e-01 1.97072670e-01 1.05075312e+00
1.45988137e-01 6.22815073e-01 -1.05554514e-01 -5.65414250e-01
9.83940661e-01 -9.56169426e-01 -7.78722048e-01 -3.43105227e-01
4.17414665e-01 3.37770246e-02 9.96589124e-01 5.96406102e-01
-9.06758189e-01 -2.89934337e-01 -1.03381014e+00 9.63996470e-01
-2.06145495e-01 -4.54952598e-01 9.59362209e-01 5.65942466e-01
-7.14387059e-01 9.12332416e-01 -1.12351990e+00 -6.66821375e-03
6.86247766e-01 3.87042910e-01 2.31611114e-02 1.91636905e-01
-1.48289013e+00 9.51296151e-01 1.01086450e+00 4.25348803e-02
-1.63236892e+00 -3.21913809e-02 -4.94440377e-01 1.74314246e-01
1.27645707e+00 -1.07532308e-01 1.17481446e+00 -6.03607535e-01
-1.73124993e+00 -1.55752748e-01 3.08684856e-01 -8.50464463e-01
5.55564106e-01 -1.43491507e-01 -8.20753574e-02 -1.24407940e-01
9.02950764e-02 8.61301720e-01 5.50782084e-01 -9.60569739e-01
-7.89811313e-01 -2.45462835e-01 6.01036727e-01 6.81202888e-01
-2.74849683e-01 -4.98187155e-01 6.46495670e-02 -2.13315398e-01
1.18092060e-01 -1.05290961e+00 -8.46993506e-01 -4.25447285e-01
-3.92378420e-01 -5.66378295e-01 6.29670084e-01 -8.86529610e-02
1.18343186e+00 -1.95548153e+00 6.72073588e-02 2.30913058e-01
1.21514320e-01 2.92420030e-01 -1.76166654e-01 3.42551768e-01
1.26207814e-01 5.32661900e-02 2.53304485e-02 3.63755941e-01
-1.96859222e-02 5.41680217e-01 -3.26149911e-01 9.29644629e-02
-2.67219961e-01 8.29169571e-01 -1.32325661e+00 -2.00982243e-01
1.46508113e-01 -3.62494916e-01 -4.96763706e-01 1.91440269e-01
-3.35651726e-01 4.49333370e-01 -9.16383266e-01 3.40346605e-01
6.76221967e-01 -9.26741064e-02 8.97539333e-02 5.07999241e-01
-2.66707271e-01 5.10975599e-01 -1.39439082e+00 1.31517172e+00
-2.54174531e-01 8.64385143e-02 -1.12587750e-01 -9.61653173e-01
8.95272732e-01 1.07436188e-01 2.26480201e-01 -7.25531459e-01
2.87250817e-01 8.19812641e-02 4.20294911e-01 -1.87280282e-01
4.89849836e-01 -1.47080734e-01 2.96920147e-02 7.65575707e-01
-2.63346970e-01 1.56209514e-01 2.84366906e-01 3.50571513e-01
1.11693048e+00 3.09148252e-01 5.85676014e-01 -2.90930539e-01
2.16564998e-01 -6.94080740e-02 9.29805636e-01 1.05774009e+00
-5.60328126e-01 -3.97506863e-01 8.90704930e-01 -2.05730736e-01
-5.67568660e-01 -9.36747670e-01 2.48765811e-01 8.59597921e-01
5.41778147e-01 -1.81965753e-01 -5.24810672e-01 -1.09413195e+00
-4.07274887e-02 9.05971944e-01 -7.69416153e-01 -5.81776500e-01
-3.85971338e-01 -7.46499181e-01 1.89312875e-01 5.53276479e-01
1.10669601e+00 -1.51916194e+00 -1.10850656e+00 2.12907299e-01
8.70601311e-02 -2.21035555e-01 -2.17355147e-01 6.87387884e-01
-9.23015654e-01 -8.68031323e-01 -6.94096327e-01 -2.46059567e-01
5.51297843e-01 1.32102728e-01 5.46335042e-01 -2.45579928e-01
2.12910950e-01 -1.01859398e-01 -4.11863744e-01 -3.73195469e-01
-4.83670197e-02 1.12528399e-01 3.48441839e-01 -4.56606984e-01
4.30755407e-01 -6.10035181e-01 -6.63434923e-01 3.23071301e-01
-6.96440876e-01 -8.23167115e-02 7.03730881e-01 1.08644509e+00
3.70882362e-01 5.00169754e-01 9.03637946e-01 -5.21900237e-01
1.07302082e+00 -7.43528366e-01 -8.71787548e-01 -3.42515074e-02
-9.39162731e-01 5.10806322e-01 5.49868166e-01 -7.51864433e-01
-1.04297721e+00 -3.08361232e-01 1.21446729e-01 -4.66607422e-01
1.92518249e-01 6.03702843e-01 2.91278865e-02 7.41900876e-02
5.85559249e-01 3.61624837e-01 1.46625698e-01 -3.25824440e-01
1.25066400e-01 3.11585248e-01 8.21828842e-02 -6.48371577e-01
5.45707822e-01 1.87109411e-01 -2.70336680e-03 -2.96688974e-02
-7.05466628e-01 -8.85273442e-02 1.57363296e-01 -1.71294615e-01
5.36327064e-01 -6.84060633e-01 -1.43054986e+00 2.37895951e-01
-6.10780478e-01 -3.94585490e-01 -5.77563524e-01 6.81854248e-01
-5.89128733e-01 3.53843451e-01 -3.08083415e-01 -1.23197949e+00
-3.86823229e-02 -1.20826852e+00 1.28921866e-01 9.08550203e-01
2.69267082e-01 -5.61283886e-01 1.24863163e-01 -2.24072844e-01
3.54075670e-01 1.93367392e-01 5.35504222e-01 -4.26872760e-01
-8.68208230e-01 3.71232331e-01 2.68163364e-02 1.13296278e-01
7.70785138e-02 -6.66478097e-01 -4.94239628e-01 -5.28651059e-01
3.13401967e-02 -6.32208705e-01 1.04049551e+00 4.33198690e-01
9.07538414e-01 -5.75303614e-01 -4.53364044e-01 2.56687850e-01
1.30752635e+00 8.48162770e-01 8.08414578e-01 7.73728371e-01
6.67344406e-02 2.97934979e-01 1.19434178e+00 8.26299906e-01
-5.20682856e-02 5.07943809e-01 1.09762776e+00 3.07101637e-01
4.29541469e-01 -4.86100614e-01 4.35735732e-01 1.41416833e-01
-1.07683204e-01 -4.34566960e-02 -4.87708539e-01 5.43736994e-01
-2.13881183e+00 -1.04072905e+00 5.23158252e-01 2.37247849e+00
1.08074009e+00 7.32543588e-01 2.73993224e-01 -8.84650946e-02
6.60195172e-01 1.68146268e-01 -1.21170294e+00 -5.08953154e-01
7.89185986e-02 -1.38241380e-01 5.73292494e-01 4.71864283e-01
-6.15184128e-01 1.09423172e+00 5.71002197e+00 1.29252744e+00
-9.24574554e-01 2.54473947e-02 5.77668130e-01 -2.32011989e-01
-2.03690246e-01 3.56822938e-01 -9.28951025e-01 7.08384275e-01
6.41953170e-01 -2.56521851e-01 9.37810183e-01 1.15255034e+00
1.09303549e-01 -7.17136204e-01 -7.15028584e-01 6.21430874e-01
-5.76485813e-01 -1.07562006e+00 -2.87653178e-01 3.98069263e-01
8.25547099e-01 -2.16529127e-02 3.18309180e-02 7.79026151e-01
1.06884074e+00 -8.49118769e-01 3.80455971e-01 2.75592119e-01
3.51795375e-01 -1.11845589e+00 7.78522730e-01 7.52962828e-01
-7.38819838e-01 -4.87456739e-01 -5.84013820e-01 -3.54202986e-01
2.06636027e-01 4.10400569e-01 -1.01613176e+00 3.37101966e-01
6.44085169e-01 2.67049104e-01 -6.77931607e-02 1.02395463e+00
-6.15517557e-01 4.68970746e-01 -4.27837282e-01 -6.26654685e-01
9.41060245e-01 -4.40522403e-01 6.81952357e-01 2.85218865e-01
2.05332652e-01 1.23119995e-01 4.71426785e-01 1.09358013e+00
1.84996411e-01 -3.56494248e-01 -3.50953966e-01 -8.81568044e-02
5.76025307e-01 1.03931081e+00 -7.28230119e-01 -3.20436597e-01
3.57792079e-01 5.19337654e-01 3.92283499e-01 2.98541695e-01
-8.02655876e-01 -5.60198486e-01 4.48053628e-01 -2.75561303e-01
2.77162671e-01 8.24654009e-03 -4.99422960e-02 -9.17571902e-01
-2.72748172e-01 -7.87124336e-01 4.94805455e-01 -6.84382856e-01
-6.57084942e-01 6.29425704e-01 1.00903593e-01 -1.17345428e+00
-4.90571499e-01 -1.73341677e-01 -7.64372766e-01 8.21873188e-01
-1.67013848e+00 -2.25301936e-01 4.06200206e-03 4.11217570e-01
5.30825615e-01 -2.85870939e-01 3.79552811e-01 -3.23325992e-01
-5.99525213e-01 4.36425507e-01 2.94874400e-01 -1.47011399e-01
2.38425851e-01 -1.36139905e+00 -7.51612242e-03 7.05360174e-01
-2.49998689e-01 5.93586028e-01 8.22836697e-01 -9.19328928e-01
-1.00428081e+00 -5.07618427e-01 -5.51462583e-02 2.03560144e-01
5.26044786e-01 2.38097529e-03 -7.25959063e-01 5.00777841e-01
1.50817871e-01 -2.45744601e-01 -8.80067889e-03 3.55595350e-01
1.51191473e-01 9.57329664e-03 -1.17734611e+00 8.77681077e-01
8.91830087e-01 3.82585749e-02 -6.08287752e-01 5.81641532e-02
6.85628116e-01 -3.50561589e-01 -2.74239272e-01 2.82711387e-01
1.87979206e-01 -9.40054059e-01 5.89064062e-01 -8.63087296e-01
2.94860452e-01 -3.23866904e-01 2.60449558e-01 -1.80511391e+00
-3.80386263e-01 -1.00117898e+00 -4.50112551e-01 7.74481535e-01
1.76844358e-01 -7.86706626e-01 1.23641920e+00 3.14050257e-01
2.88702305e-02 -1.06745851e+00 -1.15205085e+00 -1.18491507e+00
-9.92391855e-02 1.85549259e-01 8.89400423e-01 6.99754477e-01
5.21447420e-01 1.36456266e-01 -5.63621581e-01 -1.86622277e-01
5.98259449e-01 1.71200499e-01 6.04428411e-01 -6.78098083e-01
-4.96236503e-01 -4.20461357e-01 -2.10945718e-02 -1.35496533e+00
-8.96174684e-02 -5.63336253e-01 1.25583902e-01 -1.31354749e+00
3.49246383e-01 -7.07146466e-01 -8.10833812e-01 7.54989028e-01
-3.73743802e-01 -4.81434494e-01 2.56778717e-01 1.89725786e-01
-8.73715520e-01 1.00651419e+00 1.77500761e+00 9.35905054e-02
-5.67218482e-01 2.20189303e-01 -8.70489717e-01 5.57499468e-01
1.04114139e+00 -4.44723308e-01 -6.13436580e-01 1.64383188e-01
4.01421160e-01 5.36419690e-01 9.72921699e-02 -9.12814558e-01
1.61461249e-01 -8.47040474e-01 1.35064155e-01 -6.05649531e-01
3.08735609e-01 -6.13784432e-01 -1.01223655e-01 1.14713418e+00
-5.29591918e-01 -1.20109916e-01 6.11006916e-02 8.96733344e-01
2.88556013e-02 -5.62533557e-01 8.07985008e-01 -4.42898124e-01
-9.47888911e-01 2.73361385e-01 -2.94093430e-01 8.00079405e-02
1.21351564e+00 -2.76082963e-01 -4.37322557e-01 -6.30004227e-01
-6.44413233e-01 8.91869664e-01 -4.27356176e-03 1.96035206e-01
6.09821975e-01 -1.41456699e+00 -4.15736049e-01 -7.98843652e-02
-2.71758854e-01 -4.29090485e-02 3.22688192e-01 6.83834970e-01
2.36217100e-02 3.62514108e-01 -7.10670352e-01 -2.33234003e-01
-7.67015874e-01 7.81281769e-01 3.09151828e-01 -7.21712470e-01
-3.88972521e-01 7.96962380e-01 2.39393860e-01 -2.62850791e-01
3.19046885e-01 -4.68884557e-02 -4.34992760e-01 -1.85323387e-01
5.18456161e-01 6.53263330e-01 -4.98487234e-01 3.17100227e-01
-2.46738806e-01 -1.80829942e-01 -5.33056796e-01 -3.12265098e-01
1.16451836e+00 -1.52606100e-01 2.01579750e-01 2.37623721e-01
5.20682931e-01 -3.80433768e-01 -1.74998665e+00 -1.54625431e-01
-2.12342232e-01 -7.12072313e-01 3.60849559e-01 -1.13718951e+00
-1.08763182e+00 6.78288877e-01 6.86612010e-01 3.45059723e-01
9.77448285e-01 -2.06484184e-01 7.83189118e-01 6.86860800e-01
7.41393507e-01 -1.44160688e+00 2.62312323e-01 5.05536616e-01
5.92081904e-01 -1.21016431e+00 7.62607455e-02 1.84547827e-01
-8.94087613e-01 7.27043808e-01 1.16811442e+00 -2.00175375e-01
3.84384274e-01 -4.71098498e-02 -3.43564868e-01 -8.61953422e-02
-8.91687632e-01 -3.70738178e-01 -3.63090158e-01 5.43425739e-01
-3.69277835e-01 2.75742173e-01 -5.65793812e-01 3.38010639e-01
-1.12985879e-01 9.43615288e-02 7.33893335e-01 9.73989725e-01
-9.81054246e-01 -1.06791437e+00 -3.00181776e-01 6.86861873e-01
-2.13704705e-01 8.94979388e-02 1.74443841e-01 6.64391339e-01
5.62540740e-02 7.99792230e-01 -9.28430930e-02 -2.86601722e-01
-2.98862737e-02 -3.48816901e-01 2.96193570e-01 -3.63966137e-01
-4.12980109e-01 5.26757985e-02 -5.50086312e-02 -7.30913222e-01
-6.46328926e-02 -5.49309731e-01 -1.47265911e+00 -3.99425179e-01
-6.62836194e-01 6.58815801e-01 4.68843192e-01 8.86807680e-01
2.19149843e-01 6.43335164e-01 8.94969225e-01 -3.76571149e-01
-1.31226134e+00 -9.32537317e-01 -8.13899338e-01 -6.49839267e-02
2.04063892e-01 -1.06648457e+00 -4.38598156e-01 -8.99721444e-01] | [3.9158782958984375, 1.8177975416183472] |
f85df74a-8713-402a-b1bd-899b0a68fe7f | back-to-the-future-unsupervised-backprop | 2010.05906 | null | https://arxiv.org/abs/2010.05906v4 | https://arxiv.org/pdf/2010.05906v4.pdf | Back to the Future: Unsupervised Backprop-based Decoding for Counterfactual and Abductive Commonsense Reasoning | Abductive and counterfactual reasoning, core abilities of everyday human cognition, require reasoning about what might have happened at time t, while conditioning on multiple contexts from the relative past and future. However, simultaneous incorporation of past and future contexts using generative language models (LMs) can be challenging, as they are trained either to condition only on the past context or to perform narrowly scoped text-infilling. In this paper, we propose DeLorean, a new unsupervised decoding algorithm that can flexibly incorporate both the past and future contexts using only off-the-shelf, left-to-right language models and no supervision. The key intuition of our algorithm is incorporating the future through back-propagation, during which, we only update the internal representation of the output while fixing the model parameters. By alternating between forward and backward propagation, DeLorean can decode the output representation that reflects both the left and right contexts. We demonstrate that our approach is general and applicable to two nonmonotonic reasoning tasks: abductive text generation and counterfactual story revision, where DeLorean outperforms a range of unsupervised and some supervised methods, based on automatic and human evaluation. | ['Yejin Choi', 'Antoine Bosselut', 'Ronan Le Bras', 'Jena Hwang', 'Chandra Bhagavatula', 'Peter West', 'Vered Shwartz', 'Lianhui Qin'] | 2020-10-12 | null | https://aclanthology.org/2020.emnlp-main.58 | https://aclanthology.org/2020.emnlp-main.58.pdf | emnlp-2020-11 | ['text-infilling'] | ['natural-language-processing'] | [ 5.79832852e-01 5.27078569e-01 -2.00239062e-01 -5.78665078e-01
-4.06433791e-01 -6.47428632e-01 1.19119310e+00 1.42376840e-01
-5.73250592e-01 1.11875749e+00 6.69173837e-01 -5.65415740e-01
4.87321503e-02 -1.01867151e+00 -9.92280066e-01 -2.99623936e-01
1.21493317e-01 7.25092769e-01 -5.36543168e-02 -4.09837991e-01
1.77310124e-01 1.62964463e-02 -1.47544479e+00 5.06797969e-01
6.05130553e-01 5.93066037e-01 3.05691451e-01 6.39653265e-01
1.34185823e-02 1.21335101e+00 -3.55793238e-01 -8.84579062e-01
-2.03486457e-01 -7.67837882e-01 -9.08365726e-01 -2.13736102e-01
-1.94149047e-01 -4.86878276e-01 -4.35244322e-01 7.97289670e-01
2.63698667e-01 2.38593414e-01 5.44195175e-01 -9.15827513e-01
-8.36743891e-01 1.53228486e+00 -1.70922786e-01 1.45908833e-01
6.81383669e-01 1.85947642e-01 1.22393286e+00 -4.85054940e-01
7.95981169e-01 1.37741065e+00 3.63062441e-01 7.55203485e-01
-1.19766653e+00 -2.65330493e-01 6.75520301e-01 4.36053485e-01
-7.93552816e-01 -3.84498268e-01 1.07136118e+00 -2.31215373e-01
1.25409567e+00 3.07352781e-01 8.10527503e-01 1.53666139e+00
5.12141824e-01 9.90028977e-01 9.23551559e-01 -7.22918451e-01
4.37085390e-01 -2.26274073e-01 -1.97546408e-01 4.02010828e-01
2.11476609e-01 1.41761765e-01 -6.97133601e-01 -2.98059404e-01
3.75749499e-01 -2.77262554e-02 -3.23626339e-01 -1.45215109e-01
-1.58230901e+00 7.92315245e-01 -2.13031806e-02 1.93196580e-01
-5.73826551e-01 4.36564595e-01 2.21876383e-01 4.57788676e-01
2.64919430e-01 4.67660666e-01 -6.14129722e-01 -9.92575660e-02
-1.02992666e+00 5.01792312e-01 9.95605230e-01 7.79914320e-01
3.26555848e-01 -2.80420244e-01 -2.86117196e-01 4.24002647e-01
6.04982793e-01 4.80012983e-01 8.38922203e-01 -9.22688425e-01
6.24926746e-01 3.51041824e-01 3.26365739e-01 -6.04899108e-01
-2.78698504e-01 -3.01689476e-01 -3.20758671e-01 -2.71307915e-01
3.69097859e-01 -4.11285162e-01 -8.32895100e-01 2.30197072e+00
3.49632382e-01 2.72484832e-02 4.41103667e-01 7.10813880e-01
2.04740182e-01 6.15094781e-01 1.60239607e-01 -5.73294878e-01
1.14811182e+00 -5.59583902e-01 -8.24135482e-01 -8.12685430e-01
7.57963657e-01 -5.43171465e-01 9.37555313e-01 3.05197686e-01
-1.27004898e+00 1.25011206e-01 -1.16812408e+00 -2.52960950e-01
-1.09087393e-01 -3.03561985e-02 8.85762215e-01 2.77385801e-01
-7.13868916e-01 5.47163904e-01 -1.08963585e+00 -2.23659009e-01
5.40379398e-02 -1.63693614e-02 -6.76403716e-02 -1.62735239e-01
-1.54898000e+00 1.06142604e+00 8.56536329e-01 2.62761623e-01
-8.59737098e-01 -1.56789988e-01 -1.07103515e+00 -1.66314431e-02
8.20508838e-01 -1.08696771e+00 1.66905189e+00 -6.88924313e-01
-1.68174219e+00 6.82446063e-01 -2.96531111e-01 -1.04494178e+00
7.39114165e-01 -3.21160406e-01 -3.80349666e-01 -3.75170112e-02
7.27684423e-02 5.34387708e-01 5.39419115e-01 -8.31445873e-01
-4.03529793e-01 -4.94751543e-01 3.24209839e-01 5.43272614e-01
2.53926188e-01 -2.58007735e-01 -2.10395396e-01 -6.71448410e-01
4.50771183e-01 -9.87247944e-01 -1.70232385e-01 -2.81667411e-01
-4.56501126e-01 -8.39531198e-02 4.02778327e-01 -5.90285718e-01
1.25463247e+00 -1.80977261e+00 2.54215837e-01 -5.89805543e-02
-3.21368039e-01 -4.47083563e-01 2.98015207e-01 5.51960051e-01
1.93886191e-03 -1.56747758e-01 -6.09169841e-01 -2.41126329e-01
3.79432410e-01 6.50009155e-01 -9.53539491e-01 2.81307995e-01
2.69333553e-02 9.88520443e-01 -1.06092501e+00 -3.73779595e-01
3.04935090e-02 7.04829767e-03 -7.57275343e-01 7.48053640e-02
-9.95022714e-01 1.20720692e-01 -1.92923695e-01 1.96457967e-01
1.56053424e-01 -1.50613412e-01 7.92002082e-01 3.25005829e-01
1.30890399e-01 1.05359554e+00 -1.30443847e+00 1.92226374e+00
-4.97571677e-01 4.06758338e-01 -4.41083521e-01 -8.68778586e-01
4.70467269e-01 6.41463876e-01 -2.65750378e-01 -5.98466992e-01
4.84419428e-02 2.96602637e-01 6.98646232e-02 -4.57536578e-01
3.33139181e-01 -7.68548846e-01 -3.40015948e-01 1.09064460e+00
1.14003055e-01 -2.15041712e-01 4.96509314e-01 4.06399518e-01
6.49379253e-01 5.32951117e-01 7.61395395e-01 2.10762084e-01
4.53728616e-01 -2.39227265e-01 6.29860640e-01 9.60750341e-01
2.67581880e-01 4.80011046e-01 8.97952914e-01 -3.61880273e-01
-7.44342566e-01 -9.42014873e-01 1.48530170e-01 1.25051165e+00
-1.09019607e-01 -3.19849223e-01 -5.16148806e-01 -6.05443478e-01
-1.51609421e-01 1.86281478e+00 -8.08855236e-01 -3.76187116e-01
-8.20678532e-01 -9.68080163e-01 4.41232562e-01 6.53639436e-01
3.00055325e-01 -1.38863623e+00 -1.03104627e+00 5.36576629e-01
-4.18002427e-01 -8.92354250e-01 -4.34532404e-01 1.63392395e-01
-9.70108926e-01 -8.86860549e-01 -1.62991479e-01 -2.44694889e-01
3.52886617e-01 -3.52020770e-01 9.34002876e-01 -1.42366335e-01
3.82102400e-01 3.58052731e-01 -2.06648652e-02 -5.62551796e-01
-6.15800858e-01 -3.22036326e-01 -2.12333307e-01 -4.36609685e-02
1.18752055e-01 -9.51699734e-01 -2.94339657e-01 -2.17098117e-01
-9.78147388e-01 6.46872401e-01 5.77931345e-01 1.00853431e+00
5.04942477e-01 -1.83683246e-01 4.37742352e-01 -1.26424813e+00
7.76966274e-01 -6.07692301e-01 -3.57700199e-01 4.42083299e-01
-6.34765625e-01 6.05633676e-01 6.09604359e-01 -4.53683168e-01
-1.73116350e+00 -3.86974722e-01 -9.51191708e-02 1.39591739e-01
2.53494307e-02 8.72764766e-01 -3.72007906e-01 1.07011223e+00
5.90575159e-01 5.31670034e-01 -2.76412815e-01 -2.63921022e-01
8.36982250e-01 3.37586224e-01 8.96079421e-01 -8.74499381e-01
2.88279116e-01 6.55149162e-01 -2.41059333e-01 -1.77900612e-01
-1.12373734e+00 2.18304411e-01 -5.06247759e-01 -2.28517093e-02
5.43447137e-01 -7.13488042e-01 -5.04779518e-01 3.18590373e-01
-1.39026320e+00 -4.35459644e-01 -4.50873107e-01 6.30255759e-01
-9.65296268e-01 1.90032750e-01 -5.69851637e-01 -9.58501339e-01
-3.91507626e-01 -8.51961493e-01 7.34954953e-01 -9.10013467e-02
-6.18622780e-01 -1.17069983e+00 -2.47330833e-02 1.24575533e-01
5.83913997e-02 1.40441194e-01 1.06225228e+00 -8.64943564e-01
-4.20272976e-01 -1.57220677e-01 3.60532820e-01 6.61746180e-03
-2.07569376e-01 -2.80622810e-01 -8.32607627e-01 -8.50994736e-02
4.69140500e-01 -2.39390865e-01 1.02767408e+00 6.20518178e-02
7.56822586e-01 -1.00030565e+00 -2.74896502e-01 2.94844866e-01
1.12216961e+00 3.03187519e-01 5.78259826e-01 2.21823454e-01
2.04168916e-01 3.27768832e-01 4.11943644e-01 5.65331221e-01
8.23044717e-01 2.69007385e-01 3.14710379e-01 7.37820685e-01
4.05712761e-02 -8.14422965e-01 5.66136479e-01 3.49907100e-01
-2.86643393e-02 -3.95122558e-01 -7.40733027e-01 7.72134066e-01
-1.99197614e+00 -1.48470116e+00 3.26874614e-01 2.20167398e+00
1.44612992e+00 5.72983563e-01 -3.45313132e-01 1.77150562e-01
4.93721396e-01 3.25228721e-01 -6.22882426e-01 -4.53917474e-01
-1.66889191e-01 5.80756254e-02 -9.46887862e-03 6.79937303e-01
-6.46152139e-01 8.45248163e-01 6.15270710e+00 3.85772318e-01
-1.13639927e+00 1.08829483e-01 3.16136837e-01 -4.82987851e-01
-9.24662888e-01 4.69069272e-01 -4.38372821e-01 3.42098773e-01
9.91632044e-01 -3.51552188e-01 6.21479690e-01 7.00826883e-01
-1.05549112e-01 -2.56905913e-01 -1.60190082e+00 4.42603797e-01
2.78772533e-01 -1.34694934e+00 2.54249156e-01 -3.15335691e-01
5.96619785e-01 -1.24877706e-01 -1.80695802e-01 3.72927636e-01
4.95689958e-01 -6.65458143e-01 1.44704878e+00 6.44204140e-01
5.65540493e-01 -5.48625052e-01 6.21297300e-01 8.67806792e-01
-6.52020812e-01 -3.91028345e-01 -3.08545642e-02 -4.57536161e-01
4.74787503e-01 5.32962024e-01 -9.40300703e-01 3.89097780e-01
6.92064688e-02 4.03845131e-01 -1.86090544e-01 4.05594975e-01
-9.99375165e-01 6.50798082e-01 -3.00679624e-01 -1.21564187e-01
4.03950512e-02 1.18446864e-01 7.77316809e-01 1.18378210e+00
3.52298886e-01 3.69745344e-01 -2.35869080e-01 1.12700152e+00
-1.64333180e-01 -3.03527653e-01 -4.98969316e-01 3.38506736e-02
5.56808472e-01 5.20977914e-01 -4.23206359e-01 -5.21792233e-01
-2.05439940e-01 9.67514932e-01 3.46502602e-01 3.83810222e-01
-8.72778237e-01 -2.92484108e-02 1.37532204e-01 -1.07549325e-01
3.49052012e-01 -9.84865129e-02 -3.74318421e-01 -1.35089433e+00
2.11844638e-01 -6.79547012e-01 8.75840127e-01 -9.62112129e-01
-8.82537961e-01 1.89704970e-01 3.79252613e-01 -6.15156770e-01
-9.46255445e-01 -3.39638293e-01 -8.52883041e-01 6.22959018e-01
-1.30370653e+00 -1.06278634e+00 5.32981038e-01 2.57253051e-01
4.63946819e-01 2.44691849e-01 8.28368127e-01 -3.25081140e-01
-3.63220036e-01 2.07114175e-01 -3.83444488e-01 -9.52278152e-02
4.98917967e-01 -1.34424210e+00 4.62937236e-01 1.15344691e+00
3.15665245e-01 1.01686442e+00 1.08065426e+00 -7.28013754e-01
-1.23413873e+00 -8.69530439e-01 1.49647546e+00 -3.61363709e-01
6.71908081e-01 -4.24362510e-01 -8.31402540e-01 1.49904752e+00
2.66066670e-01 -5.17856300e-01 5.75698495e-01 6.50661215e-02
-4.19218361e-01 1.17590077e-01 -9.80994344e-01 1.00645328e+00
1.10178840e+00 -5.93989551e-01 -1.54705322e+00 2.67617613e-01
1.00290179e+00 -8.00866663e-01 -2.79531151e-01 1.17367133e-01
4.98122036e-01 -7.89510846e-01 5.83060145e-01 -7.67753959e-01
6.21330082e-01 -3.41791719e-01 -2.79747903e-01 -1.18287337e+00
7.14242682e-02 -9.66118574e-01 -5.68514526e-01 8.59197438e-01
8.06370735e-01 -8.17964554e-01 1.19774856e-01 9.15145755e-01
-7.36684948e-02 -7.01875448e-01 -1.01459289e+00 -2.12231964e-01
-2.56271809e-02 -1.14017081e+00 1.09107649e+00 6.98567629e-01
5.51517129e-01 5.07146835e-01 -3.59949023e-01 4.99434583e-02
2.40733907e-01 6.30733430e-01 3.62736523e-01 -8.10630322e-01
-6.01246178e-01 -2.66067266e-01 7.16691986e-02 -1.23678446e+00
2.21621975e-01 -1.00679708e+00 3.98865640e-02 -1.68374920e+00
2.33042255e-01 1.64463118e-01 -5.05999140e-02 7.11121321e-01
-1.72009155e-01 -3.42409611e-01 8.31616297e-02 -2.86780782e-02
-5.92939198e-01 6.97123826e-01 1.23724294e+00 -1.13063850e-01
-1.60472840e-01 -6.59145936e-02 -9.88282084e-01 9.95779872e-01
5.33877909e-01 -4.21978742e-01 -6.24456704e-01 -6.17965996e-01
8.26104641e-01 3.84514809e-01 5.34039915e-01 -5.83343625e-01
4.36175823e-01 -4.84667420e-01 2.96469361e-01 -5.09298384e-01
2.22086027e-01 -3.21874917e-01 1.63189560e-01 3.46530676e-01
-8.60796332e-01 4.26926930e-03 7.17830881e-02 8.13935399e-01
1.09422110e-01 -3.50748718e-01 2.44091243e-01 -4.66722816e-01
-5.47660291e-01 -2.37288818e-01 -4.15781170e-01 2.03963786e-01
8.30464303e-01 2.19648123e-01 -2.95017004e-01 -4.70316440e-01
-1.00262678e+00 2.09550217e-01 2.52664626e-01 2.21198201e-01
7.02033460e-01 -1.12023592e+00 -7.41855383e-01 1.57171637e-01
-7.05834329e-02 6.15743436e-02 2.70529926e-01 8.30874264e-01
-1.40765876e-01 6.70534790e-01 1.90678343e-01 -8.06555524e-02
-5.84712207e-01 7.01778233e-01 4.11867797e-01 -4.12493914e-01
-6.20289505e-01 7.57930160e-01 5.99509254e-02 -4.15990889e-01
1.05048064e-02 -5.99185228e-01 -1.33275703e-01 -2.21032952e-03
5.42603731e-01 4.52290066e-02 -4.47730860e-03 -2.69415021e-01
-1.81673288e-01 -2.82899290e-01 -1.31042734e-01 -8.19044352e-01
1.30641854e+00 -2.85227180e-01 -3.65776569e-01 8.75287592e-01
5.70069373e-01 -6.16509654e-02 -1.03760469e+00 -3.32675040e-01
4.05369736e-02 -1.12435624e-01 -1.67116627e-01 -1.12596369e+00
-4.46558982e-01 6.41577899e-01 -2.99052685e-01 -1.69541780e-02
1.07094312e+00 2.05099657e-01 6.96047604e-01 6.62209630e-01
4.45478231e-01 -1.07479763e+00 -2.54090697e-01 6.42345548e-01
1.16888380e+00 -6.94081485e-01 1.22625314e-01 8.87128860e-02
-7.16847181e-01 1.03902864e+00 3.53099793e-01 3.18318069e-01
2.75131881e-01 9.51388031e-02 -3.35119218e-01 -6.61896542e-02
-1.67575431e+00 1.12112790e-01 1.15840033e-01 9.45909619e-02
7.24045694e-01 2.95676261e-01 -6.10099852e-01 9.05251980e-01
-8.35635126e-01 9.22234952e-02 7.28381991e-01 1.08769345e+00
-1.62832305e-01 -9.38879311e-01 -3.90162438e-01 4.55330789e-01
-4.73398089e-01 -2.63026178e-01 -2.31523946e-01 7.51715541e-01
9.01513174e-02 8.71702433e-01 4.20098267e-02 1.31059408e-01
1.39207050e-01 5.80099285e-01 8.24514508e-01 -7.08699226e-01
-2.11009625e-02 -8.11435506e-02 3.31613749e-01 -3.18581015e-01
-2.54461199e-01 -1.09718466e+00 -1.58686543e+00 -8.93483534e-02
-2.63474226e-01 -5.41510954e-02 3.78906906e-01 1.56068480e+00
8.85725096e-02 3.97687823e-01 9.48114321e-02 -3.42009991e-01
-8.96409154e-01 -1.06518996e+00 -2.43619114e-01 3.81783485e-01
3.29637408e-01 -5.07798135e-01 -2.81642318e-01 2.99177736e-01] | [11.293642044067383, 8.887675285339355] |
6ca3d8c0-e3bf-4e78-8590-d165eb93570a | estimating-parameters-of-the-tree-root-in | 2112.13494 | null | https://arxiv.org/abs/2112.13494v1 | https://arxiv.org/pdf/2112.13494v1.pdf | Estimating Parameters of the Tree Root in Heterogeneous Soil Environments via Mask-Guided Multi-Polarimetric Integration Neural Network | Ground-penetrating radar (GPR) has been used as a non-destructive tool for tree root inspection. Estimating root-related parameters from GPR radargrams greatly facilitates root health monitoring and imaging. However, the task of estimating root-related parameters is challenging as the root reflection is a complex function of multiple root parameters and root orientations. Existing methods can only estimate a single root parameter at a time without considering the influence of other parameters and root orientations, resulting in limited estimation accuracy under different root conditions. In addition, soil heterogeneity introduces clutter in GPR radargrams, making the data processing and interpretation even harder. To address these issues, a novel neural network architecture, called mask-guided multi-polarimetric integration neural network (MMI-Net), is proposed to automatically and simultaneously estimate multiple root-related parameters in heterogeneous soil environments. The MMI-Net includes two sub-networks: a MaskNet that predicts a mask to highlight the root reflection area to eliminate interfering environmental clutter, and a ParaNet that uses the predicted mask as guidance to integrate, extract, and emphasize informative features in multi-polarimetric radargrams for accurate estimation of five key root-related parameters. The parameters include the root depth, diameter, relative permittivity, horizontal and vertical orientation angles. Experimental results demonstrate that the proposed MMI-Net achieves high estimation accuracy in these root-related parameters. This is the first work that takes the combined contributions of root parameters and spatial orientations into account and simultaneously estimates multiple root-related parameters. The data and code implemented in the paper can be found at https://haihan-sun.github.io/GPR.html. | ['Abdulkadir C. Yucel', 'Mohamed Lokman Mohd Yusof', 'Genevieve Ow', 'Chongyi Li', 'Qiqi Dai', 'Yee Hui Lee', 'Hai-Han Sun'] | 2021-12-27 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [ 3.80140692e-01 -9.17717740e-02 2.39170790e-01 -1.63444728e-01
-5.74181497e-01 -8.30501020e-02 -2.37072676e-01 2.88981676e-01
1.69810474e-01 5.35727620e-01 -2.77569830e-01 -4.87470716e-01
-4.72369999e-01 -1.12574995e+00 -1.84942782e-01 -9.68162596e-01
-4.68320698e-01 3.91664833e-01 3.04355741e-01 -2.70815194e-01
2.73954779e-01 9.42094266e-01 -1.46975160e+00 -7.92555660e-02
1.09111631e+00 1.18999493e+00 9.21198487e-01 4.45729047e-01
1.15745179e-01 4.25782531e-01 -4.91679937e-01 4.02546376e-01
-1.47321820e-01 4.41442728e-02 -2.43959442e-01 -3.05850655e-01
-6.35864288e-02 -3.58326256e-01 1.93330094e-01 1.04711306e+00
6.78524375e-01 -1.69822484e-01 9.18445349e-01 -8.08053136e-01
-4.97009039e-01 4.42184687e-01 -1.07091081e+00 9.84455273e-02
-2.54397035e-01 -4.04064506e-01 6.88620210e-01 -8.54305446e-01
2.64197718e-02 1.05358028e+00 8.83227944e-01 -3.29801857e-01
-1.01908696e+00 -1.01089096e+00 9.34140533e-02 3.46738577e-01
-1.34064615e+00 -1.53912693e-01 8.12949002e-01 -3.39849144e-01
4.81401086e-01 1.71441585e-01 6.81367815e-01 5.14230251e-01
6.09022558e-01 4.26558584e-01 9.33520079e-01 -4.00908977e-01
-7.14792684e-03 -2.51105040e-01 2.83746600e-01 4.76036251e-01
7.09290624e-01 1.78478673e-01 2.16069907e-01 -9.59119499e-02
9.34305131e-01 -3.24491799e-01 -5.65409124e-01 -1.84809834e-01
-5.78320086e-01 7.70935237e-01 5.20758927e-01 4.27684575e-01
-9.22428310e-01 -3.01184297e-01 3.10677029e-02 -1.55075684e-01
6.27533853e-01 2.88607568e-01 -5.12133598e-01 5.17099142e-01
-8.11249137e-01 1.13700420e-01 7.47370899e-01 4.65623677e-01
6.39567554e-01 4.60159242e-01 2.76539121e-02 1.29331207e+00
5.01684844e-01 1.16015053e+00 -8.78804401e-02 -8.50269139e-01
-3.45237628e-02 8.54352042e-02 1.57777756e-01 -1.59112442e+00
-7.62361526e-01 -8.47431481e-01 -1.16162694e+00 7.51759261e-02
-1.99374743e-03 -1.19173497e-01 -7.60773301e-01 1.41878510e+00
5.81681848e-01 5.14596477e-02 -5.52525222e-02 6.84457302e-01
1.23065019e+00 6.93114340e-01 2.68345982e-01 -2.13748023e-01
1.76888657e+00 -4.33801174e-01 -6.42743111e-01 -6.22315109e-01
4.78197664e-01 -8.24433029e-01 4.75476980e-01 3.46346378e-01
-5.70419610e-01 -2.70952284e-01 -1.34704494e+00 7.22298741e-01
-1.49781272e-01 7.51312613e-01 6.77870989e-01 5.53046763e-01
-5.13509035e-01 3.47716957e-01 -8.21348846e-01 -1.19402632e-01
2.88863540e-01 4.69868220e-02 -1.91570193e-01 -2.13803619e-01
-1.21552944e+00 8.47644210e-01 2.57096708e-01 9.22422409e-01
-3.95706773e-01 -6.80760622e-01 -7.58826911e-01 3.37273836e-01
2.34388664e-01 -3.69735807e-02 1.01344037e+00 -2.80592769e-01
-1.42835796e+00 3.30086082e-01 1.45364299e-01 1.48488030e-01
-2.04724565e-01 -4.09477383e-01 -7.15494335e-01 3.91788512e-01
4.05440927e-01 1.75045535e-01 5.86806953e-01 -1.65664196e+00
-4.44761187e-01 -6.14184618e-01 -5.96888483e-01 -5.64863067e-03
9.97807086e-02 -9.85097364e-02 5.94589673e-02 -4.41875666e-01
8.24635506e-01 -4.56929415e-01 -2.92233497e-01 -2.95393884e-01
-5.42535305e-01 4.11069900e-01 9.44475710e-01 -1.06727064e+00
8.36219132e-01 -1.78948605e+00 -6.50017202e-01 2.75897741e-01
1.85599923e-01 4.18655038e-01 -3.47021878e-01 2.90796399e-01
-3.91628206e-01 -2.28293002e-01 -5.24830222e-01 3.96851599e-01
-4.24499243e-01 -1.70967862e-01 4.83081490e-02 5.21560967e-01
1.76806793e-01 3.78450871e-01 -5.71152270e-01 -3.22360337e-01
2.46881709e-01 8.36346447e-01 -7.14694634e-02 -2.39906430e-01
1.61811635e-01 2.86519438e-01 -8.69828463e-01 1.07642210e+00
1.55728126e+00 -4.70045209e-02 2.03484684e-01 -7.39011168e-01
-3.47063661e-01 -3.62542808e-01 -1.47272491e+00 4.01247352e-01
-7.15637863e-01 3.49609613e-01 6.34749472e-01 -1.48569882e+00
1.48414385e+00 2.59119034e-01 7.31189013e-01 -9.29959238e-01
1.40015647e-01 4.34989780e-01 3.79401334e-02 -5.38013279e-01
4.78180707e-01 -2.14967765e-02 3.16535711e-01 1.77902564e-01
-3.97388607e-01 7.86819607e-02 -7.88174793e-02 -2.00354129e-01
9.43550229e-01 4.58586961e-02 4.18666512e-01 -4.61685181e-01
4.95488971e-01 -2.02575684e-01 8.33348334e-01 5.73947132e-01
7.31967613e-02 2.83930659e-01 2.24965781e-01 -2.47572258e-01
-5.31919003e-01 -9.43139911e-01 -6.27821386e-01 7.00425744e-01
2.65451729e-01 1.27914190e-01 -1.31824717e-01 3.26623842e-02
7.03271478e-02 6.54060125e-01 -4.57531601e-01 -3.52243297e-02
-6.18935764e-01 -1.30359089e+00 4.77337688e-01 4.50141042e-01
7.73731709e-01 -1.30256259e+00 -6.58523679e-01 6.15505815e-01
-5.06751895e-01 -1.19023252e+00 3.13208252e-01 5.34750283e-01
-7.80189753e-01 -1.13883197e+00 -7.63647497e-01 -5.23924291e-01
4.01787430e-01 5.24192452e-01 1.06954718e+00 3.27635892e-02
-7.08497882e-01 1.21841673e-02 -5.36307335e-01 -6.28976464e-01
-8.65042675e-03 7.06991628e-02 -6.61701024e-01 -2.08146930e-01
1.52792171e-01 -8.91243398e-01 -6.23712897e-01 5.16364634e-01
-6.06315017e-01 -1.72033638e-01 9.05830443e-01 7.29140520e-01
6.82182610e-01 2.26278961e-01 7.96749294e-01 -6.94432676e-01
4.28399652e-01 -7.62895584e-01 -6.59542322e-01 2.41510630e-01
-3.13164622e-01 -3.75564456e-01 2.35239312e-01 -2.85680383e-01
-1.14911342e+00 -1.78724453e-01 -2.62980133e-01 4.00376886e-01
-2.04471961e-01 1.02173364e+00 -4.34512794e-01 -1.16549075e-01
4.58755672e-01 1.09949008e-01 -3.22504282e-01 -4.60942388e-01
-1.69949725e-01 5.27792275e-01 5.91482222e-01 -5.65279365e-01
5.62256634e-01 2.50761151e-01 2.91388839e-01 -1.61402738e+00
-9.24696326e-01 -4.29564059e-01 -2.38976806e-01 -3.50485116e-01
5.02304554e-01 -8.18163276e-01 -7.14974940e-01 7.77913928e-01
-1.10327387e+00 -2.17413992e-01 2.49153316e-01 8.08480799e-01
-2.17014432e-01 4.97610122e-01 -4.54660118e-01 -1.15944827e+00
-8.25250328e-01 -8.61491680e-01 1.00691342e+00 4.51799035e-01
1.01458520e-01 -6.16318643e-01 -3.06390915e-02 8.52449387e-02
6.63490713e-01 1.99661434e-01 1.01753700e+00 -2.14730680e-01
-4.27168936e-01 -2.82361001e-01 -7.51028478e-01 8.64660293e-02
1.38471797e-01 1.41413093e-01 -7.98271298e-01 -1.35874718e-01
7.56629219e-04 3.55080627e-02 8.15335393e-01 1.30401814e+00
8.58469427e-01 2.18412369e-01 -5.00734270e-01 8.03466141e-01
1.64839303e+00 4.59526062e-01 9.33381975e-01 5.47297239e-01
5.16327381e-01 7.54802704e-01 9.27867234e-01 7.14959383e-01
4.49687660e-01 2.53780037e-01 8.47569466e-01 -3.06699544e-01
4.67174768e-01 3.76387239e-01 -4.50537689e-02 6.10938787e-01
-3.59082490e-01 -4.07810837e-01 -8.33782196e-01 4.47295010e-01
-1.35953403e+00 -1.00223148e+00 -4.55369323e-01 1.87631869e+00
2.25648850e-01 -3.10203880e-01 -3.58498842e-01 2.50772864e-01
8.98839891e-01 2.60705292e-01 -5.13460457e-01 1.58078375e-03
-2.11611703e-01 6.21465743e-01 7.42054462e-01 4.15821463e-01
-8.87282610e-01 7.69269943e-01 5.11086750e+00 5.57866573e-01
-1.35230803e+00 -2.27873430e-01 2.43317336e-01 5.68590820e-01
-1.82963222e-01 2.55073365e-02 -7.72298396e-01 -7.13161901e-02
5.09853363e-01 4.09267992e-01 -1.88205957e-01 8.43428373e-01
3.63939345e-01 -6.11889243e-01 3.26969139e-02 5.80951512e-01
-3.52336258e-01 -8.14782500e-01 -3.91316295e-01 1.03811860e-01
-3.56184021e-02 1.11397520e-01 -2.81426251e-01 2.55975336e-01
5.08111179e-01 -8.60882044e-01 3.55186403e-01 5.83200157e-01
6.02248490e-01 -8.87680233e-01 8.83795083e-01 1.20399408e-01
-1.65782142e+00 -1.04118899e-01 -5.19972265e-01 3.53491694e-01
2.25671664e-01 1.37537265e+00 -4.86368835e-01 8.76598418e-01
8.83085668e-01 4.63930041e-01 -1.76836953e-01 1.26997352e+00
-2.76480317e-01 5.55592895e-01 -5.71080744e-01 3.27025414e-01
-7.67279714e-02 -4.59129006e-01 5.52731335e-01 9.70951915e-01
1.07787645e+00 2.62194306e-01 1.37217700e-01 9.59803760e-01
6.25280499e-01 4.33956206e-01 -2.85642028e-01 3.07786074e-02
9.86519992e-01 1.68955040e+00 -1.04216778e+00 2.47638673e-01
1.16064630e-01 3.02959591e-01 -2.03051955e-01 4.98301715e-01
-5.86602151e-01 -6.09866619e-01 3.54673952e-01 2.47032300e-01
7.06765532e-01 -2.22323120e-01 -2.71740407e-01 -6.16397858e-01
8.21729824e-02 -7.13129282e-01 9.43312645e-02 -1.05909789e+00
-8.80849421e-01 6.19474173e-01 1.03107043e-01 -9.59037662e-01
1.69367433e-01 -5.67340672e-01 -7.86440611e-01 1.11654198e+00
-1.69505882e+00 -1.28067291e+00 -8.21139753e-01 -5.74346632e-03
-1.16936909e-02 -8.02630838e-03 1.10880375e+00 3.54636699e-01
-7.75113404e-01 1.23739183e-01 2.05401868e-01 -1.90042943e-01
3.88749361e-01 -6.96611404e-01 -2.09643375e-02 5.94099760e-01
-5.98749757e-01 1.65682837e-01 8.91421020e-01 -8.80980730e-01
-9.44444776e-01 -8.37709546e-01 6.11956358e-01 7.18210161e-01
3.96362931e-01 2.04647586e-01 -1.06377125e+00 4.21695560e-01
-3.92239153e-01 7.38611221e-02 5.47661960e-01 1.10654160e-01
6.64590597e-02 -1.91784978e-01 -1.13992131e+00 2.11340219e-01
5.45062959e-01 2.34807432e-02 3.06982428e-01 6.02265932e-02
1.71646863e-01 -1.10968336e-01 -1.11671627e+00 1.07078779e+00
8.74961555e-01 -9.44554329e-01 1.29338193e+00 4.17703331e-01
3.79231602e-01 -1.26442462e-01 -5.63625872e-01 -1.42256570e+00
-6.89198136e-01 2.81436861e-01 2.52822012e-01 1.14769602e+00
1.28791347e-01 -1.05960321e+00 6.00277185e-01 -4.79512334e-01
-3.11504036e-01 -7.54108250e-01 -6.18141592e-01 -4.97858733e-01
-1.32250145e-01 -4.77530926e-01 5.96346498e-01 8.23987305e-01
-6.87961876e-01 2.69565195e-01 -1.44783586e-01 9.32701707e-01
9.60282862e-01 2.98082322e-01 3.56885135e-01 -1.98571813e+00
-4.42024134e-02 -2.10249454e-01 -1.87697306e-01 -8.06827962e-01
3.58969755e-02 -5.09310961e-01 3.35000694e-01 -1.78377473e+00
-2.09750876e-01 -9.48882580e-01 -1.12409055e-01 3.66210461e-01
-1.52095914e-01 -2.08433662e-02 -1.93635091e-01 1.39063448e-01
3.47124755e-01 7.54885852e-01 1.43132341e+00 -4.92070466e-02
-2.87585169e-01 3.32780153e-01 -8.91188741e-01 8.34935904e-01
1.12347543e+00 -5.32011688e-01 -3.89390647e-01 -1.90902233e-01
1.61880441e-02 6.01743579e-01 5.60139120e-01 -1.19946218e+00
-1.15577221e-01 -2.93727815e-01 4.63974506e-01 -1.30626488e+00
4.43866104e-01 -7.27075338e-01 2.16758594e-01 4.22055274e-01
4.52241957e-01 -2.12704957e-01 4.32972163e-01 3.82244110e-01
-1.44786447e-01 -3.35080624e-01 8.34943652e-01 1.19388979e-02
-6.21774614e-01 2.48991683e-01 -6.73204958e-01 -2.22324461e-01
6.48022473e-01 -4.38385367e-01 -5.16048491e-01 -3.24158490e-01
-6.05375111e-01 1.66080013e-01 -3.45824957e-01 -9.68649387e-02
6.33878171e-01 -1.09363854e+00 -9.46337879e-01 2.67598957e-01
-6.47441372e-02 2.59378199e-02 8.21701705e-01 1.04788029e+00
-7.15154290e-01 2.19253376e-01 -5.44905782e-01 -5.82134426e-01
-1.17253888e+00 -3.37534040e-01 5.35749257e-01 -5.26669145e-01
-7.27397799e-01 5.66289783e-01 2.83081621e-01 -6.64118469e-01
-2.27616891e-01 -2.89193243e-01 -8.91704977e-01 2.30523944e-01
4.80010062e-01 4.05109257e-01 2.24015966e-01 -5.90570867e-01
-2.94376314e-01 8.29577446e-01 1.26323029e-02 2.73028225e-01
1.86768270e+00 -1.27415145e-02 -2.14865237e-01 -1.21208420e-02
5.95079899e-01 -9.45884436e-02 -7.24389374e-01 -3.26351911e-01
-1.65594414e-01 -1.09492168e-01 4.83700961e-01 -7.41659999e-01
-1.39187217e+00 7.69016445e-01 6.82274818e-01 1.81178972e-01
1.37025547e+00 -4.09469515e-01 6.61711752e-01 4.75537688e-01
2.36236349e-01 -9.57984507e-01 -2.27158397e-01 7.02067137e-01
9.79032934e-01 -9.53197658e-01 3.00076634e-01 -9.75438654e-01
-1.13402821e-01 1.15930390e+00 6.74721420e-01 6.80942908e-02
1.07709467e+00 7.32422650e-01 2.27490917e-01 -4.33780789e-01
-1.87013686e-01 -2.11705968e-01 -2.11733654e-01 1.12626350e+00
4.53224540e-01 2.87785918e-01 -1.86139449e-01 7.73494899e-01
-1.75111294e-01 3.36113223e-03 2.34513670e-01 1.02313638e+00
-9.26918507e-01 -8.97473812e-01 -7.57926464e-01 6.89455330e-01
-3.60603988e-01 -1.98886305e-01 3.00726175e-01 8.44249785e-01
-5.64541034e-02 8.04224432e-01 -1.57109678e-01 -2.45522827e-01
3.82430255e-01 -4.29003090e-01 3.66040707e-01 -3.94785821e-01
-6.92741759e-03 2.01867431e-01 2.28663027e-01 -1.37753770e-01
-2.65616894e-01 -5.99162281e-01 -1.15002394e+00 -1.54325083e-01
-7.74609149e-01 9.61110145e-02 8.84247005e-01 8.66062820e-01
4.57972102e-02 9.24665093e-01 5.24319947e-01 -1.06691241e+00
-3.98368418e-01 -1.11454213e+00 -1.25727320e+00 -5.88698089e-01
-3.24432962e-02 -1.10676050e+00 -2.53451288e-01 -3.78049850e-01] | [6.844038963317871, 1.4016773700714111] |
0b8e36d3-02f1-41a2-a3c3-97d4595a41c1 | non-parametric-cumulants-approach-for-outlier | 2305.10911 | null | https://arxiv.org/abs/2305.10911v1 | https://arxiv.org/pdf/2305.10911v1.pdf | Non-parametric cumulants approach for outlier detection of multivariate financial data | In this paper, we propose an outlier detection algorithm for multivariate data based on their projections on the directions that maximize the Cumulant Generating Function (CGF). We prove that CGF is a convex function, and we characterize the CGF maximization problem on the unit n-circle as a concave minimization problem. Then, we show that the CGF maximization approach can be interpreted as an extension of the standard principal component technique. Therefore, for validation and testing, we provide a thorough comparison of our methodology with two other projection-based approaches both on artificial and real-world financial data. Finally, we apply our method as an early detector for financial crises. | ['Jacopo Maria Ricci', 'Rosella Giacometti', 'Francesco Cesarone'] | 2023-05-18 | null | null | null | null | ['outlier-detection'] | ['methodology'] | [-2.00978145e-01 -9.17121470e-02 4.64776486e-01 -2.15798393e-01
-3.69022220e-01 -5.39472759e-01 5.67096710e-01 8.40720981e-02
-4.59214389e-01 4.57894951e-01 1.05832517e-01 -5.26140988e-01
-1.61130339e-01 -5.85018277e-01 -5.72099209e-01 -7.23149121e-01
-2.60091901e-01 3.81154746e-01 -2.69572679e-02 1.46499574e-01
6.19857669e-01 7.77957499e-01 -8.25581610e-01 -3.02777678e-01
8.56360614e-01 9.73164260e-01 -3.14693063e-01 5.47227859e-01
2.45749280e-01 3.32832813e-01 -6.48456216e-01 -5.24293125e-01
7.30450034e-01 -2.78952569e-01 -3.56713891e-01 4.75016981e-01
3.43704596e-02 -3.00176114e-01 -9.89705995e-02 1.41787422e+00
3.12594026e-01 1.76141128e-01 9.49363410e-01 -1.48954606e+00
-4.27745819e-01 4.97630239e-02 -5.41618764e-01 2.30995357e-01
3.70583504e-01 -2.33647242e-01 6.09288037e-01 -1.45398748e+00
4.87364680e-01 8.30203652e-01 6.60749972e-01 -1.65344756e-02
-1.23059845e+00 -7.79007897e-02 -1.74471319e-01 1.00346573e-01
-1.13252115e+00 -3.26952606e-01 1.00729001e+00 -6.39653623e-01
8.05716455e-01 4.71272975e-01 6.61623001e-01 6.24453187e-01
1.96795538e-01 9.16691184e-01 9.88753915e-01 -5.69389999e-01
5.49209595e-01 -1.97402030e-01 2.28014737e-01 2.79415786e-01
6.76245332e-01 8.61127898e-02 -2.59616971e-01 -7.20422983e-01
6.68558657e-01 1.65032983e-01 -6.60910606e-01 -5.46103477e-01
-9.36459363e-01 1.08862460e+00 -3.50925356e-01 3.95236276e-02
-5.69094896e-01 -3.02258372e-01 1.95928216e-02 3.46706837e-01
7.56346643e-01 3.92294466e-01 -1.44186929e-01 -1.75854877e-01
-1.11672699e+00 2.63592213e-01 1.33475959e+00 5.97152889e-01
4.18202281e-01 3.01973820e-01 1.85496688e-01 4.09061164e-01
5.65434873e-01 7.94586539e-01 5.30886829e-01 -1.05707693e+00
4.00904059e-01 5.55868983e-01 2.99943119e-01 -1.21027732e+00
-4.31756526e-01 -3.01676691e-01 -7.32902110e-01 3.83838505e-01
6.85983539e-01 -3.38044703e-01 -2.88065642e-01 1.30612326e+00
7.69620463e-02 2.43395731e-01 1.87697992e-01 7.23348379e-01
-1.25356227e-01 4.23615128e-01 -6.81838870e-01 -7.69078732e-01
7.29584992e-01 -5.75920582e-01 -6.58136129e-01 3.07323188e-01
3.14805061e-01 -8.66572440e-01 7.39298582e-01 8.18547428e-01
-8.56920719e-01 1.79978788e-01 -9.14361596e-01 5.53140521e-01
6.27943948e-02 1.25883579e-01 4.63453114e-01 7.66284347e-01
-9.25572932e-01 7.85087526e-01 -1.00395644e+00 -2.24283382e-01
-7.26217031e-02 2.50408743e-02 -3.18160713e-01 1.89201176e-01
-5.20320773e-01 6.25956953e-01 2.27144808e-01 2.20972911e-01
-6.15693271e-01 -1.25573203e-01 -6.70079768e-01 -4.00451332e-04
1.44020826e-01 -3.89546275e-01 1.05250680e+00 -5.69752097e-01
-1.51555347e+00 6.07957780e-01 -2.15166762e-01 -5.37761748e-01
8.11877489e-01 -4.67699826e-01 -3.46065283e-01 2.79717177e-01
6.98000789e-02 -5.20006657e-01 1.11199629e+00 -9.73036230e-01
-2.26291269e-01 -4.69419986e-01 -6.02173626e-01 -9.37088951e-02
-4.26728696e-01 3.48720253e-01 -4.47900817e-02 -9.75465119e-01
5.45666933e-01 -7.22710311e-01 -4.37184215e-01 -5.24948895e-01
-5.92882633e-01 -7.69652352e-02 6.31383121e-01 -9.05187547e-01
1.35713911e+00 -2.28751111e+00 4.98917289e-02 8.34600210e-01
9.09739211e-02 -1.14217453e-01 3.00742388e-01 6.72198355e-01
-3.96133929e-01 -1.22984432e-01 -8.91304672e-01 -4.53288645e-01
3.67963314e-02 3.11720315e-02 -6.88481033e-01 1.17043304e+00
1.59335230e-02 4.90853101e-01 -8.63212883e-01 3.08904075e-03
-1.11126211e-02 5.35489693e-02 -4.39055949e-01 1.05150871e-01
3.44785243e-01 1.40213653e-01 -1.70515046e-01 8.20051312e-01
9.61290717e-01 1.16787747e-01 1.00388035e-01 4.02250588e-01
-2.74148166e-01 -3.52135509e-01 -1.47352731e+00 1.13831878e+00
2.77522773e-01 5.70347428e-01 -1.65922865e-02 -1.45402789e+00
1.24620986e+00 3.15752357e-01 7.27068245e-01 -2.55349815e-01
3.90135385e-02 5.67651629e-01 -3.09559315e-01 -2.79545546e-01
3.94186884e-01 -2.00636774e-01 1.28272951e-01 7.93836117e-01
-2.25710988e-01 1.91018015e-01 4.24369007e-01 2.91008741e-01
9.36731100e-01 -1.72953218e-01 4.53377873e-01 -5.44579029e-01
3.74751329e-01 -2.96477884e-01 6.00323319e-01 6.21449888e-01
-3.28950226e-01 8.01392138e-01 1.02380478e+00 -2.43147910e-01
-1.12815368e+00 -1.18679237e+00 -4.42164540e-01 2.24434420e-01
-3.86346102e-01 -3.62884969e-01 -8.77584219e-01 -5.69683552e-01
2.31889680e-01 6.99703038e-01 -3.43662918e-01 1.15940668e-01
-5.41594088e-01 -1.44668460e+00 3.56033444e-01 3.22972089e-01
2.61649281e-01 -8.53237867e-01 -7.15030730e-01 1.87808424e-01
-5.48427925e-02 -8.43336880e-01 -3.81527156e-01 4.04298194e-02
-1.15211749e+00 -1.45481002e+00 -1.03504086e+00 -3.71578068e-01
9.07389224e-01 1.27907887e-01 1.01666236e+00 -2.84197867e-01
2.22118683e-02 6.87162459e-01 -3.37077886e-01 -3.54450911e-01
-1.16378710e-01 -7.73257434e-01 4.73632425e-01 4.08443302e-01
4.45443302e-01 -7.13931441e-01 -4.90414143e-01 2.19066426e-01
-9.66293216e-01 -6.78510010e-01 2.61961251e-01 7.31023610e-01
5.27989507e-01 1.47316411e-01 3.07651132e-01 -7.00808406e-01
1.25936127e+00 -7.01322377e-01 -1.12384880e+00 1.51498258e-01
-9.36350882e-01 -3.03232577e-04 6.27184272e-01 -3.17310005e-01
-7.23705292e-01 2.36846477e-01 3.14258993e-01 -5.76768696e-01
4.88837697e-02 8.41586471e-01 2.05646968e-03 7.17275068e-02
4.72831517e-01 4.39245433e-01 -4.78297286e-02 -8.94955158e-01
2.16421619e-01 4.91244674e-01 8.42123747e-01 -5.76199710e-01
6.56682074e-01 7.51905680e-01 2.74343282e-01 -1.03368449e+00
-3.13154787e-01 -8.47339988e-01 -7.02194989e-01 -1.17859952e-01
3.93135339e-01 -4.40196037e-01 -5.85773230e-01 7.20304012e-01
-1.10121429e+00 7.34021589e-02 -3.30208808e-01 7.83678174e-01
-7.76963770e-01 1.03841162e+00 -5.93060672e-01 -1.38515472e+00
-2.67000914e-01 -7.29364753e-01 6.88607574e-01 -4.20594886e-02
8.63769203e-02 -1.09517920e+00 7.62310863e-01 -3.12864482e-01
1.72590360e-01 7.03039348e-01 3.54173690e-01 -1.12524986e+00
-8.64415616e-02 -5.90617001e-01 -1.11539081e-01 7.79604256e-01
-3.48000556e-01 1.96319759e-01 -5.11142194e-01 -4.48712111e-01
9.95150566e-01 3.52464288e-01 8.27030599e-01 5.64231336e-01
6.28247738e-01 -5.33175528e-01 3.23949903e-02 8.99369657e-01
1.49797499e+00 4.70781960e-02 4.37890679e-01 6.60887420e-01
2.31217146e-01 5.12730777e-01 4.92486358e-01 9.08757329e-01
-1.99656576e-01 3.03411335e-01 4.54278022e-01 2.81329304e-01
7.77133703e-01 4.17040735e-02 6.84395611e-01 9.61680472e-01
-2.38869518e-01 9.03307348e-02 -1.03858888e+00 4.39318597e-01
-2.14764690e+00 -1.11529815e+00 -5.99813342e-01 2.54973125e+00
3.07175368e-01 -8.19303915e-02 4.29101914e-01 2.66181380e-01
6.97673023e-01 -3.53838295e-01 -3.28895748e-01 -2.77781874e-01
-5.06476402e-01 2.35263780e-01 5.52875996e-01 4.88460004e-01
-1.27526069e+00 3.15400660e-01 7.78179884e+00 3.16801935e-01
-7.09364891e-01 -1.25011072e-01 4.21592265e-01 1.47720858e-01
-1.09043039e-01 1.88717544e-01 -3.37846279e-01 7.08430588e-01
8.34347785e-01 -4.88285780e-01 4.03018296e-01 9.86664712e-01
3.74026209e-01 -1.08616062e-01 -8.81888270e-01 1.02756381e+00
2.00972229e-01 -6.88416243e-01 -3.03756148e-01 4.58918422e-01
8.48730505e-01 -5.34475520e-02 3.01212799e-02 -3.34424525e-01
9.85046178e-02 -7.84639776e-01 4.89440411e-01 7.43077636e-01
2.30026007e-01 -9.65261459e-01 8.60091150e-01 3.46808463e-01
-7.95042932e-01 -1.03226453e-01 -6.90510333e-01 -2.91399956e-01
4.36722875e-01 1.29155755e+00 -6.92315936e-01 6.75214112e-01
4.16604340e-01 5.74959755e-01 -3.94926697e-01 1.56672204e+00
-4.02203321e-01 6.87244475e-01 -7.06874549e-01 2.99053460e-01
-5.48033379e-02 -1.08277857e+00 1.07876205e+00 1.25490069e+00
7.77361035e-01 1.07274339e-01 1.68234512e-01 8.31472456e-01
2.18437910e-01 3.92363131e-01 -6.74309850e-01 -3.65367793e-02
1.60365999e-01 1.03995728e+00 -8.32907379e-01 -3.67051691e-01
-5.30600905e-01 1.08022618e+00 4.73781116e-02 6.56652749e-01
-3.61290425e-01 -4.96777296e-01 4.09666538e-01 -1.48040250e-01
1.50844917e-01 -4.99569416e-01 -3.99168611e-01 -1.70799446e+00
4.88027006e-01 -7.85375476e-01 4.57654476e-01 -3.80726039e-01
-1.31157994e+00 3.77992630e-01 8.96127298e-02 -1.40481818e+00
-6.08982682e-01 -9.80856299e-01 -1.05974245e+00 7.66584456e-01
-1.08227038e+00 -1.68395951e-01 2.04906583e-01 8.63383591e-01
-1.33057088e-01 -4.45131212e-01 6.67231679e-01 1.47323415e-01
-8.20102572e-01 1.07774794e-01 7.68689871e-01 3.03928792e-01
2.98298657e-01 -1.60700107e+00 3.66007090e-01 1.72864187e+00
2.76985168e-01 6.50542319e-01 8.45035493e-01 -7.59767413e-01
-9.70045865e-01 -6.62258148e-01 8.71090055e-01 -3.72244030e-01
9.34367597e-01 4.96147908e-02 -7.66603172e-01 8.62161040e-01
-5.80407977e-02 -1.25615329e-01 7.33955204e-01 -6.68718070e-02
-3.54443281e-03 4.58918124e-01 -1.19243610e+00 3.98915738e-01
4.79610920e-01 -2.64217764e-01 -8.85275126e-01 5.69919586e-01
1.71251684e-01 1.62955374e-01 -9.01148260e-01 3.89182746e-01
2.03901917e-01 -1.26084983e+00 9.45126176e-01 -7.58456111e-01
1.18485168e-01 -3.16356570e-01 -3.83525580e-01 -1.24358034e+00
-1.15955293e-01 -1.07496178e+00 -4.89128798e-01 8.00113738e-01
6.63010627e-02 -1.04166782e+00 6.60199881e-01 4.67940897e-01
-7.01420307e-02 -6.88377857e-01 -1.22605646e+00 -1.17443156e+00
5.65022938e-02 -7.29340851e-01 2.11785451e-01 1.02520561e+00
5.63615799e-01 -2.17680767e-01 -5.05020738e-01 2.52358973e-01
1.00815535e+00 1.98067233e-01 5.95861197e-01 -1.25559068e+00
-4.73584592e-01 -4.83366430e-01 -4.55939204e-01 -8.88819754e-01
-1.56798273e-01 -5.93727529e-01 -1.22173324e-01 -1.00226629e+00
2.12811694e-01 1.56742394e-01 -2.45306551e-01 -9.97970626e-02
-7.48277158e-02 -2.24405490e-02 2.22976118e-01 5.86920738e-01
-1.23564869e-01 4.95010257e-01 6.00572407e-01 3.02465111e-01
-1.68816000e-01 3.69630754e-01 -3.03070664e-01 1.10670757e+00
7.26761460e-01 -5.13062239e-01 -4.96578217e-02 2.36236434e-02
3.94420177e-01 6.50762022e-02 2.45786458e-01 -8.42353463e-01
1.48799032e-01 -7.67404214e-02 2.26964220e-01 -7.60804534e-01
5.40167242e-02 -6.35512173e-01 -1.26225144e-01 6.46145284e-01
4.39889938e-01 5.41829288e-01 -3.21006149e-01 6.07136607e-01
-3.96937847e-01 -6.91784382e-01 3.94096792e-01 -7.64758289e-02
-2.12416470e-01 1.50188595e-01 -3.97714227e-01 3.34077552e-02
9.93547916e-01 -2.06999555e-02 -1.21531293e-01 -6.31052554e-01
-8.74968767e-01 -1.90860927e-02 7.29008615e-01 -3.21129233e-01
1.03993428e+00 -1.27445066e+00 -9.23914075e-01 4.77328539e-01
-3.08392912e-01 -3.52106184e-01 -2.98432499e-01 1.26145279e+00
-7.77478516e-01 4.66038913e-01 4.57500406e-02 -4.21248436e-01
-8.52885544e-01 9.51102436e-01 1.30143374e-01 -3.53213698e-01
-6.17341220e-01 4.68981296e-01 -6.17294386e-02 -2.47413009e-01
2.28172503e-02 -2.47369409e-01 -7.83402324e-02 2.08350588e-02
4.61095273e-01 8.92953336e-01 -7.65160918e-02 -6.79355443e-01
-3.39625239e-01 2.76526332e-01 6.05706513e-01 -6.78840578e-01
1.41368091e+00 2.49210373e-02 -5.30166328e-01 3.13802272e-01
1.21479261e+00 4.28452581e-01 -1.11986017e+00 5.97508512e-02
5.74458778e-01 -5.41457951e-01 -3.53585273e-01 -2.79959068e-02
-9.37929630e-01 5.74690998e-01 3.27674419e-01 7.00165153e-01
1.29117894e+00 -3.08773667e-01 4.70903903e-01 4.66270983e-01
3.03984843e-02 -1.21903312e+00 3.18871364e-02 4.92149174e-01
9.42421794e-01 -8.90255451e-01 1.48921072e-01 -3.06348979e-01
-3.94543707e-01 1.53307819e+00 -1.68835744e-01 -6.57794714e-01
1.02898073e+00 1.39401913e-01 -5.78968935e-02 -1.85140714e-01
-3.14790487e-01 1.26398996e-01 2.64341772e-01 4.20740902e-01
2.62366354e-01 2.64221102e-01 -8.17661345e-01 5.77672362e-01
-4.14442122e-01 -1.63125724e-01 8.59136701e-01 9.52402532e-01
-5.82976520e-01 -1.02189529e+00 -8.30487847e-01 3.74364942e-01
-5.95481515e-01 -2.19066143e-01 -4.47897196e-01 4.79271412e-01
-4.87363607e-01 8.98683608e-01 5.11178486e-02 4.44236770e-02
4.03492659e-01 3.83660287e-01 1.12290472e-01 -2.37480998e-01
1.15493849e-01 3.47640902e-01 -2.79604793e-01 -5.52946031e-01
-3.14242929e-01 -1.19560993e+00 -7.54249871e-01 -2.68019646e-01
-3.06164950e-01 2.89131105e-01 6.35605931e-01 8.54644299e-01
-3.07410937e-02 -9.48615968e-02 8.96362662e-01 -6.09996319e-01
-1.17822933e+00 -1.01083457e+00 -1.00878632e+00 5.46925247e-01
1.61406174e-01 -3.30113411e-01 -8.51073742e-01 -6.54290151e-03] | [7.2955217361450195, 4.090228080749512] |
b33ffdfc-45e8-4063-b5a2-a88471ad6e49 | blind-universal-bayesian-image-denoising-with | 1907.03029 | null | https://arxiv.org/abs/1907.03029v2 | https://arxiv.org/pdf/1907.03029v2.pdf | Blind Universal Bayesian Image Denoising with Gaussian Noise Level Learning | Blind and universal image denoising consists of using a unique model that denoises images with any level of noise. It is especially practical as noise levels do not need to be known when the model is developed or at test time. We propose a theoretically-grounded blind and universal deep learning image denoiser for additive Gaussian noise removal. Our network is based on an optimal denoising solution, which we call fusion denoising. It is derived theoretically with a Gaussian image prior assumption. Synthetic experiments show our network's generalization strength to unseen additive noise levels. We also adapt the fusion denoising network architecture for image denoising on real images. Our approach improves real-world grayscale additive image denoising PSNR results for training noise levels and further on noise levels not seen during training. It also improves state-of-the-art color image denoising performance on every single noise level, by an average of 0.1dB, whether trained on or not. | ['Sabine Süsstrunk', 'Majed El Helou'] | 2019-07-05 | null | null | null | null | ['color-image-denoising'] | ['computer-vision'] | [ 5.72174549e-01 -5.20620942e-01 6.85239673e-01 -2.16649085e-01
-8.78190100e-01 -3.95393461e-01 2.51607716e-01 -3.12210590e-01
-5.98353922e-01 5.17346919e-01 7.12122954e-03 -2.84443885e-01
-4.95698266e-02 -8.50634158e-01 -8.27778995e-01 -1.09613132e+00
1.01949577e-03 -5.75112514e-02 -1.87677182e-02 -4.85693932e-01
-7.34316483e-02 4.24562842e-01 -1.56504548e+00 2.77184606e-01
8.57162118e-01 1.12712717e+00 4.50391680e-01 1.30209792e+00
1.42503604e-01 5.44424653e-01 -8.79625380e-01 -3.59909475e-01
8.42542589e-01 -6.06437147e-01 -2.46192843e-01 3.60723257e-01
5.77289283e-01 -6.20467961e-01 -4.59450960e-01 1.62489414e+00
7.87147105e-01 1.41020566e-01 3.10780972e-01 -7.32339442e-01
-9.55715656e-01 3.41478676e-01 -3.51043731e-01 1.68584645e-01
-8.78411159e-02 4.54057068e-01 3.76555741e-01 -4.64348584e-01
2.81436175e-01 1.11607361e+00 1.06941867e+00 6.08845234e-01
-1.39185226e+00 -4.46793139e-01 -2.45700836e-01 1.64585382e-01
-1.14147067e+00 -5.69821119e-01 5.97655535e-01 4.98150140e-02
5.08102894e-01 2.49741867e-01 1.40520915e-01 1.12973058e+00
3.02396834e-01 3.34889144e-01 1.52234578e+00 -4.99009013e-01
3.85970831e-01 -4.56037372e-01 -1.86872154e-01 2.19910309e-01
2.46582747e-01 2.02784553e-01 -1.86677620e-01 3.84469688e-01
1.02032781e+00 -1.93124443e-01 -5.03899753e-01 9.32722643e-04
-8.30111086e-01 4.69010293e-01 5.37896574e-01 3.03361475e-01
-6.91139162e-01 5.57326198e-01 1.48898825e-01 8.21263731e-01
4.67798293e-01 -5.45807406e-02 -4.84776944e-01 1.98006555e-02
-1.07709229e+00 -8.28092396e-02 7.46523619e-01 4.88333583e-01
7.71622598e-01 5.41540802e-01 7.86795765e-02 1.00902760e+00
1.10636637e-01 8.15386891e-01 3.79235864e-01 -1.53835142e+00
-9.14093927e-02 -1.71563953e-01 3.06678206e-01 -8.66685331e-01
-7.74841234e-02 -6.38041019e-01 -1.42952216e+00 9.46307123e-01
4.69389379e-01 -2.13932782e-01 -1.49170876e+00 1.62489140e+00
-1.48947299e-01 3.79159451e-01 2.34595180e-01 1.03390884e+00
6.59888148e-01 5.47780693e-01 -2.12037817e-01 -3.57299894e-01
1.12246716e+00 -5.68903387e-01 -1.05878961e+00 -4.58957702e-01
-2.42439881e-01 -9.89948869e-01 8.72531772e-01 9.96226132e-01
-1.25215840e+00 -8.63359451e-01 -1.17295325e+00 -1.96088925e-02
-2.97143698e-01 -9.52928215e-02 3.01417321e-01 1.32753694e+00
-1.59062755e+00 1.00741565e+00 -7.50640810e-01 -1.89656556e-01
3.31368506e-01 3.47471744e-01 -5.44748425e-01 -5.16358256e-01
-1.09753144e+00 7.99894929e-01 3.69944908e-02 6.53655767e-01
-1.27386379e+00 -4.19293314e-01 -6.97334349e-01 -9.72666033e-03
5.68318628e-02 -8.38971794e-01 1.16569769e+00 -1.29963517e+00
-1.55696809e+00 6.65297985e-01 -3.10945779e-01 -6.94118202e-01
7.36698687e-01 -3.32292885e-01 -7.01315582e-01 3.07287931e-01
-7.03361854e-02 2.54985303e-01 1.46284556e+00 -1.95699120e+00
-3.17358315e-01 -2.50358611e-01 -8.49726051e-02 -1.19602285e-01
-1.78801954e-01 -1.37729421e-01 -5.54900765e-01 -8.81600797e-01
4.30151254e-01 -3.74365687e-01 -5.10758340e-01 3.98697928e-02
-1.22329958e-01 5.65727234e-01 7.81886756e-01 -1.14861333e+00
6.94222391e-01 -2.17538285e+00 -1.21225685e-01 3.92454505e-01
1.22621454e-01 2.27941796e-01 -5.02905965e-01 1.82249710e-01
-3.05441439e-01 2.14795068e-01 -5.54429352e-01 -4.77992713e-01
-2.06217319e-02 5.45356274e-01 -1.45481110e-01 6.19380713e-01
-1.48546651e-01 4.84072298e-01 -7.57955730e-01 4.47450355e-02
4.10624027e-01 8.49481821e-01 -2.72082746e-01 2.17581064e-01
2.31710091e-01 4.16204512e-01 3.30673873e-01 5.13824522e-01
1.17702854e+00 1.50920704e-01 -5.90761453e-02 -5.68939805e-01
1.41327813e-01 -4.26536649e-01 -1.57381475e+00 1.53860176e+00
-6.67018771e-01 6.87027276e-01 9.12164092e-01 -9.33069706e-01
9.29140210e-01 3.74215841e-01 5.94485924e-02 -8.96127582e-01
2.55678862e-01 3.45092416e-01 -3.70457619e-02 -4.05954599e-01
2.78205037e-01 -4.25411254e-01 3.62790167e-01 1.23473175e-01
3.19814891e-01 -1.94289163e-01 3.00336480e-01 2.66255047e-02
1.33744228e+00 -2.28878394e-01 -6.31357506e-02 -3.73686343e-01
4.96008724e-01 -5.65267265e-01 5.13486624e-01 1.40187335e+00
-2.85988152e-01 1.07188582e+00 1.95385203e-01 -2.25359008e-01
-1.11860490e+00 -1.23030114e+00 -1.40217274e-01 9.42869186e-01
2.59769648e-01 3.58191058e-02 -9.83556926e-01 -1.59396961e-01
-3.33991438e-01 6.77955031e-01 -6.69217527e-01 -2.28455607e-02
-4.04187858e-01 -8.90252531e-01 3.75636190e-01 3.18251520e-01
9.28405702e-01 -8.71135652e-01 -1.04965620e-01 1.64515346e-01
-1.36134759e-01 -1.11741924e+00 -1.65850580e-01 5.58883190e-01
-8.34012687e-01 -9.25284684e-01 -7.87147999e-01 -7.14380383e-01
6.81033373e-01 3.57859462e-01 1.16505218e+00 1.91407442e-01
-1.15401000e-01 7.27889657e-01 -3.18282306e-01 -1.61544785e-01
-7.74568319e-01 -6.75348341e-01 9.31709930e-02 1.69083074e-01
3.90468836e-02 -9.64529574e-01 -8.32066834e-01 2.37056941e-01
-1.33410239e+00 -5.92815101e-01 6.24759972e-01 9.97362912e-01
4.81367141e-01 9.74535048e-01 3.66511524e-01 -4.48252738e-01
9.52439010e-01 -1.66309580e-01 -5.66078842e-01 -1.39831856e-04
-4.31094229e-01 -6.00414874e-04 7.64571071e-01 -2.79339254e-01
-1.22782397e+00 4.67479182e-03 -4.39862907e-01 -5.32440424e-01
-4.91605371e-01 2.82113284e-01 -3.51319760e-01 -4.74211305e-01
9.37008083e-01 3.02781224e-01 1.74310371e-01 -8.96181643e-01
4.71065521e-01 5.86132884e-01 1.33251786e+00 -3.46085161e-01
1.07216811e+00 6.46642983e-01 1.07219592e-01 -9.09881353e-01
-2.79794097e-01 -2.96266049e-01 -3.87867928e-01 -1.27507463e-01
6.37267351e-01 -1.04766428e+00 -8.00082147e-01 1.07586133e+00
-1.01571667e+00 -4.68852222e-01 -6.44022524e-02 4.93684262e-02
-3.17804635e-01 7.08700538e-01 -9.44963336e-01 -8.32425296e-01
-2.28346244e-01 -1.26876378e+00 7.17373371e-01 6.52458370e-02
4.17030424e-01 -1.05077374e+00 -4.14516658e-01 2.19757900e-01
8.44172597e-01 1.32112354e-01 3.98273915e-01 -9.08630118e-02
-4.10327256e-01 -3.50608140e-01 -2.21886948e-01 1.32085323e+00
3.14682990e-01 -2.65921324e-01 -1.13754439e+00 -4.81801182e-01
5.48982084e-01 -4.25191224e-02 1.43925381e+00 6.73450291e-01
1.19811225e+00 -1.35750458e-01 4.02835280e-01 8.15431416e-01
1.81867337e+00 -1.80450361e-02 1.22891021e+00 4.87567335e-01
5.00886202e-01 1.31931454e-01 -2.00572968e-01 1.30021244e-01
-2.82877356e-01 1.29695177e-01 9.04805481e-01 -4.53481764e-01
-4.62154955e-01 2.34992549e-01 3.22997183e-01 5.05186498e-01
-1.83300182e-01 -4.08095032e-01 -7.00243115e-01 6.98921859e-01
-1.30749857e+00 -8.07911217e-01 -3.55352461e-01 2.21658015e+00
9.40681517e-01 1.39178127e-01 -4.31934059e-01 4.13743615e-01
7.55350530e-01 -7.37383962e-02 -2.94456065e-01 -3.75513136e-01
-6.80397749e-01 7.50232816e-01 9.04522061e-01 7.77228236e-01
-1.32602632e+00 6.90820158e-01 6.70232058e+00 8.38834286e-01
-9.83808279e-01 1.58405632e-01 6.86980486e-01 2.72396445e-01
-2.49931421e-02 -2.68747002e-01 4.09301929e-02 4.48112816e-01
9.15510297e-01 2.42883950e-01 8.63371313e-01 5.35301805e-01
3.76577795e-01 -2.78043061e-01 -6.06345296e-01 1.26093733e+00
-4.50501293e-02 -9.43596125e-01 -1.01870365e-01 -1.76167473e-01
8.75049531e-01 -4.81153317e-02 1.98354840e-01 -1.11031421e-01
6.43075943e-01 -1.25437200e+00 4.95418578e-01 6.92880869e-01
6.19917810e-01 -7.47972548e-01 1.14354420e+00 1.59903020e-01
-6.17012501e-01 -8.75099748e-02 -4.13286388e-01 -1.06016822e-01
1.73175111e-01 9.05095160e-01 -1.27176225e-01 5.38014293e-01
1.02612770e+00 4.10369664e-01 -5.60328901e-01 1.17680979e+00
-5.49168468e-01 9.60350275e-01 -4.30612832e-01 7.92959630e-01
1.35189146e-01 -2.77788848e-01 5.19992888e-01 1.30888450e+00
5.31382799e-01 -7.82171637e-03 -1.61926016e-01 4.36435819e-01
-2.19902202e-01 -4.96516645e-01 -2.28306919e-01 5.05570471e-01
1.52255744e-01 1.18752933e+00 -7.53975093e-01 -1.41923130e-01
-6.53378889e-02 1.60542345e+00 -5.79594791e-01 9.18709636e-01
-4.31672782e-01 -4.32248682e-01 7.55669892e-01 -2.63581961e-01
6.73973560e-01 -1.75281271e-01 -5.54280579e-01 -7.69740939e-01
-9.21006575e-02 -1.15634120e+00 -8.63115769e-03 -9.07688618e-01
-1.67074120e+00 5.43962121e-01 -4.59542155e-01 -1.05999053e+00
-9.75842297e-04 -8.88373077e-01 -6.35697722e-01 1.18748236e+00
-1.47026944e+00 -9.09283221e-01 -5.36739409e-01 7.15166807e-01
1.73172548e-01 -5.89748658e-02 8.07505131e-01 4.33597475e-01
-3.19000393e-01 4.24129665e-01 6.02187634e-01 1.74006641e-01
8.04420412e-01 -1.40982974e+00 4.54672784e-01 1.52445519e+00
-1.80556744e-01 7.22428679e-01 1.37134886e+00 -4.79758590e-01
-1.34787166e+00 -9.85688984e-01 2.21521661e-01 -8.78028423e-02
5.19667089e-01 -2.24754170e-01 -1.13983166e+00 3.38333011e-01
6.42610848e-01 2.92191714e-01 2.35723063e-01 -2.30599582e-01
-5.04036963e-01 -5.45520186e-01 -1.53654778e+00 5.20093858e-01
8.32752526e-01 -4.78660077e-01 -3.30085367e-01 2.64476985e-01
5.22317827e-01 -3.98545295e-01 -6.99749529e-01 3.32858264e-01
2.58170635e-01 -1.26146936e+00 1.25645006e+00 -1.82909757e-01
4.06444848e-01 -6.71554208e-01 -4.56222504e-01 -1.52333283e+00
-4.20479357e-01 -8.72905791e-01 9.58379284e-02 1.04094660e+00
8.59599859e-02 -5.86240768e-01 4.55357134e-01 1.69660255e-01
-1.24781825e-01 -1.03665562e-02 -1.11178160e+00 -9.02531683e-01
6.04099222e-03 -8.28490853e-01 2.37630859e-01 7.18916714e-01
-1.03677630e+00 -8.27546567e-02 -6.97173238e-01 6.20030701e-01
1.25629056e+00 -5.05353272e-01 6.71507418e-01 -8.81103158e-01
-3.33828956e-01 -2.71746099e-01 -5.42392254e-01 -1.04041076e+00
-1.55962706e-01 -2.56889641e-01 4.15717781e-01 -1.77795422e+00
-3.17860842e-01 1.43262178e-01 -4.67399508e-01 4.58116502e-01
-2.71975338e-01 8.46696913e-01 4.78302911e-02 -7.38467276e-02
-2.95349300e-01 3.19841057e-01 1.13371289e+00 -2.12559491e-01
1.36467844e-01 -9.09511223e-02 -8.67003262e-01 6.75632775e-01
9.00811315e-01 -3.37325484e-01 -1.52199179e-01 -6.88412309e-01
2.40503587e-02 -2.24798352e-01 8.45910132e-01 -1.37971115e+00
3.28419000e-01 2.12642357e-01 6.94448173e-01 -1.67943090e-01
3.39715302e-01 -1.16201115e+00 3.35204631e-01 3.82900536e-01
3.04233339e-02 -3.93284619e-01 3.80904585e-01 7.22301662e-01
-2.57096589e-01 -2.92571783e-01 1.07303166e+00 -2.98792809e-01
-9.85766351e-01 -2.65059769e-01 -2.74334162e-01 -3.66095662e-01
3.49364460e-01 -4.33888674e-01 -3.66972029e-01 -9.33397710e-01
-9.49969232e-01 -1.77524030e-01 5.40072143e-01 5.01913503e-02
6.63257897e-01 -1.02580369e+00 -1.00171435e+00 3.38431746e-01
-4.88695979e-01 -4.26718116e-01 6.45540893e-01 5.18250525e-01
-7.93288291e-01 -4.74435568e-01 -1.72250792e-01 -4.48410690e-01
-1.12128425e+00 5.81891239e-01 6.23585880e-01 1.28791064e-01
-4.19927120e-01 1.01162589e+00 -1.76377103e-01 -1.32911623e-01
1.70442730e-01 -1.45781666e-01 2.09251150e-01 -4.00033206e-01
8.07371140e-01 4.31582659e-01 3.83281380e-01 -6.53386831e-01
3.01232692e-02 4.49548483e-01 2.18057245e-01 -1.12190023e-01
1.37408614e+00 -2.51747668e-01 -5.42910516e-01 1.35273486e-01
1.17179883e+00 5.67060784e-02 -1.43295908e+00 -1.32781208e-01
-4.48982805e-01 -6.24310374e-01 6.74347281e-01 -1.10993361e+00
-1.43760109e+00 7.34383762e-01 1.38905990e+00 4.57101882e-01
1.82185459e+00 -5.66407561e-01 7.06067562e-01 4.00086015e-01
1.16541564e-01 -9.30510342e-01 -1.55995622e-01 2.81062365e-01
1.03173137e+00 -1.22098017e+00 -7.69182444e-02 -1.54831380e-01
-7.84091353e-02 1.20684063e+00 1.54516816e-01 -3.05664420e-01
6.11717463e-01 6.45884752e-01 6.32958353e-01 1.64359733e-01
-7.13133588e-02 -3.89392346e-01 4.13597897e-02 1.00637960e+00
5.16573451e-02 -1.94182277e-01 1.16662001e-02 4.64031875e-01
2.22462472e-02 -1.13502461e-02 5.36843419e-01 6.84425414e-01
-6.05023682e-01 -1.07919610e+00 -1.02345550e+00 2.75359839e-01
-7.70090759e-01 -4.41694111e-01 -4.59980080e-03 4.68542933e-01
3.74394983e-01 1.44947469e+00 -1.77234203e-01 -2.71944612e-01
3.14952314e-01 -2.09971800e-01 3.73912841e-01 -1.88788697e-02
-5.38626373e-01 3.64340305e-01 -1.74098805e-01 -5.21189868e-01
-5.17878294e-01 -2.79130757e-01 -7.85173953e-01 -4.74044174e-01
2.69222520e-02 -1.32085472e-01 8.20197821e-01 6.63482904e-01
1.32637843e-02 7.79731810e-01 4.57380652e-01 -1.16420186e+00
-5.55067003e-01 -1.04327953e+00 -7.78317988e-01 5.33669412e-01
8.19035172e-01 -7.58897141e-02 -8.40073287e-01 4.42866504e-01] | [11.346796989440918, -2.3420865535736084] |
f592131d-1ed1-4562-9404-2641457bc278 | co-evolving-real-time-strategy-game-micro | 1803.10314 | null | http://arxiv.org/abs/1803.10314v1 | http://arxiv.org/pdf/1803.10314v1.pdf | Co-evolving Real-Time Strategy Game Micro | We investigate competitive co-evolution of unit micromanagement in real-time
strategy games. Although good long-term macro-strategy and good short-term unit
micromanagement both impact real-time strategy games performance, this paper
focuses on generating quality micro. Better micro, for example, can help
players win skirmishes and battles even when outnumbered. Prior work has shown
that we can evolve micro to beat a given opponent. We remove the need for a
good opponent to evolve against by using competitive co-evolution to evolve
high-quality micro for both sides from scratch. We first co-evolve micro to
control a group of ranged units versus a group of melee units. We then move to
co-evolve micro for a group of ranged and melee units versus a group of ranged
and melee units. Results show that competitive co-evolution produces good
quality micro and when combined with the well-known techniques of fitness
sharing, shared sampling, and a hall of fame takes less time to produce better
quality micro than simple co-evolution. We believe these results indicate the
viability of co-evolutionary approaches for generating good unit
micro-management. | ['Siming Liu', 'Sushil J. Louis', 'Navin K Adhikari', 'Walker Spurgeon'] | 2018-03-27 | null | null | null | null | ['real-time-strategy-games'] | ['playing-games'] | [-3.87805365e-02 -5.88578805e-02 5.05832434e-01 2.91794538e-01
-5.60242534e-01 -6.00128293e-01 3.98049206e-01 -1.19030468e-01
-9.15682018e-01 1.14371717e+00 -1.97831944e-01 -5.39890155e-02
-4.09408391e-01 -1.14497650e+00 -4.97947246e-01 -6.29104078e-01
-2.23491743e-01 6.97506428e-01 2.84973800e-01 -9.09794688e-01
5.14593303e-01 1.01598263e-01 -1.80524170e+00 1.10654384e-02
9.95756745e-01 1.94523737e-01 2.08788469e-01 1.36514306e+00
3.62478137e-01 8.85133743e-01 -1.32552469e+00 -6.30314112e-01
5.84981918e-01 -1.17092919e+00 -3.32382143e-01 -3.32209855e-01
-1.89436421e-01 -1.21149234e-01 2.95069188e-01 8.35592747e-01
1.01934719e+00 2.82898277e-01 3.35018307e-01 -1.35537708e+00
-1.04187347e-01 1.14919615e+00 -6.17044508e-01 4.69065338e-01
3.09867203e-01 3.97132903e-01 1.12621808e+00 6.68886378e-02
8.24450970e-01 1.26563191e+00 7.01277196e-01 5.69885314e-01
-1.31951153e+00 -8.22276831e-01 -3.37267697e-01 -1.84997156e-01
-1.37829781e+00 -1.18706912e-01 2.99400419e-01 -3.23011493e-03
1.02226996e+00 7.73223519e-01 1.61862302e+00 5.86188853e-01
7.91881680e-01 3.72581631e-01 1.15372777e+00 -5.68725944e-01
5.03673911e-01 -2.31871426e-01 -8.04515481e-01 2.36387819e-01
5.96121252e-01 5.71481466e-01 -3.14992577e-01 -3.43502820e-01
9.04686928e-01 -3.65901798e-01 5.43688126e-02 -1.68899730e-01
-8.40009868e-01 1.02673471e+00 1.25009045e-01 4.37647760e-01
-6.47820711e-01 5.86919725e-01 8.93637761e-02 7.28824496e-01
-3.38931032e-03 1.71070135e+00 -2.77276635e-01 -1.07893050e+00
-9.34670806e-01 5.79465270e-01 9.54746187e-01 2.02066064e-01
5.75515687e-01 2.60971963e-01 8.56826901e-02 6.41227961e-01
-3.05004209e-01 1.67891517e-01 5.89910388e-01 -1.32428467e+00
-2.51915097e-01 6.14566803e-01 1.15153767e-01 -1.03282905e+00
-5.71839139e-02 -7.27329195e-01 -1.58605516e-01 8.79891694e-01
2.89469182e-01 -7.98179150e-01 -3.51413518e-01 1.91469407e+00
1.05841510e-01 1.99223593e-01 -1.22086354e-01 3.45201999e-01
-3.27870809e-02 4.81394768e-01 -5.07366657e-01 -5.13569951e-01
7.22802281e-01 -8.74540567e-01 -4.11492020e-01 -1.76296726e-01
6.10483527e-01 -7.19604135e-01 9.18890178e-01 6.03763044e-01
-1.71393645e+00 -2.07580090e-01 -9.75782752e-01 1.06622696e+00
8.58662650e-02 -7.74780512e-01 4.60796475e-01 1.19765151e+00
-1.26490676e+00 8.19564104e-01 -7.05959141e-01 -2.03229234e-01
2.07429141e-01 7.16032028e-01 1.24107718e-01 4.16209817e-01
-1.00135446e+00 1.11246121e+00 1.64758801e-01 -2.57576823e-01
-7.87250102e-01 -5.79607546e-01 -3.03387403e-01 4.08177227e-02
1.05206466e+00 -7.55802214e-01 1.25374997e+00 -1.38022685e+00
-1.62407374e+00 5.01133204e-01 6.56013846e-01 -3.95806640e-01
7.30504334e-01 2.06522495e-01 -4.25241217e-02 -3.76977533e-01
3.41328867e-02 5.40390491e-01 3.97836655e-01 -8.98228109e-01
-9.97955561e-01 6.46701679e-02 5.47875106e-01 5.50501049e-01
-2.50795126e-01 3.06247294e-01 2.65540451e-01 -4.96380448e-01
-5.21280468e-01 -1.04324293e+00 -6.77376628e-01 -9.30173337e-01
4.45194095e-02 2.31559530e-01 3.94984931e-02 -9.14970785e-02
1.68075597e+00 -1.68391526e+00 7.20322013e-01 2.56450385e-01
3.46364707e-01 1.36414155e-01 -2.37210095e-01 7.35892713e-01
7.96703845e-02 3.60346019e-01 3.72580647e-01 2.43051454e-01
6.15272447e-02 3.16286951e-01 6.06802881e-01 4.05793823e-02
-1.66330189e-01 1.06970477e+00 -9.20972168e-01 -1.33360520e-01
-1.47072122e-01 -7.92920887e-02 -1.19918156e+00 3.94929992e-03
-1.60979405e-01 -5.65679744e-04 -1.47087425e-01 2.50123292e-01
1.88223168e-01 1.04401529e-01 2.77053565e-01 5.81688106e-01
-4.36426193e-01 -1.65363550e-02 -1.40053737e+00 9.69016433e-01
-2.35273436e-01 2.35582590e-01 3.72024447e-01 -5.30897975e-01
8.46851289e-01 -1.41253889e-01 5.07140517e-01 -7.15076685e-01
7.56594479e-01 7.56727457e-01 1.16519105e+00 1.59271583e-01
8.96935344e-01 -4.92538393e-01 -3.90536696e-01 9.53678846e-01
-4.00005996e-01 -7.39727080e-01 6.59912467e-01 1.82393178e-01
1.41557503e+00 -1.87103271e-01 3.97566915e-01 -2.62236238e-01
1.01525597e-01 1.32955119e-01 7.87458062e-01 1.00437164e+00
-2.93839365e-01 2.67442048e-01 8.53384256e-01 -5.49011007e-02
-1.19023275e+00 -5.30039132e-01 6.49090588e-01 1.30074275e+00
1.71066627e-01 -6.54731870e-01 -7.77097523e-01 -1.70290723e-01
4.43320423e-02 8.51157308e-01 -9.65079367e-01 -5.73515475e-01
-7.81784117e-01 -8.93698514e-01 8.55139613e-01 1.56560212e-01
3.47063512e-01 -1.21585453e+00 -1.42144620e+00 6.35351777e-01
2.73210347e-01 2.09020525e-02 -6.79203689e-01 2.81464368e-01
-2.05097675e-01 -7.89625347e-01 -6.97620451e-01 -4.24566567e-01
8.34440663e-02 9.10605267e-02 1.22708917e+00 4.33983296e-01
-6.43371224e-01 2.59170860e-01 -3.29535931e-01 -6.24077260e-01
-7.36967683e-01 -1.02644391e-01 9.26178135e-03 -4.64382768e-01
-3.77323925e-02 -7.18726099e-01 -2.95646250e-01 6.52473092e-01
-6.70941949e-01 6.71072863e-03 4.13197398e-01 1.10345948e+00
2.03895435e-01 6.38689399e-01 4.84511852e-01 -5.40538907e-01
1.38204062e+00 -5.26701100e-02 -5.63233137e-01 2.87796378e-01
-4.62489039e-01 -2.19480485e-01 2.44113550e-01 -1.01672471e+00
-7.54946351e-01 -7.36390948e-01 2.38565728e-01 -1.92155451e-01
7.06494153e-01 6.19921148e-01 9.29711908e-02 -1.51443541e-01
8.71266365e-01 1.15122774e-03 3.73233646e-01 2.66798809e-02
1.95893943e-01 2.53601372e-01 1.04360156e-01 -6.51270688e-01
6.88881278e-01 -1.32371083e-01 -2.64351904e-01 -3.90022844e-01
3.55801493e-01 2.95821100e-01 -3.73499542e-02 -4.90972012e-01
6.82005167e-01 -4.74540174e-01 -8.65044892e-01 7.32050896e-01
-3.22299898e-01 -7.11206555e-01 -1.03673255e+00 1.88188702e-01
-6.65775180e-01 -7.23488331e-02 -2.59183466e-01 -9.26737368e-01
8.59448612e-02 -1.09616482e+00 3.44849616e-01 7.15721965e-01
-5.73171437e-01 -8.00425828e-01 6.14841104e-01 1.44916773e-01
7.60151923e-01 3.82328987e-01 3.05778086e-01 -5.84000349e-01
-3.68484885e-01 -1.94888949e-01 6.51118279e-01 3.59973796e-02
2.62187213e-01 4.41764027e-01 2.05580711e-01 -4.22497272e-01
-1.75046295e-01 -3.72040063e-01 1.79749623e-01 3.92662734e-01
1.87524304e-01 -1.79720163e-01 -1.78100318e-01 3.05087954e-01
1.11466002e+00 7.65571296e-01 7.16097057e-01 1.09709942e+00
5.12340739e-02 5.62504768e-01 8.56048346e-01 8.52326691e-01
2.10758194e-01 7.96199739e-01 2.87047058e-01 1.22797862e-01
1.89505666e-01 -5.04758842e-02 3.36272418e-01 7.39150286e-01
-5.27197778e-01 -3.73119414e-01 -7.79977083e-01 4.94708747e-01
-1.87053597e+00 -1.58939838e+00 1.91327095e-01 2.01837969e+00
1.05019009e+00 4.23350334e-01 8.97887111e-01 1.23463646e-01
8.53205860e-01 -3.10986727e-01 -4.63168532e-01 -9.90254998e-01
-3.76057178e-01 5.32667100e-01 5.26199579e-01 4.71197426e-01
-1.56671241e-01 8.98083270e-01 6.79907799e+00 1.16922581e+00
-9.07772779e-01 -4.73574139e-02 6.83299541e-01 -9.27338183e-01
-6.83262169e-01 2.39715025e-01 -3.52097332e-01 6.20251298e-01
8.52904558e-01 -1.22511077e+00 7.56979406e-01 4.42844123e-01
2.86896050e-01 -4.75924313e-01 -6.14532351e-01 4.47238386e-01
-3.92001905e-02 -1.49839842e+00 -5.74231267e-01 5.13472080e-01
1.15207362e+00 -5.69374263e-01 -7.75968954e-02 4.63169724e-01
1.29453766e+00 -1.30471969e+00 8.78147066e-01 1.93257749e-01
4.88387346e-01 -1.33443403e+00 7.74250567e-01 6.32444561e-01
-1.01512766e+00 -4.09929752e-01 -2.18943819e-01 -6.31994724e-01
2.85913289e-01 -1.39867365e-01 -6.62304819e-01 5.71369529e-01
5.43891728e-01 -1.05101466e-01 -4.87173796e-01 1.22424090e+00
1.71801135e-01 2.43446380e-01 -4.55822408e-01 -4.74200606e-01
3.83813322e-01 -3.91682893e-01 5.98397851e-01 4.32205647e-01
5.24088442e-01 7.72886351e-02 7.19031096e-02 6.39050901e-01
5.01333952e-01 -4.14871015e-02 -2.73509771e-01 -5.19663095e-01
7.27633595e-01 9.54707563e-01 -8.02787900e-01 -2.78859586e-01
4.40875053e-01 8.46156180e-01 -9.70434695e-02 -3.09188426e-01
-9.69684064e-01 -7.09776521e-01 7.96696007e-01 2.49462090e-02
3.78266156e-01 -5.70465326e-02 -3.15627724e-01 -8.15357566e-01
-7.77298272e-01 -1.69927514e+00 3.35241199e-01 -6.81379139e-01
-8.88897061e-01 7.46155620e-01 4.45284620e-02 -8.33577573e-01
-7.69826293e-01 5.63800037e-02 -1.03820133e+00 8.24440956e-01
-4.93856788e-01 -5.66995740e-01 3.58905420e-02 2.73770154e-01
4.79620636e-01 -6.34246349e-01 3.67035955e-01 -1.55416861e-01
-7.06729591e-01 7.64223635e-01 2.08102927e-01 -5.30661523e-01
3.98582727e-01 -1.11646259e+00 2.57852912e-01 8.32625329e-01
-1.38607591e-01 7.38788486e-01 9.44859862e-01 -1.00087821e+00
-1.22941470e+00 -1.27201542e-01 4.64794368e-01 -1.72147021e-01
6.83421195e-01 -1.14656441e-01 -2.28831828e-01 1.14510529e-01
4.17557120e-01 -6.55118644e-01 6.24691248e-01 6.09585680e-02
4.19212997e-01 -1.65758990e-02 -1.12973332e+00 1.23701680e+00
1.17558146e+00 5.24571119e-03 -3.72296244e-01 -2.35217586e-01
5.11095643e-01 -2.77677745e-01 -5.64253032e-01 5.63950688e-02
7.63777375e-01 -1.44606972e+00 8.90456796e-01 -4.53536719e-01
3.83237779e-01 -1.30629748e-01 -5.98134510e-02 -1.77331269e+00
-5.13294399e-01 -1.40563047e+00 6.19858742e-01 9.37397718e-01
2.35408649e-01 -9.12180722e-01 9.88491416e-01 4.29687381e-01
1.58406779e-01 -6.41568959e-01 -8.75826657e-01 -1.13794708e+00
4.52838957e-01 1.39695868e-01 7.22502172e-01 7.15791583e-01
9.53656510e-02 1.39033869e-01 -6.25028193e-01 -6.20561957e-01
3.95581782e-01 -5.68404719e-02 1.17407537e+00 -9.13512886e-01
-1.29412842e+00 -1.08898795e+00 -4.55617070e-01 -2.92178184e-01
-5.33498883e-01 -2.65456825e-01 -7.36815855e-02 -1.11178315e+00
1.54661477e-01 -6.89456582e-01 -1.00870349e-01 1.81898877e-01
-3.07182133e-01 3.77087146e-01 6.20889783e-01 1.17647439e-01
-4.72709537e-01 1.82104453e-01 1.51709628e+00 2.88802564e-01
-6.28633142e-01 -3.29736955e-02 -1.39127994e+00 3.67382079e-01
6.50463343e-01 -4.79364663e-01 -6.09912336e-01 9.06725302e-02
7.23744094e-01 4.04216290e-01 -2.54882962e-01 -1.01554847e+00
7.50009418e-02 -6.48682654e-01 -1.90389872e-01 1.39003098e-01
2.69393742e-01 -2.26424426e-01 9.69990790e-01 1.13491189e+00
-1.68633029e-01 7.13382661e-01 2.11425781e-01 3.24748635e-01
6.52726665e-02 -4.42244977e-01 6.29410744e-01 -4.80348349e-01
-2.26499245e-01 -2.44353205e-01 -9.28556025e-01 1.97178200e-01
1.58590961e+00 -1.00789535e+00 -2.71746546e-01 -7.63756216e-01
-5.05370796e-01 3.35650951e-01 1.10252106e+00 3.87542211e-02
1.90833181e-01 -9.52305675e-01 -9.99168217e-01 6.89567924e-02
-4.83566344e-01 -6.99145138e-01 3.38236988e-02 7.96833932e-01
-8.39654744e-01 -3.01659614e-01 -1.03414524e+00 -6.43451363e-02
-1.75388205e+00 -8.75826087e-03 6.29824877e-01 -5.23780644e-01
-1.57756582e-01 1.43835270e+00 -8.01846981e-02 -2.37860888e-01
-1.01770230e-01 2.76830822e-01 -8.44918713e-02 3.62907261e-01
4.08739358e-01 5.32210767e-01 -3.73206735e-01 -3.48425776e-01
-4.39596981e-01 2.78262675e-01 6.40711561e-02 -8.66186976e-01
1.61110342e+00 6.82576597e-02 -2.42668033e-01 3.52825373e-01
4.44560707e-01 6.31464541e-01 -1.01206958e+00 5.93014359e-01
-3.11142534e-01 -1.04320896e+00 -2.88479865e-01 -9.62147772e-01
-7.98410177e-01 1.58103913e-01 1.08687319e-01 5.35844326e-01
1.19450688e+00 -4.65309560e-01 3.55356723e-01 1.36005759e-01
8.26797545e-01 -1.14876008e+00 6.94603384e-01 4.79701310e-01
6.09290779e-01 -2.47277573e-01 -3.38265859e-02 7.77545720e-02
-9.52688456e-01 6.26333773e-01 9.09883797e-01 -5.02306819e-01
3.81488982e-03 6.08007371e-01 -2.40956247e-02 -4.33632825e-03
-1.35007453e+00 -2.35275269e-01 -4.76510674e-01 6.41544640e-01
2.69508749e-01 1.62707195e-01 -8.42584431e-01 3.42882901e-01
-8.35684896e-01 -1.28099903e-01 1.17186749e+00 1.49863815e+00
-7.19617367e-01 -1.91044950e+00 -6.20113730e-01 5.77646434e-01
-3.74113768e-01 -1.78431738e-02 -6.99482739e-01 1.00232255e+00
5.41355252e-01 9.49358404e-01 1.89485759e-01 -6.47118628e-01
1.16779089e-01 -4.27843064e-01 8.03511620e-01 -6.78373814e-01
-1.46951640e+00 4.45177108e-01 5.50436139e-01 -3.09474915e-01
-7.54467249e-02 -9.93538260e-01 -7.56657779e-01 -1.19187438e+00
-7.88792074e-01 5.01699686e-01 1.15143768e-01 3.74554783e-01
2.11809307e-01 8.38209629e-01 5.77684641e-01 -7.67606080e-01
-7.34608710e-01 -6.68901026e-01 -8.27170849e-01 1.02622308e-01
-3.71131748e-01 -8.07411671e-01 -2.12387249e-01 -6.82233334e-01] | [3.488337516784668, 1.6006461381912231] |
7d8e5cd7-5712-4663-87cd-4b00a9ebaa39 | illumination-controllable-dehazing-network | 2306.05675 | null | https://arxiv.org/abs/2306.05675v1 | https://arxiv.org/pdf/2306.05675v1.pdf | Illumination Controllable Dehazing Network based on Unsupervised Retinex Embedding | On the one hand, the dehazing task is an illposedness problem, which means that no unique solution exists. On the other hand, the dehazing task should take into account the subjective factor, which is to give the user selectable dehazed images rather than a single result. Therefore, this paper proposes a multi-output dehazing network by introducing illumination controllable ability, called IC-Dehazing. The proposed IC-Dehazing can change the illumination intensity by adjusting the factor of the illumination controllable module, which is realized based on the interpretable Retinex theory. Moreover, the backbone dehazing network of IC-Dehazing consists of a Transformer with double decoders for high-quality image restoration. Further, the prior-based loss function and unsupervised training strategy enable IC-Dehazing to complete the parameter learning process without the need for paired data. To demonstrate the effectiveness of the proposed IC-Dehazing, quantitative and qualitative experiments are conducted on image dehazing, semantic segmentation, and object detection tasks. Code is available at https://github.com/Xiaofeng-life/ICDehazing. | ['James Tin-Yau Kwok', 'Yuan Yan Tang', 'Lei He', 'Xiaofeng Cong', 'Jie Gui'] | 2023-06-09 | null | null | null | null | ['image-dehazing', 'image-restoration'] | ['computer-vision', 'computer-vision'] | [ 4.03053820e-01 -3.00115012e-02 1.30845204e-01 -3.60155195e-01
-2.86224693e-01 -1.35869309e-02 8.45320523e-02 -5.10990798e-01
-4.19484615e-01 3.21664631e-01 -5.08302869e-03 -1.99171558e-01
-2.48977408e-01 -9.33365464e-01 -8.06404352e-01 -1.23153746e+00
7.51769006e-01 -2.91975915e-01 2.64457792e-01 -1.72255784e-01
1.95691273e-01 1.01234518e-01 -1.42669034e+00 1.08968660e-01
1.23586166e+00 1.04675210e+00 5.30559063e-01 4.23244655e-01
2.38915935e-01 5.09494483e-01 -5.94354570e-01 -1.35312289e-01
2.46106535e-01 -6.61719799e-01 -1.79152653e-01 4.31432366e-01
1.69273674e-01 -7.53736794e-01 -2.40685537e-01 1.25383115e+00
5.65552413e-01 8.84008184e-02 6.02180243e-01 -1.25700724e+00
-1.02695608e+00 1.49198428e-01 -4.67695445e-01 -1.67380720e-02
-1.39196843e-01 3.30757737e-01 7.24717498e-01 -6.89423978e-01
8.79235119e-02 1.12114072e+00 2.55620122e-01 4.86301243e-01
-9.31147516e-01 -8.07543159e-01 5.84325334e-03 3.31193864e-01
-1.47794890e+00 -3.19473505e-01 9.40882742e-01 -3.25036824e-01
3.08456719e-01 5.96856624e-02 4.80482131e-01 5.75049520e-01
2.43650511e-01 8.02880883e-01 1.05653107e+00 -3.55026126e-01
4.72342744e-02 2.87358701e-01 -7.31495023e-02 7.40257680e-01
3.06303054e-01 3.02060783e-01 -8.60829130e-02 5.55424333e-01
8.81057084e-01 3.29524368e-01 -5.71924686e-01 9.13216770e-02
-1.00873220e+00 5.54225445e-01 5.37161350e-01 1.38639241e-01
-7.34527707e-02 -4.44960035e-02 1.09926805e-01 2.25382715e-01
4.70979154e-01 2.69518435e-01 -1.25685737e-01 4.42262560e-01
-8.08731675e-01 -1.40779287e-01 2.69615769e-01 8.31250846e-01
9.22649443e-01 3.26657921e-01 -1.07813857e-01 7.55889416e-01
5.18654585e-01 4.84466374e-01 3.80438268e-01 -8.90220940e-01
3.05357873e-01 5.01743197e-01 8.90981406e-02 -1.07903588e+00
-1.50277495e-01 -3.00282121e-01 -9.33306694e-01 4.52155948e-01
1.55561611e-01 -3.27865720e-01 -1.18639612e+00 1.48909378e+00
3.43590587e-01 2.09865361e-01 -9.28373858e-02 1.25741518e+00
7.40777731e-01 1.08638668e+00 -1.04342245e-01 -2.82842100e-01
1.18458581e+00 -9.60159421e-01 -1.09354210e+00 -9.59756300e-02
2.28008300e-01 -6.85373008e-01 1.32754862e+00 6.26517296e-01
-8.61108959e-01 -5.39286911e-01 -1.45253944e+00 -2.73237795e-01
-3.61396641e-01 3.07565391e-01 2.41834417e-01 6.62435532e-01
-7.86944866e-01 1.13288522e-01 -5.75196683e-01 1.01213887e-01
4.62859780e-01 2.98672378e-01 2.53514480e-02 -1.70677438e-01
-1.30812025e+00 7.78311551e-01 7.94796944e-01 5.21126091e-01
-9.33728397e-01 -5.62730134e-01 -7.12709427e-01 1.07381403e-01
5.40723622e-01 -4.06956077e-01 8.40372920e-01 -1.25879896e+00
-1.87274730e+00 5.86500108e-01 2.09931165e-01 -3.47656310e-02
5.15760183e-01 -1.01556227e-01 -5.43512225e-01 3.09311360e-01
-2.02646017e-01 4.57193047e-01 1.12814486e+00 -1.31064618e+00
-5.89011669e-01 -2.47785851e-01 -6.95501715e-02 3.91186029e-01
-3.88404280e-01 -1.29673779e-01 -5.03087819e-01 -8.33366454e-01
9.99353733e-03 -4.46919173e-01 1.50500253e-01 2.54125118e-01
-4.14169371e-01 7.21255094e-02 1.06147420e+00 -1.04365349e+00
1.28539693e+00 -2.57881427e+00 1.36257887e-01 1.47644550e-01
1.59025982e-01 3.74117017e-01 3.79860364e-02 -1.45222032e-02
-9.89745557e-02 1.07898481e-01 -6.01347089e-01 6.04556762e-02
-8.05915520e-02 5.47740906e-02 -1.17934518e-01 5.60018182e-01
2.28292346e-01 4.56548780e-01 -6.05443299e-01 -6.31841063e-01
6.11974418e-01 5.44262648e-01 -5.01232862e-01 2.93751001e-01
-3.00368577e-01 4.84532535e-01 -4.52195853e-01 4.85991418e-01
9.18225169e-01 -1.22841140e-02 -4.02818978e-01 -4.49088573e-01
-2.34847054e-01 -2.48007163e-01 -1.21771336e+00 1.29648185e+00
-4.92024332e-01 3.84217441e-01 2.89708048e-01 -8.82954299e-01
9.09881771e-01 3.98987830e-01 1.19052291e-01 -8.50127220e-01
4.36025858e-01 3.35356653e-01 -2.49068752e-01 -5.92646301e-01
2.31796458e-01 -2.94651121e-01 2.94629961e-01 2.61165828e-01
-1.01102777e-01 -2.78256506e-01 -6.19596988e-02 -9.08876956e-02
3.30202103e-01 1.49819925e-01 5.03418446e-02 -1.97244316e-01
6.92361295e-01 -3.43268454e-01 7.22991049e-01 2.42892206e-01
-3.77574638e-02 7.65642881e-01 2.24661663e-01 -1.41739426e-02
-8.68059278e-01 -9.11991417e-01 -3.27951968e-01 7.98161685e-01
7.58133829e-01 1.07955158e-01 -1.02521682e+00 -3.15487444e-01
-4.50561941e-01 9.29650784e-01 -3.32875431e-01 -5.27629852e-01
-2.99082011e-01 -6.62633479e-01 1.73404083e-01 4.78946313e-04
1.23944283e+00 -9.00535762e-01 -3.38583291e-01 7.32758790e-02
-3.75428379e-01 -7.34375000e-01 -8.60556602e-01 -2.03319564e-01
-5.37687302e-01 -8.87986302e-01 -8.19258213e-01 -1.02465737e+00
8.40047121e-01 4.88797069e-01 3.58788908e-01 3.11470091e-01
-1.65845290e-01 5.03334105e-02 -3.55014324e-01 -5.35768390e-01
-2.55172163e-01 -1.06778078e-01 -4.06413913e-01 3.57260108e-01
2.14318812e-01 -5.83927155e-01 -9.31828916e-01 6.90742910e-01
-1.40598691e+00 3.66628885e-01 6.04965150e-01 8.10216904e-01
7.43530393e-01 6.57408178e-01 5.70488155e-01 -6.04146600e-01
4.20031875e-01 -2.16615692e-01 -8.20267439e-01 3.24449360e-01
-7.26851106e-01 -1.41307428e-01 6.48148000e-01 -4.40737128e-01
-1.42697084e+00 -1.57923147e-01 -1.52306870e-01 -6.18786335e-01
6.37012050e-02 1.40683338e-01 -7.77216971e-01 -2.64871959e-02
4.10925657e-01 4.29265141e-01 1.08103171e-01 -3.01605374e-01
2.87899256e-01 1.00907683e+00 5.25720537e-01 -2.38044903e-01
7.75880337e-01 4.55561012e-01 -2.38998979e-01 -8.66452754e-01
-6.65478706e-01 -1.17611088e-01 -9.09000859e-02 -3.88575375e-01
1.31594992e+00 -9.06254470e-01 -7.23999739e-01 9.43053126e-01
-9.39399838e-01 -4.07763094e-01 2.59228479e-02 5.61740458e-01
-3.95938665e-01 2.78936207e-01 -5.66465259e-01 -6.28250182e-01
-3.06362927e-01 -1.16254950e+00 7.24655330e-01 5.68153083e-01
5.26854992e-01 -7.92016923e-01 -5.17420530e-01 5.42733610e-01
1.61213696e-01 1.45141140e-01 8.02476585e-01 -1.17942400e-01
-9.30715978e-01 2.92984284e-02 -3.23061496e-01 9.24047053e-01
2.29758710e-01 7.43717551e-02 -9.37212467e-01 -1.46860078e-01
4.14031088e-01 -2.76166364e-03 9.27213907e-01 5.92490435e-01
1.53469753e+00 -4.51381773e-01 1.09992251e-01 9.58811164e-01
1.47813392e+00 5.39253473e-01 1.02966714e+00 4.93964493e-01
8.47743094e-01 5.13805330e-01 6.48104429e-01 2.17245623e-01
2.42358729e-01 3.91973168e-01 6.91130102e-01 -3.56753349e-01
-2.39810407e-01 -2.85964727e-01 1.65849358e-01 6.73242986e-01
-4.20940630e-02 -6.54583931e-01 -5.14061391e-01 3.74574840e-01
-1.38587117e+00 -7.83995211e-01 1.84592493e-02 2.07835674e+00
8.67527246e-01 -5.89321032e-02 -2.83521146e-01 2.23713607e-01
1.16939378e+00 2.06810251e-01 -6.46879613e-01 -2.73327470e-01
1.83574501e-02 -2.15036169e-01 6.41861737e-01 5.45623183e-01
-1.12896121e+00 7.02612996e-01 4.55232620e+00 1.08522058e+00
-1.18232977e+00 1.54233024e-01 7.01411664e-01 1.27048999e-01
-3.99604112e-01 -1.92583799e-01 -5.70869565e-01 9.46837306e-01
2.95240819e-01 -6.08069636e-02 6.40681088e-01 5.26459873e-01
4.09153163e-01 -2.42806330e-01 -5.69521844e-01 9.65285897e-01
1.51019812e-01 -8.56521010e-01 1.96651113e-03 -1.34987131e-01
6.76681042e-01 -5.27942538e-01 3.29443723e-01 -1.67420194e-01
-2.99758948e-02 -8.24652076e-01 7.67576456e-01 5.59194326e-01
8.08644533e-01 -7.68114328e-01 5.64826429e-01 2.40591854e-01
-7.93121696e-01 -6.75250888e-02 -3.35869879e-01 3.53056759e-01
4.07957807e-02 6.30030990e-01 -3.17411304e-01 5.76644480e-01
7.29877710e-01 6.83113694e-01 -2.73482770e-01 1.05274999e+00
-7.78757334e-01 7.45184779e-01 -1.06029801e-01 2.62435049e-01
9.31832567e-02 -6.17163718e-01 4.05931264e-01 9.67845559e-01
4.40548152e-01 4.64664429e-01 -1.29581168e-01 8.58530104e-01
-3.72724426e-05 -1.34922683e-01 -1.71045750e-01 9.92128327e-02
3.88706237e-01 1.12932491e+00 -4.79290277e-01 -5.97622655e-02
-2.19927430e-01 9.62242901e-01 -3.75448763e-01 6.41891539e-01
-1.02754796e+00 -9.27274466e-01 2.45677233e-01 1.12192109e-01
3.78171355e-01 -3.28600146e-02 -1.69742540e-01 -1.00376809e+00
4.03833669e-03 -8.14304531e-01 1.65323228e-01 -1.12854862e+00
-1.18456697e+00 2.71999389e-01 -5.23055941e-02 -1.37645030e+00
4.43506151e-01 -6.35721684e-01 -7.09639013e-01 9.40817654e-01
-1.65660000e+00 -1.03621328e+00 -6.18338704e-01 7.17860222e-01
4.25498813e-01 7.27614909e-02 1.78635180e-01 7.28461385e-01
-8.82058144e-01 5.45303047e-01 2.60587931e-01 1.43292084e-01
8.42358112e-01 -9.01776791e-01 -4.69445765e-01 1.20419931e+00
-6.14203632e-01 2.40107134e-01 7.25130260e-01 -4.18739527e-01
-9.64681804e-01 -1.23373675e+00 2.27521211e-01 2.03212380e-01
3.09109718e-01 -1.87914416e-01 -1.08384025e+00 5.73416948e-01
2.68096030e-01 -1.50608331e-01 3.17180514e-01 -6.22086585e-01
-1.13570437e-01 -5.42603195e-01 -1.12519825e+00 7.00447023e-01
7.25061595e-01 -3.64451975e-01 -4.03626382e-01 3.56577635e-01
1.05171943e+00 -4.21590030e-01 -6.88274443e-01 3.20183396e-01
3.28503281e-01 -9.24689710e-01 8.72665584e-01 2.09141836e-01
7.11259663e-01 -7.40384161e-01 -1.20234648e-02 -1.32165360e+00
-7.87692964e-02 -4.80795681e-01 3.03162098e-01 1.35730314e+00
1.89820975e-01 -9.29932952e-01 3.91396672e-01 3.92175376e-01
-3.66369545e-01 -6.33590102e-01 -5.12825906e-01 -7.34932244e-01
-1.01772405e-01 -6.06639273e-02 7.93724179e-01 7.28916645e-01
-5.37645400e-01 1.17170230e-01 -6.25351548e-01 5.41921616e-01
6.15779340e-01 -1.84160337e-01 3.95114869e-01 -6.10593736e-01
-3.21889162e-01 -1.91457525e-01 -1.94351315e-01 -1.05033970e+00
-1.55481234e-01 -5.87756395e-01 3.90321523e-01 -1.68843353e+00
-1.74467251e-01 -2.72362560e-01 -2.42105260e-01 3.90230507e-01
-3.30155343e-01 8.40910301e-02 8.08203295e-02 3.54635715e-01
-2.97616482e-01 8.37481558e-01 1.87193978e+00 -3.28213990e-01
-2.55272299e-01 2.57776789e-02 -7.49841690e-01 5.14879048e-01
1.08638966e+00 -3.38017315e-01 -7.13272870e-01 -8.00947905e-01
9.32630375e-02 -5.37299961e-02 5.25685072e-01 -7.77413607e-01
2.54842579e-01 -2.41196126e-01 2.35451519e-01 -2.78968662e-01
2.34590977e-01 -9.77156699e-01 2.58901604e-02 2.62586683e-01
-1.56635508e-01 -4.32656825e-01 -9.26054344e-02 6.68491721e-01
-3.13584894e-01 -3.44490260e-01 9.16110456e-01 3.70044708e-02
-7.72117376e-01 3.97336096e-01 -3.30242306e-01 -3.63593213e-02
1.09399140e+00 -5.43795824e-01 -4.54166085e-01 -3.61432552e-01
-2.59790331e-01 3.29064071e-01 3.26323509e-01 5.06937392e-02
8.21083367e-01 -1.10906279e+00 -5.87991595e-01 3.51671666e-01
-1.31510869e-01 4.71403450e-01 6.77573264e-01 6.93068266e-01
-7.49740541e-01 -1.33234769e-01 -1.30176604e-01 -3.59144926e-01
-1.10487807e+00 6.13449037e-01 6.20476067e-01 2.99909413e-01
-5.72598100e-01 6.52684748e-01 5.77215374e-01 -2.69302994e-01
1.55448407e-01 -5.32474555e-03 -2.59553999e-01 -6.94905296e-02
5.00574887e-01 4.91988122e-01 -7.10894838e-02 -2.85363555e-01
1.52222291e-02 4.79351640e-01 -6.53274134e-02 -1.13786362e-01
1.12888587e+00 -3.91792506e-01 -3.50905538e-01 1.74280599e-01
1.18838823e+00 -3.11734140e-01 -1.64080536e+00 -5.72245456e-02
-5.77203155e-01 -6.04226410e-01 4.78044450e-01 -8.08514118e-01
-1.25219154e+00 1.02924752e+00 7.71415412e-01 6.90685445e-03
1.72708368e+00 -3.67060244e-01 6.95314765e-01 2.05495477e-01
-1.82916671e-02 -1.16698635e+00 1.71120733e-01 7.95948133e-02
8.05941045e-01 -1.03267181e+00 -9.11283307e-03 -5.78854263e-01
-6.91032767e-01 8.51725817e-01 7.55095363e-01 -8.79057273e-02
6.74952149e-01 -5.15703596e-02 1.72113225e-01 -4.34160838e-03
-1.68150857e-01 4.06754762e-02 1.62515253e-01 4.15972441e-01
-1.01120152e-01 -6.40724301e-02 -4.59001780e-01 5.70614874e-01
4.64344434e-02 -5.48805157e-03 4.34714556e-01 5.48300743e-01
-5.21878421e-01 -6.09489322e-01 -3.96238267e-01 2.49896735e-01
-3.13472658e-01 -6.26160399e-05 8.01550969e-02 6.94866240e-01
5.24700403e-01 1.20720124e+00 -2.14049779e-02 -3.34996104e-01
2.14357197e-01 -3.48166227e-01 2.89029896e-01 -4.23817247e-01
-2.40585119e-01 1.82774946e-01 -3.87429506e-01 -1.97778910e-01
-4.15278405e-01 -1.79010883e-01 -1.24025738e+00 -2.55489767e-01
-6.91123486e-01 3.80443372e-02 5.99873066e-01 9.22068477e-01
2.04013586e-01 8.73576105e-01 8.03180218e-01 -4.71630812e-01
-2.69226074e-01 -7.91612923e-01 -8.63440037e-01 1.61815360e-01
3.46096307e-01 -5.07604420e-01 -6.63284004e-01 2.82247424e-01] | [10.909843444824219, -3.059062957763672] |
3b2b9b21-e3dc-4c46-b127-bb90e4237d96 | diffusion-models-in-bioinformatics-a-new-wave | 2302.10907 | null | https://arxiv.org/abs/2302.10907v1 | https://arxiv.org/pdf/2302.10907v1.pdf | Diffusion Models in Bioinformatics: A New Wave of Deep Learning Revolution in Action | Denoising diffusion models have emerged as one of the most powerful generative models in recent years. They have achieved remarkable success in many fields, such as computer vision, natural language processing (NLP), and bioinformatics. Although there are a few excellent reviews on diffusion models and their applications in computer vision and NLP, there is a lack of an overview of their applications in bioinformatics. This review aims to provide a rather thorough overview of the applications of diffusion models in bioinformatics to aid their further development in bioinformatics and computational biology. We start with an introduction of the key concepts and theoretical foundations of three cornerstone diffusion modeling frameworks (denoising diffusion probabilistic models, noise-conditioned scoring networks, and stochastic differential equations), followed by a comprehensive description of diffusion models employed in the different domains of bioinformatics, including cryo-EM data enhancement, single-cell data analysis, protein design and generation, drug and small molecule design, and protein-ligand interaction. The review is concluded with a summary of the potential new development and applications of diffusion models in bioinformatics. | ['Jianlin Cheng', 'Dong Xu', 'Duolin Wang', 'Mengrui Chen', 'Yanli Wang', 'Jian Liu', 'Zhiye Guo'] | 2023-02-13 | null | null | null | null | ['protein-design'] | ['medical'] | [ 3.33249509e-01 -4.17370796e-01 9.87721309e-02 -1.70497671e-01
-4.79615748e-01 -4.55583960e-01 4.19334441e-01 4.43774939e-01
-7.87940741e-01 7.71484077e-01 2.62108713e-01 -2.26129651e-01
-3.29840034e-01 -5.35431683e-01 -2.69222140e-01 -1.55087650e+00
-8.66530538e-02 8.84805202e-01 9.40082893e-02 -6.93424046e-02
2.49964118e-01 7.24557996e-01 -1.05466044e+00 2.74725437e-01
1.02433157e+00 4.14541870e-01 4.40825939e-01 7.90659547e-01
-2.82479405e-01 5.34585059e-01 -5.08390665e-01 -3.54809314e-01
-6.19725764e-01 -6.65116668e-01 -5.09288669e-01 -9.89532620e-02
-3.42704445e-01 1.17863715e-01 -4.32679325e-01 9.30521667e-01
1.09149992e+00 -4.26724069e-02 1.08681059e+00 -7.03849256e-01
-1.18946958e+00 1.28216714e-01 -6.71852827e-01 5.89766562e-01
2.38012657e-01 2.22866282e-01 4.07041788e-01 -1.09164655e+00
1.16930211e+00 1.32001376e+00 5.79560339e-01 8.38212371e-01
-1.39159226e+00 -5.00308946e-02 -3.82394753e-02 3.22718382e-01
-1.29951251e+00 -2.73910642e-01 6.55832469e-01 -6.45343482e-01
1.19894290e+00 4.10056338e-02 6.29170895e-01 1.12458122e+00
8.47233355e-01 9.79708493e-01 1.32665634e+00 -3.07107776e-01
5.59921384e-01 -1.18276320e-01 2.26953715e-01 4.89495665e-01
-9.50097945e-03 -3.85262519e-02 -7.83199430e-01 -7.20735669e-01
5.89562058e-01 5.01135029e-02 -9.45910141e-02 -6.53160810e-01
-1.20662081e+00 1.16484725e+00 -1.15042746e-01 5.35483539e-01
-8.56151879e-01 -1.80785373e-01 3.33889693e-01 -4.20506448e-02
7.99777389e-01 -2.84355968e-01 -2.31501639e-01 -8.84399563e-02
-1.12362540e+00 4.25280362e-01 5.94876885e-01 2.97075242e-01
1.73784509e-01 1.74248457e-01 6.08540438e-02 1.08017540e+00
6.00253880e-01 8.71838748e-01 4.80769247e-01 -7.78440177e-01
-1.98384389e-01 -1.91373646e-01 1.21064901e-01 -8.71540546e-01
-5.94163060e-01 1.74645265e-03 -1.31105125e+00 8.12477767e-02
3.70187819e-01 -2.48674929e-01 -8.09755743e-01 1.41700041e+00
6.02300167e-01 -8.20216611e-02 1.04156427e-01 8.95046711e-01
1.08636999e+00 6.03784442e-01 4.77825910e-01 -6.62731707e-01
1.24480557e+00 -4.95425165e-01 -9.59731996e-01 1.65504530e-01
4.42184836e-01 -1.00476468e+00 1.24410145e-01 5.27937114e-01
-1.16691339e+00 -1.62105754e-01 -3.82142901e-01 -2.28402242e-01
-4.34373051e-01 -1.62136585e-01 8.50276530e-01 6.64061427e-01
-9.57095265e-01 6.79939032e-01 -1.41227686e+00 -8.92413139e-01
5.68090677e-01 3.55767220e-01 -1.63738489e-01 -2.29003072e-01
-1.18446326e+00 1.00690174e+00 -4.10860717e-01 1.39213130e-01
-9.08735931e-01 -5.56511939e-01 -5.74989557e-01 -5.04384339e-01
-3.02088678e-01 -1.11989975e+00 6.77127182e-01 -3.60966772e-01
-1.43106294e+00 1.18755114e+00 -6.64182842e-01 -5.80766618e-01
3.22436094e-01 1.05337024e-01 -7.10063204e-02 1.64289623e-01
-1.65900439e-01 6.85497642e-01 2.67186433e-01 -7.13213325e-01
-1.11016981e-01 -7.28061497e-01 -7.36358881e-01 1.81494862e-01
1.44571466e-02 3.67337614e-01 -1.19021721e-01 -7.21925378e-01
-8.75232518e-02 -7.51069307e-01 -7.34694898e-01 2.10320488e-01
-1.19174391e-01 -3.10552776e-01 4.44040000e-01 -5.10571718e-01
9.32241023e-01 -1.72040582e+00 6.68917358e-01 4.62373160e-02
4.91294861e-01 3.09613675e-01 -2.03341290e-01 9.29250479e-01
-4.13524769e-02 -1.43459469e-01 -5.00014246e-01 -1.77724093e-01
-4.92015779e-01 1.66092560e-01 -2.15330511e-01 1.00708723e+00
4.74966951e-02 1.18169475e+00 -1.09934425e+00 -1.94416270e-01
2.61218965e-01 1.19881952e+00 -1.74607888e-01 -3.11169118e-01
-2.39204019e-02 8.26454997e-01 -6.07603252e-01 6.20588064e-01
8.11826110e-01 -3.56701344e-01 3.72398764e-01 3.97869423e-02
-1.31988674e-01 -1.10167161e-01 -9.28839862e-01 1.73023260e+00
3.17198247e-01 4.61621821e-01 2.73629338e-01 -9.67491925e-01
8.87125969e-01 1.89230591e-01 1.03577352e+00 -2.82239050e-01
-2.29871705e-01 8.22229683e-02 1.00348420e-01 -4.00632501e-01
1.19683193e-02 -5.14774859e-01 5.00250638e-01 5.48406661e-01
2.20325068e-02 -1.28471613e-01 2.72805512e-01 4.59485978e-01
9.25269306e-01 7.49418736e-02 2.42291763e-01 -3.09524655e-01
6.82992280e-01 1.93310380e-01 3.49996239e-01 5.94905198e-01
-5.13164997e-01 3.73941183e-01 3.82651687e-01 -5.56492746e-01
-9.84623492e-01 -7.96224773e-01 -2.90124118e-01 8.14646602e-01
-3.48398462e-02 -5.04393756e-01 -9.05877411e-01 -1.47507876e-01
1.93909965e-02 2.14195952e-01 -7.25506485e-01 2.55089384e-02
-3.00331146e-01 -1.94407618e+00 6.29647553e-01 2.11170688e-01
5.41785359e-03 -9.63642955e-01 2.47346926e-02 4.05221820e-01
-9.48830321e-02 -6.45648599e-01 -1.23510920e-01 2.42547870e-01
-1.10032177e+00 -1.08061004e+00 -1.46844590e+00 -6.95547581e-01
5.51683664e-01 3.56828511e-01 6.50538266e-01 -2.27583215e-01
-6.45739496e-01 5.93911290e-01 1.38994277e-01 -5.17993689e-01
-6.05565250e-01 -2.82421827e-01 4.80129421e-01 -4.07879323e-01
9.14305925e-01 -4.20891732e-01 -1.00015879e+00 1.11763477e-01
-1.16361010e+00 -4.64909047e-01 4.23630744e-01 1.06072044e+00
1.03128457e+00 -2.09286124e-01 4.74703670e-01 -9.83391583e-01
1.11971939e+00 -4.28343475e-01 -2.92776793e-01 1.92081615e-01
-8.59633327e-01 -1.37429908e-01 1.24547951e-01 -4.85154986e-01
-1.11673093e+00 2.16783519e-04 -7.03993082e-01 9.30099785e-02
-2.47382388e-01 7.53160775e-01 2.20845193e-01 -2.87089944e-01
8.43667090e-01 6.50162995e-01 5.42838633e-01 -5.45684457e-01
5.10752141e-01 5.22211492e-01 -8.89639631e-02 -3.76814842e-01
1.18216023e-01 9.99808490e-01 2.16557100e-01 -1.44315374e+00
-3.62100989e-01 -5.18131375e-01 -6.97901964e-01 -4.08004597e-02
9.76587892e-01 -4.86675888e-01 -6.76200092e-01 9.93443668e-01
-1.13312101e+00 -1.37631848e-01 -1.13537714e-01 4.85370189e-01
-7.59370327e-01 1.12550831e+00 -1.20525146e+00 -8.30855787e-01
-6.68274939e-01 -1.58405995e+00 8.97983730e-01 1.95127234e-01
-1.70535550e-01 -1.47532237e+00 6.86852992e-01 4.31111664e-01
5.74203789e-01 1.21298254e-01 1.07578993e+00 -4.85604733e-01
-1.68118551e-01 -7.62460008e-02 2.63524860e-01 1.04551926e-01
8.53049532e-02 3.11373144e-01 -8.46243382e-01 -8.12325552e-02
2.63361543e-01 -2.39568919e-01 1.07579184e+00 1.13696289e+00
5.09180009e-01 3.52599382e-01 -9.19747829e-01 4.42932576e-01
1.11110306e+00 2.40884081e-01 6.67750478e-01 -2.70695761e-02
2.44211301e-01 6.00125909e-01 3.48002970e-01 2.05912471e-01
1.18431345e-01 3.22619468e-01 -6.80721328e-02 -2.89408296e-01
-2.07101792e-01 1.84711933e-01 3.27367544e-01 9.90323126e-01
-4.01668817e-01 -2.44433820e-01 -7.87567198e-01 1.57605827e-01
-1.73895204e+00 -1.11631668e+00 -6.74090862e-01 2.05607009e+00
7.51410067e-01 -3.90424162e-01 2.69583881e-01 -2.71176577e-01
6.72611117e-01 -4.10300530e-02 -8.28305483e-01 -2.02776536e-01
-8.39261413e-01 4.24432419e-02 3.13474268e-01 7.23293424e-01
-8.27565372e-01 1.06362963e+00 8.23929214e+00 9.06728625e-01
-1.06284988e+00 2.92366177e-01 8.75304163e-01 3.02373338e-02
-2.45159060e-01 -2.06107184e-01 -8.37307274e-01 3.66518199e-01
9.67295110e-01 9.43553150e-02 2.47541577e-01 4.48406816e-01
7.69491076e-01 -4.40131664e-01 -4.12409365e-01 9.63645697e-01
-1.47809004e-02 -1.71106803e+00 -4.08262946e-02 2.56857753e-01
7.17006147e-01 5.66306293e-01 1.68893337e-01 -3.78223300e-01
2.58175969e-01 -7.47382939e-01 8.49768799e-03 8.78754199e-01
5.54421961e-01 -6.10724390e-01 6.73185349e-01 5.34214258e-01
-4.63545144e-01 5.48637152e-01 -8.45875263e-01 2.43114337e-01
6.29445195e-01 1.17379844e+00 -2.60407418e-01 2.49503762e-01
5.46377420e-01 7.91006744e-01 2.05231830e-01 1.01977396e+00
-5.28475456e-02 4.24251199e-01 -1.69454873e-01 -2.31007248e-01
3.51041891e-02 -8.10958982e-01 6.71167850e-01 1.41029513e+00
-7.40685910e-02 2.52898127e-01 -1.59180716e-01 8.85043621e-01
3.71002913e-01 2.55864978e-01 -3.75655919e-01 -5.57557821e-01
-4.07345518e-02 1.16062987e+00 -1.04170907e+00 -3.69059652e-01
-5.54730237e-01 1.13432920e+00 2.57407218e-01 6.58677399e-01
-6.81165695e-01 -2.02125371e-01 9.58335161e-01 8.37929025e-02
1.49407670e-01 -5.33463717e-01 3.05965059e-02 -1.11306703e+00
-7.25838542e-01 -7.97696352e-01 2.78252631e-01 -5.80731213e-01
-1.61743224e+00 2.40693986e-01 -2.64241457e-01 -3.04383785e-01
1.58782229e-01 -8.56243372e-01 -1.54230833e-01 1.16156089e+00
-1.24497366e+00 -7.45167255e-01 3.52481604e-01 2.21487969e-01
4.17514771e-01 1.20497189e-01 9.78621066e-01 2.10165471e-01
-5.23935616e-01 -1.20333008e-01 1.12862420e+00 -3.46310049e-01
8.75795007e-01 -9.78105009e-01 4.67623830e-01 2.30676994e-01
-7.02427477e-02 1.14089978e+00 8.99833739e-01 -1.00305939e+00
-1.45044792e+00 -6.25786304e-01 1.03623593e+00 -4.27774131e-01
7.02822447e-01 -1.33080646e-01 -9.59363282e-01 2.27711797e-01
1.76044852e-01 -3.33301306e-01 1.29184210e+00 -1.84693113e-01
2.88181365e-01 4.06293154e-01 -1.23454702e+00 4.91228372e-01
6.69412434e-01 -3.04079741e-01 -1.19228974e-01 9.64554071e-01
-2.11674124e-01 -3.70010734e-01 -1.29249585e+00 -5.15194722e-02
5.08938789e-01 -8.04255128e-01 1.22969174e+00 -8.80079150e-01
-2.94090398e-02 -2.93709904e-01 3.68672162e-01 -1.08415782e+00
-7.06853330e-01 -8.23229790e-01 -2.81282246e-01 7.34513700e-01
8.45187828e-02 -5.78551948e-01 8.82412136e-01 7.57759273e-01
1.94204912e-01 -1.04187846e+00 -9.86969411e-01 -2.93697447e-01
5.54891109e-01 -1.61271870e-01 -2.84343585e-02 7.77090907e-01
2.88073719e-01 4.38856512e-01 -1.72966659e-01 -4.76232499e-01
8.19554687e-01 -1.30407259e-01 3.95433277e-01 -1.19608998e+00
3.14646885e-02 -5.63574076e-01 -4.28647041e-01 -1.38356876e+00
-4.99936156e-02 -8.15783918e-01 -2.74246693e-01 -1.97499108e+00
6.58422709e-01 -5.92940301e-02 -1.57113329e-01 -1.16722889e-01
-9.78336334e-02 5.44182397e-02 -4.74007338e-01 5.32749295e-01
-4.19098556e-01 6.06067181e-01 1.29331803e+00 -4.00768816e-01
-1.25864491e-01 9.12065655e-02 -7.40124702e-01 6.43433928e-01
6.09575093e-01 -6.93845749e-01 -2.35803664e-01 -3.68608624e-01
3.55160922e-01 -4.62329127e-02 -4.29376997e-02 -2.31616020e-01
3.46309721e-01 -2.52347201e-01 6.20627165e-01 -5.95673859e-01
4.13598448e-01 -2.91310638e-01 1.75658420e-01 7.93030500e-01
-1.24832436e-01 1.51336521e-01 4.84315455e-02 9.11680222e-01
-2.53085382e-02 -1.85926422e-01 1.00989664e+00 -3.52230370e-01
-3.61645013e-01 3.85149151e-01 -1.33426476e+00 -3.64065140e-01
1.04341590e+00 -2.34536171e-01 -2.03830048e-01 -3.65908533e-01
-1.30021942e+00 -4.75805961e-02 5.63510418e-01 -1.78525746e-02
7.87262559e-01 -7.86691844e-01 -6.51059806e-01 -1.77260622e-01
-2.50196636e-01 -2.06123680e-01 8.14683855e-01 1.48896194e+00
-7.09896028e-01 8.30754459e-01 1.07927643e-01 -7.68830419e-01
-1.51400542e+00 9.50214803e-01 3.80196095e-01 -1.97226524e-01
-2.22455353e-01 7.96897531e-01 3.65875363e-01 -2.65165180e-01
-4.92246784e-02 -1.24111712e-01 -1.60621732e-01 -2.86715508e-01
6.57272696e-01 5.90552211e-01 3.79500501e-02 -8.68965924e-01
-5.32832384e-01 6.41678512e-01 -2.80050874e-01 2.97180451e-02
1.60148859e+00 -3.45320255e-01 -5.91439962e-01 2.90346384e-01
6.45263612e-01 -1.64747059e-01 -9.24325109e-01 -1.57719225e-01
-1.77104145e-01 3.35410148e-01 1.25142470e-01 -7.50452042e-01
-7.89070249e-01 1.23661816e+00 6.29680872e-01 -1.90830216e-01
7.17335820e-01 1.21959902e-01 7.74715424e-01 1.38435975e-01
1.93222359e-01 -1.04160404e+00 -1.19770281e-01 5.58526158e-01
7.02615857e-01 -8.96919966e-01 1.90366283e-01 -3.53538394e-01
-7.41354406e-01 1.08434451e+00 -1.31767184e-01 5.74334003e-02
1.02444720e+00 3.07518274e-01 2.01553091e-01 -4.61824298e-01
-6.83664978e-01 -7.49352723e-02 1.39037043e-01 1.19304049e+00
1.07374346e+00 -3.15652013e-01 -6.80179298e-01 3.67985189e-01
5.99807501e-01 2.13752165e-01 2.25561857e-01 1.06477940e+00
-5.81181407e-01 -1.53832567e+00 -5.00658512e-01 2.68202424e-01
-5.63971579e-01 -3.13378155e-01 -7.38039911e-01 9.78661329e-02
-2.68826663e-01 8.63262951e-01 -3.53956819e-01 2.82999665e-01
-1.44016370e-01 2.04344898e-01 7.31910527e-01 -1.92855969e-01
-4.01256770e-01 7.39136636e-01 -4.26932663e-01 -2.01204330e-01
-9.50171053e-01 -9.84496474e-01 -1.25419235e+00 -6.07564270e-01
-5.75945377e-01 4.40579146e-01 9.03001547e-01 9.11441624e-01
8.91416132e-01 3.12557489e-01 -1.88730523e-01 -5.87048054e-01
-3.04406971e-01 -1.06357670e+00 -9.54325378e-01 -3.23538445e-02
4.12448235e-02 -6.04563534e-01 -3.83815952e-02 3.71809512e-01] | [11.153564453125, -0.0023050648160278797] |
508b2cb5-7b58-43e5-b166-b2313b9a396b | taming-small-sample-bias-in-low-budget-active | 2306.11056 | null | https://arxiv.org/abs/2306.11056v1 | https://arxiv.org/pdf/2306.11056v1.pdf | Taming Small-sample Bias in Low-budget Active Learning | Active learning (AL) aims to minimize the annotation cost by only querying a few informative examples for each model training stage. However, training a model on a few queried examples suffers from the small-sample bias. In this paper, we address this small-sample bias issue in low-budget AL by exploring a regularizer called Firth bias reduction, which can provably reduce the bias during the model training process but might hinder learning if its coefficient is not adaptive to the learning progress. Instead of tuning the coefficient for each query round, which is sensitive and time-consuming, we propose the curriculum Firth bias reduction (CHAIN) that can automatically adjust the coefficient to be adaptive to the training process. Under both deep learning and linear model settings, experiments on three benchmark datasets with several widely used query strategies and hyperparameter searching methods show that CHAIN can be used to build more efficient AL and can substantially improve the progress made by each active learning query. | ['Tianyi Zhou', 'Xiaotian Lu', 'Jieyu Zhang', 'Linxin Song'] | 2023-06-19 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 2.87278116e-01 4.74026829e-01 -7.36163616e-01 -7.82829106e-01
-1.28040004e+00 -5.52856684e-01 1.96908936e-01 2.36510977e-01
-9.97328281e-01 8.46235335e-01 -2.11597756e-01 -2.48195097e-01
-1.48083061e-01 -8.24700058e-01 -8.63913357e-01 -6.70588493e-01
2.15409756e-01 7.96699464e-01 4.90763575e-01 2.35138997e-01
8.36953744e-02 3.69250417e-01 -1.23394918e+00 3.02651022e-02
1.17941058e+00 9.12738264e-01 2.29068473e-01 4.09252077e-01
-2.46598676e-01 7.90946305e-01 -7.60804653e-01 -3.42471361e-01
3.16660613e-01 -3.24350566e-01 -7.70058453e-01 -1.21856868e-01
7.26567507e-01 -4.33890939e-01 1.18833892e-01 8.53501856e-01
6.74749792e-01 2.87446827e-01 2.07200170e-01 -1.09422290e+00
-1.04250886e-01 1.00266182e+00 -3.65448207e-01 3.20736587e-01
-2.73363620e-01 2.57001609e-01 1.27766788e+00 -7.38727391e-01
4.98955339e-01 1.10580671e+00 6.02335036e-01 8.17072213e-01
-1.42677009e+00 -1.03020656e+00 5.82443058e-01 1.31012231e-01
-1.25279891e+00 -5.70367515e-01 6.81953728e-01 -7.05511272e-02
6.48464262e-01 3.86460394e-01 8.12840402e-01 8.84755611e-01
-2.03744218e-01 1.08831894e+00 5.89851141e-01 -4.69592959e-01
5.52675545e-01 2.96553731e-01 2.91428149e-01 8.38978589e-01
3.08101445e-01 -2.26425510e-02 -7.01295197e-01 -5.17125428e-01
5.28176010e-01 -7.17453510e-02 -2.30967119e-01 -6.00340903e-01
-6.28876448e-01 1.07461631e+00 3.71922493e-01 -1.41406149e-01
-1.47651955e-01 3.78001839e-01 2.64750212e-01 4.33322877e-01
6.33610725e-01 9.09352124e-01 -7.90063858e-01 -1.58096626e-01
-8.83642197e-01 4.11650091e-01 6.74318492e-01 8.83943498e-01
1.13491380e+00 -2.32340768e-01 -5.52250981e-01 1.02095687e+00
4.91499990e-01 1.81884170e-01 3.21292192e-01 -1.10015583e+00
4.52149481e-01 1.08312118e+00 1.72706619e-02 -3.75165939e-01
-1.63550526e-01 -5.46160400e-01 -4.49047774e-01 3.28695662e-02
5.20449221e-01 -3.20473582e-01 -9.11702394e-01 1.76143062e+00
7.37625420e-01 -2.85911709e-02 -4.48065639e-01 7.09600210e-01
6.37444556e-01 4.68824714e-01 3.10101956e-01 -3.72838020e-01
8.48858178e-01 -1.06560946e+00 -5.69966376e-01 -4.61935967e-01
1.04599953e+00 -4.67251688e-01 1.62543094e+00 4.08852726e-01
-1.22522616e+00 -2.78637469e-01 -1.01363802e+00 -2.07794189e-01
3.70833538e-02 -9.22338665e-02 7.08101273e-01 5.57827652e-01
-7.56973445e-01 3.40540171e-01 -1.05613267e+00 1.36777923e-01
9.19031799e-01 5.65808356e-01 1.59307778e-01 -9.64141116e-02
-1.05309570e+00 4.37389225e-01 4.26766247e-01 -2.02059105e-01
-1.04225719e+00 -1.26036870e+00 -5.12793243e-01 3.51508141e-01
1.01066542e+00 -5.46034634e-01 1.60549414e+00 -1.14382637e+00
-1.54877985e+00 6.35162830e-01 -1.24445699e-01 -7.08836257e-01
7.55225182e-01 -4.30200487e-01 3.43422025e-01 -9.37658325e-02
-3.40788931e-01 1.03110301e+00 8.51405084e-01 -8.71692955e-01
-4.80263323e-01 -3.25429291e-01 3.41259509e-01 4.67813879e-01
-7.85184145e-01 -2.77972102e-01 -8.21129322e-01 -2.48564884e-01
4.38692532e-02 -1.02579129e+00 -5.99276125e-01 3.31972599e-01
-1.75010487e-01 -6.74923837e-01 7.65024006e-01 5.95045835e-02
1.36394083e+00 -2.02564311e+00 -1.88955858e-01 1.15077905e-01
2.57679045e-01 3.30553293e-01 -2.15781614e-01 2.48583108e-02
3.91922742e-01 1.81669950e-01 -2.77650446e-01 -4.18558747e-01
-1.41789526e-01 4.03614789e-01 -4.07255799e-01 1.97929427e-01
2.38565385e-01 9.94577587e-01 -8.40245724e-01 -7.02519000e-01
-3.14122230e-01 2.42983535e-01 -9.94765520e-01 5.56438506e-01
-7.26027310e-01 2.09922388e-01 -5.31290352e-01 3.39887530e-01
3.94454330e-01 -5.91046572e-01 1.11872405e-01 1.59341216e-01
2.21508101e-01 5.81751287e-01 -1.07811415e+00 1.52576816e+00
-6.50259554e-01 5.28293312e-01 -7.22941533e-02 -1.02660716e+00
9.21189606e-01 1.47767141e-01 3.06916922e-01 -7.63444662e-01
-2.71258861e-01 5.21617718e-02 -3.82284001e-02 -4.31984305e-01
3.50250632e-01 4.18253064e-01 2.54376322e-01 4.93193775e-01
-1.59834430e-01 8.83446038e-02 2.11335063e-01 1.89955592e-01
1.06427801e+00 -5.05185425e-02 3.00267667e-01 -1.63298681e-01
3.19437087e-01 6.95368573e-02 8.22857618e-01 1.08766592e+00
-1.96896061e-01 2.20628738e-01 5.99087894e-01 -4.78615046e-01
-9.51339900e-01 -5.36737144e-01 -1.96909979e-01 1.65761530e+00
-9.82474685e-02 -4.18449312e-01 -8.06592643e-01 -1.22073734e+00
8.28197971e-03 6.80937588e-01 -6.28264129e-01 -4.41962332e-01
-7.95480728e-01 -1.03882170e+00 3.43741179e-01 5.58292449e-01
5.17211139e-01 -9.56115782e-01 -8.07088375e-01 9.99841690e-02
1.68237761e-01 -5.32074809e-01 -7.15598404e-01 5.71480453e-01
-1.04506528e+00 -9.25543070e-01 -5.59137285e-01 -3.82741928e-01
9.85176563e-01 -1.30096003e-01 1.38304901e+00 2.90751666e-01
-2.56187707e-01 1.74047515e-01 -1.63593903e-01 -7.81962931e-01
-1.62573963e-01 7.22367108e-01 -5.39550126e-01 -3.87418747e-01
3.75692993e-01 -1.28385723e-01 -7.81113684e-01 3.67555141e-01
-8.33088219e-01 -2.79928837e-02 6.31453097e-01 1.00032246e+00
9.32765961e-01 -3.07182670e-01 6.31830633e-01 -1.55641222e+00
5.11794269e-01 -3.13383073e-01 -1.01405668e+00 3.64668757e-01
-1.30477440e+00 2.03570679e-01 6.19953096e-01 -1.00124812e+00
-9.93640602e-01 2.34057665e-01 -2.88720205e-02 -3.16411942e-01
1.79191157e-01 3.60878080e-01 -2.92484671e-01 -2.16904040e-02
9.23046112e-01 -1.39076829e-01 -2.63215035e-01 -4.81007308e-01
1.40071794e-01 3.28390628e-01 -5.62744103e-02 -5.06957471e-01
6.49302065e-01 2.00792477e-01 -1.83745429e-01 -3.23276490e-01
-1.50653744e+00 -3.77341360e-01 -3.15586478e-01 -4.13239514e-03
2.65377492e-01 -9.56431627e-01 -6.30243123e-01 1.81513038e-02
-7.94675112e-01 -1.02065372e+00 -6.44174159e-01 2.68221885e-01
-3.50612074e-01 -1.24275200e-01 -3.14714402e-01 -8.09020102e-01
-4.87967342e-01 -1.12123132e+00 8.52028549e-01 2.61908978e-01
-3.78810495e-01 -9.78576243e-01 2.83760339e-01 4.26994801e-01
4.62823629e-01 -2.91421771e-01 9.90040720e-01 -1.00001276e+00
-8.30635071e-01 -1.05979748e-01 1.10104598e-01 1.95421651e-01
-2.83184141e-01 -2.30745748e-01 -1.04482210e+00 -5.49644232e-01
-1.13335341e-01 -7.55386770e-01 9.43735421e-01 3.58541995e-01
1.77451420e+00 -7.27616131e-01 -2.19734505e-01 7.96637475e-01
1.26994133e+00 2.38640726e-01 3.53705853e-01 3.26984286e-01
5.44746459e-01 2.62741357e-01 9.23861504e-01 1.91895500e-01
-3.18560489e-02 7.33314216e-01 3.99335742e-01 -2.13006482e-01
4.46462370e-02 -2.91818291e-01 1.13219835e-01 3.12370926e-01
4.38876122e-01 -2.99194396e-01 -1.04473770e+00 2.71274000e-01
-1.78831255e+00 -5.66557467e-01 2.46233419e-01 2.38070941e+00
1.41289079e+00 5.62804878e-01 1.41116500e-01 4.65437584e-02
1.90635443e-01 7.74644762e-02 -1.12427318e+00 -6.83468357e-02
2.20085025e-01 1.27523229e-01 6.16674483e-01 6.34450853e-01
-7.72936642e-01 1.02312183e+00 6.44817924e+00 9.73549068e-01
-1.21469736e+00 2.50760704e-01 9.75593626e-01 -4.34495121e-01
-4.24430162e-01 1.06208265e-01 -1.30273497e+00 3.56792063e-01
9.23292816e-01 -3.37116676e-03 3.38198334e-01 1.35916674e+00
9.84276608e-02 -7.72406384e-02 -1.35455561e+00 6.87763274e-01
-3.06302637e-01 -1.35944021e+00 9.03377756e-02 -7.89156258e-02
7.82127082e-01 3.62932198e-02 -6.64711669e-02 5.64395666e-01
4.33095247e-01 -7.95423090e-01 4.15264219e-01 3.70982260e-01
5.67302644e-01 -7.80169725e-01 4.79342133e-01 7.40672767e-01
-6.43548489e-01 -3.90173316e-01 -5.93565583e-01 2.72273511e-01
-2.12957963e-01 7.01899707e-01 -1.37337363e+00 -2.84448534e-01
7.37734497e-01 2.99056143e-01 -7.32635081e-01 1.10587406e+00
-1.13170870e-01 1.17713165e+00 -2.94862479e-01 -2.95462042e-01
1.04378261e-01 -5.51500022e-02 3.41577739e-01 9.72311974e-01
-3.50449234e-02 -8.69205594e-02 2.15180978e-01 9.78170872e-01
-4.03073788e-01 1.50126651e-01 -2.19364747e-01 -1.03370942e-01
9.90748882e-01 1.13490534e+00 -4.70088154e-01 -4.51155186e-01
-7.29582310e-02 5.32132089e-01 7.90585577e-01 3.84610891e-01
-6.57140136e-01 -1.38411060e-01 1.30417228e-01 3.64382982e-01
1.46186128e-01 1.90022111e-01 -3.94574761e-01 -7.51658142e-01
-7.80771747e-02 -9.14327323e-01 5.44825971e-01 -3.16920191e-01
-9.81541812e-01 1.31676346e-01 1.78398892e-01 -8.40834081e-01
-1.81105673e-01 -6.02899604e-02 -5.53173482e-01 6.01702988e-01
-1.47286665e+00 -6.11889601e-01 -3.19022685e-01 3.51215333e-01
6.88061655e-01 -8.52755308e-02 6.38939381e-01 2.86302537e-01
-7.70831048e-01 1.10109758e+00 3.97937112e-02 -3.98647040e-02
8.65808785e-01 -1.39383948e+00 2.78129309e-01 3.56206059e-01
1.24186359e-01 7.64019668e-01 6.18127644e-01 -3.54535311e-01
-1.10932171e+00 -1.08693755e+00 7.97889173e-01 -2.33270004e-01
2.25566864e-01 -5.27820766e-01 -1.18744004e+00 7.52205133e-01
-2.34831810e-01 1.75320253e-01 8.31263602e-01 4.71332997e-01
-2.99184680e-01 -5.68000376e-01 -1.09428263e+00 5.18277586e-01
1.01053417e+00 -2.49650672e-01 -1.48164555e-01 4.90554094e-01
8.71228993e-01 -5.08362234e-01 -7.67212868e-01 5.18479168e-01
4.01383877e-01 -6.80446804e-01 8.16906691e-01 -7.17448473e-01
5.07860444e-04 2.79919535e-01 2.46493518e-01 -1.03374553e+00
-2.04393387e-01 -6.09127045e-01 -5.78049183e-01 1.12664223e+00
7.97316611e-01 -5.83643615e-01 1.20832455e+00 7.97166169e-01
-1.45767212e-01 -1.32503128e+00 -7.64416754e-01 -3.73626709e-01
9.67978165e-02 -2.04174742e-01 7.20751286e-01 8.64911973e-01
-4.93818253e-01 5.25413275e-01 -7.90817961e-02 -1.95615962e-01
5.42968273e-01 -2.09110454e-01 9.81707394e-01 -1.46251392e+00
-5.61219871e-01 -1.25965044e-01 3.14755201e-01 -1.41787529e+00
1.37983426e-01 -4.41089422e-01 1.74403742e-01 -1.07617676e+00
3.86300594e-01 -1.04110980e+00 -2.40977883e-01 7.37177551e-01
-5.71635783e-01 -1.24780796e-01 -3.13663594e-02 2.84935087e-01
-8.86532009e-01 4.17504758e-01 1.33805323e+00 -1.59963116e-01
-3.95216703e-01 3.32097560e-01 -6.32991552e-01 5.39151371e-01
5.62963724e-01 -8.49864423e-01 -9.73093808e-01 -4.08836156e-01
8.08312714e-01 -8.73781070e-02 7.81592261e-03 -6.03600800e-01
3.81930828e-01 -3.28576177e-01 3.27734530e-01 -6.04349911e-01
1.76553607e-01 -8.71362388e-01 -1.18422091e-01 4.88686085e-01
-1.25848591e+00 -1.86135933e-01 2.83082551e-03 6.94681704e-01
1.65245518e-01 -5.02466857e-01 1.08103824e+00 -1.81115910e-01
-1.91581741e-01 6.18119776e-01 -7.47203827e-04 4.62687939e-01
7.88244486e-01 3.08472421e-02 -3.23460907e-01 -9.97335687e-02
-4.18185592e-01 5.24414957e-01 3.33633870e-01 9.94533002e-02
3.21727067e-01 -1.11963534e+00 -4.29220676e-01 2.16421887e-01
1.44638047e-01 7.83002377e-01 -5.79579175e-02 7.49400377e-01
-1.81922272e-01 3.89242887e-01 4.20966655e-01 -6.92646563e-01
-1.37061214e+00 3.08389872e-01 4.94427532e-01 -6.55489326e-01
-4.56698090e-01 1.23040402e+00 7.31888711e-02 -5.30584693e-01
9.76274014e-01 -1.72143698e-01 -1.31013423e-01 1.54357597e-01
4.74879026e-01 3.94264460e-01 1.82160601e-01 2.07667068e-01
1.79295149e-02 9.36439261e-02 -6.29177988e-01 7.33545050e-02
1.15874994e+00 -2.93656681e-02 2.41149142e-01 6.94938838e-01
1.17094517e+00 -2.14342788e-01 -1.74577940e+00 -5.18158436e-01
2.60765940e-01 -3.83428186e-01 4.41022307e-01 -8.49474370e-01
-1.20018804e+00 7.13420212e-01 8.07501376e-01 2.45635375e-01
9.89678919e-01 -2.10766420e-01 7.06002653e-01 1.02445042e+00
9.81223062e-02 -1.29831731e+00 2.94317037e-01 2.29398072e-01
4.63393927e-01 -1.40983152e+00 2.14106962e-01 -2.87665844e-01
-4.07592535e-01 7.96709597e-01 1.03726292e+00 1.03700720e-01
4.29188520e-01 2.38025025e-01 1.72841862e-01 -2.09800437e-01
-1.33695078e+00 3.65998119e-01 3.81612778e-01 -3.11823841e-02
5.65205574e-01 -3.22879702e-01 -2.58845419e-01 2.96250954e-02
4.31894697e-02 -2.23935936e-02 4.75774333e-03 8.36564541e-01
-6.60403073e-01 -1.16620922e+00 -6.37083724e-02 6.43500566e-01
-4.48067278e-01 2.41039693e-02 -5.44912994e-01 7.43333161e-01
4.98878062e-02 5.91772139e-01 4.01307195e-01 1.62658736e-01
2.29701295e-01 1.55460298e-01 3.46761107e-01 -8.55940819e-01
-7.25127041e-01 1.03848949e-02 5.27022816e-02 -6.60973132e-01
-3.12789053e-01 -3.82432252e-01 -1.20120585e+00 8.90025720e-02
-7.49780416e-01 3.68302673e-01 3.33456159e-01 8.34960282e-01
2.86086112e-01 2.54834741e-01 6.20706379e-01 -1.28472537e-01
-1.11922562e+00 -9.38734710e-01 -1.92626044e-01 2.55294085e-01
4.04347330e-01 -4.36436445e-01 -5.62317669e-01 -1.55484870e-01] | [9.064910888671875, 3.868990182876587] |
5c58184c-a0df-4704-a82e-e2f8948f61c1 | micronnet-a-highly-compact-deep-convolutional | 1804.00497 | null | http://arxiv.org/abs/1804.00497v3 | http://arxiv.org/pdf/1804.00497v3.pdf | MicronNet: A Highly Compact Deep Convolutional Neural Network Architecture for Real-time Embedded Traffic Sign Classification | Traffic sign recognition is a very important computer vision task for a
number of real-world applications such as intelligent transportation
surveillance and analysis. While deep neural networks have been demonstrated in
recent years to provide state-of-the-art performance traffic sign recognition,
a key challenge for enabling the widespread deployment of deep neural networks
for embedded traffic sign recognition is the high computational and memory
requirements of such networks. As a consequence, there are significant benefits
in investigating compact deep neural network architectures for traffic sign
recognition that are better suited for embedded devices. In this paper, we
introduce MicronNet, a highly compact deep convolutional neural network for
real-time embedded traffic sign recognition designed based on macroarchitecture
design principles (e.g., spectral macroarchitecture augmentation, parameter
precision optimization, etc.) as well as numerical microarchitecture
optimization strategies. The resulting overall architecture of MicronNet is
thus designed with as few parameters and computations as possible while
maintaining recognition performance, leading to optimized information density
of the proposed network. The resulting MicronNet possesses a model size of just
~1MB and ~510,000 parameters (~27x fewer parameters than state-of-the-art)
while still achieving a human performance level top-1 accuracy of 98.9% on the
German traffic sign recognition benchmark. Furthermore, MicronNet requires just
~10 million multiply-accumulate operations to perform inference, and has a
time-to-compute of just 32.19 ms on a Cortex-A53 high efficiency processor.
These experimental results show that highly compact, optimized deep neural
network architectures can be designed for real-time traffic sign recognition
that are well-suited for embedded scenarios. | ['Michael St. Jules', 'Alexander Wong', 'Mohammad Javad Shafiee'] | 2018-03-28 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 9.29684564e-02 -3.61466557e-01 -2.97974885e-01 -3.08327258e-01
6.98697418e-02 1.49426283e-02 2.83321977e-01 -5.94326138e-01
-8.38577569e-01 3.53903621e-01 -6.11750782e-01 -9.26615894e-01
-1.16217751e-02 -6.49432421e-01 -4.99310821e-01 -8.63663018e-01
6.64620548e-02 2.00997248e-01 6.01263821e-01 -1.35982811e-01
2.24735975e-01 8.48456979e-01 -1.76318324e+00 -1.12809971e-01
7.07285583e-01 1.65891862e+00 -9.36755240e-02 8.60516071e-01
-5.07181622e-02 8.26379418e-01 -5.92933476e-01 -4.56962228e-01
3.75412494e-01 -3.20453569e-02 -7.41689652e-02 -4.70597148e-01
9.02375281e-01 -6.20355010e-01 -8.38575065e-01 8.74641716e-01
4.98826832e-01 -2.48420045e-01 5.31258464e-01 -1.26789093e+00
-1.61670819e-01 1.25436857e-01 -2.27612644e-01 3.60384762e-01
-6.98163390e-01 6.88985407e-01 8.28240812e-01 -5.39793372e-01
2.86531299e-01 8.33256066e-01 6.03226602e-01 7.73396850e-01
-6.63740337e-01 -1.18531060e+00 -2.25598186e-01 5.53017974e-01
-1.11517525e+00 -7.73318172e-01 6.25912130e-01 -1.05850331e-01
1.34304714e+00 6.08858131e-02 7.99962819e-01 7.09767222e-01
4.27072972e-01 5.57438970e-01 7.60403097e-01 -2.05255136e-01
3.66197526e-01 -2.32562602e-01 5.28864861e-01 1.20646870e+00
1.02235591e+00 2.67614692e-01 -4.31755632e-01 2.75494784e-01
8.23342383e-01 -2.71125007e-02 9.21347737e-02 -8.50806832e-02
-9.20911312e-01 3.84593964e-01 7.34784067e-01 2.27355883e-01
-4.79320705e-01 1.25964534e+00 6.36574447e-01 2.72891134e-01
-5.35928071e-01 4.22499664e-02 -3.68502736e-01 -5.53906441e-01
-8.67617130e-01 8.16578269e-02 7.63996422e-01 7.80082047e-01
5.73752880e-01 9.57557797e-01 2.04674646e-01 4.88776654e-01
4.64900732e-01 1.29272139e+00 3.91438961e-01 -6.51301265e-01
5.01526177e-01 6.76508904e-01 -4.12992448e-01 -8.37639868e-01
-6.60379708e-01 -3.21874321e-01 -1.10281301e+00 7.34509945e-01
6.24609232e-01 -1.53515354e-01 -1.33469272e+00 1.35229886e+00
2.73008812e-02 6.23355433e-02 1.44635320e-01 8.35386157e-01
8.99082422e-01 4.20301497e-01 6.35488108e-02 5.61563075e-01
1.68450189e+00 -5.76384366e-01 -3.00808042e-01 -4.62645829e-01
6.76233113e-01 -3.68216813e-01 5.79969883e-01 2.13434219e-01
-6.79130018e-01 -5.86463571e-01 -1.50367892e+00 2.82307994e-03
-5.61567187e-01 5.28224051e-01 8.57896745e-01 1.40599322e+00
-8.48481357e-01 2.12322414e-01 -1.23435152e+00 -4.11840290e-01
6.87617302e-01 9.49690640e-01 3.74109596e-02 4.56216484e-02
-7.55549908e-01 7.98299193e-01 2.26025090e-01 4.94399279e-01
-5.04255712e-01 -6.30182862e-01 -5.63581109e-01 2.51702160e-01
3.49442400e-02 -5.36142886e-01 1.10451734e+00 -7.80847728e-01
-1.59317827e+00 5.51833630e-01 -1.32032514e-01 -7.00722814e-01
9.88303497e-02 3.55865628e-01 -8.26079130e-01 5.96248284e-02
-5.78135669e-01 7.48946309e-01 8.80591035e-01 -6.11404717e-01
-8.01712692e-01 -2.06241950e-01 -2.50851572e-01 -4.13491458e-01
-5.24031103e-01 2.37867564e-01 -3.48934948e-01 -2.72065669e-01
9.59374309e-02 -1.13240898e+00 -1.12917885e-01 6.25877917e-01
1.76594093e-01 4.48534414e-02 1.39213014e+00 -2.23206714e-01
1.05320907e+00 -1.94329870e+00 -8.61805439e-01 5.31821728e-01
4.51457381e-01 1.09723127e+00 -1.77063778e-01 -3.00712228e-01
2.46065110e-01 -1.25551879e-01 3.32024619e-02 1.78456455e-01
1.24243021e-01 2.81354040e-01 -6.25129715e-02 5.24894118e-01
1.48442268e-01 1.28346729e+00 -2.59102523e-01 -1.64613798e-01
4.37008947e-01 5.57939947e-01 -4.45841998e-01 -4.14233804e-01
6.72732741e-02 -3.44849765e-01 -5.26833296e-01 9.77490902e-01
6.71885133e-01 -2.53325492e-01 1.83512956e-01 -6.39406383e-01
-2.46277452e-01 2.80546993e-01 -9.33736324e-01 7.10176110e-01
-4.20425981e-01 1.51506531e+00 2.72676617e-01 -9.43868339e-01
1.11999476e+00 5.31677390e-03 3.63649368e-01 -1.26609230e+00
7.00890124e-01 7.18447030e-01 5.78274786e-01 -2.05569759e-01
5.33040047e-01 -1.34682715e-01 -1.06975600e-01 3.95887822e-01
-1.08122334e-01 2.02575594e-01 3.10613602e-01 -1.83338761e-01
1.31320488e+00 -6.73995376e-01 -1.53263807e-01 -2.21914679e-01
3.89257550e-01 -2.93496996e-02 5.44719040e-01 5.77794015e-01
-6.32933676e-01 -1.66029528e-01 3.29984725e-01 -7.08312154e-01
-1.08508241e+00 -8.91893864e-01 -1.26317516e-01 6.66605413e-01
1.81514591e-01 2.94687320e-02 -6.50593638e-01 -1.48754045e-01
1.71295851e-02 1.71421692e-01 -1.87545061e-01 -2.14686885e-01
-1.25611401e+00 -6.58442974e-01 1.38671196e+00 9.15868640e-01
1.27138865e+00 -9.57083941e-01 -1.49244308e+00 2.61757165e-01
6.68647885e-01 -1.41733646e+00 -1.41369492e-01 3.54539037e-01
-9.33131039e-01 -1.07604301e+00 -5.45227051e-01 -1.04983211e+00
6.43100858e-01 1.97091803e-01 7.02955782e-01 3.80521208e-01
-5.02996802e-01 -3.13467868e-02 5.81851276e-03 -2.91289419e-01
-3.16007555e-01 6.34045005e-02 2.37536237e-01 -1.67443268e-02
5.92895687e-01 -3.56230348e-01 -7.21619785e-01 4.42136824e-01
-6.94681168e-01 -1.01431444e-01 9.87152755e-01 1.00733197e+00
1.01377189e-01 -2.02255070e-01 4.01827604e-01 -1.68770820e-01
2.51988113e-01 3.59957188e-01 -1.19393098e+00 1.96706861e-01
-6.18902504e-01 2.02721104e-01 8.22286487e-01 -3.76874447e-01
-7.67441690e-01 -5.30176377e-03 -1.91737846e-01 -1.41380787e-01
4.42689173e-02 2.07609177e-01 -5.03751226e-02 -8.90392780e-01
4.64801848e-01 2.13607967e-01 2.57665008e-01 -2.65701823e-02
-1.99452229e-03 8.60435486e-01 5.19307673e-01 -2.71842539e-01
6.46579266e-01 5.16881824e-01 5.37265062e-01 -1.37995195e+00
1.76017717e-01 -1.98426872e-01 -2.55999416e-01 -5.27885795e-01
6.62818551e-01 -6.84979737e-01 -1.49145925e+00 9.54360247e-01
-8.59788537e-01 -5.04964173e-01 1.83031186e-02 4.53641325e-01
-1.98508531e-01 1.64128885e-01 -4.28214014e-01 -6.30395174e-01
-6.96220219e-01 -1.35768318e+00 5.13035536e-01 6.14561677e-01
-5.63719645e-02 -6.88372135e-01 -5.59072137e-01 2.01617941e-01
1.00068462e+00 -3.39045078e-02 9.78823900e-01 -2.11917028e-01
-1.20658576e+00 -4.05903012e-01 -9.31032836e-01 3.55778247e-01
-3.55410278e-02 5.42236194e-02 -8.54375839e-01 -1.99972779e-01
-4.55870032e-01 -2.94071555e-01 1.12480021e+00 4.40023214e-01
6.96124136e-01 -6.88996911e-02 -4.49665159e-01 9.52852547e-01
1.42106450e+00 6.13856375e-01 8.88297915e-01 2.23053277e-01
7.51213551e-01 -1.36664361e-01 1.43083498e-01 2.33589351e-01
1.75402015e-01 5.96061349e-01 3.16260338e-01 -6.29983377e-04
-4.96195793e-01 1.24153197e-01 2.93155164e-01 8.63272905e-01
1.40405372e-01 -3.09193701e-01 -1.15627515e+00 3.74089926e-01
-1.48829508e+00 -9.64463770e-01 -2.15898305e-01 2.06597185e+00
9.74569991e-02 5.01345217e-01 -5.59581406e-02 3.61136049e-01
3.62478435e-01 2.01983213e-01 -9.17977512e-01 -6.98215961e-01
-1.65017873e-01 6.30528927e-01 1.25311744e+00 9.23533589e-02
-8.71100545e-01 9.27231550e-01 5.61229372e+00 8.17362785e-01
-1.76035297e+00 -2.70948738e-01 3.75349700e-01 -7.13176280e-02
2.18230739e-01 -2.77527094e-01 -1.18390477e+00 3.66487801e-01
1.33006239e+00 2.58892715e-01 1.69578031e-01 8.49306941e-01
7.24358931e-02 -1.67420268e-01 -6.65685534e-01 1.23823714e+00
-1.13177635e-01 -1.70175683e+00 -5.10127842e-02 3.49758536e-01
4.71828312e-01 4.82659072e-01 2.66932487e-01 2.46950582e-01
-2.39923485e-02 -9.11444128e-01 7.15386331e-01 1.47059441e-01
1.16939867e+00 -6.46137476e-01 7.30609000e-01 -4.00010683e-02
-1.50793731e+00 -3.26308757e-01 -4.37231690e-01 8.52032751e-02
1.30838424e-01 4.73305017e-01 -4.88661706e-01 -1.89464450e-01
6.35951698e-01 3.63900244e-01 -4.04919833e-01 1.28016901e+00
2.64357477e-01 7.77094662e-01 -8.72242689e-01 -8.60286236e-01
5.40620923e-01 3.72606404e-02 1.55120537e-01 1.31669068e+00
3.81467074e-01 -9.81970876e-03 -4.19798017e-01 7.55878925e-01
-2.66557813e-01 -5.10592937e-01 -3.64346832e-01 -1.05301432e-01
5.29062390e-01 9.61454630e-01 -9.18331921e-01 -4.72937256e-01
-2.95258701e-01 6.66602910e-01 -1.09463163e-01 2.28464559e-01
-1.10693729e+00 -9.47556436e-01 1.26582658e+00 -1.25543416e-01
8.12786341e-01 -5.94233334e-01 -7.86292136e-01 -7.52070665e-01
3.36091340e-01 -8.16296518e-01 -1.06640160e-01 -2.75030494e-01
-4.67141658e-01 4.95130152e-01 -5.88370740e-01 -1.21018445e+00
1.96010415e-02 -1.61144876e+00 -6.36517644e-01 5.49040139e-01
-1.73194242e+00 -9.79151607e-01 -5.60721636e-01 2.22647697e-01
5.33473529e-02 -4.88815367e-01 4.51058358e-01 6.64832175e-01
-8.60384285e-01 1.13581026e+00 3.12383533e-01 5.37306368e-01
1.56217039e-01 -4.99119639e-01 9.42140698e-01 9.31675911e-01
-7.06891119e-02 4.08559144e-01 1.30136967e-01 -4.02787864e-01
-2.04960918e+00 -1.04763412e+00 7.24202693e-01 9.73321423e-02
7.03733206e-01 -2.80022442e-01 -5.32358289e-01 1.54300064e-01
-2.66274929e-01 2.66109824e-01 1.94899827e-01 -4.44289327e-01
-5.85709929e-01 -8.43648255e-01 -1.18394935e+00 9.35211718e-01
1.03857172e+00 -3.82981032e-01 8.01501572e-02 -2.46660352e-01
-3.31944157e-03 -3.13013643e-01 -3.73530895e-01 3.58453095e-01
1.10979271e+00 -7.44132876e-01 8.72758627e-01 -2.98138767e-01
-1.59684852e-01 -4.79459733e-01 -1.06210135e-01 -4.99234498e-01
-1.16480343e-01 -6.95378244e-01 -2.64234751e-01 6.00049555e-01
4.79225606e-01 -1.00685346e+00 1.51280785e+00 9.43355978e-01
-2.31767774e-01 -7.42962360e-01 -1.46328461e+00 -1.18169701e+00
-2.12414667e-01 -6.95448637e-01 4.21308488e-01 8.91748220e-02
-1.98702157e-01 4.94580679e-02 -9.92477462e-02 2.81809689e-03
7.74798095e-01 -1.75120547e-01 7.80969203e-01 -1.00614119e+00
1.70166820e-01 -1.12452245e+00 -1.23291862e+00 -1.64659667e+00
-1.84905820e-03 -4.91410851e-01 5.44679835e-02 -1.09941638e+00
-3.40011954e-01 -7.74017334e-01 -2.22877875e-01 5.31189680e-01
4.28922951e-01 4.83003557e-01 2.99605310e-01 -2.36810654e-01
-3.39110732e-01 4.45379019e-01 9.27692473e-01 -3.50927532e-01
7.03860968e-02 -2.15712741e-01 -2.47390881e-01 5.61761618e-01
8.22254479e-01 -2.46189028e-01 8.81403219e-03 -5.63357830e-01
-1.36372969e-01 -3.63864303e-01 5.08469641e-01 -1.58740914e+00
7.54488409e-01 3.90447937e-02 2.34898627e-01 -6.27798975e-01
5.61571956e-01 -9.84747648e-01 -1.38981476e-01 1.08602118e+00
2.83127278e-01 1.34759769e-02 5.31594157e-01 1.55307993e-01
-1.18545048e-01 -5.51420338e-02 9.66515422e-01 4.36874837e-01
-1.27293575e+00 2.51377285e-01 -8.20301116e-01 -1.48795798e-01
7.57138729e-01 -1.03072393e+00 -6.21868610e-01 4.38743606e-02
1.81104332e-01 2.12087715e-03 1.85817137e-01 1.50789201e-01
8.40422928e-01 -1.25440657e+00 -2.91281641e-01 5.21764755e-01
2.91803805e-03 -5.40001929e-01 2.18447492e-01 6.27126634e-01
-1.22806609e+00 1.05437148e+00 -5.72617769e-01 -5.70843399e-01
-1.39946258e+00 -2.20902249e-01 5.72467208e-01 -9.68405381e-02
-6.71293855e-01 7.26936519e-01 -3.67132932e-01 -4.62389514e-02
3.14284265e-01 -9.31404412e-01 4.21127498e-01 -4.04254764e-01
6.18589103e-01 6.63784206e-01 3.20779592e-01 -4.90981191e-01
-6.39815986e-01 9.13028955e-01 6.57131001e-02 1.50046855e-01
1.04667199e+00 4.80758578e-01 2.87696391e-01 -3.89893919e-01
1.24013472e+00 -6.35930240e-01 -1.22520828e+00 1.34799059e-03
3.22882235e-02 -1.49717227e-01 4.86541510e-01 -6.28607571e-01
-1.59729576e+00 9.37672913e-01 9.73033130e-01 -3.99895966e-01
1.16026640e+00 -5.01770318e-01 1.26496720e+00 1.27885067e+00
4.01147783e-01 -1.15811503e+00 -1.62209138e-01 8.83943141e-01
3.22139710e-01 -1.17011786e+00 -5.27225249e-02 -8.31491947e-02
-1.16912082e-01 1.36158609e+00 8.14244032e-01 -1.53996333e-01
8.76377225e-01 6.72232270e-01 2.07348377e-01 -3.29925716e-01
-5.79731107e-01 -4.43013191e-01 1.87816501e-01 6.06661499e-01
1.24633327e-01 2.58988917e-01 6.67125508e-02 4.96156700e-02
-4.06167358e-02 1.91356108e-01 2.67527878e-01 9.38879728e-01
-6.70355558e-01 -7.50143766e-01 -1.89677149e-01 7.57561445e-01
5.39034195e-02 -7.10458532e-02 -1.79101214e-01 8.25087607e-01
-1.17286198e-01 6.99209511e-01 4.37740088e-01 -5.69712341e-01
3.84245366e-01 -5.62483072e-02 4.98156667e-01 1.88114554e-01
-6.79173648e-01 -6.05667830e-01 3.87027651e-01 -5.24990559e-01
-2.13084088e-04 -3.39313567e-01 -1.43395221e+00 -8.51407528e-01
-3.24924618e-01 -6.58484757e-01 1.07412779e+00 8.98791015e-01
4.55212831e-01 4.47035044e-01 1.40086696e-01 -7.81904280e-01
-3.56108695e-01 -4.81595337e-01 -3.26163322e-01 -2.44253829e-01
3.74298543e-01 -4.53142107e-01 2.01582350e-02 -2.67604142e-01] | [8.000544548034668, -0.6611939668655396] |
c251f533-3900-4b29-8ea6-02d2a0327ef3 | cross-modal-self-attention-network-for | 1904.04745 | null | http://arxiv.org/abs/1904.04745v1 | http://arxiv.org/pdf/1904.04745v1.pdf | Cross-Modal Self-Attention Network for Referring Image Segmentation | We consider the problem of referring image segmentation. Given an input image
and a natural language expression, the goal is to segment the object referred
by the language expression in the image. Existing works in this area treat the
language expression and the input image separately in their representations.
They do not sufficiently capture long-range correlations between these two
modalities. In this paper, we propose a cross-modal self-attention (CMSA)
module that effectively captures the long-range dependencies between linguistic
and visual features. Our model can adaptively focus on informative words in the
referring expression and important regions in the input image. In addition, we
propose a gated multi-level fusion module to selectively integrate
self-attentive cross-modal features corresponding to different levels in the
image. This module controls the information flow of features at different
levels. We validate the proposed approach on four evaluation datasets. Our
proposed approach consistently outperforms existing state-of-the-art methods. | ['Mrigank Rochan', 'Yang Wang', 'Linwei Ye', 'Zhi Liu'] | 2019-04-09 | cross-modal-self-attention-network-for-2 | http://openaccess.thecvf.com/content_CVPR_2019/html/Ye_Cross-Modal_Self-Attention_Network_for_Referring_Image_Segmentation_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Ye_Cross-Modal_Self-Attention_Network_for_Referring_Image_Segmentation_CVPR_2019_paper.pdf | cvpr-2019-6 | ['referring-expression-segmentation', 'referring-video-object-segmentation'] | ['computer-vision', 'computer-vision'] | [ 4.37256426e-01 9.14282128e-02 -4.73185301e-01 -6.33948326e-01
-8.25307846e-01 -4.76283193e-01 5.49066067e-01 2.47796372e-01
-5.87998748e-01 3.31655949e-01 2.00108200e-01 2.01995611e-01
3.74865644e-02 -6.12160802e-01 -6.40500486e-01 -5.52300274e-01
3.11800748e-01 1.73396990e-01 4.68360633e-01 -9.43265036e-02
3.17149401e-01 4.21379209e-01 -1.55583882e+00 7.77887225e-01
9.17032659e-01 1.08478129e+00 4.80016887e-01 5.40946126e-01
-5.69843292e-01 1.02325654e+00 -4.05445576e-01 -2.24085793e-01
-2.85266310e-01 -8.16571593e-01 -1.29208815e+00 5.82669139e-01
4.60353702e-01 2.03151479e-01 -1.08138233e-01 1.37636554e+00
2.07326531e-01 3.68277162e-01 7.66805768e-01 -1.11268103e+00
-8.58010888e-01 5.26463985e-01 -9.02895391e-01 5.43774664e-01
2.61258602e-01 2.58205961e-02 9.62890446e-01 -6.37655437e-01
7.26705074e-01 1.48586500e+00 7.13470429e-02 4.68630195e-01
-1.13567948e+00 -3.21670264e-01 5.93559325e-01 3.00996780e-01
-1.24549007e+00 -2.54108369e-01 1.24382901e+00 -5.26540399e-01
7.24774539e-01 1.54319346e-01 4.78943586e-01 6.21346056e-01
8.45147297e-02 1.27811813e+00 1.33048809e+00 -5.34605563e-01
5.85501678e-02 2.45988920e-01 3.59655797e-01 7.08212256e-01
-4.44198370e-01 -3.66701126e-01 -4.88379896e-01 5.95016889e-02
6.83253527e-01 3.00036315e-02 -1.87546507e-01 -1.11636408e-01
-8.41379464e-01 7.37657666e-01 7.13034332e-01 9.14870143e-01
-6.70310020e-01 1.53557181e-01 2.82266080e-01 -1.44088820e-01
4.66843814e-01 4.89348136e-02 -2.61908323e-01 1.85097262e-01
-9.00394678e-01 -2.64998615e-01 2.76061594e-01 8.30188513e-01
8.94976199e-01 -3.44480366e-01 -6.26482069e-01 8.49906743e-01
5.00685930e-01 1.11411333e-01 2.80509472e-01 -9.58791733e-01
2.96248943e-01 9.93506849e-01 -1.19375825e-01 -1.00551784e+00
-3.46945226e-01 -3.35268468e-01 -6.37284160e-01 -7.19633186e-03
2.31008396e-01 7.81354830e-02 -1.01971102e+00 1.96116936e+00
3.97123545e-01 -2.69818585e-02 1.65832534e-01 1.08321083e+00
1.12151229e+00 7.95227587e-01 7.38497317e-01 -2.82812655e-01
1.69076753e+00 -1.21382499e+00 -9.62901771e-01 -5.26494503e-01
2.79458612e-01 -8.57867301e-01 1.20163023e+00 -1.13800675e-01
-1.42292070e+00 -7.91591525e-01 -6.53565168e-01 -4.09412116e-01
-5.44962168e-01 2.29166985e-01 5.07189989e-01 5.28264306e-02
-9.10786271e-01 1.19362138e-01 -4.21356648e-01 -5.36286056e-01
5.48524737e-01 3.65302622e-01 -3.40522528e-01 -6.99603334e-02
-1.19844580e+00 8.16743255e-01 4.60383326e-01 1.32883251e-01
-5.21893442e-01 -2.33000129e-01 -8.16670060e-01 8.71122852e-02
3.66945416e-01 -7.11388052e-01 1.05074990e+00 -1.85411084e+00
-1.12491167e+00 1.51114988e+00 -4.25803035e-01 -1.65722072e-01
2.27394342e-01 -1.96975529e-01 -2.65071571e-01 4.84477103e-01
1.49283946e-01 1.09827101e+00 6.74793541e-01 -1.71571958e+00
-8.11343729e-01 -6.47944570e-01 3.25043142e-01 5.21153748e-01
-1.95659394e-03 2.70623863e-01 -1.18017244e+00 -5.49699008e-01
1.26589864e-01 -4.82929200e-01 -6.64194599e-02 -2.88733780e-01
-4.64694142e-01 -4.46749985e-01 9.25304592e-01 -4.08229470e-01
1.21307063e+00 -2.22662616e+00 2.70131916e-01 -7.98123330e-02
-7.51100555e-02 8.15646276e-02 -1.88155904e-01 -5.27526950e-04
1.02108538e-01 2.66481221e-01 -3.20139557e-01 -3.35734159e-01
-1.70342848e-01 3.74992281e-01 -4.91025560e-02 5.96337736e-01
4.34267074e-01 1.03328156e+00 -7.93552041e-01 -1.12947893e+00
1.82171702e-01 5.63862741e-01 -2.15921387e-01 4.32746291e-01
-3.48787457e-01 6.89086616e-01 -7.97595799e-01 6.66552424e-01
5.76428056e-01 -2.73868412e-01 -1.64447382e-01 -4.67909366e-01
2.49095242e-02 -2.61627465e-01 -7.39261687e-01 1.82321990e+00
-4.95045006e-01 5.39619386e-01 1.79820910e-01 -1.08751559e+00
8.46691072e-01 -1.36233587e-02 4.42958534e-01 -9.09750342e-01
5.84061205e-01 -1.51411578e-01 -4.88652587e-02 -8.75935376e-01
3.64651144e-01 -1.52474761e-01 -3.05566281e-01 2.45317265e-01
3.44289154e-01 4.78257686e-02 2.29490653e-01 1.77982166e-01
4.67282176e-01 1.98077857e-01 4.91548449e-01 -3.25147182e-01
1.03972626e+00 -1.37077600e-01 4.65289474e-01 6.64301634e-01
-4.34969127e-01 4.28207248e-01 6.15671813e-01 -7.34030381e-02
-5.83317041e-01 -7.22444177e-01 -2.99253613e-02 1.57317722e+00
6.04302883e-01 1.12079337e-01 -9.74068105e-01 -7.24055052e-01
-2.57917702e-01 7.53686786e-01 -1.06008554e+00 -6.01881333e-02
-3.86855304e-01 -3.76231611e-01 2.94581085e-01 5.85280538e-01
7.42691934e-01 -1.40613246e+00 -7.15372741e-01 -3.02823931e-02
-2.29392260e-01 -1.36045277e+00 -6.53836310e-01 2.74563015e-01
-6.35524511e-01 -8.56524765e-01 -5.03568709e-01 -1.11450994e+00
9.58604574e-01 -2.21796110e-02 1.26384652e+00 -6.41267467e-03
-1.86481535e-01 6.28424346e-01 -2.43050754e-01 -4.76201018e-03
-2.83620417e-01 5.74987717e-02 -6.90458655e-01 4.97548491e-01
4.89255071e-01 -1.15092985e-01 -4.87776220e-01 2.09790826e-01
-1.15127659e+00 1.65996999e-01 6.14655256e-01 7.24639654e-01
1.09790099e+00 -1.36196673e-01 3.63781691e-01 -9.66701031e-01
5.60764551e-01 -5.61325967e-01 -2.89114952e-01 6.32546544e-01
-3.90139446e-02 2.19136193e-01 2.79476017e-01 -4.54224855e-01
-1.26682079e+00 3.27100396e-01 -1.85520221e-02 -3.62213105e-01
-5.91960549e-01 2.79402494e-01 -5.85238159e-01 1.07239671e-01
8.78088102e-02 1.68232724e-01 -3.62064421e-01 -1.95623487e-01
7.25737572e-01 6.97777152e-01 7.18640625e-01 -7.32964277e-01
8.86680931e-02 5.49846411e-01 -1.69866636e-01 -6.35026455e-01
-1.27044535e+00 -6.75251007e-01 -1.02970171e+00 -3.33261639e-01
1.34287298e+00 -7.00912058e-01 -7.59440780e-01 1.86378270e-01
-1.33116698e+00 -1.13488212e-01 -3.13714236e-01 7.30174631e-02
-6.97950065e-01 1.20442055e-01 -2.99756438e-01 -1.00690341e+00
-2.01952368e-01 -1.33992708e+00 1.43874252e+00 7.61141717e-01
-3.45753245e-02 -1.12969160e+00 -2.07987934e-01 5.03645718e-01
1.84023827e-01 3.66700679e-01 8.19023967e-01 -6.78491473e-01
-4.99629349e-01 -2.81749051e-02 -6.21116996e-01 1.57892108e-01
1.36022180e-01 3.86886634e-02 -1.08196485e+00 2.36415401e-01
-1.03101753e-01 -4.08301502e-01 1.20405269e+00 5.59211850e-01
1.19103634e+00 -9.73389670e-02 -3.37301940e-01 4.42216992e-01
1.32935488e+00 3.53219301e-01 4.37442064e-01 3.52704227e-02
6.89693928e-01 9.83768463e-01 8.68553698e-01 1.12834416e-01
3.88105452e-01 6.24980032e-01 4.61087227e-01 -6.34451509e-01
-9.34071019e-02 -7.33093992e-02 1.05649814e-01 4.11068559e-01
1.50139287e-01 -4.89436269e-01 -8.19105387e-01 7.03243852e-01
-2.03253746e+00 -9.53474104e-01 -1.12152033e-01 1.60832632e+00
8.25978696e-01 -6.21337481e-02 -3.34575437e-02 -2.81221390e-01
8.27467680e-01 2.20087498e-01 -6.67585194e-01 -5.91763914e-01
-4.04311925e-01 -1.32039517e-01 2.35321730e-01 6.50642097e-01
-1.50271201e+00 1.24438405e+00 5.70074940e+00 9.53963935e-01
-1.08373141e+00 3.20610046e-01 1.08127773e+00 2.27827087e-01
-3.82482618e-01 -1.61545947e-01 -6.08377218e-01 2.22082958e-01
4.59767014e-01 -3.32242325e-02 3.68825123e-02 6.90797031e-01
7.76318312e-02 -5.09220600e-01 -9.31868076e-01 8.38920772e-01
4.90319133e-01 -9.34261560e-01 1.02644868e-01 -2.58594304e-01
7.49458611e-01 -2.47430667e-01 1.97649449e-01 1.10027073e-02
-4.56895493e-02 -1.05537975e+00 6.74728811e-01 8.73452008e-01
6.25059187e-01 -8.75324607e-01 8.48587930e-01 1.71207324e-01
-1.20171010e+00 -4.60284855e-03 8.97350237e-02 3.36382538e-01
2.03235120e-01 1.63992688e-01 -2.53937542e-01 4.33993995e-01
5.77797413e-01 6.12544358e-01 -7.08781719e-01 5.46730697e-01
-3.29716146e-01 4.23923135e-01 -7.43681788e-02 1.17642075e-01
7.08861172e-01 -1.44972146e-01 2.07445338e-01 1.48531246e+00
-1.89120829e-01 2.92336047e-01 3.16867918e-01 1.12243211e+00
-1.24052130e-01 3.96358967e-01 -4.57220703e-01 5.28583815e-03
4.28572819e-02 1.27217460e+00 -9.90386248e-01 -5.16005814e-01
-5.38921118e-01 1.17785561e+00 4.22568679e-01 5.56107283e-01
-8.35699320e-01 -1.21679649e-01 3.36471766e-01 -2.25877184e-02
3.00552487e-01 7.80817866e-02 -2.49469131e-01 -8.80141675e-01
-2.56543070e-01 -4.67608184e-01 6.13761067e-01 -1.02709198e+00
-1.25286043e+00 8.42005789e-01 1.39725779e-03 -9.16587412e-01
-7.84879029e-02 -3.58676136e-01 -4.29047078e-01 1.05495608e+00
-1.62183881e+00 -1.46312976e+00 -2.93661118e-01 7.48840332e-01
7.58992374e-01 1.82740673e-01 5.61533630e-01 1.74004614e-01
-5.36811113e-01 2.30330184e-01 -3.95723641e-01 1.30417600e-01
4.35893059e-01 -1.07652462e+00 -3.06895703e-01 9.04099584e-01
2.87497878e-01 6.94020271e-01 5.62889040e-01 -4.72987711e-01
-8.89896452e-01 -9.91855860e-01 6.50797009e-01 -9.53054354e-02
5.14581442e-01 -4.32290025e-02 -1.05590522e+00 4.97453958e-01
7.70071387e-01 3.49068493e-01 5.88780582e-01 -2.34493643e-01
-1.37516767e-01 5.22062629e-02 -1.17774153e+00 3.42236459e-01
7.13536799e-01 -7.29852736e-01 -7.30553627e-01 4.73747365e-02
6.67320132e-01 -3.12610894e-01 -6.50415599e-01 4.13394094e-01
1.90027148e-01 -8.09664845e-01 8.40819180e-01 -8.16789925e-01
5.13377130e-01 -4.19343621e-01 -2.72760332e-01 -7.96325982e-01
-2.50277311e-01 -2.27645949e-01 1.91313937e-01 1.65409887e+00
2.44266734e-01 -4.87081595e-02 4.12221789e-01 5.44116974e-01
6.85414746e-02 -7.61801600e-01 -7.92124033e-01 -1.23260662e-01
-1.25112254e-02 -3.91845465e-01 1.81148171e-01 8.26608419e-01
-8.69757831e-02 6.99182272e-01 -6.28001466e-02 1.90567017e-01
3.85822356e-01 4.72881496e-01 3.33688796e-01 -9.44277644e-01
3.38630006e-02 -7.21576989e-01 -4.12786096e-01 -1.02095330e+00
7.62166321e-01 -8.57081592e-01 1.45741999e-01 -1.62711060e+00
6.54873669e-01 -6.13202006e-02 -5.78060448e-01 4.74807262e-01
-5.34439504e-01 1.75314650e-01 2.88954049e-01 -5.58198281e-02
-1.00951338e+00 4.73749101e-01 1.41172576e+00 -3.43293607e-01
-2.31472850e-01 -2.31279790e-01 -6.97286427e-01 9.91295278e-01
7.86899805e-01 -2.81204283e-01 -4.47909713e-01 -4.13522214e-01
-2.88438618e-01 1.33634835e-01 3.63255084e-01 -5.63339651e-01
3.49901706e-01 -3.39057803e-01 3.53895158e-01 -7.64326632e-01
2.53941029e-01 -9.61042225e-01 -2.76668400e-01 8.25010799e-03
-7.69499183e-01 3.02032288e-02 2.95855492e-01 4.34590161e-01
-5.92750847e-01 -2.30313614e-01 1.13683271e+00 -2.31611282e-01
-9.36996460e-01 1.60104051e-01 -2.49808967e-01 2.50553995e-01
1.27892911e+00 -3.14851329e-02 -2.16993660e-01 -2.39654899e-01
-9.63601828e-01 3.60671401e-01 4.49796289e-01 5.04430592e-01
6.38016939e-01 -1.18358016e+00 -3.14703673e-01 4.40037400e-02
2.83178270e-01 -8.01751241e-02 6.37494147e-01 9.29682195e-01
-6.22147694e-02 2.27160424e-01 -2.48619363e-01 -7.35532820e-01
-1.47758400e+00 8.22624564e-01 5.79678297e-01 -2.03401715e-01
-2.41546348e-01 9.73027766e-01 5.37012756e-01 -1.80662691e-03
2.23927617e-01 -3.73166531e-01 -6.58293009e-01 2.80500531e-01
3.72473180e-01 -2.02470168e-01 -5.31105876e-01 -1.59018993e+00
-4.62313622e-01 9.97522116e-01 1.33427680e-01 -1.43640652e-01
9.28719819e-01 -5.57911634e-01 -3.63991380e-01 7.48091757e-01
1.42589462e+00 -1.43785343e-01 -1.08503509e+00 -5.17193198e-01
-5.46866059e-02 -3.48783195e-01 2.91176319e-01 -6.84397221e-01
-1.34553838e+00 1.13311028e+00 7.36885905e-01 1.34462148e-01
1.54820693e+00 7.02632844e-01 3.72168094e-01 -2.00533450e-01
-9.22180861e-02 -1.07260871e+00 1.20931514e-01 2.22273767e-01
8.69319379e-01 -1.38530004e+00 -4.23953161e-02 -4.36534166e-01
-1.14142418e+00 9.39986229e-01 8.89757097e-01 -4.73420918e-02
3.41217041e-01 2.64058709e-01 1.94242626e-01 -3.51831824e-01
-7.45957077e-01 -9.10710752e-01 6.48981392e-01 4.52878922e-01
7.31107354e-01 -1.97397053e-01 -3.51109087e-01 7.04290152e-01
4.98283029e-01 -4.43610102e-02 4.73379306e-02 6.81635320e-01
-4.46276337e-01 -8.45756590e-01 -5.23111463e-01 3.36696580e-02
-6.45342946e-01 4.67519313e-02 -6.47558451e-01 6.58720851e-01
4.23156857e-01 9.15867150e-01 3.78022641e-01 -8.08995143e-02
3.32359403e-01 1.40280619e-01 4.64084029e-01 -3.42893660e-01
-7.01656520e-01 4.69149679e-01 -1.86305657e-01 -5.69563031e-01
-1.18630683e+00 -7.01794147e-01 -1.73333418e+00 3.52480322e-01
-9.91390049e-02 9.37989578e-02 3.86655629e-01 9.74580944e-01
3.34475935e-01 9.06574965e-01 4.22103792e-01 -6.97295666e-01
2.22091749e-01 -8.44159484e-01 -5.50947905e-01 6.78277612e-01
2.95761019e-01 -7.44508922e-01 -2.22729482e-02 3.89605075e-01] | [10.270099639892578, 1.159226655960083] |
a94c1e69-3f77-4545-8885-2f919cd1c133 | neural-natural-language-processing-for-long | 2305.16259 | null | https://arxiv.org/abs/2305.16259v4 | https://arxiv.org/pdf/2305.16259v4.pdf | Neural Natural Language Processing for Long Texts: A Survey of the State-of-the-Art | The adoption of Deep Neural Networks (DNNs) has greatly benefited Natural Language Processing (NLP) during the past decade. However, the demands of long document analysis are quite different from those of shorter texts, while the ever increasing size of documents uploaded on-line renders automated understanding of long texts a critical area of research. This article has two goals: a) it overviews the relevant neural building blocks, thus serving as a short tutorial, and b) it surveys the state-of-the-art in long document NLP, mainly focusing on two central tasks: document classification and document summarization. Sentiment analysis for long texts is also covered, since it is typically treated as a particular case of document classification. Thus, this article concerns document-level analysis. It discusses the main challenges and issues of long document NLP, along with the current solutions. Finally, the relevant, publicly available, annotated datasets are presented, in order to facilitate further research. | ['Ioannis Mademlis', 'Ioannis Gkionis', 'Dimitrios Tsirmpas'] | 2023-05-25 | null | null | null | null | ['sentiment-analysis', 'document-classification', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 2.42081150e-01 1.49227232e-01 -5.27273536e-01 -3.65975380e-01
-7.62700140e-01 -8.39082658e-01 7.14643896e-01 6.20709956e-01
-6.15410507e-01 7.62123585e-01 6.14616513e-01 -2.34189391e-01
-2.10344523e-01 -5.22547603e-01 -3.19481909e-01 -6.90450132e-01
1.63900971e-01 4.51716304e-01 -6.86960578e-01 -1.81360155e-01
8.20765615e-01 4.55806881e-01 -1.42215014e+00 6.39736652e-01
9.81955290e-01 9.56202030e-01 1.90109521e-01 9.79444861e-01
-7.58042514e-01 1.02899396e+00 -1.13917017e+00 -6.04832113e-01
-2.20445812e-01 -1.94216058e-01 -1.29321873e+00 -6.50764927e-02
5.85572124e-01 -4.44910586e-01 -1.76151216e-01 8.62568676e-01
7.97207117e-01 3.94949287e-01 6.75278425e-01 -8.21047425e-01
-8.83087754e-01 9.56424832e-01 -2.71058023e-01 3.94699395e-01
2.74619609e-01 -3.86995316e-01 1.59074938e+00 -6.92882419e-01
6.66435421e-01 1.15471339e+00 5.41808903e-01 7.39879489e-01
-5.20159125e-01 -1.61404803e-01 3.93393666e-01 2.79288113e-01
-8.27760339e-01 -5.25486648e-01 6.58037663e-01 -2.56377071e-01
1.74598300e+00 7.01436922e-02 5.02605379e-01 1.29391992e+00
3.14677864e-01 1.47442687e+00 3.04107100e-01 -6.40488029e-01
5.11727519e-02 -2.40572155e-01 1.04161525e+00 1.21813975e-01
3.90786797e-01 -4.51468945e-01 -3.18346351e-01 1.36510372e-01
1.39080480e-01 -2.16501221e-01 -4.11298484e-01 2.53425866e-01
-1.10135317e+00 1.08091271e+00 1.16340190e-01 1.09349644e+00
-4.56509113e-01 -6.31584004e-02 1.14890766e+00 2.04729289e-01
5.88967502e-01 5.96093714e-01 -6.56548738e-01 -6.21670127e-01
-1.34406638e+00 5.01086295e-01 1.30536604e+00 8.48341763e-01
3.28504443e-02 4.38051790e-01 -3.14031601e-01 1.19163477e+00
-8.72432366e-02 3.20378482e-01 8.19553196e-01 -5.41216791e-01
9.60476577e-01 3.91784906e-01 -6.97115585e-02 -1.17582691e+00
-6.87876463e-01 -4.75385666e-01 -1.25189972e+00 -4.99520093e-01
2.91228622e-01 -3.46540332e-01 -5.53250432e-01 1.24279904e+00
-4.23887074e-01 -7.26940811e-01 5.64845383e-01 4.22583580e-01
1.42703474e+00 1.30903089e+00 -1.07085571e-01 -4.87176687e-01
1.40597701e+00 -1.21408582e+00 -1.30473244e+00 -4.82030183e-01
7.63568401e-01 -6.51989937e-01 9.58295703e-01 4.78239655e-01
-1.27201092e+00 -3.90733272e-01 -1.08892441e+00 -6.68861985e-01
-8.23821545e-01 4.47967768e-01 4.79014218e-01 4.24038738e-01
-9.03745532e-01 5.84303677e-01 -6.79259658e-01 -5.66265821e-01
5.07288337e-01 3.06749165e-01 -5.06189205e-02 8.81695598e-02
-1.37349236e+00 9.99787152e-01 5.72729111e-01 3.02846342e-01
-2.68873632e-01 -5.55910528e-01 -9.72886980e-01 4.64391142e-01
2.53470421e-01 -5.55689335e-01 1.72832668e+00 -9.43944156e-01
-1.56863308e+00 8.97278428e-01 -4.41442966e-01 -7.25880682e-01
2.64492899e-01 -5.62001050e-01 -4.73891973e-01 2.10778713e-01
-4.20770496e-02 4.95667189e-01 5.03066719e-01 -8.04229379e-01
-6.55388713e-01 -4.86784250e-01 1.19647831e-02 5.89411497e-01
-6.10322952e-01 5.19754648e-01 -2.79487461e-01 -7.72066057e-01
-5.01306713e-01 -3.90433788e-01 6.21788502e-02 -7.01591969e-01
-7.47926056e-01 -7.59958029e-01 7.52197325e-01 -8.11417639e-01
1.60028481e+00 -1.80226898e+00 1.77302826e-02 -2.54493356e-01
1.74476996e-01 2.94713140e-01 -2.19536021e-01 9.78496432e-01
-1.38075083e-01 1.89825356e-01 -1.83268964e-01 -5.54942250e-01
2.35605776e-01 1.02642417e-01 -8.28653574e-01 1.44195244e-01
-8.82091299e-02 1.27053273e+00 -6.42702520e-01 -2.39603609e-01
8.21124986e-02 3.88067335e-01 1.60025671e-01 -2.74119794e-01
-3.71716410e-01 -9.70174298e-02 -2.27406308e-01 4.53728974e-01
5.28865576e-01 -3.20029289e-01 9.53366831e-02 -1.58209261e-02
-3.66220146e-01 6.43436611e-01 -6.76604867e-01 1.67002904e+00
-4.08669263e-01 1.39085424e+00 -5.90423271e-02 -1.35121655e+00
7.40258157e-01 5.04905879e-01 3.78108174e-01 -4.15322095e-01
4.55289721e-01 2.64164299e-01 -1.35967597e-01 -4.88923281e-01
1.18677390e+00 2.29975790e-01 -2.87666231e-01 7.02707052e-01
1.98611930e-01 -1.17930472e-01 9.40150559e-01 2.74850786e-01
6.08931541e-01 -4.00187492e-01 7.71950662e-01 -2.95600116e-01
7.98113108e-01 8.23082104e-02 3.91554460e-02 9.41587985e-01
-1.45348266e-01 2.98752248e-01 8.13920796e-01 -6.62675679e-01
-1.03913057e+00 -4.36679602e-01 -1.65887818e-01 1.11855781e+00
-5.12761474e-01 -4.68915999e-01 -1.02421486e+00 -6.91889405e-01
-1.34663999e-01 8.45846891e-01 -7.03619003e-01 3.34490389e-01
-6.52997255e-01 -1.01709831e+00 8.41358602e-01 7.59593844e-01
5.40003836e-01 -1.47931135e+00 -5.20699441e-01 1.71139091e-01
-4.88147497e-01 -1.12135649e+00 -1.54486775e-01 4.23215687e-01
-9.76506412e-01 -9.08428848e-01 -1.10380983e+00 -1.13573396e+00
3.51281554e-01 2.26110995e-01 1.31458306e+00 -5.99479973e-02
7.16493465e-03 2.28568137e-01 -5.04372895e-01 -6.50467277e-01
-4.85017031e-01 6.95506454e-01 4.47872765e-02 -5.28977931e-01
9.69286978e-01 -1.10115558e-01 -1.83037385e-01 -4.74423468e-01
-9.42267895e-01 -1.91431507e-01 5.65182865e-01 7.89631963e-01
3.29112053e-01 2.17155516e-01 1.04931581e+00 -7.43590534e-01
1.53367901e+00 -1.05786271e-01 -3.69392246e-01 4.33891326e-01
-5.65237582e-01 -3.72147471e-01 1.01378417e+00 -1.38593689e-01
-1.16651773e+00 -6.42752647e-01 -4.75807011e-01 5.23679495e-01
-5.15608251e-01 9.67507482e-01 -9.35854018e-03 4.31269467e-01
6.38312399e-01 3.96064192e-01 -3.23263437e-01 -4.48983908e-01
3.49420696e-01 1.22196329e+00 4.83489573e-01 -2.12843433e-01
8.36254060e-02 2.57386893e-01 -4.69566464e-01 -1.30285490e+00
-1.29020905e+00 -5.53417623e-01 -8.97175610e-01 -1.10691808e-01
8.21109235e-01 -4.32262957e-01 -5.88482916e-01 7.58224428e-01
-1.62590361e+00 -2.30875432e-01 -3.79991651e-01 1.93828836e-01
-3.95721704e-01 6.05308831e-01 -1.01284719e+00 -4.94231641e-01
-1.10290492e+00 -9.46922362e-01 9.04099345e-01 3.04619431e-01
-7.19638705e-01 -1.43163383e+00 3.54940146e-02 5.29168010e-01
2.29623318e-01 -8.71182904e-02 9.81694579e-01 -1.12264669e+00
2.21478730e-01 -6.10886395e-01 -2.00726599e-01 6.41867399e-01
-4.27075513e-02 1.04938537e-01 -1.20478010e+00 -1.83691978e-01
4.29512598e-02 -5.22879183e-01 9.63381588e-01 1.01482081e+00
1.29882050e+00 -4.50548351e-01 -1.53106451e-01 1.90902531e-01
1.06326628e+00 1.75220698e-01 4.78911549e-01 6.26959383e-01
5.74612617e-01 9.19766009e-01 4.76081043e-01 3.01921308e-01
1.55978113e-01 -2.56546866e-02 3.76731940e-02 3.17775190e-01
1.15596555e-01 1.98720187e-01 3.28778625e-01 1.34406507e+00
9.93512124e-02 -1.11580515e+00 -1.04460990e+00 5.00295579e-01
-1.81417274e+00 -1.06168509e+00 -2.62783468e-01 1.39940679e+00
6.89025581e-01 7.60698542e-02 -2.93006748e-01 4.48730618e-01
6.23334289e-01 7.24961638e-01 -5.07175207e-01 -8.96087348e-01
-5.27128637e-01 -3.54559235e-02 1.20509034e-02 4.03524488e-01
-1.29891777e+00 9.71060872e-01 6.88986206e+00 1.05306625e+00
-1.00062227e+00 -3.35534602e-01 5.70371211e-01 -2.07503900e-01
8.63397494e-02 -6.28377557e-01 -1.18423808e+00 2.90201187e-01
1.28718007e+00 -3.02880257e-01 -5.76659245e-03 7.78874874e-01
3.42094123e-01 -1.23390742e-01 -1.13688099e+00 9.40295398e-01
5.70219934e-01 -1.74541116e+00 5.50659895e-01 -1.23106703e-01
9.12812650e-01 3.10569346e-01 -5.81327043e-02 4.61034328e-01
-1.50816143e-01 -9.36033785e-01 3.97538006e-01 4.52068970e-02
5.27144790e-01 -1.02870750e+00 1.23359787e+00 4.90496486e-01
-8.58016968e-01 -2.32088506e-01 -4.15564597e-01 -1.31522417e-01
3.54715884e-01 9.02487874e-01 -2.77969122e-01 6.22451425e-01
6.76294625e-01 1.21185946e+00 -3.24715734e-01 6.75694108e-01
-4.60255116e-01 4.38052505e-01 1.09021321e-01 -5.45036137e-01
6.85555577e-01 -1.75325304e-01 5.85017323e-01 1.73824775e+00
1.40884686e-02 -9.02400166e-02 -1.63981751e-01 5.36013007e-01
-3.36969912e-01 5.36066949e-01 -5.88405669e-01 -6.74031138e-01
1.58279508e-01 1.23035192e+00 -7.08878636e-01 -6.40125990e-01
-5.21836042e-01 8.06772530e-01 2.79136151e-01 4.49081331e-01
-2.27917820e-01 -1.08600819e+00 2.40706116e-01 -5.87677240e-01
7.66830370e-02 -2.02441081e-01 -5.26486218e-01 -1.31163836e+00
4.59454395e-02 -1.06606030e+00 5.63866973e-01 -7.18973219e-01
-1.27201295e+00 6.86315238e-01 -2.14019582e-01 -9.05929565e-01
-4.39530104e-01 -8.69267106e-01 -5.42615235e-01 6.29216909e-01
-1.55593240e+00 -8.35913002e-01 -2.13869825e-01 1.26571253e-01
1.18755448e+00 -1.91477582e-01 1.13327360e+00 1.60664260e-01
-7.69354045e-01 3.42016608e-01 7.74202585e-01 3.36034954e-01
5.38029134e-01 -1.33234417e+00 5.69982052e-01 7.34988391e-01
1.24346487e-01 6.92329526e-01 5.30725241e-01 -5.19164741e-01
-9.60715532e-01 -8.23539913e-01 1.54217780e+00 -1.72566772e-01
7.74312675e-01 -2.95326054e-01 -8.22232306e-01 7.24650145e-01
8.30610573e-01 -7.98842669e-01 9.59941089e-01 1.32918298e-01
1.16161995e-01 6.20080866e-02 -8.27576816e-01 6.69580817e-01
3.56889397e-01 -5.17513216e-01 -8.24357450e-01 4.59261775e-01
5.31054258e-01 -4.43595916e-01 -6.43852472e-01 1.12946017e-03
6.27202511e-01 -7.10853100e-01 6.13130212e-01 -6.95432067e-01
7.82401443e-01 3.47442478e-01 1.49234295e-01 -1.45192814e+00
-8.21260884e-02 -4.96655852e-01 -5.58506787e-01 1.12527359e+00
3.57804358e-01 -4.28582430e-01 7.72860944e-01 2.80851454e-01
-4.18879151e-01 -9.84079599e-01 -4.98669595e-01 -5.72364092e-01
6.38971031e-01 -6.42244458e-01 3.50974232e-01 9.53513026e-01
4.26357299e-01 7.10637271e-01 -1.50447950e-01 -4.18556631e-01
1.84213087e-01 2.31415570e-01 5.46753168e-01 -1.37880242e+00
3.04412991e-01 -1.05223930e+00 2.44425219e-02 -1.52501523e+00
6.39533281e-01 -1.02746058e+00 -2.94649247e-02 -2.12292147e+00
1.92480400e-01 6.59417152e-01 1.40433284e-02 3.19843829e-01
1.26507118e-01 -2.33778462e-01 -8.67525786e-02 3.06801219e-02
-5.19613326e-01 5.01593411e-01 1.10966134e+00 -4.45321769e-01
-1.58990636e-01 2.77838130e-02 -1.02285182e+00 9.00507092e-01
1.25385189e+00 -1.54475003e-01 -2.41420463e-01 -8.12948465e-01
6.66770279e-01 -3.21968913e-01 -3.21624726e-01 -6.98759973e-01
2.98923910e-01 1.18622668e-01 3.61409456e-01 -1.48631942e+00
1.01799913e-01 -4.35131252e-01 -1.01578653e+00 1.84835896e-01
-8.42072129e-01 4.76947539e-02 4.07995790e-01 3.30711484e-01
-7.67059147e-01 -8.65534902e-01 5.81077516e-01 -1.78913981e-01
-6.61632895e-01 1.28274769e-01 -9.60605741e-01 1.36100635e-01
6.91053033e-01 -2.70206004e-01 -5.00157297e-01 -5.97454011e-01
-3.84362787e-01 4.50818688e-01 -1.10162959e-01 5.53447723e-01
5.02333164e-01 -7.73059666e-01 -6.83565617e-01 -1.45411938e-01
2.83969473e-02 2.92976469e-01 5.25088549e-01 6.45226419e-01
-5.96842825e-01 1.36493027e+00 2.09148586e-01 -2.94037670e-01
-1.37150490e+00 3.02959889e-01 1.48815587e-01 -8.11623693e-01
-7.32856572e-01 8.36226404e-01 -1.14559196e-02 -4.49108750e-01
7.07620561e-01 -4.87700552e-01 -9.63645041e-01 5.53206265e-01
9.05040443e-01 5.31858325e-01 5.19102275e-01 -4.50309485e-01
-6.35164157e-02 4.12687749e-01 -5.25510192e-01 2.41403356e-02
1.54978192e+00 -2.85406679e-01 -4.70016569e-01 7.48409390e-01
1.34070802e+00 -1.84758455e-01 -5.04144013e-01 -6.97736219e-02
2.45827287e-01 2.00477749e-01 2.73372352e-01 -9.85942841e-01
-8.07745218e-01 1.33317542e+00 -1.99430361e-01 5.51370740e-01
8.50504637e-01 -2.80939639e-01 1.08778608e+00 1.28855431e+00
-3.40648025e-01 -1.63823330e+00 -6.06030133e-03 1.40480518e+00
1.14124060e+00 -1.09423792e+00 1.24970414e-01 2.32508585e-01
-5.11533022e-01 1.81038427e+00 2.79346079e-01 9.10224542e-02
3.67190063e-01 2.40398541e-01 7.47262537e-02 -3.91729832e-01
-6.30380213e-01 3.07616949e-01 3.03145677e-01 3.96074414e-01
9.17855561e-01 -4.39176649e-01 -3.27229291e-01 7.35169709e-01
-7.43937790e-01 -1.92270540e-02 6.80007219e-01 9.24028873e-01
-4.43946481e-01 -8.70553434e-01 -3.15813482e-01 8.02550197e-01
-9.17911112e-01 -4.17877734e-01 -7.23624289e-01 6.87212765e-01
-4.61851865e-01 1.09036458e+00 3.03661972e-01 3.05244714e-01
1.58939153e-01 2.11216211e-01 2.58392513e-01 -6.81876540e-01
-7.03973889e-01 -2.33585924e-01 4.02799964e-01 -2.50647534e-02
-4.97266740e-01 -6.82614267e-01 -1.16219199e+00 -5.81217945e-01
-4.56468076e-01 3.91289651e-01 9.10613477e-01 1.05851376e+00
1.69591293e-01 6.85877621e-01 5.98442741e-03 -8.99833202e-01
-6.48364067e-01 -1.42503893e+00 -6.60922468e-01 -6.17498755e-02
5.89614153e-01 -3.04217078e-02 -4.66796070e-01 1.31129578e-01] | [12.451088905334473, 9.467552185058594] |
8a2377ed-e82e-4263-818e-af378be6037a | axiou-an-axiomatically-justified-measure-for | 2203.16062 | null | https://arxiv.org/abs/2203.16062v1 | https://arxiv.org/pdf/2203.16062v1.pdf | AxIoU: An Axiomatically Justified Measure for Video Moment Retrieval | Evaluation measures have a crucial impact on the direction of research. Therefore, it is of utmost importance to develop appropriate and reliable evaluation measures for new applications where conventional measures are not well suited. Video Moment Retrieval (VMR) is one such application, and the current practice is to use R@$K,\theta$ for evaluating VMR systems. However, this measure has two disadvantages. First, it is rank-insensitive: It ignores the rank positions of successfully localised moments in the top-$K$ ranked list by treating the list as a set. Second, it binarizes the Intersection over Union (IoU) of each retrieved video moment using the threshold $\theta$ and thereby ignoring fine-grained localisation quality of ranked moments. We propose an alternative measure for evaluating VMR, called Average Max IoU (AxIoU), which is free from the above two problems. We show that AxIoU satisfies two important axioms for VMR evaluation, namely, \textbf{Invariance against Redundant Moments} and \textbf{Monotonicity with respect to the Best Moment}, and also that R@$K,\theta$ satisfies the first axiom only. We also empirically examine how AxIoU agrees with R@$K,\theta$, as well as its stability with respect to change in the test data and human-annotated temporal boundaries. | ['Tetsuya Sakai', 'Janne Heikkila', 'Esa Rahtu', 'Yuta Nakashima', 'Mayu Otani', 'Riku Togashi'] | 2022-03-30 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Togashi_AxIoU_An_Axiomatically_Justified_Measure_for_Video_Moment_Retrieval_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Togashi_AxIoU_An_Axiomatically_Justified_Measure_for_Video_Moment_Retrieval_CVPR_2022_paper.pdf | cvpr-2022-1 | ['moment-retrieval'] | ['computer-vision'] | [-7.69197419e-02 -4.89448696e-01 -2.23954588e-01 -1.29672170e-01
-9.04480815e-01 -6.59004211e-01 4.08408821e-01 1.56368867e-01
-6.29166603e-01 4.71621484e-01 -1.37840286e-01 -8.27195644e-02
-9.07642007e-01 -6.29188418e-01 -3.87890458e-01 -6.32620394e-01
-4.64178264e-01 1.81754485e-01 5.58964193e-01 -2.46090144e-01
5.99408925e-01 7.39912927e-01 -1.82981706e+00 2.31233880e-01
6.78076386e-01 1.42808700e+00 2.86799874e-02 6.07680678e-01
1.29598781e-01 7.19158590e-01 -7.62032568e-01 -4.46285486e-01
2.19971016e-01 -4.23578560e-01 -8.59087348e-01 -2.03284696e-01
1.66700363e-01 -2.32130975e-01 -2.71853358e-01 1.16857409e+00
4.05038774e-01 4.76436824e-01 7.92256832e-01 -1.21209455e+00
-2.81454355e-01 2.42219791e-01 -6.16339982e-01 6.90869689e-01
6.66535676e-01 -2.00984493e-01 1.15669370e+00 -8.43606353e-01
8.02057087e-01 9.17604506e-01 4.65583712e-01 2.10447401e-01
-9.43868339e-01 -5.04018724e-01 8.38934332e-02 3.56022179e-01
-1.79714978e+00 -2.59531528e-01 7.57103086e-01 -4.37226713e-01
8.62239838e-01 6.44210339e-01 3.05942267e-01 4.26527560e-01
1.50904581e-01 8.27674866e-01 9.65585053e-01 -3.45159620e-01
2.42895231e-01 -9.70164016e-02 2.11776197e-01 5.48133790e-01
1.11242212e-01 -1.26313612e-01 -5.65471232e-01 -1.62796572e-01
6.48247778e-01 -2.28443384e-01 -3.58125418e-01 -3.32876503e-01
-1.31042945e+00 5.37239671e-01 -1.00407846e-01 7.49215961e-01
-1.11811854e-01 3.02229878e-02 4.22841668e-01 4.39453006e-01
3.01703840e-01 5.09173036e-01 -3.77293855e-01 -4.03321713e-01
-1.14303625e+00 2.25396261e-01 4.27506655e-01 8.25117409e-01
6.19495988e-01 -1.14952989e-01 -1.38159320e-01 9.70888972e-01
1.03021175e-01 5.73592961e-01 7.69192517e-01 -1.16622519e+00
3.51389170e-01 5.34284234e-01 1.11271918e-01 -1.33935285e+00
-3.00813138e-01 -6.46978989e-02 -6.40088677e-01 9.34931636e-03
3.89200896e-01 4.25467610e-01 -5.98936677e-01 1.82754588e+00
2.50332542e-02 -1.93004400e-01 -2.36109391e-01 8.43568623e-01
7.77502596e-01 6.09668493e-01 -2.79584080e-01 -7.91156828e-01
1.16168141e+00 -3.06361258e-01 -5.82582951e-01 1.65063187e-01
4.76441294e-01 -8.18172634e-01 1.11630702e+00 5.87574840e-01
-1.29552197e+00 -4.17783022e-01 -1.00845218e+00 3.11622918e-01
-3.89548838e-01 -7.07030073e-02 4.59757149e-01 3.51435453e-01
-1.25182271e+00 7.33307302e-01 -4.88576263e-01 -1.52338460e-01
-2.98232824e-01 4.84299093e-01 -5.32966197e-01 9.50701311e-02
-1.34959006e+00 7.56296873e-01 3.07972103e-01 9.57257301e-02
-6.18611395e-01 -2.75726408e-01 -6.71321511e-01 2.64785532e-03
5.33426583e-01 7.74202794e-02 1.02227581e+00 -6.51144385e-01
-1.25022972e+00 7.75517642e-01 -9.66431573e-02 -1.43705949e-01
5.03074288e-01 8.39253888e-02 -6.12420440e-01 5.46417356e-01
1.43155038e-01 3.01387012e-01 7.84241140e-01 -9.78278697e-01
-7.26188838e-01 -3.18931729e-01 1.65802017e-01 1.39954731e-01
-2.40352720e-01 2.03601241e-01 -6.45437360e-01 -5.84088922e-01
6.00798011e-01 -7.86840439e-01 5.02611957e-02 -5.57860136e-01
-4.46317308e-02 -4.65142787e-01 6.18451416e-01 -4.75518823e-01
1.83097303e+00 -2.29826689e+00 -9.33784842e-02 7.08869398e-01
9.85396057e-02 2.72504807e-01 -5.19258529e-02 4.68512863e-01
-2.97282130e-01 3.09188455e-01 4.15406190e-02 2.83316404e-01
-7.36226654e-03 -7.49181956e-02 -3.37461919e-01 5.66119134e-01
-1.87389150e-01 3.94682288e-01 -9.75582004e-01 -7.91815639e-01
1.97643578e-01 2.70468891e-01 -3.95367414e-01 -7.71761909e-02
2.27517672e-02 1.05879173e-01 -4.39436406e-01 6.51979864e-01
4.05070484e-01 -2.17729524e-01 3.14223878e-02 -2.25820765e-01
-2.91170239e-01 -5.56273386e-02 -1.58489203e+00 1.19104898e+00
1.60486221e-01 6.34149671e-01 -3.00928563e-01 -9.57204640e-01
8.42093110e-01 3.79595459e-01 1.05366051e+00 -1.01296079e+00
1.43162012e-01 4.20014381e-01 -6.55779243e-02 -3.76104444e-01
7.12689281e-01 4.46161777e-02 -1.83211207e-01 4.85512793e-01
-3.82124186e-02 1.40764927e-02 7.25517273e-01 3.47939163e-01
1.22145045e+00 -1.61261037e-01 4.97677267e-01 -2.98933476e-01
7.47076571e-01 -3.42938393e-01 5.49277604e-01 6.90811515e-01
-4.39536899e-01 5.69197536e-01 6.61823153e-01 -1.48528203e-01
-7.65413821e-01 -1.18566370e+00 -3.06532770e-01 1.03159976e+00
4.08624530e-01 -5.07481396e-01 -6.88281298e-01 -5.38445592e-01
-9.84171331e-02 3.52271646e-01 -4.84393388e-01 -2.32814908e-01
-6.80571735e-01 -6.59251392e-01 5.42138159e-01 2.71861970e-01
4.18305218e-01 -1.15115690e+00 -8.40499282e-01 -6.08026460e-02
-6.35060966e-01 -8.27052414e-01 -5.06799281e-01 -1.43060833e-01
-8.90265167e-01 -1.25483763e+00 -9.19123054e-01 -3.18117172e-01
5.95230281e-01 5.54306626e-01 1.15229726e+00 6.43498972e-02
-1.81439713e-01 7.00404644e-01 -6.15402043e-01 1.66889891e-01
-3.50914034e-03 -1.84793293e-01 2.36814871e-01 -1.39511198e-01
2.49340266e-01 -3.42753261e-01 -7.72561371e-01 8.76313329e-01
-1.13416207e+00 -5.60202897e-01 5.20811260e-01 5.59895039e-01
9.77703154e-01 5.16477764e-01 4.12226915e-01 -4.01417077e-01
7.45132923e-01 -7.40820840e-02 -4.97977197e-01 3.09353888e-01
-7.36212969e-01 5.44232130e-02 4.68463033e-01 -5.63743949e-01
-5.02403796e-01 -4.88165528e-01 2.53491867e-02 -4.78536844e-01
2.39390656e-01 6.50843680e-01 -5.11887595e-02 1.60888843e-02
6.01283908e-01 2.15168282e-01 -4.35635626e-01 -2.01078862e-01
8.27457458e-02 5.24032176e-01 5.90197682e-01 -6.35682762e-01
3.95475954e-01 4.49466735e-01 2.84904111e-02 -9.81741726e-01
-4.75725621e-01 -8.12919676e-01 -4.30532157e-01 -5.90334475e-01
5.22051156e-01 -4.10317302e-01 -9.93026257e-01 1.85792789e-01
-7.29724586e-01 -5.98482089e-03 -3.09407830e-01 6.69898927e-01
-8.87204170e-01 7.90258944e-01 -4.01648581e-01 -9.99555707e-01
-2.27718592e-01 -1.14177787e+00 8.47800255e-01 -7.27941841e-02
-2.83922434e-01 -4.70044047e-01 9.14549176e-03 4.86332402e-02
6.96178526e-02 2.50016421e-01 8.81510556e-01 -6.55370831e-01
-4.32248294e-01 -5.32976210e-01 -2.26313978e-01 3.90268236e-01
-1.41387939e-01 2.64781952e-01 -5.07611573e-01 -3.62656891e-01
-3.41040036e-03 1.09324493e-01 6.73279226e-01 5.33739924e-01
1.05291510e+00 -1.28553346e-01 -1.63921773e-01 1.30728140e-01
1.25850224e+00 6.47749245e-01 7.79213071e-01 4.57197398e-01
1.25284255e-01 4.87529784e-01 9.97796297e-01 5.16676843e-01
-9.89610609e-03 7.74131894e-01 3.21963519e-01 3.68338704e-01
2.35338792e-01 2.14323550e-01 4.62677926e-01 1.08702481e+00
-4.78861809e-01 -2.22985923e-01 -8.69901299e-01 5.49624205e-01
-1.85246384e+00 -1.16042531e+00 1.34333268e-01 2.69853306e+00
5.57329297e-01 4.16631311e-01 3.01870137e-01 6.06387019e-01
7.35398889e-01 2.67335922e-01 -2.59861290e-01 -3.54513288e-01
-5.96280582e-03 2.84784555e-01 3.47989082e-01 2.50225753e-01
-8.77610922e-01 5.49285769e-01 6.74253464e+00 1.15094411e+00
-9.53563631e-01 -9.37336758e-02 4.10905659e-01 -1.63398817e-01
-9.33786556e-02 -1.97500154e-01 -6.04503334e-01 5.69799185e-01
7.93731689e-01 -2.03738987e-01 3.51267189e-01 8.36229622e-01
2.09083349e-01 -4.99404311e-01 -1.03856492e+00 1.15350163e+00
1.78864270e-01 -7.72552669e-01 -3.53138782e-02 1.05685316e-01
5.10516822e-01 -3.58777106e-01 3.51944268e-01 1.40630960e-01
-5.74575588e-02 -7.40067780e-01 8.15153480e-01 5.64560294e-01
9.80313182e-01 -1.09487164e+00 7.26517797e-01 1.83583260e-01
-1.50361931e+00 1.46387711e-01 -3.15566450e-01 2.37374470e-01
-1.99425537e-02 5.29733479e-01 -3.36320102e-01 5.54076731e-01
8.88800919e-01 1.42132089e-01 -5.92979074e-01 9.95402038e-01
1.81660429e-01 1.05143011e-01 -2.89780080e-01 2.81217676e-02
1.70636818e-01 -1.45510390e-01 6.61964118e-01 1.20509899e+00
5.88586211e-01 1.71990946e-01 -7.19376653e-02 1.07817546e-01
2.74596065e-01 3.33928525e-01 -5.15896320e-01 -7.08297119e-02
3.39478582e-01 9.93186057e-01 -1.10106397e+00 -3.06129843e-01
-2.47349679e-01 7.42097437e-01 9.67669580e-03 2.42289230e-01
-8.89245272e-01 -6.23017251e-01 4.54999208e-01 2.65108377e-01
1.93366945e-01 -3.39438438e-01 1.83548197e-01 -1.16871238e+00
1.69829577e-01 -9.30181265e-01 7.85077691e-01 -7.21811533e-01
-9.74787652e-01 6.43252075e-01 4.14589524e-01 -1.66570258e+00
-5.13317585e-01 -6.92694545e-01 -7.90776610e-02 3.88218939e-01
-1.04290414e+00 -4.02108729e-01 1.50361303e-02 8.62280428e-01
2.95729071e-01 8.24013948e-02 4.91502374e-01 4.92938131e-01
-4.32832450e-01 7.52363622e-01 1.01814650e-01 1.08901836e-01
6.00948215e-01 -1.12710035e+00 -2.88011253e-01 9.06497300e-01
1.99283585e-01 8.46755743e-01 8.77090454e-01 -5.04386961e-01
-1.13161576e+00 -5.83915710e-01 8.44726861e-01 -3.02428991e-01
6.78908288e-01 1.97010756e-01 -7.97256708e-01 2.67395318e-01
-3.30081671e-01 -5.00869267e-02 7.25767493e-01 4.59623933e-02
-3.87370735e-01 -4.59495395e-01 -1.10490513e+00 7.49052107e-01
9.95210588e-01 -6.96414113e-01 -4.84198838e-01 1.65033475e-01
4.41758901e-01 -1.86913744e-01 -1.11212516e+00 8.67049932e-01
7.19120324e-01 -1.28289175e+00 1.06464803e+00 -2.70558387e-01
1.34766936e-01 -3.94718617e-01 -4.12042290e-01 -7.38651216e-01
-2.24256799e-01 -4.20048028e-01 -2.05001101e-01 1.13585496e+00
3.35040122e-01 -3.96068484e-01 5.37259996e-01 5.30811489e-01
-1.22363409e-02 -8.37502837e-01 -1.13604975e+00 -1.11988795e+00
-2.80474424e-01 -7.59572208e-01 4.81357813e-01 8.42264712e-01
2.97532618e-01 -1.22640386e-01 -2.89876640e-01 -1.12130664e-01
3.67766231e-01 1.45919085e-01 4.13972199e-01 -1.12329030e+00
-3.08364958e-01 -7.89637089e-01 -6.57801151e-01 -8.66438687e-01
-1.80365011e-01 -5.39652646e-01 5.52026294e-02 -1.36496091e+00
2.36988470e-01 -3.12266618e-01 -8.17227602e-01 1.55886933e-02
2.31723905e-01 4.64055151e-01 1.94004059e-01 6.66567028e-01
-1.08129144e+00 9.89211127e-02 1.19161296e+00 1.58661142e-01
-2.76112258e-01 4.51386310e-02 -3.61649215e-01 9.46030796e-01
4.69563037e-01 -2.31206313e-01 -4.20174152e-01 -9.17727873e-02
6.16865993e-01 3.00933242e-01 1.24650635e-01 -1.08303630e+00
4.18910831e-02 -2.88245648e-01 1.88551709e-01 -8.27212870e-01
2.50477463e-01 -6.61420763e-01 2.02744037e-01 1.53476164e-01
-1.88252300e-01 5.25428712e-01 -9.97451842e-02 3.26571733e-01
-4.48262036e-01 -4.35814261e-01 8.10491383e-01 -1.64642170e-01
-9.99223709e-01 2.55833656e-01 -2.67255455e-01 -3.12354974e-02
9.86004770e-01 -5.19707680e-01 -2.52557099e-01 -3.86388361e-01
-6.30155027e-01 6.10540435e-02 4.13295895e-01 2.41877049e-01
6.40140474e-01 -1.48150086e+00 -1.79209396e-01 -4.31069471e-02
3.24701011e-01 -3.34626108e-01 2.36823604e-01 9.73552883e-01
-4.19904768e-01 4.98488665e-01 1.46196589e-01 -6.57748103e-01
-1.24625432e+00 7.83246338e-01 5.10398299e-02 -5.87465227e-01
-2.86049306e-01 5.09026110e-01 1.11631930e-01 2.14754909e-01
2.87566453e-01 -1.62563577e-01 -3.88022810e-01 3.01298201e-01
6.22931480e-01 6.82897329e-01 2.04715461e-01 -1.00491536e+00
-4.95993882e-01 7.05247700e-01 -3.32533605e-02 -4.91204381e-01
8.69292021e-01 -2.06884220e-01 -1.64506480e-01 4.66431111e-01
1.13722932e+00 8.91934484e-02 -7.35856116e-01 1.24850739e-02
2.16665670e-01 -6.09949708e-01 -2.74762839e-01 -4.63930845e-01
-8.27907324e-01 5.22180915e-01 5.69995344e-01 6.78768575e-01
1.45336485e+00 5.83669394e-02 5.19231617e-01 3.03968728e-01
5.78411758e-01 -1.43858206e+00 3.76704991e-01 5.51454842e-01
7.63462067e-01 -7.42967963e-01 1.02510020e-01 -2.43390203e-01
-4.33925986e-01 8.36973608e-01 4.04568791e-01 -1.13404348e-01
6.52896702e-01 -1.34325907e-01 -1.86121747e-01 -1.70751587e-01
-4.48606104e-01 -4.15628672e-01 6.16730154e-01 1.84125766e-01
5.04743338e-01 -1.13764256e-01 -9.24167752e-01 5.73554337e-01
-1.36432007e-01 -1.17981151e-01 1.26582518e-01 9.32148576e-01
-6.32426739e-01 -9.17337298e-01 -4.21871185e-01 5.73289633e-01
-7.59638131e-01 1.18052550e-01 -2.33122692e-01 1.05186367e+00
1.56878069e-01 9.67696548e-01 -5.54222763e-02 -6.93548083e-01
4.82287109e-01 -7.05138594e-02 4.57099229e-01 -2.26832300e-01
-1.74356014e-01 4.00467694e-01 -2.95747727e-01 -7.44587362e-01
-5.26054323e-01 -7.18698680e-01 -1.16432655e+00 -3.96821439e-01
-4.71722454e-01 4.76615548e-01 3.81489456e-01 8.85011077e-01
1.62922382e-01 2.95709819e-01 7.67482877e-01 -6.24629736e-01
-4.53641206e-01 -7.61910141e-01 -8.58894348e-01 5.64103901e-01
7.28772208e-02 -9.67302203e-01 -6.20291352e-01 -9.91077498e-02] | [10.28726863861084, 0.6694245338439941] |
533a8da4-ba54-42fa-bbf3-894684423bf8 | editing-commonsense-knowledge-in-gpt | 2305.14956 | null | https://arxiv.org/abs/2305.14956v1 | https://arxiv.org/pdf/2305.14956v1.pdf | Editing Commonsense Knowledge in GPT | Memory editing methods for updating encyclopedic knowledge in transformers have received increasing attention for their efficacy, specificity, and generalization advantages. However, it remains unclear if such methods can be adapted for the more nuanced domain of commonsense knowledge. We propose $MEMIT_{CSK}$, an adaptation of MEMIT to edit commonsense mistakes in GPT-2 Large and XL. We extend editing to various token locations and employ a robust layer selection strategy. Models edited by $MEMIT_{CSK}$ outperforms the fine-tuning baselines by 10.97% and 10.73% F1 scores on subsets of PEP3k and 20Q. We further propose a novel evaluation dataset, MEMIT-CSK-PROBE, that contains unaffected neighborhood, affected neighborhood, affected paraphrase, and affected reasoning challenges. $MEMIT_{CSK}$ demonstrates favorable semantic generalization, outperforming fine-tuning baselines by 13.72% and 5.57% overall scores on MEMIT-CSK-PROBE. These results suggest a compelling future direction of incorporating context-specific user feedback concerning commonsense in GPT by direct model editing, rectifying and customizing model behaviors via human-in-the-loop systems. | ['Niket Tandon', 'Sarah Wiegreffe', 'Xiang Lorraine Li', 'Wenlong Zhao', 'Akshay Krishna Sheshadri', 'Debanjan Mondal', 'Anshita Gupta'] | 2023-05-24 | null | null | null | null | ['specificity', 'model-editing'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.67202580e-01 2.70685345e-01 -1.25071555e-02 -3.36036384e-01
-7.54086137e-01 -8.30378950e-01 4.17614520e-01 3.31211746e-01
-7.78035164e-01 9.42177951e-01 4.64780480e-01 -1.70930728e-01
-2.08331808e-01 -7.42376089e-01 -8.97283673e-01 3.69746536e-02
4.66261744e-01 4.50359881e-01 3.40969622e-01 -5.46519578e-01
5.30772090e-01 4.90433211e-03 -1.03282166e+00 6.39159977e-01
1.54885721e+00 6.16369307e-01 3.44912887e-01 3.66392881e-01
-2.41011679e-01 1.09542871e+00 -8.57400656e-01 -9.06810403e-01
-1.84269056e-01 -3.83449018e-01 -1.14901161e+00 -7.71150291e-01
7.69301653e-01 -1.68627948e-01 -3.73923481e-01 1.04873347e+00
6.68116927e-01 4.40266460e-01 4.66310024e-01 -8.94883633e-01
-1.66710269e+00 1.13629127e+00 6.24258630e-03 6.14507318e-01
4.77956831e-01 3.43797237e-01 9.99186695e-01 -9.62838471e-01
8.58632684e-01 1.14147842e+00 1.00053692e+00 7.26204395e-01
-1.33560324e+00 -8.39072347e-01 1.14500090e-01 7.79623985e-01
-1.17112923e+00 -4.24212575e-01 4.17626232e-01 -1.70433983e-01
1.80822873e+00 4.69652236e-01 4.84782130e-01 1.45767307e+00
-3.89460772e-02 9.57296073e-01 1.24224222e+00 -4.51690674e-01
8.67251605e-02 1.65868662e-02 4.17734683e-01 6.85875773e-01
2.36494079e-01 1.35382125e-03 -9.75665987e-01 -2.22918540e-02
6.02616787e-01 -4.16717112e-01 -4.45056736e-01 3.04186344e-01
-1.48480988e+00 6.01471186e-01 5.39596438e-01 1.31392270e-01
-3.64140362e-01 4.42155033e-01 7.96351612e-01 4.74827170e-01
2.81070948e-01 1.23734570e+00 -6.74643755e-01 -5.73768616e-01
-9.10614073e-01 4.40580010e-01 5.94019771e-01 1.27029610e+00
3.09023529e-01 -4.46781255e-02 -6.57449603e-01 1.24403119e+00
-3.32274973e-01 5.60438514e-01 7.26468623e-01 -1.30521202e+00
7.11957633e-01 5.28762817e-01 7.78265595e-02 -8.84681344e-01
-3.13119948e-01 -6.15864336e-01 -3.95090938e-01 -4.54867274e-01
1.86586320e-01 1.59545571e-01 -7.92851806e-01 2.02354455e+00
-1.47177011e-01 -1.76565498e-01 -2.35865694e-02 4.55755174e-01
9.07474399e-01 4.45483446e-01 4.85305607e-01 2.57148594e-01
1.29095793e+00 -1.04079437e+00 -7.50771165e-01 -4.84164417e-01
8.73392224e-01 -5.13215899e-01 1.90257359e+00 3.04921240e-01
-1.26163101e+00 -2.06637308e-01 -7.58593261e-01 -4.89981174e-01
-6.01448894e-01 1.21980876e-01 4.49357122e-01 4.71162826e-01
-1.11528230e+00 8.50439072e-01 -5.46236277e-01 -7.28148103e-01
5.75122893e-01 -1.12393737e-01 -1.42230570e-01 -2.26394787e-01
-1.64105618e+00 1.57629609e+00 8.20484757e-01 -1.99978516e-01
-6.87770069e-01 -1.30806684e+00 -7.12773740e-01 1.82797194e-01
3.52354944e-01 -1.00507140e+00 1.48018587e+00 -3.35715950e-01
-1.31636858e+00 9.32220995e-01 -2.49776766e-02 -4.54956502e-01
4.77191299e-01 -4.24903452e-01 -5.38043022e-01 4.34372798e-02
3.57030451e-01 6.25164986e-01 3.80294561e-01 -6.96805120e-01
-3.51552546e-01 -3.90091240e-02 1.85493603e-01 2.72816420e-01
-5.21103263e-01 1.24554142e-01 -1.53923899e-01 -1.25814056e+00
-1.80543792e-02 -8.25160265e-01 3.61386746e-01 -3.11929062e-02
-5.41991770e-01 -2.66184628e-01 2.09189713e-01 -1.25151038e+00
1.49303246e+00 -1.87139249e+00 1.59205049e-01 -9.27304178e-02
-8.52584094e-02 4.03014660e-01 -2.70248562e-01 5.13092160e-01
1.28380269e-01 3.21907520e-01 -3.92629087e-01 -2.05617756e-01
4.03612018e-01 8.75340775e-02 -4.08975929e-01 -3.76890302e-01
2.75319815e-01 1.24677646e+00 -1.33312607e+00 -4.08946514e-01
-6.33642673e-02 8.06246921e-02 -9.62403059e-01 -3.09528738e-01
-4.30313289e-01 2.53522471e-02 -4.64606546e-02 6.55584216e-01
3.32139611e-01 -2.34562889e-01 -5.33567220e-02 -3.40946972e-01
2.84673095e-01 7.08980262e-01 -7.05693007e-01 1.99470413e+00
-3.68901521e-01 5.59621096e-01 -5.71093261e-01 -5.40722370e-01
6.27057374e-01 1.28688347e-02 -3.08677495e-01 -1.08267069e+00
-2.39256144e-01 3.65653068e-01 -1.49540931e-01 -5.29113829e-01
1.02743101e+00 -3.53473842e-01 -3.07316005e-01 3.70048881e-01
1.38526738e-01 -3.17013741e-01 3.20618451e-01 6.00060463e-01
1.54682541e+00 4.69916798e-02 1.12570606e-01 -2.48439223e-01
5.00823930e-02 2.55731523e-01 4.21440244e-01 1.12390351e+00
-3.44200730e-01 2.47445494e-01 2.13184625e-01 2.31526420e-01
-8.63717318e-01 -1.26385725e+00 -4.92592715e-02 1.40767896e+00
-1.14591636e-01 -4.11132187e-01 -7.59591579e-01 -6.07092321e-01
2.98636198e-01 1.67786002e+00 -7.31720686e-01 -6.79220319e-01
-5.94902635e-01 -5.48540473e-01 1.37833500e+00 8.92547488e-01
9.11522508e-01 -1.42491746e+00 -2.55041540e-01 4.79909211e-01
-7.02657640e-01 -1.00899124e+00 -6.76964164e-01 9.09759998e-02
-8.29438806e-01 -8.51815343e-01 -4.99403566e-01 -6.32810473e-01
2.25782394e-01 -1.00518107e-01 1.36994553e+00 -1.80582516e-02
4.37242091e-02 4.15581346e-01 -5.27486682e-01 -2.07255244e-01
-5.88874876e-01 3.01935285e-01 -5.96478432e-02 -9.05714869e-01
4.57042485e-01 -7.32595444e-01 -3.94616485e-01 1.48054153e-01
-8.60447884e-01 6.30363375e-02 5.00387907e-01 8.89312088e-01
3.61726075e-01 -1.68995738e-01 1.05545187e+00 -1.00381172e+00
9.85236406e-01 -5.20962656e-01 1.17228821e-01 5.24235606e-01
-7.51487076e-01 -4.02839668e-03 5.62617779e-01 -3.68122727e-01
-1.35415053e+00 -7.96368301e-01 5.04966220e-03 -2.90316850e-01
2.88619995e-01 6.91284776e-01 -1.28880978e-01 1.72767371e-01
1.11080551e+00 5.41422129e-01 -4.44082081e-01 -5.09367585e-01
8.06817293e-01 4.40930068e-01 1.05338192e+00 -1.08075488e+00
4.21620965e-01 1.62989423e-02 -8.95194590e-01 -4.23881114e-01
-8.93670857e-01 -1.14103191e-01 -1.99157387e-01 2.64691144e-01
6.54212236e-01 -8.49637747e-01 -3.65616322e-01 6.79988265e-01
-1.31832182e+00 -7.15384662e-01 -4.58506942e-01 1.70865476e-01
-5.66110492e-01 3.97165418e-01 -9.90642428e-01 -2.86838375e-02
-6.74414635e-01 -6.86172187e-01 4.92941678e-01 3.19083184e-02
-8.77078533e-01 -1.06881940e+00 -2.36137304e-02 7.72382140e-01
8.03197622e-01 -1.12982571e-01 1.51352775e+00 -8.76849890e-01
-3.31703752e-01 5.81607074e-02 -4.75650817e-01 6.89181685e-01
1.88353788e-02 -3.18206549e-01 -7.84006834e-01 -8.48719552e-02
-4.58756328e-01 -5.83476484e-01 1.03115225e+00 -6.19647875e-02
1.34583712e+00 -4.19886023e-01 -2.16981694e-01 4.50644106e-01
1.07070541e+00 8.36266801e-02 7.23468065e-01 6.79758549e-01
5.78238726e-01 5.04399762e-02 3.67267162e-01 2.49919489e-01
6.91499650e-01 4.80539918e-01 6.71275798e-03 7.18391776e-01
-4.60113645e-01 -5.42576075e-01 4.42122936e-01 8.39492440e-01
5.14297448e-02 -2.12207407e-01 -9.22510028e-01 9.07449365e-01
-1.62821841e+00 -1.29162371e+00 2.84997880e-01 1.73960137e+00
1.53469598e+00 2.22020030e-01 -2.75880337e-01 -3.90058532e-02
7.07982183e-01 -2.36831745e-03 -9.26743150e-01 -5.72817564e-01
-4.86843407e-01 5.87055922e-01 2.55303472e-01 3.72707516e-01
-5.96433043e-01 1.34498715e+00 5.79495382e+00 1.10347033e+00
-6.83878243e-01 3.09769511e-01 2.43590027e-02 -3.91724348e-01
-6.49522662e-01 -6.10370822e-02 -7.18303323e-01 7.26813555e-01
7.24076688e-01 -3.49287093e-01 7.57798791e-01 5.54486513e-01
1.84163079e-02 -2.49364972e-03 -1.18561888e+00 7.75268793e-01
1.87653631e-01 -1.57404804e+00 3.24104607e-01 -5.57858348e-01
8.38149071e-01 2.60697514e-01 1.43291429e-01 1.01986027e+00
4.45886791e-01 -9.08201098e-01 1.00886011e+00 7.10738540e-01
7.74260700e-01 -4.49827433e-01 3.10999662e-01 2.66509920e-01
-5.99922299e-01 -7.32197315e-02 -3.03897232e-01 -2.82495469e-02
1.72618926e-01 6.26145601e-01 -8.12958479e-01 2.70533085e-01
8.08294475e-01 7.26553023e-01 -8.95367146e-01 8.41107011e-01
-5.62099934e-01 7.48076499e-01 -2.12349445e-01 -1.48846924e-01
-3.98544893e-02 4.95588303e-01 6.52034998e-01 1.72887492e+00
3.10227573e-01 2.28745043e-01 -3.59372973e-01 1.25468063e+00
-5.88111043e-01 -5.89971896e-04 -5.93057387e-02 -4.34257656e-01
1.34530401e+00 6.69395506e-01 -2.04695463e-01 -6.75392628e-01
7.11352751e-02 1.49932802e+00 8.61588180e-01 3.32393199e-01
-1.05064011e+00 -8.00186574e-01 6.10473156e-01 -1.46494374e-01
2.90731490e-01 3.79185043e-02 -5.17963529e-01 -1.31622040e+00
2.80498028e-01 -9.76348996e-01 5.62497139e-01 -1.28118324e+00
-1.65806675e+00 2.55922556e-01 1.87231287e-01 -4.61276054e-01
1.18823528e-01 -5.51279724e-01 -5.01139164e-01 9.90459621e-01
-1.34470356e+00 -1.11847413e+00 -4.21266973e-01 5.29225588e-01
5.16947389e-01 -2.54274160e-02 8.85139585e-01 2.41443783e-01
-5.09829044e-01 1.11506939e+00 3.75626862e-01 1.32643521e-01
9.11660433e-01 -1.35259366e+00 7.62436986e-01 7.20055223e-01
-3.60741436e-01 1.07440615e+00 6.77228510e-01 -1.07821727e+00
-8.69586766e-01 -1.39811015e+00 1.13813937e+00 -1.05787659e+00
7.73284495e-01 -2.19867192e-02 -1.19685960e+00 1.17023349e+00
2.12283432e-01 -3.05266052e-01 5.12153506e-01 2.23722756e-01
-9.07809854e-01 2.41882443e-01 -1.56330347e+00 7.60617614e-01
1.63300526e+00 -1.00892663e+00 -1.30155659e+00 3.22638601e-01
1.15873575e+00 -5.52588105e-01 -1.13757014e+00 1.83229670e-01
2.98184574e-01 -5.59799671e-01 9.15899277e-01 -7.62501061e-01
6.52608097e-01 -6.94561154e-02 -3.62844110e-01 -1.81916618e+00
-6.54347539e-01 -3.07925016e-01 -2.01110959e-01 1.41779423e+00
6.28071606e-01 -7.12617636e-01 4.38329130e-01 6.80297732e-01
-5.90904117e-01 -4.70513761e-01 -9.35251534e-01 -1.11135387e+00
5.47791958e-01 -5.57390690e-01 6.91732287e-01 1.35899615e+00
4.38492835e-01 2.06780374e-01 5.75960055e-02 -8.24980736e-02
4.65634704e-01 -4.15396988e-01 1.56071037e-01 -6.57613754e-01
-4.48767751e-01 -5.93631625e-01 -6.99685812e-02 -7.43990242e-01
2.94818461e-01 -1.42110646e+00 -2.62496471e-01 -1.74843860e+00
4.43227023e-01 -5.67516446e-01 -4.26799923e-01 9.24851537e-01
-5.54832935e-01 6.78973496e-02 2.59295464e-01 7.04283714e-02
-6.00140929e-01 6.20732367e-01 1.09881914e+00 -2.29008436e-01
-9.99309346e-02 -8.60444903e-01 -1.06180191e+00 4.10694927e-01
9.48119164e-01 -2.87578940e-01 -4.72719431e-01 -9.52760696e-01
3.85159075e-01 -4.85406041e-01 6.73642874e-01 -9.69055057e-01
2.30019838e-01 -2.36012876e-01 2.66625226e-01 -3.38764876e-01
3.23366523e-01 -5.64587601e-02 4.62768897e-02 2.97619343e-01
-7.08730340e-01 2.75918722e-01 6.24273062e-01 4.70872521e-01
2.14699745e-01 -2.61391103e-01 6.30468965e-01 -4.40878391e-01
-1.10026979e+00 -2.58455366e-01 -2.15164900e-01 8.79626274e-01
5.83878636e-01 -3.47978234e-01 -1.10992360e+00 -9.47066024e-03
-7.88311243e-01 1.39841333e-01 4.71716613e-01 5.20917952e-01
6.10642731e-01 -1.34136927e+00 -7.24609137e-01 -2.76133746e-01
5.85832000e-01 -2.93335170e-01 5.15645623e-01 6.32219076e-01
-3.60598624e-01 4.96728718e-01 -2.76315749e-01 -1.05450768e-03
-7.92140841e-01 3.07799429e-01 3.91058922e-01 -1.23182349e-01
-4.85111773e-01 1.32243919e+00 -3.39674443e-01 -9.18073297e-01
-1.74359232e-02 -7.12219596e-01 2.63113886e-01 -1.56145632e-01
4.40852642e-01 8.08249235e-01 3.80114675e-01 -5.10506704e-02
-4.47343171e-01 -6.12050854e-02 -4.46819574e-01 -5.79333752e-02
1.20832634e+00 -7.59236664e-02 -2.46821865e-01 2.06780255e-01
9.24290538e-01 -1.19310975e-01 -6.82222068e-01 -4.95500028e-01
2.27210492e-01 -3.01918030e-01 -2.12010384e-01 -1.92018783e+00
-5.46045780e-01 5.57743073e-01 2.24538043e-01 -4.03320283e-01
9.35604751e-01 -3.58237745e-03 1.05605435e+00 8.11212182e-01
4.52754408e-01 -1.38320172e+00 2.33207047e-01 1.01314294e+00
1.27205634e+00 -7.55299211e-01 -3.46187830e-01 -1.14577547e-01
-6.23988450e-01 5.37550747e-01 8.38438094e-01 9.24613997e-02
1.16114408e-01 -3.81430864e-01 -2.60319591e-01 -1.86500683e-01
-7.84944832e-01 2.74602771e-01 2.09420808e-02 7.08896101e-01
5.06679475e-01 1.63344502e-01 -2.47927308e-01 1.00042808e+00
-7.13197291e-01 2.62545794e-01 4.19845641e-01 9.72853422e-01
-4.88864839e-01 -6.59850717e-01 -1.24209538e-01 7.89290726e-01
-1.86523516e-02 -8.26052904e-01 -4.49945092e-01 6.93584919e-01
1.36358097e-01 7.18135834e-01 -2.11157009e-01 -2.99302012e-01
8.47890496e-01 5.87602556e-01 7.63461232e-01 -7.63793766e-01
-8.67899060e-01 -9.87349987e-01 5.66034615e-01 -5.28702199e-01
1.21773526e-01 -7.34018147e-01 -1.39503574e+00 -6.89119399e-01
-2.60180831e-01 -7.13002384e-02 2.31083065e-01 8.78659248e-01
6.83615208e-01 6.70466542e-01 -3.33986104e-01 -2.48973623e-01
-9.88861084e-01 -1.20497894e+00 -3.33555520e-01 6.18867874e-01
-1.21303186e-01 -6.30444765e-01 -2.15986073e-01 -4.52160165e-02] | [10.506449699401855, 8.164589881896973] |
48fdaa44-1231-46e8-959b-a5a4296126e4 | high-resolution-semantic-labeling-with | 1611.01962 | null | http://arxiv.org/abs/1611.01962v1 | http://arxiv.org/pdf/1611.01962v1.pdf | High-Resolution Semantic Labeling with Convolutional Neural Networks | Convolutional neural networks (CNNs) have received increasing attention over
the last few years. They were initially conceived for image categorization,
i.e., the problem of assigning a semantic label to an entire input image.
In this paper we address the problem of dense semantic labeling, which
consists in assigning a semantic label to every pixel in an image. Since this
requires a high spatial accuracy to determine where labels are assigned,
categorization CNNs, intended to be highly robust to local deformations, are
not directly applicable.
By adapting categorization networks, many semantic labeling CNNs have been
recently proposed. Our first contribution is an in-depth analysis of these
architectures. We establish the desired properties of an ideal semantic
labeling CNN, and assess how those methods stand with regard to these
properties. We observe that even though they provide competitive results, these
CNNs often underexploit properties of semantic labeling that could lead to more
effective and efficient architectures.
Out of these observations, we then derive a CNN framework specifically
adapted to the semantic labeling problem. In addition to learning features at
different resolutions, it learns how to combine these features. By integrating
local and global information in an efficient and flexible manner, it
outperforms previous techniques. We evaluate the proposed framework and compare
it with state-of-the-art architectures on public benchmarks of high-resolution
aerial image labeling. | ['Guillaume Charpiat', 'Pierre Alliez', 'Emmanuel Maggiori', 'Yuliya Tarabalka'] | 2016-11-07 | null | null | null | null | ['image-categorization'] | ['computer-vision'] | [ 5.82951486e-01 8.29263106e-02 -3.20416689e-01 -5.74674785e-01
-4.94543433e-01 -8.02259624e-01 5.48924088e-01 2.54941851e-01
-5.78997672e-01 4.43084210e-01 -5.10229021e-02 -3.42767648e-02
-3.86797488e-01 -1.22717810e+00 -7.20714271e-01 -5.85989714e-01
1.70335367e-01 4.29303080e-01 3.19730878e-01 -2.60675013e-01
3.06007043e-02 7.48406172e-01 -1.96920562e+00 2.88014114e-01
4.92943794e-01 1.33696151e+00 1.99283704e-01 2.36330941e-01
-1.95738584e-01 8.14625263e-01 -4.03967887e-01 1.49247993e-03
2.93729126e-01 -2.92608887e-01 -1.44927561e+00 3.09085250e-01
7.42740214e-01 -5.71346022e-02 1.33584812e-01 1.33110070e+00
1.67392224e-01 1.73923016e-01 8.13731372e-01 -1.03394198e+00
-7.32173741e-01 4.49459434e-01 -2.84878075e-01 4.50850725e-02
5.03151566e-02 -3.52881908e-01 1.30145669e+00 -7.02293396e-01
6.98432863e-01 1.10480440e+00 8.38790178e-01 5.86204231e-01
-1.24628758e+00 -2.03922823e-01 3.71382892e-01 1.27378047e-01
-1.42903090e+00 -4.68462743e-02 6.91528440e-01 -6.18187666e-01
5.29025972e-01 1.46326825e-01 2.79101253e-01 7.61085629e-01
-1.45713180e-01 6.11058772e-01 1.16443801e+00 -3.92139077e-01
4.35819268e-01 -9.82109681e-02 2.64676303e-01 7.62992382e-01
5.89104369e-02 -9.40896496e-02 -1.02445334e-01 1.86427325e-01
5.45959830e-01 9.46873948e-02 -2.57491380e-01 -6.51573539e-01
-1.21047854e+00 8.95006359e-01 1.09188378e+00 6.31745338e-01
-2.65724719e-01 3.66748899e-01 5.05415678e-01 1.75299302e-01
4.93119359e-01 5.48839808e-01 -6.28713906e-01 5.63518465e-01
-8.56431365e-01 1.30053237e-01 4.98039395e-01 7.69231379e-01
1.23245227e+00 -1.77972078e-01 -1.57138437e-01 8.82672846e-01
-7.88678452e-02 2.44755208e-01 3.80741835e-01 -1.04470444e+00
5.43012694e-02 8.34515929e-01 3.02713783e-03 -1.16660011e+00
-7.92477489e-01 -4.75518703e-01 -1.09516442e+00 4.18738842e-01
4.79381919e-01 1.73121795e-01 -1.07852316e+00 1.87641203e+00
9.06729624e-02 9.92023796e-02 9.53666493e-02 9.28201437e-01
8.25328827e-01 3.79393965e-01 3.03005338e-01 3.20005924e-01
1.37699378e+00 -9.36775208e-01 -4.51796889e-01 -1.36187375e-01
6.86599910e-01 -6.11266017e-01 1.03832996e+00 1.56414881e-01
-7.34490812e-01 -8.12831104e-01 -9.09647167e-01 -1.11234151e-01
-9.12141383e-01 3.70019644e-01 4.71387148e-01 5.50309479e-01
-1.39411151e+00 8.38697255e-01 -6.52046382e-01 -8.04718137e-01
5.09196758e-01 5.19732058e-01 -3.54913652e-01 1.72642395e-02
-1.09420145e+00 8.83210838e-01 5.75332582e-01 1.44389212e-01
-9.08629537e-01 -3.18956435e-01 -7.99455166e-01 1.54224306e-01
4.44638103e-01 -6.79551005e-01 1.25762606e+00 -1.42876768e+00
-1.20820200e+00 1.24046183e+00 -1.79209318e-02 -5.86418509e-01
2.26345032e-01 9.51581597e-02 -1.89667046e-01 2.98821658e-01
2.57835686e-01 1.09201634e+00 6.64592445e-01 -1.47497785e+00
-8.79413784e-01 -3.54916155e-01 6.71063542e-01 3.25863399e-02
-4.48317945e-01 -1.38866544e-01 -6.79638833e-02 -6.04571402e-01
3.79339606e-01 -1.01155746e+00 -2.75332868e-01 1.56668752e-01
-4.50333446e-01 -3.05731118e-01 9.18921709e-01 -1.98049292e-01
6.87926352e-01 -1.90980458e+00 4.02016789e-01 1.02938607e-01
2.29846373e-01 2.32815519e-01 -1.64721578e-01 1.59595430e-01
-1.86390698e-01 3.29741120e-01 -6.82860017e-01 -2.75537819e-01
7.17742518e-02 3.13387901e-01 -4.26923752e-01 5.93377948e-01
3.02302152e-01 1.03861654e+00 -7.91763008e-01 -3.45068395e-01
3.32659990e-01 4.42249507e-01 -3.60304922e-01 1.74893513e-01
-4.20787036e-01 4.64339852e-01 -3.89695019e-01 8.40202332e-01
5.64763606e-01 -4.37690228e-01 1.45075008e-01 -4.04699236e-01
-2.01638833e-01 -1.08042113e-01 -1.01462126e+00 1.69035614e+00
-5.79762995e-01 5.41845143e-01 6.92375228e-02 -1.68255460e+00
1.05324650e+00 1.61575586e-01 6.12532198e-01 -4.97398525e-01
3.36760789e-01 1.87055036e-01 -4.53079909e-01 -1.60223663e-01
4.29897696e-01 -1.63847372e-01 -2.97717690e-01 3.03817213e-01
2.67892390e-01 -1.41934931e-01 2.13444084e-01 -2.33416244e-01
8.86778712e-01 2.04274654e-01 4.47082877e-01 -6.59592271e-01
8.20456803e-01 2.35186100e-01 2.89689630e-01 5.76205313e-01
-5.53967431e-02 8.18160415e-01 2.88974792e-01 -9.62122142e-01
-8.44813645e-01 -8.88645113e-01 -3.67866099e-01 1.13808417e+00
3.86015892e-01 -2.30254218e-01 -9.69952643e-01 -8.46960068e-01
-2.27063239e-01 4.64465506e-02 -9.93304074e-01 -1.84617434e-02
-5.30362308e-01 -6.40054703e-01 5.40777028e-01 6.01821363e-01
8.81605744e-01 -1.24372363e+00 -9.16185021e-01 7.99258351e-02
-2.71096349e-01 -1.18313789e+00 1.00940699e-02 4.47458655e-01
-8.02082956e-01 -1.19287181e+00 -7.29907155e-01 -1.02593637e+00
7.79716671e-01 3.95548761e-01 1.31003201e+00 2.50803947e-01
-2.18897939e-01 4.85750675e-01 -5.88984549e-01 -6.13125078e-02
-3.01553696e-01 6.49099350e-01 -1.86033726e-01 2.40508348e-01
3.35086137e-01 -3.98182303e-01 -5.04126132e-01 4.67180580e-01
-1.31610692e+00 -1.34499073e-01 4.87340987e-01 8.27528894e-01
7.64500976e-01 1.29050419e-01 4.58191693e-01 -1.10721815e+00
1.43778071e-01 -3.69488508e-01 -6.93935692e-01 3.60008508e-01
-3.94705921e-01 1.60959750e-01 9.84370232e-01 -1.26970708e-01
-7.03503668e-01 4.93667990e-01 -4.20864165e-01 -1.53200477e-01
-7.70332992e-01 4.60339546e-01 -5.65167889e-02 -4.64841425e-01
6.89806879e-01 -3.86574380e-02 -3.02685320e-01 -4.26870137e-01
5.31876922e-01 4.21793044e-01 6.43422723e-01 -7.46434987e-01
8.50648224e-01 8.68096948e-01 2.49731585e-01 -7.21614957e-01
-1.35746801e+00 -7.35334337e-01 -1.09368873e+00 -8.21087509e-02
1.27376199e+00 -6.91484988e-01 -5.64996302e-01 4.23374951e-01
-1.08813751e+00 -4.25230771e-01 -5.47818124e-01 -1.31493583e-01
-8.57826233e-01 2.48522222e-01 -3.63745540e-01 -2.94133604e-01
-1.08336862e-02 -1.27109182e+00 1.35974956e+00 1.62288502e-01
-8.76445696e-02 -1.22455180e+00 -8.68661329e-02 1.06045730e-01
5.98491251e-01 5.58791459e-01 1.03558373e+00 -4.78113621e-01
-5.41536033e-01 9.01184455e-02 -5.84047794e-01 5.86063683e-01
2.85477370e-01 -3.06216598e-01 -1.16236234e+00 -3.46135080e-01
-3.50349277e-01 -5.36213338e-01 1.23922372e+00 4.09795254e-01
1.53812432e+00 -3.41777429e-02 -2.59973854e-01 8.60391617e-01
1.81452572e+00 -6.80254847e-02 5.96369863e-01 5.70543468e-01
7.44163990e-01 7.56239653e-01 4.84414935e-01 1.56331107e-01
2.16532320e-01 8.39210927e-01 9.07079220e-01 -4.31494951e-01
-3.04945886e-01 -5.34419455e-02 -2.87895799e-02 3.71591717e-01
-1.58770651e-01 -3.21696669e-01 -1.02173042e+00 5.80424488e-01
-1.89413118e+00 -6.68723762e-01 4.91997600e-02 1.95126903e+00
4.47350949e-01 -1.02741793e-01 -1.36576761e-02 2.34774947e-01
8.16979945e-01 3.00162494e-01 -4.47399676e-01 -1.55145273e-01
-3.17349523e-01 4.58412528e-01 7.40949333e-01 2.90936291e-01
-1.66824007e+00 1.17375112e+00 6.21470165e+00 6.53570890e-01
-1.20745027e+00 1.51398152e-01 6.65784538e-01 7.38085747e-01
2.59187613e-02 -5.98548390e-02 -8.66791308e-01 1.13744386e-01
6.60212994e-01 3.15900475e-01 2.46701375e-01 9.93834317e-01
-3.02307427e-01 6.30823001e-02 -1.01828420e+00 6.90859020e-01
5.85491620e-02 -1.44471085e+00 1.99720278e-01 -5.74447662e-02
9.20152605e-01 3.55458178e-04 2.03992620e-01 1.13525383e-01
3.86162460e-01 -1.25143790e+00 8.22965503e-01 2.84963429e-01
9.01862204e-01 -5.83379149e-01 1.05032063e+00 1.52650490e-01
-1.60931528e+00 -3.24654520e-01 -6.53692663e-01 -1.24709092e-01
-2.87700491e-03 3.94605905e-01 -2.69700676e-01 7.14909792e-01
8.40716183e-01 9.58941519e-01 -7.57448375e-01 7.67197728e-01
-2.54961461e-01 3.30840677e-01 -9.67226699e-02 3.78533870e-01
6.35575175e-01 -1.37003198e-01 -6.88928962e-02 1.16257012e+00
4.02782500e-01 -2.62205172e-02 4.36837703e-01 9.14961994e-01
-2.83469260e-01 9.15235728e-02 -6.81433141e-01 1.95594624e-01
1.40798867e-01 1.42925215e+00 -1.21695840e+00 -2.72147208e-01
-3.81874740e-01 8.84736001e-01 5.38343549e-01 2.19104528e-01
-6.33926392e-01 -2.09480107e-01 6.91121995e-01 2.87381802e-02
2.53185749e-01 -7.51256272e-02 -2.65802771e-01 -9.98076379e-01
-2.63419837e-01 -5.75075150e-01 4.77904022e-01 -6.52935565e-01
-1.27489936e+00 8.63452673e-01 5.53797334e-02 -1.22926879e+00
-2.81033684e-02 -9.56584036e-01 -2.25720897e-01 5.96746325e-01
-1.97248626e+00 -1.43886685e+00 -7.68881977e-01 5.37159801e-01
5.70842624e-01 3.71566713e-02 1.26061988e+00 2.62244642e-01
-1.83545649e-01 1.74546793e-01 -1.01733640e-01 2.69675255e-01
5.03280342e-01 -1.45910859e+00 2.23776355e-01 7.61487842e-01
3.79135877e-01 2.46324822e-01 3.36054087e-01 -1.24375805e-01
-6.68964028e-01 -1.53398311e+00 7.88212121e-01 -4.28225458e-01
4.82987583e-01 -2.26765871e-01 -8.34337175e-01 6.59033537e-01
3.07733379e-03 4.00604755e-01 2.91988462e-01 -5.39813526e-02
-5.55386722e-01 -1.46847039e-01 -9.82023418e-01 1.87504455e-01
1.21925616e+00 -4.81846005e-01 -5.12493014e-01 4.12191629e-01
6.69015706e-01 -1.90452456e-01 -7.58570135e-01 7.33732104e-01
2.11362645e-01 -1.11411142e+00 1.24815869e+00 -3.76862139e-01
5.05159914e-01 -5.82560599e-01 -3.92426342e-01 -1.22854292e+00
-4.10291523e-01 5.85839674e-02 6.42029405e-01 9.42725360e-01
1.81516886e-01 -5.29751718e-01 7.37962246e-01 2.35915836e-02
-3.29719186e-01 -5.83064318e-01 -7.87291348e-01 -8.88345718e-01
3.26099932e-01 -1.55380815e-01 5.73062181e-01 1.05964601e+00
-5.65831184e-01 1.60049558e-01 -2.33547211e-01 3.24153662e-01
5.29998779e-01 3.60093236e-01 3.90838891e-01 -1.59284806e+00
1.59344405e-01 -6.35283828e-01 -6.57696366e-01 -1.03577554e+00
6.75036609e-01 -1.07707179e+00 1.82180956e-01 -1.74526989e+00
6.11316934e-02 -6.85876131e-01 -4.32427853e-01 8.39361072e-01
2.39695922e-01 9.16421771e-01 1.58612847e-01 2.24784091e-01
-7.17096031e-01 3.36659849e-01 9.21404183e-01 -4.42297548e-01
1.91985637e-01 -6.84383810e-02 -5.21530390e-01 9.34785306e-01
1.08495712e+00 -4.11698878e-01 -2.36658573e-01 -5.39770067e-01
2.62471497e-01 -6.16822481e-01 5.98335147e-01 -1.31325126e+00
2.09530443e-01 9.54249874e-03 1.12378754e-01 -2.82820702e-01
2.12907463e-01 -1.13388097e+00 2.75533199e-02 3.56510490e-01
-6.04282200e-01 -1.02368966e-01 2.61887964e-02 3.96042109e-01
-6.02392554e-01 -4.49745893e-01 1.15922940e+00 -4.98418659e-01
-1.18122005e+00 3.49088132e-01 -2.79508621e-01 1.12898359e-02
1.03338504e+00 -1.32385224e-01 -2.45161638e-01 -5.60110025e-02
-7.94642091e-01 -1.08061939e-01 6.95434332e-01 3.36945623e-01
3.45717669e-01 -1.21595061e+00 -3.92441809e-01 -4.34678467e-03
4.16356027e-01 2.59802878e-01 2.44196445e-01 4.44695920e-01
-7.39417672e-01 5.54539800e-01 -4.43087459e-01 -7.63412058e-01
-1.07378209e+00 6.13232076e-01 4.84754652e-01 -2.06685275e-01
-5.79977274e-01 6.93484128e-01 4.62133765e-01 -6.72260702e-01
2.83868462e-01 -5.33371031e-01 -5.91792345e-01 2.25536987e-01
3.27620506e-01 1.34138139e-02 2.78184980e-01 -9.90254939e-01
-3.18583101e-01 1.34668529e+00 4.16831076e-01 3.89377743e-01
1.13058400e+00 -1.58291698e-01 -3.86340439e-01 3.93570215e-01
1.25322688e+00 -4.78002727e-01 -1.09792960e+00 -3.35717916e-01
3.16954613e-01 -2.74792880e-01 8.81224126e-02 -6.41481876e-01
-1.29364300e+00 1.03394926e+00 6.80953562e-01 5.17212570e-01
1.41140175e+00 1.51052400e-01 6.44114912e-01 4.82041270e-01
5.20965040e-01 -9.97210562e-01 9.38210040e-02 5.98465443e-01
5.98830462e-01 -1.42459726e+00 -1.22927174e-01 -6.22759283e-01
-2.56040961e-01 1.37462795e+00 4.45654333e-01 -2.90032268e-01
6.04251802e-01 5.80255538e-02 2.05335557e-01 -2.42194474e-01
-1.35670722e-01 -7.97192276e-01 3.50271076e-01 6.67553723e-01
2.95152128e-01 1.17837764e-01 -6.81583285e-02 1.69276014e-01
6.21688180e-02 -1.11187197e-01 3.59237552e-01 6.99290872e-01
-6.18734121e-01 -1.19737530e+00 -3.05867493e-01 1.24938302e-01
-3.31117809e-01 1.62642553e-01 -4.05442834e-01 8.65274549e-01
6.04332149e-01 7.06826270e-01 7.20517561e-02 -3.65030169e-01
3.94841015e-01 -1.98265295e-02 1.70638874e-01 -7.87978053e-01
-5.53932965e-01 -4.06351268e-01 -1.38281643e-01 -6.10912204e-01
-1.13516104e+00 -3.13532770e-01 -1.13684237e+00 -5.94621934e-02
-1.41029000e-01 4.70925719e-02 6.09175205e-01 9.74886537e-01
-4.81419358e-03 6.09336317e-01 3.92171174e-01 -9.65864539e-01
-4.80310887e-01 -6.67199075e-01 -6.77947402e-01 5.94631433e-01
2.10725099e-01 -7.83228993e-01 -4.31339383e-01 3.19625527e-01] | [9.733377456665039, 0.7361385822296143] |
8b7227c4-e056-4d8c-9920-435d6a4ed3b3 | a-log-linear-model-for-unsupervised-text | null | null | https://aclanthology.org/D13-1007 | https://aclanthology.org/D13-1007.pdf | A Log-Linear Model for Unsupervised Text Normalization | null | ['Yi Yang', 'Jacob Eisenstein'] | 2013-10-01 | null | null | null | emnlp-2013-10 | ['lexical-normalization'] | ['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] | [-7.391500473022461, 3.7005531787872314] |
30b1c6da-4c69-4135-95a6-976e9ee7841c | prnu-based-source-camera-identification-for | 2107.01885 | null | https://arxiv.org/abs/2107.01885v1 | https://arxiv.org/pdf/2107.01885v1.pdf | PRNU Based Source Camera Identification for Webcam Videos | This communication is about an application of image forensics where we use camera sensor fingerprints to identify source camera (SCI: Source Camera Identification) in webcam videos. Sensor or camera fingerprints are based on computing the intrinsic noise that is always present in this kind of sensors due to manufacturing imperfections. This is an unavoidable characteristic that links each sensor with its noise pattern. PRNU (Photo Response Non-Uniformity) has become the default technique to compute a camera fingerprint. There are many applications nowadays dealing with PRNU patterns for camera identification using still images. In this work we focus on video, more specifically on webcam video, because of the great importance of webcam video nowadays. Three possible methods for SCI are implemented and assessed in this work. | ['Fernando Martin-Rodriguez'] | 2021-07-05 | null | null | null | null | ['image-forensics'] | ['computer-vision'] | [ 6.03548169e-01 -6.31384909e-01 -8.14960301e-02 1.60197750e-01
-4.23238128e-01 -9.50548589e-01 5.22486389e-01 -1.83900014e-01
-3.57025415e-01 4.29792047e-01 -1.20457694e-01 -1.36973098e-01
-2.58591890e-01 -6.05634332e-01 -7.90865362e-01 -6.01220310e-01
2.20436603e-01 -1.58198327e-02 4.25431609e-01 2.88565874e-01
6.59454525e-01 8.90845478e-01 -1.78491306e+00 2.56724566e-01
1.52660087e-01 1.11442804e+00 1.94392696e-01 9.11704242e-01
1.27142984e-02 9.04768169e-01 -7.93798268e-01 -5.75118959e-01
2.69255400e-01 -4.90214616e-01 -3.02921921e-01 6.56306297e-02
5.08295655e-01 -4.80528265e-01 -1.12258382e-01 1.64169860e+00
7.06689581e-02 -2.75717378e-01 7.83799350e-01 -1.10838914e+00
-3.37176889e-01 6.25293374e-01 -5.36741436e-01 3.12941641e-01
8.96756828e-01 -2.12358892e-01 1.77915648e-01 -3.73546034e-01
6.02544188e-01 9.35549021e-01 8.60873282e-01 2.71367311e-01
-1.00841403e+00 -6.99126899e-01 -8.84852350e-01 4.70621169e-01
-1.48241055e+00 -2.37812951e-01 9.20966566e-01 -4.05855000e-01
4.02994126e-01 4.09885079e-01 8.43518674e-02 1.45554888e+00
4.53860462e-01 2.16658130e-01 1.30018222e+00 -5.16100943e-01
1.60503745e-01 6.75851703e-01 3.65513265e-01 1.38537273e-01
6.74817145e-01 -2.05132701e-02 -3.15420330e-01 -2.07264483e-01
8.59119952e-01 9.61971506e-02 -3.35663915e-01 8.14687982e-02
-6.92010462e-01 4.72966939e-01 -4.23646629e-01 6.69991076e-01
-2.30419010e-01 2.22022474e-01 2.20788762e-01 2.84793764e-01
-3.96645695e-01 2.43306085e-01 2.47886911e-01 -6.62348390e-01
-7.83415854e-01 -2.33533651e-01 8.54102015e-01 1.16411519e+00
5.99585474e-01 -1.69196531e-01 5.32265246e-01 5.49285710e-01
5.82192745e-03 9.51738894e-01 5.79584837e-01 -1.04517031e+00
1.87658250e-01 3.82244945e-01 1.63913727e-01 -1.27706325e+00
9.55934972e-02 3.29965174e-01 -5.58272779e-01 3.77217978e-01
4.79225814e-01 5.90745658e-02 -3.10429275e-01 9.48459625e-01
-1.69755697e-01 4.56745625e-01 -1.03021868e-01 7.44738758e-01
5.50583243e-01 4.74941134e-01 -2.89468229e-01 -2.45748863e-01
1.27654731e+00 -1.22634813e-01 -8.88994157e-01 4.39038903e-01
-1.62474930e-01 -1.17091489e+00 5.88825345e-01 1.03274632e+00
-5.73881567e-01 -5.87003112e-01 -1.06014311e+00 4.73648071e-01
-6.10071182e-01 1.66414708e-01 1.00387506e-01 1.44665647e+00
-6.77116573e-01 6.60932124e-01 -3.03943247e-01 -4.78785723e-01
-1.68470770e-01 4.29911524e-01 -5.83359241e-01 1.09530136e-01
-7.45713890e-01 6.87063813e-01 3.96008268e-02 -2.32538432e-01
-2.92145491e-01 -3.13896418e-01 -3.89276743e-01 -1.15747765e-01
2.85419226e-01 2.63962030e-01 8.22024703e-01 -1.13056695e+00
-1.68886828e+00 8.83727312e-01 7.40063116e-02 -3.66733164e-01
5.64370751e-01 5.68834022e-02 -8.07383955e-01 6.36543870e-01
-8.82704705e-02 -4.17628825e-01 1.28005052e+00 -1.16956937e+00
-1.78541481e-01 -2.05514282e-01 -3.14638346e-01 -7.93523610e-01
-1.87987015e-01 4.35265958e-01 -2.85910040e-01 -4.03688282e-01
8.61614943e-02 -7.42720068e-01 5.30937791e-01 -5.95191061e-01
-3.41047406e-01 8.02364498e-02 1.11811757e+00 -6.42146349e-01
1.08677781e+00 -2.03789592e+00 -5.15881360e-01 5.86182654e-01
-2.24382266e-01 5.10119140e-01 2.95926213e-01 6.87976837e-01
-1.66588295e-02 1.79372638e-01 -2.87998524e-02 7.88337216e-02
-1.71280086e-01 -1.49634734e-01 -3.62153113e-01 6.34469211e-01
-2.53434479e-01 2.08454840e-02 -5.55336177e-01 -6.70105100e-01
5.73172271e-01 5.44484735e-01 2.20824003e-01 9.07714888e-02
3.07027042e-01 2.16394946e-01 -3.76227409e-01 9.16513205e-01
1.31366634e+00 2.59125590e-01 2.17307001e-01 -4.88759726e-01
-4.42136854e-01 -5.01539826e-01 -1.50157499e+00 1.15849936e+00
-1.49504662e-01 7.80733705e-01 -1.04055904e-01 -7.12107241e-01
1.17834640e+00 3.48801047e-01 4.41646039e-01 -5.54084063e-01
3.73537183e-01 4.98212308e-01 -4.82330322e-01 -1.02111852e+00
5.92186034e-01 2.33417585e-01 1.93404347e-01 3.40357155e-01
2.77998708e-02 1.90985516e-01 2.24334672e-01 -1.68243960e-01
1.22385216e+00 -7.82698218e-04 1.49010748e-01 -3.79828990e-01
9.94967043e-01 -2.32485130e-01 8.84457082e-02 8.70943010e-01
-2.92154457e-02 1.00394464e+00 6.36654854e-01 -8.33070278e-02
-1.19293022e+00 -8.81992340e-01 -3.25236708e-01 -2.63685405e-01
3.43666941e-01 -1.51108399e-01 -1.01990843e+00 -2.21870676e-01
-1.50758207e-01 2.98677117e-01 -4.30502832e-01 3.36788654e-01
-4.58475173e-01 -3.40913922e-01 7.98157632e-01 2.90176421e-01
4.67703313e-01 -8.13171506e-01 -8.18930447e-01 2.01217979e-01
2.33450279e-01 -1.57362664e+00 -1.26895756e-01 -1.55613855e-01
-8.03370893e-01 -1.70126379e+00 -6.38385713e-01 -2.27873117e-01
5.28469980e-01 4.06098574e-01 7.29320586e-01 9.31057781e-02
-3.97391051e-01 1.00436127e+00 -6.33707762e-01 -1.92733645e-01
-7.92680204e-01 -4.67422068e-01 6.64541721e-02 4.86029834e-01
7.13862062e-01 -5.15443385e-01 -3.40265393e-01 5.73779047e-01
-1.13867569e+00 -8.29810262e-01 3.71863216e-01 2.37824187e-01
3.34899813e-01 4.41786706e-01 -2.87525177e-01 -9.71803188e-01
5.84177971e-01 -4.37784314e-01 -1.12277281e+00 4.05987829e-01
-2.74753094e-01 -3.30494761e-01 6.64207995e-01 -4.54964608e-01
-1.02237701e+00 1.56448126e-01 1.45875543e-01 -8.65548015e-01
-5.62239408e-01 9.67511535e-02 -2.84221381e-01 -7.04972208e-01
6.11038685e-01 1.04338780e-01 1.55560970e-01 -7.55372286e-01
-4.38743442e-01 1.03315926e+00 8.17888856e-01 -4.75145757e-01
7.50955820e-01 5.18900275e-01 3.54214519e-01 -1.36936915e+00
1.43468127e-01 -8.03077936e-01 -3.83214325e-01 -6.60251260e-01
9.47109878e-01 -4.25380379e-01 -1.30099750e+00 8.86061788e-01
-1.36518896e+00 2.97816455e-01 1.89534679e-01 6.55079067e-01
-1.57821283e-01 8.59898627e-01 -4.33498651e-01 -1.47662902e+00
1.85136925e-02 -1.17697656e+00 7.11357474e-01 5.39684415e-01
8.50478709e-02 -9.09772635e-01 9.93586183e-02 2.26842716e-01
2.95658201e-01 5.56026101e-01 2.34865099e-01 -2.96732783e-01
-6.97007060e-01 -8.28754187e-01 -3.22819978e-01 6.64218843e-01
-4.31240909e-02 5.10527432e-01 -1.18357742e+00 2.13985488e-01
5.36224604e-01 2.91866213e-01 4.33752090e-01 1.05532430e-01
1.09002745e+00 3.28068435e-02 -2.63838112e-01 4.31509197e-01
2.15041471e+00 5.32538712e-01 1.33010268e+00 5.24854839e-01
5.34301937e-01 4.54263955e-01 4.04422641e-01 4.08413202e-01
-5.77749550e-01 8.04717839e-01 4.89246190e-01 4.98190582e-01
6.83756769e-02 -2.34283403e-01 6.80581808e-01 3.61420423e-01
-4.06199396e-01 -5.15270054e-01 -7.90820181e-01 1.02790937e-01
-1.37616348e+00 -1.29297745e+00 -9.87208307e-01 2.63098454e+00
-1.00774914e-01 -7.46189207e-02 8.26502144e-02 4.82014120e-01
1.27322876e+00 -3.21977228e-01 2.34694444e-02 -4.31950748e-01
-3.73227894e-01 4.70441699e-01 1.21093845e+00 2.00689450e-01
-9.53107119e-01 2.10858971e-01 6.29616690e+00 8.82807016e-01
-1.22732449e+00 1.34772196e-01 2.68660963e-01 4.78782743e-01
8.49959848e-04 2.95784939e-02 -7.82011092e-01 1.15099275e+00
9.64791834e-01 4.16862577e-01 2.74794519e-01 7.83708513e-01
-1.06335077e-02 -7.83337355e-01 -8.46965134e-01 1.55354476e+00
5.11082113e-01 -1.18278337e+00 -9.74745005e-02 1.88715562e-01
4.25719470e-01 -6.03648782e-01 9.95390937e-02 -6.35548472e-01
-4.48788762e-01 -7.82345891e-01 6.12667084e-01 7.71254182e-01
7.74053693e-01 -7.41576314e-01 1.22728586e+00 -1.62145928e-01
-9.21986759e-01 -5.70769608e-02 -7.82097220e-01 9.45231467e-02
1.46233559e-01 6.00494564e-01 -4.11279172e-01 4.00253028e-01
7.41123557e-01 4.72004563e-01 -5.70401907e-01 1.38149738e+00
-3.70551795e-02 5.21749616e-01 -4.64179933e-01 -2.23592296e-01
-6.52105883e-02 -8.38032603e-01 7.54841268e-01 1.21029341e+00
9.99257445e-01 -8.51394236e-02 -7.29107916e-01 9.64618802e-01
1.48124158e-01 -7.83263445e-02 -9.01106656e-01 -5.56258261e-02
4.17272925e-01 1.05668318e+00 -9.63674784e-01 -1.43605888e-01
-4.80650455e-01 1.09932244e+00 -7.96917856e-01 1.03520811e-01
-7.70480990e-01 -5.23656487e-01 3.35382313e-01 3.76136690e-01
2.60883659e-01 -6.32348508e-02 -8.92360657e-02 -8.92482102e-01
6.07598163e-02 -6.16239250e-01 1.45171713e-02 -7.41996944e-01
-1.28348565e+00 2.82125354e-01 1.25371173e-01 -1.71168649e+00
1.55535853e-02 -1.10594308e+00 -7.50807881e-01 5.74355364e-01
-8.74172986e-01 -9.08189774e-01 -3.50678295e-01 7.89806843e-01
1.65963992e-01 -4.53982741e-01 6.70161545e-01 4.23757046e-01
-4.02575374e-01 5.79843342e-01 4.19253767e-01 1.94908857e-01
6.12306833e-01 -9.85287786e-01 -7.98383579e-02 1.09655690e+00
1.63209945e-01 7.11661935e-01 9.18484747e-01 -6.84301794e-01
-1.77490735e+00 -1.22218341e-01 5.85005939e-01 -7.59353817e-01
7.35981405e-01 -1.62423566e-01 -5.24152994e-01 2.10904613e-01
3.68190050e-01 -3.34693253e-01 6.48618042e-01 -5.06156325e-01
-2.71321237e-01 -3.48743558e-01 -1.42881680e+00 9.59926378e-03
4.76555318e-01 -7.40697443e-01 -4.00203615e-01 1.19235940e-01
-1.29197538e-01 -8.56109858e-02 -9.52984750e-01 -9.61593017e-02
8.74333739e-01 -1.61018336e+00 8.10776532e-01 3.90083760e-01
3.55494358e-02 -4.61309612e-01 -2.30114564e-01 -2.02451900e-01
3.58122528e-01 -9.15045321e-01 3.68259102e-01 1.61555529e+00
-3.54440659e-02 -6.55433595e-01 8.92671287e-01 5.75977027e-01
4.20608521e-01 5.12538791e-01 -8.42500508e-01 -1.20207810e+00
-6.22945607e-01 -5.15355706e-01 5.45359910e-01 7.91585982e-01
-4.14149053e-02 -3.05268139e-01 -5.91402352e-01 2.78610319e-01
1.13738453e+00 -3.72806489e-01 7.48404324e-01 -1.28368318e+00
-4.30620164e-01 -1.83170319e-01 -9.67978954e-01 -4.82207000e-01
-1.34794116e-01 -2.21498366e-02 -4.45635468e-01 -5.94463050e-01
1.19552612e-01 -1.05210140e-01 -1.59779981e-01 -5.30824184e-01
5.29381812e-01 6.34685934e-01 2.25552842e-01 3.55389714e-01
-3.44798863e-01 -6.02329910e-01 4.42844123e-01 1.32703260e-01
2.61703551e-01 3.13610405e-01 1.03438519e-01 5.55962086e-01
7.65118837e-01 -6.87812567e-01 -2.04314485e-01 -2.24411841e-02
3.73249173e-01 1.85075447e-01 6.17152214e-01 -1.65171456e+00
4.99627739e-01 1.91811353e-01 3.89833838e-01 -5.70545554e-01
2.17562646e-01 -1.54580057e+00 8.50042284e-01 4.44250286e-01
3.84624064e-01 2.70678043e-01 -1.99416026e-01 5.13317168e-01
-4.14617330e-01 -1.14729345e+00 6.69455409e-01 -4.53345299e-01
-7.97083497e-01 -2.80096233e-01 -8.46214533e-01 -6.23853981e-01
9.42544878e-01 -9.32169914e-01 -3.65115196e-01 -2.96253622e-01
-1.74303114e-01 -8.10467958e-01 9.66355503e-01 -9.76432208e-03
6.67223573e-01 -1.03025222e+00 1.52014077e-01 2.53362775e-01
-5.83203062e-02 -8.77652407e-01 3.29261869e-01 7.77665317e-01
-1.22051334e+00 2.28525251e-01 -5.75137734e-01 -4.93391573e-01
-1.55828881e+00 7.78806865e-01 1.98272333e-01 2.78383970e-01
-2.01370195e-01 3.95709902e-01 -5.86947024e-01 4.88118142e-01
1.54519696e-02 -8.53401143e-03 -3.30632359e-01 -4.26668152e-02
7.18112588e-01 9.37928379e-01 6.52344376e-02 -7.51878917e-01
-3.85229915e-01 1.26327765e+00 4.21296090e-01 -1.44060656e-01
9.56138372e-01 -5.94560690e-02 -3.64369005e-01 4.20878619e-01
1.40633047e+00 5.34791589e-01 -8.84521484e-01 5.91809809e-01
2.61929572e-01 -8.02507997e-01 -2.64445722e-01 -3.60484809e-01
-7.86764920e-01 7.61147082e-01 9.60157156e-01 5.16847908e-01
9.31043029e-01 -4.60554421e-01 5.68372607e-01 3.04138035e-01
7.41472602e-01 -1.51237655e+00 -8.53818804e-02 -5.80149889e-02
4.27471995e-01 -1.05947757e+00 9.20225978e-02 -5.42533636e-01
-2.43135184e-01 1.85771859e+00 -5.78136556e-02 -4.75880921e-01
7.40098655e-01 5.34091771e-01 -5.48770539e-02 8.05699304e-02
2.83507407e-01 -7.71610290e-02 -1.75443783e-01 1.00061941e+00
1.84471071e-01 -7.57341459e-02 -5.86933613e-01 4.49927032e-01
2.14959398e-01 7.11132735e-02 1.02086842e+00 7.62252510e-01
-3.01415175e-01 -1.57731354e+00 -1.10104597e+00 2.18893871e-01
-9.30307865e-01 3.60626817e-01 -6.41410589e-01 9.37418222e-01
3.64696801e-01 1.25012052e+00 -4.83926311e-02 -7.17197537e-01
4.00388896e-01 -1.72522187e-01 5.35795212e-01 1.68565288e-01
-6.61684453e-01 -7.39595518e-02 -3.58518846e-02 -6.88118279e-01
-7.75524914e-01 -8.94398570e-01 -6.65188611e-01 -6.53297484e-01
-2.68649638e-01 1.62698984e-01 1.28098297e+00 4.32271898e-01
-1.39880344e-01 -1.45678103e-01 6.44111454e-01 -5.07656038e-01
-1.73659399e-01 -6.02930665e-01 -1.09767449e+00 5.48129737e-01
9.76870582e-02 -7.09166646e-01 -7.72494614e-01 2.06228659e-01] | [12.38310432434082, 0.9978600144386292] |
8585eb57-0942-4e2b-a306-85519838f207 | eulernet-adaptive-feature-interaction | 2304.10711 | null | https://arxiv.org/abs/2304.10711v1 | https://arxiv.org/pdf/2304.10711v1.pdf | EulerNet: Adaptive Feature Interaction Learning via Euler's Formula for CTR Prediction | Learning effective high-order feature interactions is very crucial in the CTR prediction task. However, it is very time-consuming to calculate high-order feature interactions with massive features in online e-commerce platforms. Most existing methods manually design a maximal order and further filter out the useless interactions from them. Although they reduce the high computational costs caused by the exponential growth of high-order feature combinations, they still suffer from the degradation of model capability due to the suboptimal learning of the restricted feature orders. The solution to maintain the model capability and meanwhile keep it efficient is a technical challenge, which has not been adequately addressed. To address this issue, we propose an adaptive feature interaction learning model, named as EulerNet, in which the feature interactions are learned in a complex vector space by conducting space mapping according to Euler's formula. EulerNet converts the exponential powers of feature interactions into simple linear combinations of the modulus and phase of the complex features, making it possible to adaptively learn the high-order feature interactions in an efficient way. Furthermore, EulerNet incorporates the implicit and explicit feature interactions into a unified architecture, which achieves the mutual enhancement and largely boosts the model capabilities. Such a network can be fully learned from data, with no need of pre-designed form or order for feature interactions. Extensive experiments conducted on three public datasets have demonstrated the effectiveness and efficiency of our approach. Our code is available at: https://github.com/RUCAIBox/EulerNet. | ['Zhao Cao', 'Ji-Rong Wen', 'Wayne Xin Zhao', 'Ting Bai', 'Zhen Tian'] | 2023-04-21 | null | null | null | null | ['click-through-rate-prediction'] | ['miscellaneous'] | [-3.20161879e-01 -4.72057790e-01 -1.47586063e-01 -4.33221281e-01
-9.03528109e-02 -5.26083708e-01 3.03161323e-01 -5.10555916e-02
-3.84373844e-01 3.84242088e-01 -3.85522880e-02 -2.51253575e-01
-5.29979646e-01 -8.48443389e-01 -5.63071609e-01 -8.30578089e-01
-1.55534774e-01 1.19082771e-01 2.12739035e-01 -5.30776441e-01
3.14384907e-01 3.33331257e-01 -1.59708107e+00 7.73049593e-02
1.11934769e+00 1.18108189e+00 2.63218433e-01 1.41232327e-01
-1.48678795e-01 2.94338793e-01 -1.34705648e-01 -3.29313040e-01
5.23586690e-01 -2.10659772e-01 -4.50022876e-01 -2.49339372e-01
-9.90447775e-02 -3.30802739e-01 -3.96122992e-01 1.00613415e+00
3.86331141e-01 1.86298192e-01 3.34722072e-01 -1.21310818e+00
-3.52532566e-01 4.93965447e-01 -4.46734458e-01 6.25570193e-02
1.88278884e-01 1.87981918e-01 1.43953156e+00 -1.17361367e+00
3.74771118e-01 1.03432798e+00 6.07301176e-01 1.01653762e-01
-1.01105046e+00 -1.00375879e+00 4.16172594e-01 4.66335356e-01
-1.56645441e+00 2.94698458e-02 1.04642248e+00 -2.37199917e-01
6.70777678e-01 4.38909560e-01 8.94300640e-01 7.24068999e-01
2.03472614e-01 8.00356269e-01 6.78041458e-01 -7.12791830e-02
-1.47618219e-01 1.98547140e-01 5.18720806e-01 6.67001665e-01
1.46579891e-01 1.33203968e-01 -4.30035412e-01 -2.33962551e-01
9.57286775e-01 4.18665826e-01 -2.32213721e-01 -4.22259986e-01
-9.74111736e-01 1.05970275e+00 6.83723629e-01 5.42199552e-01
-2.46555194e-01 -1.82972088e-01 2.28731543e-01 4.79739994e-01
1.68593079e-01 3.94845575e-01 -8.28376174e-01 -4.36706282e-02
-1.83233082e-01 1.77406952e-01 7.31093585e-01 7.89665580e-01
8.94489110e-01 -3.74506712e-01 1.06444463e-01 6.89605832e-01
3.33263546e-01 1.33008137e-01 4.41390574e-01 -5.18781662e-01
3.45991969e-01 9.41073895e-01 5.26244044e-02 -1.45658445e+00
-5.96266866e-01 -8.88753176e-01 -1.11149347e+00 -3.06425631e-01
3.68422896e-01 -1.39451206e-01 -2.51142293e-01 1.61742854e+00
5.86242855e-01 -2.86436453e-02 -2.95356750e-01 1.02459264e+00
6.64831877e-01 6.34440601e-01 -8.81268382e-02 -3.28156471e-01
1.34823895e+00 -9.64008808e-01 -6.41082764e-01 1.89578682e-01
7.19879985e-01 -8.10689807e-01 1.25269842e+00 5.06212473e-01
-6.04041040e-01 -5.73997021e-01 -1.04200339e+00 -1.10206213e-02
-2.82285780e-01 2.45977372e-01 1.23206735e+00 2.08413646e-01
-5.17491579e-01 8.57821226e-01 -7.07904458e-01 2.84802616e-01
2.87313819e-01 8.76677334e-01 -3.08346152e-01 1.80905595e-01
-1.59047461e+00 4.44034517e-01 3.45775902e-01 5.34124434e-01
4.15995009e-02 -9.32062507e-01 -6.26059949e-01 3.16479415e-01
5.14199495e-01 -3.62821847e-01 1.08407509e+00 -9.26741719e-01
-1.53672838e+00 -1.74476892e-01 1.47765353e-01 2.01421212e-02
5.35317361e-01 -3.63128066e-01 -4.86602426e-01 -1.15844920e-01
-3.23410481e-01 2.65265226e-01 7.35623181e-01 -7.79583812e-01
-5.92036247e-01 -2.40555093e-01 2.13906676e-01 4.26953018e-01
-8.33445549e-01 -2.38984093e-01 -4.45349425e-01 -5.00044167e-01
3.25440407e-01 -8.00894499e-01 -4.91504163e-01 5.65248430e-02
-1.31888524e-01 -4.13950831e-01 7.92128444e-01 -4.74737644e-01
1.40298843e+00 -2.26827526e+00 -6.59646243e-02 4.28277373e-01
3.90915006e-01 3.61479849e-01 -2.25267127e-01 4.76402283e-01
3.83024616e-03 1.15549816e-02 1.92678943e-01 3.14802587e-01
-4.97851707e-02 -1.38026968e-01 -1.52748540e-01 2.82943815e-01
9.69899595e-02 8.32295001e-01 -8.72029245e-01 -3.28792840e-01
3.14099669e-01 7.34729290e-01 -8.79053056e-01 1.08389892e-01
-8.30037594e-02 4.50447083e-01 -9.36704099e-01 2.31572166e-01
9.21476245e-01 -6.17509365e-01 3.60914588e-01 -7.31822431e-01
-2.36147642e-01 2.59714097e-01 -1.54837525e+00 1.19643247e+00
-4.40429389e-01 5.86528406e-02 -2.07465485e-01 -1.04482496e+00
9.48019028e-01 -2.37799641e-02 8.72524023e-01 -8.18594337e-01
3.05445015e-01 3.80454600e-01 3.10435086e-01 -3.31575483e-01
2.46855751e-01 3.45380567e-02 -6.01085424e-02 1.06648922e-01
-1.52987525e-01 4.07603562e-01 1.30763128e-01 4.93220575e-02
8.23132634e-01 -1.22808941e-01 4.66123134e-01 -1.94979593e-01
8.38273525e-01 -2.56364375e-01 6.01600945e-01 4.19055343e-01
2.22261220e-01 7.30133280e-02 4.62973177e-01 -7.18182683e-01
-6.42753959e-01 -7.40518093e-01 -3.61083269e-01 9.64313984e-01
4.33468461e-01 -7.79284358e-01 -2.56815910e-01 -7.98345447e-01
7.99627975e-02 2.56348401e-01 -3.84954661e-01 -3.49868715e-01
-7.08888054e-01 -8.65498245e-01 -1.15311988e-01 3.08167100e-01
6.94601655e-01 -7.48991013e-01 -2.31940046e-01 2.99267977e-01
-5.30451946e-02 -7.33932078e-01 -8.89645576e-01 1.83759466e-01
-7.43400574e-01 -1.06834507e+00 -2.17094019e-01 -6.95528865e-01
6.52231932e-01 4.07435775e-01 6.09651446e-01 4.54108924e-01
-2.48531237e-01 -2.33240813e-01 -4.31117058e-01 -7.73466527e-02
2.22376049e-01 3.15587670e-01 4.76001725e-02 2.30082855e-01
3.77691448e-01 -6.64340794e-01 -9.04998600e-01 5.60993850e-01
-8.14270437e-01 9.60284397e-02 9.00360644e-01 1.28191853e+00
4.90334988e-01 3.98993671e-01 5.61317623e-01 -8.40157866e-01
4.90741223e-01 -4.61931467e-01 -8.14392686e-01 -5.57953641e-02
-6.83942437e-01 6.70437515e-02 1.15487266e+00 -7.23731577e-01
-8.81174326e-01 3.72846842e-01 -3.44863266e-01 -1.32196829e-01
3.90617281e-01 7.26960778e-01 -3.08531344e-01 -3.91494662e-01
3.76622617e-01 2.42179051e-01 4.54377122e-02 -5.87525964e-01
2.81120658e-01 5.02213418e-01 -1.06867865e-01 -3.28150541e-01
1.01990712e+00 1.33732155e-01 5.39669171e-02 -6.90973282e-01
-8.55421960e-01 -4.87652272e-01 -5.34177721e-01 4.46597748e-02
1.80723444e-01 -7.28223979e-01 -1.27274668e+00 4.82542723e-01
-8.30181599e-01 1.07254393e-01 -3.01775942e-03 6.15269780e-01
-6.75005987e-02 5.19937456e-01 -6.54282331e-01 -4.57546830e-01
-4.76243436e-01 -1.03002119e+00 5.48781335e-01 3.77136350e-01
9.72764939e-02 -7.65512824e-01 -2.13840201e-01 4.14457318e-04
4.72200722e-01 -2.02289224e-01 9.59334910e-01 -6.79688811e-01
-7.73700237e-01 -3.30244184e-01 -3.01992595e-01 2.28837565e-01
2.30060503e-01 -5.78961931e-02 -4.30015355e-01 -1.74058095e-01
9.15810242e-02 -1.56760309e-02 6.60893857e-01 1.93010151e-01
1.35746181e+00 -4.61764842e-01 -3.05845678e-01 6.85608745e-01
1.31269526e+00 1.91583872e-01 4.02263641e-01 1.30991414e-01
7.67134607e-01 4.40834969e-01 9.20715928e-01 6.91379905e-01
3.06718290e-01 8.29762042e-01 2.63035536e-01 -8.00229087e-02
4.30216670e-01 -3.53580981e-01 1.27175450e-01 1.17653871e+00
-7.74108395e-02 8.12696740e-02 -5.14042795e-01 -4.85886000e-02
-1.93688512e+00 -7.55384088e-01 -3.13716471e-01 2.09641814e+00
8.89575601e-01 1.07235759e-01 7.49697015e-02 1.51139915e-01
4.47572261e-01 -9.51627865e-02 -7.41118848e-01 2.69795060e-02
1.44681245e-01 7.72468969e-02 3.01783562e-01 4.36069757e-01
-1.13623083e+00 7.40255415e-01 5.19199562e+00 1.14095640e+00
-1.26654458e+00 -6.21864200e-02 4.40082639e-01 4.36254703e-02
-3.48371774e-01 1.11842295e-02 -9.68005776e-01 5.16231179e-01
4.75086123e-01 -1.22215599e-01 5.24491191e-01 9.88176167e-01
7.31309801e-02 2.67035097e-01 -8.61080647e-01 1.06804121e+00
-4.38890636e-01 -1.18041229e+00 7.20365793e-02 1.35712892e-01
2.60417879e-01 -2.53101498e-01 9.17513371e-02 6.00877047e-01
-1.65289566e-02 -7.67414033e-01 2.97060639e-01 4.37796324e-01
5.17130613e-01 -8.96077991e-01 8.23480308e-01 3.98437530e-01
-1.47748470e+00 -4.46614087e-01 -4.26778346e-01 -2.21406862e-01
-1.26249745e-01 8.23225498e-01 -4.56846505e-01 6.49112523e-01
5.72263062e-01 8.89830530e-01 -4.20180798e-01 1.02575791e+00
3.37690227e-02 3.19960028e-01 -4.36051130e-01 -4.38543200e-01
8.05725828e-02 -4.70882148e-01 2.65519351e-01 7.18370318e-01
3.94883811e-01 3.30218464e-01 3.10303122e-01 5.23839951e-01
2.26804003e-01 6.44665420e-01 -1.93784058e-01 -1.41909227e-01
4.11363870e-01 1.64620364e+00 -5.30942082e-01 3.55513282e-02
-5.65594196e-01 7.08982766e-01 3.95535529e-01 1.79063320e-01
-8.59172344e-01 -6.33116245e-01 5.48311412e-01 6.28445446e-02
3.45974624e-01 -2.19939291e-01 -1.25577107e-01 -1.39020693e+00
3.80186975e-01 -8.93463254e-01 3.98963720e-01 -8.82354453e-02
-1.46066296e+00 5.20899415e-01 -1.69800371e-01 -1.61371052e+00
-7.51878247e-02 -6.02859199e-01 -4.26152945e-01 6.44794226e-01
-1.26108420e+00 -9.01241958e-01 -2.79433608e-01 7.11942255e-01
2.32268199e-01 4.66089845e-02 6.24405086e-01 6.19893014e-01
-5.73197842e-01 8.38139951e-01 1.49132848e-01 -1.06055640e-01
5.43849587e-01 -7.76555598e-01 2.81484351e-02 2.19141662e-01
-5.68823367e-02 9.73112285e-01 5.62124431e-01 -4.04420227e-01
-1.75848007e+00 -8.07537198e-01 8.98127615e-01 -3.20461839e-02
7.45564759e-01 -6.72693670e-01 -9.39054489e-01 4.02976364e-01
-3.29032600e-01 1.73745498e-01 6.96179211e-01 4.15401399e-01
-2.60825515e-01 -4.97793943e-01 -9.32090640e-01 7.55812407e-01
1.24194741e+00 -1.10624440e-01 -1.52459964e-01 4.94616300e-01
7.76580989e-01 -2.78313220e-01 -1.00102925e+00 5.37365496e-01
8.53914917e-01 -7.90022492e-01 1.00486457e+00 -4.36834812e-01
2.22256035e-01 -2.82740146e-01 2.61797383e-02 -1.11590052e+00
-6.57536089e-01 -6.04024053e-01 -2.36499384e-01 9.75550056e-01
3.46468091e-01 -1.11069107e+00 6.22914553e-01 5.45238733e-01
5.17043546e-02 -1.28400922e+00 -7.13537931e-01 -6.70552850e-01
-1.88643709e-01 -4.03963104e-02 8.04049432e-01 9.43678796e-01
1.28448263e-01 6.15460396e-01 -4.13280576e-01 5.15877903e-02
3.46156925e-01 3.47663581e-01 7.23026037e-01 -1.53926814e+00
-5.42760313e-01 -4.22903806e-01 -3.99503052e-01 -1.25588715e+00
-1.22845508e-01 -8.36835504e-01 -2.98434347e-01 -1.02811193e+00
3.36925417e-01 -7.41765916e-01 -5.52825093e-01 5.04969656e-01
-2.23371223e-01 -1.03369400e-01 1.03826202e-01 3.69208515e-01
-5.53890586e-01 8.52152824e-01 1.65393877e+00 1.60347179e-01
-4.31541383e-01 8.58564004e-02 -1.00058448e+00 6.89976931e-01
8.49001169e-01 -3.08538169e-01 -5.28428376e-01 5.09967431e-02
3.91686350e-01 -2.67256983e-02 1.88376904e-01 -6.76328361e-01
2.37498626e-01 -3.49317938e-01 6.06514871e-01 -6.75390601e-01
2.71202803e-01 -1.04686606e+00 3.34483624e-01 5.93759120e-01
-1.55468300e-01 -6.42460808e-02 -2.19758861e-02 4.36261356e-01
-2.20420972e-01 -7.41839111e-02 4.76813465e-01 7.73506388e-02
-4.77946013e-01 6.49712622e-01 3.62581536e-02 -4.23210382e-01
9.16633308e-01 2.05933467e-01 -1.71087697e-01 -3.23212683e-01
-5.39604723e-01 3.58121932e-01 -4.44134086e-04 4.93169010e-01
4.15553391e-01 -1.51686668e+00 -2.56846249e-01 5.76050341e-01
-1.38352469e-01 -8.11898932e-02 6.48203433e-01 9.59102452e-01
-1.71435535e-01 4.35749620e-01 -1.65385395e-01 -4.72394973e-01
-1.08929276e+00 6.11888945e-01 1.43546477e-01 -6.69460058e-01
-5.81457555e-01 4.39624339e-01 3.20000589e-01 -6.53811991e-01
-7.79867619e-02 -5.56799434e-02 -5.11213481e-01 7.10090920e-02
5.69716513e-01 1.25544071e-01 3.79297957e-02 -4.59489942e-01
-2.89693356e-01 6.87195659e-01 -5.47812104e-01 3.94076318e-01
1.40031195e+00 -1.43672042e-02 -6.52035549e-02 6.73521981e-02
1.51120329e+00 5.76878488e-02 -1.15258956e+00 -3.31120670e-01
-2.84081429e-01 -5.28725266e-01 4.41377470e-03 -5.24377406e-01
-1.26637995e+00 6.12882793e-01 5.61923087e-01 1.03104308e-01
1.22512591e+00 -3.27432364e-01 9.21943784e-01 6.92879736e-01
4.40222174e-01 -9.79866624e-01 1.95739076e-01 6.59875870e-01
7.91566074e-01 -1.13037097e+00 3.98885645e-02 -9.16052580e-01
-2.86441326e-01 1.15978563e+00 7.72736788e-01 -2.76375532e-01
1.17059302e+00 7.01417252e-02 -1.53663859e-01 -1.34528100e-01
-6.31529093e-01 4.95498963e-02 4.74923253e-01 5.93019091e-02
4.56677407e-01 4.63428311e-02 -9.27402556e-01 1.20324099e+00
-2.83669591e-01 -3.18372361e-02 2.32285131e-02 6.00197494e-01
-1.64077446e-01 -1.40405381e+00 9.13017616e-02 7.87702382e-01
-2.72019207e-01 -1.72561780e-01 2.08002180e-02 9.06548262e-01
6.95019439e-02 7.34963119e-01 -1.98303878e-01 -8.85140777e-01
4.47418928e-01 -1.84592515e-01 1.85832754e-01 -3.06447566e-01
-6.61666155e-01 2.61519760e-01 -1.88158676e-01 -6.57254696e-01
3.70603167e-02 -6.68535590e-01 -1.44002438e+00 -2.25761309e-01
-6.14511728e-01 3.65493268e-01 4.50066775e-01 8.14703703e-01
6.07216835e-01 3.29069912e-01 1.16521323e+00 -7.00410604e-01
-1.09606481e+00 -8.34186375e-01 -5.84359765e-01 4.19356734e-01
1.53016433e-01 -8.92217219e-01 -4.68942106e-01 -4.68307346e-01] | [10.101983070373535, 5.4144792556762695] |
772a6237-e27f-41ba-9194-b090c507ce73 | two-birds-one-stone-a-unified-framework-for | 2304.11335 | null | https://arxiv.org/abs/2304.11335v1 | https://arxiv.org/pdf/2304.11335v1.pdf | Two Birds, One Stone: A Unified Framework for Joint Learning of Image and Video Style Transfers | Current arbitrary style transfer models are limited to either image or video domains. In order to achieve satisfying image and video style transfers, two different models are inevitably required with separate training processes on image and video domains, respectively. In this paper, we show that this can be precluded by introducing UniST, a Unified Style Transfer framework for both images and videos. At the core of UniST is a domain interaction transformer (DIT), which first explores context information within the specific domain and then interacts contextualized domain information for joint learning. In particular, DIT enables exploration of temporal information from videos for the image style transfer task and meanwhile allows rich appearance texture from images for video style transfer, thus leading to mutual benefits. Considering heavy computation of traditional multi-head self-attention, we present a simple yet effective axial multi-head self-attention (AMSA) for DIT, which improves computational efficiency while maintains style transfer performance. To verify the effectiveness of UniST, we conduct extensive experiments on both image and video style transfer tasks and show that UniST performs favorably against state-of-the-art approaches on both tasks. Our code and results will be released. | ['Libo Zhang', 'Heng Fan', 'Bohai Gu'] | 2023-04-22 | null | null | null | null | ['style-transfer', 'video-style-transfer'] | ['computer-vision', 'computer-vision'] | [ 3.71228814e-01 -1.86286584e-01 -1.18951261e-01 -3.12745124e-01
-6.87550306e-01 -5.16016781e-01 6.38917685e-01 -6.30806506e-01
-2.79621661e-01 8.20929229e-01 8.13897625e-02 -1.59202158e-01
1.99106008e-01 -6.01008654e-01 -8.99986923e-01 -7.79180646e-01
4.79994953e-01 1.03463091e-01 3.63042831e-01 -3.43764424e-02
2.22080275e-02 1.80890992e-01 -1.19811833e+00 4.58843023e-01
1.00417411e+00 1.13283682e+00 6.01903021e-01 4.32304263e-01
-2.99932152e-01 7.60645449e-01 -3.42095822e-01 -5.51524997e-01
1.69245183e-01 -6.36728466e-01 -9.44332719e-01 3.98487061e-01
3.81113648e-01 -7.48643279e-01 -5.83294809e-01 1.04442263e+00
4.43355590e-01 6.62476495e-02 7.60782957e-01 -1.25802767e+00
-8.56066346e-01 2.77482897e-01 -7.21574128e-01 6.88044578e-02
2.51379937e-01 2.71464199e-01 7.67238140e-01 -9.03430164e-01
6.64706349e-01 1.38761795e+00 2.55381197e-01 6.48144007e-01
-1.31884336e+00 -8.25133264e-01 4.29003358e-01 2.00774133e-01
-1.10538447e+00 -4.41850334e-01 1.02567697e+00 -2.96523601e-01
4.92614269e-01 1.64371468e-02 5.68661809e-01 1.31641090e+00
4.61546518e-02 1.19657373e+00 1.11008191e+00 -2.62135565e-01
4.82993834e-02 2.20758229e-01 -2.76842773e-01 5.07814109e-01
-2.15845793e-01 5.80189563e-02 -6.14310265e-01 2.32415780e-01
1.38320959e+00 4.51418161e-02 -4.27341998e-01 -4.74198133e-01
-1.19993663e+00 6.69331670e-01 5.21996498e-01 2.13059992e-01
-2.77219951e-01 3.78775522e-02 5.25687754e-01 4.76732999e-01
6.93095088e-01 -8.77949446e-02 -2.52546459e-01 -9.93062109e-02
-7.15284944e-01 1.11082830e-01 3.43607366e-01 1.30376923e+00
8.61064374e-01 2.30053626e-02 -4.51271445e-01 8.90169978e-01
1.13545716e-01 4.17900383e-01 3.11225057e-01 -1.04582965e+00
5.11927128e-01 3.32498223e-01 3.42531092e-02 -7.31484354e-01
2.70039141e-01 -2.96837509e-01 -9.42985892e-01 5.58802411e-02
3.07353586e-01 -1.66413054e-01 -8.02429855e-01 1.98824251e+00
2.44627565e-01 4.04853255e-01 -8.56016427e-02 9.70497310e-01
7.20543981e-01 7.57133842e-01 2.35023171e-01 -2.57598162e-01
1.41442811e+00 -1.23421097e+00 -7.52287090e-01 -2.56822050e-01
4.37771857e-01 -7.50436723e-01 1.32830691e+00 1.88462526e-01
-1.34496605e+00 -7.67809093e-01 -7.92001367e-01 -2.87466973e-01
8.43251944e-02 -2.21670326e-03 4.87952381e-01 1.14442125e-01
-1.11725771e+00 4.56050724e-01 -7.25829124e-01 -4.10303622e-01
5.44290185e-01 3.16125691e-01 -4.50425357e-01 -1.90510526e-01
-1.17079401e+00 6.07976913e-01 2.83276677e-01 -1.03863090e-01
-8.12522233e-01 -6.83829427e-01 -8.86020899e-01 4.09637839e-02
2.54051983e-01 -9.16171074e-01 1.25163007e+00 -1.50459981e+00
-1.75384259e+00 9.11670387e-01 -3.89795303e-01 -1.92600101e-01
6.40432358e-01 -4.97158200e-01 -2.51918852e-01 4.45367604e-01
1.05036199e-01 1.00663841e+00 1.12692249e+00 -1.27201378e+00
-6.54974759e-01 -1.81454778e-01 2.13608757e-01 5.48531771e-01
-7.07983613e-01 4.41329703e-02 -8.58780861e-01 -9.32512760e-01
-3.52747202e-01 -8.50005627e-01 1.78588152e-01 2.76873380e-01
-8.62273052e-02 -2.15963587e-01 1.14322317e+00 -6.77526236e-01
1.20019710e+00 -2.35829043e+00 6.28730893e-01 -2.33961612e-01
1.69873193e-01 4.07649577e-01 -3.08771551e-01 1.43132851e-01
-2.84666326e-02 -1.97689787e-01 -2.18773916e-01 -4.87554550e-01
-1.46509469e-01 3.14241260e-01 -3.41510266e-01 1.52233079e-01
3.01773906e-01 1.11168063e+00 -8.05018246e-01 -5.99885881e-01
2.54396409e-01 4.63089764e-01 -8.33021820e-01 6.45047367e-01
-2.32058451e-01 8.54483485e-01 -7.17266202e-01 3.77790868e-01
7.54364789e-01 -4.92756814e-01 1.08856186e-01 -4.72053975e-01
1.10211082e-01 5.68950661e-02 -7.52335012e-01 2.00560641e+00
-5.85053325e-01 5.10678470e-01 2.76403427e-01 -9.65538919e-01
7.12781131e-01 3.27105045e-01 2.68769622e-01 -8.43018770e-01
1.41606301e-01 1.64634213e-01 -3.75843287e-01 -4.52530056e-01
2.89951861e-01 -3.64351392e-01 1.61508635e-01 3.47483844e-01
3.52927059e-01 3.99801065e-04 -9.31869671e-02 2.33580753e-01
6.98552966e-01 5.12200773e-01 1.08297713e-01 -4.46537763e-01
6.70437157e-01 -4.88222957e-01 6.07287169e-01 3.44111800e-01
-3.46013695e-01 6.76023543e-01 3.60583425e-01 -2.38765597e-01
-1.04238737e+00 -1.22237289e+00 1.07517846e-01 1.20854890e+00
4.87424165e-01 -1.52029306e-01 -7.78801978e-01 -9.32589293e-01
-1.97008967e-01 4.31345850e-01 -5.38409710e-01 -1.54303178e-01
-5.88073552e-01 -2.07085282e-01 1.32853821e-01 6.68435097e-01
8.61902654e-01 -1.17827666e+00 -2.19926551e-01 1.57224000e-01
-3.55081409e-01 -1.25589263e+00 -9.90664721e-01 -2.64296532e-01
-8.55901361e-01 -7.15273082e-01 -1.20238423e+00 -9.66977775e-01
5.13450921e-01 5.35820544e-01 1.08187091e+00 -4.95684817e-02
1.22238211e-01 3.91182631e-01 -3.84408087e-01 -1.44273248e-02
-1.89704046e-01 1.25909567e-01 -9.30672809e-02 2.19442487e-01
1.84553981e-01 -8.65241885e-01 -8.16936970e-01 4.16381747e-01
-1.07090271e+00 4.20937926e-01 6.97767556e-01 1.13942981e+00
2.57597506e-01 -3.15783173e-01 5.37361443e-01 -7.52239645e-01
2.25823328e-01 -3.76888394e-01 -4.12101805e-01 3.32695425e-01
-1.06881104e-01 -2.31840964e-02 5.10633469e-01 -4.71594900e-01
-1.47086108e+00 -8.87975842e-02 -1.41696781e-01 -8.57237101e-01
-2.13998869e-01 1.86093748e-01 -4.30246294e-01 -1.20359652e-01
2.47363538e-01 5.11785388e-01 2.12943614e-01 -4.80491936e-01
3.45902652e-01 6.76587522e-01 4.15212750e-01 -8.22132885e-01
6.45728648e-01 4.67879474e-01 -3.49729747e-01 -8.02513719e-01
-8.60571802e-01 -1.17267162e-01 -6.28823400e-01 -2.28433818e-01
9.91857290e-01 -1.21974599e+00 -5.28377056e-01 7.42917001e-01
-1.23664737e+00 -5.84070861e-01 -9.57759912e-04 3.02515954e-01
-7.11170495e-01 5.48755765e-01 -9.11060154e-01 -4.38094586e-01
-2.03542620e-01 -1.35292137e+00 1.24120474e+00 1.58688188e-01
-5.49747795e-02 -1.28779423e+00 -2.74463922e-01 3.09382915e-01
4.30963069e-01 -1.50971441e-02 8.11339796e-01 -1.50012851e-01
-6.73028946e-01 3.69251460e-01 -7.08754539e-01 5.27455807e-01
3.30026031e-01 -3.43457431e-01 -9.97403264e-01 -5.34059346e-01
1.24287724e-01 -4.81042385e-01 8.15043449e-01 3.06087703e-01
1.41925681e+00 -2.47751921e-01 -2.98930079e-01 7.89900005e-01
1.29780877e+00 3.09621960e-01 7.69543588e-01 2.70567924e-01
9.31075990e-01 5.87029278e-01 5.92814684e-01 2.97629297e-01
4.04123992e-01 9.08926964e-01 1.20880924e-01 -4.11375135e-01
-3.00055981e-01 -1.87913120e-01 5.85816562e-01 7.93068767e-01
-5.71445702e-03 -2.85488307e-01 -3.40658724e-01 4.92165327e-01
-1.85673547e+00 -7.85421014e-01 2.68018872e-01 1.92235303e+00
9.54203427e-01 1.27423060e-04 2.20505372e-01 -1.77080840e-01
7.80030787e-01 1.70365483e-01 -6.54421687e-01 -1.17112167e-01
-6.70364173e-03 4.91427965e-02 9.71242189e-02 4.10961658e-01
-1.13075745e+00 1.06319702e+00 6.02811098e+00 1.16831398e+00
-1.12794995e+00 1.55440256e-01 9.02884960e-01 -6.28568381e-02
-3.70772004e-01 -1.43274873e-01 -3.99724036e-01 7.31626332e-01
4.14687127e-01 -1.41478360e-01 4.58586812e-01 6.85430765e-01
3.16704921e-02 1.65769890e-01 -1.20741808e+00 1.12513828e+00
-1.28004849e-01 -1.18464851e+00 4.76244450e-01 4.92826663e-02
7.89915323e-01 -3.41077894e-01 2.03001216e-01 3.38021457e-01
7.46139884e-02 -6.01479352e-01 7.28649914e-01 2.81346858e-01
1.16905880e+00 -7.29571998e-01 4.28328037e-01 1.54105099e-02
-1.18980360e+00 -2.70837098e-02 -2.87422478e-01 -7.21354224e-03
3.58626157e-01 4.63972330e-01 1.62617918e-02 9.26372647e-01
6.51943862e-01 1.14885151e+00 -2.17878744e-01 6.16106629e-01
-1.05718195e-01 4.21843261e-01 -1.57460757e-02 4.76203948e-01
4.35077757e-01 -4.04897213e-01 4.94661182e-01 1.10337603e+00
4.41860139e-01 1.87816501e-01 1.79734498e-01 8.59758735e-01
-1.10613041e-01 -1.00135818e-01 -7.04538405e-01 1.50651917e-01
3.88624638e-01 1.05437517e+00 -4.40776080e-01 -6.03998125e-01
-7.43335485e-01 1.53189027e+00 4.42363441e-01 7.11709976e-01
-9.55303669e-01 -2.03265831e-01 8.83091569e-01 5.07801995e-02
4.33985144e-01 -1.91617712e-01 -2.54721969e-01 -1.51738620e+00
3.07907630e-02 -7.66535759e-01 2.27163121e-01 -7.45553374e-01
-1.41420186e+00 7.75264144e-01 -3.18996981e-02 -1.47374630e+00
-8.38259794e-03 -6.07302964e-01 -6.50528431e-01 8.98482978e-01
-1.54357803e+00 -1.47188282e+00 -4.00074184e-01 9.28399920e-01
8.47198069e-01 -9.14467201e-02 4.04478073e-01 5.12037456e-01
-5.05295992e-01 8.14929664e-01 5.23081534e-02 1.59951975e-04
1.12253237e+00 -9.84548986e-01 2.46406332e-01 7.28541791e-01
-3.14012825e-01 4.45695281e-01 5.06451130e-01 -5.19428968e-01
-1.33266759e+00 -1.19154954e+00 4.28706795e-01 -3.16107929e-01
6.28395319e-01 -3.99938077e-01 -1.06594253e+00 7.24053442e-01
5.70260584e-01 -2.04275712e-01 4.89512354e-01 -1.14150852e-01
-2.87388831e-01 -1.19635083e-01 -8.62502396e-01 8.48871410e-01
1.32352376e+00 -6.14700317e-01 -4.71554011e-01 1.23920195e-01
9.69097376e-01 -2.44764760e-01 -8.10273588e-01 3.97424906e-01
4.82546687e-01 -1.10216475e+00 9.24787521e-01 -4.88410145e-01
8.17630470e-01 -2.81363070e-01 -1.93612471e-01 -1.18556774e+00
-4.81021821e-01 -6.67175531e-01 -1.39413550e-01 1.48071778e+00
-1.72523484e-02 -5.09553134e-01 5.93962967e-01 5.26629269e-01
-1.72904089e-01 -5.68374932e-01 -6.83344245e-01 -8.46964121e-01
2.65425682e-01 -9.15522724e-02 4.08356339e-01 8.53383243e-01
-1.64669335e-01 5.01652300e-01 -7.46392369e-01 -2.00908810e-01
6.18948579e-01 2.33069822e-01 7.08169281e-01 -7.54801571e-01
-4.51229006e-01 -4.65510190e-01 -1.15606584e-01 -1.51031804e+00
3.34362984e-01 -5.74421942e-01 -1.34915769e-01 -1.24730146e+00
3.93240213e-01 -2.14891970e-01 -3.31396908e-01 3.08089226e-01
-4.13110793e-01 5.22141039e-01 3.39394957e-01 2.72535920e-01
-6.83149576e-01 1.04807997e+00 1.69172990e+00 4.67209965e-02
-8.29332247e-02 -2.07266331e-01 -7.06737220e-01 6.61243558e-01
5.94336689e-01 7.92032778e-02 -7.28456020e-01 -8.23296070e-01
-4.07606602e-01 2.59639442e-01 3.41639668e-01 -7.43118942e-01
-1.22380324e-01 -1.34738877e-01 4.36844379e-01 -3.08292001e-01
3.37374151e-01 -7.38247633e-01 3.33585939e-03 2.15443969e-01
-2.41957769e-01 -9.68572274e-02 2.80995220e-01 6.64903104e-01
-5.26635885e-01 2.77411073e-01 1.00067806e+00 -2.33963076e-02
-8.92046392e-01 5.90667784e-01 -1.28157333e-01 -1.63842335e-01
1.13680625e+00 -1.93789527e-01 -6.41552359e-02 -4.82112855e-01
-7.76322901e-01 3.04231167e-01 7.32106090e-01 5.20496249e-01
5.31662047e-01 -1.74254322e+00 -5.91479361e-01 3.10569286e-01
3.51191163e-02 2.85997801e-03 7.05583811e-01 7.93326914e-01
-2.62770593e-01 3.05473089e-01 -5.72524726e-01 -7.20942974e-01
-1.22129714e+00 8.63425732e-01 3.52364480e-02 -1.67742923e-01
-6.14631832e-01 7.71135330e-01 1.34081793e+00 -1.01050638e-01
8.72393996e-02 -4.44634147e-02 1.78430062e-02 -9.98408794e-02
6.60902619e-01 1.44599393e-01 -3.84455442e-01 -4.80460018e-01
-1.83282033e-01 7.14453280e-01 -4.00561392e-01 -1.18118405e-01
1.06570911e+00 -6.16621435e-01 -4.12361622e-02 2.41417691e-01
1.31591988e+00 -4.63805914e-01 -1.91725802e+00 -5.02489150e-01
-4.33752507e-01 -6.33273780e-01 -1.18435912e-01 -6.27570987e-01
-1.17883301e+00 1.17419660e+00 4.00089473e-01 -1.14570133e-01
1.62976098e+00 -9.53675583e-02 1.08957517e+00 -1.48401976e-01
4.34085697e-01 -8.68684888e-01 4.51649696e-01 4.83053774e-01
9.75111306e-01 -1.20533550e+00 -3.92877549e-01 -5.93885601e-01
-9.09456253e-01 8.42333198e-01 9.32423294e-01 -1.45708248e-01
5.22977352e-01 4.39777767e-04 -8.82878751e-02 2.52726495e-01
-6.86263204e-01 1.18498234e-02 3.96715552e-01 4.72878277e-01
5.36182463e-01 -1.81411162e-01 -1.64155066e-01 5.24209321e-01
3.45462024e-01 1.77719772e-01 4.64008413e-02 8.46543610e-01
-2.11220995e-01 -1.37223029e+00 -1.86620235e-01 2.04606429e-01
-3.89662594e-01 -1.05603904e-01 -1.43655822e-01 6.39256656e-01
-9.93232578e-02 6.02922380e-01 6.33287802e-02 -3.45694810e-01
8.42173472e-02 -5.96927740e-02 6.72572732e-01 -3.38047117e-01
-2.24009842e-01 4.65421379e-01 -1.59312606e-01 -6.88205779e-01
-4.72776771e-01 -5.34856379e-01 -7.66489744e-01 -4.24597383e-01
1.34752048e-02 -1.61555912e-02 1.20425157e-01 1.05050540e+00
5.12786031e-01 6.47205353e-01 6.86734319e-01 -1.17820621e+00
-2.69458145e-01 -9.17456508e-01 -5.94835520e-01 6.03636682e-01
2.86809146e-01 -8.73139620e-01 -2.25740601e-03 3.54030937e-01] | [11.317412376403809, -0.7088479399681091] |
a1f80398-b779-4c84-9e6d-b5d51ce132e5 | depth-estimation-in-nighttime-using-stereo | 1909.13701 | null | https://arxiv.org/abs/1909.13701v2 | https://arxiv.org/pdf/1909.13701v2.pdf | Nighttime Stereo Depth Estimation using Joint Translation-Stereo Learning: Light Effects and Uninformative Regions | Nighttime stereo depth estimation is still challenging, as assumptions associated with daytime lighting conditions do not hold any longer. Nighttime is not only about low-light and dense noise, but also about glow/glare, flares, non-uniform distribution of light, etc. One of the possible solutions is to train a network on night stereo images in a fully supervised manner. However, to obtain proper disparity ground-truths that are dense, independent from glare/glow, and have sufficiently far depth ranges is extremely intractable. To address the problem, we introduce a network joining day/night translation and stereo. In training the network, our method does not require ground-truth disparities of the night images, or paired day/night images. We utilize a translation network that can render realistic night stereo images from day stereo images. We then train a stereo network on the rendered night stereo images using the available disparity supervision from the corresponding day stereo images, and simultaneously also train the day/night translation network. We handle the fake depth problem, which occurs due to the unsupervised/unpaired translation, for light effects (e.g., glow/glare) and uninformative regions (e.g., low-light and saturated regions), by adding structure-preservation and weighted-smoothness constraints. Our experiments show that our method outperforms the baseline methods on night images. | ['Lionel Heng', 'Loong-Fah Cheong', 'Robby T. Tan', 'Aashish Sharma'] | 2019-09-30 | null | null | null | null | ['stereo-depth-estimation'] | ['computer-vision'] | [ 3.06441605e-01 -1.28827721e-01 4.07110810e-01 -6.02016032e-01
-4.96042997e-01 -6.84343755e-01 4.62303311e-01 -6.53547585e-01
-2.49716058e-01 1.11222064e+00 2.20345676e-01 -2.11090177e-01
3.28826249e-01 -9.01757777e-01 -8.12704921e-01 -1.01449084e+00
3.92775834e-01 1.77106142e-01 2.43793219e-01 -4.50876683e-01
1.44952551e-01 3.72100085e-01 -1.95542765e+00 3.62949856e-02
1.14858580e+00 7.04554021e-01 4.63169068e-01 6.87689364e-01
-3.16977757e-03 7.59692371e-01 -3.61917764e-01 -1.63839176e-01
7.80836701e-01 -5.05802691e-01 -2.23722562e-01 4.67431128e-01
1.05505025e+00 -8.65221500e-01 -4.46032912e-01 1.39637804e+00
4.57482576e-01 2.23222628e-01 1.90580055e-01 -1.06953716e+00
-3.08408886e-01 -2.66043633e-01 -7.67417312e-01 6.77240193e-02
3.28709483e-01 3.92921895e-01 3.84986132e-01 -7.49733031e-01
3.76398295e-01 1.01546013e+00 4.57577795e-01 4.28091526e-01
-1.12621307e+00 -6.83700979e-01 9.87798497e-02 7.79698566e-02
-1.34289443e+00 -6.69846356e-01 8.75873387e-01 -2.74643481e-01
6.07925415e-01 3.12086761e-01 6.25907302e-01 9.75530982e-01
2.10621938e-01 2.90544689e-01 1.50782406e+00 -3.52988869e-01
1.98664457e-01 -5.50036989e-02 -3.15126896e-01 4.35558826e-01
2.15364069e-01 5.01025558e-01 -4.84914482e-01 1.97573811e-01
8.12245131e-01 2.54063547e-01 -7.55351722e-01 -3.17933932e-02
-8.85358453e-01 4.94845182e-01 6.31682456e-01 -2.50216037e-01
-9.41539258e-02 -2.38820881e-01 -2.48572975e-01 2.79637545e-01
6.67243063e-01 2.24843007e-02 -4.85979348e-01 9.76222008e-02
-8.91859472e-01 2.61355072e-01 5.42486608e-01 9.67281938e-01
1.46454334e+00 9.17048231e-02 2.18898788e-01 6.14186525e-01
3.01122397e-01 9.67650294e-01 2.59191006e-01 -1.24091518e+00
5.70377529e-01 1.70367405e-01 3.47028464e-01 -7.11464942e-01
-3.93611401e-01 -1.17184497e-01 -1.03375661e+00 6.98480904e-01
4.58268195e-01 -3.05949211e-01 -1.23891664e+00 1.67342579e+00
2.49245822e-01 3.25742781e-01 1.67709485e-01 1.37608671e+00
6.89318299e-01 6.20117486e-01 -6.94674373e-01 -4.23796624e-01
9.18481946e-01 -9.30270553e-01 -8.65557075e-01 -8.93100142e-01
3.05017501e-01 -9.29487944e-01 1.22948170e+00 2.33079523e-01
-1.01307380e+00 -3.28187406e-01 -9.29510891e-01 -4.51261282e-01
-3.34510863e-01 -2.03984961e-01 6.06636345e-01 1.79481283e-01
-1.23199141e+00 3.62255961e-01 -7.34806895e-01 -3.19044441e-01
-6.46065325e-02 6.02967339e-03 -2.03376710e-01 -3.80442619e-01
-1.10838199e+00 6.66869700e-01 1.62362263e-01 5.81886530e-01
-8.21649432e-01 -5.26529968e-01 -1.09624481e+00 -3.15265894e-01
1.48620635e-01 -6.57093883e-01 9.85105932e-01 -1.27645588e+00
-1.35168159e+00 1.08385670e+00 -4.21578735e-01 -1.63262621e-01
4.94260907e-01 -6.03207108e-03 -4.11791652e-01 -1.63055182e-01
2.21744120e-01 5.73872685e-01 6.84004426e-01 -1.63423097e+00
-7.40984678e-01 -5.44182420e-01 4.36655849e-01 6.80171072e-01
2.62971878e-01 -3.90399843e-01 -5.92047513e-01 -7.04128891e-02
4.60438520e-01 -9.10089910e-01 -1.43406972e-01 2.86499500e-01
-4.17680949e-01 7.27714181e-01 8.64991307e-01 -6.07276559e-01
5.71081579e-01 -2.35137773e+00 -3.98104638e-01 -1.94830656e-01
-2.64602480e-03 -2.32450828e-01 -5.85165359e-02 1.26812875e-01
-4.18350250e-02 -4.77531552e-01 -5.36760032e-01 -6.12843454e-01
-3.15522432e-01 9.03046250e-01 -5.86824536e-01 6.40086174e-01
-5.55787198e-02 5.53333342e-01 -9.15571928e-01 -3.56067896e-01
6.43808305e-01 6.04372382e-01 -4.53183413e-01 3.56899947e-01
-2.02487677e-01 8.46572638e-01 1.27362479e-02 5.78264177e-01
1.31911302e+00 1.49614751e-01 -8.65357816e-02 -1.77587420e-01
-6.94263101e-01 3.94443303e-01 -1.12411737e+00 1.70143425e+00
-6.99037790e-01 1.04719603e+00 2.37074316e-01 -2.42965937e-01
8.73999357e-01 -7.07964599e-02 -1.59263104e-01 -1.27457404e+00
2.60398742e-02 3.99060935e-01 -2.76298046e-01 -6.78497791e-01
4.35720682e-01 -4.78310615e-01 4.87795085e-01 8.45478401e-02
-5.83717763e-01 -4.92661327e-01 -1.15679629e-01 -2.01430112e-01
7.38622010e-01 3.06885302e-01 -2.14191675e-01 -1.78455234e-01
4.05672938e-01 -2.01970875e-01 9.03961360e-01 5.35251737e-01
-2.98318826e-02 1.31682122e+00 7.17403367e-02 -6.21441126e-01
-1.09589088e+00 -1.18373048e+00 -2.12385923e-01 7.41144061e-01
5.81274450e-01 1.84580773e-01 -4.21438009e-01 2.32264139e-02
-3.02998006e-01 5.74898481e-01 -4.37035024e-01 -9.37396660e-02
-4.59557503e-01 -8.86111081e-01 4.10161987e-02 3.33532631e-01
1.10294557e+00 -9.12046850e-01 -5.08575737e-01 -3.57686705e-03
-5.00975072e-01 -1.28442454e+00 -3.01069558e-01 2.98541844e-01
-8.48825991e-01 -9.88156796e-01 -5.40195107e-01 -6.01474226e-01
8.46051276e-01 8.82446945e-01 1.27981246e+00 -5.45407785e-03
-2.04088315e-01 -2.16140047e-01 -7.16065019e-02 -3.15105945e-01
-8.72295070e-03 -5.51563323e-01 3.12009845e-02 7.77428374e-02
1.48898900e-01 -8.55319560e-01 -1.03236544e+00 4.20108408e-01
-1.10233033e+00 4.50764328e-01 5.01183383e-02 7.46409535e-01
7.05745280e-01 2.72855282e-01 -3.17718774e-01 -7.67791510e-01
-8.10186639e-02 1.10911660e-01 -1.12271583e+00 -1.55342206e-01
-4.13890719e-01 -2.77330697e-01 6.14562988e-01 8.38536844e-02
-1.43018115e+00 1.18401468e-01 -1.68014586e-01 -5.71083724e-01
-3.31631035e-01 -4.15849686e-02 -3.35914731e-01 -3.11579168e-01
8.81341398e-01 3.26130003e-01 -1.45154536e-01 -4.11978096e-01
6.24346770e-02 6.69230998e-01 8.92123282e-01 -2.84494430e-01
1.20468080e+00 1.34222996e+00 9.18809623e-02 -9.71408904e-01
-1.37306571e+00 -3.84171367e-01 -3.89328808e-01 -9.04437155e-02
9.03860986e-01 -1.36577141e+00 -3.32094133e-01 6.82368219e-01
-1.10681069e+00 -7.56653190e-01 -2.65375882e-01 4.63840336e-01
-4.83941466e-01 3.78229618e-01 -3.79812032e-01 -7.27458835e-01
9.89374220e-02 -1.02271152e+00 1.31879091e+00 5.84457517e-01
4.52590406e-01 -1.13254309e+00 2.34308303e-03 2.83375531e-01
3.33395600e-01 4.02260333e-01 4.17571038e-01 7.03506768e-01
-1.00571477e+00 8.56350586e-02 -5.77118456e-01 5.27271092e-01
5.87150872e-01 2.67756097e-02 -1.63006163e+00 -2.79930860e-01
4.79297936e-01 -2.81157643e-01 8.76513839e-01 4.63994890e-01
1.08107281e+00 -2.36414909e-01 1.62232399e-01 1.46616793e+00
1.77028394e+00 5.89079671e-02 9.85315740e-01 7.04507828e-01
9.96608555e-01 6.81888223e-01 7.33892381e-01 4.43729818e-01
5.83275557e-01 4.96271670e-01 9.16970015e-01 -7.55683482e-01
-8.41568857e-02 -1.00428835e-01 2.39112094e-01 3.80499542e-01
-2.16428444e-01 -1.80791482e-01 -6.34895921e-01 3.15298259e-01
-1.70516515e+00 -7.85201252e-01 -4.96300220e-01 2.53137851e+00
7.31233776e-01 6.14343323e-02 -4.93884623e-01 -3.71209309e-02
4.63745385e-01 3.22409511e-01 -6.12547219e-01 -2.02063173e-01
-8.02933156e-01 -1.12764329e-01 1.04350471e+00 9.41329658e-01
-8.25890481e-01 9.43991780e-01 5.66280603e+00 7.36331269e-02
-1.31792033e+00 -9.75257829e-02 5.59801400e-01 -5.58848456e-02
-5.98789334e-01 3.10679227e-01 -6.80326819e-01 5.16779661e-01
3.64570349e-01 9.73947346e-03 1.01847923e+00 3.93559992e-01
7.63679564e-01 -4.75098312e-01 -8.12528372e-01 1.28267753e+00
3.53238761e-01 -9.29904997e-01 -2.92888224e-01 1.70682102e-01
1.18499458e+00 4.66422021e-01 -1.25613764e-01 2.35932181e-03
3.16505820e-01 -6.19009078e-01 4.93406892e-01 5.86879671e-01
8.11947882e-01 -4.03296709e-01 6.91117108e-01 6.26829922e-01
-9.02232230e-01 2.25736335e-01 -6.77246332e-01 -4.52443600e-01
2.44277090e-01 8.87141407e-01 -1.92221940e-01 4.53571767e-01
9.84190822e-01 9.56986010e-01 -4.51941103e-01 1.00657010e+00
-7.28877127e-01 5.70707805e-02 -5.59490144e-01 7.66318023e-01
1.52429700e-01 -8.25718462e-01 2.20900372e-01 7.38584995e-01
3.47828001e-01 8.57936591e-02 -5.17251864e-02 8.48552704e-01
1.51446342e-01 -5.41752934e-01 -9.04058874e-01 6.72998846e-01
2.60430425e-02 1.30164075e+00 -2.32838601e-01 -2.28869289e-01
-6.40372694e-01 1.27457440e+00 -2.17287913e-01 8.33959579e-01
-8.36479485e-01 -8.57732594e-02 9.78242934e-01 2.95090050e-01
-2.79029250e-01 -2.69709766e-01 -3.19389880e-01 -1.57713819e+00
3.46385926e-01 -6.38972521e-01 1.74133003e-01 -1.59735811e+00
-1.20826590e+00 4.73053455e-01 -3.16556334e-01 -1.56691420e+00
-9.54634100e-02 -5.70824385e-01 -7.44080603e-01 1.18769050e+00
-2.35895038e+00 -8.57637346e-01 -1.12455201e+00 9.60465133e-01
4.95790094e-01 4.91758347e-01 5.44776738e-01 5.90026677e-01
-3.80574614e-01 -6.39402196e-02 3.75140101e-01 -3.36319894e-01
9.96534944e-01 -1.34178591e+00 2.85145879e-01 1.04792523e+00
-3.86248380e-01 2.32963994e-01 9.27710116e-01 -2.61649847e-01
-1.30401158e+00 -1.30661893e+00 8.74242902e-01 -1.93520755e-01
5.30178666e-01 -4.56239641e-01 -1.04346490e+00 7.59235144e-01
-3.41142491e-02 2.35616565e-01 -2.27319859e-02 -2.18541026e-01
-8.74258950e-02 -4.57181782e-01 -9.84111607e-01 8.27749908e-01
1.13002610e+00 -7.97416091e-01 -3.12744260e-01 7.82168448e-01
7.83720911e-01 -8.81494641e-01 2.36388240e-02 4.18381751e-01
2.63846546e-01 -1.70285463e+00 8.56367409e-01 1.50378987e-01
4.96068001e-01 -7.37705290e-01 -4.24808860e-01 -1.26097405e+00
-4.94140312e-02 -5.78169346e-01 3.11479360e-01 1.16500056e+00
2.20288649e-01 -9.12143350e-01 8.51525307e-01 6.28388643e-01
-4.93580222e-01 5.20639420e-02 -8.99549246e-01 -7.74098814e-01
-1.78506106e-01 -2.46496707e-01 5.45339346e-01 1.11249125e+00
-8.16267133e-01 1.83187410e-01 -4.77321982e-01 6.01795912e-01
8.91842127e-01 3.66499066e-01 9.23848569e-01 -1.07597685e+00
-2.58214563e-01 5.69810383e-02 -1.83199227e-01 -1.03884947e+00
6.24278523e-02 -3.58473063e-01 5.63504577e-01 -1.70108974e+00
-2.78914701e-02 -1.81760401e-01 7.78844431e-02 4.69160646e-01
-2.17133877e-03 6.21294916e-01 -2.84812748e-01 3.17677706e-01
-1.71309114e-02 6.68062329e-01 1.53870797e+00 9.65966433e-02
-1.73625007e-01 -1.76305279e-01 -4.23535138e-01 1.09788656e+00
8.08034897e-01 -1.57675773e-01 -4.57019776e-01 -1.00687861e+00
5.58937132e-01 7.56331161e-02 4.54744548e-01 -1.13786888e+00
1.25099212e-01 -4.09151495e-01 3.87889892e-01 -5.89824736e-01
6.29051566e-01 -9.27546084e-01 1.59977257e-01 -1.17129218e-02
4.04277712e-01 -3.64789695e-01 9.84965861e-02 1.88954324e-01
-3.41272205e-01 4.63968515e-02 1.09999657e+00 -2.25083992e-01
-7.25636005e-01 3.23976517e-01 1.25459671e-01 -6.62928373e-02
5.47357142e-01 -5.60136735e-01 -8.60667884e-01 -5.12556970e-01
-3.92446220e-01 3.02192152e-01 1.14664626e+00 3.02340597e-01
5.26865065e-01 -1.07219541e+00 -4.03627276e-01 7.24993765e-01
4.11701173e-01 6.63226724e-01 3.59223038e-01 6.87216640e-01
-8.82502317e-01 -1.59197692e-02 -8.51189941e-02 -5.97215772e-01
-9.34297621e-01 2.49744579e-01 7.11344182e-01 2.68767536e-01
-8.85087848e-01 5.77476203e-01 9.93819237e-01 -5.47263622e-01
2.66294293e-02 -3.78917277e-01 2.28186309e-01 -3.76538068e-01
5.51669419e-01 2.32410431e-02 1.87422678e-01 -5.01192868e-01
-8.97619873e-02 9.17072892e-01 4.12892431e-01 -1.93483941e-02
1.06140542e+00 -6.46865726e-01 -2.77194679e-01 6.89536691e-01
1.10484993e+00 -5.05558494e-03 -1.69543719e+00 -6.69195801e-02
-8.37918103e-01 -8.98817539e-01 3.42928916e-01 -6.11795604e-01
-1.24087429e+00 9.41969872e-01 7.73108661e-01 -5.94271384e-02
1.57169926e+00 -4.07773077e-01 6.61712348e-01 3.49525899e-01
4.42907780e-01 -1.03355896e+00 -3.93027186e-01 7.81234205e-01
6.79291904e-01 -1.64743733e+00 -1.80305988e-01 -3.36547285e-01
-2.80425161e-01 1.11119735e+00 8.37842464e-01 6.50724303e-03
4.72089887e-01 3.47030789e-01 6.49196744e-01 -1.05970263e-01
-5.31669080e-01 -5.45899510e-01 -1.98743612e-01 5.74190617e-01
2.55636394e-01 -1.63518935e-01 1.78619064e-02 -4.01293606e-01
-4.03425813e-01 -1.78121142e-02 8.63613725e-01 7.71202683e-01
-5.80629766e-01 -6.91599309e-01 -6.95593297e-01 1.70064270e-01
1.93968229e-02 -2.96668261e-01 -2.77675003e-01 5.80011487e-01
4.25022215e-01 9.64499414e-01 5.67826629e-01 -1.36888057e-01
4.04436737e-01 -3.89582127e-01 3.34335983e-01 -4.59263027e-01
2.59599928e-02 3.82450432e-01 -1.02705723e-02 -6.96991265e-01
-5.74353576e-01 -4.24271435e-01 -1.07802224e+00 -4.75584328e-01
-1.61239356e-01 -2.85100132e-01 7.71112144e-01 8.33923221e-01
1.01950979e-02 3.39005321e-01 1.03403616e+00 -1.12652791e+00
2.12314174e-01 -7.04318404e-01 -1.03535581e+00 4.12353635e-01
7.81723619e-01 -5.89154601e-01 -1.22217572e+00 1.52122423e-01] | [9.81088924407959, -2.9204325675964355] |
f4b9577f-e7ee-4109-96a9-d810bc545d12 | bayesian-optimisation-for-robust-model | 2203.00551 | null | https://arxiv.org/abs/2203.00551v3 | https://arxiv.org/pdf/2203.00551v3.pdf | Bayesian Optimisation for Robust Model Predictive Control under Model Parameter Uncertainty | We propose an adaptive optimisation approach for tuning stochastic model predictive control (MPC) hyper-parameters while jointly estimating probability distributions of the transition model parameters based on performance rewards. In particular, we develop a Bayesian optimisation (BO) algorithm with a heteroscedastic noise model to deal with varying noise across the MPC hyper-parameter and dynamics model parameter spaces. Typical homoscedastic noise models are unrealistic for tuning MPC since stochastic controllers are inherently noisy, and the level of noise is affected by their hyper-parameter settings. We evaluate the proposed optimisation algorithm in simulated control and robotics tasks where we jointly infer control and dynamics parameters. Experimental results demonstrate that our approach leads to higher cumulative rewards and more stable controllers. | ['Fabio Ramos', 'Rafael Oliveira', 'Rel Guzman'] | 2022-03-01 | null | null | null | null | ['bayesian-optimisation'] | ['methodology'] | [ 2.77495414e-01 3.58145684e-02 -1.18466884e-01 1.46823883e-01
-6.44254148e-01 -5.05384505e-01 6.66221857e-01 -2.95245111e-01
-4.90026385e-01 1.01346052e+00 -3.40893082e-02 -3.99457477e-02
-8.16435695e-01 -6.00366414e-01 -7.70631611e-01 -1.04857278e+00
-1.26610473e-01 7.86375463e-01 2.09707722e-01 1.08320273e-01
2.58358359e-01 9.65393037e-02 -1.38048768e+00 -5.84450483e-01
9.15949166e-01 6.87321663e-01 6.48785591e-01 1.20607960e+00
7.35805213e-01 5.23042500e-01 -7.37485647e-01 3.21171799e-04
3.15112889e-01 -1.44243121e-01 -6.50933161e-02 1.45424813e-01
-5.00063360e-01 -9.69635770e-02 -1.35709181e-01 1.39306045e+00
4.94652063e-01 7.27529228e-01 9.49298859e-01 -1.18898773e+00
-1.97644219e-01 5.80566049e-01 -1.85313866e-01 -2.51705259e-01
-2.47527212e-01 9.31488693e-01 5.71101606e-01 -3.83768696e-03
1.74346790e-01 1.61106598e+00 4.27084804e-01 4.26343679e-01
-1.58940029e+00 -4.69161004e-01 2.69372433e-01 -1.17989652e-01
-1.52281308e+00 -2.07491353e-01 3.79681379e-01 -7.71863103e-01
6.46803737e-01 -7.51583055e-02 4.54838902e-01 1.47739005e+00
8.67907643e-01 3.77046317e-01 1.05519235e+00 -1.34942904e-01
7.21911192e-01 7.18377158e-02 -4.96849537e-01 2.56487638e-01
3.47053289e-01 9.75650847e-01 -2.17457768e-02 -2.88284957e-01
1.12868321e+00 -7.91557282e-02 -7.79351741e-02 -5.89185119e-01
-1.25483131e+00 8.73584330e-01 -1.42153993e-01 -3.92650306e-01
-7.30610371e-01 8.82784009e-01 2.18631819e-01 2.74763495e-01
-9.72828716e-02 6.47312164e-01 -5.40013552e-01 -4.98070031e-01
-1.33487344e-01 6.73956811e-01 1.01206851e+00 1.25390863e+00
2.63120949e-01 5.02227366e-01 -6.30941510e-01 7.83469796e-01
6.46637619e-01 9.37993467e-01 4.20688391e-01 -1.50652027e+00
4.13872272e-01 -1.24811158e-01 1.00345087e+00 -9.97836292e-01
-2.06934631e-01 -6.17262065e-01 -7.24272728e-01 4.68332052e-01
3.02024990e-01 -6.96989119e-01 -1.01470315e+00 1.64164221e+00
2.73873985e-01 1.58227742e-01 6.64279759e-02 9.89598274e-01
-2.71941066e-01 5.38293064e-01 9.52739790e-02 -3.44525695e-01
8.87011468e-01 -4.68851894e-01 -1.05691290e+00 -3.31435710e-01
1.49806052e-01 -5.90509057e-01 9.66010451e-01 5.73912621e-01
-9.63999450e-01 -4.06753331e-01 -9.34425831e-01 8.96776617e-01
1.68748051e-01 1.34628892e-01 -7.03672618e-02 6.38111591e-01
-4.81261402e-01 9.60812747e-01 -1.15851498e+00 7.26313591e-02
1.35299996e-01 5.35164177e-01 4.50697631e-01 3.88749212e-01
-9.25558865e-01 1.03016353e+00 8.98519933e-01 2.11298034e-01
-1.64948463e+00 -4.56371486e-01 -4.89042401e-01 -1.36835098e-01
1.10370338e+00 -7.76779830e-01 1.51002145e+00 -2.39352077e-01
-2.49081635e+00 -1.40347064e-01 3.46289963e-01 -6.07884049e-01
5.14719188e-01 -3.47199053e-01 -1.45498767e-01 -3.60171169e-01
-4.39979374e-01 4.15530980e-01 1.42854810e+00 -1.22081614e+00
-4.96476084e-01 -2.77695879e-02 -1.86216101e-01 3.18505704e-01
1.97884575e-01 -3.06355953e-01 -1.87527269e-01 -6.23276889e-01
-3.29871744e-01 -1.47349966e+00 -8.04457605e-01 -3.72264683e-01
-5.97658992e-01 2.42765583e-02 3.85808408e-01 -2.13002786e-01
1.22013938e+00 -1.89730334e+00 5.58196783e-01 5.15415788e-01
-1.06398836e-01 4.84120511e-02 1.07517444e-01 4.38628525e-01
5.35827160e-01 -1.07900910e-02 -2.45274022e-01 2.78032082e-03
4.48724210e-01 6.86634898e-01 -3.15830082e-01 4.31134850e-01
3.18421721e-01 4.88731861e-01 -9.25379097e-01 -1.40608355e-01
3.17661583e-01 3.16279352e-01 -5.05081713e-01 3.79559427e-01
-6.71200097e-01 7.31038272e-01 -9.36830640e-01 1.05682462e-01
1.89283624e-01 1.43845066e-01 2.44194523e-01 2.90871412e-01
-2.72552613e-02 -2.83662230e-01 -1.80386698e+00 8.75286758e-01
-4.28018928e-01 6.53926581e-02 4.02985960e-01 -7.58031309e-01
1.11646628e+00 2.25440785e-01 1.53894916e-01 6.75675273e-02
8.42114806e-01 -6.29307255e-02 2.65848786e-01 -4.25466925e-01
4.13005084e-01 -1.63414747e-01 -2.64964461e-01 -1.15012795e-01
1.23640075e-01 -9.74184155e-01 -1.15511969e-01 -4.52745259e-01
9.52841163e-01 3.41619909e-01 3.29751998e-01 -2.62977719e-01
2.79610127e-01 -9.38414689e-03 7.78862238e-01 1.15254593e+00
-4.38103735e-01 1.56973109e-01 5.64121246e-01 3.38531047e-01
-1.04063869e+00 -9.97297764e-01 1.57175049e-01 8.18480790e-01
-9.30181071e-02 -2.79220082e-02 -6.23428583e-01 1.49901345e-01
6.57139793e-02 1.09672785e+00 -5.08907974e-01 -6.38226330e-01
-2.59346813e-01 -7.23016918e-01 -4.02523503e-02 3.70959848e-01
-7.73748457e-02 -7.12742329e-01 -6.63857996e-01 7.58693397e-01
3.48217666e-01 -8.26970100e-01 -4.28890318e-01 5.14103651e-01
-8.36034775e-01 -8.61825347e-01 -4.53247547e-01 -2.34638348e-01
3.07589918e-01 -6.38291955e-01 7.16567576e-01 -5.65724969e-01
-1.54642850e-01 6.89817309e-01 1.89141750e-01 -8.68147612e-01
-7.41648316e-01 -2.54249573e-01 3.86765182e-01 -1.22736692e-01
-2.60007262e-01 -3.19896132e-01 -3.85320991e-01 6.34093344e-01
-6.22746766e-01 -5.31107783e-01 4.42248344e-01 1.04223037e+00
8.51437867e-01 5.31055212e-01 3.65534514e-01 -6.03667617e-01
1.06848538e+00 -4.82232720e-01 -1.47128451e+00 2.19417904e-02
-5.50351799e-01 2.36625999e-01 5.14580071e-01 -1.11791229e+00
-1.23053944e+00 3.69970173e-01 6.42685056e-01 -8.80149186e-01
-3.01942199e-01 2.44247809e-01 6.40062541e-02 1.72403708e-01
5.66828728e-01 -1.40706569e-01 3.33372831e-01 -2.01742858e-01
3.36067289e-01 4.38406438e-01 7.34527230e-01 -1.02863216e+00
9.18934703e-01 -1.67916231e-02 3.29876840e-01 -4.56250995e-01
-4.09880191e-01 -1.90195560e-01 -4.16676521e-01 -4.21470493e-01
7.87617922e-01 -9.95336235e-01 -1.26555765e+00 4.81395096e-01
-7.32524872e-01 -7.57886112e-01 -4.19632673e-01 9.57478166e-01
-1.48503673e+00 -1.66067317e-01 -2.80774474e-01 -1.70315671e+00
2.51544237e-01 -1.48455286e+00 9.57645416e-01 3.32322419e-01
-2.15280920e-01 -1.09941626e+00 3.05056661e-01 -2.13591009e-01
4.33379740e-01 4.18587923e-01 5.89718461e-01 -3.34730864e-01
-6.23562098e-01 -3.52889866e-01 5.22997737e-01 2.13417083e-01
-2.54561841e-01 1.74388126e-01 -4.76533175e-01 -2.16693506e-01
2.22069979e-01 -1.73567951e-01 3.09021115e-01 1.02718472e+00
9.36744988e-01 -5.05321026e-01 -2.98665404e-01 1.79531872e-01
1.36004233e+00 4.74468648e-01 1.74889311e-01 3.78183007e-01
4.50498700e-01 6.23166978e-01 1.09914017e+00 1.05866230e+00
-8.31445400e-03 6.75469875e-01 7.92511582e-01 8.50705087e-01
6.28998756e-01 -4.16097403e-01 6.08665228e-01 4.87146914e-01
9.70405787e-02 -3.01206291e-01 -8.90413642e-01 5.97719073e-01
-2.11803150e+00 -6.82323277e-01 -1.67526111e-01 2.63100839e+00
1.08549988e+00 2.95594573e-01 1.63163796e-01 -4.09319609e-01
9.68823075e-01 -4.18953419e-01 -8.79011273e-01 -4.24244821e-01
2.12090790e-01 -1.39449209e-01 1.02977824e+00 5.14124036e-01
-1.01562953e+00 6.23301744e-01 6.91868353e+00 1.12692368e+00
-2.78901130e-01 -2.80481100e-01 2.43710697e-01 -2.35638559e-01
2.07850382e-01 -5.91843799e-02 -9.32710707e-01 7.10133970e-01
1.25319982e+00 -4.30647403e-01 8.87172341e-01 9.21970129e-01
7.45111585e-01 -3.13360929e-01 -7.79598475e-01 6.41839981e-01
-7.52934575e-01 -6.74729168e-01 -4.92284894e-01 1.05886914e-01
1.09224832e+00 -7.81028196e-02 4.79908019e-01 4.30924118e-01
1.38568819e+00 -9.20683920e-01 8.11310768e-01 1.11605489e+00
1.25645891e-01 -1.01076972e+00 7.91156292e-01 6.54329777e-01
-6.94072664e-01 -6.95307493e-01 -5.11140347e-01 -3.54290567e-02
2.02859610e-01 3.97654563e-01 -8.62533689e-01 3.96089070e-02
4.99668211e-01 3.65288228e-01 1.51917646e-02 1.43491602e+00
-3.44873428e-01 7.62135744e-01 -5.14661551e-01 -5.61365783e-01
1.74982622e-01 -6.35291755e-01 1.25269663e+00 6.18766367e-01
2.97723323e-01 -1.15059115e-01 5.29915869e-01 1.22812104e+00
5.90477526e-01 -4.39587593e-01 -5.27652383e-01 -1.95443705e-01
7.44557858e-01 7.28788853e-01 -4.73182201e-01 -1.07482873e-01
3.76993984e-01 2.65918434e-01 -1.03435017e-01 5.91958106e-01
-8.92284632e-01 -4.22253668e-01 7.58730710e-01 -6.48985505e-01
5.35206854e-01 -5.31490266e-01 -1.75798923e-01 -4.61897194e-01
-3.58931780e-01 -8.13361228e-01 5.92602529e-02 -5.24197221e-01
-1.37475085e+00 -2.51032840e-02 4.09334898e-01 -1.30357552e+00
-8.21457982e-01 -5.25794089e-01 -5.00982642e-01 7.66073883e-01
-9.47725534e-01 -3.55924517e-01 3.14929336e-01 3.67560863e-01
3.91072422e-01 -1.16290204e-01 4.29562122e-01 -6.23078883e-01
-8.46929967e-01 2.19730195e-02 7.50911415e-01 -5.14771879e-01
5.96503854e-01 -1.39649522e+00 -1.65493160e-01 4.49411541e-01
-7.33952999e-01 6.43735707e-01 1.51851940e+00 -9.14552391e-01
-1.48293483e+00 -1.27680814e+00 -2.32699722e-01 -5.94817221e-01
1.28186655e+00 -1.17996082e-01 -7.19845355e-01 4.14303064e-01
-3.89590748e-02 -4.58103597e-01 -2.39215717e-02 -2.56830543e-01
4.61150050e-01 2.98526794e-01 -1.10544229e+00 8.00803542e-01
6.55146599e-01 -1.11850485e-01 -4.63157922e-01 4.21764761e-01
1.08665586e+00 -5.22092879e-01 -1.39572668e+00 4.51882601e-01
1.97213843e-01 -1.10553950e-01 9.62179601e-01 -5.03824651e-01
-1.10180452e-01 -3.95088017e-01 -9.95306298e-02 -1.82680535e+00
-2.60373652e-01 -1.31305158e+00 -3.56486231e-01 1.14576662e+00
2.68853813e-01 -5.85713625e-01 3.84523928e-01 7.25997388e-01
3.62245063e-03 -4.95757461e-01 -1.05463243e+00 -1.27099216e+00
5.80211729e-02 -6.74390733e-01 3.18268657e-01 2.06000403e-01
-2.74868876e-01 8.35850090e-02 -6.65756106e-01 4.94708121e-01
9.52256203e-01 -4.47203279e-01 8.51416647e-01 -9.83556569e-01
-7.17037976e-01 -5.46187997e-01 -1.70634076e-01 -7.26715267e-01
2.30045512e-01 6.58107409e-03 1.05392110e+00 -8.60265851e-01
3.42876185e-03 -1.42747104e-01 -1.04780346e-01 -2.48079505e-02
-4.22236890e-01 -5.43219686e-01 2.82374062e-02 3.27164400e-03
-6.34240031e-01 1.20423818e+00 1.21638322e+00 2.18058983e-03
-6.38858020e-01 7.13104308e-01 -4.89481017e-02 6.31508827e-01
7.94736564e-01 -7.30329931e-01 -7.37445533e-01 -2.10850406e-02
-8.99483822e-03 3.40179563e-01 4.27716941e-01 -9.72460508e-01
1.87662870e-01 -7.54474401e-01 3.87547761e-02 -4.94515270e-01
2.36335292e-01 -1.11062467e+00 5.34341276e-01 6.40171826e-01
-5.30081570e-01 -1.99908450e-01 2.71225274e-01 1.59755468e+00
1.22043207e-01 -2.85117656e-01 8.96133006e-01 3.00208554e-02
-1.64652690e-01 1.14712015e-01 -1.01166093e+00 4.50016074e-02
9.13207769e-01 -5.25911190e-02 8.60178024e-02 -5.55414140e-01
-1.08258641e+00 6.90574527e-01 1.85927883e-01 4.19882208e-01
3.66198808e-01 -1.04359734e+00 -3.61765772e-01 -1.77063286e-01
-1.56755999e-01 -1.27990380e-01 5.19985706e-02 7.58518934e-01
1.11201562e-01 3.21137428e-01 -2.31222846e-02 -9.35791254e-01
-7.43179202e-01 5.23542643e-01 4.62362260e-01 -2.33591035e-01
-2.26302609e-01 6.41976655e-01 -1.02771938e-01 -7.71481037e-01
5.29760361e-01 -5.86344779e-01 -1.84813112e-01 -3.69618207e-01
1.37098029e-01 6.96557760e-01 -4.31986928e-01 -1.79246530e-01
7.48989433e-02 3.75021726e-01 5.15370667e-01 -4.56305981e-01
1.15951002e+00 -4.04246658e-01 3.91045749e-01 8.38831544e-01
5.18140554e-01 -7.39740610e-01 -2.21072674e+00 -2.39907041e-01
2.76986092e-01 -3.74560118e-01 4.34704393e-01 -7.02064455e-01
-4.35538799e-01 4.23953295e-01 7.68091738e-01 -9.79417711e-02
6.29461467e-01 -4.50518548e-01 6.13092072e-02 5.14879942e-01
5.26849151e-01 -1.52717865e+00 -2.25985460e-02 8.07238877e-01
8.04703712e-01 -9.16137278e-01 -1.88355178e-01 -1.03143647e-01
-1.05882037e+00 9.67182159e-01 6.61767185e-01 -5.08184135e-01
6.91462874e-01 4.50010777e-01 -4.70440269e-01 1.47388577e-01
-1.27339661e+00 -2.78701693e-01 2.84687191e-01 1.07597613e+00
-3.71140331e-01 2.43211672e-01 6.86472207e-02 5.61662614e-01
-1.45024378e-02 1.61988791e-02 5.73939025e-01 8.53840649e-01
-6.68786347e-01 -7.55099237e-01 -9.21952188e-01 3.09306026e-01
-1.84547886e-01 3.06325287e-01 7.75647387e-02 6.14275157e-01
-4.77270842e-01 1.39912665e+00 -6.36974052e-02 -2.59248137e-01
6.41200900e-01 -7.71310553e-02 2.57106334e-01 -6.00785494e-01
-4.44338143e-01 6.02119803e-01 7.64488578e-02 -6.88132584e-01
7.45111033e-02 -6.95440829e-01 -9.92161214e-01 -6.72261119e-02
-6.30935967e-01 -3.55490968e-02 7.58349597e-01 8.44401658e-01
3.20798397e-01 9.80320692e-01 6.89274192e-01 -1.03423786e+00
-1.64871645e+00 -1.00139284e+00 -6.37220860e-01 -2.54525751e-01
1.92038879e-01 -1.31892276e+00 -5.16946793e-01 5.73387705e-02] | [4.906641483306885, 2.3784332275390625] |
e0f91631-79c9-401a-a95c-58c29479dc4c | neural-networks-for-stock-price-prediction | 1805.11317 | null | http://arxiv.org/abs/1805.11317v1 | http://arxiv.org/pdf/1805.11317v1.pdf | Neural networks for stock price prediction | Due to the extremely volatile nature of financial markets, it is commonly
accepted that stock price prediction is a task full of challenge. However in
order to make profits or understand the essence of equity market, numerous
market participants or researchers try to forecast stock price using various
statistical, econometric or even neural network models. In this work, we survey
and compare the predictive power of five neural network models, namely, back
propagation (BP) neural network, radial basis function (RBF) neural network,
general regression neural network (GRNN), support vector machine regression
(SVMR), least squares support vector machine regresssion (LS-SVMR). We apply
the five models to make price prediction of three individual stocks, namely,
Bank of China, Vanke A and Kweichou Moutai. Adopting mean square error and
average absolute percentage error as criteria, we find BP neural network
consistently and robustly outperforms the other four models. | ['Ren-Jie Han', 'Yu-Long Zhou', 'Yue-Gang Song'] | 2018-05-29 | null | null | null | null | ['stock-price-prediction'] | ['time-series'] | [-6.72613084e-01 -4.16966945e-01 -2.54177392e-01 -1.78113297e-01
1.01631284e-01 -4.79707837e-01 4.80391234e-01 -3.07984620e-01
-2.63826311e-01 1.18199909e+00 -3.22697163e-02 -8.98041725e-01
-3.25756490e-01 -1.02058899e+00 -1.27001852e-01 -5.23392737e-01
-1.34160504e-01 2.57396460e-01 5.06760031e-02 -5.31600237e-01
1.18001783e+00 6.41347051e-01 -1.13623405e+00 -3.77638072e-01
5.82720101e-01 1.69982648e+00 -2.16949880e-01 3.48065048e-01
-2.15774491e-01 1.44290853e+00 -3.86104763e-01 -3.27832133e-01
7.10983992e-01 -1.85221359e-02 -4.11351860e-01 -4.85351086e-01
-4.65244025e-01 -4.16900665e-01 -3.84184092e-01 9.08861279e-01
2.03895777e-01 4.55773920e-02 7.84794748e-01 -1.16180241e+00
-1.31598365e+00 5.30698717e-01 -6.43872678e-01 6.74012423e-01
-2.65641540e-01 -1.15256675e-01 9.83909190e-01 -1.17945540e+00
2.73037311e-02 5.89669228e-01 9.64442611e-01 3.73572409e-02
-9.33993399e-01 -1.06861651e+00 -7.38750547e-02 1.41748488e-01
-1.13882804e+00 -6.38377070e-02 8.36137235e-01 -5.51528811e-01
1.08934081e+00 3.46816242e-01 6.88865364e-01 2.18505397e-01
5.23520708e-01 2.35749185e-01 1.53455198e+00 -4.07041490e-01
1.70641303e-01 6.47845685e-01 3.08318228e-01 1.92221984e-01
6.21213973e-01 4.07690465e-01 -3.17813575e-01 -2.60421753e-01
9.84818399e-01 2.70289600e-01 -2.07290366e-01 5.87949872e-01
-7.21142292e-01 1.26287854e+00 3.36627990e-01 5.11134684e-01
-8.92377257e-01 -1.36433676e-01 -6.97240606e-03 6.00484431e-01
6.15991116e-01 5.85907161e-01 -9.23870265e-01 -1.25393942e-01
-1.25595200e+00 5.86841166e-01 1.20678091e+00 3.22571605e-01
3.60766858e-01 9.17854130e-01 3.72850686e-01 8.38647902e-01
5.65145195e-01 4.08347398e-01 1.13871741e+00 -5.15580893e-01
3.13223004e-01 5.86067796e-01 3.16678494e-01 -1.41221917e+00
-4.62384015e-01 -2.18467504e-01 -8.33856821e-01 5.34217536e-01
2.24209294e-01 -6.10478044e-01 -3.90160888e-01 6.49538696e-01
-3.89424711e-01 3.16883475e-01 3.70587051e-01 5.27841866e-01
5.93559146e-01 9.78861153e-01 -3.23463023e-01 -4.19448584e-01
9.85971868e-01 -9.68995988e-01 -4.97733951e-01 -1.01641014e-01
-9.12873540e-03 -6.63889945e-01 6.88427389e-02 5.18099725e-01
-1.18921757e+00 -5.05746424e-01 -8.83115232e-01 5.76637983e-01
-6.76293373e-01 3.25799361e-03 6.32531226e-01 6.18654788e-01
-8.18823755e-01 9.48911071e-01 -3.53207290e-01 4.80073035e-01
-4.87061106e-02 5.59164643e-01 1.37548167e-02 8.39024305e-01
-1.33219492e+00 1.25495481e+00 7.40656793e-01 3.91418040e-01
3.22538227e-01 -6.36924326e-01 -4.81028974e-01 3.06074601e-02
-2.12097585e-01 -4.24426757e-02 1.16867375e+00 -8.19187045e-01
-1.72589302e+00 2.24886581e-01 1.96487263e-01 -8.50738823e-01
2.52244085e-01 1.08928690e-02 -7.90333331e-01 -3.00133765e-01
-4.50294733e-01 4.48906794e-02 5.19417465e-01 -6.59848034e-01
-7.69810438e-01 -1.00911133e-01 -4.88018781e-01 -2.43031517e-01
-4.07771021e-02 3.62100303e-01 6.70917690e-01 -1.13136756e+00
4.54022050e-01 -7.43940055e-01 -1.78628907e-01 -8.00160706e-01
-2.23176345e-01 -3.90877545e-01 5.64408779e-01 -1.28430736e+00
1.54379714e+00 -1.78039455e+00 -6.50168478e-01 8.14600646e-01
-2.28307366e-01 1.78503513e-01 6.10893011e-01 2.79651791e-01
-6.18183076e-01 1.73437491e-01 -3.68489549e-02 6.03842616e-01
2.45849118e-02 6.52270168e-02 -7.05777228e-01 3.09373587e-01
2.73425579e-01 1.07057023e+00 -1.59524158e-01 -7.91406352e-03
3.87302078e-02 1.18374839e-01 7.74245430e-03 2.21343730e-02
2.22318098e-01 -2.14020267e-01 -4.36296493e-01 1.01557934e+00
6.30088568e-01 -2.88586766e-01 -4.46227551e-01 8.84114131e-02
-5.48047125e-01 1.71205476e-01 -1.55868673e+00 1.18607759e-01
4.87660393e-02 7.94812143e-01 -4.73043621e-01 -1.24838424e+00
1.82430363e+00 6.36733174e-01 2.77058959e-01 -6.94688320e-01
1.59439996e-01 9.41437006e-01 8.07579681e-02 -2.95849681e-01
7.23414481e-01 -6.55832052e-01 4.85296905e-01 4.47399884e-01
-5.20661891e-01 4.37365174e-01 -3.71147282e-02 -7.04421699e-01
3.85061502e-01 -1.62889585e-01 4.79970306e-01 -2.49959975e-01
6.51062489e-01 -1.34751964e-02 5.61700344e-01 3.96951973e-01
-2.77058452e-01 4.99726593e-01 4.70247746e-01 -7.81046569e-01
-1.00356662e+00 -5.86539686e-01 -4.03366238e-01 6.19020760e-01
-3.00166339e-01 5.89253008e-01 -1.42179877e-01 3.30686234e-02
6.00618601e-01 9.16217625e-01 -4.89987582e-01 1.33694023e-01
-5.80942333e-01 -1.12654221e+00 2.53224880e-01 6.77061319e-01
7.33529210e-01 -1.38856113e+00 -5.56105912e-01 5.43535888e-01
6.84182167e-01 -6.12846792e-01 2.24641666e-01 1.64596453e-01
-1.02235889e+00 -8.48691344e-01 -1.12903190e+00 -5.75418770e-01
1.60904676e-01 -2.08312467e-01 1.03726113e+00 3.57595198e-02
1.41227379e-01 -2.09479094e-01 -1.21192127e-01 -7.88880348e-01
-7.98405856e-02 -2.02497780e-01 1.92574933e-01 -1.08119115e-01
7.53252625e-01 -4.63215381e-01 -5.02853334e-01 2.33619973e-01
-3.38494062e-01 -4.08518672e-01 7.43922293e-01 6.73810959e-01
2.27692202e-01 7.29974031e-01 1.06038070e+00 -6.02586627e-01
1.08493865e+00 -8.77860487e-01 -1.06154668e+00 2.07742721e-01
-1.39556074e+00 -2.09582195e-01 4.26894009e-01 -2.78054297e-01
-9.38458323e-01 -5.31187177e-01 7.21380562e-02 -1.65902600e-01
1.93204269e-01 9.82285857e-01 8.31560194e-01 -2.92615652e-01
4.09187198e-01 7.58555174e-01 1.29748195e-01 -6.01758778e-01
-2.98349261e-01 8.30219388e-01 4.62441832e-01 -6.23939559e-02
1.13758063e+00 -1.42178774e-01 1.08408287e-01 -5.74983716e-01
-3.19452941e-01 -1.19062871e-01 -5.17533243e-01 -2.25340098e-01
5.72335541e-01 -8.01027477e-01 -7.14162052e-01 7.10292280e-01
-7.79808402e-01 1.40984848e-01 1.63605824e-01 8.00367236e-01
-2.66529351e-01 -2.08699927e-01 -9.49844301e-01 -1.33431101e+00
-5.55464983e-01 -7.96640694e-01 -1.56510130e-01 5.56840062e-01
-3.35137933e-01 -1.32608724e+00 1.33462518e-01 1.08945899e-01
9.47195053e-01 5.58974564e-01 6.86679184e-01 -1.32726955e+00
-1.96768865e-01 -7.71058083e-01 -5.17934859e-01 7.51426637e-01
1.41266719e-01 3.70878339e-01 -5.92956245e-01 1.31198540e-01
1.72759324e-01 -9.48574161e-04 6.00675702e-01 7.21660018e-01
7.35244930e-01 -7.59071887e-01 2.26284042e-01 2.76741266e-01
1.75226796e+00 9.70468044e-01 6.95061505e-01 1.19069076e+00
-4.34936117e-03 4.19943750e-01 2.09294900e-01 5.81883609e-01
3.00687104e-01 -1.54675990e-01 2.11855061e-02 1.74416542e-01
1.04783189e+00 1.56588275e-02 3.58661383e-01 8.10189605e-01
-8.41150701e-01 4.80041087e-01 -8.63004029e-01 -1.31916881e-01
-1.61311734e+00 -1.21037877e+00 -1.57456860e-01 1.73686934e+00
6.92062914e-01 5.77503681e-01 4.83620673e-01 6.59911871e-01
5.50394595e-01 3.80011164e-02 -3.81103188e-01 -6.21409535e-01
-2.92159706e-01 2.45645329e-01 1.15828133e+00 6.75935373e-02
-1.03623939e+00 4.26761985e-01 6.76528215e+00 4.28890169e-01
-1.53544462e+00 -5.35163939e-01 1.01604235e+00 2.70933896e-01
9.47600044e-03 -2.08161488e-01 -9.55721915e-01 8.25744152e-01
1.23194528e+00 -6.96251869e-01 5.01697779e-01 1.22825551e+00
1.85441256e-01 7.77102411e-02 -3.57478887e-01 7.78507948e-01
-3.23931098e-01 -1.60622537e+00 -3.16695780e-01 1.44958317e-01
6.96008086e-01 1.56087568e-02 2.65082419e-01 6.09189570e-01
1.73833534e-01 -1.16520655e+00 7.13773131e-01 1.09443724e+00
-1.27325267e-01 -8.30521166e-01 1.15391505e+00 3.61792028e-01
-9.62405145e-01 -3.55040610e-01 -6.98496878e-01 -7.93921411e-01
-1.00099161e-01 4.58395571e-01 -2.75037855e-01 2.59676546e-01
8.21064770e-01 5.75660586e-01 -2.89440125e-01 9.97011781e-01
2.12499380e-01 7.84508586e-01 -2.56346673e-01 -5.13361394e-01
5.47278404e-01 -5.23467541e-01 6.26639128e-02 8.32288146e-01
5.29782355e-01 6.44758761e-01 -4.66639131e-01 9.78277802e-01
3.80169421e-01 2.36835748e-01 -1.76678896e-01 -5.45915216e-02
4.96157616e-01 9.10183370e-01 -6.30101562e-01 -2.21635565e-01
-6.86979175e-01 1.96537718e-01 -2.12055549e-01 4.00283247e-01
-6.01614356e-01 -6.49634898e-01 3.47668767e-01 5.55289723e-02
3.09738934e-01 -1.66191205e-01 -9.09599245e-01 -9.74346042e-01
5.51215522e-02 -5.04948556e-01 1.50346518e-01 -7.98259497e-01
-1.44120479e+00 5.58799803e-01 -3.54818106e-02 -1.14039099e+00
-4.49782342e-01 -1.24631298e+00 -1.11288369e+00 1.40008628e+00
-1.85054493e+00 -2.21343502e-01 2.51762450e-01 2.64781296e-01
1.76602110e-01 -1.12990057e+00 4.48684037e-01 1.20921344e-01
-8.96496236e-01 9.35679525e-02 5.64431071e-01 6.05002046e-01
3.96997407e-02 -1.35224628e+00 3.56132418e-01 3.55296373e-01
-3.17665935e-01 4.83396828e-01 6.85836971e-01 -6.69408560e-01
-8.23554158e-01 -8.07378709e-01 1.12237966e+00 1.32754982e-01
1.18547559e+00 5.54074585e-01 -1.22928071e+00 8.99247885e-01
2.06789896e-01 -4.90350053e-02 6.60723746e-01 -3.34861636e-01
-1.30024299e-01 -2.40497082e-01 -1.25085318e+00 2.93875277e-01
-3.37002158e-01 -2.69681104e-02 -1.01200867e+00 -4.59904931e-02
3.18046391e-01 6.08706921e-02 -1.23381150e+00 3.60134661e-01
7.76686251e-01 -1.36072373e+00 1.08553636e+00 -5.43359339e-01
4.76544529e-01 1.79110020e-02 -1.50343582e-01 -1.06764805e+00
-6.39405549e-01 -3.34855646e-01 -2.06662938e-01 1.11149955e+00
8.91861141e-01 -1.47513187e+00 9.45905030e-01 1.43928456e+00
3.77932787e-01 -1.12940502e+00 -9.41750348e-01 -8.81805837e-01
4.27460194e-01 -2.04596430e-01 8.46138358e-01 1.17661059e+00
1.78355664e-01 -1.74651295e-01 -4.79383290e-01 -1.08297527e-01
3.46354693e-01 1.88142627e-01 2.20892772e-01 -1.55168068e+00
-3.13156396e-01 -1.03923488e+00 -3.48376960e-01 -6.05943441e-01
1.31843001e-01 -3.65575790e-01 -5.90980232e-01 -1.26670825e+00
-4.36278403e-01 -5.32315195e-01 -8.81128192e-01 3.60899210e-01
6.14195056e-02 1.05766930e-01 1.51019931e-01 7.03666747e-01
4.18584108e-01 1.14008717e-01 7.87486732e-01 1.77001625e-01
-3.28376204e-01 5.61701894e-01 -7.60785997e-01 8.87635410e-01
1.22202325e+00 -1.99171454e-01 1.97386608e-01 2.50478595e-01
4.74898040e-01 2.97109902e-01 1.34371847e-01 -6.20055854e-01
4.14878041e-01 -6.86282039e-01 9.14261401e-01 -8.93056571e-01
-2.79739983e-02 -6.81822181e-01 2.81743765e-01 8.59139085e-01
-7.28732795e-02 7.17339873e-01 1.84249759e-01 1.14915650e-02
-6.64723158e-01 -7.82311499e-01 5.51607668e-01 -3.80650997e-01
-7.43467629e-01 1.23408169e-01 -3.82838666e-01 -3.15152258e-01
1.25083220e+00 -7.36165762e-01 -2.20614478e-01 -3.01797241e-01
-4.62994039e-01 2.04431772e-01 -2.87845939e-01 1.65680826e-01
7.49971330e-01 -1.40384746e+00 -8.81447434e-01 2.71760285e-01
-5.07428110e-01 -4.87939149e-01 -1.80523753e-01 7.75607169e-01
-1.21899891e+00 8.62509131e-01 -2.95462132e-01 3.69487852e-01
-6.38314068e-01 2.79440403e-01 6.36615515e-01 -1.32221445e-01
-4.77788448e-01 8.69827032e-01 -5.47197700e-01 -1.71412416e-02
-7.25911791e-03 -4.88689005e-01 -7.68359303e-01 1.32443056e-01
6.72109544e-01 9.77524340e-01 -1.51948914e-01 -7.12288976e-01
-2.44983584e-02 6.94798529e-01 -2.37916168e-02 2.70117335e-02
1.67901421e+00 3.54431152e-01 -3.64506930e-01 5.97885966e-01
1.02207506e+00 -4.29244995e-01 -8.27114046e-01 -1.97769523e-01
7.50130951e-01 -2.64172822e-01 2.51801342e-01 -6.49745703e-01
-1.18848586e+00 3.69726390e-01 3.29307705e-01 8.79834592e-01
1.01316273e+00 -6.74396396e-01 8.83866310e-01 6.10845447e-01
-3.29581983e-02 -1.42294419e+00 -5.53545058e-01 4.15502399e-01
8.66763234e-01 -1.27327776e+00 1.15521781e-01 1.55005947e-01
-5.43697596e-01 1.69904828e+00 1.88035890e-01 -8.04452717e-01
1.61335564e+00 2.53942162e-01 1.56152949e-01 2.76494503e-01
-9.26835179e-01 4.13991541e-01 2.62633145e-01 8.38140026e-02
5.09724140e-01 -4.84139323e-02 -2.38506556e-01 1.02387941e+00
-8.40633750e-01 3.22713792e-01 6.47626698e-01 8.60442698e-01
-8.71659398e-01 -5.06465375e-01 -7.81779468e-01 1.02990127e+00
-9.80175674e-01 -3.62239361e-01 -5.03816009e-02 1.09727919e+00
-3.51866305e-01 5.92202365e-01 5.16690314e-01 -1.99963316e-01
4.41946834e-01 2.57972091e-01 -2.97989011e-01 2.63797566e-02
-6.53307498e-01 4.38616201e-02 -2.06582472e-01 2.65283942e-01
-2.48822093e-01 -8.05855393e-01 -1.19414103e+00 -7.62571871e-01
-5.17454982e-01 3.78835291e-01 7.23999321e-01 9.23794866e-01
-3.77247483e-01 2.68839926e-01 9.63628471e-01 -6.01253271e-01
-1.11560428e+00 -1.08980107e+00 -1.51423430e+00 -2.63972908e-01
3.01913708e-01 -6.89070582e-01 -7.57812679e-01 -2.99583562e-02] | [4.570211887359619, 4.168581485748291] |
66ab6fd1-13ce-44b7-9012-68ae6efa507d | pages-iai-uni-bonn-de-https-pages-iai-uni | null | null | https://pages.iai.uni-bonn.de/gall_juergen/download/jgall_forecastintention_3dv21.pdf | https://pages.iai.uni-bonn.de/gall_juergen/download/jgall_forecastintention_3dv21.pdf | Intention-based Long-Term Human Motion Anticipation | Recently, a few works have been proposed to model the
uncertainty of the future human motion. These works do
not forecast a single sequence but multiple sequences for
the same observation. While these works focused on in-
creasing the diversity, this work focuses on keeping a high
quality of the forecast sequences even for very long time
horizons of up to 30 seconds. In order to achieve this goal,
we propose to forecast the intention of the person ahead of
time. This has the advantage that the generated human mo-
tion remains goal oriented and that the motion transitions
between two actions are smooth and highly realistic. We
furthermore propose a new quality score for evaluation that
correlates better with human perception than other metrics.
The results and a user study show that our approach fore-
casts multiple sequences that are more plausible compared
to the state-of-the-art. | ['Juergen Gall', 'Chintan Zaveri', 'Julian Tanke'] | 2021-12-01 | null | null | null | 3dv-2021-12 | ['human-pose-forecasting'] | ['computer-vision'] | [-1.27669588e-01 2.31267229e-01 5.57228662e-02 -3.34538192e-01
-3.54724139e-01 -1.52785361e-01 9.38021302e-01 1.12198897e-01
-3.90811503e-01 8.78335059e-01 4.63510901e-01 9.69669670e-02
1.84896782e-01 -6.29234672e-01 -3.13405693e-01 -2.80230552e-01
-8.18795115e-02 5.78524888e-01 8.60978186e-01 -3.44855398e-01
1.53510943e-01 3.10505927e-01 -1.80802953e+00 3.40007186e-01
6.83138430e-01 6.45216286e-01 4.64924544e-01 9.61919904e-01
3.47331196e-01 7.95553148e-01 -5.56407750e-01 -4.98245239e-01
1.30710110e-01 -8.64750743e-01 -5.96138835e-01 3.79813254e-01
2.39802837e-01 -4.96041685e-01 -2.50402123e-01 7.64240384e-01
3.70728433e-01 4.22710836e-01 5.23795426e-01 -1.38982940e+00
-6.44567609e-02 4.28876936e-01 1.22208064e-02 2.59233806e-02
1.08954525e+00 4.40273881e-01 6.16909325e-01 -4.17841107e-01
9.65399146e-01 1.37587583e+00 6.26852214e-01 5.96258879e-01
-8.55396390e-01 -1.69421494e-01 3.75869334e-01 5.48674703e-01
-1.18535066e+00 -4.08206522e-01 4.86218005e-01 -5.85654140e-01
8.54102731e-01 3.95207345e-01 8.51780593e-01 1.32936203e+00
5.45473158e-01 6.64971471e-01 9.10924554e-01 -4.21333373e-01
4.00270432e-01 9.31167044e-03 -8.03310424e-02 5.09261191e-01
8.46275035e-03 3.46034914e-01 -4.23600584e-01 -6.86502433e-04
5.63737690e-01 4.31483192e-03 -3.07444900e-01 -1.60487100e-01
-1.56688917e+00 5.19637942e-01 -6.21820837e-02 3.97466511e-01
-9.39566135e-01 2.15955764e-01 3.16511363e-01 -6.20895959e-02
1.71563789e-01 3.31812680e-01 1.38187453e-01 -6.54353440e-01
-1.19507933e+00 7.98260510e-01 8.36882651e-01 7.45500684e-01
3.79795760e-01 6.00108244e-02 -4.74387586e-01 1.62607491e-01
3.96301895e-01 4.02143687e-01 4.49793279e-01 -1.25128245e+00
2.43823916e-01 2.06515595e-01 6.96628451e-01 -1.28147507e+00
-4.66228157e-01 -1.64061978e-01 -6.88556373e-01 5.26763022e-01
5.21444678e-01 -2.78682977e-01 -7.67006278e-01 1.46330357e+00
8.12814236e-02 3.66213053e-01 -3.13504376e-02 1.13714457e+00
2.68282473e-01 9.15035188e-01 7.22793117e-02 -2.24491745e-01
1.11900043e+00 -1.33890963e+00 -9.02961910e-01 -3.41325730e-01
3.07650715e-01 -6.52360737e-01 7.94763386e-01 7.16689765e-01
-1.28247488e+00 -1.00468278e+00 -9.55268383e-01 5.85335016e-01
9.77341011e-02 -1.86130553e-01 3.81678909e-01 7.82910168e-01
-1.09180725e+00 8.21324229e-01 -1.08527505e+00 -5.79456270e-01
-2.81083316e-01 -1.20698132e-01 -2.31461644e-01 -3.02329995e-02
-1.19381654e+00 1.19570339e+00 6.39932096e-01 -2.35345185e-01
-8.49014103e-01 -1.40649050e-01 -8.61360788e-01 6.02579117e-02
3.43694031e-01 -9.11227942e-01 1.41523111e+00 -1.04588234e+00
-1.86410558e+00 2.33747452e-01 -2.27730289e-01 -7.88394094e-01
1.02366376e+00 -3.05491537e-01 -4.96825278e-01 1.14659891e-01
-1.36447772e-01 8.04559052e-01 6.02024317e-01 -1.34719753e+00
-9.96098340e-01 4.37910929e-02 3.12539905e-01 3.99346977e-01
4.68104482e-02 2.55102385e-02 -6.13688827e-01 -6.38981521e-01
-2.51750439e-01 -1.15432763e+00 -6.01341605e-01 -2.75061160e-01
-1.44111142e-01 1.06456928e-01 4.99058932e-01 -6.69693649e-01
1.30583060e+00 -1.62822640e+00 2.33142972e-01 -7.90462121e-02
-8.98526683e-02 2.07408726e-01 -2.09852774e-02 7.27504313e-01
2.36337870e-01 -3.32607180e-02 9.36282575e-02 -7.30791986e-01
1.54759631e-01 2.07617342e-01 -1.31366417e-01 3.32288265e-01
-3.42466950e-01 6.73934817e-01 -9.89063025e-01 -5.37661076e-01
6.12560689e-01 4.59211111e-01 -3.85836035e-01 3.96517575e-01
-4.68506306e-01 7.07092047e-01 -2.50294358e-01 2.17305105e-02
3.55206132e-01 -1.89928770e-01 2.50729889e-01 2.79311448e-01
-2.31502146e-01 7.35834539e-02 -1.49299228e+00 1.61777043e+00
-3.30383897e-01 4.00866061e-01 -5.39620340e-01 -2.78404087e-01
6.80304408e-01 6.78380430e-01 4.67376560e-01 -5.60573518e-01
3.29674594e-02 -1.38219446e-01 -1.21050915e-02 -5.70358574e-01
7.99741745e-01 -1.67344749e-01 1.05332755e-01 3.48847479e-01
-4.26927805e-01 -3.76357697e-02 4.62922037e-01 -3.05141555e-03
7.92656958e-01 6.79689646e-01 6.09466910e-01 8.83620456e-02
5.54658532e-01 3.59747559e-02 5.53817034e-01 8.95352185e-01
-2.25719154e-01 8.17623258e-01 2.32471690e-01 -5.05820513e-01
-1.20447695e+00 -7.35183835e-01 6.45834684e-01 6.86239600e-01
4.62514758e-01 -6.44043088e-01 -9.19372380e-01 -4.45609629e-01
-5.82940280e-01 1.10586810e+00 -4.08908993e-01 1.39556631e-01
-6.53654277e-01 -2.25392282e-01 3.40305299e-01 5.02129972e-01
5.76932073e-01 -1.12021267e+00 -1.49472833e+00 4.98424798e-01
-7.94649005e-01 -1.32567465e+00 -5.40987253e-01 -6.92607284e-01
-8.38655472e-01 -7.80787945e-01 -9.42554533e-01 -2.57570565e-01
2.09642068e-01 3.76741514e-02 1.13543236e+00 -3.49369040e-03
1.44390807e-01 3.83657485e-01 -6.26200795e-01 -3.08703601e-01
-8.02178800e-01 -2.79362679e-01 1.03526033e-01 1.08777158e-01
-7.33520556e-03 -3.09921116e-01 -5.39161086e-01 5.49264073e-01
-7.48390853e-01 5.07338107e-01 3.61653239e-01 4.78214502e-01
4.20761347e-01 2.35096559e-01 3.55909407e-01 -4.41049725e-01
6.30638123e-01 -3.10081393e-01 -4.59704220e-01 1.75568879e-01
-5.77709377e-01 1.12704828e-01 7.49410987e-01 -5.45399547e-01
-1.30137575e+00 2.18140975e-01 -4.00159061e-01 -2.73654521e-01
-5.72180212e-01 2.70222843e-01 9.06118676e-02 4.90502477e-01
7.09160984e-01 1.40735701e-01 -2.39764407e-01 -2.02869207e-01
2.61060327e-01 1.73548058e-01 4.82091427e-01 -1.70929164e-01
4.06417042e-01 5.20579815e-01 -1.59373991e-02 -7.67432511e-01
-3.10561627e-01 -4.49508131e-01 -4.79505479e-01 -7.64213443e-01
9.62957740e-01 -7.01981246e-01 -7.42755234e-01 5.06577075e-01
-1.32959878e+00 -4.44099337e-01 -2.04460725e-01 7.66427279e-01
-9.53517616e-01 8.46130848e-01 -4.01685476e-01 -1.31081367e+00
1.23226970e-01 -1.24602962e+00 9.64875877e-01 1.55253127e-01
-6.51449144e-01 -9.40754890e-01 2.93026388e-01 9.11335181e-03
3.11565161e-01 6.40570343e-01 5.66608422e-02 -2.84602255e-01
-7.00392127e-01 -2.84312963e-01 2.61563063e-01 -6.29935116e-02
-9.04396847e-02 -1.06778651e-01 -5.97586095e-01 -1.42462030e-01
5.51489443e-02 1.11576885e-01 6.75212264e-01 5.06387115e-01
4.65345055e-01 -3.71548593e-01 -3.15386504e-01 3.32170948e-02
1.21570432e+00 4.40476358e-01 1.22372019e+00 4.42389637e-01
2.53120482e-01 7.54726112e-01 1.27240479e+00 6.57567799e-01
5.35301208e-01 1.18134081e+00 5.60553908e-01 1.91823110e-01
-1.57623768e-01 -3.04753572e-01 6.06508434e-01 2.65488446e-01
-7.35901713e-01 -9.07581389e-01 -8.36329460e-01 5.85579395e-01
-2.38253927e+00 -1.52306926e+00 -2.47734636e-01 2.31147528e+00
1.95855107e-02 5.01371264e-01 5.92798114e-01 1.86697036e-01
5.49189925e-01 3.85462224e-01 -3.54510248e-02 -2.52883047e-01
1.23023972e-01 -4.07958031e-01 3.52538198e-01 8.32691848e-01
-9.00689363e-01 7.94757783e-01 6.81770182e+00 3.89070481e-01
-8.24045062e-01 -2.09626239e-02 3.61799210e-01 1.85796782e-01
-1.51420072e-01 1.04116015e-01 -7.36645043e-01 6.00090444e-01
1.18314874e+00 -1.22767650e-01 3.60928252e-02 8.02569568e-01
8.25221896e-01 -4.53566879e-01 -9.72982049e-01 8.02895963e-01
2.16354102e-01 -1.32356763e+00 -1.62482113e-01 1.26515076e-01
6.35156810e-01 -4.03037131e-01 -1.85605392e-01 3.68321128e-02
2.08810732e-01 -1.01191545e+00 1.19638908e+00 1.20226479e+00
3.20814162e-01 -8.20042789e-01 9.95996654e-01 8.63077879e-01
-1.12883878e+00 7.57901445e-02 -8.67626816e-02 -3.66329968e-01
1.05773604e+00 3.75717133e-01 -1.07656574e+00 7.51485407e-01
3.98558438e-01 4.77824569e-01 -5.03282726e-01 1.23383176e+00
-1.95644855e-01 4.31784242e-01 -1.30431280e-01 -1.46671891e-01
2.93898582e-01 -1.38255637e-02 7.68992007e-01 1.48792756e+00
6.84712112e-01 3.48549709e-02 6.86186910e-01 5.44638932e-01
7.66348362e-01 -4.71566468e-02 -6.02688909e-01 1.83716878e-01
2.71855351e-02 7.01720178e-01 -6.64307714e-01 -6.16878211e-01
-1.04349606e-01 1.43432546e+00 -1.50886819e-01 1.99001476e-01
-1.09064925e+00 4.96845022e-02 4.04276133e-01 3.21848452e-01
2.72521824e-01 -5.74555218e-01 -3.97034101e-02 -1.04022920e+00
9.90651771e-02 -8.00771713e-01 3.02296460e-01 -9.22475994e-01
-7.53655493e-01 9.95635688e-01 3.36838007e-01 -1.47235680e+00
-1.09226775e+00 -1.21648639e-01 -4.12748188e-01 7.65109003e-01
-1.14283025e+00 -1.06316817e+00 -3.81959409e-01 1.81669757e-01
9.30953622e-01 4.94862050e-02 7.44379580e-01 2.88570627e-05
-4.45271768e-02 4.79169302e-02 -3.50263268e-01 -5.34381390e-01
6.17305517e-01 -1.12991142e+00 8.30711901e-01 1.17437804e+00
1.42230485e-02 3.73889327e-01 1.49403226e+00 -1.08187473e+00
-6.99644089e-01 -7.20427036e-01 1.25188243e+00 -4.24562156e-01
1.28222421e-01 1.13021366e-01 -8.33341956e-01 5.92830539e-01
2.83345670e-01 -4.02348250e-01 1.62077203e-01 -3.36442828e-01
2.39437848e-01 4.08424199e-01 -1.04538488e+00 8.16393971e-01
9.77776647e-01 9.55359861e-02 -6.05008304e-01 1.01128183e-01
4.60622817e-01 -5.23775697e-01 -6.73520148e-01 2.22603694e-01
8.30272019e-01 -1.82275367e+00 9.65616226e-01 -3.58542889e-01
3.41192931e-01 -4.43504423e-01 5.82336262e-03 -1.36257267e+00
-4.06421214e-01 -6.12761259e-01 -2.77704924e-01 6.58777833e-01
1.85122252e-01 -2.42390886e-01 8.87475252e-01 5.44750333e-01
-9.08147171e-02 -3.31419706e-01 -5.79646587e-01 -1.10033500e+00
-5.82580566e-01 -5.35417914e-01 5.64848125e-01 4.47965384e-01
-7.31344288e-03 1.27205014e-01 -1.15182149e+00 1.42542377e-01
4.78985161e-01 -3.00733484e-02 1.02038014e+00 -1.22005916e+00
-5.50170183e-01 -2.05050305e-01 -6.52028620e-01 -1.26543403e+00
-8.11793804e-02 -1.71379298e-01 4.25186783e-01 -1.81530488e+00
4.34338953e-03 1.14626840e-01 2.31064647e-01 6.43263711e-03
-2.55823165e-01 1.69495530e-02 6.24698520e-01 6.06536344e-02
-8.27399552e-01 5.31204402e-01 1.23382711e+00 3.82605076e-01
-3.42760473e-01 2.73959488e-01 -9.27525237e-02 9.38733339e-01
7.15306282e-01 -3.35297853e-01 -5.31159103e-01 -2.32858900e-02
-9.71333459e-02 5.60341358e-01 5.21403432e-01 -1.60938358e+00
2.11694404e-01 -5.33280969e-01 1.77178800e-01 -8.09196949e-01
7.59028196e-01 -9.34919298e-01 8.62635016e-01 7.84923792e-01
-2.75269777e-01 2.62356132e-01 -8.16303566e-02 8.25684607e-01
-1.69843927e-01 -2.92598307e-01 7.29986191e-01 -3.99663597e-01
-1.08816564e+00 1.45241708e-01 -8.16318750e-01 -3.99811327e-01
1.17006814e+00 -5.55519223e-01 1.24515444e-01 -1.17968261e+00
-8.61726642e-01 1.55684605e-01 7.32794940e-01 5.86284459e-01
7.30516553e-01 -1.20320868e+00 -8.73199821e-01 -3.69746715e-01
9.88926589e-02 -5.52703202e-01 4.07779753e-01 5.26668847e-01
-6.98914707e-01 5.91953576e-01 -5.56090474e-01 -5.21097183e-01
-1.32366967e+00 6.01362407e-01 1.58593506e-01 -6.66939616e-01
-9.29579139e-01 1.45373732e-01 -5.33390120e-02 9.96977165e-02
2.96014369e-01 -1.19644314e-01 -4.80672687e-01 -3.19080144e-01
8.40144992e-01 6.83541000e-01 -4.02837604e-01 -9.32424128e-01
-2.57737190e-01 2.69089431e-01 4.28296834e-01 -6.78704202e-01
1.02398241e+00 -6.18397117e-01 4.35457468e-01 7.39115298e-01
4.53909844e-01 3.03465147e-02 -1.59118903e+00 2.01678708e-01
1.90154627e-01 -7.62667656e-01 -5.05828977e-01 -8.08024347e-01
-4.06312108e-01 6.18765056e-01 6.41162574e-01 2.15176299e-01
9.60243940e-01 -3.63345951e-01 8.31512272e-01 2.49685898e-01
7.35206604e-01 -1.03388691e+00 2.41170540e-01 5.35138309e-01
9.81537163e-01 -1.06374407e+00 -7.14363530e-02 -4.69358742e-01
-1.28400767e+00 1.18519151e+00 5.77310264e-01 1.94785912e-02
1.46011651e-01 8.62015113e-02 2.92931939e-03 1.62703633e-01
-7.02444315e-01 -4.63726282e-01 4.57771033e-01 8.23515236e-01
3.49671632e-01 1.30175799e-01 -5.71733713e-01 3.16190094e-01
-1.58086687e-01 5.10377169e-01 8.93473387e-01 7.75673747e-01
-7.08839595e-01 -1.08191073e+00 -5.68548322e-01 -1.37045443e-01
-2.06738770e-01 3.98397684e-01 -1.71000198e-01 6.96729481e-01
5.17543666e-02 1.27761447e+00 -2.03880101e-01 -4.48496431e-01
4.89836156e-01 1.22214988e-01 4.09963638e-01 -4.61613208e-01
-3.88650745e-01 1.47652671e-01 6.11666262e-01 -8.65146935e-01
-6.42858565e-01 -7.38534033e-01 -1.15466309e+00 -2.62810171e-01
9.22855288e-02 6.21418208e-02 3.91613066e-01 9.49928164e-01
1.65296480e-01 6.32165551e-01 2.42921129e-01 -1.16773725e+00
-1.40848860e-01 -8.57421875e-01 -3.02733600e-01 7.07160652e-01
2.18521118e-01 -5.14719963e-01 -2.67543960e-02 3.37978989e-01] | [7.340800762176514, -0.05529000237584114] |
d4b31ec3-df03-4d1e-a3d1-401e4d7b15aa | the-kfiou-loss-for-rotated-object-detection-1 | 2201.12558 | null | https://arxiv.org/abs/2201.12558v6 | https://arxiv.org/pdf/2201.12558v6.pdf | The KFIoU Loss for Rotated Object Detection | Differing from the well-developed horizontal object detection area whereby the computing-friendly IoU based loss is readily adopted and well fits with the detection metrics. In contrast, rotation detectors often involve a more complicated loss based on SkewIoU which is unfriendly to gradient-based training. In this paper, we propose an effective approximate SkewIoU loss based on Gaussian modeling and Gaussian product, which mainly consists of two items. The first term is a scale-insensitive center point loss, which is used to quickly narrow the distance between the center points of the two bounding boxes. In the distance-independent second term, the product of the Gaussian distributions is adopted to inherently mimic the mechanism of SkewIoU by its definition, and show its alignment with the SkewIoU loss at trend-level within a certain distance (i.e. within 9 pixels). This is in contrast to recent Gaussian modeling based rotation detectors e.g. GWD loss and KLD loss that involve a human-specified distribution distance metric which require additional hyperparameter tuning that vary across datasets and detectors. The resulting new loss called KFIoU loss is easier to implement and works better compared with exact SkewIoU loss, thanks to its full differentiability and ability to handle the non-overlapping cases. We further extend our technique to the 3-D case which also suffers from the same issues as 2-D. Extensive results on various public datasets (2-D/3-D, aerial/text/face images) with different base detectors show the effectiveness of our approach. | ['Jirui Yang', 'Qi Tian', 'Xiaopeng Zhang', 'Junchi Yan', 'Wentao Wang', 'Gefan Zhang', 'Yue Zhou', 'Xue Yang'] | 2022-01-29 | the-kfiou-loss-for-rotated-object-detection | https://openreview.net/forum?id=B9LUI0pZFGc | https://openreview.net/pdf?id=B9LUI0pZFGc | null | ['object-detection-in-aerial-images'] | ['computer-vision'] | [-1.49493501e-01 4.65242518e-03 -5.02277119e-03 -1.97034195e-01
-5.02587199e-01 -4.25711423e-01 5.20142674e-01 1.13864444e-01
-5.82735538e-01 3.23572069e-01 -3.18082899e-01 -1.41828150e-01
-7.66038597e-02 -6.93323135e-01 -6.45482063e-01 -8.04636836e-01
-4.07790532e-03 2.02160060e-01 6.07890129e-01 -2.16171011e-01
3.03199977e-01 8.34839463e-01 -1.44993174e+00 -3.95555496e-01
7.43055642e-01 1.18009174e+00 -1.19742483e-01 3.19508493e-01
-9.65712070e-02 -5.56340404e-02 -4.97901976e-01 -4.44588274e-01
5.91360867e-01 -1.51090428e-01 -2.39019200e-01 1.26258686e-01
6.53122962e-01 -4.39829707e-01 -3.15956436e-02 1.03728175e+00
6.10752165e-01 -1.10525385e-01 1.01436746e+00 -1.06366611e+00
-6.20303631e-01 2.87477877e-02 -1.18059492e+00 -4.11331616e-02
3.28951985e-01 -2.87488122e-02 7.57177293e-01 -1.09065759e+00
4.77038890e-01 1.47634006e+00 1.01489019e+00 2.77628839e-01
-1.01810551e+00 -6.00691020e-01 2.26287201e-01 -1.27038017e-01
-1.85203493e+00 1.95903599e-01 6.18138909e-01 -4.30362731e-01
5.94580770e-01 2.47258190e-02 2.27979466e-01 8.12165618e-01
1.69252321e-01 7.95349240e-01 8.96885335e-01 -5.54764032e-01
1.22266755e-01 4.52923834e-01 -6.34520650e-02 5.78870773e-01
4.23038423e-01 8.38530809e-02 1.26933008e-01 -7.77331963e-02
1.02845049e+00 5.65324798e-02 -3.27435881e-01 -8.88608396e-01
-8.54155719e-01 8.42192292e-01 6.29770398e-01 1.01396352e-01
3.28596979e-02 -1.52502686e-01 3.14396441e-01 1.15640108e-02
4.12249118e-01 1.77723333e-01 -3.53329122e-01 9.04948264e-02
-1.09085917e+00 3.09355855e-01 5.73124945e-01 1.04271162e+00
7.22692609e-01 5.10085151e-02 -1.91138729e-01 8.15719545e-01
5.98612070e-01 4.67789620e-01 2.67504007e-01 -3.88707191e-01
3.58691305e-01 6.14562273e-01 1.11418568e-01 -1.07783449e+00
-4.84670162e-01 -7.49304593e-01 -7.28744745e-01 5.36585450e-01
7.98926711e-01 1.10415168e-01 -8.21233571e-01 1.74757814e+00
6.46853030e-01 -3.22354436e-02 -3.84330958e-01 9.35791016e-01
4.75738883e-01 2.35159323e-01 -7.29243383e-02 -9.87626314e-02
1.49922872e+00 -5.16430080e-01 -4.47572172e-01 8.24889988e-02
4.64454472e-01 -8.01064372e-01 1.20398498e+00 4.53773469e-01
-8.95066142e-01 -5.25570452e-01 -1.31100929e+00 -2.38567982e-02
-5.15869975e-01 3.53307039e-01 4.01741654e-01 1.02996445e+00
-1.01650858e+00 3.90828341e-01 -5.46749711e-01 -4.89478856e-01
3.01554263e-01 2.37003073e-01 -1.70176342e-01 3.18546146e-01
-8.97355855e-01 9.04419720e-01 3.02709699e-01 -1.46888658e-01
-4.47231054e-01 -8.17162633e-01 -7.30357170e-01 -2.42210105e-02
4.25135553e-01 -4.32809979e-01 9.32987571e-01 -5.81064939e-01
-1.51496124e+00 8.94587755e-01 1.78990111e-01 -4.43823218e-01
1.15991938e+00 -2.68356472e-01 -8.01964477e-02 2.35299654e-02
-1.34206563e-01 6.38884723e-01 1.10069442e+00 -1.11295414e+00
-3.97303879e-01 -5.62389493e-01 5.85394166e-02 9.92090330e-02
-5.16175807e-01 -6.78709894e-02 -6.08233392e-01 -7.80183375e-01
2.50110328e-01 -7.98760772e-01 5.36405630e-02 4.67164457e-01
-4.83453929e-01 -2.06503510e-01 1.03305566e+00 -2.65394539e-01
1.44324601e+00 -2.30483508e+00 -3.84150237e-01 2.05721766e-01
8.52279291e-02 3.72915298e-01 2.14511961e-01 3.47289890e-01
-1.11143760e-01 4.99369577e-03 -3.76629651e-01 -3.21341664e-01
3.54084671e-02 -2.47714862e-01 -2.87652016e-01 8.58498752e-01
4.24974561e-01 3.52609098e-01 -7.41458833e-01 -4.23259556e-01
2.73866683e-01 7.56410480e-01 -4.62292522e-01 -1.45112783e-01
2.28248566e-01 -7.21352175e-02 -3.21488857e-01 5.90253055e-01
1.23605633e+00 -3.83241922e-02 -3.42576802e-01 -2.68313110e-01
-2.51876473e-01 -1.99238852e-01 -1.61009598e+00 1.23372972e+00
-3.38174194e-01 4.43174452e-01 4.24131937e-02 -7.56962180e-01
1.28232193e+00 5.31273074e-02 2.37973705e-01 -2.51349986e-01
2.27382258e-01 2.02719167e-01 -2.63292670e-01 1.72982384e-02
5.15109241e-01 -1.04600132e-01 1.40931860e-01 4.55041463e-03
-1.23225503e-01 -1.85720593e-01 3.63146104e-02 -9.01112854e-02
7.06130922e-01 4.15654451e-01 5.35350442e-01 -4.00694430e-01
5.17811894e-01 -3.80851299e-01 5.69100857e-01 7.39871800e-01
-4.21998501e-01 1.09259117e+00 5.39297104e-01 -2.37951800e-02
-9.86836255e-01 -1.24901199e+00 -8.24719191e-01 5.01853287e-01
3.99754882e-01 -1.27957940e-01 -8.35730016e-01 -7.69011378e-01
2.38712743e-01 6.14219666e-01 -5.33750296e-01 -2.18678284e-02
-5.31628966e-01 -7.66876817e-01 5.44826388e-01 6.43822074e-01
5.94552994e-01 -3.36997479e-01 -6.08321965e-01 -1.27440952e-02
5.03105104e-01 -1.03990662e+00 -6.51542962e-01 2.60079242e-02
-8.30547810e-01 -1.06467235e+00 -9.86004353e-01 -3.47380400e-01
7.21854031e-01 3.32749993e-01 6.84008777e-01 -3.27265471e-01
-5.29872060e-01 3.02399307e-01 -3.34594280e-01 -5.99254847e-01
1.65919829e-02 -1.20616861e-01 1.34640187e-01 4.54361774e-02
3.75339001e-01 -2.77370065e-01 -7.85172343e-01 5.97344398e-01
-9.39972043e-01 -3.15354228e-01 6.16214097e-01 7.12828577e-01
4.82015848e-01 -9.52181369e-02 4.24280465e-01 -5.01584291e-01
3.83841455e-01 -3.84040773e-01 -9.01218653e-01 2.16731414e-01
-7.71892607e-01 9.89019498e-03 4.90548193e-01 -7.10848272e-01
-8.83196473e-01 -2.77371369e-02 1.16073608e-01 -5.74294925e-01
-6.62240610e-02 -8.03973526e-02 -1.02830686e-01 -2.21422955e-01
6.97255433e-01 -8.09785947e-02 8.60145837e-02 -4.86949205e-01
3.90032679e-01 5.97983539e-01 3.78447056e-01 -2.78806895e-01
1.04971111e+00 7.68767893e-01 2.41719916e-01 -9.65851545e-01
-4.25120234e-01 -6.05520129e-01 -4.28424835e-01 -1.11991428e-01
7.15830088e-01 -6.98219180e-01 -7.99065173e-01 4.30736154e-01
-1.00654185e+00 2.97743231e-01 -3.39888722e-01 6.44432485e-01
-4.11572695e-01 7.80236304e-01 -5.71407974e-01 -1.16165781e+00
-1.80009723e-01 -1.15757704e+00 1.33281589e+00 3.29466969e-01
5.77745661e-02 -9.40966368e-01 -1.03621498e-01 -2.17343867e-01
3.65878195e-01 4.51030254e-01 7.48635948e-01 -5.00530362e-01
-3.92898321e-01 -6.34698391e-01 -5.67519665e-01 6.86204314e-01
8.96393657e-02 3.29083771e-01 -9.59410846e-01 -4.61111426e-01
1.10537693e-01 -7.82799423e-02 9.00461853e-01 5.37839532e-01
9.52690065e-01 1.53869614e-01 -2.98734039e-01 5.56517541e-01
1.52408659e+00 2.06261710e-03 6.44298494e-01 4.06981587e-01
5.06151438e-01 5.45995057e-01 9.20719504e-01 6.03551686e-01
2.35213935e-01 1.05934322e+00 6.11361921e-01 -3.47859949e-01
-6.71439767e-02 -1.39230728e-01 4.14404958e-01 2.99448639e-01
8.31527859e-02 -9.40527841e-02 -7.18640983e-01 2.16017216e-01
-1.65832090e+00 -7.21159220e-01 -5.59623688e-02 2.67675257e+00
5.94199479e-01 3.53381693e-01 3.11687797e-01 1.97892398e-01
7.72130668e-01 1.03874415e-01 -4.63109791e-01 -4.81219292e-01
-2.17798024e-01 -1.44596267e-02 8.86374772e-01 3.38287592e-01
-1.27135408e+00 7.34043062e-01 6.07194042e+00 1.33564353e+00
-1.16297793e+00 -7.62121333e-03 4.63380635e-01 -2.20076740e-02
7.03878403e-02 -1.87293179e-02 -1.33117855e+00 3.87690663e-01
3.21783453e-01 8.45093951e-02 -1.00665741e-01 1.02073061e+00
3.29861641e-01 -1.46734044e-01 -9.41243112e-01 1.01064038e+00
1.02782100e-02 -6.56698167e-01 -9.07503739e-02 2.00590581e-01
4.83738661e-01 -2.26974830e-01 3.16486984e-01 2.15304136e-01
-1.52873665e-01 -7.54966378e-01 9.59323227e-01 1.75809368e-01
7.81716704e-01 -7.07582355e-01 4.96893167e-01 2.97387838e-01
-1.25810075e+00 -5.09687699e-02 -4.88138825e-01 1.73442647e-01
6.00646483e-03 6.50206506e-01 -1.01004922e+00 6.65390968e-01
6.58657134e-01 3.86942059e-01 -5.81563532e-01 1.36456478e+00
-1.33798927e-01 2.07600027e-01 -6.27382219e-01 -6.40755743e-02
3.70639652e-01 -5.10414898e-01 8.05143893e-01 1.38835716e+00
4.29920375e-01 -5.96529841e-01 5.12689613e-02 9.12185609e-01
1.89891353e-01 2.54194140e-01 -4.24088746e-01 5.50195634e-01
5.48719823e-01 1.29396594e+00 -8.00931275e-01 -5.71007356e-02
-4.39004451e-01 6.99004412e-01 3.61046270e-02 4.04414207e-01
-1.16033423e+00 -6.37924492e-01 5.75679541e-01 5.29581368e-01
6.51992917e-01 -2.49022529e-01 -2.74588734e-01 -8.43797028e-01
4.22831208e-01 -4.57618088e-01 2.08035365e-01 -3.03546160e-01
-1.20258284e+00 5.53274095e-01 3.63036960e-01 -1.76142526e+00
-1.43779501e-01 -1.05159628e+00 -6.16903245e-01 7.67213702e-01
-1.69451880e+00 -1.13513076e+00 -2.88000792e-01 2.42708355e-01
2.62849987e-01 1.68773279e-01 3.73466909e-01 4.98595059e-01
-5.68748593e-01 9.57446575e-01 1.46581545e-01 9.82635692e-02
9.17520583e-01 -1.29449522e+00 1.83095470e-01 8.51954401e-01
-1.70003563e-01 7.02204525e-01 8.33426356e-01 -3.56588811e-01
-1.03401816e+00 -9.06918645e-01 5.08679748e-01 -2.92695910e-01
5.15294731e-01 -4.42765325e-01 -8.82539332e-01 2.56203532e-01
-2.27423236e-01 1.18175261e-01 2.40588829e-01 -1.88805327e-01
-5.78711808e-01 -2.02412575e-01 -1.42200387e+00 5.30404985e-01
9.93947864e-01 -2.36884698e-01 -2.98656553e-01 1.43919736e-01
3.84315908e-01 -3.00756603e-01 -6.61369920e-01 6.97895229e-01
5.57738841e-01 -1.21802974e+00 1.01964378e+00 -1.30277351e-01
-1.56653285e-01 -7.07276464e-01 -6.05742000e-02 -8.49737346e-01
-1.76550254e-01 -5.21416366e-01 1.02474414e-01 1.33103943e+00
3.72647166e-01 -9.02194202e-01 6.34264350e-01 3.20276290e-01
6.06079847e-02 -9.93216693e-01 -1.11743557e+00 -1.26811814e+00
5.44239953e-02 -4.65344340e-01 4.25849408e-01 6.90447390e-01
-3.00417572e-01 -1.43655717e-01 -1.30572155e-01 1.74536198e-01
7.12351501e-01 -2.43628442e-01 7.70055711e-01 -1.13663602e+00
-3.61912847e-01 -7.66282737e-01 -7.89075613e-01 -1.28140640e+00
-3.64735335e-01 -5.16703069e-01 -2.10054144e-01 -9.89225447e-01
-1.99263152e-02 -5.08585095e-01 -1.38381749e-01 4.49226461e-02
-1.04072496e-01 1.97661459e-01 -3.93987074e-03 1.65465668e-01
-1.74604803e-01 6.55074000e-01 1.03345680e+00 1.34724513e-01
-3.32516015e-01 1.43479869e-01 -4.57417369e-01 9.90107954e-01
5.41284859e-01 -3.28900546e-01 -3.46915990e-01 -1.29942104e-01
-6.89233094e-02 -3.93284410e-01 4.53804225e-01 -9.99980271e-01
1.05732925e-01 1.78727180e-01 3.96640271e-01 -6.65668666e-01
4.47577268e-01 -7.64752209e-01 -2.06032842e-01 3.36896867e-01
1.41321898e-01 -6.72889650e-02 1.23027764e-01 5.23744643e-01
-1.93648145e-01 -2.57713795e-01 1.16400254e+00 2.22825557e-01
-4.47809964e-01 3.82081866e-01 1.90312728e-01 -1.61111385e-01
1.32374561e+00 -8.23940217e-01 -2.56245643e-01 -2.57097065e-01
-3.37424129e-01 -1.04841087e-02 6.21511817e-01 3.77549738e-01
6.06537461e-01 -1.18520367e+00 -7.19925165e-01 1.72172368e-01
2.13433519e-01 -7.65468553e-02 5.25856279e-02 1.15744781e+00
-6.49327934e-01 2.88737953e-01 1.83250517e-01 -8.23180318e-01
-1.07817590e+00 5.56359053e-01 3.19646239e-01 -2.08155289e-01
-5.30571282e-01 8.23147893e-01 5.80105960e-01 -2.45323598e-01
4.69698906e-01 -5.52440166e-01 -9.23693851e-02 1.25261128e-01
5.50596833e-01 5.63266277e-01 9.39393491e-02 -4.76630211e-01
-5.55025339e-01 8.67124677e-01 -1.68254569e-01 1.92126492e-03
8.30496669e-01 -3.54343653e-02 2.35329077e-01 2.75619894e-01
1.26469445e+00 1.18614815e-01 -1.66564763e+00 -1.80752128e-01
1.15479469e-01 -5.61495423e-01 7.02480506e-03 -3.54980677e-01
-8.44243228e-01 8.58106911e-01 9.84099329e-01 2.90235758e-01
1.16863728e+00 -1.16660148e-01 4.39787745e-01 -9.56843346e-02
2.39357665e-01 -1.24246633e+00 -2.34238990e-02 2.27540344e-01
9.36132252e-01 -1.14119375e+00 2.22114190e-01 -6.84009910e-01
-4.07159418e-01 1.11559737e+00 6.20170474e-01 -2.49793142e-01
6.76110148e-01 2.10145026e-01 2.78205816e-02 1.80960566e-01
-7.70340934e-02 -2.42865190e-01 3.58652115e-01 5.16105294e-01
3.67599100e-01 -1.08118989e-01 -5.46415567e-01 2.47907519e-01
3.60543877e-02 -2.41720259e-01 2.13252172e-01 7.23249078e-01
-5.47382355e-01 -1.03470707e+00 -5.74198127e-01 1.22300975e-01
-4.04963702e-01 1.44952193e-01 -7.16257021e-02 1.22231555e+00
1.12963296e-01 7.40057468e-01 6.38350919e-02 -1.76965237e-01
5.57454109e-01 -2.74305016e-01 4.67631966e-01 -3.18792611e-01
-2.82871008e-01 2.68414050e-01 -4.41939771e-01 -5.55922925e-01
-2.07152426e-01 -6.83976769e-01 -1.19554424e+00 -1.49810418e-01
-7.43390203e-01 -2.30786622e-01 8.86623263e-01 6.01990759e-01
2.80979633e-01 4.42881212e-02 7.82281399e-01 -7.46840179e-01
-1.17497051e+00 -9.56542194e-01 -8.36466074e-01 3.62855673e-01
2.22451612e-01 -1.01764882e+00 -6.12771749e-01 -4.16525394e-01] | [8.586236000061035, -0.8714494705200195] |
aef0b466-1b32-401f-95af-8788705d95ce | suggesting-relevant-questions-for-a-query | 2204.12069 | null | https://arxiv.org/abs/2204.12069v1 | https://arxiv.org/pdf/2204.12069v1.pdf | Suggesting Relevant Questions for a Query Using Statistical Natural Language Processing Technique | Suggesting similar questions for a user query has many applications ranging from reducing search time of users on e-commerce websites, training of employees in companies to holistic learning for students. The use of Natural Language Processing techniques for suggesting similar questions is prevalent over the existing architecture. Mainly two approaches are studied for finding text similarity namely syntactic and semantic, however each has its draw-backs and fail to provide the desired outcome. In this article, a self-learning combined approach is proposed for determining textual similarity that introduces a robust weighted syntactic and semantic similarity index for determining similar questions from a predetermined database, this approach learns the optimal combination of the mentioned approaches for a database under consideration. Comprehensive analysis has been carried out to justify the efficiency and efficacy of the proposed approach over the existing literature. | ['Onkar Litake', 'Sheetal Sonawane', 'Archana Ghotkar', 'Hrushabh Hirudkar', 'Anuj Kanetkar', 'Shriniwas Nayak'] | 2022-04-26 | null | null | null | null | ['self-learning'] | ['natural-language-processing'] | [ 1.31874084e-01 -4.37922515e-02 -1.15162559e-01 -5.58351099e-01
-8.22978854e-01 -5.64103127e-01 6.45482898e-01 9.07074273e-01
-4.89849538e-01 3.77901077e-01 3.86773169e-01 -2.49976918e-01
-1.03564978e+00 -6.37766123e-01 -1.15521505e-01 -8.09306279e-02
1.77762821e-01 5.02045870e-01 6.95356786e-01 -4.98996526e-01
1.31081593e+00 3.63405436e-01 -1.83662295e+00 3.74589026e-01
1.04278970e+00 8.35738540e-01 6.22627497e-01 5.15308917e-01
-9.43754077e-01 6.08567595e-01 -3.05879384e-01 -4.77371126e-01
1.74388587e-02 -4.84890074e-01 -1.32678306e+00 2.33348478e-02
3.66062850e-01 1.17052324e-01 1.38539329e-01 1.13656878e+00
4.35565412e-01 5.14695525e-01 8.67301822e-01 -8.32232058e-01
-6.60621524e-01 3.63293201e-01 -4.42151874e-01 6.41801476e-01
9.09479558e-01 -4.92721558e-01 1.13616931e+00 -8.44224274e-01
3.53296220e-01 1.22194302e+00 5.98156810e-01 -4.63886671e-02
-6.50881946e-01 -2.60152936e-01 -5.43995015e-02 4.20643866e-01
-1.28669882e+00 7.08470866e-02 7.82220304e-01 -2.81094521e-01
1.06714606e+00 2.97984093e-01 3.04948330e-01 3.57902050e-02
2.21953779e-01 3.73211473e-01 1.10609031e+00 -7.70584226e-01
2.68643945e-01 8.15780997e-01 7.90365577e-01 5.78000307e-01
3.45566273e-01 -4.76784647e-01 -3.67929906e-01 -3.01817954e-01
5.81261069e-02 1.88717186e-01 3.14962357e-01 -1.91991761e-01
-5.25361121e-01 9.02608633e-01 -4.70230309e-03 9.96689439e-01
-4.39561605e-01 -6.08506799e-01 6.25029325e-01 4.63005811e-01
2.74987370e-02 7.47215569e-01 -3.67032558e-01 1.10533178e-01
-9.35073733e-01 5.31085551e-01 8.33838165e-01 1.18084991e+00
8.09036195e-01 -3.98925006e-01 -1.18644811e-01 7.15145290e-01
5.14164865e-01 -6.77629113e-02 1.04549897e+00 -7.42428482e-01
3.20999980e-01 1.12091732e+00 -1.69272080e-01 -1.38541317e+00
-2.55699307e-01 -1.95980102e-01 -1.81420371e-01 -3.13606322e-01
1.79089546e-01 -4.03633192e-02 -3.55755538e-01 1.22141087e+00
4.69373763e-01 -2.94934362e-01 4.07254905e-01 5.39836168e-01
1.20832527e+00 7.25685239e-01 2.33853385e-01 -2.45071411e-01
1.66079283e+00 -8.07518005e-01 -6.85677767e-01 -1.09076656e-01
4.87591565e-01 -1.43938398e+00 1.14687920e+00 2.45504528e-01
-9.33550298e-01 -7.07674384e-01 -8.23480368e-01 -9.48378220e-02
-6.81742013e-01 -1.41993761e-01 2.23976582e-01 5.53105354e-01
-8.18468034e-01 6.29885972e-01 -8.11583176e-02 -1.21697891e+00
-2.19774202e-01 4.38074946e-01 6.09234311e-02 -1.88743547e-02
-1.12618220e+00 1.06539893e+00 7.29085922e-01 -4.92399007e-01
-1.17488056e-01 -5.27624071e-01 -4.81304556e-01 1.03880458e-01
3.57280433e-01 -5.60807467e-01 1.20833933e+00 -7.48030007e-01
-1.05139256e+00 7.98650861e-01 -2.36062512e-01 -3.61421973e-01
2.64762547e-02 -2.21427474e-02 -4.36148584e-01 3.09164971e-01
2.49131516e-01 1.95734903e-01 4.27369177e-01 -8.84337962e-01
-9.44747031e-01 -5.53690255e-01 1.83307782e-01 6.62083924e-01
-7.19722927e-01 5.08221269e-01 -3.50484222e-01 -5.89373410e-01
3.76733959e-01 -5.14531851e-01 -2.05609694e-01 -3.75396132e-01
1.01273857e-01 -7.84869730e-01 8.68566155e-01 -7.72811770e-01
1.55376303e+00 -1.49231458e+00 -3.52555722e-01 5.23174107e-01
-1.59746885e-01 2.46291578e-01 1.86530650e-01 1.09285927e+00
4.75720912e-02 1.45635486e-01 1.87248774e-02 3.95167172e-01
-5.75032830e-02 -1.56958431e-01 -1.48045952e-02 1.24082696e-02
-1.09545708e-01 2.68701732e-01 -7.71327019e-01 -1.11944795e+00
1.26661330e-01 -4.08805325e-04 -5.73763430e-01 3.35968345e-01
9.22028422e-02 -7.43093863e-02 -8.75739396e-01 4.49464858e-01
3.21528852e-01 -7.71360379e-03 9.08110589e-02 -2.78068453e-01
9.60800424e-03 2.29354292e-01 -1.39899683e+00 1.48982775e+00
-4.54634041e-01 1.58947036e-01 -3.35451931e-01 -1.26189244e+00
1.27815783e+00 3.39831144e-01 5.99728823e-01 -8.49330008e-01
3.72498959e-01 3.28849405e-01 -6.95796981e-02 -1.16692472e+00
7.30759263e-01 -2.58802623e-01 3.33639868e-02 5.99670887e-01
2.22500429e-01 -8.94900858e-02 4.74934697e-01 2.96600312e-01
8.30374718e-01 2.15748139e-02 7.29436755e-01 -6.17380083e-01
1.17359805e+00 3.45774174e-01 -7.25908652e-02 7.54186571e-01
-1.42874822e-01 8.79350901e-02 -1.14140891e-01 -6.24010265e-02
-9.37993586e-01 -6.17811322e-01 6.85756505e-02 1.18830574e+00
1.99766025e-01 -3.90596241e-01 -8.41767907e-01 -5.20295560e-01
-1.76241383e-01 7.22198486e-01 -3.01964656e-02 -2.28753015e-01
-3.60736996e-01 -2.22524796e-02 1.39738560e-01 -2.57560841e-05
4.78412390e-01 -1.08194005e+00 -7.57900536e-01 1.39078334e-01
-3.33123356e-02 -8.23733389e-01 -3.25961858e-01 5.81642240e-02
-1.19972479e+00 -1.00432777e+00 -5.46505570e-01 -1.33349979e+00
6.23925626e-01 6.01735830e-01 9.22569454e-01 1.42269522e-01
-2.85382509e-01 6.72062218e-01 -6.77509189e-01 -5.85933268e-01
-3.22687894e-01 2.97784179e-01 -1.13811031e-01 -2.57046998e-01
8.74700308e-01 -4.75342333e-01 -6.77129030e-01 1.79333881e-01
-1.05807579e+00 -6.42918825e-01 5.89049697e-01 5.31478047e-01
2.45840952e-01 4.00586486e-01 9.22639966e-01 -8.96965325e-01
1.37021244e+00 -8.79089534e-01 -2.51896471e-01 7.11145043e-01
-1.15625298e+00 3.71845096e-01 7.59256423e-01 -1.04617268e-01
-1.22362125e+00 -1.12283953e-01 -1.68926075e-01 2.24373251e-01
-2.91205525e-01 6.14436507e-01 1.30501673e-01 -2.09019586e-01
5.86774766e-01 3.92930835e-01 -7.33487457e-02 -4.83513176e-01
2.88546145e-01 8.21146071e-01 4.48436111e-01 -7.41898358e-01
5.70195198e-01 -1.48713857e-01 -2.06897318e-01 -7.61072278e-01
-6.87331319e-01 -1.28729856e+00 -5.30807018e-01 -3.27859253e-01
6.65687025e-01 -2.12208331e-01 -9.09789622e-01 -2.74131715e-01
-1.04365873e+00 9.34905052e-01 -8.65817666e-02 5.05976379e-01
-4.27967966e-01 7.22293913e-01 -1.31359892e-02 -8.48162830e-01
-6.98426306e-01 -8.34347844e-01 4.53676760e-01 6.38907492e-01
-5.92153668e-01 -1.07494330e+00 3.17352340e-02 5.96948504e-01
4.95629936e-01 -2.60351330e-01 1.42846811e+00 -1.45933020e+00
-7.25180060e-02 -3.88928622e-01 -4.90047969e-02 7.79845789e-02
4.27210301e-01 -2.34700546e-01 -3.83201033e-01 -1.10973649e-01
3.90872121e-01 -4.56925064e-01 2.19409645e-01 1.37346620e-02
9.35578585e-01 -5.92709005e-01 -1.75279856e-01 -2.32648164e-01
1.73767447e+00 6.46420062e-01 1.88809857e-01 6.56053960e-01
3.72869670e-02 1.30288088e+00 9.73674655e-01 6.22517943e-01
2.87492663e-01 3.39803398e-01 7.25486204e-02 3.99515480e-01
1.62364006e-01 -2.11532474e-01 -1.03747562e-01 8.45810115e-01
3.40739876e-01 -6.43794686e-02 -8.95660818e-01 6.82055712e-01
-1.60178983e+00 -1.09043419e+00 5.22023886e-02 2.14446902e+00
6.72148407e-01 1.12938486e-01 2.16813505e-01 4.17224288e-01
8.23416710e-01 -1.81659698e-01 -3.00466925e-01 -8.40211511e-01
3.04910183e-01 3.50452811e-01 1.25571489e-01 4.48564589e-01
-5.18768907e-01 7.58569479e-01 5.70432854e+00 9.94483471e-01
-7.90829480e-01 -8.51636976e-02 2.97040492e-01 4.27866310e-01
-6.11639738e-01 2.90226996e-01 -8.08782995e-01 2.08531007e-01
8.98216546e-01 -8.83756220e-01 1.34220734e-01 8.16276431e-01
3.74374330e-01 -3.90329570e-01 -7.44026780e-01 8.39773118e-01
3.76600176e-01 -1.09360266e+00 5.63330531e-01 -4.38869178e-01
6.45732224e-01 -5.99576712e-01 -1.44429997e-01 1.06411055e-01
4.53964211e-02 -5.79301953e-01 1.58828914e-01 5.24024189e-01
-2.25178003e-01 -1.03204072e+00 8.55349422e-01 5.21635056e-01
-1.10108709e+00 -1.76862270e-01 -4.01699394e-01 -1.78490933e-02
-1.59009635e-01 -8.76838341e-02 -1.11430025e+00 6.98844433e-01
6.56016231e-01 4.93056625e-02 -8.82550180e-01 1.12622988e+00
5.78878164e-01 3.41230124e-01 -1.44041672e-01 -6.99321151e-01
5.50710380e-01 -3.60057384e-01 4.54362035e-01 1.08558989e+00
6.06993437e-01 2.11384669e-01 1.41994819e-01 5.11219442e-01
3.41034532e-01 1.15693796e+00 -7.52194703e-01 2.04404332e-02
5.25150597e-01 1.19559705e+00 -9.12689209e-01 -3.46655399e-01
-4.43353742e-01 6.64951980e-01 1.02035508e-01 -1.32513970e-01
-3.40333849e-01 -9.37182128e-01 -1.61562413e-01 3.97087425e-01
-6.16373145e-04 6.20973036e-02 -2.32425198e-01 -2.87929058e-01
2.04018921e-01 -9.53984976e-01 9.14282382e-01 -6.21785700e-01
-1.14143288e+00 5.24339318e-01 3.84746611e-01 -1.21762240e+00
-2.44924158e-01 -4.00213629e-01 -8.00036311e-01 7.45588422e-01
-1.36752546e+00 -7.90213645e-01 -4.20583665e-01 6.76683307e-01
1.05263853e+00 -5.94530582e-01 5.71889222e-01 3.04210961e-01
-2.02170640e-01 3.63205731e-01 2.32453987e-01 -4.84703362e-01
8.41255844e-01 -1.10299277e+00 -2.09732443e-01 7.93307006e-01
1.29024401e-01 1.06676209e+00 8.55511010e-01 -6.49699986e-01
-1.32413757e+00 -5.58648944e-01 1.57696629e+00 3.40703689e-02
6.78624272e-01 5.60659707e-01 -1.03893292e+00 3.31862569e-02
7.73419023e-01 -6.84027553e-01 9.25376236e-01 -2.98788279e-01
4.82103303e-02 -3.29367787e-01 -1.28101754e+00 5.40972769e-01
3.84544075e-01 -3.43916655e-01 -1.21065617e+00 4.56546545e-01
5.32519579e-01 1.71432748e-01 -9.21689332e-01 8.04644376e-02
4.04503018e-01 -1.13671517e+00 9.09492671e-01 -8.70064557e-01
2.83555657e-01 -3.12491387e-01 -1.68922588e-01 -7.66067266e-01
-2.07146674e-01 -3.94074559e-01 3.78845066e-01 1.70888150e+00
4.19159710e-01 -3.16805959e-01 8.97924185e-01 6.77680850e-01
4.17501554e-02 -7.96312094e-01 -3.40331554e-01 -5.21035016e-01
-1.20180495e-01 -4.95748371e-02 3.48665923e-01 8.18621457e-01
1.89723045e-01 6.28753424e-01 1.17794782e-01 -1.89903267e-02
3.75012159e-01 5.48759475e-02 3.64891112e-01 -1.25631487e+00
-2.28837952e-02 -5.17808795e-01 -2.82633036e-01 -8.07791114e-01
-7.68075231e-03 -1.03572929e+00 -1.43558949e-01 -1.73424685e+00
2.21996576e-01 -2.21820712e-01 -2.94808298e-01 -1.97577730e-01
-3.03475797e-01 -2.47472182e-01 1.05259821e-01 2.32343569e-01
-6.28575265e-01 9.83014926e-02 8.78338635e-01 -1.46230338e-02
-1.88345268e-01 4.00743246e-01 -1.02998817e+00 7.70273328e-01
8.47312450e-01 -6.62343144e-01 -1.03702247e+00 -7.97443166e-02
2.57211626e-01 1.17226027e-01 -1.09607808e-01 -9.90555763e-01
8.37618530e-01 -2.59554684e-01 2.19734502e-03 -6.85328305e-01
-2.36069605e-01 -1.17398298e+00 -1.86522305e-01 5.10246396e-01
-8.81653428e-01 7.94101834e-01 -3.90732363e-02 5.91522217e-01
-6.04423404e-01 -1.35212588e+00 6.34077191e-01 -5.30911863e-01
-8.90779555e-01 -1.49750590e-01 -4.20300215e-01 3.27487111e-01
1.09483206e+00 -6.33624911e-01 3.43791425e-01 -5.14123619e-01
-3.42697829e-01 2.61211723e-01 -7.12063462e-02 5.36710858e-01
6.77720547e-01 -1.09899545e+00 -5.37330687e-01 -5.20886123e-01
5.25228381e-01 -4.81369317e-01 2.25759506e-01 4.10786122e-01
-6.74367249e-01 7.69491553e-01 -3.51042718e-01 -8.55584890e-02
-1.63303077e+00 5.90976119e-01 -1.34159103e-01 -5.23883522e-01
-6.09919950e-02 5.00388026e-01 -4.19931740e-01 -4.32998449e-01
4.49777693e-01 -3.40413637e-02 -1.05152678e+00 1.46764368e-01
3.52283388e-01 6.08462512e-01 2.39656031e-01 -5.74663579e-01
-1.84778273e-01 8.99537385e-01 -3.09093386e-01 -7.19100907e-02
9.56180692e-01 -5.37587166e-01 -1.82382077e-01 3.53651911e-01
1.36577308e+00 -1.56837523e-01 -9.23267230e-02 -5.79745054e-01
9.29151297e-01 -4.11243349e-01 -2.32028469e-01 -5.01100838e-01
-4.53376174e-01 6.30224288e-01 8.13450813e-01 4.41841125e-01
1.15646434e+00 -4.99696238e-03 8.24256420e-01 9.61732149e-01
-2.71174014e-02 -1.42383456e+00 3.35690439e-01 4.79210287e-01
6.17004275e-01 -1.35619414e+00 1.64396897e-01 -3.42041999e-01
-6.45602107e-01 1.53199744e+00 7.43287683e-01 -6.87570199e-02
6.87124610e-01 -1.58976823e-01 -1.66159660e-01 -4.36646968e-01
-4.69246119e-01 -2.62136638e-01 4.86359775e-01 3.73838753e-01
8.60150456e-01 -7.43792534e-01 -1.38979745e+00 3.90105873e-01
-2.79443353e-01 -1.68891713e-01 2.40524098e-01 1.26019752e+00
-1.03120732e+00 -1.15530753e+00 -3.92426699e-01 5.59774518e-01
-6.54206753e-01 -9.41123068e-02 -5.59145868e-01 5.60414255e-01
6.58634529e-02 1.30987263e+00 -2.06297293e-01 -2.63647199e-01
5.17433286e-01 3.18106771e-01 1.07693791e-01 -7.07654059e-01
-1.03065896e+00 -2.58669227e-01 4.61254306e-02 -1.15985833e-02
-6.65116310e-01 -6.12943769e-01 -1.26380181e+00 -1.58559814e-01
-3.95049423e-01 8.38142991e-01 7.33528554e-01 1.21364594e+00
2.94406980e-01 1.22670271e-01 8.27551246e-01 -5.19956015e-02
-7.65993774e-01 -8.78647029e-01 -3.26855183e-01 5.09544969e-01
-2.17548534e-01 -2.62020558e-01 -5.84236458e-02 3.72114144e-02] | [10.54991626739502, 7.612081527709961] |
b779652e-83cb-4c9e-bfa2-7694b3e250ea | self-chained-image-language-model-for-video | 2305.06988 | null | https://arxiv.org/abs/2305.06988v1 | https://arxiv.org/pdf/2305.06988v1.pdf | Self-Chained Image-Language Model for Video Localization and Question Answering | Recent studies have shown promising results on utilizing pre-trained image-language models for video question answering. While these image-language models can efficiently bootstrap the representation learning of video-language models, they typically concatenate uniformly sampled video frames as visual inputs without explicit language-aware, temporal modeling. When only a portion of a video input is relevant to the language query, such uniform frame sampling can often lead to missing important visual cues. Although humans often find a video moment to focus on and rewind the moment to answer questions, training a query-aware video moment localizer often requires expensive annotations and high computational costs. To address this issue, we propose Self-Chained Video Localization-Answering (SeViLA), a novel framework that leverages a single image-language model (BLIP-2) to tackle both temporal keyframe localization and QA on videos. SeViLA framework consists of two modules: Localizer and Answerer, where both are parameter-efficiently fine-tuned from BLIP-2. We chain these modules for cascaded inference and self-refinement. First, in the forward chain, the Localizer finds multiple language-aware keyframes in a video, which the Answerer uses to predict the answer. Second, in the reverse chain, the Answerer generates keyframe pseudo-labels to refine the Localizer, alleviating the need for expensive video moment localization annotations. SeViLA outperforms several strong baselines/previous works on five video QA and event prediction tasks, and achieves the state-of-the-art in both fine-tuning (NExT-QA, STAR) and zero-shot (NExT-QA, STAR, How2QA, VLEP) settings. We show a comprehensive analysis, e.g., the impact of Localizer, comparisons of Localizer with other temporal localization models, pre-training/self-refinement of Localizer, and varying the number of keyframes. | ['Mohit Bansal', 'Prateek Yadav', 'Jaemin Cho', 'Shoubin Yu'] | 2023-05-11 | null | null | null | null | ['moment-retrieval', 'video-question-answering'] | ['computer-vision', 'computer-vision'] | [-9.31785032e-02 -3.56666803e-01 -4.42656040e-01 -2.25184575e-01
-1.28261375e+00 -7.87599385e-01 4.06169444e-01 4.01309580e-02
-5.68200171e-01 3.81747991e-01 2.99977124e-01 -1.90290064e-01
3.93035382e-01 -5.04761815e-01 -1.11220539e+00 -3.97069633e-01
5.69905266e-02 2.93760210e-01 7.00135946e-01 1.35428295e-01
1.01233043e-01 1.75115719e-01 -1.50761354e+00 6.97198629e-01
2.60307461e-01 1.03338540e+00 5.28480351e-01 1.14051950e+00
-1.57203212e-01 1.55249619e+00 -4.43789124e-01 -1.95986480e-01
2.46517640e-02 -4.27653372e-01 -8.86373520e-01 1.76429331e-01
7.59041488e-01 -6.21932685e-01 -5.96232712e-01 6.44310176e-01
2.70248234e-01 4.00019616e-01 1.70187712e-01 -1.30205154e+00
-6.19434834e-01 2.12983087e-01 -5.07123470e-01 6.54830694e-01
9.06054020e-01 4.58578497e-01 9.94158030e-01 -1.09592330e+00
1.00655437e+00 1.25340080e+00 4.72731799e-01 4.71610248e-01
-9.01255012e-01 -7.33760238e-01 5.57780564e-01 6.69399559e-01
-1.54955757e+00 -5.86217940e-01 6.65364087e-01 -4.29705590e-01
9.53915477e-01 -1.71248671e-02 6.08264863e-01 1.20275187e+00
1.41569898e-01 1.00497198e+00 5.82091093e-01 -1.44547448e-01
2.05779761e-01 -7.75116906e-02 -1.41213253e-01 1.06192899e+00
-6.60295963e-01 -1.68095142e-01 -9.60102379e-01 -1.99597757e-02
8.01594615e-01 2.31935754e-01 -4.31105465e-01 -2.31073439e-01
-1.48758280e+00 6.11431658e-01 1.77412301e-01 1.57355845e-01
-5.19449592e-01 4.86619890e-01 3.89359355e-01 3.01016808e-01
4.89432454e-01 1.99827075e-01 -6.40329480e-01 -1.95914552e-01
-1.35562038e+00 2.16019556e-01 5.49460053e-01 1.06813288e+00
1.14235234e+00 -1.42583579e-01 -6.90469027e-01 3.83827120e-01
1.97946846e-01 3.71821851e-01 2.25819558e-01 -1.33717799e+00
4.69266921e-01 4.82289076e-01 1.73644111e-01 -9.83545661e-01
-9.41054821e-02 -4.92247567e-02 -3.98706496e-01 -4.31381077e-01
2.59933233e-01 8.85257050e-02 -1.08031273e+00 1.62230062e+00
5.04327297e-01 9.41927433e-01 -2.36395136e-01 1.13357174e+00
9.46853817e-01 1.12689424e+00 4.16117400e-01 -2.94690818e-01
1.44515359e+00 -1.44660842e+00 -5.97258508e-01 -1.55694723e-01
4.29166973e-01 -7.39041150e-01 1.26518953e+00 1.57385617e-01
-1.15794456e+00 -9.62849081e-01 -5.00770271e-01 -3.93072397e-01
-1.28805548e-01 1.52772933e-01 2.75301367e-01 2.34334804e-02
-1.31643152e+00 -6.15745783e-03 -7.64318585e-01 -3.94106507e-01
1.62224293e-01 -4.89061587e-02 -5.15936136e-01 -5.41052222e-01
-1.19933689e+00 4.77562845e-01 1.86740279e-01 -1.60428926e-01
-1.58056164e+00 -9.82327819e-01 -1.08250511e+00 -5.71876243e-02
7.15999961e-01 -8.59734535e-01 1.42770135e+00 -1.05265641e+00
-1.24065757e+00 7.27422833e-01 -8.45656157e-01 -6.47788942e-01
2.65131593e-01 -3.82023335e-01 -3.82476151e-01 9.50447500e-01
4.32088077e-01 1.12325954e+00 1.32230496e+00 -9.43894506e-01
-8.37102354e-01 5.86818829e-02 5.54397285e-01 1.49184763e-01
3.44079100e-02 1.53422907e-01 -1.38709092e+00 -6.03312135e-01
-3.02089274e-01 -5.45510054e-01 -8.91960487e-02 5.95335700e-02
4.44084443e-02 -3.37393999e-01 1.09823298e+00 -7.84429193e-01
1.39136982e+00 -2.08975029e+00 1.52867407e-01 -2.88402915e-01
1.06183432e-01 4.70458940e-02 -4.75734204e-01 3.11400414e-01
-2.00777892e-02 -4.18311954e-02 2.56696999e-01 -6.06013238e-01
-3.46036822e-01 2.73783684e-01 -5.07734358e-01 5.57868302e-01
2.95400143e-01 1.11086655e+00 -1.10013044e+00 -8.39072704e-01
4.60116446e-01 6.23276889e-01 -8.36977124e-01 5.66008449e-01
-7.43509054e-01 4.86620098e-01 -3.17810446e-01 9.27691579e-01
2.15549782e-01 -7.04938591e-01 -2.11336076e-01 -6.95506871e-01
-7.85912350e-02 1.49928629e-01 -1.02799559e+00 2.01369333e+00
-5.85244536e-01 7.26495147e-01 4.06455845e-02 -7.47615695e-01
2.68430501e-01 5.51901519e-01 6.15419507e-01 -6.65610731e-01
-2.36188889e-01 -1.93885013e-01 -8.35234344e-01 -8.84002388e-01
5.12883842e-01 3.80560458e-01 -1.45160733e-02 3.19254577e-01
5.09200096e-01 1.15761608e-01 4.10726726e-01 5.75826764e-01
1.29250491e+00 5.96524239e-01 1.64921403e-01 3.66550416e-01
8.36152017e-01 2.91807130e-02 3.83790493e-01 7.61304796e-01
-3.50306809e-01 8.49771738e-01 4.11379784e-01 -4.59488690e-01
-8.33169580e-01 -9.31719899e-01 5.60209751e-01 1.71451581e+00
3.67932886e-01 -8.76427412e-01 -7.07041025e-01 -9.56689239e-01
-3.46204579e-01 5.77005148e-01 -4.99887764e-01 6.39067367e-02
-6.74913287e-01 1.63384646e-01 4.53981996e-01 4.17201042e-01
4.52815264e-01 -9.60735023e-01 -7.27796435e-01 1.87467381e-01
-8.47169757e-01 -1.56470466e+00 -1.00296307e+00 -2.82973409e-01
-4.75733876e-01 -1.22699010e+00 -7.63288021e-01 -6.48993731e-01
5.25497496e-01 6.17556870e-01 1.41736567e+00 2.47007981e-01
-7.20153451e-02 1.20045769e+00 -5.53205609e-01 2.87358999e-01
-5.74474931e-02 -1.07957646e-01 -2.43829414e-01 2.04254195e-01
3.37213486e-01 -2.22007155e-01 -9.08031344e-01 3.71169895e-01
-9.25799489e-01 9.64402333e-02 4.15618271e-01 5.66895127e-01
1.07645333e+00 -2.48899087e-01 3.18934798e-01 -4.30532783e-01
1.41710164e-02 -6.95237398e-01 -5.40591419e-01 5.38787127e-01
-7.07412511e-02 -1.42661268e-02 4.06914353e-01 -6.89516544e-01
-7.78734505e-01 2.02794209e-01 -1.95045024e-01 -1.19776535e+00
-2.16082796e-01 3.16365451e-01 9.68158171e-02 -1.03868125e-03
3.49875510e-01 4.23880756e-01 -4.43891317e-01 -2.20325157e-01
5.48493505e-01 2.11430773e-01 7.24864364e-01 -5.67332268e-01
6.75363541e-01 5.83337307e-01 -4.31923449e-01 -6.52430236e-01
-1.13273251e+00 -1.00347304e+00 -4.09866005e-01 -5.28332710e-01
1.29645371e+00 -1.43921649e+00 -6.91582263e-01 -5.87288328e-02
-1.41459966e+00 -4.75857288e-01 -2.05628797e-01 3.15927833e-01
-6.30866706e-01 4.18299496e-01 -6.06365740e-01 -5.49290717e-01
-1.97212055e-01 -1.39582646e+00 1.55303574e+00 7.17361644e-02
-1.68626562e-01 -8.23805094e-01 -7.43822977e-02 6.46652400e-01
1.36951745e-01 -2.48791706e-02 5.01637459e-01 -3.81339520e-01
-1.14661694e+00 8.21623728e-02 -2.79439002e-01 5.28537529e-03
-2.00137109e-01 -2.27195933e-01 -8.45610261e-01 -2.45133698e-01
-8.07301998e-02 -4.41125184e-01 9.48637664e-01 3.51034343e-01
1.27245915e+00 -3.25516760e-01 -2.33251378e-01 6.67546749e-01
1.11293197e+00 -1.37336254e-02 4.29854870e-01 1.40765980e-01
6.99037910e-01 2.26294011e-01 9.54736590e-01 2.80196548e-01
9.36059296e-01 5.90679467e-01 3.97189498e-01 -4.42119204e-02
-4.50669914e-01 -5.87977409e-01 7.69723654e-01 5.32357454e-01
1.00708399e-02 -3.27908695e-01 -6.71801984e-01 8.17949295e-01
-1.95407856e+00 -1.23072505e+00 2.61714846e-01 1.87820029e+00
7.39283562e-01 -2.69640952e-01 9.78296250e-02 -3.59352469e-01
5.42430520e-01 4.86334860e-01 -4.82561886e-01 2.58368373e-01
2.27109212e-02 1.19308688e-01 2.91589588e-01 7.18379557e-01
-1.17513311e+00 1.27328336e+00 5.51295328e+00 7.98365474e-01
-1.10107636e+00 5.64123273e-01 7.14116573e-01 -2.81757355e-01
-2.18493238e-01 2.31948957e-01 -9.67721760e-01 3.38418156e-01
9.90863442e-01 2.98811555e-01 4.95232582e-01 7.62662053e-01
4.55108613e-01 -4.67844427e-01 -1.24926996e+00 1.24585116e+00
5.25501668e-01 -1.71694350e+00 2.26266488e-01 -4.91981745e-01
6.81237757e-01 1.02533102e-01 -1.41987160e-01 5.20595014e-01
-1.47428676e-01 -8.34228218e-01 9.46701527e-01 6.15604579e-01
9.34899688e-01 -5.14320791e-01 3.98909897e-01 2.72450387e-01
-1.66520488e+00 -6.39293194e-02 -6.83552176e-02 1.13641746e-01
5.99089503e-01 2.26211935e-01 -6.98213696e-01 3.64072114e-01
1.08541393e+00 8.33045542e-01 -6.36622071e-01 6.71383083e-01
-3.90022188e-01 7.21421242e-01 -1.11141086e-01 3.64423931e-01
3.77403021e-01 2.06213504e-01 4.63717073e-01 1.30150151e+00
3.31707627e-01 2.02684730e-01 5.38911879e-01 7.19395339e-01
-9.97820869e-02 -6.96172267e-02 -2.13414967e-01 -5.43951169e-02
4.84180003e-01 1.16802251e+00 -6.17947996e-01 -5.60134947e-01
-7.78277576e-01 1.33037245e+00 2.24667907e-01 7.86604881e-01
-1.19182110e+00 1.38826668e-01 6.46776080e-01 4.34864253e-01
6.01833045e-01 -3.56957972e-01 6.74690962e-01 -1.45684505e+00
-2.11782679e-01 -9.81854796e-01 6.85642898e-01 -1.23060668e+00
-8.27165902e-01 4.18429345e-01 1.57584503e-01 -1.05137730e+00
-5.08606493e-01 -3.28446507e-01 -2.31638014e-01 5.08257031e-01
-1.69407225e+00 -1.45918536e+00 -4.26478595e-01 1.18371224e+00
1.29541385e+00 8.22063982e-02 4.11823720e-01 5.47456741e-01
-3.04175168e-01 3.33455473e-01 -7.03664422e-01 1.54631346e-01
1.05355620e+00 -8.11323762e-01 1.03875652e-01 1.08652282e+00
5.73688507e-01 4.95117962e-01 4.80906308e-01 -6.18121088e-01
-1.64911020e+00 -1.38294041e+00 9.73274648e-01 -6.28651977e-01
7.47024536e-01 -3.82241279e-01 -7.53301620e-01 8.52353692e-01
2.76518494e-01 3.72150391e-01 3.25644135e-01 -3.57768655e-01
-4.41115856e-01 -2.78442577e-02 -6.92257345e-01 5.80926895e-01
7.81416893e-01 -1.10285699e+00 -4.06910926e-01 4.59225208e-01
1.18694353e+00 -5.58743596e-01 -7.15500176e-01 2.22490638e-01
3.25745761e-01 -9.79936421e-01 1.31614983e+00 -2.97751933e-01
3.34843814e-01 -7.56626248e-01 -2.50506490e-01 -5.76881230e-01
-1.16515867e-01 -7.68365741e-01 -7.35705793e-01 1.17580378e+00
7.24123791e-02 2.53888905e-01 6.30986571e-01 2.87931114e-01
1.58314690e-01 -6.94752574e-01 -7.60696411e-01 -2.51127839e-01
-6.55660570e-01 -9.16628897e-01 2.74986982e-01 5.85394502e-01
-4.69014108e-01 4.13292170e-01 -6.08715415e-01 4.43534523e-01
2.69153476e-01 9.33439136e-02 8.99651289e-01 -3.95438284e-01
-4.26605850e-01 1.19283788e-01 -2.00730845e-01 -1.65702021e+00
3.84061098e-01 -4.49156374e-01 1.36627361e-01 -1.58805931e+00
1.58810049e-01 1.07898928e-01 -2.09320113e-01 4.94338989e-01
-2.62593567e-01 3.75160456e-01 2.76182741e-01 2.58983701e-01
-1.50197232e+00 1.86126783e-01 1.13084126e+00 -2.19333321e-01
-1.86011404e-01 -2.03390568e-01 -1.48833439e-01 7.51588464e-01
2.68238455e-01 -3.32191467e-01 -6.38251841e-01 -5.88665485e-01
3.37827951e-01 4.76027548e-01 8.41933310e-01 -1.03118789e+00
6.45767868e-01 -1.85178131e-01 4.96756136e-01 -9.10563767e-01
5.62047958e-01 -7.29183018e-01 3.69017534e-02 2.83754431e-02
-3.48133832e-01 3.85995924e-01 1.25894621e-01 8.59959781e-01
-4.95846570e-01 -2.90928967e-02 4.41780478e-01 -3.80638629e-01
-1.40577161e+00 7.29998171e-01 -4.55711216e-01 2.25809351e-01
1.06184101e+00 -1.54581562e-01 -7.22749457e-02 -7.59567916e-01
-6.27876818e-01 4.34558719e-01 4.10528660e-01 6.27332091e-01
8.75483394e-01 -1.01491964e+00 -3.86390597e-01 -3.40789370e-02
2.35159487e-01 4.76095267e-02 6.56765223e-01 1.06242919e+00
-4.24598545e-01 3.33157241e-01 4.46613461e-01 -9.89435017e-01
-1.14008653e+00 8.93478572e-01 2.74208009e-01 -2.63265789e-01
-5.21728218e-01 1.04930270e+00 5.87887347e-01 1.66297369e-02
4.73580509e-01 -5.07986605e-01 -1.10289015e-01 1.36682928e-01
8.27115774e-01 9.16472748e-02 -2.49167964e-01 -7.91510642e-01
-3.74633342e-01 6.46226704e-01 8.62291828e-03 -1.09519131e-01
9.13500130e-01 -5.79045117e-01 -2.08651256e-02 4.21587586e-01
1.27518296e+00 -2.76085492e-02 -1.63983655e+00 -2.73781836e-01
-2.84300238e-01 -2.91452825e-01 -1.84903741e-02 -5.40300488e-01
-9.24231768e-01 8.51543605e-01 4.12803888e-01 -4.03474033e-01
1.33471429e+00 3.70688736e-01 9.64179456e-01 3.15721661e-01
4.32930619e-01 -8.61575127e-01 6.61917746e-01 5.98722577e-01
6.64102912e-01 -1.28943968e+00 -2.41526529e-01 -1.07780710e-01
-6.58243477e-01 9.86013591e-01 7.89596558e-01 -5.07747382e-02
5.16888320e-01 1.18491106e-01 3.52917202e-02 -7.23812059e-02
-1.12572122e+00 -2.62369752e-01 5.00240445e-01 4.36460584e-01
2.80429155e-01 -4.64188993e-01 3.04529965e-01 4.70687777e-01
3.60086113e-01 1.97856113e-01 9.72581059e-02 6.87797606e-01
-2.96878606e-01 -7.36616850e-01 -3.82171094e-01 -1.01922750e-01
-5.13804436e-01 -2.37800911e-01 4.16782824e-03 6.87680244e-01
2.96697795e-01 1.05636954e+00 1.90616339e-01 -3.30679983e-01
7.35711539e-03 5.50377034e-02 3.47881347e-01 -5.59357226e-01
-5.38637578e-01 2.56418169e-01 -1.65654913e-01 -1.24686360e+00
-7.87665904e-01 -4.08047199e-01 -1.24091291e+00 1.48910377e-03
-1.14041694e-01 1.77528888e-01 1.47632167e-01 1.08797419e+00
5.63889265e-01 5.62205970e-01 1.40141293e-01 -9.15441632e-01
9.47817788e-02 -6.35226011e-01 1.23764984e-01 3.25182647e-01
7.13001549e-01 -4.81623292e-01 -2.22345918e-01 6.71674907e-01] | [10.152024269104004, 0.8215510249137878] |
85e3e0bb-72fe-4e75-9486-c1d033a213c3 | unsupervised-extractive-summarization-by | 2104.08392 | null | https://arxiv.org/abs/2104.08392v1 | https://arxiv.org/pdf/2104.08392v1.pdf | Unsupervised Extractive Summarization by Human Memory Simulation | Summarization systems face the core challenge of identifying and selecting important information. In this paper, we tackle the problem of content selection in unsupervised extractive summarization of long, structured documents. We introduce a wide range of heuristics that leverage cognitive representations of content units and how these are retained or forgotten in human memory. We find that properties of these representations of human memory can be exploited to capture relevance of content units in scientific articles. Experiments show that our proposed heuristics are effective at leveraging cognitive structures and the organization of the document (i.e.\ sections of an article), and automatic and human evaluations provide strong evidence that these heuristics extract more summary-worthy content units. | ['Shay B. Cohen', 'Matthias Galle', 'Ronald Cardenas'] | 2021-04-16 | null | null | null | null | ['unsupervised-extractive-summarization'] | ['natural-language-processing'] | [ 5.54282367e-01 6.20905280e-01 -5.40911973e-01 -1.27940504e-02
-8.60331893e-01 -7.07074404e-01 5.14185786e-01 1.12204480e+00
-3.51180762e-01 1.10848534e+00 1.19137096e+00 -2.14807652e-02
-3.46419930e-01 -5.88113189e-01 -5.80737352e-01 -2.54835397e-01
-1.10405728e-01 3.89580578e-01 8.04369152e-02 -5.64083271e-02
1.40724635e+00 3.96694273e-01 -1.52320790e+00 6.38722837e-01
1.16219151e+00 4.32776868e-01 4.07233000e-01 8.08064342e-01
-5.10034561e-01 1.07078481e+00 -1.01520145e+00 -1.46017030e-01
-2.31256634e-01 -6.59201980e-01 -1.34343207e+00 2.34074906e-01
6.04593396e-01 -5.99209778e-02 -2.67648757e-01 1.13864827e+00
4.53170210e-01 3.55542570e-01 6.88305199e-01 -3.95161837e-01
-5.33725977e-01 1.21306872e+00 -3.63038659e-01 1.00471294e+00
6.30636394e-01 -3.98230433e-01 1.19832528e+00 -7.84979045e-01
1.16954434e+00 1.12636065e+00 2.49674737e-01 3.71776938e-01
-7.92687833e-01 -3.22865620e-02 2.62158722e-01 7.04800859e-02
-9.27542925e-01 -9.34320807e-01 7.08561182e-01 -2.83380687e-01
1.15775609e+00 5.35397410e-01 5.15980601e-01 9.46045160e-01
7.27818131e-01 1.19063449e+00 5.39845943e-01 -5.48776329e-01
4.62862641e-01 -4.64370325e-02 1.22662723e+00 6.07246995e-01
1.07355261e+00 -6.45100951e-01 -9.86579537e-01 -4.82276082e-01
1.23516195e-01 -1.85035527e-01 -2.25449979e-01 1.59626260e-01
-1.12293708e+00 8.55106890e-01 -6.52553216e-02 4.43897665e-01
-7.78809607e-01 -2.16352984e-01 6.83147848e-01 1.63937047e-01
4.50030059e-01 1.40214574e+00 -3.48275304e-02 -1.80299416e-01
-1.26103806e+00 2.59874582e-01 8.13777149e-01 1.19312060e+00
4.93584394e-01 -1.58226803e-01 -7.28360772e-01 5.46983480e-01
-2.54086882e-01 2.63822526e-01 8.87869000e-01 -1.03461087e+00
5.61096907e-01 7.18808413e-01 -4.73360345e-02 -1.17048061e+00
-4.93796140e-01 -6.57469213e-01 -6.22618198e-01 -7.40279973e-01
-5.33600390e-01 -8.30308944e-02 -8.09807241e-01 1.33116984e+00
-2.75225043e-01 -5.27161300e-01 2.81472176e-01 2.48344198e-01
1.03455567e+00 7.65885949e-01 1.40452787e-01 -8.12880695e-01
1.36164081e+00 -8.88841033e-01 -9.51586366e-01 -5.42672932e-01
2.60350555e-01 -7.60330617e-01 3.10205251e-01 2.84565091e-01
-1.98164296e+00 -4.29344773e-01 -1.08603287e+00 -2.52679467e-01
-1.57971144e-01 1.29646897e-01 4.04765189e-01 4.14712615e-02
-1.04711640e+00 8.12713206e-01 -4.63023335e-01 -4.05383021e-01
4.87414896e-01 1.82878766e-02 1.75020024e-02 -1.38081014e-01
-8.63181710e-01 7.93793559e-01 8.89096081e-01 -1.79715201e-01
-5.87351382e-01 -4.50508952e-01 -6.85723841e-01 3.78548533e-01
5.84850967e-01 -8.81531894e-01 1.14562845e+00 -6.48212552e-01
-7.70394862e-01 8.15939724e-01 -6.06811106e-01 -6.66243434e-01
-1.73092559e-01 -5.37642419e-01 -2.18529791e-01 8.72504592e-01
2.59590507e-01 2.54682183e-01 8.10537338e-01 -1.12012815e+00
-7.94376612e-01 -4.27758187e-01 -1.27071723e-01 3.27963084e-01
-5.45263290e-01 5.64227477e-02 -3.19000125e-01 -9.50112104e-01
2.47455269e-01 -6.55102670e-01 -2.57799476e-01 -1.00936830e+00
-8.33039224e-01 -3.61268818e-01 2.37772599e-01 -7.63642669e-01
1.81747591e+00 -1.64281309e+00 3.69050413e-01 4.53809261e-01
5.67425728e-01 1.10377096e-01 -1.70834720e-01 7.68929482e-01
3.50258648e-01 6.49005294e-01 7.35146983e-04 2.28286445e-01
-1.74388453e-01 -1.07419573e-01 -9.04445112e-01 1.33916438e-02
-2.05155760e-01 9.48003292e-01 -1.19986999e+00 -8.10307324e-01
-4.70090806e-01 -2.90573657e-01 -2.86882162e-01 1.73179373e-01
-2.60242194e-01 -7.95489922e-02 -8.33727241e-01 6.14774525e-01
1.44894853e-01 -1.77159265e-01 1.84293106e-01 3.22204418e-02
-7.15493560e-02 8.20810318e-01 -5.12137234e-01 1.44996083e+00
1.69462353e-01 8.70477259e-01 -2.21944660e-01 -9.18346584e-01
7.22636700e-01 5.66330478e-02 1.77624956e-01 -6.27855241e-01
3.60933356e-02 -2.86388770e-02 -9.67354402e-02 -6.00752056e-01
1.33300948e+00 2.95806617e-01 -3.17735344e-01 9.58604991e-01
1.59693122e-01 1.54080421e-01 9.14832234e-01 8.42559695e-01
1.25712907e+00 -6.34013534e-01 8.13412905e-01 -7.65704155e-01
3.68245691e-01 3.49456847e-01 2.74208009e-01 1.54101264e+00
1.20980546e-01 3.76126140e-01 5.70093036e-01 -1.79439321e-01
-1.03920662e+00 -7.52199709e-01 1.79192141e-01 1.39181805e+00
2.02405639e-02 -1.05538034e+00 -7.69157708e-01 -5.58763087e-01
-7.63972998e-02 1.07148051e+00 -7.51075387e-01 -6.02402627e-01
-8.29578340e-01 -4.77751046e-01 5.31119049e-01 5.62883139e-01
3.69703509e-02 -1.31845379e+00 -9.30135608e-01 3.34210157e-01
-4.03998286e-01 -7.71060467e-01 -4.58025664e-01 2.31760859e-01
-1.29306066e+00 -9.85785186e-01 -4.37590450e-01 -7.85449803e-01
9.59874988e-01 6.80252552e-01 1.45207787e+00 1.99390203e-01
-3.09582744e-02 6.32421434e-01 -5.21435440e-01 -6.22555852e-01
-5.96558452e-01 7.01080978e-01 -1.86566859e-01 -6.94682181e-01
1.85089201e-01 -3.09442878e-01 -3.79786491e-01 -2.46446013e-01
-1.00082612e+00 -1.61691681e-02 9.17568386e-01 6.45474851e-01
5.85002303e-01 1.56282321e-01 7.59282529e-01 -1.19300687e+00
1.49947727e+00 -5.81969023e-01 1.08541034e-01 6.75415277e-01
-8.00113916e-01 5.89178503e-01 4.66542631e-01 -9.26907137e-02
-1.06567192e+00 -6.04560554e-01 2.86274433e-01 9.87504199e-02
1.77092880e-01 1.10806572e+00 1.04099393e-01 4.74849015e-01
7.02343166e-01 6.79705083e-01 -4.02959377e-01 -3.98894876e-01
2.44124979e-01 4.05771106e-01 6.54423773e-01 -7.71136701e-01
2.23623186e-01 2.57100314e-01 -1.56106248e-01 -1.05604982e+00
-1.24337459e+00 -6.72462821e-01 -6.09485865e-01 -1.57097995e-01
5.02703309e-01 -6.53833210e-01 -1.73494458e-01 -3.92712533e-01
-1.11458659e+00 4.97676700e-01 -6.27412021e-01 7.16638342e-02
-3.80112737e-01 7.39216447e-01 -7.04578519e-01 -5.32720804e-01
-9.85930860e-01 -5.10668159e-01 9.54810798e-01 5.44115901e-01
-8.63260269e-01 -7.38442361e-01 1.31178245e-01 2.44189918e-01
6.80353194e-02 6.26504570e-02 1.28149235e+00 -1.15866232e+00
-3.61356020e-01 -3.86029482e-01 2.14438960e-01 -1.66908532e-01
1.34398993e-02 -7.76676163e-02 -4.75297183e-01 -3.97193789e-01
1.24888800e-01 -2.24882618e-01 1.61096478e+00 5.15167892e-01
1.09860718e+00 -9.79141474e-01 -5.08820295e-01 -2.01591216e-02
8.08526099e-01 1.05990432e-01 6.07216179e-01 3.91017526e-01
4.93088424e-01 8.11109602e-01 4.75723475e-01 5.72640479e-01
-3.52784917e-02 -1.09891787e-01 -2.01339990e-01 5.41817725e-01
1.34746194e-01 -2.40873322e-01 3.46573442e-01 1.27352202e+00
-1.73088297e-01 -3.82360011e-01 -7.83924103e-01 7.44634926e-01
-1.87938488e+00 -1.55518234e+00 1.70452774e-01 1.90820527e+00
1.07719421e+00 4.81432348e-01 -4.64031193e-03 -8.56930539e-02
7.58769989e-01 3.25246543e-01 -4.02504474e-01 -6.45722449e-01
-2.89435476e-01 2.00541049e-01 3.52591753e-01 2.63611883e-01
-8.97431314e-01 9.24490869e-01 7.46797991e+00 6.50415719e-01
-4.38942999e-01 -4.43576604e-01 4.32074904e-01 -2.88718104e-01
-4.65138555e-01 2.55645327e-02 -1.00508797e+00 1.56457424e-01
1.18893278e+00 -1.04787064e+00 -1.41085938e-01 7.78200567e-01
1.86435372e-01 -3.80118489e-01 -1.06471884e+00 4.89573628e-01
6.23617291e-01 -1.77279437e+00 8.60322714e-01 4.30020429e-02
1.02592170e+00 -4.82332587e-01 -1.53581887e-01 4.31348085e-02
3.96384597e-01 -8.59825730e-01 7.67014682e-01 9.96281147e-01
-8.24586451e-02 -9.65286613e-01 7.38052726e-01 3.06437880e-01
-6.14215136e-01 -1.56909570e-01 -9.93734121e-01 1.61986932e-01
-1.15257660e-02 7.87695765e-01 -8.94689560e-01 5.46868205e-01
2.17227444e-01 7.33977914e-01 -1.14132881e+00 1.00102508e+00
-2.46716723e-01 6.38099730e-01 3.44265163e-01 -3.03010851e-01
3.85720730e-01 1.49776101e-01 1.03600037e+00 1.61250103e+00
2.57231742e-01 3.06445658e-01 1.80586562e-01 5.62931955e-01
-3.66145223e-01 1.97532922e-01 -5.30563772e-01 -5.85352719e-01
6.22302473e-01 9.74043608e-01 -1.03277218e+00 -9.51401830e-01
3.94048780e-01 6.37651920e-01 3.55514795e-01 3.54955867e-02
-8.71111229e-02 -5.98768950e-01 1.54983357e-01 -8.61473754e-02
4.76454318e-01 -1.67614594e-01 -5.19062221e-01 -1.15430117e+00
-1.05897635e-02 -9.76394117e-01 7.14361370e-01 -7.03465343e-01
-9.97623444e-01 6.23357177e-01 2.67135836e-02 -8.41080248e-01
-4.92266864e-01 -2.10510697e-02 -8.06241393e-01 3.05135399e-01
-9.49812531e-01 -4.68013406e-01 4.60156016e-02 1.49595350e-01
1.03623509e+00 -3.31343919e-01 6.65301323e-01 -6.45070374e-01
-6.38481736e-01 2.65083820e-01 4.43041295e-01 -4.40442860e-02
3.69873732e-01 -1.30696726e+00 2.66344905e-01 1.11310184e+00
2.68748403e-01 1.38952351e+00 9.20172036e-01 -1.13515627e+00
-1.46725178e+00 -9.76673722e-01 1.11562836e+00 -4.43483770e-01
3.46072882e-01 -3.10918093e-02 -9.64222193e-01 5.51753283e-01
8.21532845e-01 -1.11719513e+00 9.85767365e-01 2.34594107e-01
-1.68266758e-01 3.36213142e-01 -6.91268504e-01 7.07528472e-01
1.09796298e+00 -3.95243198e-01 -1.63094211e+00 2.34931841e-01
6.16532683e-01 -1.64227381e-01 -4.91590828e-01 2.33959630e-02
3.52576911e-01 -5.77387393e-01 9.53016102e-01 -1.10572207e+00
8.90787184e-01 1.65665463e-01 1.51578978e-01 -1.38252532e+00
-7.06410706e-01 -7.50951707e-01 -7.61895776e-01 1.17911816e+00
4.44922000e-01 -9.67563838e-02 4.57827300e-01 4.94499385e-01
-3.93912584e-01 -4.21117932e-01 -4.61847216e-01 -4.63358611e-01
-7.32686021e-04 1.71878755e-01 1.91010818e-01 7.17634976e-01
7.06128955e-01 9.70172048e-01 -1.71294406e-01 -4.09600377e-01
6.31150067e-01 5.62271237e-01 4.40735370e-01 -1.36457849e+00
2.78829038e-01 -8.13068509e-01 -7.46939480e-02 -8.93972576e-01
2.50853390e-01 -9.88756478e-01 1.33247077e-01 -1.97608364e+00
7.25764811e-01 4.53485042e-01 -5.04089117e-01 1.32094204e-01
-5.09654641e-01 -5.12397408e-01 1.59924433e-01 7.39550412e-01
-1.30179405e+00 3.59541774e-01 9.61302936e-01 -4.55674320e-01
-5.07678688e-01 -1.89186871e-01 -1.49181509e+00 7.96991110e-01
9.64150608e-01 -5.19482195e-01 -4.13374782e-01 -2.55132049e-01
3.91718030e-01 -2.79238839e-02 -2.14516088e-01 -8.73263240e-01
7.18113899e-01 -2.75736541e-01 6.60900414e-01 -1.05280387e+00
-3.32037002e-01 4.93956283e-02 -4.11492199e-01 3.53434473e-01
-1.22187328e+00 4.08561260e-01 2.96559572e-01 7.05818295e-01
-1.38804153e-01 -6.91372633e-01 2.72605389e-01 -5.51530600e-01
-6.75441027e-01 -2.25335509e-01 -8.64135981e-01 5.73120952e-01
4.20905679e-01 -4.88827378e-02 -6.78703368e-01 -2.52617031e-01
-4.61159348e-01 1.67941153e-01 2.52482742e-01 3.36346269e-01
9.57071424e-01 -9.53664899e-01 -9.81302142e-01 -3.60085815e-01
1.90758690e-01 -2.38835096e-01 -1.38955917e-02 4.70938772e-01
-3.10472339e-01 8.80810738e-01 -4.11213100e-01 2.22888160e-02
-1.25816464e+00 6.93245530e-01 -5.30878663e-01 -5.89434087e-01
-8.39635015e-01 5.17963588e-01 5.94256772e-03 5.77749133e-01
1.48797840e-01 -2.91467249e-01 -6.74671352e-01 6.14961386e-01
1.00073135e+00 5.70433676e-01 1.45117238e-01 -4.98524606e-01
-1.90593630e-01 7.89241120e-02 -8.61747444e-01 5.79592474e-02
1.55625153e+00 -3.00758243e-01 -5.86307704e-01 3.96357238e-01
9.37985003e-01 4.62071806e-01 -4.23984408e-01 -3.22993159e-01
7.95830667e-01 -8.52623433e-02 8.10559914e-02 -6.96614206e-01
-5.30345500e-01 3.96672577e-01 -3.13278317e-01 3.65207314e-01
9.60815966e-01 2.77723402e-01 7.63349414e-01 1.07911253e+00
1.99030470e-02 -1.57434440e+00 2.12895155e-01 7.42618024e-01
1.17012119e+00 -5.75136781e-01 7.71588802e-01 -1.49560004e-01
-5.50880909e-01 1.32482576e+00 3.55775952e-01 -1.34946898e-01
1.51039734e-01 -1.33197322e-01 -4.88735884e-01 -6.48951769e-01
-1.02681112e+00 -1.91412881e-01 6.52912199e-01 1.59962416e-01
4.99516398e-01 -2.44048879e-01 -7.46155024e-01 8.08350325e-01
-4.59243536e-01 -4.19075131e-01 1.07402229e+00 1.56496000e+00
-1.19478440e+00 -4.98193532e-01 -4.85558987e-01 9.32814121e-01
-6.45401359e-01 -2.57327825e-01 -1.09930730e+00 2.22745746e-01
-5.32793701e-01 9.70823646e-01 -2.12489933e-01 -8.38940814e-02
2.59880126e-01 3.16804022e-01 4.02437717e-01 -9.02321994e-01
-7.00464070e-01 2.19685420e-01 2.59931386e-01 -2.51843601e-01
-3.90681595e-01 -7.92564154e-01 -1.20135486e+00 -3.22216481e-01
-1.24701113e-01 6.93036199e-01 9.40199494e-02 8.50851417e-01
7.39264369e-01 6.82162881e-01 3.92250597e-01 -5.83911359e-01
-5.01699984e-01 -9.76926506e-01 -3.45799714e-01 4.78088886e-01
3.92676145e-01 -3.15513343e-01 -9.81430486e-02 2.70392716e-01] | [12.516242027282715, 9.496461868286133] |
4c4c83b7-46e1-4f2d-8145-319c2a9fb4e5 | deblurgan-v2-deblurring-orders-of-magnitude | 1908.03826 | null | https://arxiv.org/abs/1908.03826v1 | https://arxiv.org/pdf/1908.03826v1.pdf | DeblurGAN-v2: Deblurring (Orders-of-Magnitude) Faster and Better | We present a new end-to-end generative adversarial network (GAN) for single image motion deblurring, named DeblurGAN-v2, which considerably boosts state-of-the-art deblurring efficiency, quality, and flexibility. DeblurGAN-v2 is based on a relativistic conditional GAN with a double-scale discriminator. For the first time, we introduce the Feature Pyramid Network into deblurring, as a core building block in the generator of DeblurGAN-v2. It can flexibly work with a wide range of backbones, to navigate the balance between performance and efficiency. The plug-in of sophisticated backbones (e.g., Inception-ResNet-v2) can lead to solid state-of-the-art deblurring. Meanwhile, with light-weight backbones (e.g., MobileNet and its variants), DeblurGAN-v2 reaches 10-100 times faster than the nearest competitors, while maintaining close to state-of-the-art results, implying the option of real-time video deblurring. We demonstrate that DeblurGAN-v2 obtains very competitive performance on several popular benchmarks, in terms of deblurring quality (both objective and subjective), as well as efficiency. Besides, we show the architecture to be effective for general image restoration tasks too. Our codes, models and data are available at: https://github.com/KupynOrest/DeblurGANv2 | ['Junru Wu', 'Tetiana Martyniuk', 'Zhangyang Wang', 'Orest Kupyn'] | 2019-08-10 | deblurgan-v2-deblurring-orders-of-magnitude-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Kupyn_DeblurGAN-v2_Deblurring_Orders-of-Magnitude_Faster_and_Better_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Kupyn_DeblurGAN-v2_Deblurring_Orders-of-Magnitude_Faster_and_Better_ICCV_2019_paper.pdf | iccv-2019-10 | ['single-image-blind-deblurring', 'blind-face-restoration'] | ['computer-vision', 'computer-vision'] | [-8.37612674e-02 -2.85911292e-01 -1.71974152e-01 2.35580772e-01
-8.50447237e-01 -6.11884534e-01 5.54983795e-01 -1.07788181e+00
-3.17064002e-02 6.57021284e-01 5.66176355e-01 -2.81255305e-01
1.90463290e-01 -6.77335441e-01 -7.64654994e-01 -9.05801833e-01
1.59452781e-01 2.26679686e-02 1.63685337e-01 -3.58116746e-01
-1.53019562e-01 2.77853459e-01 -9.87098277e-01 -1.62772253e-01
9.52026069e-01 7.95081258e-01 2.57187933e-01 7.77186632e-01
5.14250100e-01 9.04622555e-01 -4.49853182e-01 -5.12722552e-01
2.95085758e-01 -6.56932473e-01 -7.00363100e-01 -2.11886302e-01
5.10791540e-01 -8.53747606e-01 -1.09473801e+00 9.74797904e-01
7.46310294e-01 3.23162824e-02 5.52185178e-01 -1.20378315e+00
-1.33001006e+00 3.62348169e-01 -6.84122443e-01 2.84688652e-01
4.68488876e-03 7.02076554e-01 7.69381285e-01 -7.61888981e-01
6.10029519e-01 1.17993271e+00 7.83762336e-01 8.97246778e-01
-1.07713330e+00 -8.43898773e-01 -2.39355247e-02 2.71996319e-01
-1.29529619e+00 -3.96817356e-01 6.75375223e-01 -2.93357491e-01
5.80472469e-01 2.26599038e-01 3.92957598e-01 1.53473520e+00
3.35603565e-01 9.44175422e-01 9.24118161e-01 1.19154669e-01
4.30216938e-02 -6.51300430e-01 -1.40109509e-01 3.98206741e-01
6.04874380e-02 4.62077379e-01 -2.60155678e-01 1.40748069e-01
1.14877951e+00 -4.75517921e-02 -8.13333392e-01 -2.16248422e-03
-1.20410180e+00 6.32848501e-01 6.76951945e-01 1.47118524e-01
-2.73629308e-01 7.83724189e-01 2.62265980e-01 3.65213811e-01
5.90012491e-01 3.03509414e-01 -2.21067071e-01 -2.15461940e-01
-1.10978329e+00 2.02990800e-01 2.21338630e-01 8.58135283e-01
4.86157864e-01 2.64874637e-01 -4.87146646e-01 7.63653100e-01
8.03303719e-02 5.75845778e-01 5.95939636e-01 -1.15985572e+00
3.33907306e-01 -9.59344730e-02 7.11811185e-02 -7.88381577e-01
-1.82429533e-02 -5.35753608e-01 -1.42231143e+00 1.81732789e-01
4.03451473e-02 -3.14867109e-01 -1.10040522e+00 1.93614030e+00
1.40769565e-02 6.03806674e-01 -3.23430859e-02 1.26654577e+00
7.92581141e-01 9.54142094e-01 -1.78067297e-01 9.15325209e-02
1.22724533e+00 -1.44774580e+00 -6.30258441e-01 -3.72943372e-01
2.55106241e-01 -5.11853695e-01 1.08832538e+00 3.76977473e-01
-1.14593840e+00 -6.31798029e-01 -9.45751131e-01 -2.25330159e-01
2.42015108e-01 -1.44495219e-02 6.17711902e-01 4.57490146e-01
-1.39228344e+00 6.61017656e-01 -9.45132017e-01 -5.50168715e-02
4.82882440e-01 3.71089578e-02 -3.01133752e-01 -4.66898233e-01
-1.38243234e+00 6.32030725e-01 3.71715240e-02 1.54856578e-01
-1.46761763e+00 -8.32396686e-01 -7.89086461e-01 6.68717250e-02
1.42075747e-01 -1.25618958e+00 8.80613446e-01 -9.03704047e-01
-1.63161159e+00 5.96444964e-01 1.57368138e-01 -4.12082344e-01
8.73413384e-01 -4.69370842e-01 -4.58354712e-01 5.98165430e-02
-9.29896683e-02 7.86177635e-01 1.10724914e+00 -1.19829345e+00
-1.30647257e-01 4.10960391e-02 -1.32078543e-01 7.54329201e-04
-1.94915235e-01 -1.61999073e-02 -7.13165462e-01 -1.34662163e+00
-2.05128178e-01 -9.22799110e-01 -1.82566002e-01 -1.01623066e-01
-4.28922236e-01 8.92776027e-02 9.45847452e-01 -1.01794398e+00
1.27440059e+00 -2.19242311e+00 5.46060562e-01 -3.98998380e-01
5.45283675e-01 4.34940249e-01 -5.57716012e-01 2.32601896e-01
-2.64226973e-01 2.64431268e-01 -3.42577398e-01 -3.83169293e-01
-8.78241137e-02 -2.09938139e-02 -3.39300513e-01 7.29308665e-01
-6.21800795e-02 1.45126152e+00 -7.93458164e-01 1.37679368e-01
2.22138986e-01 6.79062724e-01 -6.94993973e-01 3.64852518e-01
-6.06473386e-02 6.99162781e-01 -2.56237537e-01 5.89194834e-01
8.34716141e-01 -4.62213457e-01 -3.09258759e-01 -3.18309635e-01
2.66049832e-01 8.27554837e-02 -7.23929882e-01 1.99518657e+00
-5.69839239e-01 9.35007095e-01 1.86895967e-01 -5.11429131e-01
6.03737116e-01 2.51128823e-01 2.39471674e-01 -4.53566939e-01
7.59833008e-02 2.13910371e-01 -4.22537833e-01 -3.05174232e-01
7.92838097e-01 3.44713964e-02 -1.13192312e-01 2.88036972e-01
9.50071812e-02 -6.29568398e-02 -1.90522358e-01 3.85269254e-01
1.25854695e+00 2.52485842e-01 1.64085068e-02 -2.67259032e-01
2.76844680e-01 -4.58791792e-01 4.57228601e-01 6.99263752e-01
-9.57892612e-02 1.03262091e+00 2.19903842e-01 -3.40116382e-01
-1.08703339e+00 -1.10101378e+00 1.78036034e-01 6.76314294e-01
5.68405032e-01 -3.54762375e-01 -9.52037752e-01 -5.93910992e-01
-2.63973683e-01 6.38314426e-01 -5.71388662e-01 -2.98121691e-01
-5.91269851e-01 -8.79753828e-01 7.72940695e-01 4.46164012e-01
8.77604365e-01 -7.57336795e-01 -1.72822416e-01 1.39573440e-01
-4.35061663e-01 -9.95231211e-01 -1.10999370e+00 -4.45704192e-01
-6.67982459e-01 -7.17733145e-01 -1.37804425e+00 -6.36130333e-01
3.44956994e-01 4.85535264e-01 1.29322314e+00 1.05774932e-01
-5.16341440e-02 2.29975238e-01 -5.34187436e-01 5.41168988e-01
-5.97913384e-01 1.13016814e-01 -2.10990772e-01 -1.83013082e-01
-4.33310360e-01 -9.82087016e-01 -1.29208803e+00 6.77744865e-01
-1.15566790e+00 1.83070704e-01 4.51838970e-01 1.01812851e+00
3.00572604e-01 -7.66745135e-02 5.70211828e-01 -3.54008138e-01
4.49172795e-01 -5.31256258e-01 -4.20347989e-01 5.10509089e-02
-6.35705531e-01 -7.11563304e-02 6.51238620e-01 -5.92314184e-01
-7.70070374e-01 -6.02332413e-01 -3.76308203e-01 -7.75211930e-01
1.16253309e-01 2.11606011e-01 -3.47260803e-01 -1.96466565e-01
7.13105857e-01 5.14831245e-01 -6.95562661e-02 -5.74439943e-01
7.91169405e-01 6.15093946e-01 1.06820321e+00 -3.74009520e-01
9.34645534e-01 2.94996351e-01 -2.75481761e-01 -3.50563765e-01
-4.03395146e-01 -2.16561928e-02 8.39522555e-02 -9.27651674e-02
8.93932462e-01 -1.40380883e+00 -4.56590652e-01 1.34245098e+00
-1.07124925e+00 -7.63871431e-01 -1.46963671e-01 1.58844203e-01
-4.98521447e-01 5.69221795e-01 -1.19831407e+00 -1.05499014e-01
-7.50099421e-01 -1.33256972e+00 1.04903829e+00 1.59118578e-01
1.17078051e-01 -8.83657753e-01 8.35230276e-02 5.39464653e-01
9.19303298e-01 1.93941444e-01 5.28284311e-01 4.51103561e-02
-8.01856756e-01 1.31757716e-02 -3.50654691e-01 6.48004949e-01
1.46011159e-01 -3.89635414e-01 -8.45949709e-01 -8.12936962e-01
-1.65107384e-01 -3.43275189e-01 1.35091197e+00 4.77494836e-01
1.30344200e+00 -5.27042329e-01 -1.12628296e-01 1.25593734e+00
1.23103619e+00 -1.39696419e-01 1.32997179e+00 2.43069902e-01
1.12267804e+00 -2.80922860e-01 1.41620502e-01 2.24980697e-01
5.13158917e-01 9.05184567e-01 7.90519714e-01 -2.17841074e-01
-7.82648981e-01 -3.15637767e-01 7.52483964e-01 9.25099254e-01
-1.38122514e-01 -6.32833719e-01 -5.47431946e-01 5.36435544e-01
-1.69543350e+00 -9.52061176e-01 2.06868281e-03 1.83497548e+00
9.51497197e-01 -2.27344289e-01 -1.20061473e-03 -3.53306562e-01
8.51009011e-01 7.41843343e-01 -7.00812638e-01 2.03688964e-01
-2.78925657e-01 6.82990253e-02 7.11492181e-01 6.11672878e-01
-1.01602221e+00 1.23452675e+00 5.73441553e+00 1.33932877e+00
-1.18820274e+00 4.17577654e-01 9.97815549e-01 7.76691921e-03
-5.64926922e-01 -4.23185043e-02 -5.14547527e-01 7.94599354e-01
6.16647542e-01 3.89327109e-02 1.07878327e+00 6.41228616e-01
2.30807424e-01 3.19232106e-01 -7.01068163e-01 1.07538247e+00
5.96376359e-02 -1.58150530e+00 6.47262391e-03 -1.64149655e-03
1.01880574e+00 3.80411923e-01 2.64372557e-01 2.49306023e-01
5.06517291e-01 -1.22840405e+00 7.97253549e-01 1.23112202e-01
1.37933135e+00 -6.91109359e-01 6.63256884e-01 1.21341065e-01
-9.28333223e-01 2.03847811e-01 -2.37353027e-01 3.15831453e-01
5.42925000e-01 8.11179996e-01 1.32705554e-01 7.90201247e-01
7.12527871e-01 1.07812893e+00 -3.30375314e-01 8.21313679e-01
-5.89947164e-01 8.32283974e-01 2.48524807e-02 6.48586810e-01
2.39239663e-01 -1.70370564e-01 6.95208371e-01 1.10990977e+00
7.29840219e-01 1.08762734e-01 -3.23694110e-01 1.04547381e+00
-5.23600638e-01 -4.15696591e-01 -1.71670616e-01 1.68958023e-01
6.08711004e-01 1.07601321e+00 -2.40598768e-01 -3.42462182e-01
-1.65294513e-01 1.63953531e+00 1.74930263e-02 5.08675754e-01
-1.27097797e+00 -1.62804753e-01 1.12419140e+00 2.71103010e-02
5.40098727e-01 -2.31615424e-01 7.88749531e-02 -1.57046354e+00
-1.40604496e-01 -1.31228447e+00 1.58544239e-02 -1.08145392e+00
-1.46655297e+00 8.79429996e-01 -2.96828240e-01 -1.18676043e+00
-1.16014615e-01 -3.77065957e-01 -6.92002177e-01 1.00814819e+00
-1.59324455e+00 -1.20536876e+00 -7.17483461e-01 6.75314546e-01
6.12996817e-01 -3.24372873e-02 4.94598866e-01 4.48505819e-01
-6.19101524e-01 8.57822061e-01 2.07448795e-01 1.77068368e-01
8.72839153e-01 -7.98260331e-01 1.20620406e+00 1.31481707e+00
-1.14778034e-01 2.41302654e-01 6.22143388e-01 -5.88448644e-01
-1.54446125e+00 -1.33470392e+00 7.43033364e-02 -3.62155050e-01
6.86116934e-01 -6.50355443e-02 -8.36841106e-01 6.26973808e-01
4.63511795e-01 4.42575544e-01 8.12396631e-02 -5.23696125e-01
-6.20135128e-01 1.88631397e-02 -1.12866056e+00 7.47416735e-01
1.35412741e+00 -4.65486109e-01 -1.39720351e-01 1.37571663e-01
1.12087965e+00 -7.72308648e-01 -8.31337750e-01 4.03173327e-01
4.00815219e-01 -1.05783689e+00 1.18433738e+00 -3.31246853e-01
9.59395409e-01 -3.38539809e-01 -5.34602888e-02 -1.68981206e+00
-6.47605777e-01 -1.28522933e+00 -3.45363468e-01 1.13875151e+00
1.07129231e-01 -8.76875103e-01 5.59445083e-01 7.08397180e-02
-2.95906782e-01 -7.15297222e-01 -9.72631872e-01 -9.54050064e-01
5.18442333e-01 -3.97649139e-01 6.72803283e-01 9.61178303e-01
-6.35517538e-01 1.70264855e-01 -1.12215590e+00 1.19504035e-01
6.35140777e-01 -1.07965348e-02 7.64957070e-01 -4.14453924e-01
-7.76472390e-01 -4.85108346e-01 -3.93004239e-01 -1.89986610e+00
9.02308524e-02 -8.43638361e-01 -7.35749081e-02 -1.45572543e+00
2.53299445e-01 -3.80842358e-01 -1.33317396e-01 5.02743304e-01
-5.99946618e-01 5.84989011e-01 4.32406515e-01 5.97192347e-01
-2.50736684e-01 9.09643710e-01 1.67510891e+00 -2.30971783e-01
-9.52316895e-02 -2.60686189e-01 -9.02388334e-01 4.15136784e-01
6.96867526e-01 -2.91119516e-01 -2.67930031e-01 -8.52733731e-01
4.52892222e-02 2.42302150e-01 6.23526275e-01 -9.30954695e-01
4.13912423e-02 -4.34655026e-02 7.00125471e-02 -1.63025066e-01
1.54257923e-01 -4.50691253e-01 7.20632732e-01 2.58430332e-01
-7.92446956e-02 1.57737985e-01 7.28536025e-02 5.53517222e-01
-1.55588597e-01 2.24721804e-01 9.01577890e-01 5.08259982e-02
-7.92382777e-01 5.72476625e-01 -1.11755103e-01 3.03182602e-01
7.36488342e-01 1.87483400e-01 -9.34113085e-01 -7.77353525e-01
-3.02733809e-01 3.18412781e-02 9.54305410e-01 5.10346472e-01
5.53420067e-01 -1.50211465e+00 -1.02464962e+00 1.27107874e-01
-2.66635895e-01 4.24022153e-02 5.80240905e-01 9.20924962e-01
-6.49433732e-01 1.90903489e-02 -1.94340691e-01 -3.86154681e-01
-8.17144990e-01 6.16196930e-01 4.22609568e-01 -4.03121442e-01
-9.00368154e-01 1.00604773e+00 6.24858618e-01 -2.56228019e-02
-1.27107188e-01 -1.05377264e-01 4.35064346e-01 -4.64992613e-01
7.09044456e-01 4.92675185e-01 -1.47427976e-01 -4.98327553e-01
-2.33179033e-01 6.24311924e-01 5.91958649e-02 4.85780314e-02
1.21486211e+00 -3.80169183e-01 -1.44408613e-01 -4.64694887e-01
1.21096587e+00 1.63813621e-01 -1.80808198e+00 -6.10576309e-02
-7.73263931e-01 -5.34273088e-01 4.42294151e-01 -8.76528025e-01
-1.85003722e+00 5.26424170e-01 5.47888398e-01 -2.85689980e-01
1.52275264e+00 1.05739579e-01 1.32455266e+00 -3.79031599e-01
5.48653543e-01 -1.43365517e-01 2.45770171e-01 3.18524837e-01
1.03772604e+00 -9.88543391e-01 1.88462157e-03 -3.00288290e-01
-6.52843297e-01 7.31709361e-01 3.46845895e-01 -3.87541890e-01
4.05947298e-01 4.13396001e-01 4.74564917e-02 1.97573230e-01
-5.89586914e-01 1.45907342e-01 2.60990024e-01 5.38750768e-01
8.94834325e-02 1.80840120e-01 -2.16539666e-01 6.92125022e-01
-5.98958731e-02 1.43084884e-01 5.11543095e-01 2.95216471e-01
2.49337219e-02 -9.57117558e-01 -2.81839490e-01 1.69838760e-02
-6.00491047e-01 -6.02286994e-01 1.11205459e-01 6.47332728e-01
-8.62993822e-02 8.46212387e-01 -2.02009752e-01 -6.51511729e-01
1.67764146e-02 -6.91452742e-01 4.18144017e-01 4.24601398e-02
-4.60143119e-01 -1.60097610e-02 -1.03607364e-01 -9.36423302e-01
-1.41315892e-01 -1.84836671e-01 -8.00432563e-01 -7.64278889e-01
-1.99463397e-01 -2.14581370e-01 2.56906897e-01 6.75887406e-01
7.65330017e-01 6.50728345e-01 6.98679805e-01 -1.17853868e+00
-6.54994726e-01 -9.65509355e-01 -3.90013099e-01 2.12071881e-01
5.86610854e-01 -3.11214715e-01 -5.45090079e-01 1.75216272e-01] | [11.43323802947998, -2.300601005554199] |
74db058d-1bc4-4dd7-9400-c9ac8247a213 | initiative-defense-against-facial | 2112.10098 | null | https://arxiv.org/abs/2112.10098v1 | https://arxiv.org/pdf/2112.10098v1.pdf | Initiative Defense against Facial Manipulation | Benefiting from the development of generative adversarial networks (GAN), facial manipulation has achieved significant progress in both academia and industry recently. It inspires an increasing number of entertainment applications but also incurs severe threats to individual privacy and even political security meanwhile. To mitigate such risks, many countermeasures have been proposed. However, the great majority methods are designed in a passive manner, which is to detect whether the facial images or videos are tampered after their wide propagation. These detection-based methods have a fatal limitation, that is, they only work for ex-post forensics but can not prevent the engendering of malicious behavior. To address the limitation, in this paper, we propose a novel framework of initiative defense to degrade the performance of facial manipulation models controlled by malicious users. The basic idea is to actively inject imperceptible venom into target facial data before manipulation. To this end, we first imitate the target manipulation model with a surrogate model, and then devise a poison perturbation generator to obtain the desired venom. An alternating training strategy are further leveraged to train both the surrogate model and the perturbation generator. Two typical facial manipulation tasks: face attribute editing and face reenactment, are considered in our initiative defense framework. Extensive experiments demonstrate the effectiveness and robustness of our framework in different settings. Finally, we hope this work can shed some light on initiative countermeasures against more adversarial scenarios. | ['Nenghai Yu', 'WeimingZhang', 'Wenbo Zhou', 'Jie Zhang', 'Qidong Huang'] | 2021-12-19 | null | null | null | null | ['face-reenactment'] | ['computer-vision'] | [ 5.96494973e-01 1.12137787e-01 -1.23882130e-01 -9.78611857e-02
-3.77547622e-01 -1.02094889e+00 6.29656017e-01 -5.48276246e-01
-4.79990244e-02 5.99614620e-01 -1.47524089e-01 -1.13146998e-01
2.63801962e-01 -9.85852063e-01 -6.97698414e-01 -9.91902471e-01
2.48499230e-01 -7.52166882e-02 -7.09608793e-02 -1.72189146e-01
1.22166872e-01 7.64038622e-01 -1.10176361e+00 2.65320037e-02
8.26909125e-01 8.91137362e-01 -3.84713680e-01 2.89285451e-01
1.37553275e-01 7.08264887e-01 -9.33139086e-01 -8.83672535e-01
5.03778636e-01 -6.59815609e-01 -3.47658545e-01 1.66007448e-02
7.78435916e-02 -7.78520763e-01 -5.10623872e-01 1.41227388e+00
6.16934717e-01 -3.87999684e-01 3.26419234e-01 -1.77676594e+00
-7.83927858e-01 3.55951995e-01 -9.19831932e-01 -1.69624835e-01
4.01899159e-01 7.34947503e-01 2.60302901e-01 -3.81466240e-01
5.20122766e-01 1.42434621e+00 3.70701134e-01 1.25599408e+00
-1.04831588e+00 -1.41321480e+00 -5.78988381e-02 -3.23087662e-01
-1.13546729e+00 -6.46068156e-01 1.02148998e+00 -2.34807253e-01
4.79450300e-02 3.39765429e-01 6.18122339e-01 1.59247136e+00
6.30152225e-02 7.23161876e-01 1.25839651e+00 1.03945054e-01
1.57686338e-01 2.24639952e-01 -4.44779575e-01 7.44188070e-01
2.18677714e-01 3.73691589e-01 -3.01681906e-01 -6.61896527e-01
8.53444636e-01 1.06273927e-01 -5.03117800e-01 1.95824578e-02
-5.87932408e-01 9.21034992e-01 2.71166176e-01 -3.58455181e-02
-4.19618666e-01 1.18505940e-01 4.54004645e-01 2.26026788e-01
3.24623257e-01 3.44623178e-01 6.79396733e-04 -2.45318357e-02
-5.06967187e-01 3.71804565e-01 7.83079743e-01 5.47798991e-01
4.20604140e-01 2.53766090e-01 -6.70608282e-02 4.90811259e-01
2.91465640e-01 6.83516383e-01 3.11370850e-01 -8.53504181e-01
1.60345495e-01 5.86911321e-01 -1.29504696e-01 -1.34389937e+00
2.76310712e-01 2.64255926e-02 -8.11926782e-01 4.35564071e-01
3.48552287e-01 -3.49448115e-01 -8.53776038e-01 1.94625866e+00
7.03744173e-01 2.83584982e-01 -6.23423941e-02 6.90737247e-01
4.09055829e-01 5.92443466e-01 2.33793586e-01 -3.70926231e-01
1.21947181e+00 -4.38793391e-01 -7.80904233e-01 4.06555310e-02
6.85617849e-02 -5.97676337e-01 9.60680425e-01 3.95677358e-01
-1.00288188e+00 -1.08826853e-01 -1.19287574e+00 4.85486060e-01
-2.75845170e-01 -3.31295699e-01 6.47454321e-01 1.17524350e+00
-5.79492033e-01 5.67649841e-01 -7.33665049e-01 -1.48573726e-01
1.01606274e+00 5.46499074e-01 -2.99447894e-01 7.48640895e-02
-1.21131611e+00 4.36054319e-01 -1.45710811e-01 2.32967764e-01
-1.28950775e+00 -5.68364918e-01 -5.48382699e-01 -2.09729537e-01
5.92002332e-01 -4.77027237e-01 8.97979498e-01 -1.11259115e+00
-1.70128918e+00 8.75683308e-01 2.58487850e-01 -4.54083622e-01
7.96786845e-01 -2.10282832e-01 -3.62599462e-01 1.43768117e-01
-1.59536730e-02 4.96766329e-01 1.28698754e+00 -1.26749563e+00
-9.14237872e-02 -6.10098362e-01 1.91273853e-01 -1.78507030e-01
-7.25209773e-01 4.16910350e-01 -1.96712181e-01 -9.06670630e-01
-3.78078789e-01 -9.98535931e-01 -1.06810248e-02 1.95564911e-01
-7.65338480e-01 3.92786525e-02 1.45156574e+00 -5.67542434e-01
9.92666483e-01 -2.14916897e+00 7.37717375e-02 1.54796422e-01
2.78219879e-01 8.84276569e-01 -2.17176437e-01 3.32099527e-01
1.73314124e-01 6.32854342e-01 -3.50960970e-01 3.23726647e-02
-2.30783775e-01 1.03009969e-01 -9.21753168e-01 5.97377956e-01
2.49214575e-01 9.18508649e-01 -7.35725820e-01 -2.61720330e-01
-1.75895289e-01 6.68174922e-01 -4.50140715e-01 5.06315291e-01
-2.39555463e-01 6.97456837e-01 -8.93001437e-01 1.07029748e+00
9.10344303e-01 2.34626696e-01 -1.51035143e-02 -7.84223303e-02
2.79807836e-01 -2.57280529e-01 -6.19492590e-01 1.03098857e+00
5.20875379e-02 1.84215650e-01 4.58894432e-01 -5.70728898e-01
7.66591728e-01 2.56783038e-01 4.24005568e-01 -2.79771358e-01
4.91459161e-01 4.67601307e-02 6.07563294e-02 -5.64461529e-01
-9.96676162e-02 -2.81074882e-01 -1.61136556e-02 6.34647846e-01
-3.40246528e-01 -1.21512026e-01 -2.98778027e-01 1.38188943e-01
1.06665564e+00 2.05290601e-01 8.41286927e-02 1.80198744e-01
6.82751477e-01 -2.26907030e-01 8.11507046e-01 3.62933755e-01
-5.17917514e-01 2.20032468e-01 7.57725775e-01 -2.13568315e-01
-8.16909730e-01 -1.02926588e+00 2.42384315e-01 6.34119809e-01
1.58407420e-01 -3.23725075e-01 -1.27986073e+00 -1.20506072e+00
1.35819674e-01 3.58139336e-01 -7.25429058e-01 -7.75416434e-01
-7.48950660e-01 -6.43200159e-01 1.37226200e+00 2.53245741e-01
9.43176925e-01 -1.12049794e+00 -3.92434657e-01 -9.37636420e-02
8.61217454e-02 -8.99254203e-01 -4.74304855e-01 -6.38064742e-01
-4.76729661e-01 -1.27021694e+00 -6.51053011e-01 -4.04517919e-01
6.28250241e-01 2.12328508e-01 3.17507714e-01 4.43792939e-01
-3.92745912e-01 1.67604715e-01 -1.41005099e-01 -7.61255741e-01
-7.08418906e-01 -2.98641503e-01 3.48733425e-01 3.48175049e-01
2.49084309e-01 -6.78376973e-01 -5.29990137e-01 5.17462790e-01
-1.20529974e+00 -4.54563409e-01 4.76129264e-01 6.74784064e-01
2.37146586e-01 1.83134839e-01 7.60856688e-01 -1.00673282e+00
9.87879813e-01 -4.40142661e-01 -6.36056960e-01 1.73178688e-01
-3.30900401e-01 -3.04669678e-01 9.15764749e-01 -9.17061806e-01
-9.13956285e-01 -4.46535461e-02 -2.53995836e-01 -8.10778379e-01
-8.71139243e-02 1.25221014e-02 -8.41729641e-01 -5.37524462e-01
3.93210918e-01 3.86520028e-01 5.45992374e-01 -2.27940068e-01
3.13246369e-01 6.24610245e-01 4.64194089e-01 -6.49371266e-01
1.42809248e+00 7.06820726e-01 7.81174377e-02 -8.11818540e-01
-4.99165624e-01 3.36135864e-01 -1.01579763e-01 -3.84040803e-01
8.32714736e-01 -5.14255464e-01 -1.10730433e+00 1.02051556e+00
-1.24207366e+00 3.58330235e-02 1.86832547e-01 -1.66218594e-01
-4.06500310e-01 5.15960932e-01 -5.89139819e-01 -1.02098393e+00
-4.68857557e-01 -1.20511842e+00 9.13834870e-01 5.08650601e-01
-8.05201009e-03 -5.71880758e-01 -6.75129294e-02 5.07741749e-01
6.73219085e-01 7.59896934e-01 7.84733772e-01 -4.67584223e-01
-6.60356700e-01 -5.07898510e-01 5.07977465e-03 5.40180087e-01
4.56804127e-01 4.55127321e-02 -9.53984618e-01 -3.49303812e-01
5.32046974e-01 -6.43966496e-01 4.35968429e-01 -1.92769378e-01
1.34272265e+00 -7.30692983e-01 -4.11536098e-01 7.26454020e-01
9.84933436e-01 6.27493262e-01 8.81159365e-01 8.62606913e-02
7.51032472e-01 8.30246687e-01 6.02757812e-01 3.67141664e-01
-3.06759715e-01 4.98768002e-01 8.16657960e-01 3.25593129e-02
1.72956333e-01 -5.22203505e-01 6.49209440e-01 2.14086170e-03
-1.17618732e-01 -3.15747648e-01 -3.69402260e-01 -1.08721316e-01
-1.39072990e+00 -1.08418095e+00 4.04637873e-01 2.08545661e+00
8.57678831e-01 7.59316329e-03 3.08063805e-01 -5.21884523e-02
8.81215334e-01 2.41051793e-01 -9.80843008e-01 -1.16495423e-01
4.46959250e-02 2.55256832e-01 3.47614080e-01 -4.82639531e-03
-9.82963502e-01 1.21723521e+00 5.78560448e+00 7.89466560e-01
-1.35407507e+00 4.58353721e-02 7.31253743e-01 -1.12171106e-01
-3.32188196e-02 7.40592927e-02 -6.92693710e-01 7.06686437e-01
6.19134843e-01 -2.08029956e-01 5.65273881e-01 8.24325800e-01
1.88960597e-01 3.80657464e-01 -7.56875336e-01 8.11897457e-01
2.22062394e-01 -1.12394428e+00 2.74286419e-01 4.80653793e-01
2.32256114e-01 -7.12608695e-01 3.93937558e-01 1.92330509e-01
1.42295629e-01 -1.12364757e+00 4.02294993e-01 2.50506908e-01
7.21552074e-01 -1.08646357e+00 3.16878140e-01 4.07797843e-01
-8.56201828e-01 -1.16327845e-01 -2.39901811e-01 3.07707578e-01
3.33235472e-01 1.87401280e-01 -5.08790970e-01 2.25458086e-01
3.82441044e-01 3.22120160e-01 -3.05532634e-01 4.79305536e-01
-5.42618036e-01 7.64436781e-01 -1.19637296e-01 4.53411564e-02
3.12200114e-02 -3.60055178e-01 8.28782976e-01 5.46631157e-01
6.28998727e-02 2.73488194e-01 7.59170428e-02 1.24323153e+00
-3.99445266e-01 -1.32163927e-01 -1.15157676e+00 -5.42778671e-01
5.94071925e-01 1.39297736e+00 -4.02780592e-01 1.39406987e-03
-2.92101335e-02 1.17671025e+00 8.59522074e-02 1.60326898e-01
-1.24148333e+00 -4.29089278e-01 9.67339873e-01 2.75683284e-01
-1.82797045e-01 1.90574154e-02 1.73820872e-02 -1.04305661e+00
1.55677963e-02 -1.38863933e+00 6.55534342e-02 -3.70162606e-01
-1.34335387e+00 4.86909509e-01 -1.80141002e-01 -1.06405079e+00
-8.38814750e-02 -3.78678650e-01 -9.71806467e-01 7.27188647e-01
-1.06114089e+00 -1.36053252e+00 -2.22097784e-01 6.52213275e-01
3.07927996e-01 -3.10000062e-01 5.99656999e-01 1.02677099e-01
-9.65963781e-01 8.81910920e-01 -5.14074624e-01 4.72795457e-01
5.71608007e-01 -5.06739795e-01 2.48935074e-01 8.76228273e-01
-3.09702784e-01 1.09581554e+00 6.68033421e-01 -9.87136364e-01
-1.78925800e+00 -1.15887451e+00 5.27580380e-02 -3.48606199e-01
6.35031819e-01 -5.75753629e-01 -8.76906812e-01 5.95709682e-01
1.65834129e-01 -1.65928215e-01 5.99076271e-01 -6.12697899e-01
-5.16592145e-01 -4.80628014e-02 -1.73776817e+00 7.97937632e-01
9.95337367e-01 -5.22730589e-01 -1.76733404e-01 4.15019333e-01
7.18312502e-01 -4.86010909e-02 -6.15155220e-01 4.40063089e-01
6.72231257e-01 -8.57288480e-01 9.32313502e-01 -7.90873647e-01
3.65071982e-01 -2.76384383e-01 3.19924355e-02 -9.24630344e-01
2.33573452e-01 -1.19773448e+00 -3.14477593e-01 1.77127409e+00
-1.66358408e-02 -8.96436214e-01 9.98752952e-01 6.07122242e-01
4.23460633e-01 -8.13383758e-01 -7.02277482e-01 -9.98048365e-01
1.52954802e-01 1.65293757e-02 7.36432076e-01 9.94448721e-01
-2.78762490e-01 8.90642703e-02 -8.21387708e-01 1.72121942e-01
8.39188516e-01 -2.78639793e-01 1.11523330e+00 -7.95115829e-01
-2.25092709e-01 -3.07172239e-01 -4.07902032e-01 -6.20130360e-01
4.44849402e-01 -5.24510503e-01 -1.60935104e-01 -6.33103967e-01
1.77592888e-01 -1.93058595e-01 1.35768563e-01 6.99068010e-01
-1.88558802e-01 3.55365962e-01 1.63804263e-01 3.25958461e-01
1.25689581e-01 6.60615921e-01 1.20041418e+00 -1.76669896e-01
-5.37919737e-02 8.97844806e-02 -1.01232481e+00 7.49725997e-01
9.34129894e-01 -5.99846959e-01 -4.34341103e-01 2.43568718e-02
-7.08436966e-02 -3.47721912e-02 5.72633743e-01 -7.70629346e-01
-4.92535122e-02 -4.01809216e-01 1.48944661e-01 -5.67558818e-02
2.97583103e-01 -7.10553527e-01 7.81957135e-02 6.95913851e-01
1.97964329e-02 -2.80433451e-03 1.37199806e-02 7.29950547e-01
-2.52217874e-02 3.71839143e-02 1.11738586e+00 -1.78471476e-01
-8.71793926e-02 6.29338026e-01 -2.39819661e-01 -1.57442838e-01
1.42168486e+00 -1.35346979e-01 -5.26894987e-01 -4.48859632e-01
-3.30243886e-01 1.70850798e-01 7.91714549e-01 3.97942692e-01
7.47949183e-01 -1.29035842e+00 -6.22062683e-01 3.77528250e-01
-2.03400984e-01 -4.33205158e-01 4.84647900e-02 7.35439181e-01
-2.57341295e-01 -1.52689829e-01 -3.22334200e-01 -2.61968702e-01
-1.45244503e+00 8.86748493e-01 3.07684630e-01 7.90470690e-02
-3.45216811e-01 6.95321798e-01 3.53004485e-01 -1.28340468e-01
1.23787299e-01 5.05179763e-01 -7.32404068e-02 -2.44189262e-01
7.78501093e-01 3.07776034e-01 -4.28721547e-01 -7.06958771e-01
-2.56804079e-01 3.90628308e-01 -3.94516468e-01 3.49658206e-02
1.10308349e+00 1.81014538e-01 -3.30248713e-01 -2.99370378e-01
1.16432631e+00 3.61887485e-01 -1.22843993e+00 2.96579450e-01
-3.92745823e-01 -7.66482472e-01 -3.84701997e-01 -6.97906256e-01
-1.43600380e+00 7.07041442e-01 4.69170272e-01 3.43848795e-01
1.26260996e+00 -1.94998458e-01 1.20937002e+00 8.78450274e-02
4.01595116e-01 -6.08646870e-01 2.90651709e-01 -6.38693795e-02
8.57582688e-01 -9.44550514e-01 -1.50589749e-01 -6.71764076e-01
-5.84414065e-01 8.93389046e-01 8.95224810e-01 -2.49779582e-01
3.95842344e-01 4.38978523e-01 4.70040180e-02 -3.44498098e-01
-3.38147014e-01 3.80055279e-01 -1.55489519e-01 8.53870571e-01
8.27992931e-02 -2.08876282e-01 -2.86552191e-01 6.90577090e-01
-2.80151237e-02 -6.31438866e-02 4.49246258e-01 1.10919595e+00
-2.49232560e-01 -1.40743542e+00 -4.72888082e-01 2.14230165e-01
-7.78141379e-01 2.38012910e-01 -1.11039317e+00 8.21164727e-01
1.22915812e-01 9.47863460e-01 -5.58222413e-01 -6.33875251e-01
2.34553710e-01 -1.21592201e-01 3.75350565e-01 -2.76776135e-01
-7.41724491e-01 3.29552032e-02 -2.77653843e-01 -7.08306015e-01
-1.77709848e-01 -5.35625517e-01 -9.22655165e-01 -5.80389857e-01
-2.95356214e-01 4.79004718e-02 5.21971345e-01 7.29503989e-01
4.01787251e-01 3.08726858e-02 1.17732394e+00 -6.25548422e-01
-1.00803781e+00 -7.44939387e-01 -5.04347205e-01 4.63848442e-01
6.36586547e-02 -7.14924693e-01 -5.93567133e-01 -1.47594064e-01] | [12.78032112121582, 0.9016807079315186] |
f4b49d63-177b-468d-8858-8174a05823b8 | topics-to-avoid-demoting-latent-confounds-in | 1909.00453 | null | https://arxiv.org/abs/1909.00453v2 | https://arxiv.org/pdf/1909.00453v2.pdf | Topics to Avoid: Demoting Latent Confounds in Text Classification | Despite impressive performance on many text classification tasks, deep neural networks tend to learn frequent superficial patterns that are specific to the training data and do not always generalize well. In this work, we observe this limitation with respect to the task of native language identification. We find that standard text classifiers which perform well on the test set end up learning topical features which are confounds of the prediction task (e.g., if the input text mentions Sweden, the classifier predicts that the author's native language is Swedish). We propose a method that represents the latent topical confounds and a model which "unlearns" confounding features by predicting both the label of the input text and the confound; but we train the two predictors adversarially in an alternating fashion to learn a text representation that predicts the correct label but is less prone to using information about the confound. We show that this model generalizes better and learns features that are indicative of the writing style rather than the content. | ['Noah A. Smith', 'Yulia Tsvetkov', 'Shuly Wintner', 'Sachin Kumar'] | 2019-09-01 | topics-to-avoid-demoting-latent-confounds-in-1 | https://aclanthology.org/D19-1425 | https://aclanthology.org/D19-1425.pdf | ijcnlp-2019-11 | ['native-language-identification'] | ['natural-language-processing'] | [ 3.81238729e-01 2.10150495e-01 -4.99726772e-01 -6.75667822e-01
-5.43617010e-01 -8.94187212e-01 9.15797353e-01 1.41228110e-01
-4.85135823e-01 7.57743001e-01 5.60027719e-01 -4.78624851e-01
2.37808615e-01 -7.20129907e-01 -7.45345891e-01 -5.23021281e-01
2.78161198e-01 6.29876316e-01 -3.99145484e-01 1.61310315e-01
2.92392582e-01 1.25810340e-01 -1.20563388e+00 4.98850822e-01
7.84359396e-01 6.22700095e-01 -7.37353563e-02 7.54077971e-01
-3.99097353e-01 1.00147867e+00 -8.81499648e-01 -6.66866899e-01
2.12592736e-01 -5.19313633e-01 -1.06964338e+00 -1.31606802e-01
1.17961311e+00 -1.83628261e-01 -4.36743855e-01 1.11252093e+00
-9.19230096e-03 -1.09410875e-01 1.20547473e+00 -1.03093314e+00
-8.41512799e-01 1.02226400e+00 -4.81073886e-01 2.62391210e-01
2.42779717e-01 -5.89116663e-02 1.31682312e+00 -5.81078410e-01
6.16451919e-01 1.10408008e+00 1.04919028e+00 6.82213962e-01
-1.54119718e+00 -8.91497970e-01 5.35431564e-01 -9.71451253e-02
-9.14881468e-01 -4.08404201e-01 7.32867301e-01 -8.47239494e-01
5.02679288e-01 2.09275469e-01 2.00923085e-01 2.04413819e+00
5.22814453e-01 6.38936579e-01 1.44376194e+00 -4.20632988e-01
-1.18332334e-01 5.01222908e-01 5.65261483e-01 7.44711161e-01
1.92616165e-01 9.66142491e-02 -6.96139038e-01 -1.64083079e-01
1.93411753e-01 -1.76357225e-01 -2.53505290e-01 1.19704872e-01
-1.26547015e+00 8.29151869e-01 2.19681099e-01 4.32665050e-01
5.42074256e-03 1.40646294e-01 4.12505597e-01 5.52406311e-01
6.70379579e-01 8.69175792e-01 -7.83131897e-01 1.39044486e-02
-1.31317627e+00 1.95323959e-01 1.14290178e+00 7.06860483e-01
7.86518455e-01 -2.41048336e-01 -2.30124816e-01 8.87431622e-01
-2.06343569e-02 4.94656563e-01 9.15896237e-01 -4.88454133e-01
4.16371047e-01 4.32671815e-01 -1.31423995e-01 -8.42865169e-01
-5.10042846e-01 -6.27030909e-01 -7.40210950e-01 7.83670619e-02
9.05828774e-01 -3.36128503e-01 -9.60555017e-01 2.15555501e+00
-4.38747525e-01 -1.13949478e-01 1.22131467e-01 4.11091506e-01
6.48997247e-01 5.43693066e-01 3.69667977e-01 -1.38143152e-01
1.00160468e+00 -8.37208390e-01 -6.61605895e-01 -5.86055875e-01
1.00161660e+00 -7.14314640e-01 1.27617681e+00 3.44062507e-01
-8.80921423e-01 -6.05921686e-01 -1.16920006e+00 -2.66342998e-01
-6.58436298e-01 3.21943700e-01 5.39672852e-01 5.02594173e-01
-7.92541862e-01 8.42290163e-01 -4.40785497e-01 -4.12306726e-01
2.88549036e-01 3.54614437e-01 -3.04891020e-01 1.36282608e-01
-1.13061857e+00 1.17064452e+00 1.96136057e-01 -4.24045950e-01
-5.85743129e-01 -8.63159478e-01 -4.35223669e-01 1.96294665e-01
9.11701545e-02 -5.01084626e-01 1.19498742e+00 -1.85879946e+00
-1.26843083e+00 1.27198935e+00 -3.98319989e-01 -2.33234674e-01
7.74739802e-01 -1.85776725e-01 -3.18261385e-01 -4.05184418e-01
4.80925113e-01 2.95518190e-01 9.22373891e-01 -1.03270578e+00
-5.36307037e-01 -4.99115378e-01 -1.92059666e-01 5.35461716e-02
-6.40148103e-01 -1.37555644e-01 1.54017031e-01 -9.13602710e-01
-1.35445818e-01 -8.50054502e-01 1.95165083e-01 -1.18644327e-01
-8.05999398e-01 -3.89568478e-01 7.27837682e-01 -6.82642102e-01
1.09265232e+00 -2.15564179e+00 2.79617965e-01 3.36486727e-01
3.36196899e-01 -8.06976184e-02 -7.76951984e-02 1.83307111e-01
-2.24881604e-01 5.42363703e-01 -1.49686754e-01 -4.95329738e-01
1.64540689e-02 1.22373231e-01 -6.55223370e-01 4.65549588e-01
1.67745724e-01 7.47304261e-01 -7.79018700e-01 -2.18303680e-01
-4.85137463e-01 -5.06324433e-02 -2.81658053e-01 2.50456542e-01
-4.27901566e-01 3.06801111e-01 -8.64437371e-02 1.13506012e-01
3.22753638e-01 -2.69820362e-01 2.31213391e-01 1.39974738e-02
-5.92909157e-02 7.17219472e-01 -6.97642624e-01 1.31003773e+00
-5.91470301e-01 1.13829684e+00 -2.82395948e-02 -9.13917899e-01
9.56838727e-01 2.21495971e-01 4.74516414e-02 -2.95505226e-01
8.88821483e-02 2.94549406e-01 1.08178012e-01 -4.87752318e-01
2.29187861e-01 -4.88217562e-01 -1.55689418e-01 7.88938344e-01
2.81797200e-01 2.45066032e-01 -2.51160026e-01 1.98318288e-01
1.05568051e+00 6.37580873e-03 2.61005312e-01 -5.64343750e-01
4.37045574e-01 -6.21042959e-02 4.08307999e-01 1.14021111e+00
8.14873651e-02 2.55135924e-01 8.68084073e-01 -4.29487735e-01
-1.05722177e+00 -9.98768985e-01 -2.57920921e-01 1.31769907e+00
-2.40283698e-01 -5.61799586e-01 -5.32493711e-01 -1.26082468e+00
2.89606303e-01 1.02368116e+00 -1.17271543e+00 -3.43479246e-01
-3.92002344e-01 -5.46078742e-01 7.24579692e-01 5.91536999e-01
9.56560373e-02 -7.45769143e-01 3.29533778e-02 -2.66910587e-02
-1.65925007e-02 -8.31543446e-01 -5.25944769e-01 6.68800175e-01
-5.91271698e-01 -1.02984250e+00 -5.61247528e-01 -8.61862183e-01
6.97016716e-01 -3.96185637e-01 1.28209078e+00 2.80588627e-01
-5.18280566e-02 1.75211743e-01 1.39949583e-02 -5.61472058e-01
-9.01807666e-01 5.35476267e-01 5.28335869e-02 5.02734724e-03
7.86521077e-01 -4.99006420e-01 -9.80676860e-02 -7.15264603e-02
-4.83027130e-01 1.37703195e-01 4.24731702e-01 1.17712343e+00
-5.08912168e-02 -6.07583486e-02 6.75962508e-01 -1.67643881e+00
5.82747281e-01 -7.17354059e-01 -1.68087557e-01 2.97666788e-01
-7.44359851e-01 3.00981373e-01 9.98850286e-01 -7.80271828e-01
-9.63714778e-01 4.09247465e-02 -6.46056905e-02 3.85280214e-02
-4.69826967e-01 6.15930080e-01 -5.85959181e-02 2.65108585e-01
9.09342825e-01 1.44983709e-01 -3.22966576e-02 -5.58331668e-01
4.82944809e-02 8.63155186e-01 4.52195406e-01 -8.15504313e-01
9.25003827e-01 1.09286867e-01 -1.50234979e-02 -7.72638977e-01
-1.47464657e+00 -1.73629388e-01 -9.01722133e-01 6.41539618e-02
7.13462353e-01 -7.90338993e-01 -4.35199469e-01 4.24251527e-01
-1.28821337e+00 -4.46922570e-01 -2.21227363e-01 3.73334050e-01
-4.19878036e-01 1.06172986e-01 -5.26373208e-01 -5.96634507e-01
5.39332964e-02 -8.24858546e-01 6.85636520e-01 -9.46057215e-02
-7.60639071e-01 -1.36980832e+00 -1.29981544e-02 5.69230360e-05
3.32474709e-01 1.92259415e-03 1.51853561e+00 -1.23307943e+00
-1.56714350e-01 -2.10121810e-01 -1.02494903e-01 4.93915200e-01
1.90571442e-01 2.69441754e-02 -1.24396551e+00 -8.88344869e-02
3.02182585e-02 -3.97498012e-01 1.14808512e+00 1.59694895e-01
1.31653976e+00 -7.36244440e-01 -3.98213714e-01 7.15629280e-01
1.21626925e+00 -2.03653410e-01 1.12640247e-01 2.46083602e-01
9.52673793e-01 9.06562567e-01 -1.99461684e-01 -1.88711375e-01
1.57519266e-01 5.20977080e-01 -2.21499428e-01 -5.24938852e-02
-1.04129158e-01 -5.71922243e-01 5.61287105e-01 6.40052378e-01
2.79682279e-01 -1.61848158e-01 -9.93485868e-01 3.86068523e-01
-1.54869866e+00 -9.52798247e-01 -2.03765988e-01 2.01164055e+00
1.39825845e+00 3.71550322e-01 1.25496045e-01 1.32454202e-01
3.71856391e-01 5.34231886e-02 -5.62134743e-01 -5.52713513e-01
-2.69376308e-01 3.50542098e-01 5.85140288e-01 8.25038791e-01
-1.16166842e+00 9.48409021e-01 6.91712523e+00 3.35628331e-01
-1.31168437e+00 7.66485138e-03 7.33691156e-01 5.60640066e-04
-4.13605034e-01 -1.56849369e-01 -8.97708476e-01 4.82359082e-01
1.08455980e+00 -2.15795040e-01 3.81545722e-01 6.55257881e-01
-2.68360972e-01 1.01738296e-01 -1.78518617e+00 4.94099230e-01
3.22727084e-01 -1.01727366e+00 3.19887727e-01 6.53121620e-02
7.65841544e-01 -1.89169660e-01 3.60067606e-01 3.43931258e-01
3.88151675e-01 -1.47651482e+00 7.74962306e-01 5.05792260e-01
8.04693997e-01 -3.96723121e-01 6.69686198e-01 5.53196251e-01
-2.95811117e-01 -3.19980532e-01 -2.47141287e-01 -3.88339341e-01
-5.23247778e-01 3.84770185e-01 -9.21767235e-01 -3.60889137e-02
3.84989858e-01 1.00849080e+00 -9.93790269e-01 4.20274436e-01
-4.96331841e-01 9.62784410e-01 1.92508940e-02 -1.83331132e-01
9.83603820e-02 1.57058597e-01 5.66128969e-01 1.46789384e+00
2.67255381e-02 -2.68298537e-01 2.04973742e-01 1.21257854e+00
-4.77491945e-01 2.99197018e-01 -9.48348403e-01 -3.05605650e-01
6.26675934e-02 8.30085039e-01 -3.68919998e-01 -4.20379281e-01
-6.53963268e-01 1.05598378e+00 6.70279086e-01 6.51918650e-01
-2.05281228e-01 -2.16371670e-01 5.61762094e-01 1.37913883e-01
-8.51206630e-02 -2.28761621e-02 -9.73407030e-01 -1.41730905e+00
-8.77464935e-02 -8.63175690e-01 1.94567651e-01 -5.94394624e-01
-1.91479981e+00 4.63537723e-01 -4.01491582e-01 -5.73293447e-01
-2.30316669e-01 -1.12496889e+00 -7.11275041e-01 1.44845021e+00
-1.25826943e+00 -1.15612388e+00 -5.06031737e-02 3.39872271e-01
4.48863804e-01 -4.63300347e-01 1.12316024e+00 -2.07681373e-01
-4.31312054e-01 1.01030052e+00 1.96994171e-01 5.69195807e-01
1.07094181e+00 -1.66884673e+00 2.50100583e-01 5.40915191e-01
3.87227267e-01 7.78302252e-01 8.56296122e-01 -6.93880379e-01
-8.92357588e-01 -8.89755547e-01 1.42897785e+00 -8.98850977e-01
9.78261054e-01 -6.01144910e-01 -9.36560750e-01 1.09703207e+00
1.40981510e-01 -4.02578324e-01 9.69392240e-01 7.88101792e-01
-9.54683423e-01 8.99463817e-02 -7.41039574e-01 5.58985293e-01
8.67653191e-01 -1.04255903e+00 -8.54433715e-01 4.86269236e-01
3.33824843e-01 -8.71377140e-02 -6.15273237e-01 2.05011852e-03
6.82512522e-01 -6.64853573e-01 5.28640270e-01 -1.26522124e+00
9.45855796e-01 2.89737523e-01 -1.33724719e-01 -1.48315060e+00
-3.73062968e-01 -2.04791248e-01 1.48271117e-02 1.31504869e+00
7.59253383e-01 -4.61895496e-01 7.72699773e-01 4.88712549e-01
1.46285310e-01 -6.28080666e-01 -6.81404889e-01 -6.44454062e-01
8.43435645e-01 7.58214742e-02 2.34762698e-01 1.55093431e+00
1.30007744e-01 9.42144394e-01 -3.20486426e-01 -1.08898394e-02
4.93986547e-01 4.64666076e-02 5.02666593e-01 -1.85055685e+00
-3.58455688e-01 -8.28558087e-01 -2.12764010e-01 -8.68216455e-01
9.10059631e-01 -1.41517127e+00 -6.54318370e-03 -1.00928986e+00
4.86102313e-01 -4.25891638e-01 -1.80069208e-01 4.70903963e-01
-3.47630024e-01 -4.08616401e-02 -5.48119321e-02 2.46466741e-01
-5.66434003e-02 -5.49803562e-02 1.00613022e+00 -4.68328714e-01
5.96819408e-02 1.90389842e-01 -9.91952896e-01 7.76509285e-01
7.22485304e-01 -5.15576601e-01 -2.11989626e-01 -4.71735567e-01
5.09792686e-01 -3.04585248e-01 4.75539923e-01 -7.10992694e-01
1.98057041e-01 -1.06561780e-01 9.22740996e-01 -6.86913431e-02
3.36718559e-02 -7.79370844e-01 -3.85985494e-01 2.46998489e-01
-1.33667672e+00 -3.86736207e-02 1.12884797e-01 5.76017976e-01
-2.27594897e-02 -6.38062537e-01 7.44528770e-01 -2.22742617e-01
5.79559617e-03 1.04998752e-01 -5.31299233e-01 3.26268077e-01
4.76501465e-01 6.07947335e-02 -5.54869771e-01 -3.64132971e-01
-7.82913744e-01 -7.63775855e-02 6.70963228e-01 4.48116422e-01
-1.08616233e-01 -9.30309176e-01 -8.05278838e-01 3.41243058e-01
-2.56379577e-03 -5.72103918e-01 -3.70068997e-01 5.10682344e-01
3.61222290e-02 3.15924764e-01 -1.36331156e-01 -2.36279204e-01
-1.12424803e+00 5.73207378e-01 5.28519750e-01 -2.25869000e-01
-4.91112679e-01 1.00342357e+00 4.27805781e-01 -4.18302357e-01
4.72438693e-01 -6.12302162e-02 -7.82211795e-02 2.12224320e-01
4.09722805e-01 -4.24852259e-02 -8.04509372e-02 -5.75593531e-01
-7.36642629e-02 3.63009065e-01 -5.57925880e-01 -1.37082059e-02
1.11858249e+00 6.60505444e-02 -1.38821036e-01 1.08159399e+00
1.45677495e+00 3.83712858e-01 -9.42047894e-01 -4.09356594e-01
3.28316569e-01 -2.09979564e-01 1.40931055e-01 -1.37410164e+00
-6.84691548e-01 1.09405959e+00 2.59182125e-01 3.45171601e-01
6.71813369e-01 7.04305321e-02 3.62518132e-01 3.88844639e-01
-2.13410214e-01 -1.09389293e+00 -1.80146913e-03 8.00439060e-01
7.24965394e-01 -1.20147848e+00 2.46486757e-02 -2.67478406e-01
-4.55791116e-01 1.47324824e+00 7.32976258e-01 -8.57801810e-02
6.76601350e-01 4.21799719e-01 3.17714244e-01 -1.06774226e-01
-8.50412607e-01 2.07149193e-01 4.82393891e-01 4.81786877e-01
9.16079462e-01 1.07024021e-01 -2.30674788e-01 9.35074031e-01
-6.86855435e-01 -3.14146191e-01 5.20594418e-01 4.07401025e-01
-1.46519005e-01 -1.05551147e+00 -1.26127988e-01 1.02309406e+00
-7.66122222e-01 -2.99796492e-01 -9.44280148e-01 5.65567076e-01
1.97902143e-01 4.78670835e-01 2.41355360e-01 -6.07949078e-01
2.75151227e-02 8.48733842e-01 4.59018975e-01 -9.24169660e-01
-9.12922323e-01 -5.60948551e-01 1.01538688e-01 -1.17073692e-01
-2.28432298e-01 -7.85279572e-01 -7.11371362e-01 -3.37264538e-01
-5.87650277e-02 2.00541943e-01 5.70017159e-01 1.21214545e+00
-1.26029685e-01 3.75068188e-01 5.52922726e-01 -4.68680710e-01
-8.95135462e-01 -1.20404255e+00 -5.26891530e-01 6.75266564e-01
6.62255824e-01 -5.92178643e-01 -6.83917880e-01 3.16158831e-01] | [10.548357963562012, 10.137763023376465] |
8809ac9a-2b0e-47f4-b2e8-ada2a279d684 | switchhit-a-probabilistic-complementarity | 2203.00591 | null | https://arxiv.org/abs/2203.00591v1 | https://arxiv.org/pdf/2203.00591v1.pdf | SwitchHit: A Probabilistic, Complementarity-Based Switching System for Improved Visual Place Recognition in Changing Environments | Visual place recognition (VPR), a fundamental task in computer vision and robotics, is the problem of identifying a place mainly based on visual information. Viewpoint and appearance changes, such as due to weather and seasonal variations, make this task challenging. Currently, there is no universal VPR technique that can work in all types of environments, on a variety of robotic platforms, and under a wide range of viewpoint and appearance changes. Recent work has shown the potential of combining different VPR methods intelligently by evaluating complementarity for some specific VPR datasets to achieve better performance. This, however, requires ground truth information (correct matches) which is not available when a robot is deployed in a real-world scenario. Moreover, running multiple VPR techniques in parallel may be prohibitive for resource-constrained embedded platforms. To overcome these limitations, this paper presents a probabilistic complementarity based switching VPR system, SwitchHit. Our proposed system consists of multiple VPR techniques, however, it does not simply run all techniques at once, rather predicts the probability of correct match for an incoming query image and dynamically switches to another complementary technique if the probability of correctly matching the query is below a certain threshold. This innovative use of multiple VPR techniques allow our system to be more efficient and robust than other combined VPR approaches employing brute force and running multiple VPR techniques at once. Thus making it more suitable for resource constrained embedded systems and achieving an overall superior performance from what any individual VPR method in the system could have by achieved running independently. | ['Shoaib Ehsan', 'Klaus McDonald-Maier', 'Michael Milford', 'Maria Waheed'] | 2022-03-01 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [ 7.79349580e-02 -3.92906129e-01 -1.50911257e-01 -1.55411452e-01
-4.65607941e-01 -6.71555340e-01 5.85060537e-01 2.64535099e-01
-4.85057920e-01 6.52000189e-01 -6.25083983e-01 -3.95510584e-01
-1.92262501e-01 -8.72396350e-01 -7.84112513e-01 -6.58667445e-01
-5.41014224e-02 7.22956002e-01 9.60451126e-01 -3.29842746e-01
4.24951166e-01 9.55419660e-01 -2.04574013e+00 -1.04300015e-01
5.32960653e-01 7.76310563e-01 9.59033668e-01 4.74796444e-01
-1.10910587e-01 2.02322349e-01 -4.55721974e-01 1.58725247e-01
4.88789916e-01 5.99438660e-02 -4.06882018e-01 -2.24570513e-01
8.95804986e-02 2.94297710e-02 -1.05815910e-01 1.14736891e+00
3.50804329e-01 3.16918939e-01 6.44141197e-01 -1.51243758e+00
-2.13057026e-01 1.06592521e-01 -6.89889491e-01 1.03360713e-01
7.63741732e-01 2.64318913e-01 7.11420238e-01 -7.32223094e-01
6.38678253e-01 1.10265338e+00 4.84493583e-01 2.95465440e-01
-1.28675032e+00 -4.83899832e-01 1.43408269e-01 4.66490000e-01
-1.64691758e+00 -9.07274783e-02 8.28910887e-01 -4.34314340e-01
9.63775754e-01 2.25724638e-01 5.63842952e-01 8.66010725e-01
2.65835285e-01 5.26621342e-01 1.21562350e+00 -4.40927416e-01
4.71718699e-01 2.13827133e-01 -1.65559545e-01 4.98560399e-01
2.58473307e-01 1.50371239e-01 -3.55291069e-01 -1.51101798e-01
5.67991018e-01 1.85486540e-01 -2.12987304e-01 -8.46157432e-01
-1.29668319e+00 4.60800976e-01 6.10204279e-01 4.71854001e-01
-4.61094111e-01 4.47656363e-02 5.90415858e-02 8.43501389e-02
-4.02863950e-01 2.58182734e-01 -4.08014536e-01 1.03023918e-02
-8.80664051e-01 1.72370911e-01 6.21818721e-01 9.43956792e-01
1.03390050e+00 -2.22171202e-01 1.24887221e-01 8.31033409e-01
6.57478511e-01 7.39292383e-01 4.45441812e-01 -6.44575775e-01
2.21381664e-01 6.74771607e-01 2.44164675e-01 -1.41117251e+00
-5.76923013e-01 1.48634180e-01 -5.20502925e-01 7.78005064e-01
1.81485385e-01 2.20624849e-01 -1.14094901e+00 1.38600993e+00
5.23367345e-01 -1.23024866e-01 1.21407002e-01 1.05725753e+00
6.30831718e-01 7.79535055e-01 -1.45366699e-01 3.08477487e-02
1.41020298e+00 -6.82159305e-01 -4.60879654e-01 -5.20638227e-01
2.40244061e-01 -1.02883828e+00 6.86124325e-01 3.87667835e-01
-2.85443485e-01 -3.61152470e-01 -1.49514461e+00 2.25308672e-01
-6.23540699e-01 3.00699294e-01 3.82403493e-01 4.01950687e-01
-1.16773641e+00 2.38595277e-01 -8.11056614e-01 -8.28502059e-01
-1.22179523e-01 7.29469001e-01 -7.28440583e-01 -2.13003248e-01
-8.92000794e-01 1.17196822e+00 4.33514833e-01 2.31048211e-01
-7.85504043e-01 1.19100241e-02 -8.02411914e-01 -2.55227983e-01
4.40071523e-01 -3.75221819e-01 1.02681553e+00 -7.36201167e-01
-1.54351008e+00 5.48780382e-01 -3.36058229e-01 -4.01267916e-01
5.86353242e-01 1.77865461e-01 -2.47435480e-01 -5.60742356e-02
8.20670649e-02 7.20490873e-01 6.83337510e-01 -1.40852427e+00
-6.21424079e-01 -6.09344721e-01 3.40325087e-01 3.69080126e-01
3.52823228e-01 -1.86491549e-01 -5.71479142e-01 2.15177480e-02
4.72463697e-01 -1.24483287e+00 -4.17331457e-01 1.77016661e-01
3.74324918e-05 -2.21338928e-01 9.84392881e-01 -3.75253588e-01
4.52231348e-01 -2.24942851e+00 -4.10503224e-02 3.46614242e-01
-2.08960637e-01 3.35136175e-01 6.92698732e-02 5.29614091e-01
3.01668316e-01 -2.58255690e-01 -2.53352523e-01 6.57044202e-02
-1.33001566e-01 3.57790023e-01 2.20635030e-02 7.35571861e-01
4.43531089e-02 2.19981864e-01 -9.03858483e-01 -6.19540215e-01
7.67542243e-01 4.84616548e-01 -3.14869225e-01 -1.82547085e-02
-2.51817971e-01 4.08134371e-01 -3.99161875e-01 6.66101396e-01
8.18023384e-01 1.79659903e-01 2.96621770e-01 -1.31171212e-01
-2.66234368e-01 -9.98527482e-02 -1.56748223e+00 1.39915800e+00
-6.54549241e-01 8.04710209e-01 -6.28253296e-02 -9.28857446e-01
1.43088019e+00 9.80853066e-02 2.84545213e-01 -8.16616774e-01
1.42990708e-01 4.75196451e-01 -1.28918573e-01 -3.53574932e-01
7.92179286e-01 5.43209277e-02 -1.16430603e-01 -1.65840670e-01
-1.15059979e-01 1.57600734e-02 -5.15002161e-02 -1.50023088e-01
1.26146591e+00 3.56061369e-01 5.22332013e-01 -6.36874735e-02
6.72673523e-01 2.94287950e-01 7.13673353e-01 7.11548507e-01
-5.34624577e-01 6.32286191e-01 -6.99492777e-03 -3.65883917e-01
-8.42473805e-01 -1.10698950e+00 -5.59572093e-02 6.60056412e-01
9.16218519e-01 3.16231437e-02 -6.32732809e-02 -5.41257441e-01
5.33605255e-02 5.93868613e-01 -3.35258663e-01 1.41666112e-02
-3.61486912e-01 -5.56859851e-01 2.38927484e-01 2.48386025e-01
6.48577034e-01 -9.79359806e-01 -1.12003672e+00 1.51135072e-01
-2.64614105e-01 -1.02947974e+00 1.73545122e-01 2.69857436e-01
-6.98322713e-01 -9.58811522e-01 -6.15116954e-01 -6.87213778e-01
6.70466900e-01 7.78839886e-01 6.70581698e-01 -9.81961470e-03
-2.15629905e-01 4.18192208e-01 -5.36033630e-01 -2.63561726e-01
-2.59880900e-01 1.80025247e-03 1.27981275e-01 -1.07218318e-01
3.56349975e-01 -4.61690009e-01 -4.04864490e-01 6.70701504e-01
-7.41342902e-01 -2.81950802e-01 6.46628022e-01 7.11600542e-01
8.01714182e-01 1.06302544e-01 1.30981222e-01 -3.28033209e-01
8.88625905e-02 -5.34359157e-01 -9.44329679e-01 4.58475620e-01
-3.90675873e-01 1.29492879e-01 4.16985333e-01 -5.77174962e-01
-8.42149794e-01 5.92512012e-01 2.30545327e-02 -5.97472847e-01
-2.80472934e-01 2.93948025e-01 -1.46192774e-01 -3.13603461e-01
8.16553891e-01 2.57395357e-01 -1.53667470e-02 -1.32514209e-01
1.64722607e-01 9.37665403e-01 3.48091930e-01 -2.35700473e-01
8.43381226e-01 4.80741084e-01 1.51433259e-01 -9.53237057e-01
-5.84852174e-02 -8.73654842e-01 -5.43022394e-01 -4.50618148e-01
7.28972316e-01 -7.52726853e-01 -7.27515876e-01 3.32770079e-01
-1.14794219e+00 -3.75471786e-02 2.22724199e-01 5.75900078e-01
-5.52470028e-01 3.23198438e-01 2.50190705e-01 -1.01197636e+00
5.70825450e-02 -1.62284946e+00 9.47512031e-01 5.45552671e-01
1.17372721e-02 -5.88821054e-01 1.41979828e-01 5.94984554e-02
2.86312759e-01 1.34304211e-01 2.96088606e-01 -6.17468417e-01
-8.16835105e-01 -3.83617014e-01 -1.26511723e-01 -1.04609452e-01
1.95540473e-01 2.24353358e-01 -7.28215575e-01 -2.17080802e-01
-2.37457350e-01 3.68937962e-02 4.79860783e-01 6.95892349e-02
5.22543669e-01 1.03597321e-01 -7.45113492e-01 9.79507789e-02
1.82319212e+00 6.34766281e-01 6.68410182e-01 8.43510509e-01
5.05780578e-01 4.91774619e-01 1.22219789e+00 3.32232684e-01
5.88860810e-01 1.13510287e+00 8.72694790e-01 1.34604141e-01
8.74624327e-02 9.42432147e-04 4.06856090e-01 1.63653523e-01
-5.07374257e-02 -8.37759823e-02 -1.22975159e+00 6.93202198e-01
-1.99145746e+00 -8.50778401e-01 -6.60742298e-02 2.56623983e+00
1.60616994e-01 -5.22996373e-02 -1.55473333e-02 1.51244044e-01
9.84053612e-01 6.69985861e-02 -4.89044577e-01 -4.03254002e-01
3.52665223e-02 -2.33430818e-01 8.85784447e-01 3.27191830e-01
-1.03509128e+00 7.83337474e-01 5.50553179e+00 4.99580115e-01
-1.45186830e+00 -1.17343791e-01 -4.83622216e-02 2.92576879e-01
-8.18797275e-02 2.12132961e-01 -7.76884973e-01 2.75823802e-01
5.54626763e-01 -2.65194736e-02 6.01007402e-01 1.15683568e+00
-1.06272846e-01 -7.37511754e-01 -9.61579859e-01 1.38721240e+00
2.89392591e-01 -8.92430604e-01 -2.82055050e-01 6.38068765e-02
3.33914489e-01 3.85859579e-01 -6.38191700e-02 2.17806533e-01
4.83883142e-01 -6.22643113e-01 8.79841864e-01 3.08494657e-01
1.82250232e-01 -5.31338274e-01 8.73221457e-01 4.61064845e-01
-1.23306584e+00 -1.57483354e-01 -3.72981071e-01 8.86337832e-02
2.93864142e-02 3.51250112e-01 -1.26585805e+00 7.47704744e-01
1.02914059e+00 2.16051280e-01 -5.96294463e-01 1.45083761e+00
-2.15298727e-01 -1.73308104e-01 -8.29864144e-01 -3.97936583e-01
6.02602512e-02 4.52926382e-02 8.46003532e-01 8.13881338e-01
5.65002084e-01 -1.48903266e-01 2.67796963e-01 4.39571470e-01
4.31538582e-01 1.23844557e-01 -1.05665612e+00 4.82923329e-01
6.80753291e-01 1.33775938e+00 -1.18507600e+00 -6.56325519e-02
-1.27790913e-01 9.03443694e-01 1.75697029e-01 1.17893055e-01
-9.14675951e-01 -3.89476657e-01 4.77545083e-01 5.81264943e-02
5.76406598e-01 -5.61575353e-01 -9.07533243e-02 -9.32668269e-01
2.55847633e-01 -5.63989401e-01 1.89983636e-01 -1.03345466e+00
-8.55985343e-01 7.74986923e-01 5.69231249e-02 -1.59482455e+00
-2.49667183e-01 -6.65197790e-01 -1.32399321e-01 6.28214180e-01
-1.54291666e+00 -9.25413430e-01 -3.71125400e-01 5.77579916e-01
4.47573721e-01 1.27699628e-01 6.49368703e-01 1.79898515e-01
-2.04043880e-01 3.43328863e-01 1.07974224e-01 -2.52021194e-01
7.34682918e-01 -8.85504603e-01 -1.36578411e-01 9.15432334e-01
2.74433553e-01 4.89694208e-01 1.08895528e+00 -6.20031953e-01
-1.57503498e+00 -8.27343166e-01 4.77482259e-01 -1.69001132e-01
3.95913839e-01 -1.54010311e-01 -6.72248602e-01 5.81573725e-01
-4.66919988e-02 1.94503322e-01 3.09468657e-01 -2.35345550e-02
-2.92727351e-01 -2.83257753e-01 -1.54354596e+00 7.47191489e-01
7.17980206e-01 -3.78677368e-01 -7.49038458e-01 1.09917194e-01
2.76751578e-01 -5.83134234e-01 -5.27781308e-01 7.23373771e-01
4.53806132e-01 -9.64345157e-01 8.74788582e-01 3.55562478e-01
-1.97639510e-01 -1.17594731e+00 -5.92818856e-01 -1.10927176e+00
-7.20135421e-02 -6.03735223e-02 3.81181538e-01 1.23967576e+00
4.29318458e-01 -9.41935301e-01 4.61805254e-01 5.18097639e-01
9.66728032e-02 -3.65470409e-01 -1.21961725e+00 -1.07320309e+00
-6.67053998e-01 -4.71614420e-01 5.14348328e-01 6.41269565e-01
-1.46095842e-01 -8.43350068e-02 -2.56921854e-02 8.72946262e-01
5.34341812e-01 1.71695396e-01 9.56322074e-01 -1.18299985e+00
-2.42590815e-01 -8.46402198e-02 -1.16367126e+00 -6.65166795e-01
-1.03111662e-01 -6.36687458e-01 6.91876233e-01 -1.85771966e+00
-6.41169176e-02 -8.80306184e-01 -2.23919272e-01 5.27117908e-01
2.99409688e-01 2.73149133e-01 4.37993795e-01 4.74160552e-01
-5.41006505e-01 3.41887087e-01 6.41408086e-01 -3.92634481e-01
-4.30164844e-01 -3.26921716e-02 8.54035933e-03 8.21867645e-01
7.91791797e-01 -5.05042434e-01 -4.33084995e-01 -3.57144117e-01
1.70720533e-01 1.12756848e-01 4.34219301e-01 -1.56322396e+00
5.49159765e-01 -2.43102506e-01 2.79902756e-01 -7.37716794e-01
4.89469111e-01 -1.34119153e+00 6.37323856e-01 4.21793997e-01
3.63710523e-01 3.14202935e-01 3.26123863e-01 7.03330100e-01
-1.63503960e-01 -5.14818311e-01 7.00858772e-01 -1.66283891e-01
-1.31094766e+00 -1.31473586e-01 -6.46928549e-01 -5.99540472e-01
1.69667554e+00 -7.01035917e-01 -7.16526881e-02 7.74560273e-02
-5.79818308e-01 1.96874470e-01 9.80491042e-01 7.07799673e-01
7.12316275e-01 -1.16287720e+00 5.01441816e-03 -3.12589742e-02
4.58098233e-01 1.64261162e-01 3.91151160e-01 6.22563243e-01
-8.01831782e-01 2.02825099e-01 -4.87148225e-01 -1.02378881e+00
-1.47816288e+00 7.59633601e-01 2.64144689e-01 -5.10208271e-02
-4.86179352e-01 4.29051101e-01 -8.34705457e-02 -6.79479241e-01
8.36447105e-02 -1.09109938e-01 -4.69566613e-01 -2.46154051e-03
3.75107348e-01 1.37163267e-01 2.13355944e-01 -9.79254007e-01
-8.64659131e-01 8.34151149e-01 6.51890412e-02 -3.99273962e-01
1.03971469e+00 -2.57355809e-01 3.56257632e-02 6.62277997e-01
7.56224990e-01 -1.00171581e-01 -1.00031507e+00 5.69209382e-02
3.70912924e-02 -5.29442430e-01 -3.70017961e-02 -6.24269545e-01
-8.01162183e-01 6.44104540e-01 1.00803697e+00 -3.65598612e-02
9.88288105e-01 -3.95525023e-02 2.27382198e-01 5.34472883e-01
1.32939327e+00 -9.73664403e-01 -2.52752095e-01 4.11670059e-01
7.13717759e-01 -1.32041788e+00 -3.16340514e-02 -3.98617387e-01
-5.56110561e-01 1.03626180e+00 5.96045613e-01 -2.03660026e-01
5.83404243e-01 1.26895860e-01 8.72282982e-02 2.45350003e-02
-4.05387372e-01 -3.48878920e-01 -6.34264797e-02 1.02396810e+00
-2.17361107e-01 2.16807365e-01 -2.72496879e-01 -1.63974002e-01
3.04944627e-02 3.65665592e-02 6.46966696e-01 1.22004473e+00
-4.60531563e-01 -1.08599830e+00 -7.35746920e-01 1.17524475e-01
-2.01682106e-01 3.52908552e-01 -6.90952018e-02 8.14854264e-01
2.05364347e-01 9.91431236e-01 -6.52667359e-02 -6.64681375e-01
4.93498176e-01 -3.65559667e-01 4.61105436e-01 -4.26078171e-01
-2.76420802e-01 -4.25248444e-01 -8.63759518e-02 -5.58226466e-01
-5.98413706e-01 -9.38613355e-01 -1.48573232e+00 -3.80872525e-02
-4.09107298e-01 -1.72662124e-01 1.19075668e+00 8.53287399e-01
4.36975390e-01 2.45626122e-01 6.31157815e-01 -1.20436442e+00
-2.94707268e-01 -4.86784607e-01 -2.57471502e-01 -1.18914492e-01
2.01613858e-01 -1.33284116e+00 -3.07653576e-01 -2.85482556e-01] | [7.459722995758057, -1.91754949092865] |
7fb71003-6586-414d-98df-c5663623f75d | mplug-docowl-modularized-multimodal-large | 2307.02499 | null | https://arxiv.org/abs/2307.02499v1 | https://arxiv.org/pdf/2307.02499v1.pdf | mPLUG-DocOwl: Modularized Multimodal Large Language Model for Document Understanding | Document understanding refers to automatically extract, analyze and comprehend information from various types of digital documents, such as a web page. Existing Multi-model Large Language Models (MLLMs), including mPLUG-Owl, have demonstrated promising zero-shot capabilities in shallow OCR-free text recognition, indicating their potential for OCR-free document understanding. Nevertheless, without in-domain training, these models tend to ignore fine-grained OCR features, such as sophisticated tables or large blocks of text, which are essential for OCR-free document understanding. In this paper, we propose mPLUG-DocOwl based on mPLUG-Owl for OCR-free document understanding. Specifically, we first construct a instruction tuning dataset featuring a wide range of visual-text understanding tasks. Then, we strengthen the OCR-free document understanding ability by jointly train the model on language-only, general vision-and-language, and document instruction tuning dataset with our unified instruction tuning strategy. We also build an OCR-free document instruction understanding evaluation set LLMDoc to better compare models' capabilities on instruct compliance and document understanding. Experimental results show that our model outperforms existing multi-modal models, demonstrating its strong ability of document understanding. Besides, without specific fine-tuning, mPLUG-DocOwl generalizes well on various downstream tasks. Our code, models, training data and evaluation set are available at https://github.com/X-PLUG/mPLUG-DocOwl. | ['Fei Huang', 'Ji Zhang', 'Qian Qi', 'Junfeng Tian', 'Chenliang Li', 'Guohai Xu', 'Chenlin Zhao', 'Yuhao Dan', 'Ming Yan', 'Qinghao Ye', 'Haiyang Xu', 'Anwen Hu', 'Jiabo Ye'] | 2023-07-04 | null | null | null | null | ['optical-character-recognition'] | ['computer-vision'] | [ 1.65450305e-01 -3.46559733e-01 -4.29147691e-01 -4.36954379e-01
-8.66242051e-01 -6.92311287e-01 5.65450609e-01 2.60646850e-01
-2.86331266e-01 2.22173110e-01 1.70690179e-01 -8.15343380e-01
8.21822602e-03 -5.25705278e-01 -8.58545184e-01 -1.97074801e-01
6.17866993e-01 3.56742531e-01 2.31741846e-01 -3.05832893e-01
7.17860818e-01 1.23169258e-01 -1.56623220e+00 7.61427641e-01
1.29945421e+00 7.42150784e-01 7.39995718e-01 1.07339990e+00
-5.92601895e-01 8.83223295e-01 -6.80699885e-01 -9.75800976e-02
-4.07731712e-01 -1.23759314e-01 -6.64674520e-01 -3.50676775e-02
8.93623292e-01 -6.54781640e-01 -4.89865750e-01 8.44857275e-01
2.23870903e-01 8.58449202e-04 7.78179526e-01 -1.01256061e+00
-1.38441122e+00 5.26162148e-01 -6.78830564e-01 9.57287773e-02
4.47693467e-01 3.14948916e-01 8.31312776e-01 -1.04701042e+00
3.72799307e-01 1.46890175e+00 2.15376914e-01 6.02154613e-01
-7.96129942e-01 -7.61128843e-01 3.28123689e-01 3.05822134e-01
-1.15297365e+00 -2.33409584e-01 4.34717655e-01 -4.07497019e-01
1.29626513e+00 2.29668126e-01 1.68700054e-01 1.09554827e+00
6.23521686e-01 1.54328227e+00 1.07103956e+00 -7.96664119e-01
-1.76610306e-01 7.44864717e-02 8.97548974e-01 1.05630171e+00
2.99141020e-01 -1.35913849e-01 -7.82275558e-01 5.55151165e-01
4.47792023e-01 2.34619334e-01 -4.23007935e-01 -6.51529729e-02
-9.94918764e-01 5.80185294e-01 7.66278803e-02 2.72638232e-01
5.14673829e-01 1.89866811e-01 3.45779270e-01 3.15072477e-01
3.11207086e-01 1.83967516e-01 -3.85877252e-01 -3.05817872e-01
-9.02939558e-01 -3.66650015e-01 6.13948762e-01 1.44025612e+00
7.76711762e-01 1.15738034e-01 -3.21774960e-01 9.44938064e-01
4.85068768e-01 7.28417218e-01 1.02147055e+00 -5.26482940e-01
7.31526375e-01 7.74998069e-01 -4.50743765e-01 -5.22123098e-01
-2.65396655e-01 -2.39836380e-01 -5.86211503e-01 -8.44087899e-02
2.09067762e-01 3.47793311e-01 -1.28046608e+00 1.28285110e+00
-2.53613681e-01 -2.81504422e-01 2.31242493e-01 5.66199124e-01
1.12209880e+00 1.00520074e+00 1.81673076e-02 2.22595379e-01
1.64072108e+00 -1.46065068e+00 -9.09367442e-01 -7.19338238e-01
1.26813030e+00 -8.48225057e-01 1.79892170e+00 5.30182362e-01
-5.98471105e-01 -9.51499164e-01 -1.38104773e+00 -6.84616566e-01
-9.89276052e-01 4.95769590e-01 5.85109472e-01 6.91272736e-01
-8.96625638e-01 1.45869955e-01 -6.67529285e-01 -3.86826843e-01
5.38248241e-01 1.65835209e-02 -1.12101875e-01 -7.64859617e-01
-8.19704652e-01 5.28013706e-01 6.34597540e-01 -2.73735404e-01
-1.09380043e+00 -8.47088933e-01 -1.13338602e+00 2.68954188e-01
5.58454216e-01 -2.20962822e-01 1.31620669e+00 -4.26738828e-01
-1.34345531e+00 9.18731034e-01 -3.28441113e-01 -6.13440201e-03
1.78884149e-01 -6.85203016e-01 -5.09490967e-01 1.12046354e-01
-1.45710751e-01 6.64306283e-01 7.31406629e-01 -1.35320365e+00
-3.33972126e-01 -4.50845867e-01 9.22513530e-02 1.00326605e-01
-6.55235529e-01 -2.62075454e-01 -1.06043291e+00 -5.46462119e-01
-2.13433310e-01 -5.53259075e-01 4.53830153e-01 -2.08030432e-01
-6.69689834e-01 -3.33551645e-01 1.10856545e+00 -7.13246167e-01
1.54677653e+00 -2.20085907e+00 -2.77102262e-01 -2.77215719e-01
2.27772519e-01 3.39367956e-01 -7.30732739e-01 4.12070900e-01
1.04176223e-01 4.09747601e-01 1.30186066e-01 -4.20636386e-01
3.05394053e-01 6.29076585e-02 -4.74361360e-01 -5.84571175e-02
7.41084516e-02 1.20563638e+00 -7.43885517e-01 -6.38887167e-01
3.68483901e-01 1.87518284e-01 -4.93211836e-01 1.50041834e-01
-6.21412873e-01 -1.69988766e-01 -6.22560322e-01 9.68357146e-01
4.78676200e-01 -5.23768663e-01 -8.96970257e-02 -6.44943863e-02
-6.34922907e-02 -2.63855588e-02 -7.05477059e-01 2.02173281e+00
-9.09765720e-01 1.08394170e+00 -3.43896329e-01 -5.98172903e-01
8.72652113e-01 -7.67681301e-02 -2.51354396e-01 -1.05604327e+00
3.16727422e-02 2.05526397e-01 -1.45532221e-01 -4.43461210e-01
8.49412620e-01 3.96419525e-01 -1.44180954e-01 6.80565417e-01
1.54646561e-01 -3.53158385e-01 4.89762604e-01 6.74135447e-01
8.99311364e-01 2.04279140e-01 1.10731483e-01 -2.05818728e-01
5.69906056e-01 2.10155919e-03 -2.66992688e-01 9.13614929e-01
-7.20418096e-02 4.82698858e-01 2.87584424e-01 4.31958288e-02
-6.18870139e-01 -9.12416279e-01 -1.53117076e-01 1.57556927e+00
3.26222450e-01 -7.37266541e-01 -8.25025678e-01 -6.42937720e-01
-9.27210227e-02 1.08153689e+00 -3.79287928e-01 -3.93931776e-01
-4.10027593e-01 -5.19361019e-01 4.50954080e-01 7.21214056e-01
5.86248279e-01 -8.36823881e-01 -2.02873975e-01 -5.78630902e-02
-9.45158973e-02 -1.24412131e+00 -1.02627635e+00 2.63664216e-01
-9.06076372e-01 -1.13326335e+00 -5.37166119e-01 -8.88465881e-01
7.48219490e-01 6.23464704e-01 1.33941996e+00 5.29044867e-01
-5.37635624e-01 9.16259468e-01 -4.36070412e-01 -5.30600667e-01
-5.80714941e-01 6.43322319e-02 -1.63319051e-01 -4.60129112e-01
8.83403957e-01 1.77333891e-01 -3.11530501e-01 2.87943661e-01
-1.27076471e+00 3.56146008e-01 8.22269320e-01 7.75711596e-01
5.79041958e-01 -1.69805080e-01 3.51024330e-01 -8.40227187e-01
8.59165907e-01 -1.16564676e-01 -5.59477091e-01 8.05285692e-01
-7.63271451e-01 3.85969907e-01 7.58310556e-01 -5.30467808e-01
-1.10341430e+00 -3.73229921e-01 3.16686258e-02 -3.44756752e-01
-2.82483876e-01 5.04427791e-01 -3.55470479e-01 7.89683387e-02
4.00835395e-01 6.51893973e-01 -3.44330668e-01 -5.83869100e-01
4.58316535e-01 1.01862121e+00 4.63035166e-01 -8.39727700e-01
7.34920859e-01 1.59054026e-01 -6.45004809e-01 -1.08070052e+00
-1.09531438e+00 -6.20526671e-01 -7.22975850e-01 -1.58533633e-01
9.89807725e-01 -1.01993859e+00 -7.17828572e-01 6.89770699e-01
-1.03045654e+00 -5.84585965e-01 1.90993220e-01 3.71972442e-01
-4.48620081e-01 5.37580192e-01 -8.75710726e-01 -5.65764606e-01
-3.16695780e-01 -1.45068204e+00 1.32064426e+00 5.65234065e-01
-1.37177870e-01 -1.15337348e+00 -1.68203026e-01 8.69949758e-01
1.89078465e-01 -4.21143442e-01 1.42710972e+00 -8.42895329e-01
-8.18971872e-01 -1.51712954e-01 -5.85274816e-01 3.18123132e-01
1.17547683e-01 2.86659244e-02 -1.03263283e+00 -4.41399217e-01
-4.40364569e-01 -8.82647634e-01 1.07870722e+00 2.08498910e-01
1.59967792e+00 8.36689025e-03 -4.56244469e-01 7.00010955e-01
1.51041090e+00 3.59231174e-01 5.78122199e-01 3.91900480e-01
1.06832528e+00 2.93336958e-01 7.47706890e-01 1.22504838e-01
4.91252989e-01 3.57208192e-01 2.46660650e-01 8.61916468e-02
-5.46248913e-01 -5.49463034e-01 6.77992344e-01 1.28134263e+00
3.94327551e-01 -6.36615276e-01 -1.03761804e+00 3.54714662e-01
-1.38352537e+00 -2.11724222e-01 -4.97307219e-02 1.80318928e+00
1.02195561e+00 2.60180950e-01 -6.57359004e-01 -1.74189940e-01
3.29110801e-01 4.11346644e-01 -7.73428440e-01 -7.07096756e-01
-4.20483761e-02 3.77579145e-02 2.55185038e-01 7.17263103e-01
-7.64743805e-01 1.37822270e+00 5.23387527e+00 1.38758218e+00
-9.00153339e-01 -8.50045979e-02 5.48025191e-01 -6.31465092e-02
-4.55130637e-01 -2.73174167e-01 -1.44709027e+00 2.66951352e-01
1.07019186e+00 -1.85610905e-01 1.79163948e-01 7.99658597e-01
-2.37013353e-03 -1.73661932e-01 -1.34175658e+00 9.19885814e-01
4.83341187e-01 -1.24701369e+00 5.73279679e-01 -1.25745714e-01
6.33599877e-01 -1.33116737e-01 1.66238844e-01 8.43416810e-01
4.18439984e-01 -1.21378052e+00 5.80995619e-01 2.47047395e-01
1.10958302e+00 -4.85229641e-01 3.43739003e-01 4.34182972e-01
-1.20164037e+00 1.00564793e-01 -3.85657489e-01 3.35932732e-01
-2.24809647e-01 2.15816334e-01 -4.22185540e-01 4.14234221e-01
5.24952888e-01 9.29087400e-01 -1.06682634e+00 5.93824208e-01
-3.59254271e-01 4.70116556e-01 4.21830118e-02 -3.25709403e-01
3.12771231e-01 1.09249659e-01 6.07433878e-02 1.52296138e+00
3.41183215e-01 3.38351950e-02 2.10964486e-01 8.04075122e-01
-2.45398000e-01 8.42918381e-02 -4.09369379e-01 -6.92701340e-01
3.35157424e-01 9.77184892e-01 -5.93767047e-01 -5.54532707e-01
-7.86546111e-01 1.05812192e+00 2.37882808e-01 6.45548880e-01
-7.21088231e-01 -5.69258273e-01 6.31840229e-01 -2.09543094e-01
1.65054947e-01 -4.16821033e-01 -2.42429152e-01 -1.44679403e+00
-6.70699254e-02 -1.24126637e+00 2.69979149e-01 -1.24007833e+00
-7.19734848e-01 3.46060902e-01 -6.00163043e-02 -1.11294734e+00
-1.46692932e-01 -1.40362394e+00 -5.65170944e-01 5.96403658e-01
-1.76034832e+00 -1.22806120e+00 -4.90451425e-01 5.42191565e-01
1.47453821e+00 -2.21926540e-01 6.76365852e-01 -4.13735695e-02
-8.06579173e-01 8.88857484e-01 5.49446106e-01 1.85054362e-01
9.81513739e-01 -1.28064799e+00 2.25746125e-01 8.41639459e-01
3.29840332e-01 8.72195125e-01 4.97526318e-01 -6.20230019e-01
-1.89186978e+00 -8.97515357e-01 3.35859299e-01 -8.01365077e-01
8.78844917e-01 -6.27077341e-01 -1.10682583e+00 9.12075281e-01
3.03288430e-01 -4.63299036e-01 9.23791587e-01 5.31597696e-02
-8.07781100e-01 1.44032598e-01 -4.27975982e-01 7.50743687e-01
7.38210201e-01 -8.48824322e-01 -7.63154089e-01 5.67050874e-01
1.04143929e+00 -4.43273634e-01 -7.24809706e-01 5.60607649e-02
5.59313357e-01 -6.90381646e-01 1.02186894e+00 -6.13852739e-01
9.31285739e-01 -9.23757777e-02 -2.54750729e-01 -1.25698245e+00
1.09449595e-01 -1.11181274e-01 -4.23355937e-01 1.16392827e+00
3.86056095e-01 -3.49869937e-01 3.78981918e-01 1.90054461e-01
-6.20699227e-01 -6.71758175e-01 -2.95875698e-01 -9.38315332e-01
3.91949862e-01 -7.74005830e-01 2.98135519e-01 7.24668264e-01
-6.89863460e-03 3.64136934e-01 -1.85305215e-02 1.18225761e-01
5.42548895e-01 4.19612914e-01 8.04681599e-01 -9.78973091e-01
-3.43628556e-01 -5.58263779e-01 2.05375895e-01 -1.76188052e+00
2.94544667e-01 -8.75962615e-01 1.29588395e-01 -1.65004492e+00
6.46488130e-01 5.39108403e-02 -2.31457278e-01 7.34448850e-01
-3.54532838e-01 -4.77151535e-02 3.86241257e-01 1.68352515e-01
-9.86302674e-01 4.80572790e-01 1.70253289e+00 -6.88983262e-01
1.92616396e-02 -5.35224676e-01 -6.63224995e-01 8.17198455e-01
4.79056984e-01 -1.38171881e-01 -5.99509895e-01 -8.31013024e-01
-1.86217710e-01 -1.71756148e-01 -8.72232206e-03 -9.21316206e-01
2.78059363e-01 -1.76896200e-01 5.44341385e-01 -9.12996888e-01
1.90467089e-01 -5.24327040e-01 -8.12535942e-01 5.58574498e-01
-4.93420035e-01 -2.10260972e-01 8.51716340e-01 5.50409198e-01
-3.63299400e-01 -6.61748111e-01 4.33605313e-01 1.30452355e-02
-1.42233109e+00 9.40845758e-02 -4.36220825e-01 1.25460878e-01
7.42184639e-01 -3.68583441e-01 -1.08741724e+00 -3.47620070e-01
-2.72307038e-01 5.39743602e-01 5.54493964e-01 9.98642325e-01
9.10308361e-01 -9.30636525e-01 -4.43899632e-01 3.16935092e-01
7.59024024e-01 -1.21167392e-01 4.76350337e-01 3.32899749e-01
-6.00471973e-01 1.17183220e+00 -2.95125879e-02 -6.22933984e-01
-1.47155917e+00 6.72281086e-01 2.01970749e-02 -5.90022564e-01
-4.67588335e-01 7.41865575e-01 8.60142887e-01 -5.55870891e-01
3.45117658e-01 -7.36359894e-01 -1.44749939e-01 -1.53263345e-01
1.01025307e+00 -2.61278935e-02 -7.01352656e-02 -8.71720538e-02
-1.27695724e-01 8.96860778e-01 -5.96359670e-01 3.34139228e-01
8.86394382e-01 -3.06402922e-01 2.32990175e-01 5.81429124e-01
1.34950221e+00 -1.38055179e-02 -1.14628863e+00 -1.44477919e-01
1.19330339e-01 -2.20788524e-01 1.77758366e-01 -1.05103099e+00
-7.90264964e-01 1.42348599e+00 6.18343949e-01 -2.56979555e-01
1.16080844e+00 -2.02358328e-02 7.43602931e-01 8.10368180e-01
2.20285997e-01 -1.12524056e+00 6.30164564e-01 9.32187736e-01
7.92194963e-01 -1.75004888e+00 9.51546244e-03 -3.08734417e-01
-2.42358610e-01 1.39310765e+00 1.10004771e+00 5.37105381e-01
2.02948719e-01 4.46298748e-01 1.45855621e-01 -1.87973514e-01
-8.03818464e-01 -1.80076525e-01 6.56417489e-01 4.57706511e-01
8.00664008e-01 -1.92072004e-01 1.51595443e-01 7.03639030e-01
-1.73771232e-01 9.38256970e-04 6.19062364e-01 9.80101943e-01
-7.21072257e-01 -8.80126357e-01 -3.96623731e-01 4.92398649e-01
-8.57258439e-02 -6.22173011e-01 -3.28805119e-01 1.05157661e+00
-5.59811294e-01 9.54739511e-01 1.32642403e-01 -2.89593577e-01
1.55922577e-01 4.15504575e-01 3.83425325e-01 -9.08819973e-01
-1.99089825e-01 -6.09189942e-02 -1.79709211e-01 -5.72684169e-01
1.79612532e-01 6.60213754e-02 -1.51264751e+00 -1.77940667e-01
-3.38473350e-01 -2.07181513e-01 6.60924971e-01 1.18115151e+00
1.86820358e-01 1.07490957e+00 -7.18706846e-02 -2.95711339e-01
-3.56938094e-01 -1.15770423e+00 -4.27954108e-01 1.54528126e-01
3.31609875e-01 -2.77620196e-01 -3.37191910e-01 2.45090067e-01] | [11.445191383361816, 2.158355712890625] |
f4c9c4b4-517a-4c19-830d-23c687f63606 | bubblerank-safe-online-learning-to-rerank | 1806.05819 | null | https://arxiv.org/abs/1806.05819v2 | https://arxiv.org/pdf/1806.05819v2.pdf | BubbleRank: Safe Online Learning to Re-Rank via Implicit Click Feedback | In this paper, we study the problem of safe online learning to re-rank, where user feedback is used to improve the quality of displayed lists. Learning to rank has traditionally been studied in two settings. In the offline setting, rankers are typically learned from relevance labels created by judges. This approach has generally become standard in industrial applications of ranking, such as search. However, this approach lacks exploration and thus is limited by the information content of the offline training data. In the online setting, an algorithm can experiment with lists and learn from feedback on them in a sequential fashion. Bandit algorithms are well-suited for this setting but they tend to learn user preferences from scratch, which results in a high initial cost of exploration. This poses an additional challenge of safe exploration in ranked lists. We propose BubbleRank, a bandit algorithm for safe re-ranking that combines the strengths of both the offline and online settings. The algorithm starts with an initial base list and improves it online by gradually exchanging higher-ranked less attractive items for lower-ranked more attractive items. We prove an upper bound on the n-step regret of BubbleRank that degrades gracefully with the quality of the initial base list. Our theoretical findings are supported by extensive experiments on a large-scale real-world click dataset. | ['Maarten de Rijke', 'Csaba Szepesvari', 'Branislav Kveton', 'Masrour Zoghi', 'Ilya Markov', 'Tor Lattimore', 'Chang Li'] | 2018-06-15 | null | null | null | null | ['safe-exploration'] | ['robots'] | [ 1.11381337e-01 -1.93648580e-02 -7.70577371e-01 -3.76174331e-01
-1.24382043e+00 -9.97730970e-01 9.15901661e-02 4.25781786e-01
-6.10851228e-01 9.22798574e-01 1.03644200e-01 -4.72357512e-01
-6.99203432e-01 -6.27303839e-01 -9.82618213e-01 -5.56269407e-01
-3.58867735e-01 8.96756470e-01 1.70130551e-01 -1.91821098e-01
3.10502678e-01 9.62592810e-02 -1.55135179e+00 3.12440097e-01
1.18392575e+00 1.23289812e+00 1.73610255e-01 5.69126785e-01
-3.15788612e-02 5.13658464e-01 -2.01440364e-01 -6.70762479e-01
8.23460460e-01 -3.77843142e-01 -7.99668014e-01 -1.58714816e-01
4.26980972e-01 -5.91308296e-01 -1.90481022e-01 1.15559506e+00
2.83896387e-01 4.49617416e-01 5.60442917e-02 -1.03945684e+00
-3.18104714e-01 1.21155226e+00 -5.25602818e-01 1.99600160e-01
2.69102067e-01 -3.56273770e-01 1.83002019e+00 -5.18045127e-01
4.47510481e-01 1.09434628e+00 1.82334527e-01 3.79175931e-01
-1.36258245e+00 -5.63103974e-01 7.04476058e-01 2.08198518e-01
-8.22789431e-01 -1.01652130e-01 5.96929014e-01 -2.55037010e-01
2.40050610e-02 6.46822274e-01 8.85601997e-01 6.90979958e-01
-2.57504582e-01 1.14732730e+00 1.27387023e+00 -3.39767814e-01
5.00828147e-01 2.45773107e-01 3.52815837e-01 2.57919848e-01
4.62536246e-01 3.22105020e-01 -1.03424132e+00 -4.91571486e-01
5.21117151e-01 2.01179251e-01 -3.47314656e-01 -1.03108263e+00
-8.87300134e-01 9.64612424e-01 7.58740604e-01 -2.38522559e-01
-3.39515477e-01 1.80253282e-01 2.87332535e-01 7.00901330e-01
4.07884389e-01 6.05613708e-01 -5.60270488e-01 -3.21142733e-01
-8.81518722e-01 2.00994849e-01 9.01487589e-01 7.62495518e-01
6.91744685e-01 -7.51905978e-01 -4.02286917e-01 9.09316182e-01
1.93429459e-02 4.25350308e-01 2.96036988e-01 -9.19298649e-01
6.72142267e-01 2.61173666e-01 8.31340909e-01 -5.68473637e-01
3.07485126e-02 -6.37730718e-01 -3.31099749e-01 1.77727655e-01
5.42896330e-01 1.71990190e-02 -8.18369210e-01 1.76069236e+00
3.33020180e-01 -2.72924483e-01 -4.99796420e-01 1.01956820e+00
-1.04348339e-01 4.31311727e-01 -3.93409938e-01 -4.86512750e-01
7.20679104e-01 -1.21233451e+00 -4.64328706e-01 -2.26904541e-01
4.12743449e-01 -3.39236885e-01 1.29952490e+00 9.28918362e-01
-1.21160269e+00 -8.79044682e-02 -8.49942625e-01 2.05379799e-01
-1.40162095e-01 -2.28623599e-01 8.04962993e-01 6.70574427e-01
-8.82379949e-01 8.67669582e-01 -5.99966407e-01 -1.93519872e-02
3.02970499e-01 6.03380978e-01 1.38741076e-01 -1.06960282e-01
-1.16721702e+00 4.95142430e-01 2.72984713e-01 1.43093482e-01
-7.98064291e-01 -6.22704208e-01 -1.48441449e-01 2.16027826e-01
1.08654928e+00 -3.29181284e-01 1.75404477e+00 -1.14375162e+00
-1.53846824e+00 3.44268560e-01 -5.20976149e-02 -4.36445802e-01
1.03638184e+00 -7.45408416e-01 1.70784488e-01 -2.99131185e-01
6.89596385e-02 6.99041188e-02 5.70316076e-01 -1.25752926e+00
-1.01048076e+00 -4.34261441e-01 5.63460827e-01 4.24345225e-01
-5.62657595e-01 -3.58913451e-01 -5.96990943e-01 -2.24662259e-01
7.50226080e-02 -1.12966299e+00 -3.37697059e-01 -8.31882283e-02
-2.16240361e-01 -9.98910740e-02 1.11251399e-02 -2.66099006e-01
1.40429831e+00 -1.87371445e+00 8.88361770e-04 8.65817606e-01
3.91965210e-02 -2.55858302e-02 -8.96625221e-02 4.85933959e-01
2.52674967e-01 3.09715897e-01 3.56551826e-01 -2.96487957e-02
1.24445483e-01 -2.61235312e-02 -6.05066717e-01 2.18182772e-01
-7.43849039e-01 7.53027678e-01 -1.42017162e+00 -1.34097189e-01
-2.44058341e-01 -3.78929138e-01 -8.07533443e-01 1.36861101e-01
-3.26417536e-01 1.95710018e-01 -5.01573980e-01 3.36703360e-01
5.39363921e-01 -5.28965831e-01 4.11831141e-01 2.41906643e-01
-4.85813916e-02 4.12805259e-01 -1.35307038e+00 1.29692578e+00
-6.08725429e-01 2.68309325e-01 2.84806222e-01 -6.77239180e-01
3.81885350e-01 -2.11410243e-02 3.14931691e-01 -7.53974736e-01
-1.35330111e-01 5.14169514e-01 -1.18288107e-01 -1.12721294e-01
5.71954370e-01 7.20002353e-02 -5.82683235e-02 6.96616530e-01
-5.74482024e-01 1.84251785e-01 4.77329582e-01 4.27147239e-01
9.83066142e-01 -2.50065714e-01 1.45861343e-01 -7.77052790e-02
4.00010608e-02 -1.08267784e-01 4.94582832e-01 1.52498698e+00
1.01536222e-01 2.92665184e-01 6.20719910e-01 -3.18485349e-01
-8.14693451e-01 -1.07913411e+00 2.72458326e-02 1.78570533e+00
4.72632349e-01 -3.66427958e-01 -3.23137045e-01 -9.63295937e-01
4.99771655e-01 3.84290427e-01 -7.11832762e-01 -3.43863033e-02
-2.65660793e-01 -4.00683671e-01 -3.82899761e-01 3.95173043e-01
5.99830747e-02 -6.50538445e-01 -5.66400826e-01 2.03921229e-01
-1.01982303e-01 -3.65385979e-01 -8.14998150e-01 5.37313402e-01
-1.16048265e+00 -9.92220938e-01 -5.96132815e-01 -3.98903400e-01
8.37956548e-01 4.64064449e-01 1.08698857e+00 6.31588027e-02
2.85069913e-01 2.13111117e-01 -4.77443308e-01 -3.13085675e-01
1.73943788e-01 2.92430252e-01 5.13451099e-02 1.71859473e-01
-1.39394194e-01 -2.82380015e-01 -9.40099716e-01 6.19019449e-01
-8.33306730e-01 -1.30319357e-01 8.16008806e-01 1.20097220e+00
6.01983190e-01 4.22380455e-02 5.05913317e-01 -1.34904838e+00
8.54111493e-01 -4.47478384e-01 -1.10812533e+00 4.39044327e-01
-9.35877860e-01 4.24653202e-01 6.98042810e-01 -6.45719111e-01
-7.71649182e-01 -2.09019575e-02 4.82926250e-01 -1.34767190e-01
7.94434309e-01 8.39890957e-01 3.36437561e-02 -2.15364397e-02
4.75219429e-01 -1.19467013e-01 -3.17079455e-01 -7.03847468e-01
4.90162998e-01 7.06731677e-01 3.50228012e-01 -6.91383421e-01
8.29249978e-01 5.00955462e-01 -1.50599703e-01 -6.06176965e-02
-1.36707568e+00 -7.90699899e-01 -2.48849019e-01 -2.15550676e-01
-5.86916953e-02 -6.41735554e-01 -9.64945972e-01 -1.23747205e-02
-5.16235292e-01 -5.34110069e-01 -4.06134605e-01 3.84580642e-01
-4.83300060e-01 1.34152040e-01 -2.79243559e-01 -1.09161997e+00
-2.30388641e-01 -7.51746416e-01 7.72684216e-01 2.54053771e-01
-1.02345329e-02 -7.06012666e-01 1.17828898e-01 4.06837493e-01
3.08512092e-01 -2.25020230e-01 8.07839692e-01 -7.82429695e-01
-1.02586484e+00 -3.69963914e-01 -5.47390059e-02 3.36088799e-02
-8.48559067e-02 -4.59608644e-01 -5.11139154e-01 -6.54516578e-01
-4.90168005e-01 -7.03838587e-01 1.05748010e+00 2.20239803e-01
1.33896303e+00 -5.47821701e-01 -3.71363729e-01 2.09495485e-01
1.16873467e+00 1.63250878e-01 1.43464893e-01 6.53645813e-01
2.97773182e-01 4.11198050e-01 1.20844531e+00 7.92951703e-01
-5.78325279e-02 8.41046333e-01 5.06380379e-01 2.46439964e-01
5.17952859e-01 -6.80198669e-01 2.99399197e-01 3.95337909e-01
2.09201217e-01 -3.30519229e-01 -5.29850543e-01 6.29135013e-01
-2.27597570e+00 -7.29410052e-01 3.51420224e-01 3.00969076e+00
1.26585627e+00 4.93318290e-01 3.90351564e-01 2.14217380e-01
6.66424036e-01 -1.89037979e-01 -9.27592754e-01 -2.87057400e-01
2.46104091e-01 -7.47316703e-03 8.17887187e-01 6.96636558e-01
-9.04638350e-01 5.90351343e-01 5.74722290e+00 8.17275107e-01
-1.00187993e+00 -7.77930859e-03 7.75894344e-01 -8.88248384e-01
-5.53934991e-01 1.90994292e-01 -7.67064631e-01 6.58595264e-01
7.03130722e-01 -3.70918781e-01 9.22518432e-01 1.08871639e+00
1.67301312e-01 -3.45699221e-01 -1.54543400e+00 9.50981677e-01
-3.66902798e-01 -1.08055353e+00 -1.12876311e-01 1.79512948e-01
1.10031784e+00 -7.81008378e-02 2.83192068e-01 2.43339211e-01
8.22275579e-01 -7.77686596e-01 8.63737762e-01 2.71602720e-01
7.40302801e-01 -8.96856904e-01 5.41954160e-01 4.31944758e-01
-7.84148753e-01 -6.46339118e-01 -2.81043112e-01 -3.74437645e-02
1.91865727e-01 8.54825079e-01 -6.68693066e-01 2.22855672e-01
8.57813597e-01 2.16467738e-01 -3.77553076e-01 1.82248259e+00
-2.72347987e-01 5.52088201e-01 -5.07393062e-01 -3.91679943e-01
1.94957957e-01 -3.10730577e-01 3.81550968e-01 4.95074391e-01
1.74893975e-01 -1.71920836e-01 4.18669522e-01 3.29843700e-01
-4.22679007e-01 2.69284040e-01 -1.15446053e-01 -3.10937478e-03
6.39857054e-01 1.07578576e+00 -5.52090466e-01 -3.55560005e-01
-5.37891872e-02 8.13282728e-01 5.60147583e-01 2.01751903e-01
-7.08562374e-01 -2.92389572e-01 2.32385159e-01 3.74449313e-01
4.32957768e-01 1.53778017e-01 -2.22327918e-01 -9.16192055e-01
3.54966432e-01 -8.09733748e-01 6.24660552e-01 -4.51619655e-01
-1.22285867e+00 2.03369483e-01 1.15795121e-01 -1.07914364e+00
-5.28372265e-02 -3.34285200e-01 -3.08397979e-01 4.08720195e-01
-1.44887078e+00 -5.14118910e-01 -2.24100277e-01 1.37288615e-01
5.69928288e-01 4.35906291e-01 1.72212586e-01 2.14868098e-01
-2.61697292e-01 7.90203393e-01 1.13165474e+00 -3.66038710e-01
8.79969180e-01 -1.57238746e+00 6.64046919e-03 4.74590033e-01
3.21280599e-01 8.07174444e-01 6.52120054e-01 -6.82054162e-01
-1.37198424e+00 -7.01289475e-01 7.55173743e-01 -1.96521416e-01
8.21892619e-01 -5.46728849e-01 -4.53836948e-01 3.83897990e-01
-3.32818151e-01 -9.68745202e-02 5.84503055e-01 7.81296551e-01
-3.46716523e-01 -6.75254166e-01 -7.59635985e-01 6.93439126e-01
1.17536759e+00 -3.34890276e-01 -1.52858391e-01 5.04459739e-01
4.55418140e-01 -3.85008186e-01 -4.09793198e-01 1.92716375e-01
9.53796446e-01 -9.57577527e-01 6.65659904e-01 -6.06177866e-01
8.41008052e-02 -7.26584345e-02 2.58880258e-01 -1.47119415e+00
-2.28538692e-01 -1.22657347e+00 -2.24324629e-01 7.66502202e-01
6.62662148e-01 -5.70729852e-01 1.14847469e+00 8.50194275e-01
3.20710957e-01 -9.51884687e-01 -8.75611663e-01 -9.25541461e-01
-1.82815313e-01 2.83806715e-02 4.36112374e-01 1.19960107e-01
2.91697294e-01 2.79599249e-01 -7.28930950e-01 -2.32944652e-01
6.38700604e-01 6.89565241e-01 7.43299484e-01 -1.57248867e+00
-7.09990263e-01 -3.64854842e-01 3.87663394e-01 -1.76099384e+00
-2.75400519e-01 -6.78433716e-01 2.33613968e-01 -1.30382693e+00
7.14561701e-01 -8.92582059e-01 -7.91811287e-01 2.50863522e-01
-3.24538559e-01 -6.06493205e-02 2.44637877e-01 3.50330293e-01
-1.30095804e+00 3.21140796e-01 1.06085396e+00 -4.35101464e-02
-5.15171587e-01 5.77823997e-01 -1.06360841e+00 3.78466219e-01
5.57774842e-01 -4.93432522e-01 -6.08686566e-01 -2.42848098e-01
1.00853038e+00 1.27393380e-01 -1.06062479e-01 -4.36437696e-01
3.56843919e-01 -4.21955556e-01 -2.87247356e-02 -7.29006231e-01
1.21838950e-01 -8.32278848e-01 -8.17637965e-02 3.54926229e-01
-9.89273489e-01 -4.76998650e-02 -4.05690044e-01 1.09158683e+00
-3.03549562e-02 -3.13076228e-01 6.25505090e-01 -2.11737654e-03
-2.46816322e-01 3.43153298e-01 3.02062146e-02 1.69599533e-01
8.39262843e-01 -1.19946972e-01 -2.28654489e-01 -8.69419575e-01
-6.16008282e-01 6.57231212e-01 5.34709156e-01 2.63417184e-01
3.01094890e-01 -1.24953997e+00 -2.70706773e-01 -9.42841172e-02
1.92057207e-01 -4.50195074e-02 -1.70126855e-01 8.25165391e-01
-3.54289040e-02 5.64586639e-01 1.75412983e-01 -2.53020018e-01
-1.22997069e+00 7.14707434e-01 -5.47707081e-02 -7.77508736e-01
-9.36584175e-02 9.52488780e-01 1.79136753e-01 1.37866095e-01
7.46588528e-01 -2.48581454e-01 7.53802285e-02 2.50191450e-01
4.12847847e-01 5.78370988e-01 1.54803783e-01 2.37861454e-01
-6.82287663e-02 2.29848161e-01 -6.92474961e-01 -4.59308892e-01
1.20247209e+00 -2.60652751e-01 1.80775358e-03 5.85569382e-01
8.91494691e-01 4.57829386e-01 -1.20928204e+00 -3.52389723e-01
3.46839935e-01 -1.05618596e+00 -2.37685554e-02 -1.02102172e+00
-8.53870630e-01 3.10764760e-01 4.57246602e-01 4.68584388e-01
9.22083378e-01 -2.47055329e-02 7.82174051e-01 7.87897587e-01
6.39521718e-01 -1.54984903e+00 4.36241359e-01 4.54334646e-01
6.00515783e-01 -1.22174358e+00 3.78110297e-02 -1.72796205e-01
-5.69773376e-01 7.64255464e-01 4.74768698e-01 1.32784590e-01
4.80002075e-01 -2.83290654e-01 -1.72217667e-01 1.00884728e-01
-9.31463242e-01 -1.21731654e-01 2.57756233e-01 2.74539236e-02
1.15496203e-01 1.73750460e-01 -5.28917491e-01 6.94441438e-01
-2.07987726e-01 -7.31583685e-02 4.52984035e-01 9.94783103e-01
-7.05330372e-01 -1.45683408e+00 -3.44904244e-01 9.39275622e-01
-5.61285317e-01 -1.87517524e-01 -5.68809867e-01 2.46752888e-01
-3.09075534e-01 9.82709527e-01 -1.77399322e-01 -2.00291321e-01
1.58329651e-01 -1.84941337e-01 3.24032933e-01 -4.86748070e-01
-5.52167773e-01 1.86925247e-01 1.28166586e-01 -6.95950210e-01
2.70060394e-02 -6.65124834e-01 -7.34702468e-01 -1.31596923e-01
-7.61536002e-01 8.69286299e-01 5.15230179e-01 4.40875858e-01
3.73476744e-01 -5.91502013e-03 1.02499008e+00 -5.06512284e-01
-1.42040277e+00 -5.06943047e-01 -7.66212523e-01 6.39393210e-01
6.00619197e-01 -8.13296497e-01 -5.75475276e-01 -4.43047911e-01] | [4.638023376464844, 3.3294436931610107] |
d3a7ac7b-bb90-4da0-ad3d-ad0fc6d2ba71 | sharp-deviations-bounds-for-dirichlet | 2304.03056 | null | https://arxiv.org/abs/2304.03056v1 | https://arxiv.org/pdf/2304.03056v1.pdf | Sharp Deviations Bounds for Dirichlet Weighted Sums with Application to analysis of Bayesian algorithms | In this work, we derive sharp non-asymptotic deviation bounds for weighted sums of Dirichlet random variables. These bounds are based on a novel integral representation of the density of a weighted Dirichlet sum. This representation allows us to obtain a Gaussian-like approximation for the sum distribution using geometry and complex analysis methods. Our results generalize similar bounds for the Beta distribution obtained in the seminal paper Alfers and Dinges [1984]. Additionally, our results can be considered a sharp non-asymptotic version of the inverse of Sanov's theorem studied by Ganesh and O'Connell [1999] in the Bayesian setting. Based on these results, we derive new deviation bounds for the Dirichlet process posterior means with application to Bayesian bootstrap. Finally, we apply our estimates to the analysis of the Multinomial Thompson Sampling (TS) algorithm in multi-armed bandits and significantly sharpen the existing regret bounds by making them independent of the size of the arms distribution support. | ['Michal Valko', 'Daniil Tiapkin', 'Alexey Naumov', 'Pierre Menard', 'Denis Belomestny'] | 2023-04-06 | null | null | null | null | ['thompson-sampling', 'multi-armed-bandits'] | ['methodology', 'miscellaneous'] | [ 1.09264374e-01 3.72290969e-01 -3.69563580e-01 -1.86107680e-01
-8.18790853e-01 -6.95927203e-01 2.69217253e-01 1.43685371e-01
-3.20879072e-01 1.17032325e+00 8.97012055e-02 -5.02570450e-01
-5.40935636e-01 -8.89620960e-01 -7.28041708e-01 -9.23893929e-01
1.63162246e-01 8.71711493e-01 1.03985250e-01 -5.71813583e-02
3.36804837e-01 4.28886026e-01 -1.23460245e+00 -2.55733460e-01
7.73113251e-01 1.24167597e+00 -2.48151734e-01 7.91212380e-01
3.86539325e-02 2.22833693e-01 -3.35110426e-01 -5.82395554e-01
2.15967730e-01 -3.95986944e-01 -8.96806717e-01 -9.71572995e-02
2.85621379e-02 -5.15870273e-01 2.97892690e-01 1.22678030e+00
4.08644408e-01 1.77933380e-01 1.13891375e+00 -9.93751705e-01
-6.16996348e-01 9.11147833e-01 -1.25789273e+00 2.42404386e-01
3.34664434e-01 -8.40472877e-01 1.08357298e+00 -5.28916299e-01
2.97745198e-01 1.28987134e+00 8.58007610e-01 1.86604992e-01
-1.48964942e+00 -5.63090920e-01 -4.59837019e-02 1.37505308e-01
-1.20813179e+00 -3.99045259e-01 3.61993551e-01 -4.91006613e-01
3.14385086e-01 3.53805870e-01 6.13761604e-01 9.36105907e-01
-1.27267584e-01 9.49752629e-01 1.03677893e+00 -9.13051188e-01
5.07937729e-01 -2.47087553e-01 3.99680048e-01 4.47499275e-01
9.08149064e-01 -1.13490246e-01 -2.36954749e-01 -8.20627213e-01
8.30857038e-01 -9.18270722e-02 -2.86480337e-01 -4.47510213e-01
-6.67661726e-01 1.29736817e+00 -2.30639622e-01 3.91459353e-02
-6.23387992e-01 3.35770041e-01 1.91991642e-01 -1.46729171e-01
9.85354364e-01 -2.70428151e-01 -2.50757813e-01 -2.70497680e-01
-9.27564800e-01 4.36642498e-01 1.35815752e+00 9.67773497e-01
3.43062401e-01 -3.11404794e-01 -4.30512041e-01 4.87358689e-01
5.22004783e-01 7.57077038e-01 -2.50498772e-01 -1.08964741e+00
3.40605587e-01 -5.01961410e-01 7.74730742e-01 -4.98867035e-01
-8.40759873e-02 -3.59830379e-01 -6.81092560e-01 -2.55492896e-01
9.77228940e-01 -1.72021076e-01 -4.59503204e-01 1.69563830e+00
4.31306481e-01 -7.21819699e-02 -3.56640011e-01 5.14814556e-01
-2.93047071e-01 4.92350131e-01 -1.77887797e-01 -6.00958705e-01
1.08932316e+00 -5.18260241e-01 -8.10063064e-01 5.54058373e-01
3.42988133e-01 -8.33261967e-01 5.22029877e-01 7.63669252e-01
-1.40028763e+00 1.34931207e-01 -6.68214083e-01 3.06076944e-01
1.92127779e-01 -2.57540494e-01 7.41832793e-01 1.16060662e+00
-8.42242718e-01 6.98790073e-01 -6.84996307e-01 -3.53322417e-01
5.13159692e-01 -2.57161427e-02 4.51654196e-01 -1.03380762e-01
-9.16717947e-01 6.02599442e-01 1.26512125e-01 -9.45100654e-03
-5.22369862e-01 -5.76803446e-01 -1.38026118e-01 2.03107193e-01
4.67349291e-01 -8.25951755e-01 1.52861464e+00 -7.85551131e-01
-1.72135675e+00 5.33028305e-01 -2.27853179e-01 -5.51441371e-01
7.18224704e-01 -1.68215513e-01 2.87949234e-01 4.09186363e-01
2.03758016e-01 -3.82263184e-01 7.74024665e-01 -8.01026344e-01
-5.52036285e-01 -6.82608902e-01 9.31727663e-02 -2.66227841e-01
1.12106763e-01 -1.08322620e-01 1.30018830e-01 -6.43422365e-01
4.81465496e-02 -7.91589975e-01 -2.93299645e-01 -2.42102757e-01
-2.61283189e-01 -2.49846756e-01 -1.71181142e-01 -5.24363995e-01
1.15275145e+00 -1.88961315e+00 2.45551661e-01 6.03103578e-01
-3.76809724e-02 -2.94679433e-01 2.20908567e-01 5.20090759e-01
-6.55034706e-02 2.22071722e-01 -9.29578543e-02 -3.78914326e-02
4.58324999e-01 9.47197378e-02 -5.96685052e-01 9.79389608e-01
-4.94382471e-01 5.06591558e-01 -7.46665776e-01 -3.59229058e-01
-2.05686808e-01 5.14711551e-02 -7.42555320e-01 -2.78647900e-01
-1.23565368e-01 1.35103345e-01 -6.18578136e-01 3.55880380e-01
1.08888853e+00 -3.44622284e-01 2.15531155e-01 1.29002452e-01
1.77163091e-02 -5.00972271e-02 -1.23377025e+00 1.21923196e+00
-4.23160940e-01 5.56581207e-02 6.36132061e-01 -1.29579425e+00
3.78639489e-01 2.21792117e-01 2.63003588e-01 -1.12357736e-03
4.01108414e-01 1.94432035e-01 -4.03959125e-01 -3.24824184e-01
4.19564247e-01 -7.69593537e-01 -1.16084583e-01 7.35011280e-01
-3.86072904e-01 -1.43958315e-01 3.32482964e-01 1.56412557e-01
5.87850630e-01 1.55046701e-01 5.24426103e-01 -6.66292310e-01
1.86485514e-01 -4.61708635e-01 2.10302696e-01 1.13208675e+00
3.05526536e-02 2.65851974e-01 8.75035524e-01 -4.51750821e-03
-1.17842209e+00 -1.42756641e+00 -7.00212061e-01 1.34571207e+00
-1.56727865e-01 1.83584169e-02 -8.57516587e-01 -1.05920911e-01
2.97928005e-01 1.01510584e+00 -1.01130474e+00 2.87529558e-01
6.75436184e-02 -9.84770477e-01 4.99046206e-01 5.09570122e-01
1.61165014e-01 4.55317199e-02 -4.65431452e-01 1.76665664e-01
-2.30278179e-01 -6.97581708e-01 -3.23900640e-01 1.53692901e-01
-9.74831760e-01 -1.08936250e+00 -1.27259183e+00 -6.99186176e-02
2.07820117e-01 5.17864227e-02 8.28850567e-01 -4.98469800e-01
3.14381309e-02 7.03153551e-01 -4.05323774e-01 -7.16645062e-01
-3.22800368e-01 -5.42267002e-02 9.42928568e-02 2.33511850e-02
4.53265220e-01 -5.33080578e-01 -4.45569754e-01 3.74338597e-01
-9.23690617e-01 -3.43799680e-01 2.26978824e-01 7.64960706e-01
4.66380864e-01 -2.39438891e-01 7.79137671e-01 -6.01710141e-01
7.70127773e-01 -7.98662364e-01 -1.07165480e+00 3.36117715e-01
-5.49770534e-01 2.84846723e-01 3.01427394e-01 -4.55366224e-01
-1.20815194e+00 -4.32154566e-01 2.77481526e-01 2.48868223e-02
2.31736347e-01 3.53456497e-01 1.84836730e-01 8.06143805e-02
3.07580978e-01 -1.43173411e-01 -2.34543681e-01 -6.64143264e-01
4.72196668e-01 9.15381610e-01 1.99174002e-01 -1.19140005e+00
2.26932049e-01 8.81871402e-01 4.50433820e-01 -7.19600499e-01
-1.35241067e+00 -3.38797718e-01 -1.43923238e-01 1.23572774e-01
7.37098038e-01 -4.43641782e-01 -1.28661668e+00 1.80528566e-01
-1.04521072e+00 -1.62537172e-01 -4.03127044e-01 7.57016361e-01
-1.15159214e+00 6.92840636e-01 -5.76989710e-01 -1.68008840e+00
6.06024079e-02 -5.91602325e-01 9.85915184e-01 7.80902356e-02
-1.24946594e-01 -1.04120207e+00 4.10965115e-01 4.42985058e-01
3.97835732e-01 2.20873713e-01 8.64872456e-01 -5.75157881e-01
-2.05505550e-01 -1.62624747e-01 -3.29011768e-01 1.99209452e-01
-3.81149322e-01 -5.74582331e-02 -5.67716062e-01 -6.64227977e-02
2.49176934e-01 -1.90502908e-02 1.02553260e+00 1.06454325e+00
1.05559337e+00 -2.34210268e-01 -3.67091894e-01 3.73753697e-01
1.28735566e+00 -2.67621607e-01 4.65316296e-01 4.15160179e-01
-6.19066767e-02 3.19651634e-01 5.21766305e-01 1.16723979e+00
7.61501342e-02 4.71569687e-01 2.40905359e-01 5.95982730e-01
6.23646975e-01 -8.91467556e-02 4.32235189e-02 2.60029703e-01
-5.96528947e-01 -1.29667684e-01 -5.51024437e-01 6.28594339e-01
-2.02045894e+00 -1.23900497e+00 -3.97503287e-01 2.61220956e+00
1.06307805e+00 -1.85962632e-01 6.38253450e-01 -5.95087791e-03
1.10161412e+00 -2.21605226e-01 -3.45894128e-01 -6.58006012e-01
2.26699677e-03 7.23093688e-01 1.06657314e+00 7.07666695e-01
-7.39764512e-01 3.34693313e-01 7.19575882e+00 1.21509182e+00
-1.56140169e-02 4.86581802e-01 3.82874519e-01 -5.00220716e-01
-4.02605385e-01 8.93309116e-02 -8.74244928e-01 5.59099853e-01
1.02776086e+00 -4.89954948e-01 5.60564637e-01 8.29192936e-01
5.14845848e-02 -7.55925417e-01 -9.84409511e-01 7.97657311e-01
-1.21849954e-01 -9.29336846e-01 -4.94417340e-01 5.63993454e-01
9.33373153e-01 -3.73172849e-01 1.54849753e-01 -7.24986494e-02
7.83993244e-01 -1.04638183e+00 7.11707294e-01 6.71781361e-01
5.54675281e-01 -1.10164177e+00 7.17419565e-01 2.69544035e-01
-6.72885239e-01 -8.62671062e-02 -5.13783038e-01 -4.28674072e-01
4.54528093e-01 1.15200198e+00 -4.82372046e-01 5.50698936e-01
3.85143489e-01 -5.41158253e-03 3.68376613e-01 1.10832155e+00
-1.09150834e-01 8.17392051e-01 -8.98558617e-01 -1.22215621e-01
2.67501593e-01 -5.65363228e-01 4.86851245e-01 9.12076175e-01
4.85456198e-01 1.53739259e-01 -2.84540296e-01 7.54835308e-01
2.70760637e-02 4.65599485e-02 -1.55700997e-01 1.60892844e-01
4.91093189e-01 7.21005261e-01 -7.68382847e-01 -3.38984281e-01
-2.90699989e-01 5.29840648e-01 -1.21705479e-03 4.95525271e-01
-9.91495907e-01 -3.63919884e-01 4.03984159e-01 3.32235657e-02
1.10165203e+00 -9.98807177e-02 -3.45612317e-01 -9.54712152e-01
1.32047519e-01 -3.24045509e-01 3.85288239e-01 -3.26261312e-01
-1.43318939e+00 -2.05170512e-01 6.45377755e-01 -5.59772432e-01
-4.34709787e-01 -4.23790216e-01 -2.91657746e-01 1.15784812e+00
-1.31630290e+00 -7.26090908e-01 4.44612175e-01 3.73819560e-01
-4.86436971e-02 3.05345625e-01 6.43112421e-01 4.92655374e-02
-3.71311665e-01 2.79192120e-01 1.17889321e+00 -4.10654515e-01
3.77332240e-01 -1.16809559e+00 -1.48477718e-01 5.45819163e-01
-4.75967288e-01 3.85002017e-01 1.13798749e+00 -6.50121748e-01
-1.11172140e+00 -3.95116836e-01 3.10355753e-01 -1.53508067e-01
1.08225703e+00 -1.96351167e-02 -3.48405510e-01 9.10442531e-01
3.45826484e-02 -4.26049858e-01 1.01633072e+00 5.67829430e-01
-1.62002429e-01 1.93489119e-01 -1.37859762e+00 1.58051297e-01
7.76587903e-01 5.00255041e-02 -5.88507593e-01 5.65325081e-01
1.25828907e-01 -1.65401306e-02 -9.73915398e-01 2.77138781e-02
9.73319232e-01 -1.05152071e+00 8.70558977e-01 -7.76149750e-01
3.04096609e-01 2.09001049e-01 -5.97493887e-01 -1.03037822e+00
-3.34575891e-01 -9.29433525e-01 -1.80982634e-01 1.06896877e+00
8.17740113e-02 -9.00821030e-01 4.38375235e-01 3.82621050e-01
4.29493189e-01 -7.97410488e-01 -1.24258959e+00 -1.10729349e+00
5.26243031e-01 -3.66382271e-01 6.48790777e-01 6.34619057e-01
2.40567639e-01 -5.72461449e-03 -3.17287892e-01 -8.93420950e-02
1.17671025e+00 1.53485939e-01 4.53454584e-01 -1.44626665e+00
-8.29153836e-01 -5.82459807e-01 -4.80565289e-03 -1.28871727e+00
2.59727925e-01 -7.15274334e-01 -1.74544752e-01 -1.23126733e+00
5.28629482e-01 -3.13044935e-01 9.86558199e-02 -1.45021737e-01
2.41684601e-01 4.70223390e-02 9.80271474e-02 1.39446050e-01
-4.36919123e-01 6.06919587e-01 9.58267212e-01 2.42697909e-01
1.59625471e-01 5.26012242e-01 -8.22472572e-01 9.29019511e-01
5.85707486e-01 -5.63884556e-01 -9.91502628e-02 -2.90545225e-02
8.37173939e-01 3.25509697e-01 1.59054622e-01 -2.07789823e-01
-1.22155540e-01 -3.12713236e-01 2.41831854e-01 -8.27668369e-01
2.06303999e-01 -5.51433384e-01 -3.25842611e-02 3.59453768e-01
-3.35234493e-01 -5.02976358e-01 -6.70457408e-02 1.07391167e+00
5.17048299e-01 -1.00571001e+00 8.22894871e-01 1.15941220e-03
4.80666161e-01 -1.98002771e-01 -5.62838554e-01 3.03737402e-01
1.05184186e+00 -1.03369057e-01 -9.77568850e-02 -7.68611372e-01
-1.04231596e+00 1.42951056e-01 2.11105689e-01 -7.30613530e-01
7.30354711e-02 -1.24351966e+00 -8.05987537e-01 -3.12925786e-01
-5.58895886e-01 -2.20818315e-02 1.99901089e-01 1.49033546e+00
-3.96279246e-01 6.17020190e-01 7.79095516e-02 -3.54930669e-01
-8.74856651e-01 5.65296412e-01 -1.30215272e-01 -3.34393024e-01
2.38566417e-02 7.23553836e-01 1.35428742e-01 3.74403328e-01
1.71051741e-01 -3.92644972e-01 4.22499031e-01 2.81089306e-01
7.69430161e-01 1.07082689e+00 -3.27584863e-01 -5.72573878e-02
-3.50177616e-01 4.89762902e-01 -1.64063975e-01 -6.79280281e-01
1.56333685e+00 -5.81657708e-01 -2.57329553e-01 6.48008049e-01
9.41126049e-01 2.46519268e-01 -9.41339672e-01 -2.78684229e-01
-1.38108477e-01 -3.56411308e-01 7.80661404e-02 -4.08962131e-01
-6.37010157e-01 7.70974576e-01 1.71291098e-01 7.59018481e-01
1.00681686e+00 3.69211197e-01 6.19529188e-01 2.93239355e-01
4.67999697e-01 -1.29265559e+00 -4.93144274e-01 3.49454641e-01
7.26873577e-01 -6.66231573e-01 3.58502716e-01 -2.96921492e-01
-1.78365305e-01 1.26470661e+00 -5.44018269e-01 -9.12678763e-02
7.68867493e-01 1.91515446e-01 -9.16177511e-01 2.22248659e-01
-3.15666586e-01 -2.56657898e-01 -2.12082759e-01 5.74793875e-01
4.41091657e-01 3.12476069e-01 -7.89056361e-01 8.26073825e-01
-2.87137091e-01 3.80762905e-01 8.03141296e-01 7.45511532e-01
-6.39937103e-01 -8.48097324e-01 -8.15706730e-01 4.98683333e-01
-9.73456323e-01 -1.87787771e-01 -3.76167223e-02 5.46524048e-01
-6.40477836e-01 7.74029315e-01 -5.40865958e-03 5.24302244e-01
-1.24300174e-01 3.26287776e-01 1.20272863e+00 -3.20356280e-01
1.78380013e-01 3.92304063e-01 9.17066708e-02 -5.58859808e-03
-6.14307404e-01 -9.39655304e-01 -6.02628469e-01 -8.55514228e-01
-7.93742716e-01 5.54977357e-01 7.00802326e-01 9.65837538e-01
-4.32522297e-02 9.20222402e-02 5.02971888e-01 -8.10743034e-01
-1.48171377e+00 -1.14453471e+00 -1.27944350e+00 -1.77312031e-01
1.04899846e-01 -6.72519863e-01 -6.50967538e-01 -1.33527637e-01] | [4.603140354156494, 3.324242353439331] |
aab7b9d9-3f82-4258-9177-317fdcfdbb43 | ecg-beats-classification-via-online-sparse | 2008.06672 | null | https://arxiv.org/abs/2008.06672v1 | https://arxiv.org/pdf/2008.06672v1.pdf | ECG beats classification via online sparse dictionary and time pyramid matching | Recently, the Bag-Of-Word (BOW) algorithm provides efficient features and promotes the accuracy of the ECG classification system. However, BOW algorithm has two shortcomings: (1). it has large quantization errors and poor reconstruction performance; (2). it loses heart beat's time information, and may provide confusing features for different kinds of heart beats. Furthermore, ECG classification system can be used for long time monitoring and analysis of cardiovascular patients, while a huge amount of data will be produced, so we urgently need an efficient compression algorithm. In view of the above problems, we use the wavelet feature to construct the sparse dictionary, which lower the quantization error to a minimum. In order to reduce the complexity of our algorithm and adapt to large-scale heart beats operation, we combine the Online Dictionary Learning with Feature-sign algorithm to update the dictionary and coefficients. Coefficients matrix is used to represent ECG beats, which greatly reduces the memory consumption, and solve the problem of quantitative error simultaneously. Finally, we construct the pyramid to match coefficients of each ECG beat. Thus, we obtain the features that contain the beat time information by time stochastic pooling. It is efficient to solve the problem of losing time information. The experimental results show that: on the one hand, the proposed algorithm has advantages of high reconstruction performance for BOW, this storage method is high fidelity and low memory consumption; on the other hand, our algorithm yields highest accuracy in ECG beats classification; so this method is more suitable for large-scale heart beats data storage and classification. | ['Nanyu Li', 'Duo Deng', 'Chunyu Yuan', 'Yujuan Si'] | 2020-08-15 | null | null | null | null | ['ecg-classification'] | ['medical'] | [ 7.83929499e-05 -8.84738922e-01 -1.19590893e-01 -1.41183302e-01
-4.42127347e-01 -2.40366030e-02 -3.38653088e-01 4.75621670e-01
-3.04602921e-01 2.67980963e-01 2.40379289e-01 8.17087516e-02
-2.27480873e-01 -1.05268896e+00 2.17354849e-01 -1.01422918e+00
1.95594244e-02 9.02952105e-02 3.78018796e-01 -1.49855554e-01
3.08466882e-01 4.71100390e-01 -1.42994475e+00 1.73339739e-01
7.40009367e-01 1.42708576e+00 3.80882472e-01 2.93760628e-01
-2.09610477e-01 6.95017993e-01 -5.75581253e-01 1.40654862e-01
9.69675630e-02 -7.52696455e-01 -2.65986145e-01 -8.43974799e-02
-4.54934359e-01 -4.10389096e-01 -5.32105446e-01 1.09325421e+00
1.11805832e+00 8.57937336e-03 2.72915989e-01 -7.98716545e-01
-2.95982391e-01 2.76752412e-01 -5.30026793e-01 6.10127568e-01
8.42559189e-02 -5.10618649e-02 5.53481936e-01 -6.14948213e-01
2.09501013e-01 1.09245646e+00 9.31618631e-01 1.84471846e-01
-8.01994324e-01 -9.13960040e-01 -3.11809003e-01 5.29523551e-01
-1.85573375e+00 -4.16568875e-01 9.26128745e-01 -2.80669510e-01
4.71160591e-01 5.80961883e-01 1.09906089e+00 3.35643858e-01
3.88068169e-01 6.13085270e-01 7.94884682e-01 -2.89621860e-01
4.08373475e-02 -1.76285237e-01 1.67448502e-02 7.26145446e-01
1.97524697e-01 -5.39981015e-02 -1.64754167e-01 -2.12644443e-01
9.39367235e-01 7.79286087e-01 -4.86541659e-01 9.69678909e-02
-1.38405681e+00 7.44763672e-01 3.12116027e-01 8.04455161e-01
-3.42036009e-01 -7.42488876e-02 9.09796119e-01 4.22554642e-01
6.37964671e-03 -4.33764867e-02 -1.55218929e-01 -3.54918361e-01
-8.54059517e-01 2.90574819e-01 3.72997344e-01 6.58698499e-01
7.15095758e-01 8.39801952e-02 -2.39685744e-01 9.26824927e-01
2.36697227e-01 6.63925767e-01 1.12136531e+00 -6.43566191e-01
4.84863132e-01 5.60916185e-01 -2.85345733e-01 -1.78782427e+00
-5.30411541e-01 -3.37691218e-01 -1.51375079e+00 -2.60809541e-01
-5.70034496e-02 3.14565510e-01 -5.80322325e-01 1.24042022e+00
1.50616303e-01 3.07207555e-01 8.44630972e-03 9.74253178e-01
1.12397027e+00 1.01203823e+00 -6.53680414e-03 -1.04453135e+00
1.61730981e+00 -4.50909197e-01 -1.34307241e+00 1.67286992e-01
4.57252473e-01 -9.07465518e-01 8.75322640e-01 3.65121335e-01
-8.74529660e-01 -9.49696600e-01 -1.04151058e+00 5.90228550e-02
4.12900932e-02 2.45046288e-01 5.91132820e-01 4.87844914e-01
-3.28904569e-01 5.83725989e-01 -7.09325254e-01 -5.23654111e-02
4.36585695e-01 1.26399636e-01 -1.66519377e-02 -1.20997608e-01
-1.43001938e+00 6.91072166e-01 5.14491141e-01 2.86670953e-01
-4.43312764e-01 -5.02697349e-01 -7.12027192e-01 1.79909408e-01
-3.46715376e-02 -2.08993152e-01 8.13587666e-01 -5.12419462e-01
-1.10502696e+00 4.52224880e-01 -2.03680366e-01 -6.00048080e-02
2.12700203e-01 1.89233467e-01 -8.54390502e-01 2.73217946e-01
1.14204846e-01 -4.41788137e-02 6.46540761e-01 -4.02984887e-01
-5.68136752e-01 -4.65747297e-01 -6.43368661e-01 1.49624079e-01
-6.37225389e-01 -3.37454639e-02 -4.67213422e-01 -8.87448907e-01
7.44021714e-01 -5.50235927e-01 -2.85792202e-01 6.47713803e-03
2.96930641e-01 -9.39108208e-02 9.21183407e-01 -7.78758347e-01
2.04703498e+00 -2.82302308e+00 2.83912271e-02 3.64562303e-01
1.45835221e-01 2.85973310e-01 2.14462280e-01 3.72036397e-01
5.60105368e-02 -1.58829000e-02 -2.43281350e-01 2.18601093e-01
-5.54125607e-01 3.98535430e-01 -2.50479192e-01 3.72327983e-01
-2.24176213e-01 4.87881094e-01 -6.52632475e-01 -1.18873370e+00
1.81746751e-01 4.68493432e-01 -3.99482459e-01 1.80011764e-01
5.26320398e-01 2.35198528e-01 -8.70396256e-01 5.49535692e-01
7.68454254e-01 -1.18987694e-01 -4.98248115e-02 -7.44258463e-01
-1.29640415e-01 -6.92452341e-02 -1.68861747e+00 1.93605006e+00
-1.01615578e-01 1.62228709e-04 -1.52179763e-01 -1.09175587e+00
1.33924365e+00 5.43690264e-01 7.69115567e-01 -7.66214728e-01
3.64242077e-01 3.29782188e-01 -1.22943819e-01 -9.70857263e-01
-1.21250898e-01 -4.94100541e-01 4.31083925e-02 1.78700075e-01
-2.85518050e-01 -3.69984284e-02 1.70617387e-01 -9.01900977e-02
7.84626961e-01 -4.56275523e-01 3.73537272e-01 -2.61285961e-01
8.49604666e-01 -2.56807536e-01 1.18290675e+00 1.75523639e-01
-1.65292487e-01 6.09317183e-01 9.58607420e-02 -7.88564622e-01
-6.13036513e-01 -7.01008379e-01 -4.34931040e-01 2.86213130e-01
4.64879185e-01 -6.20836437e-01 -2.67425686e-01 -1.50164053e-01
3.07815596e-02 -1.16978809e-01 -1.65355712e-01 -4.72001731e-01
-8.97458196e-01 -9.59468842e-01 4.19991404e-01 5.09523094e-01
8.95720065e-01 -1.14484167e+00 -8.96181047e-01 6.89341962e-01
-6.20193183e-01 -4.97416705e-01 -5.51389158e-01 -3.84289861e-01
-1.33034933e+00 -9.95605052e-01 -6.80686593e-01 -1.05560493e+00
2.75218248e-01 3.13993484e-01 7.14972734e-01 6.74685657e-01
-8.27955842e-01 -2.16614082e-01 -5.24750888e-01 -3.69986027e-01
4.62584049e-02 -3.68904382e-01 1.56785339e-01 7.62919337e-02
4.12769288e-01 -7.14911163e-01 -9.04465973e-01 2.95306630e-02
-8.17265093e-01 -2.38815114e-01 6.49258733e-01 1.04143786e+00
8.68500829e-01 6.68662846e-01 5.42010605e-01 -3.76698762e-01
6.70145929e-01 -3.87633443e-02 -3.53719085e-01 3.19117755e-02
-9.08206940e-01 -4.97456282e-01 8.26513946e-01 -3.16517293e-01
-4.43377882e-01 -1.53859869e-01 -2.75789857e-01 -4.00848210e-01
3.87627840e-01 5.79913914e-01 -1.32500768e-01 -4.98320349e-02
3.14143538e-01 7.61768699e-01 4.23048884e-01 -7.59349287e-01
-1.34691224e-01 9.36114371e-01 4.22860980e-01 -3.27387691e-01
3.82691979e-01 2.43272290e-01 2.91815430e-01 -4.81235087e-01
-2.84271061e-01 -4.58527297e-01 -3.69647384e-01 9.96441394e-03
9.56402361e-01 -9.99076724e-01 -9.07299757e-01 3.29933643e-01
-1.05028903e+00 5.98345518e-01 -2.68007815e-01 8.22245002e-01
-3.62483710e-01 9.50563073e-01 -9.36621249e-01 -7.70203590e-01
-7.22488105e-01 -1.00552285e+00 7.02123582e-01 3.60010624e-01
8.83629825e-03 -4.69413728e-01 -2.60076553e-01 -5.91902733e-02
3.77534002e-01 3.21104884e-01 9.38918948e-01 5.18666543e-02
-4.66442823e-01 -4.19653744e-01 -1.19504772e-01 3.04097980e-01
2.42569953e-01 -2.55258977e-01 -4.73887354e-01 -5.16257226e-01
5.07529914e-01 -2.00874567e-01 6.14742041e-01 1.92780212e-01
1.64197743e+00 -3.18077117e-01 -3.09815437e-01 6.49848282e-01
1.45010793e+00 7.43481696e-01 8.96664143e-01 2.03055114e-01
4.61277217e-01 2.45372981e-01 1.05094469e+00 8.04127216e-01
1.46492600e-01 5.03345788e-01 -3.19000669e-02 -3.06707472e-01
-2.52200924e-02 -2.22476870e-01 -1.22005992e-01 1.85143495e+00
-1.52448967e-01 3.25007498e-01 -6.34709418e-01 4.63116407e-01
-1.92950344e+00 -1.24951971e+00 -2.03012437e-01 2.19662690e+00
1.13730443e+00 8.03630042e-04 7.71178911e-03 7.62676179e-01
5.25877118e-01 1.32361710e-01 -3.21744174e-01 -7.24969953e-02
6.99625611e-02 3.02992731e-01 1.84315480e-02 1.68669643e-03
-8.83838236e-01 2.27187768e-01 5.93790960e+00 1.08612406e+00
-1.30471396e+00 2.48481214e-01 4.38297123e-01 1.01500019e-01
7.03347288e-03 -8.69190544e-02 -5.99891186e-01 9.51428056e-01
3.62676561e-01 -2.46017843e-01 2.33516440e-01 6.78587198e-01
2.14542910e-01 2.98019618e-01 -5.92305660e-01 1.86770284e+00
1.11927584e-01 -1.27702534e+00 1.41011149e-01 -2.96426743e-01
3.95508930e-02 -4.74738955e-01 -2.62907177e-01 1.79013908e-01
-8.19968283e-01 -7.66351342e-01 2.50209153e-01 6.57952607e-01
9.69405591e-01 -8.63067448e-01 1.01579154e+00 4.52161878e-01
-1.66685927e+00 -3.26339245e-01 -9.15074170e-01 -2.26573691e-01
5.22630401e-02 9.75512743e-01 3.75433154e-02 6.60484195e-01
9.28363442e-01 1.00617790e+00 -3.12120199e-01 1.15005970e+00
3.56583118e-01 2.86988139e-01 -2.83783704e-01 -8.30974951e-02
-1.90494537e-01 -4.38796461e-01 1.78409770e-01 1.06141865e+00
6.37423217e-01 7.66473234e-01 6.59305871e-01 2.89764881e-01
3.58884215e-01 6.92899048e-01 -3.39189380e-01 7.28558227e-02
7.33974934e-01 1.09281087e+00 -5.80810666e-01 -5.50578594e-01
-2.47350961e-01 7.71328926e-01 -2.37988412e-01 -1.12671241e-01
-7.17823565e-01 -1.12298000e+00 1.68638736e-01 8.14460292e-02
1.63482442e-01 -4.49150428e-02 -1.44563720e-01 -1.12444937e+00
2.93418258e-01 -9.51270223e-01 6.45491302e-01 -5.32918632e-01
-1.14659619e+00 4.20279831e-01 -4.59438026e-01 -1.77811539e+00
2.66479254e-01 -1.56907752e-01 -5.42265296e-01 9.18476462e-01
-1.34071493e+00 -6.30013525e-01 -6.11006618e-01 7.77289152e-01
3.24233800e-01 -2.86849290e-01 1.22171307e+00 9.01788354e-01
-4.21737522e-01 5.14298737e-01 1.65009692e-01 3.27891022e-01
6.63892031e-01 -5.71464837e-01 -3.97527754e-01 4.95301366e-01
-1.65586933e-01 7.14331210e-01 3.17147106e-01 -5.96959949e-01
-1.49336958e+00 -8.43083501e-01 1.03235865e+00 1.90161750e-01
1.50365382e-01 1.24166898e-01 -9.46247637e-01 5.37356287e-02
-4.96167213e-01 3.19688350e-01 9.26734388e-01 -7.31085017e-02
-5.02848551e-02 -1.01476693e+00 -1.10187089e+00 2.43865728e-01
6.82994366e-01 -5.20644188e-01 -7.05957532e-01 5.03043771e-01
5.37025392e-01 -4.22179490e-01 -1.36417770e+00 6.30854607e-01
6.87537074e-01 -8.20593297e-01 1.06790662e+00 -9.77445617e-02
1.62678808e-01 -8.45469773e-01 -4.55390438e-02 -7.02893734e-01
-8.97010326e-01 -3.32888722e-01 4.80282865e-02 1.23917544e+00
-1.88017070e-01 -7.05895603e-01 3.78961444e-01 7.69618973e-02
1.94453809e-03 -9.80948210e-01 -1.05696404e+00 -6.48529351e-01
-4.61786509e-01 1.19977035e-01 8.21031451e-01 1.05889833e+00
2.57859528e-01 2.50773489e-01 -5.97201407e-01 -2.00589314e-01
5.06753862e-01 5.55578113e-01 4.09387976e-01 -1.27819300e+00
-4.09396857e-01 -1.82435900e-01 -7.89523959e-01 -1.07474446e+00
-6.20223105e-01 -7.99820900e-01 -1.82393670e-01 -1.39182472e+00
2.84841955e-01 -9.57120240e-01 -7.49262273e-01 3.85222256e-01
-2.82371849e-01 3.13123465e-01 8.93426761e-02 7.70116866e-01
-3.44386548e-01 4.69532460e-01 1.52175212e+00 -1.88706800e-01
-1.56707674e-01 -1.37654349e-01 -5.96333027e-01 4.08819616e-01
6.40528023e-01 -4.47227150e-01 -3.48454654e-01 -3.47320467e-01
-4.71993275e-02 5.53121865e-01 2.54236092e-03 -1.20423377e+00
2.17393830e-01 -1.90913826e-02 7.63346255e-01 -5.44211447e-01
2.64509678e-01 -9.63819683e-01 3.59817266e-01 1.14363480e+00
9.35398638e-02 2.41360098e-01 -1.12122530e-02 5.62763453e-01
-5.22923112e-01 -3.41244012e-01 7.02853501e-01 -2.19296917e-01
-5.84330559e-01 6.12038970e-01 -1.62460566e-01 -1.23375617e-01
8.89931142e-01 -2.31218204e-01 1.74499065e-01 -1.05447099e-01
-6.65656686e-01 1.87911302e-01 1.75070092e-01 2.05883607e-01
8.59329760e-01 -1.77678359e+00 -7.06829309e-01 5.39543867e-01
1.67403251e-01 6.89339936e-02 5.63280702e-01 1.00564706e+00
-1.00827205e+00 -1.23245642e-01 -2.75938898e-01 -6.98086858e-01
-1.36291170e+00 7.07897604e-01 4.92092706e-02 -1.22295469e-01
-1.14153361e+00 5.38070619e-01 -1.71754330e-01 3.61790299e-01
1.27894357e-01 -2.14581877e-01 -4.82456893e-01 1.75999135e-01
9.58055913e-01 3.53728771e-01 4.00467366e-02 -4.65387851e-01
-4.34597045e-01 1.14532375e+00 1.95572436e-01 3.45599443e-01
1.16256201e+00 -1.22672983e-01 -5.26099861e-01 4.21196043e-01
1.27360094e+00 -4.02563475e-02 -3.11074585e-01 -8.82590115e-02
-4.66135979e-01 -8.48845959e-01 1.70089558e-01 -4.41831619e-01
-1.14477086e+00 1.17516696e+00 1.09226763e+00 2.64042467e-01
1.56892562e+00 -5.99840343e-01 1.47738934e+00 2.56056696e-01
4.84361649e-01 -1.11428845e+00 -5.33213615e-02 2.99117006e-02
6.17998064e-01 -8.33598316e-01 4.66956317e-01 -4.74755675e-01
-3.67881715e-01 1.38002586e+00 2.06759200e-01 -2.05810815e-01
9.54224229e-01 1.75109938e-01 2.64058828e-01 -7.57858828e-02
-2.25383937e-01 1.79876387e-01 -4.79273349e-02 3.17016453e-01
2.76328564e-01 1.10637713e-02 -1.09435797e+00 8.55014205e-01
-7.76984394e-02 2.40929022e-01 2.13413574e-02 9.11091805e-01
-7.53503621e-01 -1.20662534e+00 -4.10114199e-01 5.45735359e-01
-5.46578050e-01 -8.24208483e-02 4.40145940e-01 2.55653888e-01
4.94470835e-01 9.23591733e-01 -6.72059357e-02 -5.11395574e-01
4.83973861e-01 -1.05330991e-02 3.70269984e-01 -2.46364921e-01
-3.96541595e-01 3.95362735e-01 -4.84022051e-01 -4.49599385e-01
-3.16020817e-01 -5.92742801e-01 -1.44723952e+00 -2.09141359e-01
-4.54395741e-01 5.16536236e-01 2.46330440e-01 5.94117880e-01
3.61511230e-01 5.38876235e-01 8.59626412e-01 -1.21892445e-01
-7.27271378e-01 -8.17367017e-01 -9.76038277e-01 7.99384892e-01
1.04577422e-01 -4.84726578e-01 -2.58019209e-01 1.50465980e-01] | [14.183272361755371, 3.2294209003448486] |
2068f1ee-e1c8-44a3-b012-04b197b0f4ee | emotion-experiencer-recognition-as-a | 2305.16731 | null | https://arxiv.org/abs/2305.16731v2 | https://arxiv.org/pdf/2305.16731v2.pdf | Automatic Emotion Experiencer Recognition | The most prominent subtask in emotion analysis is emotion classification; to assign a category to a textual unit, for instance a social media post. Many research questions from the social sciences do, however, not only require the detection of the emotion of an author of a post but to understand who is ascribed an emotion in text. This task is tackled by emotion role labeling which aims at extracting who is described in text to experience an emotion, why, and towards whom. This could, however, be considered overly sophisticated if the main question to answer is who feels which emotion. A targeted approach for such setup is to classify emotion experiencer mentions (aka "emoters") regarding the emotion they presumably perceive. This task is similar to named entity recognition of person names with the difference that not every mentioned entity name is an emoter. While, very recently, data with emoter annotations has been made available, no experiments have yet been performed to detect such mentions. With this paper, we provide baseline experiments to understand how challenging the task is. We further evaluate the impact on experiencer-specific emotion categorization and appraisal detection in a pipeline, when gold mentions are not available. We show that experiencer detection in text is a challenging task, with a precision of .82 and a recall of .56 (F1 =.66). These results motivate future work of jointly modeling emoter spans and emotion/appraisal predictions. | ['Roman Klinger', 'Maximilian Wegge'] | 2023-05-26 | null | null | null | null | ['emotion-classification', 'emotion-classification'] | ['computer-vision', 'natural-language-processing'] | [ 8.36932138e-02 2.75586337e-01 1.74242958e-01 -4.69764233e-01
-4.66981441e-01 -7.85819054e-01 6.82671010e-01 6.41502202e-01
-6.45715773e-01 5.37476122e-01 4.12211508e-01 -6.20602779e-02
2.61286438e-01 -5.16103983e-01 -1.59811988e-01 -4.65301722e-01
2.32894421e-01 2.43722752e-01 -3.44117880e-01 -1.69398814e-01
2.14183599e-01 3.82298887e-01 -1.70690918e+00 4.44532573e-01
5.67768812e-01 1.18504095e+00 -1.90772429e-01 8.04967403e-01
-3.11137736e-01 9.32764113e-01 -9.93992865e-01 -7.26801753e-01
-3.72443050e-01 -4.04610932e-01 -1.06541252e+00 1.67935863e-01
8.40264708e-02 2.57544935e-01 4.47404623e-01 8.33639145e-01
4.09999847e-01 2.41640538e-01 1.00640738e+00 -1.23562074e+00
-2.44384184e-01 7.07352221e-01 -3.43601584e-01 2.17930183e-01
6.36113226e-01 -4.09949362e-01 1.19025576e+00 -7.07037270e-01
7.68742085e-01 9.08847213e-01 5.05620539e-01 4.35089290e-01
-9.15470064e-01 -4.29108948e-01 7.16202036e-02 -2.66218521e-02
-1.22884107e+00 -4.16256279e-01 6.29530907e-01 -6.71906173e-01
9.88407731e-01 4.79499936e-01 3.57724696e-01 1.14493322e+00
-2.46273071e-01 5.65148592e-01 1.40051174e+00 -4.75917548e-01
4.50396717e-01 7.64476299e-01 3.60764295e-01 4.08558667e-01
-2.44134724e-01 -6.84961557e-01 -5.43160141e-01 -1.87203214e-01
3.99126299e-02 -3.38171542e-01 -2.15394169e-01 2.90316105e-01
-8.63656640e-01 7.70305693e-01 2.39347950e-01 7.49435723e-01
-7.79687703e-01 -6.35927767e-02 8.01057220e-01 5.47980726e-01
8.59261632e-01 9.18073058e-01 -8.32409263e-01 -6.54825985e-01
-8.69526982e-01 -5.43268397e-02 1.37481821e+00 5.53927243e-01
5.87380767e-01 -4.38660830e-01 -4.52705026e-02 8.57811630e-01
2.97231749e-02 2.20745206e-02 4.26927537e-01 -5.79317927e-01
-4.58686203e-02 8.31090212e-01 3.83774847e-01 -1.08907366e+00
-6.96466863e-01 -3.38388354e-01 -5.57783067e-01 7.00248480e-02
4.35093522e-01 -7.71968365e-01 -4.96933550e-01 1.62886858e+00
4.17099983e-01 6.65529072e-02 1.35129780e-01 1.00985956e+00
1.12460101e+00 4.90240544e-01 5.10817945e-01 -3.07672232e-01
2.08280301e+00 -4.80196178e-01 -8.60251069e-01 -1.05673499e-01
8.90567958e-01 -9.41673279e-01 8.94836664e-01 3.56797904e-01
-6.59864366e-01 -2.80976355e-01 -7.41869032e-01 -3.00416560e-03
-8.00894141e-01 4.51277226e-01 8.31862986e-01 9.43642497e-01
-7.16618061e-01 5.42197585e-01 -3.43273759e-01 -7.61426687e-01
1.82425424e-01 2.02773362e-01 -5.02378404e-01 5.42487442e-01
-1.42626953e+00 1.16883910e+00 -5.40967137e-02 -1.19588487e-01
-1.35522529e-01 -6.15350604e-01 -8.30356359e-01 1.10309891e-01
1.84569195e-01 -3.86185646e-01 1.13070881e+00 -1.55503607e+00
-1.44909680e+00 1.45878053e+00 -2.36299515e-01 -3.11121911e-01
2.98100054e-01 -3.20167363e-01 -6.89338803e-01 7.50126392e-02
-5.31015731e-02 4.14580405e-01 8.33814800e-01 -1.03060889e+00
-6.11625969e-01 -3.93487275e-01 2.37013519e-01 1.87814459e-01
-5.48072040e-01 9.13465083e-01 -1.45305768e-01 -3.69943589e-01
-2.59000033e-01 -9.88066554e-01 -2.19670773e-01 -5.53976178e-01
-4.62398380e-01 -6.03870153e-01 3.21331590e-01 -7.96342432e-01
1.22221160e+00 -2.13352203e+00 -4.79184799e-02 2.57646948e-01
2.80587137e-01 -7.79732987e-02 1.30501553e-01 5.41704655e-01
-2.55725890e-01 4.06195223e-01 6.80942163e-02 -4.34173375e-01
2.95150131e-01 -2.23086998e-01 -3.55306983e-01 3.97016048e-01
2.62113243e-01 6.53332829e-01 -7.64322519e-01 -4.43148315e-01
-8.65415558e-02 6.42776191e-01 -3.40737402e-01 4.28846091e-01
-1.89912580e-02 2.99550354e-01 -4.31535959e-01 3.75431389e-01
3.18318576e-01 -1.43804610e-01 1.48336530e-01 -2.94230163e-01
-3.25992465e-01 3.76570642e-01 -1.22721314e+00 1.12560058e+00
-7.82305896e-01 7.71801054e-01 3.40668887e-01 -8.35905910e-01
1.05302870e+00 5.99987686e-01 5.31423569e-01 -1.68257907e-01
5.81433237e-01 8.68863538e-02 -6.67958800e-03 -7.04568624e-01
8.12433302e-01 -5.79940677e-01 -5.34951746e-01 4.91143107e-01
-1.64714810e-02 1.80291012e-01 1.50721982e-01 1.89190269e-01
1.38611221e+00 3.51130555e-04 4.61880267e-01 -1.24440745e-01
5.17307639e-01 -4.58070822e-02 3.90475482e-01 5.01883626e-01
-2.47340843e-01 4.18570846e-01 7.95067072e-01 -4.15002368e-02
-6.79780483e-01 -3.51801962e-01 -8.54493380e-02 1.52454925e+00
-1.97849467e-01 -6.66581631e-01 -8.51781249e-01 -7.37585306e-01
-3.55034769e-01 7.25305855e-01 -7.79317856e-01 1.64943531e-01
5.48211113e-03 -6.24343693e-01 6.86519384e-01 2.27019742e-01
1.04388155e-01 -1.34064138e+00 -9.15415704e-01 1.45748422e-01
-3.06732893e-01 -1.29826581e+00 4.25410755e-02 4.83408213e-01
-1.83899999e-01 -8.75977755e-01 -5.06993175e-01 -5.16265035e-01
4.38188672e-01 -4.62345809e-01 1.46035111e+00 1.59640223e-01
-3.03703636e-01 7.24582613e-01 -7.33596027e-01 -5.81532657e-01
-2.69088805e-01 2.61164218e-01 -2.75582355e-02 3.90468836e-01
7.65450895e-01 -3.99467200e-01 -4.64048356e-01 1.03889331e-01
-6.86508000e-01 -2.74670631e-01 3.82138997e-01 3.45092297e-01
-1.61148980e-02 -4.70093600e-02 6.27623498e-01 -1.21554673e+00
8.47939491e-01 -7.27650046e-01 2.71908522e-01 9.09932330e-02
-1.97870046e-01 -2.42951661e-01 6.64927185e-01 -4.98580307e-01
-1.14372194e+00 2.12393388e-01 -5.44477999e-01 9.91411507e-02
-9.89482284e-01 5.99226356e-01 -2.74061322e-01 3.73705566e-01
6.12618148e-01 -2.83768266e-01 -3.17568153e-01 -3.91771615e-01
3.96429598e-01 1.05616891e+00 1.17774598e-01 -5.16235292e-01
4.05000299e-01 1.91003904e-01 -2.62755781e-01 -1.18392313e+00
-1.13444865e+00 -1.10158169e+00 -5.04199922e-01 -3.33214641e-01
1.11032534e+00 -9.18273091e-01 -1.10868001e+00 1.39510170e-01
-1.20188773e+00 -2.04026207e-01 -5.85229248e-02 3.18796754e-01
-3.94238561e-01 1.27562851e-01 -7.51544356e-01 -1.10583186e+00
-3.60212356e-01 -5.08827507e-01 8.73450756e-01 1.38995543e-01
-1.04719770e+00 -9.86566305e-01 -8.82849991e-02 4.22925174e-01
3.48473549e-01 3.92498821e-01 6.60872579e-01 -1.14142215e+00
6.00320041e-01 -3.22675318e-01 -1.26039803e-01 2.32050642e-01
-8.93541202e-02 1.40363500e-02 -1.18412971e+00 2.00392321e-01
1.18845090e-01 -5.88893175e-01 6.17025375e-01 -1.04385480e-01
8.56771111e-01 -2.73465097e-01 -1.35577962e-01 -7.74643421e-02
1.09527719e+00 -9.75832865e-02 5.92849791e-01 2.53624231e-01
4.93816793e-01 1.25614154e+00 8.08457136e-01 6.88487411e-01
2.99713701e-01 6.59923613e-01 2.34500155e-01 -2.80698121e-01
3.46986800e-01 2.26373091e-01 4.83982980e-01 4.96579885e-01
-1.61741987e-01 -3.09444517e-01 -8.26512456e-01 4.96289819e-01
-1.62101364e+00 -9.55213904e-01 -7.08494842e-01 1.64836478e+00
9.15675938e-01 -1.22847073e-01 8.53420049e-02 2.08328739e-01
7.29205728e-01 6.57192394e-02 -9.17866975e-02 -8.48989606e-01
7.70635903e-02 2.83541769e-01 4.42023613e-02 4.35452312e-01
-1.28067815e+00 9.73259687e-01 4.69860077e+00 4.54475611e-01
-1.02903879e+00 2.02547073e-01 8.94439995e-01 2.06513882e-01
6.25848100e-02 -5.04072644e-02 -7.44367301e-01 3.88238728e-01
1.16326106e+00 1.79961413e-01 1.47248596e-01 9.55012679e-01
1.80984214e-01 -2.60489017e-01 -1.17252648e+00 1.13810217e+00
2.94743180e-01 -5.31350851e-01 -6.14262700e-01 -2.00360984e-01
3.05056900e-01 -3.30131322e-01 -2.20044076e-01 6.03231013e-01
9.31396782e-02 -1.03719628e+00 5.72066307e-01 4.94826943e-01
4.77663219e-01 -7.63932765e-01 8.26501071e-01 2.54238278e-01
-9.58991885e-01 2.37199962e-01 -2.26753414e-01 -4.90569711e-01
1.95092529e-01 8.48905385e-01 -8.65917146e-01 1.73251331e-01
6.72592700e-01 5.18404305e-01 -5.28035223e-01 7.10016727e-01
-5.09914041e-01 7.81172633e-01 -2.15707004e-01 -3.54896486e-01
1.64823439e-02 1.39187202e-02 4.18732643e-01 1.79878712e+00
3.09850156e-01 4.05002266e-01 7.23083690e-02 6.59355104e-01
-3.02198291e-01 7.67135084e-01 -2.76468188e-01 -2.73238093e-01
2.90603876e-01 2.13293695e+00 -1.06943345e+00 -2.92455584e-01
-1.81585953e-01 1.45618093e+00 4.79499608e-01 1.67528048e-01
-6.68606460e-01 -4.94028479e-01 8.03267121e-01 1.81571260e-01
7.38603622e-02 3.86962861e-01 -1.49041265e-01 -1.07252407e+00
-2.96089470e-01 -6.88192487e-01 4.42575365e-01 -8.03239167e-01
-1.60654521e+00 7.03764379e-01 -6.31767988e-01 -7.85263300e-01
-8.83433223e-02 -8.98189366e-01 -6.51487887e-01 8.03866982e-01
-1.13765574e+00 -1.07418776e+00 -2.60076344e-01 3.28470618e-01
1.96460634e-01 3.27479511e-01 1.14842093e+00 2.91927457e-01
-3.61437976e-01 3.47516328e-01 -5.89022577e-01 3.02212596e-01
1.02608919e+00 -1.63895142e+00 -2.72776216e-01 3.67399275e-01
1.99271798e-01 6.38725221e-01 1.24288940e+00 -4.24380898e-01
-1.04287863e+00 -8.55080426e-01 1.56797481e+00 -7.36661911e-01
9.25186157e-01 -4.87739354e-01 -7.52399921e-01 2.66143471e-01
4.45761383e-01 -2.08923504e-01 1.21784484e+00 7.96018541e-01
-3.80779117e-01 3.69264901e-01 -1.12561285e+00 4.26975191e-01
7.18141139e-01 -6.38773024e-01 -5.85993648e-01 2.75264591e-01
4.12933886e-01 -1.31807491e-01 -1.17326307e+00 3.08167450e-02
3.34913999e-01 -9.51590776e-01 5.12549400e-01 -8.73024046e-01
7.60028780e-01 -1.39054000e-01 1.50406644e-01 -1.48824573e+00
5.07960282e-02 -4.66535062e-01 6.19678646e-02 2.00442338e+00
5.28850079e-01 -3.09450150e-01 5.62733352e-01 1.09809089e+00
3.99970599e-02 -6.28043056e-01 -4.92103577e-01 -2.70672441e-01
-1.13360338e-01 -6.74569905e-01 2.86385506e-01 1.53083110e+00
4.74112004e-01 8.76257300e-01 -3.42107922e-01 8.90301615e-02
-1.02661639e-01 1.58787951e-01 5.71481347e-01 -1.49666071e+00
-1.73641309e-01 -5.24534285e-01 -4.06006813e-01 -4.20919299e-01
5.40485144e-01 -9.17848766e-01 1.55255511e-01 -1.51707590e+00
1.86126187e-01 -3.43643069e-01 -2.81072080e-01 5.84492803e-01
-2.53305554e-01 3.59746426e-01 1.39470190e-01 -1.98815674e-01
-9.10139620e-01 1.11561239e-01 4.59440142e-01 2.46424869e-01
-2.11952522e-01 -1.44608140e-01 -9.23866034e-01 8.54267299e-01
9.45105433e-01 -4.13635999e-01 1.08178534e-01 3.86313558e-01
9.17571187e-01 -1.69031456e-01 2.30190441e-01 -6.73086345e-01
6.23180717e-02 1.23415202e-01 3.04853737e-01 -1.39037013e-01
3.50329757e-01 -8.79372597e-01 -1.49137322e-02 1.57200508e-02
-5.05238473e-01 3.77238207e-02 1.71510652e-02 1.52698621e-01
-2.38756314e-01 -5.70800841e-01 6.08741701e-01 -9.88422558e-02
-7.18491673e-01 -1.34450629e-01 -7.56911218e-01 1.71288878e-01
9.48539019e-01 1.38964280e-01 -5.23678102e-02 -4.95749176e-01
-9.81370687e-01 -1.53650656e-01 4.64772463e-01 4.96090263e-01
1.95096314e-01 -7.99524009e-01 -7.08807051e-01 -4.74533856e-01
2.49892533e-01 -7.33773708e-01 1.72276586e-01 9.10994112e-01
6.06680177e-02 -7.23828450e-02 6.13175407e-02 -1.12538813e-02
-1.67956603e+00 5.11319995e-01 1.47352159e-01 -3.42113525e-01
-1.13413081e-01 9.00459409e-01 -6.95160180e-02 -4.75180537e-01
7.07668439e-02 4.32968847e-02 -9.28168714e-01 1.00159812e+00
6.41789615e-01 1.71683669e-01 2.13842481e-01 -1.09053981e+00
-4.07436818e-01 9.59616005e-02 2.99677365e-02 -1.58925638e-01
1.48683727e+00 -1.50117606e-01 -5.39996624e-01 7.70539939e-01
1.21414709e+00 3.47823083e-01 -5.00998616e-01 2.03763992e-01
3.77345204e-01 -2.28161886e-02 4.30617668e-02 -1.03684056e+00
-7.57589579e-01 7.76726007e-01 3.83077532e-01 6.18645012e-01
9.09322023e-01 2.71809131e-01 4.45467234e-01 2.19976604e-01
3.11272754e-03 -1.37383902e+00 -8.10406730e-03 7.80238807e-01
8.30791116e-01 -1.29004824e+00 -2.44550049e-01 -5.79661310e-01
-1.05940831e+00 1.02080595e+00 6.12092316e-01 1.69134662e-01
7.03082442e-01 2.46495605e-01 2.71075964e-01 -6.13973320e-01
-6.79909289e-01 -6.55891299e-01 3.06922227e-01 1.77060962e-01
1.06604588e+00 2.08370104e-01 -6.54248059e-01 1.10328770e+00
-4.37286168e-01 -2.98921019e-01 4.93699700e-01 6.50322676e-01
-2.26923645e-01 -8.71879160e-01 -2.35330299e-01 5.63224137e-01
-1.15524077e+00 -1.22200064e-01 -9.78372633e-01 2.78715014e-01
2.01384142e-01 1.31843412e+00 7.71940574e-02 -3.39643925e-01
2.83871025e-01 4.37866926e-01 4.91811819e-02 -9.60832357e-01
-1.33755708e+00 -2.93994188e-01 6.53104365e-01 -1.69321328e-01
-7.27611303e-01 -6.94576204e-01 -1.44681787e+00 -6.44544959e-02
-2.99525052e-01 5.62110484e-01 8.38157356e-01 1.12919664e+00
2.47256264e-01 4.77015346e-01 5.32713771e-01 -6.33673131e-01
1.44642889e-02 -1.11585367e+00 -6.54101789e-01 9.47922051e-01
-6.29453957e-02 -4.11096483e-01 -5.82326353e-01 1.52717963e-01] | [12.662830352783203, 6.340439319610596] |
8a1251e6-3e65-4910-8fbf-35dbc785870c | multilingual-named-entity-recognition-using-1 | 1906.09978 | null | https://arxiv.org/abs/1906.09978v1 | https://arxiv.org/pdf/1906.09978v1.pdf | Multilingual Named Entity Recognition Using Pretrained Embeddings, Attention Mechanism and NCRF | In this paper we tackle multilingual named entity recognition task. We use the BERT Language Model as embeddings with bidirectional recurrent network, attention, and NCRF on the top. We apply multilingual BERT only as embedder without any fine-tuning. We test out model on the dataset of the BSNLP shared task, which consists of texts in Bulgarian, Czech, Polish and Russian languages. | ['Ekaterina Artemova', 'Anton A. Emelyanov'] | 2019-06-21 | multilingual-named-entity-recognition-using-2 | https://aclanthology.org/W19-3713 | https://aclanthology.org/W19-3713.pdf | ws-2019-8 | ['multilingual-text-classification', 'joint-ner-and-classification', 'multilingual-named-entity-recognition'] | ['miscellaneous', 'natural-language-processing', 'natural-language-processing'] | [-7.16090679e-01 1.29895866e-01 -1.86724126e-01 -2.89328873e-01
-6.31783843e-01 -7.34163523e-01 7.21949637e-01 1.93463769e-02
-1.21675241e+00 9.64164019e-01 7.07473218e-01 -7.90665627e-01
3.30448329e-01 -4.64029968e-01 -7.41449058e-01 -3.27353209e-01
3.40918377e-02 7.06949115e-01 7.53250718e-02 -1.95308506e-01
-3.31645370e-01 5.33924580e-01 -3.09603423e-01 2.04921693e-01
6.08870745e-01 2.88927257e-01 -5.69210562e-04 6.09919488e-01
-1.86545640e-01 1.10159874e+00 -4.12782162e-01 -9.98252153e-01
7.33505264e-02 1.46039456e-01 -9.66844678e-01 -6.22703433e-01
-9.36853960e-02 4.27451059e-02 -6.22448802e-01 6.78522050e-01
8.51917922e-01 2.74104714e-01 5.62282681e-01 -8.03248703e-01
-9.90351319e-01 1.55075908e+00 -1.94553509e-01 4.16532159e-01
1.64659426e-03 -4.73716343e-03 1.49465728e+00 -1.19439590e+00
1.19562674e+00 1.26297390e+00 7.13396549e-01 4.93348390e-01
-1.21366072e+00 -5.79407275e-01 2.22846374e-01 4.50026542e-01
-1.59859741e+00 -6.16069674e-01 3.40042681e-01 -2.09148735e-01
1.91171706e+00 -2.03728557e-01 4.32030380e-01 1.35743332e+00
1.00849815e-01 1.14941478e+00 5.74253738e-01 -5.14059961e-01
-1.43286511e-01 2.97664821e-01 4.29092318e-01 3.77660185e-01
5.75889535e-02 1.22227974e-01 -1.82008162e-01 -6.86003789e-02
3.89815480e-01 -2.68388420e-01 -2.11411878e-01 -1.06008530e-01
-1.18448699e+00 8.37203801e-01 4.52668220e-01 7.24606633e-01
-5.71797609e-01 1.50467053e-01 6.31389618e-01 4.61763889e-01
6.80452764e-01 5.09867609e-01 -1.28901041e+00 -1.28240988e-01
-5.73723257e-01 -2.40989521e-01 1.12099814e+00 1.01977980e+00
5.74723363e-01 1.46367252e-01 -2.21460918e-03 1.21328461e+00
3.10561478e-01 4.32812572e-01 7.09150016e-01 -3.48444432e-02
6.59213483e-01 3.68491441e-01 -1.86115578e-01 -3.77049208e-01
-5.86134672e-01 -4.18926060e-01 -5.74878335e-01 -4.69051689e-01
-9.51396301e-02 -7.42862821e-01 -8.63231301e-01 1.63785160e+00
2.68054396e-01 3.62030804e-01 6.19270205e-01 5.50174057e-01
1.04504323e+00 9.68525052e-01 2.14743212e-01 1.57746136e-01
1.46520627e+00 -1.48573792e+00 -9.29878116e-01 -1.04120255e-01
9.89947438e-01 -6.53007805e-01 6.68477893e-01 1.11708887e-01
-8.81208122e-01 -2.68930584e-01 -6.84679985e-01 -6.51036620e-01
-9.55542803e-01 4.52361315e-01 2.27839306e-01 2.94452071e-01
-1.12149143e+00 2.84176111e-01 -9.74882007e-01 -4.00124818e-01
-1.54099450e-01 1.36854500e-01 -7.53922462e-01 3.87417860e-02
-1.68872702e+00 1.21967030e+00 9.61292088e-01 3.86586368e-01
-6.88687563e-01 -7.76210606e-01 -1.19798934e+00 -2.35448834e-02
-4.29036655e-02 -3.82919401e-01 9.99362290e-01 -4.29626644e-01
-1.65128005e+00 9.38670814e-01 -1.30325323e-02 -7.17943072e-01
3.05702269e-01 -5.02159953e-01 -8.24743330e-01 -4.30803984e-01
-6.20586574e-02 5.82358599e-01 -1.08749308e-01 -6.08157635e-01
-3.19998562e-01 -2.58192182e-01 -1.74629718e-01 1.18390515e-01
-4.65065353e-02 4.78043586e-01 -5.34074605e-01 -6.84277117e-01
-5.21468580e-01 -8.56468737e-01 -4.36154723e-01 -1.04520285e+00
-6.45207644e-01 -6.85757041e-01 5.40680587e-01 -1.26353955e+00
1.26181853e+00 -2.33872080e+00 4.04811680e-01 -1.44630885e-02
-3.23963642e-01 3.36751521e-01 -5.58423877e-01 8.06010187e-01
-2.01605588e-01 3.68626952e-01 1.41724244e-01 -4.91881579e-01
1.72703207e-01 4.64594990e-01 -2.86101282e-01 5.65351248e-01
6.36631489e-01 1.16030359e+00 -6.90796733e-01 -1.21543407e-01
7.20165670e-02 8.46659422e-01 -4.69506949e-01 3.44971687e-01
-5.96002936e-02 3.28599840e-01 -1.45108774e-01 1.18478626e-01
6.61169112e-01 3.06435049e-01 4.66531068e-01 -4.45532277e-02
-5.00270545e-01 9.86200094e-01 -9.31993365e-01 1.81286359e+00
-1.06389594e+00 4.71182019e-01 5.95263168e-02 -6.48768425e-01
8.11167061e-01 7.24990547e-01 -5.63550517e-02 -5.30050159e-01
4.64427657e-02 4.81097072e-01 -6.97071925e-02 -2.40135908e-01
5.57547212e-01 -1.61688969e-01 -4.12666500e-01 2.20464781e-01
8.15897524e-01 1.96819544e-01 1.02360867e-01 7.79835880e-02
1.20325232e+00 4.40695249e-02 6.76353097e-01 -1.84996024e-01
5.40958226e-01 -1.49525687e-01 6.63332939e-01 4.07045543e-01
-4.07563038e-02 4.11171317e-01 6.92079663e-01 -3.84836793e-01
-1.04523158e+00 -9.41240728e-01 -3.03707391e-01 1.59587502e+00
-5.44555306e-01 -3.41363639e-01 -1.53539047e-01 -1.01248789e+00
-6.97788820e-02 9.90503013e-01 -5.69194913e-01 3.45028400e-01
-1.05436563e+00 -5.76351464e-01 1.08608556e+00 4.33092386e-01
1.29758447e-01 -1.67276025e+00 3.20783794e-01 5.59472561e-01
5.16049117e-02 -1.54539084e+00 -4.86576647e-01 7.09805369e-01
-3.81146550e-01 -7.67035425e-01 -9.62717474e-01 -1.05052233e+00
-7.32018054e-02 -6.58048213e-01 1.30428171e+00 -5.33343256e-01
6.20021373e-02 -3.05168107e-02 -3.62545580e-01 -3.02630097e-01
-1.75841480e-01 9.76167738e-01 -2.86652356e-01 -2.52225697e-01
3.21327209e-01 -3.62745345e-01 -8.19297507e-02 -1.41128540e-01
-6.04909122e-01 -4.38713819e-01 5.81686199e-01 9.92873967e-01
4.00615275e-01 -5.22745252e-01 5.49170494e-01 -1.31420135e+00
3.35927814e-01 -7.62115419e-01 -6.87509537e-01 4.18234348e-01
-4.93143648e-02 3.12550157e-01 5.73297501e-01 -1.62002683e-01
-1.17115855e+00 -1.05497502e-01 -7.90095091e-01 -1.36434317e-01
-2.59382904e-01 7.37039626e-01 -5.39666355e-01 5.15486956e-01
3.77061784e-01 6.22858778e-02 -1.02502775e+00 -1.01229739e+00
9.65805411e-01 7.28462934e-01 3.04597914e-01 -2.24093571e-01
3.60468805e-01 -8.39640871e-02 -7.91188896e-01 -1.03122044e+00
-5.93469024e-01 -7.79519081e-01 -8.70192349e-01 4.62553680e-01
1.19669366e+00 -1.24063897e+00 -2.93589652e-01 1.73010439e-01
-1.77498055e+00 -4.77593482e-01 -2.02404901e-01 7.62550890e-01
-8.05488527e-02 -1.58485696e-01 -1.34795439e+00 -6.81408465e-01
-5.06305456e-01 -8.59300196e-01 8.25236082e-01 -1.53044522e-01
1.11536190e-01 -1.46927536e+00 7.10723162e-01 -2.24575087e-01
4.38627422e-01 -2.56960064e-01 7.65903175e-01 -1.42588019e+00
-6.96393102e-02 5.00676632e-02 -2.61737138e-01 4.95015502e-01
-2.56935924e-01 -1.33100748e-01 -9.77877080e-01 -3.36157262e-01
-5.33278406e-01 -3.21317732e-01 1.24550617e+00 -4.28517237e-02
3.39349389e-01 -4.29442257e-01 -2.78878510e-01 7.57875562e-01
1.55760622e+00 1.85931653e-01 3.54388416e-01 3.85518223e-01
8.72940242e-01 4.53126013e-01 -1.50321722e-01 1.79069519e-01
6.16510212e-01 4.20529395e-01 -2.59172022e-02 -1.91813827e-01
-9.13731307e-02 -2.86250472e-01 7.43010283e-01 1.66637421e+00
2.03689575e-01 -5.25096357e-01 -1.44397712e+00 8.98744345e-01
-1.61684918e+00 -5.88695407e-01 2.88058259e-02 1.68483603e+00
8.79902720e-01 -2.29810223e-01 -2.11273819e-01 -6.68427706e-01
6.50853097e-01 3.36950779e-01 -1.13837868e-01 -8.01370382e-01
-4.04225707e-01 5.61141312e-01 6.88920677e-01 6.10454559e-01
-1.35513997e+00 1.59570515e+00 5.80472374e+00 6.72438622e-01
-1.23945045e+00 7.63098359e-01 2.46948004e-01 9.22022909e-02
-3.36003333e-01 5.20709087e-04 -1.15484285e+00 1.14977591e-01
1.60233045e+00 -6.32825270e-02 3.15028787e-01 6.40349746e-01
-1.91731125e-01 6.83739543e-01 -9.58521605e-01 5.53181887e-01
-6.40233904e-02 -1.00410628e+00 -1.45768777e-01 -1.20366409e-01
6.93847299e-01 1.24037445e+00 -4.07573968e-01 9.39167976e-01
1.02331913e+00 -9.04022574e-01 5.82403421e-01 2.84315437e-01
7.19717145e-01 -1.01682913e+00 1.19318533e+00 2.84063548e-01
-1.09232831e+00 3.15424472e-01 -5.51160514e-01 5.18016100e-01
4.94488984e-01 5.44620812e-01 -1.02224100e+00 7.77131259e-01
2.98511565e-01 9.26011264e-01 -4.20702726e-01 9.55713570e-01
-8.29008639e-01 1.10904288e+00 -3.71082634e-01 -8.13494325e-02
5.41189432e-01 -9.33897123e-02 5.62655449e-01 1.92585754e+00
5.25313020e-02 -3.72460008e-01 1.61935881e-01 3.31910729e-01
-5.44060111e-01 7.05855668e-01 -8.08188081e-01 -2.55679905e-01
1.51620030e-01 1.17140436e+00 -3.04266214e-01 -4.32636261e-01
-4.64373738e-01 1.10909212e+00 1.11636162e+00 6.38191283e-01
-5.46153367e-01 -6.19494200e-01 5.45870245e-01 -4.59321737e-01
6.14583671e-01 -3.46494943e-01 2.42276862e-01 -1.62881172e+00
-3.98149222e-01 -5.13597786e-01 5.48123777e-01 -4.39783007e-01
-1.55845845e+00 1.01699257e+00 -5.08363843e-01 -3.28417152e-01
-2.99555421e-01 -9.77693856e-01 -4.64155138e-01 9.32483673e-01
-1.74562335e+00 -1.28209841e+00 7.63159990e-01 3.52616310e-01
3.36893022e-01 -1.72126800e-01 8.62874448e-01 6.99937224e-01
-1.00005245e+00 6.06638193e-01 3.45779836e-01 9.13044095e-01
9.85322237e-01 -1.43610597e+00 1.03158295e+00 4.98174191e-01
5.46872973e-01 6.30414009e-01 1.88623786e-01 -4.23091412e-01
-9.67437267e-01 -1.42469692e+00 1.87153852e+00 -3.90812606e-01
1.36219501e+00 -9.11314547e-01 -9.23556089e-01 1.28176939e+00
9.63843942e-01 3.08435917e-01 7.20407784e-01 6.45521283e-01
-5.21399856e-01 3.24112356e-01 -5.95316589e-01 3.16829681e-01
8.35122824e-01 -7.64363050e-01 -7.93860555e-01 2.55494922e-01
1.26906586e+00 -1.98429987e-01 -1.08971417e+00 1.52635083e-01
4.24298257e-01 -3.05501044e-01 5.97885847e-01 -1.12763262e+00
1.08413734e-01 2.42913857e-01 -1.40182361e-01 -1.86995196e+00
-2.83137590e-01 -2.82491446e-01 4.47848916e-01 1.70345497e+00
1.24499059e+00 -8.89405608e-01 1.83478430e-01 -9.90603641e-02
-1.96566001e-01 -3.65696639e-01 -1.22730803e+00 -8.25757861e-01
6.80506289e-01 -6.92052960e-01 6.01744890e-01 1.30958652e+00
-5.59973298e-03 8.40590954e-01 -2.53158182e-01 1.33222062e-02
1.09829828e-02 -4.66412485e-01 4.10750836e-01 -8.53390098e-01
-3.11715811e-01 -2.56159306e-01 -3.96449387e-01 -6.95394397e-01
9.24249053e-01 -1.30557060e+00 1.02681801e-01 -1.58214962e+00
3.09634265e-02 -4.57895130e-01 -9.22312498e-01 7.98137009e-01
-3.25848758e-02 1.54502681e-02 3.61800462e-01 -1.55765414e-01
-6.92565382e-01 7.80680418e-01 3.49650085e-01 -1.49701700e-01
-1.34411648e-01 -4.18441683e-01 -1.63178414e-01 4.94243056e-01
7.67025352e-01 -8.68877947e-01 3.30867678e-01 -9.93645072e-01
5.58067203e-01 2.53156036e-01 -2.01947391e-01 -5.35510719e-01
9.30847302e-02 3.29328537e-01 6.80129677e-02 -5.74294686e-01
5.36513440e-02 -7.09474027e-01 -6.79008737e-02 2.64449924e-01
-4.08724189e-01 5.68988681e-01 3.74239236e-01 3.60436827e-01
-4.73526180e-01 -1.41560122e-01 4.73825276e-01 2.72912215e-02
-6.12760246e-01 2.52148539e-01 -4.77984875e-01 3.55868071e-01
6.02344036e-01 7.67034948e-01 -3.59130383e-01 1.44176826e-01
-1.15170813e+00 4.33651060e-01 6.16427846e-02 6.60696149e-01
1.36688873e-01 -1.44997847e+00 -1.21879542e+00 4.02953684e-01
1.57278076e-01 -5.12618899e-01 8.31784531e-02 8.23281169e-01
-5.10867536e-01 7.18425632e-01 1.69154987e-01 -1.08304210e-02
-7.75426090e-01 5.89492321e-01 2.23098412e-01 -9.64594424e-01
-3.98696631e-01 8.92042875e-01 6.77827373e-02 -1.67518485e+00
-3.58680673e-02 -1.94536030e-01 -5.28568685e-01 2.78680325e-01
1.21185116e-01 4.10145640e-01 3.76709670e-01 -8.42078805e-01
-8.49635065e-01 1.85892507e-01 -1.57438457e-01 -5.70148826e-01
1.49567735e+00 3.43470909e-02 -2.73099661e-01 8.60438764e-01
1.72575378e+00 4.44608361e-01 -2.94071406e-01 -2.92228103e-01
6.30605161e-01 5.57490766e-01 2.01203171e-02 -8.30711186e-01
-1.07995713e+00 1.01956010e+00 1.70934126e-01 -1.58573076e-01
5.73957980e-01 1.94527656e-01 7.37039506e-01 6.06224000e-01
2.53059447e-01 -8.41137648e-01 -8.64536583e-01 1.42806268e+00
7.71118879e-01 -1.02780938e+00 -5.01821518e-01 2.45076150e-01
-8.25552583e-01 8.76116991e-01 3.14992666e-01 -5.33850491e-01
1.08816218e+00 4.87935841e-01 3.92124236e-01 -5.44310501e-03
-1.38281846e+00 -2.80117542e-01 2.99337983e-01 1.41380712e-01
9.67870176e-01 2.44318619e-01 -6.35168791e-01 6.10263169e-01
-2.03910291e-01 -2.11470142e-01 3.78899693e-01 6.07982755e-01
1.22751363e-01 -1.17532492e+00 1.56736061e-01 4.34292853e-02
-1.01507759e+00 -7.71050692e-01 -3.72778773e-01 1.01017082e+00
1.45417765e-01 7.54938841e-01 1.45005047e-01 -1.11671180e-01
6.39494240e-01 4.92815286e-01 -2.27890551e-01 -6.72078311e-01
-1.14513874e+00 1.00504421e-01 5.02573252e-01 -4.85931903e-01
-3.13165714e-04 -5.67360640e-01 -1.22989821e+00 6.78910837e-02
-4.52276707e-01 6.25056446e-01 7.88033009e-01 6.90726280e-01
5.36522865e-01 4.71968621e-01 3.24432611e-01 -4.11883324e-01
-1.70158416e-01 -1.41081405e+00 -6.47771001e-01 1.15935445e-01
2.75760800e-01 -8.15899149e-02 -2.68504024e-01 -3.46174508e-01] | [9.891218185424805, 9.775591850280762] |
9153510e-f19b-4399-9667-f079c8102303 | towards-offensive-language-identification-for | null | null | https://aclanthology.org/2021.dravidianlangtech-1.3 | https://aclanthology.org/2021.dravidianlangtech-1.3.pdf | Towards Offensive Language Identification for Dravidian Languages | Offensive speech identification in countries like India poses several challenges due to the usage of code-mixed and romanized variants of multiple languages by the users in their posts on social media. The challenge of offensive language identification on social media for Dravidian languages is harder, considering the low resources available for the same. In this paper, we explored the zero-shot learning and few-shot learning paradigms based on multilingual language models for offensive speech detection in code-mixed and romanized variants of three Dravidian languages - Malayalam, Tamil, and Kannada. We propose a novel and flexible approach of selective translation and transliteration to reap better results from fine-tuning and ensembling multilingual transformer networks like XLMRoBERTa and mBERT. We implemented pretrained, fine-tuned, and ensembled versions of XLM-RoBERTa for offensive speech classification. Further, we experimented with interlanguage, inter-task, and multi-task transfer learning techniques to leverage the rich resources available for offensive speech identification in the English language and to enrich the models with knowledge transfer from related tasks. The proposed models yielded good results and are promising for effective offensive speech identification in low resource settings. | ['Yashvardhan Sharma', 'Siva Sai'] | null | null | null | null | eacl-dravidianlangtech-2021-4 | ['transliteration'] | ['natural-language-processing'] | [-1.68366060e-01 -2.63682336e-01 -2.96123058e-01 -7.74507001e-02
-1.37276697e+00 -9.48036790e-01 5.88805079e-01 -3.22044700e-01
-5.06009400e-01 3.21705371e-01 4.86792326e-01 -7.58162856e-01
4.07985486e-02 -2.34562844e-01 -1.89533979e-01 -3.47289234e-01
-3.68932746e-02 4.76917744e-01 -2.51182407e-01 -6.16442680e-01
5.18898759e-03 2.24287629e-01 -8.81872058e-01 6.22375786e-01
1.06076109e+00 4.35777843e-01 2.64790028e-01 8.17374408e-01
-7.85025656e-02 9.98837888e-01 -5.30747950e-01 -9.27632093e-01
2.43219271e-01 -2.25809589e-01 -9.29819226e-01 -3.83352846e-01
3.07726115e-01 -3.17992568e-01 -3.78526151e-01 1.05713844e+00
1.00076783e+00 3.12086008e-02 5.65720081e-01 -7.81812131e-01
-9.58544075e-01 1.24198151e+00 -6.27236903e-01 4.80127573e-01
4.82511014e-01 9.48670506e-02 9.14817870e-01 -9.40781891e-01
3.99563402e-01 1.46486712e+00 8.62127304e-01 5.63508153e-01
-8.89744818e-01 -1.15227830e+00 -1.16498575e-01 3.40822488e-01
-1.47624183e+00 -8.80943000e-01 6.68555856e-01 -7.08627403e-01
1.27358711e+00 2.94760406e-01 9.77196544e-02 1.86882007e+00
-3.09063315e-01 5.70495844e-01 1.10954082e+00 -4.77142572e-01
-3.28968406e-01 6.29019439e-01 -9.27406400e-02 7.33480155e-01
-4.77492481e-01 2.43662491e-01 -6.86506510e-01 -3.47243577e-01
1.04659237e-02 -3.73328865e-01 3.19602937e-02 3.85933638e-01
-1.06033385e+00 1.17309618e+00 -2.15525597e-01 6.99383855e-01
1.76651299e-01 -6.50421143e-01 9.87695336e-01 6.26278102e-01
8.32184732e-01 6.45075917e-01 -5.94421744e-01 -6.06029153e-01
-8.69559228e-01 -5.66738188e-01 9.06912625e-01 8.53113115e-01
3.76877695e-01 4.14031118e-01 2.01953035e-02 1.73059511e+00
3.06929406e-02 7.33345151e-01 1.04895175e+00 -2.87250787e-01
9.26199734e-01 1.48930606e-02 -4.61559147e-01 -7.39519358e-01
-1.32783443e-01 -3.72268736e-01 -3.75493497e-01 -2.29355305e-01
1.51023403e-01 -6.84741139e-01 -6.88743353e-01 1.57911229e+00
-3.15896571e-02 -2.26818025e-02 3.17429125e-01 4.99458969e-01
6.60860360e-01 8.09272885e-01 1.88572332e-01 4.75217178e-02
1.35529149e+00 -1.08719254e+00 -4.15863484e-01 -1.72736511e-01
9.05727327e-01 -1.13353419e+00 1.49596488e+00 1.89872831e-01
-6.67886555e-01 -5.31127453e-01 -8.48333240e-01 9.95971560e-02
-6.25923812e-01 4.27926034e-02 2.68750578e-01 1.15264308e+00
-7.63908207e-01 1.42760321e-01 -3.88656020e-01 -5.11213481e-01
2.30855897e-01 1.13340706e-01 -5.57444811e-01 2.41920337e-01
-1.44862413e+00 1.10582471e+00 2.64831185e-01 -4.73324031e-01
-1.00949025e+00 -6.21839941e-01 -8.37618172e-01 -8.79635140e-02
1.19548291e-01 3.39966059e-01 1.12782371e+00 -1.09573400e+00
-1.61916888e+00 1.18840647e+00 4.67935830e-01 -3.32084268e-01
3.43225896e-01 -1.83652967e-01 -1.08836436e+00 -1.22153938e-01
1.82810739e-01 2.16167364e-02 8.60330701e-01 -5.22368550e-01
-5.21285295e-01 -3.77845645e-01 -1.16399579e-01 1.94262624e-01
-8.83296669e-01 9.26671743e-01 6.46668226e-02 -8.89873266e-01
-6.05066299e-01 -9.12772477e-01 3.79366934e-01 -7.91177332e-01
-4.79398400e-01 2.84054819e-02 8.55774403e-01 -1.35633099e+00
1.25481939e+00 -2.41416788e+00 7.90161490e-02 -5.87862404e-03
-2.18492940e-01 6.61520183e-01 -3.39263856e-01 6.27500057e-01
-8.96341205e-02 2.92712569e-01 2.38609403e-01 -1.69898376e-01
-6.84384331e-02 -1.73461631e-01 -4.53241646e-01 4.79966760e-01
-5.90333454e-02 5.64873755e-01 -8.43942106e-01 -4.31394070e-01
1.78139001e-01 4.94392812e-01 -3.48339975e-01 1.45582795e-01
4.05147403e-01 4.92875427e-01 -1.34023935e-01 8.03909838e-01
3.41369063e-01 4.96995121e-01 1.82767093e-01 1.76691532e-01
-3.70884329e-01 5.42080462e-01 -4.66035098e-01 1.52106464e+00
-1.07137370e+00 5.98056376e-01 3.50524008e-01 -6.12494528e-01
7.78899014e-01 5.02480507e-01 2.02510193e-01 -8.69108498e-01
3.57227474e-01 5.78669667e-01 3.69133621e-01 -6.39142275e-01
2.73334861e-01 -2.96195954e-01 -3.83465886e-01 3.92373443e-01
4.64574754e-01 4.34626732e-03 -1.23437509e-01 -7.03153387e-02
9.09269869e-01 -2.50208586e-01 3.92557591e-01 -2.86222249e-01
6.74287200e-01 -3.03827763e-01 2.55483806e-01 5.67462921e-01
-4.92990583e-01 3.15391541e-01 1.63132429e-01 -1.01376340e-01
-9.52848196e-01 -9.35415208e-01 -2.48747468e-01 2.02566528e+00
-5.42806923e-01 -4.33895260e-01 -6.53608441e-01 -7.16248512e-01
-1.84730649e-01 8.59344602e-01 -2.60928333e-01 -2.53647298e-01
-5.43949842e-01 -6.78901613e-01 1.47213805e+00 1.84401140e-01
3.22371781e-01 -8.50763202e-01 1.15212515e-01 2.17610627e-01
-3.68496120e-01 -1.26204240e+00 -9.12916660e-01 3.31850886e-01
3.59556861e-02 -8.22798789e-01 -7.78087378e-01 -9.83972311e-01
-1.80533066e-01 -6.37727790e-03 6.32625341e-01 -5.09444475e-01
-1.16380155e-01 1.46692619e-01 -5.33108473e-01 -4.17796314e-01
-1.06445909e+00 4.60882276e-01 3.70368987e-01 2.31453642e-01
3.83282632e-01 -4.06498998e-01 3.30092907e-01 2.96148539e-01
-2.89627433e-01 -2.11633787e-01 3.82604539e-01 7.69071281e-01
-2.11262792e-01 -4.10361856e-01 5.88139892e-01 -8.53396714e-01
7.88027108e-01 -7.78545439e-01 -3.46206307e-01 3.67065936e-01
-1.26255110e-01 -1.10511512e-01 8.95821929e-01 -8.72504294e-01
-1.11564481e+00 -2.28117168e-01 -5.95475018e-01 -4.67152387e-01
5.94070032e-02 2.79744357e-01 -1.21950738e-01 -1.85504898e-01
8.04841578e-01 1.19296402e-01 -1.63147554e-01 -6.10938668e-01
4.46515322e-01 1.54348528e+00 4.26878899e-01 -3.77669424e-01
8.22747827e-01 -1.15698427e-01 -1.04271662e+00 -1.16404474e+00
-6.47406697e-01 -7.88633049e-01 -5.48626781e-01 -1.43959280e-03
1.13122761e+00 -1.23474824e+00 -4.37495023e-01 7.70494401e-01
-1.07782578e+00 -2.43453324e-01 1.76051468e-01 6.09481692e-01
-1.66071132e-01 3.76275480e-01 -8.66135418e-01 -8.20881784e-01
-6.04582846e-01 -1.20697451e+00 8.00999939e-01 -4.35372859e-01
-4.55805063e-01 -1.07447720e+00 5.03997266e-01 6.21773541e-01
7.00874031e-01 -3.56776148e-01 1.10223854e+00 -1.37619603e+00
4.56411898e-01 -1.53980508e-01 -1.96373076e-04 6.61129475e-01
2.45700881e-01 -7.40783215e-02 -1.20869732e+00 -3.72877598e-01
-7.47933537e-02 -7.91363657e-01 5.11345744e-01 -1.93407536e-01
7.42224872e-01 -5.83505213e-01 -5.10395551e-03 7.14029193e-01
7.96008945e-01 2.96850711e-01 6.55787289e-02 1.82941303e-01
9.30540919e-01 6.38907135e-01 1.87448978e-01 4.50802386e-01
2.79847890e-01 8.60227108e-01 -1.48584947e-01 -9.61597543e-03
-2.16256469e-01 -4.43787277e-01 9.38581288e-01 1.59894383e+00
1.91481546e-01 4.24249358e-02 -1.32197189e+00 6.13157094e-01
-1.11420560e+00 -1.13894808e+00 4.49383527e-01 2.06494594e+00
1.06039298e+00 -1.72817245e-01 3.88729542e-01 -2.67301053e-01
9.62347984e-01 2.27872580e-01 -2.52259582e-01 -8.76511455e-01
-3.08425933e-01 2.16300532e-01 5.29652596e-01 6.55976593e-01
-1.37207198e+00 1.34361804e+00 5.83446598e+00 1.24217248e+00
-1.42126727e+00 8.24742317e-01 5.33536792e-01 -2.47042805e-01
6.07098080e-02 -2.75373369e-01 -1.04554677e+00 6.62064910e-01
1.67949665e+00 -2.97849089e-01 7.81643033e-01 9.46827292e-01
3.00631002e-02 7.35004306e-01 -8.38836432e-01 1.15710664e+00
4.78123665e-01 -8.27522337e-01 -1.94510832e-01 4.45507839e-02
7.65942335e-01 8.28244269e-01 3.82779181e-01 8.63797784e-01
6.64718688e-01 -9.71055448e-01 7.79949009e-01 -3.82198572e-01
1.11985552e+00 -8.94480824e-01 4.96848851e-01 3.96127909e-01
-9.08947825e-01 -6.01145387e-01 -2.88885385e-01 2.70068407e-01
-1.81469068e-01 -9.27924067e-02 -1.17551148e+00 2.76715875e-01
5.61649203e-01 5.03105879e-01 -5.25507808e-01 4.23892200e-01
-9.49902553e-03 8.01631689e-01 -1.15243308e-01 -1.24059066e-01
3.80241513e-01 2.26436183e-03 6.33071125e-01 1.51470304e+00
3.99723887e-01 -5.22221923e-01 2.85920113e-01 5.05754948e-01
-1.84170976e-01 7.50446975e-01 -9.16177392e-01 -4.76635635e-01
4.98641968e-01 1.28538144e+00 -1.00790814e-01 -2.09193900e-01
-7.70732880e-01 9.86479223e-01 5.62425137e-01 8.35073888e-02
-8.82082880e-01 -4.42515612e-01 7.70595431e-01 -3.21340673e-02
-2.49430779e-02 -1.41662672e-01 6.35291338e-02 -1.41370702e+00
-4.14895147e-01 -1.56829917e+00 5.36733985e-01 -3.19514245e-01
-1.46170449e+00 1.06320834e+00 -2.02939093e-01 -1.18974030e+00
-4.97817338e-01 -6.45388782e-01 -6.79123938e-01 1.01929367e+00
-1.16130328e+00 -1.59242332e+00 3.53682816e-01 8.44389677e-01
9.36221123e-01 -1.10857022e+00 1.11933851e+00 6.83191955e-01
-7.29863107e-01 1.06086957e+00 2.78495699e-01 6.01852477e-01
9.36150670e-01 -8.37427378e-01 5.02236605e-01 8.94599676e-01
3.23321521e-01 5.27399302e-01 3.21328193e-01 -6.24912977e-01
-1.17244160e+00 -1.01151943e+00 8.35723639e-01 -3.80897522e-01
1.43836713e+00 -8.31635952e-01 -6.29053235e-01 6.66915536e-01
3.81605566e-01 -4.59250599e-01 1.20334768e+00 4.70609218e-01
-8.93820345e-01 1.35094628e-01 -9.31173265e-01 5.66921771e-01
8.28778565e-01 -1.37313807e+00 -5.21794558e-01 5.09067059e-01
6.96768939e-01 9.07660648e-02 -7.82837987e-01 -1.59311026e-01
5.60842693e-01 -4.19441819e-01 8.51419270e-01 -7.99015462e-01
2.01321155e-01 4.59275663e-01 -4.80390638e-01 -1.61886573e+00
-1.56813055e-01 -7.59239137e-01 2.99567282e-01 1.58147824e+00
7.84184694e-01 -6.97085977e-01 2.59506218e-02 9.85355526e-02
-3.34612727e-01 -8.75797793e-02 -1.22186327e+00 -1.01652122e+00
5.05564213e-01 -4.49808389e-01 2.70698428e-01 1.48843300e+00
3.87216687e-01 5.78088522e-01 -7.55306423e-01 -6.19658194e-02
2.71796286e-01 -3.61206353e-01 5.46060085e-01 -8.82717967e-01
-5.63045323e-01 -3.51407439e-01 -2.39561647e-01 -2.91512609e-01
9.02145445e-01 -1.56830347e+00 -2.89004505e-01 -5.79195797e-01
2.58979440e-01 -2.69448757e-01 -2.24029273e-01 5.83624363e-01
1.25371367e-01 2.80094892e-01 2.42479593e-01 1.49568990e-01
-2.45454043e-01 3.81939888e-01 5.75315773e-01 -3.66491407e-01
-2.02854171e-01 -2.64239684e-03 -7.96753943e-01 6.42471075e-01
7.77157724e-01 -4.29844528e-01 -2.84349710e-01 -4.24132556e-01
-1.23656586e-01 2.07211692e-02 -1.04052126e-01 -7.93026149e-01
-4.54049446e-02 1.28117204e-01 -2.94447877e-02 -9.57062095e-02
4.18844134e-01 -4.95080173e-01 -2.47222424e-01 3.27345550e-01
-4.02290374e-01 1.04004316e-01 2.61107534e-01 -4.29305285e-02
-1.39645487e-01 -1.59523189e-01 1.13810050e+00 -2.54527301e-01
-7.10088730e-01 2.27949038e-01 -7.12599218e-01 3.89895111e-01
8.26417208e-01 -1.52123403e-02 -4.07146186e-01 -2.94834316e-01
-5.59794247e-01 -2.09693506e-01 1.20367952e-01 1.03504097e+00
2.29893014e-01 -1.10162890e+00 -1.23358476e+00 5.56205571e-01
4.48200732e-01 -1.34437776e+00 1.64187446e-01 8.72744203e-01
-2.85938501e-01 4.18877453e-01 -4.84103382e-01 -1.09850101e-01
-1.43404317e+00 4.96427119e-01 4.85949039e-01 -1.68632030e-01
-2.77879030e-01 9.01394367e-01 6.85966313e-02 -1.21976650e+00
2.11409613e-01 4.23055887e-01 -2.51200557e-01 4.22163159e-01
4.88783807e-01 4.89440620e-01 1.33055910e-01 -1.24601126e+00
-5.02110600e-01 3.08116704e-01 -3.17956567e-01 -3.38619769e-01
1.12011921e+00 -2.11307406e-01 3.97054255e-02 5.56605697e-01
1.56412852e+00 7.26037562e-01 -4.31589574e-01 -1.58491671e-01
6.17675632e-02 -2.93086201e-01 4.50798385e-02 -9.85923707e-01
-7.26403952e-01 9.45343316e-01 5.67699671e-01 -3.13685946e-02
6.77766621e-01 5.73284887e-02 7.97140002e-01 1.29723474e-01
3.93580824e-01 -1.42024577e+00 -1.17651768e-01 8.91792715e-01
7.99061477e-01 -1.25970674e+00 -7.31916964e-01 -5.84523343e-02
-9.40021217e-01 8.96589160e-01 4.81515497e-01 4.51102614e-01
7.31823027e-01 4.15191203e-01 4.75365937e-01 1.84305072e-01
-5.65426469e-01 -7.71894902e-02 2.37783417e-01 6.61879361e-01
7.36955583e-01 2.92731285e-01 3.41528319e-02 7.30732083e-01
-6.05769157e-01 -7.95339704e-01 3.61005127e-01 4.27598417e-01
-1.11228988e-01 -1.10499895e+00 -4.06515032e-01 3.39439720e-01
-1.03917074e+00 -7.39027262e-01 -6.97269917e-01 5.52903712e-01
4.76541929e-02 1.29368472e+00 -1.59204975e-01 -8.64381492e-01
1.21123172e-01 3.42195034e-01 1.88096222e-02 -8.64728212e-01
-1.17193627e+00 1.58161998e-01 5.09977102e-01 -9.62159261e-02
2.02340841e-01 -6.53290629e-01 -6.61943436e-01 -3.65972489e-01
-3.74891132e-01 1.54301494e-01 7.31082797e-01 8.14412236e-01
2.09245205e-01 2.19871283e-01 9.23022509e-01 -5.10102570e-01
-9.01430190e-01 -1.38993800e+00 -5.93296766e-01 3.68244320e-01
2.64866114e-01 -3.36861789e-01 -5.16278088e-01 -3.56325179e-01] | [9.2388916015625, 10.532865524291992] |
7711ecf3-1652-4691-be25-e52d4238617b | sttracker-spatio-temporal-tracker-for-3d | 2306.17440 | null | https://arxiv.org/abs/2306.17440v1 | https://arxiv.org/pdf/2306.17440v1.pdf | STTracker: Spatio-Temporal Tracker for 3D Single Object Tracking | 3D single object tracking with point clouds is a critical task in 3D computer vision. Previous methods usually input the last two frames and use the predicted box to get the template point cloud in previous frame and the search area point cloud in the current frame respectively, then use similarity-based or motion-based methods to predict the current box. Although these methods achieved good tracking performance, they ignore the historical information of the target, which is important for tracking. In this paper, compared to inputting two frames of point clouds, we input multi-frame of point clouds to encode the spatio-temporal information of the target and learn the motion information of the target implicitly, which could build the correlations among different frames to track the target in the current frame efficiently. Meanwhile, rather than directly using the point feature for feature fusion, we first crop the point cloud features into many patches and then use sparse attention mechanism to encode the patch-level similarity and finally fuse the multi-frame features. Extensive experiments show that our method achieves competitive results on challenging large-scale benchmarks (62.6% in KITTI and 49.66% in NuScenes). | ['Zheng Fang', 'Zhiheng Li', 'Yubo Cui'] | 2023-06-30 | null | null | null | null | ['object-tracking', '3d-single-object-tracking'] | ['computer-vision', 'computer-vision'] | [-5.42871617e-02 -8.35105002e-01 -6.87445924e-02 -2.88108103e-02
-5.35194278e-01 -4.79818612e-01 4.78092223e-01 -9.95077118e-02
-3.79538119e-01 2.36566827e-01 -2.24177614e-01 1.41953155e-01
1.74926743e-01 -6.91790760e-01 -7.54166722e-01 -8.60164523e-01
-5.33576347e-02 3.81955594e-01 9.87742841e-01 -4.53100093e-02
3.28482181e-01 9.36409414e-01 -1.57793987e+00 2.73238849e-02
4.34409231e-01 1.21926224e+00 7.13556528e-01 5.21304071e-01
-4.36212212e-01 4.45981652e-01 -4.06118184e-01 9.17126238e-02
6.33756816e-01 1.94355287e-02 -2.17859879e-01 4.36511070e-01
7.73438156e-01 -5.31548381e-01 -6.02734149e-01 1.31821465e+00
3.46096963e-01 1.67245910e-01 2.87556231e-01 -1.26965296e+00
-3.38689536e-01 -1.08684540e-01 -9.23590958e-01 5.59097707e-01
2.22192198e-01 4.33956474e-01 4.93309915e-01 -1.00862288e+00
7.25126743e-01 1.52465785e+00 6.96480155e-01 3.62392813e-01
-6.59903407e-01 -1.01288056e+00 5.50482631e-01 3.50178212e-01
-1.36027837e+00 -2.56055266e-01 8.23488712e-01 -5.61426103e-01
6.94422364e-01 6.51776791e-02 1.13516223e+00 5.34889877e-01
6.25469312e-02 6.96908236e-01 6.26516700e-01 -3.77518348e-02
-2.41037250e-01 -3.51668060e-01 5.05467989e-02 5.78199029e-01
2.76351690e-01 5.78101039e-01 -1.97404236e-01 -1.57211140e-01
1.03872788e+00 7.26882875e-01 -3.52104068e-01 -4.96463835e-01
-1.67959249e+00 5.97862959e-01 7.80114532e-01 2.24844426e-01
-5.37493825e-01 2.81062663e-01 2.26101354e-01 5.76525144e-02
4.17034477e-01 -4.08738315e-01 -4.71818805e-01 2.14198738e-01
-8.78917813e-01 3.29489619e-01 2.49385893e-01 1.31635380e+00
9.43672597e-01 -8.83281827e-02 -2.68143445e-01 2.45833129e-01
5.79814315e-01 1.06457901e+00 8.04012269e-02 -9.24503565e-01
5.72488427e-01 6.74423635e-01 5.28869450e-01 -1.13924134e+00
-9.04727280e-02 -2.95377523e-01 -6.63290203e-01 4.64414001e-01
2.50478059e-01 -3.85392234e-02 -1.10694969e+00 1.20116174e+00
8.72946918e-01 9.19534743e-01 -2.83877283e-01 1.21529138e+00
7.86571741e-01 9.63101983e-01 -1.39412388e-01 -4.03309405e-01
1.20393455e+00 -9.85867023e-01 -4.48293746e-01 -9.09068584e-02
2.45307788e-01 -9.81559634e-01 2.85422683e-01 -5.64986840e-02
-1.00123620e+00 -9.95503664e-01 -8.43420744e-01 9.98093411e-02
-9.82623771e-02 9.58378538e-02 4.16065991e-01 9.80877280e-02
-6.64749205e-01 5.15842617e-01 -1.13156605e+00 -1.92501128e-01
4.95200336e-01 4.13527131e-01 -2.34072849e-01 -3.63979310e-01
-7.80801058e-01 7.13605165e-01 3.52065861e-01 1.71175927e-01
-7.59459674e-01 -8.72239709e-01 -7.62347996e-01 -1.49052083e-01
2.23435208e-01 -8.40535581e-01 8.90410066e-01 -6.60064161e-01
-1.09863460e+00 6.13151133e-01 -4.50221241e-01 -2.81478137e-01
4.02185470e-01 -3.36442024e-01 -1.70509651e-01 -8.89277533e-02
2.42899209e-01 8.16868067e-01 7.82430410e-01 -1.12283015e+00
-1.34569287e+00 -5.28632760e-01 -1.49591878e-01 3.66129130e-01
2.36880973e-01 9.50186998e-02 -1.11679769e+00 -2.62429088e-01
5.30327678e-01 -1.08984137e+00 -3.61568511e-01 5.48109114e-01
4.55744155e-02 -5.03342450e-01 1.30360126e+00 -2.05181152e-01
7.51992345e-01 -2.20327902e+00 5.17176371e-03 -1.35546520e-01
2.99667679e-02 4.00508046e-01 -9.95809808e-02 5.64223737e-04
1.62754327e-01 -4.69298124e-01 1.74797356e-01 -2.76154548e-01
-3.52396458e-01 1.39251024e-01 -3.73640060e-01 7.42391229e-01
2.41863742e-01 9.28540707e-01 -1.03353810e+00 -7.22528636e-01
6.56683803e-01 7.10762143e-01 -3.98949534e-01 2.39145681e-02
-3.61015588e-01 6.18874073e-01 -8.90423536e-01 7.57666826e-01
1.05680478e+00 -3.21259230e-01 -5.59082627e-01 -4.23743337e-01
-5.27404666e-01 -6.59018457e-02 -1.34403336e+00 2.03811502e+00
9.64711681e-02 3.99005175e-01 -1.23416424e-01 -4.38410938e-01
1.02535856e+00 7.07345828e-02 9.63824034e-01 -4.23706710e-01
1.42006606e-01 -3.16701196e-02 -8.67984742e-02 -2.49670714e-01
3.19931149e-01 8.16164091e-02 2.51608074e-01 -6.68483302e-02
-2.44966865e-01 -2.22657949e-01 -1.22345127e-01 -1.22062251e-01
9.44573283e-01 4.08882529e-01 1.61057767e-02 2.56281316e-01
7.17454970e-01 5.66824079e-01 8.76555681e-01 5.13429165e-01
-4.36935067e-01 7.32127070e-01 -2.58506149e-01 -7.71153986e-01
-1.01038110e+00 -9.22098219e-01 -2.84675527e-02 5.40517449e-01
7.48757660e-01 -1.52783006e-01 -6.79022670e-02 -5.41546404e-01
2.71385223e-01 1.06774919e-01 -3.66526693e-01 -4.65413071e-02
-8.00437212e-01 -1.71582207e-01 -1.46215886e-01 4.92964238e-01
3.83261412e-01 -7.52222896e-01 -7.71085799e-01 4.09681380e-01
-4.92852256e-02 -1.03290224e+00 -7.66119182e-01 -2.05123052e-01
-1.09667003e+00 -1.06641078e+00 -7.46430576e-01 -7.41921961e-01
4.31462795e-01 1.11155963e+00 8.99093747e-01 3.20279300e-01
-1.60961896e-02 1.15390159e-01 -3.21622312e-01 -5.97330153e-01
2.76253670e-01 -4.29350138e-01 -2.85973940e-02 -2.16182128e-01
5.78547776e-01 -2.53231436e-01 -6.79163754e-01 4.63390559e-01
-3.73768508e-01 -4.31157909e-02 4.64129806e-01 7.08642244e-01
1.06343639e+00 -2.99328528e-02 -2.63428330e-01 -2.03862578e-01
-2.89976776e-01 -3.00513446e-01 -1.07837498e+00 1.34783536e-01
3.96609344e-02 -3.85663718e-01 7.03696311e-02 -6.96472704e-01
-5.40773749e-01 6.88157856e-01 3.44524607e-02 -1.55454409e+00
-2.61808392e-02 -6.02754131e-02 -7.15043992e-02 -3.25610906e-01
7.83849880e-02 3.42104495e-01 -3.89622413e-02 -4.63383675e-01
1.33930907e-01 2.19912097e-01 6.41138136e-01 -3.30312401e-01
1.21811831e+00 7.30127573e-01 7.99007118e-02 -3.50272208e-01
-9.34074223e-01 -8.22792709e-01 -8.92472386e-01 -2.41064206e-01
1.10715795e+00 -1.33747244e+00 -9.38907623e-01 3.30004454e-01
-1.66737187e+00 2.59421140e-01 -1.13554731e-01 7.58178592e-01
-2.66221106e-01 3.42283845e-01 -3.28722119e-01 -7.65506923e-01
-3.46004754e-01 -1.29053104e+00 1.60238385e+00 4.27750289e-01
5.16439319e-01 -6.02212727e-01 1.35566607e-01 -1.75478876e-01
3.13732848e-02 3.18797022e-01 2.13742852e-01 -1.86129376e-01
-1.27783489e+00 -3.84428710e-01 -4.64523762e-01 -1.69085771e-01
1.53230473e-01 9.16094109e-02 -6.07448339e-01 -3.77655149e-01
2.12431878e-01 3.47893119e-01 6.70535803e-01 6.34811163e-01
9.89047825e-01 1.98686495e-01 -8.72738421e-01 8.07427049e-01
1.47120214e+00 5.00570893e-01 3.63389313e-01 2.59195805e-01
9.91879761e-01 1.43470123e-01 1.28197551e+00 4.19812828e-01
3.33278000e-01 8.77860904e-01 6.57383740e-01 1.26504481e-01
-1.39017865e-01 -2.53515452e-01 3.06658179e-01 7.82288015e-01
-1.96102440e-01 2.17077777e-01 -7.97481418e-01 4.76258159e-01
-2.03084111e+00 -1.26884818e+00 -3.81233454e-01 2.12728190e+00
3.40154558e-01 2.17717290e-01 3.10326237e-02 -3.21803063e-01
1.11105371e+00 9.65330303e-02 -8.33436489e-01 6.70915246e-01
-4.75629121e-02 -1.01360068e-01 6.92926884e-01 2.78446406e-01
-1.26010084e+00 9.66692150e-01 5.54116201e+00 7.13946164e-01
-1.19918668e+00 1.34779692e-01 1.32030249e-02 -3.40825796e-01
1.48508504e-01 1.73909321e-01 -1.37163091e+00 7.52772570e-01
4.41138059e-01 -1.84316069e-01 2.06550620e-02 8.97770703e-01
7.29997084e-02 9.19788852e-02 -9.45846498e-01 1.29248083e+00
-2.37052351e-01 -1.46322477e+00 -6.88428581e-02 7.36757442e-02
7.30511785e-01 4.99559492e-01 -2.08773553e-01 1.68211207e-01
1.93827242e-01 -5.77478051e-01 8.59725714e-01 7.53866613e-01
3.55621874e-01 -6.41676128e-01 6.58834338e-01 5.84511459e-01
-1.83119833e+00 8.85524079e-02 -1.00535548e+00 -1.40937958e-02
2.54300058e-01 4.53155279e-01 -5.40791452e-01 6.54729784e-01
1.17498279e+00 1.29379928e+00 -4.36879903e-01 1.57970452e+00
2.94616818e-01 7.16076419e-02 -7.61076689e-01 4.49884757e-02
3.94649833e-01 -2.75264084e-01 7.29937434e-01 7.60631025e-01
7.04803884e-01 3.87557060e-01 6.99841261e-01 5.86768985e-01
3.89588624e-01 -2.80190378e-01 -6.47920668e-01 4.06162262e-01
7.04934895e-01 1.16232538e+00 -5.40682733e-01 -5.34377992e-01
-7.09314823e-01 7.35139012e-01 3.98762636e-02 4.55164239e-02
-9.30164218e-01 -1.33973688e-01 9.16620731e-01 5.72112240e-02
9.55974996e-01 -5.40372789e-01 6.73812404e-02 -1.10175562e+00
5.50872460e-02 -4.04774755e-01 2.86060065e-01 -1.06754255e+00
-1.23910737e+00 5.83633423e-01 -2.18660519e-01 -2.07581401e+00
4.86143678e-02 -4.16846246e-01 -6.51257277e-01 1.15061390e+00
-1.60948598e+00 -1.22409999e+00 -6.40570164e-01 8.48626494e-01
6.55334532e-01 1.19937435e-01 2.21676841e-01 3.64395350e-01
-1.98832378e-01 -6.88398927e-02 2.26139370e-02 1.96084082e-01
6.44491017e-01 -7.77241528e-01 6.91506982e-01 6.73348784e-01
1.54756695e-01 4.85873044e-01 4.44148749e-01 -1.00668812e+00
-1.67490304e+00 -1.27881241e+00 6.39696002e-01 -7.04541326e-01
5.92335224e-01 -7.42802620e-02 -9.85524416e-01 5.69998324e-01
-9.83075574e-02 7.51895845e-01 -4.10330705e-02 -4.90671545e-01
2.68243216e-02 -1.30207837e-01 -8.73452604e-01 3.27680171e-01
1.46577585e+00 -1.54087156e-01 -7.09985077e-01 4.36558634e-01
1.15684867e+00 -1.07051873e+00 -7.97694683e-01 5.63788831e-01
2.60203660e-01 -7.72992849e-01 1.43913090e+00 -4.41026747e-01
-4.11448395e-03 -1.14322698e+00 -1.90261662e-01 -9.21020269e-01
-6.48048043e-01 -1.39633596e-01 -1.31411269e-01 1.00852931e+00
-1.98215604e-01 -1.99484169e-01 1.05526876e+00 3.42172652e-01
-2.95932472e-01 -7.14927554e-01 -9.44588900e-01 -6.93036318e-01
-1.97969794e-01 -3.48089993e-01 9.36795354e-01 7.06147432e-01
-8.22812855e-01 3.50191556e-02 -3.77861023e-01 6.60443485e-01
8.20316851e-01 7.12936997e-01 1.06145442e+00 -1.45767081e+00
1.74361721e-01 -1.72520489e-01 -8.19062591e-01 -1.50293481e+00
1.11667484e-01 -8.19239259e-01 1.03882439e-01 -1.32837260e+00
4.59042899e-02 -6.48097336e-01 -2.99563080e-01 2.41203502e-01
-4.26623702e-01 1.06693037e-01 6.32769346e-01 5.58812082e-01
-7.42478848e-01 5.62151670e-01 1.50044382e+00 -3.27947825e-01
-1.69481233e-01 2.81407535e-01 -1.50141135e-01 6.01754546e-01
1.93002060e-01 -7.39172339e-01 -1.06931686e-01 -7.15675235e-01
-4.95821744e-01 3.38456601e-01 7.60274947e-01 -1.35486448e+00
6.93895757e-01 -4.53795791e-01 1.00456440e+00 -1.63764262e+00
6.90603316e-01 -1.41238952e+00 4.95231807e-01 6.98481083e-01
2.85626173e-01 3.44276488e-01 2.34617800e-01 9.54471946e-01
-2.17126325e-01 -1.04781792e-01 6.22808397e-01 -1.13666788e-01
-1.11685014e+00 1.20346379e+00 3.54111642e-01 -4.05957311e-01
1.31169999e+00 -4.19457525e-01 -1.61322132e-01 2.12951601e-01
-4.46189791e-01 5.69875836e-01 8.44105184e-01 6.58625484e-01
6.94198072e-01 -1.67371464e+00 -5.89185894e-01 1.24949209e-01
6.61454871e-02 5.12397349e-01 3.70310068e-01 8.25579822e-01
-3.53115976e-01 3.53360921e-01 -3.04091692e-01 -1.48818076e+00
-1.35190094e+00 8.93980384e-01 3.05910528e-01 7.75488019e-02
-9.69731271e-01 8.24317038e-01 3.43018234e-01 -6.32251147e-03
2.32648030e-01 -3.81737351e-01 -5.20256311e-02 -2.13713139e-01
6.59886420e-01 -7.15244040e-02 -1.31978691e-01 -7.92319059e-01
-6.07634962e-01 1.33537531e+00 2.64609847e-02 2.23240569e-01
1.28831995e+00 -1.07743755e-01 1.37535229e-01 4.36614543e-01
1.20802033e+00 -2.32927889e-01 -1.77540100e+00 -6.23696268e-01
-1.00622736e-01 -1.05683184e+00 7.93507695e-02 -1.41704440e-01
-1.36765325e+00 8.53184819e-01 8.55654955e-01 -2.37115741e-01
9.71429467e-01 3.84045877e-02 9.14426446e-01 2.81316936e-01
6.89149141e-01 -4.06031281e-01 -2.33086064e-01 6.81834519e-01
6.43049002e-01 -1.30910504e+00 2.40868792e-01 -5.65497339e-01
-3.77571225e-01 9.76040781e-01 6.65073514e-01 -4.72146720e-01
7.63141453e-01 1.78766936e-01 2.47026823e-04 -1.55577555e-01
-8.83078098e-01 -5.67485034e-01 3.64347816e-01 8.49843264e-01
-1.52797783e-02 -4.30375993e-01 1.94041550e-01 2.03633413e-01
9.02582258e-02 5.44839650e-02 -1.08338006e-01 1.05410087e+00
-8.21960986e-01 -9.17726159e-01 -8.15961778e-01 3.80895495e-01
-1.51730096e-03 1.85846120e-01 -4.27994691e-02 7.69944370e-01
4.26410317e-01 6.19564056e-01 4.30348486e-01 -4.40696359e-01
5.29671609e-01 -1.78909272e-01 5.02804279e-01 -5.60719609e-01
-6.17911756e-01 4.46961164e-01 -4.85338539e-01 -6.16689265e-01
-7.86798358e-01 -1.05817497e+00 -1.26423883e+00 -3.60922068e-01
-5.62167227e-01 3.14853527e-02 6.11496389e-01 7.81722546e-01
4.60402071e-01 3.31369281e-01 6.41690612e-01 -1.48219383e+00
-2.27548108e-01 -7.03844309e-01 -1.70034245e-01 4.01976883e-01
5.57231367e-01 -1.04238200e+00 -8.95385742e-02 7.87453875e-02] | [6.602785110473633, -2.336766004562378] |
354ca205-df2d-40ac-b806-a200067ab008 | em-pre-training-for-multi-party-dialogue | 2305.12412 | null | https://arxiv.org/abs/2305.12412v1 | https://arxiv.org/pdf/2305.12412v1.pdf | EM Pre-training for Multi-party Dialogue Response Generation | Dialogue response generation requires an agent to generate a response according to the current dialogue history, in terms of which two-party dialogues have been well studied, but leaving a great gap for multi-party dialogues at the same time. Different from two-party dialogues where each response is a direct reply to its previous utterance, the addressee of a response utterance should be specified before it is generated in the multi-party scenario. Thanks to the huge amount of two-party conversational data, various pre-trained language models for two-party dialogue response generation have been proposed. However, due to the lack of annotated addressee labels in multi-party dialogue datasets, it is hard to use them to pre-train a response generation model for multi-party dialogues. To tackle this obstacle, we propose an Expectation-Maximization (EM) approach that iteratively performs the expectation steps to generate addressee labels, and the maximization steps to optimize a response generation model. Theoretical analyses and extensive experiments have justified the feasibility and effectiveness of our proposed method. | ['Hai Zhao', 'Yiyang Li'] | 2023-05-21 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [ 4.73765761e-01 6.93218291e-01 1.74799293e-01 -6.53448403e-01
-9.95935202e-01 -6.56966686e-01 9.86090183e-01 -3.74823213e-02
-3.11736524e-01 9.85953987e-01 6.83745086e-01 -3.28946531e-01
2.44680524e-01 -8.53111207e-01 2.16817409e-01 -4.76086706e-01
4.90283459e-01 8.49837542e-01 2.93482449e-02 -7.78736174e-01
3.58530432e-01 -8.76989663e-02 -1.02045071e+00 4.91205961e-01
7.73029745e-01 5.53680182e-01 3.74227762e-01 9.45569932e-01
-5.24867058e-01 8.15623164e-01 -8.26654553e-01 -5.02408028e-01
-2.78296154e-02 -1.06437433e+00 -1.40652812e+00 2.95325249e-01
-4.29393768e-01 -4.35718119e-01 -2.19149783e-01 6.90328419e-01
7.32403517e-01 5.04517257e-01 6.62738502e-01 -1.11545157e+00
-4.24655408e-01 7.35778451e-01 -6.04331009e-02 -2.95465499e-01
8.92685771e-01 1.77640811e-01 1.04244947e+00 -5.15626252e-01
5.95760822e-01 1.47289979e+00 1.96033150e-01 1.23979592e+00
-9.42865908e-01 -1.26658082e-01 -3.09416950e-02 -1.94055766e-01
-6.21792495e-01 -4.96855319e-01 8.58242512e-01 -2.32345581e-01
5.99003911e-01 3.25016379e-01 3.64934504e-01 1.15337062e+00
-1.41416624e-01 7.93202102e-01 1.22046173e+00 -5.22742152e-01
7.71618783e-02 6.19693816e-01 8.31260532e-02 1.71761617e-01
-8.35097551e-01 -3.90484869e-01 -2.86794037e-01 -4.66187894e-01
5.03045619e-01 -4.22679752e-01 -3.24392825e-01 2.72830337e-01
-1.11190736e+00 1.18453336e+00 -1.40461951e-01 4.34461743e-01
-4.96126354e-01 -5.73498011e-01 5.73933601e-01 5.56267798e-01
3.78564626e-01 4.83899444e-01 -2.21885607e-01 -3.50742698e-01
-5.30979097e-01 4.00592864e-01 1.57137513e+00 7.31302500e-01
7.26967156e-01 -2.66765356e-01 -4.38663423e-01 1.16094971e+00
3.27390641e-01 1.54083192e-01 3.91948044e-01 -7.32363164e-01
7.62554944e-01 8.71988118e-01 5.10546029e-01 -9.21134889e-01
-3.99245918e-01 2.55729198e-01 -9.25108492e-01 -2.99066931e-01
6.21029556e-01 -5.55461884e-01 -9.09851268e-02 1.58285606e+00
5.69041133e-01 -4.09534156e-01 5.72664261e-01 1.01216245e+00
9.94907677e-01 1.08084846e+00 -4.21992615e-02 -5.31157434e-01
1.36390889e+00 -1.09157515e+00 -8.12097251e-01 -1.65855199e-01
5.92435181e-01 -9.56488967e-01 1.06783307e+00 7.05236122e-02
-9.85952854e-01 -4.17437792e-01 -5.55482626e-01 2.76292358e-02
2.01734632e-01 -1.24561833e-02 5.94168842e-01 6.13016427e-01
-7.78634429e-01 1.62197575e-01 -1.69676974e-01 -3.08770895e-01
-4.12019044e-01 2.03251645e-01 -2.77817965e-01 1.27093971e-01
-1.64936078e+00 9.37212884e-01 1.33492455e-01 2.80587345e-01
-6.75816357e-01 -5.15369959e-02 -7.09342837e-01 -5.64067671e-03
3.66220295e-01 -5.71603298e-01 1.74869251e+00 -6.93463862e-01
-2.20427132e+00 7.42080450e-01 -3.32478553e-01 -1.62128046e-01
7.45689869e-01 1.67833015e-01 -1.11685537e-01 7.57548511e-02
-3.87052447e-02 5.22145987e-01 5.11814356e-01 -1.19746888e+00
-7.37587988e-01 -1.34747177e-01 6.04065359e-01 6.59887731e-01
-1.68663353e-01 2.37671256e-01 -1.99423462e-01 -1.77412350e-02
1.88384420e-05 -1.07015741e+00 -5.04706562e-01 -9.23409224e-01
-4.34508473e-01 -7.25547671e-01 6.24630690e-01 -5.62719822e-01
1.19842899e+00 -1.79381561e+00 2.31386963e-02 -2.21841320e-01
4.84460825e-03 1.74650222e-01 -1.96901396e-01 1.06097484e+00
2.81987786e-01 -1.63907502e-02 -1.84605107e-01 -3.95370126e-01
3.53509516e-01 4.30200063e-02 -4.71884310e-01 5.57086617e-02
4.99392189e-02 6.43130481e-01 -1.00694478e+00 -5.59897721e-01
1.07896790e-01 1.45852551e-01 -4.23641533e-01 8.94336462e-01
-4.64818090e-01 9.18535888e-01 -7.72117615e-01 2.14232281e-02
4.35545951e-01 -1.84618533e-01 4.18608814e-01 1.17436245e-01
-1.31743193e-01 6.12097979e-01 -9.29622710e-01 1.35249555e+00
-7.60349333e-01 3.31506252e-01 2.33410239e-01 -7.75886536e-01
1.27082539e+00 8.93042207e-01 3.28398794e-01 -4.68307793e-01
2.28225946e-01 2.20976233e-01 2.44491935e-01 -5.89754045e-01
9.94734287e-01 -2.86394000e-01 -7.14966655e-01 1.17540312e+00
-2.03992099e-01 -5.25668442e-01 3.10140103e-01 2.34078169e-01
8.73256445e-01 -1.17848933e-01 2.50182003e-01 2.75523573e-01
1.04874897e+00 9.45735164e-03 3.99495900e-01 6.78505182e-01
-1.53626993e-01 5.61757803e-01 6.54154420e-01 -3.44434857e-01
-8.17055106e-01 -4.19860125e-01 2.66664833e-01 1.19531155e+00
8.78373757e-02 -1.70857623e-01 -9.14272189e-01 -7.97980487e-01
-7.58522749e-01 7.74475515e-01 -1.40873492e-01 1.98994786e-01
-8.14601123e-01 -6.87234581e-01 4.52065170e-01 -1.78961203e-01
6.50119722e-01 -1.42020130e+00 -4.42453742e-01 6.54383123e-01
-1.04226387e+00 -1.10287881e+00 -5.69209158e-01 -2.17603281e-01
-2.63404220e-01 -9.63917017e-01 -7.45728970e-01 -7.92836964e-01
6.15214825e-01 2.69103318e-01 8.56184840e-01 1.31506413e-01
3.22327673e-01 2.47317597e-01 -5.82079113e-01 -1.43488780e-01
-1.31758320e+00 3.55853051e-01 -1.34147644e-01 2.10343421e-01
3.50951135e-01 -3.78332883e-01 -5.72106481e-01 7.26395667e-01
-7.68499494e-01 3.43811035e-01 4.61426258e-01 9.36637819e-01
-1.15772523e-01 -3.01051974e-01 1.11696696e+00 -1.09811974e+00
1.49393761e+00 -4.97927189e-01 -2.60483414e-01 4.03283507e-01
-2.27073297e-01 -5.17484359e-02 8.87188613e-01 -4.29168761e-01
-1.44481730e+00 -1.07943371e-01 -4.91476089e-01 6.53191328e-01
-3.46948415e-01 3.99964303e-01 -2.33783722e-01 3.42442125e-01
4.43957061e-01 3.81936133e-01 3.76042956e-03 -2.59834141e-01
3.36216807e-01 1.26049578e+00 2.83821076e-01 -7.64114678e-01
4.55673397e-01 -6.46881908e-02 -3.93322498e-01 -6.84993088e-01
-6.72331989e-01 -5.04432917e-01 -5.96553504e-01 -3.99702936e-01
8.99248660e-01 -4.98548657e-01 -1.03810942e+00 4.31907624e-01
-1.57502770e+00 -2.57433563e-01 1.76661313e-01 3.49291325e-01
-7.67265201e-01 5.48165321e-01 -6.42387688e-01 -1.10434437e+00
-5.76365411e-01 -1.23451746e+00 5.50478280e-01 3.87391567e-01
-6.53810918e-01 -1.03164184e+00 1.58608198e-01 7.07532763e-01
4.35625464e-01 -7.53231645e-02 1.00309646e+00 -1.11882222e+00
-2.48533621e-01 -3.16802174e-01 -6.13065884e-02 1.46887690e-01
3.27216446e-01 -4.12977904e-01 -6.85715437e-01 7.62495911e-03
4.02878553e-01 -6.05685115e-01 9.18850899e-02 -2.02301949e-01
3.86636704e-01 -6.20026886e-01 2.33608242e-02 -3.97497356e-01
7.38936782e-01 3.98012877e-01 4.77822602e-01 1.93685129e-01
2.01367691e-01 1.22447133e+00 7.28838563e-01 6.82951570e-01
8.45807493e-01 6.66828573e-01 7.72836581e-02 1.87694192e-01
2.74271756e-01 -3.67540836e-01 2.22337127e-01 1.11219013e+00
2.38341928e-01 -6.16792619e-01 -6.57845795e-01 6.05096638e-01
-1.98355651e+00 -9.60018218e-01 -4.66722816e-01 2.03682685e+00
1.17997003e+00 -9.67523456e-02 6.65923953e-02 -1.35255575e-01
8.66325617e-01 3.14277709e-01 -2.55656958e-01 -6.92512393e-01
1.65273249e-02 -2.44931772e-01 -2.09772646e-01 7.40705073e-01
-6.64937854e-01 8.66990566e-01 5.78778744e+00 5.23117721e-01
-9.07373726e-01 8.04853812e-02 6.52503669e-01 4.31102514e-01
-4.16302383e-01 3.60987753e-01 -9.21501458e-01 4.58553523e-01
9.19834614e-01 -4.03370708e-01 4.00886863e-01 4.92872268e-01
6.06441557e-01 -2.70344853e-01 -1.28815091e+00 6.24871314e-01
-1.39112398e-02 -9.15217936e-01 -1.01158589e-01 -2.90520880e-02
5.95043242e-01 -6.39165103e-01 -3.74705732e-01 5.70317447e-01
5.10235488e-01 -8.04638386e-01 7.46875629e-02 2.76399225e-01
2.71274954e-01 -6.83616877e-01 8.13524187e-01 1.06285048e+00
-7.94878542e-01 1.94186479e-01 -3.28606695e-01 -2.08969489e-01
7.41213143e-01 1.86964005e-01 -1.45022392e+00 5.81970453e-01
-1.22784577e-01 -2.20881954e-01 1.02613002e-01 4.35996652e-01
-4.81175542e-01 6.29999578e-01 -1.18091352e-01 -4.14561927e-01
4.01028216e-01 -4.30345863e-01 5.04588664e-01 1.05551100e+00
1.65954158e-01 4.67618793e-01 4.76363212e-01 5.96409500e-01
-1.33694962e-01 4.83213305e-01 -3.05636346e-01 -1.06404781e-01
5.38880348e-01 1.45887196e+00 -3.75980884e-01 -3.70800227e-01
-4.47024673e-01 9.58569407e-01 2.04248037e-02 1.36630654e-01
-5.64947486e-01 -3.21652859e-01 1.46053433e-01 -3.49817462e-02
-3.77076983e-01 -3.34481411e-02 4.64309752e-02 -9.03204560e-01
-1.39009446e-01 -1.25776005e+00 2.26339430e-01 -4.33745801e-01
-1.30586386e+00 9.37246621e-01 -5.87888882e-02 -1.16215050e+00
-1.13378751e+00 -3.93171757e-02 -8.51454258e-01 1.18466842e+00
-1.30285931e+00 -1.01567721e+00 -9.16989073e-02 4.07813549e-01
9.83951211e-01 -1.59502700e-01 1.21123314e+00 1.46162063e-01
-4.16150510e-01 4.27461118e-01 -3.37364197e-01 2.16096207e-01
8.76820385e-01 -1.00601685e+00 3.76488298e-01 3.18726957e-01
-3.20712656e-01 6.90331638e-01 9.03597832e-01 -4.07695770e-01
-1.30709291e+00 -5.05894065e-01 1.33104026e+00 -1.05293781e-01
6.10002577e-01 -2.78489560e-01 -1.06403089e+00 3.59673709e-01
7.43696988e-01 -7.65469253e-01 8.86001587e-01 5.65112010e-02
2.92783111e-01 4.41632926e-01 -1.01167417e+00 8.32668066e-01
5.58676064e-01 -4.12671566e-01 -8.81662428e-01 5.03857255e-01
5.39491415e-01 -5.20862937e-01 -8.41396451e-01 1.08128749e-01
2.68148601e-01 -8.64825249e-01 3.94274861e-01 -4.22639161e-01
5.03461957e-01 -1.65470526e-01 7.54248500e-02 -1.52032351e+00
2.22053871e-01 -1.15790153e+00 3.84976685e-01 1.59494388e+00
4.73547667e-01 -6.57079756e-01 6.69974864e-01 1.25322080e+00
-9.90759507e-02 -3.93492579e-01 -7.47171462e-01 -1.66088462e-01
-1.47903096e-02 -4.74573784e-02 6.74105525e-01 8.11271787e-01
5.36312699e-01 1.05879676e+00 -8.94694209e-01 -1.68735683e-01
1.00901619e-01 4.87259209e-01 1.40314782e+00 -1.07011175e+00
-4.54518408e-01 -2.10854679e-01 3.81971836e-01 -1.79097450e+00
3.63034099e-01 -5.72024703e-01 4.66192245e-01 -1.80799007e+00
1.32965520e-01 -4.52356666e-01 4.93376315e-01 2.27131352e-01
-3.84816378e-01 -2.90275156e-01 1.28492564e-01 3.49867702e-01
-4.11712617e-01 7.70444453e-01 1.35909081e+00 -6.22087605e-02
-5.54587960e-01 6.24996841e-01 -7.21051276e-01 5.87206244e-01
7.97091186e-01 -4.48249310e-01 -5.29814243e-01 -1.13414794e-01
6.40354082e-02 9.82395947e-01 -1.47565350e-01 -3.75187606e-01
4.74452049e-01 -5.23642480e-01 -3.00165921e-01 -5.94572484e-01
4.42589462e-01 -4.25280780e-01 -1.19591527e-01 1.65057272e-01
-8.13924670e-01 -5.57409562e-02 -2.75176257e-01 3.76629084e-01
-4.80193257e-01 -7.83335984e-01 5.53484738e-01 -4.66166437e-01
-5.67705989e-01 2.29727793e-02 -8.17185163e-01 1.05578981e-01
8.61839354e-01 -1.39908656e-01 -2.23091543e-01 -9.63358998e-01
-6.22251689e-01 5.13179243e-01 1.75173596e-01 4.77439463e-01
5.12404084e-01 -1.06576622e+00 -1.10868204e+00 -1.62113383e-01
-5.63503429e-02 6.59300238e-02 6.03012919e-01 5.56392074e-01
-5.36281019e-02 4.47158813e-01 -7.22906664e-02 -2.63142169e-01
-1.33344293e+00 9.37880948e-02 7.43505135e-02 -8.03622723e-01
-3.30466926e-01 5.11426151e-01 2.79474646e-01 -9.46807146e-01
1.44171966e-02 4.21782702e-01 -6.00646257e-01 1.69139534e-01
6.15704060e-01 9.48972329e-02 -2.91222006e-01 -7.43158996e-01
2.05828086e-01 5.13045751e-02 -1.43306911e-01 -5.46661079e-01
1.05040002e+00 -5.96396267e-01 -3.59423786e-01 4.65475023e-01
9.64981079e-01 1.84807256e-01 -9.36092854e-01 -4.03965056e-01
1.83602534e-02 -4.12562907e-01 -6.27398372e-01 -6.22764707e-01
-2.83237129e-01 7.24135339e-01 -2.22999662e-01 7.14942992e-01
7.06008315e-01 -2.55216479e-01 1.22398865e+00 5.13495266e-01
4.69064593e-01 -1.26958609e+00 2.86840379e-01 8.72465909e-01
9.78542566e-01 -1.35236692e+00 -3.52374732e-01 -3.84971470e-01
-1.10281384e+00 1.18824971e+00 8.26692283e-01 3.65437180e-01
8.05617273e-02 -1.33678183e-01 4.46079195e-01 1.23002835e-01
-9.56729054e-01 1.14458822e-01 1.99993197e-02 3.51902157e-01
6.99977636e-01 -3.51321362e-02 -8.43040824e-01 6.52028620e-01
-5.41608930e-01 -2.46614397e-01 8.59756589e-01 7.39727795e-01
-4.68522698e-01 -1.62627947e+00 -3.43891144e-01 2.76931766e-02
-3.64522904e-01 1.39300451e-01 -8.23796391e-01 3.69212449e-01
-5.93423307e-01 1.78712463e+00 -3.81008953e-01 -2.67703950e-01
3.79550278e-01 3.72678965e-01 4.25780527e-02 -9.28081155e-01
-1.07949924e+00 -1.73785221e-02 6.37824297e-01 1.55837655e-01
-5.15876651e-01 -4.16234821e-01 -1.30885386e+00 -3.72195363e-01
-4.45740849e-01 7.93620825e-01 7.14459419e-01 1.05950439e+00
7.21822754e-02 9.69065651e-02 1.28112447e+00 -5.68436623e-01
-1.01531374e+00 -1.38274801e+00 -1.13976263e-01 6.12979889e-01
-1.00411847e-01 -5.32456972e-02 -2.27428436e-01 -1.72526032e-01] | [12.786247253417969, 8.0309476852417] |
41577c19-1ebf-42fa-8ca5-eb68e6cc63d2 | layered-stereo-by-cooperative-grouping-with | 2006.16094 | null | https://arxiv.org/abs/2006.16094v3 | https://arxiv.org/pdf/2006.16094v3.pdf | Level Set Stereo for Cooperative Grouping with Occlusion | Localizing stereo boundaries is difficult because matching cues are absent in the occluded regions that are adjacent to them. We introduce an energy and level-set optimizer that improves boundaries by encoding the essential geometry of occlusions: The spatial extent of an occlusion must equal the amplitude of the disparity jump that causes it. In a collection of figure-ground scenes from Middlebury and Falling Things stereo datasets, the model provides more accurate boundaries than previous occlusion-handling techniques. | ['Todd Zickler', 'Jialiang Wang'] | 2020-06-29 | null | null | null | null | ['occlusion-handling'] | ['computer-vision'] | [ 2.34630302e-01 -4.05559093e-02 -2.68767446e-01 -5.45523822e-01
-5.39939404e-01 -4.71533895e-01 1.65816650e-01 1.06276326e-01
-2.35779896e-01 8.97769153e-01 7.27552831e-01 -1.17439047e-01
2.02709034e-01 -6.61626577e-01 -7.80184269e-01 -2.06904262e-01
-8.14128518e-02 2.54208803e-01 6.36848092e-01 -3.06728870e-01
5.28995216e-01 6.40740693e-01 -1.79568148e+00 5.12095571e-01
9.60665286e-01 9.74747837e-01 -3.98172252e-02 5.75925648e-01
2.09213629e-01 5.15843928e-01 -3.93671840e-01 9.34877172e-02
6.08928859e-01 -4.67169166e-01 -6.53211474e-01 1.83426991e-01
1.34618640e+00 -5.30438185e-01 -6.44298077e-01 6.90015972e-01
3.42141807e-01 3.39013994e-01 4.94470417e-01 -1.00657284e+00
-1.90369636e-01 -2.96194553e-01 -8.38198423e-01 5.83502710e-01
7.39974976e-01 1.50184473e-03 1.05913758e+00 -6.54521108e-01
1.29562294e+00 1.34781480e+00 6.77578807e-01 1.56657577e-01
-1.69852984e+00 -2.62919754e-01 2.21456900e-01 -2.04080883e-02
-1.45403266e+00 -6.37226164e-01 6.83872819e-01 -6.70749903e-01
1.20991266e+00 2.92814344e-01 8.64519179e-01 3.95593345e-01
5.66790938e-01 4.74189997e-01 9.73496795e-01 -2.81188428e-01
1.68393612e-01 -4.69841540e-01 -1.93544943e-02 3.25486183e-01
1.61520422e-01 4.73039597e-01 -7.37698793e-01 -5.26849478e-02
1.07639718e+00 -5.74240327e-01 -5.54736733e-01 -8.03365409e-01
-7.28456080e-01 6.54191196e-01 7.11142421e-01 -7.22800791e-02
-1.57515898e-01 2.15136811e-01 5.76572828e-02 -2.48352308e-02
5.26243508e-01 5.71506858e-01 -2.70786166e-01 9.39776748e-02
-1.24573243e+00 6.04236543e-01 4.68235344e-01 9.52015400e-01
1.10799432e+00 -1.62298352e-01 4.87458184e-02 3.72971416e-01
3.10107004e-02 1.47230238e-01 -2.16262445e-01 -1.75018525e+00
6.53997123e-01 5.91785073e-01 2.51703441e-01 -1.07872391e+00
-2.65092224e-01 -6.59497678e-02 -2.54903585e-01 8.07796419e-01
6.70778811e-01 8.78792033e-02 -1.20293939e+00 1.52327824e+00
5.87304652e-01 6.56538308e-02 -4.03857887e-01 9.44315434e-01
6.51754320e-01 3.23005587e-01 -1.15006141e-01 2.43275892e-02
9.61378217e-01 -6.07126832e-01 -5.83483517e-01 -5.99009871e-01
3.16066712e-01 -1.00267041e+00 7.60049820e-01 5.05730789e-03
-1.18770802e+00 -5.52807331e-01 -1.27486515e+00 -8.15195858e-01
-1.45295486e-01 -7.45981336e-01 7.31472731e-01 1.16926849e-01
-1.16952813e+00 6.55156970e-01 -4.78985280e-01 -8.71800259e-02
2.65244156e-01 3.13041717e-01 -4.92261976e-01 -3.39861773e-02
-9.34373260e-01 9.18435216e-01 8.30443725e-02 9.09944344e-03
-3.05975854e-01 -8.36764097e-01 -1.36299777e+00 -1.49267912e-01
6.61659986e-02 -8.53660226e-01 8.02420020e-01 -8.51242721e-01
-8.25620532e-01 1.35086226e+00 -6.12492383e-01 -3.72409135e-01
7.12576747e-01 -2.60837346e-01 1.90172777e-01 1.25203222e-01
3.91138047e-01 1.22145283e+00 4.68897253e-01 -1.44918263e+00
-6.22363210e-01 -3.52472067e-01 2.53483176e-01 6.26955152e-01
5.33022404e-01 -2.49011904e-01 -6.24513924e-01 -5.74070692e-01
7.79325068e-01 -5.25041342e-01 -3.32550287e-01 3.71355355e-01
-5.62321663e-01 2.79658198e-01 6.35333419e-01 -7.40772426e-01
1.33453619e+00 -2.26911592e+00 2.18238354e-01 3.04927915e-01
2.79860944e-01 -3.39585245e-01 1.58865392e-01 1.27273306e-01
-7.09604174e-02 8.08070228e-02 -2.75190532e-01 -1.87477648e-01
-2.42135689e-01 2.53434896e-01 -3.43657523e-01 6.34836018e-01
-2.72877235e-02 5.53424299e-01 -5.03415942e-01 -8.27373564e-01
5.07184505e-01 4.94039804e-01 -1.00702131e+00 -1.12821147e-01
-1.75406083e-01 5.09987533e-01 -1.66132778e-01 7.10260391e-01
8.70438278e-01 1.74470678e-01 -3.28811258e-01 -1.64066330e-01
-5.44214487e-01 3.60735655e-01 -1.53697741e+00 1.71377444e+00
-5.36440015e-02 1.12864029e+00 1.61001682e-01 -4.06833217e-02
6.04575455e-01 -1.81457102e-01 5.04111528e-01 -1.06794441e+00
-1.64419767e-02 2.40441293e-01 -1.76429108e-01 -2.50412405e-01
6.11139059e-01 -1.80890448e-02 5.59155419e-02 -4.14874367e-02
-6.33290827e-01 -8.05795670e-01 4.06422526e-01 5.84254116e-02
9.14580822e-01 4.50389445e-01 2.29141772e-01 -4.08158749e-01
2.56583929e-01 2.57164091e-01 9.44470882e-01 6.56978488e-01
-4.10627127e-01 1.19923413e+00 2.87018687e-01 -7.11596906e-01
-1.40544927e+00 -1.21591580e+00 -7.69244134e-01 6.73909605e-01
7.39254534e-01 -2.23263532e-01 -7.37424791e-01 1.50455788e-01
3.63736719e-01 3.41161877e-01 -7.35479057e-01 2.56277353e-01
-9.60108936e-01 -9.20774788e-02 -1.26216710e-01 6.36215031e-01
5.06205678e-01 -7.56509542e-01 -8.41621399e-01 1.91472843e-01
-5.86963952e-01 -1.14283192e+00 -7.69002497e-01 1.84365228e-01
-9.62575018e-01 -1.22839260e+00 -4.97066528e-01 -7.82457411e-01
5.01783609e-01 6.58021346e-02 1.40241742e+00 1.54098570e-01
-6.62615240e-01 -2.00200155e-01 2.35920981e-01 1.18661478e-01
1.85234383e-01 -1.41525686e-01 -4.84075367e-01 -4.89328563e-01
1.69104949e-01 -4.86895114e-01 -7.92891383e-01 4.25863624e-01
-3.10800195e-01 1.84352875e-01 -2.65727937e-01 6.46361947e-01
9.12597716e-01 -2.05228269e-01 -3.31425756e-01 -2.72652656e-01
1.68348476e-02 -5.45665063e-03 -8.97393167e-01 -1.10848457e-01
6.26557395e-02 6.45327494e-02 -7.57214846e-03 6.34089559e-02
-1.03583002e+00 1.20162576e-01 9.91887003e-02 -1.15126029e-01
-1.46063849e-01 -2.15168700e-01 -1.38134912e-01 -1.03660174e-01
7.41532028e-01 -4.77064848e-01 -5.59538662e-01 -2.23482251e-01
1.66956440e-01 1.78702638e-01 8.30624104e-01 -5.82561851e-01
4.15646911e-01 1.02226853e+00 2.07075208e-01 -8.70201528e-01
-8.53449523e-01 -5.31895936e-01 -8.64011645e-01 -2.96373695e-01
1.08628678e+00 -8.25869143e-01 -5.94965935e-01 1.34240821e-01
-1.22532094e+00 -4.83692795e-01 -5.03905714e-01 5.77927351e-01
-7.60163665e-01 1.95169464e-01 -6.31467819e-01 -4.95484024e-01
3.02242726e-01 -1.17855346e+00 1.31088078e+00 3.58087927e-01
-5.70402026e-01 -8.79200041e-01 3.73013437e-01 2.93510944e-01
1.19948767e-01 7.47066200e-01 1.00345588e+00 5.35706758e-01
-9.78357077e-01 -1.20045422e-02 -2.76385188e-01 -7.55264936e-03
-2.08443627e-02 3.39863628e-01 -1.04235184e+00 -1.59775955e-03
-2.60529548e-01 -1.23257399e-01 1.04913580e+00 1.17697179e+00
7.76115239e-01 2.89184123e-01 -5.30138791e-01 1.14743769e+00
1.62016940e+00 3.77674341e-01 9.96432066e-01 7.71848798e-01
5.99169731e-01 7.34398901e-01 3.58649850e-01 1.67164117e-01
4.04251218e-01 1.02971375e+00 2.04985663e-01 -6.96468472e-01
-5.48401594e-01 -2.66153127e-01 -2.54890084e-01 -5.76580204e-02
-4.10405882e-02 -1.97182566e-01 -1.01381004e+00 5.20275235e-01
-1.70417738e+00 -9.34966683e-01 -5.55424213e-01 2.31834698e+00
7.47467935e-01 3.20339561e-01 -2.64053136e-01 1.44695684e-01
8.50367367e-01 3.45111191e-01 -7.19461918e-01 -5.97180367e-01
-5.79883695e-01 1.01602554e-01 6.38313532e-01 1.53709793e+00
-1.22731102e+00 1.03000283e+00 8.30170250e+00 3.86209458e-01
-6.90151691e-01 -4.46519107e-01 7.54281163e-01 -2.34746829e-01
-4.53335434e-01 2.67427206e-01 -8.26796830e-01 2.36579448e-01
1.49074748e-01 -1.21573605e-01 3.34270298e-01 6.22524381e-01
3.95549804e-01 -9.80672061e-01 -1.31365240e+00 8.97075117e-01
-5.96173964e-02 -1.25794339e+00 -3.78721088e-01 2.50739604e-01
1.09111762e+00 -9.14727300e-02 9.04552937e-02 -3.36818635e-01
2.93649614e-01 -1.12822437e+00 7.60250151e-01 2.66019762e-01
9.00121033e-01 -4.60978061e-01 4.49334592e-01 -1.54624462e-01
-1.41776037e+00 2.12069228e-01 -2.32869849e-01 -5.05825520e-01
6.38701677e-01 5.25992274e-01 -2.37828076e-01 3.60334776e-02
7.14912355e-01 4.67552602e-01 -4.16384786e-01 1.22057498e+00
-3.20927680e-01 -2.30315387e-01 -5.89479864e-01 8.61542583e-01
2.34416261e-01 -3.87593389e-01 5.37765443e-01 7.66931355e-01
-2.01571181e-01 2.33408242e-01 1.94816500e-01 6.88663185e-01
8.66303220e-02 -1.10022336e-01 -9.03909981e-01 9.97431576e-01
5.40590584e-01 3.59138817e-01 -7.29005396e-01 -4.33760107e-01
-2.96930850e-01 8.88170481e-01 4.03751999e-01 6.98610842e-01
-6.67286754e-01 -2.35050514e-01 1.06796825e+00 5.50119221e-01
7.54495561e-02 -2.13470504e-01 -9.52689946e-01 -9.26865757e-01
3.04365605e-01 -6.04752302e-01 2.58051425e-01 -9.73896742e-01
-7.94013917e-01 3.74379069e-01 1.13550082e-01 -9.56516862e-01
-1.94104016e-01 -2.77410537e-01 -3.34918380e-01 1.06792247e+00
-1.30898786e+00 -5.55971801e-01 -5.35317600e-01 4.35104787e-01
5.19001722e-01 6.85747683e-01 4.13089871e-01 4.50967371e-01
-2.41416931e-01 1.60372645e-01 -1.88004076e-01 -5.40746190e-02
6.27549767e-01 -1.18063009e+00 4.99211788e-01 1.11619568e+00
-2.44640365e-01 2.98140585e-01 1.02139223e+00 -8.22707891e-01
-5.20903945e-01 -3.74966174e-01 1.10830891e+00 -4.00423199e-01
2.16444954e-01 -3.50208908e-01 -9.53884482e-01 7.67739236e-01
1.37559712e-01 -1.32037438e-02 1.75753072e-01 -1.55809359e-03
-3.79613638e-01 3.39979976e-02 -1.21112895e+00 7.80221522e-01
1.49284506e+00 -6.51122689e-01 -7.42506444e-01 7.59840235e-02
3.77655536e-01 -9.22834814e-01 -5.30713797e-01 4.69426095e-01
7.30065286e-01 -1.58075380e+00 1.32769310e+00 -4.19868708e-01
4.41560775e-01 -3.22564989e-01 -4.25163716e-01 -1.01508081e+00
-2.72258848e-01 -5.58211803e-01 3.20551187e-01 6.97533727e-01
4.78022993e-02 -4.56455767e-01 1.04455662e+00 1.08591819e+00
-3.26745570e-01 -5.43026745e-01 -1.32386053e+00 -5.80981910e-01
4.40931097e-02 -1.97155014e-01 3.36495399e-01 6.35156751e-01
-9.58997849e-03 -2.41806637e-02 1.17004849e-01 -4.96348292e-02
8.65680397e-01 1.69286489e-01 7.84285188e-01 -1.09344363e+00
-2.98058558e-02 -5.01693070e-01 -7.36887276e-01 -1.34760010e+00
-3.75118032e-02 -3.51630867e-01 1.23333149e-01 -1.76608837e+00
-7.97531679e-02 -6.88819528e-01 3.51921946e-01 1.96482703e-01
-2.53031999e-01 5.28367579e-01 2.57563926e-02 1.84558928e-01
-1.05525829e-01 3.12789381e-01 1.39704204e+00 8.39166716e-02
-4.85476822e-01 -6.50037825e-01 -1.02067433e-01 1.15061784e+00
4.85116243e-01 -3.63269925e-01 -2.60956138e-01 -6.56709254e-01
1.05114892e-01 2.36144319e-01 3.40710908e-01 -1.06969285e+00
1.50538251e-01 -2.68060803e-01 6.62892878e-01 -9.54228818e-01
7.20964015e-01 -5.61512291e-01 4.25949454e-01 2.09940717e-01
-2.73423910e-01 3.04150600e-02 3.32215637e-01 1.12705395e-01
-2.61129856e-01 1.11846425e-01 1.06961286e+00 -1.69400692e-01
-8.52545381e-01 1.61526069e-01 -1.65072232e-01 7.29856074e-01
1.04248166e+00 -7.86919713e-01 -3.56826901e-01 -3.44571978e-01
-7.10269213e-01 2.49449536e-01 1.31533563e+00 2.03862563e-01
5.90344846e-01 -1.28073370e+00 -5.21433353e-01 3.78583431e-01
-4.58333977e-02 2.27631643e-01 1.16935953e-01 5.30437291e-01
-1.06541491e+00 3.57590616e-01 -2.49887973e-01 -7.12527514e-01
-1.26703811e+00 4.66440171e-01 8.07774067e-01 1.18809611e-01
-8.48336995e-01 1.00870621e+00 7.00657487e-01 2.27968439e-01
3.09974760e-01 -4.46640879e-01 1.21097013e-01 -3.03293243e-02
4.07493353e-01 6.16140902e-01 -8.55124518e-02 -9.37465370e-01
-5.05603731e-01 1.24263597e+00 2.41623208e-01 -4.31520760e-01
8.71674061e-01 -3.00958663e-01 4.33706604e-02 4.20973390e-01
1.19464517e+00 1.45412192e-01 -1.68849456e+00 2.49014243e-01
-3.03439736e-01 -1.15604377e+00 1.80632547e-01 -3.63367140e-01
-6.81120276e-01 7.70245850e-01 5.69999397e-01 -8.79925415e-02
7.72124231e-01 -5.87556697e-03 6.39079452e-01 -2.55035609e-01
5.30344427e-01 -1.39195216e+00 -1.96262419e-01 3.44241798e-01
9.49849606e-01 -1.18997622e+00 2.93088317e-01 -9.73065436e-01
-1.83591425e-01 8.14383626e-01 8.11896145e-01 -1.38941959e-01
5.28245270e-01 6.13472998e-01 1.24257542e-01 -3.85438770e-01
-2.91783780e-01 -4.36430782e-01 3.58433783e-01 7.54833937e-01
4.57902431e-01 -1.82905257e-01 -3.14995766e-01 -6.48546696e-01
-4.11953837e-01 -1.72550634e-01 2.64222860e-01 9.98445749e-01
-6.43308401e-01 -7.00737000e-01 -6.88241959e-01 1.02593042e-01
3.09007596e-02 -2.43209139e-01 -2.77857423e-01 9.72563565e-01
4.97417986e-01 1.03598285e+00 6.11956835e-01 1.10307194e-01
6.25703990e-01 -3.07244748e-01 6.83556437e-01 -5.47133327e-01
-1.15271159e-01 3.58200282e-01 3.80531549e-01 -1.14370322e+00
-3.28620136e-01 -7.68728614e-01 -1.48964071e+00 -3.88109773e-01
-2.62134314e-01 -2.85882652e-01 2.00443968e-01 6.49301469e-01
1.38733834e-01 3.55723768e-01 4.88119096e-01 -1.15359497e+00
3.62381637e-01 -4.68820959e-01 -5.01911998e-01 7.33972967e-01
5.61884046e-01 -9.39653456e-01 -7.80257165e-01 1.56267971e-01] | [8.864907264709473, -2.4267330169677734] |
91cd6695-f378-44eb-bfa7-0377cf7211fe | multilingual-unsupervised-sentence | 2005.00352 | null | https://arxiv.org/abs/2005.00352v2 | https://arxiv.org/pdf/2005.00352v2.pdf | MUSS: Multilingual Unsupervised Sentence Simplification by Mining Paraphrases | Progress in sentence simplification has been hindered by a lack of labeled parallel simplification data, particularly in languages other than English. We introduce MUSS, a Multilingual Unsupervised Sentence Simplification system that does not require labeled simplification data. MUSS uses a novel approach to sentence simplification that trains strong models using sentence-level paraphrase data instead of proper simplification data. These models leverage unsupervised pretraining and controllable generation mechanisms to flexibly adjust attributes such as length and lexical complexity at inference time. We further present a method to mine such paraphrase data in any language from Common Crawl using semantic sentence embeddings, thus removing the need for labeled data. We evaluate our approach on English, French, and Spanish simplification benchmarks and closely match or outperform the previous best supervised results, despite not using any labeled simplification data. We push the state of the art further by incorporating labeled simplification data. | ['Benoît Sagot', 'Éric de la Clergerie', 'Angela Fan', 'Louis Martin', 'Antoine Bordes'] | 2020-05-01 | null | https://aclanthology.org/2022.lrec-1.176 | https://aclanthology.org/2022.lrec-1.176.pdf | lrec-2022-6 | ['parallel-corpus-mining'] | ['natural-language-processing'] | [ 3.59047055e-02 2.56495953e-01 -1.74776822e-01 -6.70125902e-01
-9.93118465e-01 -6.26985312e-01 5.60912371e-01 7.06716657e-01
-7.74965167e-01 8.41104388e-01 8.92201126e-01 -3.29652727e-01
1.16764039e-01 -6.86847031e-01 -6.32245958e-01 4.92204390e-02
5.07525682e-01 8.01153481e-01 -3.86608511e-01 -9.76709008e-01
2.80342370e-01 3.02376568e-01 -1.08642805e+00 3.16552103e-01
1.38077879e+00 -9.00034830e-02 1.27445713e-01 7.10751235e-01
-3.89711112e-01 6.98207259e-01 -8.02517951e-01 -7.90784001e-01
2.67644972e-01 -2.13610694e-01 -9.50948298e-01 -2.64010757e-01
6.48528457e-01 5.41746840e-02 -2.87328035e-01 8.40790510e-01
6.27112329e-01 1.72808275e-01 4.75963086e-01 -7.07565546e-01
-9.38223720e-01 1.01057744e+00 4.35378551e-02 1.87822953e-01
5.83032727e-01 1.52298480e-01 1.03134418e+00 -5.27203381e-01
1.14391196e+00 1.43103111e+00 1.03920305e+00 7.18752146e-01
-1.52011192e+00 -3.29075813e-01 1.35507509e-01 7.57212415e-02
-1.16315258e+00 -6.78745866e-01 6.55835032e-01 -2.17578202e-01
1.67623901e+00 2.07975492e-01 5.63431859e-01 9.48471606e-01
2.52223104e-01 5.46352744e-01 7.40237892e-01 -9.85600829e-01
-1.53685939e-02 7.45012164e-02 3.43507081e-01 6.49686396e-01
3.02061588e-01 -3.69542867e-01 -2.08811626e-01 9.99318138e-02
-1.80121914e-01 -1.53867081e-01 7.90866539e-02 -1.50526971e-01
-9.17238116e-01 1.07844055e+00 1.52472362e-01 2.48418957e-01
1.34449288e-01 -2.15799987e-01 9.70808864e-01 7.67868757e-01
7.44011521e-01 1.28783154e+00 -7.51464844e-01 -2.67833918e-01
-1.13226628e+00 4.15599138e-01 1.17580307e+00 1.18275285e+00
8.79311800e-01 6.31153211e-02 -3.14667672e-01 1.01642215e+00
-3.07497203e-01 5.51905036e-01 4.46276128e-01 -1.24028647e+00
9.53418195e-01 6.43159986e-01 -8.85236412e-02 -5.90598702e-01
-4.76548642e-01 -1.81500182e-01 -5.66891909e-01 -2.08480835e-01
6.12128042e-02 -1.82206839e-01 -7.48164296e-01 1.89088976e+00
-1.62670925e-01 -6.75061345e-01 1.65886179e-01 2.78691947e-01
8.28250408e-01 3.55138600e-01 2.15782613e-01 -1.05421498e-01
9.31286395e-01 -1.26322722e+00 -8.06033969e-01 -2.19211295e-01
1.43501818e+00 -9.78938997e-01 1.66963792e+00 4.07249331e-01
-1.11989355e+00 -4.03688729e-01 -1.05448008e+00 -8.88327003e-01
-5.46375275e-01 -1.27422273e-01 7.91459799e-01 6.76349282e-01
-1.02410746e+00 1.02245128e+00 -6.21262729e-01 -7.31994569e-01
3.68494362e-01 3.23490739e-01 -5.17022610e-01 -4.24853474e-01
-1.13518953e+00 1.51463008e+00 1.27421126e-01 -4.54858541e-01
-4.64450210e-01 -9.23886538e-01 -1.39280856e+00 1.07591510e-01
3.22302729e-02 -1.05674207e+00 1.50799131e+00 -8.85463655e-01
-1.60163021e+00 9.43595231e-01 -3.67099643e-01 -6.64506257e-01
3.48124564e-01 -6.44188344e-01 -2.10775435e-01 -2.47609273e-01
4.14595813e-01 7.59865642e-01 4.57982510e-01 -7.05969930e-01
-2.22863741e-02 -2.97612976e-02 4.76940453e-01 3.57117921e-01
-5.87052166e-01 1.35837466e-01 -1.79439366e-01 -7.21294045e-01
-6.67437434e-01 -8.06116879e-01 -5.41191101e-01 -6.28198266e-01
-5.65988719e-01 -4.29781258e-01 3.27056915e-01 -8.24069321e-01
1.51560545e+00 -1.64761364e+00 4.83512402e-01 -3.90516341e-01
4.05843928e-02 4.31366891e-01 -6.54140949e-01 8.66099775e-01
-1.23173863e-01 5.57080090e-01 -3.98012459e-01 -1.06915402e+00
1.40820771e-01 4.14181620e-01 -1.60968632e-01 2.38890439e-01
3.51594657e-01 9.39306021e-01 -1.10778892e+00 -6.53931439e-01
4.38360572e-01 9.58786979e-02 -1.17262638e+00 9.10270435e-04
-1.87197983e-01 1.72766730e-01 -6.48832042e-03 3.36459279e-01
5.00337899e-01 4.75164592e-01 1.40654773e-01 -1.34174200e-02
-2.28610933e-01 8.99541795e-01 -5.58135569e-01 2.43912530e+00
-1.15348649e+00 6.97105408e-01 -2.27835923e-01 -9.44668829e-01
7.80314863e-01 3.61364447e-02 1.50735065e-01 -7.51093268e-01
7.94662684e-02 7.42618367e-02 -5.83581217e-02 -5.60129225e-01
6.85999870e-01 -2.86019951e-01 -6.20782197e-01 5.48429728e-01
2.49161974e-01 -1.09913158e+00 8.99300039e-01 4.72672969e-01
1.33164454e+00 4.80295569e-02 2.53261805e-01 -5.60602248e-01
5.64959586e-01 2.77620643e-01 5.68373561e-01 7.83313930e-01
-3.14442702e-02 6.42195225e-01 4.51313525e-01 -4.65129077e-01
-1.60806799e+00 -9.46429312e-01 -8.11484307e-02 1.28657413e+00
-4.26043212e-01 -1.15257382e+00 -1.06281984e+00 -6.62330687e-01
1.29230134e-02 1.17063403e+00 -4.76689965e-01 -2.66238481e-01
-8.06104541e-01 -4.73180860e-01 8.16643596e-01 2.07980245e-01
-2.26710275e-01 -1.23165643e+00 1.92963406e-02 2.42545292e-01
-2.29073852e-01 -1.07135558e+00 -6.20508671e-01 2.17047691e-01
-9.86879885e-01 -8.89029682e-01 -9.96166766e-02 -8.65123749e-01
5.40247023e-01 -4.45590802e-02 1.77394938e+00 1.06799528e-01
-3.92536938e-01 1.03629092e-02 -3.31737459e-01 -6.35080159e-01
-6.74751043e-01 9.19716299e-01 3.33608121e-01 -1.02241385e+00
4.40500945e-01 -6.49060428e-01 -1.15252092e-01 -4.58632529e-01
-7.85665870e-01 2.76279785e-02 2.46898115e-01 8.69037509e-01
3.26585591e-01 -5.50009668e-01 6.36904776e-01 -1.46022880e+00
1.18279874e+00 -2.74091542e-01 -1.10866576e-01 1.84138313e-01
-6.16375208e-01 4.30668771e-01 1.30411947e+00 8.55580792e-02
-1.12923503e+00 -1.74814492e-01 -5.33170164e-01 -6.01501465e-02
-2.21308097e-01 5.63636065e-01 -2.09509164e-01 3.23245317e-01
1.02636397e+00 -3.09502274e-01 -9.59979147e-02 -7.74294794e-01
8.47018003e-01 5.91836095e-01 4.70315963e-01 -6.56818151e-01
7.30347812e-01 2.50731081e-01 -2.05725431e-01 -8.71112525e-01
-1.32426691e+00 -2.58105159e-01 -9.04930711e-01 6.02647424e-01
5.28924108e-01 -8.91831756e-01 5.77155687e-02 1.15406914e-02
-1.43203950e+00 -4.12211150e-01 -9.01603460e-01 2.55424410e-01
-5.19991577e-01 7.05734670e-01 -9.90285516e-01 -5.38574271e-02
-8.05645704e-01 -9.25982893e-01 9.27612484e-01 -5.63562326e-02
-8.82717967e-01 -1.21549094e+00 5.75076878e-01 4.16712403e-01
5.42523146e-01 7.62527660e-02 1.19767809e+00 -7.89789438e-01
-8.67071673e-02 -3.41956973e-01 4.57975343e-02 7.39011526e-01
2.43875816e-01 2.24516764e-01 -4.94541764e-01 -2.47288436e-01
-8.20960104e-02 -6.31255507e-01 7.75255144e-01 2.67314523e-01
9.60811973e-01 -2.47537777e-01 1.40873104e-01 1.06836832e+00
1.14370370e+00 -6.53294206e-01 5.44232011e-01 6.82839394e-01
8.29337955e-01 4.87298787e-01 3.79709184e-01 1.48356304e-01
6.75032318e-01 4.90708321e-01 -1.94237635e-01 -2.43544415e-01
-4.95365202e-01 -3.96792561e-01 4.19965804e-01 1.62009668e+00
2.64887631e-01 -1.99066341e-01 -7.09209085e-01 5.97647011e-01
-1.50380743e+00 -7.69873500e-01 -1.53317437e-01 1.90614486e+00
1.41660035e+00 3.47437322e-01 -2.67296702e-01 8.89874026e-02
3.20939004e-01 1.73516423e-01 -3.99912745e-01 -1.53825772e+00
-4.07959431e-01 9.10976768e-01 4.14454460e-01 1.07795513e+00
-1.08989739e+00 1.49763060e+00 6.84874725e+00 7.61041999e-01
-6.84122384e-01 2.50746608e-01 2.34331161e-01 -3.26030105e-01
-7.66096532e-01 2.41978973e-01 -7.22037196e-01 1.48459837e-01
1.11338484e+00 -4.34710324e-01 7.35854149e-01 7.36388624e-01
4.66178983e-01 9.64178368e-02 -1.30391169e+00 6.19983077e-01
2.71417469e-01 -1.33551073e+00 3.59893948e-01 -6.01598859e-01
1.00230598e+00 3.73383015e-01 -4.19552654e-01 8.16622138e-01
6.55224144e-01 -1.26550794e+00 4.12476897e-01 5.90714276e-01
8.20520401e-01 -9.06080306e-01 6.98233485e-01 4.44831014e-01
-6.54650569e-01 2.59235144e-01 -4.61385459e-01 -6.48760200e-01
2.82941848e-01 5.79344034e-01 -6.19535506e-01 4.50748146e-01
3.86965066e-01 1.16011572e+00 -9.86954451e-01 6.50708437e-01
-7.36158788e-01 5.38833380e-01 -3.09364617e-01 1.41040131e-01
2.77999928e-03 -2.07495809e-01 6.21183693e-01 1.83833957e+00
-1.17181450e-01 -4.04674590e-01 1.04861021e-01 5.02822518e-01
-3.52623105e-01 5.41689456e-01 -8.20011318e-01 8.04154575e-02
5.05104959e-01 1.07014453e+00 -1.54545560e-01 -4.20543164e-01
-4.68366235e-01 1.03255701e+00 8.37901890e-01 9.11780968e-02
-5.39180100e-01 -8.58218312e-01 6.25364184e-01 -1.56893134e-01
2.48080958e-02 -2.99249679e-01 -8.38919044e-01 -1.64939320e+00
-5.54880612e-02 -1.01050568e+00 1.66196883e-01 -4.89081502e-01
-1.42583096e+00 5.44393957e-01 -2.00692588e-03 -7.09926784e-01
-2.90698409e-01 -2.32175380e-01 -7.39017308e-01 1.09911227e+00
-1.67829144e+00 -1.06754994e+00 7.96253532e-02 3.66942972e-01
9.11706924e-01 -1.42557859e-01 1.26512027e+00 3.25460076e-01
-7.04287827e-01 9.23157454e-01 2.56740600e-01 -7.71293640e-02
1.08685791e+00 -1.44388747e+00 1.01176465e+00 8.06435466e-01
1.03722602e-01 8.05047035e-01 9.25611556e-01 -5.04252851e-01
-1.10240734e+00 -1.13257349e+00 1.79055274e+00 -9.84165251e-01
6.01108372e-01 -6.97192609e-01 -6.62482440e-01 6.42577410e-01
6.01601303e-01 -5.63253045e-01 7.46787548e-01 7.30685830e-01
-4.22782332e-01 -1.81053117e-01 -1.03514302e+00 8.03258121e-01
1.35335469e+00 -6.02020323e-01 -1.20549464e+00 8.01751971e-01
1.05472457e+00 -3.10263485e-01 -7.81549633e-01 1.80688873e-01
4.12320532e-02 -7.15323806e-01 8.65875185e-01 -1.01344728e+00
7.92158604e-01 1.30400270e-01 7.25800321e-02 -1.71508574e+00
-3.57196301e-01 -6.65690243e-01 1.12131484e-01 1.34630013e+00
5.59835196e-01 -3.33984315e-01 7.10449994e-01 5.03017306e-01
-6.93998456e-01 -5.94790578e-01 -8.32950473e-01 -8.08895588e-01
7.51687706e-01 -3.60696971e-01 6.02009714e-01 9.47926939e-01
2.05998883e-01 8.46661687e-01 -5.07818721e-02 -6.41831756e-01
4.15885538e-01 -9.94323101e-03 1.01614237e+00 -1.08956933e+00
2.21857186e-02 -4.64484423e-01 -4.92768809e-02 -7.26348877e-01
7.95627117e-01 -1.44155598e+00 -1.59943789e-01 -1.75804234e+00
2.18800977e-01 -3.25852692e-01 1.35916442e-01 4.71576840e-01
-1.99035928e-01 7.71335065e-02 4.71905291e-01 -5.19530661e-02
-8.07678759e-01 8.24993253e-01 1.10384440e+00 -2.07364470e-01
-2.56540507e-01 -5.21135151e-01 -9.79254901e-01 8.58585954e-01
9.46336687e-01 -8.01709116e-01 -2.95074552e-01 -1.23554194e+00
2.71576136e-01 -5.00801623e-01 -3.05587023e-01 -9.56536949e-01
2.11739298e-02 -2.28147395e-02 -1.09245077e-01 -1.17066778e-01
1.11096226e-01 -2.42919520e-01 -4.30599153e-01 3.27719629e-01
-5.77726662e-01 4.35813695e-01 2.99743891e-01 1.76143110e-01
-1.14977024e-01 -5.56305051e-01 7.44637609e-01 -2.58076012e-01
-2.29422793e-01 -9.57224816e-02 -4.08298522e-01 7.29243219e-01
5.04661500e-01 1.52632773e-01 -3.61833841e-01 -3.12937230e-01
-7.23838210e-01 3.34880233e-01 8.64589632e-01 6.44181728e-01
8.91039670e-02 -9.68001246e-01 -8.98309946e-01 1.37216926e-01
9.27801952e-02 -1.17575843e-03 -7.87293240e-02 3.75431567e-01
-9.33043003e-01 6.60812378e-01 -2.71902774e-02 -1.37189299e-01
-9.22724843e-01 6.89222276e-01 7.16269165e-02 -7.74718165e-01
-5.32716870e-01 6.53693736e-01 -4.25622404e-01 -1.24428248e+00
-4.22111228e-02 -5.17525911e-01 -3.31120566e-02 -9.94480476e-02
1.58618137e-01 3.11964750e-01 5.57447493e-01 -4.30571318e-01
-2.03083172e-01 3.52949947e-01 -3.44820291e-01 8.69109556e-02
1.57632494e+00 -2.39163816e-01 -2.53546655e-01 2.60506481e-01
1.30706155e+00 3.94616544e-01 -6.02328897e-01 -1.36179715e-01
2.54855484e-01 -2.32879072e-01 -3.18747401e-01 -6.89082205e-01
-5.88992000e-01 9.29602742e-01 -5.22649363e-02 -1.60177246e-01
7.85879672e-01 -3.31949085e-01 1.05069053e+00 9.58018899e-01
1.72017515e-01 -1.44867575e+00 -2.58556247e-01 1.16524172e+00
8.63452196e-01 -1.11666930e+00 2.86947310e-01 -2.37879187e-01
-6.07617140e-01 9.61751580e-01 4.39440161e-01 -6.29631996e-01
4.13600713e-01 3.25116724e-01 1.87246799e-01 7.71631747e-02
-8.61542344e-01 -7.66447634e-02 -1.38163984e-01 7.04452574e-01
8.02966356e-01 1.56560868e-01 -8.78984392e-01 8.32022071e-01
-8.43506396e-01 3.63498107e-02 7.27993965e-01 8.99187624e-01
-1.50353715e-01 -1.82851303e+00 4.63069640e-02 6.78264201e-01
-4.07910168e-01 -7.88441241e-01 -4.39712644e-01 7.55402505e-01
1.13893941e-01 1.03062785e+00 -3.89235727e-02 -1.23659134e-01
8.98049176e-01 1.46091774e-01 6.20983481e-01 -1.36077881e+00
-9.76103127e-01 -5.66625774e-01 5.88311434e-01 -3.71222645e-01
-3.94978002e-02 -7.30803192e-01 -9.86593604e-01 -6.67627394e-01
-6.34288490e-02 3.72405559e-01 4.34389323e-01 1.05665267e+00
5.63066721e-01 4.43328291e-01 4.88290787e-01 -1.03834331e+00
-6.44829035e-01 -1.15693748e+00 -6.33033887e-02 9.04480875e-01
1.57819852e-01 -1.97306484e-01 -3.62575948e-01 1.55363739e-01] | [11.07859992980957, 10.303961753845215] |
e74e92ec-b2cf-40ed-8171-5fd5f915c58b | scale-guided-hypernetwork-for-blind-super | 2306.02398 | null | https://arxiv.org/abs/2306.02398v1 | https://arxiv.org/pdf/2306.02398v1.pdf | Scale Guided Hypernetwork for Blind Super-Resolution Image Quality Assessment | With the emergence of image super-resolution (SR) algorithm, how to blindly evaluate the quality of super-resolution images has become an urgent task. However, existing blind SR image quality assessment (IQA) metrics merely focus on visual characteristics of super-resolution images, ignoring the available scale information. In this paper, we reveal that the scale factor has a statistically significant impact on subjective quality scores of SR images, indicating that the scale information can be used to guide the task of blind SR IQA. Motivated by this, we propose a scale guided hypernetwork framework that evaluates SR image quality in a scale-adaptive manner. Specifically, the blind SR IQA procedure is divided into three stages, i.e., content perception, evaluation rule generation, and quality prediction. After content perception, a hypernetwork generates the evaluation rule used in quality prediction based on the scale factor of the SR image. We apply the proposed scale guided hypernetwork framework to existing representative blind IQA metrics, and experimental results show that the proposed framework not only boosts the performance of these IQA metrics but also enhances their generalization abilities. Source code will be available at https://github.com/JunFu1995/SGH. | ['Jun Fu'] | 2023-06-04 | null | null | null | null | ['image-super-resolution', 'image-quality-assessment', 'super-resolution'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.75483599e-01 -4.12828565e-01 4.46524285e-02 -2.71819174e-01
-7.32347906e-01 -3.87740344e-01 2.16909885e-01 -2.58870244e-01
-2.19367206e-01 4.51526821e-01 5.33204973e-01 -1.78407416e-01
-4.49276865e-01 -8.37560713e-01 -1.72696739e-01 -6.53440237e-01
1.99644774e-01 -2.64719218e-01 3.61313403e-01 -2.55004048e-01
4.66863394e-01 3.09632272e-01 -1.66159606e+00 2.44907036e-01
1.35070610e+00 1.09193599e+00 4.21345353e-01 5.74769795e-01
-4.38865572e-02 6.81531549e-01 -5.72358251e-01 -1.76954061e-01
4.52619314e-01 -5.59719741e-01 -6.76994562e-01 -1.34243086e-01
2.80118287e-01 -4.48443472e-01 -3.75163943e-01 1.65243196e+00
8.14258635e-01 8.85373875e-02 4.21536922e-01 -8.40711296e-01
-9.91823733e-01 3.83730084e-01 -3.73311400e-01 7.20029056e-01
4.83171016e-01 1.80681154e-01 9.07657862e-01 -8.85945618e-01
2.20807001e-01 1.34219992e+00 3.00847560e-01 3.91493767e-01
-8.75336111e-01 -8.09026062e-01 -1.49941919e-02 6.11328721e-01
-1.38113999e+00 -4.72411364e-01 6.92589760e-01 -3.36294085e-01
3.12600642e-01 2.33733252e-01 6.42988622e-01 5.79122722e-01
-4.26344238e-02 4.50915962e-01 1.72065437e+00 -2.66211987e-01
3.71559709e-01 -2.14652326e-02 -2.60661170e-02 3.93725842e-01
2.83883721e-01 4.66066122e-01 -5.07763445e-01 8.51871297e-02
1.31744170e+00 -1.35762364e-01 -5.80621600e-01 -1.33407682e-01
-1.06705439e+00 5.96541882e-01 9.14427757e-01 3.49471211e-01
-4.11289245e-01 -4.08730417e-01 -5.63118011e-02 3.76064599e-01
2.61437833e-01 4.84803349e-01 1.90725550e-02 2.35250667e-01
-7.20041394e-01 -5.36203524e-03 1.58151820e-01 7.03067124e-01
5.70038974e-01 2.97161937e-02 -4.22785223e-01 1.06739306e+00
1.80136457e-01 5.60953856e-01 6.20136499e-01 -1.17901957e+00
2.20988050e-01 7.25777209e-01 4.48409140e-01 -9.09253657e-01
-4.06081468e-01 -5.19443214e-01 -1.04115379e+00 7.40050793e-01
2.30168909e-01 9.45117623e-02 -1.02943385e+00 1.47059000e+00
2.18705326e-01 -1.40652314e-01 9.47118104e-02 1.40040946e+00
1.15532982e+00 4.27362084e-01 -5.44721670e-02 -4.69670892e-01
1.46204245e+00 -7.02726305e-01 -7.05300272e-01 7.35636279e-02
-3.03249508e-01 -8.89119744e-01 1.27518165e+00 3.79805595e-01
-1.32594848e+00 -8.90470326e-01 -1.12007797e+00 6.41166866e-02
-7.60544240e-02 1.11050323e-01 2.65054017e-01 4.74162698e-01
-1.42249501e+00 4.60111499e-01 -3.40644598e-01 -3.25826854e-01
4.16197389e-01 2.74804801e-01 -1.01830296e-01 -2.43461668e-01
-1.29443419e+00 7.29126334e-01 3.84528697e-01 2.83592520e-03
-7.28746235e-01 -4.64367002e-01 -4.37692076e-01 -1.25259116e-01
1.70396253e-01 -8.60125959e-01 1.13996828e+00 -8.51728857e-01
-1.39887989e+00 6.04452729e-01 -3.40267390e-01 -1.68888599e-01
4.59777296e-01 4.17545773e-02 -8.67379785e-01 4.38164145e-01
8.36666897e-02 3.54312688e-01 1.22639191e+00 -1.58066785e+00
-9.76040065e-01 -5.48372924e-01 3.27849925e-01 5.69476306e-01
-3.68317634e-01 3.16938668e-01 -4.68933403e-01 -5.98692358e-01
3.57177287e-01 -2.94126213e-01 1.69372559e-03 9.97176841e-02
-1.60742000e-01 -7.92296827e-02 2.96658427e-01 -9.49634492e-01
1.38435304e+00 -2.27304530e+00 7.20475242e-02 2.59153575e-01
5.66613615e-01 4.11872983e-01 -2.31665432e-01 -1.81229442e-01
2.15665381e-02 9.65664834e-02 -1.96691930e-01 2.34242901e-01
-2.66006351e-01 -3.29910249e-01 1.73053235e-01 2.61718363e-01
-1.64833412e-01 5.19398570e-01 -7.95411229e-01 -7.16888130e-01
2.02975258e-01 6.67866647e-01 -2.68563062e-01 4.97819364e-01
2.74156392e-01 6.90493882e-01 -4.47325200e-01 7.37045527e-01
8.67696285e-01 -5.06668389e-01 -2.94051021e-01 -5.94073534e-01
-2.79033154e-01 -2.10915990e-02 -1.28071284e+00 1.30828774e+00
-4.52480227e-01 2.41581395e-01 1.31889641e-01 -3.99278075e-01
1.07927799e+00 3.05270016e-01 1.76530808e-01 -1.22076309e+00
4.02217135e-02 1.40555501e-01 -2.52807327e-03 -5.27248502e-01
4.00315374e-01 -5.38476445e-02 6.20917559e-01 1.88215241e-01
-3.39428425e-01 6.16293140e-02 9.98878703e-02 2.75139678e-02
7.53853619e-01 -2.91658103e-01 4.01213795e-01 -2.68698973e-03
8.59079242e-01 -2.48728275e-01 5.29855907e-01 5.79544425e-01
-5.08013129e-01 7.41082191e-01 1.35542184e-01 -2.99534291e-01
-1.30209184e+00 -1.37983775e+00 -2.37133935e-01 8.92397821e-01
7.18348026e-01 -1.79613307e-02 -8.52019727e-01 -2.73346007e-01
-3.29659283e-01 3.58930022e-01 -5.61441183e-01 2.55866218e-02
-3.21082562e-01 -7.76808381e-01 -1.63844097e-02 2.86887527e-01
1.12297022e+00 -1.20957148e+00 -1.97870508e-01 -2.68770933e-01
-3.23334605e-01 -9.76919413e-01 -4.22423512e-01 -6.91690564e-01
-9.32724833e-01 -9.97958481e-01 -9.67452168e-01 -9.19097304e-01
7.17207134e-01 6.96140409e-01 8.83492172e-01 1.49625018e-01
-8.08798522e-02 1.42662600e-01 -4.86259162e-01 3.49301435e-02
-3.50184709e-01 -2.70988852e-01 1.33641198e-01 9.96449441e-02
1.45092949e-01 -7.13326156e-01 -1.38262427e+00 5.00419021e-01
-8.39552224e-01 5.49011305e-02 1.03836632e+00 4.38963115e-01
8.16211641e-01 5.93442082e-01 6.79310560e-01 -3.83577585e-01
8.72721553e-01 -1.06018484e-01 -7.87891626e-01 1.92058489e-01
-9.56892371e-01 -1.44257084e-01 5.86281955e-01 -3.59746575e-01
-1.23052168e+00 -3.77661318e-01 3.67804542e-02 -2.35971212e-01
-5.23896776e-02 2.03402713e-01 -3.80620003e-01 -4.71327275e-01
7.48451054e-01 4.69028264e-01 -1.15156593e-02 -5.39313495e-01
2.48871550e-01 9.18034673e-01 7.22429216e-01 -7.48512894e-02
1.03277242e+00 4.43672270e-01 -1.57481357e-01 -5.32491565e-01
-6.84391260e-01 -6.00304067e-01 -3.07601333e-01 -2.61643738e-01
9.42509830e-01 -1.03079844e+00 -7.63749123e-01 4.42555696e-01
-7.24268913e-01 -3.50140147e-02 -7.95161724e-02 5.01992047e-01
-4.34079111e-01 4.80362445e-01 -5.47846913e-01 -6.55530453e-01
-5.22690058e-01 -1.21644711e+00 6.10729516e-01 7.24004388e-01
2.41047651e-01 -6.10599577e-01 -1.87977880e-01 5.94212592e-01
6.93811238e-01 -1.90640256e-01 8.76492977e-01 -7.41753802e-02
-7.50981331e-01 4.99832965e-02 -9.36567605e-01 4.50296164e-01
3.58630002e-01 -5.26148617e-01 -9.05764043e-01 -4.18748528e-01
1.10007435e-01 1.03475742e-01 5.68472385e-01 5.08002639e-01
1.11032009e+00 -4.74558741e-01 2.74588048e-01 8.76137197e-01
1.66369510e+00 2.62877345e-01 8.05347919e-01 5.81170440e-01
5.97737074e-01 4.54744518e-01 5.43924689e-01 2.79340386e-01
2.99667776e-01 7.25180447e-01 5.11363804e-01 -1.42908767e-01
-5.47430277e-01 -2.41459638e-01 2.62817264e-01 7.03862667e-01
-5.70651352e-01 2.28197262e-01 -6.50280893e-01 4.23660696e-01
-1.32418799e+00 -9.29003119e-01 9.91595313e-02 2.37192822e+00
8.44876051e-01 -2.25650743e-02 2.55939901e-01 1.90583318e-01
9.99442220e-01 1.18591353e-01 -8.14471185e-01 1.00741848e-01
-2.57909119e-01 -2.50082780e-02 3.92687291e-01 5.63916564e-01
-8.73380840e-01 7.66301334e-01 5.76094723e+00 7.72663057e-01
-8.64622355e-01 1.24991462e-01 3.14477980e-01 1.68260571e-03
-2.90319473e-01 -2.13780880e-01 -5.11167943e-01 4.62328255e-01
5.16293168e-01 -5.11386096e-01 1.00172639e+00 5.26024342e-01
6.23224437e-01 1.39257327e-01 -4.43139583e-01 1.25718784e+00
5.28990757e-03 -8.66265535e-01 4.31424439e-01 -4.49669175e-02
7.12928593e-01 -1.81125224e-01 4.57296491e-01 -2.14219496e-01
3.32421213e-01 -1.17468894e+00 3.66087109e-01 6.38172686e-01
1.03192616e+00 -7.20182359e-01 8.14305782e-01 -1.02217890e-01
-1.29362488e+00 -4.77497965e-01 -5.41493833e-01 2.27891251e-01
-8.81618485e-02 5.44608474e-01 -4.17091966e-01 5.99810183e-01
1.09018826e+00 5.00566542e-01 -9.22802925e-01 1.50998199e+00
-4.81260657e-01 3.16913754e-01 3.85829896e-01 4.10036176e-01
-3.87656927e-01 -1.11061975e-01 7.78473973e-01 6.18437052e-01
4.40966427e-01 4.76344168e-01 -1.83095649e-01 9.12570000e-01
7.07536340e-02 3.78650874e-01 -1.78960294e-01 2.66165018e-01
6.94217443e-01 1.18936515e+00 -4.08872426e-01 -1.64514139e-01
-5.08394182e-01 9.24007297e-01 -1.05416030e-01 7.20348001e-01
-3.15651774e-01 -4.50471252e-01 4.92950916e-01 1.42186403e-01
1.94996610e-01 5.48481792e-02 -5.05794585e-01 -1.00092602e+00
1.08396322e-01 -1.22663653e+00 3.45563918e-01 -1.21121454e+00
-1.28312957e+00 9.60414350e-01 -2.77345985e-01 -1.63910902e+00
1.67216554e-01 -4.03846323e-01 -5.17296612e-01 1.23882210e+00
-1.74397624e+00 -9.57021654e-01 -9.01732028e-01 9.84782815e-01
2.75926292e-01 -2.79455215e-01 5.03935039e-01 9.43123773e-02
-4.18720216e-01 6.96376443e-01 8.76877755e-02 1.34301171e-01
6.74850643e-01 -1.20689535e+00 1.18540816e-01 1.11648524e+00
-3.22628856e-01 4.64935094e-01 7.32554197e-01 -5.60756385e-01
-8.91651988e-01 -9.44678187e-01 4.52207357e-01 -1.60077810e-01
4.97336268e-01 2.03218192e-01 -1.06731987e+00 -1.27098465e-03
-1.42072961e-01 -4.55766357e-02 2.62845308e-01 -8.81337300e-02
-4.53584194e-01 -3.96603078e-01 -1.36631787e+00 5.42146981e-01
1.09448075e+00 -6.34716988e-01 -5.03287494e-01 4.69931029e-03
8.98734033e-01 -1.41040459e-01 -1.15618515e+00 5.43731153e-01
4.24526751e-01 -1.25341725e+00 1.26499677e+00 2.20503341e-02
4.26469982e-01 -6.66058004e-01 -1.89321771e-01 -1.23701441e+00
-7.62101054e-01 -2.56123751e-01 -9.65401307e-02 9.89683986e-01
2.33965278e-01 -7.02993751e-01 3.42929870e-01 3.91436428e-01
9.11528319e-02 -3.54303807e-01 -7.50411272e-01 -7.37655640e-01
-1.75495699e-01 1.30185530e-01 8.44327211e-01 6.64791167e-01
-2.57725328e-01 1.80827573e-01 -1.83374539e-01 7.57828474e-01
1.12094235e+00 2.11534455e-01 3.04717362e-01 -1.30104804e+00
-1.99248776e-01 -6.73885465e-01 -4.72793162e-01 -6.51791155e-01
-5.41647494e-01 -6.16790295e-01 -1.12811789e-01 -2.00656271e+00
3.76468122e-01 -1.98393673e-01 -6.47573590e-01 2.74372809e-02
-5.22668839e-01 4.16558832e-01 1.25976592e-01 7.47200608e-01
-6.66759789e-01 3.32877934e-01 1.82386672e+00 2.89764907e-02
-5.00919402e-01 1.48430333e-01 -9.76098418e-01 6.32802904e-01
1.03319621e+00 1.44984871e-01 -4.79635119e-01 -3.22821826e-01
3.00425231e-01 3.37911814e-01 5.74274421e-01 -1.27720273e+00
2.81996340e-01 -1.17454037e-01 5.76616466e-01 -1.13861062e-01
6.68821111e-02 -6.37239337e-01 -1.99339613e-01 3.17967862e-01
-3.28251094e-01 -1.77090578e-02 -4.02617343e-02 4.38262910e-01
-3.54783654e-01 1.16201721e-01 1.14380777e+00 -1.48741215e-01
-8.89587283e-01 6.06302977e-01 1.71539843e-01 -6.24181796e-03
4.13459986e-01 -2.50212312e-01 -6.10514641e-01 -4.46703106e-01
-7.90058732e-01 2.51843929e-02 6.09134734e-01 4.34437692e-01
1.00964713e+00 -1.36215889e+00 -9.17180419e-01 1.80933088e-01
3.01721394e-01 -7.60852173e-02 6.84286594e-01 6.41111374e-01
-3.06382328e-01 1.79787040e-01 -4.16431695e-01 -3.36603880e-01
-1.33233941e+00 8.16845894e-01 4.59219217e-01 5.87858893e-02
-6.34131253e-01 6.40395284e-01 4.30976450e-01 1.40490666e-01
2.71173060e-01 -4.80271466e-02 -9.57070768e-01 -2.75923371e-01
1.06689644e+00 5.45925438e-01 -1.85953915e-01 -7.68173397e-01
-7.63031170e-02 9.52221572e-01 -1.12922020e-01 -1.76243350e-01
1.02736032e+00 -6.69116259e-01 -2.61484325e-01 1.11126304e-01
7.59560466e-01 -8.20921808e-02 -1.14120448e+00 -5.01966536e-01
-3.92216891e-01 -7.93046951e-01 3.40000361e-01 -1.17027712e+00
-1.14752579e+00 6.81196749e-01 1.33706594e+00 2.93463618e-01
1.85461915e+00 7.04348907e-02 6.92528069e-01 -1.76132098e-01
4.22987849e-01 -8.58122289e-01 2.15420842e-01 1.02609058e-03
1.09460449e+00 -1.44634473e+00 -1.22784600e-02 -4.48709518e-01
-4.72926140e-01 7.43880630e-01 6.83056295e-01 5.68766445e-02
3.71466368e-01 -2.95049369e-01 3.35920483e-01 -4.71950024e-02
-1.13792278e-01 -4.12009448e-01 5.27994156e-01 9.84155893e-01
1.14174008e-01 1.38416722e-01 -2.63953239e-01 5.83388090e-01
-4.51629937e-01 2.22028479e-01 3.82827938e-01 2.32748568e-01
-8.14797282e-01 -7.91530430e-01 -6.51754677e-01 4.94622737e-01
-3.95890474e-01 -3.63337487e-01 -8.03473815e-02 1.89447150e-01
9.90411565e-02 1.21053779e+00 -2.65667349e-01 -6.30257428e-01
4.46192712e-01 -4.67139959e-01 3.16005319e-01 -2.62759447e-01
-3.33009213e-02 -1.32135034e-01 -3.07973295e-01 -6.78458214e-01
-5.25428236e-01 -2.99909800e-01 -1.01950812e+00 -3.60638320e-01
1.34615705e-03 7.50224888e-02 5.00641823e-01 5.68687677e-01
1.92295939e-01 5.09990096e-01 8.30375969e-01 -4.67040598e-01
-3.75183970e-01 -1.03051388e+00 -6.85683429e-01 3.69678676e-01
6.34239733e-01 -5.32282710e-01 -6.79345906e-01 -1.63751677e-01] | [11.753117561340332, -1.939570665359497] |
6095fec7-28e6-4de0-b544-8286f111e4bd | large-scale-long-tailed-recognition-in-an | 1904.05160 | null | http://arxiv.org/abs/1904.05160v2 | http://arxiv.org/pdf/1904.05160v2.pdf | Large-Scale Long-Tailed Recognition in an Open World | Real world data often have a long-tailed and open-ended distribution. A
practical recognition system must classify among majority and minority classes,
generalize from a few known instances, and acknowledge novelty upon a never
seen instance. We define Open Long-Tailed Recognition (OLTR) as learning from
such naturally distributed data and optimizing the classification accuracy over
a balanced test set which include head, tail, and open classes. OLTR must
handle imbalanced classification, few-shot learning, and open-set recognition
in one integrated algorithm, whereas existing classification approaches focus
only on one aspect and deliver poorly over the entire class spectrum. The key
challenges are how to share visual knowledge between head and tail classes and
how to reduce confusion between tail and open classes. We develop an integrated
OLTR algorithm that maps an image to a feature space such that visual concepts
can easily relate to each other based on a learned metric that respects the
closed-world classification while acknowledging the novelty of the open world.
Our so-called dynamic meta-embedding combines a direct image feature and an
associated memory feature, with the feature norm indicating the familiarity to
known classes. On three large-scale OLTR datasets we curate from object-centric
ImageNet, scene-centric Places, and face-centric MS1M data, our method
consistently outperforms the state-of-the-art. Our code, datasets, and models
enable future OLTR research and are publicly available at
https://liuziwei7.github.io/projects/LongTail.html. | ['Zhongqi Miao', 'Jiayun Wang', 'Boqing Gong', 'Ziwei Liu', 'Xiaohang Zhan', 'Stella X. Yu'] | 2019-04-10 | large-scale-long-tailed-recognition-in-an-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Liu_Large-Scale_Long-Tailed_Recognition_in_an_Open_World_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Liu_Large-Scale_Long-Tailed_Recognition_in_an_Open_World_CVPR_2019_paper.pdf | cvpr-2019-6 | ['long-tail-learning-with-class-descriptors'] | ['methodology'] | [ 2.64514443e-02 -2.64867514e-01 -4.23379272e-01 -6.57070756e-01
-7.44302690e-01 -5.51470637e-01 6.04851305e-01 8.49636123e-02
-2.57569075e-01 3.05431843e-01 1.95693672e-01 1.39656335e-01
-2.55462110e-01 -6.29083395e-01 -5.25242805e-01 -5.73995411e-01
-9.49223712e-03 5.58949888e-01 -1.67995781e-01 1.02515921e-01
1.89443097e-01 4.19121921e-01 -1.94628048e+00 6.30370200e-01
5.91417432e-01 1.39081800e+00 -2.00977623e-01 5.11529803e-01
-3.74719709e-01 7.55409896e-01 -3.91338974e-01 -4.51527029e-01
4.74325746e-01 -1.86022699e-01 -5.65280795e-01 7.56664798e-02
1.05161166e+00 1.29518494e-01 -2.48319373e-01 8.91443670e-01
6.92765057e-01 1.07913993e-01 8.99112165e-01 -1.60509276e+00
-1.15342534e+00 2.95970198e-02 -5.86100399e-01 3.52670372e-01
3.91894102e-01 2.55795687e-01 1.06524634e+00 -1.38923228e+00
8.50551069e-01 9.97819006e-01 6.52687788e-01 4.72012967e-01
-1.31206262e+00 -7.35622227e-01 3.58947039e-01 6.41253889e-01
-1.61200845e+00 -4.40338820e-01 5.32347143e-01 -6.92358613e-01
8.97455215e-01 4.01115745e-01 5.76320648e-01 1.05310142e+00
3.07239264e-01 5.93025565e-01 1.14717424e+00 -9.48245823e-02
4.82287109e-01 2.88450420e-01 2.80294061e-01 5.47190368e-01
1.70192972e-01 5.85053563e-02 -8.61261249e-01 -3.90294008e-02
-2.14557406e-02 5.54184854e-01 -2.31745303e-01 -8.46616447e-01
-1.09326947e+00 6.61886334e-01 6.48388207e-01 1.25837252e-01
-1.38486475e-01 -3.68783712e-01 2.74892926e-01 6.31062508e-01
6.69326842e-01 3.37948889e-01 -4.81186628e-01 -4.57209982e-02
-9.25137877e-01 -1.41009927e-01 9.51323271e-01 7.25505650e-01
1.07131791e+00 -2.13432655e-01 -5.44427037e-02 9.74219501e-01
1.46178886e-01 5.49461603e-01 7.49853313e-01 -5.42070925e-01
1.02288894e-01 8.17461193e-01 -4.31208730e-01 -1.05770206e+00
-8.29980299e-02 -7.34185040e-01 -6.42062128e-01 2.52158344e-01
2.11698741e-01 2.89935648e-01 -1.02930868e+00 1.62949955e+00
3.41153383e-01 1.58916593e-01 8.30763578e-02 6.08836830e-01
9.98414516e-01 5.86390018e-01 -2.29574323e-01 -6.74038604e-02
9.84888494e-01 -1.04042637e+00 -3.62567395e-01 -3.53573710e-01
7.23789454e-01 -5.16176701e-01 1.22681463e+00 2.79939294e-01
-5.54968178e-01 -4.60694581e-01 -9.29830015e-01 -6.64278269e-02
-9.47425067e-01 -1.78401843e-01 3.45808536e-01 5.40535092e-01
-1.03722656e+00 2.40189761e-01 -2.93986082e-01 -5.00483751e-01
8.19595397e-01 7.93340132e-02 -9.53897238e-01 -4.66332316e-01
-6.58408582e-01 9.37121212e-01 2.43740752e-02 -2.62621921e-02
-8.76459241e-01 -1.00936162e+00 -9.18952405e-01 -1.85480248e-02
2.97161371e-01 -3.25452328e-01 8.39453340e-01 -1.21859276e+00
-7.76814342e-01 1.19546270e+00 -3.79174016e-02 -1.58512756e-01
3.00873250e-01 4.62687165e-02 -6.90111876e-01 -1.31424904e-01
1.35969028e-01 6.82326913e-01 9.99584198e-01 -1.08430684e+00
-3.55921537e-01 -8.07303071e-01 -2.75336266e-01 2.42582053e-01
-6.89405143e-01 -3.90910029e-01 8.24923441e-03 -4.18547124e-01
1.81080863e-01 -7.34557211e-01 2.45880902e-01 5.66037357e-01
-1.38068840e-01 -1.51912004e-01 9.57025230e-01 -2.46934414e-01
1.04235125e+00 -2.44618821e+00 1.40064836e-01 4.21798341e-02
4.34546858e-01 1.04733594e-01 -5.03575265e-01 4.83181864e-01
-4.74590003e-01 -1.26899168e-01 -7.94380084e-02 -1.26978070e-01
3.74860018e-02 1.78390816e-01 -4.92101759e-01 6.83398724e-01
7.48978183e-03 8.89244556e-01 -8.00095320e-01 -2.72064120e-01
7.42894337e-02 3.38992417e-01 -4.74113256e-01 3.14310402e-01
-5.67853414e-02 -4.55387533e-02 2.78240860e-01 1.05626333e+00
6.44177735e-01 -1.99775308e-01 -1.86682180e-01 -2.75314540e-01
-5.86692896e-03 -2.41158500e-01 -1.18723977e+00 1.48618436e+00
-3.89382243e-01 7.41488814e-01 -3.27722996e-01 -8.79987836e-01
1.11543858e+00 -2.43292689e-01 2.59941310e-01 -8.36300433e-01
6.81422418e-03 2.57984161e-01 2.14092452e-02 -4.39165205e-01
1.92170247e-01 -1.25202626e-01 1.27769887e-01 3.79976600e-01
4.07985955e-01 9.63128060e-02 -8.22739601e-02 2.54981965e-01
1.03956044e+00 -2.91107118e-01 3.46130997e-01 -1.72368661e-01
7.87609741e-02 -2.23137006e-01 6.04758501e-01 7.21214533e-01
-5.59655488e-01 7.37093508e-01 4.43455338e-01 -8.05033684e-01
-5.87145984e-01 -1.44138479e+00 -4.00222570e-01 1.31041789e+00
1.18630894e-01 -3.38249534e-01 -3.64208847e-01 -9.06477451e-01
4.02274907e-01 5.33643186e-01 -9.69888330e-01 -4.40298796e-01
2.94632554e-01 -3.72662455e-01 3.38692933e-01 4.15775359e-01
2.31281385e-01 -8.32208931e-01 -3.95751894e-01 -2.41473868e-01
1.17850907e-01 -6.99752152e-01 -6.83743894e-01 3.65246415e-01
-4.81318355e-01 -1.16755831e+00 -5.83049059e-01 -7.74189115e-01
6.83309138e-01 3.58358532e-01 1.13878381e+00 -1.80589795e-01
-7.91808486e-01 6.59953594e-01 -2.62734830e-01 -4.05507028e-01
2.41310969e-01 -1.62541270e-01 2.92883158e-01 6.41878188e-01
6.61495149e-01 -4.12020236e-01 -5.04982352e-01 4.53777045e-01
-7.57575631e-01 -1.52278021e-01 4.99888211e-01 8.90866637e-01
6.51052833e-01 -3.44207048e-01 6.41348183e-01 -7.36616850e-01
1.45980611e-01 -8.37641537e-01 -1.22626722e-01 4.50365156e-01
-7.86105216e-01 -1.76237509e-01 3.25753570e-01 -7.17740953e-01
-6.05344236e-01 -9.55211818e-02 1.85495079e-01 -7.03326583e-01
-2.00147197e-01 3.74453127e-01 -3.27408582e-01 -2.54785150e-01
8.40309083e-01 3.96644711e-01 1.43256292e-01 -3.14711839e-01
5.78095913e-01 8.67065310e-01 4.50543165e-01 -3.83659869e-01
7.39783287e-01 6.69527650e-01 -4.48131949e-01 -9.10598516e-01
-1.14588392e+00 -6.42887354e-01 -7.12290764e-01 -4.20429438e-01
3.49068135e-01 -1.03066373e+00 -3.96182954e-01 5.49882174e-01
-5.94037712e-01 -1.83115721e-01 -8.38097155e-01 2.23028123e-01
-4.34297502e-01 7.64442235e-02 2.20579617e-02 -5.07375777e-01
-1.76963449e-01 -7.67959654e-01 9.69263494e-01 3.06589723e-01
-2.77354449e-01 -9.15899158e-01 3.58595818e-01 3.82521421e-01
4.68705535e-01 9.43457782e-02 6.89410865e-01 -1.05050349e+00
-4.52352464e-01 -5.31448007e-01 -3.50568712e-01 4.63845879e-01
1.13380514e-01 6.58851117e-02 -1.17821479e+00 -5.74781597e-01
-1.04873933e-01 -7.91140258e-01 1.14400232e+00 -5.72312735e-02
1.24214292e+00 -2.45369226e-01 -1.51703864e-01 8.93697381e-01
1.36254489e+00 -1.16774008e-01 4.97222602e-01 1.38586044e-01
6.52570724e-01 5.40373087e-01 4.97213900e-01 3.73975009e-01
6.09492123e-01 3.17930877e-01 3.73036087e-01 3.37268203e-03
-2.16336548e-01 -2.74550378e-01 4.43289727e-01 6.83203042e-01
5.24929345e-01 -3.81883048e-02 -1.05020618e+00 7.46383429e-01
-1.58458161e+00 -1.10384333e+00 4.29381073e-01 2.24855876e+00
7.58563101e-01 -3.47270593e-02 -1.28817439e-01 1.30276456e-02
5.76128662e-01 3.53271663e-01 -8.01605165e-01 -4.45250392e-01
-3.16826254e-01 -8.31719767e-03 8.30174983e-02 1.90141365e-01
-8.87260318e-01 5.72430432e-01 5.75078344e+00 9.68972564e-01
-1.37967610e+00 2.28635520e-01 7.89238214e-01 -5.10993719e-01
-2.94507116e-01 1.69634297e-02 -7.97020435e-01 4.11381513e-01
8.04168105e-01 -3.31393868e-01 3.11147243e-01 1.07870841e+00
-4.85607147e-01 -1.63751245e-02 -1.55057561e+00 1.48553658e+00
6.80257142e-01 -1.32958591e+00 2.16027245e-01 1.44710764e-01
7.56603122e-01 3.15563321e-01 3.83012533e-01 5.29025376e-01
-8.54585096e-02 -1.18439281e+00 6.15180433e-01 7.76785195e-01
1.02142251e+00 -3.28007251e-01 4.65104163e-01 3.38088423e-01
-1.08322191e+00 -3.83574456e-01 -3.09676081e-01 -2.16122091e-01
-5.14842331e-01 5.97885609e-01 -3.62987578e-01 1.36409670e-01
7.85480142e-01 1.17876363e+00 -8.79586458e-01 1.12616169e+00
2.58379996e-01 2.06012815e-01 -1.18037097e-01 2.11268082e-01
-3.66853848e-02 -5.50568290e-02 3.50592077e-01 9.17113662e-01
4.83342744e-02 -1.79995105e-01 4.19494539e-01 6.39929712e-01
-1.69463277e-01 1.69145584e-01 -1.00567484e+00 -1.37315542e-01
3.76808047e-01 1.31663275e+00 -4.00673628e-01 -2.55401522e-01
-5.90993702e-01 8.50588679e-01 5.92867374e-01 2.67408550e-01
-5.39008558e-01 -4.79097545e-01 1.01320148e+00 9.96179953e-02
2.46636569e-01 1.09893046e-01 -3.22274864e-01 -1.52993298e+00
2.34874934e-01 -8.85613501e-01 8.85430813e-01 -6.72812819e-01
-1.74391723e+00 7.33945727e-01 -2.95563996e-01 -1.29577506e+00
1.32097870e-01 -6.87949479e-01 -4.47051316e-01 5.04400551e-01
-1.54889715e+00 -9.93040144e-01 -5.29018998e-01 7.07514048e-01
5.17176926e-01 -4.21869338e-01 8.66204262e-01 2.88589031e-01
-5.97125947e-01 9.01263893e-01 3.37790072e-01 1.38749868e-01
9.72883701e-01 -1.00020874e+00 -4.08806317e-02 4.37188953e-01
5.26748002e-01 3.86382520e-01 3.42420608e-01 -5.37058353e-01
-1.56309271e+00 -1.22935271e+00 8.59850705e-01 -8.53651464e-01
7.71964788e-01 -5.31209171e-01 -1.01097870e+00 6.27324522e-01
-1.36977270e-01 8.41263950e-01 1.02179301e+00 2.63264716e-01
-1.26455665e+00 -6.09525621e-01 -1.17622662e+00 3.60285282e-01
1.12025940e+00 -9.15830553e-01 -6.69619918e-01 2.73830652e-01
3.49630624e-01 -1.99372783e-01 -8.90113771e-01 2.75549650e-01
8.34715068e-01 -9.67952371e-01 8.68121624e-01 -8.70140851e-01
2.99237877e-01 -2.20384344e-01 -7.21665084e-01 -1.43872249e+00
-3.55532795e-01 -1.05728559e-01 -2.10620508e-01 9.73673522e-01
2.82712460e-01 -9.43310618e-01 5.50549686e-01 4.57517534e-01
1.64906576e-01 -1.04651618e+00 -1.08370900e+00 -1.09322298e+00
4.52062376e-02 -3.44588161e-01 3.39717478e-01 1.07402313e+00
-6.32104203e-02 1.74350992e-01 -2.05813646e-01 8.49469230e-02
7.82415390e-01 5.24157882e-01 7.70687044e-01 -1.41975152e+00
2.71590408e-02 -2.11090386e-01 -1.10248375e+00 -4.08928186e-01
2.88554251e-01 -1.29371703e+00 -1.47221744e-01 -1.15760052e+00
6.85305238e-01 -3.31489325e-01 -5.16255498e-01 6.61639214e-01
1.95881933e-01 5.72558701e-01 2.22828701e-01 2.92997330e-01
-1.00337088e+00 7.98442900e-01 7.84406006e-01 -3.73895079e-01
8.49959403e-02 -3.76651555e-01 -8.45966637e-01 6.35401487e-01
5.92895508e-01 -4.84266341e-01 -4.11532313e-01 -3.63942951e-01
7.53423274e-02 -3.68696243e-01 6.46937191e-01 -1.15312362e+00
4.80631500e-01 -2.33859494e-01 6.94975257e-01 -4.46202368e-01
4.52239543e-01 -8.05052042e-01 1.15701027e-01 2.35157803e-01
-5.47265351e-01 -2.90004071e-02 4.83519286e-02 7.59917736e-01
-2.01720804e-01 2.72875249e-01 9.68524635e-01 1.10223487e-01
-1.04462683e+00 8.06802630e-01 9.55833346e-02 5.31707883e-01
1.29837942e+00 -5.23138642e-01 -7.41595745e-01 -2.89514542e-01
-7.77201712e-01 2.47226924e-01 7.32294261e-01 8.97901952e-01
8.45797658e-01 -1.35594738e+00 -6.48172557e-01 6.83003306e-01
9.91625786e-01 -5.30196726e-01 4.52957302e-01 6.93156362e-01
-3.81741747e-02 7.33538112e-03 -4.28185791e-01 -8.25333655e-01
-1.17327046e+00 6.14884853e-01 4.42719877e-01 2.72020727e-01
-5.04318476e-01 8.76787245e-01 3.48618329e-01 -8.80598783e-01
4.05785173e-01 1.55087590e-01 8.26754645e-02 5.36629677e-01
9.10983205e-01 2.33294055e-01 1.74755812e-01 -8.70145202e-01
-6.10570729e-01 5.90845764e-01 -2.54707664e-01 2.68746406e-01
1.36577940e+00 -3.14759053e-02 -2.21800253e-01 1.01745272e+00
1.59549713e+00 -3.24418992e-01 -1.09738386e+00 -5.46551704e-01
-1.49340138e-01 -9.21594799e-01 3.10148932e-02 -9.17980969e-01
-1.03122437e+00 9.57972169e-01 9.91223812e-01 -3.03970221e-02
1.04339910e+00 1.86194479e-01 3.65535140e-01 4.57144588e-01
3.13980192e-01 -1.04968774e+00 3.00601780e-01 5.55316329e-01
9.99388993e-01 -1.43288946e+00 -1.80135503e-01 -2.00158614e-03
-6.99738204e-01 9.30734098e-01 9.38811839e-01 9.29074213e-02
9.40949380e-01 1.79399624e-01 3.28613102e-01 -2.61257291e-01
-1.23351121e+00 -1.98972642e-01 3.85919154e-01 7.50537694e-01
1.37615874e-01 1.37859941e-01 3.90931964e-01 3.93781185e-01
-1.41666774e-02 -3.78946625e-02 3.85752708e-01 7.62117326e-01
-4.98752475e-01 -5.94785988e-01 -1.88459814e-01 8.95260096e-01
1.22002594e-01 -5.95090576e-02 -4.64964271e-01 2.89755791e-01
2.11735904e-01 6.92577124e-01 3.62608492e-01 -5.85681498e-01
2.92659819e-01 3.10297549e-01 2.19099641e-01 -8.63058984e-01
-1.99432865e-01 -6.88726783e-01 -2.37720847e-01 -8.07638884e-01
6.86725900e-02 -6.23429239e-01 -6.50035441e-01 -3.10738891e-01
-1.83790073e-01 -1.01617143e-01 6.11813605e-01 7.28844941e-01
7.70842075e-01 5.84927388e-02 9.12212193e-01 -6.95291460e-01
-7.30650246e-01 -5.50771356e-01 -7.47238398e-01 5.55897534e-01
5.39919198e-01 -7.34107614e-01 -6.53833807e-01 -2.34119520e-01] | [9.632003784179688, 2.7644848823547363] |
168daf59-ec3c-4e87-ab14-213025d66c63 | compressing-video-calls-using-synthetic | 2210.03692 | null | https://arxiv.org/abs/2210.03692v1 | https://arxiv.org/pdf/2210.03692v1.pdf | Compressing Video Calls using Synthetic Talking Heads | We leverage the modern advancements in talking head generation to propose an end-to-end system for talking head video compression. Our algorithm transmits pivot frames intermittently while the rest of the talking head video is generated by animating them. We use a state-of-the-art face reenactment network to detect key points in the non-pivot frames and transmit them to the receiver. A dense flow is then calculated to warp a pivot frame to reconstruct the non-pivot ones. Transmitting key points instead of full frames leads to significant compression. We propose a novel algorithm to adaptively select the best-suited pivot frames at regular intervals to provide a smooth experience. We also propose a frame-interpolater at the receiver's end to improve the compression levels further. Finally, a face enhancement network improves reconstruction quality, significantly improving several aspects like the sharpness of the generations. We evaluate our method both qualitatively and quantitatively on benchmark datasets and compare it with multiple compression techniques. We release a demo video and additional information at https://cvit.iiit.ac.in/research/projects/cvit-projects/talking-video-compression. | ['C V Jawahar', 'Vinay P. Namboodiri', 'Rudrabha Mukhopadhyay', 'Anchit Gupta', 'Madhav Agarwal'] | 2022-10-07 | null | null | null | null | ['talking-head-generation', 'face-reenactment'] | ['computer-vision', 'computer-vision'] | [ 8.55101496e-02 5.36898114e-02 -1.39751792e-01 -3.68208408e-01
-7.12628782e-01 -2.30628848e-01 3.09972703e-01 -3.41866940e-01
-6.46232218e-02 4.41306472e-01 8.18541348e-01 1.37976453e-01
1.04819819e-01 -6.73906982e-01 -6.12156749e-01 -8.28326881e-01
-1.39037207e-01 3.55036616e-01 2.17213020e-01 -1.88533887e-01
1.84599474e-01 3.77886951e-01 -1.71513104e+00 5.02905965e-01
5.16182482e-01 8.80192339e-01 2.28753015e-01 8.42457294e-01
1.05739653e-01 7.90049493e-01 -2.91449517e-01 -6.53674066e-01
3.56116831e-01 -5.69616258e-01 -6.40597224e-01 3.53023767e-01
6.22994125e-01 -6.36288047e-01 -7.23359108e-01 8.65642428e-01
7.77019918e-01 1.68325111e-01 1.66706696e-01 -1.35108209e+00
-2.72923589e-01 5.12219727e-01 -7.47686982e-01 1.83880255e-01
8.85491788e-01 1.30391568e-01 7.21953571e-01 -9.04301047e-01
9.38933790e-01 1.53606939e+00 6.19801879e-01 6.56030893e-01
-8.47153485e-01 -9.15375233e-01 -1.56266987e-01 5.10217071e-01
-1.48892581e+00 -1.18796527e+00 1.11980522e+00 6.87887892e-02
2.81559438e-01 3.33335787e-01 8.70550513e-01 8.77688587e-01
2.58154403e-02 7.16982245e-01 5.33748865e-01 -2.51374811e-01
9.94687527e-02 -3.49929571e-01 -4.52737868e-01 8.25696170e-01
-2.90546268e-01 3.72927226e-02 -1.03189921e+00 4.87258174e-02
7.60701120e-01 8.83430988e-02 -6.57940149e-01 -1.37235888e-03
-8.31561685e-01 7.06203878e-01 3.29252154e-01 1.99556530e-01
-5.04234672e-01 1.73364744e-01 3.82164717e-01 1.60294384e-01
6.09853804e-01 -3.73456299e-01 6.50641695e-02 -3.04945081e-01
-1.48964381e+00 4.17633504e-01 6.21497393e-01 8.47575605e-01
6.34354830e-01 -1.74753144e-01 -3.68012160e-01 9.88900661e-01
1.28316209e-01 2.79848188e-01 2.66241312e-01 -1.68068767e+00
4.87759143e-01 3.12013198e-02 -1.57026932e-01 -1.30284274e+00
-3.54674011e-02 -1.08628251e-01 -8.31469119e-01 -1.10813752e-01
1.25237582e-02 -1.67948753e-01 -5.74826717e-01 1.62486947e+00
6.73140347e-01 7.95721591e-01 -1.91614568e-01 1.06014073e+00
9.09568906e-01 9.61744189e-01 -3.25580895e-01 -7.18026042e-01
1.45060396e+00 -9.06124651e-01 -9.19563234e-01 1.54962987e-01
2.84128487e-01 -9.71339226e-01 7.23113000e-01 4.85195965e-01
-1.69954729e+00 -5.09498000e-01 -6.63432419e-01 -2.31861129e-01
4.80349809e-01 -2.90283263e-01 1.35608360e-01 4.57338572e-01
-1.39292598e+00 5.09634316e-01 -6.32377744e-01 -8.08162440e-04
6.06543243e-01 2.19138056e-01 -3.49085420e-01 -4.38805401e-01
-9.13479388e-01 2.37521887e-01 -3.71210650e-02 -3.39047581e-01
-7.80832946e-01 -9.21005428e-01 -8.08518589e-01 6.77670911e-02
3.68466973e-01 -8.44895363e-01 1.36885130e+00 -8.32255304e-01
-1.68041325e+00 6.39748991e-01 -6.64326251e-01 -3.97586644e-01
6.89152956e-01 -1.27070159e-01 -2.69101709e-01 8.10069799e-01
5.58770858e-02 1.02734685e+00 9.98470724e-01 -1.21471334e+00
-7.99103558e-01 -2.72077441e-01 -2.71099776e-01 4.23326850e-01
-3.22928101e-01 3.53619047e-02 -1.06454933e+00 -7.47011662e-01
2.86289752e-01 -8.83372903e-01 3.16942841e-01 3.27444494e-01
-3.81197959e-01 -9.81442779e-02 1.42487574e+00 -7.10635304e-01
1.26942074e+00 -2.18894958e+00 1.94975302e-01 -7.63130039e-02
5.76205909e-01 7.86134452e-02 -2.85428818e-02 4.32930529e-01
2.32086657e-03 -2.82518119e-01 -9.73697156e-02 -9.70323324e-01
-4.86100882e-01 2.05855280e-01 -2.09146246e-01 6.11282110e-01
-3.36105764e-01 6.48780584e-01 -8.02512050e-01 -6.68945849e-01
4.40240830e-01 1.08151710e+00 -1.08365786e+00 3.06873500e-01
1.41752571e-01 5.03499985e-01 4.71797921e-02 4.93706793e-01
9.72301900e-01 7.54717812e-02 1.05174564e-01 -3.66251647e-01
-1.49194255e-01 2.52780557e-01 -1.06918252e+00 1.66923618e+00
-4.24808323e-01 7.60971367e-01 3.30886334e-01 -6.95040226e-01
6.61961555e-01 4.74307895e-01 7.74062872e-01 -4.28471178e-01
2.56810874e-01 -1.56079322e-01 -4.39210683e-01 -6.68668270e-01
6.75464392e-01 -4.83278818e-02 4.00356948e-01 2.15528592e-01
-7.83051252e-02 -4.64669764e-02 1.75370932e-01 4.03983593e-01
9.41441357e-01 -2.37698138e-01 -1.06975116e-01 2.97279749e-02
4.88232762e-01 -7.53501356e-01 6.58278346e-01 2.45675832e-01
-2.54114866e-01 1.09620500e+00 4.25485224e-01 -2.69640833e-01
-1.16167104e+00 -9.37004447e-01 -1.73732955e-02 1.10841227e+00
1.48643672e-01 -8.39746535e-01 -1.16261268e+00 -1.86768115e-01
-2.81817526e-01 5.35351932e-01 -3.79206687e-01 -2.30050571e-02
-6.92642927e-01 -3.80154438e-02 2.60105163e-01 2.55702287e-01
8.47757816e-01 -9.64361489e-01 -7.00932801e-01 2.25152418e-01
-7.69163013e-01 -1.29577589e+00 -1.06092143e+00 -5.68208754e-01
-6.57411695e-01 -8.67623389e-01 -1.00611055e+00 -8.10017288e-01
6.04337275e-01 6.26991689e-01 8.52357328e-01 2.76894450e-01
-5.25257923e-02 1.61514446e-01 -3.68233114e-01 1.77961230e-01
-4.18086380e-01 -3.44515681e-01 5.29681221e-02 3.51161480e-01
-1.47816792e-01 -9.20458734e-01 -1.09238791e+00 4.01695758e-01
-8.08962107e-01 3.86885166e-01 5.71156256e-02 5.46658039e-01
5.47714412e-01 4.20753993e-02 6.77568689e-02 -5.34215927e-01
4.48024720e-01 -3.49719197e-01 -3.26393396e-01 -1.51395947e-01
2.03033239e-02 -1.11652493e-01 4.64639604e-01 -4.57451224e-01
-9.22273874e-01 -8.58856142e-02 -3.75700772e-01 -7.82938778e-01
2.28033870e-01 -1.86106805e-02 -1.82562143e-01 -7.17805326e-03
2.07239822e-01 2.92187303e-01 2.91110259e-02 -2.39573896e-01
4.52118427e-01 6.28185928e-01 9.37078655e-01 -3.01100105e-01
6.14323616e-01 7.30251610e-01 -9.55628231e-02 -1.13860106e+00
-3.99268836e-01 -4.07216579e-01 -2.10859954e-01 -7.02749729e-01
6.17188513e-01 -8.73827875e-01 -1.04297066e+00 3.90639931e-01
-1.26528692e+00 -1.43440872e-01 -1.87807441e-01 1.58731714e-01
-6.62786007e-01 6.54787540e-01 -7.27474689e-01 -6.77805126e-01
-5.30611396e-01 -1.30280483e+00 1.25702155e+00 2.54703164e-01
-1.85094655e-01 -6.19108737e-01 5.38642481e-02 5.20359874e-01
2.68552959e-01 6.99515417e-02 2.21008092e-01 1.91105589e-01
-5.16654611e-01 1.08571433e-01 4.07084487e-02 -9.61222872e-03
2.41811555e-02 1.90224528e-01 -9.50854659e-01 -4.46634144e-01
1.13363869e-01 6.24126680e-02 8.47473085e-01 7.62909055e-01
1.18817115e+00 -8.36918056e-01 -3.46933365e-01 9.54301476e-01
8.37736309e-01 1.65817291e-01 9.30230260e-01 -1.60755262e-01
4.85219836e-01 6.23158276e-01 2.49204189e-01 7.43938923e-01
5.72164714e-01 1.06194592e+00 2.29253531e-01 1.70730591e-01
-6.43050432e-01 -4.99483466e-01 5.56178331e-01 9.47758079e-01
-1.83482349e-01 -4.19942051e-01 -4.43775058e-01 5.23792922e-01
-1.68284404e+00 -1.26657915e+00 1.54022753e-01 2.06148505e+00
8.34206641e-01 1.54125895e-02 4.10512149e-01 4.48456138e-01
1.00287437e+00 3.00432444e-01 -1.90794930e-01 2.74090916e-02
6.55486807e-02 1.36823088e-01 6.58474043e-02 9.25091326e-01
-7.22029805e-01 9.33838844e-01 5.18703270e+00 8.94577324e-01
-1.23954141e+00 2.83647984e-01 8.46459150e-01 -4.97490913e-01
-2.39803046e-01 -8.75063986e-02 -7.28218138e-01 7.37829030e-01
9.75116134e-01 -1.30625278e-01 5.05676508e-01 6.67537510e-01
7.75421143e-01 -1.47729352e-01 -6.88597083e-01 1.35924792e+00
1.80531025e-01 -1.69237757e+00 2.08485331e-02 1.45383805e-01
5.83365083e-01 -2.81988144e-01 1.62441149e-01 -2.56843090e-01
-1.41130269e-01 -6.26583278e-01 8.94820333e-01 4.31319416e-01
9.06551003e-01 -1.13098550e+00 2.23972932e-01 -4.56040241e-02
-1.41353774e+00 -6.38388619e-02 -3.19468230e-01 2.54574299e-01
5.37854671e-01 7.05016613e-01 -8.17515671e-01 2.66408235e-01
8.57075632e-01 6.59130216e-01 -1.69498771e-01 1.01044703e+00
-1.95035473e-01 5.16521275e-01 -3.79857749e-01 5.56587517e-01
-1.02846939e-02 -1.25373051e-01 8.60463083e-01 1.09388137e+00
5.35829604e-01 4.04350400e-01 -8.69382545e-02 4.43751484e-01
-4.60218757e-01 3.36110331e-02 -5.30527234e-01 6.48050010e-01
8.29025984e-01 1.09798932e+00 -7.21739054e-01 -5.02668560e-01
-2.53383577e-01 1.15267718e+00 -1.11104876e-01 1.67746246e-01
-8.64120424e-01 -1.60733908e-01 7.03764856e-01 6.52758837e-01
4.00827795e-01 -9.33643878e-02 2.14599073e-01 -1.07702136e+00
8.72955285e-03 -7.67207861e-01 2.99991786e-01 -9.64946926e-01
-6.57273889e-01 7.07168162e-01 5.21431565e-02 -1.18210530e+00
-3.58573765e-01 2.24590197e-01 -7.79219806e-01 3.28520209e-01
-1.23300660e+00 -8.57068598e-01 -5.83262682e-01 1.09101462e+00
9.14530575e-01 3.59631144e-02 2.73745686e-01 5.35039306e-01
-3.25464368e-01 9.35479939e-01 -2.48940334e-01 6.28054962e-02
7.17150688e-01 -5.07894218e-01 3.47355604e-01 8.43252659e-01
-2.92667840e-02 2.78789699e-01 8.95581424e-01 -6.19913995e-01
-1.35231125e+00 -9.87047851e-01 8.83161366e-01 2.61258841e-01
1.06544577e-01 -4.11273539e-01 -8.47445607e-01 5.50683081e-01
3.98331702e-01 5.86731061e-02 3.38926017e-01 -5.16303182e-01
-1.64076373e-01 -4.06806111e-01 -1.34708679e+00 7.04762220e-01
1.14128590e+00 -3.08350235e-01 -2.56212234e-01 2.33506814e-01
8.39588642e-01 -4.55258667e-01 -4.79536355e-01 7.04144165e-02
5.94153225e-01 -1.43326306e+00 1.04470336e+00 1.88843563e-01
6.79384828e-01 -1.93605408e-01 -1.01444185e-01 -1.06287718e+00
-1.20508283e-01 -1.45926547e+00 -3.21339101e-01 1.33896971e+00
-2.88354248e-01 -3.45903695e-01 1.04925776e+00 2.75174290e-01
-1.80683538e-01 -5.87144256e-01 -1.35528338e+00 -2.90858001e-01
-4.10051435e-01 -4.99821484e-01 4.41530436e-01 6.86017752e-01
2.34163497e-02 1.88577041e-01 -7.33222008e-01 5.32498443e-03
8.61856222e-01 -1.98654100e-01 7.75017619e-01 -7.92574406e-01
-1.46702647e-01 -1.09159380e-01 -5.10999680e-01 -1.34744596e+00
1.45238247e-02 -6.32090747e-01 -1.51497975e-01 -1.08137369e+00
1.97099552e-01 -2.98973452e-02 3.47893029e-01 2.75840163e-01
9.48827714e-02 5.68538010e-01 5.12408495e-01 3.55664879e-01
-4.47754651e-01 6.24541461e-01 1.36630440e+00 1.40660897e-01
-4.07138407e-01 2.08028741e-02 -5.85597277e-01 6.39804065e-01
7.28922129e-01 -4.63719696e-01 -5.00343919e-01 -2.78537124e-01
-3.43222708e-01 7.87136614e-01 3.22747082e-01 -1.39258313e+00
4.71169174e-01 9.43243951e-02 3.78490627e-01 -7.42837071e-01
7.78059244e-01 -6.43251419e-01 3.38293821e-01 4.68421668e-01
-3.10656190e-01 2.80287683e-01 -4.47012559e-02 3.09289694e-01
1.82055756e-02 2.33714692e-02 1.20212758e+00 6.83406442e-02
-1.84866816e-01 6.05291963e-01 -2.90140629e-01 1.22660091e-02
9.70921636e-01 -2.09332600e-01 -3.27362260e-03 -1.15492988e+00
-6.62569404e-01 -4.43054736e-02 4.09294963e-01 2.77169287e-01
9.82068717e-01 -1.27137482e+00 -9.19269145e-01 3.99005502e-01
-4.51639295e-01 -3.43336053e-02 5.11748731e-01 8.24185729e-01
-7.16671348e-01 -7.49647059e-03 -1.95278123e-01 -7.77661860e-01
-1.71522021e+00 3.80331844e-01 1.44153923e-01 1.21875525e-01
-1.10858881e+00 8.17585826e-01 2.76997447e-01 4.04583126e-01
2.83937991e-01 1.67794097e-02 5.09298965e-02 2.61260830e-02
1.01794088e+00 4.59161013e-01 -7.81039819e-02 -9.52364564e-01
-1.93073243e-01 6.75871193e-01 -6.25173301e-02 -3.77331346e-01
1.50956523e+00 -5.30347586e-01 7.87983835e-02 -2.84395307e-01
1.56021404e+00 1.99225813e-01 -1.45833254e+00 -1.39874190e-01
-6.71466529e-01 -8.35137367e-01 2.57719666e-01 -1.98309183e-01
-1.71095908e+00 6.85110271e-01 6.35361075e-01 -1.18861675e-01
1.43689394e+00 -5.74237248e-03 1.35931051e+00 -1.38301983e-01
1.84118420e-01 -7.44274139e-01 3.96030486e-01 2.23733634e-01
9.64217186e-01 -7.12214231e-01 8.25616196e-02 -7.15519726e-01
-6.76943064e-01 1.24610257e+00 1.03300907e-01 -4.82409969e-02
6.18388116e-01 5.92127085e-01 -9.85472202e-02 -8.11388567e-02
-7.78390408e-01 2.06582651e-01 -9.14018229e-02 6.33679271e-01
3.60377043e-01 -1.64113015e-01 -2.21420854e-01 -4.38958295e-02
-7.50217140e-01 2.42235005e-01 7.02110648e-01 7.50096560e-01
-4.44261551e-01 -8.50951195e-01 -7.07381248e-01 2.77502816e-02
-4.61794466e-01 -1.44680992e-01 3.12697142e-02 4.03196156e-01
2.92752706e-03 1.24803329e+00 2.08625913e-01 -5.27605474e-01
1.61060259e-01 -1.20853119e-01 4.64279681e-01 -7.92631879e-02
-3.43323141e-01 6.43524289e-01 -7.81971514e-02 -8.08545351e-01
-3.75097692e-01 -7.23251402e-01 -1.27783144e+00 -1.24303877e+00
1.82096899e-01 2.00621441e-01 4.17245001e-01 4.81860012e-01
6.43810570e-01 3.52988094e-01 9.81821179e-01 -1.38521147e+00
1.62808448e-01 -6.75091803e-01 -1.74721584e-01 4.32206959e-01
4.82218981e-01 -5.29065788e-01 -3.92909855e-01 4.13337618e-01] | [13.207052230834961, -0.4774235785007477] |
fdecc07f-0453-4aa3-afae-493417a19b88 | transferrable-end-to-end-learning-for-protein-1 | null | null | https://openreview.net/forum?id=SkgToo0qFm | https://openreview.net/pdf?id=SkgToo0qFm | Transferrable End-to-End Learning for Protein Interface Prediction | While there has been an explosion in the number of experimentally determined, atomically detailed structures of proteins, how to represent these structures in a machine learning context remains an open research question. In this work we demonstrate that representations learned from raw atomic coordinates can outperform hand-engineered structural features while displaying a much higher degree of transferrability. To do so, we focus on a central problem in biology: predicting how proteins interact with one another—that is, which surfaces of one protein bind to which surfaces of another protein. We present Siamese Atomic Surfacelet Network (SASNet), the first end-to-end learning method for protein interface prediction. Despite using only spatial coordinates and identities of atoms as inputs, SASNet outperforms state-of-the-art methods that rely on hand-engineered, high-level features. These results are particularly striking because we train the method entirely on a significantly biased data set that does not account for the fact that proteins deform when binding to one another. Demonstrating the first successful application of transfer learning to atomic-level data, our network maintains high performance, without retraining, when tested on real cases in which proteins do deform. | ['Ron O. Dror', 'Rishi Bedi', 'Raphael J. L. Townshend'] | 2018-09-27 | null | null | null | null | ['protein-interface-prediction'] | ['miscellaneous'] | [ 4.97024655e-01 2.21780241e-01 3.69877182e-02 -3.79711270e-01
-8.60734344e-01 -5.61247885e-01 4.97135252e-01 3.82775605e-01
-5.29585779e-01 1.17813301e+00 -2.06796220e-03 -4.11857724e-01
1.94015920e-01 -6.85054302e-01 -1.27714407e+00 -8.45078111e-01
-2.70343333e-01 8.52888644e-01 4.84958172e-01 -4.32441264e-01
2.58125931e-01 7.61042058e-01 -1.27722216e+00 4.84492451e-01
8.64802122e-01 7.23112822e-01 2.01359883e-01 5.62065363e-01
1.17050130e-02 3.77056628e-01 -3.05708259e-01 -1.51471347e-01
-1.72935892e-02 -4.57909316e-01 -9.00648832e-01 -4.82438087e-01
6.65900707e-01 2.63090014e-01 1.10616557e-01 6.32229090e-01
3.22461396e-01 -8.22920874e-02 1.11148655e+00 -4.21892524e-01
-8.42896998e-01 -1.88153058e-01 -3.03602755e-01 2.19633281e-02
4.56732243e-01 1.60271615e-01 1.31483269e+00 -8.64177585e-01
1.07386756e+00 1.06416368e+00 8.20094705e-01 5.07186949e-01
-1.79803991e+00 -2.71286875e-01 7.46598048e-03 1.29260674e-01
-8.91816080e-01 -1.12809412e-01 4.46164846e-01 -7.60738254e-01
1.39034534e+00 3.30776535e-02 2.95828849e-01 8.79680634e-01
5.35566747e-01 1.88815311e-01 9.02889609e-01 -2.18748987e-01
2.35928863e-01 -3.41567010e-01 1.54730186e-01 7.43957460e-01
3.21245104e-01 -4.46331576e-02 -4.01749432e-01 -5.58560550e-01
5.46699762e-01 6.23933747e-02 -2.80972034e-01 -8.82288754e-01
-1.12818480e+00 8.45763683e-01 7.21551716e-01 2.33448997e-01
-5.16252875e-01 -8.73973127e-03 1.20781176e-01 2.91721880e-01
2.94368535e-01 7.83332646e-01 -8.10418546e-01 -1.12078615e-01
-5.23133814e-01 4.84751731e-01 8.88586044e-01 4.20388311e-01
1.06675971e+00 -4.41371709e-01 2.91054755e-01 4.92963552e-01
-2.76329834e-02 2.83609390e-01 2.34725088e-01 -4.35275137e-01
1.00873619e-01 7.80748725e-01 2.59256035e-01 -7.35269487e-01
-4.52242434e-01 -2.47729439e-02 -6.58756495e-01 6.48555815e-01
7.00857401e-01 2.43218914e-02 -8.37810636e-01 1.80771804e+00
2.80536205e-01 -1.64379869e-02 1.62678108e-01 7.79674709e-01
5.02789557e-01 6.71878934e-01 1.75413296e-01 1.83766603e-03
1.26617837e+00 -6.75767660e-01 -7.68760219e-02 -5.98472320e-02
7.56466031e-01 -6.06268942e-01 9.68449116e-01 3.76335502e-01
-1.06494832e+00 -4.27959591e-01 -1.14639843e+00 -2.44440675e-01
-5.32288074e-01 -2.05665454e-01 6.30936623e-01 1.51543647e-01
-7.81020403e-01 1.11101806e+00 -8.12470138e-01 -4.56004351e-01
3.81094307e-01 7.36950338e-01 -8.99697661e-01 2.64387339e-01
-1.08277500e+00 1.14495277e+00 1.84675336e-01 -4.37478065e-01
-4.05119240e-01 -8.16965103e-01 -6.25244200e-01 9.17934477e-02
2.58684993e-01 -5.48049748e-01 1.03268564e+00 -9.93515491e-01
-1.24267805e+00 8.80092382e-01 -3.98946106e-01 -6.53102517e-01
1.57976002e-01 -1.06401056e-01 -7.60509893e-02 1.68507159e-01
-3.82571258e-02 5.19443035e-01 3.77739698e-01 -1.15906870e+00
-3.70642900e-01 -6.25008523e-01 -1.14593521e-01 -9.98099819e-02
1.22705251e-01 -2.14539841e-01 1.82196602e-01 -3.73045832e-01
2.23529451e-02 -1.00799227e+00 -2.72539705e-01 2.21518591e-01
-8.60586315e-02 -4.34237063e-01 6.28824413e-01 -4.17684078e-01
5.38578749e-01 -1.79172242e+00 6.84033096e-01 2.93553650e-01
3.39140654e-01 5.40585339e-01 -6.94877952e-02 7.89934754e-01
-3.42020720e-01 9.42912549e-02 -4.95341480e-01 1.59618944e-01
-2.79152691e-01 1.18139863e-01 -2.26395503e-01 5.46019137e-01
5.49114466e-01 9.81601596e-01 -8.77994478e-01 -1.27682269e-01
-7.39437416e-02 6.31159604e-01 -8.49089742e-01 3.63507479e-01
-4.97354567e-01 6.66536093e-01 -4.51871723e-01 2.40635708e-01
5.55607319e-01 -6.72155559e-01 4.39051330e-01 -3.29114795e-01
-8.48159939e-02 4.05895144e-01 -6.27919793e-01 1.52644730e+00
-4.96403687e-02 1.90116331e-01 -5.48123531e-02 -1.05504513e+00
8.81508291e-01 1.62790775e-01 7.59147048e-01 -5.33360362e-01
-2.96048880e-01 3.27849030e-01 4.43166256e-01 -1.74309969e-01
-1.52734742e-01 -3.35914284e-01 8.03277194e-02 4.14368272e-01
9.86812040e-02 1.30354628e-01 5.25465459e-02 1.00833945e-01
1.37840140e+00 3.75361204e-01 4.17710841e-01 -2.74114311e-01
5.17523110e-01 2.35861436e-01 4.76932555e-01 3.58651966e-01
1.42580569e-01 7.34965324e-01 6.97388172e-01 -9.26229298e-01
-1.44095182e+00 -1.09311175e+00 -2.51120716e-01 1.27208173e+00
-8.64569191e-03 -3.51131022e-01 -8.96871984e-01 -5.68036437e-01
4.37517732e-01 3.94967496e-02 -8.70436907e-01 -1.80205360e-01
-7.52508879e-01 -7.56253242e-01 2.28076860e-01 3.41546059e-01
-1.39160275e-01 -1.22639012e+00 -6.54106975e-01 5.07140338e-01
3.88738513e-01 -7.20288455e-01 -3.62773448e-01 5.09691417e-01
-7.84194231e-01 -1.41254640e+00 -6.89523101e-01 -7.65244901e-01
5.96290350e-01 -2.10842378e-02 1.36223471e+00 1.87250093e-01
-4.92106199e-01 -1.91347420e-01 -1.04374439e-01 -3.66454363e-01
-3.99382889e-01 2.96252370e-01 6.68644607e-02 -1.19040757e-01
6.10660911e-01 -8.94691229e-01 -6.93287909e-01 2.87270635e-01
-8.23148668e-01 5.31930523e-03 6.30838871e-01 1.01842570e+00
7.28910089e-01 -6.24620974e-01 5.93872070e-01 -1.36001313e+00
5.76990128e-01 -2.66257048e-01 -4.42907780e-01 2.46837497e-01
-4.49614704e-01 6.16339505e-01 9.48686540e-01 -2.21219882e-01
-6.23929024e-01 4.99642283e-01 -2.31698602e-01 4.86051664e-02
-2.41746709e-01 3.68452609e-01 3.95205431e-02 -3.01987410e-01
8.90514076e-01 9.90212262e-02 5.08329093e-01 -6.16169631e-01
1.71797425e-01 2.64783263e-01 4.48882937e-01 -8.34256530e-01
6.93585575e-01 4.50233370e-01 2.97871441e-01 -8.49069178e-01
-6.41855180e-01 -3.43718976e-01 -9.84059632e-01 5.55270553e-01
1.01532841e+00 -4.57599401e-01 -1.12220144e+00 1.72154143e-01
-1.22676766e+00 -3.51213276e-01 2.00972054e-03 1.71717674e-01
-7.45084584e-01 4.23974961e-01 -4.98115867e-01 -2.74071544e-01
-2.55993158e-01 -1.28667617e+00 1.27348340e+00 -2.73635656e-01
-4.54003334e-01 -8.04185867e-01 5.06129026e-01 1.46770880e-01
3.68336409e-01 5.84353626e-01 1.44815731e+00 -8.90278935e-01
-5.13264418e-01 -7.74155781e-02 -1.85752124e-01 6.28570169e-02
3.20033759e-01 -3.30975093e-02 -6.56018615e-01 -4.52545434e-01
-4.24466372e-01 -5.27104557e-01 9.77926552e-01 5.57041876e-02
8.91616881e-01 -9.24784914e-02 -5.00668943e-01 4.74098563e-01
1.20659804e+00 2.53069997e-01 5.97984552e-01 1.90215021e-01
7.15807796e-01 6.78011298e-01 2.51818031e-01 9.65263415e-03
-2.37535685e-02 9.14849460e-01 5.15725613e-01 -4.12088960e-01
1.95210949e-01 -1.19027965e-01 7.93129653e-02 2.70617962e-01
-4.54747468e-01 -4.98854220e-02 -1.00680792e+00 8.83983448e-02
-1.91474462e+00 -8.57054472e-01 -7.66249672e-02 2.21765709e+00
1.11527979e+00 2.57262141e-01 3.14410925e-01 -2.91490436e-01
3.43880326e-01 -1.62382603e-01 -9.94435906e-01 -3.73983175e-01
-1.16580412e-01 5.80469251e-01 3.80338490e-01 5.93770921e-01
-1.01043284e+00 9.16208386e-01 6.54994202e+00 3.31378102e-01
-1.16604388e+00 -2.12182134e-01 5.83165228e-01 2.32507363e-01
-1.38937891e-01 -5.24297617e-02 -5.51804423e-01 3.79679888e-01
1.06149685e+00 2.07637027e-01 4.32005227e-01 6.82292163e-01
-1.02517838e-02 1.87712029e-01 -1.44531488e+00 5.41322172e-01
-3.12250286e-01 -1.73878491e+00 1.46029919e-01 2.88950592e-01
4.88332540e-01 2.12549090e-01 -1.77024484e-01 5.46051823e-02
2.72232801e-01 -1.46973062e+00 4.63120677e-02 5.60533643e-01
7.28249252e-01 -6.48236871e-01 5.02957463e-01 1.94811270e-01
-9.47695732e-01 4.92107630e-01 -4.53380495e-01 3.09654661e-02
1.06960069e-02 2.15989232e-01 -8.39395463e-01 2.45432734e-01
3.91176015e-01 5.32057464e-01 -3.85166377e-01 5.47123194e-01
1.64325833e-01 4.56735939e-01 -1.97580248e-01 -8.31879154e-02
2.26943254e-01 -3.85075480e-01 1.04623288e-01 1.09387839e+00
-1.52747393e-01 2.45835185e-01 1.62524343e-01 9.10348713e-01
-1.94609746e-01 1.62432238e-01 -7.91938663e-01 9.15700849e-03
-3.80313061e-02 1.06722522e+00 -2.21040249e-01 -1.74405456e-01
-5.39941847e-01 9.61350858e-01 8.97572815e-01 3.91889215e-01
-6.24528587e-01 -3.59310269e-01 1.09715700e+00 5.09156585e-01
5.00442088e-01 -4.00492191e-01 1.47021189e-01 -9.84244406e-01
-2.58533359e-02 -1.05447841e+00 1.56754944e-02 -5.68188488e-01
-1.48000777e+00 5.58302939e-01 -4.05931056e-01 -5.74344456e-01
-2.11774200e-01 -1.24042308e+00 -5.10115683e-01 1.15708065e+00
-1.36724365e+00 -1.01248538e+00 2.70187557e-02 1.04775205e-01
2.52166629e-01 -1.96698353e-01 1.14254928e+00 -7.37487003e-02
-8.56339633e-02 3.87435228e-01 4.57149953e-01 -7.15655237e-02
8.82871091e-01 -1.49744403e+00 7.94445932e-01 6.45658523e-02
1.09325111e-01 9.23170447e-01 8.20720017e-01 -7.17908502e-01
-1.63030636e+00 -8.45278859e-01 6.51270449e-01 -6.37059450e-01
5.87796092e-01 -6.09965205e-01 -1.49211049e+00 5.67903280e-01
-4.86450568e-02 1.91107452e-01 7.62783051e-01 2.55518019e-01
-5.58433115e-01 2.60687470e-01 -1.09732664e+00 5.63541949e-01
1.13931239e+00 -4.96808827e-01 -6.37162089e-01 4.72423971e-01
7.13969409e-01 -1.80231750e-01 -1.04385400e+00 5.81424356e-01
5.57540417e-01 -1.00952959e+00 1.21357632e+00 -1.33978605e+00
3.70465666e-01 -2.46243134e-01 -5.43720908e-02 -1.09868491e+00
-5.56516945e-01 -5.39325774e-01 5.69778830e-02 4.51550841e-01
8.01960707e-01 -7.87928522e-01 1.11204922e+00 5.19904792e-01
-9.81118679e-02 -1.32285511e+00 -9.75376070e-01 -6.06885076e-01
6.41958833e-01 3.14552218e-01 3.95882040e-01 6.51195943e-01
1.25362411e-01 6.71879530e-01 -2.90759448e-02 -1.26593262e-01
3.14624727e-01 2.74749160e-01 9.07368481e-01 -1.65810895e+00
-3.76516819e-01 -2.79315174e-01 -5.77701449e-01 -1.05407190e+00
2.94757098e-01 -8.85951579e-01 -1.24934241e-01 -1.25610113e+00
2.24580497e-01 -2.67257750e-01 -4.00276423e-01 5.17535090e-01
-1.49004355e-01 2.07953066e-01 -1.91158339e-01 3.12709600e-01
-5.15049338e-01 5.32410622e-01 1.13945246e+00 -1.24067232e-01
-5.29409247e-03 -2.46825650e-01 -5.10662556e-01 5.73461771e-01
7.20004737e-01 -3.63114178e-01 -1.22309826e-01 6.13308512e-02
2.23305881e-01 -2.66772836e-01 1.66309506e-01 -9.54319715e-01
-2.18913574e-02 -1.52115002e-01 3.89973104e-01 -2.69809395e-01
5.47697484e-01 -8.04652333e-01 1.15739539e-01 6.16388738e-01
-2.56609023e-01 2.14062169e-01 1.53565735e-01 7.52024770e-01
3.97816040e-02 1.93301346e-02 9.38823044e-01 -1.75458997e-01
-2.76892275e-01 3.99166733e-01 -4.10822749e-01 7.17315301e-02
1.09830546e+00 -2.89785713e-01 -1.93795860e-01 -4.34794948e-02
-7.51788139e-01 -2.54807383e-01 1.06275189e+00 6.43710345e-02
4.43933964e-01 -9.49820220e-01 -6.32742643e-01 3.22418571e-01
1.51845053e-01 1.66298989e-02 -3.15278649e-01 4.86374676e-01
-8.25197577e-01 5.93020558e-01 -3.72010380e-01 -6.07692897e-01
-1.35817659e+00 6.40249550e-01 4.49456424e-01 -2.59747237e-01
-4.91706133e-01 5.65575421e-01 5.67889333e-01 -5.29755235e-01
-1.77873045e-01 -3.20015550e-01 1.34680927e-01 -3.11761618e-01
2.51513362e-01 -1.01166308e-01 3.65741193e-01 -6.20022774e-01
-4.19367492e-01 6.80872023e-01 -5.58144867e-01 3.85003328e-01
1.72408211e+00 6.33113384e-01 -1.60486311e-01 3.43342692e-01
1.21297741e+00 -1.36724159e-01 -1.56325829e+00 -2.47306302e-01
3.66357952e-01 -1.31579518e-01 -6.12704396e-01 -8.02684128e-01
-4.65946257e-01 9.06597316e-01 4.98395175e-01 1.20183244e-01
5.02825618e-01 1.79400563e-01 8.72184634e-01 9.28676069e-01
5.78508079e-01 -6.27519250e-01 1.94349512e-01 6.60117209e-01
7.81792760e-01 -1.37262595e+00 6.77103549e-02 -2.86613107e-01
-4.35168058e-01 1.03255773e+00 5.11126935e-01 -5.38179338e-01
4.09763604e-01 1.61652878e-01 -9.40910205e-02 -4.35721725e-01
-8.26426804e-01 2.27611456e-02 3.08221698e-01 6.96721196e-01
1.02462089e+00 -1.98374867e-01 -1.30782500e-01 2.01902956e-01
5.42056523e-02 -2.99457461e-01 1.86058618e-02 1.06893945e+00
-8.14179718e-01 -1.62433529e+00 -1.39871374e-01 4.21003431e-01
-6.46659553e-01 -1.63055286e-01 -1.00373256e+00 7.53388762e-01
-1.78499267e-01 3.89362574e-01 -1.20610349e-01 -4.37695645e-02
3.86346251e-01 1.96846336e-01 5.74503779e-01 -8.39034915e-01
-5.95100284e-01 -3.16596627e-01 -1.87798440e-01 -4.30002362e-01
-2.83631623e-01 -4.93103474e-01 -1.60191846e+00 -1.74213007e-01
-9.53159928e-02 4.44358736e-01 4.12121058e-01 8.00686955e-01
8.97591710e-01 3.41248930e-01 2.10859373e-01 -1.17000961e+00
-6.01576328e-01 -8.14844012e-01 -3.69750857e-01 9.43340003e-01
5.39533854e-01 -6.83105886e-01 -4.49525155e-02 1.78833440e-01] | [4.853334426879883, 5.645125389099121] |
556bb972-7620-4637-ab2a-83e20f492a2c | editing-text-in-the-wild | 1908.03047 | null | https://arxiv.org/abs/1908.03047v1 | https://arxiv.org/pdf/1908.03047v1.pdf | Editing Text in the Wild | In this paper, we are interested in editing text in natural images, which aims to replace or modify a word in the source image with another one while maintaining its realistic look. This task is challenging, as the styles of both background and text need to be preserved so that the edited image is visually indistinguishable from the source image. Specifically, we propose an end-to-end trainable style retention network (SRNet) that consists of three modules: text conversion module, background inpainting module and fusion module. The text conversion module changes the text content of the source image into the target text while keeping the original text style. The background inpainting module erases the original text, and fills the text region with appropriate texture. The fusion module combines the information from the two former modules, and generates the edited text images. To our knowledge, this work is the first attempt to edit text in natural images at the word level. Both visual effects and quantitative results on synthetic and real-world dataset (ICDAR 2013) fully confirm the importance and necessity of modular decomposition. We also conduct extensive experiments to validate the usefulness of our method in various real-world applications such as text image synthesis, augmented reality (AR) translation, information hiding, etc. | ['Xiang Bai', 'Chengquan Zhang', 'Junyu Han', 'Errui Ding', 'Liang Wu', 'Jingtuo Liu', 'Jiaming Liu'] | 2019-08-08 | null | null | null | null | ['scene-text-editing'] | ['computer-vision'] | [ 9.51839328e-01 -5.44122458e-02 2.26020172e-01 -1.33821398e-01
-2.57363886e-01 -4.03417677e-01 4.68435317e-01 -1.17748275e-01
-3.62836510e-01 8.05806458e-01 1.31509408e-01 -1.62476093e-01
4.25591081e-01 -7.72414267e-01 -8.71236682e-01 -7.21032023e-01
7.23129213e-01 2.87727583e-02 2.06373870e-01 -2.64026880e-01
1.77829474e-01 4.81456310e-01 -1.23296618e+00 3.46908927e-01
9.77091372e-01 8.25806141e-01 4.57566828e-01 5.28917730e-01
-3.53349477e-01 9.47655380e-01 -6.34445190e-01 -5.25574923e-01
3.43179524e-01 -6.51164591e-01 -3.52208465e-01 5.46103597e-01
5.43776453e-01 -4.40381706e-01 -6.55202210e-01 1.39205825e+00
4.01431680e-01 1.61125399e-02 3.92155051e-01 -1.04287326e+00
-9.44056034e-01 4.15539682e-01 -8.88211787e-01 -2.66518146e-01
1.41983747e-01 -1.28310034e-02 4.74430412e-01 -8.80410492e-01
7.84056544e-01 1.23981547e+00 4.21069920e-01 5.38203895e-01
-1.28620601e+00 -6.49925709e-01 1.08885512e-01 -1.66060880e-01
-1.25027454e+00 -5.24275422e-01 1.05085433e+00 -2.98346579e-01
2.52505749e-01 3.80430013e-01 4.92014199e-01 1.12997019e+00
7.47740507e-01 7.23181367e-01 1.22300398e+00 -7.42187977e-01
5.65179512e-02 3.63481462e-01 -3.39345336e-01 6.79431140e-01
2.73092568e-01 9.76484790e-02 -4.35908586e-01 2.92501420e-01
9.33357418e-01 1.43355280e-01 -5.91145694e-01 -4.15116608e-01
-1.49664664e+00 5.25039315e-01 2.03625843e-01 3.23037744e-01
-2.37145811e-01 5.57802198e-03 1.80616111e-01 4.09328431e-01
4.79822934e-01 1.10240474e-01 -2.83788778e-02 3.20852101e-01
-1.08182013e+00 -1.28840566e-01 4.95977730e-01 9.87659395e-01
6.51602447e-01 2.34115079e-01 -1.28060490e-01 9.04251099e-01
1.85190782e-01 7.72886276e-01 5.09653509e-01 -6.22075558e-01
5.47868311e-01 4.78128791e-01 4.61198762e-02 -1.28413665e+00
4.02999371e-02 -2.95037836e-01 -1.33040833e+00 5.51745832e-01
5.33594675e-02 -2.10129052e-01 -9.94274199e-01 1.51089847e+00
1.36003196e-01 -1.23701900e-01 1.12802479e-02 7.52091229e-01
8.24586093e-01 9.07193601e-01 -2.28533015e-01 -2.74702817e-01
1.18663549e+00 -1.00054455e+00 -1.15310872e+00 -4.16126698e-01
1.41800851e-01 -1.20415974e+00 1.02280915e+00 3.26695293e-01
-1.24487972e+00 -7.35070646e-01 -1.34386945e+00 -3.55330616e-01
-1.69833168e-01 5.70661604e-01 1.43356577e-01 3.95377517e-01
-8.32765698e-01 2.14511141e-01 -5.21586478e-01 -2.54860550e-01
1.37656093e-01 6.15137890e-02 -5.79266846e-01 8.79546255e-03
-1.10452902e+00 7.30477870e-01 4.12245423e-01 2.06914768e-01
-3.44048202e-01 -4.34976876e-01 -8.93776238e-01 -6.34509996e-02
3.93052518e-01 -8.22909474e-01 9.39401448e-01 -1.58606601e+00
-1.63580954e+00 8.86777163e-01 -1.72599286e-01 -2.81756699e-01
8.17928076e-01 -1.34124458e-01 -5.46224952e-01 1.12952581e-02
3.72621194e-02 7.27510035e-01 1.47046685e+00 -1.49768043e+00
-6.03131592e-01 -3.20133030e-01 -2.19087899e-01 2.59805381e-01
-2.12978110e-01 -2.06961870e-01 -6.69210732e-01 -1.37849617e+00
1.18597992e-01 -8.29743385e-01 8.92350674e-02 5.34965634e-01
-3.81831646e-01 6.99501455e-01 1.22533965e+00 -1.00316870e+00
1.03901851e+00 -2.52159166e+00 3.49905908e-01 -6.19312897e-02
2.84002870e-01 1.97310030e-01 -1.63080469e-01 3.90288085e-01
-2.04602495e-01 -1.26184389e-01 -4.84274954e-01 -5.42116880e-01
-2.77598590e-01 7.27576613e-02 -6.88998580e-01 3.86541754e-01
8.43425170e-02 8.56024623e-01 -4.82736439e-01 -6.61603093e-01
4.11906868e-01 6.27164185e-01 -2.79401660e-01 3.41767892e-02
-3.10940981e-01 5.92604935e-01 2.94402987e-02 4.56927299e-01
9.27897453e-01 1.77995071e-01 -8.40827227e-02 -4.77447569e-01
-1.22077465e-01 -2.82359928e-01 -1.32186997e+00 1.64249933e+00
-4.13241476e-01 9.84112144e-01 2.02809379e-01 -4.78990614e-01
9.80696917e-01 1.52798072e-01 2.22030386e-01 -1.01942098e+00
2.50810653e-01 -3.18026822e-03 -2.53310025e-01 -3.45647782e-01
9.66461420e-01 -2.15712816e-01 1.14749834e-01 4.17992264e-01
-3.59327495e-01 -3.38229835e-01 9.84814093e-02 1.48724526e-01
4.83145773e-01 1.04899779e-01 9.02576596e-02 1.12798281e-01
6.31312728e-01 -2.56744504e-01 3.56101722e-01 4.07073945e-01
1.70392498e-01 9.03789341e-01 4.11618978e-01 -3.36916268e-01
-1.25474453e+00 -1.02989590e+00 1.04286186e-01 6.97095335e-01
5.24099290e-01 -1.20398894e-01 -7.97395468e-01 -4.94678199e-01
-3.05696070e-01 7.46748209e-01 -6.89113498e-01 -2.63843983e-01
-6.38231695e-01 -3.29425454e-01 4.50756907e-01 3.54017138e-01
1.09958625e+00 -1.08294678e+00 -4.00318116e-01 1.77397147e-01
-5.19041777e-01 -1.18710053e+00 -1.02272594e+00 -5.28430119e-02
-7.51609504e-01 -7.35990465e-01 -8.10804904e-01 -1.03379428e+00
1.07214212e+00 5.56105793e-01 5.99886537e-01 8.72041956e-02
-1.68527573e-01 -2.61936374e-02 -3.07450593e-01 -1.91835716e-01
-7.64461994e-01 -3.62850904e-01 -3.04613471e-01 4.61051136e-01
-3.85618448e-01 -2.90297866e-01 -3.54243159e-01 2.99569398e-01
-1.56142008e+00 6.73500836e-01 7.08399117e-01 9.34316754e-01
6.76042140e-01 6.04909837e-01 -3.25109065e-02 -8.76857519e-01
5.07134795e-01 2.12233111e-01 -6.72567546e-01 3.50521684e-01
-2.80236989e-01 4.31877822e-02 7.18045115e-01 -6.29481077e-01
-1.36129892e+00 2.01858491e-01 -1.78017933e-02 -4.41025227e-01
3.79949249e-02 2.54739553e-01 -4.89738971e-01 -4.17231955e-02
4.46300656e-01 7.38870442e-01 4.05170135e-02 -4.58231300e-01
3.93547297e-01 6.34225070e-01 8.90033424e-01 -1.96828187e-01
1.12019670e+00 6.52742565e-01 -9.71566513e-02 -9.53837574e-01
-4.89775777e-01 1.75603822e-01 -6.59244299e-01 -3.03083770e-02
8.16603959e-01 -8.21181774e-01 -4.46811877e-02 7.55194068e-01
-1.30651021e+00 -2.54863650e-01 -5.31272173e-01 1.65900677e-01
-3.60948265e-01 7.38567173e-01 -3.84668499e-01 -3.77477705e-01
-3.25598508e-01 -1.08943808e+00 1.19494581e+00 1.85088903e-01
6.30486757e-02 -7.94669092e-01 -4.92545627e-02 3.26892793e-01
3.29122365e-01 1.91855088e-01 1.17185485e+00 1.00099519e-01
-5.79098582e-01 -1.09439448e-01 -3.60025465e-01 5.53210914e-01
4.73261327e-01 1.67134747e-01 -7.60796845e-01 -4.13196653e-01
1.86740860e-01 1.93105087e-01 1.02804744e+00 1.39266819e-01
9.40213501e-01 -3.22681159e-01 -8.74580890e-02 6.54455781e-01
1.35791707e+00 5.19369066e-01 1.17815268e+00 3.31876218e-01
7.73545325e-01 3.90976071e-01 4.71328378e-01 1.80244014e-01
-3.69333103e-02 6.33179963e-01 2.71956027e-01 -6.30208015e-01
-5.10491788e-01 -3.85021061e-01 5.48483193e-01 6.59451604e-01
3.99013937e-01 -6.16753161e-01 -3.30664992e-01 1.90899983e-01
-1.61100137e+00 -9.90747631e-01 -1.01038264e-02 2.29284072e+00
9.27444637e-01 2.00734347e-01 -3.59636664e-01 1.90098420e-01
9.83780026e-01 3.17237765e-01 -4.09113377e-01 -2.56017447e-01
-3.79182309e-01 4.85663004e-02 3.75941306e-01 5.29914737e-01
-1.11227787e+00 1.02327263e+00 5.09797478e+00 9.22760904e-01
-1.58336401e+00 -1.35346979e-01 6.92831218e-01 1.56483710e-01
-2.29302272e-01 7.02818530e-03 -4.02822763e-01 5.95421672e-01
2.50141680e-01 -9.55603346e-02 5.20526826e-01 3.86516988e-01
2.55847782e-01 -3.13907355e-01 -8.57479095e-01 1.03063297e+00
3.26707006e-01 -1.32772934e+00 5.06887555e-01 -1.78516388e-01
7.33617365e-01 -6.17643476e-01 2.34939635e-01 -3.42488214e-02
-1.13443397e-01 -8.89031291e-01 1.04538059e+00 7.38904774e-01
1.06414580e+00 -7.56036580e-01 5.69849610e-01 1.68909982e-01
-1.08113050e+00 2.11242795e-01 -2.75361270e-01 3.31227750e-01
1.02585316e-01 7.31292069e-01 -4.95618612e-01 5.61143041e-01
3.47549558e-01 6.65883601e-01 -7.39175618e-01 7.89649606e-01
-1.69478714e-01 7.68635646e-02 1.44671649e-03 4.43135411e-01
-1.01284415e-01 -3.95397663e-01 7.02337623e-01 1.06775606e+00
3.00572306e-01 -4.73859347e-02 -1.39612213e-01 1.04191816e+00
-3.63686979e-01 2.07720920e-01 -6.90496445e-01 -2.55955011e-01
7.26906955e-02 1.15467405e+00 -9.15210187e-01 -3.22154433e-01
-3.66862476e-01 1.65939498e+00 -2.43260860e-01 4.31988716e-01
-9.27790999e-01 -7.54981279e-01 2.25578040e-01 6.30718768e-02
3.53028864e-01 -2.19471529e-01 -4.65271324e-01 -1.42853045e+00
3.52488816e-01 -1.13839281e+00 -1.81520313e-01 -1.22264624e+00
-8.09909940e-01 7.78404474e-01 -3.80103797e-01 -1.56128907e+00
2.47095302e-01 -3.06827813e-01 -5.79710543e-01 8.33983064e-01
-1.19986451e+00 -1.17956853e+00 -5.45124650e-01 6.03518069e-01
8.31011534e-01 -1.66613594e-01 4.82460350e-01 3.83373171e-01
-5.72653711e-01 5.59888422e-01 4.72544700e-01 2.13992372e-01
9.17645037e-01 -8.98814976e-01 4.19588298e-01 1.10469592e+00
7.20030740e-02 4.36680526e-01 6.98458314e-01 -9.89980638e-01
-1.28021383e+00 -1.24263966e+00 5.77292740e-01 1.05187781e-01
2.64188170e-01 -4.50222850e-01 -9.35461342e-01 6.15059614e-01
4.41959649e-01 -3.40632260e-01 1.04842454e-01 -8.31076205e-01
-1.78704694e-01 -1.73539132e-01 -1.12485266e+00 1.00095260e+00
5.48677206e-01 -4.04730052e-01 -4.74111289e-01 1.06810898e-01
7.75101006e-01 -6.82166636e-01 -3.89808565e-01 3.48556876e-01
5.73521674e-01 -8.98851335e-01 8.51317465e-01 -7.76222944e-02
7.63203561e-01 -6.04390562e-01 -1.94616586e-01 -1.30939865e+00
-1.22809313e-01 -6.17550552e-01 2.66506046e-01 1.29874194e+00
1.60208076e-01 -4.71806884e-01 4.07622725e-01 2.81871229e-01
1.75604761e-01 -2.84510583e-01 -6.20282531e-01 -4.96000588e-01
-1.37151107e-01 -9.49344039e-02 5.11629939e-01 9.59797084e-01
-5.30095935e-01 3.49397957e-01 -8.59793842e-01 1.10171959e-01
4.99580234e-01 2.06533328e-01 8.65306437e-01 -8.26473951e-01
-2.48794436e-01 -2.69561708e-01 -2.13556081e-01 -1.05501902e+00
-1.60154365e-02 -5.59193552e-01 8.34920704e-02 -1.39043736e+00
6.49203882e-02 -2.38605458e-02 2.55160391e-01 2.41582260e-01
-8.82006809e-02 4.34240729e-01 4.04038191e-01 2.08413392e-01
-9.82954428e-02 8.24752152e-01 1.66588426e+00 -4.80465353e-01
-2.69686967e-01 -5.21829575e-02 -8.03895235e-01 7.39873171e-01
7.30945528e-01 -3.53074461e-01 -4.62455660e-01 -5.70519567e-01
3.09851766e-02 2.26201549e-01 4.43193197e-01 -9.38050926e-01
1.55924127e-01 -2.47822344e-01 7.88176715e-01 -7.30621457e-01
2.93547809e-01 -1.17968595e+00 3.79335821e-01 4.27644372e-01
-3.28350306e-01 1.01189725e-01 3.89582008e-01 5.10633290e-01
-2.14699864e-01 -1.93536177e-01 1.21430457e+00 4.35622185e-02
-4.89976138e-01 1.05003938e-01 -3.60178262e-01 -2.57295072e-01
9.96385753e-01 -4.59904462e-01 -3.56368363e-01 -5.01598597e-01
-6.08715236e-01 -2.29870513e-01 7.15292335e-01 4.99037355e-01
9.06289995e-01 -1.28442752e+00 -7.18291044e-01 6.79032743e-01
-5.85404970e-02 -9.57682207e-02 3.31845611e-01 6.79356694e-01
-9.15870249e-01 1.05620407e-01 -3.29106301e-01 -4.21886027e-01
-1.40344620e+00 7.61336386e-01 3.46979767e-01 -1.02060623e-01
-9.07608211e-01 2.95887768e-01 5.84321380e-01 -1.27690390e-01
4.53417182e-01 -3.65352005e-01 1.04343392e-01 -1.40973374e-01
6.66778684e-01 1.89065307e-01 -5.39988629e-04 -7.26664186e-01
1.20961197e-01 7.56436110e-01 -3.69908243e-01 -3.32352281e-01
1.04942870e+00 -4.42387253e-01 -4.38156247e-01 3.10502529e-01
1.03836024e+00 4.61676389e-01 -1.10048831e+00 -3.70719343e-01
-5.14725626e-01 -7.19253719e-01 1.79935992e-01 -8.77700925e-01
-1.34026015e+00 8.37203026e-01 7.19146311e-01 -4.06771824e-02
1.43347192e+00 -5.76821744e-01 1.00830042e+00 3.31439048e-01
1.04576938e-01 -1.04956198e+00 2.52165616e-01 3.54363918e-01
1.21855044e+00 -9.11652744e-01 1.25156358e-01 -3.64878565e-01
-7.25738406e-01 1.23031616e+00 6.77805483e-01 1.29918456e-01
3.52220714e-01 2.81210691e-01 1.66786969e-01 1.25383064e-01
-3.96048456e-01 1.09430514e-01 2.15132952e-01 3.63302618e-01
4.56142873e-02 -2.59191126e-01 -9.64262784e-02 1.66134074e-01
-2.14984909e-01 4.81695868e-02 7.78601050e-01 9.79055882e-01
-3.57180148e-01 -1.11879718e+00 -7.70172477e-01 9.20965448e-02
-2.97441125e-01 -2.48046964e-01 -6.83501959e-01 7.43824422e-01
2.03900531e-01 8.32740963e-01 4.79962677e-02 -3.86078805e-01
3.69354367e-01 -7.01987296e-02 5.25481462e-01 -2.62766093e-01
-3.58983159e-01 3.88386458e-01 -2.90760368e-01 -1.80084333e-01
-2.82453299e-01 -2.73688823e-01 -1.20199692e+00 -4.97651547e-01
-2.76585549e-01 -2.60977596e-01 6.52285099e-01 6.66161299e-01
1.56814396e-01 7.12990701e-01 7.14490533e-01 -6.96493328e-01
-3.21261846e-02 -6.78435683e-01 -7.17936039e-01 3.29039603e-01
5.82787454e-01 -2.85846174e-01 -2.80065984e-01 6.36429429e-01] | [11.542257308959961, -0.6181567311286926] |
ec308bb9-5324-44bd-bef5-8e7b7a2e603d | placenta-segmentation-in-ultrasound-imaging | 2206.14746 | null | https://arxiv.org/abs/2206.14746v1 | https://arxiv.org/pdf/2206.14746v1.pdf | Placenta Segmentation in Ultrasound Imaging: Addressing Sources of Uncertainty and Limited Field-of-View | Automatic segmentation of the placenta in fetal ultrasound (US) is challenging due to the (i) high diversity of placenta appearance, (ii) the restricted quality in US resulting in highly variable reference annotations, and (iii) the limited field-of-view of US prohibiting whole placenta assessment at late gestation. In this work, we address these three challenges with a multi-task learning approach that combines the classification of placental location (e.g., anterior, posterior) and semantic placenta segmentation in a single convolutional neural network. Through the classification task the model can learn from larger and more diverse datasets while improving the accuracy of the segmentation task in particular in limited training set conditions. With this approach we investigate the variability in annotations from multiple raters and show that our automatic segmentations (Dice of 0.86 for anterior and 0.83 for posterior placentas) achieve human-level performance as compared to intra- and inter-observer variability. Lastly, our approach can deliver whole placenta segmentation using a multi-view US acquisition pipeline consisting of three stages: multi-probe image acquisition, image fusion and image segmentation. This results in high quality segmentation of larger structures such as the placenta in US with reduced image artifacts which are beyond the field-of-view of single probes. | ['Julia A. Schnabel', 'Joseph V. Hajnal', 'Daniel Rueckert', 'Bernhard Kainz', 'Jacqueline Matthew', 'Karen Lloyd', 'Nooshin Ghavami', 'Shujie Deng', 'Gavin Wheeler', 'Robert Wright', 'Emily Skelton', 'Alberto Gomez', 'Veronika A. Zimmer'] | 2022-06-29 | null | null | null | null | ['placenta-segmentation'] | ['medical'] | [ 3.66511136e-01 4.39803243e-01 3.77720088e-01 -5.56312263e-01
-1.18752456e+00 -1.17950344e+00 7.38971010e-02 4.48828757e-01
-2.90651709e-01 2.49592692e-01 -1.17651843e-01 -4.52500612e-01
-9.39173400e-02 -5.90946853e-01 -8.94002318e-01 -6.83326185e-01
-4.55954522e-02 6.02306724e-01 4.01471734e-01 3.58824819e-01
-2.10169792e-01 6.92412615e-01 -1.05512917e+00 4.19482112e-01
9.07419503e-01 1.06224847e+00 2.16634348e-01 1.14503658e+00
-1.25728399e-01 2.26718903e-01 -8.09893012e-02 -6.65488303e-01
3.20215046e-01 -4.66441542e-01 -8.24553847e-01 3.65670607e-03
7.14469612e-01 -5.27622759e-01 2.68502384e-01 6.04361296e-01
8.85353386e-01 -2.15319142e-01 5.64126194e-01 -6.19272053e-01
-3.49481776e-02 4.27710891e-01 -7.75688112e-01 3.77391726e-01
1.05402187e-01 2.17472110e-03 6.06259942e-01 -3.38099062e-01
6.25133455e-01 8.24362338e-01 8.27569187e-01 2.89201468e-01
-1.29460633e+00 -4.75761235e-01 -4.13582325e-01 -3.71376634e-01
-1.08191729e+00 -5.33083677e-01 2.60414153e-01 -9.60190058e-01
5.20457685e-01 3.19877386e-01 5.73348701e-01 5.68130493e-01
2.29022011e-01 4.50175822e-01 9.77635741e-01 -1.18177891e-01
1.41888916e-01 -1.12648336e-02 -3.29914480e-01 9.69416142e-01
4.00890231e-01 -3.66451561e-01 1.66920724e-03 -1.70255899e-01
1.06545150e+00 -2.10046753e-01 -1.00691207e-01 -7.49794602e-01
-9.24861491e-01 5.33268809e-01 9.37859863e-02 4.67192620e-01
-6.15748346e-01 -9.70844366e-03 7.80323029e-01 -1.97790474e-01
2.41830036e-01 4.19492215e-01 -4.82238770e-01 -2.40128592e-01
-1.15883911e+00 -3.60019147e-01 7.50514507e-01 7.55829215e-01
5.23210466e-01 -2.82115012e-01 -1.45517424e-01 9.79629755e-01
1.22603424e-01 4.99662936e-01 2.20387310e-01 -1.33625615e+00
4.26744729e-01 2.80869097e-01 -3.13195437e-02 -6.75175011e-01
-8.48121703e-01 -3.77550334e-01 -5.92344999e-01 2.49458522e-01
7.03674614e-01 -5.52891791e-01 -1.17686713e+00 1.62949717e+00
4.72805381e-01 1.12733074e-01 -2.31391340e-01 8.02839160e-01
1.25064540e+00 1.72393441e-01 1.18458413e-01 -2.22953826e-01
1.68140328e+00 -6.85088098e-01 -5.69335938e-01 -3.28604206e-02
7.30863869e-01 -8.19115818e-01 5.61808288e-01 2.11879387e-01
-1.64336705e+00 -1.87880576e-01 -8.48510921e-01 7.19445199e-02
5.74355982e-02 3.59561294e-01 5.42756677e-01 1.08895278e+00
-1.32954371e+00 4.90329623e-01 -1.47722304e+00 -4.66442049e-01
9.33761954e-01 5.36864340e-01 -7.77156055e-01 2.35095918e-02
-3.89371157e-01 1.02687562e+00 2.01268464e-01 -8.77159908e-02
-5.80864727e-01 -1.27462137e+00 -1.12121081e+00 3.41614902e-01
2.80951500e-01 -3.07343125e-01 1.22181404e+00 -7.44352341e-01
-1.28401315e+00 1.19982111e+00 -1.67866752e-01 1.54174581e-01
8.73464644e-01 1.14085518e-01 8.26361030e-02 7.95300007e-01
3.11209232e-01 5.02941132e-01 2.24107906e-01 -1.46412373e+00
-6.11521959e-01 -3.57402354e-01 -2.62635112e-01 -1.47756547e-01
1.78953215e-01 1.76255200e-02 -5.81658423e-01 2.78441962e-02
4.74714994e-01 -9.02202964e-01 -1.41804308e-01 2.70919263e-01
3.07495832e-01 2.11332411e-01 2.95057416e-01 -1.02068007e+00
7.71651447e-01 -1.97769880e+00 -3.04660469e-01 3.20394546e-01
4.35687035e-01 5.64062119e-01 -1.09334327e-01 -1.03696272e-01
-2.09803879e-01 2.21989498e-01 -8.86527821e-02 -3.42741340e-01
-4.21926737e-01 2.75993645e-01 5.60128033e-01 6.52292907e-01
1.67466849e-01 1.11794281e+00 -1.19733977e+00 -8.22355866e-01
3.11323166e-01 3.40001494e-01 -4.30940837e-01 6.22252047e-01
5.16927719e-01 7.25870073e-01 -3.86898428e-01 6.37877107e-01
9.26940560e-01 -3.78209472e-01 3.74880552e-01 -4.54374850e-01
-2.04796419e-01 -5.29848896e-02 -9.96654451e-01 2.01133275e+00
-7.05451310e-01 5.26294649e-01 4.27818447e-01 -9.15758252e-01
6.22383893e-01 9.35021818e-01 8.55771542e-01 -4.61214095e-01
3.32754076e-01 5.01191914e-01 4.15519416e-01 -1.08678293e+00
-1.08467110e-01 -3.76225561e-01 5.76483011e-01 5.64648390e-01
4.59727407e-01 -2.14930624e-01 4.95062858e-01 -5.02115954e-03
1.04704714e+00 1.63771302e-01 1.82525754e-01 -4.67223108e-01
1.79892316e-01 -5.05763888e-01 6.06727540e-01 8.35929215e-01
-3.61398488e-01 1.42691064e+00 9.49325442e-01 -4.23159331e-01
-1.15846217e+00 -1.15498149e+00 -6.15637124e-01 7.96517909e-01
-1.06906585e-01 2.33395934e-01 -8.41878235e-01 -7.94930577e-01
-3.23258415e-02 3.46551508e-01 -1.03790033e+00 4.18438822e-01
-7.46550560e-01 -6.74444199e-01 8.23066592e-01 5.63398123e-01
-9.32690650e-02 -7.33656645e-01 -1.07025611e+00 1.77455708e-01
-2.52026230e-01 -1.32845819e+00 -1.12546697e-01 5.77114560e-02
-6.51918828e-01 -1.18225169e+00 -1.15228975e+00 -6.76108956e-01
7.83572793e-01 -1.24000907e-01 1.18441832e+00 -8.81526843e-02
-3.76492202e-01 4.12116021e-01 -3.43173087e-01 -3.86597574e-01
-6.90578818e-01 -1.15802817e-01 -4.85551149e-01 -1.79741662e-02
-1.80331349e-01 -7.31030166e-01 -9.71888006e-01 8.06064010e-02
-8.32811475e-01 -4.50759418e-02 7.92764187e-01 5.58820844e-01
1.92730710e-01 -7.90858567e-01 4.02508408e-01 -1.04799199e+00
-9.15259868e-02 -7.40668356e-01 -6.95585132e-01 5.61862648e-01
-1.25166431e-01 -4.82633650e-01 1.85864523e-01 -1.86036572e-01
-1.20785141e+00 -2.26428658e-01 -2.56820500e-01 -2.01406315e-01
-3.96863371e-01 2.87019163e-01 3.48393142e-01 -3.95771176e-01
5.53121924e-01 -6.30763173e-02 4.08287257e-01 -3.28442127e-01
-1.52703896e-01 5.15344083e-01 3.35134298e-01 -5.35958111e-01
4.32576016e-02 5.53136349e-01 3.32054496e-01 -5.39093971e-01
-6.30602539e-01 -4.27564979e-01 -1.05452645e+00 -3.28749418e-01
1.05549443e+00 -6.08723640e-01 -6.24806881e-01 -8.81611556e-02
-1.15288103e+00 -2.67134607e-01 5.56603819e-02 6.42204702e-01
-4.85286325e-01 6.07158005e-01 -8.35649610e-01 -6.30284369e-01
-4.79588181e-01 -1.63466930e+00 1.17934060e+00 4.33261275e-01
-2.24370226e-01 -1.14680898e+00 -6.12907000e-02 4.89837408e-01
5.35388649e-01 7.37251461e-01 7.86451399e-01 -7.37011075e-01
-2.03222170e-01 -1.77360624e-01 -6.08018756e-01 2.40339741e-01
5.68585157e-01 1.62008926e-01 -1.08778095e+00 -2.12791160e-01
-2.89661199e-01 -2.64190018e-01 4.74636465e-01 9.26187098e-01
8.37734044e-01 3.09668243e-01 -2.36580282e-01 7.52779126e-01
1.31891394e+00 2.87239373e-01 3.02427322e-01 -6.71944171e-02
6.58419371e-01 7.25836575e-01 4.44358140e-01 4.68989074e-01
2.51988888e-01 1.39066890e-01 3.59377146e-01 -6.38552129e-01
-8.23557600e-02 4.38879192e-01 -6.65076733e-01 3.59235972e-01
-3.23968232e-01 -2.31895581e-01 -1.36091077e+00 1.02113116e+00
-1.48768961e+00 -3.96462590e-01 -1.81751281e-01 2.03390241e+00
7.70320773e-01 -2.57464737e-01 4.34032315e-03 -4.69743043e-01
5.62104464e-01 -2.30996773e-01 -6.66823387e-02 -7.48081446e-01
1.57022282e-01 1.65690392e-01 5.74105084e-01 4.19856459e-01
-1.04020035e+00 4.64956284e-01 6.28739977e+00 4.05950874e-01
-1.31506014e+00 2.77837753e-01 9.38957989e-01 1.47918081e-02
-1.23243064e-01 -4.04556304e-01 -6.06320538e-02 3.33257735e-01
6.61654353e-01 3.80960137e-01 7.74942990e-03 2.73686230e-01
8.95421803e-02 -2.98667967e-01 -1.24344039e+00 8.11025321e-01
7.70304352e-02 -1.30132556e+00 -3.13719571e-01 -1.62477359e-01
8.29944491e-01 3.00821453e-01 -3.50106955e-01 -3.74651134e-01
-1.53533518e-01 -1.07754135e+00 3.92540276e-01 2.69468784e-01
1.21248090e+00 -4.00669485e-01 1.04041195e+00 3.91516909e-02
-8.10385406e-01 3.96934211e-01 -7.84704462e-02 5.85298419e-01
2.40115166e-01 3.28239739e-01 -1.02903354e+00 5.79144478e-01
7.96332002e-01 2.58104950e-01 -2.85162479e-01 1.36216819e+00
2.34225929e-01 6.87588751e-01 -4.83339399e-01 4.01710510e-01
3.53482306e-01 -1.96554348e-01 3.84192050e-01 1.85418260e+00
4.23972428e-01 4.43269163e-01 -2.63478965e-01 7.65114307e-01
1.98622003e-01 1.09114155e-01 -2.37925529e-01 2.58963704e-01
1.74971148e-01 1.63821101e+00 -1.12532890e+00 -2.88681507e-01
-6.95639491e-01 4.28496778e-01 2.69461274e-01 3.43963385e-01
-4.49584723e-01 3.42228333e-03 7.14129582e-02 8.93932059e-02
4.67051834e-01 1.71517253e-01 -6.93439484e-01 -6.86531842e-01
9.03696194e-02 -3.61750215e-01 4.97109115e-01 -5.12320161e-01
-7.66778409e-01 4.63786632e-01 -1.29207194e-01 -1.00448108e+00
2.31567062e-02 -3.61605316e-01 -6.20691836e-01 9.74480450e-01
-1.58400428e+00 -1.21225893e+00 -4.52059209e-01 -2.70007521e-01
1.07904986e-01 1.19562559e-01 8.61220360e-01 6.58639967e-01
-1.54046848e-01 7.42213666e-01 6.54875580e-03 3.44037801e-01
5.71174026e-01 -1.56578600e+00 -1.17981909e-02 6.00456297e-01
-6.64820731e-01 6.12823308e-01 4.41547275e-01 -4.60148871e-01
-1.07418048e+00 -7.29130208e-01 5.61759293e-01 -4.56302583e-01
4.01823670e-01 -6.36977926e-02 -9.73567605e-01 7.84763396e-01
1.56632796e-01 3.88800204e-01 1.22061086e+00 -1.15959443e-01
6.77045956e-02 -4.92080934e-02 -1.52074420e+00 3.00200731e-01
3.94734085e-01 2.38995984e-01 -8.78840685e-02 1.44187838e-01
2.00412780e-01 -1.11382067e+00 -1.36351359e+00 6.38366818e-01
1.02280962e+00 -1.11257482e+00 7.04368532e-01 -4.48875934e-01
7.87146688e-01 6.90224990e-02 3.53298992e-01 -8.82329524e-01
-2.31011093e-01 -4.68373239e-01 4.17853504e-01 1.30845392e+00
6.36223376e-01 -4.77927089e-01 7.63819695e-01 9.87854064e-01
-2.10532784e-01 -7.79687583e-01 -1.11781991e+00 1.47857042e-02
1.26740068e-01 -1.37822196e-01 2.72095680e-01 9.07759249e-01
-5.53097241e-02 -1.94828942e-01 1.82334319e-01 3.98263365e-01
5.21171451e-01 2.55850971e-01 3.32640231e-01 -9.28764045e-01
-1.13392010e-01 -4.75009471e-01 -3.64850819e-01 -5.21725178e-01
-3.01467687e-01 -8.99065077e-01 1.58362851e-01 -1.72370803e+00
5.46718597e-01 -8.93323362e-01 -2.08717555e-01 5.08808553e-01
-3.66548300e-01 4.03233558e-01 4.86674625e-03 -2.83166647e-01
-6.53278112e-01 -2.18340799e-01 1.54389930e+00 4.18238550e-01
2.03543343e-02 4.23917845e-02 -4.80091125e-01 6.09860778e-01
4.44294512e-01 -4.44400281e-01 -4.39196751e-02 -8.06627154e-01
-1.52722851e-03 6.92017913e-01 -2.86838622e-03 -5.47679007e-01
-7.27237463e-02 7.75252953e-02 6.10872984e-01 -2.80260175e-01
1.35912791e-01 -8.01212311e-01 -9.98492315e-02 4.63090807e-01
-1.66166097e-01 -3.58101308e-01 4.27827179e-01 -4.03959006e-02
1.37726992e-01 -4.30442721e-01 1.09437692e+00 -4.11348671e-01
1.23882413e-01 1.81074753e-01 -3.53194863e-01 8.67903829e-02
7.18507409e-01 -3.55559558e-01 1.86091959e-01 -2.45601177e-01
-9.33033764e-01 5.37845910e-01 2.89410263e-01 2.04000756e-01
1.50321007e-01 -6.13850713e-01 -1.21683550e+00 2.25343443e-02
-1.01097301e-01 5.13390541e-01 4.06699538e-01 1.47448492e+00
-1.25145352e+00 2.33350232e-01 -2.81803042e-01 -1.24899459e+00
-1.40868390e+00 -6.69342279e-02 6.81286573e-01 -1.46141991e-01
-5.22505045e-01 1.00212538e+00 4.68457669e-01 -5.65344989e-01
-1.08959712e-01 -1.71786815e-01 -4.38721120e-01 5.91862053e-02
3.64647657e-01 4.91476744e-01 4.06950533e-01 -7.10844934e-01
-3.30612451e-01 8.01722586e-01 -6.97956514e-03 1.92256272e-01
1.33285117e+00 -2.39703134e-01 -1.93875745e-01 1.47530347e-01
1.35271955e+00 4.86132614e-02 -1.35929108e+00 -3.40340510e-02
-2.14473039e-01 -6.55672789e-01 -1.29206717e-01 -8.63355756e-01
-1.24099469e+00 1.04358971e+00 7.52620161e-01 2.50390500e-01
8.91598701e-01 2.05659866e-01 6.19427919e-01 -2.75762945e-01
-8.33765715e-02 -6.76943839e-01 -3.80830348e-01 2.27655657e-02
5.96460879e-01 -1.56299043e+00 -3.07296626e-02 -5.70868313e-01
-6.49129331e-01 1.28737032e+00 3.46257627e-01 6.26034811e-02
3.09520990e-01 6.31508589e-01 6.43717408e-01 -1.81486100e-01
-3.41624796e-01 -3.43728028e-02 4.69218969e-01 4.46964473e-01
8.81551087e-01 9.41573270e-03 -4.96251792e-01 5.13715744e-01
2.62197331e-02 -1.05782755e-01 6.64752483e-01 1.00707185e+00
-2.84741968e-01 -6.16372943e-01 -1.01934075e-01 6.94712043e-01
-1.23208344e+00 2.87122697e-01 2.81021386e-01 4.35312778e-01
3.71582925e-01 9.17838752e-01 1.06814332e-01 5.47166228e-01
3.65662813e-01 -2.41878629e-01 6.79096043e-01 -7.91565239e-01
-7.91319668e-01 5.26897252e-01 1.04312576e-01 -6.60595655e-01
-1.43423960e-01 -1.14154541e+00 -1.39712524e+00 2.36489877e-01
-3.71236920e-01 1.05366260e-01 8.54634583e-01 1.15598154e+00
1.51042461e-01 7.16552734e-01 4.86257225e-01 -9.01449680e-01
4.28589322e-02 -8.26598227e-01 -6.28581882e-01 4.01328474e-01
6.93815649e-01 -2.91611850e-01 -3.50500755e-02 1.98697999e-01] | [14.254617691040039, -2.4278461933135986] |
9db3b3b1-18f9-48c9-895f-aade9b6c3ec7 | smoothed-multi-view-subspace-clustering | 2106.09875 | null | https://arxiv.org/abs/2106.09875v1 | https://arxiv.org/pdf/2106.09875v1.pdf | Smoothed Multi-View Subspace Clustering | In recent years, multi-view subspace clustering has achieved impressive performance due to the exploitation of complementary imformation across multiple views. However, multi-view data can be very complicated and are not easy to cluster in real-world applications. Most existing methods operate on raw data and may not obtain the optimal solution. In this work, we propose a novel multi-view clustering method named smoothed multi-view subspace clustering (SMVSC) by employing a novel technique, i.e., graph filtering, to obtain a smooth representation for each view, in which similar data points have similar feature values. Specifically, it retains the graph geometric features through applying a low-pass filter. Consequently, it produces a ``clustering-friendly" representation and greatly facilitates the downstream clustering task. Extensive experiments on benchmark datasets validate the superiority of our approach. Analysis shows that graph filtering increases the separability of classes. | ['Zhao Kang', 'Zhengrui Ma', 'Liang Liu', 'Peng Chen'] | 2021-06-18 | null | null | null | null | ['multi-view-subspace-clustering'] | ['computer-vision'] | [-2.62310058e-01 -5.26560485e-01 -7.22565576e-02 -1.60890386e-01
-5.00822663e-01 -7.41883755e-01 4.15394098e-01 1.73395798e-02
1.53064325e-01 1.63284525e-01 3.22257847e-01 8.28919932e-02
-2.89603770e-01 -5.92204928e-01 -2.16493741e-01 -1.11401916e+00
2.65906572e-01 4.02054600e-02 2.15844333e-01 9.06965509e-02
2.07754955e-01 3.27078432e-01 -1.42058074e+00 2.52202928e-01
1.03733337e+00 5.24963558e-01 2.19643593e-01 2.29793400e-01
-6.89976811e-02 1.47247821e-01 -2.24368855e-01 -1.07791252e-01
2.67655998e-01 -4.69566792e-01 -3.67282718e-01 6.05467081e-01
1.71623513e-01 1.43258765e-01 -7.70777091e-02 1.28984809e+00
4.26130563e-01 2.28308812e-01 4.80467260e-01 -1.25443852e+00
-6.09100819e-01 2.37225994e-01 -1.18045282e+00 -6.56059012e-02
3.56315166e-01 -2.29808092e-01 8.25613916e-01 -1.10360265e+00
7.05007255e-01 1.31475210e+00 5.26298344e-01 2.29078859e-01
-1.57915771e+00 -4.40750301e-01 3.72755706e-01 2.37145841e-01
-1.61892080e+00 -2.85177797e-01 1.18851459e+00 -3.68254781e-01
2.59715319e-01 4.26576465e-01 7.00196147e-01 5.49777865e-01
-2.41417084e-02 8.73083413e-01 1.27874267e+00 -3.09641398e-02
1.47338048e-01 1.20337214e-02 4.86061014e-02 4.63407844e-01
2.80682921e-01 -3.39364797e-01 -2.14976743e-01 -3.47781748e-01
4.91223365e-01 5.62760174e-01 -6.23163760e-01 -1.08939052e+00
-1.48988485e+00 7.93586433e-01 3.65497202e-01 3.94838929e-01
-2.72473484e-01 -3.64818543e-01 4.81801689e-01 2.72598006e-02
4.82683808e-01 -1.15076765e-01 4.66042571e-02 1.82174310e-01
-6.87032998e-01 -5.31557314e-02 4.16293323e-01 1.07006240e+00
7.54888415e-01 -6.62594959e-02 3.69296759e-01 9.65401709e-01
3.23903412e-01 4.96082246e-01 4.06298369e-01 -8.09944034e-01
4.78009880e-01 1.06075990e+00 -2.85031796e-01 -1.45721102e+00
-3.58142942e-01 -4.65357840e-01 -1.41633701e+00 7.79578015e-02
2.22333342e-01 -9.16897261e-04 -5.53254724e-01 1.36762965e+00
5.96257687e-01 3.11147958e-01 3.53534967e-01 9.49384332e-01
9.56708133e-01 5.27359068e-01 -1.61788955e-01 -4.79281306e-01
1.14519942e+00 -7.61159360e-01 -7.61084139e-01 1.80328369e-01
3.59019428e-01 -9.14538920e-01 8.18548262e-01 5.92793763e-01
-9.02164817e-01 -6.25667572e-01 -9.72295821e-01 3.90400410e-01
-1.93441257e-01 9.06432122e-02 5.91802478e-01 4.94946331e-01
-8.20408583e-01 4.09301758e-01 -7.96163917e-01 -4.50535417e-01
2.37367436e-01 2.44964615e-01 -6.89966321e-01 -2.32489794e-01
-5.35060406e-01 1.57741189e-01 5.29779315e-01 7.46895820e-02
-3.11313272e-01 -3.81808132e-01 -7.24863172e-01 -2.44799536e-03
6.53112113e-01 -6.31112278e-01 4.82364893e-01 -6.93670988e-01
-1.04181159e+00 5.26222110e-01 -4.45528328e-01 2.31736884e-01
2.98872232e-01 -3.29359472e-02 -6.48815095e-01 3.84758502e-01
2.54902214e-01 6.31650686e-02 9.45818484e-01 -1.84059572e+00
-5.80013156e-01 -7.55677164e-01 -3.43724161e-01 4.57910329e-01
-3.79451066e-01 -1.39629513e-01 -8.40583205e-01 -6.77201211e-01
8.74468029e-01 -9.76855695e-01 -5.04553854e-01 -4.18962836e-01
-5.17263532e-01 -1.51869580e-01 1.32810533e+00 -5.04345655e-01
1.46517324e+00 -2.48056912e+00 6.10330164e-01 4.38072324e-01
5.09414196e-01 2.04753084e-03 2.37162858e-01 6.02231562e-01
-2.14864820e-01 7.80390901e-03 -2.10857555e-01 -1.43410251e-01
-4.08304900e-01 2.63119638e-02 3.78154181e-02 8.52890253e-01
-3.99788618e-01 4.85610932e-01 -9.71892715e-01 -7.45359004e-01
3.97911072e-01 2.51233637e-01 -6.25287414e-01 3.35922800e-02
2.88804591e-01 5.86248040e-01 -5.68287492e-01 6.25992298e-01
1.06512535e+00 -4.86405134e-01 4.65667844e-01 -4.27845269e-01
-1.69317260e-01 -7.54141748e-01 -1.64434779e+00 1.90847027e+00
-7.12010115e-02 7.43429139e-02 3.79093200e-01 -1.09877467e+00
8.29766273e-01 2.28132859e-01 8.23623836e-01 -2.19489597e-02
7.68966973e-02 1.31963983e-01 -1.30852461e-01 -3.11072469e-01
2.52325326e-01 -8.90387371e-02 3.77451777e-02 3.53215247e-01
-2.04967931e-01 1.18930303e-01 1.28181443e-01 4.52043712e-01
5.21849275e-01 -4.49753553e-02 4.67510521e-01 -4.59664255e-01
9.99441445e-01 -1.01769373e-01 6.93087399e-01 2.27035414e-02
-1.17414378e-01 7.16382802e-01 1.62807688e-01 -1.19061872e-01
-7.08105028e-01 -1.24948454e+00 -3.01357377e-02 5.44958889e-01
6.16770744e-01 -6.44436300e-01 -7.94680417e-01 -8.01033616e-01
-4.16666688e-03 2.76503235e-01 -3.42208117e-01 -2.04292297e-01
-5.31848848e-01 -7.45464146e-01 -1.08850382e-01 4.97228026e-01
6.05489850e-01 -3.38160783e-01 -1.78655162e-01 1.01106592e-01
-3.08458954e-01 -9.87616897e-01 -7.48535752e-01 -3.62550169e-01
-1.21726787e+00 -1.19506252e+00 -7.97184169e-01 -8.60787332e-01
9.62860763e-01 1.24853647e+00 6.85676455e-01 1.95305869e-01
6.84894919e-02 4.13987279e-01 -5.04407883e-01 8.76405612e-02
-4.44873646e-02 -3.65119539e-02 2.76761860e-01 5.62194645e-01
4.63122457e-01 -8.15707684e-01 -7.85340011e-01 5.04557073e-01
-8.40533078e-01 9.80073661e-02 5.02986193e-01 8.27202201e-01
9.83049750e-01 5.06387234e-01 4.73260432e-01 -8.59203577e-01
5.88684022e-01 -4.61209357e-01 -3.84279937e-01 3.67133766e-01
-6.67916656e-01 -2.72508204e-01 1.20560265e+00 -1.14656001e-01
-1.08708310e+00 3.82974684e-01 3.15225124e-01 -8.64489019e-01
-3.19424540e-01 4.90236133e-01 -5.72465003e-01 -1.07415795e-01
1.73670322e-01 6.22326553e-01 2.24518821e-01 -6.33363485e-01
6.90065622e-01 6.24192059e-01 4.35699642e-01 -2.62282908e-01
1.19630063e+00 1.00611734e+00 1.67642400e-01 -9.16431367e-01
-5.78146577e-01 -9.16699946e-01 -9.92814004e-01 -3.76317173e-01
9.66802955e-01 -1.03796661e+00 -7.66478181e-01 3.64553809e-01
-7.41574228e-01 4.88847584e-01 2.33459324e-01 7.31938243e-01
-4.54002410e-01 1.04508066e+00 -2.83909976e-01 -5.98552883e-01
-1.65835649e-01 -9.99939024e-01 9.48028982e-01 3.47289175e-01
1.13361277e-01 -1.17034316e+00 -6.72428757e-02 2.18329921e-01
-9.79748070e-02 5.06515682e-01 8.81782174e-01 -4.00956780e-01
-5.22282362e-01 -6.89049438e-02 -2.08985254e-01 3.74876671e-02
4.14054990e-01 1.41000718e-01 -7.24503636e-01 -6.98370934e-01
2.08848998e-01 2.99092054e-01 6.78699851e-01 4.11790073e-01
1.27646780e+00 5.38366623e-02 -6.54080272e-01 7.89172232e-01
1.71422601e+00 3.51278096e-01 2.87767619e-01 1.19848073e-01
1.09428692e+00 6.15529537e-01 7.15276182e-01 4.98663008e-01
3.23742926e-01 6.15968645e-01 3.14143330e-01 -2.53081828e-01
1.71644703e-01 -2.10549027e-01 -7.94844516e-03 1.50213575e+00
-3.44296843e-01 1.55643607e-02 -8.37996781e-01 4.65486944e-01
-2.08040977e+00 -1.16552269e+00 -6.10097826e-01 2.14271593e+00
3.44298854e-02 -1.75965935e-01 3.37187350e-01 1.67897508e-01
1.09812868e+00 2.75933951e-01 -5.32297492e-01 5.80147505e-02
-1.79568559e-01 -4.62067515e-01 2.94666380e-01 1.09518811e-01
-1.24508357e+00 6.82541192e-01 5.51328325e+00 8.17706347e-01
-7.32275367e-01 -2.83608157e-02 2.97872394e-01 1.43238589e-01
-4.13399458e-01 7.89454356e-02 -4.79960799e-01 4.11878020e-01
3.22785437e-01 -5.18693924e-01 4.23071533e-01 9.71727550e-01
3.54863554e-01 -1.13694398e-02 -6.86779737e-01 1.32525396e+00
2.56482303e-01 -1.08109546e+00 2.29106292e-01 1.32223144e-01
7.09842145e-01 -4.55214947e-01 -1.16564054e-02 -1.48915514e-01
2.23543510e-01 -5.14499485e-01 3.24029714e-01 3.76389146e-01
6.67222261e-01 -1.24926877e+00 4.29043800e-01 4.13589299e-01
-1.84152222e+00 -1.34368345e-01 -4.22478586e-01 3.45076382e-01
2.44422331e-01 5.09786248e-01 -3.08262795e-01 1.35537934e+00
6.92086220e-01 1.32617974e+00 -7.17766762e-01 9.37709928e-01
1.54121563e-01 2.70733863e-01 -1.31145036e-02 2.56353229e-01
1.94029108e-01 -9.66225803e-01 6.58342540e-01 8.17315757e-01
3.55442226e-01 3.43556613e-01 5.73225200e-01 4.88886803e-01
4.58606482e-01 4.98694003e-01 -8.96633983e-01 2.06478953e-01
4.23870802e-01 1.46642029e+00 -1.18338275e+00 -2.93500423e-01
-8.57740045e-01 9.92836177e-01 2.12175936e-01 4.66632903e-01
-6.88399911e-01 -4.27666634e-01 4.16736454e-01 -1.25652671e-01
3.62071395e-01 -4.10767198e-01 -4.00718242e-01 -1.56290019e+00
3.01500745e-02 -9.15341377e-01 6.94677651e-01 -4.05910343e-01
-1.29769409e+00 4.53303695e-01 -6.61294088e-02 -1.93730092e+00
1.69933382e-02 -1.73624799e-01 -6.45725191e-01 6.94100142e-01
-1.05329263e+00 -1.13985407e+00 -5.23274720e-01 1.03375793e+00
4.65629369e-01 -2.12019101e-01 6.85977995e-01 2.78159976e-01
-7.59036541e-01 2.23271459e-01 6.85105562e-01 -2.60691326e-02
6.46254182e-01 -1.31233454e+00 -2.29943991e-02 1.00573552e+00
2.51140594e-02 9.00985479e-01 4.36479151e-01 -6.82796299e-01
-1.79136884e+00 -1.10408092e+00 1.21547438e-01 -1.83301941e-01
5.40745318e-01 -3.21807027e-01 -9.98472691e-01 5.72503567e-01
2.71719843e-01 -2.13917136e-01 1.15573668e+00 9.66147780e-02
-1.68836564e-01 -1.51768893e-01 -9.22089279e-01 6.79235220e-01
1.02468693e+00 -4.66056287e-01 -5.18012285e-01 2.29810119e-01
4.72345233e-01 -9.70016643e-02 -1.15667486e+00 3.25991571e-01
3.53506625e-01 -1.28071415e+00 1.15936279e+00 -4.48874921e-01
1.60492450e-01 -7.86734760e-01 -2.38340989e-01 -1.50676858e+00
-6.94897234e-01 -4.29731011e-01 -4.70445976e-02 1.53584087e+00
-1.07007280e-01 -7.89565623e-01 7.79313862e-01 1.86089545e-01
-2.23012790e-01 -6.76882386e-01 -7.32081413e-01 -9.54504609e-01
-2.40510657e-01 7.52821267e-02 6.44912124e-01 1.26706922e+00
1.25345632e-01 3.44007611e-01 -2.23385960e-01 4.65405941e-01
1.12832963e+00 9.56646681e-01 9.13948596e-01 -1.43226945e+00
-1.09323137e-01 -4.05356616e-01 -4.78620261e-01 -7.43813038e-01
1.02657318e-01 -9.70134735e-01 -3.17829281e-01 -1.62637675e+00
5.39259255e-01 -2.27450922e-01 -1.28004327e-01 -1.48677945e-01
-5.13874471e-01 4.47139982e-03 2.57116407e-01 6.03434801e-01
-7.70400763e-01 5.82464635e-01 1.42013502e+00 1.42192170e-01
-2.93097228e-01 7.76484907e-02 -7.91070461e-01 8.95898044e-01
7.52213359e-01 -5.00532277e-02 -5.30133963e-01 -2.14823619e-01
-3.23559076e-01 1.24511570e-01 1.34817168e-01 -9.94177341e-01
2.67526060e-01 -2.42679194e-01 4.14181948e-01 -9.53762591e-01
2.50711858e-01 -1.15378726e+00 5.72560728e-01 4.89439040e-01
3.56214464e-01 9.32899192e-02 -8.78036469e-02 1.08667147e+00
-5.31842172e-01 1.98952988e-01 8.51923704e-01 -2.94173449e-01
-9.22986627e-01 3.46249551e-01 -4.92500030e-02 -1.04774080e-01
1.46465635e+00 -4.21594918e-01 -7.91031569e-02 -2.87217945e-01
-7.88269460e-01 4.54024047e-01 9.67049599e-01 2.83507884e-01
7.48484969e-01 -1.78688669e+00 -4.96926159e-01 2.73766905e-01
2.66073763e-01 8.09406713e-02 5.89313567e-01 1.01548290e+00
-2.83903152e-01 2.10870877e-01 -1.17086820e-01 -9.20997679e-01
-1.68422377e+00 1.12602520e+00 -1.37351841e-01 9.79096908e-03
-9.23865914e-01 3.32531691e-01 4.58200008e-01 -2.09579512e-01
-1.79035649e-01 1.92903116e-01 -5.99068880e-01 4.01135832e-01
4.45367277e-01 6.36191368e-01 -2.74051309e-01 -1.02405155e+00
-3.92907560e-01 1.14286315e+00 8.32975209e-02 1.74371064e-01
1.45436299e+00 -5.53865373e-01 -2.86612183e-01 4.07627553e-01
1.33731782e+00 2.40394682e-01 -1.08705151e+00 -1.52978912e-01
-9.78277624e-02 -9.14706767e-01 -9.37436745e-02 1.87916204e-01
-1.35698378e+00 7.95927644e-01 3.05358440e-01 5.48833907e-01
1.39108860e+00 -2.07245280e-03 6.97377384e-01 6.80515021e-02
4.26762134e-01 -1.00538146e+00 -2.26598103e-02 -3.58765647e-02
6.02257371e-01 -1.03669548e+00 2.19775990e-01 -1.00024819e+00
-7.04982996e-01 1.11390054e+00 6.07205927e-01 -1.03620447e-01
7.11876512e-01 -3.41319203e-01 2.82549076e-02 -4.50436622e-01
-3.65056515e-01 -9.90400389e-02 2.62374967e-01 4.64222431e-01
2.32827812e-01 1.52186424e-01 -3.34150314e-01 4.59504545e-01
1.20845668e-01 -5.06322563e-01 4.52411294e-01 8.10286820e-01
-3.57035875e-01 -9.16134119e-01 -6.63622558e-01 5.09505212e-01
-2.68304974e-01 2.25545168e-01 -3.55572045e-01 7.21786559e-01
-7.83007443e-02 1.13180816e+00 -2.14756653e-01 -6.28431380e-01
5.33619881e-01 -1.87696159e-01 1.86522845e-02 -6.15520954e-01
-2.56648630e-01 6.20484591e-01 -3.52146626e-01 -6.28666282e-01
-8.24254632e-01 -1.04048383e+00 -1.20376909e+00 -3.85987818e-01
-4.31506813e-01 3.34809691e-01 2.44632632e-01 5.23988307e-01
5.04641891e-01 3.62870932e-01 1.14358187e+00 -7.20177472e-01
-3.11452389e-01 -3.10454071e-01 -9.70225155e-01 8.13068509e-01
-3.03255972e-02 -7.77626097e-01 -4.07817692e-01 1.29451528e-01] | [8.17601490020752, 4.613504409790039] |
a2d9ff5d-9ca6-41ba-9a2a-5c6013bd5503 | phenaki-variable-length-video-generation-from | 2210.02399 | null | https://arxiv.org/abs/2210.02399v1 | https://arxiv.org/pdf/2210.02399v1.pdf | Phenaki: Variable Length Video Generation From Open Domain Textual Description | We present Phenaki, a model capable of realistic video synthesis, given a sequence of textual prompts. Generating videos from text is particularly challenging due to the computational cost, limited quantities of high quality text-video data and variable length of videos. To address these issues, we introduce a new model for learning video representation which compresses the video to a small representation of discrete tokens. This tokenizer uses causal attention in time, which allows it to work with variable-length videos. To generate video tokens from text we are using a bidirectional masked transformer conditioned on pre-computed text tokens. The generated video tokens are subsequently de-tokenized to create the actual video. To address data issues, we demonstrate how joint training on a large corpus of image-text pairs as well as a smaller number of video-text examples can result in generalization beyond what is available in the video datasets. Compared to the previous video generation methods, Phenaki can generate arbitrary long videos conditioned on a sequence of prompts (i.e. time variable text or a story) in open domain. To the best of our knowledge, this is the first time a paper studies generating videos from time variable prompts. In addition, compared to the per-frame baselines, the proposed video encoder-decoder computes fewer tokens per video but results in better spatio-temporal consistency. | ['Dumitru Erhan', 'Julius Kunze', 'Santiago Castro', 'Mohammad Taghi Saffar', 'Han Zhang', 'Hernan Moraldo', 'Pieter-Jan Kindermans', 'Mohammad Babaeizadeh', 'Ruben Villegas'] | 2022-10-05 | null | null | null | null | ['video-generation', 'video-prediction'] | ['computer-vision', 'computer-vision'] | [ 6.53858066e-01 1.53274417e-01 -2.80615836e-01 -1.36545852e-01
-9.69962358e-01 -6.20611846e-01 8.30669343e-01 -3.57539803e-01
-2.84080923e-01 9.30028915e-01 4.27381068e-01 -1.91913754e-01
3.22866797e-01 -5.87300181e-01 -1.26045728e+00 -5.51874042e-01
-1.06536753e-01 1.96637183e-01 1.68957189e-01 2.68655479e-01
1.40172303e-01 3.79158147e-02 -1.77306867e+00 7.71534264e-01
3.90270948e-01 8.39022696e-01 5.36704302e-01 1.21064949e+00
-1.13861859e-01 1.34184980e+00 -7.41515934e-01 -3.26440483e-01
2.96461940e-01 -6.76127613e-01 -6.96443379e-01 3.50146979e-01
8.93649995e-01 -8.62343788e-01 -7.76259899e-01 5.48057795e-01
2.80580342e-01 3.05301875e-01 5.76971889e-01 -1.54340684e+00
-8.70070219e-01 9.96107936e-01 -2.87783116e-01 2.17400864e-01
7.49416351e-01 3.38355035e-01 9.87088263e-01 -8.62451077e-01
1.04087079e+00 1.26301455e+00 3.14088941e-01 7.11033940e-01
-1.12003851e+00 -7.32759058e-01 3.67610872e-01 4.58064526e-01
-1.20224118e+00 -5.16417980e-01 4.97143060e-01 -5.27376592e-01
1.24202585e+00 2.01913416e-01 7.89844036e-01 1.66657710e+00
1.06398381e-01 9.02886093e-01 6.27917171e-01 -4.03967679e-01
7.02934936e-02 -2.65694618e-01 -5.02153695e-01 5.50280035e-01
-2.23423585e-01 1.13144748e-01 -8.48871171e-01 1.51545510e-01
8.00575733e-01 -5.74346818e-02 -4.13080186e-01 5.57394996e-02
-1.80022454e+00 6.29241526e-01 -1.32730007e-01 -8.79188161e-03
-3.84485722e-01 7.35824168e-01 5.86612761e-01 3.73617291e-01
2.92516381e-01 2.92396724e-01 -3.52111757e-01 -6.25822246e-01
-1.41824329e+00 5.08435309e-01 6.10942841e-01 1.40896595e+00
4.95540828e-01 5.47862470e-01 -6.26644731e-01 2.78221279e-01
-1.40210301e-01 7.48596430e-01 6.73437953e-01 -1.06561291e+00
9.58577394e-01 -2.80421734e-01 1.45283312e-01 -7.37375975e-01
2.09447935e-01 4.83615369e-01 -3.81727934e-01 -6.98792413e-02
3.86255741e-01 -3.71576816e-01 -1.06167912e+00 1.77106941e+00
-5.25250547e-02 8.24810982e-01 1.63850307e-01 8.78846288e-01
6.31002188e-01 1.22114694e+00 -5.46431132e-02 -4.67185169e-01
1.13395965e+00 -9.38829482e-01 -9.85284448e-01 7.89066181e-02
4.61581588e-01 -7.91853309e-01 9.94610906e-01 4.85176742e-01
-1.30713248e+00 -6.21421218e-01 -9.09431398e-01 -3.10550302e-01
-2.34842047e-01 3.52992527e-02 2.60726899e-01 2.71695405e-01
-1.14564073e+00 5.14973879e-01 -8.64851356e-01 -2.13535726e-01
2.16479585e-01 1.37084231e-01 -3.80160958e-01 -2.17807263e-01
-1.23515880e+00 4.56866622e-01 4.69016999e-01 -2.64753550e-01
-1.25173759e+00 -7.46112764e-01 -1.16024518e+00 -3.52390110e-02
6.18418694e-01 -6.77773774e-01 1.62260044e+00 -1.52373552e+00
-1.43636751e+00 2.91519135e-01 -4.17872071e-01 -7.52664387e-01
5.74923217e-01 -3.91732752e-01 -2.97967464e-01 6.62655771e-01
1.36748031e-01 1.17213035e+00 1.45568264e+00 -8.65668774e-01
-8.35224330e-01 3.16140383e-01 2.73246199e-01 1.46064162e-01
-2.28533387e-01 7.83730149e-02 -6.13494992e-01 -1.17600107e+00
-6.16003156e-01 -8.68118763e-01 4.93009761e-02 -8.37939531e-02
-2.98105836e-01 3.11082155e-02 1.10420680e+00 -8.79346073e-01
1.16090918e+00 -2.16816878e+00 4.08591121e-01 -3.01039308e-01
-1.47698531e-02 -7.82503486e-02 -2.92985111e-01 5.72364390e-01
-3.21902573e-01 2.50370264e-01 -1.08263642e-02 -4.42503601e-01
-1.15709044e-01 3.15296441e-01 -6.97686493e-01 2.46362835e-01
4.56262231e-01 8.31274927e-01 -1.10202622e+00 -6.59279883e-01
2.36004770e-01 5.15246630e-01 -7.56927431e-01 3.71976763e-01
-7.84735739e-01 2.40294561e-01 -2.16277808e-01 4.33539480e-01
2.45621577e-01 -1.51169032e-01 -2.84650382e-02 -1.85821533e-01
-6.35115057e-02 2.44897768e-01 -1.06359935e+00 1.69267678e+00
-4.20493931e-01 1.20990086e+00 -3.78306985e-01 -8.73886466e-01
2.56201863e-01 9.67494965e-01 6.61978722e-01 -4.92892951e-01
-1.35976911e-01 -1.39978752e-01 -4.45634037e-01 -8.82311165e-01
9.88077760e-01 -1.00129865e-01 -4.10049595e-03 5.22175550e-01
1.35267243e-01 -1.95460901e-01 6.70417905e-01 5.80158830e-01
1.14707637e+00 6.20840013e-01 -6.26911595e-02 5.00551224e-01
6.00245707e-02 5.28074950e-02 4.08502758e-01 6.67745113e-01
2.08904400e-01 9.33852553e-01 6.31947339e-01 -1.34247586e-01
-1.53783727e+00 -9.49247241e-01 2.62424767e-01 9.30123150e-01
-1.49196312e-01 -8.86784077e-01 -7.23631501e-01 -4.93863702e-01
-1.43070772e-01 7.33399093e-01 -5.08093357e-01 6.83868825e-02
-8.87288213e-01 -6.52281567e-02 6.04634047e-01 7.43192792e-01
8.34362656e-02 -1.13345551e+00 -6.03245139e-01 3.14871699e-01
-5.49420714e-01 -1.63027406e+00 -8.70604336e-01 -1.21208794e-01
-5.36710382e-01 -9.18013930e-01 -8.50706100e-01 -8.82049680e-01
6.44067407e-01 4.03296292e-01 1.03247285e+00 -2.37685308e-01
-2.52584249e-01 6.98242188e-01 -7.40572035e-01 -1.35299534e-01
-6.37568057e-01 -3.50301862e-01 4.73626070e-02 4.93535176e-02
-5.23196254e-03 -2.17654943e-01 -3.64125133e-01 -4.54690382e-02
-1.32566917e+00 6.51431501e-01 3.25312138e-01 1.01564622e+00
5.92135191e-01 2.89537907e-02 3.66374433e-01 -7.73126960e-01
2.55066365e-01 -6.32071972e-01 -6.31716669e-01 6.59025833e-02
-1.90665677e-01 1.06652819e-01 1.03270829e+00 -9.44593608e-01
-1.18290997e+00 6.84659779e-02 3.26283872e-01 -9.51176465e-01
1.30639374e-02 5.70707440e-01 2.61972278e-01 7.01828718e-01
4.62083817e-01 2.35853404e-01 -1.78162545e-01 -7.77190477e-02
6.75188363e-01 5.13369143e-01 6.34064674e-01 -6.26441479e-01
8.00129831e-01 4.55931276e-01 -1.46777004e-01 -8.38011444e-01
-4.06870812e-01 -1.51820436e-01 -4.39773917e-01 -2.98856765e-01
1.11555243e+00 -1.15792811e+00 -1.68634057e-01 3.87609065e-01
-1.25985551e+00 -8.09838414e-01 -3.81619364e-01 6.54197872e-01
-9.94391620e-01 4.44015265e-01 -8.40683579e-01 -4.64965910e-01
4.21340615e-02 -1.19055784e+00 1.26883888e+00 -2.00790569e-01
-2.44998038e-01 -7.73568153e-01 -3.90664637e-01 1.81168735e-01
-4.73529920e-02 2.50410736e-01 4.36731070e-01 -9.23408121e-02
-1.07188427e+00 -1.99505389e-02 -1.40427379e-02 3.04587752e-01
2.00271636e-01 4.77582544e-01 -7.76742995e-01 -3.85293782e-01
-2.08358511e-01 -5.19056797e-01 7.40065277e-01 4.46826786e-01
1.32616377e+00 -7.18270063e-01 -1.82075366e-01 5.37644863e-01
1.26771283e+00 2.86564916e-01 6.01257801e-01 8.67118165e-02
8.54313493e-01 2.90836155e-01 9.21632349e-01 7.58669138e-01
3.32204759e-01 6.41705275e-01 2.52884597e-01 1.37106135e-01
-3.77474964e-01 -5.14963150e-01 9.14618611e-01 7.85896242e-01
5.75545207e-02 -6.46034420e-01 -6.09893918e-01 7.32006729e-01
-1.89126241e+00 -1.68401539e+00 9.45531353e-02 2.01108193e+00
9.82339203e-01 -8.08793232e-02 1.52043700e-01 1.59801394e-01
7.35900640e-01 3.15021008e-01 -3.12006205e-01 -2.75599301e-01
-4.84489463e-02 1.98081776e-01 7.17657804e-01 4.35162693e-01
-9.61821616e-01 1.01097524e+00 6.49773884e+00 7.61276007e-01
-1.32385170e+00 1.47142643e-02 3.36717665e-01 -7.82553613e-01
-3.93175006e-01 7.67822564e-02 -7.13803589e-01 8.41441810e-01
1.31918705e+00 -4.45396692e-01 6.95544600e-01 6.00479603e-01
6.68396413e-01 -1.96825638e-02 -1.51518703e+00 1.05398750e+00
5.36207318e-01 -1.72869778e+00 3.93975109e-01 -1.95363149e-01
8.59232187e-01 -1.95130974e-01 -7.05537573e-03 3.64653677e-01
3.34759206e-01 -1.11342418e+00 1.37995708e+00 2.30692759e-01
1.31639838e+00 -3.59051079e-01 2.09695756e-01 1.44718453e-01
-1.30331373e+00 -1.36877060e-01 -2.02523038e-01 -3.68883461e-02
5.62552571e-01 3.23218495e-01 -1.04082239e+00 4.64067221e-01
6.09823763e-01 1.13516212e+00 -2.46743351e-01 6.87229395e-01
-1.79209873e-01 8.22677612e-01 -1.73558787e-01 1.75909817e-01
2.75630981e-01 1.34419069e-01 3.73342901e-01 1.43359184e+00
6.49342000e-01 1.13694422e-01 3.29189628e-01 6.10429525e-01
-1.65559798e-01 -2.21656054e-01 -8.32833171e-01 -3.46270770e-01
4.80386198e-01 9.61684763e-01 -5.02960742e-01 -8.49236131e-01
-7.91522920e-01 1.28134048e+00 1.18437875e-02 6.96893990e-01
-1.26071727e+00 -2.56418318e-01 2.83480495e-01 1.54660940e-01
7.23357379e-01 -4.03132588e-01 2.01795071e-01 -1.37262058e+00
3.06525856e-01 -1.09417343e+00 3.24638963e-01 -1.24674177e+00
-8.22689950e-01 4.02767032e-01 4.72318500e-01 -1.40298128e+00
-8.11176002e-01 -4.01607394e-01 -2.82978356e-01 5.88285089e-01
-1.46640778e+00 -1.20118570e+00 -2.63844728e-01 7.68336236e-01
1.20245755e+00 -9.94038209e-02 4.13928837e-01 3.62528503e-01
-2.20964074e-01 4.16746318e-01 -8.58457312e-02 1.18699484e-01
9.48626995e-01 -1.17422175e+00 5.80001533e-01 1.05848968e+00
2.41506040e-01 2.22983688e-01 8.14409316e-01 -8.42391908e-01
-1.65113366e+00 -1.26713204e+00 7.37936318e-01 -3.57185632e-01
7.26307988e-01 -4.70232248e-01 -6.19220793e-01 1.21309674e+00
5.35857022e-01 -1.36746600e-01 3.86693865e-01 -5.38644791e-01
-3.18438530e-01 1.15748428e-01 -6.88530385e-01 8.99486244e-01
9.50396419e-01 -6.63648725e-01 -5.67804754e-01 4.73642886e-01
1.17821836e+00 -7.48024940e-01 -7.84283876e-01 -2.33464651e-02
5.41407526e-01 -5.88544190e-01 8.14834893e-01 -5.26255012e-01
9.63446021e-01 -3.39011729e-01 -1.53385207e-01 -1.17956197e+00
1.16506433e-02 -1.06387269e+00 -2.79699355e-01 1.36019886e+00
2.72645831e-01 -9.26453155e-03 6.64714396e-01 5.11446774e-01
-1.96792170e-01 -2.92600662e-01 -7.82913744e-01 -9.15011048e-01
-1.48261651e-01 -7.65776813e-01 4.73267347e-01 8.21664393e-01
1.92667559e-01 1.00617208e-01 -9.33960438e-01 -2.66761351e-02
2.84351796e-01 2.61320081e-02 8.23104441e-01 -3.81758451e-01
-4.14077967e-01 -3.08388863e-02 -2.39709094e-01 -1.23327589e+00
3.88572603e-01 -7.39658713e-01 2.84630865e-01 -1.35838902e+00
7.39067569e-02 -1.76958442e-01 2.35899433e-01 4.34956461e-01
-2.47029722e-01 1.45816773e-01 4.40639764e-01 9.08092037e-02
-4.71039206e-01 4.35261846e-01 1.31907928e+00 -1.54384509e-01
7.31461868e-02 -5.31573474e-01 -1.66157514e-01 2.44383946e-01
6.05473757e-01 -4.18208957e-01 -7.69814312e-01 -6.20655596e-01
2.05598176e-01 6.40606105e-01 3.69494736e-01 -8.81014228e-01
5.88981286e-02 -5.50729334e-01 3.09073299e-01 -6.59164429e-01
5.84824920e-01 -6.13014638e-01 3.62350166e-01 5.74499443e-02
-6.15867913e-01 4.54749465e-01 9.47999209e-02 6.48290455e-01
-3.09739232e-01 -1.86442703e-01 3.21006060e-01 -3.17349970e-01
-8.04898441e-01 4.29960519e-01 -9.32959497e-01 7.98591450e-02
1.16224468e+00 -4.82942373e-01 -2.83033133e-01 -7.30972350e-01
-4.09911782e-01 1.27949238e-01 5.91926873e-01 7.22744942e-01
8.02991033e-01 -1.44088864e+00 -6.59152329e-01 3.13519835e-02
-5.31288497e-02 5.28505594e-02 2.21071586e-01 4.70926255e-01
-5.40353954e-01 4.76501912e-01 -9.53560695e-02 -6.03044689e-01
-1.38862801e+00 8.99552286e-01 -2.61247933e-01 -1.73563007e-04
-9.36530888e-01 4.71145511e-01 2.42972568e-01 4.12254214e-01
3.28294128e-01 -6.83936059e-01 1.54438570e-01 6.00083284e-02
7.11039603e-01 1.14825815e-01 -3.45524311e-01 -6.08751357e-01
1.81198031e-01 2.78463095e-01 -5.67028634e-02 -6.16314888e-01
1.09105504e+00 -1.59690380e-01 3.47748429e-01 5.27290344e-01
1.16665995e+00 6.96138740e-02 -1.76134181e+00 3.76487919e-03
-2.97272056e-01 -6.67017758e-01 -1.83974877e-01 -3.63401473e-01
-9.92507517e-01 6.43422961e-01 1.49412215e-01 -5.02766622e-03
1.00297546e+00 -1.39396980e-01 1.12812650e+00 2.39159510e-01
1.43968001e-01 -1.26309204e+00 5.98000288e-01 4.59196985e-01
8.03500235e-01 -1.07460892e+00 -7.47189820e-02 -2.25018695e-01
-8.34447026e-01 1.18585336e+00 5.50755799e-01 4.62221913e-03
1.62943199e-01 5.14539242e-01 -2.25221831e-02 1.61675259e-01
-1.34158015e+00 6.56140000e-02 -5.49775315e-03 7.49412179e-01
4.86448854e-01 -1.18713237e-01 -4.53479588e-02 1.04072891e-01
-3.06352168e-01 4.13054466e-01 1.08271635e+00 1.01186574e+00
-4.57083024e-02 -1.02968633e+00 -5.20662546e-01 4.94798601e-01
-7.18588054e-01 -3.36307317e-01 3.21088992e-02 6.73090577e-01
1.01933321e-02 9.77746964e-01 3.83544564e-01 -1.67486548e-01
-1.60702944e-01 1.17353037e-01 5.66160202e-01 -6.99180961e-01
-3.46592277e-01 1.44830942e-01 2.29475707e-01 -5.00709653e-01
-6.35169923e-01 -9.24324095e-01 -1.40500200e+00 -3.29116315e-01
-1.01746351e-01 -3.94271649e-02 4.06434983e-01 7.20216513e-01
2.34255776e-01 6.77931428e-01 4.79209989e-01 -1.19567621e+00
-1.85866013e-01 -6.72317147e-01 -2.14862019e-01 8.23235869e-01
5.52109897e-01 -4.42286938e-01 -2.89651781e-01 1.11757994e+00] | [10.788999557495117, -0.27343907952308655] |
ffe8f7ab-c07c-4c31-9d21-f50eea9aa2ef | a-melody-unsupervision-model-for-singing | 2110.06546 | null | https://arxiv.org/abs/2110.06546v2 | https://arxiv.org/pdf/2110.06546v2.pdf | A Melody-Unsupervision Model for Singing Voice Synthesis | Recent studies in singing voice synthesis have achieved high-quality results leveraging advances in text-to-speech models based on deep neural networks. One of the main issues in training singing voice synthesis models is that they require melody and lyric labels to be temporally aligned with audio data. The temporal alignment is a time-exhausting manual work in preparing for the training data. To address the issue, we propose a melody-unsupervision model that requires only audio-and-lyrics pairs without temporal alignment in training time but generates singing voice audio given a melody and lyrics input in inference time. The proposed model is composed of a phoneme classifier and a singing voice generator jointly trained in an end-to-end manner. The model can be fine-tuned by adjusting the amount of supervision with temporally aligned melody labels. Through experiments in melody-unsupervision and semi-supervision settings, we compare the audio quality of synthesized singing voice. We also show that the proposed model is capable of being trained with speech audio and text labels but can generate singing voice in inference time. | ['Juhan Nam', 'Soonbeom Choi'] | 2021-10-13 | null | null | null | null | ['singing-voice-synthesis'] | ['speech'] | [ 1.08057044e-01 -2.58846898e-02 -5.10878190e-02 -4.00314927e-01
-1.14207506e+00 -7.84642816e-01 3.25158775e-01 -5.07237673e-01
8.16012674e-04 4.15170670e-01 3.20097089e-01 -2.76984483e-01
1.31037101e-01 -4.57738817e-01 -6.42380059e-01 -5.54803312e-01
1.80817574e-01 5.20597816e-01 1.00678034e-01 -2.19974115e-01
-1.51631296e-01 2.42882445e-01 -1.86967373e+00 4.23642844e-01
5.09804845e-01 9.21511352e-01 3.10650468e-01 1.40556669e+00
9.41436738e-02 7.49086440e-01 -9.20182705e-01 2.25968853e-01
3.87299120e-01 -9.23733830e-01 -5.68385065e-01 1.15948416e-01
7.19878495e-01 -3.45621318e-01 -2.10154921e-01 5.06879032e-01
9.95249212e-01 4.68454421e-01 4.26950514e-01 -1.08119965e+00
-2.07623899e-01 8.77307951e-01 2.02574998e-01 5.12354448e-02
2.38986030e-01 3.17377627e-01 1.14613259e+00 -1.08363831e+00
3.69225502e-01 1.20240903e+00 5.24409354e-01 9.05164242e-01
-1.07629192e+00 -8.97285700e-01 -2.58257389e-01 8.67752731e-02
-1.13953507e+00 -9.96627390e-01 8.68970335e-01 -5.00387609e-01
9.80652511e-01 5.58266819e-01 6.30541146e-01 7.66321063e-01
-1.91150039e-01 5.81777453e-01 6.73332095e-01 -7.90643156e-01
1.35854036e-01 -1.30078256e-01 -4.64095771e-01 3.79751384e-01
-1.02937436e+00 5.31201303e-01 -1.03402793e+00 1.77093312e-01
7.35259295e-01 -5.73002696e-01 -1.14155844e-01 3.29488069e-01
-1.19454312e+00 5.82296073e-01 8.56460258e-02 3.29616964e-01
-1.93885699e-01 2.54386812e-01 5.69795430e-01 5.23225904e-01
3.33719939e-01 5.50048530e-01 -2.55143136e-01 -3.10907006e-01
-1.71755016e+00 2.94086635e-01 7.30316877e-01 8.64190638e-01
6.71968237e-02 8.27014685e-01 -4.31846738e-01 1.10257971e+00
6.97949305e-02 4.25464660e-01 8.18027616e-01 -1.16524601e+00
4.00937825e-01 -2.62558907e-01 4.23043631e-02 -2.51197189e-01
-2.09503219e-01 -4.34855372e-01 -4.55320746e-01 3.38600636e-01
3.88696969e-01 -2.93625295e-01 -7.94492841e-01 1.81722450e+00
3.27752113e-01 5.10618985e-01 -5.05285561e-02 9.94148612e-01
7.31297433e-01 1.06978846e+00 -3.43028069e-01 -4.89474446e-01
1.10004866e+00 -1.59024239e+00 -1.15438867e+00 3.36963795e-02
1.06444217e-01 -1.12228835e+00 1.54516101e+00 5.25709331e-01
-1.56280828e+00 -1.09872580e+00 -1.18812501e+00 -3.11286628e-01
1.58077031e-01 3.35870147e-01 -1.27279475e-01 4.48568642e-01
-9.80509877e-01 9.35929596e-01 -7.13185966e-01 9.57008302e-02
-3.89304489e-01 3.35189760e-01 6.00628816e-02 7.03914464e-01
-1.25966251e+00 4.59880054e-01 2.74761021e-01 8.50220677e-03
-1.43915164e+00 -7.06995368e-01 -6.59680963e-01 8.27731714e-02
2.35672504e-01 -5.08674145e-01 2.12978172e+00 -8.50715220e-01
-2.16780233e+00 4.93369192e-01 -1.42414331e-01 -5.24132133e-01
4.67253983e-01 -8.06792453e-02 -8.36842954e-01 1.54991776e-01
3.69121432e-02 7.16580212e-01 1.32900560e+00 -8.06823730e-01
-8.49994004e-01 5.53347953e-02 -4.80851144e-01 4.19662774e-01
-2.55639046e-01 1.98177442e-01 -1.63313732e-01 -1.01318836e+00
-2.14025266e-02 -9.28497553e-01 1.98008299e-01 -3.98558348e-01
-4.93642747e-01 -2.67031938e-01 9.39718604e-01 -8.38069022e-01
1.52536368e+00 -2.18282080e+00 1.05632439e-01 -2.81412303e-01
-3.71035486e-01 3.52705985e-01 -1.80546284e-01 3.93905371e-01
-1.81866542e-01 -1.87067837e-01 -5.42942546e-02 -6.98628545e-01
1.23473644e-01 -6.19195588e-02 -9.31977034e-01 6.07462078e-02
2.28823945e-01 6.35206640e-01 -9.25464749e-01 -5.63753068e-01
2.41189077e-01 5.17254829e-01 -6.77000582e-01 8.68776977e-01
-5.05239904e-01 7.97209263e-01 3.36352497e-01 4.46158886e-01
-1.26719028e-01 3.84155184e-01 -1.59177948e-02 -3.01166549e-02
-3.51034075e-01 1.22676766e+00 -1.26125860e+00 1.59669936e+00
-8.04450274e-01 6.47719622e-01 2.96588004e-01 -5.04565537e-01
1.12299454e+00 1.14446867e+00 1.98881403e-01 -2.97918618e-01
1.46051459e-02 4.48450118e-01 1.19934820e-01 -4.77879763e-01
6.42872214e-01 -7.57401168e-01 4.33694944e-02 7.20595598e-01
3.14603269e-01 -7.35143840e-01 5.70403449e-02 -4.49236542e-01
7.43430078e-01 2.70705193e-01 -1.79979708e-02 1.64673895e-01
4.13691908e-01 -2.29050636e-01 5.47135472e-01 3.39015156e-01
-1.59836803e-02 9.86410618e-01 -3.80040780e-02 -1.98841952e-02
-1.31128633e+00 -1.02729821e+00 2.69853491e-02 1.54821908e+00
-6.46747231e-01 -5.21793664e-01 -9.57553864e-01 -2.00396225e-01
-4.63939071e-01 9.36248243e-01 -8.67628530e-02 6.73006028e-02
-1.02893746e+00 2.45546818e-01 9.97127891e-01 5.44655919e-01
-1.23148248e-01 -1.62940061e+00 -3.67158890e-01 6.74899757e-01
-2.13991806e-01 -9.62038040e-01 -1.22517180e+00 3.64550918e-01
-8.04708958e-01 -4.32829082e-01 -8.15367281e-01 -1.01713276e+00
4.38228510e-02 -3.49455290e-02 1.04218066e+00 -9.27367136e-02
-8.54699500e-03 -5.27363792e-02 -9.60361287e-02 -5.07256269e-01
-9.42480147e-01 1.50612324e-01 5.73957980e-01 5.04745990e-02
-1.53814942e-01 -8.38583648e-01 -3.72151881e-01 5.47642946e-01
-8.28026891e-01 1.96042106e-01 -9.28057060e-02 9.56613421e-01
7.85543084e-01 2.18191683e-01 1.17705286e+00 -3.99995476e-01
6.65750802e-01 -1.46444067e-01 -6.55467808e-01 -1.15471847e-01
-3.78101081e-01 -1.23457111e-01 1.22790527e+00 -7.77024925e-01
-7.40648270e-01 2.08164945e-01 -4.57508445e-01 -8.41541231e-01
-1.36431709e-01 2.31674984e-01 -2.49376774e-01 6.26496613e-01
7.66624868e-01 7.85038248e-02 -5.59973940e-02 -6.65571511e-01
4.50034201e-01 1.10562074e+00 1.01011515e+00 -5.74586272e-01
7.19636321e-01 4.62405160e-02 -3.87330472e-01 -9.31307793e-01
-8.59394789e-01 -4.49869365e-01 -6.62528932e-01 -3.34315628e-01
6.50739610e-01 -8.64252448e-01 -5.36822736e-01 4.47899193e-01
-1.08641493e+00 -5.88384032e-01 -7.12474108e-01 6.27504051e-01
-1.06869948e+00 7.08423480e-02 -6.84633851e-01 -1.03491652e+00
-4.94064182e-01 -1.05857623e+00 9.68477964e-01 7.44078308e-02
-5.10681689e-01 -6.07503235e-01 1.69422358e-01 4.63633060e-01
4.40881222e-01 -1.45818487e-01 6.66891217e-01 -4.80312169e-01
-3.37612629e-01 -3.82883511e-02 5.14521658e-01 6.64780259e-01
3.58508378e-01 2.00360641e-01 -1.36334705e+00 -1.34287789e-01
1.14364475e-01 -5.89834511e-01 4.76838529e-01 4.71269816e-01
9.65855181e-01 -6.18068159e-01 4.29519057e-01 4.84409213e-01
4.91482139e-01 4.32233810e-01 2.40150347e-01 -3.01095128e-01
6.17101789e-01 8.22128057e-01 8.86763394e-01 3.44203472e-01
-1.67987682e-03 1.06901109e+00 1.20388940e-01 2.57308315e-03
-8.22680473e-01 -8.40528488e-01 7.24279821e-01 1.60159528e+00
2.19273061e-01 -1.36041194e-01 -5.27709842e-01 9.30647612e-01
-1.57078135e+00 -1.12823427e+00 1.69283032e-01 2.26997471e+00
1.44752610e+00 7.87551627e-02 6.11112654e-01 7.27275133e-01
8.60744178e-01 2.82945782e-01 -5.99009216e-01 -7.23445117e-01
1.83447972e-01 5.94880462e-01 -1.93187207e-01 8.28048825e-01
-8.32249045e-01 1.16422677e+00 6.31562281e+00 9.46964383e-01
-1.54361510e+00 1.48438767e-01 1.34934142e-01 -6.02278173e-01
-3.10676724e-01 -6.04344569e-02 -8.92646611e-01 3.69573951e-01
1.41048491e+00 -1.06869072e-01 1.02085173e+00 7.61587024e-01
7.98873305e-01 4.69885826e-01 -1.55856800e+00 8.35814834e-01
-4.09837514e-02 -1.29318249e+00 -8.49652290e-02 -4.45156693e-01
6.87490225e-01 -3.37694198e-01 2.34602869e-01 3.51162374e-01
-1.08851649e-01 -1.29601920e+00 1.38216317e+00 2.09612668e-01
1.29200053e+00 -6.73775256e-01 6.96815997e-02 4.38147306e-01
-1.34188986e+00 2.10415162e-02 1.46307766e-01 -2.21914008e-01
5.66542089e-01 2.23873839e-01 -1.62998021e+00 1.07457384e-01
4.22922403e-01 3.10457945e-01 1.10698715e-01 8.61404717e-01
-4.96697426e-01 1.23850727e+00 -2.08962187e-01 2.66823471e-01
4.50523645e-02 1.54523790e-01 6.83054805e-01 1.16910243e+00
6.61938906e-01 -1.75361738e-01 1.41645715e-01 8.84774387e-01
2.01842673e-02 1.54895574e-01 -3.86250407e-01 -4.40710753e-01
8.30145836e-01 9.71016467e-01 -1.61192968e-01 -2.64737904e-01
1.48912117e-01 6.99056208e-01 -9.15327445e-02 1.90168142e-01
-6.99397266e-01 -3.99042517e-01 4.28734273e-01 3.78875732e-01
2.94444263e-01 -3.54363889e-01 -1.79973900e-01 -6.36736691e-01
-1.01719134e-01 -1.05158961e+00 1.67458370e-01 -1.03307748e+00
-9.98807728e-01 8.26210916e-01 -3.63739878e-01 -1.43353188e+00
-1.06621647e+00 -2.39704773e-01 -8.74400318e-01 1.02290535e+00
-1.30800545e+00 -9.22459424e-01 8.48458931e-02 5.10602772e-01
1.13250363e+00 -3.68672192e-01 9.80130613e-01 4.02445376e-01
-3.22065890e-01 6.11349761e-01 -2.90542245e-01 -3.21477912e-02
9.84208107e-01 -1.38706946e+00 5.84015131e-01 6.83664083e-01
6.04926467e-01 2.40637615e-01 9.53329742e-01 -3.26351583e-01
-8.30029011e-01 -1.11946189e+00 1.32113659e+00 -2.86955714e-01
5.01311719e-01 -5.15085697e-01 -8.46574366e-01 4.68845338e-01
3.96561742e-01 -5.80861345e-02 7.81544089e-01 -1.61140218e-01
-1.70956373e-01 -2.68706322e-01 -7.04230309e-01 6.73443854e-01
7.21435130e-01 -1.00915158e+00 -8.38132203e-01 3.13914448e-01
1.09802377e+00 -7.08895445e-01 -7.15992391e-01 4.04141933e-01
5.09910047e-01 -8.29574704e-01 6.57184958e-01 -5.42747438e-01
3.58407408e-01 -5.98162055e-01 -1.26506343e-01 -1.38507974e+00
6.47900775e-02 -1.29061878e+00 -1.17058799e-01 1.35860491e+00
5.54238558e-01 1.20959342e-01 4.36293244e-01 4.04838845e-02
-5.23945510e-01 -3.17241818e-01 -9.88759696e-01 -1.06905890e+00
-4.44794558e-02 -5.79832435e-01 6.98567808e-01 7.26920009e-01
-1.59220793e-03 7.08627582e-01 -7.30850875e-01 7.97309652e-02
1.71077117e-01 3.06858182e-01 7.72749066e-01 -8.47534597e-01
-7.53694177e-01 -2.01506257e-01 4.72641796e-01 -1.01154888e+00
1.17333062e-01 -8.93094540e-01 5.57833016e-01 -1.06640506e+00
-5.80876231e-01 -4.19963717e-01 -3.04132491e-01 6.35147810e-01
1.01064578e-01 2.79830664e-01 4.16296273e-01 1.40819728e-01
-1.57743573e-01 6.42338157e-01 1.28205287e+00 -1.07446633e-01
-6.65313959e-01 4.77779686e-01 6.14766404e-02 6.56204402e-01
8.01495016e-01 -5.06733954e-01 -4.47482467e-01 -1.02198072e-01
-1.47867844e-01 7.10573673e-01 2.12487772e-01 -1.15737510e+00
3.47353846e-01 -6.74094772e-03 -9.69360918e-02 -9.60639834e-01
7.79272914e-01 -5.80939412e-01 2.67901689e-01 1.22176155e-01
-7.34074056e-01 -1.05991133e-01 2.27868840e-01 5.03497086e-02
-5.28848946e-01 -4.54126716e-01 8.86770129e-01 7.03271627e-02
-7.97828212e-02 8.39429870e-02 -4.10135299e-01 1.14072941e-01
4.70036179e-01 1.39775907e-03 1.16355635e-01 -7.34863818e-01
-1.01520562e+00 -2.02558056e-01 1.07757319e-02 6.98379993e-01
4.24810350e-01 -1.64224911e+00 -7.78471828e-01 5.19649327e-01
-1.86106905e-01 1.60587933e-02 1.85458288e-01 4.10528243e-01
-2.09864572e-01 5.92547178e-01 4.90820967e-03 -6.34315252e-01
-1.38079095e+00 3.94426078e-01 4.01861489e-01 -4.43158448e-02
-4.23396170e-01 7.68706858e-01 -5.48588783e-02 -7.00057268e-01
8.10139418e-01 -3.63569826e-01 2.39269994e-02 6.48997426e-02
4.38312292e-01 4.19123173e-01 2.15225652e-01 -5.51949322e-01
7.05691501e-02 3.85380566e-01 3.27284068e-01 -7.74939477e-01
1.07069314e+00 -1.24211274e-02 2.41005704e-01 9.91558135e-01
8.49795520e-01 5.97196579e-01 -1.45485222e+00 -5.81122711e-02
-1.49542972e-01 -1.51243314e-01 1.26031205e-01 -8.73819947e-01
-7.70033360e-01 1.23672616e+00 2.56433278e-01 3.46859574e-01
1.13028955e+00 -2.77541667e-01 1.29813027e+00 8.13385472e-02
2.88118329e-02 -1.39742255e+00 3.01059186e-01 7.06168532e-01
1.19152117e+00 -7.72983730e-01 -6.46262407e-01 -1.19768545e-01
-6.83302522e-01 1.06813550e+00 5.52471519e-01 -8.51094648e-02
5.82660735e-01 4.84489858e-01 5.00172734e-01 2.91258246e-01
-9.78846967e-01 -5.67964725e-02 5.54954350e-01 3.35887492e-01
5.83986282e-01 1.82009161e-01 -6.59774318e-02 7.01099515e-01
-1.09786427e+00 -5.04452810e-02 2.73964345e-01 2.75529236e-01
-4.75202531e-01 -1.38370264e+00 -5.89334786e-01 -1.15802750e-01
-5.42787790e-01 -3.45408529e-01 -4.22978193e-01 1.27706602e-01
1.61865175e-01 1.33598018e+00 1.10843070e-01 -5.10777891e-01
3.05523843e-01 5.13798416e-01 3.29288274e-01 -9.35418427e-01
-1.07623553e+00 7.72606850e-01 2.28024513e-01 -2.17742726e-01
-2.73167738e-03 -3.48384261e-01 -1.55088258e+00 1.94267228e-01
-4.04265165e-01 3.88269573e-01 7.60575294e-01 7.61930525e-01
2.01536879e-01 9.74105895e-01 1.13056004e+00 -9.76415932e-01
-8.06317031e-01 -1.13632071e+00 -7.00007081e-01 9.42890644e-02
7.47062147e-01 -1.57869488e-01 -5.43726087e-01 3.91577870e-01] | [15.45610237121582, 6.193839073181152] |
d480aa9c-1bf9-47a0-9437-bf27919ab486 | minimal-learning-machine-for-multi-label | 2305.05518 | null | https://arxiv.org/abs/2305.05518v1 | https://arxiv.org/pdf/2305.05518v1.pdf | Minimal Learning Machine for Multi-Label Learning | Distance-based supervised method, the minimal learning machine, constructs a predictive model from data by learning a mapping between input and output distance matrices. In this paper, we propose methods and evaluate how this technique and its core component, the distance mapping, can be adapted to multi-label learning. The proposed approach is based on combining the distance mapping with an inverse distance weighting. Although the proposal is one of the simplest methods in the multi-label learning literature, it achieves state-of-the-art performance for small to moderate-sized multi-label learning problems. Besides its simplicity, the proposed method is fully deterministic and its hyper-parameter can be selected via ranking loss-based statistic which has a closed form, thus avoiding conventional cross-validation-based hyper-parameter tuning. In addition, due to its simple linear distance mapping-based construction, we demonstrate that the proposed method can assess predictions' uncertainty for multi-label classification, which is a valuable capability for data-centric machine learning pipelines. | ['Tommi Kärkkäinen', 'João P. P. Gomes', 'César L. C. Mattos', 'Amauri Souza', 'Joonas Hämäläinen'] | 2023-05-09 | null | null | null | null | ['multi-label-learning'] | ['methodology'] | [ 3.24102879e-01 2.44650200e-01 -3.86272162e-01 -8.42689872e-01
-1.31342745e+00 -6.53257549e-01 5.71013212e-01 7.53312230e-01
-4.12764519e-01 6.33901417e-01 -4.38355833e-01 -3.24118972e-01
-7.42873669e-01 -6.03341639e-01 -3.59386981e-01 -8.14029455e-01
4.42439988e-02 8.36022794e-01 2.40228072e-01 1.62302062e-01
3.67125094e-01 4.30112869e-01 -1.75205123e+00 2.89277166e-01
7.18983650e-01 1.15160203e+00 -1.84455793e-02 5.03295422e-01
1.00268878e-01 6.77148879e-01 -1.45552546e-01 -5.07797837e-01
2.40275506e-02 -2.53298283e-01 -9.48692918e-01 -4.64973629e-01
4.21440840e-01 1.35460630e-01 4.83382702e-01 9.92653131e-01
6.19175851e-01 -6.39019459e-02 1.16245782e+00 -1.17765284e+00
-3.27524602e-01 5.37452042e-01 -4.82033908e-01 -3.03484201e-01
2.19834834e-01 -3.22915405e-01 1.49326873e+00 -1.03735876e+00
2.89884925e-01 1.15779066e+00 1.04165542e+00 3.30888480e-01
-1.51865923e+00 -4.54657078e-01 -2.98362434e-01 7.59490207e-02
-1.36120641e+00 -7.20996261e-02 5.39673746e-01 -7.66778111e-01
6.72875226e-01 3.31869692e-01 1.02454990e-01 6.03774071e-01
2.95267761e-01 7.33184457e-01 1.25002968e+00 -8.41241896e-01
2.05133632e-01 6.10627055e-01 1.95668578e-01 8.23624730e-01
1.14547558e-01 1.27635568e-01 -1.67324930e-01 -4.43451464e-01
8.65840539e-02 -1.65071681e-01 1.81254014e-01 -8.31656039e-01
-1.00058889e+00 1.08181524e+00 1.19087294e-01 3.76065046e-01
5.36304452e-02 -5.62140048e-02 5.05961239e-01 2.08251610e-01
7.80419588e-01 3.11349690e-01 -5.09753227e-01 9.75618958e-02
-1.10416770e+00 1.02125190e-01 6.75768137e-01 6.25713706e-01
7.71181762e-01 -5.78476787e-01 -3.03132981e-01 9.99750614e-01
6.49456561e-01 2.04924062e-01 5.05645573e-01 -5.51091969e-01
2.26380542e-01 6.82655931e-01 -7.03179389e-02 -7.41091847e-01
-1.12409687e+00 -5.35511374e-01 -6.71989620e-01 3.33221972e-01
4.79641795e-01 5.10588586e-02 -2.05450878e-01 1.67851663e+00
5.09078562e-01 6.31185323e-02 2.13834327e-02 4.23469663e-01
6.14605606e-01 2.07710072e-01 2.60335922e-01 -4.80824560e-01
1.35188580e+00 -7.74264634e-01 -3.95762712e-01 9.47498605e-02
1.23235095e+00 -6.48825526e-01 1.00561595e+00 4.16552156e-01
-7.46753156e-01 -4.55586910e-01 -1.08392680e+00 -2.87108235e-02
-4.49653924e-01 2.54120559e-01 5.52638173e-01 8.65223765e-01
-8.78215969e-01 8.70959759e-01 -4.22589719e-01 -3.68311197e-01
1.96887478e-01 4.60540324e-01 -3.66748095e-01 -6.35396019e-02
-1.32607138e+00 1.18131673e+00 7.78170466e-01 -3.10286939e-01
-5.23831964e-01 -7.16923237e-01 -6.46240830e-01 1.49446093e-02
5.59101887e-02 -4.96656299e-01 1.22286272e+00 -4.44776446e-01
-1.39144719e+00 1.09660673e+00 2.00399235e-01 -2.65494436e-01
5.30665100e-01 1.11478880e-01 -4.83895391e-01 8.33466556e-03
3.00442837e-02 5.37933588e-01 6.66046739e-01 -1.11183679e+00
-8.48710775e-01 -4.70626205e-01 -2.48545781e-01 4.72023105e-03
-3.38586926e-01 2.09450290e-01 6.54869899e-02 -4.42343801e-01
6.38633072e-02 -1.05544639e+00 -1.78026795e-01 -1.79730475e-01
-2.96314806e-01 -5.34600794e-01 2.93565631e-01 -8.93561691e-02
1.50544786e+00 -2.09842992e+00 3.71563584e-02 5.19778371e-01
1.36961654e-01 2.67432034e-01 -1.23833917e-01 6.18691981e-01
-1.45129502e-01 6.35683462e-02 -3.35881472e-01 -4.87253219e-01
2.15816379e-01 -2.35247672e-01 1.61932439e-01 6.56872571e-01
2.99567699e-01 7.99663305e-01 -1.06524622e+00 -6.51623130e-01
2.22703055e-01 3.70015681e-01 -2.20554128e-01 7.52906799e-02
-2.33160585e-01 2.61668146e-01 -1.44204974e-01 4.55581546e-01
5.93391597e-01 -4.25852686e-01 4.57100302e-01 -2.64449984e-01
-9.80332866e-02 -5.52950092e-02 -1.34643185e+00 1.69887030e+00
-5.27688742e-01 2.77300924e-01 -3.92580479e-01 -1.11453080e+00
1.16096318e+00 1.91637114e-01 6.69645667e-01 -2.56399482e-01
9.15196985e-02 3.84480715e-01 -2.02437714e-01 -3.21012884e-01
2.00665191e-01 -3.39148402e-01 -1.29884571e-01 7.84540236e-01
2.61269093e-01 2.11748198e-01 1.38623193e-01 -7.66652450e-02
9.65860248e-01 2.17274606e-01 7.68051326e-01 -4.83055145e-01
8.54733944e-01 -3.81632239e-01 4.87554729e-01 5.51386833e-01
-9.39110294e-02 4.39447939e-01 6.15815639e-01 -2.77459174e-01
-8.69949639e-01 -8.53094459e-01 -6.97158039e-01 1.41859078e+00
-1.54234052e-01 -5.57687163e-01 -5.80732465e-01 -1.01563680e+00
3.17047745e-01 7.33352482e-01 -6.81689560e-01 -1.51928246e-01
-1.69434741e-01 -1.07873440e+00 7.89825737e-01 1.69241622e-01
-1.58463836e-01 -6.55039966e-01 -2.72638887e-01 2.16909558e-01
5.52793639e-03 -4.81487840e-01 -2.15676457e-01 6.01046860e-01
-8.99937034e-01 -1.29901314e+00 -5.36820114e-01 -7.21674144e-01
4.62179601e-01 -6.57693446e-02 1.08471012e+00 -2.94189185e-01
-1.98084697e-01 2.24334344e-01 -1.94008693e-01 -2.60522276e-01
-6.58871889e-01 2.18154639e-01 7.48691857e-02 2.44134098e-01
6.77082837e-01 -3.92198473e-01 -2.95100838e-01 4.69639957e-01
-8.61153722e-01 -2.29716077e-01 7.97603548e-01 1.07605171e+00
9.46156800e-01 -5.86710498e-03 1.03195465e+00 -1.38464296e+00
6.94468975e-01 -6.82865500e-01 -5.73144972e-01 7.88285732e-01
-1.45907438e+00 4.91344303e-01 5.32622755e-01 -2.71028340e-01
-8.07439029e-01 3.22621673e-01 -9.75825638e-02 1.16540812e-01
-1.55100808e-01 4.19107407e-01 -1.75223589e-01 -1.97661623e-01
8.51992428e-01 -1.69388428e-01 1.37921974e-01 -7.06729531e-01
6.88906670e-01 8.07832479e-01 9.63930413e-02 -3.76315147e-01
4.81824160e-01 2.27593154e-01 5.81554413e-01 -1.39314070e-01
-1.01550901e+00 -7.55341768e-01 -1.12798834e+00 -2.30993912e-01
7.53222764e-01 -7.07056403e-01 -9.91891682e-01 1.34891659e-01
-9.43480253e-01 5.97795919e-02 -6.28868565e-02 6.40338957e-01
-6.93766892e-01 4.18478578e-01 -5.82050622e-01 -8.41633797e-01
-3.54336232e-01 -9.51760113e-01 1.06905222e+00 -1.52316794e-01
-2.20208332e-01 -1.43849635e+00 4.38589960e-01 2.88367748e-01
1.99449733e-01 1.36234850e-01 1.23400295e+00 -1.25804913e+00
-1.48350168e-02 -3.64624888e-01 -3.35155040e-01 4.46004450e-01
-3.76356207e-02 -1.32980108e-01 -1.14282799e+00 -4.99540001e-01
-2.32279316e-01 -6.10562623e-01 7.89497614e-01 3.19339931e-01
8.88862312e-01 1.24762505e-01 -3.20109427e-01 2.79220462e-01
1.69079876e+00 -1.31114900e-01 1.88823134e-01 4.23652500e-01
4.34374273e-01 8.68123472e-01 1.05746174e+00 4.74713653e-01
5.56830406e-01 8.37317646e-01 3.82604718e-01 -5.20932823e-02
1.34293497e-01 -8.44079629e-02 1.59408987e-01 8.00236225e-01
2.83201218e-01 1.41674280e-01 -9.17078376e-01 -6.86128587e-02
-2.08173203e+00 -7.83458650e-01 -3.20091158e-01 2.49010563e+00
8.27585578e-01 1.70549765e-01 3.18697393e-01 2.49237105e-01
6.35572314e-01 -1.52734756e-01 -5.38773715e-01 -4.92013693e-01
-1.52016044e-01 4.15843800e-02 4.97196496e-01 5.17837584e-01
-1.37121856e+00 6.15229845e-01 6.78823566e+00 1.11213315e+00
-7.07197070e-01 3.00675154e-01 4.45981383e-01 9.72468257e-02
-3.73116255e-01 -5.20967767e-02 -1.09878838e+00 3.12427104e-01
1.31134129e+00 8.72563720e-02 -6.78333221e-03 9.02597249e-01
7.07597956e-02 -1.43487364e-01 -1.43038607e+00 9.26705062e-01
9.78689566e-02 -1.00002635e+00 -2.82420248e-01 -2.65353750e-02
4.91406918e-01 -1.33909181e-01 1.81895837e-01 2.32489169e-01
1.52717054e-01 -9.55525994e-01 4.98520672e-01 5.74604571e-01
1.08227372e+00 -9.73161578e-01 8.75154257e-01 4.32107359e-01
-9.91572976e-01 -3.42854530e-01 -3.22743088e-01 1.84101477e-01
-4.22038138e-02 1.19914579e+00 -7.98762619e-01 7.16500103e-01
1.89510912e-01 6.19842947e-01 -9.24698055e-01 8.44587982e-01
7.86498934e-02 4.11396056e-01 -6.08434007e-02 -6.82311654e-02
5.46769984e-02 -2.70530581e-01 -2.98573319e-02 1.49631906e+00
2.60282516e-01 -4.27872121e-01 1.22003272e-01 5.01511693e-01
1.87573522e-01 6.54033720e-01 -5.75126112e-01 3.47151130e-01
6.24248564e-01 1.43395424e+00 -6.44684613e-01 -7.09656700e-02
-6.47827923e-01 5.88048279e-01 7.40050793e-01 -2.54452020e-01
-6.91100895e-01 -4.67637777e-01 2.96429127e-01 -3.03225696e-01
1.39732793e-01 2.72124797e-01 -3.46165597e-01 -6.84039533e-01
-7.57101774e-02 -7.17925429e-01 6.65105343e-01 -2.16214716e-01
-1.54003215e+00 4.37639415e-01 8.00130963e-02 -1.54092300e+00
-6.37213171e-01 -7.50942290e-01 -1.04318276e-01 7.64118373e-01
-1.51070690e+00 -1.37381029e+00 1.33268639e-01 2.45843127e-01
6.33564638e-03 -3.56155694e-01 1.28026330e+00 5.41060567e-01
-5.29428065e-01 9.72039640e-01 7.80110955e-01 -3.33656788e-01
1.15147889e+00 -1.47890580e+00 -2.12372869e-01 2.80987948e-01
1.37410775e-01 3.20081830e-01 5.02931178e-01 -3.02412748e-01
-7.52117217e-01 -1.11725068e+00 1.13337767e+00 -5.49659491e-01
6.68214262e-01 -2.41964385e-01 -7.97159910e-01 3.95933867e-01
-4.55300063e-01 5.76129444e-02 1.56854212e+00 3.86541337e-01
-6.46493852e-01 -3.06901455e-01 -1.40460455e+00 -6.23159967e-02
6.71182752e-01 -5.84681213e-01 -3.40216249e-01 5.66064596e-01
3.43219519e-01 -3.01007275e-02 -1.49003029e+00 4.78615105e-01
8.23747814e-01 -1.13500309e+00 9.06064034e-01 -6.93872094e-01
2.61682987e-01 -1.90358996e-01 -2.87094027e-01 -1.38431215e+00
-6.50626063e-01 -8.67491141e-02 -1.14084475e-01 1.50663483e+00
7.91487992e-01 -6.55799866e-01 4.51864064e-01 3.53227049e-01
1.28043920e-01 -1.01113760e+00 -9.34040308e-01 -7.92998075e-01
4.51693594e-01 -4.02355820e-01 5.78083932e-01 9.54105437e-01
2.32306272e-01 4.97516185e-01 -5.25586128e-01 4.71938550e-02
8.76174927e-01 1.17459819e-01 2.41559848e-01 -1.81477141e+00
-4.62018400e-01 -4.51114655e-01 -4.19815749e-01 -6.12286091e-01
3.71147722e-01 -1.41203833e+00 -1.00815065e-01 -1.26100159e+00
2.93268085e-01 -9.17435586e-01 -7.66140342e-01 3.64115775e-01
-1.14266284e-01 2.31141403e-01 7.20500126e-02 3.81971180e-01
-8.34366500e-01 1.64247230e-01 7.59846270e-01 1.58809591e-02
-5.68203554e-02 3.01930010e-01 -6.12421393e-01 6.50402427e-01
7.90270686e-01 -8.52924287e-01 -4.96128798e-01 9.42180753e-02
3.55784684e-01 -1.06173612e-01 2.27511544e-02 -7.34161675e-01
2.03817248e-01 -3.50396037e-02 9.96857435e-02 -3.99226695e-01
1.41212083e-02 -8.46957564e-01 3.45042855e-01 4.44397956e-01
-9.51502323e-01 -1.57279357e-01 -1.88685954e-01 6.93270862e-01
-1.50761247e-01 -7.12644100e-01 1.01989841e+00 6.49413243e-02
-4.77341413e-01 1.57581672e-01 -1.13141827e-01 6.32425100e-02
1.22569799e+00 8.71167183e-02 -1.38425037e-01 1.83631465e-01
-6.36711955e-01 7.76599497e-02 2.32420385e-01 2.15795159e-01
2.78936207e-01 -1.57268417e+00 -7.81642735e-01 3.01458277e-02
6.43483937e-01 -4.15002823e-01 7.39343613e-02 9.70581949e-01
-2.02006117e-01 6.98148310e-01 6.55361637e-02 -6.60420954e-01
-1.39462936e+00 7.22808838e-01 1.66274011e-01 -7.57335007e-01
-1.80735901e-01 5.65813243e-01 -1.54377371e-01 -8.10737133e-01
2.07912832e-01 2.04236549e-03 -4.14447248e-01 4.54084933e-01
5.45495927e-01 6.22196436e-01 3.57576370e-01 -6.12653315e-01
-4.55688089e-01 7.62997448e-01 7.74360867e-03 -3.78927612e-03
1.05069065e+00 -2.21413106e-01 -2.38606378e-01 9.42791581e-01
1.51502514e+00 -2.16345772e-01 -8.11436951e-01 -4.88153040e-01
6.49894714e-01 -2.69008666e-01 1.77358743e-02 -7.97938466e-01
-4.71677810e-01 1.00050461e+00 8.46462488e-01 1.92129180e-01
9.54573214e-01 6.15675226e-02 2.01042131e-01 3.84845883e-01
2.97747314e-01 -1.27899683e+00 9.56049131e-04 2.01321065e-01
4.58789825e-01 -1.64210927e+00 1.52008370e-01 -3.75584781e-01
-4.59463596e-01 1.25627232e+00 2.88379073e-01 4.88533914e-01
9.89567339e-01 5.29023930e-02 5.02404198e-02 -9.71757621e-03
-8.07434261e-01 -1.54447272e-01 5.71568906e-01 6.30922675e-01
7.55312741e-01 2.47402236e-01 -6.87296987e-01 6.09260976e-01
1.51633531e-01 -1.81706637e-01 7.82156959e-02 4.90388304e-01
-6.28641188e-01 -1.62430215e+00 -8.81926790e-02 6.63435102e-01
-4.98717964e-01 2.40548961e-02 -3.28121111e-02 6.32295370e-01
2.58932561e-01 1.04208446e+00 -2.32450441e-01 -5.52086055e-01
2.32651561e-01 4.76466388e-01 4.37685579e-01 -6.06811285e-01
-5.13115883e-01 -2.50567049e-01 1.06233425e-01 -3.59954596e-01
-5.88480890e-01 -6.60464883e-01 -1.11323857e+00 -1.52664334e-01
-6.97062790e-01 1.86166272e-01 8.74490380e-01 1.00535679e+00
3.50681216e-01 1.45869046e-01 9.74307537e-01 -4.59846884e-01
-1.07666588e+00 -1.06663787e+00 -9.74682331e-01 4.63583261e-01
4.01926786e-02 -8.70962858e-01 -3.45641047e-01 -3.45078409e-01] | [9.312506675720215, 4.325052261352539] |
03d5d8f0-7c67-42c3-8eae-0dc10c68c7ce | robust-machine-learning-segmentation-for | 2209.02032 | null | https://arxiv.org/abs/2209.02032v2 | https://arxiv.org/pdf/2209.02032v2.pdf | Robust machine learning segmentation for large-scale analysis of heterogeneous clinical brain MRI datasets | Every year, millions of brain MRI scans are acquired in hospitals, which is a figure considerably larger than the size of any research dataset. Therefore, the ability to analyse such scans could transform neuroimaging research. Yet, their potential remains untapped, since no automated algorithm is robust enough to cope with the high variability in clinical acquisitions (MR contrasts, resolutions, orientations, artefacts, subject populations). Here we present SynthSeg+, an AI segmentation suite that enables, for the first time, robust analysis of heterogeneous clinical datasets. In addition to whole-brain segmentation, SynthSeg+ also performs cortical parcellation, intracranial volume estimation, and automated detection of faulty segmentations (mainly caused by scans of very low quality). We demonstrate SynthSeg+ in seven experiments, including an ageing study on 14,000 scans, where it accurately replicates atrophy patterns observed on data of much higher quality. SynthSeg+ is publicly released as a ready-to-use tool to unlock the potential of quantitative morphometry. | ['Juan. E. Iglesias', 'Sudeshna Das', 'Steven E. Arnold', 'You Cheng', 'Colin Magdamo', 'Benjamin Billot'] | 2022-09-05 | null | null | null | null | ['brain-segmentation'] | ['medical'] | [ 2.98641592e-01 1.91598490e-01 3.98189604e-01 -5.95729291e-01
-8.54113817e-01 -3.20187569e-01 3.59968364e-01 3.12072545e-01
-7.40467250e-01 7.10891247e-01 1.34049922e-01 -2.06845105e-01
-1.72463134e-01 -4.79867637e-01 -5.17939806e-01 -5.06652772e-01
-4.76044178e-01 1.06777442e+00 6.30920231e-01 1.09674126e-01
2.69043922e-01 5.50699890e-01 -1.26566768e+00 1.18095890e-01
7.47112989e-01 7.00889409e-01 3.60279173e-01 3.96307707e-01
2.41957232e-01 2.43634224e-01 -2.53232062e-01 -3.36763382e-01
2.13586777e-01 -3.69156927e-01 -8.64655018e-01 4.24427427e-02
6.10662162e-01 -3.29907417e-01 1.59103470e-03 9.87025499e-01
9.09214556e-01 -3.04676175e-01 5.63142359e-01 -8.63434911e-01
-5.42825982e-02 6.85793281e-01 -7.17087924e-01 9.39928651e-01
-4.81866859e-02 4.14860308e-01 4.19394463e-01 -5.93221188e-01
1.09888923e+00 8.29761446e-01 6.78490102e-01 2.08295628e-01
-1.33263159e+00 -6.25495732e-01 -4.36815321e-01 -3.49391885e-02
-1.14839590e+00 -6.12167537e-01 1.72547400e-01 -8.87843668e-01
9.37413275e-01 1.63746729e-01 9.84612525e-01 8.54945600e-01
4.83104974e-01 3.70685518e-01 1.45649052e+00 -6.11795746e-02
3.19216877e-01 -3.94991606e-01 3.01903952e-02 4.78438735e-01
4.35412318e-01 -2.67955989e-01 -3.06758940e-01 -2.57006913e-01
8.72435987e-01 -4.21513200e-01 -3.09003443e-01 -5.34805775e-01
-1.91245580e+00 5.08867145e-01 2.78713018e-01 5.62820792e-01
-5.56610465e-01 -1.32670686e-01 6.28284991e-01 1.92994043e-01
4.92401510e-01 4.56207514e-01 -3.02600980e-01 -1.46272957e-01
-1.52214897e+00 3.05185258e-01 2.04727262e-01 6.22339964e-01
1.53949171e-01 -2.41557620e-02 3.80088761e-02 9.31187809e-01
1.48359343e-01 3.94909650e-01 1.00020659e+00 -9.64117169e-01
3.50293547e-01 3.81824315e-01 -5.29532552e-01 -7.63726175e-01
-9.69077706e-01 -3.50578040e-01 -9.15275037e-01 1.78901911e-01
6.81190670e-01 -4.99313399e-02 -9.32952285e-01 1.38198054e+00
3.09177518e-01 -1.54935509e-01 -7.42042720e-01 1.08532119e+00
5.16102850e-01 -3.75983238e-01 1.37119591e-01 -2.31356472e-01
1.42839348e+00 -3.88486594e-01 -3.45948249e-01 -2.83428282e-01
6.26838446e-01 -6.16463900e-01 8.44188392e-01 5.94597757e-01
-1.35645878e+00 1.17884263e-01 -7.82858312e-01 2.35119373e-01
-2.33966917e-01 -3.76786917e-01 6.13089979e-01 7.55703151e-01
-1.34817886e+00 5.30705333e-01 -1.17015898e+00 -3.85854810e-01
9.86798048e-01 4.42825407e-01 -7.67600775e-01 -1.44299999e-01
-7.12737262e-01 1.21063697e+00 3.32560271e-01 1.64052378e-02
-6.14883900e-01 -1.11050355e+00 -5.76480448e-01 -4.34841782e-01
2.70032763e-01 -5.09689152e-01 1.01003933e+00 -6.14272594e-01
-7.95868218e-01 1.32504737e+00 9.41832587e-02 -6.20738685e-01
9.56493080e-01 4.05313194e-01 -4.47413027e-01 3.57997358e-01
3.40613365e-01 7.50294507e-01 6.17982984e-01 -6.83861315e-01
2.38926429e-02 -1.09415615e+00 -9.10399616e-01 -4.84245531e-02
3.49845856e-01 4.91920501e-01 -8.27633440e-02 -5.59421718e-01
5.12542903e-01 -7.23818779e-01 -4.54184800e-01 1.44814989e-02
-1.70327768e-01 4.76595789e-01 2.03870952e-01 -1.08029425e+00
8.09145331e-01 -1.85043776e+00 6.10601716e-02 3.09480369e-01
7.32679486e-01 9.01534874e-03 1.47284523e-01 -1.50639579e-01
-2.63812959e-01 2.65007913e-01 -9.56062078e-01 -6.85548084e-03
-1.51069656e-01 -3.21028195e-02 4.78192031e-01 1.03256750e+00
-1.25784902e-02 9.98621345e-01 -7.50970304e-01 -6.73314869e-01
1.79367572e-01 2.25980669e-01 -3.81772608e-01 -2.10072190e-01
4.05804068e-01 7.90798247e-01 -2.76344210e-01 6.87058330e-01
6.95908248e-01 -5.66445142e-02 2.21718773e-01 1.48225680e-01
-2.44412452e-01 -1.22869626e-01 -8.20366919e-01 1.87644804e+00
1.00009955e-01 5.13606131e-01 4.57880944e-01 -9.44832087e-01
6.30749643e-01 3.79908621e-01 8.20974410e-01 -9.31699455e-01
4.92621243e-01 5.93280494e-01 6.45072758e-01 -5.71921527e-01
5.55107519e-02 -4.00926858e-01 4.17672366e-01 6.34795189e-01
2.30015904e-01 -5.30123413e-01 3.04716527e-01 1.46869093e-01
1.40725183e+00 -3.38890642e-01 1.89566895e-01 -6.30615056e-01
1.03144079e-01 2.44424921e-02 3.68806660e-01 4.54023123e-01
-6.79635286e-01 1.01076448e+00 5.31678498e-01 -4.08739358e-01
-1.57458246e+00 -1.13627887e+00 -7.43309140e-01 4.60992903e-01
-5.68173110e-01 3.28640337e-03 -1.09936285e+00 -4.00352657e-01
-1.08618565e-01 2.64108121e-01 -7.56648481e-01 1.31002277e-01
-4.90352422e-01 -1.41574621e+00 6.51734471e-01 2.87331462e-01
3.61767322e-01 -1.29910207e+00 -1.11233795e+00 3.28669190e-01
2.74682846e-02 -1.11285269e+00 -1.62612006e-01 6.87089004e-03
-1.29333735e+00 -1.21513700e+00 -1.06918800e+00 -3.96936774e-01
5.78181684e-01 -1.96373612e-01 1.39677453e+00 1.89186603e-01
-9.36361372e-01 2.17835486e-01 -1.40275672e-01 -2.79992282e-01
-2.93110847e-01 8.58093053e-02 -4.51684650e-03 -3.40753227e-01
1.80813476e-01 -8.98996592e-01 -8.99500489e-01 1.92621455e-01
-9.63092685e-01 -9.37585682e-02 7.98448980e-01 4.76924062e-01
7.35201240e-01 -3.42512429e-01 7.70717204e-01 -9.38320816e-01
6.93456650e-01 -7.42589712e-01 -3.66256446e-01 4.42617275e-02
-5.10679603e-01 -2.85439879e-01 1.42759271e-02 -2.40401849e-02
-5.14339209e-01 -1.46456808e-01 -2.28919938e-01 1.68709740e-01
-5.51550388e-01 4.69122589e-01 1.87572718e-01 -2.40990415e-01
6.37191415e-01 -8.51196609e-03 5.34905851e-01 -5.31490326e-01
9.88567527e-03 3.36258322e-01 7.29606867e-01 -3.19849998e-01
1.36197791e-01 5.03355265e-01 9.49595124e-02 -8.35251391e-01
-3.65635455e-02 -3.32132876e-01 -1.33917224e+00 -3.63933206e-01
9.71550226e-01 -5.04158974e-01 -1.36256680e-01 6.28255785e-01
-6.21099710e-01 -6.17502093e-01 -1.51685342e-01 6.46232247e-01
-5.29623151e-01 2.83207417e-01 -3.72075200e-01 -4.00214881e-01
-5.69766283e-01 -1.57595873e+00 8.96863222e-01 -2.08037406e-01
-4.34235513e-01 -7.10346520e-01 1.13618672e-01 5.03896773e-01
6.29194200e-01 5.75457692e-01 9.41456139e-01 -6.51094615e-01
-1.26930296e-01 -1.52970299e-01 -2.64598191e-01 1.36298418e-01
-2.33936548e-01 -7.76253939e-02 -6.59342110e-01 -1.33003801e-01
4.44466397e-02 -2.65578777e-01 6.03058159e-01 8.18079770e-01
1.21861672e+00 2.65795141e-01 -1.02892525e-01 3.99127334e-01
1.11520147e+00 1.09303735e-01 8.36245894e-01 5.63669086e-01
3.68207872e-01 7.09781051e-01 4.61631455e-02 1.55014083e-01
3.56953800e-01 4.56895351e-01 3.96597832e-01 -1.36980250e-01
-2.83611834e-01 7.16625035e-01 -2.20955119e-01 8.66277397e-01
-3.94703388e-01 4.47795838e-01 -1.60558271e+00 6.68251395e-01
-1.28666008e+00 -7.88990915e-01 -6.55741274e-01 2.37872124e+00
6.44771159e-01 2.09225684e-01 4.66091812e-01 8.09529722e-02
7.10403800e-01 -1.02659382e-01 -6.96207762e-01 -2.26350889e-01
-1.44438267e-01 4.35927004e-01 6.91626608e-01 6.19400106e-02
-7.63978839e-01 4.77774024e-01 7.55216360e+00 3.90912354e-01
-9.79881346e-01 6.00198627e-01 9.67983902e-01 -3.45291913e-01
-1.05061665e-01 -5.20486176e-01 -1.74404250e-03 5.88314235e-01
1.12159252e+00 -5.39703518e-02 5.36303639e-01 5.24943531e-01
3.66601586e-01 -4.79279995e-01 -6.34644210e-01 7.91754723e-01
7.61813670e-02 -1.14450276e+00 -3.37876767e-01 2.39684433e-01
4.15673345e-01 8.29853833e-01 -1.22773670e-01 -2.07866251e-01
-8.13435912e-02 -1.19440639e+00 7.15744257e-01 5.51361978e-01
9.34142768e-01 -5.79774201e-01 7.90107071e-01 1.38149396e-01
-5.13896465e-01 2.39794269e-01 -2.53434956e-01 2.66218632e-01
3.53982210e-01 9.28159356e-01 -9.01571512e-01 3.18235010e-01
8.78660142e-01 3.37861001e-01 -1.07553351e+00 1.60187399e+00
2.54774809e-01 5.34255087e-01 -2.80015528e-01 6.30192637e-01
1.01825502e-02 -3.07338536e-01 5.88288426e-01 1.05974317e+00
2.74863422e-01 2.30299890e-01 -2.01945752e-01 8.21440876e-01
3.83484997e-02 3.47074777e-01 -4.24954802e-01 6.19149804e-02
1.22189790e-01 1.48099172e+00 -1.47203648e+00 -2.82759041e-01
-4.00918603e-01 6.65153384e-01 3.63276511e-01 -1.55341297e-01
-4.83939022e-01 1.77060902e-01 1.88894749e-01 7.58105040e-01
-2.64410019e-01 -4.11260456e-01 -5.27832329e-01 -9.61036205e-01
5.92230633e-02 -1.00945163e+00 3.40761691e-01 -8.50343287e-01
-1.22113001e+00 7.43428648e-01 9.81669798e-02 -5.02964735e-01
-1.26157284e-01 -4.93997514e-01 -4.28461641e-01 7.77457118e-01
-1.06170583e+00 -5.81094623e-01 -2.14221716e-01 4.28735316e-01
2.21390083e-01 5.94277419e-02 6.75126076e-01 4.85967696e-01
-5.06596029e-01 1.20386459e-01 -9.01999176e-02 4.30858023e-02
7.35191822e-01 -1.24377608e+00 5.16671419e-01 6.95306480e-01
-9.81588140e-02 5.01958668e-01 4.86932307e-01 -7.80327499e-01
-1.00735819e+00 -6.86254799e-01 4.13416356e-01 -4.51085597e-01
9.60344434e-01 -1.88074082e-01 -1.18459725e+00 6.67026460e-01
-1.01785168e-01 1.50737494e-01 6.91334844e-01 -1.56833097e-01
-3.49119566e-02 2.49705404e-01 -1.48183966e+00 2.37859532e-01
1.05719972e+00 2.61971094e-02 -8.46271634e-01 2.21953824e-01
1.16070084e-01 -5.13862133e-01 -1.30531490e+00 3.97134721e-01
6.93898380e-01 -1.23743701e+00 8.93489301e-01 -3.50745052e-01
3.71926099e-01 1.38773024e-01 2.52330244e-01 -1.33633041e+00
-1.61926150e-01 -1.45817548e-01 2.46068716e-01 8.36481392e-01
4.78172004e-01 -8.04302335e-01 6.12047851e-01 1.05342758e+00
-4.14911062e-01 -6.79275155e-01 -1.44356859e+00 -6.33379579e-01
5.19936860e-01 -4.87709224e-01 7.77088404e-01 1.05637479e+00
1.70744658e-02 -2.62128830e-01 3.24237764e-01 -3.13986748e-01
8.26763272e-01 -3.40366483e-01 2.92032361e-01 -1.58482659e+00
1.61901079e-02 -8.84487033e-01 -9.10060287e-01 1.84371874e-01
-1.10311724e-01 -1.16319108e+00 -1.04927309e-01 -1.56152654e+00
4.12049979e-01 -6.18382514e-01 -1.72591239e-01 4.61443067e-01
1.05129704e-01 8.11787128e-01 -8.57425779e-02 1.70186266e-01
-2.17916757e-01 4.80852462e-02 1.55732548e+00 9.69283655e-02
1.92813426e-02 -4.40021902e-01 -4.60575134e-01 8.11864316e-01
9.60838079e-01 -4.78185564e-01 -1.83539629e-01 -5.20050108e-01
1.32134587e-01 -7.83325061e-02 5.80068827e-01 -1.26263011e+00
-4.03335355e-02 2.48447299e-01 6.21659636e-01 -4.04082656e-01
-1.16742231e-01 -5.12341082e-01 2.63980061e-01 3.43771785e-01
2.69303117e-02 3.84600252e-01 1.91892758e-01 5.40868044e-02
7.28044733e-02 -1.51169926e-01 8.51406455e-01 -5.05831361e-01
-2.96503216e-01 5.07454276e-01 -6.03090346e-01 4.36767250e-01
8.42919409e-01 -2.59229511e-01 -2.81730801e-01 1.22886434e-01
-9.65483129e-01 1.74548328e-01 5.61191559e-01 8.81848112e-02
4.27188218e-01 -1.00409329e+00 -9.46733296e-01 1.92077681e-01
-1.42853200e-01 1.53685227e-01 6.54948711e-01 1.78843713e+00
-7.85866559e-01 1.76809832e-01 -7.72708535e-01 -6.08193696e-01
-1.14646351e+00 3.32652062e-01 4.00716782e-01 -1.93024099e-01
-1.22874379e+00 5.82018375e-01 -1.65508747e-01 -3.79639506e-01
-3.68363172e-01 -3.47321004e-01 -7.43035078e-02 3.06945205e-01
7.65481591e-01 3.74935955e-01 6.31076753e-01 -1.02515495e+00
-5.91065168e-01 2.26925805e-01 -5.51517010e-02 -4.14787263e-01
1.67713392e+00 -1.85588032e-01 -4.89857703e-01 4.30075824e-01
8.00926447e-01 -3.00383002e-01 -8.71015370e-01 2.79615700e-01
1.54861093e-01 -3.52819353e-01 3.88755918e-01 -1.05314136e+00
-1.39293730e+00 8.76075864e-01 6.93007886e-01 1.12058133e-01
8.84041011e-01 -3.79740298e-02 7.61927545e-01 -3.35490078e-01
7.45858133e-01 -9.79545355e-01 -5.56642711e-01 9.35332105e-02
9.80135679e-01 -1.11641073e+00 1.91483185e-01 -1.43603683e-01
-7.15725601e-01 9.13724184e-01 1.97666332e-01 -9.76588279e-02
5.38602114e-01 5.32436311e-01 -2.42174435e-02 -8.14139247e-01
-2.07299307e-01 -8.55153333e-03 2.04152897e-01 7.70491898e-01
5.61191082e-01 2.00882182e-01 -5.83125770e-01 5.67959785e-01
-5.69532275e-01 1.02464780e-01 6.63251698e-01 7.84010887e-01
-3.64962429e-01 -9.02470648e-01 -5.47301292e-01 1.13867521e+00
-7.42455244e-01 -1.63542882e-01 -4.28321213e-01 8.23754549e-01
8.61276537e-02 5.01682878e-01 1.02926195e-01 1.89857259e-01
2.62161404e-01 2.98939377e-01 8.30152392e-01 -6.20378375e-01
-6.35780215e-01 -4.45357934e-02 -2.81449631e-02 -7.35302687e-01
-2.61411160e-01 -1.19589055e+00 -1.38032401e+00 -2.56288737e-01
9.44660753e-02 -2.72906840e-01 8.17249775e-01 9.91331279e-01
2.87248999e-01 6.40566289e-01 -1.76444590e-01 -1.00179720e+00
3.95958349e-02 -1.11046207e+00 -9.69125986e-01 3.64621550e-01
-7.74537325e-02 -9.32649136e-01 -1.75849050e-01 -1.75067410e-02] | [14.112432479858398, -2.2738490104675293] |
529de9f7-a3af-4121-a717-a1ca4d073c98 | nonparanormal-information-estimation | 1702.07803 | null | http://arxiv.org/abs/1702.07803v1 | http://arxiv.org/pdf/1702.07803v1.pdf | Nonparanormal Information Estimation | We study the problem of using i.i.d. samples from an unknown multivariate
probability distribution $p$ to estimate the mutual information of $p$. This
problem has recently received attention in two settings: (1) where $p$ is
assumed to be Gaussian and (2) where $p$ is assumed only to lie in a large
nonparametric smoothness class. Estimators proposed for the Gaussian case
converge in high dimensions when the Gaussian assumption holds, but are
brittle, failing dramatically when $p$ is not Gaussian. Estimators proposed for
the nonparametric case fail to converge with realistic sample sizes except in
very low dimensions. As a result, there is a lack of robust mutual information
estimators for many realistic data. To address this, we propose estimators for
mutual information when $p$ is assumed to be a nonparanormal (a.k.a., Gaussian
copula) model, a semiparametric compromise between Gaussian and nonparametric
extremes. Using theoretical bounds and experiments, we show these estimators
strike a practical balance between robustness and scaling with dimensionality. | ['Barnabás Pøczos', 'Shashank Singh'] | 2017-02-24 | nonparanormal-information-estimation-1 | https://icml.cc/Conferences/2017/Schedule?showEvent=805 | http://proceedings.mlr.press/v70/singh17a/singh17a.pdf | icml-2017-8 | ['mutual-information-estimation'] | ['methodology'] | [-2.94216275e-01 9.43967849e-02 -4.62447964e-02 -3.30980837e-01
-9.59113598e-01 -6.48498833e-01 2.86004335e-01 -1.93498194e-01
-2.37514257e-01 9.21542466e-01 -3.46032947e-01 -1.73047677e-01
-5.64268589e-01 -6.94372892e-01 -8.21936190e-01 -8.53217065e-01
-3.31192374e-01 6.67580843e-01 -1.23780243e-01 2.91693002e-01
3.50800395e-01 4.06343192e-01 -1.57643914e+00 -5.67452371e-01
1.25497031e+00 1.07953107e+00 -1.13621019e-02 6.66899979e-01
-9.58021451e-03 6.27489462e-02 -4.55155134e-01 -5.72017908e-01
4.38086271e-01 -2.54745901e-01 -4.07078981e-01 -9.57541838e-02
1.73362017e-01 6.24974295e-02 1.25373472e-02 1.50434184e+00
2.59016544e-01 9.62150618e-02 1.35228872e+00 -1.51335669e+00
-5.27671158e-01 5.88597775e-01 -9.23372507e-01 -1.34182885e-01
3.46716195e-02 -1.01891190e-01 9.89115536e-01 -6.05954528e-01
2.14800119e-01 1.19317281e+00 9.99857306e-01 3.09854358e-01
-1.50226057e+00 -8.65132034e-01 -1.89135909e-01 -6.08049870e-01
-1.82593858e+00 -3.67392600e-01 6.47148192e-01 -5.39577782e-01
4.95101541e-01 7.94139355e-02 1.06867939e-01 7.76182413e-01
3.25351387e-01 4.61404741e-01 1.09689569e+00 -1.24646090e-01
5.42502880e-01 2.42111862e-01 -4.89127189e-02 4.42393780e-01
8.40131998e-01 4.28784788e-02 -9.95753258e-02 -6.79462433e-01
7.70285428e-01 -2.25782454e-01 1.32185360e-03 -7.71577358e-01
-6.09728038e-01 9.22574818e-01 -1.63123369e-01 3.64980698e-01
-2.66148537e-01 3.08730662e-01 3.34420451e-03 4.78322655e-01
6.11496925e-01 4.08757657e-01 -5.29002249e-01 -4.40076917e-01
-9.36762512e-01 5.41129768e-01 1.14913845e+00 1.20708740e+00
6.95478559e-01 3.93083654e-02 5.78096688e-01 9.03412461e-01
4.23046619e-01 9.70209122e-01 8.52045491e-02 -1.20220041e+00
5.90640426e-01 1.38729736e-01 8.41973126e-01 -9.40833569e-01
-3.32903951e-01 -2.30510682e-01 -7.35612392e-01 3.09805200e-02
1.06323433e+00 -3.51279318e-01 -3.62176120e-01 2.14169121e+00
9.81020033e-02 2.49968003e-02 1.84121937e-01 3.13342273e-01
1.15269691e-01 3.40476125e-01 -2.69859768e-02 -3.38789582e-01
9.52757418e-01 -1.11207003e-02 -5.33296049e-01 -1.97346229e-02
6.07054710e-01 -8.06599796e-01 8.74124885e-01 5.19608796e-01
-1.42359030e+00 -2.11293176e-01 -8.40769470e-01 5.74233413e-01
-1.93107635e-01 -1.77532613e-01 4.53428954e-01 1.14172626e+00
-1.11404991e+00 6.86289251e-01 -7.24684477e-01 -2.24002734e-01
1.39084816e-01 3.78217101e-01 -3.23017538e-01 -1.90187559e-01
-8.87037396e-01 5.92633188e-01 5.87211140e-02 -7.11595938e-02
-3.08107823e-01 -5.84169686e-01 -6.47579312e-01 3.72774713e-03
-1.81320906e-01 -6.68973804e-01 9.48949456e-01 -8.29991281e-01
-1.37118316e+00 6.77696288e-01 -5.50418720e-02 -3.40001076e-01
6.83216929e-01 -1.35568693e-01 -2.16002017e-01 7.60172978e-02
-7.70419510e-03 -2.97772624e-02 9.36564326e-01 -1.39658225e+00
-2.48724774e-01 -5.87910950e-01 -4.04885769e-01 3.09764203e-02
1.03575669e-01 -4.83077094e-02 -1.90672040e-01 -4.28990036e-01
3.73174340e-01 -9.33919489e-01 -8.30431208e-02 -1.77277774e-01
-5.43293893e-01 -2.21963331e-01 3.95962536e-01 -5.10556698e-01
1.01036620e+00 -2.22603035e+00 -2.54392773e-01 5.85361540e-01
8.19784477e-02 -2.77722716e-01 1.57551616e-01 4.35419142e-01
3.66588705e-04 3.76847804e-01 -6.57735467e-01 -4.91599441e-01
3.94141465e-01 1.15739204e-01 2.77566575e-02 1.07721257e+00
-1.31441519e-01 2.92539746e-01 -5.90730667e-01 -3.68162930e-01
-2.73815781e-01 3.55512679e-01 -6.87054574e-01 7.98691884e-02
5.85008487e-02 1.48666114e-01 -3.72575134e-01 6.59894764e-01
1.13815248e+00 -3.58411759e-01 1.18417032e-01 6.46635592e-02
-1.02305841e-02 -3.64583194e-01 -1.60081148e+00 1.18225574e+00
-3.26953292e-01 4.05130416e-01 5.40863872e-01 -1.37302256e+00
1.08696747e+00 1.21936224e-01 6.38530731e-01 -1.42684013e-01
3.44123095e-01 5.65576196e-01 1.49446890e-01 -3.33694577e-01
3.25825155e-01 -7.44254410e-01 -3.06908488e-01 3.66753668e-01
7.41916299e-02 -5.19344687e-01 -1.90031566e-02 -5.29235788e-02
1.15704179e+00 -1.39312193e-01 1.98090568e-01 -7.60478139e-01
-1.06092934e-02 -5.74369431e-01 5.84964871e-01 1.00823653e+00
-3.47445488e-01 6.68173909e-01 7.82054126e-01 5.06885052e-01
-1.12639725e+00 -1.61797380e+00 -6.40160501e-01 3.88604879e-01
1.26776397e-01 1.68931663e-01 -8.24444294e-01 -4.01520848e-01
4.55844730e-01 7.38353729e-01 -6.44275963e-01 -1.16503112e-01
6.11423403e-02 -8.81899953e-01 5.14635205e-01 1.97361991e-01
3.43025565e-01 -2.65197039e-01 -8.93060192e-02 6.31271303e-02
5.45133241e-02 -8.36074889e-01 -4.39411223e-01 1.67476639e-01
-8.63575995e-01 -1.28601289e+00 -9.55533028e-01 -4.97394681e-01
6.10432863e-01 -2.45286614e-01 1.11459291e+00 -5.61994731e-01
-9.13014114e-02 8.29909325e-01 -3.22278440e-02 -2.05189392e-01
-2.51302332e-01 -4.75365609e-01 3.71193051e-01 5.65034337e-02
3.68150085e-01 -6.71754122e-01 -6.94473505e-01 7.82294452e-01
-8.51337731e-01 -8.92954648e-01 4.22826260e-01 6.23358667e-01
4.16960537e-01 5.14120281e-01 8.42014074e-01 -7.61651397e-01
7.23839462e-01 -9.38762486e-01 -6.05430067e-01 4.36554700e-02
-5.58730245e-01 2.04614878e-01 4.56921309e-01 -5.66318214e-01
-9.72591639e-01 -2.85914779e-01 8.32288116e-02 -5.56024790e-01
-1.74122512e-01 5.24264634e-01 -3.43908757e-01 5.21457642e-02
5.58180094e-01 -1.17902100e-01 2.43600830e-01 -6.32484734e-01
2.52935439e-01 7.67482996e-01 4.52632755e-01 -1.02943861e+00
6.92025185e-01 5.01370072e-01 1.57532156e-01 -1.06250072e+00
-4.57662493e-01 -5.93709767e-01 -2.91271865e-01 2.20987305e-01
4.99404073e-01 -6.05003536e-01 -8.58526111e-01 3.93723667e-01
-6.94764912e-01 -3.98871839e-01 -5.49729407e-01 7.93312013e-01
-1.20607030e+00 4.55695391e-01 -4.24913347e-01 -1.41020429e+00
1.23872533e-01 -8.75583172e-01 7.85054982e-01 1.63906395e-01
-1.14517406e-01 -1.27512038e+00 3.10915083e-01 -1.34050936e-01
4.02644157e-01 3.01306218e-01 6.84565008e-01 -8.14538956e-01
6.54776692e-02 -5.86771727e-01 -2.81139046e-01 5.56776226e-01
2.25258872e-01 2.49463037e-01 -5.45293391e-01 -2.89437711e-01
2.32037917e-01 -1.68519631e-01 5.06574392e-01 8.80614102e-01
1.11331272e+00 -4.89126623e-01 -3.00245523e-01 5.36487937e-01
1.30543387e+00 2.44043216e-01 4.90061015e-01 -2.82559127e-01
-6.29199669e-02 7.26481140e-01 4.78619128e-01 7.00874567e-01
3.76911968e-01 1.45281136e-01 2.74194717e-01 3.99621099e-01
6.19906843e-01 -2.00572774e-01 1.66675240e-01 6.43215418e-01
1.00915939e-01 -2.96993464e-01 -9.47709203e-01 6.11713290e-01
-1.65280271e+00 -1.09031200e+00 -4.11877148e-02 2.64740872e+00
7.76447415e-01 1.70255706e-01 3.37247163e-01 -2.42833704e-01
8.92570198e-01 -2.97042698e-01 -6.95295691e-01 -3.06646585e-01
-1.14457652e-01 2.44124323e-01 7.88757026e-01 3.94678444e-01
-1.03735375e+00 2.60774612e-01 6.64460564e+00 7.69561410e-01
-5.54848075e-01 -4.29918431e-02 5.74827969e-01 2.94586331e-01
-3.04286271e-01 1.69952959e-01 -7.38939285e-01 8.97550941e-01
1.16473341e+00 -2.99902767e-01 -2.52064951e-02 9.04219091e-01
1.74631625e-02 -3.56283337e-01 -1.02674627e+00 1.19501281e+00
-2.97893167e-01 -6.18978560e-01 -5.67857981e-01 3.46415371e-01
7.88531363e-01 -7.91715160e-02 4.82998401e-01 2.89226770e-01
7.41190016e-01 -1.02240658e+00 2.16571987e-01 9.77036893e-01
7.61116624e-01 -1.14810896e+00 7.02739537e-01 5.67109227e-01
-1.00266302e+00 1.89154506e-01 -6.08288109e-01 1.68507323e-01
7.96335042e-02 9.62113857e-01 -2.96425462e-01 2.70565867e-01
7.99400210e-01 4.37318087e-01 4.00564447e-02 1.33843935e+00
4.56647009e-01 5.11616945e-01 -9.51488793e-01 1.33894026e-01
-3.85033339e-02 -5.21229625e-01 4.69093144e-01 1.09691191e+00
9.70225751e-01 -3.77995078e-03 -1.80925652e-01 9.97997820e-01
2.50166990e-02 6.32401034e-02 -8.41933727e-01 -1.55046314e-01
5.57006240e-01 7.34054923e-01 -5.40077031e-01 -8.62798169e-02
-3.75134319e-01 7.20625758e-01 -1.25681339e-02 4.68610704e-01
-6.20168746e-01 -7.29765415e-01 8.74433517e-01 1.46378309e-01
5.41886210e-01 -4.38338101e-01 -1.15990959e-01 -1.22143626e+00
-3.74654569e-02 -2.60509372e-01 3.81692022e-01 -2.78903306e-01
-2.05129480e+00 8.19093175e-03 3.39802355e-01 -1.25785720e+00
-4.13285017e-01 -6.27782702e-01 -5.22646666e-01 1.08795822e+00
-1.01371658e+00 -4.82317567e-01 4.65856850e-01 5.49419284e-01
-1.93046778e-01 1.90063581e-01 5.23282886e-01 -2.51746830e-02
-3.76744598e-01 7.59655178e-01 9.30116594e-01 -2.50544816e-01
7.37113714e-01 -1.49686491e+00 5.46849705e-02 3.48351628e-01
-3.90624642e-01 9.83390927e-01 1.20976889e+00 -6.87535703e-01
-1.43986285e+00 -7.01501608e-01 5.28597891e-01 -3.98255974e-01
9.40417230e-01 -2.56979376e-01 -9.25067902e-01 5.35206497e-01
-2.38393500e-01 3.43935549e-01 1.01017618e+00 2.47847155e-01
-4.19423908e-01 1.49717331e-01 -1.82717252e+00 2.81761110e-01
7.04687357e-01 -3.67396683e-01 -3.73861462e-01 2.15031222e-01
2.06008226e-01 6.21501319e-02 -1.43641889e+00 3.71877164e-01
7.11454570e-01 -1.02115941e+00 8.65008652e-01 -3.30928594e-01
-1.85001176e-02 -6.36751652e-02 -6.29281104e-01 -1.30445600e+00
3.07211936e-01 -8.10916245e-01 -8.99449177e-03 1.26354480e+00
5.28731287e-01 -1.01734149e+00 7.18252182e-01 1.13892639e+00
2.49213129e-01 -4.52621669e-01 -1.35765946e+00 -1.24734080e+00
7.23398983e-01 -6.37781680e-01 2.50399530e-01 9.34028566e-01
2.36283392e-01 -1.97863713e-01 -2.32350960e-01 1.32267654e-01
1.14572465e+00 1.23132311e-01 7.68902838e-01 -1.48962247e+00
-5.84231973e-01 -4.84096617e-01 -3.54182780e-01 -1.24496865e+00
3.30152839e-01 -4.98014539e-01 3.44355404e-01 -1.02022266e+00
2.41916806e-01 -8.94012511e-01 -7.06690326e-02 -4.79967147e-01
1.70959439e-02 -7.69053549e-02 -3.00904870e-01 2.26372540e-01
-3.93459469e-01 5.90381682e-01 8.27331662e-01 2.78610021e-01
-2.02817082e-01 5.10805190e-01 -8.11594009e-01 1.03122520e+00
9.66765881e-01 -2.83117592e-01 -3.52479577e-01 2.97455251e-01
1.66937962e-01 5.09259462e-01 7.20957294e-02 -8.75230372e-01
9.64276642e-02 -2.69849002e-01 1.19283400e-01 -4.47605789e-01
4.86987770e-01 -8.46282601e-01 1.68213069e-01 1.80275564e-03
-2.53497325e-02 -1.85330654e-03 8.07645023e-02 1.00500858e+00
-4.89083081e-02 -3.01023543e-01 1.10320425e+00 1.25893459e-01
2.97510147e-01 2.62812912e-01 -2.88325667e-01 5.79487562e-01
8.10473204e-01 -1.18282400e-01 -2.53174931e-01 -6.96447074e-01
-8.10977697e-01 1.54949082e-02 7.10664272e-01 -2.18131214e-01
3.95215094e-01 -1.08103693e+00 -4.85137582e-01 4.53156792e-02
9.44504235e-03 -2.15700746e-01 3.69238138e-01 1.05765033e+00
-1.65230781e-01 1.18964545e-01 3.78516316e-01 -6.45372689e-01
-5.22423744e-01 5.24026334e-01 3.41411799e-01 -8.42283666e-02
-1.37672231e-01 7.38347232e-01 2.70648807e-01 -5.66350281e-01
5.98518550e-02 -3.64070565e-01 2.66049087e-01 -7.08829686e-02
2.85834074e-01 3.77821058e-01 -3.96755159e-01 -6.50937915e-01
-1.89462289e-01 5.74587822e-01 9.39445645e-02 -2.42524251e-01
1.12402999e+00 -3.13603282e-01 7.23056719e-02 6.90596342e-01
1.55998981e+00 3.59207451e-01 -1.51982391e+00 -1.48491248e-01
7.70328520e-03 -5.03180623e-01 -4.25589651e-01 -2.99369365e-01
-8.03196430e-01 5.82176685e-01 5.10881603e-01 5.63379467e-01
6.57947063e-01 2.45009691e-01 3.48927170e-01 1.24153830e-01
4.04615939e-01 -1.20603895e+00 -1.61898181e-01 3.07123899e-01
6.41611040e-01 -1.12299478e+00 -1.19738795e-01 -1.50501788e-01
-3.95894498e-01 8.87072802e-01 2.06250653e-01 -5.12759268e-01
1.36663485e+00 4.26685065e-01 -5.05261123e-01 -1.34290725e-01
-3.15233767e-01 -2.43709553e-02 2.38945752e-01 6.42247438e-01
3.91829431e-01 2.96601653e-01 -3.84785801e-01 9.17890370e-01
-3.61210078e-01 -3.84833276e-01 3.19906592e-01 8.72687161e-01
-5.83227038e-01 -8.66766274e-01 -4.16104525e-01 8.49904001e-01
-6.13325655e-01 1.13779217e-01 9.89173539e-03 1.09536231e+00
-2.63437420e-01 9.91062880e-01 3.08094949e-01 3.58460024e-02
2.22380161e-01 8.64234716e-02 3.83634031e-01 -1.68372676e-01
5.75874522e-02 2.31193438e-01 -1.96042702e-01 -4.28275377e-01
-3.12633753e-01 -1.16031206e+00 -7.61665046e-01 -5.71482480e-01
-4.58951861e-01 5.81417024e-01 7.00389564e-01 7.62223244e-01
6.37243986e-02 -2.19516128e-01 6.15719736e-01 -6.20970845e-01
-1.14470601e+00 -9.20777917e-01 -1.48429072e+00 1.98369518e-01
1.63831472e-01 -9.21019018e-01 -1.08524168e+00 -2.37437665e-01] | [7.2349853515625, 4.178459644317627] |
d06593dd-5677-47f1-9b8e-b5d160dc0bd0 | ror-read-over-read-for-long-document-machine | 2109.04780 | null | https://arxiv.org/abs/2109.04780v2 | https://arxiv.org/pdf/2109.04780v2.pdf | RoR: Read-over-Read for Long Document Machine Reading Comprehension | Transformer-based pre-trained models, such as BERT, have achieved remarkable results on machine reading comprehension. However, due to the constraint of encoding length (e.g., 512 WordPiece tokens), a long document is usually split into multiple chunks that are independently read. It results in the reading field being limited to individual chunks without information collaboration for long document machine reading comprehension. To address this problem, we propose RoR, a read-over-read method, which expands the reading field from chunk to document. Specifically, RoR includes a chunk reader and a document reader. The former first predicts a set of regional answers for each chunk, which are then compacted into a highly-condensed version of the original document, guaranteeing to be encoded once. The latter further predicts the global answers from this condensed document. Eventually, a voting strategy is utilized to aggregate and rerank the regional and global answers for final prediction. Extensive experiments on two benchmarks QuAC and TriviaQA demonstrate the effectiveness of RoR for long document reading. Notably, RoR ranks 1st place on the QuAC leaderboard (https://quac.ai/) at the time of submission (May 17th, 2021). | ['BoWen Zhou', 'Xiaodong He', 'Youzheng Wu', 'Yongwei Zhou', 'Yifan Wang', 'Junwei Bao', 'Jing Zhao'] | 2021-09-10 | null | https://aclanthology.org/2021.findings-emnlp.160 | https://aclanthology.org/2021.findings-emnlp.160.pdf | findings-emnlp-2021-11 | ['triviaqa'] | ['miscellaneous'] | [ 4.23857778e-01 5.68728030e-01 -2.70329297e-01 -1.53607965e-01
-1.19006681e+00 -5.33379912e-01 5.32082915e-01 5.83452106e-01
-3.46152395e-01 7.90783942e-01 7.79518723e-01 -3.80945355e-01
5.99643029e-03 -8.73919070e-01 -7.79663146e-01 -5.99445581e-01
4.79885191e-01 6.73373461e-01 2.65091896e-01 -2.61512876e-01
5.24795413e-01 -4.01782483e-01 -1.49124980e+00 6.02806866e-01
1.37144303e+00 9.78706896e-01 7.12475061e-01 7.03029454e-01
-2.56726056e-01 1.22760177e+00 -6.25306249e-01 -4.64632928e-01
-1.19221322e-01 -7.00343132e-01 -1.33328664e+00 -1.79467350e-01
2.89155304e-01 -6.13026440e-01 -1.94078460e-01 7.75260031e-01
4.04324591e-01 2.25340888e-01 4.30571049e-01 -6.15645289e-01
-9.50435042e-01 1.28133476e+00 -3.76176298e-01 1.55766442e-01
6.78515851e-01 -1.95862025e-01 1.54460526e+00 -9.73396540e-01
4.46766973e-01 8.52186739e-01 6.89872131e-02 5.63699186e-01
-7.25477755e-01 -3.78017247e-01 6.55551627e-02 5.13504267e-01
-1.05825210e+00 -4.90689099e-01 5.39633095e-01 -1.45960942e-01
8.28204930e-01 4.94212568e-01 5.74120700e-01 6.42937481e-01
3.15506041e-01 1.22860157e+00 9.27207410e-01 -4.51052129e-01
2.46740118e-01 -3.84658575e-01 6.27357066e-01 4.00078893e-01
-1.97825029e-01 -4.14766133e-01 -9.19311523e-01 2.81177878e-01
-1.05590364e-02 -9.20402855e-02 -6.12437367e-01 1.87036991e-01
-1.33235323e+00 7.73955941e-01 2.60501325e-01 1.57421574e-01
-4.98786896e-01 -2.81481266e-01 4.37381506e-01 4.51729864e-01
4.92141604e-01 4.66978759e-01 -3.21339548e-01 -2.77920127e-01
-1.17312324e+00 3.41026336e-01 8.37430060e-01 9.04038072e-01
6.93235099e-01 -6.24071896e-01 -7.30989039e-01 8.54318619e-01
3.04273907e-02 5.15821040e-01 6.66539073e-01 -7.93013155e-01
9.02113676e-01 6.33330405e-01 1.29218698e-01 -7.68003881e-01
-1.17107496e-01 -6.84399068e-01 -1.16201687e+00 -5.04937172e-01
2.52166599e-01 1.92018285e-01 -6.60750151e-01 1.36431098e+00
1.87077105e-01 -3.51789504e-01 2.32675686e-01 9.47298646e-01
1.12830043e+00 1.09664607e+00 -2.21498325e-01 -3.72396290e-01
1.45182431e+00 -1.39891315e+00 -6.45332158e-01 -1.09319367e-01
6.93718970e-01 -7.48086810e-01 1.09186471e+00 5.71163058e-01
-1.53518045e+00 -7.47924685e-01 -9.46488857e-01 -5.50164700e-01
1.14245955e-02 2.20341578e-01 -1.06293239e-01 1.32260934e-01
-9.73217964e-01 2.14145243e-01 -4.10433352e-01 1.24196567e-01
3.06925237e-01 -3.18771563e-02 -1.49370427e-03 -4.16535944e-01
-1.32701445e+00 8.19898188e-01 6.31636977e-01 6.25117943e-02
-8.22724879e-01 -7.50821054e-01 -7.13941872e-01 4.56698418e-01
3.59093934e-01 -5.39909780e-01 1.53517485e+00 -6.92671120e-01
-1.65444398e+00 6.01926029e-01 -5.85261226e-01 -5.58581769e-01
3.73424321e-01 -5.99259615e-01 -7.97261745e-02 2.15700537e-01
1.51606694e-01 6.02306664e-01 7.48214781e-01 -8.12710643e-01
-8.97406459e-01 -3.37130189e-01 3.34741652e-01 5.27581990e-01
-3.10444206e-01 -2.84725931e-02 -3.42207879e-01 -4.42706227e-01
1.89174235e-01 -4.70170945e-01 2.51099229e-01 -7.75103509e-01
-5.12453020e-01 -7.66783834e-01 3.32453787e-01 -1.16659951e+00
1.54025304e+00 -1.92841816e+00 6.14256740e-01 -1.00513808e-01
5.76461315e-01 1.33975580e-01 -3.00348908e-01 7.14190006e-01
2.38329411e-01 -1.07438952e-01 -3.64799723e-02 -5.04392982e-01
1.47845754e-02 -3.26412804e-02 -9.73694384e-01 1.56375676e-01
-2.19084695e-01 1.21420157e+00 -1.01099205e+00 -4.99629766e-01
-1.42846420e-01 -1.90547034e-01 -4.10628825e-01 4.79378551e-01
-6.64514184e-01 4.06203151e-01 -4.17568237e-01 3.40884656e-01
6.78750694e-01 -4.22686577e-01 -2.07076028e-01 2.44752601e-01
-2.11513519e-01 9.59403217e-01 -5.03826797e-01 1.76516104e+00
-4.93138552e-01 4.46277797e-01 -2.67849803e-01 -8.43976200e-01
9.77204382e-01 3.44837159e-01 1.15380034e-01 -1.07923734e+00
6.27563894e-02 3.02998841e-01 -1.60475969e-01 -2.45652810e-01
1.15485966e+00 1.07930906e-01 -1.63394928e-01 8.85189891e-01
-2.54172206e-01 -1.53087571e-01 4.07207876e-01 5.24848163e-01
1.01598322e+00 -1.94177046e-01 2.31126517e-01 -1.42380774e-01
8.42143059e-01 5.54258004e-02 3.22818220e-01 7.46107876e-01
2.23338932e-01 7.63577402e-01 4.90038157e-01 -1.79195940e-01
-9.28921402e-01 -9.42518175e-01 1.84133917e-01 1.45435536e+00
3.72753412e-01 -7.88167477e-01 -8.08945298e-01 -7.08392680e-01
-2.54368752e-01 9.16020155e-01 -5.72070539e-01 -2.09217370e-01
-7.49994934e-01 -2.26259232e-02 4.14894968e-01 4.39975232e-01
5.82785010e-01 -1.14803696e+00 -9.31699693e-01 3.18043143e-01
-7.10583627e-01 -7.30220973e-01 -6.86809003e-01 1.06609866e-01
-5.16532004e-01 -8.51231992e-01 -9.87001657e-01 -8.30241442e-01
5.39410651e-01 3.85381341e-01 1.33445287e+00 3.97512674e-01
4.95607883e-01 -2.61500496e-02 -1.08300424e+00 -1.80189788e-01
-5.93104482e-01 5.31486511e-01 -3.42848241e-01 -1.69106260e-01
1.89499050e-01 -8.92416239e-02 -6.82747543e-01 1.65421650e-01
-7.64114499e-01 6.01685286e-01 7.58530736e-01 1.03139067e+00
5.71976364e-01 -1.35953426e-01 7.94225335e-01 -8.06683242e-01
8.90187919e-01 -4.96315122e-01 -3.38659704e-01 7.45238662e-01
-5.08920133e-01 4.60911915e-02 9.82528925e-01 -2.77268618e-01
-1.04143310e+00 -6.14154279e-01 -1.09113604e-01 8.68941694e-02
2.11794123e-01 9.60918248e-01 -9.15505216e-02 7.05166340e-01
2.18877375e-01 9.80728626e-01 -1.72711179e-01 -4.79187250e-01
4.51846629e-01 1.08412552e+00 7.33749986e-01 -4.72978383e-01
5.95705509e-01 -1.78069219e-01 -6.02944553e-01 -5.42536020e-01
-1.35144496e+00 -5.32024682e-01 -5.39887309e-01 -1.50003463e-01
8.15484822e-01 -9.38235462e-01 -5.20981133e-01 6.68906510e-01
-1.32592678e+00 -4.56624568e-01 -4.20994520e-01 2.08983392e-01
-4.15647209e-01 3.84393364e-01 -6.67683244e-01 -4.83093202e-01
-7.87794650e-01 -8.56998563e-01 1.04825127e+00 3.97912890e-01
-3.96848202e-01 -6.68154418e-01 1.52993441e-01 9.21041369e-01
2.74775416e-01 -5.33443809e-01 1.14689779e+00 -6.53629005e-01
-5.75205326e-01 4.43468727e-02 -2.64979482e-01 4.55872148e-01
-3.72637063e-02 -4.13123161e-01 -7.18770862e-01 -4.24202859e-01
8.92484486e-02 -7.99738586e-01 1.01334417e+00 -2.81310920e-02
1.26461470e+00 -4.30221915e-01 2.92289089e-02 2.37755012e-02
8.22201073e-01 -5.85849807e-02 8.57586086e-01 2.01297089e-01
6.19361103e-01 5.45311451e-01 9.25382376e-01 5.59795678e-01
9.02027190e-01 5.05389035e-01 2.96282768e-01 4.08848315e-01
-3.11986595e-01 -6.49852157e-01 4.75217372e-01 1.65821397e+00
9.96189117e-02 -5.62349737e-01 -7.92428136e-01 6.21533334e-01
-1.71201336e+00 -8.09864283e-01 -9.14706066e-02 2.19859719e+00
1.19067872e+00 8.65201876e-02 -2.37155452e-01 2.76393414e-01
4.06162292e-01 4.99637336e-01 -4.58005875e-01 -3.92749012e-01
-1.61082312e-01 3.32269847e-01 1.31489441e-01 4.26156789e-01
-5.66586316e-01 9.00870681e-01 5.22955084e+00 1.11753201e+00
-8.58118892e-01 1.03512652e-01 5.03030121e-01 9.03363824e-02
-6.98839784e-01 2.29425356e-01 -8.69800687e-01 6.24166489e-01
7.66264141e-01 -4.82455254e-01 3.74646127e-01 2.69179434e-01
4.74508442e-02 -4.21407580e-01 -1.00947332e+00 6.71785176e-01
2.95245469e-01 -1.42490876e+00 3.04610133e-01 -1.81101561e-01
7.88555384e-01 -1.41477138e-01 -2.19588657e-03 5.56243896e-01
2.54588336e-01 -1.28030229e+00 1.06614339e+00 5.56256056e-01
7.31139839e-01 -7.16133535e-01 8.34248185e-01 9.65695322e-01
-1.07151616e+00 -9.49189067e-02 -6.54816270e-01 -3.61924440e-01
1.45143837e-01 6.46770000e-01 -5.52753389e-01 8.28455806e-01
4.52206433e-01 7.67220676e-01 -5.15600979e-01 7.32012928e-01
-5.67123532e-01 7.61890590e-01 7.64275789e-02 -2.29829758e-01
1.78297803e-01 -1.27080768e-01 3.09975326e-01 8.13390791e-01
2.40038961e-01 4.06179309e-01 1.26490176e-01 6.06260717e-01
-4.60070342e-01 3.04041266e-01 8.97375792e-02 9.89120454e-02
6.38208151e-01 9.31196272e-01 -2.65148073e-01 -4.88456458e-01
-3.77406597e-01 9.67536449e-01 6.59162700e-01 1.13677263e-01
-5.71182191e-01 -4.54384327e-01 -6.99236318e-02 -5.99889532e-02
4.30232555e-01 -1.02007333e-02 -4.57280934e-01 -1.30516410e+00
3.52834523e-01 -1.28452218e+00 4.64404285e-01 -6.84509218e-01
-8.98771822e-01 4.43452775e-01 -2.44362116e-01 -1.13548076e+00
-1.86691627e-01 -2.74830293e-02 -7.49886155e-01 9.36925888e-01
-1.53031492e+00 -1.01840079e+00 -3.44914436e-01 3.23706895e-01
1.04654539e+00 -4.71967459e-02 6.97480440e-01 -9.19714794e-02
-4.17834103e-01 7.49997139e-01 3.19684565e-01 2.77316403e-02
5.04928887e-01 -1.45324612e+00 2.41322234e-01 8.04467380e-01
1.32652953e-01 4.85971361e-01 5.88102520e-01 -5.80252409e-01
-1.29737306e+00 -9.28310812e-01 1.44389927e+00 -3.67587030e-01
4.40126896e-01 -2.23505378e-01 -1.17074943e+00 5.21526933e-01
5.94660878e-01 -6.54997945e-01 6.15084112e-01 2.59592757e-02
-3.12791407e-01 -1.84516206e-01 -5.40230036e-01 4.68763322e-01
6.64037287e-01 -6.34706855e-01 -1.16020918e+00 2.13512972e-01
9.41498041e-01 -6.97761714e-01 -9.14272726e-01 2.81099617e-01
4.49556828e-01 -1.01649463e+00 4.41139072e-01 -2.95165211e-01
1.05417204e+00 -1.49456695e-01 2.95583494e-02 -1.46458375e+00
-1.30084753e-01 -6.07739866e-01 -4.48289931e-01 1.14774418e+00
3.51661593e-01 -4.07577097e-01 6.45817399e-01 2.97091663e-01
-3.15583110e-01 -1.09864938e+00 -9.61346865e-01 -5.43042958e-01
5.39248884e-01 -9.24417526e-02 8.11930776e-01 3.21306974e-01
3.89848918e-01 6.21585190e-01 -3.70195240e-01 -2.04880029e-01
4.21239942e-01 7.24414229e-01 8.06005120e-01 -9.02647495e-01
-3.55125666e-01 -3.80523115e-01 3.36077303e-01 -1.99553585e+00
1.65514156e-01 -9.55195725e-01 3.09449643e-01 -1.83163130e+00
6.17953837e-01 -4.26886797e-01 -1.76015079e-01 3.33199412e-01
-5.06468832e-01 -1.09029151e-01 2.72480905e-01 5.93026578e-01
-1.03726053e+00 8.32548022e-01 1.60257018e+00 -3.48705292e-01
-6.08538929e-03 7.63623044e-02 -6.99154437e-01 1.85911343e-01
8.17680478e-01 -1.95639655e-01 -4.35593635e-01 -8.11681926e-01
3.44649017e-01 4.39076900e-01 9.16270465e-02 -6.89656794e-01
4.68049705e-01 1.25714568e-02 6.58364445e-02 -1.10424662e+00
-3.90575230e-02 -1.60214722e-01 -3.02149147e-01 4.78586048e-01
-8.82702887e-01 6.32211268e-02 -3.61280173e-01 2.82576770e-01
-5.71808100e-01 -4.18269724e-01 5.19528806e-01 7.43399337e-02
-4.30170238e-01 3.44304889e-02 -2.91730523e-01 5.02747297e-01
7.84149647e-01 -8.35949183e-02 -7.27450788e-01 -5.53187490e-01
-2.31452197e-01 6.87964499e-01 3.06842566e-01 4.95348990e-01
7.37398863e-01 -9.12051857e-01 -1.24340034e+00 -6.38892427e-02
1.79554284e-01 5.65636814e-01 4.86318678e-01 7.62309313e-01
-3.21749449e-01 7.63481319e-01 6.38379678e-02 -4.11107123e-01
-1.22655594e+00 3.40448529e-01 -1.48736268e-01 -8.90286684e-01
-8.11499894e-01 7.77509153e-01 1.66662261e-01 -2.25164980e-01
2.05154091e-01 -4.47493225e-01 -4.50280994e-01 2.23494574e-01
9.39694226e-01 4.08524960e-01 1.85393646e-01 -5.34955084e-01
-4.53082584e-02 4.04910386e-01 -5.29926896e-01 2.12166011e-02
1.08063686e+00 -4.81247723e-01 -5.87067842e-01 2.87759334e-01
1.01026177e+00 2.45741278e-01 -9.17499721e-01 -4.65635806e-01
5.90521619e-02 -2.26017684e-01 -1.47189185e-01 -8.49034786e-01
-7.39111781e-01 9.00050163e-01 -2.18264714e-01 1.75634027e-01
1.21747279e+00 3.28781337e-01 1.45127523e+00 4.76555914e-01
3.29804063e-01 -9.47165668e-01 1.80200294e-01 1.04417086e+00
9.97246325e-01 -8.71478021e-01 -1.29853293e-01 -3.68591771e-02
-6.01406395e-01 1.02308023e+00 4.93712962e-01 2.43093327e-01
7.99921677e-02 -1.40600875e-01 -2.63293181e-03 1.10634491e-02
-9.66154933e-01 7.69719779e-02 5.06677926e-01 2.08930731e-01
2.51316071e-01 2.36931190e-01 -4.69245851e-01 8.06214154e-01
-8.26824605e-01 6.02013133e-02 6.20363891e-01 7.82618999e-01
-7.88788915e-01 -1.01216161e+00 -1.41719773e-01 5.81602037e-01
-3.58343005e-01 -3.46388519e-01 -2.02517480e-01 1.23088628e-01
-7.55594149e-02 1.17378163e+00 2.29984764e-02 -3.37152123e-01
5.01799211e-03 2.66134832e-03 3.25026363e-01 -6.67816281e-01
-7.27086902e-01 -4.11603391e-01 -7.78058618e-02 -3.50265920e-01
-1.94306187e-02 -5.62285662e-01 -1.29514527e+00 -3.69825244e-01
-3.97475302e-01 7.06938982e-01 8.16005692e-02 1.19034398e+00
1.93684474e-01 4.68691081e-01 6.16088450e-01 -2.37540737e-01
-9.43740308e-01 -1.24931955e+00 -4.33091134e-01 2.37759635e-01
5.34489751e-01 -6.21281825e-02 -2.38467768e-01 5.90792298e-02] | [11.437163352966309, 8.114126205444336] |
0cb0638c-a072-43fa-9882-2da810a2e27d | pst-plant-segmentation-transformer-enhanced | 2206.13082 | null | https://arxiv.org/abs/2206.13082v2 | https://arxiv.org/pdf/2206.13082v2.pdf | PST: Plant Segmentation Transformer for 3D Point Clouds of rapeseed plants at the podding stage | Segmentation of plant point clouds to obtain high-precise morphological traits is essential for plant phenotyping. Although the fast development of deep learning has boosted much research on segmentation of plant point clouds, previous studies mainly focus on the hard voxelization-based or down-sampling-based methods, which are limited to segmenting simple plant organs. Segmentation of complex plant point clouds with a high spatial resolution still remains challenging. In this study, we proposed a deep learning network plant segmentation transformer (PST) to achieve the semantic and instance segmentation of rapeseed plants point clouds acquired by handheld laser scanning (HLS) with the high spatial resolution, which can characterize the tiny siliques as the main traits targeted. PST is composed of: (i) a dynamic voxel feature encoder (DVFE) to aggregate the point features with the raw spatial resolution; (ii) the dual window sets attention blocks to capture the contextual information; and (iii) a dense feature propagation module to obtain the final dense point feature map. The results proved that PST and PST-PointGroup (PG) achieved superior performance in semantic and instance segmentation tasks. For the semantic segmentation, the mean IoU, mean Precision, mean Recall, mean F1-score, and overall accuracy of PST were 93.96%, 97.29%, 96.52%, 96.88%, and 97.07%, achieving an improvement of 7.62%, 3.28%, 4.8%, 4.25%, and 3.88% compared to the second-best state-of-the-art network PAConv. For instance segmentation, PST-PG reached 89.51%, 89.85%, 88.83% and 82.53% in mCov, mWCov, mPerc90, and mRec90, achieving an improvement of 2.93%, 2.21%, 1.99%, and 5.9% compared to the original PG. This study proves that the deep-learning-based point cloud segmentation method has a great potential for resolving dense plant point clouds with complex morphological traits. | ['Haiyan Cen', 'Yong He', 'Pengyao Xie', 'Zhihong Ma', 'Ruiming Du'] | 2022-06-27 | null | null | null | null | ['plant-phenotyping', 'point-cloud-segmentation'] | ['computer-vision', 'computer-vision'] | [-7.13900179e-02 -1.60526589e-01 2.06152108e-02 -1.32188722e-01
-4.69945103e-01 -5.61265111e-01 5.78731187e-02 2.59762347e-01
1.94427408e-02 3.35033745e-01 -9.67272639e-01 -2.62213469e-01
-3.27208579e-01 -1.32358849e+00 -6.76528335e-01 -8.54757428e-01
-6.40399754e-02 8.19313228e-01 4.03074682e-01 -1.68035865e-01
1.10345386e-01 8.56729865e-01 -1.60661173e+00 1.21966936e-01
1.29373729e+00 1.43991888e+00 8.06017160e-01 2.69241065e-01
-5.80624104e-01 -1.36542484e-01 -5.50294936e-01 7.75963813e-02
1.82844207e-01 3.55139971e-02 -5.99563420e-01 2.35348463e-01
-1.45338058e-01 -2.02288777e-01 4.29642975e-01 1.02310836e+00
2.56924033e-01 -1.72270730e-01 6.07502282e-01 -1.25481105e+00
-8.14663053e-01 6.29319966e-01 -1.32517612e+00 -5.09077907e-01
-3.23977798e-01 1.95109993e-01 6.72063112e-01 -7.46415496e-01
3.06405663e-01 9.36228633e-01 7.72492468e-01 7.04445019e-02
-1.11690795e+00 -7.05142438e-01 5.22293150e-02 6.75634667e-02
-1.60543728e+00 3.62484366e-01 4.05454546e-01 -6.32449567e-01
6.06612921e-01 4.50538903e-01 9.77510512e-01 1.92892045e-01
1.41468689e-01 7.32328236e-01 7.25059807e-01 2.64048623e-03
8.38595107e-02 -9.96703953e-02 7.59569257e-02 5.77702403e-01
3.19429874e-01 2.53548287e-02 4.34835076e-01 2.31582984e-01
1.13794374e+00 9.80478302e-02 -1.65326416e-01 -2.42507756e-01
-8.59611273e-01 7.83875108e-01 1.07937372e+00 5.29631615e-01
-7.58614659e-01 -3.63344729e-01 3.22759062e-01 -3.36666286e-01
5.05765498e-01 6.36377275e-01 -9.88676786e-01 2.28601515e-01
-1.03976345e+00 2.22215161e-01 6.16517842e-01 1.22601533e+00
7.41845310e-01 1.66276693e-01 -2.26520881e-01 8.54754746e-01
3.59006613e-01 8.95205438e-01 -5.95282950e-03 -9.62452471e-01
-6.45624995e-02 9.57941651e-01 1.04035102e-01 -1.18300843e+00
-5.12957036e-01 -5.74959040e-01 -1.17742765e+00 7.84854814e-02
1.12281553e-01 1.79225072e-01 -1.18483877e+00 1.38501096e+00
3.30318063e-01 6.04608767e-02 -1.70710221e-01 8.98058474e-01
1.14018357e+00 1.11804211e+00 4.32211220e-01 -1.44940332e-01
1.39873230e+00 -4.13223624e-01 -3.83485585e-01 2.96640992e-02
5.13169110e-01 -7.76971400e-01 1.01498294e+00 2.94869781e-01
-1.12515652e+00 -7.83265531e-01 -8.13807368e-01 2.75726795e-01
-4.31231350e-01 6.82848811e-01 8.49859178e-01 3.91661674e-01
-7.85894632e-01 7.22445130e-01 -8.22256207e-01 -2.99857944e-01
1.07839561e+00 4.51660216e-01 -9.68239158e-02 8.71971250e-02
-6.84980750e-01 4.85906690e-01 5.82036316e-01 3.86943817e-01
-8.44206393e-01 -1.17136419e+00 -3.54963720e-01 5.07259250e-01
4.02216136e-01 -4.05427456e-01 9.06112492e-01 -6.23152435e-01
-1.64227760e+00 9.71003771e-01 9.29280147e-02 -2.21067756e-01
1.35858469e-02 -1.31849468e-01 -6.06831275e-02 1.61061034e-01
3.10443997e-01 9.46528316e-01 2.29021475e-01 -1.36748374e+00
-8.40411961e-01 -8.48038018e-01 -2.42878959e-01 7.06829801e-02
-7.59629011e-02 -2.36294791e-01 -3.86594892e-01 -1.45922620e-02
6.30585909e-01 -8.71780932e-01 -2.80765533e-01 1.66336685e-01
-5.36464810e-01 -1.27747491e-01 1.10136962e+00 -9.20664310e-01
5.76008976e-01 -1.90837586e+00 1.82727769e-01 8.92980099e-02
1.50255993e-01 8.57695162e-01 -1.68057814e-01 1.36153745e-02
-1.13245949e-01 3.35283548e-01 -5.52912056e-01 2.08855852e-01
-4.03673649e-01 8.53631869e-02 2.97628045e-02 1.93276450e-01
3.98525357e-01 9.87925112e-01 -7.22736418e-01 -3.62552702e-01
7.60436952e-01 4.22322124e-01 -2.95283198e-01 1.87884212e-01
-3.21272105e-01 4.91438866e-01 -6.66600943e-01 1.10183704e+00
1.37172019e+00 -1.57884225e-01 -2.40557685e-01 -3.05829227e-01
-6.09135866e-01 -5.77094615e-01 -1.10594332e+00 1.52724469e+00
-2.69239575e-01 1.71370894e-01 4.37346309e-01 -1.13046002e+00
1.40311289e+00 1.63966328e-01 9.44364548e-01 -4.60140705e-01
2.56832778e-01 3.61692697e-01 -6.63269311e-02 -2.62692571e-01
1.65834576e-01 9.49907750e-02 1.47377819e-01 -5.89711666e-01
4.25027423e-02 -5.93792379e-01 -6.46964973e-03 -4.40194644e-02
5.94718277e-01 1.37272969e-01 9.15024728e-02 -5.57119846e-01
7.49609590e-01 3.06425482e-01 6.20977581e-01 3.76154631e-01
-1.46092772e-01 5.27525783e-01 4.52099860e-01 -2.05878377e-01
-8.73491645e-01 -1.00439405e+00 -4.89335358e-01 5.94092906e-01
3.45993072e-01 -8.70281681e-02 -8.43019724e-01 -1.57764599e-01
2.37796888e-01 7.08987236e-01 -1.25447005e-01 -2.54741132e-01
-1.47453502e-01 -8.25195670e-01 3.13493550e-01 6.21169746e-01
1.19191813e+00 -1.15328181e+00 -5.78370452e-01 1.22681230e-01
-1.53066710e-01 -1.17312336e+00 4.08504844e-01 4.49941941e-02
-1.03131938e+00 -9.63564992e-01 -7.26007760e-01 -9.00855720e-01
3.80355239e-01 3.60073000e-01 9.74918127e-01 1.58116832e-01
-4.28520888e-01 -3.77372682e-01 -4.92433250e-01 -5.65487564e-01
-3.37690823e-02 4.44367498e-01 -4.72557127e-01 -2.59786814e-01
2.67410308e-01 -5.50710440e-01 -4.10069227e-01 3.59154552e-01
-7.57516801e-01 1.99373141e-01 6.21535361e-01 8.13075900e-01
1.11522424e+00 2.19212890e-01 2.07026407e-01 -7.63067186e-01
7.90604332e-04 -3.54263425e-01 -1.20813501e+00 4.34309505e-02
-5.17200232e-01 -7.51373231e-01 6.39214873e-01 -3.73823009e-02
-8.17665935e-01 2.91590631e-01 -2.68562436e-01 -3.58336508e-01
-5.94644070e-01 5.56246817e-01 -6.28973126e-01 -1.22681089e-01
4.97726917e-01 9.34366286e-02 -1.09408133e-01 -4.03727084e-01
3.27564299e-01 6.35793269e-01 4.36147898e-01 -4.26519960e-01
8.31983745e-01 3.74315262e-01 1.85053378e-01 -1.14441395e+00
-6.02115095e-01 -5.01023591e-01 -8.28651667e-01 -9.79500189e-02
1.02501130e+00 -7.90447652e-01 -1.13485026e+00 8.70101392e-01
-1.19171226e+00 -2.25817233e-01 -5.04757702e-01 2.55736500e-01
-3.94084156e-01 3.03710938e-01 -5.52996039e-01 -5.99926651e-01
-7.46627390e-01 -1.28758907e+00 1.36927319e+00 7.22916305e-01
2.44744048e-01 -4.19600278e-01 -6.41626656e-01 1.59522861e-01
4.92209673e-01 5.19877136e-01 8.71286392e-01 -3.13916981e-01
-7.57908523e-01 -4.12820429e-01 -9.29877579e-01 3.55961382e-01
1.51309371e-01 6.16080642e-01 -9.06862438e-01 -1.45801194e-02
-1.44084364e-01 -1.20116882e-01 3.81528854e-01 1.10775375e+00
1.75987148e+00 4.33611184e-01 -5.81775308e-01 1.05408204e+00
1.64865005e+00 6.16660833e-01 7.63437152e-01 7.59510025e-02
9.80885327e-01 5.42294145e-01 9.11686778e-01 5.06777108e-01
-5.75398607e-03 4.06475276e-01 1.03296292e+00 -3.24029326e-01
2.77763397e-01 1.65542752e-01 -4.55091298e-01 5.26024103e-01
-2.46656463e-01 -1.63654104e-01 -1.18890750e+00 6.98143601e-01
-1.60596359e+00 -9.05633509e-01 -7.12806404e-01 1.86556244e+00
4.67970818e-01 -1.14249706e-01 -2.40924999e-01 2.87555426e-01
8.56173933e-01 5.82685843e-02 -7.47613370e-01 -1.09667085e-01
-1.99233219e-01 4.12708879e-01 7.28948712e-01 9.45816934e-02
-1.30573404e+00 1.45195198e+00 4.12387276e+00 9.19103146e-01
-1.06093180e+00 -1.07024558e-01 6.65203094e-01 4.29638386e-01
2.47404233e-01 5.61729111e-02 -8.26521516e-01 4.12332177e-01
6.65315092e-01 2.47647911e-01 2.83115000e-01 1.00830007e+00
2.23492458e-01 -9.35989320e-02 -3.91321272e-01 9.44442630e-01
-6.25274360e-01 -1.17786479e+00 -8.37468654e-02 1.68004960e-01
6.90102875e-01 7.93609023e-02 5.82163967e-02 3.02890509e-01
3.67997259e-01 -1.04947126e+00 3.33741635e-01 3.24481398e-01
8.96802783e-01 -8.57004821e-01 9.77468789e-01 5.84039807e-01
-1.55823684e+00 -7.03496337e-02 -8.54221642e-01 3.26724291e-01
-4.86131310e-02 1.08040380e+00 -6.09744370e-01 9.82850969e-01
9.39859509e-01 7.60135114e-01 -4.20219928e-01 1.10496354e+00
-1.25655875e-01 6.58758461e-01 -6.61785543e-01 1.93749238e-02
4.85015303e-01 -4.16467726e-01 2.85849273e-01 8.90141547e-01
4.85315412e-01 1.65130734e-01 4.82048899e-01 1.30551481e+00
2.12497756e-01 1.81989357e-01 -1.49936691e-01 -1.10565819e-01
5.55316210e-01 1.76922417e+00 -1.04903758e+00 -1.35361344e-01
1.45764485e-01 7.11891174e-01 -6.20092899e-02 -9.92944986e-02
-1.09606218e+00 -6.00580692e-01 4.61842805e-01 5.24898805e-02
5.34418046e-01 -1.74994290e-01 -6.15382195e-01 -6.41345024e-01
-2.62842715e-01 -5.14751494e-01 -3.21251191e-02 -1.08229733e+00
-1.00405157e+00 4.37207878e-01 1.30344159e-03 -8.09794664e-01
3.81531835e-01 -7.88610578e-01 -4.73725647e-01 1.42602777e+00
-1.34333229e+00 -1.37193942e+00 -1.05159020e+00 1.15074605e-01
5.91880023e-01 -1.17012680e-01 9.87230301e-01 1.23595119e-01
-7.34930456e-01 9.93315056e-02 2.97822475e-01 -1.21904157e-01
-6.41363710e-02 -1.09174013e+00 3.60547483e-01 6.02881134e-01
-3.49550903e-01 4.34176810e-02 1.19881727e-01 -6.27497733e-01
-1.29087734e+00 -1.37089467e+00 5.06362915e-01 -4.38861735e-03
1.65597916e-01 9.95421875e-03 -1.25815475e+00 3.77571315e-01
-3.44063312e-01 7.19853491e-02 1.47261858e-01 -7.84676448e-02
2.44712487e-01 -2.47116610e-01 -1.48805928e+00 4.04163450e-01
8.96139801e-01 1.22896908e-02 2.10452452e-01 4.67446506e-01
7.95707881e-01 -6.04873478e-01 -1.21980989e+00 8.61492157e-01
4.80357021e-01 -6.31102443e-01 1.08704841e+00 -1.95325553e-01
5.51126838e-01 -3.50909323e-01 -2.77849615e-01 -1.25771630e+00
-9.93386269e-01 1.19864918e-01 4.30291086e-01 1.43239415e+00
1.70133322e-01 -4.47823614e-01 8.10517132e-01 2.01197714e-01
-5.27517796e-01 -8.12092841e-01 -5.30509770e-01 -6.72568858e-01
2.06754655e-01 -2.94686764e-01 1.19547272e+00 9.20800686e-01
-9.41521943e-01 3.17977578e-03 3.36693585e-01 5.42844534e-01
3.74672323e-01 3.39644581e-01 6.25376403e-01 -1.91452265e+00
2.36121789e-01 -6.57695234e-01 -3.13713461e-01 -7.85310745e-01
4.33275104e-02 -9.19694424e-01 -1.45673752e-02 -1.88711965e+00
-1.04154870e-01 -7.04708219e-01 -3.62687409e-02 4.72884923e-01
-9.77803245e-02 -2.50039399e-01 1.45906672e-01 2.03564316e-01
1.96232185e-01 5.66117048e-01 1.51529324e+00 -2.15118036e-01
-4.08791304e-01 4.16591436e-01 -6.71996295e-01 7.64026344e-01
1.31818295e+00 -2.14765862e-01 -4.14355785e-01 -6.32793844e-01
-3.76441985e-01 4.38337363e-02 5.25717795e-01 -1.11558771e+00
-1.61835358e-01 -3.86333436e-01 7.34395325e-01 -1.21887183e+00
3.61153722e-01 -8.77498627e-01 2.27156460e-01 6.15465283e-01
1.00308239e-01 -3.31257582e-01 5.54740369e-01 1.18413761e-01
1.10371694e-01 -3.22292715e-01 9.66502368e-01 -2.52767086e-01
-8.56560111e-01 7.37079263e-01 2.35147551e-02 -2.97781169e-01
1.22348714e+00 -3.60167116e-01 -4.82777730e-02 1.86532974e-01
-4.87074703e-01 3.51088792e-01 1.19549677e-01 2.63484687e-01
6.12065971e-01 -1.03359902e+00 -7.52449811e-01 3.18503857e-01
-7.75295720e-02 6.52625740e-01 3.60313654e-01 6.08577371e-01
-9.87557650e-01 5.44143200e-01 -4.85733420e-01 -1.18755317e+00
-9.92040932e-01 3.79613042e-01 3.02332014e-01 -1.32803425e-01
-4.38066542e-01 1.00107455e+00 2.29264781e-01 -6.47809744e-01
-2.24153429e-01 -7.13350713e-01 -4.74043071e-01 5.06874621e-02
-1.12110004e-01 3.72936964e-01 2.24737108e-01 -7.14972377e-01
-3.49488169e-01 8.01196277e-01 2.75123566e-01 4.67438161e-01
1.46283996e+00 3.69974762e-01 -2.33435243e-01 3.71475637e-01
8.70164692e-01 -5.44008970e-01 -1.01661289e+00 2.23999731e-02
-5.44159524e-02 -5.40209889e-01 2.36837491e-01 -9.76413488e-01
-1.46569288e+00 9.66112137e-01 8.78959477e-01 6.93800986e-01
1.33985400e+00 -7.20841140e-02 8.57063353e-01 -1.90367941e-02
4.88830984e-01 -6.31688774e-01 -7.10984886e-01 6.34261906e-01
9.03316975e-01 -1.12270558e+00 2.74136141e-02 -9.50516701e-01
-3.37692946e-01 9.96270716e-01 9.18472409e-01 -2.07996011e-01
7.56298840e-01 2.00744525e-01 -3.59753847e-01 -3.31769079e-01
-2.88441926e-01 -2.66128600e-01 1.29868254e-01 8.86463344e-01
4.27184701e-01 5.21193624e-01 -1.25283211e-01 7.50320494e-01
-2.59719610e-01 9.71335080e-03 -9.85916406e-02 6.65715694e-01
-6.92229569e-01 -5.60830295e-01 -3.98212224e-01 5.90138137e-01
-4.45608124e-02 1.14209861e-01 -2.41605803e-01 1.00783098e+00
3.15066993e-01 6.73722088e-01 2.15388656e-01 -4.01246458e-01
6.49209499e-01 -2.46061295e-01 2.61310279e-01 -7.40400076e-01
-6.03237927e-01 1.64522007e-01 -5.13193488e-01 -4.62026089e-01
-9.53269526e-02 -5.64270675e-01 -1.47965980e+00 -4.70513761e-01
-7.20087230e-01 -2.52343416e-02 9.04066920e-01 6.62421227e-01
4.89830434e-01 9.43559885e-01 5.85045457e-01 -9.95371699e-01
-2.77047902e-01 -9.55797195e-01 -8.26409101e-01 5.23237213e-02
-3.66020232e-01 -6.74119413e-01 -3.65884937e-02 -1.70899227e-01] | [9.096100807189941, -1.6478341817855835] |
9bcc8580-d256-447b-9a76-d91f59a4073f | manipulating-the-distributions-of-experience | 2006.00283 | null | https://arxiv.org/abs/2006.00283v1 | https://arxiv.org/pdf/2006.00283v1.pdf | Manipulating the Distributions of Experience used for Self-Play Learning in Expert Iteration | Expert Iteration (ExIt) is an effective framework for learning game-playing policies from self-play. ExIt involves training a policy to mimic the search behaviour of a tree search algorithm - such as Monte-Carlo tree search - and using the trained policy to guide it. The policy and the tree search can then iteratively improve each other, through experience gathered in self-play between instances of the guided tree search algorithm. This paper outlines three different approaches for manipulating the distribution of data collected from self-play, and the procedure that samples batches for learning updates from the collected data. Firstly, samples in batches are weighted based on the durations of the episodes in which they were originally experienced. Secondly, Prioritized Experience Replay is applied within the ExIt framework, to prioritise sampling experience from which we expect to obtain valuable training signals. Thirdly, a trained exploratory policy is used to diversify the trajectories experienced in self-play. This paper summarises the effects of these manipulations on training performance evaluated in fourteen different board games. We find major improvements in early training performance in some games, and minor improvements averaged over fourteen games. | ['Éric Piette', 'Dennis J. N. J. Soemers', 'Cameron Browne', 'Matthew Stephenson'] | 2020-05-30 | null | null | null | null | ['board-games'] | ['playing-games'] | [ 6.70615062e-02 1.22770742e-02 -3.01714689e-01 -1.72172280e-04
-5.90018392e-01 -7.31123030e-01 3.82705480e-01 1.07958829e-02
-8.03361893e-01 8.53951752e-01 2.34134406e-01 -5.10629475e-01
-3.64542693e-01 -7.67354429e-01 -5.49379766e-01 -6.97600365e-01
-2.45679960e-01 5.29991925e-01 4.58645612e-01 -3.47282946e-01
6.26768410e-01 2.70698965e-01 -1.85090065e+00 2.92469531e-01
9.89140809e-01 5.60925126e-01 4.21786517e-01 1.12895799e+00
1.33091090e-02 1.04422200e+00 -7.01700687e-01 -1.16715893e-01
4.37103003e-01 -9.38940346e-01 -6.71295166e-01 2.46925607e-01
-2.72496969e-01 -2.77103961e-01 -2.80378282e-01 7.20475972e-01
6.72110081e-01 8.60875547e-01 3.10700595e-01 -6.74479902e-01
3.45574200e-01 6.05889440e-01 -2.99874723e-01 6.80933356e-01
6.18394017e-01 5.81634641e-01 7.91126609e-01 -1.65636525e-01
5.52730918e-01 8.59396458e-01 5.58026671e-01 7.04601228e-01
-1.12859118e+00 -6.31480217e-01 1.01398721e-01 -1.34502649e-02
-8.26789558e-01 -3.99864465e-01 4.87232059e-01 -4.68791544e-01
8.04011643e-01 2.10703298e-01 1.27624953e+00 1.01083434e+00
1.24291833e-02 9.65481937e-01 1.17698014e+00 -4.42213804e-01
8.99363399e-01 3.13066155e-01 -3.02474648e-01 4.85386670e-01
-2.73074001e-01 9.66253459e-01 -9.37321007e-01 -1.28037900e-01
8.93835425e-01 -5.05998969e-01 -1.03134148e-01 -4.01881963e-01
-4.96373922e-01 7.36753762e-01 1.03439264e-01 1.04105221e-02
-6.97835803e-01 1.43000275e-01 6.32119596e-01 6.12690866e-01
4.48191434e-01 8.89993966e-01 -4.45370018e-01 -1.12603831e+00
-1.18166292e+00 6.22002721e-01 9.30653155e-01 5.03413796e-01
6.05246961e-01 1.89747736e-01 -3.87842149e-01 6.43320918e-01
7.05519095e-02 -6.86499700e-02 7.09089339e-01 -1.16430032e+00
3.20573360e-01 3.56237143e-01 4.80534077e-01 -5.00828028e-01
-1.62503853e-01 -5.73582470e-01 8.25132281e-02 5.56616187e-01
5.65333068e-01 -6.13308430e-01 -7.82822907e-01 1.34032512e+00
4.77140278e-01 3.29444557e-01 -1.26570329e-01 7.23618209e-01
2.84806877e-01 5.32753348e-01 1.16638012e-01 -4.16386843e-01
8.23772132e-01 -6.87806427e-01 -3.05347115e-01 -3.49357277e-01
7.06511021e-01 -2.47780561e-01 1.05626893e+00 7.32983530e-01
-1.50561166e+00 -7.02098012e-01 -8.11757982e-01 7.93947637e-01
1.08854398e-01 -1.91818908e-01 6.42467856e-01 8.34701061e-01
-9.10213053e-01 1.27268875e+00 -1.15222859e+00 -1.74443766e-01
3.62971127e-01 3.44020933e-01 3.53434086e-01 2.94527471e-01
-9.88261700e-01 7.24596560e-01 6.12991393e-01 -2.35087827e-01
-1.23537767e+00 -7.89199829e-01 -5.14092088e-01 2.01737881e-03
6.98111117e-01 -3.60330641e-01 1.91200721e+00 -8.65449250e-01
-1.92414737e+00 7.38028169e-01 6.15024380e-02 -5.65753937e-01
6.30294144e-01 -4.49354351e-01 1.56128583e-02 -1.13670878e-01
1.39023051e-01 3.22155386e-01 6.81434333e-01 -1.08631349e+00
-1.11278033e+00 -1.27051458e-01 1.22489274e-01 9.05813813e-01
1.66798592e-01 2.02218294e-02 -5.72862208e-01 -2.23665670e-01
-1.09879442e-01 -8.80031466e-01 -5.91201365e-01 -9.20432627e-01
5.94706610e-02 -2.34882399e-01 1.39461905e-01 -3.58584374e-01
1.65271783e+00 -2.13082719e+00 1.49100885e-01 4.54577982e-01
-6.72852546e-02 2.44017676e-01 1.92824736e-01 5.57371259e-01
7.93453455e-02 -3.22912484e-01 2.46433616e-01 -1.02438085e-01
-1.73803940e-01 3.11435223e-01 -3.10507119e-01 5.66049553e-02
-3.49087983e-01 5.91940761e-01 -1.42100906e+00 -9.62359235e-02
2.11757794e-01 -4.59802628e-01 -9.47820961e-01 4.79332119e-01
-5.11957586e-01 7.97943890e-01 -5.70375085e-01 1.71656057e-01
-2.12784819e-02 1.42853767e-01 1.20909370e-01 4.67951268e-01
-3.19329739e-01 6.00901186e-01 -1.11652815e+00 1.53873670e+00
-5.82915425e-01 4.91082877e-01 -1.45754203e-01 -9.49021757e-01
9.17066455e-01 2.06612512e-01 5.71433246e-01 -7.40869641e-01
2.00377598e-01 -1.25805130e-02 2.17502639e-01 -6.74229622e-01
6.16894722e-01 -3.11898254e-02 1.13648064e-01 6.99711859e-01
3.45047414e-02 -4.36004341e-01 7.01389015e-01 1.69606969e-01
1.23196113e+00 6.17660999e-01 2.82343566e-01 4.10375558e-02
4.87873517e-02 3.30127925e-01 4.62234735e-01 1.35242355e+00
-2.14401498e-01 4.27749194e-02 5.36096156e-01 -4.98439163e-01
-8.75885963e-01 -1.00134122e+00 3.42119217e-01 1.70262694e+00
-1.71194896e-01 -6.19024813e-01 -6.70308650e-01 -6.43817842e-01
-1.74728423e-01 8.86356175e-01 -4.85861123e-01 -3.67729217e-01
-4.44169074e-01 -4.79904234e-01 2.23025173e-01 4.44341600e-01
6.42325401e-01 -1.65691245e+00 -1.15747654e+00 6.47980988e-01
1.64331958e-01 -3.66639555e-01 -2.60753691e-01 6.20965123e-01
-1.23821235e+00 -1.07078958e+00 -3.14046443e-01 -4.23242569e-01
4.40879285e-01 -2.69919246e-01 1.31872451e+00 9.14271474e-02
-2.01679721e-01 6.31351471e-01 -4.79397237e-01 -3.88173312e-01
-5.63677132e-01 1.83151320e-01 2.32258644e-02 -3.39469731e-01
4.79575932e-01 -7.67194211e-01 -6.89690530e-01 2.57489443e-01
-3.93913895e-01 -5.63029721e-02 4.44043785e-01 8.10936511e-01
4.18923587e-01 4.22011107e-01 2.91135162e-01 -9.61721718e-01
1.16862679e+00 -5.18476665e-01 -6.22614861e-01 -1.45229265e-01
-6.41420484e-01 6.83959350e-02 3.56150568e-01 -6.81360722e-01
-1.24684238e+00 -4.37269807e-02 1.51159037e-02 -2.09368154e-01
-2.04568699e-01 5.60684144e-01 4.17697877e-01 2.97104508e-01
1.29654336e+00 3.75156492e-01 -6.05710819e-02 -1.96313888e-01
3.25830400e-01 4.42237228e-01 5.12298346e-01 -1.02957880e+00
4.29973751e-01 -8.82923882e-03 -5.87602019e-01 -5.66151738e-01
-8.34156692e-01 -5.83529651e-01 -2.50055373e-01 -7.38532186e-01
6.09396935e-01 -7.06782460e-01 -8.07108104e-01 5.22235572e-01
-4.08288747e-01 -1.18426144e+00 -7.93352783e-01 5.54022849e-01
-8.88033450e-01 -2.63252079e-01 -5.00774682e-01 -1.23259759e+00
9.11814719e-02 -8.85366321e-01 4.67626989e-01 8.55607033e-01
-5.78369558e-01 -1.03941810e+00 5.31099021e-01 2.05247611e-01
1.46866322e-01 -1.95536107e-01 1.84905455e-01 -6.61075473e-01
-2.69826591e-01 -2.06336126e-01 5.21002650e-01 1.36919618e-01
1.47425070e-01 -1.52030796e-01 -7.53582120e-01 -4.52630788e-01
3.48097831e-02 -5.22243798e-01 5.39344788e-01 6.00977242e-01
8.66643608e-01 1.05866999e-03 -2.92898118e-01 4.76608068e-01
9.55774009e-01 6.79636002e-01 5.64729035e-01 6.45717084e-01
1.16340816e-01 4.38146353e-01 9.12291884e-01 7.76828885e-01
-2.23771125e-01 3.92665118e-01 5.91450222e-02 2.80595422e-01
4.10378247e-01 -5.66999137e-01 4.08348411e-01 3.77577543e-01
-4.55979079e-01 1.27543330e-01 -6.52531683e-01 6.48659229e-01
-1.81109238e+00 -1.31603706e+00 4.17916864e-01 2.23916006e+00
1.13391626e+00 7.42739797e-01 7.55597413e-01 5.85447736e-02
4.13719207e-01 2.40159985e-02 -9.27163303e-01 -6.76481366e-01
6.54204547e-01 6.72928274e-01 5.30435205e-01 6.24035537e-01
-6.93438530e-01 1.23875630e+00 6.87454224e+00 8.28267753e-01
-7.41014063e-01 -2.64300019e-01 4.32495415e-01 -6.19032323e-01
1.43415136e-02 1.58481598e-01 -7.32756078e-01 5.79308212e-01
9.32265699e-01 -3.99279654e-01 8.54616523e-01 9.29000318e-01
5.81859887e-01 -8.43858719e-01 -8.30599308e-01 5.45447826e-01
-5.99156439e-01 -1.13573873e+00 -5.41105866e-01 1.01666465e-01
9.68889713e-01 -1.03032485e-01 1.26370579e-01 8.33712697e-01
1.12207782e+00 -7.60227680e-01 4.82488483e-01 5.29545724e-01
3.95310134e-01 -1.03959644e+00 2.25693494e-01 7.57430613e-01
-9.15729523e-01 -3.61652851e-01 -2.27486745e-01 -5.72984338e-01
2.96532493e-02 3.97127122e-01 -9.14053857e-01 2.76164144e-01
7.22957075e-01 3.44754040e-01 -2.48270825e-01 1.36480319e+00
-4.42047268e-01 1.21345675e+00 -2.50831455e-01 -2.28523985e-01
4.14416462e-01 -4.69745159e-01 6.45385802e-01 8.45745385e-01
9.71098319e-02 2.35699996e-01 2.90098250e-01 6.61726117e-01
5.71089923e-01 -1.80038616e-01 -5.57173848e-01 -4.87661771e-02
6.39690578e-01 9.66029167e-01 -8.26531291e-01 -3.39461744e-01
9.02641192e-02 9.02988434e-01 1.81944102e-01 4.26566660e-01
-5.67249954e-01 -2.78122365e-01 5.83650053e-01 1.45532340e-01
4.48523134e-01 -1.50326518e-02 -1.97860509e-01 -5.34787357e-01
-3.31825048e-01 -1.19992363e+00 3.70621979e-01 -7.94238925e-01
-6.75145984e-01 2.99839795e-01 2.28316352e-01 -1.06769586e+00
-8.08544159e-01 -7.03784078e-02 -1.06907177e+00 9.59425926e-01
-5.97336769e-01 2.07060114e-01 -9.10017714e-02 2.95540333e-01
7.94795036e-01 -3.01699311e-01 5.08797169e-01 -2.93553621e-01
-4.90728319e-01 3.77486646e-01 1.85885832e-01 -2.28326634e-01
2.26117864e-01 -1.52995467e+00 5.52751362e-01 5.32264829e-01
2.00344637e-01 5.52783549e-01 9.65439379e-01 -9.10688400e-01
-7.49850273e-01 -3.93478185e-01 3.29411812e-02 -3.48420560e-01
6.61262214e-01 -3.34623419e-02 -7.31897771e-01 4.12155896e-01
3.38969491e-02 -6.12536967e-01 6.14465654e-01 3.27181071e-01
5.12244046e-01 3.84356380e-02 -7.86152840e-01 9.78513300e-01
1.09102821e+00 -3.91546905e-01 -5.48443913e-01 -2.65551377e-02
1.56624347e-01 -1.00641525e+00 -5.19534171e-01 -8.50479454e-02
4.19376582e-01 -1.41450536e+00 7.18328059e-01 -7.85261273e-01
2.74646938e-01 5.22378944e-02 5.18493533e-01 -1.92432845e+00
-2.32618287e-01 -1.16944730e+00 -1.80989861e-01 9.11706269e-01
2.86250055e-01 -2.28193447e-01 1.39358795e+00 5.41134834e-01
-2.43808907e-02 -1.06754935e+00 -4.49983507e-01 -7.01343179e-01
-2.14216679e-01 -5.64212978e-01 2.52169997e-01 3.03944588e-01
3.59433979e-01 1.54992893e-01 -2.52034992e-01 -2.39149258e-01
5.10018051e-01 -2.84573495e-01 9.02771533e-01 -8.80610168e-01
-8.16265106e-01 -4.34995651e-01 1.92733370e-02 -1.54908013e+00
-3.21686000e-01 -4.76454437e-01 3.98894846e-01 -1.29943466e+00
-2.62671888e-01 -4.79914129e-01 -5.08204028e-02 1.06001921e-01
-4.82162416e-01 -3.30231965e-01 9.63158384e-02 4.09611277e-02
-6.95042193e-01 3.78733903e-01 1.46727788e+00 5.38791478e-01
-1.06775403e+00 8.01534116e-01 -6.26689374e-01 6.27988517e-01
1.07114494e+00 -3.92884195e-01 -9.31270063e-01 2.38801271e-01
5.05199552e-01 5.00937939e-01 -5.52477241e-02 -1.32393157e+00
4.52773124e-01 -2.76111752e-01 3.22635978e-01 -5.65375566e-01
2.11868882e-02 -3.01374286e-01 -2.05442645e-02 5.87632596e-01
-4.76479739e-01 -4.66100574e-02 4.25244480e-01 5.17270744e-01
1.14208482e-01 -5.86492240e-01 4.31492388e-01 -4.86612320e-01
-7.23189890e-01 3.93800773e-02 -1.05350947e+00 5.25137961e-01
1.02461374e+00 -7.80881882e-01 4.94950771e-01 -6.19574010e-01
-1.19078231e+00 5.67588627e-01 1.41033813e-01 7.35387579e-02
2.27424592e-01 -9.19820905e-01 -4.09152359e-01 2.93744624e-01
-3.83296341e-01 1.95344612e-01 1.74717769e-01 5.42212784e-01
-3.79109830e-01 -3.43005620e-02 -1.49079055e-01 -4.95521694e-01
-1.04029524e+00 2.24153042e-01 6.70640290e-01 -7.09232450e-01
-5.10195732e-01 1.07321370e+00 -2.57950217e-01 -3.51594508e-01
5.27391851e-01 -1.29898682e-01 -1.86281651e-01 1.80338249e-01
4.15802330e-01 6.32153094e-01 -1.01878308e-01 3.80912840e-01
9.71084014e-02 1.07747331e-01 -1.79338500e-01 -6.82339132e-01
1.14983428e+00 3.95855233e-02 5.61168730e-01 7.02516854e-01
3.14632446e-01 -1.30428880e-01 -2.10145116e+00 -2.10441858e-01
8.18731934e-02 -4.70534474e-01 3.13322432e-02 -8.58953476e-01
-7.15687633e-01 2.16426685e-01 3.30826551e-01 4.58757639e-01
1.09607100e+00 -1.76912472e-01 3.50091547e-01 2.56960928e-01
4.81067926e-01 -1.53424907e+00 3.97126466e-01 5.61996281e-01
3.65431845e-01 -7.63404429e-01 -2.72315107e-02 2.59967387e-01
-9.77183461e-01 7.99704254e-01 8.70085180e-01 -3.42813551e-01
3.80144686e-01 6.51483610e-02 3.78003973e-03 -4.28957462e-01
-9.25884366e-01 -4.19097006e-01 1.27680097e-02 6.71496093e-01
8.48899931e-02 -2.47957915e-01 -2.56259233e-01 3.22300822e-01
-6.30380213e-01 4.30563301e-01 3.50168169e-01 1.03750527e+00
-7.11394429e-01 -9.15975273e-01 -3.69947016e-01 7.36231446e-01
-1.92015961e-01 4.70990352e-02 -2.26645276e-01 6.40200615e-01
4.19466533e-02 8.20780277e-01 3.73676002e-01 -4.98869389e-01
4.49406445e-01 1.09854423e-01 6.30811214e-01 -9.09704864e-01
-9.43117023e-01 2.14744685e-03 2.30744123e-01 -8.07098508e-01
-5.59111051e-02 -8.72735977e-01 -1.11399543e+00 -3.72936815e-01
-3.00123155e-01 7.02562511e-01 2.83724934e-01 9.64896977e-01
-3.79955806e-02 8.50164354e-01 8.22648406e-01 -1.05815482e+00
-7.81341374e-01 -8.98253381e-01 -7.92764068e-01 7.99588785e-02
-1.47168767e-02 -7.18961596e-01 -5.14706016e-01 -2.24502116e-01] | [3.7298192977905273, 1.6371930837631226] |
9a6570a9-1aae-464f-b791-7a7d0bdef7d7 | analyzing-and-diagnosing-pose-estimation-with | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/He_Analyzing_and_Diagnosing_Pose_Estimation_With_Attributions_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/He_Analyzing_and_Diagnosing_Pose_Estimation_With_Attributions_CVPR_2023_paper.pdf | Analyzing and Diagnosing Pose Estimation With Attributions | We present Pose Integrated Gradient (PoseIG), the first interpretability technique designed for pose estimation. We extend the concept of integrated gradients for pose estimation to generate pixel-level attribution maps. To enable comparison across different pose frameworks, we unify different pose outputs into a common output space, along with a likelihood approximation function for gradient back-propagation. To complement the qualitative insight from the attribution maps, we propose three indices for quantitative analysis. With these tools, we systematically compare different pose estimation frameworks to understand the impacts of network design, backbone and auxiliary tasks. Our analysis reveals an interesting shortcut of the knuckles (MCP joints) for hand pose estimation and an under-explored inversion error for keypoints in body pose estimation. Project page: https://qy-h00.github.io/poseig/. | ['Angela Yao', 'Qiuxia Lin', 'Kerui Gu', 'Linlin Yang', 'Qiyuan He'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['hand-pose-estimation'] | ['computer-vision'] | [-3.26755643e-02 2.27049053e-01 -3.60575020e-01 -3.24568093e-01
-7.37299919e-01 -7.54323542e-01 3.83665830e-01 -3.65037173e-01
-1.64850727e-01 5.70397198e-01 7.58350790e-01 -1.88995525e-02
-3.17693442e-01 -4.21064079e-01 -7.33668864e-01 -1.30995616e-01
-1.34081140e-01 4.21857953e-01 -4.19724584e-02 -1.91568837e-01
1.56088457e-01 5.28012097e-01 -1.08056808e+00 7.76259303e-02
3.32003832e-01 7.90525198e-01 -4.80319671e-02 7.20078647e-01
4.75209653e-01 3.27001572e-01 -4.78536487e-01 -6.77204549e-01
2.40389496e-01 -1.26232058e-01 -8.27405989e-01 -2.98697084e-01
9.51893568e-01 -5.06574154e-01 -4.83785063e-01 7.33669758e-01
8.35131466e-01 -7.64892250e-02 6.90531313e-01 -1.52367246e+00
-3.18954021e-01 5.80152094e-01 -6.07744157e-01 1.69026226e-01
6.67641699e-01 5.60474992e-01 1.31084132e+00 -1.01480484e+00
1.18377566e+00 1.59885955e+00 1.16440427e+00 3.40943187e-01
-1.33282518e+00 -5.58666646e-01 2.88976759e-01 -4.06098925e-02
-1.33351231e+00 -2.35102668e-01 8.06206465e-01 -6.28234386e-01
5.09782553e-01 3.73438329e-01 8.34800541e-01 1.59467268e+00
2.38428965e-01 8.78178596e-01 1.02260351e+00 -4.51290868e-02
-2.05102518e-01 -4.05258983e-01 -3.73923145e-02 9.21009302e-01
1.36374235e-01 1.09685078e-01 -1.08162701e+00 -1.11801170e-01
1.19952488e+00 -1.37285382e-01 -2.73555011e-01 -4.94198561e-01
-1.48130906e+00 6.21666551e-01 8.76794875e-01 -4.28145707e-01
-1.87805712e-01 8.42060328e-01 2.95832038e-01 -6.14122562e-02
5.49740255e-01 4.34227735e-01 -5.04162967e-01 -1.70905471e-01
-8.61334503e-01 7.39490092e-01 7.08094060e-01 8.61911595e-01
6.20116591e-01 -1.90819934e-01 -5.65223515e-01 5.40026963e-01
6.96850955e-01 3.85704607e-01 -2.93976605e-01 -1.37977552e+00
6.92695081e-01 3.62256795e-01 3.11067905e-02 -1.20336437e+00
-7.68894613e-01 -7.85893083e-01 -3.19903791e-01 3.17682654e-01
8.64015281e-01 -3.12619001e-01 -7.87279963e-01 1.86447155e+00
4.04994607e-01 -4.74187106e-01 -9.46838200e-01 1.17361283e+00
6.81858838e-01 -1.55052254e-02 1.80493936e-01 7.18572080e-01
1.41736090e+00 -1.12766516e+00 -2.85562128e-01 -4.41588283e-01
2.80760884e-01 -5.62100470e-01 1.42070854e+00 2.93935865e-01
-9.89325881e-01 -4.70904857e-01 -1.01597023e+00 -3.13557237e-01
-2.87119359e-01 5.07427573e-01 5.92636347e-01 4.57133323e-01
-1.05239713e+00 9.39244092e-01 -9.34532523e-01 -4.32337284e-01
5.10859191e-01 2.24931315e-01 -3.67750019e-01 3.80630225e-01
-9.76685286e-01 1.02067983e+00 -6.50605187e-02 4.51063693e-01
-9.09983456e-01 -7.59945989e-01 -7.21879363e-01 -3.43238473e-01
2.47580484e-01 -9.48679030e-01 1.21066904e+00 -4.69253361e-01
-1.38313794e+00 8.70147347e-01 1.20338127e-01 9.57452729e-02
1.10272336e+00 -6.89232707e-01 2.59346396e-01 2.55353421e-01
1.85078993e-01 1.17248452e+00 9.20002699e-01 -1.02098656e+00
-1.38196707e-01 -6.12932444e-01 -1.07100122e-02 3.52884501e-01
-8.60752165e-02 -9.81770232e-02 -6.22066259e-01 -8.43187034e-01
2.42912740e-01 -1.03705585e+00 5.36655486e-02 7.65738547e-01
-8.55864763e-01 1.34685352e-01 4.41308498e-01 -1.25785649e+00
1.23858905e+00 -1.77959001e+00 5.69150686e-01 4.12888408e-01
5.47287941e-01 -6.65694773e-01 -1.11303575e-01 4.36831266e-01
-5.08275032e-02 9.95704308e-02 -1.57026321e-01 -6.78443968e-01
4.52145249e-01 -1.68243423e-01 -1.03777677e-01 7.64155567e-01
1.54496863e-01 1.27873051e+00 -7.64954925e-01 -5.38610518e-01
2.80738980e-01 7.18056023e-01 -6.83123112e-01 -1.54024273e-01
-2.82835305e-01 6.65334046e-01 -2.98168629e-01 8.89487326e-01
4.21392709e-01 -2.59939075e-01 1.13320351e-01 -7.06569612e-01
1.56034948e-02 3.64097834e-01 -1.12859702e+00 2.05368972e+00
9.52911843e-03 6.53620780e-01 2.47634977e-01 -1.80271477e-03
5.19503951e-01 7.05247698e-03 3.47346514e-01 -3.79930139e-02
2.11047277e-01 7.42863193e-02 -2.23792315e-01 -4.38194200e-02
3.84580493e-01 2.07617119e-01 -2.13829279e-01 3.57394189e-01
3.26113939e-01 -1.80485770e-01 -7.51951337e-02 3.85807753e-02
9.17133272e-01 9.79415655e-01 -7.46400058e-02 -3.33496541e-01
-1.46591380e-01 -4.98281419e-02 1.13849066e-01 5.08738875e-01
-3.19108933e-01 1.00413692e+00 6.74706995e-01 -3.78746301e-01
-1.01752579e+00 -1.52032864e+00 -7.04889596e-02 1.44890845e+00
-4.47111279e-02 -6.74870431e-01 -8.53561759e-01 -5.80921769e-01
5.89461327e-01 1.24049962e-01 -8.98730516e-01 7.05498010e-02
-5.06475866e-01 -4.44912761e-01 1.02654064e+00 8.13143730e-01
1.85562775e-01 -1.01055753e+00 -7.86387801e-01 -1.55942932e-01
-1.61878437e-01 -6.13703728e-01 -6.06742084e-01 1.01892270e-01
-8.65278661e-01 -1.00061464e+00 -9.29836035e-01 -2.36640394e-01
4.63639289e-01 -4.87680316e-01 1.35322654e+00 -1.81149274e-01
-4.11429286e-01 7.14586914e-01 -4.78015281e-02 -2.26418406e-01
7.88516477e-02 5.09607911e-01 7.44673237e-02 -7.11616158e-01
-1.79284215e-01 -6.34859860e-01 -9.94294107e-01 3.72940153e-01
-2.75797606e-01 9.14738029e-02 4.52377021e-01 5.55531323e-01
4.16537166e-01 -9.35415208e-01 1.28636464e-01 -4.07070726e-01
7.93430388e-01 -1.32225260e-01 -3.32243502e-01 1.75003871e-01
-4.31846231e-01 2.07679346e-01 -1.06209867e-01 -1.12660021e-01
-7.93511510e-01 1.34061888e-01 -2.23217100e-01 -3.64280373e-01
9.50763747e-02 4.11182404e-01 1.39525253e-02 -1.92467511e-01
8.43089104e-01 -2.37409830e-01 4.00267094e-01 -7.20151484e-01
7.61293828e-01 1.96508273e-01 6.86838448e-01 -8.83904755e-01
6.93109453e-01 4.99931425e-01 -7.99382553e-02 -4.17727053e-01
-7.27114797e-01 -8.50697756e-02 -1.07236993e+00 -6.28512979e-01
8.67472112e-01 -8.14221025e-01 -9.94091988e-01 4.12963420e-01
-1.12461841e+00 -5.82465351e-01 -1.59487218e-01 3.22822869e-01
-7.23166347e-01 3.12748313e-01 -8.06893826e-01 -4.71219361e-01
-4.64349002e-01 -1.12043190e+00 1.50643146e+00 -5.69453351e-02
-9.26746607e-01 -9.25974488e-01 2.87846893e-01 3.33121777e-01
2.10657567e-01 5.53938150e-01 4.84543025e-01 -3.63613628e-02
-4.54208553e-01 -1.36353686e-01 -2.51561195e-01 -7.39452168e-02
-2.47141659e-01 1.97680965e-01 -1.13047969e+00 -3.49499822e-01
-5.85120678e-01 -2.88359612e-01 7.71556914e-01 6.55108392e-01
1.24849474e+00 -1.42900333e-01 -3.97045493e-01 9.01191711e-01
1.02370012e+00 -6.96130931e-01 4.53975677e-01 4.46067363e-01
1.02497208e+00 7.29805708e-01 4.21313971e-01 4.17814881e-01
3.53631377e-01 8.64717424e-01 4.57156301e-01 -1.49837434e-01
-4.75168973e-01 -7.29196072e-01 4.23010170e-01 2.45068610e-01
-4.80673730e-01 -1.81509729e-03 -9.35769200e-01 1.55011266e-01
-1.68377399e+00 -5.71650326e-01 6.75510243e-02 1.99432373e+00
8.13359976e-01 9.11133736e-02 5.65299034e-01 -2.02923313e-01
5.37627757e-01 4.58763480e-01 -6.09545052e-01 1.10820815e-01
1.98559433e-01 2.09498867e-01 7.35269368e-01 7.03571558e-01
-1.05101526e+00 8.83319616e-01 7.54600286e+00 6.68371916e-01
-1.00010431e+00 1.19422734e-01 5.06431162e-01 -2.96497375e-01
-3.77362311e-01 -8.04717466e-02 -6.98254585e-01 3.07369590e-01
3.73270214e-01 3.83841127e-01 3.91992867e-01 8.36721897e-01
2.08830103e-01 5.09667769e-02 -1.15067983e+00 8.30853283e-01
-1.62263498e-01 -1.05770063e+00 -2.35943034e-01 -4.51500341e-03
2.52527505e-01 1.24583244e-01 3.43913108e-01 -7.48147219e-02
1.23223171e-01 -1.09777641e+00 1.20515537e+00 6.25918210e-01
9.85403717e-01 -4.29032922e-01 1.06441811e-01 -2.68578231e-01
-1.19035339e+00 -5.89497574e-02 4.82716896e-02 -3.05121243e-01
3.38102847e-01 3.74417692e-01 -7.63843119e-01 2.12728411e-01
7.71313667e-01 6.93706214e-01 -8.74783874e-01 8.48398447e-01
-7.33705461e-01 3.34265977e-01 -5.04683495e-01 1.79430917e-01
-7.89470449e-02 -2.02400070e-02 8.10818791e-01 1.25367749e+00
-7.07774013e-02 -6.81698263e-01 -8.97039622e-02 1.38007891e+00
7.26027191e-02 -2.74795204e-01 -3.70953739e-01 1.68026611e-01
5.54131806e-01 1.33038116e+00 -8.37966561e-01 1.13529690e-01
1.64826035e-01 1.31632078e+00 4.49917465e-01 4.89162147e-01
-9.49607253e-01 -1.69564933e-01 8.47496450e-01 4.20287430e-01
-1.07672833e-01 -4.86007720e-01 -6.12340689e-01 -8.03830683e-01
2.68830091e-01 -6.30769730e-01 3.26665252e-01 -1.06756604e+00
-1.08506978e+00 -4.78474945e-02 4.05453473e-01 -8.78024340e-01
-3.18942130e-01 -9.59180415e-01 -5.16036570e-01 8.43668938e-01
-8.13514709e-01 -1.45470965e+00 -5.31096518e-01 2.41112068e-01
1.13948479e-01 3.78891170e-01 5.43172956e-01 2.71366030e-01
-6.18942201e-01 7.79567659e-01 -2.19757885e-01 2.32233152e-01
8.49322915e-01 -1.36384690e+00 8.75795484e-01 5.08655846e-01
5.11441007e-02 9.49258327e-01 7.86573112e-01 -8.71374965e-01
-1.39048994e+00 -6.46312475e-01 2.37092957e-01 -1.19531393e+00
5.71179032e-01 -6.10986948e-01 -1.59252569e-01 1.06076968e+00
-1.81248128e-01 -1.89607307e-01 1.95267126e-01 5.18061161e-01
-4.74911809e-01 2.23145023e-01 -1.00397587e+00 7.97941864e-01
1.51277530e+00 -8.02436769e-01 -3.56960297e-01 2.49714658e-01
5.57300150e-01 -7.22156167e-01 -1.07275534e+00 2.20522895e-01
1.31649375e+00 -8.00104797e-01 1.50515664e+00 -4.26923871e-01
6.37805521e-01 -2.37104103e-01 3.53114344e-02 -1.10844553e+00
-4.85796988e-01 -4.79779363e-01 -2.48190969e-01 7.67034054e-01
5.65506876e-01 -4.84835327e-01 1.16142523e+00 5.50647020e-01
-7.11569795e-03 -8.86958838e-01 -8.10414255e-01 -4.89894390e-01
5.12029007e-02 -4.51286137e-01 3.26072752e-01 4.99817193e-01
6.44868240e-02 1.52279750e-01 -4.57364827e-01 -1.48993134e-01
7.32636869e-01 -1.78279743e-01 7.83026218e-01 -9.94130611e-01
-4.85396266e-01 -5.89067042e-01 -3.96283358e-01 -1.10709906e+00
-1.85586914e-01 -8.17176282e-01 -1.77610498e-02 -1.63097894e+00
1.45569950e-01 -1.91510975e-01 -2.35989895e-02 7.44190633e-01
-1.61564961e-01 7.72137821e-01 4.06254351e-01 2.86748528e-01
-3.96859586e-01 3.15597266e-01 1.43831563e+00 8.61844495e-02
-5.40037379e-02 -1.39473677e-01 -5.35074294e-01 9.27051306e-01
7.29512036e-01 -3.04474503e-01 -2.23233759e-01 -6.99741125e-01
5.53727984e-01 -1.54175803e-01 1.05794430e+00 -9.88674760e-01
-6.63462132e-02 1.20539121e-01 8.70570958e-01 -7.27219462e-01
5.84314942e-01 -4.24072713e-01 3.88432533e-01 6.37882650e-01
-3.91183645e-01 2.63515115e-01 7.53469393e-02 3.88784319e-01
2.75316209e-01 2.44861886e-01 3.57431322e-01 -1.71848059e-01
-7.05091000e-01 1.80902958e-01 7.82887563e-02 8.95517543e-02
4.25688356e-01 -5.73503792e-01 -3.00514102e-01 -5.32527268e-01
-1.09182560e+00 2.35027269e-01 5.57812631e-01 3.93922091e-01
4.76277918e-01 -1.32651460e+00 -5.75492322e-01 -2.24064887e-01
9.55177769e-02 -1.17177933e-01 6.87732399e-02 1.07162011e+00
-9.07412708e-01 1.52636513e-01 -4.09436166e-01 -6.25331163e-01
-1.09838951e+00 -2.29615629e-01 5.51537812e-01 -6.53042942e-02
-6.07375622e-01 1.26271319e+00 -7.23967925e-02 -1.00239635e+00
3.35400790e-01 -5.44112146e-01 3.01551521e-01 2.02948421e-01
3.00022587e-02 8.25627804e-01 -2.14126036e-01 -4.48734105e-01
-5.83206713e-01 6.52246773e-01 4.06223029e-01 -4.80303437e-01
1.17972016e+00 -1.38083130e-01 6.96208328e-03 3.49150985e-01
1.06871128e+00 7.91290551e-02 -1.73558843e+00 2.65411586e-01
-2.14580938e-01 -3.23498756e-01 -2.56615579e-01 -1.13922811e+00
-8.49321425e-01 8.47885668e-01 7.41579115e-01 -5.48443973e-01
5.80086648e-01 9.41759124e-02 2.67228454e-01 1.35843381e-01
3.75413597e-01 -1.18578875e+00 2.00516120e-01 3.99800599e-01
1.29002643e+00 -9.13442373e-01 4.76597995e-01 -5.92096090e-01
-4.96103287e-01 1.01750624e+00 6.18963480e-01 -1.67205825e-01
4.99270827e-01 3.10160190e-01 -1.19874226e-02 -6.11775517e-01
-1.09829158e-02 1.28619839e-02 6.26872778e-01 5.75052857e-01
7.62940049e-01 2.29955941e-01 -2.32061654e-01 4.34888721e-01
-7.06956983e-01 -1.28363296e-01 -1.75515518e-01 8.66832912e-01
-1.63640752e-01 -9.77934897e-01 -6.04321599e-01 5.00348270e-01
-2.06528425e-01 -1.50312099e-03 -7.87319422e-01 9.15467381e-01
-1.04577661e-01 2.76606083e-01 -6.84550628e-02 -6.42755985e-01
3.49374384e-01 1.90045133e-01 9.29124355e-01 -4.40584213e-01
-5.81768095e-01 -4.82782312e-02 2.03993112e-01 -9.54234838e-01
-7.43666366e-02 -4.59859729e-01 -9.12121534e-01 -2.75278300e-01
-1.48321301e-01 -5.17152548e-01 6.39375091e-01 5.74300468e-01
3.46122116e-01 5.99909425e-01 -1.42358243e-01 -1.60657823e+00
-6.84664786e-01 -1.17653704e+00 -4.68666524e-01 2.97856569e-01
1.07020445e-01 -1.02330601e+00 -3.55921149e-01 -1.21423252e-01] | [7.00507926940918, -0.9483263492584229] |
bc82a5f0-d3c7-4fbf-9298-cbdee8139887 | architecture-representations-for-quantum | 2210.15073 | null | https://arxiv.org/abs/2210.15073v3 | https://arxiv.org/pdf/2210.15073v3.pdf | Hierarchical quantum circuit representations for neural architecture search | Machine learning with hierarchical quantum circuits, usually referred to as Quantum Convolutional Neural Networks (QCNNs), is a promising prospect for near-term quantum computing. The QCNN is a circuit model inspired by the architecture of Convolutional Neural Networks (CNNs). CNNs are successful because they do not need manual feature design and can learn high-level features from raw data. Neural Architecture Search (NAS) builds on this success by learning network architecture and achieves state-of-the-art performance. However, applying NAS to QCNNs presents unique challenges due to the lack of a well-defined search space. In this work, we propose a novel framework for representing QCNN architectures using techniques from NAS, which enables search space design and architecture search. Using this framework, we generate a family of popular QCNNs, those resembling reverse binary trees. We then evaluate this family of models on a music genre classification dataset, GTZAN, to justify the importance of circuit architecture. Furthermore, we employ a genetic algorithm to perform Quantum Phase Recognition (QPR) as an example of architecture search with our representation. This work provides a way to improve model performance without increasing complexity and to jump around the cost landscape to avoid barren plateaus. Finally, we implement the framework as an open-source Python package to enable dynamic QCNN creation and facilitate QCNN search space design for NAS. | ['Francesco Petruccione', 'Carsten Blank', 'Daniel K. Park', 'Ilya Sinayskiy', 'Matt Lourens'] | 2022-10-26 | null | null | null | null | ['genre-classification'] | ['computer-vision'] | [ 4.73833144e-01 -1.47186518e-01 -1.58140317e-01 -1.54953733e-01
-5.28233469e-01 -6.96768761e-01 3.29390943e-01 2.75468919e-02
-2.90324390e-01 5.50928056e-01 -4.26384658e-01 -5.89571953e-01
-4.64442104e-01 -1.32726014e+00 -7.93016255e-01 -7.89974868e-01
6.62274659e-02 2.99356252e-01 9.04171839e-02 -5.28571069e-01
6.71742499e-01 4.79735941e-01 -1.86864495e+00 5.01364291e-01
7.89503813e-01 1.16498518e+00 -1.10091204e-02 7.11022735e-01
-2.38597598e-02 6.33854806e-01 -5.65738380e-01 -4.56102669e-01
2.61339545e-01 -6.98708773e-01 -8.28414142e-01 -7.06490397e-01
2.76139230e-01 3.13215733e-01 -5.22315085e-01 9.92115319e-01
5.98463774e-01 -1.83930755e-01 6.32312775e-01 -1.00304377e+00
-6.38373375e-01 8.94692421e-01 3.58586699e-01 6.43708333e-02
-4.81598191e-02 3.28517646e-01 1.49875438e+00 -1.69859663e-01
4.80369568e-01 8.24477196e-01 7.23963380e-01 8.70661199e-01
-1.49962223e+00 -7.75492430e-01 -9.00447547e-01 8.54526818e-01
-1.67016101e+00 -2.98977613e-01 7.43907690e-01 -7.46853724e-02
1.27510381e+00 1.80417120e-01 1.01789439e+00 8.94160807e-01
3.48486871e-01 5.74383199e-01 9.64960933e-01 -6.87176704e-01
7.71361828e-01 -4.10380870e-01 1.14801815e-02 9.11790729e-01
2.13452101e-01 5.76297760e-01 -8.82453024e-01 -1.19856916e-01
6.05161667e-01 -3.85099620e-01 4.90901619e-02 -4.03421223e-01
-9.15414512e-01 9.15499210e-01 1.04671454e+00 4.86134380e-01
-1.29028693e-01 8.71196449e-01 1.46276146e-01 3.75563145e-01
-4.17721897e-01 1.11932385e+00 -2.18338609e-01 -2.49123499e-01
-9.21199024e-01 2.72211373e-01 7.38009453e-01 7.24984586e-01
9.91363525e-01 3.02438121e-02 -2.89533079e-01 4.56180125e-01
4.86291572e-03 3.08142096e-01 3.74241084e-01 -9.45307136e-01
-1.01712741e-01 8.51648271e-01 -5.81470191e-01 -6.08018279e-01
-5.66809952e-01 -8.81341457e-01 -9.30121064e-01 1.17235459e-01
2.58165479e-01 3.41536105e-01 -8.52059782e-01 1.56904542e+00
-1.35948256e-01 -1.04980124e-03 3.60867471e-01 5.69410622e-01
7.79877484e-01 5.60183883e-01 -4.78065491e-01 1.85376927e-01
1.50842392e+00 -7.57589936e-01 -3.77370059e-01 1.78974509e-01
9.38742816e-01 -3.37674707e-01 8.61228466e-01 6.54420316e-01
-7.15135515e-01 -5.15320420e-01 -1.62932956e+00 1.08468533e-01
-6.21000171e-01 1.10979252e-01 1.04474676e+00 1.26322150e+00
-1.15523458e+00 1.31981611e+00 -6.45747066e-01 -3.88198793e-02
6.54566646e-01 9.58839178e-01 1.08321518e-01 1.22114070e-01
-1.30302787e+00 5.50798416e-01 9.07407403e-01 8.33902061e-02
-8.43175590e-01 -3.76592219e-01 -4.41327125e-01 3.76792848e-01
1.09049559e-01 -8.87634039e-01 1.19070196e+00 -7.68799007e-01
-1.90344584e+00 4.73014504e-01 1.12275243e-01 -8.71959627e-01
-2.23301560e-01 4.53762442e-01 -3.19788754e-01 1.11616716e-01
-4.53974754e-01 7.41390526e-01 8.18818092e-01 -4.51111585e-01
-3.06469083e-01 -1.88834459e-01 1.68254480e-01 -2.91364372e-01
-3.78108561e-01 -3.78273636e-01 -1.43773571e-01 -2.57244319e-01
4.50102299e-01 -1.17915022e+00 -3.41167718e-01 -5.25786400e-01
-3.83266658e-01 -3.05841148e-01 3.02877843e-01 3.98245871e-01
1.52693832e+00 -1.95316756e+00 2.38149509e-01 4.95306879e-01
1.35965183e-01 2.86356121e-01 -1.50011361e-01 3.53233278e-01
9.11400989e-02 3.11770678e-01 -3.67860287e-01 2.72440493e-01
1.57286018e-01 2.55513608e-01 8.80928934e-02 1.79282218e-01
4.18178707e-01 1.31323981e+00 -7.34437108e-01 -9.66731086e-02
-9.71913189e-02 1.43085554e-01 -1.05679798e+00 -2.85377830e-01
-4.31127936e-01 3.56102854e-01 -2.25835115e-01 8.51287901e-01
3.73013020e-01 -5.82008362e-01 1.99708149e-01 -3.57048512e-01
-2.91298181e-01 4.68195498e-01 -9.55885231e-01 1.83160090e+00
-2.70063996e-01 8.04925919e-01 -5.60210586e-01 -1.07733130e+00
1.07926953e+00 2.48804912e-02 1.82912573e-01 -1.02345598e+00
3.31378281e-01 6.91987693e-01 5.95581651e-01 -8.49358961e-02
4.75603908e-01 -7.79358149e-02 -1.18681498e-01 4.25334871e-01
4.35758352e-01 -4.08016533e-01 3.49604815e-01 -1.94913909e-01
1.30072570e+00 -1.31161073e-02 1.69755861e-01 -2.46747255e-01
5.87853193e-01 1.62849754e-01 2.57640272e-01 1.09718192e+00
-1.74495205e-01 5.02007127e-01 5.04423857e-01 -8.93490970e-01
-1.15319288e+00 -8.71595442e-01 -4.32326943e-01 8.48974288e-01
3.13083567e-02 -8.62335205e-01 -7.78085589e-01 -1.08246952e-01
-4.78429914e-01 3.61821920e-01 -5.57795882e-01 -4.90786403e-01
-5.52360654e-01 -1.08310390e+00 1.11711061e+00 1.56594902e-01
6.25521958e-01 -1.30684686e+00 -9.18986857e-01 3.26755673e-01
2.38862652e-02 -9.01767612e-01 3.15814853e-01 8.35569561e-01
-9.24296796e-01 -1.08680212e+00 -2.66646117e-01 -7.49893963e-01
2.74124801e-01 -4.31782395e-01 1.15261793e+00 1.44396216e-01
-5.97924232e-01 -1.21648736e-01 -4.25030559e-01 -2.49426901e-01
-7.34173238e-01 6.16271734e-01 6.21087551e-02 -1.94301277e-01
3.25658560e-01 -7.24290729e-01 -7.90738404e-01 -3.87930274e-02
-9.92884517e-01 6.86066747e-02 8.92178893e-01 1.19273114e+00
6.91560566e-01 2.11641282e-01 4.26757663e-01 -5.79670072e-01
5.41941285e-01 3.81793417e-02 -8.72067511e-01 3.41315657e-01
-8.83342087e-01 7.11120725e-01 7.21554101e-01 -1.70959234e-01
-1.37883872e-01 2.27041423e-01 -3.05411190e-01 -2.03782678e-01
-4.12485674e-02 5.62030613e-01 3.79601508e-01 -6.35851741e-01
1.15617216e+00 3.58383864e-01 -1.84847906e-01 -2.00493429e-02
3.79191250e-01 6.38367772e-01 3.84230435e-01 -5.64184368e-01
6.93214655e-01 2.36017749e-01 7.21823514e-01 -7.33802676e-01
-5.70210457e-01 -6.91008866e-02 -6.76682413e-01 8.16141665e-02
8.51478159e-01 -4.76830512e-01 -1.28297496e+00 2.08310366e-01
-1.01934028e+00 -3.03578347e-01 -2.35718533e-01 2.27631643e-01
-7.55834520e-01 -8.22232454e-04 -5.98441780e-01 -8.14888775e-01
-6.51431441e-01 -1.37715542e+00 7.24713266e-01 4.83749390e-01
7.39562139e-02 -5.57477355e-01 -4.71228696e-02 2.54697725e-02
7.27330506e-01 1.19553484e-01 1.34452808e+00 -4.24646318e-01
-1.22354221e+00 1.10958844e-01 -5.94050959e-02 2.48159051e-01
-3.85810941e-01 -1.84495747e-02 -1.07567203e+00 -1.61387965e-01
-3.05119693e-01 -5.00106633e-01 1.06348395e+00 3.14877927e-01
1.36973023e+00 -8.57962370e-02 -1.02018915e-01 1.02204001e+00
1.58195138e+00 3.04396838e-01 8.23720992e-01 7.31146216e-01
4.18578446e-01 2.52257198e-01 -8.24413970e-02 2.45171741e-01
5.97115979e-02 7.85878420e-01 8.10644329e-01 3.17821980e-01
7.06104636e-02 -6.16819523e-02 2.02667013e-01 8.40339661e-01
-2.24640459e-01 9.63864699e-02 -1.04032826e+00 1.03236794e-01
-1.55937517e+00 -8.61841917e-01 -1.45484045e-01 2.00654793e+00
6.85924232e-01 2.17424005e-01 -1.42228186e-01 4.57659870e-01
3.30694646e-01 -1.37442619e-01 -4.87255096e-01 -7.55428016e-01
-3.57834876e-01 1.11956608e+00 5.64776242e-01 -1.39248222e-01
-1.06361222e+00 1.13902605e+00 6.09879780e+00 1.17945623e+00
-1.24399102e+00 -9.00344998e-02 4.76289153e-01 1.09956942e-01
-2.21775919e-01 3.42072994e-01 -6.91609740e-01 1.23632345e-02
1.43139017e+00 1.41325876e-01 1.07529449e+00 6.55291855e-01
-4.52734381e-01 3.07915747e-01 -1.19174111e+00 1.37318134e+00
-3.11791450e-01 -2.08641219e+00 7.20134974e-02 9.19574350e-02
7.30802059e-01 2.88786739e-01 1.06248565e-01 5.07989287e-01
-1.97646976e-01 -1.48382759e+00 8.00472856e-01 2.52634585e-01
9.45836306e-01 -8.90242159e-01 6.98546767e-01 7.84755871e-02
-1.16978586e+00 -6.69058979e-01 -5.75411856e-01 -1.62076011e-01
-3.23286206e-01 8.23389068e-02 -6.49738312e-01 5.80780804e-01
7.46945739e-01 4.79251713e-01 -9.73235846e-01 1.20432568e+00
-1.86497793e-01 5.54581106e-01 -2.58534729e-01 -7.86781073e-01
4.21381265e-01 -3.10611397e-01 2.44441718e-01 8.90617669e-01
5.52749395e-01 3.60801630e-02 -5.47154903e-01 1.44594812e+00
-1.16748698e-01 4.18735035e-02 -6.01237595e-01 -4.87942606e-01
4.43919271e-01 1.09373081e+00 -9.25503969e-01 -6.66721463e-02
1.44370675e-01 7.85997093e-01 1.30989522e-01 -2.59054508e-02
-6.53275967e-01 -7.29676008e-01 2.81943262e-01 -3.55669409e-01
4.81754571e-01 5.00373021e-02 -5.10019302e-01 -1.04402053e+00
-3.97887081e-01 -9.53657091e-01 8.29979107e-02 -4.96458828e-01
-8.65864754e-01 8.05117548e-01 -4.83966827e-01 -1.29821670e+00
-2.40799278e-01 -9.98197258e-01 -5.09024680e-01 3.98099959e-01
-1.31911600e+00 -9.16523218e-01 -5.76451495e-02 4.27281767e-01
-7.66986310e-02 -3.64825040e-01 1.28544080e+00 2.70660192e-01
-3.85827154e-01 8.88358295e-01 3.18330377e-01 -6.66434094e-02
7.57941157e-02 -1.29025388e+00 4.53864455e-01 3.55189443e-01
6.46623313e-01 8.52473140e-01 4.19641614e-01 1.34977289e-02
-1.80005550e+00 -7.22556531e-01 5.77142000e-01 -3.44058275e-01
5.92859447e-01 -6.24198139e-01 -4.67634678e-01 -3.40050757e-02
-9.16109756e-02 -2.01609179e-01 8.61376107e-01 3.14821184e-01
-3.83984208e-01 -2.66488165e-01 -8.29591751e-01 5.95482171e-01
1.06014216e+00 -7.43034542e-01 -1.81427062e-01 1.62243575e-01
6.84881687e-01 -2.62017012e-01 -7.92594075e-01 6.99131250e-01
9.81600285e-01 -1.29119813e+00 9.49135482e-01 -3.02716821e-01
5.04376113e-01 -3.54111522e-01 -3.30904543e-01 -1.07808495e+00
-3.89318287e-01 -8.91331255e-01 -2.35470504e-01 6.33400857e-01
6.85728431e-01 -4.82720941e-01 1.19180143e+00 -7.65189677e-02
-2.70076543e-01 -9.71010029e-01 -1.26065159e+00 -8.61441135e-01
1.60088077e-01 -6.73749864e-01 8.66723955e-01 5.51599622e-01
-9.06058401e-02 5.09505928e-01 -5.50756790e-02 -6.86157867e-02
5.38351774e-01 3.40634376e-01 4.84562397e-01 -1.24299455e+00
-6.38573110e-01 -1.01646745e+00 -8.55265677e-01 -7.69132137e-01
-1.12910159e-01 -1.20967245e+00 -1.39731929e-01 -1.01314092e+00
7.49514699e-02 -7.21374631e-01 -6.15744829e-01 2.84405768e-01
4.63635802e-01 6.69610381e-01 1.53006434e-01 3.21232498e-01
-5.70737660e-01 7.66324401e-01 1.07926023e+00 -3.92234594e-01
-2.84322590e-01 8.90035853e-02 -5.09270668e-01 2.77592093e-01
9.21104014e-01 -6.74059808e-01 -9.86988097e-02 -1.10894710e-01
8.39986503e-01 -1.50559321e-01 3.40607285e-01 -1.61994743e+00
4.51697797e-01 4.40573871e-01 2.77750283e-01 -5.83224356e-01
3.08298230e-01 -4.87204522e-01 2.15813532e-01 1.05410218e+00
-2.78893113e-01 -6.54683858e-02 7.69815296e-02 4.47544724e-01
-1.42844632e-01 -6.02924049e-01 7.13771403e-01 -1.14728898e-01
-6.01603687e-01 1.37507111e-01 -3.62515599e-01 -4.25747752e-01
5.66986918e-01 -3.69049489e-01 -5.19039452e-01 2.89205797e-02
-6.76313162e-01 -5.49418256e-02 3.76195937e-01 1.86612800e-01
4.45157647e-01 -1.23357546e+00 -2.12405920e-01 4.63189542e-01
4.54349995e-01 -1.61066920e-01 2.13711560e-01 5.59514701e-01
-1.02893639e+00 1.00721300e+00 -4.06034350e-01 -9.66630757e-01
-8.08341324e-01 3.34693909e-01 6.37454629e-01 -3.14875185e-01
-1.68713480e-01 9.16482747e-01 -3.68830949e-01 -7.54616857e-01
7.72457719e-02 -4.96487528e-01 -1.79844633e-01 -2.60576099e-01
2.69564480e-01 7.92103633e-02 3.59072715e-01 -3.25491667e-01
-3.51322979e-01 5.98640621e-01 4.74534601e-01 -8.12900513e-02
1.30365837e+00 5.25135040e-01 -4.57912385e-01 1.25152782e-01
1.07981861e+00 -5.24464309e-01 -7.12268054e-01 3.47295888e-02
3.93941104e-01 1.95172250e-01 3.36256027e-02 -6.12775624e-01
-7.59458005e-01 1.08822548e+00 1.08531630e+00 2.60878563e-01
1.14663815e+00 -4.89784814e-02 5.96882105e-01 1.12089384e+00
6.78224742e-01 -9.04733837e-01 2.98792273e-01 9.58996356e-01
3.17047268e-01 -8.77655327e-01 -2.17149377e-01 8.43556523e-02
7.79944807e-02 1.73916328e+00 3.57900411e-01 -1.41760275e-01
5.84719598e-01 -2.38978252e-01 -2.96726942e-01 -5.63481152e-01
-8.03851426e-01 -5.26883721e-01 3.50426108e-01 3.59410793e-01
3.29970479e-01 1.65420175e-01 -2.73784757e-01 5.81647038e-01
-8.43154550e-01 3.50017026e-02 5.39235771e-01 9.47211802e-01
-4.01889950e-01 -1.58166289e+00 -9.57606882e-02 3.56689990e-01
-4.85289603e-01 -4.95348573e-01 -3.29945534e-01 3.47583443e-01
3.67702961e-01 7.17434049e-01 -1.42420053e-01 -9.87149239e-01
1.21324144e-01 1.35038525e-01 9.05605912e-01 -5.31884432e-01
-8.39037716e-01 -2.42174014e-01 -1.21773407e-01 -4.96089220e-01
-3.50632578e-01 -7.02518746e-02 -1.11555851e+00 -2.12074891e-01
-7.35978961e-01 2.60079563e-01 1.01217592e+00 9.14356351e-01
4.68718112e-01 7.78381765e-01 4.35774028e-01 -7.61418998e-01
-5.50286233e-01 -7.52029538e-01 -4.57704276e-01 -1.21208407e-01
6.88849464e-02 -4.04136270e-01 4.55809794e-02 -4.27354753e-01] | [5.5676398277282715, 4.9798665046691895] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.