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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
8b33e049-9a37-486f-99bb-0e3f6cf91fee | scalable-scene-flow-from-point-clouds-in-the | 2103.01306 | null | https://arxiv.org/abs/2103.01306v5 | https://arxiv.org/pdf/2103.01306v5.pdf | Scalable Scene Flow from Point Clouds in the Real World | Autonomous vehicles operate in highly dynamic environments necessitating an accurate assessment of which aspects of a scene are moving and where they are moving to. A popular approach to 3D motion estimation, termed scene flow, is to employ 3D point cloud data from consecutive LiDAR scans, although such approaches have been limited by the small size of real-world, annotated LiDAR data. In this work, we introduce a new large-scale dataset for scene flow estimation derived from corresponding tracked 3D objects, which is $\sim$1,000$\times$ larger than previous real-world datasets in terms of the number of annotated frames. We demonstrate how previous works were bounded based on the amount of real LiDAR data available, suggesting that larger datasets are required to achieve state-of-the-art predictive performance. Furthermore, we show how previous heuristics for operating on point clouds such as down-sampling heavily degrade performance, motivating a new class of models that are tractable on the full point cloud. To address this issue, we introduce the FastFlow3D architecture which provides real time inference on the full point cloud. Additionally, we design human-interpretable metrics that better capture real world aspects by accounting for ego-motion and providing breakdowns per object type. We hope that this dataset may provide new opportunities for developing real world scene flow systems. | ['Jonathon Shlens', 'Zhifeng Chen', 'Nichola Abdo', 'Chris Sweeney', 'Philipp Jund'] | 2021-03-01 | null | null | null | null | ['scene-flow-estimation'] | ['computer-vision'] | [-9.04723778e-02 -2.93819994e-01 -2.41347522e-01 -3.46437544e-01
-4.45767611e-01 -7.17050016e-01 6.27196729e-01 8.87402594e-02
-4.93811876e-01 6.03596687e-01 -1.81452513e-01 -6.77115798e-01
1.28063664e-01 -1.06631577e+00 -7.95497119e-01 -1.48490444e-01
-3.59739304e-01 7.18547821e-01 9.06855464e-01 -1.13362566e-01
3.09106708e-01 1.12164307e+00 -1.87649405e+00 -1.59504056e-01
5.87297559e-01 9.70156074e-01 1.13892019e-01 1.05556798e+00
-3.77682060e-01 5.01946390e-01 -4.18230116e-01 -2.04745546e-01
6.18065596e-01 8.88801664e-02 -5.65494895e-01 4.86294478e-02
9.76610303e-01 -8.04611742e-01 -4.33847159e-01 5.85444629e-01
5.94972335e-02 1.96753770e-01 4.35541570e-01 -1.79417419e+00
2.43225411e-01 -2.63812721e-01 -5.08104444e-01 6.29633665e-01
1.83271930e-01 6.74314678e-01 6.83107615e-01 -7.76545584e-01
8.61632228e-01 1.32995486e+00 6.80455327e-01 2.50233829e-01
-1.11662340e+00 -6.86458886e-01 2.13738397e-01 3.52190524e-01
-1.21271336e+00 -6.81465507e-01 5.44974685e-01 -6.17124557e-01
1.23796070e+00 2.12490991e-01 8.68059397e-01 4.51791078e-01
2.08115373e-02 4.99673933e-01 5.50095499e-01 1.08123526e-01
2.30055034e-01 -1.71662658e-01 1.69568025e-02 7.31680691e-01
5.72298884e-01 2.82196850e-01 -4.89922136e-01 -1.98097497e-01
7.15296984e-01 -1.35610104e-01 -8.00461601e-03 -8.74493062e-01
-1.33896172e+00 7.57260025e-01 3.87921214e-01 -3.17795664e-01
-1.78299263e-01 4.96876001e-01 4.67885941e-01 2.48704273e-02
5.15444756e-01 2.72583365e-02 -4.49363142e-01 -4.74052161e-01
-1.06618524e+00 5.01745224e-01 7.59403348e-01 1.40447652e+00
1.21303642e+00 6.59944490e-02 3.57065916e-01 1.27059996e-01
3.01712871e-01 7.69981205e-01 -3.24784636e-01 -1.77019739e+00
7.19481766e-01 6.24217331e-01 3.18254143e-01 -9.06327069e-01
-3.78586620e-01 3.29823978e-02 -3.03499550e-01 6.48067772e-01
6.76027298e-01 -3.20623741e-02 -7.86024153e-01 1.40909159e+00
6.00455880e-01 6.17273390e-01 -7.54654706e-02 8.51837575e-01
2.93062896e-01 6.69683754e-01 -5.44814914e-02 -1.01021387e-01
1.11543202e+00 -7.03195572e-01 -2.29009628e-01 -4.19856161e-01
7.07974136e-01 -6.00011826e-01 7.95519173e-01 1.31055519e-01
-1.07656956e+00 -5.80131650e-01 -9.10737813e-01 -2.55195051e-01
-3.43827039e-01 -4.46126282e-01 8.50403607e-01 6.00423694e-01
-1.03872538e+00 4.20049578e-01 -1.22211528e+00 -6.16564095e-01
7.36606538e-01 3.24465722e-01 -3.24896663e-01 -2.10515410e-01
-5.79300523e-01 1.21897423e+00 1.18828498e-01 -1.29020093e-02
-8.00512850e-01 -9.83477235e-01 -9.00920749e-01 -1.81037486e-01
4.07408416e-01 -7.89332330e-01 1.18788230e+00 -6.39264286e-02
-9.55531895e-01 7.03870773e-01 -5.61690927e-01 -7.28125513e-01
7.26214707e-01 -1.87786385e-01 -1.80806950e-01 2.52737731e-01
2.60687530e-01 9.33870435e-01 4.96780515e-01 -1.24014735e+00
-1.02096474e+00 -4.94644791e-01 2.62928694e-01 1.70203909e-01
1.62424251e-01 -2.09086210e-01 -6.80032849e-01 7.37352446e-02
7.92666897e-02 -1.11188257e+00 -4.96572137e-01 7.62680411e-01
-9.57240816e-04 4.70140623e-03 1.37802339e+00 7.66066387e-02
7.25671768e-01 -2.00977755e+00 -4.52431053e-01 5.06314524e-02
2.57707447e-01 3.28652114e-01 6.28391206e-02 2.40044206e-01
5.26734531e-01 1.80907562e-01 -1.74392238e-01 -3.65449697e-01
-1.14995450e-01 5.75052202e-01 -4.88356054e-01 7.33315766e-01
2.92608231e-01 6.50400102e-01 -1.01558554e+00 -7.94980586e-01
6.82973862e-01 3.89321834e-01 -7.29826510e-01 -1.53960332e-01
-2.01575965e-01 2.90921301e-01 -4.82630819e-01 5.89750409e-01
9.87654030e-01 -3.11887506e-02 -1.82908073e-01 -1.29929215e-01
-4.60729331e-01 3.26645344e-01 -1.32942235e+00 1.70058465e+00
-4.30335253e-01 1.09543991e+00 1.89963058e-01 -6.43471360e-01
7.48508632e-01 -1.16710477e-01 8.42109919e-01 -3.38416189e-01
-8.43490288e-02 1.60909534e-01 -1.63501069e-01 -3.36503148e-01
9.19492900e-01 1.11269066e-03 1.26850396e-01 2.30896100e-01
-4.41209048e-01 -7.77412057e-01 5.74722707e-01 3.03035408e-01
1.35244131e+00 2.87384689e-01 9.13025439e-02 -1.71787649e-01
1.86341822e-01 8.25123131e-01 5.91604233e-01 6.75583899e-01
-6.15216136e-01 2.80964136e-01 2.40323141e-01 -6.62776411e-01
-1.24742806e+00 -1.11482811e+00 -3.49924415e-01 5.62431991e-01
6.02980733e-01 -4.16786015e-01 -2.76219666e-01 -5.00201464e-01
5.16105533e-01 6.63983583e-01 -1.15069084e-01 3.39266181e-01
-8.05947602e-01 -3.89955729e-01 4.91551757e-01 6.25481963e-01
3.60532165e-01 -5.43932319e-01 -1.45399523e+00 3.99466455e-01
-3.18994671e-02 -1.61225677e+00 -1.84963763e-01 -1.03357458e-03
-1.22132051e+00 -1.25724888e+00 -1.46959201e-02 6.06969232e-03
4.78460222e-01 7.01188385e-01 1.19664061e+00 6.77282661e-02
-3.37605953e-01 4.38102424e-01 -7.83368498e-02 -5.55314481e-01
-2.95345753e-01 -1.12224020e-01 1.49814546e-01 -5.43944716e-01
5.87489188e-01 -5.74101210e-01 -6.43719375e-01 4.29678917e-01
-5.90405047e-01 7.93642327e-02 1.52328014e-01 1.49455965e-01
6.30064726e-01 6.71210289e-02 -7.80961663e-02 -5.44476986e-01
-4.96972390e-02 -4.78926808e-01 -1.05323112e+00 -3.90951186e-01
-3.62741709e-01 6.77819178e-02 3.16328317e-01 -1.99722052e-01
-7.38573670e-01 2.75953323e-01 1.75087854e-01 -8.60558450e-01
-2.35487595e-01 4.10089791e-02 2.18832448e-01 -9.56401676e-02
5.24183333e-01 -1.79706335e-01 1.31235182e-01 -1.66771621e-01
5.16999364e-01 2.77656496e-01 7.34301209e-01 -5.23888826e-01
1.04716897e+00 1.12755716e+00 5.22799253e-01 -8.91990066e-01
-4.93364602e-01 -7.21456349e-01 -8.49011958e-01 -3.67851973e-01
7.60202348e-01 -9.50936973e-01 -9.54574227e-01 1.30123511e-01
-1.30768180e+00 -4.09044445e-01 -4.46154386e-01 6.45544648e-01
-8.67911398e-01 2.46678993e-01 -2.75149763e-01 -9.96610701e-01
1.37555853e-01 -1.29799616e+00 1.25474930e+00 -1.04753889e-01
-2.02848971e-01 -7.33176172e-01 3.04728150e-02 2.68441111e-01
4.04851496e-01 3.69722694e-01 5.01985848e-01 -4.85854484e-02
-1.29479086e+00 -7.96236396e-02 -4.36374128e-01 -1.31319731e-01
-8.24060738e-02 4.85973865e-01 -8.20621431e-01 -1.87697917e-01
-2.54961878e-01 9.38544348e-02 7.01818883e-01 3.39047968e-01
1.00060797e+00 3.85710150e-02 -6.98826015e-01 6.71968162e-01
1.51868248e+00 6.46365136e-02 4.39005703e-01 3.73791397e-01
7.47087836e-01 5.44782281e-01 8.73595536e-01 4.56411153e-01
7.51498342e-01 7.64015973e-01 7.70505309e-01 -3.21563967e-02
-2.18420878e-01 -2.79216141e-01 6.60949647e-02 4.68484163e-01
-6.26833960e-02 -1.94141895e-01 -1.18940163e+00 7.92252243e-01
-1.76192462e+00 -1.21137905e+00 -6.20048463e-01 2.15973687e+00
1.59653425e-01 5.30611634e-01 1.92779794e-01 7.99712315e-02
4.54929531e-01 2.18022287e-01 -8.55274975e-01 -2.82156289e-01
1.94516420e-01 4.39508632e-02 1.21074784e+00 6.82059169e-01
-8.94077241e-01 1.03224337e+00 6.74394560e+00 2.50206560e-01
-1.01410043e+00 -1.17608652e-01 1.58228606e-01 -4.00965333e-01
-1.06589720e-01 3.96696419e-01 -1.12486053e+00 4.76402044e-01
1.22095835e+00 -3.91216278e-01 1.12143956e-01 1.00509512e+00
6.23117447e-01 -4.77992713e-01 -1.10827971e+00 9.56639290e-01
-2.42719084e-01 -1.55005956e+00 -9.61223170e-02 5.54607451e-01
3.58830065e-01 6.34585977e-01 -4.79995400e-01 1.79173887e-01
6.28130615e-01 -6.19378030e-01 6.73821926e-01 3.52100611e-01
7.14435637e-01 -6.34922206e-01 2.29884118e-01 4.78766769e-01
-1.60687888e+00 9.69504938e-02 -4.44277167e-01 -2.33903185e-01
6.42273664e-01 4.98225421e-01 -1.09096336e+00 3.93308371e-01
7.89439619e-01 7.60642409e-01 -5.37648261e-01 1.27619064e+00
3.86275291e-01 3.81175518e-01 -1.00141811e+00 1.57523140e-01
2.18826443e-01 -2.57084109e-02 7.86714137e-01 1.12023807e+00
4.72700357e-01 9.24770012e-02 2.85579890e-01 7.52913058e-01
2.16985255e-01 -4.94105041e-01 -1.09352636e+00 4.60855335e-01
7.84946144e-01 1.05895698e+00 -8.29247296e-01 -4.73845154e-01
-5.46355188e-01 3.22742611e-01 8.30828175e-02 2.05826059e-01
-9.24982965e-01 -4.26620916e-02 1.22234225e+00 5.61868668e-01
3.75405550e-01 -9.38761055e-01 -4.29747283e-01 -1.09540176e+00
-7.82572478e-02 -1.74532279e-01 1.52975932e-01 -8.31113100e-01
-8.17038834e-01 3.59374136e-01 4.58685517e-01 -1.63685262e+00
-3.51110131e-01 -5.92041790e-01 -3.39490563e-01 7.20488548e-01
-1.82855046e+00 -7.08737969e-01 -6.17241502e-01 5.65537155e-01
6.82555497e-01 2.47436017e-01 3.02022725e-01 2.92879641e-01
-1.61128908e-01 -6.24370798e-02 -2.87594348e-01 -1.73160300e-01
3.90553594e-01 -8.92336607e-01 9.11813080e-01 1.03637934e+00
1.30935073e-01 2.92600304e-01 8.36044073e-01 -6.32568240e-01
-1.74304473e+00 -1.13863659e+00 5.50282240e-01 -9.82488394e-01
6.57135129e-01 -2.67017782e-01 -7.54279852e-01 7.56626666e-01
-3.99576128e-01 5.81545115e-01 1.49081796e-01 -1.80079654e-01
-1.02546461e-01 -2.67991126e-01 -1.19115341e+00 5.23570478e-01
1.55484164e+00 -3.22215796e-01 -2.92378426e-01 -2.08401103e-02
7.47797966e-01 -6.67996407e-01 -7.82661974e-01 6.34737968e-01
4.21711683e-01 -1.02234316e+00 1.05523229e+00 -3.51347774e-01
9.93393511e-02 -7.39256978e-01 -2.98664600e-01 -8.76939595e-01
-2.28353534e-02 -3.68728518e-01 -4.38645363e-01 7.51939416e-01
6.88198283e-02 -5.24428189e-01 1.21386874e+00 8.22848380e-01
-3.42283875e-01 -4.06779319e-01 -1.09463692e+00 -8.74604225e-01
-1.40260667e-01 -1.04886663e+00 7.71035850e-01 6.89810574e-01
-4.37003702e-01 1.18011050e-01 8.94082040e-02 2.86003590e-01
8.00202489e-01 1.40015244e-01 1.48864615e+00 -1.33725142e+00
3.23445618e-01 -4.22810137e-01 -9.94089723e-01 -1.35456252e+00
6.64128438e-02 -5.67216694e-01 3.59708034e-02 -1.60844946e+00
-2.69920200e-01 -9.43994403e-01 3.15571159e-01 2.99525559e-01
7.01373070e-02 2.90463507e-01 4.14620250e-01 3.23809654e-01
-5.22354364e-01 2.99302369e-01 1.11068022e+00 -3.81973796e-02
-1.31298125e-01 2.95318151e-03 -1.96580142e-01 7.28363812e-01
5.46758592e-01 -3.18806380e-01 -5.34880042e-01 -8.21444929e-01
-1.27196372e-01 1.11155637e-01 6.19265556e-01 -1.30896199e+00
4.61682767e-01 -5.72514713e-01 1.96468830e-01 -1.27325940e+00
8.42047632e-01 -1.09270561e+00 2.25244358e-01 3.14533412e-01
1.67548344e-01 4.14284199e-01 5.22991598e-01 6.26571655e-01
6.58446774e-02 1.26112357e-01 6.51768446e-01 -2.21027941e-01
-1.09603536e+00 7.15338707e-01 -2.38946661e-01 1.81240454e-01
1.35562360e+00 -7.06163526e-01 -4.48039681e-01 -2.05317706e-01
-1.31936073e-01 4.19196784e-01 8.08196068e-01 4.42909002e-01
4.79702234e-01 -1.12015271e+00 -6.38336360e-01 1.16107590e-01
7.61843771e-02 5.60615599e-01 -7.85926450e-03 4.84699756e-01
-9.52317178e-01 5.46504200e-01 -1.56683683e-01 -1.12709928e+00
-1.20594037e+00 4.38090712e-01 1.35675371e-01 5.29086404e-02
-8.80362689e-01 3.83521408e-01 -7.93474913e-02 -2.73973346e-01
-1.26108035e-01 -4.53693151e-01 2.97615111e-01 -6.89040497e-02
3.27399254e-01 7.37880528e-01 1.32582814e-01 -8.97821426e-01
-6.05971038e-01 6.74259186e-01 3.02625000e-01 -1.97127625e-01
1.11286390e+00 -2.93942451e-01 1.72910303e-01 3.79648417e-01
1.12731516e+00 -6.77767918e-02 -1.82698798e+00 5.49032446e-03
-5.14743589e-02 -8.97852659e-01 1.10061958e-01 -3.37644219e-02
-9.98459339e-01 9.19294119e-01 5.05247772e-01 2.25046694e-01
7.28595674e-01 -1.55341430e-02 8.75016749e-01 4.48680192e-01
8.70208204e-01 -8.11195552e-01 -4.81534362e-01 5.42668402e-01
3.05142283e-01 -1.22304356e+00 3.28966379e-01 -5.61501145e-01
-2.92013645e-01 1.03158307e+00 7.12597668e-01 -9.54219103e-02
5.37302613e-01 5.26743352e-01 -4.99979630e-02 -2.71717995e-01
-8.66968632e-01 -4.76345807e-01 -2.50806689e-01 7.61888087e-01
-1.97047442e-01 -1.11666672e-01 1.52148634e-01 -5.05966127e-01
-5.38753688e-01 1.81787297e-01 6.19330704e-01 1.19286859e+00
-5.94468594e-01 -8.63039315e-01 -4.03315634e-01 5.25195718e-01
-2.96159252e-03 2.42694393e-01 1.02902271e-01 1.02827859e+00
7.80850872e-02 1.07016575e+00 6.22955501e-01 -4.12310250e-02
5.57462573e-01 -2.75705546e-01 4.99990880e-01 -5.03600478e-01
7.81387612e-02 -4.08213824e-01 2.04684034e-01 -9.18743372e-01
-5.55301666e-01 -7.99603343e-01 -1.31849468e+00 -7.90306389e-01
-8.74213204e-02 -9.87974256e-02 9.70434189e-01 7.84121335e-01
6.13898754e-01 1.08514540e-01 4.09342468e-01 -1.38781512e+00
-7.51194879e-02 -4.01119530e-01 -5.94892725e-02 2.15477362e-01
5.62978148e-01 -8.53465974e-01 -3.48391831e-01 1.86467588e-01] | [7.990083694458008, -2.3660106658935547] |
4e7fc372-c802-4dc7-af62-d2cf12958293 | tbrats-trusted-brain-tumor-segmentation | 2206.09309 | null | https://arxiv.org/abs/2206.09309v3 | https://arxiv.org/pdf/2206.09309v3.pdf | TBraTS: Trusted Brain Tumor Segmentation | Despite recent improvements in the accuracy of brain tumor segmentation, the results still exhibit low levels of confidence and robustness. Uncertainty estimation is one effective way to change this situation, as it provides a measure of confidence in the segmentation results. In this paper, we propose a trusted brain tumor segmentation network which can generate robust segmentation results and reliable uncertainty estimations without excessive computational burden and modification of the backbone network. In our method, uncertainty is modeled explicitly using subjective logic theory, which treats the predictions of backbone neural network as subjective opinions by parameterizing the class probabilities of the segmentation as a Dirichlet distribution. Meanwhile, the trusted segmentation framework learns the function that gathers reliable evidence from the feature leading to the final segmentation results. Overall, our unified trusted segmentation framework endows the model with reliability and robustness to out-of-distribution samples. To evaluate the effectiveness of our model in robustness and reliability, qualitative and quantitative experiments are conducted on the BraTS 2019 dataset. | ['Huazhu Fu', 'Meng Wang', 'Xiaojing Shen', 'Xuedong Yuan', 'Ke Zou'] | 2022-06-19 | null | null | null | null | ['brain-tumor-segmentation'] | ['medical'] | [-1.44003540e-01 8.95958483e-01 -4.68581349e-01 -8.20248604e-01
-9.14913654e-01 -4.21226531e-01 3.34416062e-01 2.73457140e-01
-3.72176975e-01 8.57969701e-01 -7.76169375e-02 -2.82711804e-01
-1.99125126e-01 -8.20280552e-01 -7.54243970e-01 -8.83359671e-01
2.66502887e-01 5.58456719e-01 3.75004619e-01 3.33282322e-01
5.81168383e-02 1.81891799e-01 -9.15833950e-01 -2.71210093e-02
1.26615870e+00 1.36858273e+00 -2.90337175e-01 1.05733559e-01
5.67353927e-02 6.75174236e-01 -6.77092612e-01 -7.13299632e-01
9.31779519e-02 2.06224844e-01 -6.83656633e-01 5.00558577e-02
-2.44465292e-01 -4.91338283e-01 -2.36196676e-03 1.59699142e+00
7.46553540e-02 -3.26816231e-01 8.21372211e-01 -1.40784335e+00
-4.18289185e-01 1.33857799e+00 -3.88617218e-01 -4.38289225e-01
-2.75679417e-02 1.64546594e-01 7.80550659e-01 -4.00451154e-01
4.09843951e-01 8.56232822e-01 6.92050338e-01 4.40166444e-01
-1.04031265e+00 -5.17563760e-01 2.80946851e-01 -1.17299624e-01
-1.53101039e+00 -3.34160775e-01 5.73119283e-01 -3.70114714e-01
1.98122874e-01 -5.09345084e-02 7.60731399e-01 1.14398277e+00
7.85127223e-01 8.15320849e-01 1.02078581e+00 -1.76892072e-01
7.80429721e-01 4.88551885e-01 3.96900266e-01 5.84328890e-01
6.74080908e-01 7.99148157e-02 -4.06314641e-01 -2.64125198e-01
5.13689995e-01 -1.82281181e-01 -2.63264447e-01 -2.51948386e-01
-7.58024514e-01 8.80759656e-01 4.65065271e-01 7.03397766e-02
-3.78275931e-01 4.53403860e-01 2.43368939e-01 -3.21009338e-01
6.60215259e-01 -1.59588858e-01 -2.43368894e-01 1.55694276e-01
-1.19212234e+00 -2.36577448e-02 8.77159297e-01 8.94332170e-01
2.54443228e-01 -1.28011480e-01 -4.27523792e-01 2.07846820e-01
1.01177645e+00 5.81480742e-01 3.13535929e-01 -1.18943703e+00
-2.95641005e-01 3.08563381e-01 4.01046649e-02 -8.64655674e-01
-2.84132004e-01 -5.56741953e-01 -7.94283569e-01 2.01842681e-01
4.87780273e-01 -3.00042808e-01 -1.26808619e+00 1.76624715e+00
2.63649315e-01 2.04996645e-01 -3.24624553e-02 7.56618798e-01
5.80410182e-01 2.24114191e-02 2.11099405e-02 -9.78711694e-02
1.03438544e+00 -6.60418749e-01 -7.93316603e-01 3.30564976e-02
2.26338878e-01 -2.69753993e-01 5.95798552e-01 4.98660117e-01
-1.00955284e+00 1.08262673e-01 -1.01804698e+00 4.00503337e-01
9.56061035e-02 -1.48089558e-01 6.76272154e-01 1.07418251e+00
-1.30077040e+00 5.16854525e-01 -1.12697875e+00 1.30130693e-01
8.15427423e-01 4.38721657e-01 2.90687699e-02 8.85288268e-02
-1.31464839e+00 8.38326991e-01 5.31637847e-01 1.63241670e-01
-1.24293673e+00 -7.65093267e-01 -6.50322139e-01 5.85450698e-03
4.39052254e-01 -6.48229718e-01 1.37069416e+00 -9.49680269e-01
-1.72780776e+00 5.96192658e-01 9.43804756e-02 -7.42766798e-01
9.58019972e-01 2.92716891e-01 9.42474678e-02 3.90511721e-01
5.35408594e-02 9.65970039e-01 9.38126802e-01 -1.51853263e+00
-5.03164649e-01 -4.09636259e-01 -3.75020444e-01 -2.26039588e-01
-6.06586486e-02 -3.09816450e-01 -5.38299203e-01 -2.82908827e-01
4.70232248e-01 -9.03942168e-01 -5.40024221e-01 1.59582973e-01
-7.40085721e-01 -1.01912141e-01 2.20904410e-01 -5.18526733e-01
8.48138392e-01 -2.01456594e+00 -1.84659347e-01 1.03617907e+00
5.68725765e-01 -3.36757272e-01 4.23469692e-01 -5.92541575e-01
3.23185325e-01 5.72432697e-01 -4.20625806e-01 -4.23871160e-01
2.39832729e-01 4.42553133e-01 -2.66195118e-01 5.80980062e-01
-1.49369821e-01 7.90177107e-01 -7.16118395e-01 -7.97727644e-01
-9.74064544e-02 5.46772242e-01 -3.27632457e-01 -1.06612988e-01
-3.26477855e-01 4.42336679e-01 -7.50954568e-01 8.22465599e-01
6.20618880e-01 -1.92123771e-01 1.35234520e-01 -9.55975950e-02
3.32940996e-01 -2.84204096e-01 -7.10073769e-01 1.40715516e+00
1.52166381e-01 1.33965105e-01 1.31846488e-01 -6.68429375e-01
8.19405258e-01 3.79650027e-01 5.45745850e-01 -2.19539866e-01
6.74073637e-01 1.84063211e-01 -1.92830876e-01 -2.64860570e-01
4.51041907e-01 -1.44326240e-01 -2.10319415e-01 4.32424784e-01
-6.81100935e-02 -6.41789377e-01 -2.63251334e-01 3.26514810e-01
7.93007195e-01 -1.43257707e-01 -3.87462564e-02 -3.65020663e-01
4.92689237e-02 -7.94688985e-02 8.93135130e-01 8.08839321e-01
-8.01859260e-01 4.33844924e-01 1.00967705e+00 1.17543496e-01
-7.02733219e-01 -1.32476723e+00 -4.27763760e-01 1.84861705e-01
1.14455506e-01 -1.67368597e-03 -1.25468624e+00 -9.54743922e-01
-2.71155778e-02 1.08003366e+00 -8.23398232e-01 -2.50498533e-01
2.63912737e-01 -6.67809248e-01 8.16436291e-01 5.48531830e-01
5.16467035e-01 -4.35485035e-01 -4.12639439e-01 -5.89613952e-02
-2.32940435e-01 -1.18770361e+00 -1.18592240e-01 1.00110970e-01
-1.04428744e+00 -9.38808203e-01 -5.54997087e-01 -2.13023737e-01
1.12335932e+00 -3.90854120e-01 8.14311624e-01 4.49334197e-02
1.40087828e-01 3.94482851e-01 -9.64936242e-02 -5.85504115e-01
-5.32886684e-01 -1.33482128e-01 1.12161905e-01 -1.76774696e-01
1.11898795e-01 -3.07549506e-01 -5.70378482e-01 3.57064456e-01
-1.04231203e+00 -1.39997065e-01 4.06953543e-01 7.18932807e-01
9.21641707e-01 5.38343310e-01 5.81377625e-01 -8.99934351e-01
7.07103968e-01 -4.38170344e-01 -7.48812795e-01 3.63669753e-01
-1.14385557e+00 -2.43328474e-02 2.02918991e-01 -1.16998196e-01
-1.23875797e+00 -4.55908384e-03 6.39028400e-02 -3.72634858e-01
3.70368175e-02 7.59920835e-01 -4.43908945e-02 -1.87332869e-01
6.52643919e-01 -2.59749681e-01 2.61127502e-01 1.59924403e-01
3.07274729e-01 5.41801095e-01 4.53969777e-01 -8.17370355e-01
4.44041550e-01 6.16134942e-01 -1.35246873e-01 4.09947708e-02
-1.00096357e+00 3.35412383e-01 -3.63559425e-01 -5.74905097e-01
8.33205402e-01 -6.29315376e-01 -8.35468471e-01 7.65928507e-01
-9.29491878e-01 -3.27148765e-01 -3.87310237e-01 4.66438085e-01
-5.29079199e-01 2.17355162e-01 -6.46514416e-01 -7.88158298e-01
-3.92234474e-01 -1.67730463e+00 9.04830217e-01 3.50620925e-01
-2.48974830e-01 -1.00230384e+00 -3.90017688e-01 5.93149483e-01
3.38042587e-01 3.71845454e-01 9.43413317e-01 -7.30908215e-01
-6.64063275e-01 -5.16893864e-01 -1.74721405e-01 5.29993296e-01
-1.95202067e-01 5.19592345e-01 -1.18406010e+00 3.09535712e-02
3.34965408e-01 -1.34185538e-01 7.76067615e-01 8.62277687e-01
1.47703958e+00 -1.05337054e-01 -5.17467082e-01 3.10755104e-01
1.10890746e+00 -9.10624564e-02 5.30183554e-01 3.87354055e-04
2.24393398e-01 6.61510587e-01 5.56913912e-01 4.29785043e-01
5.29151261e-01 -1.94247309e-02 7.64144123e-01 2.62887299e-01
2.32534915e-01 -1.23374283e-01 2.68370718e-01 4.71264333e-01
1.74123436e-01 -2.76925862e-01 -1.26194310e+00 4.19823200e-01
-1.95382786e+00 -4.36614901e-01 1.73020691e-01 2.17383361e+00
1.15073955e+00 6.78665340e-01 -3.67586106e-01 3.92471142e-02
8.04799497e-01 -3.60989362e-01 -7.37492085e-01 9.36748832e-03
2.61349261e-01 -2.13567168e-01 7.17504442e-01 5.19845188e-01
-5.79081416e-01 8.81376445e-01 6.94523335e+00 1.13043952e+00
-8.89333665e-01 2.40904883e-01 1.19647765e+00 6.64127469e-02
-7.21961379e-01 -3.53045389e-02 -4.69540566e-01 5.01146495e-01
1.03226352e+00 -1.36028692e-01 7.69338235e-02 7.40380704e-01
3.02521944e-01 -6.33607984e-01 -1.00946498e+00 4.83776987e-01
-3.52691375e-02 -1.22846818e+00 -1.83281824e-01 9.72393006e-02
8.57731521e-01 4.32864502e-02 2.98664689e-01 -1.18812323e-01
8.25608969e-01 -1.23466086e+00 1.08799684e+00 1.10090983e+00
4.55431908e-01 -9.85469162e-01 1.13656962e+00 3.25262129e-01
-2.54614651e-01 1.23262957e-01 -2.02812076e-01 6.64318025e-01
1.14562690e-01 1.36424994e+00 -1.24540365e+00 4.47646648e-01
5.80981374e-01 1.89172432e-01 -4.44896042e-01 9.65319037e-01
-6.93606853e-01 8.32214355e-01 -5.53350985e-01 1.35708570e-01
-1.63905621e-02 5.70893176e-02 4.50557321e-01 6.68554723e-01
3.22878897e-01 -1.79022640e-01 1.29411310e-01 1.25537896e+00
-1.00647815e-01 -4.28941905e-01 -2.10640147e-01 -6.30054176e-02
5.87739766e-01 1.22445166e+00 -1.25411880e+00 -2.80546606e-01
2.83143491e-01 3.37413520e-01 4.17959988e-02 3.50021034e-01
-1.08417988e+00 4.80717458e-02 1.49774536e-01 -2.27880225e-01
-1.96301639e-01 9.10822153e-02 -9.88369942e-01 -9.99210715e-01
-1.70890301e-01 -7.02648759e-01 9.37092230e-02 -7.74277925e-01
-1.34922051e+00 4.85343426e-01 4.79775891e-02 -7.36267626e-01
-6.75201118e-02 -5.19463837e-01 -5.07212400e-01 5.32917202e-01
-1.23235059e+00 -1.15154171e+00 -7.83638358e-02 5.18180728e-01
-1.15734980e-01 2.99970526e-02 7.06179619e-01 -2.59447873e-01
-7.38454223e-01 7.93538511e-01 -1.04005560e-01 1.62999973e-01
8.06266725e-01 -1.14049816e+00 -3.48414928e-01 7.72443831e-01
-6.06624126e-01 5.18216908e-01 9.20427442e-01 -1.04442203e+00
-8.42564464e-01 -8.87441695e-01 2.26060197e-01 -4.48548555e-01
7.95977116e-01 3.58225480e-02 -7.56726980e-01 5.30887008e-01
-8.72483477e-02 9.63881165e-02 8.84629965e-01 -9.36853662e-02
-8.67364928e-02 -8.67088661e-02 -1.81120074e+00 5.46022654e-01
3.76883298e-01 -3.20600361e-01 -2.69607991e-01 2.37878844e-01
1.00899220e+00 -5.35666108e-01 -1.16667652e+00 6.37059450e-01
5.09491265e-01 -8.54580700e-01 5.00647306e-01 -1.42074544e-02
2.16192350e-01 -2.76781917e-01 -3.35519522e-01 -1.31973672e+00
1.73352152e-01 -2.50864863e-01 7.34698698e-02 1.14214611e+00
7.88588226e-01 -7.87829995e-01 9.24405575e-01 1.52092063e+00
-1.90848783e-01 -7.54783452e-01 -1.22787750e+00 -3.91279608e-01
2.54704446e-01 -9.74660337e-01 7.91082442e-01 6.14998639e-01
5.45621552e-02 -4.60763335e-01 1.48812935e-01 4.18902814e-01
1.06497490e+00 -3.98102671e-01 4.58282754e-02 -1.27457416e+00
-4.74787354e-02 -5.17830551e-01 -3.99292707e-01 -2.29050964e-01
4.79660481e-01 -7.58611202e-01 5.22135377e-01 -1.28420508e+00
2.00049251e-01 -4.95138586e-01 -3.33837956e-01 6.01208746e-01
1.87469050e-01 -5.62027730e-02 -2.02514812e-01 1.27602220e-01
-5.63602686e-01 4.99486595e-01 1.19289172e+00 -2.55803466e-01
1.00986622e-01 2.34523669e-01 -8.72695625e-01 1.17107368e+00
8.16570103e-01 -6.82362795e-01 -5.62266350e-01 -3.48708600e-01
5.14882267e-01 3.01110178e-01 4.87596124e-01 -6.75952375e-01
6.31221473e-01 -1.37884542e-01 4.38939065e-01 -4.65716898e-01
-7.74685480e-03 -1.01781511e+00 3.22275348e-02 3.03813547e-01
-4.24916625e-01 -5.85290253e-01 -2.91823465e-02 7.52007902e-01
-1.58087201e-02 -6.41114891e-01 7.12007225e-01 1.14090428e-01
-2.62657069e-02 3.66257191e-01 -4.55983788e-01 -2.10387141e-01
1.12057126e+00 -1.13432921e-01 -4.03060317e-01 -3.39736909e-01
-9.55382824e-01 6.68290138e-01 5.02778471e-01 -2.22498313e-01
7.20024049e-01 -9.91180718e-01 -5.46835244e-01 1.04328252e-01
-1.30888730e-01 5.31259954e-01 1.76922351e-01 1.11257315e+00
-2.49740556e-01 6.12532124e-02 1.29362836e-01 -8.88859451e-01
-6.23409271e-01 1.36435166e-01 6.09861672e-01 -2.02023417e-01
-1.16178282e-01 7.20410883e-01 -3.30689222e-01 -3.71675700e-01
5.41739166e-01 -5.98918736e-01 -9.29934531e-02 2.71095186e-02
1.76543891e-01 2.90726483e-01 3.02827395e-02 -2.89457828e-01
-2.47187912e-01 7.27138715e-03 -1.88498095e-01 -6.68661475e-01
1.01235318e+00 -2.34822199e-01 -5.12460828e-01 2.93823272e-01
3.73877168e-01 -1.79667905e-01 -1.38479197e+00 -8.87696967e-02
-2.42750812e-03 -1.77052438e-01 6.88665092e-01 -1.13671815e+00
-1.43048036e+00 5.22886574e-01 5.89844882e-01 8.81563947e-02
9.06790793e-01 -5.69542032e-03 5.72850227e-01 3.36915672e-01
6.32906020e-01 -1.34541452e+00 -1.97900817e-01 2.58856446e-01
5.21309197e-01 -1.26892030e+00 -6.64088950e-02 -3.43069732e-01
-7.72955775e-01 1.00166678e+00 5.63287079e-01 9.35941935e-02
1.06093597e+00 7.17708647e-01 1.11202583e-01 -2.01818734e-01
-5.44060111e-01 3.69867265e-01 1.41174570e-01 5.56379557e-01
1.82307772e-02 4.47777480e-01 -1.00568600e-01 1.28801739e+00
-2.30426729e-01 3.58379185e-01 5.92250586e-01 6.64490879e-01
-5.44961512e-01 -8.24006379e-01 -3.74525219e-01 4.02341247e-01
-6.72022939e-01 7.32940063e-02 -3.13478529e-01 3.20995539e-01
-2.10577529e-02 1.13173914e+00 -2.20390081e-01 -3.36421460e-01
-1.76448360e-01 -3.77527700e-04 2.01086387e-01 -4.04192209e-01
-3.52010638e-01 1.43684790e-01 -2.92318147e-02 -5.50349712e-01
-3.27304989e-01 -5.82706988e-01 -1.82974863e+00 -2.41307810e-01
-5.87182283e-01 2.08339676e-01 9.62522388e-01 1.14423192e+00
5.03603555e-02 6.61930561e-01 4.45378035e-01 -4.17788267e-01
-1.00990546e+00 -6.08458757e-01 -7.12479353e-01 -2.34651819e-01
2.36291289e-02 -4.18232292e-01 -4.99067307e-01 -3.59539181e-01] | [14.405646324157715, -2.088986873626709] |
c43171cf-7f7c-4d35-8978-8c9f06edc65f | single-stage-uav-detection-and-classification | null | null | https://scholar.google.com/citations?view_op=view_citation&hl=en&user=ZztEI20AAAAJ&citation_for_view=ZztEI20AAAAJ:qUcmZB5y_30C | https://ieeexplore.ieee.org/iel7/9663572/9663735/09663841.pdf | Single-stage uav detection and classification with yolov5: Mosaic data augmentation and panet | In Drone-vs-Bird Detection Challenge in conjunction with the 4th International Workshop on Small-Drone Surveillance, Detection and Counteraction Techniques at IEEE AVSS 2021, we proposed a YOLOV5-based object detection model for small UAV detection and classification. YOLOV5 leverages PANet neck and mosaic augmentation which help in improving detection of small objects. We have combined the challenge dataset with one of the publicly available UAV air to air dataset having complex background and lighting conditions for training the model. The proposed approach achieved 0.96 Recall, 0.98 mAP0.5, and 0.71 mAP0.5:0.95 on the 10% randomly sampled dataset from the whole dataset. | ['Miodrag Bolic', 'Varun Mehta', 'Vaibhav Patel', 'Fardad Dadboud'] | 2021-11-16 | null | null | null | 2021-17th-ieee-international-conference-on | ['2d-object-detection'] | ['computer-vision'] | [-4.80470990e-05 -2.72894382e-01 7.83190429e-02 1.41841723e-02
4.16468866e-02 -1.14242876e+00 5.59760153e-01 -8.61894526e-03
-4.85418111e-01 2.91360170e-01 -4.03163195e-01 1.42270382e-02
6.97432235e-02 -7.06569791e-01 -5.79601288e-01 -3.73858601e-01
-4.91700888e-01 -1.60823151e-01 8.98450315e-01 -4.64910090e-01
-2.67488033e-01 6.40294611e-01 -1.56069171e+00 2.29300454e-01
2.28944138e-01 1.23332393e+00 -5.30840121e-02 1.56015682e+00
1.23552537e+00 6.62606895e-01 -1.09158933e+00 -2.33500853e-01
1.16922128e+00 1.59585088e-01 -3.76201682e-02 1.09242298e-01
1.25750077e+00 -6.66012764e-01 -3.42733085e-01 1.00251091e+00
3.98957163e-01 2.63724446e-01 3.71898890e-01 -1.64178133e+00
-2.18767285e-01 -8.94735828e-02 -1.00000763e+00 9.31945503e-01
3.55025232e-01 4.91016477e-01 6.01612747e-01 -6.88474298e-01
3.19989473e-01 9.67583477e-01 9.33361411e-01 4.02199835e-01
-9.42977786e-01 -1.07943714e+00 9.97171104e-02 -1.95449889e-01
-1.68071496e+00 -1.34808108e-01 -1.41844004e-01 -6.65827334e-01
1.10032046e+00 4.63048995e-01 1.03722239e+00 7.01300323e-01
8.13190639e-02 5.50357163e-01 7.82142282e-01 -7.61287659e-02
-4.44856405e-01 3.23001564e-01 8.64917561e-02 1.06228352e+00
5.02065718e-01 8.93497765e-01 -1.15478262e-01 -1.91846490e-01
4.56566811e-01 1.28320634e-01 -2.74297982e-01 -1.89513400e-01
-1.27845478e+00 6.84714079e-01 8.47964466e-01 -4.25184697e-01
-2.16364592e-01 9.78605002e-02 2.66612440e-01 3.44912291e-01
2.49131694e-01 6.62189305e-01 -3.46019179e-01 3.90916228e-01
-1.01574123e+00 5.63177586e-01 4.85206544e-01 1.51310623e+00
1.34101763e-01 3.49287570e-01 -2.76809752e-01 3.16818804e-01
3.45107913e-01 1.07691193e+00 -2.65295416e-01 -7.81502604e-01
3.53705794e-01 6.18075669e-01 7.11368084e-01 -1.23116565e+00
-3.32393706e-01 -6.04346216e-01 -3.46625328e-01 4.25528288e-01
5.14209678e-04 -6.45950377e-01 -1.17387080e+00 1.31013358e+00
5.21930814e-01 4.56616729e-01 6.20629676e-02 1.04937327e+00
9.82226133e-01 9.10701752e-01 -1.39156938e-01 1.73903078e-01
1.20173109e+00 -1.24515486e+00 -3.85462344e-01 -4.40191239e-01
6.46688819e-01 -8.07393849e-01 2.70693064e-01 6.80391431e-01
-3.78772020e-01 -7.36724138e-01 -1.50580239e+00 6.96423769e-01
-5.05459666e-01 8.26234579e-01 2.06204042e-01 8.81930411e-01
-1.09784043e+00 -1.31733552e-01 -4.62183565e-01 -7.38248467e-01
3.90916914e-01 3.53346288e-01 -2.93766737e-01 -8.34580362e-02
-7.13898838e-01 8.95141602e-01 6.06718421e-01 7.84758478e-02
-1.83743119e+00 -4.67842638e-01 -6.05187058e-01 -4.21026230e-01
8.43864501e-01 -3.64922523e-01 1.20600140e+00 -3.80756229e-01
-6.11532092e-01 9.42371547e-01 6.34070754e-01 -1.12842584e+00
4.96588975e-01 -7.03965068e-01 -8.81808758e-01 1.50795370e-01
1.46451388e-02 1.14828086e+00 1.11765885e+00 -1.03285968e+00
-1.48324335e+00 -2.28813887e-01 5.52881122e-01 1.81095019e-01
2.15090781e-01 2.20405057e-01 -7.25094527e-02 -7.16613054e-01
-5.13148844e-01 -1.21806896e+00 -1.64824545e-01 -1.91137753e-02
-1.16573516e-02 1.98528156e-01 1.35521138e+00 -5.19847333e-01
1.16590285e+00 -2.08065867e+00 -2.76834726e-01 1.28851738e-02
5.44259787e-01 9.91598725e-01 -2.51504540e-01 3.37891519e-01
1.70480758e-01 -1.34860352e-01 2.08659157e-01 2.10485026e-01
-4.39593971e-01 -2.54250944e-01 -3.46513957e-01 6.95674598e-01
2.74811417e-01 3.82102191e-01 -9.89591181e-01 -9.96950269e-02
4.47896928e-01 1.97124630e-01 -5.36800504e-01 4.44610476e-01
-7.28335530e-02 -2.08340555e-01 -2.16352493e-02 1.15689754e+00
9.27797079e-01 5.44948518e-01 -5.35757542e-01 -2.08806619e-01
-4.78881299e-01 -4.55864519e-01 -1.02899253e+00 9.19883788e-01
1.85286373e-01 1.04890585e+00 3.92939866e-01 -1.01082012e-01
9.23040986e-01 2.14440584e-01 7.71574602e-02 2.43505672e-01
4.58287954e-01 -3.89077276e-01 2.76570469e-01 -1.09785348e-01
8.37092757e-01 3.59571040e-01 -4.83092293e-02 -2.10351616e-01
2.80297369e-01 -4.52284545e-01 5.59801161e-01 3.59621555e-01
1.21242702e+00 -2.21711203e-01 5.34217179e-01 -3.75729799e-01
2.90823728e-01 5.62043846e-01 3.59679669e-01 9.64727283e-01
-7.51382291e-01 4.33355868e-01 -5.97483329e-02 -6.67278767e-01
-6.27206564e-01 -8.65058959e-01 3.55672501e-02 1.09088290e+00
3.40193123e-01 -9.08855140e-01 -8.75386417e-01 -8.86141121e-01
-5.50957285e-02 5.92096984e-01 -7.08615303e-01 -3.05444270e-01
-7.10168704e-02 -9.64301169e-01 9.38417435e-01 3.89889419e-01
6.34407520e-01 -5.68244457e-01 -1.30889487e+00 -7.71293566e-02
4.85427290e-01 -1.48603570e+00 -4.19796556e-01 3.40559781e-02
-2.60631233e-01 -1.47236264e+00 -3.63111377e-01 -4.49119687e-01
5.48563540e-01 9.62065518e-01 7.94536710e-01 1.81183681e-01
-7.85975575e-01 6.05102956e-01 -6.33694172e-01 -9.89543617e-01
-1.13525473e-01 -3.04081738e-01 5.32831311e-01 -2.56021827e-01
2.96095073e-01 2.83783734e-01 -5.48068762e-01 7.60851324e-01
-6.38281822e-01 -3.73206139e-01 4.49948639e-01 6.77277803e-01
1.29049435e-01 2.61042416e-01 -1.73015490e-01 -3.28584075e-01
3.55372548e-01 -3.34875792e-01 -1.46399772e+00 1.24575362e-01
-4.25512969e-01 -8.96885395e-01 3.64778191e-01 -4.98947710e-01
-3.02838773e-01 3.30192208e-01 2.59403586e-01 -7.91195035e-01
-3.35673928e-01 -3.77998352e-01 1.53049529e-01 -6.33459449e-01
9.05987144e-01 -7.78962150e-02 -4.62461591e-01 -1.31347775e-02
2.03531131e-01 4.98530477e-01 6.58561110e-01 2.92814851e-01
1.09315360e+00 3.01725388e-01 -5.91447763e-02 -1.12820208e+00
-8.54123056e-01 -7.56888807e-01 -6.31734192e-01 -4.94908839e-01
9.15359378e-01 -1.34025455e+00 -5.06850362e-01 4.77032989e-01
-9.79397476e-01 -5.42530045e-02 -2.38920338e-02 4.41559374e-01
9.58444402e-02 -5.20910174e-02 -1.33951372e-02 -1.12855744e+00
-3.89881253e-01 -9.54361618e-01 1.14813638e+00 2.41469130e-01
8.75984579e-02 -3.34048450e-01 6.42114580e-02 3.03942919e-01
2.80157924e-01 6.39295101e-01 -5.10906696e-01 -5.28953254e-01
-6.54502213e-01 -9.88956928e-01 -3.85269314e-01 5.50813913e-01
-1.63559224e-02 4.34353590e-01 -8.37804496e-01 -8.08949471e-01
-3.34010273e-01 -2.25557044e-01 1.11116457e+00 2.20181048e-01
4.47188705e-01 -1.77716210e-01 -5.47158420e-01 7.98616588e-01
1.23611426e+00 4.56232905e-01 9.10388492e-03 1.01564869e-01
6.96946442e-01 1.50289133e-01 1.56485331e+00 6.17921412e-01
9.94424000e-02 5.71989536e-01 1.05969489e+00 -1.55085802e-01
6.52022567e-03 -2.04625338e-01 5.55304825e-01 -3.73533785e-01
-3.05737108e-01 -7.93394387e-01 -1.00866890e+00 3.92933816e-01
-1.50870657e+00 -1.07248092e+00 -2.46147767e-01 2.15405560e+00
-2.04609051e-01 3.34166795e-01 5.21430731e-01 -8.33195373e-02
7.57046521e-01 4.59466338e-01 -4.86739203e-02 2.46591084e-02
-6.98101372e-02 -4.80246276e-01 1.36948955e+00 4.04935211e-01
-1.96922898e+00 1.32386911e+00 6.46492863e+00 5.37755847e-01
-8.71582150e-01 4.87331897e-02 -9.26865563e-02 -3.19563240e-01
1.01413214e+00 -3.24780494e-01 -1.50508547e+00 2.25166842e-01
8.04833651e-01 -4.58199270e-02 4.68516737e-01 1.08816302e+00
-1.36137024e-01 -3.28202605e-01 -7.00039029e-01 8.32366943e-01
4.18478131e-01 -9.14096057e-01 -1.27621964e-01 -7.25600719e-02
8.52596819e-01 3.60664636e-01 1.54503062e-01 3.21844280e-01
5.93599319e-01 -9.75570977e-01 6.85702085e-01 3.55698690e-02
8.12182605e-01 -8.27053607e-01 1.18265498e+00 4.32247549e-01
-1.55005550e+00 -3.03518206e-01 -5.42684495e-01 -9.06069353e-02
-2.20826194e-01 -4.09556329e-01 -1.58125019e+00 7.49839023e-02
1.22773790e+00 8.53296101e-01 -9.83697832e-01 1.32562470e+00
-9.13276002e-02 6.18910730e-01 -6.45027280e-01 -6.62958026e-02
5.41164637e-01 2.11288020e-01 1.18977308e+00 1.39069927e+00
2.28038922e-01 3.37529004e-01 5.62094510e-01 2.56751388e-01
9.13260598e-03 -5.16812325e-01 -1.04392004e+00 -3.00448705e-02
4.26869512e-01 1.41022170e+00 -5.50674796e-01 -3.92073482e-01
-2.74462491e-01 7.04558134e-01 -3.89842749e-01 -1.37312431e-02
-1.29478800e+00 -3.09968710e-01 7.19284415e-01 2.46011123e-01
5.09962440e-01 -2.38784343e-01 6.70179784e-01 -9.10218537e-01
-5.19024551e-01 -9.97051179e-01 6.15383685e-01 -6.35318339e-01
-8.03745449e-01 1.03483582e+00 3.72370899e-01 -1.51769900e+00
6.57968074e-02 -8.01940084e-01 -4.33134615e-01 3.13832343e-01
-1.01928806e+00 -1.54248190e+00 -8.81272316e-01 3.64490122e-01
7.33806610e-01 -6.47760451e-01 5.86673498e-01 2.53355682e-01
-6.11393452e-01 6.10998929e-01 -4.72770274e-01 1.15802318e-01
6.75854325e-01 -1.11689222e+00 7.49333858e-01 1.50648642e+00
6.96064159e-02 2.72156775e-01 6.88676953e-01 -7.24188745e-01
-1.37687266e+00 -1.70991194e+00 -1.52450874e-01 -9.25121784e-01
7.52093554e-01 -5.39760590e-01 -1.61286891e-01 9.96981263e-01
3.18242937e-01 4.74031746e-01 2.96506971e-01 -6.67914689e-01
1.20504238e-01 -2.69421399e-01 -1.36007166e+00 2.34220758e-01
9.58079815e-01 1.36329859e-01 -4.92654324e-01 4.46988344e-01
8.35370898e-01 -7.30130315e-01 -6.37170494e-01 8.30361009e-01
6.22810245e-01 -8.63416374e-01 1.15838015e+00 -5.33961535e-01
-4.36945319e-01 -9.48017418e-01 -5.15486240e-01 -1.25446987e+00
-4.65193242e-01 -6.53668344e-01 -9.81900096e-02 8.08995187e-01
3.97554040e-01 -2.23404408e-01 5.06196499e-01 -1.06966771e-01
-2.03226998e-01 -1.06622629e-01 -5.44769585e-01 -9.48885381e-01
-8.93349707e-01 -1.80793464e-01 2.96185702e-01 4.34477419e-01
-8.23184609e-01 1.61287948e-01 -1.07368290e+00 9.85402167e-01
7.31266975e-01 -1.62184939e-01 1.36413741e+00 -1.09144998e+00
-3.86339575e-02 2.58275092e-01 -9.03837621e-01 -9.27856147e-01
-5.75579405e-01 -3.51544887e-01 4.60724458e-02 -9.33961451e-01
-1.83718681e-01 5.02380682e-03 -2.60519654e-01 4.12921309e-01
-1.25356928e-01 6.80218279e-01 4.41430449e-01 -9.75863636e-02
-9.75282252e-01 1.12102352e-01 6.92134857e-01 -1.41551167e-01
-5.89926764e-02 5.18596172e-01 -1.42625794e-01 7.76296973e-01
8.69442761e-01 -7.15895772e-01 -2.87775040e-01 -2.30572745e-01
-2.40279466e-01 -2.20070288e-01 8.54973435e-01 -1.73262978e+00
1.23629935e-01 2.61751302e-02 6.79095984e-01 -1.09711730e+00
6.54155016e-01 -1.28631902e+00 -6.89340234e-02 9.82817233e-01
8.85647759e-02 3.80242229e-01 8.89026284e-01 7.98720598e-01
-1.25482395e-01 1.02687471e-01 1.02816606e+00 -1.11373968e-01
-1.13729298e+00 4.05759662e-01 -6.32200241e-01 -5.80473654e-02
1.74150026e+00 -1.02747358e-01 -5.38085520e-01 6.85886666e-03
-4.14730161e-01 6.06588066e-01 4.58885819e-01 7.32622564e-01
7.82159388e-01 -8.46725643e-01 -8.57130051e-01 5.09262085e-01
5.09516180e-01 -2.71684259e-01 -1.18247919e-01 7.22480536e-01
-9.41021323e-01 5.53340435e-01 -5.11029601e-01 -8.16665411e-01
-2.03605318e+00 5.94305456e-01 4.97032017e-01 2.36473352e-01
-6.08156137e-02 1.34893560e+00 2.76565433e-01 -2.68293440e-01
2.61003405e-01 -5.08442760e-01 -1.54487982e-01 3.79441045e-02
7.22617865e-01 7.16001391e-01 -2.34068304e-01 -8.57878625e-01
-6.59977853e-01 4.13248837e-01 -6.86317729e-03 3.13613683e-01
1.09070098e+00 1.17825344e-01 3.91156197e-01 -2.15504378e-01
4.23962355e-01 1.08206570e-01 -1.34185922e+00 3.18806708e-01
-2.03006014e-01 -8.02779496e-01 5.09869158e-01 -9.75248575e-01
-8.13557267e-01 6.73156679e-01 1.20979977e+00 4.01283622e-01
1.11246943e+00 -3.37282538e-01 6.17187917e-01 5.66528976e-01
3.76238763e-01 -5.86655080e-01 -2.82962769e-01 3.83014351e-01
9.82079208e-01 -1.66707253e+00 3.72693807e-01 -5.57882607e-01
-7.27739513e-01 8.61817360e-01 1.15765321e+00 -5.34262061e-01
2.30797872e-01 4.94434923e-01 7.44574741e-02 -3.51746857e-01
-9.49219406e-01 -4.48787779e-01 5.73889256e-01 9.27007854e-01
-2.48639330e-01 3.28672677e-01 3.56562108e-01 2.63378263e-01
-1.61113232e-01 -2.83031136e-01 4.27060932e-01 1.10523009e+00
-7.40069747e-01 -3.31064463e-01 -6.70337915e-01 3.70604545e-01
-3.39062542e-01 -1.88864559e-01 -1.03591824e+00 1.18810046e+00
6.52450442e-01 1.16402793e+00 -2.34212622e-01 -1.11154521e+00
6.76379442e-01 -8.09169888e-01 3.17372084e-01 -7.20952094e-01
-9.07270133e-01 -6.00537211e-02 3.34962487e-01 -5.73892951e-01
-2.85767704e-01 -3.75694275e-01 -5.46660841e-01 -1.25072777e-01
-8.64053547e-01 6.22569919e-02 3.21593672e-01 1.32040605e-01
2.08652258e-01 4.01337326e-01 4.61103439e-01 -8.61755133e-01
-4.10554826e-01 -9.79001284e-01 -5.65229654e-01 -4.94710833e-01
8.63388360e-01 -8.78298163e-01 -4.80258495e-01 2.50381660e-02] | [8.556339263916016, -0.9421209692955017] |
94382065-f888-4459-ac09-96a655e4672c | ell-p-slack-norm-support-vector-data | 2203.08932 | null | https://arxiv.org/abs/2203.08932v1 | https://arxiv.org/pdf/2203.08932v1.pdf | $\ell_p$ Slack Norm Support Vector Data Description | The support vector data description (SVDD) approach serves as a de facto standard for one-class classification where the learning task entails inferring the smallest hyper-sphere to enclose target objects while linearly penalising any errors/slacks via an $\ell_1$-norm penalty term. In this study, we generalise this modelling formalism to a general $\ell_p$-norm ($p\geq1$) slack penalty function. By virtue of an $\ell_p$ slack norm, the proposed approach enables formulating a non-linear cost function with respect to slacks. From a dual problem perspective, the proposed method introduces a sparsity-inducing dual norm into the objective function, and thus, possesses a higher capacity to tune into the inherent sparsity of the problem for enhanced descriptive capability. A theoretical analysis based on Rademacher complexities characterises the generalisation performance of the proposed approach in terms of parameter $p$ while the experimental results on several datasets confirm the merits of the proposed method compared to other alternatives. | ['Shervin R. Arashloo'] | 2022-03-16 | null | null | null | null | ['one-class-classification'] | ['miscellaneous'] | [ 3.72666180e-01 3.93212497e-01 -1.85210153e-01 -2.34337494e-01
-5.57990789e-01 -2.38292500e-01 5.12060106e-01 2.63803631e-01
-5.58372557e-01 7.59349704e-01 -2.51784205e-01 -1.20691746e-01
-7.52655864e-01 -5.63715756e-01 -4.40957218e-01 -1.05372810e+00
-3.48371774e-01 2.07091704e-01 -2.34156638e-01 -1.75421104e-01
3.68914813e-01 4.33131754e-01 -1.78013909e+00 -2.34921709e-01
9.81499910e-01 1.57601666e+00 4.03661467e-02 3.41763534e-02
2.46333018e-01 3.78862500e-01 -3.26101184e-01 -4.00405556e-01
6.87942624e-01 -7.65973181e-02 -6.06985092e-01 2.57117093e-01
3.07557374e-01 3.28826487e-01 1.84767604e-01 1.14500690e+00
2.74770886e-01 3.22346717e-01 9.73304391e-01 -1.06721151e+00
-1.59383491e-01 2.06987076e-02 -6.44456685e-01 2.90924579e-01
4.81429100e-01 6.01503588e-02 1.38707185e+00 -1.15113521e+00
4.01174664e-01 6.50150359e-01 7.29034841e-01 -3.75442728e-02
-1.49547112e+00 -2.82438368e-01 1.00188695e-01 2.95326784e-02
-1.66287541e+00 -3.39940429e-01 1.01241040e+00 -6.42519534e-01
6.46578610e-01 5.72395742e-01 4.78862882e-01 6.38382971e-01
8.12678263e-02 6.64390802e-01 1.26432276e+00 -4.62665170e-01
3.18751812e-01 6.79154515e-01 1.43859431e-01 6.49562180e-01
4.90793437e-01 -4.14865538e-02 -1.88562840e-01 1.49326900e-03
5.47887802e-01 -3.26304793e-01 -3.13369274e-01 -8.58412683e-01
-9.15088356e-01 9.50441778e-01 3.50618422e-01 4.67880994e-01
-3.71701807e-01 -2.40843907e-01 5.38261354e-01 2.09000949e-02
5.89638412e-01 4.59849596e-01 -2.14835908e-02 9.98815894e-02
-1.13224566e+00 2.87896991e-01 6.64712369e-01 8.14147890e-01
5.44723809e-01 3.71501058e-01 6.00033775e-02 8.37777376e-01
4.29549128e-01 3.01296026e-01 3.64638329e-01 -6.12446606e-01
6.29982769e-01 9.30826604e-01 1.67747229e-01 -1.44791377e+00
-3.34538549e-01 -8.87367487e-01 -9.87559676e-01 3.80645633e-01
3.25844467e-01 2.18261793e-01 -1.87655017e-01 1.57219589e+00
4.67121273e-01 3.65140438e-02 1.39142454e-01 9.94433582e-01
4.32302058e-01 3.34515154e-01 -9.40488577e-02 -6.01671994e-01
1.12104273e+00 -4.11216170e-01 -4.96661603e-01 -4.60261703e-02
4.32070196e-01 -6.21034920e-01 1.08043957e+00 7.30952621e-01
-1.20138574e+00 -5.34986496e-01 -1.15751457e+00 4.44757134e-01
-1.74635231e-01 3.73079240e-01 4.59459692e-01 7.34550536e-01
-7.34462678e-01 3.98434669e-01 -4.23376352e-01 -9.34876502e-03
2.88005739e-01 5.69790900e-01 -4.63730633e-01 1.19051404e-01
-1.02987504e+00 9.01734233e-01 8.30664039e-01 3.93594980e-01
-3.03696096e-01 -7.46403158e-01 -9.12836790e-01 8.21077526e-02
4.69405979e-01 -2.84010410e-01 4.17419642e-01 -8.49616289e-01
-1.31948400e+00 1.03403795e+00 2.17926919e-01 -4.69161987e-01
8.92365515e-01 1.29942584e-03 -2.59172291e-01 3.14684331e-01
2.88522919e-03 -4.92752064e-03 1.07180560e+00 -1.35589683e+00
-3.73066843e-01 -4.49065030e-01 2.02149898e-01 3.03487539e-01
-4.65783209e-01 -3.40157121e-01 8.51327181e-02 -9.17411327e-01
2.76736170e-01 -6.60643935e-01 -2.24954545e-01 -1.62591025e-01
-3.13505590e-01 -1.94205388e-01 5.48480868e-01 -3.20991933e-01
1.60944200e+00 -2.32406807e+00 4.55092311e-01 7.76867151e-01
2.59177417e-01 3.72686714e-01 2.08385319e-01 3.36793840e-01
-3.38862240e-01 -3.84771079e-01 -6.25293911e-01 -1.91189960e-01
8.55391845e-02 1.78688303e-01 -1.42098991e-02 1.03596079e+00
1.76513523e-01 1.46016553e-01 -7.61451364e-01 -4.46773559e-01
3.76005083e-01 5.29972136e-01 -6.29531026e-01 -8.56667608e-02
5.85258156e-02 7.45153800e-02 -4.49128062e-01 3.69952619e-01
6.59366786e-01 -2.09524959e-01 9.79122594e-02 -9.43342969e-02
-2.30486006e-01 -2.91405201e-01 -2.02047133e+00 1.46500373e+00
-4.00421292e-01 2.72294469e-02 5.26506484e-01 -1.87889016e+00
1.19765067e+00 1.81908011e-01 9.64928746e-01 -3.69770437e-01
3.64452869e-01 3.85114938e-01 -2.33912140e-01 -3.53507698e-01
4.02998656e-01 -4.47150052e-01 -1.18879244e-01 1.02294959e-01
1.15737710e-02 -3.31279263e-03 3.74570400e-01 -2.19318017e-01
6.39209092e-01 3.98763046e-02 6.79617882e-01 -8.96864831e-01
1.26860690e+00 -1.01164319e-01 5.31296551e-01 2.75655776e-01
-5.57700992e-02 1.00403816e-01 5.34768343e-01 -3.48648220e-01
-8.20984304e-01 -8.80640030e-01 -7.00548112e-01 6.42937481e-01
8.77451822e-02 -6.21065423e-02 -6.35134399e-01 -5.53002238e-01
2.23262444e-01 8.13447833e-01 -6.17612898e-01 -4.59119752e-02
-5.45561254e-01 -1.00928760e+00 3.81433696e-01 2.38460332e-01
3.23419720e-01 -5.66492260e-01 -8.98992956e-01 1.94507420e-01
1.40308648e-01 -7.74395943e-01 -1.63749680e-01 5.26716411e-01
-7.91631103e-01 -1.09772313e+00 -7.59220898e-01 -6.56535208e-01
7.94444859e-01 -1.06232785e-01 7.28538990e-01 -1.79392055e-01
-4.62373972e-01 4.59715903e-01 -4.08195108e-01 -2.49584302e-01
-1.86123371e-01 -1.65183589e-01 3.76589954e-01 4.63722259e-01
2.22808778e-01 -4.66668874e-01 -7.10079730e-01 3.80430520e-01
-9.80765224e-01 -3.36066484e-01 5.30421913e-01 1.04326069e+00
7.86894083e-01 2.39531144e-01 6.00114346e-01 -6.92454278e-01
5.26815474e-01 -5.84923267e-01 -8.45391631e-01 -4.29599360e-02
-1.10070002e+00 2.79843993e-02 9.09057260e-01 -6.10604770e-02
-7.98615515e-01 8.58038068e-02 -8.60932246e-02 -3.80045950e-01
3.37955356e-02 6.31993353e-01 -2.13065892e-01 -3.34179580e-01
6.31758988e-01 5.41387200e-01 2.93726653e-01 -4.08695906e-01
3.03237021e-01 3.64462107e-01 3.08570296e-01 -6.71250701e-01
7.15129673e-01 3.90219986e-01 5.74903190e-01 -9.77375627e-01
-6.61281407e-01 -6.37197912e-01 -6.30267978e-01 -1.60433620e-01
5.64031541e-01 -6.36995971e-01 -1.07439089e+00 4.83664498e-02
-5.12048364e-01 3.28675896e-01 -7.49814212e-01 5.49757421e-01
-8.92525792e-01 6.02239370e-01 1.54385462e-01 -9.77779210e-01
-1.40888393e-01 -1.19155312e+00 6.54127717e-01 -2.26515725e-01
-4.09746543e-02 -1.13292444e+00 -1.56223390e-03 4.20704424e-01
1.31679654e-01 6.83088362e-01 8.88196290e-01 -6.92502320e-01
-1.73855294e-02 -4.52727437e-01 -1.80237204e-01 8.51993144e-01
-2.59648502e-01 -3.86082590e-01 -8.25118363e-01 -5.50014973e-01
5.74356675e-01 -9.30531323e-02 6.15333974e-01 3.19008052e-01
9.09178972e-01 -7.24241734e-01 7.42135744e-04 5.60360610e-01
2.00268292e+00 2.32924987e-02 1.98501572e-01 3.27817053e-01
3.13357651e-01 6.23356044e-01 7.44081020e-01 9.47801173e-01
1.16589397e-01 7.43318379e-01 7.95353889e-01 -3.88689712e-02
3.11106145e-01 3.40186507e-01 9.41582993e-02 6.13945603e-01
-2.35914394e-01 1.20482497e-01 -8.49759817e-01 5.09549499e-01
-1.69793272e+00 -9.07130420e-01 -2.57676154e-01 2.44995737e+00
4.61808115e-01 3.66305590e-01 1.08112670e-01 9.15785670e-01
4.14429069e-01 2.45524600e-01 -3.10128897e-01 -5.46027541e-01
-6.01688661e-02 2.60921180e-01 5.41247606e-01 4.90101635e-01
-1.16043556e+00 1.99410811e-01 5.50184917e+00 1.13096857e+00
-1.15120661e+00 -2.40353569e-01 3.76580328e-01 -2.21491426e-01
-1.83419824e-01 -3.53221416e-01 -7.60198772e-01 6.19757116e-01
6.51275039e-01 -1.88273385e-01 2.16984615e-01 1.06200707e+00
3.02402020e-01 -3.70798707e-01 -1.07408452e+00 1.08399415e+00
3.69209409e-01 -1.03630507e+00 -2.00111240e-01 3.02997738e-01
5.35024166e-01 -5.86746395e-01 3.70891750e-01 1.19267635e-01
-3.53551716e-01 -9.57739294e-01 8.43545198e-01 3.23747188e-01
8.44676435e-01 -8.74797225e-01 7.82831192e-01 4.63164777e-01
-1.29015052e+00 -2.70382315e-01 -1.02237821e-01 -2.07988769e-01
-3.33689228e-02 7.15596259e-01 -8.24421167e-01 8.47233653e-01
3.28583628e-01 6.54890418e-01 -2.91877627e-01 7.00849712e-01
1.39827400e-01 1.94683895e-01 -3.59411210e-01 2.97705308e-02
4.11635637e-01 -6.83429897e-01 9.60154295e-01 1.31935394e+00
2.60013044e-01 9.62367803e-02 3.02653134e-01 7.22030878e-01
3.56562674e-01 6.41506791e-01 -5.03563702e-01 2.01453343e-01
1.02557763e-01 9.48034227e-01 -6.94694042e-01 -3.21074687e-02
-4.27674919e-01 7.20425487e-01 3.65806073e-02 7.00604618e-02
-6.56924248e-01 -4.61345971e-01 6.70605838e-01 2.74803847e-01
4.20317113e-01 -1.37987122e-01 -5.86235583e-01 -1.09314489e+00
5.10036767e-01 -6.95174336e-01 4.51229066e-01 -6.11508749e-02
-1.22566152e+00 6.02552652e-01 1.39223441e-01 -1.62636685e+00
-8.94783959e-02 -8.32967222e-01 -2.05965623e-01 1.01412296e+00
-1.58207035e+00 -9.97545481e-01 4.96372953e-02 6.31869018e-01
2.55277246e-01 -1.79683194e-01 8.62265050e-01 3.81004900e-01
-7.60814369e-01 6.32334650e-01 4.53725040e-01 -2.24822477e-01
1.74001142e-01 -1.33674550e+00 -7.04511642e-01 8.33328009e-01
-3.16887110e-01 5.48596561e-01 1.05211842e+00 -1.33740023e-01
-1.24398637e+00 -7.18972564e-01 7.06643462e-01 -8.20126384e-02
6.83856547e-01 -1.33593038e-01 -6.91724718e-01 3.06495965e-01
-2.04383060e-01 2.93240935e-01 8.73560727e-01 -2.83949613e-03
-1.30768001e-01 -3.29146951e-01 -1.52665484e+00 2.77610011e-02
6.76482797e-01 -2.63926536e-01 -6.48862958e-01 1.15431063e-01
6.56740367e-02 -2.08197117e-01 -1.36090612e+00 4.94624197e-01
4.20938641e-01 -1.09715796e+00 1.21842217e+00 -2.61554062e-01
2.86031425e-01 -2.87714571e-01 -3.24448496e-01 -9.68679428e-01
-2.14523405e-01 -3.29014570e-01 -3.56642269e-02 9.89731193e-01
2.14119852e-01 -6.86436176e-01 8.73611569e-01 5.88667631e-01
-1.32276356e-01 -1.28377759e+00 -1.47029817e+00 -9.86450195e-01
4.17717807e-02 -4.64703083e-01 2.48448923e-02 8.60480905e-01
1.96152061e-01 -1.07893042e-01 -2.51952976e-01 -2.26019323e-02
9.34117496e-01 -7.65908584e-02 3.80081475e-01 -1.42197919e+00
-4.33544129e-01 -6.93023682e-01 -7.38493383e-01 -7.73140907e-01
1.53790534e-01 -1.20600510e+00 -1.40490383e-01 -8.97377789e-01
-2.53723979e-01 -7.61575341e-01 -5.84174037e-01 7.26157576e-02
4.63030078e-02 2.52540439e-01 1.67428762e-01 1.77464589e-01
-3.61605942e-01 5.76198757e-01 1.02221799e+00 -3.04782875e-02
-1.40972063e-01 1.55231506e-01 -7.00349867e-01 8.91815066e-01
3.88446748e-01 -2.04236403e-01 -4.77791488e-01 -3.69154066e-02
3.69801670e-01 9.71972719e-02 2.86616445e-01 -9.39241052e-01
-1.40371069e-01 -1.91519544e-01 1.25341445e-01 -2.45837197e-01
4.18152392e-01 -1.30348194e+00 7.09289089e-02 5.81093788e-01
-2.70316333e-01 -2.11095437e-01 2.33747195e-02 5.36722422e-01
-3.51393998e-01 -5.35538316e-01 1.15393889e+00 -3.45606878e-02
-5.40437341e-01 -9.41439718e-02 -2.47010626e-02 1.21351972e-01
1.59166372e+00 -8.24336231e-01 3.41287762e-01 3.38066459e-01
-8.67166936e-01 -7.66810449e-03 3.04812908e-01 -8.50185528e-02
4.52490270e-01 -1.20434511e+00 -5.87603211e-01 4.14045841e-01
2.05241412e-01 -3.65493707e-02 2.82263637e-01 1.12067115e+00
-4.20410752e-01 6.87461913e-01 -2.13751093e-01 -7.09831357e-01
-1.11391103e+00 7.03602612e-01 5.06066903e-02 -5.09250522e-01
-5.95359802e-01 1.01760936e+00 -1.41080141e-01 -4.67577875e-02
4.35641944e-01 -1.91358626e-01 -4.55041140e-01 4.38063353e-01
2.11684436e-01 8.10779154e-01 2.89654166e-01 -9.63392735e-01
-5.90938568e-01 4.45611924e-01 2.29688838e-01 7.69693926e-02
1.47867930e+00 1.17867738e-02 7.89338574e-02 1.84516028e-01
1.28297091e+00 -1.08224787e-01 -1.24785233e+00 -2.95505553e-01
8.65712836e-02 -6.69441998e-01 1.05340309e-01 -5.07287741e-01
-8.06270301e-01 5.07764399e-01 5.70661366e-01 3.21199328e-01
1.20116651e+00 -3.03409189e-01 1.81862399e-01 7.30156824e-02
4.13165003e-01 -1.29516292e+00 -1.74609665e-02 2.15946734e-01
1.06154764e+00 -1.25678897e+00 1.22383475e-01 -5.30816793e-01
-4.89581823e-01 1.04199994e+00 2.93444067e-01 -4.49533969e-01
8.23456645e-01 5.98871335e-02 -3.78739387e-01 -1.77529171e-01
-1.17389455e-01 -5.56878522e-02 4.59300578e-01 3.99766952e-01
3.67213666e-01 7.32096955e-02 -1.00067556e+00 6.22726738e-01
-1.22015499e-01 -9.97271612e-02 2.52246261e-01 7.92006731e-01
-3.18295687e-01 -9.22591329e-01 -5.52841127e-01 5.20808399e-01
-4.03569877e-01 9.88981053e-02 1.73307061e-01 8.18619549e-01
5.20223737e-01 8.26909482e-01 -2.16351986e-01 1.32982107e-02
6.70791864e-01 1.05125479e-01 3.74989510e-01 -5.53836882e-01
-6.99533224e-01 4.77312468e-02 -2.13014975e-01 -4.82947081e-01
-6.65570498e-01 -1.06561613e+00 -9.91730928e-01 2.35615927e-03
-3.48471373e-01 3.62426102e-01 7.48904765e-01 8.43818545e-01
1.35070207e-02 3.66831571e-01 8.46335292e-01 -6.97911203e-01
-1.05717969e+00 -6.85473979e-01 -8.84962142e-01 4.71145630e-01
2.62308627e-01 -9.86149669e-01 -5.97074986e-01 1.54698296e-02] | [7.8457746505737305, 4.111073017120361] |
08ad331a-834d-4a76-ab43-0256aa5f0f40 | generating-natural-language-adversarial-3 | 2003.10388 | null | https://arxiv.org/abs/2003.10388v1 | https://arxiv.org/pdf/2003.10388v1.pdf | Generating Natural Language Adversarial Examples on a Large Scale with Generative Models | Today text classification models have been widely used. However, these classifiers are found to be easily fooled by adversarial examples. Fortunately, standard attacking methods generate adversarial texts in a pair-wise way, that is, an adversarial text can only be created from a real-world text by replacing a few words. In many applications, these texts are limited in numbers, therefore their corresponding adversarial examples are often not diverse enough and sometimes hard to read, thus can be easily detected by humans and cannot create chaos at a large scale. In this paper, we propose an end to end solution to efficiently generate adversarial texts from scratch using generative models, which are not restricted to perturbing the given texts. We call it unrestricted adversarial text generation. Specifically, we train a conditional variational autoencoder (VAE) with an additional adversarial loss to guide the generation of adversarial examples. Moreover, to improve the validity of adversarial texts, we utilize discrimators and the training framework of generative adversarial networks (GANs) to make adversarial texts consistent with real data. Experimental results on sentiment analysis demonstrate the scalability and efficiency of our method. It can attack text classification models with a higher success rate than existing methods, and provide acceptable quality for humans in the meantime. | ['Jun Zhou', 'Yankun Ren', 'Siliang Tang', 'Xiang Ren', 'Jianbin Lin', 'Yuan Qi', 'Shuang Yang'] | 2020-03-10 | null | null | null | null | ['adversarial-text'] | ['adversarial'] | [ 2.00020224e-01 2.12926954e-01 3.27144355e-01 -9.33139324e-02
-5.01561046e-01 -9.24734890e-01 9.06761885e-01 -2.78947502e-01
-1.17134564e-01 7.89998412e-01 4.91219945e-02 -2.41751298e-01
4.81124073e-01 -1.26155782e+00 -8.12869012e-01 -7.20057130e-01
6.62735939e-01 6.53302610e-01 -1.00810803e-01 -6.73820913e-01
-1.95990726e-01 9.96713266e-02 -1.17118120e+00 1.82186946e-01
1.18358731e+00 6.14855826e-01 -1.41703531e-01 7.83928573e-01
-2.10487634e-01 9.57245827e-01 -1.28682280e+00 -1.01918924e+00
2.12351203e-01 -7.21790254e-01 -5.96188188e-01 -1.02076963e-01
3.96432653e-02 -6.08384728e-01 -7.43741870e-01 1.21695817e+00
4.63827938e-01 2.36059278e-01 7.71124542e-01 -1.45538652e+00
-1.23606634e+00 8.12422216e-01 -2.04509929e-01 -3.37160170e-01
2.32775643e-01 2.56707698e-01 7.06410646e-01 -5.50889432e-01
4.00940120e-01 1.30465567e+00 5.10980368e-01 1.07978988e+00
-9.43796277e-01 -9.80281711e-01 1.87316224e-01 -2.18839407e-01
-9.94346559e-01 -1.43697485e-01 1.05457485e+00 -2.88594574e-01
3.91193420e-01 4.27681059e-01 4.96325552e-01 1.72207224e+00
3.54440123e-01 7.78069019e-01 5.76066315e-01 -1.82404593e-01
2.36175433e-01 2.84810185e-01 -5.48150003e-01 4.85609204e-01
6.22449741e-02 1.27823129e-01 1.00082643e-02 -2.89437324e-01
5.37227631e-01 4.00798529e-01 -3.40907782e-01 5.36207594e-02
-1.03232276e+00 1.29299176e+00 4.58827972e-01 2.40802497e-01
-1.75159931e-01 1.49025142e-01 5.59987247e-01 4.17086333e-01
6.13277197e-01 3.54452610e-01 -1.00093134e-01 1.00196317e-01
-5.33762991e-01 4.50555176e-01 7.92997062e-01 9.85884666e-01
3.04449081e-01 4.70707923e-01 -1.19662233e-01 7.89738655e-01
2.18086496e-01 8.18284094e-01 8.76746953e-01 -3.54909718e-01
6.46821976e-01 4.13936645e-01 1.36466667e-01 -1.09845126e+00
8.95216540e-02 -1.27423242e-01 -1.28440523e+00 2.18098670e-01
2.56896198e-01 -6.18573129e-01 -8.49462628e-01 1.69104946e+00
3.37168604e-01 3.98486815e-02 3.84570688e-01 4.91314501e-01
7.27393150e-01 1.05967009e+00 -3.62180881e-02 -7.83675686e-02
9.15233314e-01 -9.83936667e-01 -8.77907693e-01 -3.61925215e-01
4.29587513e-01 -8.15867066e-01 1.31330740e+00 3.41823339e-01
-9.42625761e-01 -5.76543152e-01 -1.00933087e+00 1.55883253e-01
-5.05906582e-01 -3.73926848e-01 2.97295243e-01 8.43805015e-01
-3.33103836e-01 4.61533457e-01 -6.80885434e-01 2.85564572e-01
5.49831390e-01 9.82708558e-02 -9.21268314e-02 2.88462862e-02
-1.60116029e+00 6.97890162e-01 4.82788563e-01 6.98665977e-02
-9.43383396e-01 -5.63645005e-01 -8.88729751e-01 -5.12422100e-02
1.07531965e-01 -7.06800163e-01 1.19840181e+00 -1.42533696e+00
-1.74524474e+00 4.47540283e-01 1.33939207e-01 -3.77498537e-01
1.07794321e+00 -1.46243066e-01 -5.29474258e-01 -7.22301304e-02
-1.37790874e-01 2.04338968e-01 1.32495904e+00 -1.32562971e+00
-3.44910353e-01 -4.83451448e-02 2.72526443e-01 6.00148849e-02
-8.02584827e-01 -2.89847493e-01 1.80834364e-02 -1.19580579e+00
-4.83851433e-01 -8.64017487e-01 -3.03585052e-01 -1.94881007e-01
-5.30945659e-01 -1.93742007e-01 1.17914391e+00 -6.29536808e-01
1.01994109e+00 -2.14193106e+00 1.38778716e-01 -4.43391241e-02
2.84918100e-01 5.36074042e-01 -1.10523835e-01 3.68217796e-01
-5.45284562e-02 6.08297944e-01 -4.51827049e-01 -3.66267323e-01
1.18911296e-01 3.89663935e-01 -9.54716384e-01 1.78435415e-01
1.92277193e-01 1.00791621e+00 -1.11660159e+00 -3.53307456e-01
2.30277240e-01 4.64800626e-01 -6.12646759e-01 4.28507566e-01
-5.55969119e-01 4.76332635e-01 -6.71059251e-01 2.01136470e-01
6.74244523e-01 6.81081787e-02 -2.24290509e-02 3.11106682e-01
6.03097975e-01 -1.85489610e-01 -9.25681591e-01 1.13942504e+00
-7.00831234e-01 6.23807609e-01 -3.68464321e-01 -9.92541969e-01
8.93849194e-01 4.16123420e-01 -1.49140572e-02 -1.20127939e-01
3.86277616e-01 4.41900752e-02 -1.63990423e-01 -2.96326786e-01
3.14725161e-01 -4.43278044e-01 -3.83979470e-01 5.92427731e-01
-1.41434848e-01 -6.61458611e-01 -6.57958090e-02 2.63861477e-01
8.09854329e-01 -2.52589047e-01 1.61063850e-01 3.20282578e-01
6.05717838e-01 -1.56222671e-01 4.00502414e-01 7.41339326e-01
-1.72358546e-02 7.57351220e-01 4.99646008e-01 -3.91528428e-01
-1.25880206e+00 -9.11300063e-01 8.76030251e-02 8.02832246e-01
9.55850855e-02 -3.00591558e-01 -1.00412989e+00 -1.21260953e+00
-1.99555039e-01 9.79790270e-01 -7.12594330e-01 -5.39890230e-01
-6.54000998e-01 -7.22510159e-01 8.73334289e-01 5.26258469e-01
6.23839259e-01 -1.29117799e+00 -4.41155359e-02 2.64863640e-01
-2.44798914e-01 -1.02098882e+00 -6.79124534e-01 -2.64361501e-01
-5.20704567e-01 -9.52536285e-01 -7.76932001e-01 -9.49138045e-01
9.16133046e-01 -6.96761012e-02 1.06094170e+00 2.97594965e-01
8.76330063e-02 1.86248776e-03 -6.16088867e-01 -7.65226781e-01
-1.26656330e+00 -3.84160355e-02 1.19379587e-01 8.53132308e-02
6.73374906e-02 -4.65034455e-01 -3.93554479e-01 2.57932991e-01
-1.49397945e+00 -6.66326359e-02 2.51505911e-01 1.38933086e+00
3.67402345e-01 5.90071201e-01 7.44866967e-01 -1.15680933e+00
8.37935507e-01 -5.75227678e-01 -4.74983960e-01 1.42403170e-01
-3.96263242e-01 -2.55624831e-01 1.81393158e+00 -9.99659896e-01
-1.00724721e+00 -3.20641220e-01 -4.48688060e-01 -6.78710580e-01
-1.60194352e-01 2.83835381e-01 -4.57289875e-01 1.55812904e-01
8.26625943e-01 4.18554366e-01 -1.39914155e-01 1.16071872e-01
5.49515545e-01 8.41775239e-01 3.28677893e-01 -4.41384405e-01
1.45029235e+00 4.05151576e-01 -2.95090348e-01 -5.91792583e-01
-7.40958929e-01 4.90169376e-01 -3.07971060e-01 -4.10704650e-02
7.16791272e-01 -6.23248160e-01 -4.36405748e-01 8.52100611e-01
-1.14643180e+00 -3.88636678e-01 -4.35205996e-01 1.42385796e-01
-3.67031425e-01 4.41888005e-01 -6.33958519e-01 -6.12079084e-01
-6.20680928e-01 -1.01948631e+00 7.22330689e-01 3.53570193e-01
-8.27815384e-02 -1.26898384e+00 -5.39882891e-02 3.23765486e-01
3.40912551e-01 7.03631639e-01 8.77371073e-01 -9.51213598e-01
-2.02774808e-01 -6.16579473e-01 4.13783520e-01 6.78908050e-01
2.16222510e-01 2.51502514e-01 -7.27316856e-01 -3.64276260e-01
1.67418301e-01 -5.41060269e-01 5.16522825e-01 -1.92163274e-01
1.23319244e+00 -7.98843145e-01 -1.49570987e-01 5.69654882e-01
9.72304046e-01 3.69483232e-01 6.95634186e-01 9.98657569e-02
9.76774275e-01 4.56042111e-01 3.44652057e-01 4.39793944e-01
-4.16883864e-02 1.77097708e-01 5.36205053e-01 -1.66674685e-02
1.87604666e-01 -6.29164696e-01 4.85923618e-01 8.44173610e-01
2.58202314e-01 -7.57881165e-01 -6.91235960e-01 3.38920414e-01
-1.57104445e+00 -1.14676273e+00 -2.19480116e-02 1.89725804e+00
1.01995230e+00 3.88922542e-01 -1.84632123e-01 2.95570076e-01
8.61325800e-01 2.46772245e-01 -7.60908842e-01 -3.92637253e-01
-7.32697174e-02 3.24516326e-01 1.09032758e-01 4.34201688e-01
-1.04025316e+00 1.05266559e+00 5.66910458e+00 9.49522734e-01
-1.12243116e+00 2.03399837e-01 5.97992539e-01 -2.15677798e-01
-6.35603607e-01 -2.19096020e-01 -5.04199326e-01 9.79335785e-01
7.32394278e-01 -5.21324337e-01 4.62415189e-01 1.08167768e+00
-1.14765264e-01 8.22521985e-01 -9.91236448e-01 7.42305815e-01
1.42761141e-01 -1.23274243e+00 5.01876056e-01 -2.36984447e-01
9.83490467e-01 -4.39708591e-01 3.29780906e-01 6.79309785e-01
7.22526371e-01 -1.21097195e+00 7.02879012e-01 2.77373165e-01
7.25583613e-01 -1.24170732e+00 8.87046754e-01 6.82857335e-01
-7.67206550e-01 -5.06386198e-02 -4.96672839e-01 1.60458252e-01
2.03493442e-02 4.53536630e-01 -4.73114341e-01 4.72315788e-01
3.74362439e-01 3.99169862e-01 -3.21193874e-01 2.40293443e-01
-5.62175274e-01 6.65591896e-01 -1.23731963e-01 -3.80169064e-01
2.98995554e-01 -9.04040486e-02 5.60739636e-01 8.14260364e-01
2.66738176e-01 1.74358577e-01 1.05721951e-01 9.31413174e-01
-5.68875968e-01 1.62289545e-01 -8.21214199e-01 -2.08098799e-01
4.55150157e-01 1.01485693e+00 -3.59690607e-01 -3.38399529e-01
-2.03876376e-01 1.35613692e+00 6.71749711e-02 3.17319632e-01
-1.12202072e+00 -8.84996116e-01 4.24777865e-01 -2.31187016e-01
1.62239358e-01 2.31746122e-01 -1.52158752e-01 -1.49240875e+00
1.36068717e-01 -1.40666497e+00 1.40461698e-01 -7.11498439e-01
-1.72592652e+00 9.06059623e-01 -4.78399754e-01 -1.25622869e+00
-5.68353534e-01 -4.85282868e-01 -1.04369402e+00 9.03472900e-01
-1.00371945e+00 -1.15750468e+00 -3.25974971e-01 9.33399796e-01
6.75020635e-01 -4.03977066e-01 9.52043295e-01 -9.44521651e-03
-6.69078290e-01 1.07386577e+00 5.67051530e-01 5.53985953e-01
6.26096547e-01 -1.25282669e+00 7.42071211e-01 9.99254882e-01
1.13403492e-01 4.57461059e-01 7.48813331e-01 -6.36837006e-01
-1.12174070e+00 -1.49406183e+00 4.92283434e-01 -6.42893970e-01
7.71783829e-01 -6.42605364e-01 -1.02854657e+00 8.93434703e-01
1.47211134e-01 -8.18885714e-02 7.18414187e-01 -3.95970166e-01
-2.91892886e-01 8.55451152e-02 -1.31942296e+00 1.03572285e+00
6.60804689e-01 -5.68081558e-01 -6.98389769e-01 6.88385487e-01
1.02674627e+00 -5.92561245e-01 -7.40462899e-01 1.29401281e-01
2.45323479e-01 -6.40664279e-01 8.26340079e-01 -7.50347853e-01
9.52464938e-01 -2.16676757e-01 7.31735826e-02 -1.67050076e+00
2.24368721e-02 -9.12714183e-01 -2.49672249e-01 1.44152319e+00
2.19849244e-01 -8.59237254e-01 7.12056875e-01 5.03820419e-01
6.38691187e-02 -5.79125464e-01 -6.67924166e-01 -7.15228796e-01
8.62547636e-01 -3.25687140e-01 9.83115852e-01 1.20916450e+00
-2.98284113e-01 3.31399977e-01 -6.03727877e-01 2.17771605e-01
4.59685177e-01 -1.10727273e-01 9.77519751e-01 -1.00558770e+00
-3.41521412e-01 -5.01436532e-01 -2.37801969e-01 -8.69722486e-01
5.05740821e-01 -7.27898061e-01 1.34418681e-01 -1.14705718e+00
-1.66264147e-01 -4.35934246e-01 1.67657226e-01 3.77656221e-01
-7.76669860e-01 2.48366728e-01 2.06183434e-01 2.63201594e-01
1.23157881e-01 9.10834551e-01 1.48125899e+00 -4.05568600e-01
6.40818775e-02 1.77480742e-01 -8.17483425e-01 8.37836266e-01
8.90195787e-01 -6.36216581e-01 -5.65933347e-01 -3.20893198e-01
2.24342033e-01 9.02663684e-04 2.72185981e-01 -6.28054559e-01
2.01212801e-02 -2.64361769e-01 5.81457376e-01 -2.82569766e-01
-1.77524779e-02 -7.47502625e-01 -4.36964966e-02 4.70144033e-01
-2.86129892e-01 -2.29907557e-01 3.90411052e-03 7.28578389e-01
-2.16047630e-01 -4.89227593e-01 9.52424049e-01 2.64408253e-02
-3.87070291e-02 4.72188354e-01 -2.59734362e-01 3.22546124e-01
1.24723947e+00 3.28871310e-01 -3.40375006e-01 -6.45804286e-01
-5.43159425e-01 1.68748319e-01 5.27931273e-01 5.35249174e-01
5.63258529e-01 -1.39287317e+00 -8.75203729e-01 2.56454438e-01
-1.31912038e-01 2.96444654e-01 3.47367674e-01 -1.22024454e-01
-5.99264920e-01 -7.54963532e-02 4.85025421e-02 -1.44500718e-01
-9.72135246e-01 9.80063140e-01 4.97440666e-01 -4.63804662e-01
-6.21371806e-01 8.24176252e-01 3.59700799e-01 -5.96809566e-01
9.57593322e-02 6.86462745e-02 -1.25386000e-01 -1.64705932e-01
6.66989148e-01 2.86480598e-02 -2.67599195e-01 -3.94529164e-01
1.45462900e-01 2.36291498e-01 -3.30706149e-01 4.13816646e-02
9.37857330e-01 2.04896644e-01 1.42707065e-01 2.39644989e-01
1.10451734e+00 3.67470413e-01 -1.20138323e+00 -1.56784937e-01
-8.02051485e-01 -3.60295206e-01 -2.44797692e-01 -5.36573470e-01
-1.23692942e+00 9.25489068e-01 1.73970357e-01 7.60345697e-01
9.88186181e-01 -3.07921648e-01 1.22452223e+00 4.98371750e-01
-7.01372698e-02 -8.71308744e-01 3.45913440e-01 4.36472595e-01
9.82791364e-01 -1.14209962e+00 -3.64008427e-01 -1.87893465e-01
-8.94115269e-01 1.10524666e+00 6.79106593e-01 -3.37627262e-01
4.73204672e-01 2.38802880e-01 1.79112211e-01 3.51373374e-01
-4.91720349e-01 3.92735869e-01 1.97222475e-02 7.12416530e-01
4.32825908e-02 1.98876224e-02 4.59588170e-02 8.74043941e-01
-7.26599693e-01 -3.98177594e-01 6.76259339e-01 7.91023195e-01
-1.19196810e-01 -1.22329783e+00 -5.16345322e-01 3.47138435e-01
-7.52114475e-01 -2.09365547e-01 -3.89156640e-01 7.25213230e-01
1.36651278e-01 1.14994895e+00 -1.08604275e-01 -5.40546536e-01
3.54428291e-01 1.10946909e-01 1.22259796e-01 -4.94831353e-01
-7.63241708e-01 -3.93286169e-01 -3.55245471e-01 1.58549622e-01
1.26439452e-01 -4.06771362e-01 -1.14059818e+00 -8.31041515e-01
-4.35617357e-01 2.91308850e-01 3.74170542e-01 1.05448651e+00
1.27172917e-01 7.58748710e-01 1.24898767e+00 -4.98210192e-01
-8.89812648e-01 -9.69247103e-01 -3.87510389e-01 7.07213759e-01
2.30417296e-01 -2.13773042e-01 -7.39048123e-01 3.68304491e-01] | [5.958520412445068, 8.079705238342285] |
86199ef8-8363-4568-b4e2-39296f3b814c | federated-learning-in-the-presence-of | 2305.19971 | null | https://arxiv.org/abs/2305.19971v1 | https://arxiv.org/pdf/2305.19971v1.pdf | Federated Learning in the Presence of Adversarial Client Unavailability | Federated learning is a decentralized machine learning framework wherein not all clients are able to participate in each round. An emerging line of research is devoted to tackling arbitrary client unavailability. Existing theoretical analysis imposes restrictive structural assumptions on the unavailability patterns, and their proposed algorithms were tailored to those assumptions. In this paper, we relax those assumptions and consider adversarial client unavailability. To quantify the degrees of client unavailability, we use the notion of {\em $\epsilon$-adversary dropout fraction}. For both non-convex and strongly-convex global objectives, we show that simple variants of FedAvg or FedProx, albeit completely agnostic to $\epsilon$, converge to an estimation error on the order of $\epsilon (G^2 + \sigma^2)$, where $G$ is a heterogeneity parameter and $\sigma^2$ is the noise level. We prove that this estimation error is minimax-optimal. We also show that the variants of FedAvg or FedProx have convergence speeds $O(1/\sqrt{T})$ for non-convex objectives and $O(1/T)$ for strongly-convex objectives, both of which are the best possible for any first-order method that only has access to noisy gradients. Our proofs build upon a tight analysis of the selection bias that persists in the entire learning process. We validate our theoretical prediction through numerical experiments on synthetic and real-world datasets. | ['Pengkun Yang', 'Jiaming Xu', 'Lili Su'] | 2023-05-31 | null | null | null | null | ['selection-bias'] | ['natural-language-processing'] | [-0.31104746 0.18078372 -0.10299103 -0.21432748 -1.2020569 -0.97462326
-0.08346 0.13333924 -0.7268543 0.7892822 -0.34468332 -0.6095022
-0.47577986 -0.8173212 -1.1391151 -1.1224045 -0.58559674 0.45936057
-0.13368347 0.06232153 -0.07940972 0.38036478 -0.9391768 0.03744726
0.715499 1.1039414 -0.37430176 0.74742764 0.14150386 0.70517653
-0.44241837 -0.7731525 0.7308736 -0.33533 -0.7508125 0.07075401
0.45500252 -0.39931405 -0.47080705 1.4798789 0.5372668 0.08993747
0.14816996 -1.4031435 -0.1967421 1.0129489 -0.42166987 0.20785347
-0.18798186 0.47383684 1.119339 -0.43065053 0.29613793 0.8757539
0.58585066 0.50719696 -1.1988838 -0.9072312 0.4030616 -0.01477761
-1.1608855 -0.12451698 0.54862446 -0.38071176 0.65962636 0.54077315
-0.02404588 0.789174 -0.19842643 0.72076946 1.231138 -0.2656282
0.5655343 0.396221 0.08621993 0.80032694 0.43624762 0.12673523
-0.330687 -0.5399344 0.5126021 0.19366923 -0.5048547 -0.30905703
-0.75142634 1.1310699 0.28659332 -0.00519249 -0.31257507 0.39731145
0.2864019 0.74984676 0.33826098 0.20752046 -0.7381214 -0.18039127
-0.52680975 0.07604737 1.0771462 0.86731404 0.8291709 0.05779064
0.05812947 0.44382492 -0.04695571 0.7228414 0.11404596 -1.2450104
0.8418217 0.42666805 0.40827248 -0.72962254 -0.09762114 -0.62497604
-0.8039204 0.37669048 0.86708367 -0.7000052 -0.38731042 2.068797
0.3156592 -0.05592924 -0.16748907 1.0007534 -0.13389406 0.3470414
-0.20815516 -0.54047054 0.949702 -0.7009115 -0.2625824 0.02683308
0.70726794 -0.40265077 1.1723614 0.50924075 -1.3089808 0.32340235
-0.6167561 0.5308114 -0.07915854 -0.41176334 0.5889167 1.0737207
-0.945145 0.6239333 -0.813567 0.21409 0.49883607 0.7793779
-0.3359444 -0.05406866 -0.7093279 0.24552143 -0.07305992 -0.04500527
-1.1912678 -0.66591996 -0.34843844 0.22690894 0.7261422 -0.49668035
1.2282323 -0.9310454 -1.295374 0.67235905 0.14433482 -0.5650058
1.0124915 0.12743406 0.04263146 0.04276219 -0.21791774 -0.4943998
0.59297884 -1.0964277 -0.8217342 -0.6571682 0.4557141 -0.10175937
-0.71925086 0.02971711 -0.16278777 -0.3557957 -0.09219868 -1.0322324
-0.5235531 0.15993045 -0.29750925 -0.08631256 0.6658914 -0.16963561
0.8173718 -1.9825683 -0.1872455 0.5130458 0.4190607 0.22062196
-0.16876154 0.5533849 0.38284364 0.38072172 -0.2747853 -0.44690463
0.33803576 0.20633596 -0.25038597 0.97435373 -0.6280822 0.47334105
-0.8471552 -0.03099579 -0.1325686 0.23914188 -0.83549964 0.12612198
-0.30104515 0.17957886 -0.7186288 0.46362507 0.8693915 -0.5591434
0.5247972 0.24365187 0.09316254 0.07729792 -1.365008 1.1031957
-0.65563035 0.1854385 0.81712455 -1.0871499 0.29503974 0.3990918
0.6489801 -0.3786088 0.44163126 0.55670154 -0.32441875 -0.43239835
-0.17210782 -0.3853222 -0.02168872 0.8551632 -0.08625843 0.6209527
-0.06313197 0.27130288 1.5011328 -0.6934461 -0.28199893 -0.26828024
0.23341449 -0.34347275 0.7655487 1.1980253 -0.34703884 0.22967902
0.9157696 -0.35928592 -0.901938 -1.1308806 0.3279524 1.3378928
0.195856 0.01992471 -0.8419283 -1.133353 0.24120075 0.4882761
-0.7343669 0.14402945 -0.519146 -1.0429645 0.51997 0.2689652
0.41630703 -0.76430696 -0.3763585 0.17996491 0.02414632 -0.871779
-0.86939335 0.25125954 -0.8417347 -1.1470795 -0.503643 -0.45477185
0.9069253 0.16992664 0.76462835 0.09606663 -0.07493196 0.5906679
-0.1884095 -0.27231312 -0.30294845 0.16933045 0.06156719 0.36360043
0.05334822 -0.8650237 -0.87838227 0.27871972 -1.0432048 -0.52610546
0.3825476 0.770645 0.5077888 0.14889683 0.3559342 -1.2457013
0.59401035 -0.63331527 -1.1097523 0.16031429 -0.92673415 0.17320862
1.1484007 -0.68401366 -0.7331635 -0.07519192 0.09856431 -0.6390225
0.22709195 0.24823426 -0.15460266 -0.4054543 0.63069445 0.28597236
-0.01160046 -0.6031254 0.45254228 0.47695285 0.3371488 -0.7947599
0.9357041 0.78287673 -0.05001998 -0.35635248 -0.59202987 -0.05647947
0.22978015 -0.03701912 -0.00638753 -0.5375265 -1.6552671 0.3034628
-0.7184343 -0.5299264 -0.39850459 0.4853448 -0.48066986 0.39210683
-0.714345 -1.1942298 -0.62350667 -1.0727358 0.17047645 0.18506917
0.42808348 -1.0237409 -0.01944374 0.49179262 0.7964976 0.258022
0.7735272 -0.8210751 -0.7653764 -0.47115576 -0.07285751 0.38157603
-0.14984338 -0.44718382 -0.7517274 -0.82914156 0.24253483 -0.4349929
0.61068076 0.18116403 1.3175753 -1.4527947 -0.11972626 0.78385276
1.815298 0.07689262 0.23378544 0.32097426 0.41205105 0.2414517
0.07950226 1.0574112 0.22614448 0.26656422 1.0026222 0.10234918
0.562481 -0.09643386 0.40834334 0.4003093 0.07938235 -0.40676755
-0.5344387 0.7600617 -1.6640992 -0.84617066 0.04428826 2.5519965
1.0563225 0.08052897 0.2719979 0.03331245 0.5545137 -0.06038096
-0.78018737 -0.54590505 -0.27541313 0.3977804 1.1751388 0.6268658
-0.77819633 0.5904462 5.1229706 0.8466675 -1.1721064 0.47919348
0.7648895 -0.6483463 -0.34591806 0.09592668 -0.47053775 0.71817106
0.8726602 -0.32093233 1.0072904 1.1805882 0.07143104 0.2628426
-1.0542823 0.7740936 -0.47021094 -1.2574191 -0.43016 0.3469819
0.89634734 0.32625967 0.33756515 0.05161734 0.9648053 -0.84427834
0.43456775 -0.08957148 0.69834703 -1.02165 0.64858764 0.5055874
-0.8420729 -0.6532065 -0.18222712 0.17193629 -0.01496657 0.63655305
-0.55069995 0.2474665 0.80309796 -0.37561285 0.04519545 1.0003152
-0.04148211 0.7437663 -0.851906 0.02624777 0.3421811 -0.14946459
0.590568 1.1307976 0.13646328 0.37347993 0.2814267 0.53110486
-0.7005327 0.22872543 -0.25056133 0.1558006 0.5814231 1.1043932
-0.29941007 -0.19842297 -0.30732688 0.9517869 0.47386083 0.46456522
-0.84559745 -0.5063107 0.94666433 0.12526006 0.5019298 0.03156272
-0.33733544 -1.1404872 0.42302492 -0.988848 0.8605554 0.25137103
-1.5341209 0.5564069 -0.46235144 -0.7466597 0.04725103 -0.3120408
-0.6882078 0.73164076 -1.3980302 -0.8632847 0.02126116 1.0119691
-0.05925115 -0.04722862 0.6709944 0.4583727 -0.71753067 1.3651298
0.6311675 0.24586825 0.26248264 -1.0691526 -0.13862684 0.96966106
-0.08013339 0.47912428 0.85844076 -0.15348832 -1.733356 -1.1155317
0.51752883 -0.10906599 0.7236172 -0.29999092 -0.76263225 0.85036814
-0.14384706 0.69109267 0.54367423 -0.16961165 -0.77096164 -0.46950397
-1.6909896 0.5154677 0.9489053 -0.5322712 0.36139172 0.4903817
0.58050895 -0.22297712 -0.8355371 0.18986157 0.328126 -0.93929684
0.64958096 -0.94378334 -0.30738518 0.11583076 -0.32241136 -0.7567666
-0.02158031 -1.3624328 -0.42037076 0.9212671 0.5121446 -1.1397469
1.330883 0.9995888 0.30586344 -0.7795206 -1.386811 -1.0285096
0.57276416 -0.417105 0.32409987 0.94418144 0.10913752 -0.4012429
-0.42117533 0.5492972 1.0248058 0.10346299 0.62067664 -0.67497313
-0.85505366 -0.5359063 -0.00558278 -0.76008695 -0.0556745 -0.7628817
-0.18931186 -0.8799108 0.37752664 -0.9014091 -0.67806864 0.79456013
0.07247408 0.04852988 0.2948839 0.07612026 -0.64105964 0.35234624
0.7778749 -0.14495979 -0.23231658 0.40256262 -0.7345891 0.5879658
0.8133182 -0.88709074 -0.35228965 -0.61173624 0.24294811 0.3665104
0.3419694 -0.5247937 0.28307152 -0.537609 -0.395772 0.02308647
0.13137977 -0.87034214 0.1315642 0.5910779 -0.42849684 -0.12434802
-0.24922545 0.62990457 0.3678034 -0.10058583 1.0098774 -0.11187452
0.3998421 0.7410256 -0.2652706 0.3079916 1.0592822 0.21747886
-0.453078 -0.695469 -0.550856 0.32136363 0.43550882 -0.22034551
0.30010024 -0.8217846 -0.71314037 -0.03377257 -0.42176905 -0.19396977
0.36693797 1.0411967 -0.3281693 0.18407044 0.3916156 -0.10437525
-1.0726992 0.76141906 0.66299033 -0.46079993 -0.5278962 1.081238
0.01207932 -0.23792341 0.74993443 0.03990186 0.69207656 -0.33270425
0.50149167 0.6412508 0.12091946 -0.09721217 -0.3735708 0.0804872
-0.27894917 -0.39077 1.4131961 -0.09294637 -0.02236014 -0.2532864
1.5294441 0.2126622 -1.5916109 -0.5156535 -0.14906932 -0.55193144
-0.096487 -0.79393554 -1.6693583 0.7684245 0.7039233 0.33807686
1.1653066 -0.06130156 0.81615865 0.45205852 0.7320972 -0.9858697
-0.16887026 0.23316285 0.40678802 -0.9681829 -0.39697737 -0.16814299
-0.3176694 0.78798735 0.5489854 -0.1104359 0.65433055 0.31041166
-0.07762419 0.14331038 -0.92581314 0.15039067 -0.4737627 0.18510327
-0.11877545 0.10398392 -0.32726395 0.9074852 0.04503678 -0.0438805
0.5524273 1.0958198 -0.3215573 -1.1318914 -0.3297898 0.32807675
-1.1347126 0.12040825 -0.06090251 0.50669223 -0.07734589 1.0327266
-0.21993256 -0.2522951 0.13694085 -0.20672315 0.45133373 -0.13574743
-0.85286945 -0.01113582 -0.2585925 -0.57048655 0.13880216 -0.37185982
-1.2880893 -0.94379956 -0.33926934 0.49247658 0.61256623 0.77988744
0.35297588 -0.21071567 1.2721915 -0.18601184 -1.2251925 -0.6022447
-0.7490471 0.3573371 0.45525417 -0.08037035 -1.0132705 -0.2560986 ] | [6.070937633514404, 5.2093329429626465] |
71684804-d698-4dde-9cc0-0f72bd78a238 | a-knowledge-graph-perspective-on-supply-chain | 2305.08506 | null | https://arxiv.org/abs/2305.08506v1 | https://arxiv.org/pdf/2305.08506v1.pdf | A Knowledge Graph Perspective on Supply Chain Resilience | Global crises and regulatory developments require increased supply chain transparency and resilience. Companies do not only need to react to a dynamic environment but have to act proactively and implement measures to prevent production delays and reduce risks in the supply chains. However, information about supply chains, especially at the deeper levels, is often intransparent and incomplete, making it difficult to obtain precise predictions about prospective risks. By connecting different data sources, we model the supply network as a knowledge graph and achieve transparency up to tier-3 suppliers. To predict missing information in the graph, we apply state-of-the-art knowledge graph completion methods and attain a mean reciprocal rank of 0.4377 with the best model. Further, we apply graph analysis algorithms to identify critical entities in the supply network, supporting supply chain managers in automated risk identification. | ['Volker Tresp', 'Martin Berbalk', 'Dagmar Beyer', 'Emanuel Weigel', 'Roger Wernert', 'Daniela Inzko', 'Maximilian Buchner', 'Marcel Hildebrandt', 'Bailan He', 'Yushan Liu'] | 2023-05-15 | null | null | null | null | ['knowledge-graph-completion'] | ['knowledge-base'] | [-2.24172205e-01 3.88485372e-01 -4.21193123e-01 7.33012054e-03
-1.75638497e-01 -1.20998967e+00 6.99826106e-02 6.83193624e-01
6.68909773e-02 3.19708645e-01 7.55141795e-01 -5.38429976e-01
-7.43118227e-01 -1.18124652e+00 -4.58465129e-01 -2.10820973e-01
-2.53403068e-01 8.00983846e-01 -5.09707667e-02 -2.54341632e-01
2.16021553e-01 7.25911915e-01 -2.44505450e-01 2.56368935e-01
7.72480249e-01 8.63697112e-01 -2.83614676e-02 1.76219180e-01
-1.67910799e-01 1.06919897e+00 -2.26269275e-01 -1.10212040e+00
6.24637306e-01 -7.83653259e-02 -7.65850127e-01 7.50265494e-02
-4.33995575e-01 -2.45894462e-01 -3.83999228e-01 1.20901716e+00
-1.11267090e-01 -3.90392989e-01 5.05010128e-01 -1.33176243e+00
-7.62249708e-01 1.33778703e+00 -6.60294890e-01 1.70399010e-01
4.25576359e-01 2.23227262e-01 1.34218204e+00 -7.68931270e-01
8.55213284e-01 9.88696754e-01 7.60077417e-01 -3.26736361e-01
-1.50522709e+00 -6.62702560e-01 2.85239518e-01 1.22278452e-01
-1.18380034e+00 -1.01848226e-02 8.55158329e-01 -7.37986803e-01
1.00087070e+00 1.08292267e-01 9.32675838e-01 7.25494742e-01
5.00831664e-01 2.04458795e-02 1.01657069e+00 1.11871377e-01
1.23872804e-02 1.56633675e-01 5.18770516e-02 3.66703987e-01
9.19953525e-01 -2.16047764e-01 -5.23167074e-01 -2.12159440e-01
7.40888417e-01 4.83138531e-01 -2.81871885e-01 -3.58184665e-01
-1.33544421e+00 7.44461536e-01 6.81612611e-01 4.45161074e-01
-9.50383186e-01 9.33198109e-02 1.15931526e-01 7.39107728e-01
2.06888422e-01 7.82941043e-01 -5.82783997e-01 1.34656280e-01
-5.22846818e-01 -9.73132700e-02 1.54131782e+00 9.84131217e-01
6.47310257e-01 -1.04557633e-01 5.15918493e-01 -2.69445945e-02
4.89229262e-01 4.05571640e-01 -8.01758528e-01 -9.56724167e-01
8.61360967e-01 9.87356186e-01 1.65946990e-01 -1.56083572e+00
-6.55351639e-01 -6.37201905e-01 -8.57128441e-01 -1.53277874e-01
6.33263946e-01 -1.21976532e-01 -3.54718626e-01 1.06527293e+00
5.52078597e-02 -3.70087117e-01 1.17712654e-01 8.41030419e-01
1.63123354e-01 7.06648886e-01 7.39739602e-03 -5.42170584e-01
1.15361440e+00 -5.64632475e-01 -1.08622468e+00 3.72787528e-02
5.95385730e-01 -7.46807337e-01 4.42990124e-01 7.45907545e-01
-8.91738355e-01 2.70474523e-01 -6.63939655e-01 4.43497747e-01
-3.31140071e-01 -7.40885079e-01 8.00830364e-01 4.22515094e-01
-5.61767697e-01 8.39335501e-01 -5.92364550e-01 5.99460900e-02
2.71224946e-01 2.24274710e-01 -6.95625007e-01 -4.32622850e-01
-1.38938940e+00 1.18582773e+00 2.57839531e-01 4.75132763e-01
-8.58242333e-01 -1.02326596e+00 -6.83003306e-01 3.71008664e-01
9.86481249e-01 -4.06081229e-01 6.91436470e-01 -3.58280331e-01
-5.86962521e-01 -7.90742412e-02 7.48063862e-01 -3.56459804e-02
4.34508175e-01 6.37441874e-02 -5.97144902e-01 1.60458833e-01
-1.13817722e-01 -4.17342126e-01 2.43969560e-01 -1.50743961e+00
-4.77808803e-01 -4.35491860e-01 2.50353031e-02 -1.46325886e-01
-3.22444253e-02 1.99170262e-01 8.79283026e-02 -4.86020088e-01
5.04419565e-01 -7.74483025e-01 -5.23316801e-01 -4.51572001e-01
-6.29218340e-01 1.97282583e-01 1.60804883e-01 -1.33215845e+00
1.38261557e+00 -1.67227888e+00 2.08859459e-01 9.94670153e-01
6.11088693e-01 -3.21264654e-01 4.77155037e-02 1.32517052e+00
1.37846202e-01 4.44913477e-01 2.86063505e-03 1.81808084e-01
3.42330486e-01 3.91239613e-01 -1.70301229e-01 5.62523723e-01
-9.78567526e-02 7.65943110e-01 -1.07043922e+00 -3.08464300e-02
-1.54450675e-02 1.18335910e-01 -3.52269173e-01 -6.00311048e-02
-5.71611375e-02 5.68321049e-01 -5.00369549e-01 1.03858364e+00
5.63148081e-01 -3.06689620e-01 8.45899999e-01 -3.13565850e-01
-2.16562852e-01 2.20353991e-01 -1.34612024e+00 1.26827705e+00
-3.65958095e-01 8.35044682e-02 5.73041677e-01 -5.93381822e-01
8.66989672e-01 3.41302127e-01 7.46592939e-01 -5.90600789e-01
4.33770604e-02 2.07082674e-01 2.23986134e-01 -3.76837492e-01
2.26970628e-01 -6.71778098e-02 -3.70250583e-01 6.69736564e-01
-1.63820311e-01 2.46563070e-02 1.56715110e-01 5.29208302e-01
1.38881540e+00 -3.72717261e-01 1.40331343e-01 -2.89606452e-01
-2.58613005e-02 2.75007486e-01 8.75539184e-01 2.49092489e-01
1.04875796e-01 -4.60940674e-02 1.17155230e+00 -6.16945446e-01
-8.93924773e-01 -1.10349381e+00 3.55331421e-01 6.72020435e-01
2.15230677e-02 -4.96454805e-01 -2.50188321e-01 -8.92853320e-01
5.85050285e-01 7.29723930e-01 -4.04135227e-01 1.81467794e-02
-3.43643874e-01 -4.12583172e-01 -9.34991986e-02 5.49814105e-01
-7.31502799e-03 -6.88664615e-01 7.18444660e-02 6.72278643e-01
-1.38349026e-01 -1.07555735e+00 -4.62050021e-01 1.38036594e-01
-6.47410750e-01 -1.58302617e+00 7.48258159e-02 -3.77724200e-01
8.37448120e-01 1.03212930e-01 1.16032434e+00 -2.97658015e-02
7.10214078e-02 1.78975537e-01 -3.04271817e-01 -2.07134619e-01
-6.95158780e-01 -3.77192944e-02 2.19235197e-01 -1.15414605e-01
8.49191546e-02 -7.71686971e-01 -6.93081200e-01 5.17613173e-01
-5.58155477e-01 -2.36718148e-01 7.75617480e-01 3.67648184e-01
4.59662020e-01 4.85265434e-01 9.84423876e-01 -9.52129662e-01
6.91585422e-01 -8.37902963e-01 -1.01097381e+00 4.82906252e-01
-1.36866856e+00 -2.48383075e-01 4.92025673e-01 -2.51655340e-01
-9.43811595e-01 1.00287184e-01 3.97555888e-01 -1.76665336e-01
4.59123909e-01 1.19300985e+00 -2.55916029e-01 -3.74942906e-02
1.96305126e-01 -3.37973207e-01 -1.30188629e-01 -7.09202528e-01
7.49973178e-01 1.09303683e-01 2.56771415e-01 -1.62581444e-01
1.04568172e+00 1.62612811e-01 2.90872455e-01 -1.01010367e-01
-5.04942656e-01 -3.36469114e-01 -9.41675186e-01 -4.43530083e-01
5.92730224e-01 -8.95056367e-01 -1.10757184e+00 -9.12516937e-02
-1.09095049e+00 -4.15681861e-03 -2.16028646e-01 6.01678729e-01
-1.58102244e-01 1.87554210e-01 -1.00205684e+00 -7.20739722e-01
-2.49340519e-01 -7.44403660e-01 1.02289729e-01 -3.29020590e-01
-3.23221177e-01 -1.13317752e+00 9.24272910e-02 6.36011899e-01
4.91308570e-01 5.96773028e-01 1.12841010e+00 -8.53837967e-01
-1.23495758e+00 -3.05094093e-01 -2.27492183e-01 1.86077029e-01
5.15729308e-01 -1.83321759e-01 -1.46620467e-01 -2.95112371e-01
-1.71922043e-01 2.96336204e-01 4.20074224e-01 -6.11961484e-02
2.08654329e-01 -1.01641023e+00 -2.19995782e-01 3.53455991e-01
1.42629933e+00 1.83698207e-01 4.20040905e-01 1.03446431e-01
9.53476846e-01 1.49400437e+00 5.96060574e-01 6.76556349e-01
8.56641293e-01 1.65353268e-01 7.25481987e-01 1.94313094e-01
2.86573648e-01 -4.58973706e-01 -2.76211519e-02 1.18604243e+00
-1.83423579e-01 -6.25418872e-02 -1.27267683e+00 6.97395802e-01
-2.02508497e+00 -9.38076794e-01 -7.18605161e-01 1.73987162e+00
6.96432114e-01 3.09927762e-01 2.42722128e-02 1.41880624e-02
5.08845985e-01 -1.54318333e-01 -1.74033791e-01 -1.51412800e-01
1.92557454e-01 -5.72502017e-01 1.10483134e+00 4.46503520e-01
-3.86634171e-01 6.59668803e-01 6.73310852e+00 -2.91572902e-02
-3.35956007e-01 -1.51672130e-02 3.80374402e-01 -7.51311779e-02
-9.92655575e-01 6.68033481e-01 -2.85865843e-01 2.61641264e-01
9.24932003e-01 -4.65461731e-01 8.64533126e-01 5.21118462e-01
2.93706149e-01 -6.47623241e-02 -1.02038991e+00 4.99108106e-01
-5.21368861e-01 -1.62344837e+00 -2.26302579e-01 3.99865299e-01
9.98602152e-01 -1.74869463e-01 -3.71378928e-01 -3.72266889e-01
9.69821632e-01 -8.15328002e-01 6.72834277e-01 9.38171566e-01
1.78279877e-01 -1.20796180e+00 8.62819612e-01 2.00934589e-01
-1.44199908e+00 -5.49433351e-01 -2.62341976e-01 -1.15249436e-02
6.27762496e-01 8.91474903e-01 -8.85358036e-01 9.09709990e-01
5.33497870e-01 5.07313907e-01 -1.30680710e-01 5.97200394e-01
-3.50688368e-01 4.61446285e-01 4.50594872e-02 3.97221655e-01
6.24342076e-03 -8.18387508e-01 2.32509568e-01 9.18566167e-01
3.46510172e-01 2.45350435e-01 2.44149953e-01 9.74745154e-01
-3.66182566e-01 -5.95701858e-02 -8.26718092e-01 -6.15276456e-01
7.12201595e-01 1.32806396e+00 -6.75047755e-01 2.60937601e-01
-6.85061455e-01 3.66139561e-01 6.57856986e-02 4.67052132e-01
-2.04608038e-01 -3.64793003e-01 8.08265984e-01 4.21570867e-01
-2.26969849e-02 -6.27060652e-01 -3.44652057e-01 -8.20077300e-01
-4.30971980e-02 -8.00067604e-01 3.98379683e-01 -3.57181668e-01
-1.69428670e+00 6.61082193e-02 -2.30876893e-01 -7.98163533e-01
-1.05978198e-01 -3.00183684e-01 -3.01389635e-01 8.88368011e-01
-1.53369069e+00 -1.22661698e+00 1.13987356e-01 5.26832759e-01
-2.50738531e-01 -1.64045542e-02 3.89057666e-01 2.80874044e-01
-3.77360851e-01 -8.84913281e-02 1.57559067e-01 1.37415141e-01
4.88130778e-01 -1.29662704e+00 4.11374927e-01 7.84391224e-01
-1.57194987e-01 7.99792945e-01 5.71094990e-01 -1.49863315e+00
-2.23929000e+00 -9.70483541e-01 1.17380869e+00 -6.04800165e-01
1.52667189e+00 -5.96511185e-01 -1.05212533e+00 9.39388752e-01
2.07876667e-01 -2.11393058e-01 8.16007674e-01 2.46834084e-01
-5.46557724e-01 -4.83772099e-01 -1.20610189e+00 3.19481045e-01
1.18990386e+00 -5.90160787e-01 -4.75987643e-01 4.80286211e-01
1.04103494e+00 2.70985216e-01 -1.77961087e+00 1.18886098e-01
3.70053500e-01 -6.07526422e-01 7.39110529e-01 -4.02308166e-01
5.19139506e-02 -1.57874927e-01 -1.62331402e-01 -1.41369212e+00
-9.04227197e-01 -9.77086425e-01 -1.92170531e-01 1.41384566e+00
6.93167269e-01 -9.17494416e-01 6.80645406e-01 8.90140235e-01
-7.41536617e-02 -4.49558675e-01 -5.97059011e-01 -8.61914933e-01
-2.25780532e-01 -2.03040853e-01 9.71685588e-01 1.53718054e+00
5.43571770e-01 8.77649561e-02 -4.07466561e-01 5.74998677e-01
9.78466451e-01 3.73375177e-01 3.90627027e-01 -1.45970356e+00
4.67711873e-02 -2.27154329e-01 -3.78007703e-02 -1.78195134e-01
-1.37842253e-01 -9.63016033e-01 -4.83407229e-01 -2.32668209e+00
1.44535918e-02 -5.24399757e-01 -5.23413062e-01 5.73769808e-01
4.76536900e-01 -1.96066663e-01 4.46804553e-01 4.48175371e-01
-2.53801554e-01 1.61454782e-01 1.42241454e+00 -5.33126108e-02
-1.79806337e-01 -1.07714839e-01 -1.19230151e+00 6.21682942e-01
7.89115369e-01 -7.60430098e-01 -3.82773191e-01 -2.27891400e-01
1.24324894e+00 7.49571562e-01 -1.32144457e-02 -2.27948517e-01
5.36308765e-01 -7.69923270e-01 1.49516329e-01 -7.53726482e-01
-2.37406850e-01 -1.48782325e+00 1.20341182e+00 8.74538004e-01
-3.10453046e-02 6.05339885e-01 -4.37123686e-01 7.34976292e-01
-7.24298656e-02 -5.86178750e-02 -9.90494154e-04 -1.48347989e-01
-2.25817114e-01 3.20480078e-01 -4.84445453e-01 -4.85882759e-01
9.21486557e-01 2.04076558e-01 -8.69247794e-01 -5.57443619e-01
-8.89849663e-01 7.40128279e-01 5.99437237e-01 3.05353731e-01
6.46254659e-01 -1.27173769e+00 -7.34787107e-01 -1.49982870e-01
2.12002974e-02 -2.51483560e-01 2.34211013e-01 9.49321866e-01
-6.52839482e-01 3.01309139e-01 -1.65793106e-01 2.67423272e-01
-5.73374152e-01 1.03821743e+00 -1.22423343e-01 -5.81801832e-01
-4.31196004e-01 3.03954214e-01 -9.60396901e-02 -1.80544972e-01
1.00673214e-01 -2.59912372e-01 -1.57747135e-01 6.38824284e-01
3.53044569e-01 8.09413671e-01 -2.08124369e-01 -4.12893742e-01
-5.35840392e-01 3.58856618e-01 1.46411672e-01 9.51689035e-02
1.70165479e+00 -4.12679553e-01 -5.73010564e-01 -7.23114908e-02
7.66030550e-01 3.29482079e-01 -1.29693782e+00 -3.01701635e-01
6.05646074e-01 -8.18066895e-01 -1.43141627e-01 -1.08571410e+00
-1.39933157e+00 3.76055777e-01 -2.47095317e-01 6.93196833e-01
9.68913972e-01 1.65328562e-01 5.78159451e-01 5.33391774e-01
5.46744883e-01 -1.13389170e+00 7.69801289e-02 2.76650071e-01
1.10485089e+00 -7.97541499e-01 2.45068371e-01 -9.69100773e-01
-8.83872509e-01 9.36787248e-01 4.33822833e-02 1.31270736e-01
7.42147386e-01 3.91001612e-01 8.53629317e-03 -5.93939841e-01
-7.38000453e-01 5.98188899e-02 2.92986669e-02 7.29678035e-01
-1.59763902e-01 2.98215002e-01 -2.39969287e-02 8.86027694e-01
4.82485332e-02 -4.53227490e-01 8.52450371e-01 8.34557056e-01
-2.92253882e-01 -1.18682194e+00 -2.62299448e-01 5.14811993e-01
-6.37176037e-01 -1.37462586e-01 -5.32592952e-01 7.96039999e-01
-1.84855148e-01 1.34084380e+00 -2.20400035e-01 -5.96149504e-01
8.91192615e-01 -1.71242282e-01 1.12519920e-01 -3.92602980e-01
-9.84514654e-01 2.10883200e-01 6.06582820e-01 -7.02792704e-01
2.44363640e-02 -7.12553203e-01 -1.41352201e+00 -8.41893554e-01
-4.38066781e-01 1.05200283e-01 8.52445960e-01 4.76211816e-01
4.53589201e-01 5.52342772e-01 1.00499058e+00 -2.15946838e-01
-6.47368193e-01 -6.03116274e-01 -1.03053987e+00 3.09984922e-01
1.59973130e-02 -4.04017299e-01 -2.48048991e-01 -9.54190791e-02] | [6.5946760177612305, 5.935716152191162] |
ac101df0-3527-4eed-8c11-4a60684a380f | mllp-vrain-upv-systems-for-the-iwslt-2022 | null | null | https://aclanthology.org/2022.iwslt-1.22 | https://aclanthology.org/2022.iwslt-1.22.pdf | MLLP-VRAIN UPV systems for the IWSLT 2022 Simultaneous Speech Translation and Speech-to-Speech Translation tasks | This work describes the participation of the MLLP-VRAIN research group in the two shared tasks of the IWSLT 2022 conference: Simultaneous Speech Translation and Speech-to-Speech Translation. We present our streaming-ready ASR, MT and TTS systems for Speech Translation and Synthesis from English into German. Our submission combines these systems by means of a cascade approach paying special attention to data preparation and decoding for streaming inference. | ['Alfons Juan', 'Albert Sanchis', 'Jorge Civera Saiz', 'Joan Albert Silvestre-Cerdà', 'Pau Baquero-Arnal', 'Gonçal Garcés Díaz-Munío', 'Adrián Giménez Pastor', 'Alejandro Pérez-González-de-Martos', 'Javier Jorge Cano', 'Javier Iranzo-Sánchez'] | null | null | null | null | iwslt-acl-2022-5 | ['simultaneous-speech-to-text-translation', 'speech-to-speech-translation'] | ['natural-language-processing', 'speech'] | [ 3.39170069e-01 1.70460761e-01 -1.55481875e-01 -5.19002914e-01
-1.44977653e+00 -7.03256428e-01 7.79122114e-01 -7.93863758e-02
-2.49077410e-01 7.23027468e-01 4.64271516e-01 -1.18433273e+00
4.93767709e-01 9.16026309e-02 -7.17832863e-01 1.38668092e-02
3.08965504e-01 1.00378978e+00 4.35787328e-02 -4.79676127e-01
-3.28094214e-01 3.52818251e-01 -8.97929430e-01 9.20493543e-01
5.99575460e-01 4.25118089e-01 6.37158573e-01 1.38128042e+00
-1.96676642e-01 6.48379087e-01 -6.88123286e-01 -4.82363880e-01
1.39618486e-01 -7.10402369e-01 -1.02907217e+00 -1.67693987e-01
1.95287794e-01 -1.49598032e-01 -4.46506053e-01 5.92102826e-01
8.47169161e-01 3.08248162e-01 2.13150397e-01 -8.74012709e-01
-4.12510604e-01 1.41813970e+00 1.12857707e-01 6.41614795e-01
8.17750692e-01 7.35379085e-02 1.08745909e+00 -1.34780896e+00
8.61842930e-01 1.55546749e+00 1.43496124e-02 6.27617836e-01
-1.16434944e+00 -2.58631349e-01 7.86154717e-02 2.35667318e-01
-1.38661551e+00 -1.55460691e+00 2.61924148e-01 -2.13932209e-02
2.09368253e+00 6.49526000e-01 5.14828563e-01 1.43153644e+00
-4.84990627e-02 1.25927007e+00 8.06438625e-01 -6.55276477e-01
1.03873581e-01 2.04335466e-01 -4.97679502e-01 1.52905628e-01
-5.94529331e-01 1.06350921e-01 -9.85362947e-01 1.95476055e-01
3.93491089e-01 -9.89024103e-01 1.05132302e-02 8.05790305e-01
-1.76230276e+00 6.01262867e-01 -4.41297531e-01 2.79827237e-01
-4.00852412e-01 1.27412692e-01 8.81656110e-01 9.69323158e-01
8.62519681e-01 1.72718033e-01 -7.70495176e-01 -6.53067410e-01
-1.32578897e+00 2.12175786e-01 9.42670465e-01 1.52478862e+00
-1.59345821e-01 6.44902945e-01 -5.22773206e-01 1.06459808e+00
6.14680767e-01 9.77149069e-01 4.91797805e-01 -8.59010875e-01
1.41714013e+00 -5.98333597e-01 8.08639918e-03 -2.89910346e-01
1.87853687e-02 -1.45507231e-01 -2.91783005e-01 -7.12666631e-01
2.48645544e-02 -6.12526119e-01 -9.67243910e-01 1.36980057e+00
2.53045917e-01 2.40227610e-01 6.22656345e-01 6.29346430e-01
1.06449425e+00 1.45944488e+00 -1.55171782e-01 -7.21303999e-01
1.06564486e+00 -1.18947363e+00 -1.18202126e+00 -5.60854316e-01
9.14962769e-01 -1.42712510e+00 8.06835890e-01 1.68328717e-01
-2.01142716e+00 -5.34281731e-01 -7.42733955e-01 -2.36506715e-01
1.01351805e-01 2.97169238e-01 1.13173358e-01 3.07528645e-01
-1.51670599e+00 3.55408937e-01 -1.03444552e+00 -3.99154931e-01
-3.90403241e-01 1.96922973e-01 -1.59681648e-01 2.06932411e-01
-1.50787210e+00 1.07913625e+00 2.87599206e-01 3.37394506e-01
-5.54550827e-01 -3.43075156e-01 -8.86107981e-01 -1.91639513e-01
3.45645584e-02 -6.12490237e-01 1.98913586e+00 -6.99728489e-01
-2.29641342e+00 8.24080825e-01 -8.07156324e-01 -7.34098256e-01
6.33966506e-01 -4.16155994e-01 -8.22054267e-01 -2.45719478e-02
-2.19904423e-01 4.75843877e-01 6.68601930e-01 -5.03520727e-01
-8.17981303e-01 -1.60593703e-01 -7.86872685e-01 6.51769519e-01
3.81071180e-01 1.06084406e+00 -2.82031238e-01 -7.05287755e-01
-5.58078177e-02 -9.80026603e-01 -2.18941823e-01 -1.04257095e+00
-6.85361147e-01 -2.50183672e-01 6.03290796e-01 -1.31345129e+00
1.21528983e+00 -1.73814809e+00 8.60683084e-01 -6.91100955e-02
-6.36477292e-01 4.63047206e-01 -3.11699390e-01 9.61369693e-01
7.54583552e-02 1.55154958e-01 1.89524397e-01 -1.03084671e+00
5.30328639e-02 3.65469277e-01 -9.65563953e-01 1.78721309e-01
2.81081110e-01 1.26420033e+00 -7.89024115e-01 -6.33151710e-01
3.81923765e-01 2.19445661e-01 1.96637157e-02 4.83910292e-01
-4.35188413e-01 6.72604203e-01 -3.29362415e-02 4.75068778e-01
4.22711186e-02 5.65388680e-01 3.22037667e-01 3.41521770e-01
-4.98453140e-01 1.43134379e+00 -8.08949053e-01 1.96040523e+00
-6.65138960e-01 7.96172082e-01 3.23307902e-01 -6.81915700e-01
7.51765192e-01 1.14180553e+00 -1.15799494e-02 -6.18499696e-01
1.67845264e-01 7.10881472e-01 -4.61392030e-02 -3.39458525e-01
8.83103192e-01 -5.28250784e-02 -1.61469299e-02 2.28164658e-01
4.97838616e-01 -7.61190832e-01 8.25003013e-02 3.01968127e-01
8.62359107e-01 2.14946494e-01 1.83506876e-01 -1.85930710e-02
2.67611772e-01 6.44317046e-02 2.22198188e-01 5.33080280e-01
5.33268005e-02 7.43649065e-01 -3.95471929e-03 1.46165758e-01
-1.32192421e+00 -1.02166021e+00 3.30751061e-01 1.23500037e+00
-5.31515419e-01 -6.27172410e-01 -8.19741666e-01 -3.42029601e-01
-4.45096672e-01 1.33213878e+00 5.06498106e-02 2.41328850e-01
-1.17300594e+00 -2.18885448e-02 9.31402802e-01 2.36103579e-01
-3.10730040e-01 -1.29670072e+00 2.55341560e-01 5.90884686e-01
-6.84716165e-01 -1.66524005e+00 -8.04042101e-01 1.47955611e-01
-8.79808068e-01 -5.69705628e-02 -8.56023788e-01 -9.53040302e-01
-2.41271585e-01 1.06769346e-01 1.06472969e+00 -2.36325398e-01
1.98517978e-01 -2.85094064e-02 -5.53731680e-01 -4.83001977e-01
-1.36034977e+00 5.22454619e-01 1.89554408e-01 -2.29931176e-01
-5.12105003e-02 -3.35314751e-01 4.49569449e-02 8.38030651e-02
-3.41996670e-01 3.60751659e-01 2.90070981e-01 4.36568707e-01
6.30072534e-01 -1.06659484e+00 5.55844247e-01 -7.82081962e-01
7.50431776e-01 -4.12923306e-01 -6.92005873e-01 3.49205196e-01
-1.13502890e-01 -7.74633810e-02 6.52352273e-01 -3.93887073e-01
-1.15693665e+00 2.02281654e-01 -8.17912281e-01 -3.99843514e-01
6.84207603e-02 5.65337360e-01 -2.54271626e-01 5.86471915e-01
3.61992657e-01 5.88933945e-01 -2.43361369e-01 -5.38618445e-01
8.54254663e-01 1.25418568e+00 5.77686250e-01 -3.50972384e-01
7.31463492e-01 -3.73194426e-01 -5.38520098e-01 -1.00480533e+00
-3.19510311e-01 -4.03373688e-01 -5.29089987e-01 -2.53781192e-02
7.72338510e-01 -1.10856390e+00 -4.86646891e-02 4.12911773e-01
-1.93539226e+00 -6.95685804e-01 -5.83253145e-01 7.20730066e-01
-9.06629622e-01 1.61923319e-01 -1.03598773e+00 -7.47333884e-01
-6.79940403e-01 -1.33578312e+00 1.40194368e+00 -4.72920448e-01
-5.61322868e-01 -9.66870844e-01 3.23375195e-01 4.46099848e-01
5.03217041e-01 -7.61759520e-01 4.13144380e-01 -9.69045818e-01
-3.13481033e-01 2.76304595e-02 2.56136507e-01 3.98358434e-01
-2.53925145e-01 1.30084068e-01 -7.74594367e-01 -1.95479736e-01
-3.55241776e-01 -3.33075017e-01 5.00240088e-01 5.15270829e-01
1.65653944e-01 -5.84626436e-01 -1.00584030e-01 7.48029470e-01
7.57047653e-01 3.12879354e-01 5.74186683e-01 -2.43430361e-01
7.46478617e-01 6.78068697e-01 4.86826211e-01 2.69139081e-01
5.12151599e-01 9.92382407e-01 -2.42719844e-01 -1.20992158e-02
-5.20128965e-01 -2.88391888e-01 1.10760105e+00 2.12942958e+00
3.10271621e-01 -9.09832358e-01 -9.30992901e-01 6.36327207e-01
-1.85963130e+00 -6.12352133e-01 -4.71388072e-01 1.94894552e+00
9.22570646e-01 -2.79533595e-01 -1.89325642e-02 -2.91655511e-01
7.14285314e-01 4.72166836e-01 7.64999837e-02 -1.26766467e+00
-2.95753181e-01 3.30029756e-01 4.39450741e-01 1.11705124e+00
-4.61434364e-01 1.72810328e+00 7.57702160e+00 1.10614002e+00
-1.06876600e+00 5.61118841e-01 3.01111877e-01 -2.60016322e-01
-5.51385462e-01 4.33420464e-02 -1.20500183e+00 3.11188251e-01
2.32913947e+00 -4.27820206e-01 8.50942075e-01 2.51137644e-01
5.76201260e-01 3.30686986e-01 -1.18287015e+00 8.57685149e-01
-4.93147150e-02 -1.51934350e+00 8.70278180e-02 -3.04164827e-01
5.60705602e-01 9.12094653e-01 -9.05633718e-02 2.09060416e-01
4.91000354e-01 -1.03480685e+00 1.23972821e+00 2.12791383e-01
1.04386115e+00 -7.02221274e-01 4.73565370e-01 3.71661782e-01
-1.27033925e+00 5.20425379e-01 -2.61950605e-02 8.95796493e-02
8.78952384e-01 7.25629330e-02 -1.26490963e+00 6.62094831e-01
-9.21124443e-02 5.24099708e-01 -2.25526560e-02 4.93049026e-01
-4.62527961e-01 1.34335601e+00 -3.62529606e-01 -1.85067460e-01
1.23282582e-01 7.37372460e-03 1.11474824e+00 1.87106013e+00
5.44557691e-01 -3.13493982e-02 3.12003911e-01 3.31816703e-01
-1.56717494e-01 4.65587825e-01 -5.66676319e-01 -3.69130224e-01
7.30159223e-01 7.93781817e-01 -5.17082810e-01 -7.33264208e-01
-1.22545436e-01 1.35972154e+00 5.52332364e-02 4.12629724e-01
-3.76791090e-01 -2.99825698e-01 5.20623267e-01 -1.05771869e-01
2.82793403e-01 -7.05039263e-01 -1.72126099e-01 -1.29067755e+00
-6.33604228e-02 -1.20287716e+00 -1.24127537e-01 -7.73622453e-01
-7.49531806e-01 1.17462885e+00 -4.36252989e-02 -6.97546721e-01
-1.19889057e+00 -4.84749936e-02 -3.09512764e-01 1.33436501e+00
-1.27743697e+00 -1.14150155e+00 7.59907961e-01 4.63969678e-01
1.20559204e+00 -4.93360519e-01 8.99869204e-01 4.06246960e-01
-4.34275180e-01 5.55154204e-01 1.94428265e-01 -1.02535218e-01
6.43341601e-01 -1.06735790e+00 1.69886923e+00 1.09552300e+00
6.25935018e-01 2.14194715e-01 9.86381888e-01 -7.86803067e-01
-1.71231639e+00 -1.17451620e+00 1.74088013e+00 -6.19407892e-01
8.74000371e-01 -8.38457346e-01 -7.66690671e-01 1.04127359e+00
7.86497712e-01 -1.76373005e-01 1.86387599e-01 -6.28036037e-02
1.57836631e-01 1.07674368e-01 -5.13121367e-01 6.33727312e-01
9.90305722e-01 -8.02462161e-01 -5.97664893e-01 7.58793771e-01
1.64503562e+00 -9.16877925e-01 -7.38071799e-01 -3.01472265e-02
2.89845437e-01 8.58568251e-02 6.50871038e-01 -7.21242189e-01
2.73142725e-01 3.53321061e-02 -5.82416415e-01 -1.76167333e+00
1.91772223e-01 -1.59486079e+00 -1.33934915e-01 1.27545178e+00
9.61858809e-01 -3.20294946e-01 2.72323906e-01 -1.56601086e-01
-7.20464766e-01 -2.19344557e-03 -1.44529355e+00 -9.73940074e-01
4.97007556e-03 -8.55804086e-01 3.77107263e-01 5.55311918e-01
1.84213877e-01 8.46603811e-01 -5.10205209e-01 1.78263009e-01
1.99672893e-01 -4.08185959e-01 6.70095384e-01 -5.52066803e-01
-6.43586934e-01 -7.27443993e-02 -2.07875222e-02 -1.11346853e+00
2.53971606e-01 -1.34731054e+00 5.46367884e-01 -1.66183650e+00
-4.71364409e-01 3.96017507e-02 2.31725797e-01 3.06878656e-01
2.07027256e-01 -6.73646405e-02 5.49198031e-01 2.44673118e-01
-4.17704850e-01 5.75579464e-01 8.81596625e-01 1.15753219e-01
-5.60338914e-01 2.76347905e-01 -1.46867529e-01 -3.22991759e-02
6.02487028e-01 -6.49783850e-01 -2.71102399e-01 -1.02086246e+00
1.95441976e-01 1.00294578e+00 -2.44751409e-01 -5.33378422e-01
3.70614648e-01 -2.17914596e-01 -3.02571505e-01 -7.05660105e-01
5.84376276e-01 -1.34657696e-01 2.05319077e-01 2.23853916e-01
-6.03754580e-01 6.59400225e-01 5.38284063e-01 -5.41294478e-02
-4.13237482e-01 4.40095216e-02 3.86510938e-01 -6.94287717e-02
-2.19282970e-01 8.37885514e-02 -1.00699997e+00 3.34437877e-01
5.31835318e-01 1.06566794e-01 -8.50768015e-02 -6.46818042e-01
-8.89386117e-01 1.85724393e-01 -2.28859901e-01 7.38638103e-01
7.78671086e-01 -9.86473680e-01 -1.57196236e+00 3.41222852e-01
-2.86487281e-01 -3.79204750e-01 -2.80045897e-01 9.62399840e-01
-4.02777582e-01 9.01502073e-01 3.35432172e-01 -4.93675232e-01
-1.48823512e+00 5.95930703e-02 1.33960277e-01 -1.93227306e-01
-4.91343141e-01 1.08767498e+00 -6.00746393e-01 -7.36277640e-01
1.32190421e-01 -2.95789927e-01 3.68752748e-01 -1.90477386e-01
5.69614351e-01 3.46305490e-01 5.58365524e-01 -9.25320446e-01
-4.79201674e-01 -1.79631576e-01 -1.68681696e-01 -1.24904442e+00
1.01160693e+00 -6.33070529e-01 -5.26685221e-03 8.60847831e-01
1.03634071e+00 2.25373924e-01 -6.22791886e-01 -3.42453033e-01
2.08347172e-01 1.46462277e-01 2.45413959e-01 -9.43319678e-01
-5.15283048e-01 1.11982954e+00 -3.64710912e-02 4.54879403e-02
5.29030204e-01 3.44545133e-02 1.35246611e+00 4.28747356e-01
6.75708726e-02 -1.46687388e+00 -6.52864873e-01 1.32916069e+00
1.02506685e+00 -1.04407585e+00 -5.69622159e-01 -3.07068139e-01
-8.00706506e-01 1.01806450e+00 -7.05284104e-02 1.75599471e-01
2.23557413e-01 5.49578905e-01 3.09870511e-01 3.87862563e-01
-1.55803525e+00 6.99400017e-03 5.79069853e-01 5.48914850e-01
7.38612235e-01 3.63432556e-01 -3.40392888e-01 1.49533808e-01
-5.53101778e-01 -1.06930189e-01 3.33829910e-01 5.88912964e-01
-4.07450169e-01 -1.31921160e+00 -1.42272070e-01 -1.22152992e-01
-5.32998979e-01 -9.25659359e-01 -5.31854331e-01 4.03561592e-01
-6.38269305e-01 1.20426941e+00 4.63478528e-02 -4.26915854e-01
5.11548936e-01 2.59594083e-01 6.14759147e-01 -1.01367915e+00
-8.81445289e-01 5.23804843e-01 8.50073576e-01 -7.58860826e-01
1.52749881e-01 -1.07119763e+00 -1.16470003e+00 -3.11207891e-01
-1.74722359e-01 4.49440449e-01 1.27271652e+00 1.15876400e+00
3.89463395e-01 6.45080924e-01 6.08588994e-01 -9.22010243e-01
-6.51968718e-01 -1.27624178e+00 -2.77340800e-01 -4.71973985e-01
3.94956678e-01 5.93484938e-01 -2.54194081e-01 1.28217682e-01] | [14.498761177062988, 7.175112724304199] |
fd06bad4-59b0-4e4d-ae1f-2badc686feab | a-multi-horizon-quantile-recurrent-forecaster | 1711.11053 | null | http://arxiv.org/abs/1711.11053v2 | http://arxiv.org/pdf/1711.11053v2.pdf | A Multi-Horizon Quantile Recurrent Forecaster | We propose a framework for general probabilistic multi-step time series
regression. Specifically, we exploit the expressiveness and temporal nature of
Sequence-to-Sequence Neural Networks (e.g. recurrent and convolutional
structures), the nonparametric nature of Quantile Regression and the efficiency
of Direct Multi-Horizon Forecasting. A new training scheme,
*forking-sequences*, is designed for sequential nets to boost stability and
performance. We show that the approach accommodates both temporal and static
covariates, learning across multiple related series, shifting seasonality,
future planned event spikes and cold-starts in real life large-scale
forecasting. The performance of the framework is demonstrated in an application
to predict the future demand of items sold on Amazon.com, and in a public
probabilistic forecasting competition to predict electricity price and load. | ['Ruofeng Wen', 'Kari Torkkola', 'Balakrishnan Narayanaswamy', 'Dhruv Madeka'] | 2017-11-29 | null | null | null | null | ['time-series-regression'] | ['time-series'] | [ 7.92675614e-02 -3.34235787e-01 -4.15489376e-01 -7.93428600e-01
-9.85507965e-01 -7.50799775e-01 6.15813553e-01 6.23172801e-03
-1.55914858e-01 9.03796673e-01 4.59907591e-01 -5.85553765e-01
-3.31454366e-01 -6.83547199e-01 -7.07389176e-01 -5.73447227e-01
-8.20796013e-01 3.33674431e-01 -2.53178895e-01 -3.01298320e-01
-2.19842941e-02 7.35432446e-01 -1.25849688e+00 3.89183640e-01
2.15561956e-01 1.35717404e+00 -3.06339916e-02 7.41089165e-01
3.41707398e-03 9.55296695e-01 -4.30430025e-01 -2.40543559e-01
2.86153138e-01 8.03762823e-02 -1.81545228e-01 -3.87238920e-01
-1.92992374e-01 -4.20342505e-01 -2.05785364e-01 3.36625665e-01
4.35200125e-01 4.99038696e-01 6.55721784e-01 -1.46616769e+00
-5.06123662e-01 1.15495622e+00 -6.51916802e-01 5.75189471e-01
-1.34876490e-01 6.98913485e-02 1.23287845e+00 -5.53310692e-01
2.09585741e-01 1.21809971e+00 1.32447779e+00 2.42181778e-01
-1.39131844e+00 -6.51237845e-01 5.86546779e-01 -8.80140513e-02
-1.01736224e+00 -2.95463324e-01 7.46846080e-01 -5.42992413e-01
1.37076545e+00 1.81199029e-01 2.32766941e-01 1.45244443e+00
5.78508556e-01 9.16729093e-01 8.27126682e-01 3.79847400e-02
3.18606228e-01 -2.70356864e-01 5.12494221e-02 -9.05747563e-02
-5.69784284e-01 7.36655593e-01 -5.10213375e-01 -4.76742119e-01
3.56458634e-01 3.95700961e-01 3.64665180e-01 4.43066023e-02
-9.05323982e-01 9.43578601e-01 1.01303920e-01 1.26968458e-01
-7.25282907e-01 5.89412510e-01 9.75113511e-01 4.41892475e-01
1.04387057e+00 4.05806266e-02 -1.22639358e+00 -2.99978107e-01
-1.51898992e+00 4.35857207e-01 9.76537049e-01 8.52767467e-01
4.25243050e-01 6.84476495e-01 -4.46756661e-01 5.28813183e-01
2.38475744e-02 5.09354293e-01 5.77265739e-01 -6.58306420e-01
4.41955864e-01 -3.07934970e-01 2.64411926e-01 -7.55479932e-01
-8.95261884e-01 -4.76391613e-01 -1.12536335e+00 -6.91617876e-02
1.66818023e-01 -5.75770438e-01 -8.05567443e-01 1.81626892e+00
-1.28219932e-01 5.45177221e-01 -2.07046121e-01 3.09774160e-01
1.63338095e-01 8.97136688e-01 3.69254589e-01 -4.81718361e-01
8.49741340e-01 -7.57366538e-01 -6.71114147e-01 1.06773257e-01
3.85517061e-01 -4.23527151e-01 4.43100035e-01 4.28814948e-01
-1.00921512e+00 -5.04581630e-01 -5.78851521e-01 2.50432551e-01
-6.61526024e-01 1.13619246e-01 7.31001318e-01 4.33053881e-01
-9.38759327e-01 1.04392719e+00 -1.01897430e+00 8.07482228e-02
2.12055594e-01 3.50597680e-01 2.79675663e-01 4.77825075e-01
-1.54624343e+00 9.49244142e-01 5.72219670e-01 4.21066225e-01
-9.45621789e-01 -1.07589281e+00 -6.59812629e-01 4.49539602e-01
7.87257329e-02 -2.41948634e-01 1.66542184e+00 -1.10954595e+00
-1.59500194e+00 8.21232796e-02 -1.09811701e-01 -1.03014398e+00
5.75185180e-01 -1.92869872e-01 -9.11295295e-01 -5.39398730e-01
-1.46043092e-01 1.49096757e-01 8.12379479e-01 -4.63960111e-01
-8.42016578e-01 -6.82033747e-02 -2.99801201e-01 -2.75182307e-01
2.08485350e-02 4.36895669e-01 4.66504157e-01 -9.44524109e-01
-5.97510219e-01 -7.45060682e-01 -8.54932964e-01 -7.17808306e-01
-1.32739007e-01 -5.91177344e-01 5.84526658e-01 -9.58368659e-01
1.36705518e+00 -2.11461592e+00 -2.88615525e-01 5.68206072e-01
-3.09029013e-01 -3.56433004e-01 -2.18242645e-01 9.28120971e-01
-6.11907065e-01 3.89766172e-02 -3.86735424e-02 -5.15828252e-01
4.27010149e-01 2.94016510e-01 -1.08444285e+00 4.37372595e-01
3.74653220e-01 1.29920030e+00 -6.52395785e-01 2.20546678e-01
1.43811136e-01 2.98170418e-01 -9.59794000e-02 3.72156277e-02
-7.35754311e-01 1.84227914e-01 -2.03808039e-01 5.99390686e-01
6.00902438e-01 -3.27184051e-01 9.12814662e-02 3.31493318e-01
-4.69378650e-01 4.39021915e-01 -1.06951654e+00 1.33619010e+00
-6.98155880e-01 4.37981367e-01 -2.51002252e-01 -9.40717518e-01
8.38389874e-01 5.99405646e-01 7.71479964e-01 -6.44685805e-01
-1.76471680e-01 2.38122493e-01 -4.16350842e-01 2.29613595e-02
8.54334712e-01 -1.96270734e-01 -3.49914700e-01 7.17945457e-01
-4.08971822e-03 1.43250406e-01 2.42810339e-01 -2.33763784e-01
7.74948895e-01 5.29709160e-01 3.29564989e-01 -3.19172382e-01
1.28171131e-01 -2.93251485e-01 6.44527793e-01 9.43404078e-01
8.30645710e-02 3.91485870e-01 6.54497445e-01 -9.16638970e-01
-1.39933479e+00 -9.83858764e-01 -1.00702725e-01 1.91921616e+00
-7.48283386e-01 -1.20705485e-01 1.26705244e-01 -2.66935796e-01
3.83853137e-01 1.19308984e+00 -8.70765686e-01 2.09302619e-01
-7.07189560e-01 -1.12458956e+00 3.62200737e-01 9.67997909e-01
-5.02892613e-01 -1.47738636e+00 -3.82458448e-01 1.03660083e+00
2.75323391e-01 -8.08826625e-01 -5.35046041e-01 7.04783440e-01
-8.02944362e-01 -4.18786049e-01 -6.04679465e-01 -2.26491690e-01
-3.53614718e-01 -3.42189670e-01 1.52063370e+00 -8.27878952e-01
5.63694760e-02 3.29420507e-01 6.44724350e-03 -8.64992619e-01
-3.15720856e-01 2.66265452e-01 -2.55230162e-02 -1.29646778e-01
2.95245022e-01 -1.00036979e+00 -4.51174349e-01 9.67110172e-02
-1.07485807e+00 -3.40813279e-01 2.52318084e-01 1.05127645e+00
4.12558943e-01 1.01689175e-01 1.15430546e+00 -6.43307209e-01
9.00663376e-01 -1.00890863e+00 -9.91065204e-01 3.45177680e-01
-8.46662700e-01 -2.41658539e-01 8.40127587e-01 -7.39218831e-01
-9.04774666e-01 3.99994552e-02 -1.92995951e-01 -3.61817032e-01
-3.35124147e-04 1.03545058e+00 7.59532809e-01 5.90144277e-01
4.91175413e-01 4.21812057e-01 -2.25151613e-01 -3.66171300e-01
5.35338223e-01 2.46089965e-01 5.38946450e-01 -6.32288098e-01
6.10468388e-01 3.97333086e-01 -1.33488383e-02 -3.32790166e-01
-9.62609470e-01 -4.67715442e-01 -4.71676350e-01 1.46271084e-02
3.75576943e-01 -1.24572635e+00 -1.08197725e+00 3.90065342e-01
-1.17202735e+00 -5.01795888e-01 -5.87737262e-01 3.68890524e-01
-8.15056026e-01 -1.32688835e-01 -9.97753918e-01 -1.37286079e+00
-6.30636334e-01 -5.53702950e-01 9.87941146e-01 9.87378731e-02
-1.77661151e-01 -1.31265831e+00 2.71701723e-01 -6.85839295e-01
9.11027730e-01 4.20007318e-01 9.44352925e-01 -1.05324590e+00
-2.31050640e-01 -2.06981018e-01 -6.11627214e-02 4.30234522e-01
-1.61303192e-01 2.20878914e-01 -9.94679570e-01 -3.73757899e-01
-1.64211437e-01 -1.97923690e-01 8.63594830e-01 8.38432312e-01
1.53511643e+00 -4.13836151e-01 -8.87692124e-02 4.96878535e-01
1.34060454e+00 2.47879028e-01 3.42150748e-01 2.88848817e-01
4.63887393e-01 6.17056191e-01 1.96751118e-01 1.01098943e+00
4.33454037e-01 2.87627041e-01 2.76844680e-01 -6.53744116e-02
7.13458538e-01 -2.63595879e-01 6.02451086e-01 9.34595227e-01
1.05006889e-01 -3.29209268e-01 -9.13846612e-01 7.22112834e-01
-2.10105467e+00 -1.40065193e+00 3.48561145e-02 1.98218954e+00
7.17701256e-01 2.02078536e-01 4.39101040e-01 -2.42906839e-01
5.13948143e-01 3.93697202e-01 -7.27597654e-01 -8.36807430e-01
-4.85023320e-01 5.14618754e-01 8.29013646e-01 1.03087507e-01
-1.35247612e+00 6.49642587e-01 7.44543743e+00 1.09467959e+00
-1.06675875e+00 4.03443873e-02 1.13677967e+00 -2.66172975e-01
-4.19571310e-01 -3.30558181e-01 -9.18683529e-01 6.04383111e-01
1.83246958e+00 -3.99907440e-01 6.33075535e-01 8.00062537e-01
5.35245061e-01 4.09938246e-01 -1.27670455e+00 4.87413883e-01
-4.18660551e-01 -1.62536275e+00 -4.00110006e-01 -2.47569636e-01
8.99973094e-01 7.32156932e-01 3.08598042e-01 6.98346138e-01
9.11847472e-01 -1.38164449e+00 7.35893130e-01 8.35957229e-01
5.29829144e-01 -1.17540908e+00 6.97676420e-01 3.68908107e-01
-1.24497390e+00 -4.98001724e-01 -7.92368501e-02 1.64095666e-02
5.32542825e-01 8.88687313e-01 -6.43245816e-01 7.17986882e-01
7.82489538e-01 9.76428211e-01 1.15551334e-02 8.33649278e-01
1.82091326e-01 8.88913155e-01 -7.04187572e-01 -1.26612606e-02
6.66562378e-01 -1.74602598e-01 2.34540910e-01 1.61375177e+00
3.41317475e-01 -3.63149524e-01 2.85741746e-01 6.09481275e-01
1.07434630e-01 -3.59653682e-02 -5.72321773e-01 -1.26664847e-01
9.85867158e-02 1.10743964e+00 -3.09461564e-01 -2.57088363e-01
-6.38367772e-01 3.64023000e-01 9.33067054e-02 7.77366817e-01
-9.07839179e-01 -1.39103774e-02 5.02588093e-01 -3.58042002e-01
8.61446977e-01 -3.85776699e-01 -3.26636523e-01 -8.93716693e-01
1.00801443e-03 -7.62370348e-01 7.14855790e-01 -7.45330453e-01
-2.05140805e+00 4.56206143e-01 4.69500795e-02 -1.02831626e+00
-1.00827861e+00 -5.88106990e-01 -8.96244884e-01 1.30312705e+00
-1.78572345e+00 -1.25035524e+00 6.92713618e-01 5.51304460e-01
5.36715567e-01 -1.84074551e-01 7.89410710e-01 1.19771302e-01
-3.49390984e-01 3.26332182e-01 8.65505219e-01 -1.63445532e-01
3.31867844e-01 -1.33098841e+00 1.08603537e+00 5.46108842e-01
-9.13371295e-02 3.03202927e-01 7.16488957e-01 -4.86908048e-01
-1.02844501e+00 -1.16765463e+00 1.05749679e+00 -2.70330489e-01
1.49114823e+00 -4.56047237e-01 -9.38685060e-01 9.09114957e-01
3.16019386e-01 -2.40894243e-01 6.96274757e-01 5.35595119e-01
-4.98715401e-01 -2.06290156e-01 -8.81439328e-01 3.75655979e-01
2.15594381e-01 -7.26183772e-01 -2.88304746e-01 5.86210728e-01
9.03304935e-01 -3.36148679e-01 -1.07679427e+00 4.56648827e-01
8.44591260e-01 -5.69713831e-01 8.36171150e-01 -1.07497251e+00
4.70726579e-01 1.38266563e-01 -2.10234836e-01 -1.23033905e+00
-4.85319376e-01 -1.14248908e+00 -5.35639286e-01 1.10565829e+00
7.18674839e-01 -6.79466546e-01 5.23965716e-01 7.11744189e-01
-3.87013033e-02 -6.76308334e-01 -1.22671771e+00 -8.19579303e-01
3.07029307e-01 -8.89802575e-01 9.27087903e-01 9.43984807e-01
-4.33039010e-01 -2.42167279e-01 -9.35627759e-01 2.63305604e-01
2.71275997e-01 4.61937606e-01 2.53464222e-01 -1.13197351e+00
-4.15023386e-01 -5.78517795e-01 1.27307087e-01 -8.84053051e-01
3.73216093e-01 -6.40452087e-01 -2.98412088e-02 -7.31505275e-01
-1.64337546e-01 -2.64875263e-01 -1.04495358e+00 5.47212005e-01
2.03904063e-01 -1.95115730e-01 6.43969625e-02 2.66173342e-03
-4.54083145e-01 7.30917573e-01 4.45309460e-01 5.25400527e-02
-2.67210841e-01 7.48482466e-01 -2.10167453e-01 3.50176424e-01
1.14154649e+00 -6.36359811e-01 -9.70673412e-02 5.81088886e-02
6.19565547e-01 5.85892081e-01 3.14847291e-01 -5.12887657e-01
1.69244617e-01 -4.91011620e-01 5.63127160e-01 -1.27215314e+00
2.42565542e-01 -7.63157964e-01 3.84245098e-01 1.89012125e-01
-7.52563298e-01 7.49587834e-01 5.76795995e-01 8.62085223e-01
-1.31151766e-01 8.15709531e-02 2.13884354e-01 -4.32698894e-03
-5.77656031e-01 6.75124645e-01 -4.86798853e-01 -1.84947535e-01
6.40037894e-01 2.18146130e-01 4.71259421e-03 -6.88919961e-01
-8.52925658e-01 5.88865399e-01 -4.41168398e-01 6.25825763e-01
1.74954966e-01 -1.30468321e+00 -1.03162122e+00 -2.09848657e-02
-1.69815466e-01 -4.62133974e-01 4.03454214e-01 7.97494531e-01
1.15341917e-01 4.10622507e-01 1.01858899e-01 -5.05978942e-01
-5.45654476e-01 7.04018593e-01 3.85275304e-01 -7.87757218e-01
-5.02724290e-01 5.24966002e-01 -3.23756039e-02 -5.20622432e-01
2.79377133e-01 -5.58677256e-01 -5.55496812e-02 3.69755805e-01
4.56986248e-01 3.45733941e-01 1.11830913e-01 -1.39703631e-01
-8.36476460e-02 -1.83482468e-01 -1.28043771e-01 -1.60773352e-01
1.94694841e+00 -1.61398008e-01 -2.65422106e-01 1.27624190e+00
1.04361415e+00 -3.87468874e-01 -1.56925416e+00 -4.45339024e-01
5.84428430e-01 2.45515347e-01 -1.79838732e-01 -1.03249085e+00
-8.98930252e-01 6.91481829e-01 4.39940065e-01 7.85289466e-01
1.02060294e+00 -4.22740549e-01 7.54073381e-01 3.48785311e-01
4.11460400e-02 -1.28432584e+00 -7.30580449e-01 9.19521987e-01
8.24847043e-01 -1.01088309e+00 -3.22335474e-02 4.51559961e-01
-6.63609684e-01 1.50044096e+00 -2.78946459e-02 -2.48971850e-01
1.02540016e+00 4.88830864e-01 -8.58579725e-02 1.84467942e-01
-1.38920486e+00 1.47119105e-01 3.89869660e-01 3.46760720e-01
4.88932759e-01 4.11489427e-01 5.44498935e-02 6.19052351e-01
-1.11266688e-01 8.04890469e-02 3.16612899e-01 6.14912868e-01
1.62941933e-01 -7.98556983e-01 -3.58024687e-02 7.08008170e-01
-8.19176853e-01 -4.69015598e-01 5.40499091e-01 7.47442245e-01
-3.17189217e-01 7.01654434e-01 2.65562624e-01 7.96527490e-02
1.37270719e-01 3.49764645e-01 -2.15105385e-01 -1.36098728e-01
-1.15801811e+00 2.41098210e-01 9.06470045e-02 -6.75350785e-01
-4.06149000e-01 -8.56846988e-01 -7.52875209e-01 -3.74196500e-01
-1.27777457e-05 -8.81161541e-02 9.27260518e-01 8.79181266e-01
2.87272751e-01 7.34192789e-01 1.08249784e+00 -1.04726672e+00
-1.10924196e+00 -1.02120721e+00 -8.88940036e-01 4.94760275e-03
6.93613052e-01 -1.23005703e-01 -3.83488864e-01 -9.45049971e-02] | [6.89016580581665, 3.1092891693115234] |
a471f019-4839-4189-a643-74dc7945d6cb | deepflorist-rethinking-deep-neural-networks | 2307.01806 | null | https://arxiv.org/abs/2307.01806v1 | https://arxiv.org/pdf/2307.01806v1.pdf | DeepFlorist: Rethinking Deep Neural Networks and Ensemble Learning as A Meta-Classifier For Object Classification | In this paper, we propose a novel learning paradigm called "DeepFlorist" for flower classification using ensemble learning as a meta-classifier. DeepFlorist combines the power of deep learning with the robustness of ensemble methods to achieve accurate and reliable flower classification results. The proposed network architecture leverages a combination of dense convolutional and convolutional neural networks (DCNNs and CNNs) to extract high-level features from flower images, followed by a fully connected layer for classification. To enhance the performance and generalization of DeepFlorist, an ensemble learning approach is employed, incorporating multiple diverse models to improve the classification accuracy. Experimental results on benchmark flower datasets demonstrate the effectiveness of DeepFlorist, outperforming state-of-the-art methods in terms of accuracy and robustness. The proposed framework holds significant potential for automated flower recognition systems in real-world applications, enabling advancements in plant taxonomy, conservation efforts, and ecological studies. | ['Afshin Khadangi'] | 2023-07-04 | null | null | null | null | ['classification-1'] | ['methodology'] | [-1.21702418e-01 -5.19002259e-01 -2.58753777e-01 -3.50605547e-01
1.14712231e-01 -8.31961095e-01 4.38359320e-01 8.55486095e-02
2.32134700e-01 6.98266625e-01 -3.38496417e-01 -2.65640020e-01
-2.09389806e-01 -1.07812309e+00 -4.96860862e-01 -9.61131275e-01
-1.80706352e-01 -2.88907200e-01 -2.09990382e-01 -2.71790236e-01
2.04534426e-01 8.92349243e-01 -1.84082103e+00 3.06413025e-01
1.22604251e+00 1.47994721e+00 1.16143771e-01 5.29405355e-01
-1.73350945e-01 7.25877523e-01 -4.59620237e-01 -4.27579790e-01
4.26146351e-02 -9.29047093e-02 -5.03216147e-01 -1.48596177e-02
4.50414091e-01 -1.52299985e-01 -3.14560562e-01 5.77858210e-01
3.27779382e-01 -1.12999074e-01 7.05491900e-01 -1.10765934e+00
-1.11078811e+00 3.67796212e-01 -3.24897945e-01 6.74343631e-02
-2.72323102e-01 5.05403399e-01 7.08803654e-01 -9.16331768e-01
3.63787413e-02 1.10100389e+00 1.01800001e+00 2.38887355e-01
-1.22264838e+00 -8.67150068e-01 2.22467735e-01 -1.15620770e-01
-1.11765504e+00 -4.26172942e-01 4.94248658e-01 -2.56159127e-01
5.84453821e-01 8.62678885e-02 6.87773764e-01 1.06402886e+00
5.74949443e-01 5.93553662e-01 1.05551481e+00 -1.74983650e-01
1.39921933e-01 -1.70830607e-01 2.11134955e-01 9.95706856e-01
3.89454961e-01 1.59129262e-01 -3.35216850e-01 2.68467069e-02
8.91993344e-01 2.87873954e-01 1.32975522e-02 1.80577919e-01
-7.45795965e-01 6.14747465e-01 1.15224123e+00 4.45049167e-01
-5.27525187e-01 1.58477142e-01 4.27735209e-01 1.59167767e-01
4.81903255e-01 5.44825435e-01 -4.86463785e-01 4.17368084e-01
-7.86678433e-01 1.10943623e-01 6.31185472e-01 7.05904245e-01
4.20500636e-01 4.22589123e-01 -4.51139480e-01 7.97884524e-01
2.92897373e-01 4.10422087e-01 3.42611611e-01 -7.26971686e-01
-1.18616112e-01 1.09212160e+00 3.10837887e-02 -1.19976711e+00
-2.25937799e-01 -1.02111745e+00 -1.21757841e+00 1.51292190e-01
9.95502025e-02 1.51185453e-01 -9.68693733e-01 1.36548030e+00
2.29749784e-01 6.28065228e-01 1.14767648e-01 4.49119240e-01
1.24203980e+00 2.64653087e-01 2.42440790e-01 2.02404603e-01
1.19489181e+00 -8.91793370e-01 -3.93183142e-01 1.88759834e-01
3.08688074e-01 -4.76626009e-01 5.71090937e-01 3.96922112e-01
-6.30594611e-01 -9.98158753e-01 -1.15724194e+00 2.29118347e-01
-5.62785983e-01 7.62879133e-01 1.04736686e+00 7.98916221e-01
-7.99677312e-01 8.15406859e-01 -6.25475883e-01 -2.82312125e-01
9.97563183e-01 5.72129190e-01 -5.00906646e-01 4.82766405e-02
-6.61535382e-01 6.43947661e-01 5.43591082e-01 6.99244440e-01
-9.94639575e-01 -6.03693068e-01 -6.04951262e-01 3.15399200e-01
-3.52753475e-02 -5.74309886e-01 1.03977299e+00 -9.88612354e-01
-1.44927704e+00 3.99048418e-01 -7.24679753e-02 -4.49416786e-01
-6.36417121e-02 -8.67488235e-02 -6.65754318e-01 -3.67008597e-02
-2.11082205e-01 4.92270559e-01 9.41415370e-01 -1.25783396e+00
-7.06185639e-01 -6.42161667e-01 -2.72648305e-01 -5.03320992e-01
-7.73314595e-01 -6.63978159e-01 2.16988713e-01 -5.61803102e-01
2.80301750e-01 -9.39750016e-01 -1.09564267e-01 1.31886989e-01
-1.87285230e-01 -4.66055721e-01 1.55684638e+00 -3.22387934e-01
1.16363168e+00 -1.98089373e+00 -1.11736141e-01 2.49214217e-01
4.11119074e-01 8.58980596e-01 -4.64160889e-01 3.99659187e-01
1.41528934e-01 1.25497803e-01 -7.05912411e-02 -2.78418809e-01
-4.75823641e-01 3.57818305e-01 -2.51632601e-01 3.11088771e-01
1.02556646e+00 1.07891250e+00 -9.41464663e-01 -1.16403289e-01
4.73108202e-01 6.85538292e-01 -1.36187077e-01 1.96055457e-01
-1.78181574e-01 6.03641391e-01 -5.38381338e-01 1.28207517e+00
9.49967146e-01 -3.73688579e-01 1.24537177e-01 -3.66713315e-01
-1.99766114e-01 -2.38591671e-01 -7.39266992e-01 1.33331430e+00
-7.01372206e-01 3.65663290e-01 -6.84549287e-02 -8.79467487e-01
1.40960956e+00 6.54075965e-02 1.80027947e-01 -1.69436634e-01
3.02441180e-01 3.96834224e-01 4.99954633e-02 -3.52932841e-01
4.03161645e-01 6.35606527e-01 2.78476387e-01 -9.98762995e-02
7.10288763e-01 1.55874401e-01 2.23955497e-01 -2.08745450e-01
9.17606950e-01 2.96537548e-01 1.39284402e-01 -4.81281966e-01
7.50716031e-01 -3.21676254e-01 5.12914658e-01 7.73466349e-01
-3.17125142e-01 1.48070809e-02 1.73100874e-01 -8.56552482e-01
-4.76400137e-01 -8.14105153e-01 -3.85672301e-01 1.05114090e+00
7.48481527e-02 -3.33618909e-01 -5.64854264e-01 -8.39255035e-01
4.44251984e-01 2.91599572e-01 -9.69424069e-01 -2.74887204e-01
-2.74481148e-01 -8.49713206e-01 1.08054757e+00 8.93532634e-01
1.04807246e+00 -1.13977253e+00 -4.83344525e-01 3.85475576e-01
2.40549706e-02 -9.48464036e-01 1.92222610e-01 4.00048524e-01
-9.57835555e-01 -1.12825394e+00 -2.02434957e-01 -6.75042391e-01
4.16532755e-01 3.43874007e-01 1.09724820e+00 6.14073277e-01
-4.64325488e-01 -2.01315075e-01 -5.19937336e-01 -5.41795850e-01
-2.20742673e-01 3.14589083e-01 -8.10825527e-02 2.25874126e-01
1.26111552e-01 -7.46319532e-01 -7.93174803e-01 1.63356572e-01
-6.18511558e-01 -5.15111089e-01 7.71632195e-01 1.35028005e+00
3.58779073e-01 1.16999879e-01 1.14324629e+00 -6.07578516e-01
4.21175390e-01 -7.69981980e-01 -6.74063861e-01 5.13448119e-01
-4.72954452e-01 -1.09229609e-01 8.24962795e-01 -3.06640685e-01
-9.04497683e-01 1.40006810e-01 2.01589745e-02 -3.76007527e-01
-2.92981304e-02 6.79990351e-01 -7.19020367e-02 -6.47337198e-01
6.50571048e-01 2.22247958e-01 1.08183451e-01 -2.59231687e-01
3.26525629e-01 8.83578002e-01 7.70718157e-01 -4.95585442e-01
3.00639093e-01 2.19039172e-01 3.89784366e-01 -8.11883748e-01
-7.75412142e-01 -2.01944172e-01 -7.30588198e-01 -5.30655980e-01
3.77037048e-01 -8.79017591e-01 -1.18875790e+00 8.02044213e-01
-8.66196394e-01 8.88651088e-02 1.90202534e-01 3.85826081e-02
-1.57645091e-01 -1.29311875e-01 -5.20590067e-01 -1.03172529e+00
-7.04753041e-01 -1.13561773e+00 1.05918539e+00 8.06597173e-01
2.70367652e-01 -8.92351687e-01 -3.45865667e-01 3.73053104e-02
7.36069143e-01 7.28656173e-01 6.66718781e-01 -4.51040953e-01
-5.31949103e-01 -2.83823371e-01 -3.16886991e-01 4.50107694e-01
4.27895248e-01 6.41255260e-01 -1.35608637e+00 -2.51937479e-01
-5.17859697e-01 -5.71563303e-01 1.22091484e+00 2.72111475e-01
1.74850333e+00 -1.70911282e-01 -5.05402684e-01 6.67403698e-01
1.60970545e+00 -3.70435305e-02 5.44834256e-01 -4.70699705e-02
5.14239132e-01 3.15244943e-01 1.28077686e-01 4.73894417e-01
2.42422536e-01 4.96678241e-02 1.04993808e+00 1.37163801e-02
7.24288672e-02 -7.61452690e-02 -1.22102074e-01 3.95298451e-01
-2.89884657e-01 -5.20249426e-01 -8.46267700e-01 4.77491409e-01
-1.88790298e+00 -1.13736701e+00 -2.43173942e-01 1.98532665e+00
8.62520486e-02 -3.06424916e-01 -2.21651811e-02 2.08492935e-01
5.27157187e-01 -1.83133364e-01 -9.12310064e-01 -4.24406439e-01
-3.90466660e-01 5.43389857e-01 1.77421287e-01 -3.98119777e-01
-1.51416445e+00 1.03741920e+00 6.39941406e+00 5.87855279e-01
-1.52640331e+00 -2.20925972e-01 9.09776449e-01 3.44880372e-01
3.06303203e-01 -3.61037582e-01 -8.83018434e-01 3.81827235e-01
7.75440335e-01 4.60895002e-01 2.24990308e-01 8.04942489e-01
-8.58653635e-02 1.25323415e-01 -6.51377261e-01 5.49358070e-01
-1.22336991e-01 -1.71317756e+00 -1.11113705e-01 -4.91981730e-02
8.20705473e-01 7.08059371e-02 3.43105078e-01 4.03671920e-01
4.94478792e-01 -1.42564213e+00 1.97583884e-01 6.36068404e-01
6.25059605e-01 -1.05178213e+00 8.65895748e-01 2.81431586e-01
-1.64134562e+00 -6.25557005e-01 -4.28769678e-01 6.45995289e-02
-6.44923568e-01 5.86722791e-01 -5.46305418e-01 1.10938036e+00
1.08847988e+00 1.20612013e+00 -9.50439692e-01 9.51858461e-01
1.76686689e-01 5.45014143e-01 -1.97089612e-01 -5.07080182e-02
3.72226179e-01 1.48920715e-02 -3.04217022e-02 1.16659498e+00
3.89336199e-01 -8.39531943e-02 2.97342122e-01 8.63387585e-01
-4.49135244e-01 -1.43566849e-02 -6.91615343e-01 -4.23212558e-01
7.47942388e-01 1.91859090e+00 -8.18888485e-01 -1.26807347e-01
-7.14382902e-02 6.64745688e-01 5.86280525e-01 -6.21257089e-02
-4.83585566e-01 -3.62362891e-01 5.71389735e-01 -4.66988355e-01
7.13129818e-01 5.63308503e-03 -5.23730874e-01 -9.23745215e-01
-1.84846982e-01 -7.96408057e-01 2.57171303e-01 -4.76063073e-01
-1.75532043e+00 9.11629796e-01 -5.43828547e-01 -1.21750176e+00
1.56932414e-01 -9.72746789e-01 -9.26963151e-01 8.96371186e-01
-1.42266369e+00 -2.03488135e+00 -7.78199613e-01 3.39647233e-01
2.46133491e-01 -6.26249194e-01 1.15769446e+00 -3.38586211e-01
-8.19245696e-01 6.17327929e-01 4.71274108e-01 2.12504238e-01
3.71521354e-01 -1.01832449e+00 2.40044668e-01 7.22968459e-01
-9.19880420e-02 6.37103438e-01 1.27202287e-01 -4.43488836e-01
-1.48228669e+00 -1.67742884e+00 4.36312765e-01 -3.04399669e-01
3.64569217e-01 -1.04793116e-01 -8.53539586e-01 2.32345328e-01
1.85949206e-01 6.42632127e-01 7.62710154e-01 2.67869264e-01
-6.10491097e-01 -5.01006484e-01 -1.49958193e+00 2.15048313e-01
7.81181574e-01 -3.61831844e-01 1.20258376e-01 1.72215372e-01
2.96726227e-01 -2.17783853e-01 -1.27664936e+00 8.80382597e-01
1.16186702e+00 -9.32739615e-01 7.52770066e-01 -6.63853168e-01
5.43453813e-01 -3.78836691e-01 -2.85927713e-01 -1.36023068e+00
-5.56455374e-01 -2.06582800e-01 -4.32135463e-01 1.28523088e+00
1.63799852e-01 -5.26322246e-01 5.83833158e-01 1.43478392e-02
-1.50241345e-01 -1.11554015e+00 -5.46062171e-01 -6.33976161e-01
1.35101661e-01 5.69856204e-02 1.02915406e+00 8.08052421e-01
-4.66784090e-01 6.10020757e-02 -6.86588585e-02 3.75511795e-01
4.42432284e-01 3.11861128e-01 6.51137173e-01 -1.70719111e+00
1.08017221e-01 -4.78224039e-01 -6.69814467e-01 -2.73725688e-01
4.38545734e-01 -8.95634949e-01 -2.15897352e-01 -1.15854776e+00
-1.13101035e-01 -5.79411447e-01 -5.91085553e-01 7.06220567e-01
-3.11714858e-01 5.20172477e-01 2.95495898e-01 8.60620290e-02
-3.34326804e-01 7.48130202e-01 1.18580151e+00 -1.36923864e-01
-2.39096820e-01 5.58370687e-02 -8.06823075e-01 4.55734879e-01
9.31740701e-01 -3.21964882e-02 -1.51720420e-01 -4.91325736e-01
-5.02419472e-01 -2.36104369e-01 6.73519015e-01 -1.31376922e+00
1.22963436e-01 2.48308942e-01 1.17432654e+00 -5.96801162e-01
1.10192567e-01 -7.44370639e-01 1.63718779e-02 6.36123180e-01
-9.44456309e-02 2.53057837e-01 5.69187045e-01 6.61247671e-01
-8.16647410e-02 1.50933355e-01 8.74190032e-01 8.68477579e-03
-6.83699608e-01 4.89794403e-01 -2.54746437e-01 -7.31315196e-01
1.01495242e+00 -1.19542830e-01 -5.49374759e-01 2.33019546e-01
-5.35248697e-01 3.57992530e-01 1.34416729e-01 5.32228053e-01
6.94902658e-01 -1.23297286e+00 -8.39964151e-01 5.61135352e-01
4.08775121e-01 -1.74642026e-01 2.43502796e-01 6.17068529e-01
-4.30371523e-01 2.85032183e-01 -5.79159200e-01 -9.32919264e-01
-1.02888262e+00 1.36781767e-01 5.52642107e-01 -1.98057160e-01
-1.27448395e-01 7.39081919e-01 -4.30383921e-01 -7.28491366e-01
1.79827332e-01 -1.33928493e-01 -5.00928640e-01 -1.43910618e-02
5.99834919e-01 2.78548688e-01 4.76706773e-01 -5.77139974e-01
-2.96127111e-01 6.03887327e-02 -2.41279826e-01 7.03578174e-01
1.48583496e+00 2.70026416e-01 -2.20789775e-01 2.40393564e-01
7.53764331e-01 -6.89315557e-01 -1.19994986e+00 -7.98313394e-02
4.05418836e-02 -2.99975991e-01 3.48072082e-01 -1.16087294e+00
-1.25104535e+00 7.74136424e-01 9.63109314e-01 1.60186961e-01
1.40545392e+00 -6.32663012e-01 3.70620489e-01 8.19348097e-01
2.91195780e-01 -5.93550265e-01 3.93625833e-02 4.96179104e-01
7.83586562e-01 -1.50523388e+00 -2.27272317e-01 -2.22285897e-01
-1.29157156e-01 1.56286633e+00 9.31267917e-01 -5.33244073e-01
9.34454978e-01 3.17683458e-01 -1.46775961e-01 1.02515355e-01
-9.96084869e-01 -2.80723929e-01 3.11401427e-01 6.42297685e-01
6.65174901e-01 1.78848967e-01 1.77680373e-01 5.89862645e-01
2.77496427e-01 3.61729205e-01 -3.78451236e-02 8.86439741e-01
-2.62528956e-01 -1.11519825e+00 -1.73084050e-01 9.04668331e-01
-2.96077371e-01 -1.66978443e-03 -6.29044175e-01 6.90337002e-01
4.57981884e-01 1.13683963e+00 1.78522617e-01 -6.74773693e-01
8.87934342e-02 -9.43495929e-02 6.59019530e-01 -2.02039570e-01
-1.35459507e+00 -3.48363519e-01 -2.54612207e-01 -4.36799616e-01
-6.36271834e-01 -2.36006722e-01 -4.90118772e-01 -6.25181973e-01
-7.41002083e-01 -1.25658110e-01 6.78141594e-01 1.04932821e+00
6.98284209e-01 6.36710942e-01 1.03365684e+00 -9.92070913e-01
-5.02064347e-01 -1.06553686e+00 -4.08963084e-01 -1.78814381e-01
3.44794065e-01 -7.42737710e-01 1.34102032e-02 -1.90604657e-01] | [9.435924530029297, -1.2391856908798218] |
bfcda236-1177-4569-8b00-95019e84703e | deep-learning-based-novel-cascaded-approach | 2301.06226 | null | https://arxiv.org/abs/2301.06226v1 | https://arxiv.org/pdf/2301.06226v1.pdf | Deep Learning based Novel Cascaded Approach for Skin Lesion Analysis | Automatic lesion analysis is critical in skin cancer diagnosis and ensures effective treatment. The computer aided diagnosis of such skin cancer in dermoscopic images can significantly reduce the clinicians workload and help improve diagnostic accuracy. Although researchers are working extensively to address this problem, early detection and accurate identification of skin lesions remain challenging. This research focuses on a two step framework for skin lesion segmentation followed by classification for lesion analysis. We explored the effectiveness of deep convolutional neural network based architectures by designing an encoder-decoder architecture for skin lesion segmentation and CNN based classification network. The proposed approaches are evaluated quantitatively in terms of the Accuracy, mean Intersection over Union and Dice Similarity Coefficient. Our cascaded end to end deep learning based approach is the first of its kind, where the classification accuracy of the lesion is significantly improved because of prior segmentation. | ['Sanjay Talbar', 'Ujjwal Baid', 'Bhakti Baheti', 'Prasad Dutande', 'Shubham Innani'] | 2023-01-16 | null | null | null | null | ['skin-lesion-segmentation'] | ['medical'] | [ 5.10108173e-01 1.68435857e-01 -4.61731762e-01 -3.44464749e-01
-7.82454431e-01 -4.26196337e-01 2.71421403e-01 3.43915671e-01
-4.85651374e-01 4.38920945e-01 -1.23394154e-01 -3.85580122e-01
-5.52223362e-02 -6.66652739e-01 -1.39124528e-01 -7.87612557e-01
1.38737470e-01 1.10551454e-01 1.67664632e-01 2.06026852e-01
3.52801770e-01 6.21586204e-01 -9.80744362e-01 4.03016925e-01
1.15963423e+00 9.76577938e-01 5.14786318e-03 1.10871780e+00
-1.54504940e-01 7.06454456e-01 -3.73413742e-01 -7.05088913e-01
2.04239979e-01 -4.97385651e-01 -1.09222126e+00 3.09844613e-01
4.47544485e-01 -5.79928815e-01 3.03086881e-02 9.84302104e-01
6.46562755e-01 -4.31881428e-01 9.14768457e-01 -7.33182371e-01
-4.40866202e-01 2.00854212e-01 -8.17334592e-01 1.19990125e-01
-2.93890480e-03 1.88720345e-01 4.61118162e-01 -3.37316006e-01
5.13290524e-01 6.72661781e-01 9.41467583e-01 7.31058776e-01
-7.49461055e-01 -2.02368155e-01 -3.98653209e-01 1.74997985e-01
-1.25005543e+00 -8.67437646e-02 2.60664582e-01 -4.65350270e-01
6.92991078e-01 4.39359933e-01 6.85240924e-01 5.64105451e-01
2.31249541e-01 9.37446475e-01 1.04622269e+00 -4.80954766e-01
-5.16779721e-03 2.30745673e-01 5.90439253e-02 9.08705711e-01
3.79950970e-01 -1.96949005e-01 1.78011164e-01 2.71835059e-01
9.19838965e-01 1.13604024e-01 9.64037627e-02 1.88750803e-01
-5.34420490e-01 6.71733677e-01 8.53825152e-01 2.92537838e-01
-5.66137612e-01 2.02225879e-01 7.40224600e-01 -7.12215602e-02
4.25171643e-01 1.83687583e-01 -6.07064972e-03 7.51367584e-02
-1.16655588e+00 -2.16165289e-01 6.17291927e-01 2.10288480e-01
1.03614263e-01 -3.62680256e-01 -4.04669166e-01 8.81824255e-01
1.18210271e-01 2.63450325e-01 3.67346734e-01 -6.74159884e-01
-5.47024645e-02 8.07862461e-01 -1.74785644e-01 -5.70252597e-01
-4.72942591e-01 -2.36447722e-01 -8.61030757e-01 2.28344381e-01
4.24946576e-01 -2.57117748e-01 -1.48440075e+00 9.83423591e-01
4.82926309e-01 -3.74029018e-02 -1.82242617e-01 1.04583085e+00
6.97493553e-01 1.55511037e-01 5.21892607e-01 5.58432899e-02
1.25867248e+00 -9.99862492e-01 -7.43515968e-01 1.97170407e-01
9.26518679e-01 -8.74545038e-01 5.04359007e-01 3.47423702e-01
-1.03364062e+00 -2.21603796e-01 -8.14139545e-01 -2.36040041e-01
-3.19436491e-01 5.88518262e-01 7.72341251e-01 8.91431451e-01
-1.08359003e+00 4.23377037e-01 -1.07194269e+00 -8.89371157e-01
1.06692147e+00 4.96185362e-01 -4.05429572e-01 -1.37579441e-01
-6.30748808e-01 9.77378666e-01 5.48369512e-02 3.05024564e-01
-6.92754388e-01 -6.34766400e-01 -5.74500144e-01 -3.48555535e-01
5.06286547e-02 -5.13497651e-01 1.38538861e+00 -1.17549038e+00
-1.33295918e+00 1.21560037e+00 -2.34909102e-01 -5.58379591e-01
6.29474521e-01 2.26829406e-02 -1.45298943e-01 4.78803575e-01
-1.83482647e-01 7.32304513e-01 5.50203562e-01 -9.00996268e-01
-8.30955386e-01 -4.56874669e-01 -1.57798633e-01 3.22450697e-01
-3.53915781e-01 -1.77764356e-01 -2.79372931e-01 -7.86656588e-02
-2.65678108e-01 -6.54901981e-01 -6.09932303e-01 5.76142311e-01
-6.25408471e-01 1.51217717e-03 8.04171443e-01 -1.17556882e+00
1.12045491e+00 -1.82591438e+00 -4.68732893e-01 3.25238258e-01
3.84089619e-01 9.66407657e-01 -3.02399006e-02 4.65167195e-01
2.55760550e-02 4.84691113e-01 -1.66857630e-01 -6.88562319e-02
-4.47389573e-01 -2.65033901e-01 4.41145897e-01 6.61534309e-01
5.44087529e-01 9.61206019e-01 -6.55212522e-01 -7.04092443e-01
5.86211383e-01 7.13686228e-01 -6.42990917e-02 1.78857267e-01
-4.75091562e-02 5.77636398e-02 -3.34556580e-01 9.10030782e-01
7.82107830e-01 -3.49986255e-01 2.97773004e-01 -4.78123218e-01
1.37447089e-01 -4.70439680e-02 -6.80063903e-01 1.51083243e+00
-5.12438297e-01 7.88047254e-01 3.21090490e-01 -7.53466308e-01
5.16963184e-01 4.26115066e-01 6.12176239e-01 -7.26960242e-01
4.52003211e-01 1.70149624e-01 2.65119851e-01 -1.08713222e+00
8.48537609e-02 -6.98708892e-02 5.50261319e-01 1.92576736e-01
-3.96029353e-01 2.13247091e-01 2.38852963e-01 4.98731062e-02
1.04884112e+00 -1.57947123e-01 5.36738575e-01 -5.61098987e-03
5.79084814e-01 3.96543533e-01 -1.32008055e-02 1.40253559e-01
-6.44184649e-01 4.54388976e-01 4.60311651e-01 -3.19428056e-01
-1.00540304e+00 -9.83765006e-01 -3.91580552e-01 5.02889812e-01
1.99800581e-01 1.59553111e-01 -1.23807871e+00 -8.87603223e-01
-3.64725441e-02 1.92334712e-01 -8.95900130e-01 1.38727605e-01
-2.46330172e-01 -7.73038030e-01 6.69591725e-01 7.53793836e-01
7.24010706e-01 -6.76779449e-01 -6.14550650e-01 -3.97214256e-02
1.47093505e-01 -7.56325483e-01 -2.55057752e-01 -3.83615419e-02
-7.35686004e-01 -1.63838816e+00 -9.53506291e-01 -1.00977170e+00
1.04544985e+00 2.40677193e-01 4.48821843e-01 4.08730626e-01
-1.39369857e+00 4.45264317e-02 -2.31957793e-01 -4.59383041e-01
-3.89156520e-01 1.80697277e-01 -6.11866057e-01 -6.61708340e-02
6.51036561e-01 -2.96236202e-02 -1.19949257e+00 -6.23448230e-02
-1.03163433e+00 3.07539832e-02 1.09985101e+00 6.73575342e-01
5.85860908e-01 4.38394472e-02 4.26534295e-01 -1.23669660e+00
8.05918872e-01 -3.44907016e-01 -1.28200531e-01 4.44162637e-01
-3.38962823e-01 -3.39816719e-01 5.31503022e-01 -8.20006803e-02
-1.06687081e+00 3.66469502e-01 -4.33670044e-01 -1.94798820e-02
-3.81942064e-01 4.71620739e-01 4.63380188e-01 -4.03169423e-01
7.68125534e-01 -3.17637101e-02 3.98547322e-01 -1.55152351e-01
1.28045142e-01 1.08694386e+00 2.95696259e-01 3.15080762e-01
2.63580889e-01 5.25932670e-01 2.68419951e-01 -9.76301312e-01
-6.21624410e-01 -6.86166346e-01 -8.43580484e-01 -3.94209087e-01
1.21765912e+00 -6.99550867e-01 -5.91994107e-01 5.54779470e-01
-9.36243117e-01 -2.47910440e-01 8.14764798e-02 2.30215043e-01
5.82048334e-02 5.36760807e-01 -8.34501088e-01 -1.01101303e+00
-7.98445880e-01 -9.83432233e-01 9.70552623e-01 7.24114597e-01
-3.51334482e-01 -1.51917136e+00 -7.04516694e-02 4.17778760e-01
4.71786112e-01 5.19899487e-01 8.63564312e-01 -4.89142895e-01
-1.85357451e-01 -9.67909455e-01 -5.21242142e-01 4.21745807e-01
3.79852921e-01 4.74548250e-01 -9.72425759e-01 -5.99196330e-02
-5.54256678e-01 -2.29644075e-01 9.34533119e-01 8.58976364e-01
1.37798941e+00 3.45840789e-02 -5.60800135e-01 5.51625192e-01
1.85780466e+00 1.41599566e-01 7.14529335e-01 -1.94332525e-01
6.89564466e-01 7.66423941e-01 4.14111048e-01 -4.31432053e-02
2.55174816e-01 3.30922054e-03 5.71114063e-01 -7.70799160e-01
-5.60390592e-01 -1.97627153e-02 -1.08163089e-01 4.56119120e-01
-7.18042329e-02 -2.21690565e-01 -9.05171633e-01 8.94473970e-01
-1.37909412e+00 -7.49678552e-01 -5.35833359e-01 2.11295724e+00
6.65457368e-01 -2.01562345e-01 2.70958722e-01 2.16482803e-01
9.51848149e-01 -4.00307119e-01 -6.58216953e-01 -8.11662853e-01
3.48672092e-01 4.59171861e-01 8.20990205e-01 5.73053896e-01
-1.22993231e+00 7.86584377e-01 6.75367117e+00 8.54966879e-01
-1.52575362e+00 -1.66702226e-01 1.01276505e+00 8.60729888e-02
1.82676911e-01 -5.50749421e-01 -4.79665965e-01 3.02020371e-01
6.60190165e-01 2.40910560e-01 -3.86418700e-02 6.16520464e-01
2.21071616e-01 -6.67501450e-01 -8.73081565e-01 6.59587324e-01
-5.71806468e-02 -1.34031034e+00 -1.85126495e-02 1.75584555e-01
8.31093967e-01 -2.95786709e-01 2.77035654e-01 -4.35265958e-01
6.13604300e-02 -1.44923246e+00 -2.86927432e-01 7.64916301e-01
1.25699902e+00 -7.48367190e-01 1.07633781e+00 7.54597783e-02
-9.51665401e-01 8.39176252e-02 -9.74150095e-03 2.05191135e-01
-1.25772968e-01 5.65769434e-01 -1.52733481e+00 2.33608529e-01
9.53063518e-02 5.01060843e-01 -7.42396235e-01 1.55634069e+00
-6.76196441e-02 7.50889599e-01 -9.48184654e-02 -3.67828995e-01
3.72503161e-01 -8.94871820e-03 1.79966111e-02 1.44166505e+00
-3.22651304e-02 -3.24184820e-02 -6.52876273e-02 5.66442788e-01
7.72103518e-02 2.13018015e-01 -2.02590778e-01 -2.36610368e-01
3.42136532e-01 1.70509636e+00 -1.09244645e+00 -1.28481194e-01
-1.52602479e-01 1.04172683e+00 7.96626657e-02 8.96492973e-02
-6.90881968e-01 -6.09188735e-01 4.80648428e-01 4.73105669e-01
3.01750354e-03 2.66488165e-01 -8.29896271e-01 -4.07681495e-01
-1.38608739e-01 -4.09876049e-01 2.79574811e-01 -1.66342184e-01
-1.36932957e+00 3.47007424e-01 -6.36110663e-01 -8.03380907e-01
-1.04997670e-02 -9.03737545e-01 -9.96516824e-01 9.74695444e-01
-1.60996854e+00 -1.45300496e+00 -6.86241627e-01 2.68050164e-01
4.65333551e-01 1.97424963e-01 8.90229940e-01 2.08732411e-01
-9.74974632e-01 7.72945642e-01 1.47842571e-01 4.28506821e-01
6.79825664e-01 -1.47061193e+00 5.02148606e-02 8.46108735e-01
-5.58710098e-01 2.99904615e-01 1.84549272e-01 -7.47533262e-01
-1.15126610e+00 -1.32081628e+00 5.24262428e-01 -7.95825347e-02
2.89536208e-01 1.25930011e-01 -5.70004284e-01 3.32150981e-02
3.17599505e-01 -1.03187717e-01 1.03363812e+00 -2.08811745e-01
5.28027676e-02 -1.79661170e-01 -1.64912570e+00 6.23714149e-01
3.90660703e-01 -4.69629914e-01 2.75251210e-01 5.97037196e-01
5.69902770e-02 -3.77526224e-01 -9.94563818e-01 2.02040493e-01
6.09347820e-01 -8.99911463e-01 7.13309228e-01 -5.18856823e-01
9.91774201e-01 7.69862682e-02 3.46078604e-01 -1.05658627e+00
-1.70319155e-01 -1.09528296e-01 2.14983806e-01 8.69962335e-01
4.51987952e-01 -2.02627152e-01 1.34443188e+00 6.49787128e-01
6.91686869e-02 -1.36715281e+00 -4.34561849e-01 -1.21333972e-01
2.09684968e-01 -2.54837256e-02 5.81368357e-02 6.31403983e-01
-8.33060667e-02 -1.47563517e-01 3.95189822e-02 3.16145387e-03
6.17886126e-01 -3.01764816e-01 3.11849684e-01 -7.44562864e-01
3.25216264e-01 -6.00625932e-01 -6.19740963e-01 -3.88633937e-01
-4.52919275e-01 -8.93967032e-01 -1.34152129e-01 -2.17895293e+00
3.95657659e-01 -1.31947368e-01 -2.66465843e-02 5.60559332e-01
-4.18276370e-01 5.51735222e-01 -1.94309026e-01 -1.43130615e-01
-4.37164694e-01 -3.11560363e-01 1.61190534e+00 -4.41161133e-02
2.90291514e-02 1.74030033e-03 -6.91149533e-01 4.76524413e-01
1.06813252e+00 4.43841200e-05 -2.82480419e-01 -4.22338843e-01
-1.92333698e-01 1.02785667e-02 5.15379488e-01 -9.88615692e-01
5.12809634e-01 -9.70446020e-02 7.85799682e-01 -4.50010687e-01
1.48913860e-01 -6.85199201e-01 -3.33997369e-01 9.57205117e-01
-5.67124188e-01 -6.02488101e-01 1.61438748e-01 3.94613981e-01
-1.40349045e-01 -3.18067729e-01 1.19239914e+00 -1.48595616e-01
-6.92963600e-01 2.98187852e-01 -5.35311818e-01 -4.41925615e-01
1.57878089e+00 -5.81763923e-01 -2.57007062e-01 -1.17746145e-01
-6.44847274e-01 1.21605575e-01 2.08982259e-01 -2.91911513e-01
5.94530284e-01 -9.49267983e-01 -8.14529777e-01 -2.55189054e-02
-9.48774293e-02 -1.99173018e-02 5.24915338e-01 1.22583878e+00
-1.32852113e+00 4.50480431e-01 -3.77359211e-01 -5.37784934e-01
-1.59063649e+00 7.46380985e-02 7.56744623e-01 -4.00399715e-01
-8.60803723e-02 1.32118976e+00 -2.69971132e-01 -4.32808138e-02
3.38956237e-01 -1.61070764e-01 -2.73480356e-01 -2.16543406e-01
6.00854337e-01 6.08452678e-01 1.47669315e-01 -2.30753630e-01
-1.63771376e-01 5.04974902e-01 -6.07270718e-01 3.00961375e-01
8.89016986e-01 -6.50296733e-02 -1.93522319e-01 -2.38711044e-01
1.29574716e+00 -4.26548809e-01 -1.06071675e+00 1.88617617e-01
-1.78983435e-01 -4.98503506e-01 4.08516884e-01 -1.50189960e+00
-1.16075563e+00 1.32306063e+00 1.00979972e+00 2.56668061e-01
1.31941807e+00 -5.35188258e-01 1.08999562e+00 7.19220936e-02
-2.55675942e-01 -1.21281588e+00 -1.62526637e-01 -1.06587425e-01
2.76521057e-01 -1.69501543e+00 1.88630745e-01 -8.54690552e-01
-7.19480634e-01 1.34264350e+00 8.33689392e-01 -4.06488806e-01
5.65042078e-01 3.85318339e-01 3.81379217e-01 -2.41336539e-01
-4.58572954e-01 -3.93465906e-01 4.33988035e-01 8.24804604e-01
9.02806580e-01 2.67688334e-01 -6.48566067e-01 2.91409373e-01
2.18091041e-01 1.01107597e-01 2.87332326e-01 8.82408500e-01
-5.49208820e-01 -9.21838820e-01 -8.86221603e-02 1.07157409e+00
-8.88924897e-01 1.28322929e-01 -9.32712913e-01 7.51931727e-01
2.22399727e-01 7.37366498e-01 -3.85774747e-02 -2.63578564e-01
-1.57546103e-01 -2.40766436e-01 6.55277312e-01 -5.62398076e-01
-8.29910159e-01 -6.74630404e-02 1.49381474e-01 -3.00479293e-01
-2.76992619e-01 -3.44557285e-01 -1.20256019e+00 -2.62541205e-01
-3.92139435e-01 -2.45503396e-01 1.18624568e+00 8.65695179e-01
6.61688522e-02 6.64305389e-01 6.42011106e-01 -2.98461944e-01
-5.02684236e-01 -8.52222025e-01 -6.33109212e-01 3.05793375e-01
2.89269984e-01 -4.93949018e-02 2.31069818e-01 2.71635234e-01] | [15.64657974243164, -3.0121207237243652] |
d1940ceb-2341-401a-a621-c14f02a61518 | ortho-ode-enhancing-robustness-and-of-neural | 2305.09179 | null | https://arxiv.org/abs/2305.09179v1 | https://arxiv.org/pdf/2305.09179v1.pdf | Ortho-ODE: Enhancing Robustness and of Neural ODEs against Adversarial Attacks | Neural Ordinary Differential Equations (NODEs) probed the usage of numerical solvers to solve the differential equation characterized by a Neural Network (NN), therefore initiating a new paradigm of deep learning models with infinite depth. NODEs were designed to tackle the irregular time series problem. However, NODEs have demonstrated robustness against various noises and adversarial attacks. This paper is about the natural robustness of NODEs and examines the cause behind such surprising behaviour. We show that by controlling the Lipschitz constant of the ODE dynamics the robustness can be significantly improved. We derive our approach from Grownwall's inequality. Further, we draw parallels between contractivity theory and Grownwall's inequality. Experimentally we corroborate the enhanced robustness on numerous datasets - MNIST, CIFAR-10, and CIFAR 100. We also present the impact of adaptive and non-adaptive solvers on the robustness of NODEs. | ['Vishal Purohit'] | 2023-05-16 | null | null | null | null | ['irregular-time-series'] | ['time-series'] | [-2.09067732e-01 4.94100988e-01 4.40619051e-01 1.07959248e-02
-1.30363375e-01 -7.41987050e-01 4.92672801e-01 -4.20698613e-01
-7.05314219e-01 8.15455616e-01 -1.01680897e-01 -2.63216645e-01
-3.52340281e-01 -4.36173618e-01 -8.83739233e-01 -9.52552021e-01
-6.53445423e-01 -1.25436842e-01 9.00651440e-02 -4.49399799e-01
1.71764210e-01 7.62489140e-01 -9.82027888e-01 -2.32319623e-01
5.10949671e-01 1.11862147e+00 -6.09136224e-01 9.37345505e-01
4.00682390e-01 9.39855814e-01 -5.38801491e-01 -2.88460642e-01
7.60727286e-01 -3.10558915e-01 -6.60398781e-01 -4.58994985e-01
2.75151581e-01 -3.74729037e-01 -7.29823589e-01 1.14209592e+00
7.02274382e-01 2.68156528e-01 5.85749209e-01 -1.64101255e+00
-4.06018972e-01 7.71786630e-01 -3.27880681e-01 6.41304612e-01
-2.60886908e-01 3.08450907e-01 7.08059430e-01 -3.94471258e-01
6.27055228e-01 1.03560722e+00 1.44140863e+00 8.35264564e-01
-1.10497546e+00 -5.32318294e-01 7.53982514e-02 -2.79612213e-01
-1.32116771e+00 -3.36605042e-01 5.84343731e-01 -3.33946913e-01
9.66406345e-01 1.25652090e-01 5.09949088e-01 1.34973824e+00
4.55667794e-01 3.39418918e-01 8.82987559e-01 8.90607536e-02
4.74142700e-01 -1.38430849e-01 2.01344222e-01 6.65959597e-01
3.11245263e-01 5.10184467e-01 -2.46709228e-01 -4.20297205e-01
1.04212463e+00 -3.75732511e-01 -2.73313731e-01 7.27298930e-02
-8.32000136e-01 9.84315038e-01 3.28744620e-01 3.47046912e-01
-2.13239446e-01 6.15056694e-01 7.49402463e-01 6.47920966e-01
2.42544264e-01 5.94871581e-01 -7.36350834e-01 -7.02666044e-02
-7.45157003e-01 5.58349252e-01 1.21662867e+00 8.05370629e-01
2.05605805e-01 5.08942544e-01 3.52082342e-01 1.86729699e-01
2.62023389e-01 2.55274832e-01 5.00939906e-01 -1.52394140e+00
-9.48377885e-03 4.39095832e-02 -1.74568653e-01 -1.10560203e+00
-7.35184133e-01 -8.39678288e-01 -1.27466559e+00 4.80370402e-01
7.33561754e-01 -8.32349598e-01 -5.15386939e-01 2.15031242e+00
2.90637195e-01 4.20206398e-01 2.71493793e-01 7.59409606e-01
3.42845291e-01 6.06674969e-01 -2.38277212e-01 -1.73861399e-01
6.25220716e-01 -4.48659569e-01 -6.13867998e-01 4.78674084e-01
8.45763803e-01 -1.00776888e-01 5.63657284e-01 4.59360093e-01
-1.24900031e+00 -1.65600941e-01 -1.10446489e+00 9.50519592e-02
-2.51026869e-01 -6.38243854e-01 4.44580615e-01 5.38941741e-01
-1.35554945e+00 1.35543168e+00 -1.02013743e+00 2.88589410e-02
5.54299116e-01 6.26994610e-01 -3.12826365e-01 8.46976876e-01
-1.37016225e+00 8.70231092e-01 -4.11052816e-03 1.66075677e-01
-7.60114491e-01 -1.16008449e+00 -4.49421227e-01 1.63354054e-02
-3.79201807e-02 -5.56831777e-01 1.03612435e+00 -1.04203510e+00
-1.57343602e+00 5.94967306e-01 3.80894661e-01 -1.15197849e+00
1.09398341e+00 -6.75100014e-02 -1.60341904e-01 1.90944716e-01
-3.52941543e-01 3.54085535e-01 8.88399422e-01 -9.74470913e-01
-6.39108475e-03 -1.38293281e-01 -2.57460494e-02 -3.47098768e-01
-3.63928944e-01 -1.21508546e-01 4.25977647e-01 -8.43484223e-01
-2.13452894e-02 -1.19806635e+00 -4.53470170e-01 3.59239250e-01
-3.22715044e-01 -1.33885697e-01 8.43248487e-01 -4.79934663e-01
8.09498727e-01 -2.16623497e+00 -1.01490347e-02 4.37269002e-01
4.48251933e-01 3.75885457e-01 -1.64286181e-01 3.70976269e-01
-2.92167962e-01 4.33473706e-01 -2.97721028e-01 -1.85897171e-01
1.55516058e-01 3.10037315e-01 -6.07996404e-01 1.14308691e+00
5.81393614e-02 7.78281987e-01 -4.24374849e-01 -2.36103415e-01
-4.65946108e-01 7.74857163e-01 -8.24423492e-01 -2.20189199e-01
1.86689384e-02 4.59672600e-01 -3.26570511e-01 4.98126782e-02
6.25783980e-01 -6.82280064e-02 -3.01619768e-01 3.95907685e-02
-1.23182230e-01 -2.00990304e-01 -1.24516606e+00 1.13306451e+00
-1.24238685e-01 9.30177987e-01 4.10489708e-01 -1.26032948e+00
5.92839897e-01 5.38740158e-01 6.73550963e-01 -5.54872692e-01
6.05856299e-01 2.91554123e-01 2.21443832e-01 -5.42208374e-01
-7.95546100e-02 -2.06848532e-01 1.68116674e-01 5.62129378e-01
-5.82838021e-02 -5.11080250e-02 3.96897160e-02 3.09300572e-01
1.32975781e+00 -3.32583440e-03 -2.06252992e-01 -8.51221740e-01
4.19227719e-01 -6.29396886e-02 5.55160820e-01 7.55855322e-01
-5.04944146e-01 4.30889040e-01 1.08275473e+00 -6.45481884e-01
-1.36121786e+00 -8.51786196e-01 -3.81403416e-01 7.54842639e-01
-2.40199953e-01 -1.99341699e-02 -1.05593026e+00 -1.26634046e-01
1.88425303e-01 3.05391431e-01 -1.08032131e+00 -4.12159085e-01
-9.16569531e-01 -7.04543948e-01 1.47737300e+00 5.89501321e-01
6.30266309e-01 -7.07935035e-01 -8.34978998e-01 1.52877765e-02
1.69292718e-01 -1.23161161e+00 -2.37042874e-01 2.12631315e-01
-1.06850350e+00 -8.84116352e-01 -6.90170646e-01 -4.47148979e-01
4.38989937e-01 -5.14109492e-01 8.76760960e-01 1.96869642e-01
-2.42747366e-01 4.57490325e-01 1.48598418e-01 -4.80270386e-01
-6.36659980e-01 6.71427026e-02 6.69606626e-01 -1.49140179e-01
-4.48887169e-01 -1.05267382e+00 -4.07868832e-01 2.36357853e-01
-1.06447494e+00 -5.72170198e-01 8.80396739e-02 4.86555636e-01
1.94027752e-01 2.26800025e-01 7.62544930e-01 -6.32846713e-01
7.31639266e-01 -6.40725911e-01 -8.37034404e-01 -2.20622584e-01
-6.22193873e-01 3.96272808e-01 1.05277193e+00 -9.60348547e-01
-6.85920298e-01 -6.56281784e-02 -2.46163249e-01 -5.52692354e-01
2.98489898e-01 3.11565220e-01 3.42791945e-01 -7.44568229e-01
9.05504048e-01 -1.13345914e-01 1.74295157e-01 -1.58966824e-01
7.07847625e-02 9.68034286e-03 6.88405335e-01 -5.95278919e-01
1.00104523e+00 7.34530687e-01 5.53145587e-01 -9.51580882e-01
-3.33035976e-01 1.91856742e-01 -4.26622272e-01 -1.62708104e-01
4.95095164e-01 -6.23947263e-01 -1.16334093e+00 7.69895911e-01
-1.19761205e+00 -7.65262306e-01 -4.90407825e-01 2.67643809e-01
-6.85712636e-01 2.04896078e-01 -9.97059941e-01 -9.16835725e-01
-2.89018184e-01 -6.90650761e-01 2.49028280e-01 3.99498433e-01
-3.12954247e-01 -1.33721888e+00 4.18109030e-01 -3.75138551e-01
7.93394744e-01 8.40051174e-01 5.18414021e-01 -7.76881576e-01
-1.34236082e-01 -2.00862572e-01 -7.58789107e-02 3.79748434e-01
-7.29481995e-01 4.19972837e-01 -1.14378202e+00 -3.44106555e-01
4.74203706e-01 -2.50162631e-01 6.55051053e-01 4.07421142e-01
9.16932046e-01 -7.15008199e-01 1.37636855e-01 1.20583236e+00
1.60284877e+00 -2.02682689e-01 6.30554974e-01 3.87807906e-01
4.61075962e-01 5.52487075e-01 -3.04691017e-01 5.34117043e-01
-1.78501070e-01 3.26496288e-02 7.18263447e-01 8.21110830e-02
3.91326815e-01 2.01227546e-01 4.74056542e-01 4.73723263e-01
-4.68349189e-01 1.01283394e-01 -8.90676737e-01 4.14915800e-01
-1.53754485e+00 -9.31873918e-01 -3.30575407e-01 1.61611784e+00
9.01868999e-01 3.12015176e-01 2.76702911e-01 3.97203982e-01
3.56319785e-01 -9.25327167e-02 -8.66087794e-01 -6.56732619e-01
-4.85259175e-01 1.08621553e-01 9.53798294e-01 5.04712701e-01
-8.39282393e-01 4.46792781e-01 7.11320782e+00 3.88381958e-01
-1.23019695e+00 -1.18511706e-03 5.55055618e-01 -2.94307739e-01
7.68105313e-02 -4.74195033e-01 -4.93486434e-01 2.86216080e-01
1.22217548e+00 -4.68796998e-01 4.42027032e-01 6.91500187e-01
2.71223664e-01 2.39898413e-01 -9.82244551e-01 7.31708229e-01
-2.58498430e-01 -1.30265296e+00 -5.72500229e-01 2.87241608e-01
9.23317909e-01 3.77255976e-01 3.28015417e-01 1.50114000e-01
4.61112678e-01 -1.19306660e+00 6.02158189e-01 6.08100533e-01
3.09603184e-01 -7.12252319e-01 5.68654776e-01 4.27609771e-01
-8.69406343e-01 -1.81663901e-01 -4.00058627e-01 -2.65470386e-01
-7.58476779e-02 5.39340556e-01 -1.38415024e-01 4.37204465e-02
7.38483012e-01 4.85503674e-01 -3.35411459e-01 8.05188119e-01
2.75704920e-01 6.05224013e-01 -8.09157431e-01 8.25046152e-02
5.34176767e-01 -1.05872341e-01 7.52286196e-01 1.09346449e+00
6.41933158e-02 3.42126489e-01 -5.49939990e-01 1.08380365e+00
-1.60431340e-02 -3.37035477e-01 -5.71649790e-01 -6.30662590e-02
2.13365722e-02 9.85742569e-01 -7.32583344e-01 -8.20338503e-02
-1.06059872e-01 7.86215961e-01 2.81703174e-01 7.02001810e-01
-1.15242791e+00 -4.53448206e-01 9.72618520e-01 -2.73240078e-02
2.08522320e-01 -5.22979677e-01 -5.37573695e-01 -9.84947562e-01
1.96871776e-02 -7.74287283e-01 1.59042165e-01 -2.74832398e-01
-1.15795636e+00 4.68089849e-01 -8.97077024e-02 -7.96908498e-01
-1.54755756e-01 -6.88394308e-01 -9.00089860e-01 6.80437922e-01
-1.13570654e+00 -3.27887923e-01 2.09106117e-01 8.54056716e-01
-1.55768111e-01 7.48841465e-02 5.65550864e-01 2.22736180e-01
-7.11875439e-01 7.68282294e-01 3.63428891e-01 4.18972194e-01
1.74598381e-01 -1.06664920e+00 6.83932662e-01 8.75521421e-01
-3.36162776e-01 6.39308870e-01 1.15485394e+00 -3.78057510e-01
-1.35557306e+00 -8.93369913e-01 5.45981586e-01 -2.99735278e-01
1.09068048e+00 -1.98145896e-01 -1.17128456e+00 6.60066187e-01
1.27583027e-01 4.38836753e-01 2.37680152e-01 -5.93735456e-01
-4.90686357e-01 1.14441633e-01 -1.41605830e+00 5.35167754e-01
1.15031087e+00 -3.75426888e-01 -2.69365013e-01 1.77407071e-01
7.06202984e-01 -4.54162300e-01 -9.83146071e-01 3.25255722e-01
6.49265349e-01 -9.81067538e-01 9.92776811e-01 -9.15532053e-01
4.69546229e-01 2.51500160e-01 8.93999711e-02 -1.08337355e+00
-2.48004645e-02 -1.33340561e+00 -5.29136300e-01 7.75293350e-01
3.01380724e-01 -1.13601911e+00 6.87719345e-01 9.60672140e-01
3.53275947e-02 -9.27821457e-01 -1.11869550e+00 -1.17254972e+00
7.47792363e-01 -2.65575528e-01 1.35828897e-01 1.17619979e+00
-3.02974377e-02 -3.71260822e-01 -8.65515247e-02 2.75240481e-01
7.53416479e-01 -7.25275636e-01 3.52687210e-01 -1.28110564e+00
-3.09224516e-01 -6.59190059e-01 -5.67731559e-01 -6.04138494e-01
3.85257691e-01 -6.14373744e-01 -2.33599514e-01 -7.44856477e-01
-5.17350972e-01 -4.33858812e-01 -1.47865564e-01 2.28445068e-01
3.27580899e-01 3.18515122e-01 7.35685825e-02 1.24890238e-01
-2.13447675e-01 2.12246969e-01 1.07018006e+00 2.63922781e-01
-2.43003011e-01 3.24283540e-02 -4.57900822e-01 9.66283798e-01
1.16048265e+00 -5.18141329e-01 -3.59133631e-01 -4.82076317e-01
8.19189250e-01 -1.64677605e-01 6.31641567e-01 -1.22584677e+00
6.34429336e-01 1.10202551e-01 1.47559062e-01 -4.17258777e-02
-1.81714986e-02 -7.93286026e-01 3.34792808e-02 9.88539755e-01
-7.12926328e-01 3.76844823e-01 4.33059782e-01 4.97636259e-01
2.63742864e-01 -1.62302971e-01 1.21713006e+00 6.52241632e-02
-5.39213046e-02 4.15581405e-01 -6.07476056e-01 6.43725276e-01
9.06754315e-01 6.86494634e-02 -2.96949506e-01 -4.34159726e-01
-9.31592107e-01 1.32043615e-01 1.94392547e-01 -7.09174499e-02
2.95881867e-01 -1.08633029e+00 -7.03457475e-01 3.23460549e-01
-8.50514233e-01 -2.51229964e-02 9.74537730e-02 1.15884018e+00
-8.31883729e-01 9.78009552e-02 -2.03718692e-01 -3.15772802e-01
-8.88208926e-01 4.86137867e-01 1.12576771e+00 8.06970173e-04
-7.53947139e-01 1.12546575e+00 -2.16991425e-01 -1.51263013e-01
5.77460229e-01 -4.61055726e-01 8.84512514e-02 -8.59448686e-02
4.08736885e-01 7.68241107e-01 -1.16106823e-01 -4.83177453e-01
-2.87206233e-01 5.60007393e-01 1.58543512e-01 -3.97471368e-01
1.53320491e+00 8.86923913e-03 -1.06169075e-01 2.82560140e-01
1.59965897e+00 -2.47292742e-01 -1.64559090e+00 -1.18771039e-01
-9.97740850e-02 3.17921221e-01 -6.78960755e-02 -2.35611185e-01
-1.38771546e+00 7.74561584e-01 4.28068221e-01 7.06974864e-01
8.51121843e-01 -3.72855276e-01 9.31609750e-01 6.78813994e-01
-7.87501261e-02 -1.22130370e+00 -1.89531833e-01 8.66126597e-01
8.16397309e-01 -8.27395916e-01 -7.87048638e-02 -7.01293498e-02
-2.33410031e-01 1.19480729e+00 3.25767398e-01 -9.73228216e-01
1.06313014e+00 9.57362115e-01 6.36814237e-02 -8.48798361e-03
-7.43531346e-01 3.70130569e-01 -1.70181364e-01 5.16060948e-01
1.46540105e-01 -5.72272122e-01 -5.56139573e-02 6.09005094e-01
-4.37466234e-01 8.14366899e-03 6.92310035e-01 7.27823436e-01
-2.09270701e-01 -3.40279073e-01 -3.58231992e-01 3.19490656e-02
-9.31206524e-01 3.13817784e-02 -4.78682518e-01 9.67613637e-01
-1.68747723e-01 5.04558980e-01 -7.54531845e-02 -1.59602463e-01
9.16870236e-02 8.04332718e-02 3.76606286e-01 2.93658018e-01
-9.11310852e-01 -4.23442870e-01 -3.40219587e-01 -6.12058759e-01
-2.86593556e-01 -5.54541051e-01 -1.65558493e+00 -8.88858676e-01
-6.74717948e-02 5.53478226e-02 7.40548313e-01 1.08973062e+00
1.65688127e-01 5.07761717e-01 5.83672822e-01 -7.85372674e-01
-1.06193948e+00 -5.45238435e-01 -5.90736210e-01 2.18565270e-01
8.99949014e-01 -1.86538339e-01 -1.18463433e+00 3.63046117e-02] | [6.885190963745117, 3.557990789413452] |
8e572cd8-8e58-41a7-9c56-abfc5cfe831a | analyzing-vietnamese-legal-questions-using-1 | 2304.14447 | null | https://arxiv.org/abs/2304.14447v1 | https://arxiv.org/pdf/2304.14447v1.pdf | Analyzing Vietnamese Legal Questions Using Deep Neural Networks with Biaffine Classifiers | In this paper, we propose using deep neural networks to extract important information from Vietnamese legal questions, a fundamental task towards building a question answering system in the legal domain. Given a legal question in natural language, the goal is to extract all the segments that contain the needed information to answer the question. We introduce a deep model that solves the task in three stages. First, our model leverages recent advanced autoencoding language models to produce contextual word embeddings, which are then combined with character-level and POS-tag information to form word representations. Next, bidirectional long short-term memory networks are employed to capture the relations among words and generate sentence-level representations. At the third stage, borrowing ideas from graph-based dependency parsing methods which provide a global view on the input sentence, we use biaffine classifiers to estimate the probability of each pair of start-end words to be an important segment. Experimental results on a public Vietnamese legal dataset show that our model outperforms the previous work by a large margin, achieving 94.79% in the F1 score. The results also prove the effectiveness of using contextual features extracted from pre-trained language models combined with other types of features such as character-level and POS-tag features when training on a limited dataset. | ['Ngo Xuan Bach', 'Tu Minh Phuong', 'Hoang Thi Thu Uyen', 'Nguyen Anh Tu'] | 2023-04-27 | analyzing-vietnamese-legal-questions-using | https://link.springer.com/chapter/10.1007/978-3-030-92270-2_44 | https://link.springer.com/chapter/10.1007/978-3-030-92270-2_44 | iconip-2021-12 | ['dependency-parsing'] | ['natural-language-processing'] | [ 7.03457296e-02 3.47724736e-01 -5.09623706e-01 -6.81320012e-01
-1.25554681e+00 -5.31020582e-01 4.43974316e-01 3.88940752e-01
-8.14428210e-01 5.42591810e-01 8.41847122e-01 -6.65359855e-01
9.21365172e-02 -9.82616663e-01 -6.54066205e-01 -2.95842707e-01
6.29578009e-02 3.23475182e-01 2.24182755e-01 -5.46531737e-01
3.77360225e-01 4.20374513e-01 -9.39624786e-01 3.70697439e-01
1.13988626e+00 7.77031660e-01 4.41921800e-02 5.58042288e-01
-5.26912987e-01 1.19144917e+00 -6.77962542e-01 -9.66574669e-01
3.77288759e-02 -4.45157826e-01 -1.03103173e+00 -3.23295593e-01
3.35118115e-01 -5.24610877e-01 -7.09236324e-01 8.40114892e-01
2.12276548e-01 7.98926502e-02 4.59766358e-01 -5.53923368e-01
-7.03523815e-01 4.78058457e-01 -2.48605669e-01 5.99244237e-01
3.34988952e-01 1.57867178e-01 1.62938035e+00 -7.63208985e-01
9.60037887e-01 1.18977857e+00 2.55400479e-01 6.01073146e-01
-7.77715206e-01 -2.63234437e-01 2.89530575e-01 5.49829721e-01
-8.20956051e-01 -2.34356225e-01 8.92678142e-01 -3.12982351e-01
1.18480217e+00 -2.40547404e-01 3.04403722e-01 9.98468220e-01
7.37333298e-01 8.65439892e-01 6.09317422e-01 -4.18216884e-01
7.48179108e-02 -1.95048064e-01 7.82281101e-01 9.49671388e-01
4.92892936e-02 -1.77184120e-01 -1.86095327e-01 -2.65610993e-01
2.44947106e-01 -1.37824476e-01 1.32493317e-01 1.42086715e-01
-6.89007580e-01 1.53022861e+00 4.74758267e-01 5.47555327e-01
-5.30819535e-01 1.08108014e-01 3.88295114e-01 2.89374352e-01
3.13456237e-01 3.75602186e-01 -3.71867061e-01 -2.31086060e-01
-4.30156082e-01 3.49787951e-01 8.56634319e-01 7.33026922e-01
7.79812574e-01 -4.31199312e-01 -5.51218033e-01 7.99288452e-01
2.55210847e-01 2.64715880e-01 2.68492579e-01 -9.17369545e-01
1.07633162e+00 8.34412694e-01 -3.52203637e-01 -8.57555509e-01
-2.49867797e-01 -3.23636144e-01 -3.04088384e-01 -2.10745901e-01
4.72410738e-01 -3.44029367e-01 -1.09464681e+00 1.68588114e+00
1.91287518e-01 -3.79613727e-01 2.47344404e-01 6.80084109e-01
7.27333426e-01 8.57277095e-01 2.64083922e-01 2.85099328e-01
1.61716533e+00 -9.76545990e-01 -7.17109919e-01 -5.44689059e-01
8.24904740e-01 -5.00022948e-01 1.01697159e+00 -9.52395797e-02
-9.54561472e-01 -5.42131543e-01 -9.02703881e-01 -7.71031082e-01
-4.48135525e-01 -1.41873628e-01 3.86279672e-01 3.86025459e-01
-6.99597120e-01 2.72238493e-01 -5.67438960e-01 -1.64258890e-02
5.33924222e-01 -2.22966224e-02 -2.96091318e-01 -4.05588865e-01
-1.61152053e+00 1.01318336e+00 2.48346329e-01 1.36692286e-01
-4.97246325e-01 -3.84478837e-01 -1.44014001e+00 3.01586688e-01
4.25420284e-01 -4.07375097e-01 1.06233835e+00 -5.07387757e-01
-1.27497339e+00 6.81379080e-01 -6.83001697e-01 -5.88072956e-01
2.12400779e-02 -2.69842535e-01 -3.65922004e-01 5.67391992e-01
4.72313225e-01 5.55218756e-01 5.29020667e-01 -7.84863234e-01
-6.75597548e-01 -3.61383379e-01 5.29222786e-01 -1.81499720e-01
-2.49098390e-01 2.71769345e-01 -7.44690359e-01 -4.78707075e-01
-2.29484349e-01 -6.61118150e-01 -4.45226312e-01 -3.76757622e-01
-2.25637496e-01 -7.05757737e-01 5.32657206e-01 -1.22181499e+00
1.26693475e+00 -1.87903082e+00 1.22194868e-02 3.22867632e-01
2.29792595e-01 3.50572705e-01 -4.48557079e-01 8.14644754e-01
9.70373899e-02 3.40672657e-02 -4.92215335e-01 -2.17949346e-01
2.22422034e-02 5.91484666e-01 -4.55672771e-01 2.10824922e-01
8.37833524e-01 1.25502551e+00 -9.18390870e-01 -7.32697964e-01
-1.44989371e-01 3.96351337e-01 -7.03814626e-01 2.46176988e-01
-3.89495969e-01 2.28862837e-01 -6.73699856e-01 4.23260808e-01
3.62093836e-01 -5.40869543e-03 2.09996551e-01 2.77456105e-01
2.48951346e-01 9.16334510e-01 -3.90355408e-01 1.70390403e+00
-7.21754491e-01 6.61963880e-01 -1.67293534e-01 -1.16448891e+00
9.43710327e-01 1.50247157e-01 -1.99871231e-03 -1.29108620e+00
2.24745333e-01 1.90683842e-01 1.68875724e-01 -9.61068511e-01
5.09114027e-01 -1.02407746e-01 -6.57765150e-01 4.11008477e-01
1.88778371e-01 1.74631670e-01 5.85359275e-01 4.32249755e-01
1.30578566e+00 -8.78793672e-02 1.42329141e-01 1.39537957e-02
7.45389104e-01 -3.08059677e-02 8.16844940e-01 5.88050723e-01
-1.91104561e-01 4.91738051e-01 1.14352393e+00 -4.32188749e-01
-7.96939373e-01 -8.29164922e-01 1.90604702e-02 1.01734924e+00
-7.69583955e-02 -5.10734439e-01 -7.98114777e-01 -1.32421017e+00
-7.84999803e-02 9.96678054e-01 -8.43618274e-01 -9.58774388e-02
-1.27101374e+00 -2.42056638e-01 5.19949913e-01 8.12809348e-01
2.05053598e-01 -1.22714388e+00 -5.64662337e-01 2.84890801e-01
-3.24002326e-01 -1.34224391e+00 -5.33342600e-01 -1.02928087e-01
-5.86239815e-01 -1.18828702e+00 -5.96738994e-01 -1.05704236e+00
6.29739881e-01 -3.17660630e-01 1.02173781e+00 4.01173413e-01
-5.72175793e-02 -7.28667974e-02 -4.56250876e-01 4.45684232e-03
-3.25316250e-01 3.50968421e-01 -6.60789073e-01 1.54372901e-01
6.67800486e-01 -3.15151453e-01 -5.15091181e-01 -2.46669605e-01
-1.11774969e+00 -4.04248059e-01 7.72561371e-01 7.28229344e-01
3.23776513e-01 -5.08673906e-01 6.72279119e-01 -1.10680616e+00
1.01789784e+00 -7.28150785e-01 -3.72065723e-01 3.85803878e-01
-3.13232362e-01 4.81822878e-01 8.40359688e-01 1.53850103e-02
-1.10334301e+00 -4.24590230e-01 -8.39577615e-01 2.54048228e-01
-1.42928466e-01 9.66305494e-01 -3.26636821e-01 6.02281868e-01
3.96430790e-01 1.87450439e-01 -1.70588315e-01 -3.81731182e-01
5.18827021e-01 8.10027838e-01 4.99740839e-01 -5.45934916e-01
6.73389912e-01 2.98920065e-01 -1.41606241e-01 -7.43646502e-01
-1.08242369e+00 -4.44586873e-01 -8.73061657e-01 1.96847335e-01
1.28452897e+00 -5.41705728e-01 -4.67962265e-01 -9.11192503e-03
-1.39753556e+00 -1.94616467e-01 -2.89749086e-01 2.54509509e-01
-1.72004580e-01 6.64263487e-01 -9.07140434e-01 -5.61280847e-01
-3.80309522e-01 -1.12657070e+00 1.10329521e+00 3.67976278e-01
-8.05430040e-02 -1.05994678e+00 3.86431068e-03 9.38701153e-01
2.73115993e-01 1.50013819e-01 1.52973795e+00 -9.97771263e-01
-4.14051235e-01 -5.33586264e-01 -3.07418466e-01 5.45574725e-01
-2.30034441e-01 -3.78040224e-01 -4.62698042e-01 1.09614156e-01
1.08761732e-02 -3.19416434e-01 1.20150161e+00 2.29447186e-02
8.66126776e-01 -2.77464926e-01 -2.88280062e-02 2.70348489e-01
1.13322186e+00 1.87274918e-01 7.99101233e-01 2.59814799e-01
6.28717601e-01 8.15743208e-01 5.35262525e-01 -7.80168548e-02
1.09470880e+00 3.45519453e-01 2.28127956e-01 9.48355049e-02
-2.10744083e-01 -6.17677748e-01 3.99022579e-01 8.50632429e-01
2.76721120e-01 -1.38352558e-01 -1.11519980e+00 8.25771213e-01
-1.71579504e+00 -8.97957087e-01 -2.09893256e-01 1.75378096e+00
7.13441074e-01 3.97736788e-01 -8.76408815e-02 -1.24952286e-01
5.26759446e-01 4.33009684e-01 -4.69682217e-01 -7.71825969e-01
6.90176785e-02 6.32288814e-01 3.05656135e-01 7.62565076e-01
-8.93370748e-01 1.31611395e+00 5.30697727e+00 6.47247910e-01
-8.35757792e-01 1.95935130e-01 6.05514824e-01 3.08335215e-01
-6.81527078e-01 2.64892757e-01 -8.48272920e-01 3.68614197e-01
1.32687235e+00 3.87652442e-02 -1.22838840e-01 5.10000825e-01
7.37047493e-02 -1.39210209e-01 -8.39932203e-01 3.95259678e-01
4.33979928e-01 -1.40174580e+00 9.66525972e-02 3.22323501e-01
4.90254581e-01 7.14950487e-02 -3.13162863e-01 6.76796138e-01
2.41317734e-01 -1.03104389e+00 4.98436093e-01 5.10122716e-01
4.11196232e-01 -9.70693707e-01 9.72359478e-01 5.65602899e-01
-9.23140585e-01 -2.98992932e-01 -3.94460618e-01 -1.70135856e-01
6.91994905e-01 4.01372135e-01 -5.76299071e-01 5.99048674e-01
1.79669976e-01 6.28311396e-01 -5.43203831e-01 5.88763177e-01
-1.04931128e+00 8.22563112e-01 7.42833093e-02 -1.84767261e-01
6.70082331e-01 -2.37927184e-01 2.03345209e-01 1.27911162e+00
-8.49396437e-02 1.75732195e-01 -6.92109764e-02 7.61368275e-01
-4.23666358e-01 1.91722393e-01 -3.87239128e-01 -7.37167895e-04
1.94118947e-01 1.08456981e+00 -3.78085852e-01 -5.68489254e-01
-6.77902699e-01 9.58406210e-01 7.76807070e-01 4.20750886e-01
-8.05829942e-01 -9.18581009e-01 5.21972239e-01 1.44171426e-02
5.43133736e-01 -4.93262887e-01 -4.90601435e-02 -1.27508163e+00
3.33589911e-01 -7.42748141e-01 6.82086110e-01 -3.85531813e-01
-1.15423322e+00 8.55897963e-01 -1.86021924e-01 -7.06953526e-01
-5.47677815e-01 -6.51017249e-01 -8.45366359e-01 1.07015872e+00
-2.01443672e+00 -1.27222252e+00 2.29709744e-01 4.17211145e-01
7.60595202e-01 -1.06501102e-01 6.14206791e-01 2.32970282e-01
-6.34658813e-01 5.67080259e-01 -3.20965111e-01 9.38503265e-01
3.24804515e-01 -1.10421634e+00 8.15944195e-01 1.13315535e+00
2.97039241e-01 8.49391937e-01 1.40916288e-01 -7.33150661e-01
-1.35092473e+00 -9.57810462e-01 1.88662386e+00 -4.39906120e-01
7.28992343e-01 -5.73802948e-01 -9.69650626e-01 7.30580628e-01
3.69674921e-01 1.09766452e-02 8.26334596e-01 1.37458473e-01
-5.47862947e-01 8.09805542e-02 -9.06120479e-01 4.10268843e-01
9.25313711e-01 -8.42046499e-01 -1.39661074e+00 1.87229842e-01
7.80652821e-01 -2.12372750e-01 -8.43918562e-01 1.30895853e-01
2.79444695e-01 -8.19303274e-01 6.37159944e-01 -1.00377750e+00
8.72669160e-01 2.20628634e-01 -1.03280440e-01 -9.19755995e-01
-9.88916531e-02 -3.46339047e-01 -1.04704536e-02 1.11411083e+00
9.50794876e-01 -7.35980153e-01 6.46153748e-01 6.52300596e-01
-2.54627496e-01 -9.54154491e-01 -1.12635589e+00 -6.18657231e-01
5.81228077e-01 -4.84742880e-01 4.64328110e-01 5.24985969e-01
1.89750880e-01 8.24748874e-01 -6.30617961e-02 7.97950625e-02
3.97074878e-01 3.79577219e-01 5.29544473e-01 -1.07135570e+00
-3.86682376e-02 -2.26666480e-01 -2.41246283e-01 -1.40978324e+00
7.88858652e-01 -9.75528181e-01 6.44675642e-02 -1.99917901e+00
2.08272524e-02 -9.67514366e-02 -1.55945659e-01 4.33680326e-01
-5.45823693e-01 -1.69697016e-01 1.72826290e-01 -4.91199270e-02
-4.64365512e-01 6.15909696e-01 1.05245948e+00 -1.91880092e-01
4.41903509e-02 -1.82479858e-01 -9.20785189e-01 5.52881598e-01
7.84888744e-01 -4.73677278e-01 -1.20399714e-01 -8.36676419e-01
1.68346345e-01 2.28840157e-01 5.45438863e-02 -5.49494505e-01
2.07643464e-01 -8.43005031e-02 4.20520157e-02 -4.48296815e-01
3.28165919e-01 -6.01954460e-01 -9.81300652e-01 6.00770772e-01
-6.81655109e-01 1.33011386e-01 -1.11277342e-01 5.01881063e-01
-5.52128196e-01 -5.24657130e-01 3.31391931e-01 4.47821803e-02
-7.21937537e-01 2.87110329e-01 -4.36754942e-01 4.69927341e-01
9.41942811e-01 1.29724815e-01 -3.98871303e-01 -5.41791677e-01
-4.45322871e-01 4.70015764e-01 -9.98206884e-02 7.34931827e-01
7.23074794e-01 -1.02848840e+00 -8.90216470e-01 3.42834413e-01
2.90861309e-01 -2.83938855e-01 1.02547094e-01 4.83111739e-01
-5.72616041e-01 7.12828875e-01 2.16349632e-01 -3.89825553e-02
-9.89085913e-01 4.42414403e-01 9.47226398e-03 -5.90831757e-01
-6.43278241e-01 7.76663303e-01 -2.42656350e-01 -5.04931331e-01
-4.79142442e-02 -5.75861692e-01 -5.50334811e-01 2.06497192e-01
5.47522008e-01 -1.56980798e-01 8.67157150e-03 -9.84663665e-01
-3.38812768e-01 5.53181887e-01 -3.08989942e-01 -2.03814417e-01
1.60376620e+00 1.55547410e-01 -1.68345541e-01 -4.92364056e-02
1.53482592e+00 3.56518328e-01 -1.01536679e+00 -4.27839339e-01
4.74251717e-01 -3.24294418e-01 -8.77710730e-02 -5.50088525e-01
-1.10860372e+00 1.35557866e+00 -7.21436292e-02 7.25920871e-02
6.66549027e-01 4.08215940e-01 1.52280712e+00 4.10190046e-01
4.12191041e-02 -1.15442955e+00 -3.08592618e-02 9.21704054e-01
7.76377439e-01 -1.11422634e+00 -5.54606855e-01 -2.58793682e-01
-6.65244520e-01 1.24190319e+00 4.55553502e-01 -2.76216328e-01
4.60325956e-01 -2.32999865e-03 3.05259585e-01 -3.30192387e-01
-7.13448584e-01 -5.13841152e-01 3.13412905e-01 2.61080265e-01
3.33935022e-01 -3.31717610e-01 -7.59379566e-01 5.81165791e-01
-2.56368816e-01 -1.35384202e-01 4.82767314e-01 9.54796076e-01
-5.78521371e-01 -1.48404884e+00 1.20510988e-01 5.28502584e-01
-7.34835684e-01 -2.74517894e-01 -2.98178792e-01 6.00772023e-01
8.09010863e-02 1.28931558e+00 1.43778995e-01 -2.09850848e-01
6.24294162e-01 1.73521042e-01 8.60723257e-02 -8.36474597e-01
-6.40263736e-01 -3.84239554e-01 4.61972892e-01 -5.66782475e-01
-1.62643604e-02 -6.06370330e-01 -1.58124542e+00 1.09920904e-01
9.67415273e-02 4.48583424e-01 6.22223794e-01 1.40396750e+00
4.64139462e-01 4.05326277e-01 3.80964816e-01 -1.19210027e-01
-6.23250365e-01 -8.65262508e-01 -2.42809709e-02 5.80538154e-01
4.54289734e-01 -1.50209755e-01 -1.38863847e-01 -1.94119632e-01] | [11.054187774658203, 8.047601699829102] |
d06719af-a145-44cf-8710-deab01f6e9aa | weakly-supervised-segmentation-with-point | 2304.03572 | null | https://arxiv.org/abs/2304.03572v1 | https://arxiv.org/pdf/2304.03572v1.pdf | Weakly supervised segmentation with point annotations for histopathology images via contrast-based variational model | Image segmentation is a fundamental task in the field of imaging and vision. Supervised deep learning for segmentation has achieved unparalleled success when sufficient training data with annotated labels are available. However, annotation is known to be expensive to obtain, especially for histopathology images where the target regions are usually with high morphology variations and irregular shapes. Thus, weakly supervised learning with sparse annotations of points is promising to reduce the annotation workload. In this work, we propose a contrast-based variational model to generate segmentation results, which serve as reliable complementary supervision to train a deep segmentation model for histopathology images. The proposed method considers the common characteristics of target regions in histopathology images and can be trained in an end-to-end manner. It can generate more regionally consistent and smoother boundary segmentation, and is more robust to unlabeled `novel' regions. Experiments on two different histology datasets demonstrate its effectiveness and efficiency in comparison to previous models. | ['Yalin Zheng', 'Ke Chen', 'Sarah E Coupland', 'Abhik Mukherjee', 'Declan Sculthorpe', 'Yanda Meng', 'Liam Burrows', 'Hongrun Zhang'] | 2023-04-07 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhang_Weakly_Supervised_Segmentation_With_Point_Annotations_for_Histopathology_Images_via_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhang_Weakly_Supervised_Segmentation_With_Point_Annotations_for_Histopathology_Images_via_CVPR_2023_paper.pdf | cvpr-2023-1 | ['weakly-supervised-segmentation'] | ['computer-vision'] | [ 3.93894136e-01 2.02622592e-01 -2.55219430e-01 -5.01508892e-01
-1.10765648e+00 -4.91476923e-01 2.80359417e-01 4.45172071e-01
-5.72304904e-01 7.64770508e-01 -2.46766284e-01 -1.00213006e-01
1.77760780e-01 -4.43963438e-01 -5.63554287e-01 -1.09913671e+00
2.44750187e-01 6.53382897e-01 3.93106103e-01 -7.80449554e-05
3.95370089e-02 4.68444794e-01 -9.92617071e-01 2.10146070e-01
9.86621380e-01 8.85727823e-01 5.67709684e-01 3.60228479e-01
-2.12552935e-01 4.24923658e-01 -2.67378747e-01 -2.70331740e-01
1.80217698e-01 -4.26499575e-01 -1.07467723e+00 4.95562792e-01
1.05569996e-01 3.67987156e-02 1.18979178e-01 1.20499182e+00
4.83169734e-01 -1.27882928e-01 8.22104394e-01 -7.62817383e-01
-2.55457819e-01 3.81675541e-01 -9.85538781e-01 3.78573269e-01
-3.17008108e-01 9.15714949e-02 7.54572153e-01 -7.48993695e-01
7.52012551e-01 6.34822428e-01 6.31441057e-01 5.79735398e-01
-1.58309722e+00 -2.80729353e-01 -1.97512712e-02 -1.46082014e-01
-1.47193766e+00 -2.94400007e-01 8.02080929e-01 -4.59929258e-01
2.31248632e-01 1.60815969e-01 5.29911757e-01 7.83721268e-01
2.50329673e-01 9.83568847e-01 1.22141826e+00 -3.25207680e-01
2.77733117e-01 2.10808426e-01 -2.91166827e-02 7.50492692e-01
1.05772778e-01 -2.76345074e-01 -2.10926891e-03 -4.85160910e-02
9.89543915e-01 2.10875213e-01 -4.32480037e-01 -4.77666795e-01
-1.30878258e+00 8.07447910e-01 7.44498432e-01 3.55644047e-01
-4.40151811e-01 1.14530036e-02 3.95019740e-01 -1.94755033e-01
7.14387178e-01 2.50702739e-01 -2.16226384e-01 1.91504598e-01
-1.13900840e+00 8.65283329e-03 3.90440613e-01 7.66544044e-01
7.27353394e-01 -1.13672733e-01 -1.74038276e-01 9.80688393e-01
4.01619077e-01 2.12069377e-01 4.40530866e-01 -9.61603224e-01
-5.63975163e-02 7.45983660e-01 -1.67865112e-01 -6.22011423e-01
-4.57032830e-01 -6.16699815e-01 -1.15771675e+00 2.12733969e-01
6.60404205e-01 -5.67672327e-02 -1.24722314e+00 1.45325923e+00
6.24424517e-01 1.97784960e-01 -8.31264406e-02 1.05507040e+00
8.78868699e-01 4.00437772e-01 1.94987923e-01 -2.89361715e-01
1.25666070e+00 -8.60201120e-01 -8.35820258e-01 -6.87246025e-02
7.96045661e-01 -7.28027761e-01 9.40851927e-01 1.29185975e-01
-9.07568753e-01 -2.86841840e-01 -6.67532265e-01 -5.15192412e-02
-7.87926242e-02 1.52474508e-01 5.70859849e-01 4.31160003e-01
-1.13002586e+00 3.31521451e-01 -1.23151457e+00 -2.17686936e-01
1.11188233e+00 5.23429096e-01 -3.03197980e-01 -3.98093611e-02
-6.02471292e-01 5.37596762e-01 5.03434122e-01 3.42595994e-01
-8.02276969e-01 -7.11238325e-01 -8.63084614e-01 -3.05193424e-01
2.34740674e-01 -5.98949611e-01 1.16573608e+00 -9.24009740e-01
-1.26552761e+00 1.24159813e+00 -3.07192981e-01 -3.01979780e-01
6.88278675e-01 4.42316204e-01 2.13431165e-01 4.61142629e-01
2.01528594e-01 1.04072022e+00 6.36618793e-01 -1.54015243e+00
-4.66622561e-01 -5.09873629e-01 -3.58093172e-01 8.79335031e-02
-1.25396186e-02 -2.00541034e-01 -4.47866470e-01 -7.39420056e-01
1.69973895e-01 -9.64143693e-01 -7.94175446e-01 2.35131979e-01
-4.67975140e-01 -1.73157349e-01 8.74674916e-01 -6.09305024e-01
7.46738613e-01 -1.98984516e+00 1.34367198e-01 3.23362291e-01
3.38535130e-01 2.24215046e-01 2.01641873e-01 3.99885289e-02
1.56077728e-01 2.26535678e-01 -7.08985031e-01 -3.52730513e-01
-3.73625904e-01 3.01265478e-01 2.44103521e-01 7.58168638e-01
3.40546101e-01 1.03712237e+00 -9.21415746e-01 -1.05062461e+00
4.50741388e-02 4.02009100e-01 -5.09241223e-01 2.81703174e-01
-2.78662056e-01 1.26193106e+00 -5.39574981e-01 8.17777038e-01
5.33796787e-01 -6.20805740e-01 1.52628347e-01 -1.33524477e-01
3.04926708e-02 -4.64549750e-01 -7.10715652e-01 1.88042748e+00
-2.42320865e-01 4.97339875e-01 3.51761878e-01 -1.25352752e+00
7.59300828e-01 3.22779417e-01 7.63753414e-01 -2.71419078e-01
5.18944144e-01 3.23620558e-01 1.20779490e-02 -6.45877302e-01
-2.66141742e-02 -5.01342952e-01 6.34227395e-02 1.90186337e-01
-5.96289448e-02 -3.87932539e-01 2.60788649e-01 4.25497033e-02
7.91709602e-01 1.30257577e-01 6.35888651e-02 -4.52894658e-01
6.72695279e-01 2.69607782e-01 7.12252736e-01 2.55808324e-01
-2.44200915e-01 9.09228981e-01 3.80136162e-01 -3.40661198e-01
-1.00299525e+00 -8.39328468e-01 -6.23564601e-01 5.59078634e-01
2.99423039e-01 4.96251620e-02 -8.89081597e-01 -8.32738757e-01
-4.10831481e-01 1.37595251e-01 -5.63451946e-01 1.05474100e-01
-5.34706414e-01 -9.96703684e-01 3.99889261e-01 5.98288178e-01
4.53717470e-01 -1.07569480e+00 -4.58458245e-01 1.65151641e-01
-3.86469930e-01 -1.11942196e+00 -3.15743119e-01 2.70213097e-01
-1.04423440e+00 -1.03568852e+00 -1.22280610e+00 -1.35251129e+00
1.39101255e+00 2.06916898e-01 8.91038895e-01 3.75129163e-01
-5.08581221e-01 -4.40375432e-02 -2.35069782e-01 -4.88198876e-01
-3.85903656e-01 -1.35208210e-02 -4.30193365e-01 1.72261775e-01
-2.63870752e-04 -2.29941279e-01 -8.91645312e-01 3.15422237e-01
-1.27849662e+00 2.91554946e-02 7.78375149e-01 1.33565080e+00
1.26296985e+00 9.32692215e-02 4.47208136e-01 -1.34826207e+00
2.81013191e-01 -3.22785676e-01 -4.72731471e-01 1.53678179e-01
-2.98752815e-01 -2.16279119e-01 5.85002482e-01 -1.99469388e-01
-1.17409480e+00 3.56434047e-01 -3.54385197e-01 -3.27949494e-01
-4.07197326e-01 7.78969169e-01 1.07288107e-01 -4.19087321e-01
5.34494758e-01 1.71327159e-01 4.54134166e-01 -2.99996406e-01
8.67518317e-03 4.60648656e-01 5.01589239e-01 -4.09220994e-01
5.18483341e-01 9.23154593e-01 7.09099621e-02 -8.68847191e-01
-8.18405926e-01 -7.26950049e-01 -8.62535715e-01 -1.80529982e-01
8.95096481e-01 -6.59055829e-01 -3.23825330e-01 6.05797648e-01
-9.22684133e-01 -6.05610907e-01 -3.10020596e-01 4.17098194e-01
-5.42728961e-01 5.15703440e-01 -8.36897075e-01 -3.48662794e-01
-4.59436059e-01 -1.53988826e+00 1.24099231e+00 4.12514538e-01
-1.00644328e-01 -1.29507220e+00 5.30421361e-02 5.54879248e-01
1.50991455e-01 5.03144562e-01 9.76461470e-01 -5.59045613e-01
-5.29822171e-01 -2.28422523e-01 -2.66188085e-01 3.96181703e-01
1.87469885e-01 -6.81717843e-02 -7.04028845e-01 -3.93835753e-01
-1.12738945e-02 -5.02560377e-01 8.89072955e-01 6.84160888e-01
1.51483727e+00 -3.36355418e-02 -5.75046837e-01 6.66367054e-01
1.46386480e+00 -8.69676694e-02 4.56598431e-01 8.76538828e-02
7.51898289e-01 8.32286000e-01 7.40154803e-01 1.30806476e-01
2.93463618e-01 4.49957758e-01 3.26877683e-01 -8.83726716e-01
-8.74920338e-02 2.09440187e-01 -2.66001970e-01 6.86276436e-01
-1.43816009e-01 -1.08776659e-01 -1.09093165e+00 7.70667315e-01
-1.74157619e+00 -5.65915167e-01 -4.40895319e-01 1.86636055e+00
1.11672008e+00 -4.34804969e-02 7.61741698e-02 8.32703933e-02
8.28578174e-01 -1.89023167e-01 -5.29515862e-01 9.74296853e-02
1.96191758e-01 3.04660082e-01 3.71002108e-01 2.06415564e-01
-1.11454296e+00 9.09598291e-01 6.03492546e+00 8.66939008e-01
-1.26135457e+00 1.69628143e-01 1.20431519e+00 2.41015047e-01
-1.01126127e-01 -1.24015518e-01 -5.49254477e-01 5.20256341e-01
4.79083329e-01 2.83831716e-01 -3.01853567e-01 4.89686966e-01
3.95572841e-01 -1.75440118e-01 -1.02246857e+00 8.63668621e-01
-1.47136897e-01 -1.44197762e+00 -1.39917076e-01 1.89907089e-01
1.08871555e+00 -1.38536796e-01 1.18446484e-01 -1.64909288e-02
1.99005812e-01 -9.92338359e-01 3.15688282e-01 2.60496557e-01
8.10754001e-01 -7.26960897e-01 1.31052625e+00 5.39962292e-01
-8.18867743e-01 3.81107420e-01 -4.23956037e-01 4.50165510e-01
1.91333756e-01 5.78561723e-01 -1.10603356e+00 4.35414761e-01
5.89225650e-01 6.82567060e-01 -4.69854981e-01 1.27430785e+00
-8.05107728e-02 5.91448605e-01 -1.44816563e-01 7.86039606e-02
4.77426946e-01 -3.95712465e-01 2.86611557e-01 1.08283925e+00
5.30655906e-02 1.32668942e-01 5.62824607e-01 7.98966229e-01
-9.17317718e-02 3.79236788e-01 -2.80187964e-01 4.00613882e-02
4.41364106e-03 1.54335368e+00 -1.42869425e+00 -1.59067497e-01
-2.57239550e-01 7.86915004e-01 3.75995904e-01 3.37064445e-01
-7.85550535e-01 -1.53423948e-02 1.11219157e-02 2.41407543e-01
1.68578282e-01 -4.98178415e-02 -3.04149806e-01 -8.68140042e-01
-1.96116239e-01 -6.05547726e-01 4.70052332e-01 -3.46401304e-01
-1.35460353e+00 5.05315959e-01 -1.49195954e-01 -1.19718325e+00
1.42470315e-01 -3.35019141e-01 -6.61763251e-01 6.31312370e-01
-1.75547445e+00 -1.38041997e+00 -3.80551100e-01 4.46674705e-01
5.72699726e-01 1.98210716e-01 7.31820524e-01 2.15339199e-01
-6.45688474e-01 4.82745647e-01 2.30005383e-01 2.70331860e-01
5.27386785e-01 -1.47038627e+00 -3.38876188e-01 5.78952253e-01
2.41576079e-02 4.54270124e-01 5.50236344e-01 -5.39378703e-01
-1.01447189e+00 -1.34664321e+00 3.85998487e-01 -1.04965374e-01
3.88513625e-01 -7.47625455e-02 -1.02036166e+00 4.99981403e-01
2.62175322e-01 4.19457972e-01 1.07960713e+00 -2.71646082e-01
3.82481605e-01 -2.59726401e-02 -1.36120546e+00 5.42328656e-01
5.20562351e-01 -9.70396250e-02 -2.52405167e-01 5.77114105e-01
2.82116383e-01 -8.03691924e-01 -1.06535578e+00 5.21823049e-01
1.44271851e-01 -6.26929939e-01 8.33154738e-01 -2.79236972e-01
3.61147314e-01 -3.28046113e-01 4.24819529e-01 -1.21714854e+00
-2.15789676e-01 -3.90032709e-01 3.94068241e-01 1.19762909e+00
5.71844339e-01 -3.97527695e-01 1.17306161e+00 4.19291586e-01
-3.75998378e-01 -1.15601265e+00 -9.55183685e-01 -5.09061575e-01
2.31362522e-01 9.25132632e-02 1.00272089e-01 8.65721166e-01
-1.14711031e-01 1.60350606e-01 3.68312374e-02 3.65213566e-02
7.78991222e-01 2.51800299e-01 5.14803886e-01 -1.29654551e+00
6.97584525e-02 -5.23033202e-01 -5.01491129e-01 -7.87166893e-01
3.08065832e-01 -1.33014965e+00 3.68071765e-01 -1.67996836e+00
3.80211592e-01 -8.16959679e-01 -3.73173445e-01 4.83300120e-01
-3.13109607e-01 7.83385754e-01 -4.17644352e-01 4.40243274e-01
-6.87531888e-01 4.83216137e-01 1.76048553e+00 -4.13552284e-01
-1.28395200e-01 1.70117691e-01 -6.26673639e-01 8.56326580e-01
8.51362526e-01 -4.83111888e-01 -3.77912879e-01 -2.32065424e-01
-2.48112872e-01 4.40322272e-02 4.16823685e-01 -7.40688980e-01
2.41386414e-01 -1.04186414e-02 5.87379277e-01 -6.21437132e-01
1.19081683e-01 -8.33726048e-01 5.80023639e-02 4.90976810e-01
-3.89996260e-01 -5.01808107e-01 2.42866091e-02 8.36668432e-01
-4.34832960e-01 -3.31216276e-01 1.20422542e+00 -4.28478390e-01
-4.51819569e-01 6.15540326e-01 -2.96169549e-01 7.56723136e-02
1.26090360e+00 -3.21642876e-01 1.42141432e-01 3.62974033e-02
-8.08498919e-01 4.41235036e-01 5.03492415e-01 -9.36416909e-02
7.59784818e-01 -1.08566916e+00 -8.83661985e-01 -1.09962756e-02
1.44823015e-01 8.19542408e-01 4.04300034e-01 1.22788072e+00
-9.27695215e-01 1.64554268e-01 -5.97558655e-02 -1.23112977e+00
-1.18394268e+00 3.21827799e-01 2.45890886e-01 -4.20685023e-01
-6.79007232e-01 1.09294403e+00 4.30589408e-01 -3.86433452e-01
1.15670629e-01 -2.67978191e-01 -4.43886727e-01 -1.60980299e-01
2.40942955e-01 -4.13520150e-02 1.39531896e-01 -8.11259806e-01
-2.60957718e-01 5.20430624e-01 -4.90529001e-01 2.35874295e-01
1.36943638e+00 -1.07035413e-01 -3.16873401e-01 2.48289019e-01
1.15740871e+00 -2.91437626e-01 -1.40142334e+00 -3.45181733e-01
1.69255853e-01 -3.68665338e-01 2.96937287e-01 -5.26203871e-01
-1.49260652e+00 8.70252907e-01 5.76822639e-01 5.91312982e-02
9.15833771e-01 1.98358178e-01 1.02731621e+00 -5.11747263e-02
4.16190296e-01 -9.56192911e-01 2.09848851e-01 -1.68158691e-02
5.99629045e-01 -1.56143606e+00 -2.00598948e-02 -7.76276648e-01
-6.71936929e-01 9.96566713e-01 5.61905026e-01 2.36845724e-02
5.51387012e-01 1.72551796e-01 3.46643388e-01 -3.97828907e-01
-3.72184455e-01 -3.12314183e-01 2.03115374e-01 6.23691559e-01
5.60804129e-01 -7.08158240e-02 -3.41097981e-01 3.60150158e-01
2.66962826e-01 -1.19482584e-01 3.46763819e-01 1.09353805e+00
-2.68941611e-01 -1.24091232e+00 -2.50384003e-01 5.66004634e-01
-1.07523203e+00 1.63569599e-01 -1.94554254e-01 6.43863142e-01
3.71418111e-02 7.80081511e-01 -1.33568108e-01 2.90389478e-01
-4.54430655e-02 -3.38321030e-01 5.44092357e-01 -9.13818002e-01
-4.74375159e-01 5.76905906e-01 -4.11726236e-01 -2.10268170e-01
-6.44352376e-01 -6.69143558e-01 -1.72366488e+00 1.77597657e-01
-5.14465213e-01 1.44491524e-01 4.83125150e-01 9.19707477e-01
7.89348334e-02 5.84965289e-01 7.62629271e-01 -8.54861379e-01
-3.07200372e-01 -6.82762206e-01 -6.42390728e-01 5.43988168e-01
1.96738869e-01 -5.31612337e-01 -1.52765453e-01 6.39166474e-01] | [14.673062324523926, -2.4502429962158203] |
80fb69d2-4f21-447e-9e0e-23fddb0e0ee9 | a-new-approach-for-texture-based-script | 2009.07435 | null | https://arxiv.org/abs/2009.07435v1 | https://arxiv.org/pdf/2009.07435v1.pdf | A New Approach for Texture based Script Identification At Block Level using Quad Tree Decomposition | A considerable amount of success has been achieved in developing monolingual OCR systems for Indic scripts. But in a country like India, where multi-script scenario is prevalent, identifying scripts beforehand becomes obligatory. In this paper, we present the significance of Gabor wavelets filters in extracting directional energy and entropy distributions for 11 official handwritten scripts namely, Bangla, Devanagari, Gujarati, Gurumukhi, Kannada, Malayalam, Oriya, Tamil, Telugu, Urdu and Roman. The experimentation is conducted at block level based on a quad-tree decomposition approach and evaluated using six different well-known classifiers. Finally, the best identification accuracy of 96.86% has been achieved by Multi Layer Perceptron (MLP) classifier for 3-fold cross validation at level-2 decomposition. The results serve to establish the efficacy of the present approach to the classification of handwritten Indic scripts | ['Ram Sarkar', 'Pawan Kumar Singh', 'Mita Nasipuri', 'Supratim Das'] | 2020-09-16 | null | null | null | null | ['tree-decomposition'] | ['graphs'] | [-3.87548804e-02 -5.11046410e-01 1.75116599e-01 -1.54066190e-01
-4.41789269e-01 -6.93696022e-01 6.24311388e-01 -5.40379528e-03
-3.52634847e-01 9.76449370e-01 7.61273205e-02 -4.29473460e-01
-2.36471385e-01 -5.89422762e-01 -1.02995355e-02 -1.02701080e+00
-6.10702559e-02 2.83737838e-01 -1.25435337e-01 -2.66247809e-01
8.77860129e-01 9.24084902e-01 -1.04756272e+00 2.58504868e-01
7.34261513e-01 5.74185312e-01 5.31885885e-02 1.11389887e+00
2.97293991e-01 1.07576561e+00 -8.29234838e-01 -6.16666734e-01
8.64206254e-03 -5.76218724e-01 -1.05646420e+00 2.43829429e-01
1.61460310e-01 -3.63333404e-01 -3.50331366e-01 7.18773186e-01
5.22910118e-01 7.94675946e-03 9.22302663e-01 -3.36901635e-01
-6.93629503e-01 6.12335920e-01 -7.23504961e-01 3.35018456e-01
1.88167512e-01 -4.31134075e-01 7.19986260e-01 -8.16547334e-01
5.46793997e-01 9.83632505e-01 8.53818476e-01 -1.21747121e-01
-9.44204867e-01 -1.20146275e-01 -6.69062078e-01 4.36269790e-01
-1.25198007e+00 -8.14363956e-02 7.45101750e-01 -2.98631459e-01
1.03341556e+00 3.84903610e-01 3.09475541e-01 4.82719511e-01
2.88437307e-01 4.02570486e-01 1.97431588e+00 -8.55121791e-01
-2.29578987e-02 2.43189096e-01 1.04804397e-01 5.87682605e-01
2.03689307e-01 -4.52930301e-01 -2.53797352e-01 7.28228167e-02
6.97300017e-01 -6.56251252e-01 1.11154355e-01 5.19990325e-01
-5.79144597e-01 7.79908180e-01 -2.84945637e-01 9.91595507e-01
-5.15868425e-01 -3.37015152e-01 7.60347486e-01 3.19618523e-01
4.09580529e-01 1.67752996e-01 -3.06346089e-01 -7.91787386e-01
-1.20568490e+00 -5.29277734e-02 6.08755291e-01 4.68862623e-01
1.88292935e-01 4.58833575e-01 4.61873204e-01 1.37869751e+00
1.54581204e-01 3.67500454e-01 3.62997979e-01 -3.83635342e-01
3.62180680e-01 1.42390013e-01 8.32390562e-02 -1.09211051e+00
-5.74638098e-02 -9.68821719e-02 -5.51736414e-01 9.33696851e-02
4.74785507e-01 -3.93929005e-01 -9.94667709e-01 8.18563879e-01
-2.74281558e-02 -6.63253784e-01 4.07245278e-01 7.25298107e-01
2.85115659e-01 8.83803606e-01 -9.20838118e-02 -2.11156681e-01
1.28400648e+00 -7.21982777e-01 -7.19267011e-01 4.50851768e-01
-1.12718970e-01 -1.51282132e+00 6.98432446e-01 9.05223250e-01
-8.68984461e-01 -6.15902603e-01 -1.11419153e+00 3.44663978e-01
-2.48232946e-01 8.24895918e-01 5.23749828e-01 1.14856017e+00
-5.93249679e-01 4.68106538e-01 -9.23918724e-01 -6.91101551e-01
2.32586294e-01 3.55293721e-01 -5.40397227e-01 9.59358141e-02
-6.87762320e-01 1.12824023e+00 5.80879807e-01 2.00388610e-01
-3.34842980e-01 7.94165730e-02 -3.28911364e-01 -3.29304069e-01
-4.80353564e-01 8.63430023e-01 5.13484418e-01 -8.23114812e-01
-1.50457370e+00 7.61695206e-01 9.23705027e-02 -7.34083280e-02
7.13101983e-01 3.40647325e-02 -6.44904852e-01 3.89948428e-01
-3.26049596e-01 -1.98700838e-02 5.72449505e-01 -9.82062340e-01
-5.95323861e-01 -4.41489518e-01 -5.70312403e-02 4.17466126e-02
-3.34234476e-01 7.13812411e-01 -1.25338450e-01 -7.93793797e-01
2.37674057e-01 -6.80077851e-01 2.15499997e-01 -9.57320213e-01
-2.59743720e-01 2.08912250e-02 1.03680646e+00 -1.71327531e+00
1.29297352e+00 -2.25292420e+00 -4.28886674e-02 3.86090338e-01
-6.82000816e-01 2.06590280e-01 3.05280656e-01 8.29252660e-01
1.05367638e-01 -1.60562277e-01 -1.86709955e-01 3.41600627e-01
-1.59953520e-01 1.95099548e-01 3.76883820e-02 5.68796098e-01
3.41349512e-01 -5.70020266e-02 -4.14468437e-01 -8.81916225e-01
3.45991939e-01 5.50776362e-01 2.03311861e-01 -1.42005637e-01
5.81580102e-01 1.08028635e-01 -3.05460393e-01 1.23192072e+00
9.04950261e-01 5.26366413e-01 3.04684222e-01 -3.24935496e-01
-5.25711536e-01 -5.69640212e-02 -1.33651733e+00 7.74730861e-01
-4.93724585e-01 9.98425364e-01 -1.05455071e-01 -1.13742709e+00
1.31717384e+00 4.31866258e-01 3.01038414e-01 -3.98287803e-01
3.14084351e-01 5.79086661e-01 8.73106271e-02 -7.07592249e-01
9.01543319e-01 5.31996451e-02 1.11098774e-01 2.45218322e-01
2.11706668e-01 -1.01331929e-02 5.59354365e-01 -4.87085372e-01
4.32929099e-01 3.39573830e-01 3.30162615e-01 -2.95294404e-01
7.82653987e-01 5.19879609e-02 3.02408665e-01 5.00764906e-01
-3.93030822e-01 6.16702616e-01 5.18667936e-01 -2.92863250e-01
-1.39102757e+00 -7.77385294e-01 -7.97062337e-01 7.30217278e-01
-4.45131779e-01 2.49845222e-01 -6.53328240e-01 -3.41636807e-01
-3.97910386e-01 5.05229771e-01 -3.10195267e-01 6.49146199e-01
-9.11024511e-01 -1.21680510e+00 9.73409474e-01 1.87878326e-01
1.04792595e+00 -1.11828387e+00 -7.10476637e-01 4.30233866e-01
-1.12463869e-02 -1.03821540e+00 6.59444034e-02 3.64159673e-01
-9.82932985e-01 -8.38585138e-01 -9.82175231e-01 -9.73480225e-01
6.38357341e-01 -3.30007195e-01 2.86810040e-01 -5.16822696e-01
-5.96985757e-01 4.84693162e-02 -4.29227799e-01 -2.04706311e-01
-5.91547966e-01 -2.74975281e-02 -1.99521869e-01 -1.52228057e-01
5.91306150e-01 -2.28798717e-01 -1.00178741e-01 2.92458683e-01
-5.86863875e-01 -3.79759610e-01 6.97689950e-01 7.52480686e-01
1.35927439e-01 4.37926114e-01 5.41156471e-01 -5.92310488e-01
6.33554578e-01 -1.11780450e-01 -4.42244381e-01 5.12610734e-01
-3.03760052e-01 -3.65428358e-01 7.82632947e-01 -2.04258040e-01
-1.39361989e+00 5.85296862e-02 -3.46346438e-01 5.62814891e-01
-3.58408988e-01 5.07831156e-01 1.78234383e-01 -4.58998322e-01
5.94242513e-01 6.58379078e-01 -2.67769337e-01 -4.54712600e-01
-2.16255873e-01 1.43178833e+00 4.86943334e-01 -7.62985528e-01
4.18987840e-01 2.76278481e-02 -1.40856460e-01 -1.39843619e+00
4.05275822e-02 -3.21533442e-01 -8.40269864e-01 -6.48504198e-01
9.83974516e-01 -5.39131761e-01 -4.97455955e-01 1.21850049e+00
-9.87482905e-01 1.71398178e-01 3.20952594e-01 7.10051179e-01
-3.02205384e-01 6.44526243e-01 -9.82054710e-01 -1.16502082e+00
-3.74494851e-01 -1.03547955e+00 4.37647462e-01 5.24154365e-01
-1.52244747e-01 -1.03038108e+00 4.88938531e-03 4.91877049e-01
3.77592862e-01 3.29074383e-01 1.11393750e+00 -6.31708682e-01
3.51681337e-02 -4.65933532e-01 -2.54725724e-01 9.51131463e-01
5.19001186e-01 5.92971861e-01 -9.68148232e-01 -4.67032827e-02
-8.32588673e-02 -6.38901412e-01 3.47892910e-01 1.20917268e-01
7.37631023e-01 -3.87705043e-02 5.53041339e-01 8.38170424e-02
1.82255948e+00 9.28062260e-01 5.87493241e-01 4.53266352e-01
2.10380092e-01 2.90555954e-01 5.80973268e-01 6.29894555e-01
-4.49396065e-03 2.86944926e-01 -4.40643877e-01 1.81604445e-01
9.47218984e-02 2.25079477e-01 3.85649085e-01 1.04433322e+00
-8.29359174e-01 1.70704097e-01 -7.87318289e-01 5.85875452e-01
-1.01141405e+00 -7.36396968e-01 -1.60520181e-01 2.00614285e+00
1.16155052e+00 8.48682597e-02 2.46520817e-01 8.39952648e-01
6.10619068e-01 8.95078108e-02 1.70511872e-01 -1.29470503e+00
-3.09179395e-01 5.07965446e-01 6.24108315e-01 4.55373079e-01
-1.37843025e+00 8.66616607e-01 5.14442587e+00 9.33440864e-01
-1.44344556e+00 -1.81361243e-01 6.63410008e-01 4.77579772e-01
5.06431401e-01 -1.48667738e-01 -5.44377744e-01 4.85084385e-01
9.32410359e-01 1.90900654e-01 4.25341666e-01 5.52093446e-01
2.11269990e-01 -5.29956758e-01 -2.56176233e-01 6.92023754e-01
2.21755236e-01 -9.35156465e-01 -4.11682606e-01 -7.16870800e-02
8.34838390e-01 -8.86070058e-02 7.40953758e-02 -2.06260800e-01
-1.01752840e-02 -1.01779413e+00 4.65638906e-01 1.19962752e-01
6.02513969e-01 -1.00209987e+00 1.05684805e+00 -3.30554843e-02
-7.14888871e-01 -7.79236257e-02 -5.02412081e-01 1.65708780e-01
-1.49765447e-01 2.16682613e-01 -8.44104528e-01 6.36123061e-01
5.22979200e-01 1.56451628e-01 -4.99369532e-01 9.32603121e-01
-1.27256438e-01 1.24134624e+00 -4.91492748e-01 -4.63548094e-01
5.76873481e-01 -3.95007610e-01 1.86024383e-01 1.60378540e+00
2.89486706e-01 -3.84209752e-02 -5.78088582e-01 4.73954016e-03
4.87148702e-01 6.02521837e-01 -1.45280406e-01 -4.02624428e-01
2.79023916e-01 1.08168054e+00 -1.08426261e+00 -1.06564667e-02
-4.47423697e-01 9.21212077e-01 -2.35735670e-01 2.68592358e-01
-6.81811154e-01 -1.06939304e+00 -1.34965815e-02 -4.73993510e-01
4.78525758e-01 -5.43155432e-01 -4.84367967e-01 -7.01871991e-01
-1.77992489e-02 -1.26954925e+00 3.21394622e-01 -2.79570788e-01
-7.63246477e-01 5.78900456e-01 1.03419200e-01 -8.28844845e-01
-9.07827914e-02 -1.07049835e+00 -5.60592353e-01 1.14881790e+00
-8.75263691e-01 -1.30790091e+00 2.73014486e-01 1.35750175e-01
6.32008076e-01 -5.65279067e-01 9.50779200e-01 4.47096407e-01
-5.44802427e-01 7.12396324e-01 8.04994226e-01 5.50517380e-01
3.37983996e-01 -1.25191808e+00 -1.85935035e-01 9.29468036e-01
2.32235685e-01 3.28091502e-01 8.35696936e-01 -4.99524325e-01
-1.09299517e+00 -4.26992297e-01 1.04710293e+00 1.88749582e-01
6.67950630e-01 -5.03386296e-02 -4.91923362e-01 3.42170715e-01
3.99164617e-01 -3.93007189e-01 6.48946166e-01 -5.39938956e-02
-7.52550587e-02 -1.31019816e-01 -1.34463120e+00 3.08510721e-01
-1.01569176e-01 -7.49278128e-01 -6.07443154e-01 3.35311681e-01
-8.16173971e-01 -2.25102156e-01 -1.26704895e+00 -2.18999982e-01
9.53759849e-01 -1.03612483e+00 6.48023009e-01 -5.57174206e-01
7.16676176e-01 -4.47210595e-02 -3.80699307e-01 -9.54135478e-01
9.00130793e-02 -4.71609354e-01 5.32601535e-01 1.52662432e+00
3.93552333e-01 -5.68306446e-01 7.94542909e-01 3.44532244e-02
2.28742093e-01 -5.31844378e-01 -1.15689635e+00 -4.84121770e-01
1.42637566e-01 -1.25776112e-01 -7.80269206e-02 7.98978090e-01
-1.00251369e-01 -1.46293148e-01 -6.00666940e-01 3.68520878e-02
5.09446323e-01 1.14098355e-01 2.82665730e-01 -6.38781667e-01
-3.33757043e-01 -7.02165067e-02 -7.99687505e-01 -2.29228228e-01
-2.66465992e-01 -3.16979438e-01 -1.98019519e-01 -1.30968082e+00
8.82624984e-02 -4.04274434e-01 -1.20639175e-01 3.56183887e-01
-1.24213109e-02 6.10302806e-01 3.54305029e-01 1.30798742e-01
1.82080537e-01 -1.92638561e-01 8.97083342e-01 7.73912016e-03
6.20504133e-02 1.28791541e-01 -1.92457318e-01 5.85475683e-01
8.39119554e-01 -1.69367671e-01 -4.46106642e-02 -3.04120719e-01
-2.06840023e-01 1.19819567e-01 -1.93757281e-01 -1.02511418e+00
-2.32056335e-01 -1.11828804e-01 7.14763403e-01 -6.76725507e-01
1.95660651e-01 -6.74709082e-01 1.58114463e-01 4.67762291e-01
-2.01764092e-01 1.53274864e-01 1.71638414e-01 -1.44306362e-01
-5.39417207e-01 -6.87600493e-01 1.01708817e+00 -1.04977421e-01
-8.27671230e-01 -4.63057905e-01 -7.25428641e-01 -3.41183156e-01
1.05598128e+00 -7.10306227e-01 -1.22519635e-01 -9.94609296e-02
-6.31706059e-01 -3.49058509e-01 1.49619341e-01 -1.02274507e-01
4.29289341e-01 -9.28059280e-01 -9.46130335e-01 3.22493329e-03
-2.14894116e-01 -9.69825685e-01 2.55535096e-01 8.45041990e-01
-1.44114459e+00 5.68002403e-01 -9.10717428e-01 -1.51492134e-01
-1.77183342e+00 -3.71968687e-01 1.53808653e-01 -1.42437354e-01
-1.05409734e-01 8.18614721e-01 -9.72815335e-01 -1.00238189e-01
-1.48677649e-02 1.74560919e-01 -5.31002700e-01 3.32589179e-01
-4.12534773e-02 6.83683991e-01 3.07426333e-01 -1.02011466e+00
-1.99129105e-01 6.78827882e-01 -1.85554937e-01 -3.64257962e-01
1.28110540e+00 2.45842144e-01 -5.28945267e-01 4.86701936e-01
1.05065024e+00 3.74832571e-01 -9.27679121e-01 4.66980547e-01
1.51750028e-01 -5.95369399e-01 -1.90324128e-01 -1.05509615e+00
-4.19335216e-01 8.62763584e-01 6.80007696e-01 2.10921586e-01
1.39006889e+00 -5.14913857e-01 2.50291228e-01 4.19514269e-01
4.34303761e-01 -1.68587255e+00 -4.92153287e-01 5.42465508e-01
6.08495116e-01 -9.01148856e-01 2.26693496e-01 1.58020288e-01
-7.34820902e-01 1.90902805e+00 1.17530055e-01 -1.20170027e-01
4.09483820e-01 1.61415309e-01 2.54705995e-01 4.32841837e-01
-1.35836244e-01 1.81010008e-01 -3.94575410e-02 4.43486571e-01
1.01256609e+00 1.75804481e-01 -1.14187837e+00 1.49417505e-01
-2.67438769e-01 -1.42730340e-01 1.06272948e+00 1.26200807e+00
-2.73104489e-01 -9.45874155e-01 -7.81645596e-01 5.91776848e-01
-1.21886241e+00 -6.66849269e-03 -3.98111701e-01 1.22215188e+00
3.42645533e-02 8.90298724e-01 6.36252761e-02 -2.22366571e-01
-1.65598601e-01 3.26238304e-01 7.72399724e-01 9.96822417e-02
-7.24430919e-01 3.80207628e-01 6.25050426e-01 4.08071876e-01
-4.49107200e-01 -9.35897827e-01 -9.32973921e-01 -3.27367097e-01
-2.85422355e-01 5.39625585e-01 1.28309238e+00 8.61442864e-01
-5.61177194e-01 1.53013960e-01 7.77352989e-01 -2.60299057e-01
-7.77214706e-01 -1.37503803e+00 -7.18855321e-01 -7.36486018e-02
-3.70083712e-02 -2.07207456e-01 -1.44766167e-01 4.51978266e-01] | [11.844944953918457, 2.677154541015625] |
7c348b7d-43c1-4a03-b7be-4937b87fdadc | is-self-attention-powerful-to-learn-code | 2212.10017 | null | https://arxiv.org/abs/2212.10017v2 | https://arxiv.org/pdf/2212.10017v2.pdf | Are Code Pre-trained Models Powerful to Learn Code Syntax and Semantics? | Analysis of pre-trained code models also has revealed that they can effectively learn program syntax. However, these works are limited in analyzing code syntax and their distance-based approaches are not accurate due to the curse of high dimensionality. Furthermore, the study of the learnt program semantics of these models is rarely discussed. To further understand the code features learnt by these models, in this paper, we target two well-known representative code pre-trained models (i.e., CodeBERT and GraphCodeBERT) and devise a set of probing tasks for the syntax and semantics analysis. Specifically, on one hand, we design two probing tasks (i.e., syntax pair node prediction and token tagging prediction) to manipulate AST for the understanding of learnt program syntax. On the other hand, we design two tasks (i.e., semantic relationship prediction and semantic propagation prediction(inGraph) ) on the constructed control flow graph (CFG), data dependency graph (DDG) and control dependency graph (CDG) for the learnt program semantic analysis. In addition, to understand which kind of program semantics these pre-trained models can comprehend well, we conduct the statistical analysis for attention weights learnt by different heads and layers. Through extensive analysis in terms of program syntax and semantics, we have the following findings: 1) Both CodeBERT and GraphCodeBERT can learn the program syntax well. 2) Both CodeBERT and GraphCodeBERT can learn program semantics to different extents. GraphCodeBERT is superior to CodeBERT in learning program control flow and data dependency information but has a similar capability to CodeBERT in learning control dependency information. 3) Both CodeBERT and GraphCodeBERT can capture program semantics in the final layer of representation, but different attention heads and layers exhibit different roles in learning program semantics. | ['Yang Liu', 'Wenhan Wang', 'Jie Zhang', 'Shangqing Liu', 'Qiang Hu', 'Xiaofei Xie', 'Mengjie Zhao', 'Wei Ma'] | 2022-12-20 | null | null | null | null | ['program-synthesis', 'code-search', 'code-search'] | ['computer-code', 'computer-code', 'computer-vision'] | [-2.53103197e-01 1.57905370e-02 -5.49408138e-01 -6.28070712e-01
-5.24776317e-02 -5.42731166e-01 9.21362415e-02 5.20908892e-01
1.57330140e-01 -6.99264109e-02 2.37632632e-01 -8.55220675e-01
8.78788456e-02 -1.03175867e+00 -1.01552784e+00 -1.71375170e-01
-2.63626724e-01 4.52118069e-02 3.85309905e-01 -2.44150802e-01
5.11852741e-01 -3.99634652e-02 -1.54030323e+00 4.95270610e-01
1.28251612e+00 6.32207990e-01 6.24583066e-01 6.27702296e-01
-7.38852859e-01 1.44058955e+00 -3.55308294e-01 -2.52957284e-01
-5.31892359e-01 -3.11153293e-01 -1.18051863e+00 -5.71712673e-01
-2.24227421e-02 -2.02646136e-01 -3.12592387e-01 1.36561894e+00
2.29665097e-02 -2.09403902e-01 4.37587231e-01 -1.36247635e+00
-1.09210134e+00 1.22857714e+00 -5.28512061e-01 1.76333934e-01
4.85638171e-01 8.62563327e-02 1.25884807e+00 -5.76083302e-01
3.74326140e-01 1.36114514e+00 8.23038399e-01 6.93141341e-01
-9.75831985e-01 -6.62308216e-01 2.72659987e-01 2.20789567e-01
-1.05791295e+00 7.03993812e-02 8.99902582e-01 -8.54885161e-01
1.20611453e+00 3.41167860e-02 5.24152815e-01 8.42338204e-01
2.48142987e-01 1.02996361e+00 7.40186453e-01 -2.57894486e-01
-2.92138904e-02 1.60876125e-01 9.54066455e-01 1.26403975e+00
2.78701652e-02 2.65502244e-01 -1.18633315e-01 -1.03744015e-01
3.19456488e-01 1.71944112e-01 -3.83272529e-01 -3.42152417e-01
-8.65581870e-01 1.04909647e+00 9.55838859e-01 4.23760623e-01
2.97038019e-01 5.14924467e-01 8.16455066e-01 3.50947231e-01
-1.70341562e-02 4.09468085e-01 -6.63166225e-01 -2.21756041e-01
-3.68530601e-01 -1.18590936e-01 9.07527387e-01 1.42188752e+00
1.24691331e+00 3.01630080e-01 -4.78956364e-02 6.87686026e-01
7.23489106e-01 4.21846122e-01 6.91381454e-01 -3.35125029e-01
6.92772090e-01 1.23965383e+00 -6.24437988e-01 -1.23964155e+00
-4.63753313e-01 -1.34791821e-01 -5.37678063e-01 2.65557487e-02
1.29273176e-01 2.15007439e-02 -6.48187220e-01 1.76878369e+00
-3.03548872e-01 4.64606024e-02 2.13861279e-02 6.27106905e-01
1.19913971e+00 7.03447402e-01 3.49797308e-01 4.21976328e-01
1.34097636e+00 -1.00059426e+00 -3.81216079e-01 -4.73709852e-01
1.35328662e+00 -3.49717408e-01 1.45580888e+00 -1.39969617e-01
-6.32633865e-01 -9.07341838e-01 -1.11200798e+00 -1.84671417e-01
-5.74219525e-01 1.07325032e-01 8.77884924e-01 5.81180573e-01
-1.00569558e+00 4.70783770e-01 -9.79943693e-01 -4.72678393e-01
5.59252322e-01 3.85125041e-01 -3.04089356e-02 -1.83580622e-01
-9.50338364e-01 6.30244732e-01 6.69069171e-01 -3.56183082e-01
-1.23051202e+00 -7.98289418e-01 -1.16359305e+00 4.82077330e-01
1.86705559e-01 -3.59251201e-01 1.10149968e+00 -1.28181708e+00
-1.01595378e+00 9.53660548e-01 -1.89354554e-01 -1.34381667e-01
-3.50875884e-01 3.06112668e-03 -2.60514826e-01 -3.67896408e-01
4.41759340e-02 2.63181657e-01 5.06723821e-01 -1.39205861e+00
-5.21820545e-01 -4.17418569e-01 3.75372767e-01 -2.97587484e-01
-5.58198750e-01 1.61336511e-01 -3.12688202e-01 -5.70016325e-01
-9.42539945e-02 -6.52515709e-01 2.29619995e-01 -3.07100564e-01
-4.76166189e-01 -3.97419393e-01 1.14852333e+00 -7.40340829e-01
1.68173087e+00 -2.45661211e+00 2.14065924e-01 1.30160615e-01
6.03948951e-01 3.53651553e-01 -1.74161151e-01 3.23538750e-01
-2.45900601e-01 6.42763972e-01 -3.91256511e-01 7.91509449e-02
2.12860420e-01 5.01755357e-01 -4.77860838e-01 1.22579947e-01
1.81228146e-01 1.17049551e+00 -1.05651987e+00 -3.52265686e-01
3.55741475e-03 -8.07140116e-03 -1.08107853e+00 6.31069601e-01
-7.12551832e-01 2.05801561e-01 -1.01063967e+00 6.38871849e-01
4.22357380e-01 -4.59559381e-01 2.47866541e-01 -1.14751846e-01
-1.73882693e-02 3.26637357e-01 -3.57868969e-01 1.71529579e+00
-8.67814898e-01 7.11420357e-01 -1.29031807e-01 -1.36190987e+00
1.11787629e+00 3.59426364e-02 -4.28282209e-02 -6.06628299e-01
2.01494232e-01 2.36441791e-01 2.26517007e-01 -1.11617231e+00
1.89823151e-01 1.31920874e-01 -4.75529611e-01 5.83302975e-01
2.54410923e-01 1.54254735e-01 -1.27281666e-01 2.33117253e-01
1.25979435e+00 2.58740574e-01 2.28196129e-01 -6.27629519e-01
7.92973816e-01 1.80404827e-01 5.16809583e-01 5.69973409e-01
-1.68877974e-01 1.76615313e-01 9.72396374e-01 -4.92801309e-01
-7.01436043e-01 -8.28715503e-01 1.79751754e-01 1.47300243e+00
3.80630165e-01 -8.55290830e-01 -7.98625410e-01 -1.07277417e+00
1.06977839e-02 8.12266052e-01 -5.77802837e-01 -5.87501466e-01
-9.21246767e-01 -6.16626382e-01 7.47572839e-01 1.01827645e+00
7.02320218e-01 -1.07267332e+00 -6.20049477e-01 -5.02746478e-02
-1.23678176e-02 -6.51302636e-01 -5.79748392e-01 3.24828565e-01
-9.35966909e-01 -1.37457681e+00 -3.87507714e-02 -1.25299191e+00
8.52312624e-01 -6.98041394e-02 1.50739801e+00 9.30032551e-01
-1.94809176e-02 3.22834432e-01 -6.07269526e-01 -3.25270504e-01
-7.06135988e-01 3.21201146e-01 -6.16391718e-01 -4.80351895e-01
6.77028716e-01 -5.42506874e-01 -5.46298437e-02 2.05657423e-01
-8.00170183e-01 -4.68777977e-02 5.10967791e-01 7.44169474e-01
8.96630883e-02 3.05506259e-01 2.51610637e-01 -1.23684919e+00
6.47023499e-01 -8.09309125e-01 -7.32050240e-01 2.87454933e-01
-6.88601911e-01 5.38042367e-01 1.11868024e+00 -1.87073529e-01
-1.18388510e+00 -2.11508512e-01 -2.32762977e-01 -3.56895089e-01
-2.01785430e-01 1.04732788e+00 -3.51994336e-01 -1.15777262e-01
8.12462449e-01 3.01930130e-01 -1.09273702e-01 -5.03641129e-01
2.00515479e-01 5.13761938e-01 2.88041592e-01 -1.08916306e+00
8.30242336e-01 5.71404025e-02 -4.15349722e-01 -2.94853032e-01
-6.52465165e-01 -1.37132764e-01 -5.56683540e-01 2.05992907e-01
1.21450579e+00 -5.18726289e-01 -7.97407806e-01 5.32603085e-01
-1.25046730e+00 -5.62493384e-01 1.35097608e-01 8.60629156e-02
-4.84234214e-01 3.65668178e-01 -6.93677127e-01 -4.25965905e-01
3.81627865e-02 -1.54796636e+00 9.06095743e-01 1.42144889e-01
-2.67991930e-01 -1.47364759e+00 -1.51949406e-01 -1.04411364e-01
5.45636594e-01 1.90745555e-02 2.11174464e+00 -7.24297464e-01
-8.69032741e-01 1.25610769e-01 -5.48919618e-01 1.60636336e-01
1.43989464e-02 1.10435300e-01 -9.51673448e-01 -1.49655953e-01
1.36513352e-01 -3.26426893e-01 6.31854534e-01 2.33231470e-01
1.46720147e+00 -3.06581229e-01 -6.08799934e-01 9.19535995e-01
1.61729681e+00 5.36223531e-01 6.04326904e-01 2.39572972e-01
1.10208094e+00 6.07504487e-01 2.53341675e-01 8.60161334e-02
5.56682408e-01 3.49646896e-01 6.48333549e-01 1.85567424e-01
-1.96136460e-01 -7.14404464e-01 6.69944286e-01 1.13180363e+00
2.19223961e-01 -1.18447609e-01 -1.39998639e+00 2.87914515e-01
-1.91244805e+00 -5.08432329e-01 -3.97571355e-01 1.81114852e+00
7.15396166e-01 -3.79328690e-02 -1.43726304e-01 -1.37347519e-01
6.44979894e-01 2.01211914e-01 -4.48926210e-01 -6.56403720e-01
2.77689070e-01 9.10928920e-02 2.97747642e-01 3.12035948e-01
-7.75976956e-01 1.06073308e+00 5.27537155e+00 6.35955215e-01
-1.06302893e+00 8.59906524e-02 4.10441399e-01 8.64480972e-01
-7.07658827e-01 4.22122180e-01 -6.96951091e-01 4.88937706e-01
9.87777650e-01 -1.82102621e-01 6.20822847e-01 1.25850248e+00
-3.50525647e-01 1.57733902e-01 -1.52956164e+00 7.87007153e-01
-1.10706083e-01 -1.20115948e+00 3.89220454e-02 -3.23820204e-01
5.53396940e-01 -5.17703556e-02 -2.26973027e-01 1.01591921e+00
6.03424013e-01 -1.23420274e+00 6.01626635e-01 4.57373679e-01
4.98868644e-01 -5.53317130e-01 7.66616642e-01 4.12515402e-01
-1.54367530e+00 -5.70341349e-01 -4.48018432e-01 -8.87121707e-02
-4.57754731e-01 3.01197618e-01 -2.99808025e-01 4.79467034e-01
7.01598644e-01 1.23588729e+00 -8.43873739e-01 6.24377191e-01
-5.81958294e-01 7.46206522e-01 4.16693091e-01 -2.87352294e-01
3.04791242e-01 1.10814497e-01 1.02610126e-01 1.09321678e+00
3.61511737e-01 -1.37019604e-01 3.36497098e-01 1.78902459e+00
-1.81720182e-01 -8.60593319e-02 -6.62540495e-01 -3.27444673e-01
4.07768160e-01 7.60996282e-01 -5.75495958e-01 -3.00065577e-01
-8.85702312e-01 4.38833594e-01 6.30498588e-01 4.11190778e-01
-1.03775024e+00 -6.68917835e-01 6.05289400e-01 -4.15657759e-02
1.76289260e-01 -6.46040514e-02 -4.15688306e-01 -1.18489206e+00
-8.56772065e-02 -7.72621393e-01 5.28842568e-01 -8.34815502e-01
-1.10927761e+00 4.18389350e-01 9.93545502e-02 -7.73600817e-01
5.61123416e-02 -7.89105833e-01 -1.11631346e+00 9.91546631e-01
-1.42066681e+00 -9.11176860e-01 -5.12569785e-01 5.22659719e-01
3.06195885e-01 -3.79096150e-01 8.64539742e-01 3.44003439e-01
-5.93971848e-01 6.59351408e-01 -1.16849408e-01 6.07173979e-01
2.64586192e-02 -1.29842615e+00 3.50892097e-01 7.22280860e-01
-1.98146164e-01 1.06536937e+00 2.50227988e-01 -5.80773115e-01
-1.86098719e+00 -1.29344594e+00 7.05176651e-01 -6.28908813e-01
7.94301927e-01 -4.33527946e-01 -1.31792724e+00 1.05248809e+00
-1.62875317e-02 -2.18586531e-02 4.54885185e-01 2.60698348e-01
-7.31042325e-01 1.37833804e-01 -7.89212465e-01 3.01626801e-01
1.05736518e+00 -9.24991131e-01 -7.60859489e-01 1.79594103e-02
1.11261547e+00 -2.86744058e-01 -8.18859875e-01 3.10895741e-01
1.57851174e-01 -1.14407694e+00 6.96490943e-01 -7.81218827e-01
9.16284859e-01 -3.40567559e-01 -3.18489760e-01 -1.13562834e+00
-2.58413911e-01 -8.74149948e-02 -2.01262712e-01 1.51503730e+00
3.36290210e-01 -5.98902941e-01 7.19507873e-01 3.97490501e-01
-6.07915640e-01 -6.46884739e-01 -1.39512375e-01 -7.08049476e-01
3.87437671e-01 -5.40514112e-01 8.35191727e-01 1.17683136e+00
2.33049735e-01 3.96870166e-01 2.34359950e-01 1.75055712e-01
1.39524817e-01 5.86983263e-01 5.54062009e-01 -1.24965227e+00
-5.15860379e-01 -7.33636677e-01 -3.05850744e-01 -1.09645677e+00
7.67442286e-01 -1.61884809e+00 -8.52768049e-02 -1.29986739e+00
3.27636749e-01 -6.03286386e-01 -1.06346264e-01 6.15202427e-01
-2.52110183e-01 -7.49213994e-01 -9.58548635e-02 2.82449096e-01
-1.71299070e-01 4.32330966e-01 1.06258619e+00 -5.27715206e-01
-3.56698554e-04 -8.88893902e-02 -8.66723657e-01 7.94410944e-01
6.00475729e-01 -6.07451737e-01 -7.78985441e-01 -8.57553244e-01
4.39524412e-01 2.44327709e-01 4.21405882e-01 -7.73069143e-01
1.28004387e-01 1.14158504e-02 -1.59338266e-01 -2.00082779e-01
-5.74052572e-01 -8.32253158e-01 -2.87909985e-01 7.49229372e-01
-4.45419073e-01 1.36706933e-01 3.87912631e-01 5.00401795e-01
-4.31087524e-01 -8.17522943e-01 6.90991282e-01 -3.90713602e-01
-1.32647669e+00 3.00852329e-01 -3.73120368e-01 4.64197010e-01
6.87591195e-01 5.01533737e-03 -6.74087346e-01 -2.47850548e-02
-4.82400537e-01 3.51496249e-01 4.43297684e-01 7.75775075e-01
5.39020419e-01 -1.27070260e+00 -1.96312755e-01 4.17451829e-01
4.96811807e-01 -4.59660403e-03 6.81214258e-02 6.26593649e-01
-6.14876151e-01 3.04992586e-01 -1.29068524e-01 -6.50435328e-01
-9.82553780e-01 9.94843364e-01 4.07236487e-01 -2.72993356e-01
-4.16077822e-01 7.98263490e-01 6.58662975e-01 -1.03769410e+00
1.11020178e-01 -8.64057124e-01 -2.88998276e-01 -3.95659506e-01
3.81553024e-01 1.32949650e-01 -2.76189774e-01 -4.66278166e-01
-2.66950190e-01 7.35352039e-01 1.84643622e-02 9.35325861e-01
1.18411779e+00 2.65232623e-01 -6.48765743e-01 3.18421632e-01
1.67382979e+00 -2.13289380e-01 -8.58763933e-01 -4.16406197e-03
5.83234489e-01 -3.28432918e-01 -1.24120608e-01 -6.56243503e-01
-1.32922447e+00 1.41308212e+00 3.65183711e-01 2.66082942e-01
9.51761723e-01 2.36375302e-01 6.41527414e-01 3.69803727e-01
4.45346385e-01 -5.24064600e-01 2.25283012e-01 9.75065470e-01
4.72428679e-01 -1.13754869e+00 -6.10610604e-01 -4.54554021e-01
-1.85114413e-01 1.39855134e+00 1.13015854e+00 -1.01274617e-01
8.29330206e-01 5.09187579e-01 -2.88989842e-01 -5.80552280e-01
-5.44968069e-01 -5.47948852e-02 2.99598664e-01 7.67853618e-01
8.25652897e-01 -1.22328565e-01 1.76566750e-01 9.16860461e-01
-1.38256818e-01 -1.69534445e-01 4.38041717e-01 1.05744088e+00
-6.12538397e-01 -1.01904309e+00 -1.12082541e-01 5.83988369e-01
-1.22221492e-01 -3.69933993e-01 -3.07116389e-01 8.71875703e-01
4.07301933e-02 7.78402328e-01 -1.37415648e-01 -7.65779555e-01
2.93538541e-01 9.34748799e-02 1.77008420e-01 -1.14467180e+00
-8.33204389e-01 -6.23765945e-01 -1.83944151e-01 -7.03074932e-01
-1.43507093e-01 -2.86909658e-02 -1.78143311e+00 -3.06279719e-01
-1.05385423e-01 1.82165846e-01 3.15795064e-01 9.01620269e-01
2.43146390e-01 1.06275511e+00 4.06533509e-01 -2.51528889e-01
-3.28337461e-01 -5.92250884e-01 -3.99453759e-01 6.75022840e-01
2.37556621e-01 -4.67266828e-01 -2.72455454e-01 1.28108606e-01] | [7.527394771575928, 7.893374443054199] |
99dd238f-7d39-43af-83cf-dbebb7e81944 | multilingual-dictionary-based-construction-of | null | null | https://aclanthology.org/2020.lrec-1.519 | https://aclanthology.org/2020.lrec-1.519.pdf | Multilingual Dictionary Based Construction of Core Vocabulary | We propose a new functional definition and construction method for core vocabulary sets for multiple applications based on the relative coverage of a target concept in thousands of bilingual dictionaries. Our newly developed core concept vocabulary list derived from these dictionary consensus methods achieves high overlap with existing widely utilized core vocabulary lists targeted at applications such as first and second language learning or field linguistics. Our in-depth analysis illustrates multiple desirable properties of our newly proposed core vocabulary set, including their non-compositionality. We employ a cognate prediction method to recover missing coverage of this core vocabulary in massively multilingual dictionary construction, and we argue that this core vocabulary should be prioritized for elicitation when creating new dictionaries for low-resource languages for multiple downstream tasks including machine translation and language learning. | ['Winston Wu', 'Garrett Nicolai', 'David Yarowsky'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['cognate-prediction'] | ['natural-language-processing'] | [ 1.76974759e-01 3.42095308e-02 -7.70487785e-01 -2.19946414e-01
-9.13954973e-01 -1.16658294e+00 6.31214321e-01 3.22412282e-01
-6.01769984e-01 9.15689528e-01 6.89959466e-01 -9.58961844e-01
2.01174729e-02 -5.44457257e-01 -4.42479759e-01 -1.17836669e-01
4.23008829e-01 9.76987123e-01 -2.75695115e-01 -8.55329216e-01
2.80982256e-03 1.05783746e-01 -1.28953981e+00 2.46359289e-01
1.11071670e+00 1.81036800e-01 7.49168158e-01 -1.32963091e-01
-4.02191728e-01 2.15409815e-01 -1.28747210e-01 -6.76030576e-01
3.29697192e-01 -3.48975003e-01 -7.81262398e-01 -3.47179741e-01
5.62877178e-01 8.06484073e-02 -5.94963133e-02 1.21959889e+00
4.68760282e-01 -1.31433532e-01 3.83977950e-01 -4.84640092e-01
-6.72688127e-01 1.43489611e+00 -5.98595701e-02 5.36375284e-01
5.52111983e-01 -1.61309406e-01 1.92448461e+00 -1.42384803e+00
1.10894585e+00 8.64925325e-01 5.26295662e-01 4.72781658e-01
-1.28068411e+00 -1.00455773e+00 3.47138911e-01 -1.14621639e-01
-1.79732943e+00 -6.97481632e-01 5.19888878e-01 -6.59243047e-01
1.35733247e+00 2.68900365e-01 7.85796940e-01 1.03471899e+00
4.36003618e-02 5.25649905e-01 9.98287201e-01 -7.50458360e-01
-2.36464188e-01 2.02749893e-01 3.01251002e-02 5.98108113e-01
7.26682484e-01 9.84595791e-02 -7.44653523e-01 -2.53424346e-01
5.20227373e-01 -4.00310695e-01 -1.93920985e-01 -1.30793184e-01
-1.66130364e+00 1.03273213e+00 -4.93026227e-01 7.22542882e-01
-1.48097739e-01 -3.42016101e-01 6.14684880e-01 5.49233496e-01
5.99122226e-01 7.56724656e-01 -8.50107014e-01 1.31814972e-01
-9.56123292e-01 6.94516972e-02 7.31314898e-01 1.64649379e+00
8.54983270e-01 -5.17559126e-02 1.53902560e-01 1.02467251e+00
2.56711513e-01 7.23206699e-01 7.06181169e-01 -4.08982277e-01
3.39174062e-01 4.75316077e-01 -2.31938213e-01 -6.04238749e-01
-4.53223646e-01 -6.42131686e-01 -4.53506619e-01 -9.19710219e-01
9.97247770e-02 1.93256676e-01 -5.80504417e-01 1.96169388e+00
4.09026116e-01 -2.10566208e-01 1.70258388e-01 7.00385094e-01
1.10150266e+00 1.46029502e-01 3.07812750e-01 -5.71725011e-01
1.78956914e+00 -4.36034322e-01 -6.20402694e-01 -4.24904376e-01
8.19089890e-01 -9.62999225e-01 1.17396700e+00 1.82572991e-01
-8.18650126e-01 -4.30361986e-01 -8.99512470e-01 -3.45730990e-01
-2.93141574e-01 -3.31331789e-03 9.97813284e-01 7.27722287e-01
-1.02098095e+00 1.10322863e-01 -2.51693100e-01 -4.95060891e-01
-7.87731707e-02 3.10059994e-01 -5.06953180e-01 -1.35264352e-01
-1.61475039e+00 1.02006137e+00 8.25134754e-01 -4.68844056e-01
-9.77695644e-01 -7.89803922e-01 -1.10924900e+00 -2.61166185e-01
2.26023525e-01 -7.01945364e-01 8.97769928e-01 -9.61188436e-01
-9.30066347e-01 1.56133819e+00 -1.30376279e-01 -6.15241230e-02
-3.22888732e-01 -7.07468193e-04 -7.52628326e-01 -2.05909073e-01
6.36064231e-01 4.51569498e-01 5.20466447e-01 -8.39999437e-01
-9.88010764e-01 -1.32808343e-01 1.25992149e-01 4.59534585e-01
-3.09771180e-01 4.03022766e-01 -3.63046527e-01 -1.15897572e+00
1.70538381e-01 -7.74614513e-01 -5.02698757e-02 -6.09705687e-01
1.70407563e-01 -2.07315177e-01 -8.69730935e-02 -6.60865724e-01
1.61191320e+00 -1.68039739e+00 4.09282058e-01 2.50139564e-01
3.65139693e-01 6.05729362e-03 -3.34943086e-01 6.78220272e-01
-1.47163212e-01 1.41294718e-01 8.76441691e-03 6.92037046e-02
-8.78847763e-02 3.60960662e-01 -6.60268068e-01 4.39016610e-01
-1.37091041e-01 1.01170254e+00 -1.36824787e+00 -4.70878541e-01
-4.73775938e-02 1.03636436e-01 -6.90116704e-01 -8.18621293e-02
-3.91344994e-01 4.67428029e-01 -1.86766133e-01 8.47695351e-01
1.63535371e-01 -9.89286974e-02 1.07221210e+00 -1.12577155e-01
-4.99616414e-02 1.15491271e+00 -8.70449603e-01 2.08787036e+00
-5.89098871e-01 2.52023250e-01 -1.87706910e-02 -7.86148071e-01
9.46233213e-01 5.57853758e-01 4.89703894e-01 -6.09790921e-01
5.35589159e-02 6.99782550e-01 3.20651084e-01 -1.70173211e-04
8.01816225e-01 -7.15705693e-01 -2.44842380e-01 6.47155941e-01
5.86127520e-01 -1.34938180e-01 3.76781046e-01 3.18282783e-01
6.35387897e-01 5.28439181e-04 9.19300377e-01 -1.17336595e+00
7.09165931e-01 5.24854898e-01 7.07192123e-01 5.03376007e-01
1.61516909e-02 3.27483788e-02 -6.45152107e-02 -3.05756480e-01
-1.29109776e+00 -7.63207912e-01 -4.67538536e-01 1.68475783e+00
5.88417314e-02 -1.03180993e+00 -1.77613035e-01 -5.62382996e-01
-2.06173938e-02 6.30424559e-01 -8.66556764e-02 5.72991520e-02
-8.33006799e-01 -4.67969358e-01 8.25466394e-01 9.76840407e-02
-3.84710163e-01 -8.62169445e-01 2.87971407e-01 4.07444090e-01
-5.61701834e-01 -1.18340063e+00 -1.05517697e+00 2.98411399e-01
-6.42812967e-01 -1.09157550e+00 -3.26484650e-01 -1.30157936e+00
3.96130025e-01 3.71931821e-01 1.62097764e+00 1.65934637e-01
2.89009750e-01 2.83149660e-01 -4.54873681e-01 -2.54485607e-01
-7.26346374e-01 4.73569185e-01 7.27461755e-01 -4.65643704e-01
1.09781480e+00 -5.93609989e-01 -1.53634667e-01 1.36439160e-01
-7.17243075e-01 1.86557427e-01 2.83807814e-01 1.02270234e+00
8.09733927e-01 -6.34940088e-01 6.79169476e-01 -1.09303510e+00
8.36857200e-01 -6.67700291e-01 -7.08340526e-01 3.83146048e-01
-7.58899331e-01 1.30828127e-01 3.88520688e-01 -6.91648185e-01
-8.32507312e-01 -1.70388296e-01 -4.02697086e-01 6.44341186e-02
2.98705995e-01 8.80248427e-01 -2.99980074e-01 1.24917597e-01
7.14373529e-01 3.82639974e-01 -2.74721861e-01 -5.63955545e-01
8.05229545e-01 6.95675492e-01 2.94665843e-01 -1.09603059e+00
7.06813514e-01 1.06054507e-01 -5.19448459e-01 -7.41985559e-01
-8.25335324e-01 -8.39434266e-01 -7.98191488e-01 2.34499961e-01
4.38109368e-01 -1.67597914e+00 -3.04210275e-01 -3.00994962e-01
-1.10957360e+00 6.91623241e-02 -2.57974893e-01 6.92828536e-01
-3.06717277e-01 4.78412390e-01 -5.85047066e-01 -1.73597217e-01
-7.64958441e-01 -1.20422304e+00 1.33865201e+00 -5.96178114e-01
-6.33865476e-01 -1.37600362e+00 5.98279655e-01 2.93103933e-01
1.45913765e-01 -4.95983154e-01 1.19809413e+00 -9.62413132e-01
-2.12650448e-01 4.98579830e-01 8.21088850e-02 -6.12651482e-02
2.51961857e-01 -7.28372633e-01 -5.73916316e-01 -6.36339366e-01
-1.62524074e-01 -3.68086874e-01 5.62904716e-01 -2.43494399e-02
1.25968263e-01 -1.62541226e-01 -2.94342637e-01 7.62432098e-01
1.57907736e+00 3.01499609e-02 7.54756704e-02 3.02761734e-01
7.58455515e-01 3.96010220e-01 5.37193120e-01 1.54429570e-01
5.58982074e-01 7.13669240e-01 -3.91919315e-01 9.62462202e-02
-1.66413084e-01 -5.62716424e-01 6.41847968e-01 1.94617093e+00
3.14968884e-01 1.90489918e-01 -1.26979506e+00 1.01075685e+00
-1.39031994e+00 -6.81303084e-01 8.06789994e-02 1.94024980e+00
1.57904160e+00 4.31151409e-03 -2.82330550e-02 -4.77983952e-01
5.91551960e-01 -3.18959877e-02 -9.01295766e-02 -2.58881330e-01
-6.25160456e-01 7.56413698e-01 6.30667865e-01 8.72146428e-01
-7.31890798e-01 1.79116118e+00 7.46814346e+00 1.12705338e+00
-9.58149850e-01 6.41830146e-01 -1.89664021e-01 -6.50503533e-03
-1.15724826e+00 5.65225840e-01 -1.34657252e+00 9.74218175e-02
6.92284167e-01 -6.16514921e-01 5.68164527e-01 7.29848206e-01
-1.99671656e-01 3.92160028e-01 -1.16852248e+00 1.01924169e+00
2.19204918e-01 -1.52564692e+00 5.03311872e-01 2.22527031e-02
8.72295320e-01 5.74470937e-01 -4.75008458e-01 4.71554428e-01
8.41752410e-01 -9.64010000e-01 8.36539030e-01 -9.52931046e-02
1.76735795e+00 -5.30399978e-01 2.96616644e-01 2.27509990e-01
-1.31417322e+00 2.79983938e-01 -5.22210300e-01 -1.99679062e-01
2.46331364e-01 6.66307628e-01 -6.74193263e-01 3.76920462e-01
1.51868239e-02 6.88627660e-01 -3.32806736e-01 1.56657681e-01
-3.82417679e-01 6.59247100e-01 -1.63699418e-01 3.41687500e-02
7.46017471e-02 -2.80469388e-01 8.80412102e-01 1.43648696e+00
1.33727267e-01 2.10671917e-01 5.36855042e-01 6.72550738e-01
-2.69421935e-01 8.41619253e-01 -1.00492775e+00 -2.36638144e-01
8.37227941e-01 1.14387345e+00 -5.86850882e-01 -4.77205157e-01
-7.96699703e-01 6.00371540e-01 6.59308076e-01 9.19377804e-02
-4.21077162e-01 -1.37634967e-02 1.16398239e+00 -7.96862245e-02
1.66632280e-01 -6.06916428e-01 -2.83649385e-01 -1.86273754e+00
-8.55674893e-02 -1.56218517e+00 5.78925312e-01 -2.37939268e-01
-1.37222683e+00 7.19011128e-01 4.72040176e-02 -9.82461572e-01
-9.99791324e-02 -6.97377324e-01 1.67055592e-01 1.03515124e+00
-1.25420225e+00 -1.56229591e+00 4.54106152e-01 8.10508788e-01
7.04830468e-01 -7.95957506e-01 1.04097462e+00 6.71595335e-01
-2.10034192e-01 7.26286530e-01 1.08336307e-01 1.81642666e-01
7.86744416e-01 -9.69926000e-01 6.07643664e-01 9.25693214e-01
6.37277305e-01 1.30022430e+00 5.05836308e-01 -1.10342073e+00
-1.44030702e+00 -9.08292055e-01 1.52105856e+00 -8.50332022e-01
1.14621675e+00 -8.62509549e-01 -5.35943568e-01 1.13563192e+00
6.30864128e-02 -7.06370711e-01 1.20762730e+00 7.52724826e-01
-7.16683924e-01 1.06462516e-01 -5.39305449e-01 7.33679056e-01
1.62250233e+00 -1.05461967e+00 -9.98581827e-01 5.55186391e-01
1.13104582e+00 -2.52901435e-01 -1.06593311e+00 3.62742782e-01
5.75527430e-01 -3.96260805e-02 7.79972315e-01 -8.86511087e-01
1.34928301e-01 -1.12740308e-01 -6.11581385e-01 -1.15747392e+00
-6.49922609e-01 -7.23139822e-01 2.09842980e-01 1.10804474e+00
6.65911376e-01 -4.49232519e-01 2.35175356e-01 3.24285403e-02
-2.10411027e-01 -2.86843687e-01 -1.04492104e+00 -7.41934359e-01
4.71239537e-01 -6.68424547e-01 6.72071278e-01 1.53368795e+00
6.34794116e-01 1.21358001e+00 -3.35632294e-01 -2.46130273e-01
3.21148187e-01 2.00106606e-01 6.46589100e-01 -1.22439659e+00
-2.47699112e-01 -4.72570270e-01 -9.42955092e-02 -1.34219301e+00
5.62488616e-01 -1.84150004e+00 -3.09110463e-01 -1.08620799e+00
4.23333913e-01 -1.01635063e+00 -7.08998665e-02 4.49058741e-01
-1.90064609e-01 2.24376187e-01 -2.33189702e-01 4.55256790e-01
-3.18411618e-01 5.07354379e-01 9.74853218e-01 -1.98953435e-01
-1.14358477e-01 -8.03947210e-01 -1.40388465e+00 3.92857134e-01
3.49509656e-01 -6.70247078e-01 -3.77844453e-01 -7.01064527e-01
8.37471843e-01 -2.33969197e-01 -4.55555886e-01 -2.78438061e-01
3.75094973e-02 -7.22278297e-01 -3.41695160e-01 -1.73254550e-01
-2.41331056e-01 -7.20029116e-01 3.79902363e-01 2.88419127e-01
-1.71160892e-01 2.00817734e-01 2.72670329e-01 7.16188625e-02
-1.07767276e-01 -1.53908953e-01 5.95758855e-01 -4.74688351e-01
-1.04356909e+00 3.24367613e-01 -3.57997417e-01 6.54806733e-01
6.11933768e-01 1.20238448e-02 -8.21612999e-02 -4.90046740e-02
-5.52844286e-01 5.19830212e-02 4.98803735e-01 8.16092789e-01
3.91135126e-01 -1.55245149e+00 -1.15480947e+00 4.63480741e-01
7.02862084e-01 -5.05719841e-01 -5.26275814e-01 6.55627608e-01
-3.21758598e-01 7.79490888e-01 -1.17225058e-01 -3.67601693e-01
-1.04457045e+00 7.87215829e-01 1.72240689e-01 -5.36912918e-01
-5.77993631e-01 6.04043305e-01 4.19718921e-01 -9.13157940e-01
-1.72361866e-01 -8.43828470e-02 -1.39562100e-01 1.34959593e-02
4.51073349e-01 -2.39679277e-01 2.10972488e-01 -1.07257402e+00
-5.68037927e-01 3.59240651e-01 -1.84986666e-01 -2.24229932e-01
1.09429228e+00 -5.13917208e-01 -5.37290692e-01 3.63546163e-01
7.46045411e-01 7.39094436e-01 4.36491659e-03 -7.26138532e-01
3.27989399e-01 -1.57095388e-01 -1.11826181e-01 -4.84982073e-01
-6.26702964e-01 3.11076880e-01 5.60991988e-02 -6.30057693e-01
9.03820336e-01 2.37637535e-01 8.33337367e-01 3.94502193e-01
8.79239261e-01 -1.14537549e+00 -4.71278757e-01 1.26912737e+00
6.44575238e-01 -1.10639608e+00 1.27508581e-01 -3.22462738e-01
-5.67180693e-01 6.97498083e-01 3.18527043e-01 2.31976882e-01
8.96985531e-01 1.77641079e-01 2.01323420e-01 -6.15234196e-01
-7.41747499e-01 -6.65262640e-01 6.25049829e-01 5.96615314e-01
8.34325314e-01 2.07954407e-01 -1.20882130e+00 9.63000059e-01
-5.25241256e-01 -5.39452374e-01 1.39537379e-01 5.82464039e-01
-4.08122480e-01 -1.64769709e+00 -6.27697408e-02 4.30770189e-01
-6.22276068e-01 -1.17376006e+00 -3.67815822e-01 5.49705744e-01
4.86661971e-01 6.84831619e-01 -1.04966730e-01 -3.44913870e-01
-1.51893795e-02 2.39689693e-01 5.06692946e-01 -1.32260406e+00
-7.01060355e-01 3.07388812e-01 4.90504295e-01 -2.01771349e-01
-5.48570335e-01 -5.59316814e-01 -7.40662336e-01 -2.80790210e-01
-7.32352674e-01 3.51555318e-01 2.64625043e-01 1.08561516e+00
2.57288277e-01 -1.75687522e-01 6.60356507e-02 -2.29498997e-01
-2.00104862e-01 -1.01186752e+00 -5.84194779e-01 5.71977615e-01
-9.09897387e-02 -7.28969991e-01 6.03765622e-03 1.86452731e-01] | [10.913200378417969, 9.890629768371582] |
4f0ebc4c-ae1a-4df9-9b44-eec080301909 | natural-scene-recognition-based-on | 1506.07271 | null | http://arxiv.org/abs/1506.07271v1 | http://arxiv.org/pdf/1506.07271v1.pdf | Natural Scene Recognition Based on Superpixels and Deep Boltzmann Machines | The Deep Boltzmann Machines (DBM) is a state-of-the-art unsupervised learning
model, which has been successfully applied to handwritten digit recognition
and, as well as object recognition. However, the DBM is limited in scene
recognition due to the fact that natural scene images are usually very large.
In this paper, an efficient scene recognition approach is proposed based on
superpixels and the DBMs. First, a simple linear iterative clustering (SLIC)
algorithm is employed to generate superpixels of input images, where each
superpixel is regarded as an input of a learning model. Then, a two-layer DBM
model is constructed by stacking two restricted Boltzmann machines (RBMs), and
a greedy layer-wise algorithm is applied to train the DBM model. Finally, a
softmax regression is utilized to categorize scene images. The proposed
technique can effectively reduce the computational complexity and enhance the
performance for large natural image recognition. The approach is verified and
evaluated by extensive experiments, including the fifteen-scene categories
dataset the UIUC eight-sports dataset, and the SIFT flow dataset, are used to
evaluate the proposed method. The experimental results show that the proposed
approach outperforms other state-of-the-art methods in terms of recognition
rate. | ['Guanghui Wang', 'Shanshan Zhang', 'Jinfu Yang', 'Jingyu Gao'] | 2015-06-24 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 3.53645116e-01 -6.37502253e-01 -1.13318503e-01 -5.64353585e-01
-4.02510881e-01 8.97335187e-02 5.53959846e-01 -3.39306325e-01
-8.85748386e-01 2.81442851e-01 -1.53370872e-01 -1.00625753e-01
-2.33026911e-02 -7.00720429e-01 -5.19177794e-01 -1.18957007e+00
3.03339064e-01 2.32136309e-01 4.64295566e-01 5.21084666e-01
6.06714964e-01 4.78667676e-01 -1.43915689e+00 3.73036891e-01
9.60785031e-01 1.27519119e+00 6.49660528e-01 5.03662586e-01
-4.65500116e-01 1.43601191e+00 -4.70646590e-01 -4.78700921e-02
2.40688417e-02 -2.92549223e-01 -6.84983313e-01 5.79371929e-01
3.94334227e-01 -3.97884488e-01 -4.90797341e-01 1.19308555e+00
3.53702396e-01 5.74200630e-01 7.56884098e-01 -8.82984638e-01
-5.50518394e-01 4.50565636e-01 -7.18736112e-01 7.37988874e-02
-3.56857002e-01 -2.33431533e-02 4.92869526e-01 -1.06520236e+00
1.53972087e-02 1.31530416e+00 2.21263170e-01 4.64852720e-01
-8.85591447e-01 -5.38324952e-01 2.07875833e-01 5.67173183e-01
-1.49862182e+00 -2.85000026e-01 1.08159733e+00 -5.09636462e-01
8.13987374e-01 1.17197596e-01 5.57413220e-01 5.89619398e-01
7.60710910e-02 1.21359742e+00 1.13485086e+00 -4.75899071e-01
4.49597180e-01 9.75666866e-02 5.38955808e-01 6.64461672e-01
9.59828403e-03 -1.69648409e-01 -3.26999307e-01 1.64984375e-01
9.53497469e-01 3.22708547e-01 8.17362871e-03 -4.72733766e-01
-1.04252541e+00 6.48120880e-01 8.67264688e-01 2.89995521e-01
-4.12295789e-01 6.92749396e-04 3.46596509e-01 -4.24675196e-01
2.52980918e-01 -1.06170639e-01 1.01426370e-01 9.94592626e-03
-1.04593778e+00 -1.12255640e-01 4.70509738e-01 5.50377250e-01
5.54002941e-01 3.29350054e-01 4.02842797e-02 1.37499237e+00
5.99979818e-01 2.79383332e-01 9.10249889e-01 -4.23668116e-01
4.65684503e-01 8.12716663e-01 -1.43214971e-01 -1.11710727e+00
-1.23819195e-01 -2.82171816e-01 -1.21603799e+00 1.44103259e-01
2.30817974e-01 2.36871958e-01 -1.42466760e+00 9.70325947e-01
2.07862735e-01 4.58199620e-01 3.47074717e-01 1.21437967e+00
1.05172467e+00 1.06379116e+00 2.89688736e-01 -7.64022814e-03
1.04013753e+00 -1.28473592e+00 -3.81939650e-01 -5.28112233e-01
3.03254336e-01 -6.01325393e-01 9.19627786e-01 6.26367331e-01
-6.54531777e-01 -1.11187506e+00 -1.11130095e+00 1.35633156e-01
-2.91123092e-01 6.22133970e-01 7.15535820e-01 6.17771149e-01
-5.45565546e-01 3.23716909e-01 -1.09340394e+00 -2.19647929e-01
8.31410527e-01 2.95268148e-01 -1.12958312e-01 -4.09699768e-01
-7.49991775e-01 6.05997622e-01 9.96706069e-01 5.30447483e-01
-9.65701044e-01 -1.84693471e-01 -8.75772893e-01 1.07803993e-01
1.53682172e-01 -9.93901640e-02 9.28162694e-01 -1.09352922e+00
-1.72391081e+00 7.30667531e-01 4.84102638e-03 -3.50500941e-01
2.21879900e-01 -2.23727133e-02 -2.39754260e-01 8.69534463e-02
-4.11380947e-01 6.93915606e-01 8.23604345e-01 -1.05676126e+00
-6.67276561e-01 -3.97790283e-01 -2.52981782e-01 5.03737748e-01
-5.26890874e-01 1.07952893e-01 -7.20656931e-01 -5.31701326e-01
4.69422191e-01 -6.12178981e-01 -4.69212204e-01 -4.19270009e-01
-5.33043027e-01 -2.63395458e-01 8.33001971e-01 -7.18373537e-01
9.69728947e-01 -2.27853036e+00 3.19191068e-01 3.59809250e-01
-3.89272630e-01 4.35486257e-01 2.92087123e-02 -2.39130855e-01
1.28295168e-01 -2.71856785e-01 -5.86772799e-01 -2.90633619e-01
-2.07389072e-01 2.96847522e-01 -2.95309961e-01 4.33586299e-01
-1.37794269e-02 4.80464131e-01 -4.72202182e-01 -7.57658124e-01
6.57223523e-01 2.19278052e-01 -1.74023598e-01 4.45780545e-01
-3.87616977e-02 1.46376133e-01 -3.60129625e-01 6.17745042e-01
9.00264084e-01 -1.37161806e-01 4.97357883e-02 -2.81284451e-01
-1.09178253e-01 -1.85969457e-01 -1.38033366e+00 1.63397706e+00
-8.28458965e-02 3.26313347e-01 -4.21705186e-01 -1.40744352e+00
1.25178337e+00 -2.61137635e-01 9.71398503e-02 -6.67464077e-01
1.67405367e-01 1.61473751e-01 7.70262778e-02 -5.46621263e-01
3.82157236e-01 7.54968524e-02 1.76440005e-03 1.47430912e-01
1.11011259e-01 -1.01228021e-01 1.88470304e-01 -7.70919621e-02
4.17387456e-01 1.90903872e-01 -4.79359217e-02 -2.17001289e-01
8.34137797e-01 8.86773020e-02 7.25752532e-01 5.71118295e-01
-1.20452374e-01 4.13681746e-01 -4.26412337e-02 -6.32498682e-01
-7.91491628e-01 -9.43119466e-01 -9.37326327e-02 7.86115408e-01
5.83976805e-01 1.93606794e-01 -9.47480559e-01 -5.81453025e-01
-7.16307014e-02 7.28432238e-01 -3.24523956e-01 -1.96162760e-01
-3.48530293e-01 -1.06445312e+00 3.88905138e-01 6.92458451e-01
1.30611169e+00 -1.36142910e+00 -7.16306746e-01 9.36162397e-02
5.09825088e-02 -1.11520052e+00 -1.83563754e-01 -8.64194483e-02
-1.14990592e+00 -9.16678250e-01 -8.74138951e-01 -1.26148903e+00
9.53792274e-01 3.77139509e-01 4.17126775e-01 -1.54884681e-01
-5.77909470e-01 3.75942141e-02 -2.02049062e-01 -1.92465827e-01
-3.18352059e-02 -1.18264139e-01 1.01016924e-01 6.13099575e-01
5.35589635e-01 -2.52507269e-01 -5.56766927e-01 2.83594489e-01
-1.07089972e+00 4.99170065e-01 9.78095472e-01 9.62190509e-01
7.90751755e-01 3.15482020e-01 1.75122842e-01 -7.98076689e-01
3.42755228e-01 1.86137948e-02 -8.16780627e-01 3.48580271e-01
-2.22803131e-01 -8.56632590e-02 6.42845869e-01 -5.78894079e-01
-1.57089126e+00 3.54404718e-01 4.27755229e-02 -4.61369127e-01
-4.95401889e-01 8.70979488e-01 -3.95111769e-01 -2.43599489e-01
3.57279181e-01 8.18450809e-01 -2.36226872e-01 -5.83173633e-01
2.04556495e-01 1.13395095e+00 7.07641721e-01 -4.18622911e-01
3.64260226e-01 2.80301213e-01 -3.07258427e-01 -1.18928337e+00
-5.07794321e-01 -4.97212261e-01 -8.29587042e-01 -3.40105504e-01
1.06065261e+00 -7.77266800e-01 -4.24288064e-01 1.36545849e+00
-9.56237078e-01 -5.28616130e-01 2.37275109e-01 8.70394945e-01
-4.55579847e-01 5.21323621e-01 -8.28846395e-01 -1.06358147e+00
-2.80698240e-01 -1.16226065e+00 4.99518156e-01 8.08262885e-01
4.34605360e-01 -6.30464494e-01 -1.78137422e-01 6.36952639e-01
6.51702955e-02 -1.56832129e-01 8.14073443e-01 -4.92988527e-01
-6.43600404e-01 -2.28532106e-01 -4.63610053e-01 7.87451386e-01
1.83455080e-01 1.48965828e-02 -9.61133361e-01 -1.44281954e-01
-1.19487029e-02 -6.21789515e-01 1.14965451e+00 4.32326674e-01
1.63180423e+00 -4.55776304e-02 -1.52932540e-01 6.62055612e-01
1.51236606e+00 7.61000276e-01 9.67943132e-01 3.81485730e-01
9.78622973e-01 2.81335980e-01 6.29054785e-01 3.31873983e-01
2.37192169e-01 2.61266977e-01 2.72490650e-01 -2.10919634e-01
9.35756639e-02 -3.76477502e-02 2.03875899e-01 1.21001780e+00
-1.70012832e-01 1.42416075e-01 -1.05725145e+00 4.03774172e-01
-1.96923387e+00 -8.46516013e-01 6.47095591e-02 2.16916275e+00
7.20588088e-01 3.15634757e-01 -2.25395545e-01 2.07598090e-01
7.31021821e-01 2.44724587e-01 -6.87358379e-01 -1.69077560e-01
-5.58289140e-02 -2.68030874e-02 4.17121798e-01 2.16071248e-01
-1.50831854e+00 1.30430675e+00 5.36702681e+00 1.13649714e+00
-1.44827545e+00 -2.31078222e-01 9.01933134e-01 2.39068687e-01
5.62559426e-01 -2.37271249e-01 -6.51965261e-01 5.66867888e-01
4.27848488e-01 2.98729777e-01 4.57619041e-01 1.18318534e+00
9.14677903e-02 -3.56592238e-01 -6.98117554e-01 1.34768188e+00
3.79974484e-01 -1.13248348e+00 3.03441256e-01 -2.51155198e-01
7.04248190e-01 -1.24863975e-01 3.63287479e-02 3.29567403e-01
1.94767073e-01 -8.59249234e-01 5.55074632e-01 8.53586018e-01
4.75678533e-01 -9.08844113e-01 7.55835176e-01 5.90228200e-01
-1.07762325e+00 -2.24336088e-01 -8.86709392e-01 5.30181080e-03
-2.89082259e-01 4.95200902e-01 -4.06991363e-01 4.04426664e-01
8.38300765e-01 7.63095260e-01 -5.31638145e-01 1.46773434e+00
-2.23630890e-01 8.38211477e-01 -1.87870547e-01 -4.06698585e-01
4.94170010e-01 -4.56947237e-01 2.22434998e-02 1.30468929e+00
5.12610637e-02 2.00355217e-01 4.75733250e-01 8.04308355e-01
-3.61224287e-03 2.23794982e-01 -7.75184333e-02 -6.55986220e-02
2.21881822e-01 1.32705510e+00 -9.30324852e-01 -6.79022789e-01
-2.79298455e-01 1.13460219e+00 3.78876895e-01 4.81705278e-01
-7.65680075e-01 -6.40421093e-01 1.37535900e-01 -6.12386644e-01
4.06718463e-01 -3.42193604e-01 -2.55813956e-01 -1.24794674e+00
-9.60790813e-02 -6.59113586e-01 4.10426408e-01 -9.65326428e-01
-1.10767114e+00 4.11771089e-01 -2.16192529e-01 -1.18773758e+00
1.31094575e-01 -8.28531742e-01 -7.34829128e-01 7.80856431e-01
-1.38132823e+00 -9.99610245e-01 -5.49462616e-01 7.08266914e-01
7.60790229e-01 -3.24218363e-01 6.20776176e-01 2.81721383e-01
-1.16512859e+00 3.29501003e-01 2.80580014e-01 6.66908562e-01
3.06309998e-01 -9.83512402e-01 -6.52486756e-02 1.13834858e+00
2.85538167e-01 4.12197381e-01 8.01248252e-02 -4.70057577e-01
-1.38909781e+00 -1.33262455e+00 3.36264551e-01 3.78073096e-01
2.57102937e-01 -1.85356811e-01 -1.00294530e+00 3.56557608e-01
-3.94488238e-02 1.29288703e-01 7.21389890e-01 -2.93830693e-01
-1.31143793e-01 -4.68300432e-01 -9.82649624e-01 5.44990659e-01
5.02475977e-01 -4.05764937e-01 -7.49715388e-01 1.21626057e-01
1.99294806e-01 -4.35766190e-01 -6.68563008e-01 2.97291815e-01
5.64913452e-01 -7.30111063e-01 8.68548274e-01 -5.27951002e-01
4.38403189e-01 -5.29589772e-01 -2.79943705e-01 -1.13478136e+00
-4.51764643e-01 3.31730455e-01 -6.92064166e-02 1.37814009e+00
7.11393282e-02 -3.90361100e-01 9.30394650e-01 6.73793256e-01
-1.49990708e-01 -7.66841710e-01 -8.22627544e-01 -6.99467778e-01
-2.66257972e-01 -4.12333041e-01 2.92352349e-01 8.49576056e-01
-1.79644972e-01 1.94080040e-01 -3.08295190e-01 2.34967664e-01
9.26247656e-01 3.67397994e-01 7.05635250e-01 -9.47023690e-01
-3.46658468e-01 -4.68232334e-01 -6.85025752e-01 -1.37606800e+00
-7.80940279e-02 -6.58742249e-01 5.07902741e-01 -1.60721397e+00
5.07509470e-01 -5.01088262e-01 -7.58536577e-01 4.08994645e-01
-3.34998518e-01 2.76977628e-01 2.45248735e-01 5.17892480e-01
-7.42035627e-01 7.15606451e-01 8.92426133e-01 -6.50517523e-01
-2.74935782e-01 7.16721779e-03 -1.91444963e-01 9.17717576e-01
7.69451618e-01 -1.38529688e-01 -3.25530440e-01 -4.94135082e-01
-6.99650645e-01 -2.28632301e-01 3.82734179e-01 -1.27573156e+00
5.88330805e-01 -3.13141733e-01 8.47249746e-01 -7.91113019e-01
3.92087400e-01 -7.30283558e-01 -2.09076256e-01 6.37301028e-01
-3.37053150e-01 -7.38544226e-01 2.70815313e-01 3.97988260e-01
-4.65527743e-01 -4.44257230e-01 9.95682657e-01 -1.96878895e-01
-1.35151458e+00 4.84803319e-01 -3.73325914e-01 -4.56919760e-01
1.06026578e+00 -4.07254905e-01 -2.28369189e-03 6.48980066e-02
-4.20258492e-01 1.26632228e-01 -3.63137908e-02 3.87605399e-01
1.10084736e+00 -1.50904000e+00 -5.39488673e-01 2.56069690e-01
1.03171930e-01 3.08493078e-01 6.35773361e-01 5.93703330e-01
-8.12361240e-01 3.26421142e-01 -3.35765809e-01 -9.30045724e-01
-1.41355252e+00 5.08272111e-01 3.05700123e-01 -1.62967801e-01
-3.30468565e-01 1.03454792e+00 2.69678235e-01 -3.82892549e-01
5.77574670e-01 -3.37459058e-01 -4.91138726e-01 -2.56314069e-01
6.47330821e-01 1.84491947e-01 -1.74152717e-01 -7.64955699e-01
-3.73608112e-01 7.07716465e-01 -2.22544953e-01 1.73489511e-01
1.18125856e+00 1.97477236e-01 -2.45552287e-01 4.69832659e-01
9.76423144e-01 -7.01310337e-01 -1.32655525e+00 -4.57554817e-01
-4.83492725e-02 -5.07865429e-01 2.45674938e-01 -5.75026631e-01
-1.28069079e+00 1.19247985e+00 7.83986688e-01 -1.84708968e-01
1.23547244e+00 -4.63330925e-01 6.21019900e-01 6.48102164e-01
3.86384547e-01 -1.22459292e+00 -4.83236834e-03 3.19828659e-01
6.08779490e-01 -1.08078802e+00 -4.05140482e-02 -2.58863121e-01
-7.30942428e-01 1.21116436e+00 9.67350185e-01 -3.62182379e-01
5.01036108e-01 -1.97484754e-02 1.41697759e-02 3.93160582e-01
-9.18004215e-02 3.51001173e-02 3.00035626e-01 2.04369709e-01
1.04736798e-01 1.17444605e-01 -4.77339290e-02 8.60781968e-01
1.98711991e-01 2.24622265e-01 1.83633536e-01 9.95353520e-01
-7.10787952e-01 -6.63319468e-01 -4.77058381e-01 5.48625886e-01
-1.27519771e-01 -7.82996193e-02 -2.96178699e-01 3.85816246e-01
1.17817156e-01 9.51743543e-01 2.96533406e-01 -6.90432429e-01
6.01262711e-02 -6.67968858e-03 5.13186634e-01 -4.00467217e-01
-1.25591516e-01 2.66701151e-02 -3.94674897e-01 -3.42096686e-01
-5.85266888e-01 -3.33028972e-01 -1.34189081e+00 7.26203173e-02
-4.18333709e-01 1.73721671e-01 6.93727016e-01 1.03951669e+00
5.43814786e-02 5.19582391e-01 6.80391073e-01 -1.13022470e+00
-3.66656810e-01 -1.11744952e+00 -7.60889173e-01 3.64748955e-01
-1.75411358e-01 -6.38700783e-01 -8.65202546e-02 4.38617438e-01] | [9.923293113708496, -0.36613568663597107] |
537d095f-d552-4f78-9a1d-75fcb4e2d37b | scalable-mutual-information-estimation-using | 1801.09125 | null | http://arxiv.org/abs/1801.09125v2 | http://arxiv.org/pdf/1801.09125v2.pdf | Scalable Mutual Information Estimation using Dependence Graphs | The Mutual Information (MI) is an often used measure of dependency between
two random variables utilized in information theory, statistics and machine
learning. Recently several MI estimators have been proposed that can achieve
parametric MSE convergence rate. However, most of the previously proposed
estimators have the high computational complexity of at least $O(N^2)$. We
propose a unified method for empirical non-parametric estimation of general MI
function between random vectors in $\mathbb{R}^d$ based on $N$ i.i.d. samples.
The reduced complexity MI estimator, called the ensemble dependency graph
estimator (EDGE), combines randomized locality sensitive hashing (LSH),
dependency graphs, and ensemble bias-reduction methods. We prove that EDGE
achieves optimal computational complexity $O(N)$, and can achieve the optimal
parametric MSE rate of $O(1/N)$ if the density is $d$ times differentiable. To
the best of our knowledge EDGE is the first non-parametric MI estimator that
can achieve parametric MSE rates with linear time complexity. We illustrate the
utility of EDGE for the analysis of the information plane (IP) in deep
learning. Using EDGE we shed light on a controversy on whether or not the
compression property of information bottleneck (IB) in fact holds for ReLu and
other rectification functions in deep neural networks (DNN). | ['Alfred O. Hero III', 'Morteza Noshad', 'Yu Zeng'] | 2018-01-27 | null | null | null | null | ['mutual-information-estimation', 'information-plane'] | ['methodology', 'methodology'] | [ 5.68026416e-02 -3.62219065e-02 -1.79268539e-01 -4.71091837e-01
-8.58050704e-01 -2.25669459e-01 1.11426122e-01 1.54760808e-01
-9.05580997e-01 9.07596350e-01 -2.09752783e-01 -3.43784660e-01
-7.35374510e-01 -7.37541556e-01 -9.94644821e-01 -1.09489572e+00
-7.42600143e-01 4.81555343e-01 1.06744424e-01 1.35717675e-01
5.36484540e-01 6.45228803e-01 -1.64423645e+00 -3.13646615e-01
7.25872517e-01 1.33290517e+00 1.12250678e-01 7.25787699e-01
1.85380690e-02 5.94820976e-01 -4.60560858e-01 -5.63827276e-01
2.75584340e-01 -4.55099821e-01 -8.86166573e-01 -7.91099548e-01
5.97738862e-01 -4.01497215e-01 -6.04929924e-01 1.35424185e+00
6.15188181e-01 1.88128039e-01 1.10365987e+00 -1.35474205e+00
-5.60896039e-01 9.54030693e-01 -6.38758481e-01 7.56252646e-01
-2.97456682e-01 -5.09045064e-01 1.23869169e+00 -7.05453753e-01
4.00308788e-01 9.06071424e-01 7.40011215e-01 4.01414454e-01
-1.23062205e+00 -1.01707304e+00 -3.98815274e-01 6.31223738e-01
-1.78598428e+00 -2.35202402e-01 7.00719059e-01 -1.27636984e-01
8.67158532e-01 -7.49548227e-02 3.16511959e-01 6.27514839e-01
4.50953484e-01 8.16822231e-01 9.06915724e-01 -2.80955315e-01
2.87670493e-01 8.60850811e-02 2.80241877e-01 9.64340210e-01
6.98875010e-01 2.59016678e-02 -7.32029438e-01 -1.71864145e-02
8.29298913e-01 -1.53464139e-01 -3.49157333e-01 -3.68844748e-01
-6.38339162e-01 1.09409368e+00 6.32303059e-01 2.84103215e-01
4.65480387e-02 5.57461500e-01 5.10983467e-01 5.84610224e-01
3.18857282e-01 -2.02362128e-02 -5.66900969e-01 -1.78068057e-01
-8.42737854e-01 -1.66642889e-02 1.00221539e+00 1.13030207e+00
9.25574958e-01 -8.62682983e-02 3.89089882e-01 6.16661310e-01
3.20226759e-01 7.59706497e-01 3.40486228e-01 -9.32827950e-01
4.77473497e-01 1.72053814e-01 -3.74494970e-01 -1.08988535e+00
-3.87766868e-01 -5.50012708e-01 -1.30044973e+00 8.22726917e-03
4.53016460e-01 1.92860924e-02 -2.70370305e-01 2.03381848e+00
4.41583581e-02 -2.70907283e-01 1.11022741e-02 5.54386020e-01
7.96514094e-01 2.45770052e-01 -3.29858571e-01 -2.13700876e-01
1.05266011e+00 -2.39338368e-01 -6.23035729e-01 1.26810744e-02
9.42891657e-01 -2.82655984e-01 5.74157596e-01 2.73497701e-01
-1.08508623e+00 -2.06467122e-01 -1.18226504e+00 -1.33258849e-01
-3.51401955e-01 -2.49941975e-01 5.76817691e-01 9.02705133e-01
-1.23606324e+00 8.54662418e-01 -7.22047091e-01 -1.11042606e-02
7.93547928e-01 8.36057901e-01 -5.28151691e-01 -1.38540212e-02
-1.23164594e+00 6.36961222e-01 2.63285160e-01 -9.18527544e-02
-7.54603922e-01 -4.98038411e-01 -8.48729670e-01 1.27414033e-01
-9.92758349e-02 -3.60575736e-01 6.40703797e-01 -6.42196834e-01
-1.13424909e+00 7.37569749e-01 -5.67430332e-02 -7.13219583e-01
1.93663612e-01 -2.51510739e-01 1.60809606e-01 4.29340541e-01
-2.36122325e-01 7.59563088e-01 6.85003698e-01 -7.96783686e-01
-3.41503382e-01 -8.71810615e-01 -2.61542767e-01 3.11121810e-02
-4.53917086e-01 -2.83699274e-01 8.32377523e-02 -2.44477972e-01
3.61192375e-01 -8.29991996e-01 2.54427433e-01 1.37477443e-01
-1.25643834e-01 -1.78270370e-01 5.29832244e-01 -4.77212608e-01
1.14708996e+00 -2.19229889e+00 1.96836084e-01 2.76532024e-01
3.57775778e-01 7.60312527e-02 -2.63468977e-02 2.82460630e-01
-1.78617816e-02 2.22871363e-01 -4.85897005e-01 -1.92212507e-01
-7.73686841e-02 2.40668535e-01 -2.04867553e-02 1.15443432e+00
-3.17213356e-01 6.99245930e-01 -6.47082746e-01 -6.68608427e-01
-1.91441551e-01 8.04779351e-01 -7.59562314e-01 -1.16953641e-01
3.86299044e-01 1.18648484e-01 -1.12511963e-01 1.57567382e-01
1.11685491e+00 -3.43878269e-01 1.37137368e-01 -3.08196694e-01
1.12053677e-01 2.07115918e-01 -1.17012537e+00 1.44716585e+00
-2.71138400e-01 1.11792564e+00 3.93317267e-02 -1.45196867e+00
9.40764487e-01 3.34346779e-02 4.90716070e-01 -5.63972473e-01
4.11464751e-01 2.98872352e-01 6.78007379e-02 -1.97270483e-01
1.97748914e-01 -1.98358506e-01 3.46541740e-02 5.21097183e-01
4.97743696e-01 2.41069302e-01 -1.09922618e-01 2.46060520e-01
1.35005546e+00 -5.34484744e-01 3.51891339e-01 -6.72059536e-01
4.69191849e-01 -5.93810022e-01 2.37535968e-01 1.02806640e+00
-6.01615489e-01 3.44025224e-01 6.86542571e-01 -1.55424729e-01
-1.08907366e+00 -1.09182811e+00 -7.36255229e-01 1.07944739e+00
2.59863853e-01 -1.21676959e-01 -9.77091789e-01 -6.99402630e-01
-8.60425308e-02 4.04426396e-01 -7.94316351e-01 -2.65635014e-01
-4.58648086e-01 -7.78297782e-01 8.33943546e-01 5.06514609e-01
1.04082739e+00 -1.01941454e+00 -4.86658782e-01 -1.61342949e-01
-1.48367912e-01 -8.85748923e-01 -4.38391089e-01 7.99440622e-01
-1.16592491e+00 -8.88981581e-01 -4.97476876e-01 -9.47226048e-01
6.42429650e-01 1.68182626e-01 9.99270678e-01 -2.18785211e-01
-2.14455858e-01 4.50795263e-01 -1.41602248e-01 -1.19097568e-01
-4.15625796e-02 4.05338764e-01 1.86717317e-01 -3.74917567e-01
5.94449937e-01 -7.96167612e-01 -8.50749373e-01 4.64073390e-01
-1.16070855e+00 -5.69449067e-01 6.82142437e-01 9.73809719e-01
6.48859084e-01 2.10468322e-01 5.14905155e-01 -7.95875251e-01
2.78309435e-01 -4.84317392e-01 -5.89788556e-01 -2.32437942e-02
-7.54931867e-01 8.53944838e-01 7.18729854e-01 -3.02109241e-01
-6.51583552e-01 -2.23568127e-01 -2.90345818e-01 -3.60026538e-01
2.85335839e-01 4.51317787e-01 -6.86428919e-02 -5.04787266e-01
7.58242309e-01 3.85234237e-01 1.74377888e-01 -2.89354503e-01
2.27489114e-01 6.31119549e-01 3.24525267e-01 -3.04720283e-01
5.22392809e-01 6.19307697e-01 5.90934396e-01 -9.07115459e-01
-8.52541327e-01 -4.73059446e-01 -6.87394083e-01 1.27804101e-01
6.63684785e-01 -5.39988339e-01 -1.24388123e+00 4.04507905e-01
-8.83650422e-01 -2.36006647e-01 -1.24179095e-01 6.74362600e-01
-8.26651752e-01 4.11484778e-01 -6.30599439e-01 -1.03705430e+00
-5.63936412e-01 -8.98113608e-01 5.56129634e-01 -6.59106765e-03
3.57903361e-01 -1.03140426e+00 3.68378535e-02 2.71798912e-02
4.65520114e-01 -2.17301935e-01 9.24470603e-01 -7.90684879e-01
-4.45759863e-01 -1.78495273e-01 -5.47511458e-01 5.46754718e-01
-2.23775372e-01 -3.17659140e-01 -8.31850052e-01 -4.10624892e-01
4.26932156e-01 -1.77152306e-01 1.20753860e+00 7.19131529e-01
1.36450887e+00 -6.48370683e-01 -1.93880305e-01 9.83439684e-01
1.66239929e+00 -1.01017170e-02 6.93427145e-01 3.59411612e-02
5.13935268e-01 3.70904922e-01 2.06819043e-01 5.27382910e-01
2.53984094e-01 2.30520070e-01 4.90213364e-01 5.25119543e-01
1.34719566e-01 -1.27424151e-01 3.23803127e-01 1.11360276e+00
5.37098870e-02 -1.84172705e-01 -6.42116308e-01 4.88595814e-01
-1.56013489e+00 -1.05032492e+00 7.28237629e-02 2.51541305e+00
9.70448673e-01 1.85158253e-01 -3.25980447e-02 3.38099867e-01
7.71734774e-01 7.87680373e-02 -6.89491034e-01 -4.81868237e-01
-3.61310430e-02 5.09313762e-01 1.20384109e+00 4.36387837e-01
-9.93897259e-01 4.83314067e-01 6.07395840e+00 1.06722116e+00
-7.36923933e-01 2.13020220e-01 6.70711815e-01 1.30280433e-02
-2.73800284e-01 -2.50202298e-01 -1.09901023e+00 3.62831950e-01
1.29590833e+00 2.02527195e-02 4.99270439e-01 9.01493669e-01
-5.77319622e-01 -2.03198418e-02 -1.06216669e+00 1.34965348e+00
1.90472588e-01 -1.16223896e+00 -1.40978992e-01 4.49524552e-01
6.21411562e-01 3.55691820e-01 4.74168301e-01 6.32011816e-02
2.15243131e-01 -1.11770880e+00 2.13803336e-01 2.88099885e-01
9.56640720e-01 -1.27886903e+00 9.02892709e-01 4.12362874e-01
-1.19690752e+00 -1.62267551e-01 -9.25753415e-01 2.85732448e-01
-2.50180006e-01 7.43781507e-01 -5.15384138e-01 9.66297090e-02
8.89620364e-01 5.65386593e-01 -2.90205777e-01 9.20461953e-01
1.15371853e-01 4.20438737e-01 -7.11561620e-01 -4.61282939e-01
3.10876817e-02 -1.77163392e-01 5.02691269e-01 1.12529790e+00
3.91031116e-01 -5.99381141e-02 -7.72916794e-01 5.96935749e-01
-3.71263027e-01 1.74037382e-01 -8.50586474e-01 1.18209369e-01
6.26124561e-01 8.66075277e-01 -6.62458897e-01 7.59708732e-02
-1.91459954e-01 9.42272246e-01 6.83259368e-01 3.92346680e-02
-7.16368496e-01 -7.46891201e-01 5.36201775e-01 -5.24890646e-02
8.53386223e-01 -4.36650962e-01 -3.24973851e-01 -9.02177393e-01
-7.08252937e-02 -4.15275007e-01 5.05129158e-01 -3.21846716e-02
-1.24499810e+00 4.00582463e-01 1.97930366e-01 -9.78064477e-01
2.54181605e-02 -7.98730075e-01 5.22838868e-02 4.49184537e-01
-1.35232151e+00 -1.79935798e-01 4.55492474e-02 5.97257316e-01
-2.12720841e-01 -1.32156640e-01 7.68161833e-01 3.26664835e-01
-5.10624111e-01 1.20289588e+00 7.16848850e-01 2.49393135e-01
4.40327823e-01 -1.25785911e+00 1.65495962e-01 3.03729624e-01
2.29061961e-01 5.81111848e-01 5.98960638e-01 -2.49414220e-01
-1.46452105e+00 -7.56227493e-01 7.64557123e-01 -7.60006756e-02
5.12946069e-01 -3.04530650e-01 -9.89065886e-01 5.75475633e-01
-2.01594830e-01 3.25203598e-01 8.66709173e-01 -1.09925613e-01
-6.91609681e-01 -4.75834936e-01 -1.54967332e+00 3.85605901e-01
1.25545669e+00 -7.36071110e-01 -1.35181084e-01 1.62314847e-01
4.72575158e-01 -3.56752127e-02 -1.14959979e+00 3.92396390e-01
7.33782172e-01 -1.31986129e+00 1.01962078e+00 -1.54740065e-01
2.38404155e-01 2.12723851e-01 -5.92824161e-01 -8.62954974e-01
-8.65126103e-02 -5.68493247e-01 -3.14423293e-01 7.34040976e-01
3.32292706e-01 -9.11706269e-01 8.26158285e-01 3.59997720e-01
2.46762156e-01 -8.36667597e-01 -1.62492168e+00 -8.59205544e-01
3.88201058e-01 -4.41493094e-01 1.52134925e-01 5.18341422e-01
9.38275978e-02 2.01375529e-01 -2.70299856e-02 3.00094625e-03
1.11505985e+00 -3.68128479e-01 3.70689899e-01 -1.31187952e+00
-3.08244228e-01 -4.98636425e-01 -9.31454837e-01 -1.34868610e+00
2.09164932e-01 -9.96192694e-01 1.44818902e-01 -1.02272213e+00
4.80713993e-01 -4.35771614e-01 -5.84844291e-01 -8.86599794e-02
1.79201588e-01 1.25161722e-01 -3.14501464e-01 2.45425507e-01
-6.10242546e-01 7.27812588e-01 1.09084225e+00 1.18862234e-01
5.21794409e-02 -1.40745446e-01 -6.56456053e-01 6.60895288e-01
9.07160759e-01 -8.09833169e-01 -3.51239622e-01 -2.24234253e-01
5.74397802e-01 3.43096964e-02 1.36842251e-01 -1.20568538e+00
4.34217542e-01 3.81969899e-01 1.69787079e-01 -6.36674047e-01
1.06513523e-01 -7.42253423e-01 -2.86941200e-01 5.15996575e-01
-4.94843513e-01 1.58305764e-01 -3.23789902e-02 6.83288217e-01
-6.68721795e-02 -6.52822495e-01 1.01657593e+00 2.31469441e-02
-2.07963333e-01 7.17698276e-01 -9.79928002e-02 3.49955738e-01
6.11429036e-01 -7.85763841e-03 -4.73071456e-01 -5.40143430e-01
-1.83498293e-01 -4.00934309e-01 8.26439410e-02 -4.15134549e-01
8.15250754e-01 -1.25851798e+00 -5.14114380e-01 2.78280348e-01
-2.01771051e-01 3.42133008e-02 2.51195699e-01 1.01741982e+00
-6.94906831e-01 4.37334865e-01 -8.63317028e-02 -6.73662722e-01
-9.69481111e-01 5.50312102e-01 3.41256976e-01 -2.95453131e-01
-4.99706656e-01 1.29659390e+00 9.14335251e-02 -1.35567024e-01
5.26111901e-01 -3.17304283e-02 -3.50274257e-02 -3.18533033e-02
5.38223922e-01 6.13347173e-01 3.99673656e-02 -6.42759621e-01
-4.68365759e-01 6.10647738e-01 -2.78837800e-01 -3.51385534e-01
1.28215671e+00 -1.51939616e-01 -2.70226657e-01 4.68671501e-01
2.22506356e+00 -4.18435812e-01 -1.23301339e+00 -3.41406018e-01
-2.31595054e-01 -1.82275265e-01 2.28239730e-01 -1.59261525e-02
-1.37277865e+00 1.10020578e+00 9.35005546e-01 3.13246310e-01
1.09952521e+00 1.93642229e-01 7.80654371e-01 7.13667393e-01
4.83472973e-01 -9.57058430e-01 9.54738408e-02 5.33113122e-01
6.36739671e-01 -1.05612409e+00 3.66142802e-02 9.25492793e-02
-2.71497756e-01 9.83464539e-01 2.23341435e-01 -4.45898920e-01
1.38296950e+00 3.45853418e-01 -5.25605321e-01 -2.03242391e-01
-5.37906051e-01 -6.85761496e-02 2.25458503e-01 5.65375745e-01
3.39283615e-01 -1.15008697e-01 -5.02015769e-01 2.38732770e-01
-3.75309557e-01 -2.35964730e-01 1.12668455e-01 6.25376284e-01
-6.81978285e-01 -8.22425246e-01 1.52726442e-01 5.81891596e-01
-6.41249955e-01 -4.88802582e-01 1.41334504e-01 6.98723853e-01
-2.34124169e-01 6.86407387e-01 3.12301666e-01 -4.12717193e-01
-3.28107029e-01 -1.81954317e-02 7.99095571e-01 9.81738642e-02
-1.31138608e-01 -6.24958098e-01 -4.41507101e-01 -4.05037433e-01
-4.30085033e-01 -5.17763197e-01 -1.31143570e+00 -7.45950699e-01
-4.64736938e-01 1.08132705e-01 9.48750615e-01 9.23123062e-01
3.70558977e-01 -6.25468716e-02 6.98677778e-01 -5.86020350e-01
-6.61784053e-01 -9.30287480e-01 -9.27185297e-01 2.47878864e-01
4.92655158e-01 -6.96828961e-01 -1.08753514e+00 -4.27552730e-01] | [7.815914154052734, 3.675879955291748] |
94afdbe6-6bc1-47dc-a3ab-81d722cac5ff | very-lightweight-photo-retouching-network | 2104.06279 | null | https://arxiv.org/abs/2104.06279v2 | https://arxiv.org/pdf/2104.06279v2.pdf | Very Lightweight Photo Retouching Network with Conditional Sequential Modulation | Photo retouching aims at improving the aesthetic visual quality of images that suffer from photographic defects, especially for poor contrast, over/under exposure, and inharmonious saturation. In practice, photo retouching can be accomplished by a series of image processing operations. As most commonly-used retouching operations are pixel-independent, i.e., the manipulation on one pixel is uncorrelated with its neighboring pixels, we can take advantage of this property and design a specialized algorithm for efficient global photo retouching. We analyze these global operations and find that they can be mathematically formulated by a Multi-Layer Perceptron (MLP). Based on this observation, we propose an extremely lightweight framework -- Conditional Sequential Retouching Network (CSRNet). Benefiting from the utilization of $1\times1$ convolution, CSRNet only contains less than 37K trainable parameters, which are orders of magnitude smaller than existing learning-based methods. Experiments show that our method achieves state-of-the-art performance on the benchmark MIT-Adobe FiveK dataset quantitively and qualitatively. In addition to achieve global photo retouching, the proposed framework can be easily extended to learn local enhancement effects. The extended model, namely CSRNet-L, also achieves competitive results in various local enhancement tasks. Codes are available at https://github.com/lyh-18/CSRNet. | ['Yu Qiao', 'Chao Dong', 'Hengyuan Zhao', 'Zhengwen Zhang', 'Xiangyu Chen', 'Jingwen He', 'Yihao Liu'] | 2021-04-13 | null | null | null | null | ['photo-retouching', 'image-retouching'] | ['computer-vision', 'computer-vision'] | [ 4.17227477e-01 -1.53114110e-01 -7.03505874e-02 -7.45303556e-02
-5.09580016e-01 -1.61520541e-01 2.20153570e-01 -6.82459325e-02
-5.80109775e-01 6.73439980e-01 -5.75694963e-02 -3.57810348e-01
2.27723122e-01 -7.15753436e-01 -1.07938623e+00 -9.19078350e-01
2.29132652e-01 -6.38145566e-01 2.55576938e-01 -1.10705331e-01
2.82950133e-01 4.56863552e-01 -1.54236078e+00 3.08381498e-01
8.26081753e-01 9.84881997e-01 3.96257102e-01 7.98403561e-01
2.89165378e-01 8.88980627e-01 -3.69278103e-01 -4.29587543e-01
3.16983759e-01 -3.33802521e-01 -6.89894378e-01 5.73183373e-02
5.90327203e-01 -6.80018544e-01 -6.37925148e-01 1.11596572e+00
7.04749942e-01 2.84413993e-01 3.46640497e-01 -8.99766088e-01
-1.13358593e+00 3.99373591e-01 -9.90055084e-01 3.36832136e-01
-1.16472980e-02 3.07718724e-01 9.47070122e-01 -1.00893068e+00
2.88821399e-01 1.00992262e+00 6.27285361e-01 4.64670002e-01
-1.13557947e+00 -7.14925945e-01 3.79084386e-02 4.46570456e-01
-1.40116239e+00 -3.76359493e-01 8.09297860e-01 -1.71498898e-02
9.30345178e-01 3.84880126e-01 5.03091335e-01 7.86880195e-01
4.50222671e-01 9.41224396e-01 1.32908762e+00 -4.91791934e-01
-1.44373532e-02 9.62187871e-02 -1.95278823e-01 6.65706336e-01
6.03782013e-02 3.41536701e-01 -4.68280405e-01 3.76687407e-01
9.93403971e-01 2.96849459e-01 -5.15142798e-01 -5.07741133e-05
-9.97851372e-01 3.38873476e-01 9.30071890e-01 6.94781095e-02
-1.29070967e-01 4.42462832e-01 2.56082982e-01 1.79177567e-01
3.46768945e-01 2.12394878e-01 -4.38925236e-01 8.06157067e-02
-7.45603681e-01 -1.44342676e-01 2.55639553e-01 7.32134402e-01
1.21854794e+00 -3.88411172e-02 -1.31386250e-01 9.77234721e-01
-6.20707944e-02 3.84383589e-01 5.38683534e-01 -8.83224249e-01
2.58540630e-01 4.52220410e-01 -4.13580909e-02 -9.38713968e-01
-3.66480500e-01 -4.67342257e-01 -1.19254863e+00 6.03826880e-01
1.28648922e-01 -2.18920946e-01 -9.76454318e-01 1.38178504e+00
1.82857379e-01 3.60469639e-01 -5.23970984e-02 8.77555966e-01
7.96616256e-01 7.52367496e-01 1.80825725e-01 -1.54122636e-01
1.26771307e+00 -1.24311066e+00 -5.75934470e-01 -1.76742107e-01
4.35994685e-01 -7.96485841e-01 1.32576215e+00 4.06346142e-01
-1.23834980e+00 -7.17133820e-01 -1.35121250e+00 -4.28681076e-01
-3.77219051e-01 3.98509026e-01 5.85162699e-01 5.25926054e-01
-1.24142945e+00 7.06374586e-01 -6.48224950e-01 -7.82867894e-02
4.90098089e-01 3.32636446e-01 -3.29518855e-01 -1.02113195e-01
-9.06858027e-01 8.38727415e-01 4.19809729e-01 3.45176727e-01
-6.57680333e-01 -7.67739892e-01 -6.98143005e-01 2.03723863e-01
4.27963883e-01 -5.45017302e-01 1.24810004e+00 -1.13683522e+00
-1.67922449e+00 7.32818902e-01 -7.08689690e-02 -3.78951609e-01
4.43272650e-01 -3.07194173e-01 -5.30479610e-01 2.56219804e-01
-4.06612277e-01 6.67454600e-01 1.03443623e+00 -1.23004222e+00
-6.95030689e-01 -4.07050289e-02 2.16005504e-01 3.06338370e-01
-7.11171508e-01 -1.05614156e-01 -6.37703001e-01 -6.85576856e-01
-1.83287233e-01 -7.95628369e-01 -2.05982238e-01 2.68498361e-01
-4.24563080e-01 1.53704047e-01 8.45979571e-01 -7.49380469e-01
1.25111628e+00 -2.36511421e+00 -2.69062370e-01 -2.09916219e-01
4.33222145e-01 6.28128290e-01 -2.20582977e-01 2.86351353e-01
-3.27031881e-01 1.36256620e-01 -2.96361625e-01 -2.77397066e-01
-2.03393549e-01 -1.00131162e-01 -1.78245097e-01 5.54999411e-01
1.30494609e-01 1.14899766e+00 -7.21306503e-01 -4.05263543e-01
5.43378115e-01 5.51632285e-01 -4.37106073e-01 -3.35287228e-02
1.20896071e-01 9.95389000e-02 -1.21402211e-01 6.15070701e-01
1.01145351e+00 -3.34272414e-01 -1.29775837e-01 -4.68075663e-01
-2.57202029e-01 -1.64968938e-01 -9.91825521e-01 1.56834781e+00
-7.19970405e-01 8.14207613e-01 -2.78686315e-01 -7.35586941e-01
7.24871755e-01 3.29298042e-02 1.64147809e-01 -1.02593589e+00
2.47626841e-01 2.20224440e-01 -3.51207584e-01 -4.80058223e-01
6.84584916e-01 -1.27225816e-01 3.01644742e-01 4.62892532e-01
-2.38760840e-03 2.74991661e-01 -2.21997709e-03 6.82565244e-03
1.12766778e+00 -7.09427372e-02 4.78427857e-01 1.20633155e-01
4.94436651e-01 -4.99424785e-01 4.77904439e-01 6.42577827e-01
-3.42280477e-01 6.71582520e-01 1.20419651e-01 -4.13731188e-01
-1.06183696e+00 -1.06270695e+00 -1.42937779e-01 1.03925729e+00
5.36004186e-01 -1.89981014e-01 -6.28070831e-01 -4.72849876e-01
-2.78986126e-01 4.54506189e-01 -5.74886262e-01 -2.94650704e-01
-5.44795156e-01 -7.83379138e-01 4.23706979e-01 5.93770504e-01
1.05012393e+00 -1.18127227e+00 -5.28632879e-01 -3.12457904e-02
1.08577833e-01 -1.05339146e+00 -6.68997347e-01 4.40614447e-02
-7.57567763e-01 -8.89284670e-01 -9.09285843e-01 -9.09833729e-01
6.84095204e-01 8.36407483e-01 8.21225584e-01 2.78014868e-01
-4.02403206e-01 2.20081747e-01 -3.38325232e-01 -1.24021202e-01
-2.59421289e-01 -2.11456269e-02 -2.96264082e-01 9.09798741e-02
1.82033315e-01 -8.11020434e-01 -1.11601949e+00 2.21222296e-01
-1.30636132e+00 3.54294151e-01 1.18509817e+00 8.67200971e-01
7.00196207e-01 5.09953022e-01 3.17907810e-01 -7.22414136e-01
4.22982275e-01 -2.25734636e-02 -4.79439557e-01 3.87133688e-01
-5.16424417e-01 -1.11095428e-01 6.38886988e-01 -5.23015738e-01
-1.33163846e+00 8.82734824e-03 -2.29123682e-02 -4.94315267e-01
-7.41579682e-02 2.62131304e-01 -8.11922699e-02 -6.61804855e-01
5.92875540e-01 3.76234204e-01 -2.69871086e-01 -4.35188979e-01
5.77592015e-01 6.96291566e-01 7.38819301e-01 -1.25393674e-01
8.83869171e-01 6.24802589e-01 -1.35283351e-01 -8.53556693e-01
-5.60449541e-01 -3.57986391e-01 -3.12780976e-01 -3.21437776e-01
6.24460399e-01 -9.87588584e-01 -9.95424867e-01 8.33665967e-01
-9.72952247e-01 -6.02748573e-01 -1.75781861e-01 2.44402498e-01
-2.68304378e-01 5.24626553e-01 -8.16535950e-01 -6.88170135e-01
-5.01147866e-01 -9.71174359e-01 8.09746683e-01 6.78510010e-01
3.15439880e-01 -9.31844234e-01 -1.89638138e-01 1.77446187e-01
2.88959831e-01 9.12535284e-03 6.65332437e-01 4.25520390e-02
-7.57524252e-01 -1.44943655e-01 -6.88369215e-01 8.93956900e-01
8.40228125e-02 -8.37898105e-02 -1.11460209e+00 -3.63665313e-01
-1.54740199e-01 -2.73044258e-01 1.19204867e+00 4.22703981e-01
1.58587015e+00 -3.49709988e-01 -1.19975850e-01 8.91803384e-01
1.74105501e+00 1.65283725e-01 1.10928845e+00 3.61072719e-01
7.04021275e-01 1.69484928e-01 3.90861452e-01 4.07520115e-01
2.69933820e-01 4.18521762e-01 4.54489827e-01 -6.76255405e-01
-5.53891838e-01 -3.10656995e-01 3.58025968e-01 6.64875209e-01
-8.14016759e-02 -2.76825964e-01 -4.74921703e-01 5.04071653e-01
-1.58411562e+00 -7.71479189e-01 1.38535611e-02 2.16390395e+00
8.66803944e-01 1.30318087e-02 -2.38070786e-01 1.63469151e-01
8.76137793e-01 3.18720162e-01 -6.47626996e-01 -4.73957658e-01
-3.30034018e-01 3.53903294e-01 1.03705859e+00 3.47614467e-01
-1.05635858e+00 9.43486273e-01 6.12497711e+00 1.18257487e+00
-1.24105942e+00 1.77206203e-01 9.26154494e-01 -5.97844087e-02
-1.86955228e-01 -6.86162114e-02 -5.58025539e-01 4.81023699e-01
4.96536881e-01 3.67997191e-03 6.15745664e-01 6.39310420e-01
2.91174024e-01 -2.91185766e-01 -7.15844631e-01 1.17669570e+00
9.65287313e-02 -1.40788531e+00 -9.13684294e-02 -1.26119882e-01
9.72930193e-01 -4.60749269e-02 4.99985069e-01 7.65206069e-02
3.04643437e-02 -9.01367068e-01 4.57024664e-01 3.09102327e-01
1.02221215e+00 -5.83251178e-01 5.40124714e-01 7.81002687e-03
-1.15171194e+00 -2.55688310e-01 -5.04268706e-01 -5.44045493e-02
-2.75840256e-02 7.99989462e-01 -3.84056389e-01 4.51499254e-01
9.36892509e-01 7.68985629e-01 -6.27398729e-01 1.29112041e+00
-5.50016105e-01 3.99626374e-01 -8.01087245e-02 3.81225459e-02
6.74223155e-03 8.10883120e-02 3.66778493e-01 1.20252323e+00
3.71627808e-01 1.80166632e-01 -3.62091839e-01 6.34900212e-01
-4.51031804e-01 9.34456885e-02 -2.01570526e-01 1.58580482e-01
2.36334130e-01 1.48274612e+00 -5.50120175e-01 -2.51321465e-01
-5.50520480e-01 1.38677979e+00 2.42481604e-01 6.94859684e-01
-8.39261115e-01 -6.40046418e-01 4.10106868e-01 -3.06078158e-02
4.28416491e-01 -1.03155017e-01 -4.37780499e-01 -1.07243657e+00
2.25186571e-01 -6.99319780e-01 2.01365143e-01 -1.13067901e+00
-1.17975485e+00 5.62897861e-01 -4.00280774e-01 -1.25728273e+00
5.79997122e-01 -7.42171407e-01 -7.08256960e-01 7.49466479e-01
-1.93642282e+00 -1.16940582e+00 -6.25551820e-01 6.59031212e-01
3.26492518e-01 1.88848689e-01 3.62216890e-01 4.93256241e-01
-8.16364825e-01 7.66180813e-01 8.90607759e-02 -9.13826227e-02
9.07818377e-01 -1.13651204e+00 2.55017400e-01 1.18711698e+00
-1.32251000e-02 3.07353258e-01 5.74745834e-01 -2.73524523e-01
-1.18597090e+00 -1.19793832e+00 5.98370492e-01 1.34228915e-01
6.44732356e-01 -1.00961141e-01 -9.15653944e-01 5.29020727e-01
3.87890309e-01 1.76154777e-01 4.91739780e-01 -2.46771589e-01
-3.27859074e-01 -5.07229090e-01 -8.96911204e-01 8.58639836e-01
9.97496843e-01 -5.28888047e-01 -9.62557197e-02 1.78568333e-01
8.79674733e-01 -4.94773388e-01 -6.70744777e-01 1.99236646e-01
5.44626951e-01 -1.04111755e+00 1.06380427e+00 -1.13464303e-01
7.84446537e-01 -2.53772616e-01 1.76445954e-02 -1.23334491e+00
-4.11085814e-01 -8.40457797e-01 -9.28309485e-02 1.11904907e+00
2.52258867e-01 -8.34985971e-01 5.71325362e-01 4.50892955e-01
-2.15904951e-01 -7.62715936e-01 -7.53512561e-01 -6.49964154e-01
-9.34116915e-02 -2.40280256e-01 5.19137561e-01 6.04841471e-01
-2.03249142e-01 1.00835316e-01 -7.67630935e-01 2.99857527e-01
4.72592384e-01 -1.39921919e-01 5.96016169e-01 -4.98506218e-01
-4.95261818e-01 -4.99337018e-01 -3.39136332e-01 -1.28022182e+00
-7.12165982e-02 -5.84223270e-01 1.59712002e-01 -1.27799213e+00
3.56116980e-01 -3.30855042e-01 -6.71446741e-01 7.32304513e-01
-4.93420720e-01 8.34766388e-01 2.24648625e-01 2.03717232e-01
-5.07476866e-01 6.53067470e-01 1.60684741e+00 -2.04409286e-01
-3.20186406e-01 -2.09757000e-01 -8.87910187e-01 6.15878522e-01
9.47627008e-01 -1.50984645e-01 -3.11785907e-01 -6.03676796e-01
2.96233177e-01 -2.87821949e-01 5.88748693e-01 -1.12999904e+00
3.63584101e-01 -1.67933837e-01 4.02769148e-01 -3.02951038e-01
3.54247302e-01 -5.81014037e-01 -7.25571066e-02 3.59217346e-01
-2.81593472e-01 -1.78597480e-01 4.06990170e-01 5.74831009e-01
-1.62929416e-01 -1.33177340e-01 9.33892608e-01 5.23032472e-02
-1.10805225e+00 4.22078252e-01 -2.19896927e-01 -3.63760322e-01
1.01499271e+00 -2.79501498e-01 -6.98761821e-01 -2.39714622e-01
-2.96807557e-01 -1.77774087e-01 4.26152706e-01 1.41330928e-01
8.20652485e-01 -1.19478071e+00 -5.15165687e-01 8.27532262e-02
-9.84708313e-03 -2.72969514e-01 9.12503183e-01 9.98514116e-01
-6.96923554e-01 7.78679401e-02 -2.79936850e-01 -2.51652598e-01
-1.28846061e+00 7.06440032e-01 2.97146142e-01 -9.57708061e-02
-8.40607941e-01 9.49645579e-01 5.55853188e-01 1.81849197e-01
1.30753256e-02 -2.73501366e-01 6.05779774e-02 -4.49117631e-01
7.69770801e-01 5.12466073e-01 1.22572929e-01 -3.19150686e-01
-8.46430734e-02 5.87665796e-01 -2.52667248e-01 1.34486109e-01
1.26700962e+00 -3.54945272e-01 -1.23662725e-01 1.61152393e-01
1.48156345e+00 -8.58148634e-02 -1.51871002e+00 -3.52556169e-01
-6.27344072e-01 -6.10450864e-01 4.27667350e-01 -8.36437583e-01
-1.29054642e+00 8.87736022e-01 8.33774328e-01 3.59246433e-02
1.84002936e+00 -2.53428489e-01 1.08370316e+00 2.88529903e-01
2.45521832e-02 -1.16033351e+00 3.00388783e-01 -1.45734753e-02
8.36329997e-01 -1.19032693e+00 1.10480957e-01 -4.26256746e-01
-5.57233214e-01 1.03227532e+00 5.92845917e-01 -2.23009720e-01
4.89720196e-01 1.46938011e-01 6.25804514e-02 5.50497919e-02
-5.95951080e-01 -1.50017157e-01 2.41073504e-01 4.28734928e-01
1.59363031e-01 1.02233693e-01 -2.33961642e-01 2.45342538e-01
1.30760729e-01 8.12632218e-02 7.01319039e-01 7.19675541e-01
-4.17950809e-01 -8.99453104e-01 -2.40446478e-01 3.14444453e-01
-3.63115758e-01 -3.37195426e-01 7.30716297e-03 6.91853940e-01
1.91905975e-01 8.51317346e-01 -6.44327030e-02 -4.92163301e-01
1.42682180e-01 -5.53131938e-01 4.43102390e-01 -1.73358098e-01
-3.73785019e-01 2.52866298e-02 -2.72060573e-01 -6.54697895e-01
-4.89953816e-01 -2.56744534e-01 -9.48421836e-01 -6.11775339e-01
-3.44595194e-01 -4.12050605e-01 5.87871432e-01 6.90953135e-01
2.94992328e-01 8.52160573e-01 6.33608401e-01 -8.52603912e-01
-1.96556598e-01 -8.71479630e-01 -8.58502746e-01 1.55870304e-01
4.03250426e-01 -4.48175997e-01 -4.11327571e-01 1.62872642e-01] | [10.91137409210205, -2.122277021408081] |
3885366f-d32c-42a6-aa5e-6a71c3cbf397 | ssul-semantic-segmentation-with-unknown-label | 2106.11562 | null | https://arxiv.org/abs/2106.11562v3 | https://arxiv.org/pdf/2106.11562v3.pdf | SSUL: Semantic Segmentation with Unknown Label for Exemplar-based Class-Incremental Learning | This paper introduces a solid state-of-the-art baseline for a class-incremental semantic segmentation (CISS) problem. While the recent CISS algorithms utilize variants of the knowledge distillation (KD) technique to tackle the problem, they failed to fully address the critical challenges in CISS causing the catastrophic forgetting; the semantic drift of the background class and the multi-label prediction issue. To better address these challenges, we propose a new method, dubbed SSUL-M (Semantic Segmentation with Unknown Label with Memory), by carefully combining techniques tailored for semantic segmentation. Specifically, we claim three main contributions. (1) defining unknown classes within the background class to help to learn future classes (help plasticity), (2) freezing backbone network and past classifiers with binary cross-entropy loss and pseudo-labeling to overcome catastrophic forgetting (help stability), and (3) utilizing tiny exemplar memory for the first time in CISS to improve both plasticity and stability. The extensively conducted experiments show the effectiveness of our method, achieving significantly better performance than the recent state-of-the-art baselines on the standard benchmark datasets. Furthermore, we justify our contributions with thorough ablation analyses and discuss different natures of the CISS problem compared to the traditional class-incremental learning targeting classification. The official code is available at https://github.com/clovaai/SSUL. | ['Beomyoung Kim', 'Sungmin Cha', 'Taesup Moon', 'Youngjoon Yoo'] | 2021-06-22 | null | http://proceedings.neurips.cc/paper/2021/hash/5a9542c773018268fc6271f7afeea969-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/5a9542c773018268fc6271f7afeea969-Paper.pdf | neurips-2021-12 | ['overlapped-100-5', 'overlapped-10-1', 'disjoint-15-5', 'disjoint-10-1', 'disjoint-15-1', 'overlapped-15-5', 'class-incremental-semantic-segmentation', 'overlapped-15-1', 'overlapped-50-50', 'overlapped-100-50', 'continual-semantic-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 5.79309642e-01 3.28283370e-01 -3.77880692e-01 -3.76113027e-01
-7.57384837e-01 -5.73996246e-01 3.84532124e-01 1.67834148e-01
-4.07784373e-01 8.91424596e-01 -2.72118092e-01 -1.43385842e-01
-5.02300411e-02 -4.96074528e-01 -1.05115950e+00 -7.45181561e-01
2.25073814e-01 6.39854133e-01 8.84738505e-01 1.92943200e-01
3.95117491e-01 2.12240338e-01 -1.72702301e+00 2.72193849e-01
1.28361750e+00 1.12069345e+00 7.57364333e-02 1.94794774e-01
-1.80319846e-01 7.67491102e-01 -4.20336306e-01 -4.72661138e-01
1.50000498e-01 -3.40919077e-01 -1.27335215e+00 -1.06530838e-01
5.62133551e-01 -4.89817932e-02 -1.04547366e-01 1.09679091e+00
3.73277158e-01 3.53147417e-01 5.73002100e-01 -1.40174007e+00
-5.99170864e-01 7.57001817e-01 -5.65832257e-01 1.66911215e-01
-2.81746507e-01 1.36011198e-01 6.52618408e-01 -8.48396897e-01
8.11578035e-01 1.10827279e+00 1.11281502e+00 9.44000661e-01
-1.19324374e+00 -6.92566752e-01 5.90252638e-01 4.88781661e-01
-1.33694863e+00 -3.69129330e-01 5.47719598e-01 -3.40427399e-01
8.54140282e-01 1.41674921e-01 7.47036636e-01 1.31700850e+00
-1.70873597e-01 1.46918023e+00 1.39697218e+00 -3.59051585e-01
5.41672051e-01 2.03948632e-01 8.67856562e-01 8.29563677e-01
1.30365297e-01 2.66974363e-02 -7.53583372e-01 -6.13169409e-02
1.98849455e-01 -1.42607130e-02 -1.83001205e-01 -5.47323763e-01
-9.58060026e-01 6.27891600e-01 3.44676584e-01 2.97060937e-01
1.90554187e-02 1.68254480e-01 4.19059813e-01 1.26654636e-02
7.32159257e-01 2.09919915e-01 -9.08691287e-01 -9.79639143e-02
-1.38917017e+00 1.18688107e-01 5.57835281e-01 1.01556921e+00
6.86959982e-01 -1.86059609e-01 -3.60029370e-01 7.89860964e-01
-3.47147658e-02 2.27647945e-01 7.07603574e-01 -1.01189101e+00
3.20892371e-02 5.90785682e-01 -9.26247835e-02 -3.04490298e-01
-5.06375313e-01 -7.33568668e-01 -5.70641160e-01 -1.21298641e-01
3.33923191e-01 2.17415318e-01 -1.59790671e+00 2.06177092e+00
4.71967906e-01 7.80764818e-01 3.59469093e-02 4.42747712e-01
5.81692696e-01 3.07943493e-01 3.86204451e-01 -1.84576064e-01
1.05935729e+00 -1.57309532e+00 -5.41527629e-01 -3.77594054e-01
5.31397521e-01 -2.75592923e-01 1.25768256e+00 2.40670711e-01
-9.96003270e-01 -4.40545708e-01 -1.14156544e+00 1.01496056e-01
-8.16543639e-01 -2.37915829e-01 7.79214740e-01 6.43928230e-01
-1.24955285e+00 9.54487681e-01 -1.07837105e+00 -4.31634158e-01
8.98796320e-01 2.49503821e-01 9.21251103e-02 -2.77917124e-02
-1.28065479e+00 8.09275985e-01 7.16084719e-01 -1.80208385e-01
-9.68557119e-01 -1.09308183e+00 -5.25075912e-01 -2.15459406e-01
6.14087403e-01 -9.59589422e-01 1.30244005e+00 -1.16998041e+00
-1.28288245e+00 1.05572975e+00 -1.95778579e-01 -7.13224769e-01
6.21527374e-01 -4.43069130e-01 -1.15371928e-01 1.49378091e-01
2.27761373e-01 1.00539994e+00 9.19943690e-01 -1.47744894e+00
-6.82952046e-01 -4.17295784e-01 -2.85221159e-01 1.69007391e-01
-3.44868124e-01 -6.71136141e-01 -6.69972301e-01 -8.95319223e-01
3.86690438e-01 -1.04308915e+00 5.30624092e-02 -3.02376866e-01
-5.43714821e-01 -2.39558592e-01 7.27982342e-01 -6.55784369e-01
1.39095283e+00 -2.11122179e+00 2.11487845e-01 -2.48594448e-01
4.38311249e-02 6.20397985e-01 -1.33261710e-01 -2.11582184e-01
-5.85135480e-04 1.51384532e-01 -8.41169059e-01 -6.79006517e-01
-1.06423400e-01 3.16386729e-01 -4.79879797e-01 8.71018991e-02
1.63062558e-01 1.15897536e+00 -1.10044241e+00 -4.49159950e-01
-1.49308011e-01 3.70538831e-01 -3.00711334e-01 -1.76893651e-01
-5.14280915e-01 4.85003054e-01 -2.36930907e-01 1.11282456e+00
6.99065924e-01 -3.35095823e-01 -1.79002494e-01 -1.15867972e-01
2.72785068e-01 6.05977178e-02 -9.71889734e-01 2.01194644e+00
-1.93629041e-02 2.33912040e-02 -3.16342056e-01 -1.04270113e+00
5.80236673e-01 -1.61257014e-01 3.67641598e-01 -7.99236059e-01
-1.60399333e-01 4.82228607e-01 -5.50679326e-01 -1.22720890e-01
5.54319143e-01 -1.53558835e-01 8.35733563e-02 2.90775895e-01
3.83638382e-01 7.33124763e-02 1.66307658e-01 4.72109169e-01
1.16950607e+00 6.24752700e-01 -1.74546331e-01 -3.65673333e-01
2.98670322e-01 6.50680363e-02 9.66381073e-01 9.79104578e-01
-7.24917650e-01 5.95213771e-01 2.49215052e-01 -3.09226334e-01
-6.91856265e-01 -1.33372200e+00 -1.73348874e-01 9.79895830e-01
4.34842855e-01 -8.50888267e-02 -1.08321571e+00 -1.13688159e+00
2.36085042e-01 1.07571793e+00 -8.04589272e-01 -6.27325892e-01
-4.10627306e-01 -1.16727734e+00 7.12402642e-01 6.72237158e-01
8.09665442e-01 -1.04532194e+00 -4.63066459e-01 1.69072598e-01
-4.05548900e-01 -9.13646042e-01 -3.49526614e-01 4.09463674e-01
-1.25229239e+00 -1.19006526e+00 -7.48513103e-01 -7.21458673e-01
5.21997213e-01 1.38255060e-01 1.15395761e+00 -1.27549693e-02
-3.35185170e-01 6.77801788e-01 -3.32660288e-01 -2.73220003e-01
-2.53352225e-01 4.89628166e-01 -4.65872958e-02 -2.70913858e-02
3.80003601e-01 -6.52274847e-01 -7.16039360e-01 8.62408206e-02
-8.13286483e-01 2.74075031e-01 4.69711840e-01 8.77309084e-01
1.01479137e+00 7.79027268e-02 8.66758466e-01 -1.37145567e+00
3.75886038e-02 -6.00674212e-01 -2.33975336e-01 5.95410109e-01
-1.34586716e+00 4.91728187e-02 1.57805219e-01 -4.90708143e-01
-1.16240990e+00 1.04133575e-03 6.54915944e-02 -4.18434292e-01
-9.13843289e-02 2.94929817e-02 1.07630648e-01 -5.75697683e-02
3.67095530e-01 3.91498208e-01 -2.82267570e-01 -6.72906578e-01
4.70550895e-01 3.76921535e-01 6.89540148e-01 -7.23938644e-01
3.87405396e-01 6.07520342e-01 -1.77201197e-01 -2.68328369e-01
-1.58477628e+00 -4.61060405e-01 -9.11295354e-01 -1.46160111e-01
5.97983062e-01 -8.97061586e-01 -2.90735573e-01 1.09686852e+00
-9.21416819e-01 -7.39451170e-01 -6.03546798e-01 -1.54945135e-01
-7.18946099e-01 2.81466007e-01 -7.83939362e-01 -7.41786242e-01
-5.42263448e-01 -8.21174622e-01 9.70774949e-01 4.82939661e-01
-4.48176041e-02 -9.80573118e-01 -5.24286963e-02 8.16971183e-01
4.39173311e-01 4.15742755e-01 1.00417483e+00 -6.15715802e-01
-6.14647031e-01 2.33366072e-01 -2.21536532e-01 4.26506460e-01
-2.45673358e-01 -3.12808514e-01 -1.16829574e+00 -4.20559198e-01
-1.73528925e-01 -5.25620699e-01 1.66894484e+00 3.15783858e-01
1.17718482e+00 1.75054781e-02 -6.17124856e-01 6.41561329e-01
1.37155235e+00 2.58361369e-01 5.43312490e-01 6.42048597e-01
7.29646504e-01 3.91232967e-01 8.43439221e-01 9.93212983e-02
5.62707186e-01 3.59188080e-01 2.38782138e-01 6.15033656e-02
-7.40350127e-01 -4.23848510e-01 3.21828723e-01 9.54329371e-01
3.85890722e-01 -1.01361386e-01 -9.63172197e-01 9.33989763e-01
-2.21684313e+00 -4.72939014e-01 1.42697627e-02 2.06939745e+00
1.17839813e+00 3.34320188e-01 -1.58785358e-01 1.05116501e-01
8.53057563e-01 5.36495335e-02 -1.34315109e+00 -7.83574954e-02
-3.79561841e-01 3.52924138e-01 4.93556231e-01 4.70006526e-01
-1.33447397e+00 1.48331845e+00 5.66057348e+00 1.19224751e+00
-7.34659135e-01 7.01292992e-01 9.98894572e-01 -1.09464556e-01
-1.38536319e-01 1.98961943e-01 -1.19617927e+00 6.21994317e-01
9.60051000e-01 2.09093511e-01 3.91296178e-01 7.17851579e-01
-5.36350191e-01 -4.65024590e-01 -9.04529154e-01 7.31656313e-01
1.85197294e-01 -1.13484693e+00 1.58606350e-01 -5.86232424e-01
1.00998390e+00 1.10963866e-01 3.15386802e-01 4.94913101e-01
2.08709866e-01 -5.44198632e-01 1.16640913e+00 6.90510511e-01
7.57927060e-01 -5.34155846e-01 4.57965165e-01 1.92961589e-01
-9.41243291e-01 -2.03362152e-01 -1.19356424e-01 3.70837361e-01
7.46577755e-02 9.37978983e-01 -3.91728252e-01 5.56132913e-01
9.37344432e-01 8.31093729e-01 -9.97226954e-01 8.48699331e-01
-2.77474195e-01 9.43369329e-01 -2.92603552e-01 4.83933240e-01
3.36494446e-01 1.35295197e-01 4.37540442e-01 1.13974297e+00
8.29869509e-02 -6.26086295e-02 -6.25049472e-02 9.48558092e-01
-1.32997096e-01 -2.21613407e-01 1.32366335e-02 -6.96918042e-03
6.55120730e-01 8.16324353e-01 -1.27157509e+00 -4.82280731e-01
4.48986627e-02 1.40292120e+00 5.31230390e-01 3.16220641e-01
-9.07333374e-01 -3.13727319e-01 3.95900369e-01 -5.82137369e-02
3.88420194e-01 1.85708329e-01 -6.80211961e-01 -1.07087374e+00
7.92288780e-02 -3.86485517e-01 7.20114648e-01 -5.84683120e-01
-1.23348439e+00 4.34160918e-01 -9.70205218e-02 -5.34469366e-01
3.70783895e-01 -2.09985077e-01 -1.95602715e-01 3.50886911e-01
-1.92595947e+00 -1.41698277e+00 -1.04871936e-01 5.98843515e-01
6.35790288e-01 -7.96265975e-02 7.20373690e-01 4.37776774e-01
-6.94853306e-01 7.34333754e-01 3.31304431e-01 -4.02571797e-01
8.20628703e-01 -1.33060241e+00 4.56800342e-01 8.16698670e-01
-1.31487213e-02 4.12036061e-01 3.99185658e-01 -9.44117785e-01
-7.87030935e-01 -1.33787453e+00 8.94280732e-01 -7.22073913e-01
5.57986557e-01 -2.74256766e-01 -1.10749876e+00 7.18265474e-01
-2.62996227e-01 -9.23430622e-02 4.84696060e-01 6.74408861e-03
-5.99342525e-01 -4.23510820e-02 -1.30783844e+00 4.34450209e-01
1.47428989e+00 -3.03643197e-01 -6.16098940e-01 4.10739422e-01
1.13213944e+00 -3.16886008e-01 -5.16004324e-01 7.46861041e-01
3.18688452e-01 -1.07579279e+00 9.47150350e-01 -3.83479089e-01
9.91012082e-02 -2.19655037e-01 9.76311974e-03 -9.84451711e-01
-2.17681512e-01 -4.07855511e-01 -5.18462837e-01 1.57396507e+00
3.75708133e-01 -7.12195098e-01 9.41712260e-01 5.16217649e-01
-3.19382071e-01 -9.81048465e-01 -1.12253964e+00 -1.02555931e+00
3.33532900e-01 -2.88087368e-01 4.71287936e-01 1.13680136e+00
-4.17344511e-01 1.77903175e-01 -2.01266035e-01 -1.96793769e-02
7.73948014e-01 2.33077630e-01 7.72545338e-02 -1.25902510e+00
-1.49731681e-01 -4.25103247e-01 -1.04363933e-01 -8.11373651e-01
3.26444417e-01 -1.21801388e+00 8.59444812e-02 -1.38853288e+00
5.23617148e-01 -8.32843125e-01 -8.24654698e-01 9.17637527e-01
-4.35943067e-01 3.27607989e-01 2.38225624e-01 3.45357567e-01
-1.19610214e+00 6.42525613e-01 1.05431509e+00 -2.66117364e-01
-1.94120556e-01 -5.64140230e-02 -7.85664082e-01 6.77223206e-01
7.98965216e-01 -7.75620759e-01 -5.42221963e-01 -3.30981374e-01
7.92922676e-02 -4.69571233e-01 4.08316731e-01 -1.15032136e+00
2.77619362e-01 1.31454885e-01 6.93387091e-02 -6.54683113e-01
2.39468366e-01 -4.12608534e-01 6.66902065e-02 7.45652735e-01
-2.32926637e-01 -4.35989767e-01 4.16962683e-01 8.84435952e-01
2.78817024e-02 -3.26375872e-01 1.11088550e+00 -2.84027576e-01
-1.22622001e+00 3.24401289e-01 4.46598232e-02 4.02483076e-01
1.17412591e+00 -2.46826440e-01 -5.42666733e-01 3.07179928e-01
-9.72679257e-01 3.55543554e-01 5.63391805e-01 5.81527114e-01
4.06147540e-01 -1.05034256e+00 -2.70511091e-01 1.78468358e-02
2.15864601e-03 1.80839777e-01 5.25559604e-01 8.42931032e-01
-2.10886508e-01 2.75821149e-01 6.73845857e-02 -5.10210156e-01
-9.30541158e-01 6.34955406e-01 2.53641158e-01 -3.73224676e-01
-6.35973096e-01 1.34933519e+00 6.10576607e-02 -6.24239981e-01
5.78819811e-01 1.36682978e-02 1.41849414e-01 2.26216495e-01
3.21570456e-01 6.24159038e-01 2.04901129e-01 -2.98561484e-01
-5.15973806e-01 5.07605910e-01 -3.87063622e-01 1.28771558e-01
1.31768119e+00 -3.69678587e-01 -3.53104882e-02 7.96050787e-01
8.02550316e-01 -7.13581502e-01 -1.59185565e+00 -4.37855989e-01
4.11564916e-01 -9.06148255e-02 3.58982086e-02 -1.48808515e+00
-1.09681380e+00 8.37504685e-01 9.68897820e-01 -3.03183377e-01
1.16808498e+00 3.59740667e-02 1.38525248e+00 4.02804278e-02
5.46024203e-01 -1.57876611e+00 1.95709378e-01 6.97983921e-01
3.78000051e-01 -1.15924144e+00 -1.79395616e-01 -3.97336841e-01
-7.55656183e-01 7.34648466e-01 6.59021497e-01 3.18054199e-01
7.47730553e-01 1.47170022e-01 -3.38422018e-03 -7.08478391e-02
-6.02250338e-01 -2.46879652e-01 -3.55216600e-02 5.65795779e-01
-1.45288780e-01 3.09869722e-02 -2.78537333e-01 9.89610672e-01
1.88710243e-01 1.55759275e-01 1.66813418e-01 1.08102453e+00
-4.78809088e-01 -1.01696539e+00 -9.34789479e-02 4.05304074e-01
-3.09967399e-01 -2.66205847e-01 -4.16587710e-01 3.81339520e-01
5.73925674e-01 6.26401722e-01 -1.51847169e-01 -4.21849459e-01
2.77619481e-01 7.08858430e-01 5.56177974e-01 -4.40966725e-01
-5.37856102e-01 -3.18069398e-01 -3.31427783e-01 -7.25233734e-01
-5.15149653e-01 -8.60229731e-01 -1.42232704e+00 -1.54500484e-01
-3.53594571e-01 -7.27544427e-02 5.38402796e-01 8.97194982e-01
6.53269768e-01 6.68869793e-01 1.22236423e-01 -6.01344585e-01
-5.08004844e-01 -6.31805122e-01 -6.66381955e-01 5.64991474e-01
1.33648545e-01 -1.01630068e+00 -4.29364562e-01 1.62430093e-01] | [9.392409324645996, 2.242628812789917] |
34e2694f-fc2f-445a-891c-029143d49680 | mitos-rcnn-a-novel-approach-to-mitotic-figure | 1807.01788 | null | http://arxiv.org/abs/1807.01788v1 | http://arxiv.org/pdf/1807.01788v1.pdf | MITOS-RCNN: A Novel Approach to Mitotic Figure Detection in Breast Cancer Histopathology Images using Region Based Convolutional Neural Networks | Studies estimate that there will be 266,120 new cases of invasive breast
cancer and 40,920 breast cancer induced deaths in the year of 2018 alone.
Despite the pervasiveness of this affliction, the current process to obtain an
accurate breast cancer prognosis is tedious and time consuming, requiring a
trained pathologist to manually examine histopathological images in order to
identify the features that characterize various cancer severity levels. We
propose MITOS-RCNN: a novel region based convolutional neural network (RCNN)
geared for small object detection to accurately grade one of the three factors
that characterize tumor belligerence described by the Nottingham Grading
System: mitotic count. Other computational approaches to mitotic figure
counting and detection do not demonstrate ample recall or precision to be
clinically viable. Our models outperformed all previous participants in the
ICPR 2012 challenge, the AMIDA 2013 challenge and the MITOS-ATYPIA-14 challenge
along with recently published works. Our model achieved an F-measure score of
0.955, a 6.11% improvement in accuracy from the most accurate of the previously
proposed models. | ['Siddhant Rao'] | 2018-07-04 | null | null | null | null | ['small-object-detection'] | ['computer-vision'] | [ 3.23491126e-01 3.52213651e-01 -5.00620723e-01 -1.65022060e-01
-1.27533507e+00 -4.02241230e-01 5.56993306e-01 8.16472113e-01
-6.09550834e-01 7.34140694e-01 4.21577096e-02 -7.84070909e-01
7.59177357e-02 -8.47593307e-01 -4.58503038e-01 -8.03231359e-01
2.74846256e-02 6.40811384e-01 -1.04386069e-01 5.27985059e-02
1.17059527e-02 7.55393863e-01 -1.04298151e+00 2.24542245e-01
6.43917978e-01 1.22656417e+00 -1.08484402e-01 1.24589944e+00
1.48221046e-01 6.52688384e-01 -3.58072072e-01 -3.47985595e-01
-1.91149384e-01 -1.59388408e-01 -6.66928232e-01 -4.28508550e-01
7.11704254e-01 -4.29525167e-01 -3.58097553e-01 8.27779889e-01
4.61942434e-01 -7.36913025e-01 1.07568014e+00 -8.78740311e-01
-4.35480416e-01 5.71564496e-01 -8.85863006e-01 5.65650523e-01
-2.09388718e-01 3.74055579e-02 6.95400894e-01 -5.96265137e-01
7.78800309e-01 7.31983304e-01 1.16112769e+00 5.10287941e-01
-1.00386238e+00 -6.79616153e-01 -5.64670861e-01 4.51733656e-02
-1.40751946e+00 -5.08846462e-01 -8.72519333e-03 -2.62169421e-01
7.16579854e-01 4.15484220e-01 8.08565438e-01 7.00282395e-01
7.05823243e-01 4.45801169e-01 8.60423505e-01 -3.24482173e-01
1.08761974e-01 -5.03409095e-02 4.49176244e-02 8.56469929e-01
7.87355304e-01 -1.17088377e-01 -1.08361244e-01 -2.24051788e-01
8.03332329e-01 1.92849562e-01 1.25760008e-02 5.96930198e-02
-1.08939064e+00 5.86367249e-01 6.11632347e-01 3.87498170e-01
-2.30120849e-02 5.53857267e-01 6.45896792e-01 -2.07958236e-01
4.94050801e-01 3.93734783e-01 -1.43685147e-01 2.48581655e-02
-1.11323357e+00 4.89255711e-02 3.83856505e-01 5.27118564e-01
2.07264662e-01 -3.67659330e-01 -2.43504867e-01 4.89787966e-01
2.10668668e-01 4.23212290e-01 5.81656337e-01 -8.58113825e-01
-9.84019414e-02 8.85115147e-01 -7.44535169e-03 -9.81256366e-01
-1.04084265e+00 -7.25026786e-01 -1.22307968e+00 -1.42491627e-02
8.83927643e-01 2.43826136e-01 -1.26822257e+00 1.27682221e+00
3.92598584e-02 9.43940412e-03 -1.20428003e-01 5.09726107e-01
1.03481209e+00 1.81777284e-01 2.89910167e-01 5.33668995e-02
1.59310222e+00 -5.24372041e-01 -4.33508068e-01 -5.11949472e-02
1.23247313e+00 -4.35526580e-01 5.85730016e-01 1.74931362e-01
-9.99227464e-01 7.34545756e-03 -1.23849881e+00 -3.39544535e-01
-5.30739963e-01 6.23811066e-01 9.69064176e-01 7.87405550e-01
-1.21759093e+00 4.22326803e-01 -1.18744171e+00 -7.83993244e-01
1.13720524e+00 3.73053044e-01 -5.40440857e-01 -8.49688649e-02
-5.98443151e-01 8.79351258e-01 1.83379110e-02 1.53682396e-01
-1.12939239e+00 -1.25768065e+00 -5.74205041e-01 2.59642508e-02
-1.76830843e-01 -7.02004850e-01 1.34645116e+00 -4.64045078e-01
-7.17299759e-01 1.38988650e+00 -2.71721900e-01 -7.11555123e-01
6.35626376e-01 2.33481362e-01 1.02663590e-02 3.19347173e-01
8.49948749e-02 8.87217820e-01 -1.60986081e-01 -6.94313645e-01
-8.16589653e-01 -5.93756080e-01 -3.71290356e-01 1.05154358e-01
-5.17940104e-01 -1.54517189e-01 -4.41942066e-01 -3.27822626e-01
3.17808776e-03 -5.91805458e-01 -4.38668221e-01 5.47073185e-01
-2.71537006e-01 -1.92475721e-01 5.22265315e-01 -7.33119190e-01
8.77677619e-01 -1.92327523e+00 -4.25949126e-01 2.58782599e-02
6.30385518e-01 1.23776846e-01 5.25709353e-02 -1.15488119e-01
1.17120240e-02 5.61619759e-01 -2.11573718e-03 -1.38458759e-01
-1.75123110e-01 -2.02363536e-01 2.00780362e-01 1.03765464e+00
1.53434783e-01 1.15149438e+00 -9.37173247e-01 -6.07218623e-01
3.49832959e-02 6.30193412e-01 -2.85525054e-01 -1.32925183e-01
1.12547301e-01 2.12127669e-03 5.19499294e-02 1.18598473e+00
6.05416298e-01 -5.23808300e-01 1.34416670e-01 -2.25493893e-01
3.10846359e-01 -1.15528004e-02 -3.44374597e-01 1.24991572e+00
-1.77240863e-01 9.02331829e-01 1.31184652e-01 -7.51252592e-01
5.72778046e-01 1.88201100e-01 6.32191241e-01 -5.72327971e-01
4.52430874e-01 3.45739931e-01 1.82450816e-01 -2.37562448e-01
2.92083442e-01 -7.06407279e-02 1.54719278e-02 1.40705213e-01
-1.85775995e-01 -8.41318816e-02 2.34723300e-01 3.11550796e-01
1.80217993e+00 -5.09254038e-01 6.87217355e-01 -5.00946224e-01
1.47681355e-01 2.98202455e-01 6.85034215e-01 6.47996366e-01
-6.79473341e-01 7.50882685e-01 8.43201220e-01 -7.60679781e-01
-9.84357953e-01 -1.03837991e+00 -3.77013117e-01 7.39224017e-01
-1.59753516e-01 -1.40614986e-01 -5.04452646e-01 -5.52417934e-01
8.82696584e-02 1.85478896e-01 -1.20472813e+00 -9.43223909e-02
-4.15486008e-01 -1.16271889e+00 1.14782524e+00 8.33119929e-01
3.77306312e-01 -5.48406243e-01 -5.39236665e-01 4.23240848e-02
-1.13746658e-01 -9.14723814e-01 -1.58700362e-01 3.24084550e-01
-7.55689204e-01 -1.43065059e+00 -1.12397301e+00 -7.88020432e-01
9.98686135e-01 -1.73704237e-01 1.05698609e+00 3.30722570e-01
-1.21087205e+00 -1.40955314e-01 -3.59118357e-02 -7.57503390e-01
-5.84029078e-01 2.02641010e-01 -3.16052228e-01 -5.58997691e-01
7.99047410e-01 -2.03957800e-02 -8.96638334e-01 2.96500232e-02
-7.03661501e-01 1.27127975e-01 8.82636130e-01 7.61895061e-01
7.68468440e-01 -1.21707059e-01 4.57589835e-01 -9.53102410e-01
6.15661405e-02 -4.10361499e-01 -3.38180959e-01 3.22966315e-02
-5.15363693e-01 -6.81100011e-01 4.33726221e-01 -1.61144525e-01
-5.62790573e-01 1.89811707e-01 -1.89459115e-01 2.72862576e-02
-1.70537949e-01 4.95081663e-01 3.49631846e-01 -3.54142003e-02
8.69165838e-01 -5.27104326e-02 1.02851510e-01 1.13185391e-01
-2.59943843e-01 5.30026913e-01 9.39509511e-01 1.71833903e-01
5.36010206e-01 9.11473393e-01 4.14859712e-01 -5.46779394e-01
-9.51674461e-01 -6.12910748e-01 -2.94223249e-01 -1.21852897e-01
8.86745989e-01 -9.65028167e-01 -8.00441921e-01 6.41129375e-01
-9.10206020e-01 -4.15798575e-01 -9.88428816e-02 2.39432588e-01
-4.56373036e-01 -3.30915153e-02 -1.05988729e+00 -4.84049141e-01
-6.36101425e-01 -8.42446208e-01 1.29346097e+00 5.18609345e-01
-4.19679701e-01 -8.65195751e-01 -4.37516049e-02 4.05504614e-01
6.47306859e-01 6.90453470e-01 9.48193729e-01 -5.14843822e-01
-9.69266072e-02 -7.32669175e-01 -6.94029152e-01 -3.19819540e-01
2.58517474e-01 4.33409780e-01 -9.16602612e-01 -3.01965654e-01
-5.92236996e-01 -3.83683056e-01 1.17218983e+00 8.50016296e-01
1.37213957e+00 -4.90114801e-02 -1.16762888e+00 7.95026779e-01
1.42641699e+00 1.64575234e-01 7.76816368e-01 5.07201016e-01
2.92867273e-01 1.77516595e-01 3.80032092e-01 1.59946099e-01
4.12068546e-01 7.60889873e-02 5.43814123e-01 -5.07710159e-01
-3.62309784e-01 -2.59075277e-02 -3.35236937e-01 -4.31066751e-02
-5.69227003e-02 -2.69011915e-01 -1.46389008e+00 9.11956489e-01
-1.35755277e+00 -7.35235155e-01 -3.72561999e-02 1.77303755e+00
8.90122950e-01 2.40605667e-01 -2.18654498e-01 1.24820098e-01
7.13259816e-01 -3.21916103e-01 -5.76872408e-01 -3.06702614e-01
2.62348559e-02 1.28893659e-01 8.68168175e-01 -8.19903612e-03
-1.21004200e+00 5.30757725e-01 7.31747866e+00 6.74072802e-01
-1.09873593e+00 -2.95477509e-02 1.60228884e+00 -4.58932780e-02
9.09555554e-02 -4.72191483e-01 -9.74164426e-01 7.48750642e-02
1.16156101e+00 -2.64829159e-01 -2.77728170e-01 5.71577847e-01
1.73395470e-01 -3.80545348e-01 -1.12760878e+00 7.85637856e-01
1.33677736e-01 -1.79344285e+00 -1.95854083e-01 4.44377273e-01
5.83980680e-01 1.60031751e-01 1.12042017e-01 2.15509042e-01
3.39577734e-01 -1.62310028e+00 7.56056681e-02 7.34415054e-01
1.33714426e+00 -7.33187258e-01 1.19210076e+00 7.68405646e-02
-8.95593584e-01 6.26264187e-03 -3.55867088e-01 3.21895659e-01
-3.67008865e-01 6.87293887e-01 -1.44394886e+00 -2.41863076e-02
7.22531974e-01 5.54763198e-01 -9.74383891e-01 1.23041177e+00
4.42038864e-01 8.17627490e-01 -4.86441612e-01 -1.84916958e-01
1.00146811e-02 8.47852051e-01 6.34632036e-02 1.25899422e+00
6.02126122e-01 -7.90437609e-02 -3.53730530e-01 6.33427024e-01
-4.17413473e-01 -7.17433169e-02 -2.14955360e-01 -2.99050421e-01
4.60184902e-01 1.82347357e+00 -1.30853367e+00 -2.74106264e-01
1.64950248e-02 4.62844580e-01 1.23794265e-01 -1.23400286e-01
-8.10198665e-01 -4.56275672e-01 1.58303633e-01 3.97819370e-01
1.05358876e-01 2.78236061e-01 -6.00367963e-01 -4.85712588e-01
-3.73727351e-01 -6.03422463e-01 4.24507707e-01 -5.30530214e-01
-1.16835821e+00 2.87463874e-01 -5.34674942e-01 -8.87502789e-01
-1.14119928e-02 -8.51070046e-01 -8.29242289e-01 5.08212447e-01
-1.60305560e+00 -1.22862375e+00 -6.19524896e-01 -8.40231255e-02
2.56871045e-01 4.86635743e-03 9.89982724e-01 1.30025491e-01
-5.49855411e-01 1.10332775e+00 1.60160571e-01 3.70627552e-01
9.26184297e-01 -1.53094041e+00 1.86230496e-01 2.85407156e-01
-7.45979011e-01 3.26806694e-01 6.09544873e-01 -4.61483091e-01
-1.35976982e+00 -1.35369384e+00 8.19286644e-01 -4.48382705e-01
6.83555961e-01 -1.84713393e-01 -3.72091502e-01 6.27085567e-01
-1.85233444e-01 4.25879627e-01 8.96358848e-01 -1.36586398e-01
-3.49310458e-01 -2.98336983e-01 -1.34027636e+00 3.86009693e-01
5.94563425e-01 -8.65608454e-02 8.47991481e-02 4.25648361e-01
2.60748446e-01 -5.47881007e-01 -1.17546296e+00 8.94852221e-01
8.49668801e-01 -7.67955482e-01 6.96041107e-01 -1.05107084e-01
8.08800995e-01 -4.47553396e-02 1.42998844e-02 -8.94336700e-01
-4.12435114e-01 -9.01399404e-02 1.57845289e-01 9.78390217e-01
7.83196330e-01 -4.16042298e-01 1.37273431e+00 3.31902444e-01
-1.87309951e-01 -1.25819468e+00 -1.04879379e+00 -3.53101403e-01
6.30358577e-01 2.22854633e-02 4.45920408e-01 6.82451904e-01
-1.91499724e-03 -1.32229373e-01 4.39783841e-01 4.09690849e-02
5.50860882e-01 -4.02887136e-01 6.03234291e-01 -1.24331701e+00
1.17827140e-01 -9.68019068e-01 -9.16976094e-01 -2.18990043e-01
-4.09826487e-01 -9.60961998e-01 -1.13591105e-02 -1.80271971e+00
8.92795444e-01 -3.94083083e-01 -4.60828364e-01 6.29805684e-01
-1.14184201e-01 8.67682755e-01 -3.96809757e-01 4.25115004e-02
-5.62352598e-01 -1.32196069e-01 1.01085246e+00 -5.73367774e-01
2.99010456e-01 -3.95135790e-01 -1.17223632e+00 7.94674993e-01
7.40947604e-01 -2.45052233e-01 1.42912656e-01 -3.55030484e-02
3.44490916e-01 8.32995325e-02 4.59009260e-01 -1.21347654e+00
3.00970972e-01 -1.58111647e-01 1.21670568e+00 -9.32589412e-01
2.02479199e-01 -4.36039299e-01 4.64395285e-02 7.91676641e-01
-3.93756866e-01 -2.55093157e-01 5.10897279e-01 4.01291281e-01
1.45371035e-01 1.28070321e-02 8.60312402e-01 7.71438470e-03
-2.54832674e-02 3.88542384e-01 -6.78008199e-01 -3.92562866e-01
1.28329194e+00 -4.05848831e-01 -9.73618090e-01 -5.60384206e-02
-3.59976619e-01 2.48556450e-01 5.54211259e-01 -1.02984831e-01
2.59712517e-01 -1.11618328e+00 -1.00696540e+00 -2.80181199e-01
3.63596857e-01 1.53801516e-01 2.26813659e-01 1.09886849e+00
-1.02328098e+00 5.28678119e-01 -9.56895873e-02 -7.50630677e-01
-1.29006350e+00 1.10451423e-01 7.09151268e-01 -8.34678233e-01
-5.01867771e-01 1.00067830e+00 8.65917429e-02 -2.23650381e-01
3.79908234e-01 -3.22121471e-01 -3.91477257e-01 -3.52581553e-02
8.27397764e-01 4.33881700e-01 3.64562958e-01 -3.34318370e-01
-4.21456844e-01 1.38021573e-01 -6.11073017e-01 3.22672307e-01
1.40142155e+00 2.48696238e-01 -1.84586406e-01 1.35857955e-01
1.25122678e+00 -4.39363778e-01 -7.80378222e-01 2.87899315e-01
-1.13730006e-01 7.73931593e-02 3.05302829e-01 -1.13995516e+00
-9.82254028e-01 5.16921401e-01 1.01520705e+00 6.07232340e-02
1.15414834e+00 1.56107068e-01 7.30800271e-01 3.34993154e-01
-6.43998832e-02 -7.77215064e-01 -2.61910468e-01 1.19080938e-01
5.37718415e-01 -1.16510665e+00 2.46520594e-01 -3.22229803e-01
1.03006542e-01 1.27176559e+00 6.46099091e-01 -1.40936658e-01
5.20783246e-01 7.52639294e-01 1.76097050e-01 -4.11593765e-01
-9.75169301e-01 2.46265903e-01 2.01938264e-02 4.93224949e-01
7.62880206e-01 2.66086668e-01 -4.30368900e-01 6.85500503e-01
-3.05963904e-01 2.27931142e-01 6.64402902e-01 8.40111077e-01
-5.53447425e-01 -3.44788820e-01 -2.20636442e-01 1.16508842e+00
-8.99379313e-01 1.46303415e-01 -6.23600125e-01 8.24386418e-01
-1.23591974e-01 5.38158655e-01 5.17931044e-01 4.30921614e-02
-1.77563697e-01 -3.16703737e-01 3.41263801e-01 -4.50479478e-01
-3.48406553e-01 -2.56835725e-02 -5.16331941e-03 -3.79440099e-01
-3.16959202e-01 -5.89480102e-01 -1.45666111e+00 -3.35592479e-01
-3.69164944e-01 -2.68402934e-01 7.31221855e-01 6.39023185e-01
1.65058687e-01 6.82314754e-01 1.59780473e-01 -6.40421093e-01
-1.70107365e-01 -1.13311601e+00 -6.33272171e-01 -6.22933097e-02
3.53133291e-01 -3.52915317e-01 -4.25927460e-01 9.49718729e-02] | [15.133021354675293, -3.0637359619140625] |
0e413068-9f21-4888-bc71-2b21c4961623 | self-supervised-super-plane-for-neural-3d | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Ye_Self-Supervised_Super-Plane_for_Neural_3D_Reconstruction_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Ye_Self-Supervised_Super-Plane_for_Neural_3D_Reconstruction_CVPR_2023_paper.pdf | Self-Supervised Super-Plane for Neural 3D Reconstruction | Neural implicit surface representation methods show impressive reconstruction results but struggle to handle texture-less planar regions that widely exist in indoor scenes. Existing approaches addressing this leverage image prior that requires assistive networks trained with large-scale annotated datasets. In this work, we introduce a self-supervised super-plane constraint by exploring the free geometry cues from the predicted surface, which can further regularize the reconstruction of plane regions without any other ground truth annotations. Specifically, we introduce an iterative training scheme, where (i) grouping of pixels to formulate a super-plane (analogous to super-pixels), and (ii) optimizing of the scene reconstruction network via a super-plane constraint, are progressively conducted. We demonstrate that the model trained with super-planes surprisingly outperforms the one using conventional annotated planes, as individual super-plane statistically occupies a larger area and leads to more stable training. Extensive experiments show that our self-supervised super-plane constraint significantly improves 3D reconstruction quality even better than using ground truth plane segmentation. Additionally, the plane reconstruction results from our model can be used for auto-labeling for other vision tasks. The code and models are available at https: //github.com/botaoye/S3PRecon. | ['Ming-Hsuan Yang', 'Xueting Li', 'Sifei Liu', 'Botao Ye'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['3d-reconstruction'] | ['computer-vision'] | [ 5.40951908e-01 5.61085403e-01 -1.49552330e-01 -5.33154368e-01
-8.72270107e-01 -3.74314487e-01 5.43222547e-01 -3.00010949e-01
-4.50455509e-02 4.11519587e-01 -1.57074872e-02 -1.63639978e-01
1.64246976e-01 -9.40856695e-01 -1.11013770e+00 -5.50461292e-01
2.02326536e-01 7.12442040e-01 6.76710546e-01 -1.32796690e-02
2.41955847e-01 5.13078272e-01 -1.33959365e+00 3.65807563e-01
9.95836616e-01 1.00735307e+00 5.07032752e-01 2.47403413e-01
-1.56282827e-01 3.26223165e-01 -1.06636159e-01 -2.82556564e-01
6.25334740e-01 -1.65263880e-02 -9.02322352e-01 4.19193625e-01
7.95086086e-01 -4.16898817e-01 -5.46090379e-02 1.05798614e+00
7.44267832e-03 8.77351984e-02 5.30511379e-01 -6.59147322e-01
-3.34082752e-01 9.94600356e-02 -8.14127564e-01 -3.92540812e-01
2.42267013e-01 1.48376033e-01 8.67103875e-01 -1.15060437e+00
8.05392206e-01 1.17264581e+00 7.57571757e-01 4.06717032e-01
-1.33100224e+00 -4.01148677e-01 4.37054932e-01 -1.98216334e-01
-1.37425685e+00 -2.55211771e-01 1.17757916e+00 -4.65799838e-01
7.38969624e-01 1.15444608e-01 5.58407545e-01 9.23225343e-01
-1.32522881e-01 7.50472903e-01 1.00297856e+00 -4.01059896e-01
9.65444669e-02 -9.94282812e-02 6.78381994e-02 9.23684180e-01
1.59410432e-01 4.44101803e-02 -3.95193636e-01 1.44634277e-01
1.32592618e+00 5.16095199e-02 -4.90750551e-01 -9.38807607e-01
-1.13802552e+00 5.19677222e-01 6.52423799e-01 5.68081886e-02
-4.32763249e-01 2.32170746e-01 2.66101155e-02 -2.88641423e-01
8.55547488e-01 4.03843433e-01 -5.96471548e-01 3.48353356e-01
-1.10230315e+00 4.10986871e-01 5.17019272e-01 9.29026544e-01
1.23524475e+00 -3.17935720e-02 9.49987173e-02 8.82139385e-01
4.35732782e-01 4.04674977e-01 -1.94844492e-02 -1.57280028e+00
5.61427712e-01 6.41388416e-01 2.26065457e-01 -1.01167250e+00
-3.45601916e-01 -4.87943977e-01 -6.24262452e-01 2.74375111e-01
7.07086980e-01 1.95656791e-01 -1.30157959e+00 1.38926685e+00
6.59786105e-01 2.07541168e-01 -2.35981643e-01 9.76139545e-01
6.61010087e-01 5.25102556e-01 -3.51647705e-01 2.18307227e-01
1.38534641e+00 -1.28081810e+00 -3.85021061e-01 -3.86547118e-01
4.18589473e-01 -5.80905139e-01 1.11379623e+00 5.70982516e-01
-1.04617357e+00 -4.17526931e-01 -8.90308201e-01 -2.87860483e-01
-2.55967565e-02 7.06060231e-02 8.21317255e-01 3.68007332e-01
-1.17373073e+00 5.97125053e-01 -9.50112522e-01 -1.15027361e-01
8.63585770e-01 3.14199388e-01 -3.96232367e-01 -2.91887522e-01
-4.91343081e-01 3.65872771e-01 2.71505117e-01 3.51066105e-02
-7.67198622e-01 -7.54562616e-01 -1.09872854e+00 -2.73145407e-01
7.30479419e-01 -9.54527617e-01 1.15441751e+00 -1.00860202e+00
-1.60907912e+00 1.20201635e+00 -3.89048040e-01 -2.63731718e-01
4.70673293e-01 -4.27173644e-01 4.30267125e-01 4.05832887e-01
2.67361343e-01 9.68278468e-01 7.92631745e-01 -1.89968479e+00
-3.50354075e-01 -5.49253941e-01 1.90766782e-01 2.65882581e-01
2.34746650e-01 -5.23185074e-01 -9.04220283e-01 -5.75381696e-01
7.02013433e-01 -7.45887160e-01 -6.35676324e-01 2.76314139e-01
-7.24674940e-01 -1.27393275e-01 5.93094409e-01 -6.92302883e-01
6.64971948e-01 -1.91143429e+00 5.34952544e-02 4.62474346e-01
3.02234679e-01 -6.28041998e-02 -6.22398295e-02 -8.80138278e-02
4.38162591e-03 2.13919625e-01 -7.46934056e-01 -6.67244792e-01
-2.55534083e-01 3.42432618e-01 -1.80417269e-01 4.11530703e-01
3.87939930e-01 7.98500836e-01 -7.46102214e-01 -4.56428766e-01
4.16980535e-01 5.88720024e-01 -9.63208020e-01 5.77797443e-02
-5.10248065e-01 8.34260762e-01 -5.26399553e-01 7.61159182e-01
8.88151646e-01 -5.45756340e-01 6.04093708e-02 -3.91557366e-01
-1.36690915e-01 3.74708295e-01 -1.09096336e+00 2.07646203e+00
-2.42860079e-01 3.50144535e-01 2.52000332e-01 -9.45483267e-01
9.80639040e-01 -4.45379969e-03 5.92882812e-01 -5.45211911e-01
1.07777633e-01 2.44976148e-01 -4.68079358e-01 -3.23145539e-01
3.19715828e-01 3.99272218e-02 4.14375842e-01 -6.22356907e-02
3.21911164e-02 -8.07287693e-01 -1.16883211e-01 -1.41963959e-01
6.85851693e-01 8.03045630e-01 4.54757325e-02 -2.22570598e-01
5.46509326e-01 6.04780801e-02 7.39153445e-01 3.53429079e-01
2.68384777e-02 1.20250583e+00 2.85080492e-01 -6.04994953e-01
-1.21740019e+00 -1.00862026e+00 -4.82117236e-01 4.99368489e-01
5.36831200e-01 -3.42061907e-01 -1.00757039e+00 -6.41855121e-01
-8.92465636e-02 4.05155897e-01 -5.79634726e-01 3.81218761e-01
-7.83778965e-01 -4.81816679e-01 1.48412380e-02 6.18275642e-01
7.30703771e-01 -7.24709094e-01 -4.83068377e-01 -2.46468615e-02
-3.66768420e-01 -1.30621290e+00 -2.63806164e-01 2.84231633e-01
-1.12109637e+00 -9.35748875e-01 -8.34311545e-01 -7.34359622e-01
9.65510428e-01 3.06717962e-01 1.17535746e+00 2.23778501e-01
2.15893239e-02 4.77625608e-01 -1.12327941e-01 -7.37363622e-02
2.99275696e-01 -9.42502767e-02 -2.09461614e-01 5.31188957e-02
-1.24586500e-01 -8.71730208e-01 -7.94295073e-01 4.47547346e-01
-5.82665205e-01 6.36482775e-01 4.15767521e-01 7.54548490e-01
1.35336506e+00 -2.43174881e-01 2.49020141e-02 -1.16157079e+00
-2.59516597e-01 -3.60106885e-01 -7.46004939e-01 -7.01558813e-02
-3.93261760e-01 1.15848161e-01 2.31472999e-01 -5.19470498e-02
-1.22459745e+00 4.34433460e-01 -3.94673586e-01 -5.76288044e-01
-3.91424507e-01 9.38195810e-02 -3.28042418e-01 -1.78561732e-01
5.21819174e-01 9.55888033e-02 -1.46780655e-01 -6.72021270e-01
3.26788098e-01 4.34739776e-02 5.06541967e-01 -8.74278605e-01
1.04392767e+00 9.76541162e-01 1.23901695e-01 -9.38974261e-01
-1.09449756e+00 -4.42809522e-01 -1.00605929e+00 -1.15573443e-01
1.17711103e+00 -8.83934319e-01 -3.14442039e-01 2.88030922e-01
-1.35528600e+00 -9.25340831e-01 -3.42599928e-01 3.46072704e-01
-7.51354992e-01 5.94880879e-01 -5.18993855e-01 -7.22379625e-01
-8.54750648e-02 -1.06871772e+00 1.56032312e+00 1.59914181e-01
-6.65921196e-02 -8.68658423e-01 1.50682600e-02 1.01060283e+00
-1.71032175e-02 3.82207751e-01 5.92699051e-01 -7.53711388e-02
-1.29853678e+00 2.03448489e-01 -3.39987814e-01 3.11294436e-01
2.18712296e-02 -1.24425963e-01 -1.20189202e+00 1.70205787e-01
-7.83730019e-03 -2.67093539e-01 9.51798618e-01 5.97727537e-01
1.67981958e+00 -1.50688350e-01 -3.37298900e-01 1.39043391e+00
1.32446408e+00 -7.65727535e-02 7.72968590e-01 5.23984849e-01
1.17073143e+00 8.68582308e-01 5.20304680e-01 1.69966415e-01
6.66288793e-01 6.42311394e-01 6.63377821e-01 -4.29952800e-01
-2.44674802e-01 -4.02040780e-01 1.45898564e-02 3.78385752e-01
-5.21620095e-01 -1.94514006e-01 -1.02686191e+00 3.45238149e-01
-1.66265261e+00 -5.32041132e-01 -3.80511373e-01 2.06575775e+00
7.99838960e-01 1.91661000e-01 -1.40358001e-01 2.93188784e-02
3.03064138e-01 1.69638738e-01 -6.35764539e-01 1.45530939e-01
-1.61955297e-01 2.33006954e-01 5.61239660e-01 8.61444116e-01
-1.24286008e+00 1.28534007e+00 5.74882317e+00 8.86136055e-01
-1.04451191e+00 3.04215904e-02 9.21470761e-01 2.67302066e-01
-5.98208487e-01 1.20641507e-01 -6.60304010e-01 1.19873777e-01
2.14479983e-01 4.77675855e-01 2.97806770e-01 9.13928449e-01
2.16163680e-01 -2.69385427e-01 -1.01938558e+00 1.16285098e+00
1.19197730e-03 -1.39515460e+00 2.01541483e-01 2.23682404e-01
9.74861562e-01 1.82079837e-01 -4.56652753e-02 -9.61482972e-02
2.50684172e-01 -9.51496959e-01 9.40310597e-01 6.10200286e-01
8.64089608e-01 -3.02449882e-01 4.67081845e-01 4.18981105e-01
-1.25465298e+00 3.92021656e-01 -2.47798860e-01 1.38181090e-01
2.55288899e-01 7.75821269e-01 -4.87839550e-01 8.61574292e-01
9.31142449e-01 8.06535482e-01 -4.54147846e-01 8.32489848e-01
-4.14501458e-01 5.36219478e-01 -5.23111701e-01 5.59577525e-01
1.56340823e-01 -5.73627412e-01 5.19607961e-01 9.22916353e-01
4.36212607e-02 1.36605963e-01 3.42304260e-01 1.29918718e+00
-3.55611965e-02 -5.67571856e-02 -6.62576675e-01 4.43468392e-01
1.24330975e-01 1.19727457e+00 -9.81788993e-01 -3.14393014e-01
-3.15680206e-01 1.15766454e+00 5.26087701e-01 6.88841105e-01
-6.23150289e-01 3.10984999e-01 4.59185064e-01 6.47391796e-01
2.92633146e-01 -5.06468892e-01 -8.62438023e-01 -1.18368018e+00
1.71498746e-01 -4.77209508e-01 -5.46048768e-02 -8.56346488e-01
-1.09820211e+00 5.01259744e-01 1.87449157e-02 -1.00859511e+00
9.09909140e-03 -7.70577967e-01 -4.84109789e-01 7.38374889e-01
-1.79069591e+00 -1.28800619e+00 -5.58451295e-01 5.45259297e-01
7.02817976e-01 4.06271666e-01 5.89044511e-01 1.32980749e-01
-4.89005178e-01 3.09562445e-01 -3.67632449e-01 6.99451193e-02
3.52260381e-01 -1.31937087e+00 2.78987050e-01 8.36692631e-01
-2.55606370e-03 5.76199591e-01 3.64922404e-01 -8.48678231e-01
-1.10273778e+00 -1.11182654e+00 3.54518443e-01 -6.61962390e-01
1.19649358e-01 -4.86590564e-01 -1.01846969e+00 6.58297241e-01
-1.52604550e-01 1.71278521e-01 3.37928087e-01 -2.64130235e-02
-3.04985911e-01 3.03296708e-02 -1.07799363e+00 6.05165899e-01
1.39410508e+00 -3.45404983e-01 -3.20044190e-01 3.94884348e-01
8.64210367e-01 -7.27436960e-01 -5.54540455e-01 6.58146083e-01
2.38272071e-01 -1.12474287e+00 1.28114700e+00 -7.23706484e-02
6.77331448e-01 -5.14226019e-01 -2.87789434e-01 -7.93708920e-01
-2.47940987e-01 -3.62582654e-01 -2.72278208e-02 9.00610805e-01
4.10263300e-01 -6.07789934e-01 1.25929451e+00 6.63282216e-01
-6.02717042e-01 -1.22150993e+00 -7.65771866e-01 -6.34070873e-01
8.70831683e-03 -5.75831115e-01 3.99715662e-01 9.82118189e-01
-6.84837520e-01 4.33317609e-02 -1.16870172e-01 4.05080289e-01
9.08194125e-01 1.14274330e-01 1.04062450e+00 -1.17194510e+00
-3.30184996e-01 -3.39670420e-01 -1.42249078e-01 -1.91862655e+00
2.19914280e-02 -9.05479372e-01 2.71932006e-01 -1.79870701e+00
-6.99694827e-02 -7.91970730e-01 2.13216051e-01 6.02296889e-01
-1.50000095e-01 5.47025442e-01 -9.40151066e-02 3.02329481e-01
-5.11696637e-01 8.36537004e-01 1.64020419e+00 3.06450240e-02
-2.79695958e-01 -1.24222949e-01 -5.21568418e-01 1.27171314e+00
6.34587824e-01 -2.05988646e-01 -3.84309947e-01 -6.76174819e-01
8.17956328e-02 -1.31584093e-01 5.84682345e-01 -9.39797282e-01
7.59443501e-04 -2.52677828e-01 5.58029413e-01 -8.07640016e-01
7.28727758e-01 -8.44275892e-01 -4.92392294e-02 1.25744596e-01
1.43547997e-01 -4.61744994e-01 4.61588725e-02 5.78048348e-01
-5.56697473e-02 -1.48623422e-01 7.24782765e-01 -3.13943893e-01
-5.16616106e-01 5.70409060e-01 2.20916718e-02 8.37237760e-03
8.51250112e-01 -6.43536270e-01 1.15162283e-01 -2.01437354e-01
-6.61699831e-01 2.44777441e-01 9.13855493e-01 2.55022552e-02
8.62337053e-01 -1.01919901e+00 -3.22116435e-01 2.85293967e-01
-1.45255804e-01 1.11166978e+00 1.19853146e-01 8.21195543e-01
-9.44952250e-01 1.78750888e-01 1.47491410e-01 -1.13147998e+00
-8.91464472e-01 1.60880238e-01 6.46921992e-01 -3.04208994e-01
-1.13975573e+00 1.05689168e+00 9.39001441e-01 -7.23955452e-01
1.71843320e-01 -7.52869368e-01 -2.11199019e-02 -3.95128965e-01
5.01726680e-02 9.10129529e-05 -6.09736815e-02 -5.67779243e-01
-2.80072778e-01 1.07086694e+00 1.38886854e-01 -1.42438129e-01
1.63406873e+00 -1.03251129e-01 -5.76370470e-02 3.36860090e-01
8.01893175e-01 7.52391517e-02 -1.88442874e+00 -3.07485223e-01
-1.08685687e-01 -6.74484134e-01 4.89470363e-02 -5.62593639e-01
-1.16400671e+00 8.07450473e-01 3.62003356e-01 -3.35127354e-01
8.59515727e-01 1.41668946e-01 7.52681375e-01 3.32875520e-01
7.86420524e-01 -7.39357948e-01 2.24012405e-01 5.85928500e-01
9.85151350e-01 -1.33247530e+00 -6.87202439e-02 -1.17164254e+00
-3.55389059e-01 1.07477665e+00 7.55618513e-01 -5.22193491e-01
7.11009383e-01 1.08066112e-01 -9.75291729e-02 -4.72780615e-01
-1.87396184e-01 -1.27705365e-01 5.48834264e-01 7.99079001e-01
2.44196340e-01 1.05593475e-02 7.95199126e-02 6.31330073e-01
-4.21465755e-01 -2.76202947e-01 1.56076565e-01 6.58975780e-01
-2.99269110e-01 -1.08401442e+00 -5.33186018e-01 2.87615776e-01
-2.29336634e-01 -5.05292378e-02 -3.29417348e-01 6.57862902e-01
4.51675713e-01 5.74771464e-01 1.96401760e-01 -2.34867245e-01
3.04606527e-01 -2.22966388e-01 6.85880184e-01 -9.21545923e-01
-1.40127465e-01 2.56748736e-01 1.08778261e-01 -1.01399267e+00
-3.70213419e-01 -6.49319112e-01 -1.48939431e+00 4.57714312e-02
-1.84975892e-01 -2.29348496e-01 5.24474382e-01 8.40203702e-01
2.22059488e-01 5.13988733e-01 4.21869576e-01 -1.55391157e+00
1.37568325e-01 -6.97142899e-01 -3.39419425e-01 4.39156055e-01
2.39698797e-01 -9.76109862e-01 -4.03653562e-01 2.68962860e-01] | [8.815401077270508, -2.905479669570923] |
970d7c1f-4c5e-4a3a-9d1d-4255bdaa0d49 | animesr-learning-real-world-super-resolution | 2206.07038 | null | https://arxiv.org/abs/2206.07038v3 | https://arxiv.org/pdf/2206.07038v3.pdf | AnimeSR: Learning Real-World Super-Resolution Models for Animation Videos | This paper studies the problem of real-world video super-resolution (VSR) for animation videos, and reveals three key improvements for practical animation VSR. First, recent real-world super-resolution approaches typically rely on degradation simulation using basic operators without any learning capability, such as blur, noise, and compression. In this work, we propose to learn such basic operators from real low-quality animation videos, and incorporate the learned ones into the degradation generation pipeline. Such neural-network-based basic operators could help to better capture the distribution of real degradations. Second, a large-scale high-quality animation video dataset, AVC, is built to facilitate comprehensive training and evaluations for animation VSR. Third, we further investigate an efficient multi-scale network structure. It takes advantage of the efficiency of unidirectional recurrent networks and the effectiveness of sliding-window-based methods. Thanks to the above delicate designs, our method, AnimeSR, is capable of restoring real-world low-quality animation videos effectively and efficiently, achieving superior performance to previous state-of-the-art methods. Codes and models are available at https://github.com/TencentARC/AnimeSR. | ['Ying Shan', 'Gen Li', 'Xintao Wang', 'Yanze Wu'] | 2022-06-14 | null | null | null | null | ['video-super-resolution'] | ['computer-vision'] | [ 6.98108897e-02 -4.57059145e-01 -2.16593042e-01 1.39676496e-01
-6.89952970e-01 -1.89858899e-01 2.73650557e-01 -6.34787261e-01
-6.49698526e-02 6.34056985e-01 5.74012458e-01 -3.40881459e-02
1.68485388e-01 -5.38854897e-01 -6.81244552e-01 -6.62855625e-01
-3.12198400e-01 -1.76967442e-01 5.52582622e-01 -5.62231481e-01
-5.28393909e-02 4.21458632e-01 -1.54283857e+00 2.78293490e-01
8.37523103e-01 6.72278881e-01 4.14974898e-01 1.00092804e+00
3.99054527e-01 1.03952444e+00 -6.31228268e-01 -8.88239518e-02
1.17190026e-01 -6.30268753e-01 -4.29135114e-01 -1.55335426e-01
3.20844024e-01 -8.19354951e-01 -9.67043817e-01 1.16413140e+00
7.24003673e-01 2.13496312e-01 2.53339231e-01 -8.45199823e-01
-9.40763056e-01 6.78605795e-01 -6.90861762e-01 5.75746894e-01
4.46838379e-01 5.82388520e-01 8.58702064e-01 -7.43549705e-01
7.70840168e-01 1.37560833e+00 5.72231054e-01 8.75940859e-01
-1.06408274e+00 -7.74364412e-01 -9.78773236e-02 5.16854823e-01
-1.26139808e+00 -5.81689596e-01 9.20129120e-01 -1.15214258e-01
6.39008939e-01 2.07794800e-01 6.98545635e-01 1.45857120e+00
4.44202870e-02 7.02365935e-01 9.22322631e-01 -3.79178524e-02
1.01617545e-01 -3.51093322e-01 -3.88931096e-01 4.51107919e-01
-1.61047578e-01 3.76392663e-01 -6.83640003e-01 -3.26069701e-03
1.45092309e+00 -1.83149353e-01 -7.87066936e-01 1.02555938e-01
-1.13050938e+00 4.84546393e-01 3.55675578e-01 3.49699825e-01
-4.08646435e-01 3.75320613e-01 5.92440188e-01 1.85634613e-01
6.34215057e-01 2.87567437e-01 -7.20050111e-02 -4.72573698e-01
-1.07975435e+00 7.56783038e-02 2.69573092e-01 7.38658905e-01
3.06074739e-01 8.14927518e-01 -1.63514838e-01 9.44160283e-01
-1.02554291e-01 3.63721639e-01 6.41750276e-01 -1.60533476e+00
3.19187254e-01 -7.50614256e-02 2.38694966e-01 -1.03023767e+00
-1.81144327e-01 -2.85027653e-01 -1.33018291e+00 3.38996440e-01
1.29856840e-01 -6.94369376e-02 -8.08411777e-01 1.72166026e+00
1.25580594e-01 8.22104692e-01 5.61898127e-02 1.33139145e+00
7.60943115e-01 9.24797177e-01 -2.52713650e-01 -5.14113307e-01
1.17447412e+00 -9.24417436e-01 -9.95156705e-01 2.40163580e-01
-6.55465424e-02 -5.15627623e-01 1.35735071e+00 4.34668213e-01
-1.52100480e+00 -7.59872854e-01 -1.12943327e+00 -1.12745285e-01
2.86734641e-01 2.92383246e-02 3.29580069e-01 8.88488367e-02
-1.35771620e+00 1.10043931e+00 -9.32262897e-01 1.02558486e-01
3.64044815e-01 -1.40797555e-01 1.49416318e-02 1.04965486e-01
-1.56042552e+00 5.77908158e-01 1.19780950e-01 1.91657022e-01
-1.21824372e+00 -8.56741011e-01 -7.24688232e-01 2.15258077e-01
4.41041887e-01 -6.13823771e-01 1.24274766e+00 -9.50735986e-01
-1.89362419e+00 2.81474948e-01 -1.25283331e-01 -5.23749471e-01
6.07298553e-01 -4.56650347e-01 -5.73308170e-01 6.07703924e-01
-4.20720935e-01 3.48003447e-01 1.23868465e+00 -1.28179777e+00
-3.64740074e-01 1.19353577e-01 4.22330350e-02 1.18974432e-01
-4.92414445e-01 2.83462018e-01 -5.31906009e-01 -1.09165800e+00
-3.31986099e-01 -6.84692502e-01 -2.12079495e-01 -1.29453465e-02
-7.49362186e-02 3.74517753e-03 9.05699670e-01 -1.20860195e+00
1.56450438e+00 -2.25097513e+00 4.88445908e-01 -3.25244308e-01
5.42552233e-01 5.34876645e-01 -3.40544075e-01 1.35091633e-01
-2.19654262e-01 9.49663222e-02 -6.10392541e-02 -2.14009494e-01
-2.81598747e-01 -1.96240796e-03 -5.27853191e-01 2.84744948e-01
1.74523935e-01 7.34257221e-01 -9.16011572e-01 -4.54349667e-01
1.86206639e-01 8.85812998e-01 -3.96532148e-01 3.63308132e-01
-2.42808983e-01 5.69052637e-01 -1.70646623e-01 5.54485679e-01
4.64742154e-01 -3.76437157e-01 6.26295358e-02 -4.53519166e-01
-1.53652489e-01 4.25891615e-02 -1.06792676e+00 1.71769023e+00
-4.46488708e-01 8.64380360e-01 2.79522449e-01 -6.20079517e-01
7.52933741e-01 4.62612212e-01 4.93718475e-01 -7.99237728e-01
-2.03721765e-02 1.82478443e-01 -3.30743670e-01 -6.17800474e-01
8.49294543e-01 1.92891330e-01 4.36172247e-01 2.94538766e-01
-1.59011945e-01 6.04516044e-02 2.36618683e-01 3.81406695e-01
1.16194141e+00 3.29019368e-01 -5.93921430e-02 1.69650137e-01
3.96052450e-01 -4.55600023e-01 7.36191928e-01 3.68888289e-01
-2.84185171e-01 8.75926733e-01 4.48372126e-01 -2.64411032e-01
-1.53358948e+00 -8.31154764e-01 3.04928839e-01 9.38718617e-01
2.06789389e-01 -4.38493371e-01 -8.52134168e-01 -4.19540738e-04
-4.40947443e-01 4.62624162e-01 -3.84844571e-01 -1.72900870e-01
-1.16872716e+00 -5.11952996e-01 5.95790684e-01 5.08755922e-01
5.70967615e-01 -1.06932342e+00 -6.08732045e-01 1.65517733e-01
-3.93729746e-01 -1.15768516e+00 -7.28526473e-01 -4.96412605e-01
-9.54463303e-01 -7.83811092e-01 -1.08261847e+00 -4.86491472e-01
1.46814495e-01 3.80985260e-01 1.14496148e+00 2.98976541e-01
-1.28311709e-01 6.81080967e-02 -3.99874032e-01 5.41105509e-01
-7.54377186e-01 -1.77431852e-01 1.79786101e-01 -1.52465448e-01
-4.08857107e-01 -1.08536601e+00 -7.80662179e-01 2.47780308e-01
-9.57506835e-01 3.21462572e-01 3.99782836e-01 8.80849779e-01
6.32457852e-01 2.54063487e-01 5.05611300e-01 -4.75712717e-01
7.05405176e-01 -3.58087093e-01 -5.38312137e-01 8.18549693e-02
-3.49553823e-01 -3.16854678e-02 1.03451955e+00 -8.82047296e-01
-1.17009044e+00 -4.61424530e-01 -2.69767672e-01 -1.16030633e+00
9.44834575e-02 2.32441038e-01 -5.63018173e-02 8.57647210e-02
6.89387798e-01 4.20386434e-01 -8.27844441e-02 -5.75979173e-01
4.64677304e-01 5.48845053e-01 9.62345779e-01 -4.52123672e-01
9.75112319e-01 3.84685129e-01 -1.41636342e-01 -8.28784466e-01
-3.69411618e-01 -7.35409185e-02 -1.96047440e-01 -3.96361768e-01
7.23009944e-01 -1.15464735e+00 -6.90243661e-01 7.89429128e-01
-1.09208405e+00 -8.14946234e-01 -3.53127748e-01 3.61173749e-01
-7.75313675e-01 5.27620435e-01 -1.36252546e+00 -6.21950448e-01
-4.47939277e-01 -1.09452975e+00 8.15085232e-01 3.90359610e-01
1.71595246e-01 -7.81029761e-01 2.19170246e-02 1.42358586e-01
7.85988629e-01 3.09947163e-01 5.67143738e-01 1.87190145e-01
-8.17222118e-01 4.13061202e-01 -3.49232912e-01 5.23577034e-01
1.85686648e-01 4.45630819e-01 -5.78526080e-01 -5.46255648e-01
4.61752079e-02 -3.22815925e-01 8.93244147e-01 4.84049261e-01
1.27071905e+00 -4.47626352e-01 1.92794368e-01 9.61112440e-01
1.26792455e+00 3.86844389e-02 8.59700918e-01 1.83203712e-01
9.19836521e-01 8.21697786e-02 5.15255392e-01 5.95008731e-01
3.12922776e-01 8.70140612e-01 3.36817205e-01 -2.17412770e-01
-6.91911161e-01 -2.41457790e-01 6.88942671e-01 1.28519833e+00
-6.17668509e-01 -1.37506142e-01 -5.24333596e-01 3.91583979e-01
-1.74192917e+00 -1.27031827e+00 1.16889589e-01 1.81682336e+00
1.11737525e+00 7.39759654e-02 2.43593857e-01 -5.93883032e-03
9.17289138e-01 7.36003697e-01 -8.10012698e-01 -1.33566827e-01
-3.59646857e-01 1.88317914e-02 2.27197528e-01 3.76150429e-01
-8.83853376e-01 1.01485860e+00 6.15652275e+00 1.18385422e+00
-1.30756295e+00 2.05926269e-01 5.94363809e-01 -2.22040996e-01
-3.33821237e-01 -2.93561071e-01 -4.56946552e-01 6.34998322e-01
1.19640815e+00 -2.82174885e-01 1.06552720e+00 6.45177901e-01
8.12023759e-01 3.08259577e-01 -6.66376889e-01 9.39095378e-01
3.47114541e-02 -1.62689233e+00 7.07417950e-02 -3.49952579e-01
7.50929594e-01 -3.47030126e-02 2.71227986e-01 7.27449208e-02
4.27414924e-01 -8.15670192e-01 6.31205916e-01 5.84849954e-01
1.25600839e+00 -6.90200210e-01 4.24877852e-01 -5.66313928e-03
-1.42863905e+00 -1.46244280e-02 -3.59760284e-01 1.21056005e-01
5.34157932e-01 4.19184774e-01 -2.97309197e-02 5.28205156e-01
8.73344302e-01 1.12224650e+00 -4.59456712e-01 7.89706171e-01
-3.64760697e-01 8.77640247e-01 -3.46606772e-04 4.09367561e-01
-1.43558025e-01 -1.10907800e-01 8.28061163e-01 1.18281257e+00
4.22851771e-01 3.39577764e-01 -1.94256693e-01 7.69812524e-01
-2.19960392e-01 -3.57999116e-01 -1.79157451e-01 -6.77078366e-02
5.04025042e-01 1.11824358e+00 -3.18632334e-01 -5.41420877e-01
-3.47896397e-01 1.10259545e+00 1.64921641e-01 7.08150864e-01
-1.15150404e+00 -1.52936012e-01 8.28920662e-01 2.87089925e-02
4.18682337e-01 -4.15813446e-01 6.44192323e-02 -1.47235203e+00
3.94793693e-03 -1.32037520e+00 1.59252420e-01 -1.10511494e+00
-1.07167614e+00 8.28096747e-01 -4.86515760e-02 -1.44171965e+00
-3.67724389e-01 -1.29910618e-01 -6.79634333e-01 6.38416231e-01
-1.54530108e+00 -8.65579367e-01 -4.33284849e-01 6.21626019e-01
9.65450525e-01 -9.73832011e-02 3.09162438e-01 4.63563353e-01
-7.51503706e-01 4.95007068e-01 1.99901208e-01 3.74440625e-02
8.00209522e-01 -9.94650304e-01 6.78039670e-01 1.25102615e+00
-1.09761462e-01 3.78106654e-01 1.06961870e+00 -6.42310083e-01
-1.39784336e+00 -1.04011595e+00 3.94045450e-02 -3.84170935e-02
1.03406668e+00 -9.19877142e-02 -1.41375959e+00 3.59241843e-01
3.65789831e-01 1.54924572e-01 -1.75777692e-02 -4.27611083e-01
-4.03070569e-01 -8.62528011e-02 -8.34490001e-01 8.65887403e-01
1.26900971e+00 -6.26325786e-01 -3.00464004e-01 9.23150256e-02
1.33697259e+00 -7.35706568e-01 -1.05451322e+00 4.21549141e-01
4.48322922e-01 -1.01813245e+00 1.13837337e+00 -4.99741703e-01
1.01797616e+00 -4.22527373e-01 9.52115208e-02 -1.49757421e+00
-4.24772412e-01 -1.02519965e+00 -8.25559556e-01 1.15034688e+00
-7.52404612e-03 -3.56280565e-01 5.85363269e-01 3.06869328e-01
-1.20905876e-01 -7.30123639e-01 -8.14900875e-01 -7.33653724e-01
-1.35571465e-01 -1.46324337e-01 5.24271667e-01 8.40830505e-01
-4.22263145e-01 9.56242159e-02 -1.05983555e+00 2.79291153e-01
7.44071126e-01 8.80862679e-03 5.77882588e-01 -5.63016951e-01
-5.92701435e-01 -4.06950563e-01 -1.58897251e-01 -1.16696072e+00
1.05056114e-01 -2.03515962e-01 -4.86827604e-02 -1.31001329e+00
2.05913275e-01 -1.45156518e-01 -2.73339301e-01 1.31137267e-01
-4.44996446e-01 3.31927568e-01 4.22296226e-01 5.16104996e-01
-6.16146624e-01 8.85215402e-01 1.59454620e+00 1.18949883e-01
-3.22806448e-01 -2.13364810e-01 -3.01534086e-01 7.23264337e-01
9.34225619e-01 -1.74074918e-01 -3.17286462e-01 -4.88143504e-01
-5.31939790e-02 5.61512828e-01 4.65423584e-01 -9.73147631e-01
1.72971457e-01 -2.50652075e-01 1.32037431e-01 -3.49451721e-01
4.86140877e-01 -4.46344227e-01 3.81796032e-01 3.50210428e-01
-4.49323118e-01 2.83138633e-01 4.43129614e-02 4.95873839e-01
-3.00962389e-01 1.45089209e-01 1.05002344e+00 -9.94699895e-02
-7.44455516e-01 4.09361392e-01 -1.85544819e-01 1.90676108e-01
5.31233311e-01 5.62390275e-02 -6.69542611e-01 -6.90150499e-01
-6.09694719e-01 8.71092603e-02 7.55659521e-01 4.44820166e-01
9.29317117e-01 -1.38681591e+00 -9.14069712e-01 -1.49286687e-01
-5.84039629e-01 -9.95302573e-02 7.24720120e-01 6.46700323e-01
-6.26248300e-01 -1.48631305e-01 -3.32335740e-01 -3.60522538e-01
-1.35270905e+00 7.76525438e-01 2.31438607e-01 -2.86041379e-01
-1.08532763e+00 5.09554505e-01 5.42240143e-02 3.03355128e-01
1.30358726e-01 -1.35266319e-01 -2.25136787e-01 -2.96107978e-01
9.42570031e-01 6.51290834e-01 -5.35952806e-01 -6.40741348e-01
1.03028812e-01 4.95136291e-01 8.30668509e-02 -1.41773254e-01
1.47421980e+00 -3.38328302e-01 3.96484770e-02 3.77426475e-01
9.25477743e-01 -8.77465829e-02 -1.85250628e+00 -2.82867223e-01
-4.77028459e-01 -6.25325322e-01 4.12938446e-02 -3.90140176e-01
-1.49093843e+00 7.81744540e-01 5.50696552e-01 1.19428448e-01
1.58996487e+00 -2.04455838e-01 1.14536285e+00 -4.71835323e-02
2.61635989e-01 -1.05105197e+00 5.10310888e-01 3.02618474e-01
1.05508649e+00 -9.60191548e-01 2.75418404e-02 -2.62443185e-01
-7.44937658e-01 1.11295724e+00 6.44402087e-01 -2.59267867e-01
2.01005682e-01 5.10530412e-01 7.44389370e-02 1.96306840e-01
-1.06940448e+00 1.10626124e-01 3.69487368e-02 5.08531690e-01
1.70765266e-01 -1.55990422e-01 -4.83419597e-02 3.26783717e-01
-4.78246659e-02 2.45006248e-01 9.10385847e-01 4.91070926e-01
-3.14833790e-01 -6.56425357e-01 -2.94623226e-01 2.50250101e-02
-4.84835654e-01 -1.40020177e-01 2.56406426e-01 4.85231072e-01
-3.14511567e-01 5.94716132e-01 -1.41704932e-01 -5.54881930e-01
2.25922465e-01 -4.95073199e-01 2.40272835e-01 -1.14704873e-02
-2.85730451e-01 2.14303717e-01 -4.32670191e-02 -9.65085685e-01
-5.46188474e-01 -4.69518661e-01 -1.04927015e+00 -6.00935161e-01
-7.10136667e-02 -6.66804239e-02 2.69748509e-01 4.03293043e-01
4.65407908e-01 9.62410986e-01 8.83663058e-01 -1.27550495e+00
-5.63926995e-01 -8.89151931e-01 -4.40002710e-01 4.42659080e-01
6.55114830e-01 -4.18934792e-01 -5.40409982e-01 3.59802902e-01] | [11.089649200439453, -1.9186567068099976] |
82272371-ba28-459a-804b-df3f7734e105 | connecting-the-dots-between-audio-and-text-1 | null | null | https://openreview.net/forum?id=o1LLnx21iVj | https://openreview.net/pdf?id=o1LLnx21iVj | Connecting the Dots between Audio and Text without Parallel Data through Visual Knowledge Transfer | Machines that can represent and describe environmental soundscapes have practical potential, e.g., for audio tagging and captioning. Prevailing learning paradigms of audio-text connections have been relying on parallel audio-text data, which is, however, scarcely available on the web. We propose VIP-ANT that induces Audio-Text alignment without using any parallel audio-text data. Our key idea is to share the image modality between bi-modal image-text representations and bi-modal image-audio representations; the image modality functions as a pivot and connects audio and text in a tri-modal embedding space implicitly.In a difficult zero-shot setting with no paired audio-text data, our model demonstrates state-of-the-art zero-shot performance on the ESC50 and US8K audio classification tasks, and even surpasses the supervised state of the art for Clotho caption retrieval (with audio queries) by 2.2% R@1. We further investigate cases of minimal audio-text supervision, finding that, e.g., just a few hundred supervised audio-text pairs increase the zero-shot audio classification accuracy by 8% on US8K. However, to match human parity on some zero-shot tasks, our empirical scaling experiments suggest that we would need about $2^{21} \approx 2\text{M}$ supervised audio-caption pairs. Our work opens up new avenues for learning audio-text connections with little to no parallel audio-text data. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['audio-tagging'] | ['audio'] | [ 4.60200906e-01 6.53572083e-02 -1.42900869e-01 -2.32913867e-01
-1.53912795e+00 -6.43263638e-01 5.57010353e-01 2.48291507e-01
-4.91987765e-01 3.64445210e-01 3.35648954e-01 1.49225052e-02
-9.24816206e-02 -2.70328045e-01 -9.27728891e-01 -5.43955147e-01
-1.44155696e-01 4.51206714e-01 2.17952102e-01 -2.44722471e-01
-2.01913208e-01 -1.74450815e-01 -1.91149688e+00 6.59183621e-01
1.83149129e-01 1.40373957e+00 2.33723536e-01 1.04400051e+00
-1.48741990e-01 5.26753128e-01 -3.54884505e-01 -5.93487144e-01
1.98740900e-01 -3.48318011e-01 -8.41661096e-01 -2.69925177e-01
1.08185899e+00 -2.49433756e-01 -6.15875959e-01 7.15026557e-01
8.79159391e-01 7.12052733e-02 6.26455784e-01 -1.48092532e+00
-5.51371872e-01 8.21623802e-01 -5.16598642e-01 1.45126104e-01
5.22322714e-01 6.30248785e-02 1.52493477e+00 -7.81008720e-01
3.04517090e-01 9.69607830e-01 8.12205315e-01 5.03454685e-01
-1.38498175e+00 -9.81218457e-01 -1.07919686e-01 3.03750277e-01
-1.54693878e+00 -7.60835171e-01 8.35148573e-01 -4.34393674e-01
9.95191753e-01 3.46853375e-01 6.77335739e-01 1.23615766e+00
-1.27232939e-01 6.88852370e-01 7.16453552e-01 -5.13870120e-01
-1.14106476e-01 1.85597494e-01 -2.07220733e-01 4.68167812e-01
-4.16133195e-01 -5.74560501e-02 -1.37346101e+00 -1.26571029e-01
4.95454103e-01 -3.49219501e-01 -5.23663983e-02 -1.12926953e-01
-1.36906040e+00 6.54741228e-01 2.84352690e-01 3.00419837e-01
7.91511312e-02 6.49144530e-01 7.65176773e-01 6.56017900e-01
3.67735505e-01 3.77431840e-01 -3.62942129e-01 -3.50356996e-01
-1.07243145e+00 1.16449200e-01 5.03330708e-01 1.01911700e+00
6.48150623e-01 1.31226242e-01 8.44681263e-03 1.06755197e+00
7.77482912e-02 8.08907568e-01 5.52282572e-01 -9.90084648e-01
6.24610543e-01 -1.18390270e-01 -1.14420362e-01 -8.50034654e-01
-1.20285928e-01 -2.76942290e-02 -6.63340688e-01 -3.80658716e-01
3.41890693e-01 -6.49620546e-03 -6.14575326e-01 1.83762372e+00
3.59521946e-03 3.69372934e-01 -1.99041404e-02 8.03823590e-01
8.19438398e-01 7.63153851e-01 8.82614125e-03 -4.03220318e-02
1.65831387e+00 -7.30239987e-01 -4.50464189e-01 -1.86834335e-01
5.52677453e-01 -1.08412921e+00 1.60164988e+00 2.53342927e-01
-1.01854217e+00 -6.74010992e-01 -8.21399808e-01 -4.12480533e-02
-3.62395048e-01 6.79836422e-02 5.18362761e-01 4.08071786e-01
-9.93192017e-01 4.10371333e-01 -5.00931740e-01 -5.40787458e-01
-1.06028514e-02 4.19943362e-01 -5.02348781e-01 1.31595552e-01
-1.38594115e+00 3.51742744e-01 2.57834315e-01 -1.98160231e-01
-1.15204787e+00 -8.35490167e-01 -8.15440953e-01 5.68778887e-02
4.43639725e-01 -4.99703079e-01 1.34275603e+00 -1.00011218e+00
-1.37432969e+00 9.88736749e-01 6.75540939e-02 -6.06337547e-01
1.35071605e-01 -3.27490807e-01 -6.23469412e-01 6.07049406e-01
6.86767399e-02 1.30025733e+00 1.13398457e+00 -1.00233185e+00
-5.95652223e-01 -5.91973551e-02 1.39330789e-01 2.10959390e-01
-9.27337229e-01 1.06722869e-01 -4.12288010e-01 -7.47770429e-01
-6.05722032e-02 -1.05336320e+00 2.64883786e-01 2.64303029e-01
-3.34757805e-01 -1.15330897e-01 5.57923734e-01 -4.28257793e-01
1.00643873e+00 -2.40125251e+00 7.64804706e-03 5.46273477e-02
-8.36493224e-02 -7.27611408e-02 -5.02070367e-01 5.93914449e-01
-1.72320306e-01 -8.03878903e-03 -3.13789435e-02 -4.69543815e-01
1.85854629e-01 4.84702103e-02 -6.67053044e-01 2.18392596e-01
1.30304933e-01 7.59022057e-01 -8.02515388e-01 -7.33127415e-01
1.12804033e-01 5.42753577e-01 -7.18133867e-01 1.32267922e-01
-2.47778490e-01 1.74717337e-01 -2.81539448e-02 5.09859145e-01
-3.90653126e-02 -2.28240132e-01 -1.10537047e-02 -3.40585113e-01
-5.23178373e-03 3.68113786e-01 -1.09558320e+00 2.20485473e+00
-5.59577644e-01 9.37687516e-01 8.50709006e-02 -9.76716161e-01
6.63405359e-01 7.46388733e-01 7.13203132e-01 -6.65967643e-01
-2.46815123e-02 2.56522596e-01 -2.40196869e-01 -5.35773039e-01
5.61354935e-01 -2.73726732e-01 -4.02854353e-01 4.17934567e-01
4.37884063e-01 -3.45069528e-01 -1.28243789e-01 2.98279464e-01
1.03270304e+00 -1.48479268e-01 -2.20183730e-01 7.57752061e-02
1.23552278e-01 -1.70597330e-01 -5.18141985e-02 7.77462184e-01
7.54129738e-02 8.79279554e-01 3.25192899e-01 -6.50269678e-03
-1.31428945e+00 -1.04627562e+00 -1.02708392e-01 1.69937146e+00
6.19652905e-02 -7.74259984e-01 -5.93781769e-01 -2.79413581e-01
-8.79233032e-02 3.09636265e-01 -4.38146025e-01 -1.02865718e-01
-2.81987935e-01 -3.16157162e-01 1.29168797e+00 5.34854054e-01
1.18795104e-01 -9.46473539e-01 -2.42843390e-01 1.87503755e-01
-5.15955031e-01 -1.29733598e+00 -7.03424871e-01 4.46817189e-01
-3.97278965e-01 -6.03276908e-01 -9.72893417e-01 -9.35529828e-01
1.81889787e-01 2.84364611e-01 9.95243967e-01 -3.67020100e-01
-3.77729863e-01 9.23535109e-01 -4.84321922e-01 -1.59839869e-01
-2.03049317e-01 1.69654742e-01 3.64928156e-01 4.63725254e-02
8.63237083e-02 -8.44496429e-01 -5.67177892e-01 3.81016374e-01
-8.88215899e-01 -1.79055263e-03 4.53899711e-01 9.32868958e-01
5.83297670e-01 -1.98981613e-01 6.64464712e-01 -1.80796564e-01
2.35284761e-01 -3.49850059e-01 -2.89152473e-01 3.53572927e-02
-1.99958980e-01 -2.76237465e-02 6.07973337e-01 -9.66241360e-01
-5.35833597e-01 2.54761696e-01 -1.05016619e-01 -8.93575966e-01
-1.87005460e-01 3.17431301e-01 9.92678180e-02 4.73477244e-02
6.72381759e-01 1.97114065e-01 -1.67364672e-01 -5.58006465e-01
6.36753917e-01 1.01579595e+00 8.22977960e-01 -7.48445332e-01
7.66563535e-01 4.07327622e-01 -1.92672551e-01 -1.07506204e+00
-6.90378129e-01 -7.70074308e-01 -4.68899608e-01 -3.18121642e-01
9.94851053e-01 -1.20772529e+00 -8.32461417e-01 1.96891487e-01
-9.44392443e-01 -3.41471642e-01 -5.14946818e-01 5.46335280e-01
-8.98324251e-01 3.16302985e-01 -7.70867765e-01 -8.10326755e-01
-3.20462167e-01 -9.90582228e-01 1.52117956e+00 -4.31178987e-01
-3.59219670e-01 -3.81860763e-01 5.94018698e-02 5.44406712e-01
1.58777803e-01 -3.21077794e-01 7.66279995e-01 -6.88513100e-01
-3.34538072e-01 -1.79752201e-01 -1.05147675e-01 2.88985699e-01
-1.31527424e-01 -3.63884538e-01 -1.42898488e+00 -3.12447220e-01
-5.16046047e-01 -9.47779119e-01 8.40908647e-01 3.14903520e-02
1.32302284e+00 -2.32988268e-01 -2.63988879e-02 2.58533925e-01
1.06342924e+00 -8.52011368e-02 3.78895491e-01 -8.94265771e-02
6.93727255e-01 6.89570785e-01 5.49886525e-01 4.88375545e-01
2.13093430e-01 1.12381840e+00 3.85516763e-01 1.21883489e-02
-3.08986932e-01 -6.43319964e-01 5.62006235e-01 1.28024244e+00
3.88722420e-01 -2.44148061e-01 -8.92265499e-01 8.04559350e-01
-1.56294787e+00 -9.01427746e-01 2.26084128e-01 2.19132113e+00
1.05298519e+00 1.99141100e-01 2.72209138e-01 4.28165734e-01
6.38457537e-01 2.23490253e-01 -2.33442545e-01 -8.00737292e-02
1.14389043e-02 2.38729745e-01 3.01922321e-01 2.20300063e-01
-1.15032876e+00 8.40980649e-01 5.99552536e+00 1.21143496e+00
-1.24185765e+00 3.04745495e-01 3.86425197e-01 -5.94435453e-01
-2.99716204e-01 -1.61319181e-01 -6.61618114e-01 5.43680727e-01
1.29764235e+00 1.20134197e-01 5.94865799e-01 6.44293308e-01
-6.63312823e-02 1.75925076e-01 -1.49765003e+00 1.45022094e+00
3.04942787e-01 -1.15157199e+00 1.19208872e-01 4.30895090e-02
4.39628482e-01 1.38684079e-01 2.32918009e-01 2.54347056e-01
-3.34467366e-02 -8.94565225e-01 1.04985321e+00 1.42308891e-01
1.48195899e+00 -4.67737317e-01 4.28758323e-01 -8.44287220e-03
-1.56579602e+00 -6.57325387e-02 -1.41527072e-01 1.46975324e-01
2.76836306e-01 1.92786559e-01 -7.98641682e-01 3.66720885e-01
1.06807923e+00 5.05305827e-01 -2.52142400e-01 7.72402048e-01
2.19725832e-01 7.00553358e-01 -5.87220907e-01 9.18203816e-02
2.98157513e-01 3.77567291e-01 6.15470350e-01 1.25590038e+00
4.71512914e-01 -6.55721799e-02 3.14754874e-01 3.35492939e-01
-3.48365307e-01 3.49193841e-01 -8.16682458e-01 -4.34189975e-01
5.74989319e-01 9.13196027e-01 -6.34488165e-01 -3.29543293e-01
-4.09051329e-01 1.10726643e+00 -3.21451984e-02 6.33663610e-02
-8.72906446e-01 -5.80906749e-01 6.75307989e-01 3.09310168e-01
3.28495532e-01 -1.22660495e-01 7.79381022e-02 -8.60769689e-01
1.05020091e-01 -8.12884450e-01 3.80690694e-01 -1.17492127e+00
-1.41839993e+00 5.15188038e-01 1.10078767e-01 -1.53378797e+00
-3.57905686e-01 -4.35092509e-01 -2.51400083e-01 4.16712523e-01
-1.23157167e+00 -1.37145758e+00 -1.82410195e-01 8.21464777e-01
5.47517419e-01 -2.58718789e-01 1.05150664e+00 7.35552132e-01
2.41819844e-02 9.74249363e-01 1.07410088e-01 1.87424392e-01
1.05275035e+00 -9.45522666e-01 2.37455949e-01 2.48435244e-01
8.22980404e-01 2.76284605e-01 7.98374236e-01 -2.89555192e-01
-1.47928691e+00 -8.35500360e-01 8.94814253e-01 -3.68794531e-01
1.13860404e+00 -7.68551886e-01 -6.57702208e-01 5.71071208e-01
2.09339797e-01 4.38124649e-02 9.64092612e-01 2.51290143e-01
-7.56117105e-01 -5.95262170e-01 -6.56802356e-01 5.94669223e-01
1.04092455e+00 -1.34451759e+00 -5.37205994e-01 4.47991878e-01
1.20079160e+00 -1.50877550e-01 -1.02918363e+00 1.62748128e-01
8.32461417e-01 -2.93224335e-01 1.12366843e+00 -5.86975336e-01
4.58808422e-01 -2.03958377e-01 -6.19356871e-01 -9.78430629e-01
8.65122974e-02 -8.16747487e-01 2.69693047e-01 1.47482991e+00
3.96329075e-01 -1.71640381e-01 6.81657732e-01 -2.99508162e-02
-2.69185007e-01 -3.59286100e-01 -1.37861824e+00 -8.89807403e-01
-2.20216259e-01 -9.04783964e-01 2.64987409e-01 1.01350939e+00
3.17811042e-01 5.57341754e-01 -8.08544874e-01 8.32055509e-02
3.12994927e-01 -1.31968349e-01 7.00457096e-01 -1.14140725e+00
-6.94820285e-01 -2.03319937e-01 -4.80375141e-01 -1.18243194e+00
3.08543593e-01 -1.01038480e+00 2.05304265e-01 -9.34607506e-01
1.83248192e-01 -4.88387167e-01 -4.11245704e-01 7.58915544e-01
4.30137515e-01 8.80371988e-01 3.65450650e-01 1.40507519e-01
-8.20945144e-01 5.95529795e-01 7.44214594e-01 -4.88849998e-01
-1.17344163e-01 -4.01572704e-01 -4.95306969e-01 4.96106505e-01
5.44512987e-01 -6.00976050e-01 -4.33237195e-01 -5.30164719e-01
3.80907923e-01 2.49099761e-01 5.35533369e-01 -1.16909349e+00
3.76388937e-01 1.92139342e-01 1.14181124e-01 -4.17991430e-01
1.08794415e+00 -9.38137710e-01 1.77984312e-01 2.67598405e-02
-7.86447287e-01 -5.69502898e-02 3.41850489e-01 7.80131876e-01
-4.18826193e-01 -2.48205498e-01 4.75424021e-01 -4.96500172e-03
-3.99489552e-01 1.73547670e-01 -4.11258489e-01 1.15284383e-01
6.35984838e-01 -1.13765597e-01 -1.85948014e-01 -6.81858480e-01
-8.06819975e-01 7.02659041e-02 7.88059980e-02 5.34242392e-01
4.40329015e-01 -1.57185459e+00 -5.38114846e-01 9.24520344e-02
5.27489007e-01 -4.10141736e-01 3.27257097e-01 6.50788844e-01
3.55026647e-02 5.46885252e-01 -9.36715826e-02 -8.29880476e-01
-1.59193289e+00 2.17081442e-01 1.11768812e-01 1.53821185e-01
-5.52765667e-01 1.03490233e+00 2.06178308e-01 -7.93791637e-02
6.84184849e-01 -7.24169835e-02 1.59341037e-01 3.98745775e-01
4.07140225e-01 9.97220278e-02 6.97274953e-02 -7.74349153e-01
-2.21039459e-01 7.71749437e-01 2.91983038e-02 -5.93406439e-01
1.14468253e+00 -5.85669987e-02 2.50938982e-01 1.00248122e+00
1.27444208e+00 -1.15409851e-01 -9.77404237e-01 -3.63726050e-01
-2.95420825e-01 -2.47960448e-01 -7.39614144e-02 -5.60413122e-01
-8.00794005e-01 1.29179275e+00 8.04544747e-01 2.71368444e-01
1.07083189e+00 4.73356724e-01 1.02041459e+00 6.83179498e-01
2.48430401e-01 -1.16333544e+00 6.89589500e-01 3.37680101e-01
7.96101928e-01 -1.14661956e+00 -4.19242084e-01 -2.07171082e-01
-6.34414554e-01 7.35173881e-01 4.79143828e-01 1.34385064e-01
5.23850322e-01 2.21326083e-01 -3.13975662e-02 -1.50732920e-01
-8.96330118e-01 -2.58759379e-01 4.09834713e-01 5.70023894e-01
4.14974988e-01 -1.60203371e-02 3.81598085e-01 5.76397300e-01
-5.70820093e-01 -3.60108763e-01 7.04076961e-02 6.66703105e-01
-4.08073604e-01 -9.21721697e-01 -3.45508307e-01 4.38034803e-01
-6.21960580e-01 -4.43263859e-01 -4.14573640e-01 5.09374321e-01
-2.08466798e-02 9.53575492e-01 3.98783416e-01 -6.49728060e-01
2.49160126e-01 4.43498820e-01 4.11897808e-01 -4.66301531e-01
-6.80206537e-01 3.75260293e-01 1.04555354e-01 -4.95440722e-01
-3.40342909e-01 -4.72297847e-01 -1.01500344e+00 -2.21683294e-01
-4.01268184e-01 2.88249969e-01 6.60396039e-01 6.54110551e-01
3.43442321e-01 4.82270658e-01 4.61801171e-01 -1.07051885e+00
-3.61014843e-01 -8.97396743e-01 -5.91411352e-01 4.34244812e-01
4.67011720e-01 -5.19228458e-01 -4.43875968e-01 3.94751668e-01] | [15.261110305786133, 5.0199360847473145] |
a87c0e42-9318-4c3b-95b8-027c3bdab91e | group-invariant-global-pooling | 2305.19207 | null | https://arxiv.org/abs/2305.19207v1 | https://arxiv.org/pdf/2305.19207v1.pdf | Group Invariant Global Pooling | Much work has been devoted to devising architectures that build group-equivariant representations, while invariance is often induced using simple global pooling mechanisms. Little work has been done on creating expressive layers that are invariant to given symmetries, despite the success of permutation invariant pooling in various molecular tasks. In this work, we present Group Invariant Global Pooling (GIGP), an invariant pooling layer that is provably sufficiently expressive to represent a large class of invariant functions. We validate GIGP on rotated MNIST and QM9, showing improvements for the latter while attaining identical results for the former. By making the pooling process group orbit-aware, this invariant aggregation method leads to improved performance, while performing well-principled group aggregation. | ['Pietro Liò', 'Chaitanya K. Joshi', 'Yonatan Gideoni', 'Kamil Bujel'] | 2023-05-30 | null | null | null | null | ['rotated-mnist'] | ['computer-vision'] | [ 4.75552976e-01 3.87909919e-01 -5.58812581e-02 -4.55725789e-01
-5.32350540e-01 -6.72618091e-01 1.01590312e+00 -1.45080969e-01
-3.27143103e-01 8.64003479e-01 6.12814486e-01 1.78898778e-02
-3.97174984e-01 -6.57677948e-01 -6.25879467e-01 -8.50241303e-01
-5.60630918e-01 1.95648402e-01 2.55568564e-01 -2.75811076e-01
2.38729194e-01 1.00389814e+00 -1.31929314e+00 5.78248620e-01
3.25871617e-01 6.22353911e-01 -2.37419575e-01 7.03500569e-01
3.62681776e-01 6.95752800e-01 -4.03862983e-01 -3.34071398e-01
5.48113048e-01 -4.80086386e-01 -1.06917560e+00 -2.20406860e-01
8.27103376e-01 1.05065428e-01 -3.20045888e-01 6.69788718e-01
6.98061287e-01 5.38637161e-01 6.11469507e-01 -9.00406599e-01
-7.70622492e-01 5.64271152e-01 -4.50301245e-02 1.99917078e-01
1.41430184e-01 1.30131558e-01 1.55170143e+00 -4.16474134e-01
1.03390121e+00 1.43690979e+00 5.09579659e-01 6.08953178e-01
-1.68883264e+00 -5.05418658e-01 -3.12343426e-02 -2.78587222e-01
-1.15967345e+00 -5.88112772e-01 7.02291250e-01 -9.98124778e-02
1.43840635e+00 5.11098683e-01 3.73433471e-01 8.42909694e-01
7.24578202e-01 3.03239435e-01 1.11930037e+00 4.60803211e-02
2.35068068e-01 -4.87191826e-01 1.56238943e-01 6.54846728e-01
4.12678957e-01 -2.46372595e-01 -8.17121744e-01 -2.27620184e-01
8.74919772e-01 -8.37720707e-02 -2.85514981e-01 -5.62284827e-01
-1.37256074e+00 6.11198664e-01 8.37518752e-01 4.81383383e-01
-3.39456797e-01 4.97321546e-01 4.58396494e-01 5.44863224e-01
2.55568564e-01 1.42533743e+00 -4.76251990e-01 2.48978004e-01
-9.19963896e-01 6.31383598e-01 6.83315635e-01 7.71928430e-01
7.89493322e-01 9.75146592e-02 -6.19761646e-01 4.26166862e-01
-1.24580890e-01 -6.56175315e-02 4.37964469e-01 -9.61794913e-01
9.90423467e-03 6.71362996e-01 -5.79203255e-02 -1.05878568e+00
-9.21578586e-01 -4.88299966e-01 -7.55945444e-01 3.04725260e-01
3.15479815e-01 1.52351350e-01 -1.05171824e+00 2.14800763e+00
-1.09915942e-01 -4.32656854e-01 1.06740512e-01 4.96061027e-01
6.80420399e-01 4.44910139e-01 2.14261010e-01 1.64223015e-01
1.31066704e+00 -6.30967021e-01 -3.99828047e-01 9.52099934e-02
5.75247526e-01 -5.35119653e-01 6.78518414e-01 1.97560087e-01
-1.06549323e+00 -3.79017025e-01 -1.17733526e+00 -4.57050592e-01
-6.64693415e-01 -3.97597879e-01 1.30920720e+00 5.53364456e-01
-1.28683305e+00 1.08066928e+00 -8.51470411e-01 -2.90580124e-01
6.27918720e-01 1.01499569e+00 -1.01378810e+00 2.35924914e-01
-9.61265445e-01 8.04505587e-01 7.75526166e-01 -2.47366920e-01
-6.07771754e-01 -9.25400615e-01 -6.12903357e-01 1.61743000e-01
-5.32463193e-02 -8.19593608e-01 8.30133975e-01 -7.87497222e-01
-1.60770798e+00 8.04449677e-01 -4.78049964e-02 -6.46589518e-01
3.32392246e-01 2.95829982e-01 -2.06448510e-01 1.89461872e-01
-3.01587403e-01 1.19813430e+00 7.49413371e-01 -6.64971352e-01
1.86459452e-01 -4.39462781e-01 2.96043813e-01 -2.22247168e-02
-2.21241936e-01 2.51107216e-01 3.94196101e-02 -7.10728586e-01
3.10320646e-01 -1.08094764e+00 -4.39430296e-01 -1.86802074e-01
-4.62484598e-01 -3.35889637e-01 5.43303907e-01 -3.84327501e-01
8.53178501e-01 -1.96609449e+00 4.85554099e-01 3.19562882e-01
5.15465915e-01 1.85850695e-01 -3.00446212e-01 6.31329954e-01
-3.49234372e-01 5.16653478e-01 -3.43467027e-01 -2.81403363e-01
2.56058276e-01 3.97821009e-01 -4.79111671e-01 4.27834392e-01
5.31540632e-01 1.20867729e+00 -5.92589080e-01 -6.77698702e-02
-1.58143923e-01 6.24232173e-01 -1.14688587e+00 -2.36959457e-01
-1.59612492e-01 5.00847697e-01 -5.06802857e-01 4.71577793e-01
6.38803720e-01 -1.99273050e-01 4.03057158e-01 -1.34955302e-01
-2.19460025e-01 3.84448618e-01 -9.09905136e-01 1.66100836e+00
-1.16679154e-01 5.57331026e-01 -2.32833520e-01 -6.67197466e-01
1.00154698e+00 1.76766247e-01 5.23863971e-01 -4.22455907e-01
5.18486910e-02 2.23582670e-01 4.55500573e-01 9.05009210e-02
7.71587610e-01 -3.32862824e-01 -1.53368726e-01 3.53975415e-01
7.39263475e-01 -3.81019950e-01 3.08272690e-01 8.92787576e-02
1.19135666e+00 1.77678883e-01 3.93927574e-01 -8.87907863e-01
4.52676266e-01 -5.00483990e-01 3.41744453e-01 8.67229104e-01
-8.28204080e-02 8.03909004e-01 7.75287926e-01 -9.01486754e-01
-1.11617160e+00 -8.65007579e-01 -4.85184878e-01 1.19724405e+00
-2.66123772e-01 -8.23898494e-01 -6.56258106e-01 -3.36440057e-01
-6.28610477e-02 9.00906920e-02 -9.81837153e-01 -3.11492860e-01
-6.74389243e-01 -1.13541985e+00 9.37798679e-01 4.71343637e-01
7.32787669e-01 -1.16275549e+00 -5.22236943e-01 2.58574337e-01
4.74669784e-01 -6.63165331e-01 -2.12954327e-01 6.45522296e-01
-7.29084253e-01 -9.58634973e-01 -5.04324436e-01 -3.62902075e-01
7.11791158e-01 9.09663085e-03 8.95222723e-01 -2.61268675e-01
-2.89672017e-01 -7.50504136e-02 9.46780443e-02 -5.00996709e-01
-2.09418356e-01 4.67780322e-01 2.45498553e-01 -1.12351194e-01
1.98903292e-01 -8.42110157e-01 -4.67143446e-01 4.13794100e-01
-1.17596340e+00 2.18692683e-02 5.23769438e-01 5.61944485e-01
6.25312567e-01 -3.11722696e-01 6.00095332e-01 -7.45837688e-01
6.77863955e-01 7.73124173e-02 -3.37302268e-01 -1.03612803e-01
-1.11732624e-01 6.99876785e-01 8.36352229e-01 -1.82207033e-01
-6.43067956e-01 -9.75217670e-02 -1.71011895e-01 1.43998161e-01
8.96991324e-03 2.98511624e-01 -1.83248580e-01 -6.79007649e-01
9.02270794e-01 -1.16251551e-01 9.04392898e-02 -4.13296193e-01
3.70752424e-01 1.07786633e-01 4.50142950e-01 -8.98917437e-01
6.43666804e-01 7.12034166e-01 6.57522857e-01 -9.55643773e-01
-6.74027026e-01 -1.84266627e-01 -8.24605227e-01 4.61962253e-01
1.09655964e+00 -5.64213157e-01 -6.50975764e-01 3.33337605e-01
-7.67066538e-01 -2.17940167e-01 -5.87279439e-01 2.79196650e-01
-8.91636133e-01 1.87640980e-01 -3.75213116e-01 -2.50403821e-01
-3.34514260e-01 -1.25471306e+00 1.06162953e+00 -3.07925958e-02
-6.50458097e-01 -8.45834732e-01 3.27376872e-01 -1.29572749e-01
8.20332587e-01 6.99942172e-01 8.27205598e-01 -9.02096033e-01
-8.67337584e-01 -3.73233676e-01 -1.15467153e-01 3.32783818e-01
-2.23777883e-04 -5.44361165e-03 -1.24875963e+00 -3.92610133e-01
-4.72721994e-01 -3.14761132e-01 1.30093408e+00 1.21866934e-01
1.32462072e+00 -2.98861295e-01 -2.19476894e-02 8.78584146e-01
9.73792911e-01 -1.47714242e-01 8.50694835e-01 1.80451617e-01
6.59859359e-01 4.78172839e-01 -4.08707321e-01 -1.95643995e-02
-1.60605237e-01 8.71639550e-01 1.17916539e-01 -3.40912491e-02
-2.24011406e-01 -2.58016616e-01 4.82860714e-01 5.31080484e-01
-8.36090028e-01 -3.91799025e-02 -5.85390389e-01 1.46015957e-01
-1.71737611e+00 -1.26988351e+00 3.37523431e-01 1.91240489e+00
7.56673157e-01 1.36005655e-01 -8.77793431e-02 -2.09669210e-02
8.10685009e-03 6.37172639e-01 -2.95978546e-01 -8.23294222e-01
-5.63373268e-01 7.32090890e-01 5.89900792e-01 3.77781242e-01
-1.33108318e+00 8.46674144e-01 7.80369997e+00 6.87523127e-01
-1.13300169e+00 -2.55464226e-01 3.99477959e-01 -3.47534101e-03
-4.64909494e-01 6.75462335e-02 -6.85093999e-01 1.38642490e-02
9.02238190e-01 -2.73476213e-01 3.88992578e-01 5.76808512e-01
-1.06622487e-01 4.42393482e-01 -1.24537110e+00 5.49269199e-01
2.46141199e-02 -1.59967101e+00 5.44759870e-01 4.02430117e-01
9.82459903e-01 1.53594762e-01 1.41396075e-01 1.35474026e-01
2.65907854e-01 -1.37701392e+00 3.30349088e-01 6.46063387e-01
5.15351892e-01 -9.13885772e-01 4.49591547e-01 -2.44491473e-01
-9.70937848e-01 1.01459466e-01 -8.13089788e-01 -2.31944874e-01
-3.15889597e-01 2.25383282e-01 -5.94780087e-01 7.19878316e-01
4.54786330e-01 6.93234682e-01 -7.14646101e-01 8.07713866e-01
-2.72374421e-01 3.14342082e-01 -4.60830361e-01 2.05572680e-01
5.74314296e-01 -9.10878107e-02 7.88950324e-01 1.27158833e+00
5.57653494e-02 -1.91950545e-01 8.64326432e-02 9.28976834e-01
-6.59699738e-01 1.29714414e-01 -9.82229590e-01 -2.46019393e-01
-3.32257092e-01 1.40443373e+00 -8.77682507e-01 -1.74013451e-01
1.16704656e-02 7.45537162e-01 4.20629680e-01 2.41080865e-01
-4.71113652e-01 -5.69888413e-01 9.55512881e-01 -1.01569377e-01
3.96107405e-01 -5.72915375e-01 -7.52874929e-03 -1.16297340e+00
-2.27053836e-01 -7.29080796e-01 4.01105553e-01 -5.81199110e-01
-1.06830418e+00 5.58733821e-01 -4.08546813e-02 -6.50811315e-01
-1.26756594e-01 -1.17536128e+00 -6.28974676e-01 7.50922084e-01
-1.15207922e+00 -1.34114337e+00 8.42121169e-02 3.62779647e-01
-1.36583999e-01 -3.72905023e-02 1.28285980e+00 7.28776902e-02
-3.54091078e-01 6.69402122e-01 -1.97305158e-01 -1.69969738e-01
7.17890263e-01 -1.31660736e+00 4.33748841e-01 8.20181429e-01
2.24107459e-01 1.36697376e+00 6.92049384e-01 -2.60788918e-01
-1.41110587e+00 -1.07173204e+00 8.33430886e-01 -8.32188904e-01
5.64702451e-01 -6.97903752e-01 -9.06248510e-01 8.86256695e-01
1.75729945e-01 1.10839806e-01 8.10301244e-01 2.75793821e-01
-8.49692881e-01 1.45268977e-01 -1.13442874e+00 7.33180940e-01
1.37123477e+00 -7.20055282e-01 -5.25425196e-01 4.22266662e-01
8.35600734e-01 -1.92913115e-01 -1.26287258e+00 4.47818130e-01
6.24130487e-01 -7.88550079e-01 1.02525043e+00 -1.26948631e+00
2.37940431e-01 -5.56115568e-01 -3.05317342e-01 -1.18307686e+00
-8.37515771e-01 -1.02776825e+00 1.31536409e-01 6.98093593e-01
4.18865353e-01 -1.01381886e+00 4.46861267e-01 5.52128911e-01
-4.88391340e-01 -5.66540241e-01 -1.02616119e+00 -9.08684194e-01
3.64087194e-01 8.34866718e-04 8.71574163e-01 9.60239768e-01
1.11802019e-01 5.12800291e-02 -2.96689242e-01 -7.37362653e-02
2.55456436e-02 1.05535440e-01 7.99174130e-01 -1.17134345e+00
-3.15845549e-01 -7.16568649e-01 -1.10427177e+00 -4.76508677e-01
2.01708540e-01 -1.55295789e+00 -3.78369063e-01 -1.13492823e+00
3.46166044e-01 1.43921077e-01 -7.08168685e-01 9.28196609e-01
1.06435359e-01 7.47794449e-01 2.83805162e-01 8.83932114e-02
-7.82617033e-01 3.70277226e-01 1.38905251e+00 4.72037122e-03
-1.01571016e-01 -4.84685600e-01 -1.15572011e+00 6.13431692e-01
1.01279581e+00 -1.08850099e-01 -2.30313033e-01 -7.69430175e-02
4.81757581e-01 -1.01756597e+00 4.51193988e-01 -1.21067858e+00
-1.31650165e-01 1.03999116e-01 5.40564299e-01 -5.90406582e-02
2.28366852e-01 -3.63320023e-01 2.99726337e-01 3.76704425e-01
-5.43838620e-01 9.24609751e-02 3.88283670e-01 2.16483623e-01
-1.42186850e-01 8.48171264e-02 8.51665258e-01 -1.76086590e-01
-6.55064702e-01 6.36998296e-01 -2.22280264e-01 -8.79795998e-02
6.55373991e-01 2.22543664e-02 -5.53139687e-01 -7.10328743e-02
-7.39633560e-01 -1.71864465e-01 6.11802220e-01 3.16415757e-01
6.77225217e-02 -1.23440695e+00 -5.96145749e-01 4.25740093e-01
2.10215271e-01 -2.15828717e-01 5.58654703e-02 6.92886233e-01
-5.94492376e-01 6.86349213e-01 -7.68561184e-01 -2.27786213e-01
-1.01785946e+00 4.76047784e-01 5.13866901e-01 -5.32007337e-01
-5.93025923e-01 6.54431522e-01 3.27969462e-01 -5.95809639e-01
-2.73769438e-01 -6.14911914e-01 -8.84272158e-03 7.67285749e-02
6.81699336e-01 2.42059261e-01 2.86107361e-01 -5.57883501e-01
-3.93231332e-01 5.40855646e-01 -1.21980749e-01 1.69021890e-01
1.51833379e+00 7.01988935e-01 -4.28713679e-01 -1.82921141e-02
1.39874983e+00 2.51985431e-01 -1.10261667e+00 2.43447870e-01
-1.56870410e-01 -7.46809840e-02 -1.44967601e-01 -5.38064718e-01
-6.51236653e-01 7.19982803e-01 7.99124613e-02 2.75770515e-01
7.50708759e-01 -8.74346122e-02 2.45733395e-01 1.10211039e+00
4.81728137e-01 -6.89211309e-01 -1.08720072e-01 6.23343766e-01
1.20487928e+00 -7.57115424e-01 3.52221072e-01 -1.69140220e-01
-2.66550153e-01 1.08138847e+00 1.43628865e-01 -5.47068059e-01
3.22599143e-01 1.07676648e-02 -4.27183688e-01 -4.97425139e-01
-4.81573433e-01 -1.29255250e-01 6.49262369e-01 3.77058536e-01
6.20030701e-01 3.02186012e-01 -2.79010296e-01 2.99900472e-01
-5.42838812e-01 -1.61081269e-01 2.35735908e-01 9.08834398e-01
-4.48178440e-01 -1.28907871e+00 4.10146564e-02 3.81779999e-01
-7.02496827e-01 -1.45047858e-01 -7.54454494e-01 9.89131570e-01
5.07620573e-02 2.14400575e-01 1.24087147e-01 -6.99368045e-02
2.05504015e-01 4.62010056e-01 7.91183054e-01 -5.26553452e-01
-6.67337298e-01 -4.00215089e-01 2.43816659e-01 -8.44813108e-01
-6.05518997e-01 -5.52724540e-01 -1.26540506e+00 -3.09374720e-01
2.14194864e-01 -4.77502383e-02 3.75319451e-01 5.50369799e-01
7.07045019e-01 6.69954300e-01 2.29739830e-01 -1.20642781e+00
-3.51025254e-01 -7.76074231e-01 -4.97005612e-01 6.66997850e-01
3.15677345e-01 -6.26615524e-01 -2.40104467e-01 -3.92932355e-01] | [8.762832641601562, 2.5842156410217285] |
d8934b9b-a4df-47f0-9442-514a0c80f8cf | fast-preprocessing-for-robust-face-sketch | 1708.00224 | null | http://arxiv.org/abs/1708.00224v1 | http://arxiv.org/pdf/1708.00224v1.pdf | Fast Preprocessing for Robust Face Sketch Synthesis | Exemplar-based face sketch synthesis methods usually meet the challenging
problem that input photos are captured in different lighting conditions from
training photos. The critical step causing the failure is the search of similar
patch candidates for an input photo patch. Conventional illumination invariant
patch distances are adopted rather than directly relying on pixel intensity
difference, but they will fail when local contrast within a patch changes. In
this paper, we propose a fast preprocessing method named Bidirectional
Luminance Remapping (BLR), which interactively adjust the lighting of training
and input photos. Our method can be directly integrated into state-of-the-art
exemplar-based methods to improve their robustness with ignorable computational
cost. | ['Linchao Bao', 'Yibing Song', 'Qingxiong Yang', 'Jiawei Zhang'] | 2017-08-01 | null | null | null | null | ['face-sketch-synthesis'] | ['computer-vision'] | [ 5.37759006e-01 -4.85950351e-01 4.76598740e-02 -3.95380020e-01
-4.78304356e-01 -5.34995914e-01 5.91355741e-01 -4.58190739e-01
-6.20641299e-02 8.33523452e-01 -2.82890588e-01 -3.25417668e-02
3.10014412e-02 -8.49250317e-01 -7.56997168e-01 -6.81892633e-01
6.26296103e-01 -2.78536715e-02 2.43424281e-01 -5.63824698e-02
5.66370785e-01 8.82950425e-01 -1.68784976e+00 4.16582853e-01
7.08391547e-01 8.05476904e-01 3.36310305e-02 4.03457284e-01
-4.38040018e-01 1.24602988e-01 -6.50791705e-01 -5.08221626e-01
4.56983447e-01 -8.53983462e-01 -3.91615033e-01 3.36313874e-01
9.56701636e-01 -8.77851918e-02 -2.58496672e-01 1.16601694e+00
4.54853505e-01 4.86204252e-02 7.34577060e-01 -1.31210554e+00
-9.50335801e-01 -4.69266027e-02 -7.56193817e-01 -1.75765544e-01
4.01574552e-01 1.04156986e-01 4.71067101e-01 -1.37525427e+00
8.13336313e-01 1.19209003e+00 5.28772712e-01 8.28444779e-01
-1.44655049e+00 -7.53715754e-01 1.79113775e-01 3.80903780e-01
-1.79308009e+00 -6.85761631e-01 1.32674611e+00 -1.27265036e-01
4.52032804e-01 5.20157993e-01 6.94966793e-01 8.86581063e-01
-9.97468755e-02 4.40807253e-01 1.36645234e+00 -4.68844980e-01
2.52182245e-01 1.81253433e-01 -5.25991321e-01 7.83198774e-01
7.80639201e-02 -1.50505058e-03 -6.40600920e-01 -1.75905213e-01
1.23071158e+00 2.34664142e-01 -4.52521712e-01 -3.65552455e-01
-1.10190797e+00 4.74435925e-01 3.75395834e-01 2.15016931e-01
-1.80652618e-01 3.91662419e-02 -6.36642575e-02 4.34936523e-01
2.25573465e-01 2.89871275e-01 -1.45628557e-01 3.18934828e-01
-1.18132544e+00 1.06989823e-01 3.82137388e-01 1.02689469e+00
1.05543876e+00 7.82929957e-02 -1.56451792e-01 9.07311618e-01
1.55459911e-01 5.18387020e-01 2.09881753e-01 -9.33315277e-01
1.41748786e-01 5.96678793e-01 1.48671746e-01 -1.20770729e+00
1.57210320e-01 1.63310736e-01 -6.79069161e-01 5.15627503e-01
4.35946733e-01 2.78084606e-01 -1.12147999e+00 1.33810627e+00
4.83247072e-01 4.53246653e-01 -1.94156304e-01 9.11118925e-01
8.19474638e-01 6.51951492e-01 -2.78015286e-01 -3.95222932e-01
1.01046586e+00 -8.72228742e-01 -8.16103995e-01 -2.99250521e-02
-2.73056746e-01 -1.15725744e+00 1.43305373e+00 4.99549091e-01
-1.32471740e+00 -7.07469940e-01 -1.25487804e+00 -6.82715923e-02
-3.90866041e-01 1.97888091e-01 2.74043858e-01 7.01413870e-01
-9.99658823e-01 6.49122715e-01 -2.23445863e-01 -2.31526062e-01
4.48011041e-01 3.62519979e-01 -5.03419995e-01 -2.90381759e-01
-6.53622925e-01 7.09572375e-01 -1.70692161e-01 3.06028485e-01
-6.28613949e-01 -7.19191253e-01 -4.76378381e-01 -1.88083082e-01
2.57077992e-01 -4.28647161e-01 6.97810233e-01 -1.30560029e+00
-1.93919098e+00 9.05241549e-01 -4.04428720e-01 2.21376777e-01
5.28777957e-01 1.65406138e-01 -4.27252114e-01 1.72065854e-01
-3.38597029e-01 6.80346429e-01 1.62892830e+00 -1.57872427e+00
-3.17380011e-01 -2.27961645e-01 -2.58520305e-01 7.83903226e-02
-1.03697360e-01 -5.27440608e-02 -6.86871886e-01 -8.94021094e-01
2.09604234e-01 -5.96852124e-01 5.75083196e-02 4.90738422e-01
-1.30925000e-01 -6.46679476e-02 1.24390399e+00 -3.33399415e-01
1.04285610e+00 -2.19653630e+00 4.90006581e-02 3.55187356e-01
-2.32687190e-01 4.29054320e-01 -3.34986061e-01 4.30023253e-01
-3.24007049e-02 4.10169773e-02 -3.76186699e-01 -2.12914065e-01
-1.85824513e-01 2.14682460e-01 -5.84206462e-01 5.76233149e-01
4.84494030e-01 7.25556552e-01 -7.43939161e-01 -7.28290021e-01
3.92028123e-01 7.63273776e-01 -3.30141276e-01 2.13462949e-01
-3.18137445e-02 5.97435772e-01 -7.66920671e-02 8.59765768e-01
1.01034617e+00 1.90530941e-01 -1.41905649e-02 -4.61897373e-01
-1.69195682e-01 -6.63394481e-02 -1.38620555e+00 1.80796051e+00
-3.79876256e-01 7.19292641e-01 -3.68264206e-02 -6.80717885e-01
1.22078431e+00 1.74789771e-01 1.45210132e-01 -7.09961712e-01
-8.79370347e-02 2.08178774e-01 -2.91339397e-01 -2.64575243e-01
2.97820181e-01 -1.13038681e-01 4.81510848e-01 3.44361812e-01
-2.84005433e-01 -4.33725119e-01 1.16969896e-02 -2.56507516e-01
6.58054888e-01 4.57600266e-01 2.11182296e-01 -1.44411594e-01
8.28614771e-01 -2.75425136e-01 7.23293841e-01 4.91111189e-01
-1.19176067e-01 1.11696208e+00 1.73598737e-01 -5.87955594e-01
-1.18993044e+00 -1.08756042e+00 -4.19773936e-01 6.40929401e-01
3.63625467e-01 -2.75129765e-01 -8.08634162e-01 -8.27598035e-01
-4.82316129e-02 2.58102894e-01 -7.43551075e-01 5.37540056e-02
-9.07266140e-01 -2.93864012e-01 3.10940176e-01 2.59338945e-01
5.74074328e-01 -1.12595856e+00 -3.95793110e-01 -1.98785197e-02
2.11880088e-01 -7.34703839e-01 -7.82965064e-01 -4.99653310e-01
-6.24625504e-01 -1.02704465e+00 -8.89900446e-01 -8.67012262e-01
1.25698221e+00 5.78641653e-01 9.03345823e-01 4.35664892e-01
-5.62175214e-01 3.79533827e-01 -2.31075883e-01 -4.13296632e-02
-2.33008444e-01 -4.37484026e-01 -1.34320082e-02 4.31210518e-01
4.69343178e-02 -5.95546544e-01 -1.04940057e+00 8.21859419e-01
-1.01865995e+00 1.15034118e-01 4.28788632e-01 9.08221245e-01
1.02278543e+00 -1.15466677e-01 5.80368459e-01 -7.46858954e-01
4.20216262e-01 4.68370281e-02 -5.12755334e-01 6.80637002e-01
-4.22157377e-01 -2.00063676e-01 7.85695493e-01 -7.44276583e-01
-9.87783074e-01 2.61755228e-01 1.36990964e-01 -6.99700832e-01
-2.04692245e-01 -1.69443548e-01 -4.19586241e-01 -5.88606119e-01
4.40463185e-01 3.56017709e-01 -2.42420197e-01 -3.21469098e-01
5.34491301e-01 5.15615642e-01 5.37866592e-01 -5.27352810e-01
1.01235485e+00 6.92145646e-01 2.21018955e-01 -9.88112926e-01
-1.27349272e-01 6.46147039e-03 -7.60113716e-01 -4.25530463e-01
4.03623283e-01 -5.03790736e-01 -5.58785677e-01 2.65215993e-01
-1.11967623e+00 -2.09813401e-01 -2.67590851e-01 3.50106955e-02
-3.43079507e-01 4.07861739e-01 -8.51210281e-02 -6.87103212e-01
-2.28342250e-01 -9.94795322e-01 1.07758999e+00 4.98366743e-01
1.25405177e-01 -6.59093022e-01 -6.45064712e-02 -3.43906805e-02
4.14740831e-01 4.92804125e-02 7.67899334e-01 1.34178177e-01
-7.03120410e-01 -2.58768976e-01 -4.34431404e-01 2.34997615e-01
5.92507958e-01 5.89918554e-01 -9.85281527e-01 -3.59440207e-01
-1.91841945e-01 4.90794592e-02 6.09593332e-01 1.31705627e-01
1.32297492e+00 -3.39775771e-01 -3.58963400e-01 7.16995001e-01
1.48540771e+00 3.14570427e-01 9.52296317e-01 -5.23919575e-02
7.81085849e-01 6.12584054e-01 5.53962171e-01 1.15677044e-01
-1.36267737e-01 8.65739107e-01 -4.22932357e-02 -3.45608264e-01
-5.43142438e-01 -4.87631083e-01 1.48251995e-01 6.06228828e-01
2.83192694e-02 -6.31171390e-02 -5.20428717e-01 3.54474574e-01
-1.69134748e+00 -1.00166333e+00 1.15557373e-01 2.56790257e+00
8.53824317e-01 -2.91876972e-01 -9.13475528e-02 2.74541914e-01
8.12987089e-01 5.53800464e-02 -4.76953268e-01 -5.20376801e-01
-1.86621696e-01 3.96225780e-01 2.72697002e-01 5.15027165e-01
-7.45127320e-01 9.03004169e-01 6.84339380e+00 8.91890228e-01
-1.33029544e+00 -1.60116449e-01 5.92044830e-01 -7.74379000e-02
-5.11113584e-01 9.10067260e-02 -5.86704969e-01 5.03684402e-01
2.46543407e-01 1.69526443e-01 7.28601754e-01 6.11872971e-01
4.32937555e-02 -2.10717425e-01 -1.09141624e+00 1.40722966e+00
4.01500136e-01 -1.30654991e+00 2.76069611e-01 -2.56793588e-01
9.56046522e-01 -7.41140008e-01 3.56042355e-01 -1.42073184e-01
-3.74546736e-01 -9.94764328e-01 4.75165933e-01 7.60460675e-01
1.04391253e+00 -7.67069340e-01 -1.04651293e-02 -1.12096518e-01
-1.23964238e+00 2.13357285e-01 -5.99974155e-01 2.35001028e-01
-3.99373956e-02 2.90221334e-01 -5.60999930e-01 2.68660784e-01
5.69783509e-01 5.12823761e-01 -6.65165484e-01 1.04842210e+00
-3.95216942e-01 3.36838901e-01 -1.79404065e-01 1.13388859e-01
-1.36381522e-01 -4.85979855e-01 4.73152399e-01 8.80733609e-01
3.77834111e-01 1.54910088e-01 -1.42743930e-01 1.06876445e+00
-1.02580965e-01 3.28663707e-01 -8.12502503e-01 2.39744470e-01
6.25382125e-01 1.31266856e+00 -8.50103140e-01 -2.06512257e-01
-5.65788388e-01 1.52790296e+00 2.52712250e-01 5.75519085e-01
-7.67751694e-01 -4.97964412e-01 5.24832129e-01 2.56584406e-01
5.18155336e-01 -1.98587775e-01 -3.33966076e-01 -9.60269034e-01
1.47399932e-01 -8.24781001e-01 -2.07624827e-02 -8.56773078e-01
-1.46741307e+00 5.24014890e-01 -2.48143762e-01 -1.41385853e+00
3.13460618e-01 -5.24407685e-01 -9.39106107e-01 8.45064282e-01
-1.65770411e+00 -1.24579275e+00 -4.16851372e-01 8.34946156e-01
6.93238497e-01 -3.83336768e-02 7.29354858e-01 3.75548720e-01
-6.46300972e-01 8.37716281e-01 2.91322023e-02 -1.63912833e-01
1.04631436e+00 -9.45320606e-01 3.75339866e-01 8.74232292e-01
3.67480963e-01 7.66507506e-01 5.19363284e-01 -5.86398602e-01
-1.57982528e+00 -9.67623532e-01 5.74032485e-01 -2.18879759e-01
1.38325378e-01 -4.00494933e-01 -1.07404208e+00 1.40365571e-01
4.27690774e-01 3.59110296e-01 4.98256356e-01 -3.01399946e-01
-4.38934565e-01 -4.69540447e-01 -1.41732395e+00 9.54382837e-01
1.10240793e+00 -5.96461952e-01 -3.72745752e-01 -1.48887159e-02
2.24812284e-01 -2.49187529e-01 -6.40981138e-01 5.44110119e-01
7.29378223e-01 -9.20873046e-01 1.02160966e+00 -3.32638621e-01
5.55159301e-02 -8.18890154e-01 -9.48224738e-02 -1.09219730e+00
-2.05525741e-01 -9.06773567e-01 1.85617119e-01 1.36988223e+00
2.51514882e-01 -4.80531365e-01 6.78364456e-01 5.72926819e-01
1.88132390e-01 -6.85592175e-01 -9.80672717e-01 -5.29627383e-01
-1.49148032e-01 1.38594404e-01 8.63611639e-01 9.27600980e-01
-7.18098357e-02 5.48455864e-02 -4.27684665e-01 7.45237842e-02
6.70471430e-01 3.96365851e-01 9.39017475e-01 -9.66929734e-01
4.84005064e-02 -5.37040949e-01 -3.20543557e-01 -7.32874036e-01
-2.31253877e-02 -4.56521988e-01 1.04276754e-01 -1.28614712e+00
-3.15823741e-02 -5.86720943e-01 -2.43986189e-01 3.73624951e-01
-3.63404095e-01 8.63733530e-01 1.30206555e-01 2.38912523e-01
-1.76489845e-01 5.80171108e-01 1.44682634e+00 -1.84894279e-01
-3.45605135e-01 -2.15988696e-01 -3.00081640e-01 6.25139892e-01
8.40041697e-01 -3.61900330e-01 -4.46502924e-01 -3.24005812e-01
1.08697638e-01 -1.89096838e-01 4.57591593e-01 -9.99411404e-01
1.53078601e-01 -4.95257109e-01 7.67994940e-01 -4.77171838e-01
4.04807538e-01 -1.02390063e+00 4.73015785e-01 1.11346900e-01
-1.78092778e-01 2.80384183e-01 1.82214007e-01 4.50678259e-01
-7.58863091e-02 -2.05347955e-01 9.95630741e-01 -7.67864063e-02
-5.68569839e-01 4.72835988e-01 -8.38988647e-03 -2.70069808e-01
1.09199750e+00 -6.48580015e-01 -4.16789055e-01 -9.84503627e-02
-3.26819181e-01 -4.09684658e-01 8.72774124e-01 4.59538698e-01
1.16739595e+00 -1.66334510e+00 -6.04475439e-01 7.53119469e-01
2.06968952e-02 -3.39677244e-01 2.88072616e-01 6.10699475e-01
-6.10359430e-01 8.60284921e-03 -4.59245235e-01 -5.16597211e-01
-1.46799016e+00 7.40783513e-01 3.55762243e-01 6.03482962e-01
-5.80325127e-01 7.72954464e-01 1.88125297e-01 -1.17440280e-02
1.69698149e-01 1.19496211e-02 1.60671487e-01 -2.23208338e-01
6.72978222e-01 3.55358571e-01 4.26859967e-03 -6.88962817e-01
-3.09541702e-01 1.16659617e+00 1.90269947e-02 -1.73007429e-01
1.00748408e+00 -8.39589089e-02 -1.27574235e-01 2.93241233e-01
1.20808184e+00 2.29776382e-01 -1.38102627e+00 -1.07757270e-01
-5.95754683e-01 -1.25893450e+00 -8.19983333e-02 -5.93905628e-01
-1.14088178e+00 8.07289064e-01 7.85596907e-01 -6.21900223e-02
1.35646510e+00 -2.88052887e-01 6.98363662e-01 5.68943024e-02
4.00724679e-01 -1.16814709e+00 1.24153249e-01 -2.15367705e-01
1.22084880e+00 -1.11015940e+00 2.77826041e-01 -4.44198877e-01
-5.09269357e-01 1.13521111e+00 5.90146959e-01 -1.99812323e-01
5.60572922e-01 1.45527482e-01 9.23254937e-02 -2.72872020e-02
-4.99642938e-01 -3.11720129e-02 6.43538594e-01 6.34077847e-01
2.34739557e-01 -2.89391547e-01 -5.28829336e-01 -4.06225510e-02
3.30108404e-01 -2.31447637e-01 1.33412555e-01 9.00603771e-01
-2.91921228e-01 -1.42075098e+00 -5.48660338e-01 6.85084164e-02
-2.34762579e-01 3.42854597e-02 -6.06330931e-01 5.40090322e-01
2.63373643e-01 7.58430302e-01 6.87289163e-02 -2.61977673e-01
5.19306660e-01 -5.04771993e-03 1.03542757e+00 -2.97875643e-01
-3.65140647e-01 1.35795146e-01 -5.15834332e-01 -5.35838604e-01
-6.22534633e-01 -6.37605190e-01 -1.08684766e+00 -2.81653076e-01
-2.75503933e-01 -1.16089456e-01 7.42641449e-01 4.95422989e-01
3.57657611e-01 6.74916729e-02 1.02290916e+00 -1.05152929e+00
4.25221287e-02 -5.66310525e-01 -3.72510910e-01 5.25202334e-01
2.54092485e-01 -5.94172955e-01 -2.07267806e-01 2.95491934e-01] | [12.67293930053711, -0.018292201682925224] |
cc272a36-83d5-4efb-814b-5e81e623ccd2 | xdoc-unified-pre-training-for-cross-format | 2210.02849 | null | https://arxiv.org/abs/2210.02849v1 | https://arxiv.org/pdf/2210.02849v1.pdf | XDoc: Unified Pre-training for Cross-Format Document Understanding | The surge of pre-training has witnessed the rapid development of document understanding recently. Pre-training and fine-tuning framework has been effectively used to tackle texts in various formats, including plain texts, document texts, and web texts. Despite achieving promising performance, existing pre-trained models usually target one specific document format at one time, making it difficult to combine knowledge from multiple document formats. To address this, we propose XDoc, a unified pre-trained model which deals with different document formats in a single model. For parameter efficiency, we share backbone parameters for different formats such as the word embedding layer and the Transformer layers. Meanwhile, we introduce adaptive layers with lightweight parameters to enhance the distinction across different formats. Experimental results have demonstrated that with only 36.7% parameters, XDoc achieves comparable or even better performance on a variety of downstream tasks compared with the individual pre-trained models, which is cost effective for real-world deployment. The code and pre-trained models will be publicly available at \url{https://aka.ms/xdoc}. | ['Furu Wei', 'Cha Zhang', 'Lei Cui', 'Tengchao Lv', 'Jingye Chen'] | 2022-10-06 | null | null | null | null | ['semantic-entity-labeling'] | ['natural-language-processing'] | [ 1.86365675e-02 -2.64007479e-01 -2.05634683e-01 -3.32769483e-01
-7.86221027e-01 -8.28319848e-01 8.13480616e-01 1.82515025e-01
-4.89014357e-01 3.75169009e-01 4.39227879e-01 -5.56133151e-01
-1.42386422e-01 -6.64020061e-01 -4.86308306e-01 -3.43959153e-01
1.92660019e-01 2.75069445e-01 2.34925792e-01 -2.59289861e-01
1.62780449e-01 6.86675832e-02 -9.53285456e-01 7.12097287e-01
9.32984948e-01 7.78804123e-01 5.81518590e-01 7.21471429e-01
-5.95996201e-01 5.58404863e-01 -5.98449290e-01 -8.14629376e-01
4.56750486e-03 9.37875658e-02 -8.87611330e-01 -4.46990728e-02
4.41986859e-01 -7.11683214e-01 -6.82397425e-01 9.27954316e-01
7.10943401e-01 -8.93310308e-02 3.31168175e-01 -9.90887165e-01
-1.08297944e+00 1.25534868e+00 -3.62601459e-01 1.91214442e-01
1.04355760e-01 -2.19053887e-02 1.26124394e+00 -1.08293128e+00
4.01362717e-01 1.14971304e+00 5.82955718e-01 4.93108690e-01
-7.23716438e-01 -7.17946053e-01 3.90120506e-01 2.42669791e-01
-1.15429592e+00 -4.10572290e-01 3.79865199e-01 -2.38884971e-01
1.30527914e+00 1.60781115e-01 1.70462236e-01 1.39490843e+00
9.19499174e-02 1.19795370e+00 8.12834501e-01 -3.96477163e-01
-4.11433369e-01 9.65870693e-02 4.54622000e-01 3.69937301e-01
3.43415052e-01 -4.32541013e-01 -4.81026739e-01 2.26202399e-01
4.10274774e-01 2.51279294e-01 -3.81590724e-01 2.03486800e-01
-1.06558001e+00 6.84027135e-01 1.06932476e-01 4.86816555e-01
4.91397418e-02 -4.99683693e-02 7.50222325e-01 3.75784010e-01
4.42960203e-01 3.78708869e-01 -7.16547608e-01 -3.60781342e-01
-8.39682698e-01 2.15757981e-01 6.46137774e-01 1.34898841e+00
3.79986048e-01 -1.83820605e-01 -3.73656958e-01 1.15174496e+00
2.81135082e-01 3.41571331e-01 7.38181710e-01 -3.80287439e-01
1.30157721e+00 8.00346434e-01 -2.99753975e-02 -7.29919076e-01
-3.28338981e-01 -7.04737663e-01 -9.01734233e-01 -7.05931365e-01
3.37927103e-01 -2.39195779e-01 -8.90601218e-01 1.35380173e+00
5.42287491e-02 1.82232708e-02 1.77861869e-01 6.24499559e-01
9.32964444e-01 1.03960490e+00 4.09741653e-03 2.33290821e-01
1.50805843e+00 -1.44117415e+00 -9.48165059e-01 -3.68478358e-01
9.78122771e-01 -1.18117690e+00 1.43954360e+00 2.79074609e-01
-1.11974454e+00 -6.19911849e-01 -1.19270945e+00 -5.65802217e-01
-6.74423218e-01 4.84472096e-01 5.07086396e-01 6.39039338e-01
-8.99102926e-01 3.89615119e-01 -7.86325037e-01 -3.36638570e-01
3.90360773e-01 1.66923746e-01 -2.36526191e-01 -4.41921204e-01
-1.31490433e+00 6.21100605e-01 5.26136339e-01 1.86345831e-01
-4.70622241e-01 -9.88476157e-01 -5.73700190e-01 4.12179232e-01
3.15385908e-01 -4.31049705e-01 1.36258125e+00 -5.08621812e-01
-1.53518343e+00 5.35058022e-01 -2.99238674e-02 -2.02676654e-01
4.96123016e-01 -8.70911896e-01 -8.29339683e-01 -7.99104273e-02
-9.13268551e-02 2.58302033e-01 7.62381196e-01 -7.99116910e-01
-6.21523261e-01 -2.25321427e-01 2.61181295e-01 2.14322463e-01
-1.04134107e+00 4.04668272e-01 -1.05992103e+00 -8.68676901e-01
-2.64054924e-01 -6.73458695e-01 2.19520226e-01 -2.43807748e-01
-5.82311392e-01 -4.02920842e-01 1.04992628e+00 -6.64655924e-01
1.67646527e+00 -2.28109145e+00 -8.91787186e-02 -1.08773969e-01
1.89435065e-01 5.96019924e-01 -4.23213989e-01 8.55016530e-01
2.08132759e-01 2.26545930e-01 7.26456419e-02 -6.40304148e-01
5.07596135e-01 -3.53111041e-04 -6.26293540e-01 9.46780145e-02
3.13452244e-01 8.72054160e-01 -6.60955608e-01 -4.40784276e-01
1.33307174e-01 6.60213292e-01 -5.15432715e-01 3.48446518e-01
-2.82710165e-01 -5.71167246e-02 -6.57928109e-01 4.64653373e-01
9.10406291e-01 -5.72564125e-01 3.56280446e-01 -2.60837674e-01
2.23405343e-02 6.80561185e-01 -1.25345218e+00 1.76640856e+00
-6.22524440e-01 5.54240644e-01 -6.78342432e-02 -7.94617474e-01
7.39394724e-01 4.02686238e-01 1.02637753e-01 -6.85384750e-01
3.05832118e-01 2.21938074e-01 -6.22137403e-03 -4.53830719e-01
9.19898450e-01 4.42533791e-01 -1.34526789e-02 6.67849779e-01
3.08424771e-01 1.02612182e-01 5.90269864e-01 3.47177982e-01
1.10620928e+00 -7.54404515e-02 -5.91166429e-02 1.73185747e-02
5.38011789e-01 -2.97410488e-01 1.16286464e-02 6.78755462e-01
2.28700206e-01 5.18164456e-01 2.04029649e-01 -1.79719359e-01
-9.51271296e-01 -9.22037125e-01 -2.79593468e-01 1.51448309e+00
9.38805938e-02 -1.01262271e+00 -6.47247672e-01 -6.10845149e-01
1.21963464e-01 5.12221575e-01 -3.47644687e-01 -4.42577824e-02
-6.77413583e-01 -7.94745505e-01 6.98704422e-01 7.32152104e-01
6.61532462e-01 -7.14934468e-01 -3.63066792e-02 4.38021243e-01
-6.18727282e-02 -1.36754560e+00 -7.64712274e-01 1.13869965e-01
-7.09075093e-01 -8.31012845e-01 -7.01106191e-01 -8.09355557e-01
4.77701932e-01 4.87064034e-01 9.53239799e-01 3.02249491e-01
3.89068760e-02 1.96623489e-01 -7.69269943e-01 -2.93660998e-01
-2.53150403e-01 7.52330482e-01 -1.92631245e-01 -8.87694582e-02
4.07996595e-01 -1.76392972e-01 -5.75430274e-01 1.38635144e-01
-1.24995661e+00 6.68785349e-02 5.95436752e-01 8.54864120e-01
1.37021139e-01 3.11677698e-02 3.97134423e-01 -1.07799304e+00
9.40589666e-01 -6.03264511e-01 -4.12845433e-01 5.02103627e-01
-6.94728732e-01 3.15570161e-02 9.53773558e-01 -4.80987191e-01
-1.17944884e+00 -5.54420948e-01 -3.75053972e-01 3.95365991e-02
1.22371756e-01 8.20681810e-01 -3.01947087e-01 4.93063688e-01
1.89133525e-01 3.16174209e-01 -4.20664877e-01 -9.22754705e-01
5.45355380e-01 1.14083719e+00 4.10233766e-01 -7.11576641e-01
9.48730946e-01 9.22559798e-02 -7.91830063e-01 -5.25601208e-01
-1.07499361e+00 -5.25931358e-01 -5.63333869e-01 8.85649174e-02
5.17502367e-01 -1.14309227e+00 -2.13570803e-01 6.05829477e-01
-1.25424206e+00 -3.60211432e-01 2.01792166e-01 4.05601919e-01
9.29249600e-02 4.00420815e-01 -8.02479327e-01 -1.87291175e-01
-6.87948823e-01 -1.15346825e+00 1.08828032e+00 2.06878543e-01
-1.27073377e-01 -1.24145317e+00 -2.07498714e-01 5.66681802e-01
6.19601250e-01 -5.84614754e-01 1.02056670e+00 -8.63344014e-01
-4.48303908e-01 -2.61314034e-01 -5.52614868e-01 4.18494493e-01
2.38631114e-01 1.78663149e-01 -8.83420587e-01 -7.04452813e-01
-3.43412638e-01 -4.58440721e-01 8.04089189e-01 -7.12726861e-02
1.44494903e+00 -2.41406322e-01 -3.56633514e-01 7.71210253e-01
1.29160929e+00 -9.46003571e-03 4.97955948e-01 7.80848563e-01
7.23904610e-01 2.57315516e-01 4.14403737e-01 5.48005521e-01
5.27610481e-01 5.33522189e-01 9.43444073e-02 -2.22146232e-02
-2.36609012e-01 -4.60387498e-01 4.34027076e-01 1.36088634e+00
2.38188580e-01 -8.27641249e-01 -9.58007812e-01 4.38792139e-01
-1.60556006e+00 -7.52010822e-01 -1.34844393e-01 1.58163798e+00
1.18554997e+00 3.92088443e-01 -1.82510242e-01 1.23652816e-01
4.84909296e-01 3.27183306e-01 -3.71496707e-01 -4.11571383e-01
-1.83741048e-01 3.51071179e-01 4.91141856e-01 2.53860235e-01
-1.09971714e+00 1.10915685e+00 6.12367630e+00 1.05891299e+00
-1.38404667e+00 1.58020720e-01 1.40931293e-01 -8.50103870e-02
-3.59459579e-01 -7.96177909e-02 -1.44020236e+00 7.68301547e-01
1.00484157e+00 -3.45782220e-01 9.22698304e-02 6.56491220e-01
2.77413223e-02 5.91530561e-01 -9.32680070e-01 7.10967600e-01
-1.36168972e-01 -1.51743281e+00 4.19675648e-01 -1.69582680e-01
5.67644954e-01 2.30038777e-01 3.11906040e-01 7.00918853e-01
4.38451350e-01 -9.30362403e-01 7.48101413e-01 -9.42606572e-03
9.11041141e-01 -5.69444835e-01 6.88498974e-01 2.47170478e-01
-1.10979772e+00 -9.40578505e-02 -5.50896347e-01 2.73895729e-02
2.36714557e-02 5.39497972e-01 -6.38498425e-01 7.82079577e-01
6.82977259e-01 9.54199612e-01 -7.74529696e-01 7.88328588e-01
-9.78596509e-02 8.39912951e-01 -2.33812705e-01 -1.05240583e-01
5.83747208e-01 -2.42245272e-02 9.21121985e-02 1.63223994e+00
5.92558861e-01 -2.65368551e-01 4.31115404e-02 4.38960761e-01
-4.41464454e-01 3.80367100e-01 -1.84622884e-01 -4.26066816e-01
7.02844739e-01 1.25553143e+00 -3.60067397e-01 -3.86494696e-01
-9.74200726e-01 1.01097012e+00 5.44968426e-01 2.54535139e-01
-1.02184069e+00 -7.45031655e-01 6.03087246e-01 -2.85358969e-02
5.21979034e-01 -4.57855642e-01 -5.10695763e-02 -1.44974542e+00
2.87036240e-01 -1.02356887e+00 4.38850880e-01 -4.94094640e-01
-1.23371506e+00 7.54353225e-01 -1.02247171e-01 -1.20453441e+00
1.38759285e-01 -9.56484735e-01 -5.48903465e-01 8.14075649e-01
-1.82734537e+00 -1.27248359e+00 -1.58548191e-01 4.90936249e-01
7.92159498e-01 -3.03644866e-01 7.47927904e-01 6.51854813e-01
-1.01867163e+00 1.07818568e+00 8.15807045e-01 4.13621247e-01
1.13115239e+00 -1.27466297e+00 5.90985239e-01 8.64218891e-01
1.35655284e-01 9.98253644e-01 3.92936081e-01 -3.73935699e-01
-1.71690392e+00 -1.32421339e+00 9.76014435e-01 -4.61502224e-01
9.95367229e-01 -5.63449323e-01 -1.16548693e+00 8.59550595e-01
6.03301883e-01 -7.43700489e-02 8.41637850e-01 2.47145221e-01
-6.40190125e-01 -1.90769792e-01 -5.02031863e-01 6.69147432e-01
9.69468653e-01 -6.13255978e-01 -5.64327478e-01 4.04242069e-01
9.09124136e-01 -6.34231031e-01 -1.11427045e+00 2.13304296e-01
4.61041272e-01 -5.56307256e-01 9.20705974e-01 -7.10440695e-01
6.30892456e-01 -5.74571267e-02 -7.41570443e-02 -1.24213469e+00
-3.14382821e-01 -7.05775917e-01 -3.73316437e-01 1.58346379e+00
5.90971947e-01 -5.38533986e-01 4.74796116e-01 3.14976752e-01
-3.90519232e-01 -7.86780894e-01 -5.18629014e-01 -8.85220468e-01
4.38487738e-01 -5.33275843e-01 1.05910873e+00 1.09556913e+00
7.57510364e-02 6.00865364e-01 -2.55080611e-01 1.53927162e-01
2.29193211e-01 2.20645174e-01 8.43239307e-01 -7.98535109e-01
-3.74030024e-01 -4.62555200e-01 1.13289528e-01 -1.64380515e+00
7.28000514e-03 -1.26411414e+00 -3.39489669e-01 -1.55341697e+00
3.11820060e-01 -4.63332295e-01 -4.26857919e-01 5.82410812e-01
-5.48440695e-01 -8.89359713e-02 2.85809487e-01 2.06242695e-01
-6.82958841e-01 5.35840929e-01 1.36776924e+00 -4.08306867e-01
1.09921090e-01 -2.38951445e-01 -9.71057415e-01 3.65888178e-01
8.48323703e-01 -3.37683082e-01 -5.20632982e-01 -1.33740091e+00
1.25350252e-01 -4.73474383e-01 -1.33436039e-01 -7.59311616e-01
2.39064902e-01 9.95507017e-02 3.49026203e-01 -6.77921236e-01
8.86007026e-02 -6.82987571e-01 -3.31099123e-01 1.44950047e-01
-4.33638960e-01 2.73201168e-01 4.32188481e-01 1.97254255e-01
-3.43410701e-01 -3.55968118e-01 4.54850018e-01 2.61869848e-01
-6.33924246e-01 3.73019427e-01 -2.17786774e-01 2.53655910e-01
6.38427198e-01 1.05524570e-01 -8.20544064e-01 -1.94853157e-01
-2.82079428e-01 4.45019424e-01 1.12358592e-01 8.48915756e-01
4.76343334e-01 -1.18020093e+00 -7.84468055e-01 1.86722681e-01
1.47184700e-01 3.64995040e-02 3.65262032e-01 5.96597493e-01
-4.58952874e-01 7.96740890e-01 4.93452437e-02 -3.13844740e-01
-1.19092071e+00 2.77593940e-01 5.94308935e-02 -7.35537052e-01
-6.87689543e-01 6.63272679e-01 -5.72566912e-02 -5.98768771e-01
3.46318662e-01 -6.02202833e-01 -2.19813392e-01 -3.76067460e-02
8.66963387e-01 1.81292191e-01 3.15792829e-01 -3.35997611e-01
-9.12875608e-02 4.13557291e-01 -8.36702347e-01 2.81018883e-01
1.44943070e+00 -2.35782385e-01 5.83919063e-02 1.89821407e-01
1.48347211e+00 7.75370523e-02 -1.08667529e+00 -6.74295545e-01
9.50726122e-02 -3.27943712e-01 2.40088582e-01 -8.90271246e-01
-1.00299203e+00 9.89162624e-01 2.33540311e-01 2.13631704e-01
1.03444827e+00 -2.29149282e-01 1.21760654e+00 5.85495889e-01
-4.38448191e-02 -1.21588778e+00 1.30861076e-02 9.81226623e-01
7.61046052e-01 -1.08539772e+00 1.10992953e-01 -3.39993536e-01
-4.10441548e-01 1.33041298e+00 6.92586482e-01 2.68523693e-01
6.89507484e-01 6.04697824e-01 2.22516194e-01 8.15379620e-02
-9.77188110e-01 -1.28272688e-02 3.85986745e-01 2.47565344e-01
9.25644338e-01 -2.19092280e-01 -2.32489347e-01 9.88842487e-01
-3.71946812e-01 -4.27836664e-02 3.30250770e-01 1.09807694e+00
-1.68141991e-01 -1.58769882e+00 -7.12266564e-02 5.69945753e-01
-7.67989159e-01 -3.79027307e-01 -8.13960433e-02 9.05457497e-01
-2.16231361e-01 8.10666144e-01 8.49462226e-02 -2.19673544e-01
4.40652758e-01 1.45844176e-01 4.53913718e-01 -7.43330777e-01
-6.73973560e-01 -6.09371029e-02 6.75592124e-02 -3.09764981e-01
-1.42207861e-01 -4.81080085e-01 -1.14852726e+00 -5.52499771e-01
-3.80630255e-01 7.46836886e-02 3.62730056e-01 8.55737507e-01
5.08580923e-01 6.13380790e-01 4.17951643e-01 -4.04119372e-01
-8.50018263e-01 -1.27869153e+00 -3.48255336e-01 3.41229171e-01
1.23871990e-01 -4.62985367e-01 -1.59524798e-01 1.64481804e-01] | [11.01025390625, 8.413555145263672] |
346f0094-40f3-4ee2-a20d-55e722cd8556 | visual-grounding-of-inter-lingual-word | 2209.03714 | null | https://arxiv.org/abs/2209.03714v2 | https://arxiv.org/pdf/2209.03714v2.pdf | Visual Grounding of Inter-lingual Word-Embeddings | Visual grounding of Language aims at enriching textual representations of language with multiple sources of visual knowledge such as images and videos. Although visual grounding is an area of intense research, inter-lingual aspects of visual grounding have not received much attention. The present study investigates the inter-lingual visual grounding of word embeddings. We propose an implicit alignment technique between the two spaces of vision and language in which inter-lingual textual information interacts in order to enrich pre-trained textual word embeddings. We focus on three languages in our experiments, namely, English, Arabic, and German. We obtained visually grounded vector representations for these languages and studied whether visual grounding on one or multiple languages improved the performance of embeddings on word similarity and categorization benchmarks. Our experiments suggest that inter-lingual knowledge improves the performance of grounded embeddings in similar languages such as German and English. However, inter-lingual grounding of German or English with Arabic led to a slight degradation in performance on word similarity benchmarks. On the other hand, we observed an opposite trend on categorization benchmarks where Arabic had the most improvement on English. In the discussion section, several reasons for those findings are laid out. We hope that our experiments provide a baseline for further research on inter-lingual visual grounding. | ['R. Harald Baayen', 'Hendrik P. A. Lensch', 'Hassan Shahmohammadi', 'Wafaa Mohammed'] | 2022-09-08 | null | null | null | null | ['word-similarity'] | ['natural-language-processing'] | [-3.93714756e-01 -8.74971896e-02 -3.60077798e-01 -2.53126711e-01
-4.35510129e-01 -7.68495858e-01 1.13081551e+00 6.26805246e-01
-7.38547504e-01 1.75430790e-01 5.94585419e-01 -3.81380349e-01
3.22074950e-01 -7.69101858e-01 -6.07797384e-01 -5.07805169e-01
7.76032582e-02 -9.74912494e-02 1.59792274e-01 -4.34354097e-01
2.10503787e-01 3.64697933e-01 -1.46236598e+00 6.51475966e-01
5.98792613e-01 6.03006840e-01 2.04625234e-01 3.15025717e-01
-6.72272801e-01 6.20858848e-01 -4.17809457e-01 -6.82856262e-01
5.18680923e-02 -4.83653069e-01 -6.29253864e-01 2.48186767e-01
9.21151161e-01 1.02659687e-01 -4.38300878e-01 1.12293398e+00
3.03813279e-01 1.10153118e-02 7.24061370e-01 -1.42213142e+00
-1.58447337e+00 5.80756068e-01 -6.53834999e-01 2.61712968e-01
2.84611106e-01 1.27631113e-01 1.15909648e+00 -1.38447952e+00
8.98057640e-01 1.65896714e+00 6.27048910e-01 2.36449108e-01
-1.17110014e+00 -2.89789975e-01 4.51253712e-01 2.98869371e-01
-1.56884468e+00 -1.79604754e-01 6.42482460e-01 -6.37177706e-01
8.92354667e-01 -1.86054677e-01 6.16254210e-01 1.12462544e+00
2.19100103e-01 5.59193552e-01 1.34635925e+00 -7.90670872e-01
-5.56820296e-02 6.88358665e-01 2.60007769e-01 8.99326324e-01
5.94160020e-01 1.56180756e-02 -7.32425749e-01 1.76678836e-01
5.43504655e-01 -2.11691931e-01 -1.02037668e-01 -6.42885327e-01
-1.45521092e+00 1.21276414e+00 7.85163701e-01 7.15188861e-01
-2.48435825e-01 1.61063209e-01 5.95292628e-01 1.87142938e-01
7.72221267e-01 5.17036915e-01 4.83549573e-02 2.12517574e-01
-7.14872301e-01 -1.33637771e-01 2.92516649e-01 6.61918402e-01
1.03418863e+00 9.01254192e-02 -1.99176565e-01 8.26725781e-01
4.04814392e-01 5.87952614e-01 5.53241253e-01 -5.19106686e-01
4.87178266e-01 7.25242376e-01 -3.72526385e-02 -1.36048281e+00
-2.60089695e-01 -1.80509582e-01 -2.90368736e-01 2.09940195e-01
5.64180911e-01 8.42320826e-03 -8.59392107e-01 1.70428860e+00
-1.06834099e-01 -1.63157359e-01 3.04251224e-01 1.04330432e+00
1.06929493e+00 6.89159274e-01 3.98604810e-01 1.81794301e-01
1.85874081e+00 -9.80605006e-01 -8.03517461e-01 -4.47880268e-01
8.57769370e-01 -1.07371402e+00 1.56182432e+00 -1.83441252e-01
-8.08396280e-01 -6.67736888e-01 -1.11260355e+00 -3.56401473e-01
-1.01507759e+00 1.63187653e-01 4.76708591e-01 6.21433318e-01
-1.15572608e+00 3.48037034e-02 -6.41183436e-01 -9.67678964e-01
9.28996205e-02 -3.76495242e-01 -6.62783921e-01 -3.26463789e-01
-1.10608292e+00 1.44920063e+00 3.91084045e-01 -2.72516936e-01
-3.71103525e-01 -5.06883681e-01 -1.27004290e+00 -3.42494816e-01
1.43601373e-01 -3.39983135e-01 6.52060390e-01 -1.12374663e+00
-7.18231440e-01 1.47356141e+00 -2.54008353e-01 -4.75626528e-01
2.77489759e-02 -1.65704817e-01 -5.83915412e-01 1.67296082e-01
2.23350346e-01 1.09607422e+00 5.95801115e-01 -1.49994385e+00
-5.13459325e-01 -3.51836711e-01 3.54793698e-01 2.20098570e-01
-7.83119917e-01 1.96692012e-02 -5.40018201e-01 -9.61571872e-01
-1.59083053e-01 -9.58482265e-01 2.80389160e-01 2.28414953e-01
1.00959040e-01 -3.28463823e-01 8.44168842e-01 -6.23724520e-01
1.13836598e+00 -2.32409883e+00 1.54669508e-01 -6.04261272e-02
1.53252825e-01 -1.09416761e-01 -5.05160689e-01 7.25548804e-01
-6.11652769e-02 4.24049407e-01 1.53443456e-01 -2.52878994e-01
1.38175547e-01 4.70197111e-01 -2.54806161e-01 5.81876159e-01
3.32831353e-01 1.13649845e+00 -1.02368629e+00 -7.46801555e-01
2.22838044e-01 8.67625117e-01 -4.39592689e-01 -6.78084716e-02
6.02197833e-02 -1.13063656e-01 1.11542679e-01 5.63052893e-01
3.63422692e-01 -2.16359168e-01 9.32770818e-02 -6.66819334e-01
-2.76683480e-01 1.17637575e-01 -8.83148313e-01 1.70198989e+00
-7.13874757e-01 1.22261214e+00 -3.51581722e-01 -8.58168304e-01
8.94754291e-01 2.39986405e-01 9.63824838e-02 -1.18256688e+00
-1.67793799e-02 5.68200760e-02 1.78435147e-01 -3.92537415e-01
8.55912149e-01 -3.21351141e-01 -1.63327187e-01 4.92280334e-01
6.62305802e-02 -2.69168198e-01 3.80804002e-01 3.98523390e-01
3.15408826e-01 1.41618192e-01 2.73508787e-01 -4.21690375e-01
2.85118133e-01 2.57517874e-01 -8.61540139e-02 2.86390692e-01
-2.54731506e-01 4.40962672e-01 2.55499363e-01 -4.50098395e-01
-1.08009911e+00 -1.10505593e+00 -2.37864152e-01 1.37551463e+00
3.70616436e-01 -9.24473882e-01 -5.00267446e-01 -6.44721150e-01
5.59465326e-02 8.58686566e-01 -9.90981519e-01 -1.60568237e-01
-2.59314984e-01 -4.96978045e-01 5.28679192e-01 6.27540886e-01
2.48692870e-01 -9.99976575e-01 -4.27792460e-01 -2.80048490e-01
-1.04542412e-01 -1.49168372e+00 -5.47842026e-01 7.28189051e-02
-5.78851581e-01 -8.36594641e-01 -9.26082671e-01 -1.11191928e+00
6.79266155e-01 6.39972687e-01 1.35740960e+00 9.53887850e-02
-1.47409812e-01 7.99013138e-01 -6.23030365e-01 -5.56162000e-01
-3.08159113e-01 -1.80765510e-01 8.18252712e-02 5.59864054e-03
6.70843065e-01 1.99839026e-01 -2.94024438e-01 2.20075160e-01
-1.01097107e+00 -9.16364416e-02 3.00154001e-01 6.12703919e-01
4.90726441e-01 -3.08311671e-01 9.80848894e-02 -4.77065951e-01
7.36588120e-01 -3.45822692e-01 -3.09201479e-01 2.43364513e-01
-4.28584784e-01 3.66757959e-01 8.77261385e-02 -5.22334874e-01
-6.35740161e-01 -3.66209298e-01 3.78395528e-01 -4.37613517e-01
1.30346585e-02 8.31510663e-01 3.15526843e-01 1.72748536e-04
8.38019013e-01 -1.23177713e-03 -4.30523343e-02 -2.74349958e-01
8.83417189e-01 4.95487869e-01 4.13261741e-01 -6.81360483e-01
7.96378136e-01 4.41580802e-01 -3.29472363e-01 -1.24738526e+00
-5.79553962e-01 -3.41915369e-01 -6.39321566e-01 -7.25115910e-02
1.36245787e+00 -1.06165898e+00 2.00899206e-02 2.27999706e-02
-1.13770187e+00 -1.74632415e-01 -3.08547080e-01 5.50015986e-01
-1.39847428e-01 4.61719751e-01 -3.36685300e-01 -4.16354328e-01
1.45229444e-01 -1.11345756e+00 1.06453180e+00 -1.46307185e-01
-5.59943736e-01 -1.44940567e+00 9.81274471e-02 1.31825536e-01
2.15549976e-01 1.86128393e-01 1.11275077e+00 -5.00112176e-01
-1.36282578e-01 4.82787639e-02 -4.80089277e-01 3.14884454e-01
3.57150644e-01 5.60260676e-02 -9.06326294e-01 -4.11950797e-01
-5.09857118e-01 -5.50682664e-01 8.44608068e-01 1.12176530e-01
6.80502355e-01 -6.49307594e-02 -1.01483673e-01 3.18870932e-01
1.50260842e+00 4.90530171e-02 4.74753916e-01 6.82519138e-01
8.40802729e-01 8.62321556e-01 7.06516445e-01 6.36009052e-02
5.18263459e-01 9.40897465e-01 3.35773259e-01 -4.74892497e-01
-4.82512057e-01 -3.10073972e-01 5.92313886e-01 9.38457429e-01
1.59247480e-02 -2.92174548e-01 -1.18760324e+00 9.16646481e-01
-1.40871060e+00 -8.36450279e-01 3.41071375e-02 2.12481999e+00
7.83129632e-01 -1.86610445e-01 2.72614688e-01 2.63797697e-02
6.80654883e-01 2.66441494e-01 1.01428792e-01 -7.31691897e-01
-3.57924908e-01 -1.04860654e-02 3.02775860e-01 6.38303578e-01
-9.84488130e-01 1.22770333e+00 5.84129620e+00 5.08286655e-01
-1.56224597e+00 1.56798899e-01 3.27855885e-01 2.03188002e-01
-6.19355798e-01 -8.69484842e-02 -6.54673457e-01 2.57651418e-01
7.97556520e-01 -2.28199571e-01 9.29466486e-02 6.19382381e-01
8.78627505e-03 1.52900368e-01 -1.16711283e+00 1.06907582e+00
4.46936011e-01 -1.35768628e+00 4.53514040e-01 1.65665209e-01
7.09201276e-01 5.21646403e-02 2.67408103e-01 2.96518922e-01
1.42814606e-01 -1.20685840e+00 1.02001643e+00 2.23751366e-02
7.79908419e-01 -7.27432609e-01 6.22026920e-01 -2.25967899e-01
-1.32553613e+00 3.83057654e-01 -4.38640803e-01 -2.77255289e-02
-3.60397659e-02 -6.99516339e-03 -6.43355846e-01 4.10428375e-01
6.84468865e-01 9.16651130e-01 -1.16876090e+00 5.93614519e-01
-1.90271720e-01 3.69856715e-01 1.48771316e-01 3.08012962e-03
7.44035721e-01 -9.93838981e-02 2.22975150e-01 1.59182799e+00
1.42404675e-01 -4.34125632e-01 3.69917721e-01 6.26791358e-01
1.22483566e-01 5.10532558e-01 -1.02771068e+00 -6.37000263e-01
2.71309257e-01 1.07081544e+00 -9.95079100e-01 -3.90241057e-01
-1.05111301e+00 8.41746211e-01 3.14692497e-01 4.47085679e-01
-8.57838035e-01 -2.43172586e-01 7.33824611e-01 3.18257451e-01
2.43958294e-01 -5.22499025e-01 -3.43478769e-01 -1.22240150e+00
-2.50974178e-01 -8.99536014e-01 3.83995831e-01 -1.18110251e+00
-1.45316744e+00 7.23561883e-01 1.32461369e-01 -1.11585629e+00
-2.49755960e-02 -9.98490989e-01 -1.49043828e-01 8.61976743e-01
-1.30116844e+00 -1.32879102e+00 -3.91395360e-01 7.16476858e-01
6.05128169e-01 -1.81552351e-01 8.70566845e-01 9.81352255e-02
-1.54951677e-01 6.57000422e-01 -2.62851685e-01 3.85316223e-01
9.73582447e-01 -1.17000985e+00 2.70800352e-01 9.18452501e-01
1.06515694e+00 7.92340577e-01 7.34259963e-01 -4.73814785e-01
-1.38065410e+00 -8.99184704e-01 1.10906959e+00 -4.77526516e-01
9.60466802e-01 -3.03498864e-01 -9.82686281e-01 7.10754037e-01
9.16487515e-01 7.63258785e-02 7.68381178e-01 2.37633854e-01
-9.09356534e-01 2.71597207e-01 -6.56043947e-01 9.75882888e-01
7.56949782e-01 -1.05017960e+00 -9.17977870e-01 1.96020290e-01
7.46648967e-01 -5.49289510e-02 -7.56800830e-01 -1.35776801e-02
3.99603933e-01 -6.14451945e-01 1.22514451e+00 -5.65797806e-01
5.98341763e-01 -3.10610294e-01 -6.07159615e-01 -1.52196634e+00
-3.35876971e-01 1.35186523e-01 3.38605464e-01 1.14842844e+00
4.12037581e-01 -5.52645147e-01 3.95868182e-01 2.71497853e-02
-8.91265720e-02 -5.06977320e-01 -5.25239468e-01 -9.99740422e-01
5.37353039e-01 -5.11030078e-01 1.71160311e-01 1.25328302e+00
-1.46896318e-02 5.75007617e-01 -3.47036570e-02 7.84300081e-03
2.48046964e-01 3.27520631e-02 6.36016130e-01 -8.35657239e-01
2.45375946e-01 -5.14804065e-01 -6.69775724e-01 -5.45462906e-01
4.42962170e-01 -1.37938094e+00 -1.79921865e-01 -1.90963125e+00
1.61354885e-01 -2.68294718e-02 -4.32646364e-01 6.45341039e-01
-2.38222569e-01 8.01690161e-01 5.94874382e-01 -5.91183677e-02
-4.13947970e-01 3.79453421e-01 1.22890031e+00 -4.56116766e-01
1.72969196e-02 -9.81498361e-01 -7.93327093e-01 6.80080354e-01
6.42644405e-01 -2.90084660e-01 -5.38587928e-01 -8.42661500e-01
2.18325183e-01 -5.92130423e-01 4.48208094e-01 -6.55148566e-01
-1.37068227e-01 -1.82524949e-01 3.26099038e-01 -3.24979573e-01
3.40839684e-01 -8.48419547e-01 -4.07189399e-01 4.11942780e-01
-3.47328931e-01 6.74359083e-01 6.50137126e-01 2.93584168e-01
-5.21335661e-01 -4.21833098e-02 7.92516530e-01 2.09617931e-02
-1.24322999e+00 -1.85297862e-01 -5.32281220e-01 4.66614842e-01
8.86767507e-01 -5.42015135e-01 -3.61368895e-01 -3.48044723e-01
-5.51709235e-01 -9.48515236e-02 8.27586949e-01 1.02283418e+00
5.85721672e-01 -1.69452238e+00 -8.43364477e-01 -2.50533447e-02
7.03279257e-01 -9.79368687e-01 -2.84392595e-01 6.99587643e-01
-4.86698627e-01 6.50760770e-01 -4.22245145e-01 -8.24403584e-01
-1.49761498e+00 8.45411181e-01 1.31441966e-01 1.90655828e-01
-3.86893064e-01 7.20712423e-01 5.72409987e-01 -1.49876684e-01
2.39610016e-01 -5.06386638e-01 -3.34018379e-01 6.91660166e-01
3.33826512e-01 1.26459375e-01 -1.48053737e-02 -1.43388367e+00
-6.82833076e-01 1.01859999e+00 1.09368488e-01 -2.91183144e-01
1.01401103e+00 -2.09182724e-01 -1.69621296e-02 8.69569480e-01
1.37914062e+00 2.71303684e-01 -6.56428516e-01 -2.36908972e-01
1.36803091e-01 -4.43565339e-01 1.56054571e-01 -6.11391008e-01
-1.05619621e+00 1.18675220e+00 8.07122469e-01 9.80186090e-02
7.48098135e-01 2.70268917e-01 3.32201272e-01 1.75806329e-01
2.64039129e-01 -8.23979914e-01 5.02229154e-01 5.77408671e-01
1.17930222e+00 -1.39214599e+00 9.97073054e-02 -1.76071689e-01
-9.81966257e-01 9.61572409e-01 7.40458012e-01 -4.48035374e-02
5.05040824e-01 -2.85397768e-02 4.59232926e-01 -4.02650207e-01
-4.04684603e-01 -5.95170140e-01 8.92010212e-01 6.98142827e-01
1.02102196e+00 3.66896354e-02 -3.58233660e-01 2.29299113e-01
-1.12412743e-01 -4.32607532e-01 2.93612033e-01 8.80229592e-01
-2.38695070e-01 -1.07955921e+00 -6.16771340e-01 -1.24135375e-01
-3.59216630e-01 -4.61870581e-01 -7.30446458e-01 1.32171977e+00
3.05324584e-01 8.78092051e-01 4.93993759e-01 -2.67347008e-01
2.10301742e-01 2.85507709e-01 6.15448654e-01 -8.32719505e-01
-4.50541914e-01 -1.21586643e-01 2.02112883e-01 -3.77553344e-01
-4.96216655e-01 -4.18744177e-01 -1.23102582e+00 -2.04355657e-01
4.00365219e-02 7.67919496e-02 6.23796403e-01 8.26445162e-01
1.91925839e-01 2.79650956e-01 -2.87100505e-02 -6.75861835e-01
1.39096856e-01 -8.17835152e-01 -3.68009478e-01 7.88067579e-01
9.93514508e-02 -9.71798539e-01 -2.80844331e-01 3.61351073e-01] | [10.79731273651123, 1.8134679794311523] |
5990a4d8-001a-47ca-878a-36b1df345a95 | authorship-attribution-by-consensus-among | null | null | https://aclanthology.org/C18-1234 | https://aclanthology.org/C18-1234.pdf | Authorship Attribution By Consensus Among Multiple Features | Most existing research on authorship attribution uses various lexical, syntactic and semantic features. In this paper we demonstrate an effective template-based approach for combining various syntactic features of a document for authorship analysis. The parse-tree based features that we propose are independent of the topic of a document and reflect the innate writing styles of authors. We show that the use of templates including sub-trees of parse trees in conjunction with other syntactic features result in improved author attribution rates. Another contribution is the demonstration that Dempster{'}s rule based combination of evidence from syntactic features performs better than other evidence-combination methods. We also demonstrate that our methodology works well for the case where actual author is not included in the candidate author set. | ['Jagadeesh Patchala', 'Raj Bhatnagar'] | 2018-08-01 | authorship-attribution-by-consensus-among-1 | https://aclanthology.org/C18-1234 | https://aclanthology.org/C18-1234.pdf | coling-2018-8 | ['author-attribution'] | ['natural-language-processing'] | [-6.22148626e-03 -1.91653937e-01 -3.43209594e-01 -5.45690775e-01
-2.82985419e-01 -8.28118265e-01 9.60361719e-01 4.71371949e-01
-3.72096479e-01 6.28963530e-01 4.02330905e-01 -4.51829582e-01
-4.67523545e-01 -5.16115665e-01 7.13800043e-02 1.79335643e-02
3.50421369e-01 5.61841846e-01 1.03370376e-01 -1.25503168e-01
1.15111840e+00 8.34048688e-01 -1.52035248e+00 -1.50636602e-02
6.53787792e-01 5.34615695e-01 -2.75507066e-02 7.84641743e-01
-7.14210749e-01 8.29057753e-01 -1.06171548e+00 -7.83343971e-01
4.32209764e-03 -5.06844342e-01 -8.16595078e-01 2.27947518e-01
6.98202968e-01 -6.49181977e-02 -2.84847729e-02 7.13820100e-01
1.61064908e-01 -2.23132089e-01 9.78511631e-01 -1.22586584e+00
-6.39583647e-01 6.32851720e-01 -1.29393518e-01 5.16135395e-01
6.40302300e-01 -5.11402190e-01 1.28212273e+00 -7.33576834e-01
8.70353937e-01 1.52849376e+00 8.19473028e-01 2.03455985e-01
-8.56322587e-01 -7.28474379e-01 8.37586150e-02 1.07067972e-01
-1.03197873e+00 -3.42783898e-01 1.09586263e+00 -5.62241018e-01
7.71010995e-01 2.54534364e-01 3.74453038e-01 9.25971687e-01
4.16545600e-01 4.43188071e-01 1.65486276e+00 -9.15355444e-01
-4.78400923e-02 3.18542272e-01 9.73439455e-01 9.92062926e-01
6.33869827e-01 -1.43600866e-01 -7.61350572e-01 -7.39180505e-01
5.46453595e-01 -4.08997327e-01 3.02507639e-01 2.67717451e-01
-1.15296662e+00 8.57988358e-01 -3.75347495e-01 7.02708066e-01
-5.26390038e-02 2.87620306e-01 3.22987169e-01 1.89510420e-01
7.85181671e-02 9.13846970e-01 -4.30472642e-01 -3.41793835e-01
-1.22290194e+00 4.86753285e-01 1.16394842e+00 9.18153346e-01
4.01080996e-01 -1.97221920e-01 -1.52236253e-01 7.17312098e-01
3.80745888e-01 4.09654886e-01 4.57577020e-01 -1.34527206e+00
2.36552179e-01 7.21788585e-01 2.50241637e-01 -1.26237106e+00
-2.39178330e-01 -4.78618890e-01 -1.91806667e-02 -7.33225718e-02
6.14695847e-01 1.39220938e-01 -5.47614157e-01 1.50497496e+00
-2.19324976e-01 -4.72668797e-01 -6.78561479e-02 4.31905299e-01
8.38552237e-01 1.76814839e-01 1.91463739e-01 -3.54697198e-01
1.66967273e+00 -4.38703507e-01 -1.05081034e+00 1.22204483e-01
4.91732657e-01 -1.13327861e+00 6.65239751e-01 4.38035041e-01
-8.51447344e-01 -3.93576205e-01 -8.15585256e-01 -1.60236850e-01
-5.39443672e-01 3.49928111e-01 7.60730267e-01 9.57317173e-01
-7.89541781e-01 7.29401708e-01 -4.72221076e-02 -4.47647244e-01
3.50162148e-01 5.06742060e-01 -3.38997424e-01 4.28861350e-01
-7.34865367e-01 9.31877553e-01 1.12749059e-02 -4.58037645e-01
1.27292052e-01 -2.00046837e-01 -4.47243482e-01 9.48661119e-02
1.55782208e-01 -4.77356285e-01 1.13782203e+00 -1.14006519e+00
-1.24084055e+00 9.97049928e-01 -6.62706554e-01 9.73720476e-02
3.36365938e-01 1.27695519e-02 -3.39644909e-01 3.25316042e-01
4.02249992e-01 2.14839876e-01 8.71610641e-01 -1.22789025e+00
-5.88944733e-01 -4.43184733e-01 -1.97387382e-01 -1.35105625e-01
-6.61984921e-01 9.37341154e-01 -1.56787530e-01 -1.06218863e+00
-5.06864898e-02 -7.28678226e-01 2.68348426e-01 -2.29434937e-01
-4.59631920e-01 -9.23062265e-01 6.83726490e-01 -7.12981224e-01
1.44991589e+00 -1.73547506e+00 9.96395126e-02 5.91526806e-01
3.94522816e-01 -1.33339718e-01 1.04198821e-01 4.65933710e-01
2.98420101e-01 5.97436845e-01 9.96800587e-02 -4.12993789e-01
2.24393144e-01 2.80595988e-01 -2.33701572e-01 2.19711989e-01
-1.70127064e-01 6.18097603e-01 -6.60178185e-01 -1.01000595e+00
-1.87644020e-01 1.35898873e-01 7.75332004e-02 3.21969762e-02
2.48065263e-01 -2.79159218e-01 -5.27809262e-01 7.79606998e-01
1.61347792e-01 -2.94271976e-01 4.86451149e-01 3.55000019e-01
-2.37735018e-01 4.75918770e-01 -1.02921677e+00 9.37712073e-01
-2.88967341e-02 8.89379442e-01 -2.86541581e-01 -3.41850758e-01
1.24353242e+00 1.78080961e-01 -1.51180308e-02 -3.13560218e-01
4.16554540e-01 5.53236246e-01 1.66594893e-01 -2.45151728e-01
5.98081946e-01 -2.11517781e-01 -3.37815017e-01 8.04134309e-01
9.20500904e-02 2.07314432e-01 4.42748278e-01 4.14692968e-01
9.98281896e-01 -1.30924804e-03 5.39366424e-01 -5.95646620e-01
5.80131531e-01 -1.46802425e-01 6.46477878e-01 1.13318849e+00
-1.35679990e-01 1.76584482e-01 8.33150983e-01 -2.82568097e-01
-1.11714160e+00 -6.84843957e-01 -3.43178332e-01 1.10846472e+00
-1.72115520e-01 -8.10198069e-01 -6.56938910e-01 -1.03041649e+00
3.49647582e-01 7.29533613e-01 -6.98279262e-01 4.81532305e-01
-6.46815181e-01 -6.12714469e-01 8.66346300e-01 6.26037359e-01
-7.95370061e-03 -7.44899631e-01 -7.20915198e-01 2.67224759e-01
1.17248587e-01 -9.07825291e-01 -3.05737704e-01 1.70346141e-01
-9.94497538e-01 -1.24419737e+00 -3.50257277e-01 -5.77881396e-01
6.19655788e-01 -9.16338339e-02 7.83306897e-01 4.80478883e-01
-1.21983804e-01 5.59966743e-01 -3.64995599e-01 -5.55336416e-01
-6.47774398e-01 3.64758670e-02 -6.25424013e-02 -3.71516436e-01
6.34897232e-01 -1.97400674e-01 1.73886418e-01 1.25452355e-01
-3.20426941e-01 -5.06680965e-01 2.78093427e-01 7.70993054e-01
3.05166170e-02 -9.82237756e-02 5.11748314e-01 -1.02463293e+00
1.08117259e+00 -1.59178689e-01 -1.33701801e-01 3.26891214e-01
-1.07324898e+00 3.09127718e-01 5.44873357e-01 -2.92776644e-01
-8.98986399e-01 -3.83324802e-01 2.70345092e-01 -6.91442788e-02
-2.31401622e-01 2.81999260e-01 -5.15091531e-02 -3.08290929e-01
3.79618198e-01 4.96269204e-02 1.44097701e-01 -7.87244499e-01
-3.57179344e-01 8.31919014e-01 3.48727137e-01 -8.65813494e-01
7.49676228e-01 2.21801311e-01 4.17572558e-01 -6.45706534e-01
-7.98218787e-01 -3.03866655e-01 -7.96691895e-01 -3.04046035e-01
3.64667803e-01 -3.49029511e-01 -7.93110251e-01 2.53532827e-01
-1.34480691e+00 3.13797057e-01 1.21526219e-01 3.19309503e-01
-3.40612322e-01 8.13934863e-01 -5.62232971e-01 -8.99484396e-01
-1.38520107e-01 -8.95716190e-01 9.97168541e-01 2.74017274e-01
-9.40915167e-01 -1.27583814e+00 -1.30295724e-01 5.81549406e-01
1.57720968e-01 -1.96133517e-02 1.15444720e+00 -1.22364640e+00
-9.81711298e-02 -4.15415913e-01 -2.13672116e-01 1.31281577e-02
1.99232131e-01 4.47003186e-01 -6.54760361e-01 1.92002073e-01
-2.14591280e-01 2.13080365e-02 7.20891237e-01 -6.93239495e-02
1.07454562e+00 -4.55530256e-01 -3.91084731e-01 1.40180215e-01
1.21641684e+00 3.42048593e-02 1.89414307e-01 7.31769085e-01
7.36368001e-01 8.80623937e-01 4.00352865e-01 4.75215554e-01
5.90524435e-01 7.35909224e-01 -2.76714414e-01 5.40302396e-01
-1.59543514e-01 -7.43891597e-02 2.02076197e-01 5.96435010e-01
-4.25015748e-01 -4.69795018e-01 -1.09006786e+00 2.45018274e-01
-1.66066444e+00 -1.04027116e+00 -7.26633370e-01 1.79221880e+00
7.54289508e-01 2.71360397e-01 3.38412136e-01 4.57920402e-01
7.23549545e-01 -9.87030491e-02 2.11253136e-01 -9.30823207e-01
-4.71678019e-01 2.50928998e-01 5.22232115e-01 5.87830663e-01
-9.45602000e-01 7.68275917e-01 7.98056746e+00 5.49983621e-01
-4.81445432e-01 -4.28447180e-04 1.59912750e-01 3.48517597e-01
-3.95817757e-01 2.80718267e-01 -1.25545108e+00 6.98485017e-01
1.01782441e+00 -3.02712947e-01 -1.40712574e-01 4.46677893e-01
7.08393753e-02 -1.73336551e-01 -1.07747436e+00 6.09567523e-01
6.39781475e-01 -1.30190027e+00 2.09899142e-01 5.71191549e-01
1.93186402e-01 -8.35877001e-01 -1.78360105e-01 -2.86677063e-01
2.65019238e-01 -9.36976790e-01 1.05133128e+00 6.11697078e-01
3.12676728e-01 -5.46526551e-01 8.56311619e-01 4.48107347e-02
-7.29564428e-01 -2.50228018e-01 -4.49769646e-02 -3.47889841e-01
-1.16480604e-01 5.07906377e-01 -7.62071073e-01 2.74659365e-01
2.87890226e-01 6.79812014e-01 -1.03217959e+00 6.83924139e-01
-3.03461373e-01 6.84685886e-01 -1.69079587e-01 -3.86601150e-01
1.32274628e-01 1.46230310e-01 8.63452673e-01 1.61543894e+00
2.55035549e-01 -5.40385060e-02 -4.27465253e-02 6.42087102e-01
-2.03185081e-01 4.29494768e-01 -4.54914629e-01 -1.82815731e-01
7.07622766e-01 1.07691693e+00 -1.00711143e+00 -5.82377732e-01
-3.33250284e-01 7.83495784e-01 2.78584749e-01 -7.38338605e-02
-1.89336360e-01 -5.05401909e-01 2.06235752e-01 1.64718598e-01
3.16266149e-01 -3.71955544e-01 -8.33023429e-01 -1.09903586e+00
3.37639190e-02 -7.85312474e-01 5.78582883e-01 -6.18432581e-01
-1.60113883e+00 1.52671874e-01 1.98890120e-02 -6.10397458e-01
-9.32835191e-02 -1.02570212e+00 -7.80386388e-01 9.25481737e-01
-1.17629194e+00 -1.12243235e+00 -4.08303328e-02 1.42659560e-01
4.73372042e-01 -6.89821005e-01 8.19019794e-01 -2.95972645e-01
-5.44575930e-01 6.80935621e-01 -2.57537551e-02 3.49692047e-01
8.01859558e-01 -1.63633370e+00 -1.20463595e-01 6.19396865e-01
1.00360192e-01 1.10704958e+00 8.58025372e-01 -1.02508736e+00
-1.21624243e+00 -4.61385965e-01 1.96614182e+00 -9.72696364e-01
9.13734972e-01 -9.48030651e-02 -8.11014771e-01 5.35988092e-01
4.14031357e-01 -5.12645781e-01 1.26030636e+00 4.43264365e-01
-7.21320868e-01 1.68616518e-01 -1.04068482e+00 2.02352419e-01
9.08467233e-01 -4.92426872e-01 -1.38667822e+00 2.23967120e-01
-1.90484300e-01 2.36544043e-01 -1.16469193e+00 -7.10951835e-02
8.48636925e-01 -8.01598549e-01 6.43965006e-01 -8.11641514e-01
6.12635612e-01 -6.97286651e-02 -1.22800745e-01 -6.03981733e-01
-5.90749681e-01 -6.50735974e-01 -1.41182318e-01 1.49808741e+00
2.45448515e-01 -5.39737046e-01 4.70436513e-01 5.64086676e-01
1.38648182e-01 -3.78452957e-01 -9.32118893e-01 -9.91846979e-01
2.37773791e-01 -6.53063739e-03 3.69847298e-01 1.14329278e+00
3.36330950e-01 4.20824200e-01 -1.16875479e-02 -4.70911980e-01
1.02764237e+00 4.03663784e-01 5.18043935e-01 -2.00655746e+00
-9.47534293e-02 -7.89078236e-01 -5.73862731e-01 -2.95596033e-01
7.85895765e-01 -8.77766430e-01 -4.54484820e-01 -1.44986570e+00
6.17554247e-01 -6.34106636e-01 -1.26410857e-01 5.58978081e-01
-3.27164531e-01 3.92321087e-02 3.20637077e-01 7.78406441e-01
-3.72743487e-01 -5.80231920e-02 7.71677077e-01 2.26503685e-01
-1.21779309e-03 -3.48252535e-01 -9.10928607e-01 8.11331749e-01
9.17237997e-01 -7.15131938e-01 1.45368978e-01 -3.26456525e-03
2.80923873e-01 -2.43176892e-01 4.79272664e-01 -4.36149627e-01
1.28545165e-01 -3.17537844e-01 6.77227736e-01 -2.48760119e-01
6.68463409e-02 -4.26953703e-01 -1.56323135e-01 3.36272031e-01
-4.91086245e-01 4.98734921e-01 -6.61686584e-02 5.50207734e-01
2.80621145e-02 -9.11204219e-01 4.78136569e-01 -3.39383274e-01
-4.30068284e-01 -5.03748775e-01 -8.07154834e-01 -2.11238652e-01
7.78470337e-01 -5.78845739e-01 -4.84917849e-01 -3.59982923e-02
-2.90102959e-01 -2.51671046e-01 8.74960065e-01 4.06526893e-01
3.23324382e-01 -9.21407640e-01 -5.54669499e-01 -2.98157930e-01
-1.30315020e-03 -1.13556051e+00 -6.77718043e-01 6.91843748e-01
-4.84730542e-01 6.23186409e-01 -3.84240150e-01 6.33927137e-02
-1.79801583e+00 2.67065704e-01 -6.62286058e-02 -1.45545542e-01
-4.75751430e-01 5.40769696e-01 -5.02428055e-01 -6.65358603e-02
1.61632046e-01 1.37916788e-01 -4.23302442e-01 3.44689667e-01
3.07472616e-01 6.38145328e-01 -2.31054038e-01 -9.67972636e-01
-7.15971768e-01 4.41338420e-01 -1.96562693e-01 -4.87270832e-01
1.12061000e+00 -5.16673438e-02 -4.64542389e-01 6.87502742e-01
7.64668167e-01 7.59706140e-01 -2.89815694e-01 -1.50617316e-01
5.16823232e-01 -7.49685407e-01 1.19174540e-01 -1.00166464e+00
-3.72660965e-01 6.03175819e-01 1.17815673e-01 2.65749484e-01
6.00105226e-01 2.22728789e-01 2.36535817e-01 3.87731403e-01
1.91123709e-01 -1.43293202e+00 -3.75899673e-02 3.67349118e-01
6.11227393e-01 -9.72341776e-01 6.02710664e-01 -5.36871195e-01
-4.71790165e-01 1.77555990e+00 3.26498955e-01 3.56657244e-02
4.21190232e-01 2.22224712e-01 -3.68289351e-02 -3.19635630e-01
-6.38540685e-01 4.51979740e-03 3.78271341e-01 2.84909904e-01
8.00588310e-01 1.63548499e-01 -1.06240165e+00 9.05439198e-01
-4.78709757e-01 -1.30335927e-01 1.04779041e+00 1.26758254e+00
-7.07678914e-01 -1.76664174e+00 -7.27379918e-01 8.30278814e-01
-7.92842209e-01 1.05455937e-02 -1.38986802e+00 8.13165843e-01
2.83421844e-01 8.46217632e-01 6.48186058e-02 -1.25030890e-01
-1.03351772e-01 5.20806015e-01 8.67303610e-01 -7.00374782e-01
-1.10065162e+00 1.31451726e-01 5.62534809e-01 1.04622822e-02
-6.59486234e-01 -1.35038185e+00 -1.05416369e+00 -5.75640202e-01
-2.84001559e-01 3.74912262e-01 7.41439939e-01 1.27130425e+00
2.59261400e-01 2.44138479e-01 4.51891661e-01 -2.53342658e-01
-4.58936185e-01 -7.97301292e-01 -5.33493698e-01 2.39091426e-01
-2.26539429e-02 -9.76658642e-01 -4.54062432e-01 2.80925065e-01] | [9.579859733581543, 10.571958541870117] |
fc5d4c31-b4d8-40cd-8cb1-eb41c8b304cb | efficient-contrastive-learning-via-novel-data | 2109.05941 | null | https://arxiv.org/abs/2109.05941v2 | https://arxiv.org/pdf/2109.05941v2.pdf | Efficient Contrastive Learning via Novel Data Augmentation and Curriculum Learning | We introduce EfficientCL, a memory-efficient continual pretraining method that applies contrastive learning with novel data augmentation and curriculum learning. For data augmentation, we stack two types of operation sequentially: cutoff and PCA jittering. While pretraining steps proceed, we apply curriculum learning by incrementing the augmentation degree for each difficulty step. After data augmentation is finished, contrastive learning is applied on projected embeddings of original and augmented examples. When finetuned on GLUE benchmark, our model outperforms baseline models, especially for sentence-level tasks. Additionally, this improvement is capable with only 70% of computational memory compared to the baseline model. | ['Alice Oh', 'Jiseon Kim', 'Seonghyeon Ye'] | 2021-09-10 | null | https://aclanthology.org/2021.emnlp-main.138 | https://aclanthology.org/2021.emnlp-main.138.pdf | emnlp-2021-11 | ['continual-pretraining'] | ['methodology'] | [ 4.17018414e-01 -6.39426038e-02 -3.21430087e-01 -3.59883517e-01
-9.63762164e-01 -6.62721395e-01 5.85367560e-01 5.14599085e-01
-1.09015179e+00 6.50178373e-01 3.70917380e-01 -7.71405160e-01
3.05640548e-01 -6.11830235e-01 -8.06855798e-01 -4.17361617e-01
-4.22496982e-02 6.72380567e-01 1.46417201e-01 -1.74704492e-01
3.52594495e-01 3.35257351e-01 -1.27071774e+00 5.93566895e-01
9.55185354e-01 6.55622125e-01 2.56696463e-01 9.19341564e-01
-5.41504443e-01 6.98482513e-01 -7.16235042e-01 -4.46788937e-01
5.38029000e-02 -9.47299972e-03 -8.49771738e-01 6.83142291e-03
6.59800947e-01 -4.61296499e-01 -2.72962451e-01 7.35465884e-01
2.48198628e-01 5.37140906e-01 4.78461504e-01 -9.75030482e-01
-7.39657104e-01 7.74484694e-01 -7.07832813e-01 3.95561993e-01
1.18062243e-01 7.10773915e-02 9.97702003e-01 -1.47333276e+00
2.62481451e-01 1.02394807e+00 4.91371214e-01 7.33853459e-01
-1.46433735e+00 -6.44814193e-01 5.79602957e-01 2.27954909e-01
-8.13465238e-01 -3.24974597e-01 7.98782349e-01 -2.32932806e-01
1.27940726e+00 1.78541064e-01 5.91169834e-01 1.27312446e+00
-1.45923659e-01 1.09496033e+00 9.97502983e-01 -6.81178331e-01
2.90266901e-01 1.06705740e-01 6.75893009e-01 5.21968782e-01
-2.44802479e-02 1.03203520e-01 -5.85756183e-01 -2.20294058e-01
4.59199429e-01 -2.44308617e-02 1.29557744e-01 -1.04218051e-01
-9.63447034e-01 8.94470751e-01 2.34127507e-01 -2.23547220e-02
-1.25256568e-01 1.16553359e-01 6.44793868e-01 3.99808824e-01
6.77571535e-01 8.23109567e-01 -9.36294794e-01 -5.54014564e-01
-8.63736868e-01 2.89768994e-01 5.10529757e-01 1.02446604e+00
6.64704740e-01 1.43973619e-01 -3.59280318e-01 9.05967653e-01
-1.31693676e-01 3.12966585e-01 7.61283457e-01 -7.57899761e-01
7.80336261e-01 7.32349396e-01 -2.46141434e-01 -4.05766964e-01
-3.41508389e-01 -5.81054330e-01 -8.63074064e-01 9.23234411e-03
2.04819649e-01 -3.63677412e-01 -1.24919724e+00 1.55038357e+00
1.41302645e-02 4.65783358e-01 1.87911868e-01 3.99395108e-01
1.02247596e+00 7.64980614e-01 5.22009730e-01 -2.40825161e-01
1.11151457e+00 -1.58283353e+00 -8.67421746e-01 -6.51280582e-01
9.86632109e-01 -6.92017436e-01 1.82335865e+00 4.64881837e-01
-1.25589275e+00 -7.99492776e-01 -1.20552504e+00 -2.50397146e-01
-3.57170373e-01 4.30439413e-02 6.84443593e-01 5.53577781e-01
-9.68259752e-01 5.01490295e-01 -1.03217697e+00 2.00193003e-01
3.65956664e-01 3.41697276e-01 -2.06675738e-01 -1.03137463e-01
-9.73952055e-01 8.15207899e-01 5.75121641e-01 -3.63545954e-01
-7.64347076e-01 -1.10003448e+00 -1.03083134e+00 2.05774724e-01
3.39847326e-01 -4.44152623e-01 1.52555656e+00 -5.96809030e-01
-1.48839283e+00 5.00358462e-01 -2.84804851e-01 -6.17066145e-01
1.51440531e-01 -7.61539459e-01 -2.57885456e-01 -1.78128645e-01
-3.09556723e-01 9.17914152e-01 8.07933390e-01 -1.16307604e+00
-4.22243446e-01 -5.29206172e-02 -4.12662933e-03 4.46088433e-01
-9.62480605e-01 -1.03946120e-01 -6.50053740e-01 -8.34981441e-01
-4.38464135e-02 -7.30022669e-01 -4.83279139e-01 -4.41080391e-01
-7.17203170e-02 -3.34335864e-01 8.59490514e-01 -6.31193161e-01
1.61085260e+00 -2.28509808e+00 1.74571455e-01 -6.44697770e-02
1.73490927e-01 4.96766388e-01 -7.88511992e-01 2.09177166e-01
-2.39324600e-01 2.67221659e-01 -3.38351667e-01 -8.73467088e-01
-1.90250203e-01 3.13172102e-01 -5.27059972e-01 -9.09438133e-02
5.13140321e-01 8.80253434e-01 -9.10328329e-01 -3.46394777e-01
1.11447409e-01 1.82018816e-01 -1.14889014e+00 5.51291883e-01
-4.37634081e-01 3.11276257e-01 1.79161616e-02 2.94071198e-01
6.55690253e-01 -2.22340122e-01 2.00870603e-01 1.18865132e-01
5.04900403e-02 6.29627585e-01 -9.72694278e-01 1.97464561e+00
-6.52688444e-01 5.98751664e-01 -4.64579701e-01 -8.45393658e-01
8.63009930e-01 2.30262838e-02 -9.12046991e-03 -6.98029041e-01
-1.14212841e-01 -1.41305894e-01 -1.07877322e-01 -3.56999218e-01
9.43253219e-01 4.07669783e-01 9.03514959e-03 4.57187027e-01
1.44008204e-01 -2.10456401e-01 3.04413199e-01 3.84105265e-01
1.25269699e+00 1.25135660e-01 -6.62433282e-02 8.85928571e-02
3.98473382e-01 8.27544183e-02 3.90704751e-01 7.14862943e-01
1.28504142e-01 3.82331491e-01 5.19099832e-01 -5.10299563e-01
-1.23581922e+00 -1.14885545e+00 -2.18733288e-02 1.81510603e+00
-3.83069366e-01 -8.78552973e-01 -7.52100885e-01 -1.04770374e+00
-7.07248747e-02 8.76077592e-01 -6.32421970e-01 -3.02444786e-01
-8.78665447e-01 -8.93224418e-01 1.97575718e-01 1.02807713e+00
4.42211628e-01 -1.12047768e+00 -1.74863651e-01 1.77939013e-01
-1.80022642e-02 -9.75764990e-01 -5.81375062e-01 6.22117400e-01
-1.30308425e+00 -7.70420194e-01 -2.03357011e-01 -9.20960844e-01
8.63482594e-01 1.48943514e-01 1.38472080e+00 3.19990516e-01
-3.83891985e-02 5.16409315e-02 -5.13374627e-01 -1.47879213e-01
-2.56602764e-01 4.49497402e-01 -1.18608110e-01 -5.72370768e-01
3.92764419e-01 -5.77646971e-01 -4.14407760e-01 -2.10028321e-01
-7.08122730e-01 1.46833658e-01 7.75245607e-01 1.19366169e+00
5.78235030e-01 -2.96605051e-01 4.78262037e-01 -1.10475802e+00
8.41753364e-01 -5.15066385e-01 -5.63878596e-01 5.77058569e-02
-6.59204245e-01 1.67087585e-01 6.51643872e-01 -7.35603154e-01
-9.68564391e-01 -7.59593770e-02 -3.71530831e-01 -3.70579481e-01
-1.65368035e-01 6.66471601e-01 1.97873525e-02 3.97740901e-01
7.29209900e-01 9.74543095e-02 -1.79338992e-01 -6.09840870e-01
6.04941607e-01 4.32475597e-01 8.12941432e-01 -7.97874749e-01
7.38490403e-01 -2.27662295e-01 -5.44072211e-01 -6.62904978e-01
-1.01710296e+00 -3.13313663e-01 -7.44684935e-01 2.75163442e-01
6.50055289e-01 -8.06478262e-01 -3.84210467e-01 2.35158622e-01
-1.25449705e+00 -7.41036057e-01 -4.85845923e-01 4.19730186e-01
-2.16286898e-01 1.80692464e-01 -7.43025720e-01 -5.59506059e-01
-4.36343461e-01 -1.07226217e+00 6.63424492e-01 5.41075226e-03
-3.18377018e-01 -1.10346460e+00 3.02998126e-01 3.56414795e-01
4.75513875e-01 -3.09050232e-01 1.19518566e+00 -1.12274778e+00
-1.83086336e-01 -1.21377930e-01 -4.59591895e-02 6.19367480e-01
-1.44074976e-01 -3.01828049e-02 -9.74776864e-01 -5.06011248e-01
-2.26152629e-01 -7.61539221e-01 9.47814703e-01 2.05958672e-02
1.72890520e+00 -3.14929813e-01 -1.48224756e-01 7.53373384e-01
1.05822480e+00 3.53177905e-01 3.55338156e-01 5.75029254e-01
6.20198011e-01 6.15038015e-02 5.43693960e-01 2.99555898e-01
2.23022580e-01 3.48867446e-01 3.47150385e-01 -2.67606288e-01
-5.08785322e-02 -3.21621954e-01 3.29813957e-01 1.15686035e+00
1.53926253e-01 -8.77840966e-02 -1.04684699e+00 5.21596789e-01
-1.72902966e+00 -7.16556311e-01 3.98325920e-02 2.14866686e+00
1.16247153e+00 5.92970729e-01 3.55755836e-02 2.18187273e-01
2.61219710e-01 1.11629941e-01 -5.16343057e-01 -7.22740829e-01
-5.35641424e-03 7.19906747e-01 1.81610510e-01 6.96958959e-01
-1.25806689e+00 1.31774378e+00 7.39312267e+00 6.86694026e-01
-8.59715462e-01 2.90841877e-01 7.09673405e-01 -3.56218278e-01
-4.33056444e-01 -1.71047583e-01 -1.03525424e+00 3.05754840e-01
1.27728462e+00 4.67505679e-02 3.29193473e-01 1.06826949e+00
-2.59988129e-01 -1.14871664e-02 -1.20262516e+00 5.78237653e-01
1.02110125e-01 -1.50606227e+00 8.48975703e-02 -2.95305818e-01
8.41174483e-01 3.18132155e-02 3.28418791e-01 1.03398848e+00
5.97570956e-01 -9.36701655e-01 2.32291237e-01 4.76969071e-02
5.51800251e-01 -1.13352442e+00 7.94430494e-01 3.23797375e-01
-7.82079875e-01 -1.81031048e-01 -3.66410106e-01 -3.44883025e-01
-1.33374836e-02 3.12538564e-01 -7.81017542e-01 7.72595108e-02
4.65050489e-01 3.74370754e-01 -8.45598102e-01 8.58446121e-01
-6.35897636e-01 9.57261860e-01 -1.96502563e-02 -1.56370271e-02
2.34710231e-01 -5.50107062e-02 2.50575691e-01 1.54321682e+00
1.49327908e-02 -2.76388098e-02 4.92744952e-01 4.39386278e-01
-2.88751513e-01 1.27881050e-01 -3.35530907e-01 1.50260285e-01
9.43709970e-01 1.24307680e+00 -2.70962089e-01 -6.73553944e-01
-5.27440250e-01 1.04099309e+00 9.51223254e-01 3.18175763e-01
-7.61491179e-01 -2.64370352e-01 5.08154392e-01 -1.33973807e-01
2.97101438e-01 -4.93149966e-01 -6.15796566e-01 -9.29267764e-01
-5.69399521e-02 -7.80995607e-01 4.61307406e-01 -5.51655531e-01
-1.06674767e+00 8.69523287e-01 -6.72619045e-03 -8.61344337e-01
-2.74279535e-01 -6.41778946e-01 -7.67471313e-01 6.95117295e-01
-1.53475797e+00 -8.75541210e-01 -2.11535901e-01 4.01220858e-01
8.51650774e-01 -3.84538203e-01 1.25613403e+00 2.16763765e-01
-7.77154326e-01 1.00651217e+00 4.29126620e-02 2.91489065e-01
8.26021314e-01 -1.64127994e+00 8.37085903e-01 7.56326675e-01
1.06327429e-01 7.39863098e-01 5.75315654e-01 -6.00900173e-01
-1.14012170e+00 -1.35290408e+00 7.90257514e-01 -4.86879647e-01
8.85105610e-01 -7.21446395e-01 -1.25647318e+00 9.10145044e-01
5.55795014e-01 1.17001578e-01 1.06621671e+00 7.25916088e-01
-5.47284245e-01 1.99255198e-02 -6.93753481e-01 7.26440310e-01
7.40744948e-01 -4.84929293e-01 -1.05765843e+00 4.32024419e-01
1.21748006e+00 -7.99217224e-01 -8.56427252e-01 4.35627371e-01
1.80605114e-01 -3.94705594e-01 8.31561565e-01 -1.04312205e+00
8.16490889e-01 1.66850269e-01 -3.35252509e-02 -1.60388339e+00
-3.11590970e-01 -4.79365110e-01 -6.50610209e-01 1.25759542e+00
7.10915565e-01 -2.52563298e-01 1.09921813e+00 4.27542835e-01
-5.27263463e-01 -9.68994498e-01 -5.51336646e-01 -8.02235246e-01
5.46655655e-01 -6.17892504e-01 5.33728004e-01 1.02907503e+00
2.04880372e-01 5.99344492e-01 -1.78832859e-01 -5.52905072e-03
4.79649901e-01 -2.55110443e-01 7.89357126e-01 -9.07211304e-01
-4.02067304e-01 -3.86260241e-01 2.39247024e-01 -1.19012523e+00
2.85211176e-01 -8.25924039e-01 -7.66307786e-02 -1.17015409e+00
3.31357718e-01 -4.63732719e-01 -5.26478052e-01 7.55859196e-01
-6.96784139e-01 1.20570675e-01 2.29170561e-01 -1.04975097e-01
-6.14419103e-01 6.04347110e-01 1.13539791e+00 -2.30843604e-01
-4.55618739e-01 -2.67040104e-01 -5.20862222e-01 6.09629512e-01
9.94729340e-01 -3.16408396e-01 -5.84676683e-01 -7.74033189e-01
8.13370645e-02 -1.82666227e-01 5.39247692e-02 -1.11644983e+00
2.92102277e-01 -1.16888471e-02 4.35955346e-01 -9.98055577e-01
3.55906844e-01 -5.35463333e-01 -6.17865682e-01 3.69340152e-01
-7.76480019e-01 6.24544442e-01 8.20762813e-01 3.03581595e-01
-1.56422272e-01 -2.92866975e-01 5.75629771e-01 1.18334986e-01
-6.54509306e-01 3.55996430e-01 -6.29530132e-01 2.06102908e-01
7.19162226e-01 1.20727763e-01 -5.06176233e-01 6.71725646e-02
-1.18462920e+00 4.10669506e-01 7.19854012e-02 4.41244543e-01
7.42089033e-01 -1.46313369e+00 -6.62469268e-01 4.83368516e-01
1.14814334e-01 2.78289050e-01 2.05614865e-01 5.21589041e-01
-1.12194806e-01 2.64203459e-01 -1.57660469e-02 -4.98160392e-01
-1.36501932e+00 9.04311180e-01 -1.39342576e-01 -6.81649089e-01
-4.90764439e-01 1.06303430e+00 -1.26434490e-02 -6.55784249e-01
7.09431410e-01 -4.12509382e-01 -2.66524136e-01 2.48737633e-05
8.11616659e-01 2.58859813e-01 3.18432510e-01 2.92540908e-01
-4.29784209e-02 1.34828985e-01 -8.91945124e-01 -2.18440592e-01
1.41193330e+00 2.94005573e-01 1.05347022e-01 3.70139241e-01
1.17076993e+00 -2.75858101e-02 -1.41579950e+00 -4.10892487e-01
1.60832703e-01 -1.64710909e-01 1.66016936e-01 -1.01153636e+00
-8.93006444e-01 1.04586267e+00 5.13606429e-01 -1.23315878e-01
1.17919135e+00 -2.14836299e-01 5.43093443e-01 6.78680420e-01
-1.76050365e-01 -1.20009661e+00 6.80741429e-01 1.01750827e+00
9.31823015e-01 -1.24137414e+00 -5.02168238e-02 -2.70373136e-01
-6.14988029e-01 9.27334964e-01 1.23303533e+00 -2.37493306e-01
5.36955416e-01 8.27126801e-01 -4.74845618e-03 2.62051553e-01
-1.25717926e+00 1.26962513e-01 3.49985570e-01 4.59373564e-01
7.77804971e-01 -1.28739968e-01 -2.60852158e-01 7.46462107e-01
-3.83465767e-01 -2.53224730e-01 3.29906583e-01 1.12806892e+00
-4.61406291e-01 -1.35016823e+00 -1.11716427e-01 3.94450486e-01
-1.69417590e-01 -5.65036178e-01 -1.62760466e-01 9.36485171e-01
-9.55606997e-02 8.00835073e-01 7.28386760e-01 -5.30373931e-01
5.10398507e-01 2.62614131e-01 3.49052966e-01 -9.60869968e-01
-7.92028129e-01 1.35221183e-02 1.77222103e-01 -2.93584645e-01
8.86172727e-02 -5.17778933e-01 -1.35762870e+00 -2.68941522e-01
-2.12033480e-01 3.21896493e-01 4.83958066e-01 9.73420918e-01
3.31768513e-01 9.13065016e-01 5.51066399e-01 -6.30600333e-01
-6.22176170e-01 -1.25956929e+00 1.00663178e-01 2.49047548e-01
2.07115009e-01 -4.82710361e-01 -3.11551094e-01 -4.59712967e-02] | [10.842473983764648, 8.362937927246094] |
98b887a2-b475-4d9f-91ec-2c77ed5febf8 | zero-shot-slot-filling-with-dpr-and-rag | 2104.08610 | null | https://arxiv.org/abs/2104.08610v1 | https://arxiv.org/pdf/2104.08610v1.pdf | Zero-shot Slot Filling with DPR and RAG | The ability to automatically extract Knowledge Graphs (KG) from a given collection of documents is a long-standing problem in Artificial Intelligence. One way to assess this capability is through the task of slot filling. Given an entity query in form of [Entity, Slot, ?], a system is asked to `fill' the slot by generating or extracting the missing value from a relevant passage or passages. This capability is crucial to create systems for automatic knowledge base population, which is becoming in ever-increasing demand, especially in enterprise applications. Recently, there has been a promising direction in evaluating language models in the same way we would evaluate knowledge bases, and the task of slot filling is the most suitable to this intent. The recent advancements in the field try to solve this task in an end-to-end fashion using retrieval-based language models. Models like Retrieval Augmented Generation (RAG) show surprisingly good performance without involving complex information extraction pipelines. However, the results achieved by these models on the two slot filling tasks in the KILT benchmark are still not at the level required by real-world information extraction systems. In this paper, we describe several strategies we adopted to improve the retriever and the generator of RAG in order to make it a better slot filler. Our KGI0 system (available at https://github.com/IBM/retrieve-write-slot-filling) reached the top-1 position on the KILT leaderboard on both T-REx and zsRE dataset with a large margin. | ['Alfio Gliozzo', 'Gaetano Rossiello', 'Michael Glass'] | 2021-04-17 | null | null | null | null | ['knowledge-base-population', 'zero-shot-slot-filling'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.92072168e-01 5.98555803e-01 -3.16503912e-01 -6.59644306e-02
-1.09790289e+00 -7.05622256e-01 7.92068064e-01 4.51605380e-01
-4.02982265e-01 1.03555179e+00 2.57032067e-01 -4.19303536e-01
-3.52355719e-01 -1.06942463e+00 -6.89470112e-01 -1.71413377e-01
1.03151731e-01 1.11327374e+00 5.52902341e-01 -6.05745494e-01
1.77725837e-01 -1.54190958e-02 -1.45145833e+00 5.33868551e-01
1.07219350e+00 8.53339314e-01 2.77612597e-01 4.49425370e-01
-9.00917709e-01 5.81687510e-01 -5.52445471e-01 -7.35074937e-01
1.27638668e-01 -2.98136741e-01 -1.23643243e+00 -3.38192970e-01
1.20078668e-01 2.53703475e-01 1.03697423e-02 7.06438661e-01
4.98939157e-01 -1.69217363e-01 4.86468256e-01 -1.04168987e+00
-2.30740115e-01 1.35746324e+00 -1.95378020e-01 -3.91292684e-02
5.28309703e-01 -2.64786422e-01 1.20341742e+00 -8.65216255e-01
1.08187258e+00 1.11599576e+00 3.24548036e-01 3.29993278e-01
-1.04250014e+00 -3.88290077e-01 -2.08206475e-01 2.34554663e-01
-1.44249213e+00 -3.90381843e-01 3.03310364e-01 -1.99078068e-01
1.43039441e+00 4.29173917e-01 4.72490132e-01 6.26105130e-01
-3.01515430e-01 9.79605079e-01 8.81204724e-01 -9.67029035e-01
2.26258840e-02 4.58564609e-01 1.79531127e-01 5.36281586e-01
4.42252219e-01 -3.22128147e-01 -7.63364553e-01 -2.12145850e-01
3.57057571e-01 -5.55086970e-01 -2.33309537e-01 -1.76197484e-01
-1.20111585e+00 6.34051442e-01 2.67596573e-01 5.24165571e-01
-4.41258430e-01 -2.06428841e-01 5.02101243e-01 2.61000663e-01
5.71747839e-01 9.97615516e-01 -8.98670137e-01 -2.57598668e-01
-1.07838809e+00 5.62353015e-01 1.20584559e+00 1.14068294e+00
6.79921210e-01 -5.91830611e-01 -6.03603721e-01 8.08959365e-01
4.84587252e-02 1.99054733e-01 3.53335470e-01 -4.34149176e-01
8.21571946e-01 9.57342327e-01 1.57811910e-01 -5.34095645e-01
-3.89380485e-01 -5.81372738e-01 -4.57553595e-01 -2.50829160e-01
5.34645259e-01 -2.14338349e-03 -1.08706808e+00 1.58185697e+00
4.62664008e-01 -1.26437217e-01 3.97916764e-01 4.77032065e-01
1.00451934e+00 6.67995811e-01 2.82506943e-01 -1.47717789e-01
1.69808865e+00 -7.83468306e-01 -7.89070845e-01 -4.08475876e-01
1.03409612e+00 -9.55630600e-01 9.13352013e-01 2.29517445e-01
-9.49255347e-01 -3.32221478e-01 -8.65730166e-01 -2.07717940e-01
-8.39284956e-01 9.56942663e-02 7.02129900e-01 5.31363428e-01
-9.07417178e-01 4.53435093e-01 -4.50512886e-01 -5.08788466e-01
1.33858904e-01 6.04539365e-03 -5.05245447e-01 -1.99386284e-01
-1.73008084e+00 1.06050885e+00 9.92783010e-01 -1.59086838e-01
-2.18434781e-01 -7.27375329e-01 -8.38820338e-01 1.21743716e-01
8.93242240e-01 -9.29678261e-01 1.25362396e+00 -2.13289022e-01
-9.51113582e-01 9.01970446e-01 -2.08846211e-01 -7.54225433e-01
4.60233033e-01 -3.10512066e-01 -3.49563390e-01 -1.87872708e-01
3.98322314e-01 7.05994368e-01 5.08762002e-01 -9.86268103e-01
-7.76975095e-01 -3.67173046e-01 7.79860020e-02 3.35398018e-01
-6.74751550e-02 8.42787977e-03 -8.77422094e-01 -3.18640709e-01
1.39822528e-01 -8.95654917e-01 -1.06609456e-01 -7.78946042e-01
-8.00441623e-01 -5.97740471e-01 4.24950182e-01 -8.92866969e-01
1.72233260e+00 -1.78220332e+00 -1.12453163e-01 3.03876609e-01
-1.35897845e-01 4.98240918e-01 1.41400799e-01 9.36859310e-01
2.07016952e-02 4.95445460e-01 -8.35137367e-02 -2.03025222e-01
2.44454026e-01 1.17961712e-01 -5.71172476e-01 -4.21130955e-01
2.40509227e-01 1.22944176e+00 -8.41080427e-01 -7.31961429e-01
-2.66584098e-01 1.89139470e-01 -4.64056253e-01 2.87670325e-02
-7.67696738e-01 4.15848531e-02 -4.86291647e-01 6.52690053e-01
2.58615553e-01 -2.74332255e-01 2.96788633e-01 -9.03565288e-02
-1.19109135e-02 7.72924900e-01 -1.30719435e+00 1.76007235e+00
-3.53483111e-01 3.04282069e-01 -4.00653899e-01 -6.75852597e-01
8.29203665e-01 4.46306974e-01 2.30135202e-01 -9.37450409e-01
-2.25778341e-01 6.75826311e-01 -9.79224145e-02 -3.69182914e-01
1.04718888e+00 4.69218083e-02 -4.02890176e-01 2.48742476e-01
1.51916355e-01 -1.97951242e-01 9.82771933e-01 5.08538842e-01
1.27349651e+00 2.67958552e-01 3.31456929e-01 -8.91927257e-02
4.67942506e-01 5.63631773e-01 3.15434158e-01 8.61590981e-01
5.18794000e-01 4.03403491e-01 6.11006021e-01 -8.22901651e-02
-9.84702349e-01 -6.14655554e-01 -1.48997992e-01 9.21211123e-01
-2.18820661e-01 -8.11325371e-01 -7.13493466e-01 -7.68437266e-01
1.21418320e-01 9.96018171e-01 -3.12711507e-01 -6.99890777e-02
-4.44292635e-01 -5.36675215e-01 7.34476507e-01 3.19850981e-01
4.98157531e-01 -1.37189245e+00 -3.76341999e-01 5.09512722e-01
-6.02877498e-01 -1.22853613e+00 2.97333617e-02 4.53864902e-01
-5.91264904e-01 -8.93660426e-01 -6.33582473e-01 -5.50244808e-01
4.12753582e-01 -1.26056179e-01 1.71294355e+00 1.28294811e-01
-2.03447968e-01 1.05773799e-01 -7.27826178e-01 -7.32353449e-01
-4.45524096e-01 7.54574597e-01 -3.54607403e-01 -5.58502495e-01
5.39439976e-01 -2.16530800e-01 -4.04320836e-01 1.77920341e-01
-1.26878214e+00 2.20599383e-01 7.23692119e-01 7.33631968e-01
6.46501124e-01 4.09124792e-01 8.26853096e-01 -1.25980330e+00
8.95308554e-01 -4.74053919e-01 -5.10297537e-01 5.93460560e-01
-8.70462120e-01 5.16751111e-01 2.13988736e-01 7.87456110e-02
-9.88741100e-01 -1.78460225e-01 -3.17541838e-01 1.78413808e-01
8.45979527e-02 1.20179856e+00 -1.73859730e-01 5.53091526e-01
8.19903195e-01 1.76273093e-01 -3.70320439e-01 -6.03880644e-01
6.29279256e-01 5.77958822e-01 3.70215714e-01 -7.83998728e-01
7.20706642e-01 -1.73059981e-02 -2.09621325e-01 -5.48442304e-01
-9.52215970e-01 -8.23585629e-01 -4.15047139e-01 1.38083160e-01
4.90634412e-01 -8.12453985e-01 -1.85569733e-01 1.49660870e-01
-1.00707090e+00 -3.23037565e-01 -6.24406874e-01 5.42671382e-02
-3.81262958e-01 1.53851792e-01 -2.60872245e-01 -6.85638309e-01
-7.26560235e-01 -7.21646130e-01 1.27764165e+00 2.98315197e-01
-3.61483335e-01 -7.33877361e-01 1.02516003e-01 6.09347761e-01
4.32615310e-01 -5.13267927e-02 1.22793996e+00 -9.39446032e-01
-7.17301011e-01 -2.92705715e-01 -2.45941505e-01 5.79730123e-02
-1.18702978e-01 -2.86707789e-01 -6.84833527e-01 4.00428567e-03
-6.39236093e-01 -3.72726917e-01 9.95021105e-01 2.23194640e-02
7.64190912e-01 -1.37494788e-01 -4.86100286e-01 1.34922057e-01
1.39040434e+00 1.80332929e-01 9.26034629e-01 7.17792511e-01
2.48835757e-01 6.63433433e-01 1.06549740e+00 1.20047234e-01
4.23748493e-01 8.44470203e-01 9.80490521e-02 6.30467534e-02
-2.14907691e-01 -6.18814468e-01 3.77978459e-02 4.79183972e-01
1.46369845e-01 -3.99760008e-01 -1.16680634e+00 7.83000171e-01
-1.93489599e+00 -9.00083303e-01 -1.15294242e-02 2.23329496e+00
1.31237042e+00 5.41251898e-01 -6.66252226e-02 4.06823725e-01
2.64685273e-01 -2.06510231e-01 -2.78891116e-01 -2.65878439e-01
-2.14735314e-01 4.57258850e-01 3.98492843e-01 4.11294162e-01
-8.81943822e-01 1.50452018e+00 4.92884541e+00 1.11820996e+00
-7.34831214e-01 -3.01679432e-01 4.13965791e-01 4.41830635e-01
-4.07674134e-01 4.41762775e-01 -1.46725678e+00 2.94062167e-01
1.06631434e+00 -5.30241251e-01 1.82724878e-01 7.77322173e-01
-2.14453563e-01 -5.10774195e-01 -8.24398577e-01 5.51799953e-01
-1.94255441e-01 -1.47674906e+00 1.71599314e-01 1.78763401e-02
4.48759019e-01 -1.61592647e-01 -2.20705092e-01 8.53812337e-01
2.37856761e-01 -9.69748795e-01 6.63536906e-01 5.86566627e-01
7.80315459e-01 -5.66763043e-01 9.02342200e-01 6.29424155e-01
-1.12592709e+00 2.09578142e-01 -2.70535707e-01 2.59852946e-01
3.82458806e-01 9.67900038e-01 -1.79251337e+00 8.80430043e-01
5.88704288e-01 8.73584598e-02 -8.90625000e-01 1.24408555e+00
-3.65961760e-01 4.99346823e-01 -5.42642832e-01 -6.12694025e-02
7.56051987e-02 6.89478964e-02 4.87575889e-01 1.31707013e+00
4.19679374e-01 -1.38207644e-01 2.98401318e-03 7.67108083e-01
-1.64462492e-01 4.34744120e-01 -5.82763672e-01 -3.75677675e-01
5.78842640e-01 1.31336129e+00 -8.11859250e-01 -5.38036346e-01
-9.37134922e-02 7.17547834e-01 4.20442641e-01 1.01617716e-01
-4.46075290e-01 -5.77540219e-01 2.07484528e-01 2.78700978e-01
1.93087548e-01 -1.16688378e-01 -1.20846719e-01 -9.34607327e-01
3.40656817e-01 -9.93038058e-01 5.96536338e-01 -8.41275990e-01
-8.13556373e-01 7.48989284e-01 2.85159349e-01 -8.85650456e-01
-8.66864622e-01 -3.80587459e-01 -1.15063377e-02 1.11391318e+00
-1.62898242e+00 -1.12243700e+00 -9.17418301e-02 4.75601584e-01
4.60711122e-01 1.08811609e-01 1.07397580e+00 4.97806966e-01
-1.45804405e-01 3.88047904e-01 -2.19857007e-01 1.74395770e-01
6.99185491e-01 -1.53227150e+00 6.82645917e-01 7.13246286e-01
5.74235976e-01 8.14068019e-01 8.43271852e-01 -8.96595716e-01
-1.11906302e+00 -8.76028061e-01 1.65153146e+00 -7.35485733e-01
5.97205698e-01 -3.15932602e-01 -1.02534366e+00 4.94452149e-01
1.32629007e-01 -3.21886480e-01 4.90362495e-01 5.29781163e-01
-2.87601829e-01 1.48362309e-01 -8.28561425e-01 5.04207611e-01
9.97341931e-01 -4.84341621e-01 -7.46728301e-01 3.47810060e-01
8.39072347e-01 -6.23249888e-01 -8.27840984e-01 6.03822351e-01
2.33312100e-01 -6.91871762e-01 9.11883056e-01 -5.02041698e-01
1.75041094e-01 -4.91150975e-01 8.70532840e-02 -1.34472764e+00
4.50893082e-02 -6.00877285e-01 -2.19838813e-01 1.59756768e+00
1.11307013e+00 -4.79731411e-01 9.22012806e-01 6.34344876e-01
-6.54654056e-02 -7.57145226e-01 -7.61134684e-01 -7.50177145e-01
-6.16404116e-02 -5.20264626e-01 7.69386351e-01 6.19657874e-01
2.79785037e-01 6.51719391e-01 -3.52572277e-02 -1.86645687e-01
1.96747273e-01 2.23564789e-01 9.19103205e-01 -1.29316700e+00
-1.27946451e-01 -2.32520074e-01 -4.87868264e-02 -7.79886842e-01
-2.53291223e-02 -1.27951181e+00 -1.36201978e-01 -2.07476926e+00
1.63554579e-01 -7.44225800e-01 -1.38321057e-01 7.98525989e-01
-3.76610488e-01 -8.17808658e-02 1.13791592e-01 1.34890333e-01
-7.32976496e-01 2.11686611e-01 9.77962792e-01 -6.27011955e-02
-3.64401728e-01 4.51915786e-02 -9.76077974e-01 2.37751752e-01
8.47423255e-01 -4.63946640e-01 -4.85191345e-01 -1.88030064e-01
8.09473813e-01 2.26157650e-01 6.90590590e-02 -7.79740393e-01
4.45443630e-01 2.00413972e-01 1.47333309e-01 -7.50580251e-01
1.90400630e-01 -5.68832695e-01 3.61805975e-01 8.92593414e-02
-2.69111067e-01 8.82919878e-02 3.99076492e-01 7.09486082e-02
-4.90824401e-01 -3.83969843e-01 1.79131344e-01 -2.84453690e-01
-9.41363513e-01 1.47652701e-01 8.52314755e-02 2.53582925e-01
4.36329007e-01 1.87867116e-02 -6.20975494e-01 -1.12342171e-01
-5.77013791e-01 3.31852466e-01 1.47845238e-01 5.51226914e-01
2.72639453e-01 -1.00949574e+00 -7.39205420e-01 9.87695158e-02
5.41612267e-01 2.57408589e-01 4.39094426e-03 6.99634016e-01
-3.88922751e-01 9.66504514e-01 1.14054784e-01 -3.12078208e-01
-1.09982955e+00 3.15633327e-01 1.21477591e-02 -1.18804622e+00
-3.86750787e-01 6.90070689e-01 -4.55100477e-01 -4.79215473e-01
2.67511457e-02 -2.67089695e-01 -5.56552351e-01 2.95913637e-01
5.56448519e-01 -6.97751343e-02 6.56521142e-01 -3.26841891e-01
-1.43703803e-01 -1.30836805e-02 -3.51257235e-01 -3.43436807e-01
1.18548071e+00 1.19076908e-01 -2.77771920e-01 4.24943939e-02
7.29212523e-01 1.39449716e-01 -4.18902397e-01 -3.88500243e-01
8.04737151e-01 -2.65319258e-01 -8.32847357e-02 -1.26226330e+00
-5.81901312e-01 3.84505779e-01 1.19525857e-01 7.04156280e-01
9.34753418e-01 2.65173286e-01 7.34056711e-01 5.25363445e-01
7.46074438e-01 -1.19898927e+00 -3.27451229e-01 7.80652642e-01
8.99854302e-01 -1.19086599e+00 -7.94907957e-02 -6.01726532e-01
-5.45610309e-01 7.61309683e-01 4.07830119e-01 5.06655276e-01
4.01268899e-01 8.45230445e-02 -2.34997142e-02 -4.72231358e-01
-9.22431231e-01 -7.65734673e-01 4.40085024e-01 3.98233175e-01
5.99881172e-01 -4.99396771e-02 -6.56785965e-01 5.13520002e-01
-5.68799913e-01 1.38779849e-01 8.84063393e-02 1.07756746e+00
-5.08213580e-01 -1.64033878e+00 -1.58526838e-01 5.89802027e-01
-6.82042837e-01 -4.21568930e-01 -4.70504820e-01 8.95512640e-01
-1.06370717e-01 8.54223847e-01 -4.91678983e-01 -1.51433975e-01
6.71769381e-01 5.16415417e-01 1.99671596e-01 -9.09379900e-01
-4.93949652e-01 -1.12093883e-02 6.68878555e-01 -3.73614162e-01
-1.95326656e-01 -5.95221937e-01 -1.30972767e+00 3.36325401e-03
-4.90910679e-01 7.19279528e-01 7.53801465e-01 8.95151556e-01
4.22445387e-01 4.43613768e-01 6.11449964e-03 -2.90541768e-01
-3.07575852e-01 -1.08702457e+00 -5.03467977e-01 4.31762189e-01
-3.23143363e-01 -5.26028931e-01 -1.08055077e-01 -9.31319296e-02] | [9.684585571289062, 8.472850799560547] |
d63b2b49-0a95-44ec-8042-289eaa9edbe7 | on-complex-valued-convolutional-neural | 1602.09046 | null | http://arxiv.org/abs/1602.09046v1 | http://arxiv.org/pdf/1602.09046v1.pdf | On Complex Valued Convolutional Neural Networks | Convolutional neural networks (CNNs) are the cutting edge model for
supervised machine learning in computer vision. In recent years CNNs have
outperformed traditional approaches in many computer vision tasks such as
object detection, image classification and face recognition. CNNs are
vulnerable to overfitting, and a lot of research focuses on finding
regularization methods to overcome it. One approach is designing task specific
models based on prior knowledge.
Several works have shown that properties of natural images can be easily
captured using complex numbers. Motivated by these works, we present a
variation of the CNN model with complex valued input and weights. We construct
the complex model as a generalization of the real model. Lack of order over the
complex field raises several difficulties both in the definition and in the
training of the network. We address these issues and suggest possible
solutions.
The resulting model is shown to be a restricted form of a real valued CNN
with twice the parameters. It is sensitive to phase structure, and we suggest
it serves as a regularized model for problems where such structure is
important. This suggestion is verified empirically by comparing the performance
of a complex and a real network in the problem of cell detection. The two
networks achieve comparable results, and although the complex model is hard to
train, it is significantly less vulnerable to overfitting. We also demonstrate
that the complex network detects meaningful phase structure in the data. | ['Nitzan Guberman'] | 2016-02-29 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [ 3.46149921e-01 3.25475842e-01 -1.19932063e-01 -3.30036670e-01
3.57587226e-02 -2.20982164e-01 5.25816083e-01 8.69824290e-02
-7.99895763e-01 7.31720984e-01 -3.15205991e-01 -2.94109464e-01
-1.06006801e-01 -6.43105209e-01 -7.89149582e-01 -9.00457859e-01
-1.03889056e-01 2.51453161e-01 4.82730478e-01 -3.61996710e-01
2.40156174e-01 8.28127682e-01 -1.34928524e+00 5.89116178e-02
3.78179312e-01 1.05648649e+00 -4.78746593e-02 5.74308932e-01
7.59898424e-02 7.09850848e-01 -5.78904212e-01 -6.99782968e-02
3.35011423e-01 -2.60891140e-01 -7.53904104e-01 8.73490348e-02
4.02735382e-01 3.67393941e-01 -6.21904694e-02 1.11802876e+00
4.94937629e-01 -1.14289902e-01 7.68774331e-01 -1.24710035e+00
-3.18188488e-01 1.43593788e-01 -6.07576966e-01 3.47666562e-01
-2.28713959e-01 7.57888332e-03 6.65609360e-01 -8.09468567e-01
5.13864100e-01 1.16332197e+00 1.24682403e+00 8.25729489e-01
-1.33463573e+00 -4.46913570e-01 -1.60556048e-01 -1.21099822e-01
-1.42859292e+00 -2.37730414e-01 7.53892422e-01 -4.52932566e-01
8.92623425e-01 1.62759423e-01 5.92469752e-01 7.06118643e-01
1.84789822e-01 4.10638064e-01 1.06637692e+00 -7.09247053e-01
2.28755429e-01 2.67453521e-01 3.23237240e-01 1.05138016e+00
5.76579571e-01 8.39560330e-02 -1.71041071e-01 -1.32928535e-01
1.14777446e+00 -2.55721092e-01 -5.22063792e-01 -3.11567545e-01
-1.04868090e+00 8.93011570e-01 5.61092377e-01 5.76986969e-01
-1.68826669e-01 2.83361197e-01 2.90782452e-01 2.02828482e-01
4.00023460e-01 8.37062180e-01 -4.02605772e-01 2.74642467e-01
-7.16424942e-01 2.34001860e-01 9.35931802e-01 6.82312250e-01
8.30523670e-01 1.49323627e-01 1.18931577e-01 7.86609173e-01
8.51736739e-02 1.43323898e-01 4.73125905e-01 -8.95141840e-01
3.42493206e-02 6.17671192e-01 -8.70798379e-02 -1.10402703e+00
-9.13266957e-01 -5.31758368e-01 -1.31900179e+00 5.73440552e-01
1.02229488e+00 -2.72775497e-02 -9.88902807e-01 1.67212856e+00
-1.09621368e-01 5.41097485e-02 -3.29685472e-02 6.56999826e-01
7.24693537e-01 5.73654473e-01 -1.59898311e-01 -1.99188069e-01
1.17600787e+00 -5.02145171e-01 -6.56990528e-01 -5.31683601e-02
6.49353325e-01 -6.41214490e-01 8.61568034e-01 5.55602431e-01
-9.61543143e-01 -4.11144078e-01 -1.30322516e+00 8.26242566e-02
-5.26130378e-01 1.90422252e-01 6.92829907e-01 7.04240680e-01
-1.09885967e+00 9.40486908e-01 -7.11622179e-01 -5.18733978e-01
4.58996236e-01 7.71171510e-01 -3.30741614e-01 2.43611053e-01
-9.75243628e-01 9.74718750e-01 3.50386620e-01 3.71499747e-01
-5.70353270e-02 -4.17308390e-01 -7.72846699e-01 5.79669140e-02
-5.05822785e-02 -3.39758158e-01 1.08803618e+00 -1.11528730e+00
-1.14761400e+00 1.07680857e+00 -1.34719014e-01 -5.38875103e-01
3.81748766e-01 3.45716983e-01 -2.14825168e-01 6.50685430e-02
-3.09568584e-01 6.76810324e-01 8.46746087e-01 -1.22494447e+00
-3.00563753e-01 -1.28394023e-01 5.50434813e-02 -4.73437190e-01
-3.86666179e-01 2.32691108e-03 -1.00825004e-01 -6.49079859e-01
2.86841929e-01 -9.94399250e-01 -5.51760674e-01 4.54326302e-01
-2.62178749e-01 -2.82283217e-01 8.00754726e-01 -8.27030372e-03
8.35199475e-01 -1.99096203e+00 -3.61513317e-01 5.32121778e-01
3.66933763e-01 5.61139107e-01 -4.73672375e-02 4.67842594e-02
-4.48498338e-01 3.81573260e-01 -2.96463549e-01 -7.63667375e-02
-3.81238520e-01 4.27794188e-01 7.29364082e-02 7.92111754e-01
7.51244485e-01 8.23646545e-01 -7.70609021e-01 -4.21915889e-01
-2.31679797e-01 6.92206800e-01 -4.46371078e-01 -1.01440445e-01
-4.31742817e-02 8.49035606e-02 -1.57394797e-01 3.51282299e-01
8.21021378e-01 -5.59220016e-01 2.87918635e-02 -3.98810834e-01
-1.66194379e-01 -7.23545328e-02 -1.31019425e+00 7.29077816e-01
-5.72170243e-02 1.24625003e+00 1.36402741e-01 -1.63906229e+00
1.01432848e+00 3.40336353e-01 4.26392496e-01 -5.53171098e-01
1.88781396e-01 2.98583657e-01 2.42709070e-01 -6.19485199e-01
2.30094790e-01 -2.80833989e-01 4.81515259e-01 2.21644253e-01
1.12525903e-01 -1.01460695e-01 2.59684950e-01 -1.22981325e-01
9.83241558e-01 -3.80757183e-01 3.69699121e-01 -6.97442889e-01
5.10046780e-01 -8.76142308e-02 7.25478888e-01 6.58461273e-01
-1.30780295e-01 8.25861216e-01 7.91924059e-01 -9.19715047e-01
-9.79035556e-01 -6.59659028e-01 -3.71633321e-01 4.39773440e-01
9.83393714e-02 -2.70881653e-01 -7.18142211e-01 -2.57794917e-01
-3.08879972e-01 -1.81623459e-01 -8.94289017e-01 -1.14902951e-01
-9.57862973e-01 -1.20230114e+00 8.09297383e-01 7.06865311e-01
6.27797961e-01 -1.07660425e+00 -6.86569631e-01 1.32567570e-01
6.59827217e-02 -1.36131239e+00 -1.11002862e-01 6.90450013e-01
-9.33268011e-01 -1.48768365e+00 -6.98418260e-01 -1.17786920e+00
9.88398135e-01 1.51775897e-01 1.21040905e+00 6.15872025e-01
-3.98050100e-01 2.42736876e-01 -4.57139947e-02 -7.06673682e-01
-4.79363710e-01 -3.16711664e-02 1.60690024e-01 3.58675830e-02
3.51535082e-01 -3.16667318e-01 -3.36759269e-01 2.96660304e-01
-1.12973630e+00 -1.49754047e-01 4.78598654e-01 1.28779888e+00
4.73045468e-01 2.36781567e-01 4.66948897e-01 -9.94807065e-01
6.80698276e-01 -2.42476165e-02 -6.82769537e-01 1.70772076e-02
-6.46337509e-01 2.92849034e-01 7.83708215e-01 -7.36551642e-01
-5.70897937e-01 3.70164871e-01 -2.49550208e-01 -7.37257302e-02
-9.89618897e-02 4.14375365e-01 2.19934449e-01 -8.39199841e-01
9.70858276e-01 -2.14505538e-01 1.81931451e-01 -1.97481140e-01
-4.86572087e-01 2.59548008e-01 4.16499078e-01 -4.17684227e-01
8.49802494e-01 6.09517395e-01 6.37861729e-01 -1.48373663e+00
-6.34300709e-01 -5.01543522e-01 -5.67063153e-01 -2.20922977e-01
7.27069378e-01 -5.13714612e-01 -8.00175607e-01 6.45635962e-01
-1.44975710e+00 -4.18139011e-01 -4.60199296e-01 4.76247221e-01
-4.83427733e-01 2.43480176e-01 -6.87921941e-01 -7.26576507e-01
5.70099354e-02 -1.00493526e+00 6.82371020e-01 4.52206284e-01
-6.72668442e-02 -1.33576000e+00 -1.44951612e-01 -3.58306259e-01
5.37665546e-01 3.59019578e-01 1.00982010e+00 -5.36255240e-01
-3.44373763e-01 -2.73150206e-01 -3.64868104e-01 5.72442532e-01
-2.32650563e-01 2.55479038e-01 -1.20315504e+00 -2.31809676e-01
2.83221692e-01 -2.41358593e-01 1.09707117e+00 6.51573241e-01
1.22990251e+00 -1.72612563e-01 -2.45123252e-01 7.34071255e-01
1.54783857e+00 1.07961111e-01 9.30050433e-01 3.30263376e-01
3.90407205e-01 7.17580438e-01 3.03316079e-02 5.57273366e-02
-2.45561659e-01 5.52038193e-01 4.96791154e-01 -6.69037282e-01
1.06714861e-02 2.12403029e-01 -3.97904739e-02 6.80627525e-01
-4.10419524e-01 2.23332457e-03 -8.26221228e-01 3.02904367e-01
-1.69214725e+00 -9.34679985e-01 -4.64298695e-01 2.04938364e+00
6.47627532e-01 4.58489120e-01 2.92737801e-02 5.94269156e-01
5.88771462e-01 -8.73439461e-02 -1.81495875e-01 -3.80943865e-01
-4.95584369e-01 5.13364553e-01 6.81292832e-01 4.15814161e-01
-1.08630216e+00 5.63511252e-01 7.10827351e+00 7.06156671e-01
-1.50019014e+00 -2.63922453e-01 6.77872896e-01 5.31238675e-01
1.24159366e-01 -2.23175108e-01 -8.12837780e-01 4.27562632e-02
4.78017479e-01 2.88620681e-01 5.03465831e-02 4.77307647e-01
1.37041345e-01 -2.16578588e-01 -1.10815263e+00 9.74371314e-01
2.97753215e-02 -1.63618624e+00 -1.24756113e-01 1.69617310e-01
5.04554510e-01 -1.44645557e-01 -2.42886506e-02 7.14156032e-02
-2.55598187e-01 -1.37238336e+00 3.15220058e-01 3.97577107e-01
6.31635249e-01 -5.73600411e-01 9.52011764e-01 3.56247902e-01
-1.18501234e+00 2.59176176e-02 -6.00719213e-01 -2.44691089e-01
-3.76679122e-01 7.18110144e-01 -7.11603701e-01 1.08916704e-02
6.28221989e-01 5.51299393e-01 -7.51877069e-01 1.33257484e+00
3.03900063e-01 5.97743928e-01 -4.82979596e-01 -2.65628219e-01
2.68978477e-01 -4.85881492e-02 3.41652453e-01 1.42615485e+00
1.36878550e-01 1.93074979e-02 -5.21697663e-02 8.97335768e-01
2.58578118e-02 -5.82954660e-02 -8.14506173e-01 -1.14965923e-02
-1.26798570e-01 1.27707243e+00 -1.25398326e+00 -3.67105678e-02
-3.97292465e-01 3.70073378e-01 3.19279373e-01 6.04495347e-01
-3.42550337e-01 -5.46189666e-01 3.46279353e-01 3.31569970e-01
1.95095882e-01 -1.76424935e-01 -3.45020235e-01 -1.06746387e+00
4.19306867e-02 -6.12224221e-01 -1.16241863e-02 -5.78810692e-01
-1.17721438e+00 7.21577704e-01 -1.46190360e-01 -1.17888081e+00
-1.51119456e-01 -1.39032161e+00 -7.64641285e-01 6.65406823e-01
-1.53551793e+00 -6.54702246e-01 -3.37815315e-01 4.65420455e-01
6.92650974e-02 -3.71568799e-02 9.37095821e-01 2.39858568e-01
-4.41336423e-01 5.59656858e-01 -1.85416043e-01 4.92148399e-01
5.20136237e-01 -1.29207206e+00 2.44277924e-01 6.40386939e-01
-1.56321615e-01 6.10363901e-01 8.19761932e-01 -1.44806519e-01
-9.53883231e-01 -9.81409431e-01 1.05116045e+00 1.07391123e-02
3.36614370e-01 -3.80043387e-01 -1.07471228e+00 3.78043950e-01
-2.61576287e-02 4.04305965e-01 4.37348187e-01 -8.04315135e-02
-3.64493430e-01 -1.29874563e-02 -1.22830808e+00 4.88618404e-01
5.44692516e-01 -2.91787118e-01 -2.01509252e-01 3.73863459e-01
3.06957066e-01 -3.30517925e-02 -5.61381817e-01 5.41980386e-01
5.55402994e-01 -1.11916888e+00 9.12138224e-01 -5.04108667e-01
1.55820847e-01 -1.67811751e-01 2.33582094e-01 -1.24296796e+00
-3.46723527e-01 -7.17941225e-01 9.16992351e-02 6.60480380e-01
3.67504179e-01 -7.87968993e-01 1.12320280e+00 2.24182755e-01
2.92194821e-02 -1.13739705e+00 -9.15956855e-01 -9.98354316e-01
2.03963682e-01 -2.48342380e-01 1.22446176e-02 1.00062263e+00
-1.02668077e-01 5.06373823e-01 -3.64631750e-02 9.81490407e-03
5.71945548e-01 -4.16566014e-01 4.89853591e-01 -1.75994337e+00
-7.69646019e-02 -8.14100921e-01 -7.70394266e-01 -9.91907656e-01
9.33163688e-02 -5.86446166e-01 7.82244802e-02 -1.19138288e+00
-9.07858834e-02 -5.85659564e-01 2.17593350e-02 3.54430467e-01
2.48940051e-01 6.00463867e-01 2.39796508e-02 -6.75716847e-02
-8.56122300e-02 5.92247583e-02 1.29474533e+00 -2.14146808e-01
-1.15573257e-01 2.89986700e-01 -4.93391991e-01 1.17498827e+00
9.90612268e-01 -5.41985035e-01 -3.92301008e-02 -5.70813287e-03
5.86845040e-01 -3.60520571e-01 6.25752926e-01 -1.31992197e+00
4.49088484e-01 1.89279020e-02 6.41751766e-01 -2.78554648e-01
4.13147420e-01 -9.76011157e-01 -2.54502982e-01 6.91956699e-01
-3.29392195e-01 1.65552422e-01 2.11129606e-01 3.56700391e-01
-2.46720374e-01 -7.48076916e-01 1.24133551e+00 -3.16002339e-01
-4.94987726e-01 2.21859217e-01 -5.91790080e-01 9.70834717e-02
6.11388445e-01 -5.22931695e-01 -3.23101610e-01 -4.30589736e-01
-6.05490267e-01 -6.80213049e-02 2.55437672e-01 -1.83056295e-01
5.79145074e-01 -1.15251064e+00 -2.82122284e-01 5.48720419e-01
-1.50240704e-01 1.39448285e-01 -4.81999457e-01 9.05526757e-01
-8.35539997e-01 4.58544612e-01 -3.58171135e-01 -6.81196868e-01
-1.19389617e+00 3.15347105e-01 8.40065718e-01 -3.52949858e-01
-3.81286114e-01 7.80330956e-01 1.58575743e-01 -1.80897161e-01
5.48512757e-01 -7.80827701e-01 -4.71519142e-01 -1.66744903e-01
5.49911797e-01 1.67286575e-01 1.67171016e-01 -5.78507900e-01
-1.75978258e-01 8.33619475e-01 8.71456489e-02 1.42886311e-01
1.41527355e+00 3.64816248e-01 -2.83142865e-01 3.94802988e-01
1.30272043e+00 -2.29472607e-01 -1.08867693e+00 -1.97261378e-01
2.02439278e-01 5.96033409e-02 -1.34042010e-01 -2.47534156e-01
-1.19031775e+00 1.20284021e+00 5.63761294e-01 7.25385547e-01
1.12654221e+00 -2.40684897e-01 4.98276472e-01 7.93697834e-01
-7.17931241e-02 -1.08656800e+00 3.06048572e-01 9.22456563e-01
7.52084732e-01 -1.26818264e+00 -4.84202895e-03 -6.56411648e-01
-5.86592928e-02 1.56840742e+00 6.84159875e-01 -5.75865567e-01
1.04033601e+00 5.30212760e-01 1.21205837e-01 -4.09124047e-01
-5.09682953e-01 -3.17039222e-01 2.98768133e-01 8.41427445e-01
5.51269710e-01 -3.15415442e-01 -2.79340029e-01 3.06006104e-01
-4.75647561e-02 6.55757189e-02 7.81037033e-01 9.34421778e-01
-5.09846449e-01 -8.35038006e-01 -3.45345289e-01 4.91275221e-01
-6.93889856e-01 3.41281593e-02 -3.35985243e-01 1.02824855e+00
2.82740653e-01 5.79030991e-01 1.64136291e-01 -1.88405052e-01
3.58278334e-01 -1.76198602e-01 7.32649863e-01 -5.08441031e-01
-6.26533985e-01 -4.54825051e-02 -1.11037314e-01 -1.98107570e-01
-8.25450003e-01 -1.53604254e-01 -1.25603557e+00 -1.27781376e-01
-5.19384027e-01 1.65156484e-01 5.80277264e-01 1.01727879e+00
-1.47190362e-01 4.74223405e-01 3.15948993e-01 -1.12201452e+00
-5.51760972e-01 -8.00347447e-01 -7.08704233e-01 2.68855721e-01
6.59993887e-01 -6.29886925e-01 -5.26280642e-01 1.81446895e-01] | [9.076118469238281, 2.391167163848877] |
5aeef28a-7539-450d-b899-1d5ef649ac57 | 3d-shapenets-a-deep-representation-for | 1406.5670 | null | http://arxiv.org/abs/1406.5670v3 | http://arxiv.org/pdf/1406.5670v3.pdf | 3D ShapeNets: A Deep Representation for Volumetric Shapes | 3D shape is a crucial but heavily underutilized cue in today's computer
vision systems, mostly due to the lack of a good generic shape representation.
With the recent availability of inexpensive 2.5D depth sensors (e.g. Microsoft
Kinect), it is becoming increasingly important to have a powerful 3D shape
representation in the loop. Apart from category recognition, recovering full 3D
shapes from view-based 2.5D depth maps is also a critical part of visual
understanding. To this end, we propose to represent a geometric 3D shape as a
probability distribution of binary variables on a 3D voxel grid, using a
Convolutional Deep Belief Network. Our model, 3D ShapeNets, learns the
distribution of complex 3D shapes across different object categories and
arbitrary poses from raw CAD data, and discovers hierarchical compositional
part representations automatically. It naturally supports joint object
recognition and shape completion from 2.5D depth maps, and it enables active
object recognition through view planning. To train our 3D deep learning model,
we construct ModelNet -- a large-scale 3D CAD model dataset. Extensive
experiments show that our 3D deep representation enables significant
performance improvement over the-state-of-the-arts in a variety of tasks. | ['Jianxiong Xiao', 'Fisher Yu', 'Aditya Khosla', 'Zhirong Wu', 'Xiaoou Tang', 'Shuran Song', 'Linguang Zhang'] | 2014-06-22 | 3d-shapenets-a-deep-representation-for-1 | http://openaccess.thecvf.com/content_cvpr_2015/html/Wu_3D_ShapeNets_A_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Wu_3D_ShapeNets_A_2015_CVPR_paper.pdf | cvpr-2015-6 | ['3d-shape-representation'] | ['computer-vision'] | [-1.55027444e-02 1.55079484e-01 -1.61631033e-01 -5.78818262e-01
-6.83887124e-01 -7.23790944e-01 6.18084669e-01 -2.21779067e-02
2.18595304e-02 -1.41784619e-03 3.07421476e-01 -1.50133759e-01
5.84826954e-02 -9.54153657e-01 -8.53936374e-01 -3.04744214e-01
1.17650636e-01 1.26460302e+00 1.96411118e-01 1.21317901e-01
2.10497692e-01 1.11793220e+00 -1.62921453e+00 2.12000504e-01
3.43262136e-01 1.32285047e+00 2.83301324e-01 4.33316976e-01
-4.90282506e-01 1.38751641e-01 -1.19886607e-01 -1.32099152e-01
5.22593200e-01 2.84678191e-01 -4.68034446e-01 6.50281310e-01
6.06387377e-01 -7.79946744e-01 -3.56510580e-01 8.87300134e-01
3.16856027e-01 -1.59806535e-01 8.70088458e-01 -1.02905822e+00
-7.16111898e-01 -2.46904101e-02 -5.64785719e-01 -4.25503641e-01
4.04782802e-01 2.42795467e-01 9.47825134e-01 -1.24312258e+00
7.33231187e-01 1.53511035e+00 5.17690837e-01 5.40603995e-01
-1.40825748e+00 -3.29044670e-01 2.76717693e-01 -2.78814256e-01
-1.27181506e+00 -2.40865022e-01 1.06603575e+00 -7.24959970e-01
1.23549128e+00 -7.13170990e-02 1.01062894e+00 9.46237862e-01
1.30483091e-01 9.74704683e-01 9.08866405e-01 -8.17931816e-02
4.42597032e-01 -4.64365095e-01 -2.53170758e-01 9.29180503e-01
2.43512362e-01 1.45291062e-02 -3.73365879e-01 -9.79911163e-02
1.59300506e+00 6.03358686e-01 5.86938020e-03 -1.30821228e+00
-1.04599845e+00 6.64269030e-01 6.77247524e-01 -2.21225411e-01
-4.72994238e-01 3.88430476e-01 -4.93559502e-02 -2.75309775e-02
7.07530320e-01 1.91205069e-01 -6.40194178e-01 -1.44500405e-01
-3.78525972e-01 4.04567629e-01 6.65952325e-01 1.07719338e+00
7.27934599e-01 -4.75777611e-02 5.18973768e-01 6.56549037e-01
7.85512865e-01 7.93764651e-01 1.60829758e-03 -1.07422185e+00
3.04153174e-01 1.29892957e+00 -2.96526775e-02 -5.95445395e-01
-3.82012069e-01 -2.09501430e-01 -7.95163989e-01 7.95346618e-01
2.77279496e-01 5.35759568e-01 -1.36352348e+00 1.30529511e+00
4.19778705e-01 -2.39382684e-01 -4.17751908e-01 1.11386180e+00
9.81015146e-01 2.51455903e-01 -3.40065897e-01 4.71311808e-01
1.05506873e+00 -3.42693359e-01 -6.28050044e-02 -3.64842981e-01
1.82921693e-01 -4.82249767e-01 8.67794991e-01 4.52037960e-01
-1.12887025e+00 -4.51526999e-01 -8.88167858e-01 -4.08872008e-01
-3.25002968e-01 -2.46501744e-01 8.57895434e-01 3.88134807e-01
-8.91663074e-01 4.24688756e-01 -1.22148573e+00 -2.05343127e-01
9.47351277e-01 4.45383400e-01 -7.62481213e-01 -5.91032743e-01
-2.80053318e-01 7.53588498e-01 5.78165278e-02 -1.14943482e-01
-1.00457323e+00 -7.27906466e-01 -1.20853579e+00 -1.50254771e-01
2.05805197e-01 -9.03703213e-01 1.29282355e+00 -1.98592216e-01
-1.37088788e+00 1.40757978e+00 -1.29696310e-01 7.00869889e-04
4.59899843e-01 -2.57280111e-01 4.11928147e-01 5.49053401e-02
-4.60599437e-02 9.16549325e-01 8.66234958e-01 -1.60689926e+00
-1.88305616e-01 -1.07481301e+00 1.89660385e-01 3.54926646e-01
3.16876203e-01 -6.23847246e-01 -6.38283193e-01 -2.77166307e-01
9.98639464e-01 -7.04964280e-01 -2.90383965e-01 7.41722882e-01
-2.97222525e-01 -3.45027447e-01 6.84347928e-01 -4.05762553e-01
1.30537108e-01 -2.01965141e+00 3.81195754e-01 1.26463085e-01
5.82957804e-01 -2.26671249e-01 7.78349265e-02 3.45837101e-02
1.54236838e-01 5.87955639e-02 -3.30741256e-01 -6.45975053e-01
2.24568456e-01 5.20375371e-01 -3.61917883e-01 4.81110454e-01
3.73272330e-01 1.15793061e+00 -8.22793424e-01 -1.33499876e-01
6.13093197e-01 7.37357378e-01 -6.58600569e-01 4.37513530e-01
-6.12134695e-01 3.99322271e-01 -5.57864606e-01 1.19207621e+00
7.15281546e-01 -4.19523954e-01 -2.17341781e-01 -1.50729626e-01
-1.87062621e-02 3.01176965e-01 -8.56425703e-01 2.36557651e+00
-3.34479213e-01 3.06197643e-01 1.07823133e-01 -8.59528661e-01
1.26578021e+00 9.12916213e-02 6.38392448e-01 -6.54990971e-01
2.15009734e-01 2.62991130e-01 -2.73525596e-01 -3.04780722e-01
1.77772939e-01 -2.86752552e-01 -1.34515479e-01 4.65238005e-01
8.36626217e-02 -1.11074567e+00 -5.68758905e-01 -9.74802151e-02
1.09690416e+00 6.19187772e-01 3.82284880e-01 1.54523179e-01
-4.12045941e-02 1.16360020e-02 2.56535113e-01 3.98389757e-01
-1.06315926e-01 9.65785921e-01 3.44073325e-01 -8.08006465e-01
-1.21321285e+00 -1.66848838e+00 -1.99808389e-01 4.01198387e-01
1.85764551e-01 7.51456320e-02 -1.27151936e-01 -6.24820888e-01
5.94936132e-01 3.18550766e-01 -6.35763109e-01 3.37480754e-02
-5.38808167e-01 -7.13187754e-02 -3.95450816e-02 7.21127570e-01
3.62443745e-01 -8.67416024e-01 -8.56468141e-01 6.02040105e-02
3.36860418e-01 -1.01986325e+00 -6.17315397e-02 3.60242814e-01
-1.43727076e+00 -1.14142883e+00 -6.19641900e-01 -7.50704169e-01
7.86465108e-01 3.94703895e-01 1.37726080e+00 -2.04880849e-01
-5.11318862e-01 8.18406582e-01 -2.43631452e-01 -6.44093752e-01
-1.27832696e-01 -1.27871022e-01 -7.64428452e-02 -2.56022751e-01
4.71048564e-01 -9.62972462e-01 -6.07382596e-01 6.95755631e-02
-7.30176210e-01 3.48550439e-01 7.23063886e-01 4.82196689e-01
1.22560477e+00 -3.67233694e-01 -3.33535746e-02 -4.17144060e-01
-2.32827794e-02 -2.29699701e-01 -6.87954843e-01 -1.12226538e-01
-7.05500767e-02 2.14469969e-01 2.26861537e-01 -2.55861789e-01
-7.04818964e-01 4.43770319e-01 -3.02543759e-01 -1.02222931e+00
-6.52118146e-01 2.09855258e-01 -4.60949212e-01 1.88573986e-01
3.94388884e-01 2.63545632e-01 3.65523040e-01 -8.57868969e-01
4.65071559e-01 4.98128802e-01 4.81617659e-01 -6.03563547e-01
8.19820821e-01 9.50592995e-01 1.72432616e-01 -8.23863208e-01
-7.90459991e-01 -4.35879976e-01 -1.08984840e+00 -2.13684589e-01
9.47201371e-01 -1.09539127e+00 -8.44854772e-01 5.53493798e-01
-1.34190714e+00 -5.76243460e-01 -3.50906491e-01 3.16314876e-01
-8.71340394e-01 4.02395874e-02 -3.42024446e-01 -7.62732327e-01
-1.93092763e-01 -1.11887050e+00 1.72406828e+00 -3.55835482e-02
-1.91716269e-01 -7.67005742e-01 -1.24644917e-02 3.81247759e-01
3.23745161e-02 5.98665595e-01 1.01894009e+00 -1.84148014e-01
-1.20559347e+00 -3.43795836e-01 -3.38092566e-01 1.50599614e-01
2.39079744e-01 -3.28969330e-01 -1.04563868e+00 -2.63674539e-02
-2.20465567e-02 -5.67233205e-01 7.19852686e-01 4.73005503e-01
1.35619938e+00 2.13043720e-01 -2.13562578e-01 9.04706538e-01
1.34128404e+00 8.41385424e-02 3.87605548e-01 -9.19326693e-02
8.27871025e-01 4.69250202e-01 2.44730711e-01 5.04388869e-01
6.37619734e-01 4.70289409e-01 1.18457580e+00 1.38462692e-01
-2.63330400e-01 -5.31359673e-01 -2.57739574e-01 7.74603188e-01
-2.66938601e-02 1.86261818e-01 -1.17508006e+00 4.74761069e-01
-1.43775785e+00 -4.57958370e-01 1.29558593e-01 2.17959404e+00
6.34842634e-01 3.93581301e-01 -1.77466795e-01 1.03171699e-01
2.40729302e-01 8.05529803e-02 -1.14047360e+00 -4.26129699e-02
1.38065070e-01 3.45225781e-01 1.86511904e-01 3.95406902e-01
-9.37752783e-01 7.26703703e-01 5.27874899e+00 2.89093912e-01
-1.11352694e+00 -1.83147609e-01 4.47126001e-01 -1.03219002e-01
-5.56729913e-01 -3.19253206e-01 -6.36565924e-01 -5.36519056e-03
3.26681621e-02 1.56569257e-01 3.02483290e-01 1.17256427e+00
-1.65243432e-01 -5.23770228e-02 -1.57404304e+00 1.51746857e+00
2.42666572e-01 -1.29163492e+00 2.13115841e-01 4.02467936e-01
5.85274756e-01 3.54451030e-01 -6.80611506e-02 1.88994616e-01
4.19656605e-01 -1.16455972e+00 1.02818692e+00 4.94883627e-01
1.13768160e+00 -4.99610722e-01 2.62204736e-01 6.27655745e-01
-1.24087703e+00 2.32522205e-01 -4.90289330e-01 -2.26255909e-01
1.54145017e-01 6.71877563e-01 -9.01585221e-01 2.55041957e-01
6.24276578e-01 8.58898342e-01 -1.68708503e-01 1.07481480e+00
-2.05719560e-01 8.95251334e-02 -5.59289694e-01 6.77098408e-02
-6.34884909e-02 -3.73122320e-02 4.10327196e-01 5.34430802e-01
3.88881952e-01 3.62761199e-01 3.05829644e-01 1.18055654e+00
-2.22895414e-01 -4.69776243e-01 -1.11087203e+00 -4.75459397e-02
4.71543372e-01 9.51431096e-01 -9.02215779e-01 -1.28656194e-01
-3.65751386e-01 9.03082907e-01 4.27230626e-01 4.19428758e-02
-3.10670942e-01 7.43490532e-02 1.01780879e+00 1.42505616e-01
4.49131697e-01 -8.40981126e-01 -8.35990846e-01 -1.10051501e+00
8.64787251e-02 -4.02955621e-01 -1.38877064e-01 -1.00528657e+00
-1.48816836e+00 2.70425767e-01 6.81247190e-03 -1.34406698e+00
-1.07684895e-01 -1.07191670e+00 -1.44493401e-01 8.34707618e-01
-1.31185782e+00 -1.23092616e+00 -5.03579855e-01 4.41187114e-01
7.36943662e-01 3.01037333e-03 9.45484221e-01 -1.30806655e-01
1.15258031e-01 2.26448011e-02 -3.75469297e-01 2.59573162e-01
1.80929497e-01 -1.22358787e+00 9.55881417e-01 2.52375036e-01
3.79088551e-01 4.54601556e-01 2.74210364e-01 -7.47524321e-01
-2.01993227e+00 -9.16215718e-01 5.08172810e-01 -9.83739018e-01
2.17675462e-01 -6.79414272e-01 -7.26725161e-01 6.89439476e-01
-6.15028560e-01 2.85993755e-01 4.40929443e-01 -9.50912535e-02
-6.88044369e-01 1.36248887e-01 -1.15294790e+00 4.71009195e-01
1.73332798e+00 -7.77259469e-01 -8.91492724e-01 7.31005296e-02
7.44860947e-01 -8.66925299e-01 -8.85962307e-01 5.15690923e-01
8.35914552e-01 -9.84755397e-01 1.32711101e+00 -3.89453560e-01
5.85008621e-01 -1.40478909e-01 -6.68682277e-01 -1.16812611e+00
-2.43818849e-01 1.31638255e-02 -5.16258478e-01 4.95037526e-01
-5.99944964e-02 -3.76429588e-01 1.33051872e+00 8.83743644e-01
-4.23091680e-01 -1.04889524e+00 -1.13301671e+00 -6.47541344e-01
3.45815644e-02 -8.89669359e-01 7.40700901e-01 6.25701308e-01
-6.06734753e-01 1.92348361e-01 2.82213062e-01 1.64003357e-01
1.03421402e+00 6.07113242e-01 1.01206315e+00 -1.79452705e+00
-2.48818234e-01 -6.48056269e-01 -8.29725146e-01 -1.71751285e+00
1.33142546e-01 -9.83789384e-01 5.42457402e-02 -1.95363414e+00
9.88268480e-02 -4.77084488e-01 1.86498865e-01 5.61848342e-01
3.63801211e-01 3.74381483e-01 1.97944283e-01 1.29554242e-01
-3.92104745e-01 8.44015896e-01 1.57821989e+00 -2.54557163e-01
-1.09933384e-01 -6.01408742e-02 -4.80449408e-01 8.99310946e-01
4.91428822e-01 -1.66618913e-01 -3.23980510e-01 -9.55120742e-01
2.84013689e-01 9.21972021e-02 5.67524254e-01 -6.17530823e-01
-5.99093735e-02 -1.04726814e-01 7.88782835e-01 -1.27266502e+00
8.79731894e-01 -1.06690311e+00 -1.08288385e-01 3.18012059e-01
7.32194819e-03 -1.87824354e-01 1.51796535e-01 6.23211384e-01
9.14529338e-02 2.72106826e-01 5.00691831e-01 -5.85219383e-01
-7.85732388e-01 8.50155532e-01 1.37213811e-01 -5.53576499e-02
6.30788743e-01 -6.69646263e-01 9.59166884e-02 -9.15442482e-02
-6.60491109e-01 9.12921131e-02 8.35231006e-01 6.15003526e-01
1.21893692e+00 -1.59479511e+00 -4.65149373e-01 7.21799016e-01
2.86246777e-01 1.00464690e+00 1.95781052e-01 3.03474128e-01
-5.11567473e-01 3.18513542e-01 -3.29509616e-01 -1.23648500e+00
-8.66748393e-01 1.77753612e-01 2.75385201e-01 1.07012711e-01
-1.03497052e+00 1.05630159e+00 5.81682622e-01 -8.78963292e-01
4.26774144e-01 -7.53302336e-01 2.13723317e-01 -2.89374709e-01
3.11706126e-01 -1.91641115e-02 1.92031413e-01 -4.83585000e-01
-3.02608460e-01 9.41189289e-01 2.55725265e-01 -1.90160319e-01
1.69745326e+00 5.73780313e-02 -2.18835920e-02 6.80190444e-01
1.26772916e+00 -4.21581835e-01 -1.83375657e+00 -3.36068630e-01
-2.49801993e-01 -7.23015189e-01 6.42581610e-03 -5.85473895e-01
-9.18397427e-01 1.16157067e+00 3.83369297e-01 -2.31967960e-02
6.37286067e-01 4.82592493e-01 5.76908350e-01 5.06085515e-01
9.05199587e-01 -4.71825570e-01 3.37640107e-01 8.09549809e-01
1.33324468e+00 -1.38665175e+00 3.54409814e-02 -3.79297167e-01
-2.35613123e-01 1.18792665e+00 5.58049023e-01 -3.73141915e-01
9.07913446e-01 2.43141413e-01 -5.89187257e-03 -6.02132320e-01
-5.70954442e-01 -1.93463653e-01 4.19257253e-01 8.87352228e-01
2.22660065e-01 1.98734462e-01 6.15650058e-01 4.44198817e-01
-2.65541703e-01 -9.74317417e-02 1.56217426e-01 8.93233597e-01
-4.59475130e-01 -9.00503874e-01 -4.18577939e-01 4.99524415e-01
8.94216523e-02 4.41559374e-01 -4.98838395e-01 6.85023546e-01
7.08286539e-02 3.22646022e-01 4.59415942e-01 -2.61225075e-01
5.24503708e-01 2.10573241e-01 9.56182718e-01 -1.02040708e+00
1.39735192e-01 5.70173673e-02 -3.38458329e-01 -8.12242746e-01
-2.43286490e-01 -5.95917225e-01 -1.32787311e+00 -1.46121413e-01
2.16664329e-01 -6.93037987e-01 1.02207601e+00 8.02736104e-01
3.47676039e-01 1.74719647e-01 4.48820829e-01 -1.76859605e+00
-5.28404415e-01 -6.76547468e-01 -6.08412206e-01 4.86310095e-01
3.61675769e-01 -1.04196215e+00 -1.19428918e-01 -5.49638234e-02] | [8.31525993347168, -3.276215076446533] |
d4d71e2a-e4fa-4918-b885-b9eba3d714b9 | deepigeos-a-deep-interactive-geodesic | 1707.00652 | null | http://arxiv.org/abs/1707.00652v3 | http://arxiv.org/pdf/1707.00652v3.pdf | DeepIGeoS: A Deep Interactive Geodesic Framework for Medical Image Segmentation | Accurate medical image segmentation is essential for diagnosis, surgical
planning and many other applications. Convolutional Neural Networks (CNNs) have
become the state-of-the-art automatic segmentation methods. However, fully
automatic results may still need to be refined to become accurate and robust
enough for clinical use. We propose a deep learning-based interactive
segmentation method to improve the results obtained by an automatic CNN and to
reduce user interactions during refinement for higher accuracy. We use one CNN
to obtain an initial automatic segmentation, on which user interactions are
added to indicate mis-segmentations. Another CNN takes as input the user
interactions with the initial segmentation and gives a refined result. We
propose to combine user interactions with CNNs through geodesic distance
transforms, and propose a resolution-preserving network that gives a better
dense prediction. In addition, we integrate user interactions as hard
constraints into a back-propagatable Conditional Random Field. We validated the
proposed framework in the context of 2D placenta segmentation from fetal MRI
and 3D brain tumor segmentation from FLAIR images. Experimental results show
our method achieves a large improvement from automatic CNNs, and obtains
comparable and even higher accuracy with fewer user interventions and less time
compared with traditional interactive methods. | ['Jan Deprest', 'Tom Vercauteren', 'Sebastien Ourselin', 'Michael Aertsen', 'Maria A. Zuluaga', 'Guotai Wang', 'Wenqi Li', 'Rosalind Pratt', 'Anna L. David', 'Tom Doel', 'Premal A. Patel'] | 2017-07-03 | null | null | null | null | ['placenta-segmentation'] | ['medical'] | [ 2.59513795e-01 5.67159891e-01 9.23480541e-02 -8.14367890e-01
-6.29099727e-01 -3.14582348e-01 1.35266259e-01 3.47503990e-01
-7.02683449e-01 5.71547270e-01 -8.02302361e-02 -4.64379400e-01
-5.58262281e-02 -9.56524968e-01 -7.43303478e-01 -5.20799994e-01
2.97651570e-02 8.36215317e-01 7.26259351e-01 1.82669461e-02
7.11055025e-02 6.79193020e-01 -1.02096283e+00 3.43681157e-01
1.38733065e+00 8.88029516e-01 3.02976519e-01 4.99989718e-01
-2.61002421e-01 3.61414969e-01 -2.33407557e-01 -3.23552489e-01
2.20281988e-01 -4.67099309e-01 -1.12668431e+00 6.20834157e-02
-2.22117044e-02 -3.25556397e-01 1.59692496e-01 1.16770184e+00
4.99196231e-01 7.97973350e-02 5.61982572e-01 -6.46118045e-01
-4.01832342e-01 7.93997765e-01 -6.29997730e-01 2.14766953e-02
-2.37186681e-02 -6.44486547e-02 3.06297243e-01 -5.60157120e-01
6.68561339e-01 9.74459767e-01 7.54954994e-01 6.34557426e-01
-1.00848269e+00 -6.46755338e-01 -5.58740906e-02 -1.53092081e-02
-1.25051892e+00 -1.20298550e-01 4.81016159e-01 -4.72753227e-01
4.78768617e-01 3.62651557e-01 8.84216607e-01 4.45519239e-01
2.25669488e-01 7.22014487e-01 8.11084986e-01 -2.79526889e-01
1.18481494e-01 1.00094974e-01 1.02693059e-01 1.13353384e+00
5.91823757e-02 -2.47157872e-01 2.33445376e-01 3.79420184e-02
1.24203610e+00 3.88388261e-02 -4.36649531e-01 -3.41567844e-01
-1.19608390e+00 8.47437441e-01 9.11735594e-01 6.28256321e-01
-4.62185472e-01 -1.35325387e-01 1.65338144e-01 -2.23436952e-01
6.60707891e-01 4.17086720e-01 -3.35507095e-01 1.13388203e-01
-1.22538137e+00 3.16470973e-02 6.79618001e-01 6.34005845e-01
6.94914281e-01 -2.89540708e-01 -3.83767217e-01 7.95659006e-01
1.22306854e-01 9.23505723e-02 4.59525228e-01 -8.19001377e-01
3.64006758e-02 6.45097315e-01 -6.18298277e-02 -9.24325168e-01
-1.00668359e+00 -7.40717471e-01 -1.04479527e+00 2.61223227e-01
4.36924726e-01 -1.69937506e-01 -1.33016694e+00 1.26522231e+00
4.20340091e-01 3.29929084e-01 -2.44333297e-01 1.06074345e+00
1.01129198e+00 3.55107516e-01 -3.44278999e-02 -2.70904988e-01
1.15040445e+00 -1.06493461e+00 -8.07624936e-01 9.21990350e-02
1.03785145e+00 -5.70704579e-01 7.54539192e-01 4.82688755e-01
-1.31274080e+00 -3.18554908e-01 -7.66040921e-01 -1.12657100e-02
-9.65122059e-02 1.39113784e-01 4.87185985e-01 6.38368368e-01
-1.24622822e+00 8.69070828e-01 -1.48727787e+00 -2.21144497e-01
7.88411140e-01 7.39984989e-01 -2.51659930e-01 1.13072067e-01
-9.13650692e-01 1.05008936e+00 4.79729384e-01 3.33562702e-01
-4.65107024e-01 -8.55343163e-01 -9.91475344e-01 -9.51428488e-02
4.15727973e-01 -5.96100867e-01 1.13596511e+00 -1.01124442e+00
-1.55525148e+00 8.53353143e-01 -1.89208717e-03 -4.63729471e-01
8.02041352e-01 -4.72154170e-02 -8.61681625e-02 2.04524159e-01
2.19283327e-02 8.62325847e-01 2.92330831e-01 -1.17690361e+00
-3.65332544e-01 -3.37365627e-01 -2.87507791e-02 2.50607252e-01
-1.29380494e-01 -4.77069765e-02 -8.79662216e-01 -4.30813015e-01
5.35017192e-01 -9.83069241e-01 -6.61453307e-01 1.47928059e-01
-5.93777239e-01 1.03207581e-01 6.47526026e-01 -1.00817394e+00
1.00271964e+00 -1.70020163e+00 2.94469118e-01 4.92219210e-01
4.83028740e-01 3.57889146e-01 2.16309771e-01 -5.11072993e-01
-4.36917692e-02 2.20186189e-01 -8.17574203e-01 -4.05282825e-01
-6.32971346e-01 1.59320682e-01 6.82825267e-01 5.73981822e-01
6.22780100e-02 9.64400411e-01 -8.51911366e-01 -7.37914681e-01
3.86282831e-01 6.64626122e-01 -8.56048226e-01 3.25198531e-01
-1.60393342e-02 1.15450883e+00 -5.62860310e-01 3.20701063e-01
7.74464250e-01 -3.48186761e-01 3.01990509e-02 -3.88427436e-01
-2.69168884e-01 -2.43260562e-01 -8.57595921e-01 2.04407525e+00
-4.17135149e-01 3.99796486e-01 2.14034617e-01 -1.26976752e+00
8.32116365e-01 2.78440446e-01 6.24167264e-01 -4.62541074e-01
6.02124274e-01 3.70755613e-01 2.43274525e-01 -7.25862026e-01
1.02738231e-01 -1.15278766e-01 3.76440972e-01 2.16295615e-01
2.36875135e-02 -3.80635560e-01 -1.64971016e-02 5.78116365e-02
8.17285895e-01 2.32994139e-01 2.33583987e-01 -5.30630767e-01
7.18594074e-01 -1.36700138e-01 5.60649037e-01 4.56576437e-01
5.99671118e-02 1.11854088e+00 4.79568243e-01 -6.22646093e-01
-7.42236614e-01 -4.90694851e-01 -3.65019590e-01 6.25267208e-01
2.57880867e-01 1.45235941e-01 -1.36469066e+00 -1.04847527e+00
-3.81941348e-01 6.69694722e-01 -8.70716393e-01 1.91969708e-01
-7.48927057e-01 -8.84709775e-01 3.14777434e-01 5.62596977e-01
5.06029785e-01 -1.15642953e+00 -6.40268803e-01 3.33333433e-01
-3.24811608e-01 -1.00903821e+00 -3.88685673e-01 3.79330814e-02
-9.98969376e-01 -9.30809200e-01 -1.26727164e+00 -7.93683231e-01
1.09632516e+00 -4.49633896e-01 9.00551319e-01 4.86304700e-01
-4.03506428e-01 -2.06424743e-01 -3.31215262e-01 -2.35568240e-01
-5.85729301e-01 3.07106525e-01 -2.62989372e-01 8.18544179e-02
-3.61998498e-01 -4.96836305e-01 -7.72935212e-01 2.53236383e-01
-9.62153792e-01 5.48158050e-01 5.80788255e-01 5.97394586e-01
7.10636854e-01 -3.32659125e-01 1.44912496e-01 -1.45865035e+00
3.22638482e-01 -3.23482484e-01 -5.70836127e-01 1.94680721e-01
-4.70614344e-01 1.50352120e-01 4.54864711e-01 -2.89996803e-01
-1.16717565e+00 2.46773392e-01 -7.27674484e-01 -3.30676585e-01
-3.58621210e-01 6.13068879e-01 2.72305518e-01 -3.70933682e-01
7.49888301e-01 -8.26281905e-02 4.96286750e-02 -4.85670149e-01
1.33028165e-01 3.23547423e-01 5.62476814e-01 -2.50951409e-01
4.18235481e-01 4.76181030e-01 -5.73280379e-02 -5.23796260e-01
-8.40224504e-01 -2.46972114e-01 -1.16095841e+00 -2.97328472e-01
1.48170412e+00 -2.87617207e-01 -4.19701725e-01 2.79721260e-01
-1.34617984e+00 -4.93899465e-01 -5.48289716e-03 5.58202326e-01
-4.30785537e-01 3.38094622e-01 -6.42715514e-01 -4.11298096e-01
-4.37794030e-01 -1.75188732e+00 8.88666689e-01 5.41915953e-01
2.27884240e-02 -1.17882836e+00 -1.69485822e-01 2.76039660e-01
4.82026219e-01 5.63349664e-01 8.16510320e-01 -8.31115842e-01
-4.11492199e-01 -1.29454568e-01 -4.57254112e-01 1.35973215e-01
2.10340798e-01 -5.68477809e-02 -7.34892428e-01 -7.27828741e-02
-1.26616135e-01 2.05127031e-01 7.51115739e-01 8.62875819e-01
1.70341444e+00 -7.99878314e-02 -6.41240060e-01 1.03923714e+00
1.26813376e+00 4.00050431e-01 6.51766598e-01 5.56209683e-02
9.70001280e-01 5.61249554e-01 3.53329241e-01 1.63856670e-01
2.89232552e-01 3.52164805e-01 6.13652527e-01 -8.05789769e-01
-2.26362199e-02 3.06247503e-01 -4.77758318e-01 1.00386715e+00
-4.90439266e-01 -8.03094059e-02 -1.33417070e+00 5.09064138e-01
-1.91420281e+00 -3.26531768e-01 -4.30671275e-01 2.05795264e+00
8.74655068e-01 1.81521818e-01 -1.28255785e-01 -7.00744987e-02
6.91547692e-01 -3.97685587e-01 -4.05342937e-01 -2.30530038e-01
4.40629870e-01 5.16741872e-01 5.52479446e-01 7.60451317e-01
-1.08613956e+00 8.94977212e-01 5.41991043e+00 7.70145118e-01
-1.31837332e+00 4.54192370e-01 1.08204615e+00 3.41719538e-01
-1.88337564e-01 -3.73000205e-01 -3.13647747e-01 1.49691820e-01
4.84526575e-01 2.48795092e-01 7.75443092e-02 5.74591041e-01
2.93135881e-01 -2.79975891e-01 -1.09238446e+00 7.81932294e-01
-2.48564966e-02 -1.62810683e+00 -6.94806427e-02 -8.77830982e-02
8.82115662e-01 -1.24942608e-01 -2.67369598e-01 -1.20259486e-02
1.32817373e-01 -1.20399404e+00 4.10252690e-01 7.70576715e-01
7.98962176e-01 -8.73956263e-01 1.15530884e+00 5.89963317e-01
-9.64400053e-01 3.61341685e-01 -5.42382002e-02 3.07636678e-01
2.30002791e-01 6.04654670e-01 -1.04178798e+00 6.62809968e-01
6.23555303e-01 4.39827472e-01 -3.82589698e-01 1.37027359e+00
1.01394868e-02 4.35252279e-01 -2.17844471e-01 5.35032339e-02
3.24016124e-01 -2.91156948e-01 4.55564201e-01 1.33855557e+00
2.58811444e-01 4.98946041e-01 2.95939922e-01 9.99226034e-01
4.00428456e-04 5.15580177e-01 -2.36001924e-01 4.44622785e-01
-1.31806880e-01 1.50847971e+00 -1.52442002e+00 -4.77459937e-01
-1.41258776e-01 1.10242283e+00 3.36717129e-01 1.92137957e-01
-9.56255674e-01 -3.88672799e-01 4.84996289e-02 3.44016224e-01
-1.20071240e-01 1.33221755e-02 -5.88504791e-01 -9.69186485e-01
-3.84030104e-01 -3.74394834e-01 2.02733353e-01 -5.62737703e-01
-7.44294226e-01 9.23948765e-01 -7.33924955e-02 -8.80279541e-01
-2.31459737e-02 -3.34935308e-01 -6.74075127e-01 8.91788125e-01
-1.35741770e+00 -8.90007734e-01 -5.94425559e-01 4.35261577e-01
6.26186490e-01 3.50462437e-01 8.02569985e-01 4.57791090e-01
-4.07348454e-01 5.15474796e-01 -3.27720106e-01 3.11791718e-01
4.06084061e-01 -1.24273300e+00 6.22972362e-02 7.43258417e-01
-3.18170071e-01 3.00131887e-01 4.35450643e-01 -8.32298696e-01
-5.86922407e-01 -1.27743065e+00 4.53409016e-01 -7.00684916e-03
9.85166952e-02 -1.82254184e-02 -1.07192934e+00 6.51778460e-01
1.75048277e-01 2.48267606e-01 3.57735187e-01 -3.15908164e-01
5.24221361e-01 1.93679079e-01 -1.46660972e+00 5.12313008e-01
8.79958987e-01 1.18090905e-01 -2.69522071e-01 4.98183578e-01
8.43721092e-01 -1.11053967e+00 -7.84131706e-01 8.09553623e-01
3.87374133e-01 -1.04156339e+00 6.97917998e-01 -1.61361516e-01
4.12926704e-01 -9.23238173e-02 6.35217130e-01 -1.27498984e+00
-2.58479893e-01 -4.80745941e-01 4.86615628e-01 5.67744851e-01
7.37575650e-01 -4.51227397e-01 8.25388789e-01 9.12139416e-01
-4.89338100e-01 -1.08722818e+00 -6.85434401e-01 -2.26353392e-01
1.37085140e-01 -5.26224911e-01 4.91438568e-01 1.13195467e+00
-1.96826130e-01 -3.12910855e-01 -3.17634866e-02 2.94785678e-01
5.70957899e-01 -1.67634845e-01 2.87682384e-01 -1.36230373e+00
3.29939276e-02 -5.21584392e-01 -2.65502095e-01 -9.28394735e-01
-9.52533484e-02 -1.08799601e+00 3.07227433e-01 -1.94100511e+00
2.71442652e-01 -7.54651546e-01 -1.08691603e-01 6.80082619e-01
-1.58420339e-01 6.57472074e-01 -1.40980110e-01 -1.11966796e-01
-4.08514827e-01 2.80946285e-01 1.78773916e+00 1.06355563e-01
-5.14897585e-01 2.13333294e-01 -3.82128000e-01 1.15179479e+00
6.93870842e-01 -4.23458368e-01 -6.96652830e-02 -5.62782168e-01
-1.71603292e-01 4.46292013e-01 1.53852805e-01 -9.93307650e-01
3.68401766e-01 6.68950975e-02 5.39837956e-01 -4.72671837e-01
-7.35690584e-03 -8.04931223e-01 7.48100830e-03 6.60876691e-01
-4.02040035e-01 -3.46552670e-01 2.48981595e-01 2.14378424e-02
-4.27598022e-02 -3.24845105e-01 1.04495263e+00 -3.44087958e-01
-3.53433676e-02 6.69958830e-01 -3.64500940e-01 -1.65378869e-01
9.90042388e-01 -2.39204839e-01 5.13061106e-01 -3.02381843e-01
-1.57357085e+00 1.88626662e-01 2.28141040e-01 4.56549451e-02
6.65242314e-01 -9.71148849e-01 -7.67981172e-01 2.12567851e-01
-4.69244957e-01 6.25633299e-01 2.97817320e-01 1.28881586e+00
-1.09359086e+00 1.79834485e-01 -1.77361667e-01 -9.22039866e-01
-1.31923258e+00 7.24332482e-02 6.60448670e-01 -4.04026151e-01
-7.35519588e-01 1.09571552e+00 4.27284151e-01 -7.36270368e-01
1.96390718e-01 -7.97342360e-01 -5.99060655e-01 -2.09065899e-01
3.14968526e-01 1.73872575e-01 3.76665473e-01 -6.17572963e-01
-3.12684983e-01 6.14185989e-01 -6.45546988e-03 -6.91471100e-02
1.43176174e+00 7.31340796e-02 -2.06650510e-01 8.91259965e-03
1.06338692e+00 -2.32463941e-01 -1.03912985e+00 -1.35479122e-01
-1.75210282e-01 -2.84030437e-01 3.01495820e-01 -8.98276031e-01
-1.65807557e+00 9.71314847e-01 7.39104211e-01 2.11980000e-01
1.11881161e+00 -2.06994694e-02 8.27741444e-01 -6.94558546e-02
2.42509991e-01 -7.05621541e-01 -2.75764704e-01 3.13255370e-01
8.62559021e-01 -1.23743165e+00 -1.65661484e-01 -7.51337945e-01
-5.85111380e-01 1.41333079e+00 7.52263367e-01 -9.55638438e-02
8.96902621e-01 5.06481290e-01 1.04403935e-01 -1.78629056e-01
9.15475115e-02 -1.52111933e-01 6.59475625e-01 3.67619723e-01
6.19063914e-01 1.41366601e-01 -4.89563167e-01 7.13986337e-01
-1.01856396e-01 1.13836594e-01 3.84014040e-01 6.86810255e-01
-3.88615072e-01 -9.87313926e-01 -2.55661368e-01 6.13929391e-01
-7.20502794e-01 -1.24398753e-01 1.20484814e-01 6.10220432e-01
4.63179827e-01 4.39634562e-01 5.88727705e-02 -1.27153486e-01
2.24523813e-01 -6.91222176e-02 4.66994673e-01 -8.95225823e-01
-8.34212422e-01 2.42767438e-01 -1.27254248e-01 -6.95391297e-01
-3.24502021e-01 -4.90537494e-01 -1.78132272e+00 1.55676484e-01
-5.62419057e-01 1.17659986e-01 8.28832448e-01 1.23362160e+00
6.60929382e-02 9.25708473e-01 3.25714201e-01 -1.13052225e+00
2.14164004e-01 -9.30561602e-01 -1.59898371e-01 2.48859644e-01
1.83746353e-01 -4.64016706e-01 1.79442652e-02 2.38997892e-01] | [14.506421089172363, -2.6016182899475098] |
69d1d23e-34b5-46ce-af7a-59dde353c2a7 | deep-learning-of-high-order-interactions-for | 2007.09334 | null | https://arxiv.org/abs/2007.09334v1 | https://arxiv.org/pdf/2007.09334v1.pdf | Deep Learning of High-Order Interactions for Protein Interface Prediction | Protein interactions are important in a broad range of biological processes. Traditionally, computational methods have been developed to automatically predict protein interface from hand-crafted features. Recent approaches employ deep neural networks and predict the interaction of each amino acid pair independently. However, these methods do not incorporate the important sequential information from amino acid chains and the high-order pairwise interactions. Intuitively, the prediction of an amino acid pair should depend on both their features and the information of other amino acid pairs. In this work, we propose to formulate the protein interface prediction as a 2D dense prediction problem. In addition, we propose a novel deep model to incorporate the sequential information and high-order pairwise interactions to perform interface predictions. We represent proteins as graphs and employ graph neural networks to learn node features. Then we propose the sequential modeling method to incorporate the sequential information and reorder the feature matrix. Next, we incorporate high-order pairwise interactions to generate a 3D tensor containing different pairwise interactions. Finally, we employ convolutional neural networks to perform 2D dense predictions. Experimental results on multiple benchmarks demonstrate that our proposed method can consistently improve the protein interface prediction performance. | ['Yi Liu', 'Hao Yuan', 'Lei Cai', 'Shuiwang Ji'] | 2020-07-18 | null | null | null | null | ['protein-interface-prediction'] | ['miscellaneous'] | [ 1.02342516e-01 -2.28934377e-01 -1.66078851e-01 -4.89609361e-01
-1.86093107e-01 -1.83552116e-01 1.34387031e-01 3.41276348e-01
-1.75234184e-01 7.67794847e-01 2.11340934e-01 -3.96393895e-01
-1.52426036e-02 -7.43705571e-01 -9.85060513e-01 -7.86483943e-01
-3.70136827e-01 6.57498419e-01 1.69783115e-01 -1.12542026e-01
1.57632783e-01 5.64400136e-01 -8.51279855e-01 4.23868805e-01
1.10926783e+00 9.10323322e-01 3.46855193e-01 3.39634180e-01
-2.16945127e-01 6.68614984e-01 -1.25389233e-01 -1.83374345e-01
6.87838569e-02 -3.55941772e-01 -6.84814632e-01 -2.84403022e-02
8.09940323e-03 -2.84819752e-01 -3.20264757e-01 8.74145031e-01
3.12089741e-01 4.23943475e-02 6.48275554e-01 -1.05117810e+00
-7.99258411e-01 2.56728292e-01 -7.22043633e-01 -2.57071741e-02
4.27397102e-01 2.63467461e-01 1.37764680e+00 -1.10527349e+00
6.81306541e-01 1.22436678e+00 6.51613176e-01 2.40690817e-04
-1.35510576e+00 -5.36313891e-01 4.01653826e-01 4.22526330e-01
-1.01145089e+00 2.10213006e-01 9.50213253e-01 -6.06265903e-01
1.30585694e+00 -1.03497654e-01 7.10428298e-01 8.08492243e-01
4.65529770e-01 6.17034912e-01 6.59913838e-01 1.22195724e-02
1.77645553e-02 -5.93571007e-01 3.68386477e-01 8.50074291e-01
-3.38659296e-03 -3.20421576e-01 -3.53913218e-01 -6.63571835e-01
4.82211232e-01 6.76893175e-01 -3.84689987e-01 -5.74982643e-01
-1.18795788e+00 9.79725122e-01 7.38032043e-01 -4.31303196e-02
-6.17633462e-01 -1.18287906e-01 3.30336809e-01 1.14834063e-01
4.92651731e-01 2.03968510e-01 -7.48508215e-01 1.61303312e-01
-2.67059624e-01 3.27285379e-01 9.18608010e-01 6.50092304e-01
1.28960359e+00 -6.75416827e-01 9.24658254e-02 9.36740518e-01
5.42784333e-01 8.58123302e-02 2.71610498e-01 -6.34835303e-01
3.44411969e-01 9.71493840e-01 2.17535987e-01 -1.28910434e+00
-4.43377495e-01 -5.03691956e-02 -1.11548483e+00 -9.19683576e-02
4.93409447e-02 -1.57921202e-02 -8.03991556e-01 1.34524322e+00
4.66776639e-01 1.87787041e-01 -1.40325159e-01 1.01360810e+00
6.39208198e-01 8.76650989e-01 -4.45063040e-02 -1.69488162e-01
1.10353661e+00 -1.01206648e+00 -5.27762055e-01 1.97490826e-01
8.06103468e-01 -6.87729895e-01 6.59641445e-01 1.00441873e-01
-8.81590605e-01 -4.92694646e-01 -9.57371473e-01 -2.69080460e-01
-1.26267985e-01 8.78035724e-02 8.21540892e-01 -3.02240461e-01
-6.82069302e-01 9.62633371e-01 -9.56381202e-01 -9.68032423e-03
2.30494410e-01 7.82984316e-01 -5.76088667e-01 -3.83567773e-02
-1.33952439e+00 5.98460913e-01 3.01485598e-01 2.09270492e-01
-2.51047671e-01 -5.42688251e-01 -9.10505533e-01 1.78470075e-01
1.11877821e-01 -7.20580876e-01 8.40281546e-01 -8.75074327e-01
-1.32648683e+00 2.88566798e-01 -5.38670897e-01 -4.23402011e-01
-7.94684663e-02 -2.55395740e-01 2.20694337e-02 1.72755137e-01
-5.43496348e-02 5.26915729e-01 1.59011215e-01 -9.52027142e-01
-3.20723087e-01 -4.65501875e-01 2.07210883e-01 3.91789645e-01
-5.23241274e-02 -2.13263869e-01 -2.51299232e-01 -4.08741951e-01
2.42856860e-01 -1.07888460e+00 -6.28043354e-01 9.79026332e-02
-5.41678905e-01 -3.70450944e-01 5.52661836e-01 -6.06178820e-01
8.29449475e-01 -1.89945877e+00 5.77669978e-01 4.05710876e-01
8.10874343e-01 2.85232842e-01 -2.50650048e-01 7.35772610e-01
-1.91449910e-01 -3.42730992e-02 -1.54431820e-01 -1.39203042e-01
-8.31847116e-02 2.22435430e-01 -4.29148041e-02 4.07910794e-01
4.43643451e-01 8.48525226e-01 -7.73497403e-01 -2.30211005e-01
1.06311932e-01 8.07350814e-01 -8.21214676e-01 5.37246287e-01
-3.90962273e-01 3.82097930e-01 -7.75473118e-01 1.72834054e-01
9.61956978e-01 -8.63638818e-01 6.56710446e-01 -6.78655565e-01
-2.74663921e-02 5.90349257e-01 -5.56845427e-01 1.53420103e+00
1.83873181e-03 9.27957222e-02 -1.14089750e-01 -1.13802671e+00
1.01654553e+00 7.10014403e-02 8.93780887e-01 -2.56320596e-01
-5.01137562e-02 1.98503494e-01 4.37485933e-01 -3.14093947e-01
-1.58828013e-02 8.35914686e-02 4.18451667e-01 3.45123917e-01
-6.73571676e-02 5.85848629e-01 -3.92647833e-02 3.16486359e-01
1.24842393e+00 1.00226693e-01 2.33657405e-01 -2.59935632e-02
6.06444895e-01 2.39028316e-03 9.73481536e-01 2.00606897e-01
1.92559257e-01 4.45597172e-01 7.53044248e-01 -9.97887254e-01
-1.16354215e+00 -7.02303946e-01 2.28018209e-01 9.00590301e-01
3.71079296e-01 -6.03850722e-01 -5.46311796e-01 -7.60072231e-01
1.37669683e-01 -2.32302070e-01 -5.21613240e-01 -2.61579394e-01
-6.13697350e-01 -8.71192455e-01 -1.89986944e-01 4.58039433e-01
1.68458700e-01 -9.76262689e-01 9.19062421e-02 5.40226758e-01
-1.86878502e-01 -7.83536971e-01 -8.04965854e-01 4.94194508e-01
-7.15113521e-01 -1.22164440e+00 -6.25277996e-01 -1.09651852e+00
7.42570817e-01 4.81599361e-01 8.79319906e-01 3.86821985e-01
-1.66626140e-01 -4.19153750e-01 -3.43394667e-01 2.01446176e-01
5.95827885e-02 7.43838623e-02 4.60079871e-02 1.36818632e-01
7.19480991e-01 -8.59685719e-01 -9.94574666e-01 4.11254078e-01
-6.83539033e-01 3.42449367e-01 8.06689322e-01 1.19511664e+00
8.55868936e-01 -3.56329888e-01 2.30895013e-01 -1.07143843e+00
4.96721953e-01 -6.37591541e-01 -3.54855806e-01 1.85753584e-01
-2.59998292e-01 5.29597938e-01 8.62717032e-01 -3.39112133e-01
-7.49241412e-01 4.97065127e-01 -2.90563673e-01 -4.36799496e-01
-1.30507186e-01 8.97050142e-01 -3.44791383e-01 -2.90767014e-01
-4.57167104e-02 1.95359752e-01 1.09675087e-01 -7.94064462e-01
-8.53047147e-02 4.07326221e-01 -1.13086119e-01 -7.12117612e-01
4.55895603e-01 2.37283528e-01 1.05302799e-02 -5.93250155e-01
-7.12328434e-01 -4.14310098e-01 -7.98167229e-01 2.77198583e-01
9.57460701e-01 -7.29096472e-01 -1.17686391e+00 5.42178869e-01
-1.52746820e+00 -1.94497392e-01 5.60478330e-01 5.53794801e-01
-3.32889378e-01 8.73960793e-01 -8.81007791e-01 -2.00625986e-01
-3.50352436e-01 -1.59515774e+00 1.12533331e+00 -4.11607176e-02
8.75064917e-03 -8.71409953e-01 5.06310225e-01 2.93621123e-01
-9.38526168e-02 1.30666196e-01 1.22623026e+00 -9.25503910e-01
-6.96069360e-01 -5.22393791e-04 -5.11033356e-01 7.35840993e-03
3.43320072e-01 5.72526129e-03 -1.51828855e-01 -1.85718417e-01
-4.34238732e-01 -2.63745517e-01 9.08591509e-01 3.70944291e-01
1.11177087e+00 -3.35962445e-01 -6.60773575e-01 7.08183944e-01
1.07969308e+00 3.70258361e-01 3.65384191e-01 -6.56593964e-03
1.25735343e+00 5.64094067e-01 5.11981606e-01 4.40705985e-01
4.90514606e-01 7.35546887e-01 3.69551927e-01 -1.85547009e-01
4.37701494e-01 -2.27721080e-01 1.28335267e-01 1.08046544e+00
-1.44558042e-01 -2.49451473e-01 -1.00168514e+00 9.01112109e-02
-2.14507651e+00 -5.15112877e-01 -3.31887275e-01 1.86134112e+00
9.32200253e-01 1.80272207e-01 -1.18372083e-01 -3.90838861e-01
7.68564641e-01 -5.37644140e-02 -9.03295994e-01 -2.34700888e-01
6.51168749e-02 1.18386209e-01 2.81453073e-01 5.39950848e-01
-9.04299915e-01 8.32372665e-01 5.64603519e+00 4.09209728e-01
-9.75651085e-01 -3.57940465e-01 7.30431139e-01 1.51304722e-01
-4.89504129e-01 1.18400283e-01 -7.95515716e-01 5.88724732e-01
5.96326768e-01 2.64017969e-01 3.22401702e-01 5.23040950e-01
2.04255536e-01 2.46107429e-01 -1.03048611e+00 7.03068733e-01
-2.45830983e-01 -1.48364377e+00 2.90422916e-01 3.09589744e-01
5.39800942e-01 1.22937441e-01 -2.67880023e-01 1.65968668e-03
5.97454250e-01 -9.81061339e-01 -1.97842941e-01 6.63014948e-01
3.15208822e-01 -8.83274615e-01 7.66929150e-01 2.57753015e-01
-1.36243343e+00 3.21696639e-01 -6.58845603e-01 -3.30170631e-01
3.17929126e-02 8.66951466e-01 -7.98387229e-01 3.98713887e-01
4.54618782e-01 1.24311864e+00 -8.75004679e-02 8.80152702e-01
8.18523541e-02 3.06425899e-01 -2.34486893e-01 -1.72938123e-01
2.61074692e-01 -7.78211355e-01 1.84305102e-01 6.77873313e-01
-1.73113290e-02 4.49508011e-01 6.21653676e-01 6.62864327e-01
-1.87827364e-01 3.10504824e-01 -5.34394026e-01 -1.11267596e-01
2.66733915e-01 1.08679855e+00 -3.52764755e-01 -1.12125322e-01
-8.02434742e-01 1.22266817e+00 9.66172218e-01 4.84363794e-01
-7.18392789e-01 -5.00016451e-01 1.27340198e+00 9.40138847e-02
4.50884908e-01 -4.84606534e-01 2.80143738e-01 -1.23847461e+00
1.68981582e-01 -7.87192702e-01 3.47868763e-02 -6.58411503e-01
-1.62424743e+00 4.72086430e-01 -5.93617022e-01 -8.54965866e-01
6.38757348e-02 -8.96347284e-01 -5.92646241e-01 1.03550577e+00
-1.44586158e+00 -1.08666039e+00 -1.06988580e-03 2.43017673e-01
3.71929139e-01 4.05069590e-02 7.41777182e-01 2.06598431e-01
-7.10899830e-01 3.52152586e-01 3.73708546e-01 2.22065106e-01
7.69695163e-01 -1.24491763e+00 7.49772727e-01 7.16338754e-02
-2.44052306e-01 9.73181665e-01 4.78206456e-01 -9.41365719e-01
-1.56724191e+00 -1.12353885e+00 9.81459618e-01 4.52198349e-02
7.65331388e-01 -3.50724071e-01 -1.26836550e+00 6.75418377e-01
5.43752685e-02 1.70142725e-01 1.02406144e+00 2.25950569e-01
-3.99874359e-01 8.55419636e-02 -6.86394393e-01 5.62033355e-01
1.17693603e+00 -5.19844472e-01 -3.86860937e-01 6.05331242e-01
1.00634170e+00 -1.47608235e-01 -1.11066854e+00 6.22659564e-01
6.06693029e-01 -8.55001748e-01 1.08319604e+00 -9.00655568e-01
6.03522718e-01 -2.59266943e-01 4.82684076e-02 -1.20590258e+00
-6.26720667e-01 -2.80625880e-01 -1.14997439e-01 6.26774669e-01
6.28181279e-01 -6.84802115e-01 8.64134014e-01 6.34314954e-01
-1.68003812e-01 -1.26667595e+00 -4.59015876e-01 -1.55800208e-01
-5.25260679e-02 2.69145489e-01 5.93474627e-01 8.75948548e-01
1.69601142e-01 5.93596637e-01 -6.08786583e-01 1.61035255e-01
4.06107903e-01 6.37417734e-01 7.38887370e-01 -1.38270783e+00
-4.95626301e-01 2.71627121e-02 -4.80590224e-01 -1.61684942e+00
4.84596461e-01 -8.34932148e-01 1.08231649e-01 -1.45613670e+00
7.33556509e-01 -2.81546891e-01 -6.09994888e-01 4.54424739e-01
-5.20790756e-01 1.77399889e-02 -2.90934294e-01 4.18055862e-01
-8.08165967e-01 8.28232050e-01 1.30738330e+00 -1.92755312e-01
-5.12556620e-02 -2.59787381e-01 -3.78026724e-01 5.39589465e-01
7.63390362e-01 -3.23487282e-01 -2.10843980e-01 -3.11163038e-01
3.54919314e-01 2.18416322e-02 2.49614388e-01 -7.47802854e-01
1.70514375e-01 -3.24038416e-01 6.25835061e-01 -7.64495492e-01
4.29311931e-01 -9.21679974e-01 1.19281173e-01 5.03783345e-01
-3.77018303e-01 1.62026390e-01 -1.72277793e-01 8.62969458e-01
-2.21012995e-01 2.48123199e-01 4.33674365e-01 -8.63069110e-03
-3.55920613e-01 8.87558937e-01 -3.79895419e-01 -3.24458301e-01
9.00770247e-01 1.15993127e-01 1.55492797e-02 -6.36241809e-02
-9.20700133e-01 3.50275576e-01 7.85993636e-01 1.98683724e-01
7.17455685e-01 -1.34333730e+00 -4.60472167e-01 2.88909614e-01
-7.92675540e-02 -5.44856526e-02 2.59305149e-01 6.28275096e-01
-7.37896800e-01 3.94709587e-01 -3.92464221e-01 -5.33828914e-01
-1.37406003e+00 6.19625866e-01 1.99699044e-01 -4.21987265e-01
-5.21603107e-01 8.36139917e-01 6.35108829e-01 -6.56344473e-01
-9.35866237e-02 -2.83611000e-01 -1.60457790e-01 -3.08309346e-01
3.73920649e-01 -2.24133953e-01 -1.30234271e-01 -5.34560621e-01
-4.39638406e-01 5.93126893e-01 -7.57842422e-01 5.97713947e-01
1.57060444e+00 2.10920677e-01 -5.69583952e-01 3.16648096e-01
1.66727734e+00 -3.01455885e-01 -1.37887549e+00 -3.91169012e-01
1.24950055e-02 -2.38631055e-01 -2.93554425e-01 -4.72247422e-01
-1.07013655e+00 9.41335380e-01 1.79301322e-01 -2.13478506e-01
8.48294854e-01 -1.08949877e-01 1.18267596e+00 6.96925402e-01
2.70205379e-01 -5.59641719e-01 4.61638495e-02 7.37003863e-01
5.85256398e-01 -1.35090625e+00 -6.44275919e-02 -6.71524823e-01
-4.79690939e-01 1.04358053e+00 7.16150761e-01 -2.98991114e-01
7.75647342e-01 1.47532955e-01 -1.68667197e-01 -3.50395530e-01
-9.98010159e-01 5.52102700e-02 3.10078353e-01 3.78113598e-01
9.03242052e-01 4.76443321e-02 -4.90820974e-01 5.91002226e-01
4.05752450e-01 -7.78571442e-02 -8.30551833e-02 8.03125203e-01
-4.93356168e-01 -1.59512758e+00 9.29020494e-02 7.26890087e-01
-3.43196034e-01 -4.54041451e-01 -8.33168209e-01 1.50867403e-01
-9.12548825e-02 5.77350736e-01 2.38276049e-02 -6.41190529e-01
-1.21244870e-01 3.18220928e-02 1.85829848e-01 -6.98310673e-01
-4.79926080e-01 6.32318556e-02 -1.24302424e-01 -5.15908420e-01
-3.55678350e-01 -4.26921219e-01 -1.43362582e+00 -3.44758004e-01
-2.12940365e-01 5.87704778e-01 3.17385375e-01 9.08324957e-01
1.00324941e+00 3.28564972e-01 6.77230120e-01 -9.60351706e-01
-3.73928666e-01 -8.64555776e-01 -4.59423959e-01 5.51596224e-01
3.57741117e-01 -7.41050065e-01 -1.95161849e-01 -1.61564261e-01] | [4.800875663757324, 5.7116923332214355] |
9c978e15-4461-4f84-8f16-5ccfe01305b6 | prioritized-trajectory-replay-a-replay-memory | 2306.15503 | null | https://arxiv.org/abs/2306.15503v1 | https://arxiv.org/pdf/2306.15503v1.pdf | Prioritized Trajectory Replay: A Replay Memory for Data-driven Reinforcement Learning | In recent years, data-driven reinforcement learning (RL), also known as offline RL, have gained significant attention. However, the role of data sampling techniques in offline RL has been overlooked despite its potential to enhance online RL performance. Recent research suggests applying sampling techniques directly to state-transitions does not consistently improve performance in offline RL. Therefore, in this study, we propose a memory technique, (Prioritized) Trajectory Replay (TR/PTR), which extends the sampling perspective to trajectories for more comprehensive information extraction from limited data. TR enhances learning efficiency by backward sampling of trajectories that optimizes the use of subsequent state information. Building on TR, we build the weighted critic target to avoid sampling unseen actions in offline training, and Prioritized Trajectory Replay (PTR) that enables more efficient trajectory sampling, prioritized by various trajectory priority metrics. We demonstrate the benefits of integrating TR and PTR with existing offline RL algorithms on D4RL. In summary, our research emphasizes the significance of trajectory-based data sampling techniques in enhancing the efficiency and performance of offline RL algorithms. | ['Changjie Fan', 'Tangjie Lv', 'Yan Zheng', 'Yujing Hu', 'Jianye Hao', 'Yi Ma', 'Jinyi Liu'] | 2023-06-27 | null | null | null | null | ['offline-rl', 'd4rl'] | ['playing-games', 'robots'] | [-1.24284647e-01 -1.01793438e-01 -9.75840688e-01 -2.48788998e-01
-8.54912341e-01 -7.03141034e-01 6.86904311e-01 1.84754536e-01
-7.28228748e-01 9.23762918e-01 5.22452772e-01 -5.59661686e-01
-4.28703547e-01 -7.15579748e-01 -5.75623393e-01 -6.08559310e-01
-2.23731801e-01 1.78190634e-01 2.21576408e-01 -6.78484067e-02
4.63065177e-01 8.24735403e-01 -1.54639733e+00 -1.41053833e-02
8.96016002e-01 6.13922656e-01 3.73649865e-01 4.92519200e-01
-2.09570661e-01 1.22854853e+00 -6.55064106e-01 1.71496496e-01
3.16335201e-01 -6.75075352e-01 -6.74911559e-01 -1.45911127e-01
-2.71605819e-01 -9.39754665e-01 -5.69936395e-01 4.12777215e-01
4.89981383e-01 7.66167939e-01 1.45081237e-01 -1.37095737e+00
-1.93600014e-01 8.13727856e-01 -1.81631327e-01 3.38370144e-01
2.89997280e-01 5.80135584e-01 8.96937847e-01 -6.83642864e-01
5.48196673e-01 1.20567369e+00 4.79567558e-01 6.33603692e-01
-1.13875413e+00 -5.50602973e-01 5.13758540e-01 1.96737334e-01
-8.70945871e-01 -5.96795738e-01 7.37512946e-01 -5.20324372e-02
1.29106677e+00 5.49597144e-02 9.87284720e-01 1.11032641e+00
-3.35136987e-02 1.39990795e+00 1.06654406e+00 -3.39873910e-01
6.53582633e-01 -5.21663837e-02 -4.60831411e-02 4.43663448e-01
7.16153160e-02 7.07855105e-01 -7.11102188e-01 -2.31924448e-02
7.44065523e-01 1.09335490e-01 3.70208509e-02 -4.42751914e-01
-1.09307086e+00 8.06736350e-01 1.85560837e-01 -7.81312659e-02
-4.23788458e-01 3.10819060e-01 4.71056908e-01 3.69752437e-01
3.07080716e-01 5.87334394e-01 -2.18703434e-01 -9.17310596e-01
-1.03108573e+00 5.55932879e-01 4.91927534e-01 1.28341186e+00
7.34794080e-01 4.46581542e-01 -5.39978445e-01 5.82240701e-01
2.12055951e-01 5.27858555e-01 4.96516019e-01 -1.23924303e+00
7.01310098e-01 7.95296907e-01 4.63080436e-01 -4.44918752e-01
-3.97967011e-01 -1.48435190e-01 -1.43850431e-01 9.30197760e-02
3.66980106e-01 -3.40817571e-01 -6.26171947e-01 1.76093781e+00
4.27308798e-01 1.56001270e-01 2.71819979e-01 9.20199454e-01
7.70716965e-02 5.62000275e-01 1.60652667e-01 -2.64066279e-01
3.13145310e-01 -1.09701014e+00 -6.88094497e-01 6.01384649e-03
9.02770042e-01 -5.08287176e-02 1.34200287e+00 4.19869632e-01
-1.10880351e+00 -3.90985519e-01 -8.61409187e-01 7.09416270e-02
-2.39749208e-01 2.42892638e-01 6.05148911e-01 5.61064899e-01
-8.88721824e-01 1.04363775e+00 -1.34684563e+00 -2.99615651e-01
4.21597958e-01 4.81671989e-01 1.65826440e-01 6.08600602e-02
-1.04731941e+00 9.68675315e-01 4.31590527e-01 -1.25196248e-01
-1.40970635e+00 -6.40372992e-01 -6.92946494e-01 -2.66514003e-01
8.31441045e-01 -2.24452335e-02 1.57534552e+00 -7.33018816e-01
-2.12741876e+00 -2.28162512e-01 -1.22570567e-01 -7.47888207e-01
5.20754278e-01 -4.31336761e-01 -3.38261664e-01 2.82264918e-01
-2.26682797e-01 4.92015809e-01 7.42636085e-01 -1.01961040e+00
-8.84622157e-01 -1.42302886e-01 1.81224123e-01 5.73047161e-01
-3.54059696e-01 -2.93727845e-01 -3.47281575e-01 -4.01270777e-01
-5.04201055e-01 -9.67931688e-01 -3.76700789e-01 -3.96657377e-01
8.09451286e-03 -3.50489497e-01 9.41865206e-01 -4.28081870e-01
1.69165099e+00 -2.06263995e+00 1.64056435e-01 2.01142147e-01
-1.18509360e-01 3.96633238e-01 -2.10953951e-01 9.80923533e-01
5.72593451e-01 -2.38525674e-01 -7.18886554e-02 -2.24133819e-01
1.75185762e-02 4.98642236e-01 -6.17302418e-01 3.66312325e-01
1.46677107e-01 1.08726668e+00 -1.37396908e+00 -2.49375299e-01
5.12452185e-01 -1.57246925e-02 -6.23771966e-01 2.89114952e-01
-4.95822340e-01 7.25438237e-01 -6.86321616e-01 5.72086275e-01
2.28153557e-01 3.03137023e-03 3.17767322e-01 3.06093693e-01
-6.06101394e-01 5.32986104e-01 -8.46347332e-01 1.64638305e+00
-4.46801931e-01 3.95159215e-01 -3.40809762e-01 -5.46347439e-01
1.05278134e+00 6.26161620e-02 8.11858654e-01 -1.10648787e+00
-1.68578401e-01 8.91699269e-03 -2.03933492e-01 -5.09585202e-01
1.05294490e+00 3.59544247e-01 -3.39209884e-02 8.18311274e-01
-1.38965532e-01 2.33085871e-01 2.25331843e-01 2.44451433e-01
1.17701435e+00 9.20986056e-01 2.92280585e-01 3.66514409e-03
1.86508328e-01 5.15771091e-01 5.50462008e-01 8.38782668e-01
-5.35949171e-01 -7.10036755e-02 3.23207587e-01 -2.08765883e-02
-1.00379574e+00 -9.11179185e-01 4.43092167e-01 1.36093819e+00
1.90427497e-01 -4.59227681e-01 -6.32167399e-01 -9.85125661e-01
3.35411102e-01 1.21809494e+00 -3.36878657e-01 -5.01448512e-01
-9.43364203e-01 -2.40633845e-01 7.86616266e-01 7.89716065e-01
4.44782317e-01 -1.44290829e+00 -1.06413698e+00 5.32992840e-01
1.06609397e-01 -5.45794427e-01 -5.21123946e-01 2.54233420e-01
-1.32468152e+00 -9.55145538e-01 -5.83799124e-01 -7.47619718e-02
5.90574861e-01 4.15336341e-01 7.24242687e-01 -3.16667914e-01
2.09526941e-01 7.79544830e-01 -6.64125323e-01 -2.50094444e-01
-4.80165660e-01 2.83131868e-01 1.73252106e-01 -3.31803650e-01
2.70798028e-01 -2.57522732e-01 -5.74561477e-01 2.60783166e-01
-7.63050795e-01 -1.87332943e-01 6.36405289e-01 6.13112628e-01
6.45920992e-01 -1.61052749e-01 1.07855296e+00 -7.88619399e-01
9.13319528e-01 -6.51622176e-01 -6.80602074e-01 2.43329987e-01
-1.05257499e+00 3.32992107e-01 9.21555877e-01 -7.40003288e-01
-1.16734076e+00 -7.29668140e-02 9.20659304e-03 -7.35608280e-01
-2.88241897e-02 5.37413418e-01 1.90429032e-01 7.54541010e-02
4.66244519e-01 5.56511998e-01 3.05788577e-01 -3.64999622e-01
4.60162044e-01 5.05487621e-01 -5.45206890e-02 -6.80428743e-01
3.42500687e-01 3.04517299e-01 -1.34047493e-01 -6.12641752e-01
-6.53018355e-01 -4.40984011e-01 -4.75527942e-01 -5.40487766e-01
2.92382896e-01 -7.24931836e-01 -9.61186290e-01 1.71133712e-01
-3.81578356e-01 -1.07278895e+00 -8.48523319e-01 5.64171493e-01
-1.04189765e+00 8.18829983e-02 -3.21029156e-01 -1.36954331e+00
-1.76797003e-01 -1.22125661e+00 7.01102555e-01 1.45195290e-01
-1.36765182e-01 -9.42617297e-01 2.47596189e-01 -1.81261197e-01
4.46592033e-01 6.05104975e-02 6.86397970e-01 -6.30491018e-01
-7.15149760e-01 1.97666526e-01 2.01378927e-01 1.50431156e-01
1.27298549e-01 -2.50577658e-01 -6.17253482e-01 -7.17526555e-01
-3.87637407e-01 -5.52705646e-01 5.82945108e-01 2.27707848e-01
1.15207803e+00 -5.21591902e-01 -3.14241201e-01 2.52776682e-01
1.44805586e+00 5.99014699e-01 3.75683635e-01 5.05254507e-01
5.39893389e-01 3.08916181e-01 1.37740302e+00 7.75561631e-01
4.81124938e-01 5.98666787e-01 3.56236219e-01 3.54399681e-01
-7.92654902e-02 -8.88046563e-01 8.49268973e-01 6.97995305e-01
-4.08356600e-02 -1.18152775e-01 -6.33489609e-01 6.72624111e-01
-2.11782575e+00 -1.00061941e+00 2.67064601e-01 2.52539396e+00
8.20679426e-01 -6.93463981e-02 7.46905088e-01 3.83411869e-02
2.09848911e-01 1.86303809e-01 -1.26266205e+00 -3.85562927e-01
4.02397871e-01 1.71080846e-02 8.79180968e-01 5.02562463e-01
-6.17037714e-01 1.18093610e+00 6.76528311e+00 8.24678361e-01
-1.06928074e+00 -7.44305030e-02 1.36125758e-01 -4.28387880e-01
-3.77571970e-01 1.53202623e-01 -1.22095227e+00 4.62800801e-01
1.22171950e+00 -2.73116857e-01 1.04596019e+00 1.07172501e+00
7.08580613e-01 -2.59548098e-01 -1.17853081e+00 4.97860938e-01
-3.94356579e-01 -1.24739230e+00 3.88798341e-02 1.09567150e-01
5.63208938e-01 -3.81914377e-02 5.82250878e-02 8.83718312e-01
6.55627191e-01 -7.41993189e-01 8.03273916e-01 7.08503425e-01
6.25893712e-01 -1.25098014e+00 5.38505018e-02 6.26787186e-01
-1.27889645e+00 -6.09806001e-01 -1.77540302e-01 9.97014269e-02
7.51071349e-02 4.11859415e-02 -1.07706451e+00 5.65767884e-01
2.61689097e-01 9.93710339e-01 -2.95142889e-01 8.19856822e-01
-1.25321269e-01 8.16566408e-01 -5.83250560e-02 -3.77540231e-01
5.87755561e-01 -2.97327787e-01 6.52283013e-01 1.01203191e+00
1.84815973e-01 -4.82486933e-03 3.39543551e-01 7.10755408e-01
2.74463236e-01 -1.02850109e-01 -7.57310450e-01 -3.43495637e-01
9.42411005e-01 8.85119677e-01 -3.97140563e-01 -1.23730466e-01
-2.34925061e-01 6.22089446e-01 6.20911121e-01 4.65884387e-01
-8.63267303e-01 -2.47171119e-01 5.34805894e-01 -5.82895316e-02
2.81453192e-01 -6.99512005e-01 1.21662885e-01 -6.70949996e-01
-3.79093081e-01 -9.79929864e-01 2.93841481e-01 -2.12457567e-01
-8.35707843e-01 9.89173353e-02 3.57371032e-01 -1.48422313e+00
-4.33988541e-01 -3.35876308e-02 -2.89213985e-01 5.03618121e-01
-1.57241035e+00 -6.11014843e-01 5.00863791e-02 7.10110307e-01
8.57664645e-01 -1.45872355e-01 4.94827092e-01 9.06617790e-02
-7.88345516e-01 7.19615698e-01 4.02637213e-01 -3.49584162e-01
4.15066183e-01 -1.26680636e+00 4.22815800e-01 6.65833652e-01
-7.82362595e-02 7.76701152e-01 3.11422884e-01 -9.95402873e-01
-2.00712347e+00 -1.49246812e+00 2.78247178e-01 -2.26878077e-01
4.66365993e-01 -1.27747089e-01 -6.51064634e-01 7.50659943e-01
-1.79192662e-01 -3.83850157e-01 3.43996823e-01 -1.08199723e-01
2.12604985e-01 -1.94217831e-01 -1.24243438e+00 9.91099417e-01
1.04228806e+00 -4.02562976e-01 -1.92866713e-01 -5.29953353e-02
8.05032969e-01 -4.60283041e-01 -1.01157463e+00 7.73860887e-02
4.38290685e-01 -8.10141981e-01 8.73635411e-01 -5.36964595e-01
-6.66914135e-02 -1.26422390e-01 9.65806246e-02 -1.37982285e+00
-1.61605358e-01 -8.06779802e-01 -7.86200702e-01 1.00207400e+00
2.39273444e-01 -5.29802263e-01 1.00091171e+00 6.12562656e-01
-2.81189978e-01 -1.05549991e+00 -7.05526352e-01 -1.15640199e+00
-7.95603245e-02 -4.67853248e-01 8.84931147e-01 5.29795706e-01
2.05712095e-01 -2.33024687e-01 -5.06546438e-01 -1.74791917e-01
4.93289232e-01 8.41421187e-02 9.16292369e-01 -5.63934803e-01
-3.55308264e-01 -3.25714380e-01 4.43091124e-01 -1.51777971e+00
1.24136113e-01 -8.26273859e-01 1.93182245e-01 -1.49392438e+00
-1.77810594e-01 -8.09931993e-01 -3.05473804e-01 5.44249237e-01
-1.94256678e-02 -3.86386693e-01 3.62203181e-01 4.13706809e-01
-8.92340064e-01 9.76589143e-01 1.55838084e+00 4.17150855e-01
-8.41797471e-01 2.15662703e-01 -2.49459594e-01 2.31321320e-01
1.08771920e+00 -4.27552849e-01 -9.19339120e-01 -1.70502901e-01
-7.97742754e-02 5.55550158e-01 3.28094848e-02 -9.57835495e-01
3.75493556e-01 -6.20252073e-01 3.46795321e-02 -8.95680308e-01
3.10984343e-01 -8.36585760e-01 -1.76677704e-01 5.20217419e-01
-7.92770207e-01 2.39433080e-01 3.76583397e-01 9.05593753e-01
1.54549822e-01 -2.31069222e-01 3.77673149e-01 -1.68518856e-01
-8.41140687e-01 4.28726137e-01 -6.85010314e-01 1.93356827e-01
1.27996540e+00 -3.65801573e-01 -1.58136293e-01 -3.12226087e-01
-4.88355190e-01 4.96505171e-01 3.09394330e-01 3.89523029e-01
8.12330127e-01 -1.26047158e+00 2.56495029e-02 1.84617937e-01
-1.05888732e-01 -1.24153353e-01 -8.80236030e-02 9.23658133e-01
-8.68710279e-02 6.03928447e-01 -2.06418976e-01 -3.42215478e-01
-8.83066773e-01 5.87092757e-01 1.78161815e-01 -4.25898701e-01
-9.55462635e-01 2.49594644e-01 -7.47219026e-01 -5.52806735e-01
6.48084044e-01 -5.14737844e-01 -2.39697024e-01 -3.37150395e-02
3.65215659e-01 1.05280590e+00 -2.32449383e-01 -1.98052041e-02
-8.17459300e-02 -1.07874259e-01 -1.06544532e-01 -4.35677081e-01
1.21285224e+00 -2.16308072e-01 5.18588245e-01 8.09151649e-01
6.86163723e-01 -3.23564857e-02 -2.05051780e+00 -9.58470851e-02
3.55283082e-01 -4.26274985e-01 1.27594411e-01 -1.01327872e+00
-8.11347783e-01 4.77253228e-01 3.43426168e-01 6.84271306e-02
1.04239058e+00 -5.98040223e-01 9.92599130e-01 5.58090150e-01
8.67015779e-01 -1.41452646e+00 2.44133860e-01 6.74587190e-01
4.61415619e-01 -9.63177562e-01 9.67916567e-03 3.84372085e-01
-1.09644699e+00 9.89566863e-01 6.74607337e-01 -2.52831906e-01
2.00855985e-01 2.03252584e-01 -3.01916897e-01 1.69779718e-01
-9.47171152e-01 -3.58242810e-01 -1.05585694e-01 5.30064583e-01
5.14405891e-02 -1.71267185e-02 -1.82342634e-01 2.21979007e-01
2.17809469e-01 4.00760829e-01 2.87862241e-01 1.35962546e+00
-5.79032779e-01 -1.47170615e+00 -2.22873148e-02 4.77794141e-01
-3.72461826e-02 1.60401821e-01 -1.92559198e-01 9.54046190e-01
-4.03335571e-01 9.67937231e-01 1.41481012e-02 -5.16067684e-01
2.84622014e-01 3.28813083e-02 6.00461662e-01 -3.13626885e-01
-9.35288668e-01 -2.01569602e-01 2.50598729e-01 -1.01075637e+00
-2.43111297e-01 -8.64727855e-01 -1.71557593e+00 -2.92427719e-01
-1.94261014e-01 2.47343212e-01 6.10416353e-01 1.03184104e+00
7.16803908e-01 5.09464443e-01 8.40167403e-01 -7.07028329e-01
-1.03304231e+00 -6.20454490e-01 -4.07894880e-01 -8.15078244e-02
5.46681285e-01 -8.41975570e-01 -4.86419722e-02 -5.01209378e-01] | [4.061812400817871, 2.1586945056915283] |
7622551b-d20a-4cc1-a798-43090170e88f | kvasir-instrument-diagnostic-and-therapeutic | 2011.08065 | null | https://arxiv.org/abs/2011.08065v1 | https://arxiv.org/pdf/2011.08065v1.pdf | Kvasir-Instrument: Diagnostic and therapeutic tool segmentation dataset in gastrointestinal endoscopy | Gastrointestinal (GI) pathologies are periodically screened, biopsied, and resected using surgical tools. Usually the procedures and the treated or resected areas are not specifically tracked or analysed during or after colonoscopies. Information regarding disease borders, development and amount and size of the resected area get lost. This can lead to poor follow-up and bothersome reassessment difficulties post-treatment. To improve the current standard and also to foster more research on the topic we have released the ``Kvasir-Instrument'' dataset which consists of $590$ annotated frames containing GI procedure tools such as snares, balloons and biopsy forceps, etc. Beside of the images, the dataset includes ground truth masks and bounding boxes and has been verified by two expert GI endoscopists. Additionally, we provide a baseline for the segmentation of the GI tools to promote research and algorithm development. We obtained a dice coefficient score of 0.9158 and a Jaccard index of 0.8578 using a classical U-Net architecture. A similar dice coefficient score was observed for DoubleUNet. The qualitative results showed that the model did not work for the images with specularity and the frames with multiple instruments, while the best result for both methods was observed on all other types of images. Both, qualitative and quantitative results show that the model performs reasonably good, but there is a large potential for further improvements. Benchmarking using the dataset provides an opportunity for researchers to contribute to the field of automatic endoscopic diagnostic and therapeutic tool segmentation for GI endoscopy. | ['Pål Halvorsen', 'Dag Johansen', 'Håvard D. Johansen', 'Peter T. Schmidt', 'Thomas de Lange', 'Michael A. Riegler', 'Enrique Garcia-Ceja', 'VajiraThambawita', 'Steven A. Hicks', 'Krister Emanuelsen', 'Sharib Ali', 'Debesh Jha'] | 2020-10-23 | null | null | null | null | ['instrument-recognition'] | ['audio'] | [-2.20424756e-01 2.60516673e-01 -2.04113498e-01 8.08337629e-02
-3.61375600e-01 -1.07402670e+00 5.59330583e-02 4.38227892e-01
-4.38256145e-01 3.52760226e-01 1.54118054e-02 -5.99546194e-01
-2.07801044e-01 -6.72270656e-01 -5.37080407e-01 -6.97670877e-01
-4.08034116e-01 3.36983562e-01 4.44648683e-01 1.32455658e-02
4.79247272e-02 2.48563677e-01 -9.17808175e-01 1.92815468e-01
7.32498705e-01 8.77825201e-01 3.90255481e-01 7.98212409e-01
2.70747542e-01 4.01856154e-01 -4.35830176e-01 -4.42646593e-01
7.07364500e-01 -5.33177495e-01 -9.04207349e-01 1.20676719e-01
1.43702447e-01 -1.68325827e-01 -5.17447405e-02 1.11525559e+00
3.83652449e-01 8.45247656e-02 4.88433242e-01 -6.05855465e-01
-2.83383846e-01 4.84309554e-01 -4.82406616e-01 3.48017275e-01
4.86053079e-01 2.79414535e-01 5.32175064e-01 3.34143676e-02
1.13276672e+00 5.96136272e-01 1.03773952e+00 3.24289232e-01
-9.45531011e-01 -2.53500432e-01 -2.17457175e-01 -2.23527625e-01
-7.50118613e-01 3.87207896e-01 6.45445660e-02 -6.17436528e-01
6.69160783e-01 4.40285742e-01 1.17749631e+00 5.80464840e-01
2.05864519e-01 6.08579814e-01 7.32388556e-01 -5.20836711e-01
-6.44083917e-02 2.51859754e-01 -1.03806138e-01 9.29459274e-01
7.33175755e-01 4.52770829e-01 4.35605764e-01 -8.48815218e-02
1.08164382e+00 3.42344940e-01 -6.99099004e-01 -6.49146795e-01
-1.61003733e+00 8.26735198e-01 6.77718759e-01 6.83444440e-01
-4.77431178e-01 -3.01773399e-01 8.20713043e-01 3.97622913e-01
6.98816329e-02 7.57376313e-01 -3.02857995e-01 -3.13694887e-02
-9.29655015e-01 -2.78714627e-01 1.10555136e+00 7.38313198e-01
5.42109162e-02 -4.35541242e-01 2.00865239e-01 6.22657120e-01
1.16194084e-01 -2.82618194e-03 9.11077201e-01 -8.00048709e-01
1.52166337e-01 9.95605588e-01 2.39004076e-01 -1.01836503e+00
-6.99548602e-01 -3.73296738e-01 -7.51288652e-01 2.42878348e-01
7.51412630e-01 -3.11535329e-01 -1.04072595e+00 8.57990026e-01
3.24139595e-01 -1.09929807e-01 -1.68944731e-01 1.13347089e+00
9.82318938e-01 1.61495239e-01 -8.25622007e-02 -9.02075507e-03
1.58318377e+00 -1.29698038e+00 -5.76457858e-01 7.71179572e-02
1.15429866e+00 -1.05570173e+00 7.90966809e-01 6.25207186e-01
-1.27592695e+00 -4.13129866e-01 -1.02313423e+00 2.88492203e-01
-2.99507052e-01 4.90936279e-01 5.60016215e-01 8.16090584e-01
-9.57198918e-01 7.88845122e-01 -1.20889544e+00 -4.28701907e-01
2.59953856e-01 4.40595239e-01 -5.34174144e-01 -3.43374200e-02
-6.73589647e-01 1.13364995e+00 3.56290728e-01 -6.23500422e-02
-4.77106035e-01 -5.97611547e-01 -9.98808444e-01 -1.95846647e-01
3.31391931e-01 -8.56247544e-01 1.15483391e+00 -8.13294649e-01
-1.37826657e+00 1.24974453e+00 4.36843485e-01 -7.48049617e-01
7.96984553e-01 1.62978202e-01 -1.34025082e-01 3.11255485e-01
-1.60425037e-01 5.38060009e-01 2.17189223e-01 -8.63841832e-01
-9.90557134e-01 -3.34767759e-01 1.67482987e-01 1.57111526e-01
1.46318540e-01 -1.30395293e-01 -5.85329294e-01 -7.81876504e-01
8.26973543e-02 -1.34120142e+00 -4.15825814e-01 1.97717085e-01
-2.31819272e-01 3.68442804e-01 2.86507130e-01 -1.11707354e+00
1.37877154e+00 -2.08721590e+00 -4.67948988e-02 1.66153312e-01
-2.40106210e-02 5.42049944e-01 3.78531158e-01 6.68886974e-02
-1.83029637e-01 2.43338421e-01 -1.81778334e-02 1.76177800e-01
-3.19838494e-01 -5.80174923e-02 4.27275896e-01 9.05131340e-01
-6.08637631e-01 7.18117297e-01 -1.12387455e+00 -4.07338440e-01
8.60796094e-01 3.05697143e-01 -6.04218423e-01 -6.54027984e-02
1.08051069e-01 5.60007691e-01 -1.03080697e-01 7.42487311e-01
4.96656865e-01 -4.16801304e-01 3.34598869e-01 -4.17246878e-01
-1.21710710e-01 1.32585630e-01 -1.29043770e+00 1.98943853e+00
-4.55775023e-01 4.32452738e-01 4.20412868e-01 -7.94210970e-01
4.51694041e-01 7.35663176e-01 6.82755172e-01 -4.87450629e-01
3.73478353e-01 5.72943747e-01 3.47957432e-01 -6.24774694e-01
1.82696596e-01 6.10997267e-02 2.11722553e-01 2.86498427e-01
9.14652925e-03 -3.46498489e-01 6.04260027e-01 -2.98938811e-01
9.07780230e-01 1.87040120e-01 7.15608239e-01 -5.24780750e-01
4.68348354e-01 7.69105136e-01 1.43381000e-01 3.90608847e-01
-6.22827828e-01 5.43426633e-01 1.76278308e-01 -7.60851502e-01
-9.74612474e-01 -7.01565504e-01 -3.37648511e-01 3.06922615e-01
5.06550074e-01 8.02274719e-02 -7.94927776e-01 -7.90148914e-01
-1.87363625e-01 1.86462134e-01 -6.12998307e-01 1.70702279e-01
-5.43851912e-01 -7.02580929e-01 1.16692990e-01 3.17975610e-01
1.73204675e-01 -1.02548385e+00 -1.06637454e+00 3.74493986e-01
-6.18070476e-02 -7.77061582e-01 -5.34664869e-01 -8.76936223e-03
-1.19951844e+00 -1.69583189e+00 -1.17343903e+00 -1.41320717e+00
8.35784733e-01 3.14070225e-01 1.18709338e+00 8.59273076e-02
-7.09637225e-01 1.69936240e-01 -4.88450468e-01 -2.71757960e-01
-5.76197922e-01 -4.11309339e-02 -4.27712023e-01 -8.15801263e-01
-4.08221036e-02 -3.22171673e-02 -1.33077729e+00 7.67792284e-01
-9.18764532e-01 -1.10202864e-01 2.93129355e-01 9.54987168e-01
6.85488284e-01 -8.98095146e-02 -1.60997007e-02 -6.99857175e-01
6.76464319e-01 -2.87311584e-01 -7.82940686e-01 2.48311087e-01
-5.46727955e-01 -3.79323035e-01 4.69287813e-01 -5.16468108e-01
-7.26961613e-01 -1.31084591e-01 6.02115840e-02 -1.96919873e-01
-1.59490451e-01 3.83877963e-01 9.15473282e-01 -2.46908873e-01
8.33244205e-01 -5.08480728e-01 2.93066144e-01 -1.98528349e-01
6.65254295e-02 4.37524825e-01 3.85861546e-01 6.21538260e-04
1.35031968e-01 7.49881685e-01 -2.77190506e-01 -2.99468964e-01
-2.85211295e-01 -1.07885957e+00 -3.18665206e-01 6.11935817e-02
7.59421408e-01 -6.41087174e-01 -6.89491689e-01 -2.69705653e-02
-7.11318016e-01 -3.41859400e-01 -6.50574386e-01 1.08785033e+00
-3.29809874e-01 5.31806469e-01 -1.09053278e+00 -3.03140849e-01
-6.40330911e-01 -1.64950466e+00 6.29331529e-01 3.72341037e-01
-4.54167545e-01 -1.58316195e+00 2.16244787e-01 1.90061301e-01
3.63681346e-01 6.24422312e-01 5.47809005e-01 -6.84104323e-01
-3.32557499e-01 -6.38167441e-01 -1.02833077e-01 2.74745345e-01
4.14708287e-01 4.90842275e-02 -2.84413874e-01 -3.90473008e-01
2.34683141e-01 3.25677752e-01 5.75924754e-01 1.10924554e+00
8.52265954e-01 -1.83469370e-01 -5.03639102e-01 7.24508524e-01
1.51393580e+00 7.20319510e-01 4.05994922e-01 7.47071385e-01
2.18925044e-01 7.08687961e-01 8.11341405e-01 1.09084256e-01
-4.00555395e-02 4.64336514e-01 6.76933289e-01 -4.53944206e-01
-1.92487106e-01 2.00197384e-01 -6.93370402e-02 6.67476058e-01
-2.10267857e-01 -1.54717654e-01 -8.89624119e-01 8.99530590e-01
-1.46278453e+00 -8.01826596e-01 -2.55261213e-01 2.37526298e+00
5.51111281e-01 -1.79343358e-01 2.60394394e-01 -6.78984076e-03
7.03768373e-01 -2.90611416e-01 -1.53383613e-01 -4.30743366e-01
2.36382067e-01 7.51782535e-03 9.04370904e-01 3.86094540e-01
-1.24695492e+00 2.58832216e-01 6.62514400e+00 6.00635588e-01
-1.14975870e+00 -1.92714915e-01 5.40923178e-01 4.72726719e-03
1.91512287e-01 -3.30440551e-01 -3.59384924e-01 4.71757352e-01
6.36592269e-01 2.17531864e-02 3.01793724e-01 8.30238104e-01
1.82103977e-01 -3.54465097e-01 -7.81282127e-01 8.09352279e-01
-1.33715555e-01 -1.54878998e+00 -5.77632189e-01 1.70023009e-01
1.02442122e+00 3.42328250e-01 -1.59570530e-01 3.29407863e-02
2.59889930e-01 -6.74326241e-01 1.46036908e-01 1.31028906e-01
8.18080544e-01 -2.06562802e-01 1.20500815e+00 2.48050727e-02
-1.08842790e+00 1.39186293e-01 -6.98579177e-02 3.74271631e-01
2.06763342e-01 2.03677610e-01 -1.31298828e+00 5.85623026e-01
7.46853650e-01 5.65014422e-01 -1.54877797e-01 1.84889245e+00
6.30968809e-02 2.73988545e-01 -4.37470257e-01 1.39913693e-01
4.24181014e-01 -5.16817749e-01 5.94271302e-01 1.22518373e+00
4.83919322e-01 -4.13142353e-01 -5.40814959e-02 4.22102660e-01
1.74940601e-01 4.40005898e-01 -4.90883410e-01 7.15210885e-02
-4.93399426e-02 1.39129555e+00 -1.29142916e+00 -4.70589817e-01
-5.42745531e-01 9.20997500e-01 -6.64066911e-01 7.49323815e-02
-7.71773815e-01 -8.85965079e-02 -4.96998010e-03 3.23047966e-01
1.73719883e-01 3.11125427e-01 -2.80310541e-01 -9.98640716e-01
1.90005694e-02 -7.93989062e-01 7.73904145e-01 -4.64178830e-01
-9.85455573e-01 6.64880157e-01 -1.60427451e-01 -1.66663742e+00
-3.26316059e-01 -1.08771396e+00 -4.93843794e-01 6.14149749e-01
-1.30627418e+00 -7.04747498e-01 -5.16975820e-01 3.00872177e-01
4.21009421e-01 2.24351972e-01 1.00136614e+00 3.91532689e-01
-2.44347975e-01 2.25476161e-01 3.72930139e-01 1.32693052e-01
7.13693321e-01 -1.47704577e+00 -3.23386528e-02 4.31878746e-01
-1.47140607e-01 6.72500432e-01 6.13624871e-01 -6.79536343e-01
-9.00403917e-01 -7.77211308e-01 2.82141149e-01 -2.51498431e-01
6.99225247e-01 5.52177787e-01 -4.96625960e-01 7.65725970e-01
2.70255268e-01 5.70685081e-02 7.84386754e-01 -2.42476195e-01
3.20579529e-01 3.38809878e-01 -1.54466462e+00 4.46357399e-01
5.21312058e-01 2.97535080e-02 -5.94004154e-01 6.36479497e-01
4.10822928e-01 -1.03332233e+00 -1.36442828e+00 4.90641564e-01
7.36793101e-01 -1.11423683e+00 1.09199500e+00 -1.49203530e-02
1.89312667e-01 -5.58369383e-02 5.07315457e-01 -1.28834510e+00
4.66738492e-02 -5.33885419e-01 2.31727138e-01 3.41807455e-01
5.57532609e-01 -6.16542041e-01 9.79066014e-01 5.82953513e-01
-4.23437119e-01 -7.84013569e-01 -6.20228648e-01 -6.05829477e-01
-9.80430096e-02 3.60295549e-02 1.55841023e-01 1.07334483e+00
2.31848061e-01 -5.34606457e-01 2.69920021e-01 -1.40527934e-01
2.78416216e-01 2.85308778e-01 4.34815526e-01 -1.05614519e+00
-1.37618586e-01 -7.68955290e-01 -4.10904318e-01 -6.20605886e-01
-8.40599954e-01 -8.43601406e-01 -3.75445902e-01 -1.94104278e+00
3.12322509e-02 -2.71382183e-01 -3.73355001e-02 1.14124395e-01
-5.32716699e-02 1.71787068e-01 -9.55868289e-02 4.51914161e-01
-2.45805070e-01 -6.22755706e-01 1.88704371e+00 1.76956609e-01
-6.83982790e-01 3.62251222e-01 -3.02357167e-01 1.08021390e+00
8.53738010e-01 -2.87422985e-01 -2.08104655e-01 -8.54813010e-02
-1.50665203e-02 2.00219885e-01 3.45831424e-01 -7.67760813e-01
2.05767736e-01 3.56696367e-01 1.16052449e-01 -4.97350305e-01
-1.18633881e-01 -1.22253549e+00 4.60577846e-01 1.16617537e+00
2.21442077e-02 6.76854625e-02 2.10372582e-01 2.19090015e-01
-5.72670162e-01 -5.91708362e-01 8.76846969e-01 -8.09596777e-01
-3.21643949e-01 1.44534901e-01 -2.82795541e-02 -1.02491729e-01
1.34840512e+00 -7.11826384e-01 -1.45004749e-01 -6.04373068e-02
-1.16883457e+00 1.57251224e-01 9.64730620e-01 -4.33201604e-02
1.05309397e-01 -7.28186727e-01 -3.80161971e-01 1.63370907e-01
-1.00390270e-01 4.07093614e-01 5.51653922e-01 1.60017729e+00
-1.55896974e+00 6.18055820e-01 -7.83303827e-02 -6.51574194e-01
-1.24331605e+00 6.24897778e-01 8.06022525e-01 -7.07901001e-01
-8.33759844e-01 7.27153063e-01 1.27376854e-01 -3.25007528e-01
3.07733089e-01 -1.10405660e+00 -2.29204312e-01 2.31344193e-01
3.02101851e-01 5.23172200e-01 2.11815193e-01 -3.33251923e-01
1.53916674e-02 5.23021042e-01 1.62951760e-02 3.28898430e-01
1.16501176e+00 -2.44100362e-01 1.50063306e-01 -4.79974896e-02
9.06117916e-01 -3.10493950e-02 -9.37274814e-01 1.04596473e-01
-1.37370065e-01 -4.02626812e-01 1.78986147e-01 -1.09524298e+00
-1.26965046e+00 5.62892556e-01 9.17271197e-01 6.06165230e-01
1.07512975e+00 -3.47569227e-01 7.75137305e-01 -3.63800526e-01
4.97623116e-01 -7.82608032e-01 -3.97917479e-01 -1.18627265e-01
6.45643711e-01 -1.40397382e+00 -6.55930042e-02 -4.03796345e-01
-6.00152671e-01 1.22173333e+00 1.25137120e-01 -3.85638684e-01
5.10972321e-01 2.68790483e-01 2.08338991e-01 -3.52459073e-01
1.63405135e-01 2.27626503e-01 4.60710019e-01 4.63855624e-01
6.08381808e-01 1.77406311e-01 -6.28445745e-01 3.36255014e-01
-1.30131900e-01 2.13492662e-01 3.11853409e-01 9.61453795e-01
-2.07808942e-01 -8.86759460e-01 -3.57707232e-01 4.58562851e-01
-1.07593465e+00 -7.60240629e-02 1.50097936e-01 1.30745840e+00
1.43238649e-01 6.90270185e-01 -2.79283673e-01 2.66047001e-01
5.03637433e-01 -4.09617484e-01 5.48198819e-01 -4.98384565e-01
-1.36008716e+00 5.18381417e-01 7.26803467e-02 -5.08983612e-01
-6.49791181e-01 -5.51628351e-01 -1.23488915e+00 -8.91480669e-02
-6.64121211e-01 3.56547117e-01 9.29607391e-01 2.77274162e-01
-1.26129156e-02 7.16485381e-01 9.50112641e-02 -7.44752705e-01
-3.53824377e-01 -8.96292448e-01 -5.58102727e-01 7.23383904e-01
2.64776886e-01 -3.03624839e-01 -7.83717334e-01 2.83160299e-01] | [14.083377838134766, -3.1371309757232666] |
414585cf-a088-4af9-bea6-8c7a7f036270 | lr-to-hr-face-hallucination-with-an | 2109.14690 | null | https://arxiv.org/abs/2109.14690v1 | https://arxiv.org/pdf/2109.14690v1.pdf | LR-to-HR Face Hallucination with an Adversarial Progressive Attribute-Induced Network | Face super-resolution is a challenging and highly ill-posed problem since a low-resolution (LR) face image may correspond to multiple high-resolution (HR) ones during the hallucination process and cause a dramatic identity change for the final super-resolved results. Thus, to address this problem, we propose an end-to-end progressive learning framework incorporating facial attributes and enforcing additional supervision from multi-scale discriminators. By incorporating facial attributes into the learning process and progressively resolving the facial image, the mapping between LR and HR images is constrained more, and this significantly helps to reduce the ambiguity and uncertainty in one-to-many mapping. In addition, we conduct thorough evaluations on the CelebA dataset following the settings of previous works (i.e. super-resolving by a factor of 8x from tiny 16x16 face images.), and the results demonstrate that the proposed approach can yield satisfactory face hallucination images outperforming other state-of-the-art approaches. | ['Rama Chellappa', 'Jun-Cheng Chen', 'Nitin Balachandran'] | 2021-09-29 | null | null | null | null | ['face-hallucination'] | ['computer-vision'] | [ 2.90606797e-01 1.44105032e-01 9.23675671e-02 -4.95273560e-01
-1.03839350e+00 -2.05717623e-01 2.99874991e-01 -7.31045783e-01
-1.11296527e-01 8.57621610e-01 4.44760978e-01 5.25448143e-01
-4.60610092e-02 -4.66271192e-01 -6.33474112e-01 -8.24565530e-01
2.97297984e-02 2.39965007e-01 -1.06198534e-01 -1.91064328e-01
3.36912870e-02 5.10793984e-01 -1.87125361e+00 4.54587936e-01
8.01543176e-01 9.08847034e-01 1.95855483e-01 1.88883603e-01
2.07675979e-01 6.40960515e-01 -4.16601986e-01 -3.62650871e-01
4.05228466e-01 -4.49072540e-01 -5.79818308e-01 3.97584856e-01
9.98264313e-01 -4.99654442e-01 -2.96046287e-01 1.19071722e+00
7.12006807e-01 2.32304245e-01 2.83706278e-01 -9.71100926e-01
-1.05962062e+00 1.97432756e-01 -1.17390597e+00 2.80553669e-01
5.75518310e-01 -2.71680355e-01 6.33713841e-01 -1.60179663e+00
5.48771203e-01 1.78692234e+00 4.76886004e-01 6.90464020e-01
-1.40074074e+00 -1.11884713e+00 -1.65390242e-02 2.02107370e-01
-1.87060606e+00 -1.04814076e+00 6.73708916e-01 -9.97717157e-02
2.64334768e-01 2.03918085e-01 9.51566175e-02 1.19174647e+00
-2.40959767e-02 2.51819849e-01 1.43851829e+00 -1.45266633e-02
1.64465662e-02 2.12487906e-01 -6.08473420e-01 6.71759486e-01
6.80414066e-02 8.37023631e-02 -7.33708322e-01 -2.28195116e-01
1.15820849e+00 -4.14253287e-02 -3.68258208e-01 -1.75385281e-01
-8.21197748e-01 6.36413872e-01 4.41463262e-01 -6.42698407e-02
-4.12554324e-01 -3.73324573e-01 -3.28078941e-02 1.84302494e-01
6.30617976e-01 2.84907818e-01 -1.12973198e-01 3.85977566e-01
-1.04060805e+00 3.25386375e-01 2.13647589e-01 7.05339134e-01
8.30087185e-01 1.46574274e-01 -2.54637718e-01 1.25339127e+00
1.11501701e-01 3.88905913e-01 2.17112049e-01 -1.27408040e+00
3.75454575e-01 1.13191873e-01 1.66135848e-01 -8.72380674e-01
-1.14743397e-01 -4.08296168e-01 -9.73063231e-01 3.55430871e-01
1.18144758e-01 -7.60270432e-02 -1.00411999e+00 1.94593775e+00
5.69403946e-01 5.96005142e-01 1.38123691e-01 1.30085993e+00
9.25206900e-01 5.43961406e-01 -2.45086905e-02 -7.35486984e-01
1.45698237e+00 -8.12870741e-01 -1.03384209e+00 -2.16683745e-01
-4.52620387e-01 -9.65190232e-01 1.03981936e+00 3.17006588e-01
-1.36145735e+00 -9.59288657e-01 -9.66069937e-01 -2.35951245e-01
1.85598359e-01 1.57024413e-01 2.94855654e-01 2.58501798e-01
-1.14308548e+00 3.96801531e-01 -4.26026642e-01 -1.74657837e-01
6.36286974e-01 2.54212439e-01 -7.12056577e-01 -4.07561183e-01
-1.11752224e+00 8.32334459e-01 1.17642932e-01 1.71398908e-01
-9.41905320e-01 -7.66297996e-01 -8.04038227e-01 -4.02250960e-02
5.19093275e-01 -5.15068710e-01 8.54372561e-01 -8.00986826e-01
-1.33987200e+00 9.07818496e-01 -5.43016970e-01 1.35373563e-01
5.20413220e-01 -3.57837886e-01 -7.07738996e-01 2.93484509e-01
3.48421097e-01 6.99970782e-01 1.28443861e+00 -1.64159334e+00
-4.90488827e-01 -7.04801559e-01 -1.62810802e-01 5.31825125e-01
-2.56524354e-01 3.58265251e-01 -4.99901146e-01 -7.37978101e-01
3.00898969e-01 -8.21759284e-01 1.49986874e-02 1.45355389e-01
-9.16974843e-02 7.85918385e-02 9.35277104e-01 -9.08653915e-01
7.97234535e-01 -2.40928483e+00 3.32801700e-01 -2.24634007e-01
3.21957409e-01 1.77971318e-01 -3.22055042e-01 -2.30244473e-01
-3.51041645e-01 -2.86924243e-02 -1.64201215e-01 -6.23754680e-01
-4.03889567e-01 7.42478222e-02 -3.83982956e-01 5.54677486e-01
4.42639261e-01 5.54209709e-01 -5.98334432e-01 -6.07006609e-01
-8.16498175e-02 1.04134393e+00 -3.40118080e-01 4.85520542e-01
2.97744900e-01 8.11728716e-01 -1.57546461e-01 8.46314847e-01
1.27721727e+00 -2.96432108e-01 5.58596291e-02 -4.76347744e-01
5.30584902e-03 -1.25527099e-01 -1.20918906e+00 1.64507842e+00
-3.43643576e-01 3.88008684e-01 4.27671939e-01 -3.84937763e-01
1.07598281e+00 3.22868377e-01 4.06284422e-01 -8.49743605e-01
-2.15222329e-01 -1.74598899e-02 -3.16970438e-01 -3.81303698e-01
5.17853856e-01 -3.91934365e-01 2.46710435e-01 2.49597162e-01
3.55969481e-02 3.20042312e-01 -3.01001787e-01 -1.56563967e-01
4.67402875e-01 1.10287510e-01 1.68682918e-01 -5.27198426e-02
5.72880507e-01 -6.43123090e-01 9.38201010e-01 2.47681081e-01
-1.65001109e-01 1.07056236e+00 2.88220674e-01 -1.75053760e-01
-1.04386544e+00 -1.21526837e+00 -3.85892242e-01 1.17997313e+00
2.34933048e-01 -1.35887206e-01 -8.60935330e-01 -2.70382434e-01
-3.52119505e-02 3.11961830e-01 -7.12723792e-01 -7.63422996e-02
-5.98311782e-01 -7.86030412e-01 3.98612112e-01 5.22953808e-01
7.35574543e-01 -9.74710166e-01 -1.11272454e-01 -1.01654492e-01
-4.47764695e-01 -1.43502045e+00 -6.12966418e-01 -4.50794846e-01
-6.07744992e-01 -6.62176847e-01 -7.95956731e-01 -7.71149576e-01
7.67193556e-01 6.28238559e-01 9.22739983e-01 -1.82291925e-01
-3.13794136e-01 -1.06068663e-01 -6.96523115e-02 1.56535238e-01
-1.85563520e-01 -4.15743828e-01 3.62726986e-01 3.48058343e-01
4.65357304e-02 -6.51570261e-01 -6.26089752e-01 4.13061023e-01
-7.68900990e-01 -9.56305489e-03 7.15436161e-01 8.38088989e-01
7.90792465e-01 2.80374944e-01 7.02683568e-01 -7.52603531e-01
5.01603782e-01 -4.56335396e-01 -4.39873666e-01 2.69541353e-01
-5.66090703e-01 -2.83203013e-02 5.99476874e-01 -6.11604810e-01
-1.53108633e+00 3.20968553e-02 7.43380785e-02 -9.13654983e-01
2.55427603e-02 9.45879444e-02 -4.16113555e-01 -2.87740916e-01
5.89524388e-01 1.57630235e-01 2.50265300e-01 -6.22003496e-01
2.71347195e-01 6.22746587e-01 8.65256667e-01 -3.50421429e-01
1.02066827e+00 6.35097623e-01 2.48428881e-02 -7.02001154e-01
-1.03267324e+00 -9.50569063e-02 -6.77200615e-01 -1.50122521e-02
8.00804496e-01 -1.47200191e+00 -5.75709164e-01 1.88134000e-01
-9.00429189e-01 2.40857542e-01 -1.00444229e-02 3.16579580e-01
-3.97705525e-01 2.29633346e-01 -6.70069754e-01 -8.02355886e-01
-1.06629744e-01 -1.04516077e+00 1.24399734e+00 5.53274035e-01
-5.25175035e-03 -4.00840342e-01 -2.16100231e-01 7.57846892e-01
4.35152113e-01 2.56794184e-01 5.18713415e-01 -4.63467129e-02
-5.52795351e-01 2.63962328e-01 -6.69776320e-01 2.38803580e-01
3.65826458e-01 -3.46560478e-01 -1.22072625e+00 -7.33137727e-01
2.02157438e-01 -4.98204082e-01 8.12672019e-01 -4.53196466e-02
1.11383176e+00 -4.53260541e-01 4.84488532e-02 8.34451079e-01
1.32610190e+00 -1.80257782e-01 8.95577490e-01 -2.17740685e-02
9.00275528e-01 7.77557194e-01 1.00994503e+00 5.86256862e-01
4.58138615e-01 8.90059054e-01 3.77610952e-01 -4.24193591e-01
-4.79220808e-01 -3.53199899e-01 3.85685414e-01 1.80060580e-01
-2.67892510e-01 3.40167224e-01 -2.46849775e-01 2.34521851e-01
-1.65831614e+00 -1.14822710e+00 3.47753793e-01 2.25414157e+00
1.02899897e+00 -4.30092841e-01 2.72178818e-02 -2.19133332e-01
9.46439207e-01 4.59710181e-01 -8.18827450e-01 1.35627210e-01
-2.92791516e-01 2.63354123e-01 8.25785324e-02 5.09063780e-01
-9.94199097e-01 1.17834294e+00 6.70974207e+00 7.29808271e-01
-1.31995010e+00 1.92675367e-01 8.70122373e-01 -2.48833269e-01
-1.36530161e-01 -3.56038094e-01 -8.50885391e-01 3.33585620e-01
5.70991397e-01 -8.54982883e-02 8.96863818e-01 6.59803987e-01
1.94629580e-01 9.04055536e-02 -8.94625306e-01 1.31475663e+00
4.78647560e-01 -9.39490557e-01 1.36157155e-01 2.08000869e-01
8.32010090e-01 -4.31701660e-01 5.60453296e-01 1.98655844e-01
-4.80552725e-02 -1.53816760e+00 5.10265231e-01 5.31319618e-01
1.33353007e+00 -9.06289458e-01 4.88670170e-01 -9.34489444e-02
-1.27089977e+00 -2.17994004e-01 -7.77270138e-01 3.21867049e-01
-3.83085683e-02 3.60424131e-01 -5.65712571e-01 5.12906015e-01
1.00962305e+00 4.36466575e-01 -6.32355630e-01 4.63233411e-01
-5.76894581e-02 6.93764538e-02 1.03960961e-01 9.26185966e-01
-3.86772126e-01 -1.32577181e-01 5.15405178e-01 7.53627062e-01
4.63108420e-01 5.14439464e-01 1.92419246e-01 1.12256038e+00
-2.66235709e-01 -1.37431756e-01 -2.22678453e-01 3.65374804e-01
6.43171012e-01 1.50051498e+00 -2.23753780e-01 -2.61113755e-02
-4.04128343e-01 1.24390364e+00 3.55547786e-01 5.40663481e-01
-9.69202220e-01 8.79685432e-02 9.46637273e-01 3.28784853e-01
4.01717871e-01 6.64888620e-02 -5.28926216e-02 -1.15101171e+00
1.45071015e-01 -1.08541059e+00 3.57304066e-01 -9.19530749e-01
-1.32109606e+00 9.19644237e-01 -1.13228306e-01 -1.11947751e+00
-2.68236876e-01 -1.69793308e-01 -1.64497375e-01 1.23451316e+00
-1.78862071e+00 -1.28093970e+00 -6.40222907e-01 8.30924273e-01
5.08466065e-01 -3.00197631e-01 7.41458952e-01 3.95779371e-01
-4.87375796e-01 9.33540404e-01 -2.73097605e-01 -8.67298916e-02
1.15799356e+00 -7.40287066e-01 1.05506331e-01 8.40422213e-01
-9.67096165e-02 6.16496384e-01 7.50437617e-01 -6.33841813e-01
-1.28184080e+00 -1.33624721e+00 5.93135417e-01 -2.64829099e-01
2.28123382e-01 -3.31938595e-01 -1.31447017e+00 4.95618463e-01
6.63711578e-02 5.72701633e-01 5.58545113e-01 8.47334042e-02
-7.32472718e-01 -3.94171923e-01 -1.55843234e+00 5.52067518e-01
1.21646011e+00 -7.71160066e-01 -3.42228413e-01 -7.11788088e-02
6.09594405e-01 -4.83059645e-01 -9.75031018e-01 6.89619243e-01
6.34845257e-01 -1.17382908e+00 1.29408419e+00 -3.93164784e-01
4.97200519e-01 -5.03092527e-01 -2.98881859e-01 -1.12339282e+00
-7.60210633e-01 -6.17531061e-01 -2.11014926e-01 1.33521080e+00
1.94482859e-02 -4.74661976e-01 5.72841108e-01 4.72590953e-01
2.95629829e-01 -6.37755215e-01 -9.43382084e-01 -4.93302256e-01
-1.92195639e-01 5.23645163e-01 7.08530724e-01 1.12994730e+00
-3.69448572e-01 3.70530754e-01 -8.55699241e-01 6.68578207e-01
1.13438511e+00 2.00772673e-01 5.12708366e-01 -1.20030808e+00
-1.24474689e-01 -1.56844575e-02 -1.54661223e-01 -7.90354133e-01
3.94029617e-01 -4.12175715e-01 -9.72851925e-03 -9.65708852e-01
5.96375167e-01 -2.17489213e-01 -3.73852402e-01 4.48224694e-01
-5.14749646e-01 8.10347736e-01 5.06366491e-02 3.58372211e-01
-5.59781253e-01 6.62034214e-01 1.51865327e+00 2.57558346e-01
-6.14969619e-02 -2.50258386e-01 -1.17080283e+00 4.54262346e-01
5.49842179e-01 -1.86236426e-01 -5.08182466e-01 -3.12211871e-01
-2.35425621e-01 5.04821599e-01 4.31610703e-01 -7.87832856e-01
1.64515898e-01 -2.31446356e-01 9.48376536e-01 -4.12768781e-01
7.92638063e-01 -5.72244644e-01 2.79097378e-01 -1.18416689e-01
-3.66388172e-01 -5.42128272e-02 1.56212017e-01 4.65718627e-01
-3.42916161e-01 4.30061907e-01 1.37160015e+00 -9.47110131e-02
-4.88477081e-01 5.59118629e-01 2.78671503e-01 -4.98863868e-02
9.54847217e-01 -2.51871169e-01 -3.83866340e-01 -4.99468148e-01
-7.95479894e-01 5.49460351e-02 6.30674422e-01 7.17866957e-01
9.35083926e-01 -1.48852968e+00 -1.12678778e+00 5.35944223e-01
-8.42232350e-03 -1.36207370e-02 5.57001591e-01 5.86878181e-01
9.43578556e-02 1.74803752e-02 -6.68264270e-01 -4.32310194e-01
-1.61886132e+00 5.67259729e-01 1.92795858e-01 8.23557079e-02
-5.46866715e-01 8.64166439e-01 4.57044572e-01 -2.38542408e-01
1.42174870e-01 5.02712548e-01 -5.09113967e-01 9.02855843e-02
1.00060475e+00 3.83915812e-01 -2.32398584e-01 -1.17097366e+00
-3.21524233e-01 7.46347010e-01 -4.93319452e-01 -2.51745433e-01
1.38991332e+00 -4.99302864e-01 -2.42000654e-01 1.80493936e-01
1.19529080e+00 2.38306984e-01 -1.59044683e+00 -4.43319827e-01
-3.92378390e-01 -9.88366425e-01 2.36648950e-03 -7.29606509e-01
-1.26949954e+00 5.36364973e-01 8.26289237e-01 -3.41017812e-01
1.41925538e+00 2.04677209e-01 6.03105128e-01 -1.52448475e-01
4.42681760e-01 -9.23071742e-01 4.52883810e-01 1.77243516e-01
1.27731216e+00 -1.43120182e+00 2.53903329e-01 -8.39930832e-01
-8.22585046e-01 7.86509812e-01 9.25844729e-01 -2.08296757e-02
2.83658564e-01 2.21230716e-01 1.27129033e-01 -2.63128746e-02
-7.37647533e-01 -4.39424887e-02 2.85731643e-01 5.15227199e-01
3.62992257e-01 -8.31693560e-02 1.06579915e-01 4.25383836e-01
-2.16160402e-01 -1.36523232e-01 3.59558553e-01 3.19106549e-01
-3.48253012e-01 -8.67287278e-01 -7.90776491e-01 3.64454463e-02
-5.44916630e-01 -3.04180644e-02 -1.50512040e-01 5.05078673e-01
1.38733223e-01 9.89143252e-01 1.13208525e-01 -4.11056429e-01
1.91877171e-01 -2.46071398e-01 6.54108465e-01 -6.04850829e-01
1.72050798e-03 3.33557367e-01 -2.27375194e-01 -7.46646643e-01
-3.80232483e-01 -6.23427749e-01 -1.20892167e+00 -3.85796934e-01
-1.99746471e-02 -1.03847995e-01 1.56296045e-01 5.17033577e-01
5.68752706e-01 4.08840209e-01 1.12174726e+00 -1.03096032e+00
-6.70571566e-01 -9.49216902e-01 -7.75173604e-01 5.86768866e-01
6.10938370e-01 -8.34733784e-01 -4.63017702e-01 2.65257861e-02] | [12.833574295043945, -0.08613074570894241] |
7334556e-71c0-458d-b677-ac1e25cebfa6 | fast-online-clustering-with-randomized | 1506.03425 | null | http://arxiv.org/abs/1506.03425v1 | http://arxiv.org/pdf/1506.03425v1.pdf | Fast Online Clustering with Randomized Skeleton Sets | We present a new fast online clustering algorithm that reliably recovers
arbitrary-shaped data clusters in high throughout data streams. Unlike the
existing state-of-the-art online clustering methods based on k-means or
k-medoid, it does not make any restrictive generative assumptions. In addition,
in contrast to existing nonparametric clustering techniques such as DBScan or
DenStream, it gives provable theoretical guarantees. To achieve fast
clustering, we propose to represent each cluster by a skeleton set which is
updated continuously as new data is seen. A skeleton set consists of weighted
samples from the data where weights encode local densities. The size of each
skeleton set is adapted according to the cluster geometry. The proposed
technique automatically detects the number of clusters and is robust to
outliers. The algorithm works for the infinite data stream where more than one
pass over the data is not feasible. We provide theoretical guarantees on the
quality of the clustering and also demonstrate its advantage over the existing
state-of-the-art on several datasets. | ['Xiaofeng Liu', 'Krzysztof Choromanski', 'Sanjiv Kumar'] | 2015-06-10 | null | null | null | null | ['online-clustering'] | ['computer-vision'] | [-3.51134866e-01 -1.87858954e-01 1.18086375e-02 -3.79735649e-01
-7.47528315e-01 -5.38246870e-01 2.29470983e-01 6.94947362e-01
-3.65740806e-01 2.45472461e-01 -1.14390813e-01 -4.11881357e-02
-5.16860604e-01 -9.25151825e-01 -6.74132109e-01 -9.86834288e-01
-6.46234155e-01 1.23470759e+00 7.12261856e-01 4.04308379e-01
2.18301997e-01 6.69995368e-01 -1.54621708e+00 1.62959937e-02
7.72555113e-01 1.16788399e+00 -8.89155120e-02 8.29749823e-01
-2.94440925e-01 1.65264606e-01 -6.65290952e-01 -4.11277153e-02
3.71650428e-01 -4.69708383e-01 -3.03883642e-01 5.49425244e-01
-2.49054655e-01 -1.03923030e-01 -1.69361562e-01 8.09414089e-01
5.12127042e-01 2.78546095e-01 8.94881129e-01 -1.44338775e+00
-1.87050939e-01 6.89392924e-01 -8.91004086e-01 2.47333512e-01
1.56814530e-01 -3.22036892e-01 5.11501729e-01 -6.89373374e-01
2.38011435e-01 1.06941497e+00 6.98019981e-01 5.41758984e-02
-1.23998153e+00 -5.72361708e-01 8.16342011e-02 1.13332964e-01
-1.83143306e+00 -3.73180658e-01 4.59974021e-01 -3.83049786e-01
2.76531398e-01 4.90887351e-02 6.03058100e-01 2.58439600e-01
-2.56700933e-01 6.29883289e-01 4.74943876e-01 -3.89924675e-01
1.02950120e+00 -2.74143040e-01 -1.28617898e-01 3.37014288e-01
3.91536236e-01 -2.17236191e-01 -4.74309623e-01 -6.48060024e-01
6.77312672e-01 3.40916008e-01 -1.51891649e-01 -8.07884753e-01
-1.00551748e+00 7.02201366e-01 5.08767217e-02 9.73047912e-02
-5.11079073e-01 2.21916258e-01 3.83087933e-01 -1.87793802e-02
3.77432376e-01 -4.12033528e-01 -1.46790996e-01 -3.86410892e-01
-1.56405020e+00 -5.72513156e-02 8.74351442e-01 1.13346767e+00
7.63989806e-01 -8.61444920e-02 3.35956693e-01 6.76366568e-01
3.52767467e-01 4.79373246e-01 4.17371333e-01 -1.07928669e+00
1.72446072e-01 5.53951681e-01 1.54719815e-01 -9.19447124e-01
-2.12552294e-01 -3.47724408e-02 -1.07212472e+00 -6.98594972e-02
4.48145568e-01 -9.74749699e-02 -7.09606469e-01 1.20105362e+00
7.65026510e-01 6.04953766e-01 -7.34988973e-02 5.15961170e-01
6.63527623e-02 6.87029779e-01 -4.27906513e-01 -6.41463459e-01
7.32927144e-01 -2.23040953e-01 -6.55397892e-01 4.89638388e-01
2.55338669e-01 -5.42151928e-01 4.39105392e-01 8.15433145e-01
-1.17092216e+00 -3.93896550e-01 -7.38693476e-01 6.45083129e-01
-1.16112396e-01 -3.71644408e-01 3.68377388e-01 9.64843750e-01
-1.04570675e+00 5.08010805e-01 -1.40649652e+00 -5.62431455e-01
4.21724975e-01 3.70460212e-01 -4.05443236e-02 -1.25773013e-01
-3.12294066e-01 2.65429430e-02 7.24521458e-01 -2.15469807e-01
-7.81681836e-01 -5.70069194e-01 -5.75638354e-01 5.12289666e-02
3.67893159e-01 -4.29780632e-01 8.77886176e-01 -6.58958673e-01
-1.36471033e+00 4.52991426e-01 -4.54189271e-01 -6.33751988e-01
7.56935894e-01 2.06585884e-01 -2.50725806e-01 5.54641187e-01
-8.16624016e-02 3.50467086e-01 8.84958029e-01 -1.54313684e+00
-9.01845098e-01 -4.65070188e-01 -7.94977248e-01 2.69025419e-04
-3.14018369e-01 -7.72646368e-02 -8.35916817e-01 -4.63820785e-01
5.06594002e-01 -6.95278525e-01 -3.67502064e-01 5.64700440e-02
-2.30397180e-01 -2.10951224e-01 1.10733342e+00 -1.38241485e-01
1.46173453e+00 -2.53447247e+00 -8.49567354e-02 8.52199316e-01
1.88347146e-01 -2.30774954e-01 5.35228431e-01 6.00492954e-01
3.34868997e-01 -3.94777395e-02 -5.85146904e-01 -4.90857869e-01
1.48963422e-01 5.44783056e-01 -1.96189448e-01 1.06148911e+00
-5.14164269e-01 1.43989399e-01 -9.61854696e-01 -6.80516899e-01
2.46526539e-01 3.01960498e-01 -5.65849423e-01 1.87878206e-01
-3.53430235e-03 1.29248947e-01 -1.26916409e-01 4.33329195e-01
9.82244909e-01 -1.96650907e-01 2.81870693e-01 3.27577323e-01
5.31913899e-02 -4.18885469e-01 -1.71633804e+00 1.46702993e+00
-2.28692610e-02 1.34191424e-01 3.23004276e-01 -1.17068577e+00
1.09186542e+00 4.09528196e-01 1.00961900e+00 6.62337989e-02
-3.43039818e-02 2.24891394e-01 -2.77232200e-01 -9.94317979e-02
3.44985873e-01 -1.24311164e-01 -1.68194741e-01 6.21467352e-01
-1.15653276e-01 1.51885256e-01 3.28470826e-01 3.54711175e-01
1.20618987e+00 -3.92583907e-01 1.99610457e-01 -3.63586277e-01
1.96250990e-01 -2.03320503e-01 6.21578097e-01 7.94480562e-01
-9.72003117e-02 7.35663831e-01 2.10376427e-01 -1.96788654e-01
-1.00388944e+00 -1.53831041e+00 -1.32231802e-01 9.12342310e-01
1.95833325e-01 -4.27161753e-01 -8.93980026e-01 -3.45688999e-01
3.06853712e-01 4.87355381e-01 -5.40791452e-01 7.69860223e-02
-1.33955702e-01 -8.28113198e-01 4.33630288e-01 4.33785051e-01
3.21812212e-01 -6.22195601e-01 -7.50363648e-01 4.24862862e-01
-8.55197608e-02 -9.35362101e-01 -5.32042027e-01 5.97304888e-02
-1.23200905e+00 -1.23325443e+00 -4.98881072e-01 -5.59122682e-01
9.04985011e-01 3.04088712e-01 8.31666470e-01 -5.19573353e-02
-1.20473042e-01 7.04309344e-01 -6.07436180e-01 -3.09408605e-01
-4.46724266e-01 -2.44162008e-01 3.02533746e-01 4.74248201e-01
5.90950489e-01 -8.22493732e-01 -6.44344687e-01 5.05580485e-01
-1.03977346e+00 -6.50892735e-01 1.27117276e-01 2.78686076e-01
9.42893624e-01 9.85343814e-01 6.10764682e-01 -7.39266813e-01
3.53529990e-01 -8.98317993e-01 -6.54778242e-01 -3.76196839e-02
-5.43726265e-01 -1.66492179e-01 8.50859165e-01 -3.64720345e-01
-7.14881837e-01 2.32796013e-01 3.37295681e-01 -7.40670323e-01
-4.76235896e-01 1.75164580e-01 -6.27947375e-02 3.97550672e-01
5.05711257e-01 3.21590155e-01 2.39206087e-02 -5.28645277e-01
5.03642559e-01 9.14083004e-01 8.41738224e-01 -6.36995792e-01
8.65918696e-01 1.15848684e+00 -8.22592750e-02 -1.12070334e+00
-2.94313908e-01 -1.17064798e+00 -1.17324924e+00 -3.53367358e-01
6.44159853e-01 -6.46356225e-01 -9.41827536e-01 3.62369746e-01
-6.23453856e-01 -3.74878973e-01 -4.95013416e-01 3.59111995e-01
-8.27072084e-01 6.68366253e-01 -2.71379381e-01 -1.28378975e+00
-7.03065172e-02 -4.96460527e-01 8.10576797e-01 -1.11933820e-01
-7.25051165e-02 -1.11113763e+00 8.73688012e-02 -2.99741328e-01
8.20878148e-02 6.78101778e-01 5.16732335e-01 -7.31797755e-01
-3.41866404e-01 -2.30855972e-01 1.70725927e-01 -2.57467926e-02
1.77465916e-01 4.79017973e-01 -4.17032957e-01 -6.17161930e-01
-5.24553470e-02 1.22583978e-01 6.47399604e-01 6.36414111e-01
1.40527380e+00 -3.89455557e-01 -5.54706812e-01 5.66867113e-01
1.46037793e+00 3.98158938e-01 5.03352046e-01 1.56446155e-02
4.37243521e-01 3.92795652e-01 5.44515014e-01 1.19285762e+00
5.40145218e-01 3.81471306e-01 4.08928752e-01 1.35867164e-01
3.80929053e-01 -1.15230061e-01 2.33204529e-01 9.01615381e-01
3.04454684e-01 -1.70010701e-01 -1.04001319e+00 1.00362897e+00
-2.12790108e+00 -1.04113126e+00 -2.71495372e-01 2.84527969e+00
6.40237749e-01 1.64666951e-01 7.58549571e-01 6.59144640e-01
9.50972974e-01 -3.28578025e-01 -5.64070523e-01 -1.86943963e-01
2.02182040e-01 1.15650348e-01 6.69878900e-01 3.16258311e-01
-8.31356883e-01 5.02312064e-01 6.55505657e+00 8.00581157e-01
-4.72879797e-01 -2.25219205e-02 4.94905144e-01 -2.39744782e-01
-8.26417655e-03 -5.75874560e-02 -5.26228368e-01 7.74389982e-01
1.10419893e+00 -4.08379436e-01 3.89359266e-01 9.18707430e-01
4.47717309e-01 -4.77845669e-01 -1.10130584e+00 1.12024009e+00
1.95876863e-02 -1.00574207e+00 -2.71819919e-01 2.65104145e-01
7.08450258e-01 -9.70447212e-02 -1.67148352e-01 -2.35161975e-01
8.19187105e-01 -6.83880568e-01 5.81701398e-01 6.56347513e-01
7.62606323e-01 -1.40567029e+00 4.75676447e-01 6.46121383e-01
-1.33181202e+00 -2.36936659e-01 -5.41557789e-01 2.05563456e-01
3.43543380e-01 1.18410051e+00 -7.75156617e-01 5.49331307e-01
1.04150891e+00 4.65801477e-01 -3.49839658e-01 1.49919963e+00
3.71919036e-01 9.50811982e-01 -1.11620855e+00 3.17123264e-01
-5.92973502e-03 -3.92376155e-01 2.76734591e-01 1.33170235e+00
7.14179635e-01 1.01989120e-01 6.45652115e-01 3.91563982e-01
5.90908267e-02 1.20408386e-01 -4.25877392e-01 4.23523843e-01
1.15514982e+00 9.35923457e-01 -1.22805274e+00 -5.97952783e-01
-1.00996159e-01 1.00929475e+00 5.95857331e-04 1.62229836e-01
-5.37436545e-01 -4.33139563e-01 2.98703730e-01 5.42349100e-01
8.06940675e-01 -4.78973418e-01 -1.59185275e-01 -7.19215930e-01
-1.86979964e-01 -3.74236494e-01 8.76045406e-01 -2.39965394e-01
-1.51468468e+00 4.38332856e-02 2.12303445e-01 -1.22568130e+00
-2.80905783e-01 6.18884386e-03 -5.98320186e-01 2.00602561e-01
-8.78090024e-01 -4.34583217e-01 -3.87338430e-01 9.36984062e-01
2.46097133e-01 -1.90032590e-02 7.00535834e-01 2.34197944e-01
-2.51442075e-01 4.33745742e-01 9.23373997e-01 9.46197137e-02
6.26750529e-01 -1.41026270e+00 1.51699975e-01 8.42800319e-01
4.15320881e-02 3.59875858e-01 7.69760370e-01 -6.00950480e-01
-1.03034651e+00 -1.15585577e+00 2.91956216e-01 -1.06144838e-01
6.61007762e-01 -4.87700373e-01 -1.01675451e+00 4.04275417e-01
-1.88042998e-01 2.92996049e-01 1.07011616e+00 -9.26714018e-02
-1.58692420e-01 -4.00488943e-01 -1.44345260e+00 -2.74180286e-02
7.76129484e-01 6.49189577e-02 -2.27282554e-01 2.06714794e-01
2.49465793e-01 -5.49216755e-02 -1.11134875e+00 6.12310767e-02
1.87752500e-01 -1.28325725e+00 6.81796491e-01 -9.09486786e-02
-3.19621265e-01 -6.97682500e-01 -2.60208964e-01 -1.12411749e+00
-3.59289378e-01 -9.59638476e-01 -5.45139432e-01 1.28008783e+00
-6.51948079e-02 -4.68801796e-01 8.82361889e-01 2.09263399e-01
7.07256123e-02 -3.78301531e-01 -1.30782008e+00 -1.07138467e+00
-9.90210846e-02 -5.34450948e-01 5.97888649e-01 8.46155465e-01
2.81160861e-01 -3.16748381e-01 -3.72362919e-02 4.42699254e-01
1.27221644e+00 7.19995797e-02 1.07775366e+00 -1.62800395e+00
-4.50247407e-01 -2.71875173e-01 -5.33332467e-01 -1.05246651e+00
-1.59088656e-01 -5.17980993e-01 4.47610505e-02 -1.43704557e+00
1.29721507e-01 -7.72587419e-01 -1.15050443e-01 1.65211499e-01
1.41903847e-01 1.33546054e-01 -6.71556666e-02 4.91622090e-01
-1.08292437e+00 4.66925800e-01 3.35483521e-01 4.44784641e-01
-4.37299073e-01 2.75588125e-01 -3.15277725e-01 7.08278060e-01
8.06610465e-01 -6.57536626e-01 -4.23634648e-01 8.36631879e-02
-1.67673796e-01 1.51392743e-01 -1.05183609e-01 -1.16044438e+00
5.80416620e-01 -1.29734397e-01 3.01576823e-01 -1.06463933e+00
2.24807374e-02 -1.14648664e+00 3.60187858e-01 2.68000156e-01
1.27253562e-01 1.30797699e-01 -3.01215127e-02 1.18485093e+00
-9.87520069e-02 6.87338263e-02 1.05293310e+00 -2.27757916e-02
-8.67807567e-02 5.41694999e-01 -5.72908938e-01 1.17483668e-01
1.48857200e+00 -5.97974122e-01 3.18977118e-01 -6.38785839e-01
-8.74096096e-01 4.86131936e-01 7.86452413e-01 -6.28307788e-03
6.07286096e-01 -1.35660863e+00 -6.22713566e-01 7.56431445e-02
-4.56362329e-02 5.93166709e-01 9.99925286e-02 7.85983086e-01
-6.16391778e-01 -7.55326375e-02 2.61240304e-01 -9.13374305e-01
-1.09574044e+00 1.00973344e+00 1.41449928e-01 1.29641518e-01
-7.64129698e-01 3.87525141e-01 -1.45834610e-01 -1.51639998e-01
5.19057989e-01 8.39975476e-03 2.21132472e-01 7.45263770e-02
5.66743910e-01 8.00325155e-01 -4.41831946e-02 -4.92615610e-01
-3.70399147e-01 4.00187612e-01 1.78099647e-01 -3.65300059e-01
1.45035672e+00 -4.57605898e-01 -1.25820443e-01 8.52100670e-01
9.53386486e-01 -5.96002005e-02 -1.47588778e+00 -3.09233010e-01
2.12984264e-01 -7.03878105e-01 -3.08847636e-01 -1.96175039e-01
-1.01453257e+00 6.65646136e-01 5.60877562e-01 6.66931629e-01
1.23700678e+00 8.75263140e-02 6.41754210e-01 1.43348604e-01
6.16115570e-01 -1.34342885e+00 3.36163566e-02 1.14437558e-01
1.91468850e-01 -8.57660353e-01 6.14814200e-02 -2.15919018e-01
-4.28984046e-01 1.15582192e+00 -4.83724894e-03 -3.38274986e-01
1.09783685e+00 5.83465099e-01 -1.91317707e-01 -1.06033929e-01
-5.67138016e-01 -9.33310017e-02 -2.63183832e-01 8.76292706e-01
-4.06373963e-02 1.77330315e-01 8.08893293e-02 3.28500837e-01
-3.18061978e-01 -1.06115274e-01 7.43481398e-01 8.37106466e-01
-9.15691972e-01 -8.44444513e-01 -9.14216876e-01 5.52711785e-01
-4.78951633e-01 3.83181214e-01 -1.52169272e-01 5.19345641e-01
-2.90977079e-02 1.19743252e+00 5.77031374e-01 -4.39379700e-02
1.18009679e-01 1.17948994e-01 1.92214146e-01 -4.41610932e-01
-1.01261497e-01 5.41910410e-01 -5.47051549e-01 -3.77203345e-01
-3.75649899e-01 -1.00838983e+00 -1.71527100e+00 -5.91512203e-01
-2.50279456e-01 5.43516695e-01 5.69671094e-01 6.17797852e-01
4.15137380e-01 8.25961977e-02 9.37095523e-01 -8.40052843e-01
-2.89278686e-01 -6.47164464e-01 -9.82557118e-01 3.16816390e-01
1.70243293e-01 -5.95762193e-01 -5.39157152e-01 4.92006809e-01] | [7.309330940246582, 4.419173717498779] |
8a0c251c-02ef-4d26-a4ac-7dc7b0cedf70 | naist-at-the-hoo-2012-shared-task | null | null | https://aclanthology.org/W12-2033 | https://aclanthology.org/W12-2033.pdf | NAIST at the HOO 2012 Shared Task | null | ['Yuji Matsumoto', 'Yuta Hayashibe', 'Tomoya Mizumoto', 'Lis Kanashiro', 'Shuhei Kondo', 'Mamoru Komachi', 'Keisuke Sakaguchi'] | 2012-06-01 | null | null | null | ws-2012-6 | ['grammatical-error-detection'] | ['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] | [-7.231553077697754, 3.8342316150665283] |
942a8472-a698-43fc-81dc-2a7e65b340be | neurocs-neural-nocs-supervision-for-monocular-1 | 2305.17763 | null | https://arxiv.org/abs/2305.17763v1 | https://arxiv.org/pdf/2305.17763v1.pdf | NeurOCS: Neural NOCS Supervision for Monocular 3D Object Localization | Monocular 3D object localization in driving scenes is a crucial task, but challenging due to its ill-posed nature. Estimating 3D coordinates for each pixel on the object surface holds great potential as it provides dense 2D-3D geometric constraints for the underlying PnP problem. However, high-quality ground truth supervision is not available in driving scenes due to sparsity and various artifacts of Lidar data, as well as the practical infeasibility of collecting per-instance CAD models. In this work, we present NeurOCS, a framework that uses instance masks and 3D boxes as input to learn 3D object shapes by means of differentiable rendering, which further serves as supervision for learning dense object coordinates. Our approach rests on insights in learning a category-level shape prior directly from real driving scenes, while properly handling single-view ambiguities. Furthermore, we study and make critical design choices to learn object coordinates more effectively from an object-centric view. Altogether, our framework leads to new state-of-the-art in monocular 3D localization that ranks 1st on the KITTI-Object benchmark among published monocular methods. | ['Manmohan Chandraker', 'Enrique Dunn', 'Buyu Liu', 'Samuel Schulter', 'Bingbing Zhuang', 'Zhixiang Min'] | 2023-05-28 | neurocs-neural-nocs-supervision-for-monocular | http://openaccess.thecvf.com//content/CVPR2023/html/Min_NeurOCS_Neural_NOCS_Supervision_for_Monocular_3D_Object_Localization_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Min_NeurOCS_Neural_NOCS_Supervision_for_Monocular_3D_Object_Localization_CVPR_2023_paper.pdf | cvpr-2023-1 | ['monocular-3d-object-localization', 'object-localization'] | ['computer-vision', 'computer-vision'] | [ 1.30162016e-01 -4.87890206e-02 -6.04505576e-02 -4.08512473e-01
-8.06539714e-01 -7.65748858e-01 6.72957957e-01 -3.41371417e-01
-2.04475373e-01 5.52830100e-01 -2.30706736e-01 -2.22037479e-01
-1.43387690e-01 -5.05535185e-01 -1.16042125e+00 -5.81274152e-01
2.73947537e-01 7.53093004e-01 1.79945931e-01 1.29336417e-01
3.80591840e-01 9.37060595e-01 -1.73628771e+00 -5.72558399e-03
8.14453781e-01 1.21023428e+00 4.72520620e-01 3.81028533e-01
-4.40909304e-02 4.32948589e-01 -1.96525887e-01 -1.54269129e-01
6.95459962e-01 1.44084483e-01 -2.54269570e-01 3.39702219e-01
1.09914696e+00 -3.54791343e-01 -3.03126633e-01 1.15053022e+00
2.38068298e-01 -8.26657638e-02 9.04180527e-01 -1.33739638e+00
-6.55243278e-01 -2.51703709e-01 -6.15750730e-01 -9.19029936e-02
1.36446923e-01 3.47773343e-01 9.90386546e-01 -1.31881881e+00
6.31109238e-01 1.20102811e+00 5.14546573e-01 3.29759479e-01
-1.36998701e+00 -4.90558118e-01 3.74577433e-01 2.00861871e-01
-1.43487060e+00 -4.29638326e-01 8.69391739e-01 -7.35030949e-01
9.83181357e-01 1.01925023e-01 6.17609084e-01 9.49344516e-01
2.00441740e-02 8.96881759e-01 1.16458237e+00 -1.84297234e-01
2.73919880e-01 1.48416162e-01 -5.41097708e-02 5.15591025e-01
3.20632935e-01 4.29781824e-01 -7.04107642e-01 6.33692965e-02
9.44697261e-01 9.57490653e-02 -2.53394276e-01 -1.15261877e+00
-1.19368267e+00 6.00521088e-01 7.00286746e-01 -2.07659334e-01
-3.22360903e-01 3.66040915e-01 -2.39710450e-01 5.98319918e-02
6.54726028e-01 4.78342235e-01 -5.58213055e-01 -4.55847792e-02
-8.05769622e-01 3.94801885e-01 3.25163007e-01 1.40161169e+00
9.90143478e-01 4.18370403e-02 8.92506505e-04 4.82646108e-01
2.85376638e-01 7.94210076e-01 -3.12340677e-01 -1.14378929e+00
6.45214796e-01 5.75414598e-01 3.23523849e-01 -8.05832863e-01
-3.07552695e-01 -6.71224594e-01 -5.01106441e-01 5.96165657e-01
4.57244396e-01 2.74412036e-01 -1.00689018e+00 1.51049602e+00
4.17940408e-01 2.17536971e-01 -1.84705198e-01 1.16489458e+00
6.58609509e-01 4.10286516e-01 -4.23490793e-01 2.85223305e-01
1.16745567e+00 -7.76511848e-01 -1.31618366e-01 -5.76022744e-01
4.28379208e-01 -6.82356417e-01 9.94521737e-01 3.79016131e-01
-1.03397512e+00 -6.81242287e-01 -8.69915545e-01 -3.98513734e-01
-4.26532984e-01 4.27088737e-01 5.01984298e-01 5.08023500e-01
-9.86276269e-01 9.71630812e-02 -8.88114810e-01 -8.43041297e-03
6.51890993e-01 3.50356013e-01 -4.43351775e-01 -4.33961809e-01
-5.90098202e-01 1.08419013e+00 9.04690996e-02 1.93421364e-01
-9.23707545e-01 -1.12801623e+00 -1.25435627e+00 -1.26475647e-01
7.95227706e-01 -8.97990167e-01 9.47650611e-01 -2.56611496e-01
-1.19603276e+00 1.08782983e+00 -2.73426086e-01 -3.81883413e-01
7.52745271e-01 -3.13128173e-01 2.59704113e-01 -5.28985709e-02
2.30045989e-01 1.02690375e+00 9.23621476e-01 -1.57684720e+00
-5.95102549e-01 -7.92789102e-01 2.70565659e-01 3.09022397e-01
3.60981137e-01 -6.19176090e-01 -5.12964427e-01 -2.53566563e-01
5.61535060e-01 -9.81770813e-01 -2.25836352e-01 5.17209411e-01
-4.93274719e-01 -1.25408366e-01 7.85103798e-01 -4.61919188e-01
2.98965812e-01 -2.19669604e+00 2.01423064e-01 2.52247918e-02
1.59754857e-01 -1.01603918e-01 6.02569804e-02 1.31937526e-02
1.45453125e-01 -2.46906325e-01 -3.13034862e-01 -6.20248795e-01
2.07334593e-01 2.23016158e-01 -3.86234075e-01 7.41052806e-01
7.14502394e-01 1.15690875e+00 -7.21550643e-01 -9.41888392e-02
5.63723922e-01 6.89039409e-01 -7.30118275e-01 1.77202299e-01
-5.07749856e-01 6.04354918e-01 -3.98515284e-01 8.94683421e-01
1.02568376e+00 -2.18297020e-01 -4.78422165e-01 -2.43084267e-01
-4.12768245e-01 3.19395095e-01 -1.23992503e+00 1.96997952e+00
-5.53983569e-01 6.31344080e-01 1.84634268e-01 -7.79237211e-01
9.28901434e-01 -1.51928380e-01 2.93069035e-01 -6.06906354e-01
3.56315188e-02 3.09236109e-01 -2.69209534e-01 -2.55907595e-01
4.73188102e-01 1.10349329e-02 3.62996757e-02 2.08727885e-02
-1.44140467e-01 -8.97866130e-01 -2.08504453e-01 -1.39186233e-01
7.87727118e-01 5.09836674e-01 1.37672186e-01 -2.16971755e-01
4.57374394e-01 1.45531133e-01 4.27016675e-01 6.76042080e-01
8.31930041e-02 8.62807453e-01 3.28750491e-01 -3.17681283e-01
-1.07535541e+00 -1.20856130e+00 -4.44464684e-01 2.15497866e-01
2.64478743e-01 1.58844516e-01 -2.80408889e-01 -6.03845358e-01
5.20478666e-01 6.31649911e-01 -5.20291567e-01 4.03651297e-02
-5.11563778e-01 -2.11098775e-01 -1.40074603e-02 6.25373125e-01
2.58401215e-01 -7.66898692e-01 -9.68361676e-01 -6.31323457e-02
1.26916736e-01 -1.52572417e+00 -3.97688329e-01 5.43576241e-01
-8.20398569e-01 -1.08533466e+00 -7.30171323e-01 -5.46454430e-01
7.71741450e-01 6.99662924e-01 1.04636061e+00 -3.85683537e-01
-4.13554668e-01 3.82537037e-01 1.21315330e-01 -4.35734481e-01
5.34013994e-02 8.47871900e-02 1.47787975e-02 -4.84876297e-02
5.20155609e-01 -5.84311545e-01 -4.59849894e-01 3.54475796e-01
-5.13262391e-01 2.27021545e-01 7.00187266e-01 7.37595320e-01
8.81641746e-01 -2.33622909e-01 1.95571303e-01 -6.42828882e-01
-2.48811498e-01 -1.82186484e-01 -1.43682015e+00 -1.66552991e-01
-3.26826394e-01 -8.13530944e-03 2.94412792e-01 -3.11429035e-02
-7.54094481e-01 5.50245404e-01 8.64185989e-02 -9.35311615e-01
-4.36475217e-01 2.53082383e-02 -3.54824632e-01 -2.44289711e-01
5.18770158e-01 2.70258546e-01 4.60886918e-02 -5.80685496e-01
4.92810875e-01 4.20550436e-01 7.14743614e-01 -6.01637542e-01
1.17579627e+00 9.29817319e-01 3.54024023e-01 -8.99912119e-01
-1.09873986e+00 -7.48508334e-01 -1.06318223e+00 -9.62663814e-02
8.82492542e-01 -1.32944930e+00 -7.79780447e-01 2.12290838e-01
-1.36948216e+00 -3.23777437e-01 -3.15603018e-01 3.77351969e-01
-8.41523945e-01 1.75587445e-01 -7.22357482e-02 -9.28566933e-01
2.46609747e-01 -1.24483716e+00 1.68536699e+00 -1.27762303e-01
2.14676261e-01 -5.16870856e-01 -4.00818616e-01 4.96486634e-01
1.28918633e-01 2.75355190e-01 7.62104690e-01 1.24856502e-01
-1.52764022e+00 -5.97782619e-02 -6.02152288e-01 3.69213760e-01
-9.65547748e-04 -3.33037645e-01 -1.31810677e+00 -2.54953597e-02
4.91367802e-02 -3.07584107e-01 7.55861342e-01 5.13765037e-01
1.20454347e+00 2.55492300e-01 -2.95288593e-01 8.65922928e-01
1.46256471e+00 -2.49697138e-02 2.15438142e-01 1.53097168e-01
9.49550748e-01 7.98866451e-01 7.78392196e-01 2.57244229e-01
5.35499990e-01 8.93968463e-01 9.25771892e-01 -7.58492872e-02
-1.28743872e-01 -2.68957824e-01 -6.02534451e-02 2.34263763e-01
2.67976820e-01 7.49642104e-02 -9.81698453e-01 6.16928756e-01
-1.60371780e+00 -5.26995063e-01 -3.40853393e-01 2.13583302e+00
5.42654753e-01 2.92468071e-01 -9.75702927e-02 7.18347132e-02
3.18920523e-01 3.22287343e-02 -8.60675454e-01 4.06524122e-01
-2.02988535e-01 6.46951795e-03 7.35135734e-01 5.75801253e-01
-1.05885816e+00 9.25861478e-01 4.96999073e+00 5.46103120e-01
-1.04909253e+00 -1.20086353e-02 4.56197411e-01 -2.13811353e-01
-2.09590867e-01 5.35729434e-03 -1.28987837e+00 1.75208062e-01
2.48798251e-01 3.59031409e-01 4.69180435e-01 1.02501309e+00
1.49438530e-01 -1.94340616e-01 -1.48996651e+00 1.46737480e+00
1.64534166e-01 -1.42269683e+00 -1.36929721e-01 3.96280348e-01
9.02842283e-01 4.62544262e-01 2.16410786e-01 -1.25748888e-02
-4.09503095e-02 -9.34404671e-01 1.17261982e+00 2.61601567e-01
8.81334007e-01 -5.03198981e-01 4.04774845e-01 7.95662940e-01
-1.00969923e+00 1.10609852e-01 -5.65405488e-01 -1.49852470e-01
2.12492064e-01 7.99438894e-01 -8.12310278e-01 5.43631136e-01
5.39245605e-01 9.26892757e-01 -3.91673297e-01 1.18160880e+00
-3.20383430e-01 4.12671193e-02 -5.84559739e-01 2.11293697e-01
3.55689764e-01 -1.86113566e-01 5.35074949e-01 7.21662104e-01
2.97854960e-01 -5.03346743e-03 2.79125959e-01 1.47558129e+00
-4.45424020e-02 -3.66649270e-01 -9.72448289e-01 3.42295617e-01
3.64269376e-01 1.15649676e+00 -5.11068583e-01 1.26064140e-02
-4.68379557e-01 5.91444314e-01 3.11177641e-01 3.16338062e-01
-6.34105861e-01 1.66651607e-01 9.55985069e-01 3.89914840e-01
5.78923285e-01 -7.63043165e-01 -6.98407352e-01 -1.14341676e+00
5.24373889e-01 -4.90031004e-01 -2.05328777e-01 -9.09648836e-01
-1.33966327e+00 4.06658739e-01 6.15301393e-02 -1.42503297e+00
-1.58398245e-02 -1.02437913e+00 -7.57578090e-02 1.09032953e+00
-2.19727921e+00 -1.17498076e+00 -6.86271131e-01 3.76279920e-01
5.77347636e-01 1.35986447e-01 3.41472268e-01 3.49179149e-01
-4.27275710e-02 2.28915125e-01 -2.21502855e-01 -2.48492852e-01
5.11019826e-01 -1.22057033e+00 6.89542651e-01 7.00295031e-01
3.40554506e-01 3.35368633e-01 3.81566405e-01 -4.90106881e-01
-1.84025407e+00 -1.14060843e+00 7.57727921e-01 -1.09468627e+00
4.18287307e-01 -8.94646108e-01 -7.35207379e-01 6.13299191e-01
-4.27841008e-01 3.51316869e-01 -3.12428102e-02 -1.36630058e-01
-2.32948571e-01 -3.50902118e-02 -9.41988170e-01 4.68984902e-01
1.49047554e+00 -6.76217616e-01 -3.86115730e-01 3.95304888e-01
7.65065253e-01 -9.56564069e-01 -4.28390712e-01 4.84906644e-01
2.28804037e-01 -8.62937093e-01 1.28152823e+00 -2.19705045e-01
4.24723715e-01 -7.10003436e-01 -4.13362533e-01 -1.08192456e+00
-2.38825399e-02 -3.82218868e-01 -9.90838557e-02 8.53959501e-01
2.90116847e-01 -5.37213445e-01 1.17325175e+00 4.08787191e-01
-4.50188875e-01 -7.10796535e-01 -1.19320905e+00 -9.77061749e-01
-1.56704232e-01 -7.36580253e-01 5.66610456e-01 7.24938154e-01
-8.60054731e-01 2.67885566e-01 -1.81997731e-01 6.88594282e-01
8.26289713e-01 2.75560498e-01 1.20534790e+00 -1.29261220e+00
-2.29224816e-01 -3.92498136e-01 -6.19177997e-01 -1.75392890e+00
3.21614474e-01 -9.24433231e-01 2.69293547e-01 -1.29152381e+00
-2.07839459e-01 -7.21268356e-01 2.35651687e-01 2.50358969e-01
9.14337113e-02 5.17067492e-01 2.15137213e-01 7.43544027e-02
-2.52500117e-01 8.46884191e-01 1.26776290e+00 -2.33918324e-01
-9.51573346e-03 1.43389747e-01 -5.06024539e-01 6.16259634e-01
2.64663339e-01 -4.19440329e-01 -4.58877057e-01 -7.81580687e-01
2.30200067e-01 -3.79796587e-02 9.47727621e-01 -8.79902780e-01
2.16880500e-01 -9.74682197e-02 4.34732676e-01 -1.41626370e+00
9.37736809e-01 -1.15903914e+00 -6.92876130e-02 1.19921334e-01
1.21839307e-01 -2.06253007e-01 3.91510651e-02 5.71025014e-01
-1.31099030e-01 -6.10204637e-02 7.95080960e-01 -1.60955548e-01
-7.29327977e-01 5.99587500e-01 2.94507325e-01 1.93731144e-01
9.08993304e-01 -4.81193930e-01 -8.62109065e-02 -2.57714391e-02
-4.95016754e-01 4.03474420e-01 8.06579411e-01 5.07428944e-01
5.66830039e-01 -1.26977706e+00 -6.33742511e-01 5.91212988e-01
4.04887468e-01 7.97012389e-01 1.03049837e-01 7.22211182e-01
-3.69466811e-01 7.89026082e-01 -5.64926416e-02 -1.34010005e+00
-9.54794049e-01 5.84243834e-01 3.07140201e-01 2.74025053e-01
-8.64042699e-01 8.95630956e-01 7.02703178e-01 -6.72748566e-01
4.41602886e-01 -8.20213139e-01 2.18854770e-01 -2.72596806e-01
1.57945246e-01 3.20644910e-03 3.11796635e-01 -5.44207811e-01
-4.42439824e-01 9.74866211e-01 2.69704312e-01 -1.03483237e-02
1.34633863e+00 -9.51104015e-02 1.91541359e-01 5.38271606e-01
1.08389211e+00 -2.89577246e-01 -1.84438920e+00 -4.19370115e-01
2.17326172e-02 -8.79853725e-01 1.76547125e-01 -5.97588003e-01
-7.54049599e-01 1.15756607e+00 3.96928221e-01 -2.01493651e-01
6.09218717e-01 2.39764914e-01 5.16090155e-01 4.66720104e-01
6.74356103e-01 -7.37027943e-01 -1.37615819e-02 6.65103316e-01
9.56545472e-01 -1.58045197e+00 -8.06868449e-02 -7.39463031e-01
-2.15136200e-01 9.20037031e-01 6.27649188e-01 -2.07183436e-01
6.64256036e-01 1.96042627e-01 -1.67561509e-02 -4.27965045e-01
-4.84163225e-01 -2.43666396e-01 6.72119319e-01 7.16823280e-01
-1.05210152e-02 -5.74618801e-02 4.39152837e-01 1.72394201e-01
-3.54367673e-01 -3.19243103e-01 1.52963415e-01 8.57576311e-01
-2.19164178e-01 -8.30466390e-01 -3.97968203e-01 2.39387184e-01
1.76533297e-01 -6.39442876e-02 -3.30039024e-01 7.91933060e-01
2.52857357e-01 5.27929187e-01 1.14147827e-01 1.92765072e-02
6.49480224e-01 -1.01417094e-01 7.38351583e-01 -9.04184461e-01
-8.21825564e-02 4.07177582e-02 -1.46872684e-01 -5.78059018e-01
-2.53225714e-01 -9.37023759e-01 -9.41406488e-01 1.07750759e-01
-3.11276257e-01 -3.56125265e-01 1.02029455e+00 1.02718055e+00
4.24307555e-01 2.69656003e-01 5.52446306e-01 -1.46113908e+00
-6.86335266e-01 -5.58298409e-01 -4.30361629e-01 -2.27686646e-03
6.76663995e-01 -1.06837928e+00 -3.72381836e-01 -1.02404475e-01] | [7.845257759094238, -2.6394922733306885] |
0fdceb76-4d03-46bd-a883-100b62fddc01 | scalable-sparse-subspace-clustering-by | 1507.01238 | null | http://arxiv.org/abs/1507.01238v3 | http://arxiv.org/pdf/1507.01238v3.pdf | Scalable Sparse Subspace Clustering by Orthogonal Matching Pursuit | Subspace clustering methods based on $\ell_1$, $\ell_2$ or nuclear norm
regularization have become very popular due to their simplicity, theoretical
guarantees and empirical success. However, the choice of the regularizer can
greatly impact both theory and practice. For instance, $\ell_1$ regularization
is guaranteed to give a subspace-preserving affinity (i.e., there are no
connections between points from different subspaces) under broad conditions
(e.g., arbitrary subspaces and corrupted data). However, it requires solving a
large scale convex optimization problem. On the other hand, $\ell_2$ and
nuclear norm regularization provide efficient closed form solutions, but
require very strong assumptions to guarantee a subspace-preserving affinity,
e.g., independent subspaces and uncorrupted data. In this paper we study a
subspace clustering method based on orthogonal matching pursuit. We show that
the method is both computationally efficient and guaranteed to give a
subspace-preserving affinity under broad conditions. Experiments on synthetic
data verify our theoretical analysis, and applications in handwritten digit and
face clustering show that our approach achieves the best trade off between
accuracy and efficiency. | ['Chong You', 'Rene Vidal', 'Daniel P. Robinson'] | 2015-07-05 | scalable-sparse-subspace-clustering-by-1 | http://openaccess.thecvf.com/content_cvpr_2016/html/You_Scalable_Sparse_Subspace_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/You_Scalable_Sparse_Subspace_CVPR_2016_paper.pdf | cvpr-2016-6 | ['face-clustering'] | ['computer-vision'] | [-2.45739833e-01 -3.73322845e-01 -2.00933382e-01 -2.15584740e-01
-5.20726383e-01 -5.10618091e-01 -6.29370064e-02 -3.18107277e-01
-1.51367858e-01 5.34095645e-01 -5.71819430e-04 -1.28444089e-02
-3.54439884e-01 -4.17891443e-01 -5.63675582e-01 -1.23643661e+00
5.02297245e-02 2.35859066e-01 -1.42835617e-01 1.18724205e-01
1.38882339e-01 6.24692082e-01 -1.31396651e+00 -3.29329371e-01
1.05582440e+00 8.67821574e-01 -4.67105508e-02 5.35555445e-02
4.00709584e-02 3.17486644e-01 -1.33053243e-01 -8.50514993e-02
5.28221786e-01 -3.68153960e-01 -4.62068081e-01 6.11635149e-01
3.22490990e-01 1.71714306e-01 -4.50864404e-01 1.52103305e+00
4.86265957e-01 3.41531873e-01 6.31780505e-01 -1.28671157e+00
-5.91722012e-01 2.44605333e-01 -1.10019970e+00 -2.04360023e-01
2.09398791e-02 -1.57424957e-01 7.74148405e-01 -1.26374912e+00
4.82016891e-01 9.77115989e-01 6.90642178e-01 3.78915459e-01
-1.50562572e+00 -7.20360696e-01 2.67531276e-01 6.34915754e-02
-1.94907010e+00 -5.19415796e-01 1.04533327e+00 -5.09012997e-01
3.12400341e-01 5.39363265e-01 2.80304223e-01 5.67931056e-01
-2.04714820e-01 7.45221794e-01 9.68922198e-01 -1.89045191e-01
4.76501226e-01 1.56336427e-01 4.24529493e-01 4.63137865e-01
4.02186692e-01 -1.54320121e-01 -3.50319117e-01 -3.65340859e-01
8.49795818e-01 -1.94732510e-02 -6.99237764e-01 -9.46934044e-01
-1.20333600e+00 8.99471700e-01 2.88577467e-01 4.15737003e-01
-3.20025980e-01 -2.11418450e-01 1.80390477e-01 5.23236766e-02
1.62485853e-01 1.71675891e-01 6.36114040e-04 2.75043845e-01
-1.11346745e+00 8.47920775e-03 5.11374235e-01 1.06765628e+00
7.72559524e-01 2.26926237e-01 1.65712863e-01 1.07169580e+00
3.40400904e-01 7.08147645e-01 2.82579243e-01 -1.18659496e+00
2.86274850e-01 4.02904212e-01 1.85954690e-01 -1.47501528e+00
-3.61596256e-01 -5.13345003e-01 -1.40695584e+00 1.45461127e-01
6.51224852e-01 3.28470208e-02 -5.18820941e-01 1.96748805e+00
3.46752018e-01 3.89688760e-01 -2.36476064e-02 1.26391280e+00
4.62258816e-01 5.41558146e-01 -3.54087114e-01 -8.78195107e-01
1.06661510e+00 -5.13344526e-01 -9.39074874e-01 -7.18779117e-02
4.51750994e-01 -7.67943799e-01 1.00950825e+00 4.61316615e-01
-1.04306674e+00 -3.59756321e-01 -8.36360812e-01 3.86670977e-01
1.45888761e-01 2.65727401e-01 4.77730513e-01 6.93462551e-01
-8.15374494e-01 3.96758765e-01 -8.57636690e-01 -2.80649960e-01
1.16543494e-01 4.98865694e-01 -6.22630835e-01 -3.13635528e-01
-5.91912389e-01 2.18452394e-01 2.56187320e-01 2.16672659e-01
-2.63602138e-01 -4.03687716e-01 -6.78053617e-01 -5.13871387e-02
3.28871429e-01 -2.65448302e-01 5.10296285e-01 -7.18296945e-01
-1.10079360e+00 7.66375482e-01 -4.05293852e-01 3.16666886e-02
3.72913510e-01 -5.07538859e-03 -4.25643831e-01 1.69016019e-01
2.07000971e-01 2.49431670e-01 8.99294496e-01 -1.28739035e+00
-3.42319846e-01 -7.50799537e-01 -5.31928122e-01 2.24656239e-01
-8.34499717e-01 -2.35153735e-02 -8.82668316e-01 -7.61197329e-01
9.15134966e-01 -1.08382142e+00 -3.50768119e-01 -1.20739639e-01
-4.82225925e-01 6.72592269e-03 9.63291287e-01 -6.76628947e-01
1.36399782e+00 -2.54634285e+00 3.04353982e-01 7.71084666e-01
1.48476198e-01 2.54744962e-02 1.73753686e-02 5.80191538e-02
-3.12589556e-01 -5.32947443e-02 -6.60314262e-01 -1.06961027e-01
-1.35858133e-01 1.35851085e-01 -1.02862842e-01 1.13596773e+00
-3.13181162e-01 3.32187921e-01 -6.84881508e-01 -4.09886926e-01
1.72019079e-01 5.22281408e-01 -6.40946984e-01 -9.30427015e-02
3.96817327e-01 5.54396987e-01 -3.67527932e-01 5.94707131e-01
1.00028217e+00 -3.29080731e-01 2.82831997e-01 -4.38378453e-01
-1.06541522e-01 -6.45374238e-01 -1.87920964e+00 1.66309893e+00
3.48653458e-02 5.00401795e-01 7.57640064e-01 -1.30375552e+00
9.62186217e-01 3.51354837e-01 8.33767593e-01 -2.73805350e-01
6.33068830e-02 4.83079672e-01 -9.91204381e-02 -8.52104351e-02
9.17663872e-02 -2.67435819e-01 5.88091910e-02 3.53833139e-01
-3.01877320e-01 2.21924156e-01 1.84321664e-02 2.24714130e-01
8.87863994e-01 -2.63109624e-01 1.77227139e-01 -8.14884782e-01
8.06388199e-01 -1.67695701e-01 1.15668821e+00 4.35736597e-01
-3.57176691e-01 8.53651166e-01 3.18951726e-01 -2.08195001e-02
-9.12862659e-01 -8.57397497e-01 -4.22072709e-01 6.02118909e-01
6.09327793e-01 -2.28940040e-01 -7.88269460e-01 -4.52742159e-01
-6.30246252e-02 5.17744720e-01 -3.31544787e-01 -1.17416792e-01
-7.18887269e-01 -9.18701947e-01 1.74437210e-01 5.08094907e-01
5.05065262e-01 -4.49574351e-01 -1.54979736e-01 1.29152134e-01
-3.27508777e-01 -1.06367922e+00 -8.37339997e-01 -1.08790375e-01
-1.00017917e+00 -1.05927503e+00 -9.84528184e-01 -8.54212105e-01
1.24458158e+00 7.13845432e-01 6.63461924e-01 -3.48988734e-02
-1.12113513e-01 5.20821273e-01 -1.08720019e-01 2.17166051e-01
1.37222558e-01 -4.66381997e-01 5.70023000e-01 4.67990190e-01
3.53871852e-01 -7.97708869e-01 -4.60588634e-01 8.01558614e-01
-9.53895390e-01 -3.46617341e-01 3.39298338e-01 9.26369250e-01
9.72921252e-01 5.18281400e-01 2.17648774e-01 -5.74281573e-01
6.03637636e-01 -2.47731000e-01 -6.54334545e-01 2.65085101e-01
-6.26835525e-01 1.56601053e-02 7.50469625e-01 -4.25238609e-01
-7.97580302e-01 4.14190829e-01 1.51938573e-01 -1.06361032e+00
-7.08604380e-02 4.85417545e-01 -5.84510684e-01 -1.36763707e-01
7.33553529e-01 4.75690544e-01 1.48360193e-01 -6.15699589e-01
2.76190609e-01 5.43442667e-01 6.70804560e-01 -6.79041028e-01
1.01988411e+00 9.09277022e-01 3.98085751e-02 -1.08126044e+00
-4.47799027e-01 -8.41673017e-01 -7.28775263e-01 -1.75956398e-01
5.82207978e-01 -8.20782125e-01 -6.38155401e-01 2.98668951e-01
-7.19471455e-01 2.06583098e-01 -1.60457224e-01 7.43949056e-01
-5.86628258e-01 9.06850159e-01 -3.67004335e-01 -8.48616481e-01
-1.03929333e-01 -1.05954587e+00 5.88097095e-01 6.78942949e-02
-3.97330113e-02 -8.51574719e-01 -1.30772963e-01 2.65286624e-01
1.66818127e-01 2.64811397e-01 6.07877135e-01 -3.03532928e-01
-2.18313888e-01 -3.39766532e-01 -2.52341270e-01 3.51771981e-01
3.15756917e-01 -7.79715776e-02 -7.11603940e-01 -7.52503216e-01
3.06003094e-01 1.29118696e-01 6.73945487e-01 5.80051363e-01
1.20157301e+00 -4.63874936e-01 -4.05247927e-01 8.12004685e-01
1.31516731e+00 2.40773633e-01 5.58072448e-01 -3.67207043e-02
8.05852115e-01 7.79151142e-01 5.05415201e-01 5.03672600e-01
-1.85541496e-01 6.66352391e-01 2.25046977e-01 -8.13590083e-03
1.04694970e-01 1.75763965e-01 2.78190404e-01 1.03006077e+00
-1.31924674e-01 1.88053176e-01 -8.77273560e-01 5.31455934e-01
-2.15316010e+00 -9.14114416e-01 -3.25410903e-01 2.76093531e+00
6.06210649e-01 -2.96966612e-01 2.29638070e-01 4.26012903e-01
1.10961580e+00 -2.78529264e-02 -7.57158399e-01 2.68896639e-01
-4.55914170e-01 -1.62329555e-01 4.69171703e-01 5.42988300e-01
-1.14774001e+00 6.68521345e-01 5.28172874e+00 1.08654618e+00
-9.91875410e-01 -1.25816427e-02 4.25849646e-01 -1.21627085e-01
-2.06455618e-01 -7.02489987e-02 -5.90952575e-01 5.00296235e-01
2.50138670e-01 -2.84269810e-01 5.93180060e-01 8.18242133e-01
1.39158323e-01 2.42444351e-01 -8.82947206e-01 1.57704353e+00
8.51752385e-02 -1.04965377e+00 -2.17195749e-01 2.88904488e-01
6.95652902e-01 -4.01915669e-01 3.62393595e-02 -1.13375753e-01
7.52449706e-02 -9.15823340e-01 5.43801188e-01 1.57884836e-01
8.29104424e-01 -9.47926462e-01 5.22757769e-01 5.51978886e-01
-1.23221111e+00 3.01106032e-02 -6.84981763e-01 2.75906026e-01
1.26304522e-01 9.49575424e-01 -3.80218122e-03 5.67208052e-01
7.75749445e-01 8.26962471e-01 -6.09056763e-02 9.32339609e-01
-8.00195336e-02 4.47486252e-01 -5.16764224e-01 3.88977885e-01
3.28034312e-02 -9.33832526e-01 8.69979620e-01 1.06264484e+00
4.77510512e-01 6.12794757e-01 3.31806153e-01 8.04022968e-01
7.84113780e-02 4.56629425e-01 -5.62224567e-01 1.39718011e-01
5.14104962e-01 1.13250208e+00 -9.13768649e-01 -6.02811910e-02
-4.49225456e-01 1.00961316e+00 1.92978889e-01 7.03450680e-01
-6.75837219e-01 -3.36405426e-01 8.32078218e-01 1.42428771e-01
2.46521026e-01 -3.90771389e-01 -5.62714458e-01 -1.52508819e+00
1.38545185e-01 -9.29142237e-01 4.57600921e-01 -1.26949579e-01
-1.35215282e+00 2.85363942e-01 -3.51700902e-01 -1.43300426e+00
1.67929217e-01 -5.91498137e-01 -4.45653647e-01 7.46010303e-01
-1.01009166e+00 -5.72497785e-01 -3.12924922e-01 1.12277687e+00
8.43681768e-02 -2.25143716e-01 5.73352396e-01 4.54733640e-01
-1.06927681e+00 6.88487887e-01 6.50137246e-01 2.72196472e-01
6.11666203e-01 -9.85645950e-01 -5.50594985e-01 1.02232087e+00
2.66539007e-01 9.00606573e-01 7.42249727e-01 -4.53111619e-01
-1.66614628e+00 -9.31018293e-01 4.08432931e-01 -7.13978987e-03
3.89128983e-01 -2.46929035e-01 -1.15651453e+00 4.50659394e-01
-1.33928239e-01 1.55360967e-01 8.89688611e-01 2.44779870e-01
-3.91283453e-01 -1.56359479e-01 -1.26898324e+00 6.57475710e-01
1.01055562e+00 -3.79465222e-01 -2.53895611e-01 4.44957763e-01
1.34985849e-01 -1.24765404e-01 -1.01795602e+00 6.74580038e-01
4.23073202e-01 -1.01149237e+00 1.18911374e+00 -4.47122067e-01
-2.09813371e-01 -6.67946398e-01 -5.92161596e-01 -9.62508082e-01
-7.42981017e-01 -6.60118818e-01 -2.21065909e-01 1.12675989e+00
1.76541999e-01 -6.92093134e-01 9.50483978e-01 9.43235219e-01
3.47449556e-02 -5.89588702e-01 -1.13804770e+00 -1.30185723e+00
1.19003370e-01 -3.67031991e-01 2.44856551e-01 1.35928750e+00
1.54125765e-01 2.08708778e-01 -5.24170458e-01 4.34021384e-01
1.16272962e+00 4.16186005e-01 6.59721196e-01 -1.31822443e+00
-1.55234143e-01 -5.35386443e-01 -3.24080646e-01 -1.19893730e+00
3.30120385e-01 -6.58565283e-01 1.46148894e-02 -1.18827593e+00
3.34986299e-01 -7.34750032e-01 -4.89211380e-01 2.52521604e-01
-2.40100473e-01 3.42176974e-01 5.50780594e-02 8.09094608e-01
-3.93191814e-01 6.22506320e-01 1.01538062e+00 -1.71002477e-01
-4.20523822e-01 -4.85612713e-02 -6.91354096e-01 9.41452444e-01
7.87394047e-01 -1.62172824e-01 -3.31836581e-01 -2.75930732e-01
-7.24767894e-02 1.00163691e-01 1.37142047e-01 -8.93932164e-01
3.94337088e-01 -2.61709362e-01 2.84948587e-01 -3.52386713e-01
4.97113317e-01 -1.07928097e+00 3.38380069e-01 3.90944779e-01
7.26292059e-02 -3.89516115e-01 2.81262193e-02 8.14997196e-01
-3.90037745e-01 -1.21984757e-01 1.14857864e+00 9.53838825e-02
-5.19876063e-01 4.11968023e-01 2.10853368e-02 -7.40587413e-02
1.11313415e+00 -6.17448568e-01 2.40720719e-01 -5.68107426e-01
-7.99945056e-01 2.27644414e-01 7.86453485e-01 5.79667613e-02
5.96789360e-01 -1.66479027e+00 -8.07495534e-01 3.65984410e-01
-6.53198287e-02 -6.22988567e-02 1.90174177e-01 1.22062850e+00
-1.83818415e-01 4.72605199e-01 1.13556251e-01 -8.75661075e-01
-1.42330503e+00 8.57910991e-01 9.25571620e-02 1.41633645e-01
-5.32253444e-01 8.91450822e-01 5.28680563e-01 -3.62347066e-01
3.86578590e-01 1.54649034e-01 3.19962925e-03 -8.75562243e-03
4.04713422e-01 5.94181359e-01 -1.78164586e-01 -9.10573840e-01
-5.45223892e-01 8.99891257e-01 1.80925384e-01 -1.07890248e-01
1.18935180e+00 -1.39593095e-01 -3.68912101e-01 1.62608594e-01
1.21533728e+00 3.01387489e-01 -1.25915778e+00 -3.95157009e-01
-7.56737441e-02 -8.29873085e-01 1.27097294e-01 -1.59856137e-02
-1.37066603e+00 7.41896033e-01 4.68653023e-01 1.27659678e-01
1.35022616e+00 -1.43628374e-01 6.22520030e-01 3.79201770e-01
5.88359475e-01 -1.26502669e+00 -1.10942058e-01 1.85543209e-01
8.75822127e-01 -1.07247949e+00 1.17794551e-01 -7.94036746e-01
-3.92503470e-01 8.95454228e-01 5.28682232e-01 -1.92698270e-01
7.64870286e-01 -3.65965329e-02 -1.02939673e-01 -1.39652193e-01
-5.31974360e-02 1.04920231e-01 3.64830792e-01 3.51054192e-01
3.81694645e-01 9.34939235e-02 -5.92008173e-01 5.60784698e-01
3.17667797e-02 -4.48711544e-01 1.32884115e-01 6.69593394e-01
-3.61623228e-01 -9.34870839e-01 -9.80751634e-01 3.09227228e-01
-2.86880583e-01 1.54594094e-01 -2.89082289e-01 6.57212794e-01
-1.23051383e-01 1.05575967e+00 -1.57977432e-01 -2.42754295e-01
2.69872636e-01 -1.73346981e-01 4.16744381e-01 -3.10594976e-01
-2.84569291e-03 6.48021340e-01 -4.53003377e-01 -5.80451906e-01
-4.77757186e-01 -9.77946997e-01 -1.32829845e+00 -4.20047045e-01
-5.20381927e-01 4.84664500e-01 3.20621133e-01 5.58766484e-01
4.65734631e-01 -2.66646314e-02 7.80280948e-01 -4.46867794e-01
-6.09374404e-01 -6.09928012e-01 -1.18675256e+00 5.29166698e-01
-1.88439842e-02 -5.69868743e-01 -6.43189371e-01 1.13746777e-01] | [7.675570487976074, 4.417294025421143] |
c2aa4081-bd69-4743-9356-b2109be55d9e | ps-nerv-patch-wise-stylized-neural | 2208.03742 | null | https://arxiv.org/abs/2208.03742v1 | https://arxiv.org/pdf/2208.03742v1.pdf | PS-NeRV: Patch-wise Stylized Neural Representations for Videos | We study how to represent a video with implicit neural representations (INRs). Classical INRs methods generally utilize MLPs to map input coordinates to output pixels. While some recent works have tried to directly reconstruct the whole image with CNNs. However, we argue that both the above pixel-wise and image-wise strategies are not favorable to video data. Instead, we propose a patch-wise solution, PS-NeRV, which represents videos as a function of patches and the corresponding patch coordinate. It naturally inherits the advantages of image-wise methods, and achieves excellent reconstruction performance with fast decoding speed. The whole method includes conventional modules, like positional embedding, MLPs and CNNs, while also introduces AdaIN to enhance intermediate features. These simple yet essential changes could help the network easily fit high-frequency details. Extensive experiments have demonstrated its effectiveness in several video-related tasks, such as video compression and video inpainting. | ['Cairong Wang', 'Chao Dong', 'Yunpeng Bai'] | 2022-08-07 | null | null | null | null | ['video-inpainting'] | ['computer-vision'] | [ 3.49711329e-01 -3.57200690e-02 -3.87494326e-01 -1.32942557e-01
-4.23911721e-01 3.00566736e-03 3.46321940e-01 -2.23237410e-01
-2.86384225e-01 7.92822123e-01 1.85693890e-01 -8.11770093e-03
1.56554312e-01 -1.00438881e+00 -1.28221524e+00 -7.18534589e-01
1.95798799e-01 -2.68540204e-01 1.22432731e-01 -9.99971405e-02
3.91744941e-01 6.09853745e-01 -1.33959210e+00 5.22516966e-01
5.97290099e-01 1.19301939e+00 4.49445277e-01 4.57809180e-01
-1.90446243e-01 1.14667583e+00 -4.07435834e-01 -3.09027165e-01
4.72514600e-01 -4.13829178e-01 -4.42352265e-01 5.75041883e-02
3.85694891e-01 -7.11233974e-01 -1.06958818e+00 1.06945169e+00
4.09174770e-01 5.02643250e-02 4.43643451e-01 -9.51792777e-01
-9.90723968e-01 4.35690910e-01 -8.60844553e-01 -2.18006913e-02
3.71964723e-01 1.70488223e-01 6.62867367e-01 -9.53863859e-01
6.10479474e-01 1.19837153e+00 1.03743744e+00 4.84288812e-01
-1.18590534e+00 -4.82115000e-01 2.48977169e-02 5.07856309e-01
-1.49793863e+00 -4.28435385e-01 1.12309778e+00 -1.02463827e-01
8.93928945e-01 3.14250618e-01 8.33336592e-01 8.08461964e-01
1.49443537e-01 9.93078411e-01 8.00569415e-01 -4.37615722e-01
-5.36524542e-02 -5.89512177e-02 -3.45412225e-01 6.72188878e-01
2.64819060e-03 -7.74776097e-03 -5.87907672e-01 1.98017418e-01
1.39609718e+00 4.69813883e-01 -6.54893398e-01 -1.37434587e-01
-1.33581829e+00 6.52927160e-01 6.74053073e-01 2.79456198e-01
-5.90065718e-01 5.89476228e-01 2.93657690e-01 3.09718251e-01
3.20910364e-01 2.20687106e-01 -2.91657776e-01 -2.70578768e-02
-1.20339203e+00 2.67441154e-01 3.14442217e-01 1.00175953e+00
1.04826832e+00 2.34238625e-01 -2.04145074e-01 8.30335319e-01
-2.39139237e-02 1.13134861e-01 6.00974083e-01 -1.24446988e+00
6.06436729e-01 3.12420160e-01 3.51433307e-02 -1.66154051e+00
6.74430206e-02 -9.60211307e-02 -1.21679509e+00 8.29470083e-02
-1.02953576e-01 8.15571286e-03 -7.17786849e-01 1.58962131e+00
-5.45002110e-02 6.53874993e-01 -1.78025827e-01 8.33626568e-01
8.98570657e-01 1.18243897e+00 -1.56231090e-01 -1.82625949e-01
9.13454115e-01 -1.29353428e+00 -9.17266607e-01 3.28339031e-03
2.08429441e-01 -6.42725408e-01 8.47897947e-01 1.91927001e-01
-1.52970684e+00 -9.52735305e-01 -1.17608428e+00 -4.72261369e-01
-1.34162247e-01 2.85058767e-01 4.06998843e-01 2.74730772e-02
-1.32363796e+00 9.55156624e-01 -6.98349833e-01 8.09307620e-02
4.91371185e-01 2.87143856e-01 -5.24053216e-01 -1.16298445e-01
-1.12862647e+00 6.11205697e-01 4.06594306e-01 3.45639676e-01
-5.33891380e-01 -5.54549634e-01 -1.01779258e+00 2.59433031e-01
1.89397290e-01 -7.17982233e-01 9.24681544e-01 -1.28840947e+00
-1.63000941e+00 4.18213308e-01 -3.93174142e-01 -5.06233037e-01
3.75830680e-01 -2.28681073e-01 -1.69029891e-01 4.71864074e-01
-1.30636483e-01 1.26317966e+00 1.19948351e+00 -1.14767349e+00
-4.18499053e-01 6.93612099e-02 1.15282208e-01 2.73310333e-01
-4.50925469e-01 -3.18582088e-01 -7.75266111e-01 -1.17538345e+00
3.48512828e-01 -4.30595517e-01 -3.03550005e-01 6.15744233e-01
-2.98376948e-01 9.87959951e-02 8.78637433e-01 -1.04849696e+00
1.19124019e+00 -2.27720141e+00 4.18084592e-01 -1.13455512e-01
1.86282545e-01 3.31459403e-01 -2.88647771e-01 4.12602782e-01
-1.97451189e-01 -3.62737179e-02 -2.10014835e-01 -4.02552158e-01
-2.63647258e-01 3.35349828e-01 -3.53980422e-01 2.97359824e-01
5.89755654e-01 1.20992553e+00 -6.37021959e-01 -5.78069925e-01
3.35171968e-01 7.59682894e-01 -8.40909302e-01 2.18584448e-01
-1.56871736e-01 3.21704358e-01 -2.13521004e-01 5.90626657e-01
8.26878548e-01 -9.49113518e-02 6.63317293e-02 -6.01310432e-01
-9.30284560e-02 1.80048540e-01 -9.51053500e-01 1.92679369e+00
-3.89536470e-01 8.84494305e-01 4.93539833e-02 -1.39310777e+00
7.69774735e-01 2.20730856e-01 4.35944587e-01 -7.97101557e-01
-5.89530095e-02 9.04266015e-02 -5.29734552e-01 -6.64360285e-01
7.83256650e-01 -4.47659899e-04 4.07502681e-01 8.07701610e-04
-9.87079367e-03 1.84240192e-01 -1.70878664e-01 3.27442400e-02
8.96964610e-01 6.16537571e-01 2.88645476e-01 2.14294776e-01
4.29436862e-01 -1.48257017e-01 5.48863590e-01 6.13408506e-01
9.93584394e-02 1.00425518e+00 5.14373600e-01 -6.16122127e-01
-1.17774522e+00 -7.79727995e-01 4.94045392e-02 5.37636995e-01
4.29196507e-01 -5.45280635e-01 -7.51353920e-01 -2.74506629e-01
-2.09002033e-01 2.95290798e-01 -5.27308464e-01 -8.63160193e-02
-8.47063661e-01 -3.56852829e-01 4.18931305e-01 5.84574699e-01
8.02951038e-01 -1.25936711e+00 -4.40395325e-01 4.28119719e-01
-3.32530618e-01 -1.02190948e+00 -5.07306755e-01 1.68982986e-02
-1.19430757e+00 -8.40540826e-01 -1.11581230e+00 -1.26257968e+00
9.05125916e-01 6.09403729e-01 8.74943912e-01 3.17775786e-01
-1.71229899e-01 -2.49811374e-02 -4.33404237e-01 3.66812646e-01
-1.40603676e-01 -1.61696315e-01 -2.45802358e-01 -4.63842712e-02
8.67063552e-02 -9.10079956e-01 -8.52033734e-01 -9.31743998e-03
-1.11638689e+00 4.74288166e-01 8.93275917e-01 1.07068264e+00
9.03942108e-01 1.80332258e-01 3.11527729e-01 -7.10670412e-01
4.12619025e-01 -3.04951191e-01 -2.03929156e-01 1.15243763e-01
-1.57307863e-01 -1.16232753e-01 9.68838751e-01 -4.37769741e-01
-6.70121670e-01 5.46625778e-02 -4.86716330e-01 -7.91221797e-01
-1.14047699e-01 4.73850340e-01 -1.63770065e-01 -3.53972137e-01
2.72592962e-01 7.43304849e-01 1.82562992e-01 -5.88341773e-01
3.74063998e-01 5.30134737e-01 6.50547683e-01 -3.90880436e-01
7.59735048e-01 3.46471876e-01 -1.82344541e-02 -7.37888873e-01
-3.41772407e-01 1.24377676e-03 -4.95984972e-01 -6.35543317e-02
6.15864038e-01 -1.10211647e+00 -4.93920058e-01 4.60114926e-01
-1.50613916e+00 -1.74433053e-01 -3.94759536e-01 2.28106439e-01
-6.66767001e-01 7.71925211e-01 -1.20686829e+00 -2.73007661e-01
-1.87010348e-01 -1.30357075e+00 9.49148893e-01 1.66176736e-01
7.78596327e-02 -7.45164931e-01 -3.22900951e-01 -1.12503253e-01
4.55562741e-01 1.96790144e-01 9.03279781e-01 2.49382034e-01
-8.78511131e-01 -2.15179957e-02 -4.71091390e-01 6.67735040e-01
1.14359699e-01 -9.60364193e-02 -7.02488184e-01 -3.10436696e-01
6.00166023e-02 -1.36605322e-01 8.76668215e-01 5.15400052e-01
1.88350034e+00 -7.58238137e-01 -1.93966195e-01 1.03869939e+00
1.67590153e+00 2.93513536e-01 1.18464398e+00 4.52169359e-01
8.32024634e-01 2.85256326e-01 2.74043530e-01 2.73807049e-01
1.33337721e-01 7.01235652e-01 4.21054691e-01 -3.76076430e-01
-3.40302169e-01 -6.04239106e-01 4.09271479e-01 1.26287782e+00
-1.60075337e-01 -8.71400759e-02 -1.55875161e-01 2.57179230e-01
-1.89770901e+00 -1.08432186e+00 2.25949600e-01 1.68873513e+00
9.62686360e-01 -4.90859412e-02 -2.53101170e-01 1.83126152e-01
7.44635224e-01 4.99015033e-01 -5.37420273e-01 -2.36821294e-01
-2.66176671e-01 2.25492924e-01 5.30577242e-01 3.01905125e-01
-1.02318215e+00 8.21051061e-01 6.90928602e+00 1.21895254e+00
-1.26679182e+00 5.14450222e-02 9.55117047e-01 -1.25143945e-01
-3.03625554e-01 -2.08527178e-01 -3.76208901e-01 6.40273869e-01
4.77279812e-01 2.44249538e-01 7.03854382e-01 7.33204007e-01
2.03473687e-01 9.45595205e-02 -1.02471793e+00 1.43597913e+00
1.44594431e-01 -1.94164038e+00 4.92861658e-01 -2.12901771e-01
7.63318002e-01 -4.65107679e-01 1.75970599e-01 9.78910848e-02
-5.45235872e-01 -1.10546744e+00 9.28536952e-01 6.91939592e-01
1.15721643e+00 -6.83516145e-01 6.23182893e-01 3.02893609e-01
-1.16114795e+00 1.23397522e-01 -7.50774086e-01 -2.14559287e-01
1.46305218e-01 3.92204612e-01 -1.77565396e-01 6.43080235e-01
7.76989996e-01 1.31554949e+00 -3.16834748e-01 9.73938465e-01
-1.48808479e-01 3.18351299e-01 -8.78215358e-02 2.41258129e-01
3.28180015e-01 -3.89750719e-01 2.84542441e-01 1.11083341e+00
6.11586154e-01 1.46067053e-01 -2.43987277e-01 8.30898225e-01
-2.76215315e-01 4.04309891e-02 -7.36873090e-01 1.52944848e-02
3.42677236e-01 9.45606232e-01 -4.91995931e-01 -4.80196476e-01
-5.65136433e-01 1.34469056e+00 2.75046408e-01 5.41617155e-01
-9.04095948e-01 -6.28804743e-01 5.35345197e-01 3.79144013e-01
5.96395612e-01 -2.76847959e-01 -1.60545334e-01 -1.10182381e+00
2.42103741e-01 -9.19429064e-01 -2.65982836e-01 -1.13914144e+00
-9.40512061e-01 5.53264320e-01 -1.49841800e-01 -1.69726706e+00
-4.23435196e-02 -6.26144230e-01 -5.74390233e-01 7.57306874e-01
-1.68166387e+00 -8.71038198e-01 -2.33920306e-01 6.44773066e-01
7.13279605e-01 -1.41712397e-01 5.43323994e-01 6.64020181e-01
-5.29907107e-01 5.87290525e-01 3.33026052e-01 1.25187635e-01
4.99511272e-01 -6.42026782e-01 3.05288374e-01 7.99857795e-01
1.14286907e-01 7.67099380e-01 5.54020822e-01 -3.55309784e-01
-1.57297993e+00 -1.18720973e+00 5.69851696e-01 2.96438515e-01
1.95392147e-01 8.19085911e-02 -8.19286346e-01 6.72484219e-01
4.47449267e-01 6.67202994e-02 2.14909554e-01 -5.44054389e-01
-1.34983897e-01 -2.78738588e-01 -1.01060772e+00 8.23478162e-01
1.14529598e+00 -6.09992862e-01 -4.95750099e-01 2.42716640e-01
9.51817155e-01 -4.61256087e-01 -6.81279898e-01 3.60966444e-01
4.71038580e-01 -1.24996769e+00 1.27198613e+00 -3.45425844e-01
1.13298690e+00 -4.31755185e-01 -2.85919756e-01 -1.09450388e+00
-5.66192806e-01 -5.30422151e-01 -3.61334860e-01 1.00621068e+00
-8.18165485e-03 -5.00043809e-01 7.97361672e-01 1.02393620e-01
-1.64834678e-01 -1.10488391e+00 -8.30718517e-01 -4.30273622e-01
-2.42252752e-01 -1.86708361e-01 6.66491568e-01 9.27083969e-01
-2.70355016e-01 -1.94389448e-01 -8.58978748e-01 2.40654070e-02
2.60082096e-01 4.55767177e-02 5.83505452e-01 -5.18615127e-01
-5.34125328e-01 -4.09297347e-01 -5.88079154e-01 -1.87050176e+00
6.34394353e-03 -6.15604639e-01 8.08360502e-02 -1.54911315e+00
1.87035307e-01 -4.36233997e-01 -3.38297486e-01 4.77176517e-01
-3.15650851e-02 6.75993800e-01 1.87713057e-01 3.64051700e-01
-4.70879853e-01 7.73157656e-01 1.46512854e+00 -3.11595380e-01
4.58608605e-02 -4.52233583e-01 -6.78470492e-01 6.06464505e-01
7.53518701e-01 -2.71742731e-01 -2.61239260e-01 -7.99607217e-01
1.03629835e-01 2.59884089e-01 4.61527258e-01 -1.11811972e+00
2.47814119e-01 -6.54519275e-02 7.48481572e-01 -7.63296127e-01
6.62721157e-01 -8.32332075e-01 5.63960016e-01 4.27850306e-01
-3.06752294e-01 2.65086532e-01 1.12323940e-01 6.26991034e-01
-7.96957135e-01 -3.38268816e-01 6.56459987e-01 -2.86668241e-01
-8.87651384e-01 5.72932005e-01 -1.58927590e-01 -5.65926433e-01
8.64674747e-01 -6.59779787e-01 -9.10107866e-02 -5.26257694e-01
-5.26332080e-01 -3.16705197e-01 5.97521007e-01 1.48495406e-01
1.02776134e+00 -1.56977606e+00 -3.33577275e-01 5.09478748e-01
-4.33662772e-01 1.88873783e-01 4.18743640e-01 7.17002988e-01
-1.20751250e+00 3.39530259e-01 -5.03513813e-01 -4.98439074e-01
-9.56275403e-01 6.52831852e-01 1.40962586e-01 1.01787142e-01
-9.63249266e-01 8.73759747e-01 4.01286244e-01 -2.35899594e-02
4.51840103e-01 -3.41800421e-01 -8.96175951e-02 -2.51195520e-01
6.41958177e-01 6.97270632e-02 -3.38810474e-01 -6.30188406e-01
8.68268460e-02 8.18111300e-01 3.22090164e-02 1.32003382e-01
1.43714106e+00 -1.10962763e-01 -4.23663229e-01 1.99679255e-01
1.48295677e+00 -8.82938951e-02 -1.68386376e+00 -2.51683652e-01
-5.92864573e-01 -7.88458049e-01 1.85339805e-02 -1.77700222e-01
-1.45755315e+00 1.05118108e+00 2.60036796e-01 1.18259229e-01
1.49744105e+00 -5.28473377e-01 1.13115108e+00 2.37668604e-01
3.03681076e-01 -8.28953028e-01 1.51963875e-01 2.38136053e-01
9.79305506e-01 -8.45074832e-01 1.75673351e-01 -4.96303976e-01
-2.83007234e-01 1.34802806e+00 5.62194586e-01 -4.66845125e-01
5.30587912e-01 1.37481168e-01 -2.27275014e-01 3.26146603e-01
-6.01446152e-01 2.72842258e-01 1.56367396e-03 5.15741885e-01
2.96136409e-01 -1.62758291e-01 -3.54988873e-01 2.74023831e-01
1.20235048e-01 1.73229009e-01 2.97624886e-01 9.09183741e-01
-3.42751265e-01 -1.12924588e+00 -3.51220459e-01 4.37126815e-01
-2.89243460e-01 -3.26650918e-01 1.73215911e-01 6.85718179e-01
3.68035525e-01 3.68566513e-01 2.10928574e-01 -6.01851106e-01
1.51303187e-01 -2.98834592e-01 5.18290043e-01 -2.81807125e-01
-3.19248110e-01 1.09954976e-01 -3.99812967e-01 -7.10313857e-01
-6.85145855e-01 -3.05020094e-01 -9.28317130e-01 -5.47021508e-01
5.51017784e-02 -1.80489197e-01 4.21143293e-01 5.28160632e-01
5.26220739e-01 6.39512420e-01 6.32785738e-01 -1.31611693e+00
-2.07101926e-01 -7.12308586e-01 -5.49458086e-01 3.34542453e-01
3.82387608e-01 -3.75750899e-01 -9.86073613e-02 3.13173115e-01] | [11.248608589172363, -1.5030841827392578] |
392996d4-6dbd-4d9b-b121-3c2fdf3c9178 | fine-grained-classification-of-solder-joints | 2209.09857 | null | https://arxiv.org/abs/2209.09857v1 | https://arxiv.org/pdf/2209.09857v1.pdf | Fine-grained Classification of Solder Joints with α-skew Jensen-Shannon Divergence | Solder joint inspection (SJI) is a critical process in the production of printed circuit boards (PCB). Detection of solder errors during SJI is quite challenging as the solder joints have very small sizes and can take various shapes. In this study, we first show that solders have low feature diversity, and that the SJI can be carried out as a fine-grained image classification task which focuses on hard-to-distinguish object classes. To improve the fine-grained classification accuracy, penalizing confident model predictions by maximizing entropy was found useful in the literature. Inline with this information, we propose using the {\alpha}-skew Jensen-Shannon divergence ({\alpha}-JS) for penalizing the confidence in model predictions. We compare the {\alpha}-JS regularization with both existing entropyregularization based methods and the methods based on attention mechanism, segmentation techniques, transformer models, and specific loss functions for fine-grained image classification tasks. We show that the proposed approach achieves the highest F1-score and competitive accuracy for different models in the finegrained solder joint classification task. Finally, we visualize the activation maps and show that with entropy-regularization, more precise class-discriminative regions are localized, which are also more resilient to noise. Code will be made available here upon acceptance. | ['Dincer Gokcen', 'Atila Yilmaz', 'Seniha Esen Yuksel', 'Furkan Ulger'] | 2022-09-20 | null | null | null | null | ['fine-grained-image-classification'] | ['computer-vision'] | [ 1.72029182e-01 1.87866524e-01 1.93196252e-01 -3.87009919e-01
-9.49373603e-01 -1.07248381e-01 5.81053942e-02 5.88584458e-03
-7.85510540e-02 9.63569522e-01 -3.12834322e-01 1.95033535e-01
-4.70690936e-01 -5.87711215e-01 -7.56453037e-01 -9.39001739e-01
1.28024340e-01 4.18552279e-01 6.30905926e-01 1.55195937e-01
6.00182474e-01 6.20197475e-01 -1.42800152e+00 5.04038930e-01
1.09696221e+00 1.42970812e+00 4.37331647e-01 3.06750953e-01
3.23401868e-01 4.76719737e-01 -1.07856834e+00 -4.42337722e-01
-2.41699561e-01 -1.75547987e-01 -8.12118828e-01 -1.96587548e-01
2.22291760e-02 2.22323239e-01 1.27993315e-01 1.31127667e+00
2.56580740e-01 2.83882394e-02 1.27861893e+00 -1.17008412e+00
-5.62637389e-01 2.67283052e-01 -6.62704885e-01 1.75302535e-01
1.56769142e-01 -1.60464361e-01 9.46315825e-01 -9.30299461e-01
4.57783341e-01 1.21328497e+00 7.60687828e-01 4.54674274e-01
-8.90464604e-01 -6.62959099e-01 8.44945014e-02 2.96832919e-01
-1.45969915e+00 4.41388972e-02 8.37656558e-01 -4.68487650e-01
9.60125208e-01 5.45181453e-01 3.47586930e-01 1.07281566e+00
9.40320790e-01 5.56340575e-01 1.07254457e+00 -4.86068398e-01
3.75836730e-01 2.03106910e-01 5.14895320e-01 5.97648561e-01
3.08909625e-01 -1.96488306e-01 -2.84088701e-01 5.83017394e-02
7.04282701e-01 -1.97022423e-01 -1.07776716e-01 9.12065879e-02
-8.45717609e-01 8.19571853e-01 2.65786558e-01 3.26091051e-01
-5.72805144e-02 1.13247260e-01 3.56894493e-01 2.99224891e-02
7.17155039e-01 5.13234138e-01 -3.94804388e-01 -1.00884058e-01
-8.03850472e-01 9.65449437e-02 7.66575098e-01 8.99708807e-01
7.14084983e-01 -1.11870989e-01 -2.50805497e-01 1.26247084e+00
6.24085605e-01 1.83890730e-01 3.93655062e-01 -1.06341672e+00
1.97507188e-01 4.22087342e-01 -3.02950650e-01 -7.76321113e-01
-3.31810296e-01 -3.13113570e-01 -1.00661397e+00 4.98804241e-01
2.09015086e-01 2.02336192e-01 -7.30536342e-01 1.39898372e+00
9.26300064e-02 -3.11466724e-01 -5.76797187e-01 8.67476940e-01
4.78266209e-01 6.17661178e-01 1.79500714e-01 -5.52057028e-02
1.53652787e+00 -8.63353848e-01 -6.93471014e-01 -1.66939512e-01
2.07576960e-01 -7.72902191e-01 1.06970191e+00 5.17417371e-01
-6.12556100e-01 -5.73123932e-01 -1.32928896e+00 2.93686688e-01
-4.29094851e-01 2.30182052e-01 4.04636264e-01 8.37593079e-01
-8.22920799e-01 1.00408113e+00 -9.25205588e-01 -3.80702019e-01
5.34455895e-01 5.30985415e-01 -3.67970526e-01 1.46717995e-01
-1.00749028e+00 8.17643344e-01 1.88219607e-01 2.08161056e-01
-7.18668878e-01 -5.90481699e-01 -7.92275488e-01 -3.35283168e-02
1.31895185e-01 -1.83901042e-01 7.82053947e-01 -5.18346429e-01
-1.34986711e+00 7.76041448e-01 2.23481372e-01 -1.16726018e-01
2.17368960e-01 -2.85082012e-01 -6.80800807e-03 5.75658530e-02
2.43852243e-01 4.51210916e-01 9.49498475e-01 -1.31664217e+00
-1.88172802e-01 -4.18052524e-01 -1.70844942e-01 -4.76432592e-01
-2.21597910e-01 1.60406440e-01 -2.05157921e-01 -9.18484986e-01
1.73168629e-01 -9.41964328e-01 -1.12816997e-01 -1.01138569e-01
-7.87185490e-01 -4.78422195e-01 8.99836242e-01 -7.60966718e-01
1.13755059e+00 -2.13376236e+00 -3.08425784e-01 3.06818664e-01
1.00816391e-01 -7.74786174e-02 1.44721746e-01 4.32565026e-02
-1.53178528e-01 2.44548649e-01 -3.68596345e-01 -1.58326644e-02
2.79386461e-01 6.07972685e-03 3.35601240e-01 5.07944226e-01
5.98312616e-01 5.19351840e-01 -4.09820706e-01 -6.19479835e-01
1.78244412e-01 1.37616694e-01 -4.93346274e-01 2.12446988e-01
5.82592040e-02 2.96485126e-01 -5.69201052e-01 7.62000382e-01
7.56689727e-01 -3.05591851e-01 -1.30036771e-01 -7.60830224e-01
1.39580920e-01 -1.99826032e-01 -9.79237735e-01 1.28674841e+00
-4.30479497e-01 4.89617318e-01 8.37045014e-02 -1.15137446e+00
1.13044214e+00 1.03115477e-01 3.61090690e-01 -5.94846845e-01
1.57211363e-01 1.95073441e-01 1.26787812e-01 -2.12700933e-01
7.73163959e-02 -1.24263264e-01 -4.25207138e-01 2.21154392e-01
2.91711360e-01 -1.74667805e-01 2.11206719e-01 -1.35699928e-01
1.04525948e+00 1.10312633e-01 -2.41588857e-02 -7.96447635e-01
3.67548376e-01 -3.28064352e-01 8.37136269e-01 6.84487700e-01
-4.29631919e-02 1.04328442e+00 5.27309418e-01 -4.48491909e-02
-9.81155992e-01 -1.22291565e+00 -4.98033047e-01 7.15154707e-01
3.01162958e-01 -1.00935057e-01 -1.22565603e+00 -7.85912097e-01
-5.16941249e-02 5.78027546e-01 -6.40640795e-01 -5.14284492e-01
-4.20533329e-01 -1.13863134e+00 3.35653961e-01 6.36068225e-01
4.51188564e-01 -1.16249216e+00 -2.57768929e-01 1.49164215e-01
2.96517368e-02 -9.59674835e-01 -4.16751742e-01 8.16527426e-01
-7.56414711e-01 -1.05265653e+00 -8.15090120e-01 -9.88611519e-01
7.81049252e-01 -4.57245022e-01 1.19310701e+00 -9.79741290e-02
-7.70022154e-01 2.04512447e-01 -4.54766244e-01 -2.54013300e-01
-4.07986283e-01 -1.02639981e-02 -9.38234478e-02 -1.83921665e-01
1.02044076e-01 -3.65229309e-01 -5.80502629e-01 5.42631388e-01
-6.65470302e-01 -2.14586198e-01 6.59702718e-01 9.51373279e-01
6.77526474e-01 2.54322946e-01 8.67229462e-01 -9.16299105e-01
6.26957774e-01 -2.68050283e-01 -2.92450368e-01 3.41556072e-01
-6.04547262e-01 1.25384003e-01 5.50586820e-01 -1.18823104e-01
-1.25345588e+00 -3.12204123e-01 -6.29954875e-01 -2.71509856e-01
-3.76567841e-01 7.89323375e-02 -5.63900650e-01 -3.19083869e-01
2.49798000e-01 -2.00046241e-01 -2.04401374e-01 -7.12125659e-01
-2.41341949e-01 7.28600264e-01 8.97970498e-02 -8.27767134e-01
2.75084645e-01 -6.53389990e-02 -1.09267212e-01 -9.10091996e-01
-5.75311959e-01 -6.50906935e-02 -5.64350367e-01 -3.60340536e-01
1.14677954e+00 -4.11712706e-01 -8.27272892e-01 7.47732878e-01
-1.34702384e+00 -1.83064133e-01 -2.25141034e-01 3.01537424e-01
-4.50443089e-01 6.29236519e-01 -1.01409507e+00 -7.86069632e-01
-3.98093253e-01 -1.44003224e+00 1.33282363e+00 2.13178620e-01
-4.61449832e-01 -9.15473640e-01 -2.67572612e-01 2.94383109e-01
1.97167769e-01 1.45431176e-01 1.43422389e+00 -7.78148413e-01
-3.52175623e-01 -3.57339114e-01 -2.94649035e-01 7.07125068e-01
1.47261545e-01 2.78489962e-02 -1.01531911e+00 -1.75304964e-01
2.65015036e-01 -2.21338436e-01 8.18894267e-01 6.18011177e-01
1.56097960e+00 -1.17098518e-01 -3.37032586e-01 1.44694731e-01
1.28409433e+00 6.02701068e-01 7.60728300e-01 2.80533761e-01
5.01459122e-01 7.60368466e-01 8.93092275e-01 6.06270611e-01
-1.20261043e-01 7.32989311e-01 5.15574396e-01 6.21097535e-02
-3.52292925e-01 3.10553670e-01 1.61597475e-01 1.03905499e+00
-5.71074411e-02 -2.71131784e-01 -6.27228439e-01 4.38112974e-01
-1.56797147e+00 -4.80172127e-01 -5.74211478e-02 2.03802204e+00
7.40761518e-01 4.28032368e-01 -1.87155619e-01 2.92680919e-01
9.44385409e-01 -7.23979175e-02 -4.25914824e-01 -5.43138564e-01
2.91574951e-02 4.47761685e-01 2.60102063e-01 4.20244098e-01
-1.13786554e+00 6.37647390e-01 5.91636705e+00 1.43666780e+00
-7.53816307e-01 1.90372482e-01 1.14849019e+00 1.84187964e-01
-1.77799240e-01 -3.71958286e-01 -8.46604705e-01 1.06463456e+00
7.64122784e-01 5.08589029e-01 -1.85723856e-01 8.73087943e-01
-1.20931976e-01 -3.51480573e-01 -1.26679873e+00 8.25329840e-01
-1.46795899e-01 -1.17299020e+00 -2.04354271e-01 -1.55445427e-01
5.97511947e-01 -5.09292006e-01 4.72177975e-02 -8.44417978e-03
1.23994365e-01 -9.56468642e-01 8.77145231e-01 5.23151755e-01
7.76643872e-01 -7.02564478e-01 1.01882267e+00 -6.40385970e-02
-1.17399812e+00 8.14965889e-02 -5.78138947e-01 3.46205533e-01
-1.08398370e-01 1.15842342e+00 -3.92066121e-01 5.72173715e-01
1.21613503e+00 4.74165678e-01 -4.91998792e-01 1.03791165e+00
1.50050838e-02 5.51559269e-01 -2.28530794e-01 -3.11736256e-01
-1.52618036e-01 -4.12499636e-01 2.80877441e-01 1.29288864e+00
4.27222461e-01 -3.60850006e-01 -5.66367470e-02 1.04150069e+00
1.58359736e-01 -2.80377299e-01 -2.97509640e-01 3.34447265e-01
4.38365787e-01 9.82662380e-01 -1.05821192e+00 1.40871763e-01
-2.87499398e-01 1.14569438e+00 2.43272576e-02 2.54233360e-01
-8.77432346e-01 -8.11917901e-01 6.93348050e-01 3.05480137e-02
2.48987451e-01 6.04354627e-02 -7.32074261e-01 -6.29616976e-01
1.75677970e-01 -5.44964433e-01 1.59266099e-01 -7.30875909e-01
-1.67663777e+00 7.80302465e-01 5.78616932e-02 -1.11410153e+00
7.83033818e-02 -1.02141857e+00 -6.20575309e-01 6.18490815e-01
-1.18328905e+00 -9.95371997e-01 -2.39451528e-01 2.93647647e-01
8.51642907e-01 -2.15012655e-02 6.01222575e-01 4.40779299e-01
-7.48827219e-01 7.84147143e-01 2.48053789e-01 7.64532089e-02
5.18917918e-01 -1.21866608e+00 1.27291054e-01 3.82656157e-01
-3.17904383e-01 3.45147341e-01 6.82062566e-01 -8.19349647e-01
-1.10825622e+00 -1.16954184e+00 4.13153440e-01 1.95699073e-02
6.04981422e-01 -4.87997413e-01 -9.66968656e-01 2.96798527e-01
1.78987354e-01 -1.51852909e-02 5.46233475e-01 -6.30814210e-02
1.12099856e-01 -4.57253993e-01 -1.42071462e+00 2.66157001e-01
7.20013022e-01 -3.14226002e-01 -3.51668447e-01 4.87024516e-01
4.48989511e-01 1.47589564e-01 -1.40799749e+00 5.72704673e-01
4.86922145e-01 -7.76913404e-01 9.86565471e-01 -3.37530449e-02
3.25957566e-01 -2.80155912e-02 -1.72775418e-01 -1.12458038e+00
-5.38791716e-01 -1.02000780e-01 3.15838903e-01 1.49815273e+00
5.52049339e-01 -6.14275753e-01 1.02048218e+00 5.80108166e-01
-5.26610494e-01 -1.07027435e+00 -1.33426726e+00 -9.03048754e-01
1.65412933e-01 -4.16038632e-01 2.83213943e-01 2.87029088e-01
-1.23580292e-01 -1.06509455e-01 -2.39053324e-01 1.38695017e-01
8.34384918e-01 -1.41393811e-01 -1.54082060e-01 -1.52395141e+00
-7.92054012e-02 -3.38598967e-01 -7.76211023e-01 -6.58360124e-01
1.46654934e-01 -6.82228684e-01 5.44773161e-01 -1.39386928e+00
3.55014861e-01 -6.81141794e-01 -5.38008571e-01 4.61718529e-01
-1.32305130e-01 5.80966711e-01 -2.94552706e-02 1.16718642e-01
-6.99056566e-01 5.98527491e-01 1.21910894e+00 -4.59436715e-01
4.55299675e-01 9.33556706e-02 -3.17708582e-01 6.91395819e-01
6.35697842e-01 -6.46164894e-01 -1.35204419e-01 9.37867239e-02
2.19191946e-02 1.69799346e-02 2.79356211e-01 -1.31469738e+00
6.50938079e-02 1.47131801e-01 5.84131420e-01 -5.04825413e-01
3.96614462e-01 -8.06821823e-01 2.54680425e-01 5.08210242e-01
1.79279200e-03 -3.62981737e-01 1.60631999e-01 4.27065849e-01
-3.71919185e-01 -7.17946470e-01 1.05599606e+00 -2.05285445e-01
-6.72675252e-01 8.38772431e-02 -8.09304535e-01 -5.43439500e-02
1.08903265e+00 -4.28900987e-01 -1.08472191e-01 1.32271722e-01
-8.14742327e-01 -2.50796169e-01 3.49045098e-01 2.78086394e-01
8.21972609e-01 -1.45721447e+00 -5.36910057e-01 2.53247470e-01
-8.33557695e-02 -1.80542618e-01 4.12004650e-01 8.68109405e-01
-6.20982707e-01 2.74682760e-01 -3.15284312e-01 -7.56068230e-01
-1.18032169e+00 1.28220618e-01 1.86202705e-01 -4.78606492e-01
-3.48039597e-01 1.16606832e+00 6.38130486e-01 -2.82310337e-01
1.48912415e-01 -5.23950636e-01 -1.79872274e-01 -5.90640008e-02
-2.52970052e-03 6.03532732e-01 5.12110591e-01 -1.81012660e-01
-7.47346818e-01 9.03444648e-01 -2.63146996e-01 4.25109744e-01
1.33491600e+00 5.66292219e-02 -4.04124320e-01 3.25906336e-01
1.37411582e+00 -4.09245908e-01 -1.44813788e+00 4.80443120e-01
3.64170104e-01 -2.29067728e-01 -2.80384928e-01 -8.66823256e-01
-1.24044871e+00 1.04911041e+00 9.51670706e-01 2.01376155e-01
1.12443686e+00 3.47600788e-01 6.82639956e-01 8.89319256e-02
5.50454855e-01 -1.35431314e+00 3.26108128e-01 4.54113513e-01
9.79625583e-01 -1.16035771e+00 2.17229612e-02 -8.97883534e-01
-5.31569898e-01 1.19329667e+00 7.01545477e-01 -3.18270028e-01
8.50229621e-01 7.73865342e-01 -2.29290754e-01 -1.88245386e-01
-3.45823735e-01 4.68069017e-01 4.17807400e-01 8.49967062e-01
4.77853209e-01 -7.63702393e-02 -2.48339087e-01 1.31990409e+00
1.86819211e-01 -4.16154444e-01 1.56627640e-01 8.64145935e-01
-5.25371552e-01 -9.87704933e-01 -4.89127576e-01 7.83911049e-01
-6.52210653e-01 6.20459504e-02 -9.26466733e-02 6.82443857e-01
1.36441439e-01 9.70058978e-01 1.06151603e-01 -4.13852364e-01
1.76502854e-01 -1.52277663e-01 6.11117184e-01 -4.75057214e-01
-5.49705505e-01 -3.73478723e-03 2.15945765e-02 -6.16298854e-01
4.38402407e-02 -5.12600183e-01 -1.15623546e+00 1.89969435e-01
-6.17485523e-01 3.09445053e-01 6.09857380e-01 1.05403149e+00
1.66879594e-01 1.04868448e+00 5.07778645e-01 -1.12445569e+00
-3.89022350e-01 -1.05852199e+00 -1.10770166e+00 2.46683091e-01
-9.78470221e-02 -1.03779376e+00 -3.49062502e-01 -7.52387494e-02] | [7.578438758850098, 1.8695734739303589] |
4f86aec7-745b-47e2-9f61-be9bb5bcc109 | physics-informed-machine-learning-for-3 | 2306.13867 | null | https://arxiv.org/abs/2306.13867v1 | https://arxiv.org/pdf/2306.13867v1.pdf | Physics-Informed Machine Learning for Modeling and Control of Dynamical Systems | Physics-informed machine learning (PIML) is a set of methods and tools that systematically integrate machine learning (ML) algorithms with physical constraints and abstract mathematical models developed in scientific and engineering domains. As opposed to purely data-driven methods, PIML models can be trained from additional information obtained by enforcing physical laws such as energy and mass conservation. More broadly, PIML models can include abstract properties and conditions such as stability, convexity, or invariance. The basic premise of PIML is that the integration of ML and physics can yield more effective, physically consistent, and data-efficient models. This paper aims to provide a tutorial-like overview of the recent advances in PIML for dynamical system modeling and control. Specifically, the paper covers an overview of the theory, fundamental concepts and methods, tools, and applications on topics of: 1) physics-informed learning for system identification; 2) physics-informed learning for control; 3) analysis and verification of PIML models; and 4) physics-informed digital twins. The paper is concluded with a perspective on open challenges and future research opportunities. | ['Draguna L. Vrabie', 'Wenceslao Shaw Cortez', 'Melanie N. Zeilinger', 'Andrea Carron', 'Joel A. Paulson', 'Stefano Di Cairano', 'Ankush Chakrabarty', 'Biswadip Dey', 'Roland Schwan', 'Zoltan Nagy', 'Colin Jones', 'Ján Drgoňa', 'Truong X. Nghiem'] | 2023-06-24 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [-9.77077633e-02 -8.36158618e-02 -7.32177615e-01 1.24042153e-01
-1.30978078e-01 -4.76264596e-01 7.39898264e-01 2.79772162e-01
3.52920681e-01 8.50466967e-01 -4.71093029e-01 -3.48346770e-01
-5.95082402e-01 -5.24581313e-01 -8.19013238e-01 -9.00430918e-01
-2.85724908e-01 3.64184648e-01 -1.66643992e-01 -1.67479336e-01
2.61053711e-01 1.06345737e+00 -1.66273916e+00 -3.31873029e-01
8.59875381e-01 9.76466298e-01 1.04940735e-01 8.57081354e-01
-7.55039826e-02 8.90942216e-01 -1.29233852e-01 4.06139076e-01
1.07761003e-01 -5.39405704e-01 -4.39338863e-01 1.15447557e-02
1.84585124e-01 -5.02262171e-03 -6.06156647e-01 8.75517547e-01
2.45198622e-01 4.35203314e-01 8.78403127e-01 -1.63849378e+00
-7.47621715e-01 2.54859895e-01 -1.72321200e-01 -8.32213983e-02
1.60105452e-01 4.50053990e-01 7.16403723e-01 -9.01438236e-01
2.23993644e-01 1.45229769e+00 7.13269532e-01 8.15519869e-01
-1.34659195e+00 -4.56578642e-01 4.14636493e-01 1.95743084e-01
-1.13162029e+00 -1.79134130e-01 1.04380929e+00 -8.26545656e-01
7.94128597e-01 3.03601950e-01 8.30661535e-01 9.20012593e-01
8.02333653e-01 7.09429264e-01 1.08158791e+00 -7.95363486e-01
7.21735895e-01 4.18265045e-01 2.75337309e-01 9.18899477e-01
4.31558341e-01 8.26566517e-01 -6.98347151e-01 -2.09363908e-01
8.17944825e-01 -5.38121760e-01 1.06385842e-01 -9.97392952e-01
-8.67981374e-01 7.59145498e-01 9.46608838e-03 -1.53496370e-01
-1.53444886e-01 2.76251048e-01 2.63411283e-01 -8.35549310e-02
1.72389165e-01 8.26783299e-01 -8.89441907e-01 2.73206711e-01
-6.63707972e-01 6.38250828e-01 9.94944692e-01 9.99582648e-01
6.70608222e-01 7.16071188e-01 2.15133190e-01 4.95867193e-01
6.21835768e-01 8.39288652e-01 1.91167369e-01 -1.34945738e+00
-2.35170558e-01 3.56742173e-01 4.22346234e-01 -7.90544927e-01
-5.42930424e-01 -6.12014681e-02 -8.32044601e-01 6.42002404e-01
3.51522155e-02 -2.55233228e-01 -7.82579005e-01 1.61851597e+00
2.16801748e-01 4.40337181e-01 -7.33885691e-02 7.02112019e-01
8.40828121e-01 9.77655768e-01 1.05853088e-01 -4.81998444e-01
8.77445877e-01 -6.11980140e-01 -8.85274649e-01 -5.50293662e-02
4.23973501e-01 -2.47935683e-01 1.05484104e+00 2.87078232e-01
-1.20071781e+00 -8.27766359e-01 -1.20921969e+00 1.64365664e-01
-7.26798117e-01 -2.24664986e-01 6.48902357e-01 6.32200956e-01
-6.44453943e-01 1.12238932e+00 -1.32820010e+00 -3.35906148e-01
7.64226832e-04 3.93220633e-01 2.41890773e-01 6.57549620e-01
-1.17437899e+00 1.37652349e+00 2.13302627e-01 -1.40085161e-01
-8.12819183e-01 -1.46087468e+00 -1.02630579e+00 -3.40173990e-01
1.39750957e-01 -7.86594152e-01 1.33689010e+00 -2.37249941e-01
-2.16841364e+00 5.31667531e-01 -7.26746172e-02 -5.20491123e-01
1.95545375e-01 -2.95065045e-01 -3.37429494e-01 -1.39905632e-01
-5.18190622e-01 2.80442715e-01 6.71498060e-01 -1.44751275e+00
-4.15753484e-01 1.76151261e-01 -2.53161136e-02 -1.19157396e-01
-1.08759679e-01 -3.26240003e-01 1.48688760e-02 -2.58352131e-01
-9.79152918e-02 -9.02287841e-01 -4.44121212e-01 4.36949134e-01
-5.56562185e-01 -1.86968595e-01 1.28336561e+00 -4.67043459e-01
1.04080653e+00 -1.57804585e+00 4.90025282e-01 6.89825639e-02
-1.29691279e-02 5.45072079e-01 2.49936625e-01 7.11028337e-01
-1.88917428e-01 1.18908860e-01 -7.67016709e-02 7.09520373e-03
3.43062997e-01 2.12618992e-01 -5.75341105e-01 6.33057237e-01
3.56440514e-01 8.87615085e-01 -8.72598588e-01 -4.95397359e-01
1.40646899e+00 3.73928219e-01 -2.22533360e-01 1.03110708e-01
-5.04425466e-01 9.65134740e-01 -6.80409014e-01 6.09775841e-01
4.67984319e-01 -1.68350697e-01 -6.23795725e-02 -4.17573065e-01
-3.56426328e-01 8.55529383e-02 -1.26841831e+00 1.16467106e+00
-7.73950756e-01 5.28386772e-01 2.35680163e-01 -1.32798207e+00
8.52119267e-01 1.68502480e-01 1.04800093e+00 -5.78994215e-01
2.30276257e-01 8.59311968e-02 -3.02700460e-01 -5.91417909e-01
-5.42499945e-02 -1.11355200e-01 -1.72285035e-01 1.03020728e-01
5.68628199e-02 -8.51004779e-01 3.14134397e-02 -2.00396270e-01
5.04956126e-01 6.02078199e-01 7.34497190e-01 -6.71152055e-01
9.83259201e-01 1.65835440e-01 6.32933140e-01 6.73550427e-01
-3.65081996e-01 -1.98742151e-01 -1.49409458e-01 -5.22977948e-01
-1.06541705e+00 -1.35459793e+00 -3.46369058e-01 6.54915452e-01
4.78310704e-01 -2.89345473e-01 -4.64680612e-01 1.74174979e-02
6.21010125e-01 9.56755519e-01 -4.69890893e-01 -5.95526874e-01
-7.70301461e-01 -7.31888652e-01 -1.03086784e-01 6.42631531e-01
1.43998504e-01 -8.19313109e-01 -3.24480087e-01 1.55858561e-01
4.31101620e-01 -1.12768400e+00 1.21725924e-01 -3.73936519e-02
-1.00172687e+00 -1.31389356e+00 -5.22802100e-02 -5.41108668e-01
3.45472306e-01 -2.17975467e-01 1.05756128e+00 -1.34637997e-01
-6.54629707e-01 7.51812637e-01 2.07382031e-02 -7.73074806e-01
-7.83760369e-01 -4.21562403e-01 7.64961064e-01 -4.39171374e-01
-6.41137213e-02 -5.88348806e-01 -2.19170898e-01 8.42537954e-02
-2.24526584e-01 1.28691137e-01 3.36152434e-01 6.58847630e-01
7.51244009e-01 3.45815718e-01 5.85663140e-01 -2.34862491e-01
3.30566585e-01 -2.88562868e-02 -1.09233725e+00 2.01304644e-01
-7.09343135e-01 5.12468480e-02 1.02061856e+00 -5.14708340e-01
-1.08121562e+00 3.54930580e-01 8.82966295e-02 -6.57126784e-01
-2.51205027e-01 2.15745002e-01 -4.21323448e-01 -4.27931190e-01
4.89367038e-01 2.29769573e-01 1.38617054e-01 -6.23234034e-01
6.14779890e-01 3.00123364e-01 5.40897429e-01 -1.03147614e+00
1.06950212e+00 3.21767241e-01 5.95314801e-01 -1.31229436e+00
-9.71825242e-01 -1.75095171e-01 -9.40055490e-01 -3.04196537e-01
6.53481305e-01 -3.95313084e-01 -1.03815508e+00 5.17908394e-01
-7.24817872e-01 -4.99627590e-01 -8.23975146e-01 6.13703012e-01
-1.22781837e+00 1.82961792e-01 -6.06961787e-01 -1.35735619e+00
-1.34335250e-01 -1.01749480e+00 7.24257588e-01 4.13531840e-01
-3.41122746e-01 -1.79207802e+00 2.06086650e-01 -1.02846816e-01
1.28181815e-01 7.44863689e-01 1.36643338e+00 -1.61358133e-01
-3.54378492e-01 -2.40075931e-01 3.58153135e-01 3.46761495e-01
4.12349403e-01 6.76396549e-01 -1.01222503e+00 -4.52763915e-01
7.19652101e-02 -2.63160735e-01 4.56282079e-01 9.24556136e-01
1.30884504e+00 -2.14488193e-01 -7.85321772e-01 3.77565056e-01
1.38444436e+00 3.82639140e-01 -1.70986533e-01 -2.42742077e-02
6.57985449e-01 5.40915251e-01 5.59610665e-01 4.27262098e-01
-9.23707783e-02 6.11416817e-01 3.70292455e-01 -1.26431212e-01
-2.96700686e-01 -3.30052257e-01 3.44432801e-01 1.00552750e+00
1.78322613e-01 -1.56860948e-02 -6.92385554e-01 2.35246539e-01
-1.82828534e+00 -8.65346134e-01 -1.31447613e-01 2.12310958e+00
6.06236935e-01 -7.78893232e-02 -6.83444028e-04 1.72932342e-01
1.00312054e+00 -1.70608342e-01 -1.22875071e+00 -6.21426344e-01
1.62028242e-02 6.24280088e-02 4.66295391e-01 7.32322991e-01
-1.24577010e+00 7.45391846e-01 7.94033384e+00 6.55562818e-01
-1.21426725e+00 -1.90012679e-01 1.29655138e-01 1.89852640e-01
2.28721201e-01 -9.65426564e-02 -1.01591229e+00 4.68653142e-01
1.11087549e+00 -6.50344014e-01 3.58502746e-01 9.74340856e-01
9.46003795e-01 1.31799996e-01 -1.46072960e+00 7.25934029e-01
-5.54289877e-01 -1.76630247e+00 2.73898423e-01 7.36062005e-02
8.56973350e-01 -2.23114565e-01 1.54335737e-01 1.99220702e-01
3.76649708e-01 -8.17811012e-01 9.34137225e-01 9.16750908e-01
5.96243858e-01 -5.21085024e-01 3.22507024e-01 5.70999324e-01
-1.26286674e+00 -3.13414842e-01 -2.65123695e-01 -1.92356870e-01
2.07997277e-01 6.87579870e-01 6.20702356e-02 6.60610735e-01
4.88327295e-01 1.18109405e+00 -3.56178172e-02 1.02941561e+00
-6.86944276e-02 4.85026062e-01 -2.91911900e-01 -2.30863556e-01
-2.03962550e-01 -2.16667250e-01 8.00625503e-01 1.13921285e+00
-7.83449858e-02 -2.73315031e-02 3.93102765e-01 1.38723052e+00
5.46327591e-01 -3.69520992e-01 -5.72839797e-01 -1.72176152e-01
6.82791233e-01 1.10402787e+00 -2.90634841e-01 -2.48263717e-01
-3.58163625e-01 2.37020850e-02 -1.93001211e-01 2.64350057e-01
-1.14367199e+00 -1.16314329e-01 8.72352123e-01 -4.16713953e-02
-1.06635980e-01 -5.87795079e-01 -3.95917684e-01 -9.14863765e-01
-4.78492737e-01 -4.63543355e-01 9.52918604e-02 -5.95102608e-01
-1.54523337e+00 -2.10384250e-01 4.98084128e-01 -1.17374086e+00
-2.79746771e-01 -1.28224802e+00 -7.28208780e-01 6.72159255e-01
-1.41752779e+00 -1.07105553e+00 -8.09469149e-02 1.82956204e-01
5.57383120e-01 -1.86271980e-01 9.86392319e-01 -1.06808439e-01
-8.96984458e-01 1.14924230e-01 5.35598934e-01 -3.81498814e-01
2.14943409e-01 -1.39466763e+00 3.56182545e-01 4.46880370e-01
-3.21037024e-01 4.62911010e-01 1.17693341e+00 -5.77070892e-01
-2.17707658e+00 -1.27930427e+00 2.47953400e-01 -5.50447702e-01
9.72626746e-01 -3.50817442e-01 -7.75919735e-01 3.63120973e-01
-1.39864072e-01 -7.49104768e-02 2.97950566e-01 -1.46488860e-01
2.62524933e-01 -2.21463591e-01 -1.24722719e+00 5.40205240e-01
7.45848238e-01 -2.30394766e-01 -4.88906890e-01 7.31043339e-01
5.46352327e-01 -4.36582625e-01 -1.00965035e+00 5.63246489e-01
6.26697004e-01 -4.13683891e-01 1.30625796e+00 -7.80191123e-01
-1.56804398e-01 -2.77003467e-01 -5.89263961e-02 -1.19356406e+00
-5.39501071e-01 -9.03766096e-01 -1.13824630e+00 9.34644341e-01
1.90277815e-01 -6.35468781e-01 6.53614759e-01 7.45375633e-01
-2.99912184e-01 -8.68486166e-01 -7.03670263e-01 -1.47935808e+00
7.64916897e-01 -4.29895580e-01 1.78536341e-01 8.01148295e-01
3.38092357e-01 -2.61466652e-02 -1.06800348e-01 4.91436630e-01
8.16632450e-01 2.91714400e-01 5.11644781e-01 -1.49045980e+00
-6.04613172e-03 -6.43229783e-01 -2.28252932e-01 -5.83198369e-01
5.35144508e-01 -5.16726911e-01 8.15667585e-02 -1.63979936e+00
-3.79595235e-02 -3.69621098e-01 -3.30210596e-01 1.30977944e-01
2.70164222e-01 -2.91580617e-01 2.51946086e-03 1.97963268e-01
-1.49893820e-01 7.74745286e-01 1.25705755e+00 -9.52313244e-02
-5.67602932e-01 2.06525996e-01 -2.63547540e-01 1.03093839e+00
1.20505333e+00 1.50877357e-01 -6.51334882e-01 2.44869366e-01
-2.10433275e-01 -5.43495268e-02 8.21523666e-01 -1.07352221e+00
2.48202384e-01 -9.25614119e-01 3.66078019e-01 -8.02697659e-01
6.34632707e-01 -7.03079820e-01 -2.12159783e-01 8.67375910e-01
-2.37151682e-01 -3.20902258e-01 6.69604719e-01 5.86161435e-01
2.62471497e-01 8.05008337e-02 1.40718758e+00 -1.26285002e-01
-7.72563636e-01 2.77051747e-01 -8.27687144e-01 -1.24465480e-01
1.25942254e+00 -2.12347612e-01 -1.84838533e-01 -2.23631829e-01
-8.40030551e-01 3.09225827e-01 1.90849751e-01 4.46860552e-01
3.61428231e-01 -1.16183674e+00 -3.72494400e-01 1.64279968e-01
-1.86781287e-01 -4.46302295e-01 1.16972979e-02 5.22451341e-01
-3.25221092e-01 9.74856973e-01 -2.39298329e-01 -8.39948416e-01
-8.55305016e-01 8.47010314e-01 8.95200014e-01 2.31228992e-01
-6.31494224e-01 4.08215612e-01 1.51656508e-01 -7.50058413e-01
1.49475172e-01 -5.75996995e-01 1.88061550e-01 -6.85068369e-01
2.86343426e-01 8.62582088e-01 -7.89549872e-02 -5.07633507e-01
-5.31951010e-01 1.05542791e+00 4.30106223e-01 6.18564449e-02
1.04245424e+00 -2.07948476e-01 1.44049555e-01 7.49816179e-01
6.21828020e-01 -3.62707645e-01 -1.52560103e+00 1.68816224e-02
4.29110527e-02 2.51429677e-01 3.06377858e-01 -9.29997087e-01
-8.40431511e-01 9.87536788e-01 5.91115475e-01 1.44045308e-01
7.47900724e-01 -1.24693610e-01 5.50499856e-01 5.48983276e-01
4.14995044e-01 -1.34583402e+00 -6.30732253e-03 5.10554910e-01
1.00243592e+00 -9.45734322e-01 3.87808502e-01 -5.45679152e-01
9.49465707e-02 1.29028034e+00 8.74695361e-01 -3.73072386e-01
1.40679789e+00 6.92485273e-01 -4.30466533e-01 1.19618185e-01
-7.21960247e-01 2.19956920e-01 5.88709652e-01 9.64071035e-01
2.16167986e-01 -1.47133887e-01 1.24539003e-01 5.52549601e-01
-1.28710553e-01 -1.42941624e-02 2.07205966e-01 1.19159019e+00
-7.50225842e-01 -1.07491231e+00 -4.13732201e-01 1.59902588e-01
2.05540985e-01 4.61736590e-01 -4.64960575e-01 1.26292479e+00
1.53715879e-01 1.14672959e+00 -4.76469882e-02 -3.63357484e-01
4.22572672e-01 2.95526441e-02 7.51361012e-01 -4.35343295e-01
1.00810476e-01 -2.10617527e-01 -2.53823131e-01 -5.92039764e-01
-3.89174074e-01 -6.42264962e-01 -1.56288970e+00 -6.79569542e-01
-4.58835989e-01 5.85712306e-02 8.40586603e-01 1.17734921e+00
3.78046811e-01 6.71895087e-01 5.31066298e-01 -1.23335695e+00
-5.05645216e-01 -4.20794129e-01 -5.24246573e-01 -2.14814618e-01
6.19698584e-01 -1.20773423e+00 -5.98664522e-01 3.84815454e-01] | [6.4650702476501465, 3.592512845993042] |
703acb15-1b58-437f-a7f0-cfa71759a407 | chatgpt-as-a-mapping-assistant-a-novel-method | 2306.03204 | null | https://arxiv.org/abs/2306.03204v1 | https://arxiv.org/pdf/2306.03204v1.pdf | ChatGPT as a mapping assistant: A novel method to enrich maps with generative AI and content derived from street-level photographs | This paper explores the concept of leveraging generative AI as a mapping assistant for enhancing the efficiency of collaborative mapping. We present results of an experiment that combines multiple sources of volunteered geographic information (VGI) and large language models (LLMs). Three analysts described the content of crowdsourced Mapillary street-level photographs taken along roads in a small test area in Miami, Florida. GPT-3.5-turbo was instructed to suggest the most appropriate tagging for each road in OpenStreetMap (OSM). The study also explores the utilization of BLIP-2, a state-of-the-art multimodal pre-training method as an artificial analyst of street-level photographs in addition to human analysts. Results demonstrate two ways to effectively increase the accuracy of mapping suggestions without modifying the underlying AI models: by (1) providing a more detailed description of source photographs, and (2) combining prompt engineering with additional context (e.g. location and objects detected along a road). The first approach increases the suggestion accuracy by up to 29%, and the second one by up to 20%. | ['Boyuan Guan', 'Hartwig H. Hochmair', 'Peter Mooney', 'Levente Juhász'] | 2023-06-05 | null | null | null | null | ['prompt-engineering'] | ['natural-language-processing'] | [ 1.44780293e-01 8.80249083e-01 3.92229706e-01 -4.19170409e-01
-1.03534067e+00 -6.86907113e-01 8.55249882e-01 4.39998180e-01
-4.29320514e-01 6.69238687e-01 6.83784664e-01 -4.79056954e-01
-1.06979571e-01 -1.04263508e+00 -6.36117816e-01 -2.96784997e-01
-1.24847509e-01 9.55734313e-01 2.48084053e-01 -3.19772780e-01
5.61903179e-01 5.56818187e-01 -1.72384274e+00 4.21872824e-01
1.05187356e+00 2.72896588e-01 6.34050548e-01 6.82936013e-01
-5.81508160e-01 1.04763627e+00 -7.51410127e-01 -4.18062806e-01
2.03417525e-01 -2.25485489e-02 -7.97272921e-01 -3.54209021e-02
7.00492799e-01 -3.80235791e-01 -8.00404623e-02 7.71295905e-01
6.19009972e-01 2.41654992e-01 4.65920478e-01 -1.47021997e+00
-8.74974191e-01 8.07431161e-01 -3.88867825e-01 1.11936018e-01
5.24858713e-01 6.30295128e-02 3.99372220e-01 -1.11537302e+00
7.08073139e-01 1.35244036e+00 6.51041925e-01 7.97629952e-02
-8.11614215e-01 -7.47816980e-01 1.15622424e-01 -1.18207462e-01
-1.99660456e+00 -4.38904017e-01 6.47963285e-01 -6.53559089e-01
1.06067073e+00 2.96326995e-01 3.75729203e-01 5.39808869e-01
-2.42594063e-01 2.99164176e-01 1.20088971e+00 -5.99871159e-01
4.05358344e-01 6.53438509e-01 -2.76117653e-01 8.20199490e-01
9.32018161e-02 -1.67069823e-01 -8.46438468e-01 -6.56766772e-01
6.34695530e-01 -4.26107705e-01 1.94775179e-01 1.24980845e-01
-1.29994738e+00 7.72689164e-01 5.47398925e-01 3.09297264e-01
-5.68106115e-01 3.34252650e-03 -1.66495577e-01 -3.11671823e-01
8.74483049e-01 6.77478731e-01 1.88920468e-01 -1.30572066e-01
-9.86018002e-01 4.73904669e-01 8.45005095e-01 1.00847316e+00
1.06954110e+00 -1.26767889e-01 -5.37041686e-02 7.52746820e-01
4.45763171e-01 7.94272482e-01 -7.32575282e-02 -1.01593769e+00
8.07533622e-01 7.93693364e-01 6.91287220e-01 -1.58885515e+00
-3.17550868e-01 -2.94296257e-02 -2.05874026e-01 2.27545351e-01
2.61545300e-01 -3.91809165e-01 -8.45412970e-01 1.42705333e+00
3.66907686e-01 8.56137946e-02 -9.87209678e-02 8.79228294e-01
8.04357648e-01 8.10078681e-01 6.53467238e-01 4.38231140e-01
1.24147844e+00 -7.14241803e-01 -6.09220028e-01 -6.88731968e-01
7.75498748e-01 -6.98035419e-01 9.52933729e-01 -5.73543757e-02
-1.04058611e+00 -6.37470365e-01 -7.53319740e-01 -9.09267217e-02
-1.06904066e+00 4.40590918e-01 6.23467088e-01 5.09014726e-01
-1.63144457e+00 1.22907087e-01 -3.83076310e-01 -8.43593419e-01
2.23620728e-01 1.53373390e-01 -3.84780586e-01 -1.45009220e-01
-1.12755871e+00 1.23304558e+00 1.29102468e-01 -9.04890373e-02
-4.19957489e-01 -7.58329034e-01 -8.04356694e-01 -9.00435224e-02
1.12742774e-01 -2.74817437e-01 1.00378752e+00 -7.19672978e-01
-9.53815460e-01 9.59688067e-01 -3.14041764e-01 -4.70230542e-02
5.29289544e-01 -6.68552592e-02 -5.78669727e-01 2.19752118e-01
8.78514528e-01 1.28862083e+00 3.45834762e-01 -1.73207068e+00
-9.25197840e-01 -3.27360183e-01 2.77335733e-01 4.77733493e-01
-6.34876713e-02 1.77164391e-01 2.10517235e-02 -5.20652175e-01
-4.83996756e-02 -8.80437434e-01 -2.99512655e-01 -8.34937841e-02
-3.21834356e-01 -1.55885130e-01 6.03283286e-01 -1.14977145e+00
1.24602795e+00 -1.90097678e+00 -4.38115716e-01 5.24921179e-01
1.93875000e-01 1.87552482e-01 -2.65884399e-01 1.04201317e+00
2.85034567e-01 6.00743890e-01 -1.19602568e-01 -3.31361592e-01
-2.03635693e-02 2.09054891e-02 -4.54179257e-01 -1.13854811e-01
4.08382080e-02 8.60599339e-01 -1.23121786e+00 -6.06891096e-01
1.91294357e-01 3.87290359e-01 5.04882745e-02 -9.62349921e-02
-1.59276035e-02 2.24631131e-01 -2.73717761e-01 5.10900974e-01
6.11675918e-01 -1.29946291e-01 8.53925198e-02 8.54273289e-02
-4.69603240e-01 9.35587659e-02 -1.21537161e+00 1.53114951e+00
-2.98591018e-01 8.38364303e-01 -9.19833034e-02 -1.10266939e-01
1.27908278e+00 2.19699442e-01 1.77393600e-01 -6.96201980e-01
-4.68264014e-01 8.57947320e-02 -4.36856985e-01 -6.66949630e-01
1.12256694e+00 3.90393406e-01 -1.06672518e-01 9.39290106e-01
-5.46166420e-01 -5.89832813e-02 6.61302060e-02 5.27257264e-01
8.27603936e-01 8.82299393e-02 1.82411730e-01 -3.08730990e-01
1.22546352e-01 7.14311957e-01 -8.35485235e-02 1.03319478e+00
1.42926306e-01 3.55863512e-01 -8.06448534e-02 -5.77560961e-01
-1.08142483e+00 -7.24969983e-01 3.54718536e-01 1.23450911e+00
1.02502458e-01 -5.76544225e-01 -8.56638372e-01 -4.44230467e-01
5.61452247e-02 1.33233428e+00 -7.39843488e-01 2.45449260e-01
-3.46724838e-01 -4.55926716e-01 7.39854097e-01 6.58881485e-01
7.01693356e-01 -9.66620445e-01 -9.11305487e-01 1.66612312e-01
-5.86757183e-01 -9.65447962e-01 -1.15486778e-01 -4.46065843e-01
-3.94447356e-01 -6.62302077e-01 -5.71572363e-01 -5.04795253e-01
1.13485467e+00 6.80375099e-01 9.72740591e-01 3.64837557e-01
-9.55554917e-02 7.14896917e-01 -3.36847544e-01 -6.44791961e-01
-4.72979784e-01 1.38336182e-01 -3.38914841e-01 -2.57194489e-01
5.35813808e-01 -4.62397903e-01 -1.89077526e-01 5.72913945e-01
-6.48182869e-01 5.73313653e-01 6.13975406e-01 2.51804832e-02
1.64632469e-01 -2.05638528e-01 5.46707034e-01 -8.36775541e-01
9.00235474e-01 -7.24826217e-01 -3.87624174e-01 6.82070553e-01
-4.77459550e-01 -2.34405607e-01 -9.33382884e-02 -2.83532053e-01
-1.25312591e+00 2.87230164e-01 4.12150264e-01 1.76019251e-01
-3.97612363e-01 7.45995879e-01 1.63690239e-01 -4.29362923e-01
1.11898017e+00 -4.77814600e-02 -2.44856685e-01 -5.72322533e-02
4.77018327e-01 9.12439167e-01 7.71027267e-01 -5.60317934e-01
7.89216936e-01 2.98506290e-01 -3.79389852e-01 -9.60610092e-01
-4.05873686e-01 -3.13174814e-01 -7.80365407e-01 -6.87530160e-01
6.75128162e-01 -1.02113867e+00 -4.88636285e-01 1.12464838e-01
-1.17571843e+00 -6.87039196e-01 1.21853314e-02 2.24102989e-01
-2.39063159e-01 3.14935818e-02 4.95752133e-02 -1.19883883e+00
-1.49266735e-01 -5.29049993e-01 1.29824376e+00 3.84416640e-01
-5.21222711e-01 -7.76193678e-01 5.73298968e-02 4.66368794e-01
8.29351306e-01 3.21211338e-01 7.02023983e-01 -6.79893076e-01
-5.94354749e-01 -2.75717348e-01 -5.40643573e-01 -6.67268693e-01
1.61363766e-01 6.64116740e-02 -1.16228926e+00 2.51740962e-01
-6.70202851e-01 1.29852489e-01 1.63823351e-01 3.46419886e-02
5.24452567e-01 -4.04622585e-01 -7.32640982e-01 -1.22399114e-01
1.31986141e+00 4.10903037e-01 7.57671595e-01 4.61896390e-01
7.29409456e-01 9.95921135e-01 7.72256672e-01 4.05919194e-01
1.10795951e+00 6.20089471e-01 1.62294254e-01 -1.66925058e-01
4.73159924e-02 -6.55131817e-01 -2.60036346e-02 2.89011630e-03
-2.84579843e-01 -4.59928602e-01 -1.65871179e+00 9.30319667e-01
-2.02866745e+00 -9.12804306e-01 -8.95214304e-02 2.07185125e+00
2.96676129e-01 -5.11737049e-01 1.06651537e-01 -2.77334124e-01
9.18327093e-01 -2.25090440e-02 -1.22476563e-01 -1.77194569e-02
-9.35721025e-03 -2.90976375e-01 7.77121842e-01 9.50469017e-01
-8.05016458e-01 1.25444949e+00 6.37988281e+00 3.61721843e-01
-7.99442351e-01 1.90271765e-01 5.77857852e-01 2.37072155e-01
-6.22470260e-01 2.23238423e-01 -7.55396366e-01 5.08208156e-01
1.10676670e+00 -2.35790804e-01 6.59203529e-01 7.90090501e-01
2.81491399e-01 -8.61742735e-01 -5.63308120e-01 6.50303602e-01
3.64238143e-01 -1.68795693e+00 1.31234616e-01 2.47697324e-01
6.33887053e-01 -1.01079553e-01 -3.68849725e-01 1.31465802e-02
6.83139861e-01 -9.69396532e-01 9.22794521e-01 9.29276168e-01
6.69965029e-01 -6.85765564e-01 7.34034896e-01 5.67417145e-01
-1.12221503e+00 -5.19166142e-02 -2.04490572e-01 -1.30936399e-01
4.54630375e-01 8.39588902e-05 -1.51227725e+00 4.54645038e-01
6.90208018e-01 9.43280905e-02 -1.02090788e+00 9.34107602e-01
-3.64028811e-01 3.80268574e-01 -5.45021057e-01 -2.22453296e-01
3.23917955e-01 -7.78219942e-03 4.82450008e-01 1.34400618e+00
5.09726763e-01 5.78297377e-01 3.45916003e-01 8.04570377e-01
2.57960111e-01 2.34769434e-02 -1.11010933e+00 -9.19418931e-02
1.14029133e+00 1.22107732e+00 -8.48870337e-01 -5.23288608e-01
1.77636798e-02 6.12946987e-01 2.08966821e-01 3.99081737e-01
-5.45896351e-01 -3.76881987e-01 9.20009613e-02 6.01507068e-01
-1.60925105e-01 -3.97080779e-01 -3.41943562e-01 -2.29156166e-01
-1.47370026e-01 -4.57328469e-01 -1.61667094e-02 -1.75033069e+00
-9.15517151e-01 7.19056845e-01 5.83140016e-01 -8.58981907e-01
-5.26276946e-01 4.43919376e-02 -5.60325027e-01 1.33934855e+00
-1.20917559e+00 -1.72087359e+00 -1.10672855e+00 3.51678520e-01
3.31956685e-01 -1.35091305e-01 9.43894625e-01 1.74462706e-01
9.44088027e-03 4.89082246e-04 -5.53042173e-01 4.81660739e-02
6.95271194e-01 -1.14790213e+00 9.31336641e-01 8.67173851e-01
8.28732923e-02 7.47563660e-01 5.37446976e-01 -1.29293680e+00
-8.90103400e-01 -1.08177650e+00 1.51518941e+00 -6.52785480e-01
4.56186414e-01 -3.65761548e-01 -9.21256900e-01 7.89619863e-01
3.17970276e-01 -5.27541161e-01 7.11752653e-01 -2.78678387e-01
1.45735174e-01 1.45488024e-01 -1.42022943e+00 6.79446697e-01
1.05659771e+00 -7.54902899e-01 -5.21223843e-01 7.30704725e-01
4.69209105e-01 -4.43198532e-01 -4.42256808e-01 -1.49723455e-01
2.93187469e-01 -6.90139771e-01 7.95890570e-01 -3.21633518e-01
2.96522588e-01 -5.88594675e-01 -2.09478408e-01 -1.37053561e+00
-4.75478530e-01 -5.96642315e-01 5.10221899e-01 1.41794300e+00
5.30817389e-01 -7.97125876e-01 4.47755992e-01 1.35482132e+00
-1.63616955e-01 -1.15160152e-01 -5.17809749e-01 -2.73832858e-01
-5.03816962e-01 -4.34982777e-01 7.87814319e-01 9.86379504e-01
1.51696071e-01 -3.70866954e-02 -2.51334578e-01 6.93254411e-01
2.47885257e-01 -3.84080976e-01 1.07946146e+00 -1.34046185e+00
4.44308102e-01 -1.43848181e-01 -2.31574386e-01 -4.37432647e-01
5.64641273e-03 -6.40367568e-01 7.14413598e-02 -2.08861923e+00
-3.27841967e-01 -8.02990139e-01 5.39827228e-01 7.99833596e-01
-7.72720501e-02 7.40561709e-02 2.32674316e-01 3.64794761e-01
-6.73068687e-02 -1.32471263e-01 4.55104172e-01 1.30058184e-01
-3.37815762e-01 -3.28540027e-01 -1.01252842e+00 5.06620646e-01
8.70578945e-01 -5.56365907e-01 -6.76927447e-01 -6.59055293e-01
6.50761545e-01 -2.55612526e-02 8.21927130e-01 -1.00810099e+00
9.64816689e-01 -1.25682041e-01 3.18961501e-01 -9.09318089e-01
3.26851010e-01 -7.37403214e-01 5.62364936e-01 -1.30662292e-01
-3.96530449e-01 5.75416565e-01 6.51037931e-01 4.49640721e-01
4.62975837e-02 -2.34635174e-01 -7.26496726e-02 -3.79596949e-01
-1.19157279e+00 -3.26657474e-01 -6.58730090e-01 -4.01810348e-01
1.00171077e+00 -5.59527695e-01 -8.75934601e-01 -6.71992898e-01
-4.80171651e-01 2.87156194e-01 6.49543643e-01 3.94474328e-01
5.34876466e-01 -1.12806380e+00 -7.61435390e-01 8.64217505e-02
2.04757825e-01 -1.24680534e-01 2.45607778e-01 5.06200671e-01
-6.41144931e-01 2.89800137e-01 -3.49441707e-01 -2.27594316e-01
-9.28925872e-01 1.00628048e-01 2.29222365e-02 1.84209555e-01
-4.53605264e-01 7.80883133e-01 -2.06967741e-01 -4.92225111e-01
8.67453218e-02 1.08421892e-01 -2.46694595e-01 2.25364491e-01
8.03016186e-01 5.67680418e-01 3.55084501e-02 -7.90162146e-01
-4.11633462e-01 4.58711058e-01 3.83377194e-01 -7.01733172e-01
1.29638445e+00 -3.69350761e-01 -1.04115702e-01 3.75785023e-01
2.57713646e-01 2.52434909e-01 -9.90002930e-01 -1.34140953e-01
1.51624575e-01 -6.56752646e-01 2.41020625e-03 -1.39557636e+00
-4.97477084e-01 6.13350689e-01 4.80372906e-01 2.44679093e-01
5.86667538e-01 -2.88204011e-02 2.56228834e-01 5.71426809e-01
7.41869211e-01 -1.11644208e+00 -3.52296799e-01 3.88360433e-02
1.17719793e+00 -1.21050036e+00 1.20211400e-01 -4.15619016e-01
-1.20013130e+00 1.14236820e+00 5.28219223e-01 2.44388148e-01
4.14445847e-01 1.80858478e-01 2.07440197e-01 -6.60356760e-01
-3.68732184e-01 -2.05033422e-01 2.36093208e-01 9.27574933e-01
1.69572517e-01 1.87430561e-01 2.56034642e-01 2.01293260e-01
-3.31478655e-01 1.43112078e-01 5.33941507e-01 1.05549181e+00
-5.81935525e-01 -3.75865638e-01 -5.81820965e-01 2.77110547e-01
2.85180151e-01 -7.75521845e-02 -9.01935339e-01 9.68195617e-01
1.11332290e-01 1.36814153e+00 2.90534854e-01 -3.87741089e-01
1.42347798e-01 9.98267904e-02 -1.82521030e-01 -7.17724204e-01
-7.32566595e-01 -4.17068690e-01 5.68332076e-01 -3.70871603e-01
-5.09831190e-01 -7.60994852e-01 -1.16696572e+00 -3.09654266e-01
7.59546533e-02 2.25527942e-01 1.12583470e+00 8.32059920e-01
8.73005867e-01 -5.71238771e-02 1.87013865e-01 -1.19971228e+00
2.53627449e-01 -1.09631479e+00 -2.10895956e-01 -8.32546577e-02
-1.31354213e-01 -6.83582246e-01 -9.28343087e-02 1.58572108e-01] | [9.366415023803711, 9.160676956176758] |
cac7ce69-bf05-4fcb-bf12-db32064c22fc | full-capacity-unitary-recurrent-neural | 1611.00035 | null | http://arxiv.org/abs/1611.00035v1 | http://arxiv.org/pdf/1611.00035v1.pdf | Full-Capacity Unitary Recurrent Neural Networks | Recurrent neural networks are powerful models for processing sequential data,
but they are generally plagued by vanishing and exploding gradient problems.
Unitary recurrent neural networks (uRNNs), which use unitary recurrence
matrices, have recently been proposed as a means to avoid these issues.
However, in previous experiments, the recurrence matrices were restricted to be
a product of parameterized unitary matrices, and an open question remains: when
does such a parameterization fail to represent all unitary matrices, and how
does this restricted representational capacity limit what can be learned? To
address this question, we propose full-capacity uRNNs that optimize their
recurrence matrix over all unitary matrices, leading to significantly improved
performance over uRNNs that use a restricted-capacity recurrence matrix. Our
contribution consists of two main components. First, we provide a theoretical
argument to determine if a unitary parameterization has restricted capacity.
Using this argument, we show that a recently proposed unitary parameterization
has restricted capacity for hidden state dimension greater than 7. Second, we
show how a complete, full-capacity unitary recurrence matrix can be optimized
over the differentiable manifold of unitary matrices. The resulting
multiplicative gradient step is very simple and does not require gradient
clipping or learning rate adaptation. We confirm the utility of our claims by
empirically evaluating our new full-capacity uRNNs on both synthetic and
natural data, achieving superior performance compared to both LSTMs and the
original restricted-capacity uRNNs. | ['Thomas Powers', 'Scott Wisdom', 'John R. Hershey', 'Les Atlas', 'Jonathan Le Roux'] | 2016-10-31 | full-capacity-unitary-recurrent-neural-1 | http://papers.nips.cc/paper/6327-full-capacity-unitary-recurrent-neural-networks | http://papers.nips.cc/paper/6327-full-capacity-unitary-recurrent-neural-networks.pdf | neurips-2016-12 | ['sequential-image-classification'] | ['computer-vision'] | [ 4.06994253e-01 3.17683637e-01 -2.11838618e-01 1.12909205e-01
-7.66998708e-01 -5.41822910e-01 4.58875924e-01 -3.22313815e-01
-5.80961764e-01 6.86359942e-01 2.82887191e-01 -8.55571926e-01
8.23997557e-02 -6.36307955e-01 -7.14863837e-01 -6.58194244e-01
-1.58824146e-01 1.77151226e-02 1.23692073e-01 -3.61539364e-01
1.11486457e-01 5.29481053e-01 -1.36589289e+00 2.96000242e-02
7.19604254e-01 7.71072567e-01 2.31439009e-01 9.28496778e-01
1.14796147e-01 8.45989764e-01 -3.99137229e-01 4.40525934e-02
4.69602108e-01 -6.43228292e-01 -7.88766742e-01 -2.47644842e-01
2.03669593e-01 -2.24031329e-01 -6.45338356e-01 1.07867110e+00
4.79965270e-01 3.49644423e-01 6.58832610e-01 -6.03875399e-01
-8.88890147e-01 9.58056390e-01 -1.03070647e-01 4.60593373e-01
-6.42113239e-02 -3.01722109e-01 1.41189981e+00 -6.76267207e-01
5.67469597e-01 1.11225128e+00 6.38035536e-01 8.83689165e-01
-1.52838886e+00 -6.27987266e-01 1.25074670e-01 -3.43894470e-03
-1.43279767e+00 -7.30322838e-01 6.18294060e-01 -1.95151940e-01
1.52291894e+00 4.34226632e-01 3.47712338e-01 1.10772228e+00
4.29897346e-02 9.10979092e-01 7.69448698e-01 -6.22618079e-01
8.22565779e-02 -9.92821753e-02 3.60914230e-01 8.85595381e-01
1.84259027e-01 -2.92738024e-02 -8.29324871e-02 -2.47743856e-02
1.18779111e+00 -7.64009431e-02 -5.28034925e-01 -3.02826673e-01
-1.24837792e+00 1.08519077e+00 5.10379732e-01 6.19433880e-01
-2.37377018e-01 3.54534596e-01 3.36862475e-01 7.85209715e-01
2.95607299e-01 6.03506148e-01 -2.24670410e-01 -3.29113036e-01
-6.27666831e-01 -2.26939112e-01 1.01257598e+00 7.82188416e-01
6.29553854e-01 5.58676541e-01 -5.51272742e-02 8.99429142e-01
-6.58118948e-02 4.78282511e-01 8.93520713e-01 -8.05023491e-01
4.51855838e-01 2.27153212e-01 -9.83444005e-02 -6.54701054e-01
-6.20288014e-01 -3.69773984e-01 -1.07550883e+00 -2.38262922e-01
3.00053984e-01 -4.52558786e-01 -9.03295398e-01 2.09625030e+00
-1.57131448e-01 7.92745575e-02 3.97554010e-01 8.07861090e-01
3.35547090e-01 8.07049096e-01 -5.68446159e-01 -3.92913818e-01
9.48761225e-01 -9.06509161e-01 -6.75639629e-01 -1.09918840e-01
1.12691104e+00 -5.56877434e-01 1.07904565e+00 7.32997507e-02
-8.18578959e-01 -1.29676297e-01 -1.42053854e+00 -1.20639421e-01
-2.10149959e-01 2.05139428e-01 9.02711749e-01 8.87576222e-01
-1.28834295e+00 9.64879870e-01 -9.41937387e-01 -4.43528801e-01
-4.14333194e-01 4.06928778e-01 -3.29148859e-01 4.25642610e-01
-1.57020319e+00 1.08273602e+00 6.48206532e-01 3.85774583e-01
-4.68172431e-01 -1.53661698e-01 -9.83029783e-01 8.23611915e-02
3.67222160e-01 -6.13575101e-01 1.29292691e+00 -7.08981633e-01
-1.77498555e+00 2.78744340e-01 -2.05025300e-01 -1.01538801e+00
2.87810266e-01 -8.16780701e-02 -2.44773686e-01 2.16795187e-02
-3.72480303e-01 3.21834803e-01 7.87627518e-01 -6.38698399e-01
-2.69870788e-01 -5.61275743e-02 2.35794336e-01 3.09167653e-01
-5.33232570e-01 -2.63328850e-01 -1.67998284e-01 -5.86677194e-01
5.45307517e-01 -1.13809538e+00 -3.84480685e-01 -5.81415355e-01
-4.42630619e-01 -1.80101514e-01 4.92651403e-01 -2.66762912e-01
1.57701254e+00 -1.89244545e+00 4.21706319e-01 1.43528447e-01
6.84072301e-02 3.79227728e-01 -3.50255370e-01 4.03762549e-01
-1.19824298e-01 4.31683600e-01 -2.04465076e-01 -2.75061131e-01
1.12197548e-01 5.17521024e-01 -6.25698924e-01 5.53777277e-01
2.02621520e-01 1.00470436e+00 -7.43848443e-01 1.75063610e-02
-3.75183783e-02 6.30227327e-01 -6.29620433e-01 -9.04944241e-02
5.97453974e-02 -1.25649974e-01 -1.63287014e-01 1.88563079e-01
3.29514332e-02 -2.90307969e-01 4.34641987e-01 -1.83628947e-01
-2.07043096e-01 6.42919660e-01 -1.24385870e+00 1.25572050e+00
-5.25336325e-01 8.32002223e-01 -2.16601565e-01 -1.13079476e+00
8.08471143e-01 3.75827998e-01 2.72868693e-01 -6.44692063e-01
3.15811425e-01 4.87927616e-01 2.51882970e-01 -6.87591136e-02
6.72195196e-01 -1.31607503e-01 -3.41721959e-02 9.63817120e-01
1.31076530e-01 1.61505222e-01 4.51404333e-01 2.86935776e-01
1.19239700e+00 -1.43392161e-01 1.85648426e-01 -2.85017252e-01
2.62616396e-01 -7.53317356e-01 4.29021806e-01 8.12262416e-01
-1.57694831e-01 6.57857299e-01 8.55043948e-01 -2.53543049e-01
-1.38940561e+00 -1.01682770e+00 -1.98982030e-01 1.01362121e+00
-1.76613271e-01 -5.35285950e-01 -6.26081645e-01 -1.43034071e-01
-3.89834642e-01 5.46415627e-01 -9.37924087e-01 -3.06158930e-01
-7.10979164e-01 -9.02223289e-01 8.71903837e-01 5.81823409e-01
3.00330609e-01 -8.70859563e-01 -5.54917395e-01 2.99769849e-01
-8.64901319e-02 -1.08534694e+00 -7.80587077e-01 7.34237313e-01
-1.19013584e+00 -9.02188420e-01 -8.22674036e-01 -6.40294731e-01
6.98793292e-01 3.20177376e-01 6.24520719e-01 -1.10140398e-01
1.75602823e-01 3.37512702e-01 -2.50330091e-01 1.70565099e-01
-5.15405774e-01 5.13609588e-01 4.61776733e-01 -2.66688522e-02
9.22012329e-02 -7.94184089e-01 -3.01030934e-01 1.39659613e-01
-1.13267660e+00 1.78076133e-01 7.55967379e-01 1.02563488e+00
4.20719534e-01 -4.48531687e-01 5.57265520e-01 -8.22096646e-01
1.16089082e+00 -1.76142186e-01 -4.67697829e-01 2.60769963e-01
-5.61477482e-01 8.97228420e-01 6.57482028e-01 -7.83982337e-01
-5.19482911e-01 1.40052019e-02 -2.23383680e-01 -6.34740055e-01
5.35876513e-01 7.20259488e-01 3.50667655e-01 -4.07984509e-04
8.47706079e-01 4.18008834e-01 1.02854975e-01 -3.43443364e-01
6.41611993e-01 5.29769957e-01 2.82987297e-01 -5.90362251e-01
5.29018283e-01 2.10324749e-01 -1.75222740e-01 -1.28991139e+00
-7.61423230e-01 -3.07588339e-01 -6.96186125e-01 2.22774282e-01
6.46750152e-01 -5.36149204e-01 -6.14013135e-01 6.99163154e-02
-1.22669363e+00 -5.40432870e-01 -3.22727323e-01 5.46035290e-01
-6.76273525e-01 6.01725221e-01 -1.19975984e+00 -1.05676210e+00
-7.22505689e-01 -1.01711106e+00 5.20726204e-01 1.98478680e-02
-1.45498261e-01 -9.60261285e-01 1.05640545e-01 -3.41846555e-01
6.24483824e-01 -1.37555704e-01 8.22236896e-01 -6.35594308e-01
-3.94021690e-01 -1.78715780e-01 -2.95502007e-01 6.47092760e-01
3.55997197e-02 -1.61541030e-01 -7.34895468e-01 -3.87442112e-01
1.17603816e-01 -3.09051096e-01 1.15218306e+00 1.33964255e-01
5.54109156e-01 -4.09959406e-01 1.08892573e-02 6.16703033e-01
1.10632837e+00 1.80498466e-01 5.06651700e-01 9.48743895e-02
6.44777238e-01 1.62374705e-01 -2.86135465e-01 1.30063310e-01
2.48996057e-02 3.05315226e-01 1.59371749e-01 2.29230627e-01
1.51023120e-01 -3.00059050e-01 7.45065272e-01 1.52148008e+00
-3.03980947e-01 2.05124989e-02 -4.72637564e-01 3.78896654e-01
-1.93530226e+00 -9.20328856e-01 2.40409568e-01 2.36987591e+00
6.90298438e-01 4.25061464e-01 -5.37968194e-03 1.76250651e-01
6.12804472e-01 3.25596482e-01 -7.31081247e-01 -8.40188503e-01
-4.18455213e-01 1.22427523e-01 8.25983942e-01 6.36579394e-01
-1.08523250e+00 1.04379714e+00 7.56862211e+00 6.53041244e-01
-1.33211029e+00 2.03981981e-01 1.99747264e-01 -2.35223874e-01
-4.02818650e-01 -7.89720565e-02 -8.19430768e-01 2.96795927e-02
1.55556870e+00 -2.58424431e-01 5.67158997e-01 7.46630073e-01
-1.61828741e-01 2.56093681e-01 -1.06024885e+00 1.01127267e+00
2.66162213e-02 -1.39279699e+00 7.06765577e-02 1.46762326e-01
6.52520716e-01 4.23039317e-01 1.30694360e-01 5.96306384e-01
4.63516384e-01 -8.72012138e-01 5.20612538e-01 8.78326148e-02
7.15802968e-01 -8.42881680e-01 3.85804117e-01 2.62669832e-01
-1.20955026e+00 -6.84353411e-02 -7.50855982e-01 -2.57339209e-01
2.11279139e-01 3.31667036e-01 -7.17417538e-01 2.14582562e-01
1.04136169e-01 4.77760226e-01 -2.58285850e-01 6.13999128e-01
-2.11718455e-01 6.39207065e-01 -7.03626931e-01 -3.20298553e-01
5.71945250e-01 -3.41718942e-01 5.20786464e-01 1.29873240e+00
3.88039887e-01 1.14154749e-01 -1.41969010e-01 7.79739559e-01
-1.96849167e-01 2.00746894e-01 -8.31696510e-01 -5.58500051e-01
1.60926655e-01 9.68352318e-01 -7.37085462e-01 -2.51701266e-01
-1.90959945e-01 1.18835855e+00 7.84397244e-01 5.31910539e-01
-6.25833035e-01 -6.36121511e-01 7.03562260e-01 -2.40962729e-01
3.56252939e-01 -6.56052589e-01 6.05918393e-02 -1.77909839e+00
-5.02256565e-02 -4.51455981e-01 3.07692647e-01 -3.63796055e-01
-7.44633675e-01 9.85130668e-01 -3.30564618e-01 -1.04204881e+00
-6.01562977e-01 -7.00118840e-01 -2.33553395e-01 7.71495104e-01
-1.09575844e+00 -8.12309861e-01 6.35401547e-01 4.64916974e-01
3.63247782e-01 -3.57810557e-02 9.88624573e-01 3.83228749e-01
-8.64537120e-01 7.73309112e-01 3.20819944e-01 2.37360731e-01
1.45363539e-01 -1.24836254e+00 5.41554689e-01 1.06756759e+00
4.97350574e-01 1.13836980e+00 7.60107636e-01 -3.29666018e-01
-1.69652450e+00 -7.09819019e-01 7.21919835e-01 -3.31762016e-01
1.21530402e+00 -6.03975534e-01 -9.58541632e-01 1.18505919e+00
-4.76733036e-02 -4.18782234e-02 5.57739615e-01 4.06167805e-01
-4.95209575e-01 3.49024713e-01 -4.77347553e-01 1.03223956e+00
1.19306493e+00 -9.61648166e-01 -6.03380859e-01 2.27477238e-01
1.11727130e+00 -4.42824125e-01 -8.37095201e-01 4.13287699e-01
6.75783873e-01 -7.84179330e-01 6.43850863e-01 -7.39344537e-01
3.50156464e-02 8.60756729e-03 -3.97801399e-01 -1.34753501e+00
-3.56510103e-01 -9.18738663e-01 -4.16563720e-01 5.34469843e-01
8.21120083e-01 -1.07911563e+00 5.85670650e-01 5.69206953e-01
-2.43282273e-01 -8.23781490e-01 -1.10401475e+00 -1.00908315e+00
3.18691641e-01 -7.12279677e-01 2.17269883e-02 8.10880840e-01
2.92798400e-01 8.75517368e-01 -6.42718434e-01 -1.21974811e-01
2.22616017e-01 -1.63702462e-02 3.91257733e-01 -9.13380027e-01
-2.78948903e-01 -7.63323724e-01 -3.79512280e-01 -1.63412249e+00
1.66640371e-01 -1.01217031e+00 -1.57109976e-01 -1.28859925e+00
-1.01518355e-01 -3.59528005e-01 -4.73551601e-01 4.01470423e-01
1.23781979e-01 2.40068898e-01 3.36840361e-01 3.82624656e-01
-4.92624879e-01 6.72855914e-01 1.14622962e+00 1.60477385e-02
-4.41800714e-01 -2.95592975e-02 -6.53743923e-01 4.68473554e-01
6.60943389e-01 -4.02008779e-02 -5.73028386e-01 -3.01798582e-01
5.21798313e-01 1.09866254e-01 -2.19254076e-01 -9.48383391e-01
3.58037144e-01 2.16193974e-01 -1.20522127e-01 -3.38452607e-01
4.87049252e-01 -3.61527771e-01 -6.97483495e-02 5.81713974e-01
-4.64610904e-01 1.49231389e-01 6.36350214e-02 5.12883723e-01
-1.54891508e-02 -3.42076898e-01 7.52144694e-01 -1.93611775e-02
-5.01601279e-01 1.73923537e-01 -6.01596832e-01 -1.90856650e-01
6.76043093e-01 -7.96790048e-02 -1.10975020e-01 -5.51281273e-01
-9.43461716e-01 -5.93110733e-02 1.85945809e-01 4.75485027e-01
6.14797056e-01 -1.40009010e+00 -4.10551131e-01 3.84058475e-01
-2.49662086e-01 -2.78659403e-01 1.39255434e-01 9.30946112e-01
-4.67124134e-01 9.73796487e-01 1.39442831e-02 -3.24349493e-01
-8.52733910e-01 7.22556591e-01 5.61159968e-01 -4.04136419e-01
-7.90881753e-01 6.85179293e-01 7.09815696e-02 -6.43699646e-01
1.53364241e-01 -7.35613763e-01 -8.27075168e-02 1.73705190e-01
5.96873343e-01 1.20636858e-01 5.32146171e-02 -6.74865186e-01
1.46937385e-01 3.89515996e-01 -1.99798107e-01 -3.74602944e-01
1.09686422e+00 -9.71377715e-02 -2.47370359e-02 8.03803325e-01
1.46634364e+00 -2.76001543e-01 -9.83646154e-01 -4.52551663e-01
-1.57783702e-02 2.12575600e-01 -5.65535901e-03 8.89586136e-02
-9.56963062e-01 9.59884167e-01 4.25477445e-01 6.61861122e-01
7.27230906e-01 -2.55481094e-01 8.97692800e-01 8.81288230e-01
3.14915866e-01 -1.02741408e+00 -4.79431823e-02 1.13383043e+00
8.33287418e-01 -1.01128983e+00 -3.48829210e-01 5.42801730e-02
-5.19293010e-01 1.22145283e+00 3.08685601e-01 -3.11170220e-01
5.91090679e-01 3.37122679e-02 -2.81717107e-02 7.94695914e-02
-1.08286417e+00 -4.74605471e-01 2.17193961e-01 1.51388541e-01
5.39442956e-01 1.97624415e-01 -4.89828885e-01 2.28968710e-01
-4.71409053e-01 -3.68554205e-01 8.56828511e-01 8.51130486e-01
-7.05418348e-01 -9.73919809e-01 -2.78982252e-01 5.29652119e-01
-4.18461055e-01 -3.16041678e-01 -9.49652418e-02 5.10318935e-01
-7.69934356e-01 6.23118877e-01 9.30240229e-02 -5.76973617e-01
2.55994737e-01 1.75905406e-01 5.81058919e-01 -5.99344909e-01
-1.56069264e-01 -3.79038826e-02 -3.20675671e-02 -4.76731509e-01
-1.98626774e-03 -4.15431142e-01 -1.26501191e+00 -2.59228826e-01
-4.87356961e-01 2.52549022e-01 6.96495831e-01 8.79188657e-01
2.85833299e-01 5.62253416e-01 2.80525178e-01 -9.88438189e-01
-1.06822908e+00 -1.12922537e+00 -6.87091291e-01 7.18664005e-02
5.33839762e-01 -4.84300762e-01 -5.94692111e-01 -3.82110029e-01] | [7.804925918579102, 3.5862061977386475] |
e19af55d-5ed1-4d58-bae8-519b0f31dd64 | unsupervised-and-semi-supervised-anomaly | 1710.09207 | null | http://arxiv.org/abs/1710.09207v1 | http://arxiv.org/pdf/1710.09207v1.pdf | Unsupervised and Semi-supervised Anomaly Detection with LSTM Neural Networks | We investigate anomaly detection in an unsupervised framework and introduce
Long Short Term Memory (LSTM) neural network based algorithms. In particular,
given variable length data sequences, we first pass these sequences through our
LSTM based structure and obtain fixed length sequences. We then find a decision
function for our anomaly detectors based on the One Class Support Vector
Machines (OC-SVM) and Support Vector Data Description (SVDD) algorithms. As the
first time in the literature, we jointly train and optimize the parameters of
the LSTM architecture and the OC-SVM (or SVDD) algorithm using highly effective
gradient and quadratic programming based training methods. To apply the
gradient based training method, we modify the original objective criteria of
the OC-SVM and SVDD algorithms, where we prove the convergence of the modified
objective criteria to the original criteria. We also provide extensions of our
unsupervised formulation to the semi-supervised and fully supervised
frameworks. Thus, we obtain anomaly detection algorithms that can process
variable length data sequences while providing high performance, especially for
time series data. Our approach is generic so that we also apply this approach
to the Gated Recurrent Unit (GRU) architecture by directly replacing our LSTM
based structure with the GRU based structure. In our experiments, we illustrate
significant performance gains achieved by our algorithms with respect to the
conventional methods. | ['Suleyman Serdar Kozat', 'Tolga Ergen', 'Ali Hassan Mirza'] | 2017-10-25 | null | null | null | null | ['supervised-anomaly-detection', 'semi-supervised-anomaly-detection'] | ['computer-vision', 'computer-vision'] | [ 3.60762924e-01 -2.67096072e-01 5.34456298e-02 -2.65859425e-01
-3.15842479e-01 -4.52964008e-01 3.83012146e-01 4.36384737e-01
-7.43061364e-01 5.01252353e-01 -3.36897075e-01 -4.79617089e-01
-1.36334538e-01 -6.84683144e-01 -5.34192681e-01 -6.95695400e-01
-4.94593084e-01 3.98895621e-01 1.92854017e-01 -3.18162560e-01
3.66282582e-01 8.33369851e-01 -1.52017295e+00 -4.58715186e-02
8.91845286e-01 1.30841923e+00 -2.91081727e-01 1.10041344e+00
-2.95351684e-01 7.88883567e-01 -4.06496227e-01 2.05682144e-02
3.17594469e-01 -2.55408794e-01 -8.19436371e-01 1.09858498e-01
2.74219900e-01 -2.67061979e-01 -2.02347860e-01 1.11385477e+00
2.26533100e-01 7.33154714e-01 7.44316757e-01 -1.38480103e+00
-4.85725880e-01 2.64158607e-01 -3.32234263e-01 6.54304504e-01
2.01493576e-01 -2.35909030e-01 9.22977030e-01 -9.47737873e-01
1.79797381e-01 9.01366353e-01 8.01011205e-01 4.28935230e-01
-1.22125900e+00 -9.32695791e-02 4.02166545e-01 5.37442625e-01
-1.05178070e+00 -2.90679723e-01 6.37354910e-01 -5.79280913e-01
1.34724188e+00 2.08036691e-01 4.95386094e-01 1.18994856e+00
3.05201650e-01 1.00244403e+00 5.31004846e-01 -6.87916756e-01
4.35574502e-01 -2.63830632e-01 7.15544879e-01 9.37557459e-01
1.09496333e-01 1.53909177e-02 -1.13274470e-01 -3.52272004e-01
5.87237537e-01 3.73492509e-01 1.56747960e-02 -2.64990509e-01
-1.12270784e+00 9.68513668e-01 2.82227639e-02 7.36276150e-01
-5.11819005e-01 -1.73970889e-02 9.10452008e-01 8.81161869e-01
6.46293342e-01 4.18942660e-01 -2.62314409e-01 -1.09345227e-01
-9.72497344e-01 6.68647587e-02 8.24455082e-01 6.06457651e-01
5.04356086e-01 6.67435646e-01 -1.26526237e-01 8.47664356e-01
1.83075592e-01 1.68409809e-01 1.14498889e+00 -4.71313357e-01
2.88052589e-01 3.74184161e-01 -2.45830435e-02 -1.15943170e+00
-4.36571777e-01 -4.62714583e-01 -9.35417056e-01 5.32701127e-02
3.98623914e-01 -3.20833884e-02 -1.02250421e+00 1.62665308e+00
-9.50643569e-02 4.13225293e-01 2.18281001e-01 4.75413799e-01
1.59475431e-02 7.80191064e-01 -1.73106909e-01 -5.75047910e-01
7.06101418e-01 -1.00557911e+00 -8.93939018e-01 8.60708859e-03
1.24428272e+00 -3.67433935e-01 1.05015743e+00 6.35720551e-01
-7.67817318e-01 -3.86668593e-01 -1.32744169e+00 3.08817625e-01
-6.61485434e-01 5.90381473e-02 1.49913654e-01 4.43713248e-01
-1.09702098e+00 9.80030656e-01 -1.18057930e+00 -5.61647594e-01
-3.96615174e-03 4.31338608e-01 -2.25938275e-01 5.25124848e-01
-1.22819102e+00 8.54955256e-01 7.51152515e-01 4.84025836e-01
-6.25385225e-01 6.24018069e-03 -1.02535653e+00 -1.99021041e-01
3.44443887e-01 -2.17828616e-01 1.29819453e+00 -1.17431200e+00
-1.49820077e+00 7.30119348e-01 -2.65341014e-01 -1.05866563e+00
3.39545727e-01 -2.95114011e-01 -5.32783806e-01 8.31600651e-02
-2.86193341e-01 -1.74171001e-01 1.06619477e+00 -6.05600715e-01
-6.05494559e-01 -3.23247701e-01 -3.54199380e-01 -2.23960310e-01
-6.00683451e-01 1.07397042e-01 2.85975754e-01 -1.05977345e+00
3.09601635e-01 -7.50431299e-01 -4.78729874e-01 -2.84161419e-01
-4.54117119e-01 -3.34913075e-01 1.05302930e+00 -6.11936748e-01
1.46515381e+00 -2.13919878e+00 2.95863032e-01 6.64978266e-01
6.68074265e-02 5.08552969e-01 -1.56818792e-01 3.94303858e-01
-3.21863472e-01 -1.05350427e-01 -7.83903062e-01 -5.33397377e-01
-5.40225208e-02 8.02805662e-01 -7.60847151e-01 6.21637762e-01
1.79101288e-01 6.81481063e-01 -8.76800954e-01 -3.15823317e-01
2.25310266e-01 2.89438944e-03 -3.90005141e-01 2.95499712e-01
-1.11008100e-01 1.07270218e-01 -4.08769786e-01 5.81582963e-01
2.81282365e-01 9.55430493e-02 -2.23089650e-01 2.41382778e-01
-2.32723027e-01 -1.16848238e-01 -1.10131764e+00 1.45901978e+00
-4.59221721e-01 7.23607481e-01 -2.07596928e-01 -1.73000813e+00
1.18900037e+00 3.71188223e-01 5.08684814e-01 -3.97494495e-01
-1.12237111e-02 7.12312996e-01 -3.67902927e-02 -5.62911451e-01
5.88789344e-01 -1.85167879e-01 1.28275439e-01 5.02947748e-01
1.26361936e-01 5.97567677e-01 2.05793679e-01 -2.90102158e-02
1.14180660e+00 -5.39528131e-02 4.41930324e-01 -7.79509395e-02
1.00048971e+00 -2.79034287e-01 2.82968611e-01 1.05372131e+00
-1.99008778e-01 3.46943915e-01 5.63631475e-01 -5.79916120e-01
-1.10995090e+00 -9.36300397e-01 -2.21677184e-01 1.22982550e+00
-5.78564644e-01 -2.01781362e-01 -5.57256401e-01 -6.17006600e-01
-7.72149563e-02 7.62170494e-01 -6.51690125e-01 -2.27573529e-01
-8.20975363e-01 -7.64745772e-01 8.69180143e-01 5.87852299e-01
2.45652974e-01 -1.24401689e+00 -4.36693132e-01 5.75174809e-01
4.27424870e-02 -9.49657381e-01 -2.55047619e-01 5.46566725e-01
-1.07576787e+00 -6.43832922e-01 -6.63140476e-01 -7.50070333e-01
5.00452578e-01 -3.96480411e-01 5.76495349e-01 4.76244055e-02
-1.16657577e-01 4.26205367e-01 -5.21999478e-01 -2.32222676e-01
-5.31180918e-01 1.23595357e-01 6.31512702e-01 4.93102282e-01
3.62479746e-01 -8.67842674e-01 -1.20367400e-01 1.62007287e-01
-1.12720394e+00 -6.32771969e-01 3.18517208e-01 9.27320838e-01
6.36340320e-01 -1.88648969e-01 6.42189085e-01 -6.22018874e-01
6.40630305e-01 -5.84936261e-01 -6.73629940e-01 6.05654791e-02
-8.21822762e-01 3.59869331e-01 1.04411530e+00 -4.85139966e-01
-4.89041239e-01 5.31511828e-02 -4.42142934e-01 -9.61621463e-01
-1.12876222e-01 5.39883375e-01 2.48071864e-01 -2.23444313e-01
5.78208208e-01 6.20149195e-01 4.98285778e-02 -6.33404195e-01
-1.53703941e-02 6.70680344e-01 6.94598734e-01 -5.45990348e-01
6.64901674e-01 3.96751702e-01 -5.91661818e-02 -1.15376341e+00
-7.46845543e-01 -5.58166623e-01 -9.83862996e-01 -6.50867298e-02
7.71624267e-01 -1.85919672e-01 -5.29382765e-01 7.54426360e-01
-1.25590253e+00 -1.97046757e-01 -3.91848832e-01 4.79004890e-01
-7.29848862e-01 8.45775485e-01 -7.99926639e-01 -1.26622534e+00
-4.55912262e-01 -8.33778679e-01 6.76998615e-01 -3.43407452e-01
-2.37441182e-01 -1.43229055e+00 3.87938976e-01 -5.62478840e-01
4.55767125e-01 5.57189465e-01 9.09031630e-01 -1.58016014e+00
1.03314444e-01 -4.36002553e-01 1.46785542e-01 8.45243752e-01
-4.57150005e-02 -5.70450276e-02 -7.86292315e-01 -3.97700816e-01
4.01403815e-01 7.35089481e-02 1.06683862e+00 2.38513440e-01
1.37539828e+00 -5.26659489e-01 -7.43658021e-02 7.03679979e-01
1.40244901e+00 3.30859393e-01 4.37673211e-01 7.59441435e-01
7.58868515e-01 5.13052821e-01 4.48497564e-01 4.46964890e-01
-2.84789622e-01 6.25958145e-01 2.49073669e-01 1.85498282e-01
8.32961380e-01 -5.67525662e-02 7.74218082e-01 1.08513021e+00
-3.61621343e-02 -1.72609031e-01 -1.03591490e+00 6.42810464e-01
-2.24975133e+00 -1.07352388e+00 -2.77709097e-01 2.45419550e+00
3.76904756e-01 1.84972689e-01 3.99533778e-01 6.62690818e-01
7.09706903e-01 1.79203793e-01 -5.08755803e-01 -1.14126968e+00
-2.04841614e-01 4.13468838e-01 4.46552485e-01 5.10284722e-01
-1.38457549e+00 6.50678873e-01 6.41148853e+00 7.29709983e-01
-1.15913880e+00 -3.22204791e-02 2.84550428e-01 -5.99872172e-02
3.27677466e-02 -2.68492371e-01 -5.25263965e-01 4.28961903e-01
1.39631426e+00 -2.26559758e-01 1.37054473e-01 9.52113688e-01
2.62767166e-01 2.40183040e-01 -1.30170572e+00 9.32946444e-01
1.22390561e-01 -7.88398504e-01 3.03529322e-01 8.82572234e-02
4.32816505e-01 -9.59792640e-03 3.42347845e-02 2.91190267e-01
-1.51004389e-01 -1.02867270e+00 4.33362335e-01 4.95427370e-01
2.45375469e-01 -7.47660279e-01 7.85394192e-01 1.82874307e-01
-1.15348959e+00 -3.68354380e-01 -4.09933120e-01 -1.75487161e-01
1.35673434e-01 7.03742862e-01 -5.45325756e-01 5.56615591e-01
3.87692899e-01 8.54020000e-01 -3.88945848e-01 9.67319250e-01
1.35615021e-01 6.65585279e-01 -4.75441277e-01 -7.85114989e-02
8.21284711e-01 -5.02742350e-01 1.20574534e+00 1.32834291e+00
4.33348596e-01 -2.60186583e-01 2.70896316e-01 5.57032883e-01
3.08290720e-01 3.99410099e-01 -8.70854557e-01 -1.62957907e-01
-8.01535137e-03 9.23962057e-01 -5.91837645e-01 -3.48563999e-01
-2.69268245e-01 1.20554662e+00 2.73999065e-01 4.22148943e-01
-4.59800988e-01 -7.87289262e-01 5.95913768e-01 -1.87059879e-01
3.77223462e-01 -3.53414446e-01 -1.62958711e-01 -1.41500258e+00
4.07593638e-01 -8.09814215e-01 8.53329539e-01 -9.33048204e-02
-1.39454830e+00 8.04332256e-01 -9.84555483e-02 -1.29771733e+00
-8.02085340e-01 -8.05789232e-01 -8.42025340e-01 5.72991669e-01
-1.28812373e+00 -6.57120943e-01 3.07762176e-01 8.08786273e-01
6.85582936e-01 -4.23362643e-01 9.21705008e-01 2.33191967e-01
-9.15107727e-01 7.30139852e-01 5.79035521e-01 2.81227648e-01
2.98463494e-01 -1.45369923e+00 5.07609785e-01 1.35138381e+00
1.55380338e-01 5.95196724e-01 7.04037488e-01 -5.31910777e-01
-1.03129792e+00 -1.19792652e+00 6.98540926e-01 -2.14742675e-01
1.11496115e+00 -2.30000138e-01 -1.42882311e+00 1.07287157e+00
-3.17427635e-01 5.98570006e-03 5.83015025e-01 2.06654742e-02
-6.87367022e-02 1.89282745e-01 -1.01013565e+00 4.60506797e-01
8.74819517e-01 -5.31895220e-01 -9.26703215e-01 2.77762800e-01
5.45947611e-01 -1.47562444e-01 -6.75737679e-01 3.75175089e-01
3.99162412e-01 -7.48958528e-01 6.65809035e-01 -1.07002842e+00
1.70523807e-01 -2.89320469e-01 -4.10460860e-01 -1.20359027e+00
5.88753913e-03 -6.94525719e-01 -6.12508953e-01 8.38008106e-01
4.86110538e-01 -1.21992922e+00 4.79299366e-01 2.35958010e-01
-3.72325957e-01 -9.14991260e-01 -1.13701868e+00 -1.20703387e+00
-3.44857909e-02 -7.72365451e-01 6.23697750e-02 9.61605012e-01
7.58259147e-02 -2.09625706e-01 -5.23843646e-01 2.63403565e-01
5.64054489e-01 -3.27159390e-02 2.56157398e-01 -1.17098439e+00
-4.23748493e-01 -5.24990261e-01 -9.28434074e-01 -9.07554090e-01
3.31322998e-01 -9.89656746e-01 2.33847240e-04 -9.40648615e-01
-4.49054599e-01 -1.54267505e-01 -8.45194161e-01 5.88406622e-01
1.41756922e-01 2.32898723e-02 -1.98489815e-01 2.88476527e-01
-5.18686354e-01 5.19574106e-01 3.58580232e-01 1.91957861e-01
-5.12711287e-01 2.69268036e-01 3.64618488e-02 9.06759977e-01
7.65748382e-01 -3.76874655e-01 -9.07830149e-02 -1.58457249e-01
4.26115654e-02 -2.38845393e-01 3.65651131e-01 -1.09439850e+00
2.64423966e-01 1.51908696e-01 2.58194864e-01 -4.65414882e-01
-3.84344743e-03 -5.33904254e-01 -5.86870551e-01 7.08009124e-01
-5.48623204e-01 4.84971523e-01 1.34043396e-01 7.09696531e-01
-3.55121613e-01 -5.98782361e-01 7.81557262e-01 1.06880188e-01
-7.44240463e-01 3.59038681e-01 -9.73089039e-01 -1.46390945e-01
9.00198758e-01 -3.68187904e-01 3.20651352e-01 -3.74476641e-01
-1.18949568e+00 3.04368228e-01 1.93360522e-01 2.93535292e-01
7.97507584e-01 -1.39971852e+00 -4.86082315e-01 5.33972085e-01
1.31235093e-01 -3.20644289e-01 -5.77998683e-02 1.09011650e+00
-4.78862196e-01 4.61379617e-01 -3.09407830e-01 -7.13686287e-01
-1.04724622e+00 6.53048694e-01 5.73642850e-01 -3.45970631e-01
-7.26681232e-01 3.24580252e-01 -2.19502002e-01 -3.50510925e-01
3.84155154e-01 -4.46746051e-01 -2.74793804e-01 5.99640682e-02
5.44857800e-01 6.13790631e-01 2.91959763e-01 -4.66416687e-01
-3.06760579e-01 4.56884235e-01 -2.30248019e-01 -2.21886218e-01
1.31097448e+00 -1.33691933e-02 -1.94777966e-01 1.04160726e+00
1.47790956e+00 -3.52281243e-01 -4.31462079e-01 -4.92109776e-01
5.45263767e-01 -1.27162173e-01 -8.53986740e-02 8.78377631e-02
-7.11611629e-01 8.01476121e-01 7.29572058e-01 5.57616174e-01
1.16364014e+00 -4.51061726e-01 9.69087660e-01 9.27508175e-01
1.66494902e-02 -1.14299321e+00 -4.10637185e-02 9.34004486e-01
6.41201615e-01 -9.60544050e-01 -5.70971668e-01 1.80107221e-01
-2.04655543e-01 1.56796980e+00 3.89411479e-01 -5.03233969e-01
5.70973873e-01 6.69056028e-02 -7.87673369e-02 4.31058705e-02
-7.16640234e-01 -2.79582053e-01 2.27327719e-01 3.26920986e-01
2.36988485e-01 -2.79541433e-01 -3.16993147e-01 2.27830037e-01
-1.14668891e-01 -2.46061563e-01 5.46038687e-01 9.05910611e-01
-5.00915170e-01 -1.08779156e+00 -3.18540812e-01 6.10791445e-01
-5.66159725e-01 -3.93849649e-02 -1.54019371e-01 5.23734748e-01
-4.26786214e-01 7.17063665e-01 3.15972358e-01 -4.27663267e-01
4.60471779e-01 7.14476347e-01 1.01637512e-01 -4.60209459e-01
-4.14139032e-01 -2.46045619e-01 -9.82729793e-02 -4.48461622e-01
-2.21273974e-01 -5.96576869e-01 -1.19224787e+00 -8.04811046e-02
-2.25737035e-01 2.49654636e-01 5.10996222e-01 1.36941373e+00
-3.28584807e-03 3.04787338e-01 8.60640049e-01 -7.68411100e-01
-7.65961707e-01 -8.41247439e-01 -6.27654314e-01 2.50597119e-01
8.24162424e-01 -4.56490815e-01 -6.92602098e-01 -2.60361522e-01] | [7.5516767501831055, 2.458076238632202] |
272c94fa-0faa-48d6-bb76-f9fc22a6aafe | simplistic-collection-and-labeling-practices | 2301.07015 | null | https://arxiv.org/abs/2301.07015v2 | https://arxiv.org/pdf/2301.07015v2.pdf | Simplistic Collection and Labeling Practices Limit the Utility of Benchmark Datasets for Twitter Bot Detection | Accurate bot detection is necessary for the safety and integrity of online platforms. It is also crucial for research on the influence of bots in elections, the spread of misinformation, and financial market manipulation. Platforms deploy infrastructure to flag or remove automated accounts, but their tools and data are not publicly available. Thus, the public must rely on third-party bot detection. These tools employ machine learning and often achieve near perfect performance for classification on existing datasets, suggesting bot detection is accurate, reliable and fit for use in downstream applications. We provide evidence that this is not the case and show that high performance is attributable to limitations in dataset collection and labeling rather than sophistication of the tools. Specifically, we show that simple decision rules -- shallow decision trees trained on a small number of features -- achieve near-state-of-the-art performance on most available datasets and that bot detection datasets, even when combined together, do not generalize well to out-of-sample datasets. Our findings reveal that predictions are highly dependent on each dataset's collection and labeling procedures rather than fundamental differences between bots and humans. These results have important implications for both transparency in sampling and labeling procedures and potential biases in research using existing bot detection tools for pre-processing. | ['Philipp Zimmer', 'Erin Walk', 'Manish Raghavan', 'Zachary Schutzman', 'Chris Hays'] | 2023-01-17 | null | null | null | null | ['twitter-bot-detection'] | ['miscellaneous'] | [-8.81168768e-02 -3.14689815e-01 -6.99049532e-01 -9.67978127e-03
-3.21537733e-01 -1.01687396e+00 7.64922440e-01 3.77550274e-01
-7.65814185e-01 6.24077797e-01 1.51403069e-01 -1.01424348e+00
3.02981049e-01 -7.79150784e-01 -3.46257597e-01 -1.32635012e-01
2.71282405e-01 3.78047466e-01 2.97954321e-01 -1.77225649e-01
7.44669199e-01 4.59942162e-01 -1.04379380e+00 1.79598957e-01
4.37951237e-01 3.51439893e-01 -3.36267710e-01 6.94272578e-01
1.52351946e-01 1.19214809e+00 -9.65239644e-01 -7.57795274e-01
4.94142175e-01 -7.06400201e-02 -7.55306780e-01 -2.78451890e-01
2.71572977e-01 -8.26778471e-01 -1.99184433e-01 1.03371704e+00
1.58572748e-01 -4.93589073e-01 6.53938055e-01 -1.21208537e+00
-7.37626255e-01 6.50472522e-01 -7.19038248e-01 7.00721681e-01
7.35589713e-02 8.71408165e-01 1.16868818e+00 -2.39276722e-01
8.37799251e-01 1.15501916e+00 9.96007800e-01 2.15684980e-01
-1.39761555e+00 -1.03595281e+00 -6.61632642e-02 -6.16139127e-03
-7.59096980e-01 -6.58018529e-01 2.42444694e-01 -1.21052349e+00
8.68763804e-01 1.91010386e-01 5.10113716e-01 1.38579667e+00
2.64373720e-01 1.00086272e-01 1.36927319e+00 -8.63610804e-02
-1.04983605e-01 4.73962367e-01 4.57223833e-01 5.44146538e-01
1.25156260e+00 -3.04044262e-02 -1.88867927e-01 -7.52942562e-01
6.93139076e-01 -3.28086987e-02 3.68954867e-01 3.75304013e-01
-9.16876733e-01 1.33676088e+00 3.53491247e-01 4.81388688e-01
-4.19280142e-01 4.71188910e-02 7.72045732e-01 9.15208757e-02
5.84977686e-01 8.60231698e-01 -3.61129284e-01 -4.09363568e-01
-8.60702097e-01 3.84668648e-01 8.86081398e-01 2.41072834e-01
5.59652686e-01 -4.87481505e-02 2.63924780e-03 3.86549771e-01
4.84413393e-02 4.07625854e-01 4.73705381e-01 -9.25932646e-01
4.68940347e-01 6.93037271e-01 4.52589363e-01 -1.47656965e+00
-4.38435882e-01 -3.31901819e-01 -1.60916895e-01 2.59146035e-01
1.04675972e+00 -3.70048255e-01 -5.64473093e-01 1.49460268e+00
1.96450427e-01 -3.90815288e-01 -6.32218063e-01 5.58974445e-01
2.54318476e-01 1.27271727e-01 3.45903665e-01 -7.99519643e-02
1.48073077e+00 -2.80259758e-01 -4.81724054e-01 -7.36879468e-01
7.24890232e-01 -6.88616753e-01 1.06037843e+00 1.05288297e-01
-6.98582590e-01 -3.48849110e-02 -7.11981952e-01 4.86683007e-03
-6.21173739e-01 -3.12569678e-01 7.90339649e-01 1.13243151e+00
-5.79975486e-01 5.83823919e-01 -7.37242937e-01 -6.76762938e-01
7.49627173e-01 4.70773578e-02 -9.91822779e-02 1.76680401e-01
-6.59294665e-01 1.23901725e+00 -7.50400126e-02 -2.16552794e-01
-6.43223345e-01 -4.28284407e-01 -3.16748351e-01 -1.87032804e-01
1.67106032e-01 -1.38584033e-01 1.37341106e+00 -6.69080019e-01
-8.27995539e-01 9.82424855e-01 -3.07704117e-02 -6.28231645e-01
4.86341059e-01 2.12570861e-01 -6.84425160e-02 -2.72318460e-02
8.06622088e-01 6.71950206e-02 8.45387697e-01 -1.03489614e+00
-6.28063560e-01 -6.94815159e-01 1.41728848e-01 -4.25749004e-01
-4.99108493e-01 7.09106982e-01 5.48467815e-01 -3.13377202e-01
-2.70289004e-01 -8.85191083e-01 -8.47969577e-02 -5.45113683e-01
-3.44026387e-01 -1.10408291e-01 9.56016421e-01 -1.30469453e+00
1.29645157e+00 -1.75404203e+00 -6.84426546e-01 1.11627825e-01
5.36520064e-01 4.90342885e-01 2.12890953e-01 4.07012701e-01
1.76180661e-01 9.57789958e-01 3.54307175e-01 -2.37233862e-01
-5.10355979e-02 -1.84707537e-01 -4.23987180e-01 8.44228506e-01
2.17912301e-01 7.98701167e-01 -8.27643216e-01 -3.74759644e-01
1.63665742e-01 -4.15168107e-02 -5.22437453e-01 -2.39579514e-01
1.13755018e-01 4.65350717e-01 -2.68767416e-01 9.49939251e-01
2.91019022e-01 -2.90615797e-01 3.30720514e-01 3.02787691e-01
-4.50115025e-01 9.34217095e-01 -4.42975670e-01 5.02747715e-01
-2.27663338e-01 9.10783589e-01 4.50498253e-01 -5.45857489e-01
4.76995111e-01 -1.59618959e-01 1.68949962e-01 -4.42656100e-01
5.16344488e-01 4.47237492e-01 5.08071125e-01 -3.97671849e-01
6.27751291e-01 5.57418764e-02 9.67816114e-02 9.79350567e-01
-4.57844794e-01 1.03743179e-04 3.79764646e-01 7.14102983e-02
1.50122631e+00 -5.01537263e-01 3.56402308e-01 -8.24760124e-02
-1.43644586e-01 5.63449740e-01 7.16908336e-01 8.84297609e-01
-7.56782770e-01 -7.31761009e-02 1.00953221e+00 -3.57477307e-01
-1.27716112e+00 -4.54926550e-01 -1.53066143e-01 1.27015209e+00
-2.65740246e-01 -2.93921620e-01 -8.37749898e-01 -9.17904973e-01
4.00370717e-01 6.90145373e-01 -5.49375474e-01 -2.67010387e-02
-4.27184641e-01 -9.26898360e-01 9.02160168e-01 1.25395507e-01
2.26687804e-01 -8.44596744e-01 -8.24140370e-01 2.56881893e-01
-2.88684875e-01 -1.36854720e+00 -1.99220344e-01 -1.79727286e-01
-6.06231809e-01 -1.64697289e+00 -7.44258314e-02 -2.09533781e-01
5.39610982e-01 6.15799844e-01 7.47303486e-01 5.05844057e-01
-2.71718502e-01 8.60710442e-02 -2.38881022e-01 -7.24994600e-01
-8.77751172e-01 3.69844168e-01 9.35779735e-02 -2.23627940e-01
6.41811907e-01 -5.77099264e-01 -2.93511659e-01 4.09046501e-01
-4.47961092e-01 -5.94291031e-01 5.73128343e-01 4.64342773e-01
-3.53007138e-01 1.40547886e-01 5.92198968e-01 -1.12782609e+00
8.93621862e-01 -7.73671746e-01 -6.19575977e-01 -2.63759136e-01
-6.85368955e-01 -5.16114712e-01 4.04268116e-01 -5.48767209e-01
-5.41280866e-01 -5.00368536e-01 3.19211692e-01 -8.18974823e-02
-9.83500928e-02 3.19132030e-01 5.87759912e-01 -2.11974040e-01
1.20894647e+00 -5.44340909e-01 3.14171135e-01 -3.35238874e-01
-1.69782013e-01 1.14568782e+00 8.36385489e-02 -2.15461910e-01
9.10430789e-01 6.93637669e-01 -3.64053160e-01 -7.79408157e-01
-6.46184742e-01 -5.04405558e-01 -4.41391230e-01 -6.69463351e-02
8.32832873e-01 -7.79102981e-01 -1.00787926e+00 6.34000242e-01
-1.06006646e+00 -5.41656673e-01 2.20299497e-01 3.59908551e-01
4.22023274e-02 4.08569932e-01 -1.01123691e+00 -1.15294325e+00
-2.78716981e-01 -1.14914656e+00 5.11648774e-01 1.01214303e-02
-6.04957402e-01 -8.04011703e-01 -6.25381842e-02 8.67299259e-01
5.85247159e-01 3.33940238e-01 7.12489426e-01 -1.01348639e+00
-3.63256365e-01 -7.45919347e-01 -4.87738341e-01 2.41064772e-01
3.46572131e-01 5.31895459e-01 -1.02454627e+00 -2.31968358e-01
5.54578342e-02 -4.72927898e-01 7.57771850e-01 3.35038394e-01
7.04249442e-01 -7.93899000e-01 -3.92527342e-01 -1.10916294e-01
1.07638013e+00 -1.43249065e-01 3.20493281e-01 7.19965160e-01
6.30181193e-01 7.48367250e-01 2.18818128e-01 4.02786911e-01
3.73640150e-01 5.23755074e-01 3.97348076e-01 3.75292331e-01
2.05106437e-01 -4.35932845e-01 5.81254542e-01 2.76622862e-01
-7.29178451e-03 2.50915021e-01 -1.10981524e+00 5.28896034e-01
-1.68649638e+00 -1.44580996e+00 -3.47021610e-01 2.29514527e+00
7.00728238e-01 5.66411257e-01 1.03868592e+00 5.55586666e-02
1.09173965e+00 1.10376850e-01 -4.86274391e-01 -6.24870420e-01
2.72962928e-01 3.54473889e-02 1.22554851e+00 5.73782802e-01
-1.07127893e+00 1.05221033e+00 7.23949957e+00 4.98511881e-01
-1.14313829e+00 4.99901563e-01 6.82057679e-01 9.21805799e-02
1.45905212e-01 1.54685825e-01 -9.01524544e-01 6.39704406e-01
1.21839428e+00 -1.44350410e-01 7.47858763e-01 9.00489390e-01
5.62999666e-01 -1.93081766e-01 -4.79673237e-01 5.60075223e-01
-1.60729274e-01 -1.36443102e+00 -5.48122227e-01 7.99801350e-01
4.20483917e-01 2.70744532e-01 9.90912244e-02 3.49834114e-01
9.18939650e-01 -9.90238965e-01 9.67573941e-01 -1.39575630e-01
5.42998552e-01 -2.01851711e-01 6.67628944e-01 6.36262417e-01
-5.77882648e-01 -7.34774768e-01 -3.64038348e-01 -6.84957087e-01
2.09717900e-02 6.52615488e-01 -1.40168393e+00 -1.99462801e-01
4.65891659e-01 5.77223957e-01 -9.05084670e-01 6.71747983e-01
-2.15960875e-01 1.04579592e+00 -3.17191988e-01 -1.72421664e-01
1.68955356e-01 1.49826497e-01 3.52790982e-01 1.10098732e+00
-1.93729192e-01 -2.35744253e-01 5.29805906e-02 1.00666118e+00
-6.38408819e-03 -1.07570015e-01 -9.24492657e-01 -7.45834529e-01
8.41988444e-01 1.52282715e+00 -9.63902891e-01 -9.66826379e-02
-4.58903044e-01 2.67146736e-01 5.71773589e-01 -9.48927924e-03
-7.45405138e-01 -3.37398723e-02 9.41927969e-01 6.25816524e-01
8.20810255e-03 -4.73516822e-01 -9.54846382e-01 -1.00597680e+00
-7.24660605e-02 -1.28685153e+00 4.27080482e-01 -1.90910250e-01
-1.46599662e+00 -8.89727026e-02 -1.94640666e-01 -6.51496649e-01
-3.34442496e-01 -7.86084831e-01 -6.18991792e-01 7.72370279e-01
-1.09828508e+00 -1.16405213e+00 -2.67265290e-02 8.68080407e-02
-7.54336221e-03 -3.34356166e-02 5.37791014e-01 -6.03845576e-03
-8.57193530e-01 4.41512465e-01 -1.44378558e-01 5.81595600e-01
5.65631509e-01 -7.06429482e-01 6.84192657e-01 9.87839282e-01
-1.14233904e-01 9.47725654e-01 6.22437179e-01 -1.13498247e+00
-1.12375808e+00 -8.60627115e-01 8.09466362e-01 -1.04411006e+00
1.20091379e+00 -5.12383044e-01 -4.69421744e-01 1.11623728e+00
-1.20184511e-01 -5.12251019e-01 4.99353349e-01 4.81492728e-01
-8.15604270e-01 3.31396848e-01 -1.49796581e+00 4.68465865e-01
9.01148081e-01 -6.96716905e-01 -2.78361946e-01 6.61160529e-01
4.09669220e-01 4.62874025e-03 -4.79817957e-01 -1.35027111e-01
6.92826271e-01 -1.11044765e+00 5.18864274e-01 -9.88151968e-01
3.71146291e-01 1.83290765e-01 5.11978716e-02 -7.77217150e-01
-5.66421211e-01 -6.29544735e-01 3.41649055e-01 1.25534213e+00
5.69221795e-01 -1.09301007e+00 7.68145144e-01 9.31336641e-01
4.51221853e-01 -1.81297213e-01 -6.48683667e-01 -7.48814881e-01
2.62223840e-01 -3.86518449e-01 4.04541075e-01 1.39972162e+00
2.16767371e-01 2.14029491e-01 -6.41169772e-02 1.86131448e-01
5.89452446e-01 -3.55906516e-01 1.34840405e+00 -1.23045468e+00
-2.53336072e-01 -5.98946691e-01 -5.78870714e-01 -4.80449051e-02
3.47175628e-01 -5.54695606e-01 -5.54038882e-01 -1.29926836e+00
4.14300740e-01 -6.62584484e-01 4.06823128e-01 7.95682013e-01
-2.40699098e-01 4.32830185e-01 3.80814672e-01 7.45658278e-01
7.19784945e-02 -3.57740015e-01 6.94242477e-01 1.97309516e-02
-3.21690500e-01 6.03545122e-02 -1.22412300e+00 9.29908991e-01
1.01316845e+00 -6.89510822e-01 1.61985159e-01 -3.59310418e-01
1.36766553e-01 -6.59697652e-01 6.93263292e-01 -8.50520849e-01
-1.35470450e-01 -5.08761466e-01 1.73022762e-01 3.87319624e-02
1.93899497e-01 -5.53418159e-01 -1.31006658e-01 7.40850627e-01
1.54332323e-02 2.70175725e-01 9.75044295e-02 4.35276806e-01
5.27028024e-01 -4.84393567e-01 8.43236983e-01 -4.46593136e-01
-7.41994902e-02 -1.11993045e-01 -7.05189824e-01 2.22849652e-01
8.92891347e-01 -2.14770988e-01 -1.05653584e+00 -6.03760600e-01
-1.20853789e-01 -2.96076626e-01 8.57706785e-01 9.85594168e-02
-4.10316855e-01 -7.77157605e-01 -7.28756130e-01 -1.09894246e-01
-1.81114629e-01 -7.87748694e-01 -4.17082220e-01 9.39789236e-01
-6.13805711e-01 3.45956624e-01 -1.25792608e-01 -2.35890269e-01
-1.11957550e+00 3.99485439e-01 2.07062513e-01 -4.60881889e-01
-2.15109274e-01 3.83061826e-01 -3.36538821e-01 -4.42118585e-01
-1.83273420e-01 -5.20233251e-02 5.11582270e-02 2.23768875e-01
7.24948347e-01 5.85092127e-01 -1.45127669e-01 -7.41931319e-01
-2.65955687e-01 -2.24076614e-01 -3.18188280e-01 -1.86609581e-01
1.22457361e+00 1.37577191e-01 -3.68228376e-01 3.96275938e-01
7.19902098e-01 4.65080112e-01 -6.41970634e-01 1.17621198e-01
2.23252952e-01 -9.04891849e-01 -2.90050179e-01 -5.60838640e-01
-6.78280532e-01 5.21102428e-01 -3.56753841e-02 9.43185091e-01
1.86515138e-01 -1.90274015e-01 6.49433970e-01 1.34122878e-01
4.66926396e-01 -1.10309887e+00 -3.68578807e-02 3.08335960e-01
3.75660956e-01 -1.12039828e+00 3.64393532e-01 -4.31806296e-01
-4.06852871e-01 6.97800219e-01 5.86808860e-01 -1.21299244e-01
3.72546643e-01 3.19541618e-02 3.57319345e-03 -2.20974132e-01
-4.99587715e-01 -8.82342830e-02 -5.80832005e-01 7.10110366e-01
4.10834074e-01 2.88928300e-01 -7.06282914e-01 5.12094855e-01
-4.82624859e-01 -2.23789647e-01 9.33065772e-01 8.98348093e-01
-5.77904761e-01 -1.03420925e+00 -7.27302790e-01 1.04184282e+00
-8.80985975e-01 -2.72481218e-02 -1.14961433e+00 1.03407800e+00
4.62573674e-03 1.46119595e+00 -7.63819972e-03 -5.38678408e-01
6.16451688e-02 1.05587348e-01 1.12971485e-01 -9.43103909e-01
-1.05100679e+00 -4.55751330e-01 5.75297713e-01 -4.16565478e-01
-5.08352481e-02 -1.05645514e+00 -5.65947831e-01 -1.22532690e+00
-6.99453950e-01 -1.47009119e-01 7.91261315e-01 9.38957334e-01
5.43409288e-01 -2.62215316e-01 6.71104848e-01 -8.84185791e-01
-1.08902299e+00 -1.29871166e+00 -5.18807054e-01 3.41472179e-01
1.18981838e-01 -5.80574632e-01 -8.27757716e-01 -3.19685251e-01] | [8.232937812805176, 10.131983757019043] |
11c4366c-2cdb-4a8d-9b20-e2f3fb9982bf | keep-the-primary-rewrite-the-secondary-a-two | null | null | https://aclanthology.org/2021.findings-acl.50 | https://aclanthology.org/2021.findings-acl.50.pdf | Keep the Primary, Rewrite the Secondary: A Two-Stage Approach for Paraphrase Generation | null | ['Nigel Collier', 'Yan Wang', 'Simon Baker', 'David Vandyke', 'Yixuan Su'] | null | null | null | null | findings-acl-2021-8 | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', '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] | [-7.3846917152404785, 3.682955503463745] |
7603c3a5-cc08-498e-99cf-ee0455a7fe65 | the-elements-of-temporal-sentence-grounding | 2201.08071 | null | https://arxiv.org/abs/2201.08071v3 | https://arxiv.org/pdf/2201.08071v3.pdf | Temporal Sentence Grounding in Videos: A Survey and Future Directions | Temporal sentence grounding in videos (TSGV), \aka natural language video localization (NLVL) or video moment retrieval (VMR), aims to retrieve a temporal moment that semantically corresponds to a language query from an untrimmed video. Connecting computer vision and natural language, TSGV has drawn significant attention from researchers in both communities. This survey attempts to provide a summary of fundamental concepts in TSGV and current research status, as well as future research directions. As the background, we present a common structure of functional components in TSGV, in a tutorial style: from feature extraction from raw video and language query, to answer prediction of the target moment. Then we review the techniques for multimodal understanding and interaction, which is the key focus of TSGV for effective alignment between the two modalities. We construct a taxonomy of TSGV techniques and elaborate the methods in different categories with their strengths and weaknesses. Lastly, we discuss issues with the current TSGV research and share our insights about promising research directions. | ['Joey Tianyi Zhou', 'Wei Jing', 'Aixin Sun', 'Hao Zhang'] | 2022-01-20 | null | null | null | null | ['moment-retrieval'] | ['computer-vision'] | [ 2.63756990e-01 -1.35726333e-01 -6.96428359e-01 -2.17909798e-01
-8.69768202e-01 -9.03424382e-01 6.70168042e-01 -1.25961423e-01
-2.62459278e-01 2.64320195e-01 6.45939529e-01 -1.66105870e-02
-1.56615913e-01 -1.22286774e-01 -4.69810724e-01 -4.25619543e-01
-2.55403131e-01 -2.84850180e-01 2.04664618e-01 -8.11632946e-02
4.63256598e-01 3.52692485e-01 -1.70270109e+00 8.21752071e-01
2.79928029e-01 9.52731252e-01 2.69136101e-01 1.03522146e+00
-3.81747156e-01 1.45131636e+00 -4.74082679e-01 -4.64096040e-01
-1.76145703e-01 -6.52012706e-01 -1.20091498e+00 3.63791078e-01
6.28477156e-01 -4.42903191e-01 -9.85547304e-01 8.04356456e-01
2.38889769e-01 3.83185804e-01 4.06862974e-01 -1.73884344e+00
-4.53009963e-01 3.38533133e-01 -2.57728606e-01 6.27956152e-01
1.38018143e+00 -1.65562838e-01 8.72672141e-01 -8.98812175e-01
1.14584291e+00 1.17203534e+00 4.92959261e-01 6.71070099e-01
-7.25757360e-01 6.53035427e-03 3.62258762e-01 6.53958797e-01
-1.53248727e+00 -7.42416084e-01 6.86738014e-01 -5.57312191e-01
1.11165953e+00 5.52756727e-01 6.30191863e-01 1.18918777e+00
1.76036641e-01 1.39217269e+00 5.09941757e-01 -6.67080581e-01
-1.40820429e-01 2.63599330e-03 -5.92011549e-02 8.26730192e-01
-7.01902270e-01 -3.00171226e-01 -1.31335938e+00 -1.74370781e-01
6.86908782e-01 -1.62223831e-01 -4.93141860e-01 -4.21253741e-01
-1.79820204e+00 6.56888008e-01 -1.89574301e-01 4.50909406e-01
-9.89983156e-02 2.61315733e-01 7.10546017e-01 4.55353826e-01
2.14467958e-01 4.86987904e-02 -3.73535305e-02 -6.11578643e-01
-1.11257243e+00 2.71360576e-01 7.19464779e-01 1.20556688e+00
3.51122320e-01 5.59097901e-03 -2.32285634e-01 5.15263677e-01
4.26684797e-01 5.31583488e-01 4.43668783e-01 -1.42530251e+00
4.82955188e-01 1.08867012e-01 6.15394711e-02 -1.34203947e+00
-5.69581054e-02 5.72057545e-01 -1.30907550e-01 -4.06363875e-01
1.92432046e-01 9.86337364e-02 -4.23127234e-01 1.49947548e+00
5.14277853e-02 3.64226401e-01 1.63665310e-01 8.76734436e-01
1.36021638e+00 1.01728177e+00 5.31609543e-02 -6.85272574e-01
1.40336609e+00 -8.52377772e-01 -1.05529881e+00 -5.53531535e-02
7.39503443e-01 -8.58330667e-01 8.99104178e-01 3.38489920e-01
-1.21364689e+00 -2.06306696e-01 -6.87707782e-01 -3.57445151e-01
-4.40084189e-01 1.88537594e-02 4.89830106e-01 1.09314919e-01
-1.42694485e+00 2.86895871e-01 -8.11946869e-01 -1.12211502e+00
-2.37564996e-01 1.17705099e-01 -5.51197529e-01 3.07172742e-02
-1.20944643e+00 7.01909244e-01 4.03015435e-01 -2.71877442e-02
-7.17574477e-01 -2.03555420e-01 -1.16301823e+00 -5.37030220e-01
6.04990721e-01 -4.94367480e-01 1.60074842e+00 -9.56341147e-01
-1.39144325e+00 1.21335101e+00 -7.29912162e-01 -4.33013827e-01
2.88621366e-01 -3.67453456e-01 -7.07274497e-01 1.10236263e+00
1.36421099e-01 9.37082112e-01 8.53975654e-01 -8.23970020e-01
-8.40243936e-01 6.60423636e-02 3.46170306e-01 6.06950223e-01
-2.09022105e-01 6.60033405e-01 -1.00028479e+00 -8.60546947e-01
1.19360901e-01 -6.53460622e-01 1.59145132e-01 -8.98711085e-02
-9.70100611e-02 -3.30650419e-01 1.05657589e+00 -6.15414321e-01
1.85648680e+00 -2.12901473e+00 4.79691684e-01 -2.30092943e-01
2.66470641e-01 -2.81071931e-01 -6.80563897e-02 1.12091053e+00
-1.39304891e-01 3.12076882e-02 3.64665359e-01 -8.15411210e-02
-4.12376076e-02 1.35974899e-01 -8.41004193e-01 6.24528944e-01
-2.97357589e-01 1.10574639e+00 -1.08176231e+00 -9.33635354e-01
5.39673448e-01 3.92198861e-01 -1.41269550e-01 1.30369455e-01
-2.80924767e-01 3.38970304e-01 -5.11141002e-01 9.27402318e-01
-1.93642259e-01 -1.98460147e-01 1.41248807e-01 -5.03752410e-01
-3.72502446e-01 -3.68177854e-02 -7.46181846e-01 1.93847120e+00
4.55421284e-02 1.39968574e+00 -1.31990954e-01 -8.09510469e-01
3.42611790e-01 8.88343513e-01 9.05356586e-01 -5.82776964e-01
-1.40304789e-01 -8.44808221e-02 -7.80512631e-01 -1.15621340e+00
8.18467021e-01 1.89595908e-01 -4.16318625e-01 3.75393510e-01
2.90813267e-01 1.82394043e-01 4.68169093e-01 7.45904148e-01
8.90395224e-01 4.36266989e-01 5.95647275e-01 2.16882035e-01
6.73674047e-01 2.12167099e-01 9.41934288e-02 8.16639423e-01
-5.11831343e-01 4.25463885e-01 3.98852855e-01 -3.85300577e-01
-6.79027379e-01 -9.64223504e-01 3.71463060e-01 1.34315264e+00
3.72257411e-01 -1.02210784e+00 -7.53052533e-01 -4.88387316e-01
-4.10698473e-01 3.91848922e-01 -4.89488721e-01 1.55336872e-01
-5.75984955e-01 -5.89474477e-02 6.02672517e-01 5.00726819e-01
3.44928503e-01 -1.20519292e+00 -7.25139618e-01 -2.63201922e-01
-9.67695892e-01 -1.66471386e+00 -5.98015189e-01 -5.43997943e-01
-8.43901098e-01 -1.04544616e+00 -7.02275157e-01 -9.03680682e-01
3.14476550e-01 7.82247066e-01 1.12482226e+00 -3.18585485e-02
-2.13946775e-01 1.62301528e+00 -7.78360248e-01 9.47367549e-02
-3.55080068e-02 -4.34303969e-01 3.31557333e-01 2.97672227e-02
6.09158814e-01 -1.92830846e-01 -4.15570080e-01 3.23111564e-01
-1.04107201e+00 6.12660758e-02 -4.48903516e-02 3.49451393e-01
5.37749946e-01 -2.03286842e-01 1.93263385e-02 -1.94620967e-01
5.95492899e-01 -4.64197129e-01 -3.24238122e-01 4.74261314e-01
-2.28545852e-02 -3.85451943e-01 5.69127314e-02 -3.83117497e-01
-9.12043452e-01 -1.13036493e-02 8.90377611e-02 -7.15345025e-01
-3.04558188e-01 7.15884507e-01 1.33714536e-02 -9.25338939e-02
3.85875851e-01 6.09455526e-01 -1.55944407e-01 -2.55565733e-01
6.78799510e-01 3.74107510e-01 6.80491745e-01 -3.53912801e-01
3.15896273e-01 7.13302314e-01 -3.05926979e-01 -1.30345333e+00
-6.77705526e-01 -9.35120821e-01 -7.33381629e-01 -9.27817047e-01
1.01160574e+00 -7.78132617e-01 -8.87500703e-01 2.38418147e-01
-1.25114930e+00 -1.19602278e-01 -2.42325798e-01 7.10464478e-01
-1.07539701e+00 8.55128765e-01 -8.16922724e-01 -8.63023221e-01
9.40234773e-03 -9.98222172e-01 1.30887187e+00 1.87942564e-01
-4.45320845e-01 -1.16738498e+00 -2.74043139e-02 4.13199544e-01
-1.55402303e-01 2.68200070e-01 4.92050141e-01 -3.98176968e-01
-6.17914319e-01 -3.08778793e-01 2.36238725e-02 -1.44929364e-01
-6.45932108e-02 1.26314834e-01 -7.24088788e-01 -1.75830111e-01
1.14046104e-01 -3.38196933e-01 5.09497821e-01 7.25393176e-01
8.49833965e-01 -2.44049489e-01 -5.73427618e-01 4.10249591e-01
1.25193417e+00 5.46783447e-01 5.78522086e-01 3.77545983e-01
5.37964940e-01 6.89854205e-01 1.01024866e+00 3.37274522e-01
3.28439146e-01 7.59036362e-01 1.89511672e-01 3.49866003e-01
1.55739151e-02 -3.02738398e-01 9.94033217e-01 1.04027855e+00
-3.11851352e-01 -5.48113167e-01 -8.34667861e-01 4.04929876e-01
-2.08198428e+00 -1.51693070e+00 8.62015933e-02 1.85994482e+00
2.57631600e-01 -3.96833181e-01 3.72084379e-01 -4.14684117e-02
6.94368541e-01 6.18492305e-01 -1.71717867e-01 -1.55889690e-01
-2.77926236e-01 -6.51129901e-01 9.01722237e-02 5.57400823e-01
-1.27310836e+00 1.18603444e+00 7.85869980e+00 6.62908792e-01
-1.02997291e+00 7.95769021e-02 1.42412022e-01 -2.24868730e-01
-9.29887593e-03 9.64239516e-05 -5.64520121e-01 1.10205822e-01
1.01279807e+00 -3.88860613e-01 4.96893346e-01 5.88479578e-01
6.46530449e-01 -3.19418132e-01 -1.32345486e+00 1.71181536e+00
7.37617671e-01 -1.54955876e+00 1.57437921e-01 -2.46194243e-01
2.39896983e-01 -6.84111714e-02 7.46342763e-02 -1.07926158e-02
-4.39126432e-01 -7.98777163e-01 9.40518498e-01 7.72615790e-01
8.63737345e-01 -5.11060178e-01 3.71838301e-01 2.23371506e-01
-1.63764596e+00 7.41546378e-02 1.00262716e-01 1.43019691e-01
5.33080101e-01 -1.93546712e-01 -4.77544159e-01 6.63044870e-01
9.02286708e-01 1.02491271e+00 -3.45749587e-01 9.02387083e-01
7.68756941e-02 2.35657915e-01 4.27985080e-02 9.48207676e-02
2.10332260e-01 -1.73192039e-01 9.02138531e-01 1.45647323e+00
2.10937664e-01 3.46032768e-01 4.76064146e-01 1.64171278e-01
2.02689126e-01 1.34225398e-01 -9.82227087e-01 -5.73980749e-01
3.23900342e-01 7.99774945e-01 -9.71146703e-01 -4.35191572e-01
-8.03567529e-01 1.42982197e+00 -1.67137980e-01 6.90751493e-01
-9.04564381e-01 -3.79584700e-01 4.37164456e-01 -2.35764473e-03
3.06287743e-02 -5.00214279e-01 5.84288716e-01 -1.65273631e+00
-2.87307124e-03 -8.49898398e-01 7.87957311e-01 -1.25499427e+00
-9.41566229e-01 5.55403948e-01 5.77754200e-01 -1.67460179e+00
-6.78648770e-01 -5.34879684e-01 -1.32325441e-01 1.42208233e-01
-9.81634676e-01 -9.67672765e-01 -1.50802702e-01 8.98451269e-01
1.06877971e+00 -7.37904431e-03 6.76066279e-01 2.55769044e-01
-2.73844033e-01 1.44137383e-01 -2.71180063e-01 1.60704955e-01
6.87932432e-01 -9.08835590e-01 6.57483861e-02 9.34437752e-01
4.22088295e-01 6.72963262e-01 7.98370183e-01 -6.72164738e-01
-1.91198063e+00 -6.49334311e-01 1.16199005e+00 -6.93232298e-01
8.34276855e-01 -1.35920033e-01 -4.41704482e-01 1.14335299e+00
6.12398505e-01 -2.46108219e-01 8.32684398e-01 -4.45233434e-01
-1.36398405e-01 9.96428356e-02 -7.89242029e-01 9.75370884e-01
1.08994710e+00 -1.10862458e+00 -7.23946393e-01 5.47661722e-01
6.72263026e-01 -5.10280132e-01 -8.71854007e-01 1.21951960e-02
7.90461957e-01 -1.00142121e+00 1.04451525e+00 -6.25857472e-01
1.27778664e-01 -3.94957751e-01 -4.60676342e-01 -5.16231596e-01
1.41483009e-01 -1.08424044e+00 -5.89840472e-01 1.03155959e+00
-1.40871197e-01 3.21201840e-03 6.48669839e-01 5.68809748e-01
-6.32248521e-02 -4.76293564e-01 -7.67232478e-01 -6.03703320e-01
-5.91484725e-01 -1.08337450e+00 -1.58749342e-01 9.91588056e-01
7.44676232e-01 2.59801447e-01 -7.06096292e-01 3.30331475e-02
3.72294545e-01 7.96618462e-02 5.63013554e-01 -6.55378282e-01
1.77809656e-01 -3.13644320e-01 -8.39109242e-01 -1.48632777e+00
1.85049325e-01 -6.28338218e-01 -2.90308669e-02 -1.66555488e+00
2.16945380e-01 8.72216344e-01 4.97133918e-02 2.44358748e-01
5.48135698e-01 3.66079509e-01 4.21016157e-01 3.55239868e-01
-1.32108462e+00 1.91652492e-01 8.90712798e-01 3.63449715e-02
-2.37785041e-01 -1.66491747e-01 -2.78145820e-01 1.00318861e+00
3.15701514e-01 8.78108740e-02 -8.50000441e-01 -4.67923045e-01
4.02738661e-01 6.27127945e-01 3.90880853e-01 -6.60520732e-01
5.25026739e-01 -3.64060730e-01 1.10497288e-01 -1.03574812e+00
6.40111744e-01 -7.93181002e-01 3.02116033e-02 2.35871356e-02
-4.26100641e-01 5.76577008e-01 7.73101300e-02 6.44068420e-01
-5.96959293e-01 -8.01488534e-02 2.83115774e-01 -2.62177557e-01
-1.66845310e+00 2.96888910e-02 -8.89342785e-01 -9.38465446e-02
1.22474849e+00 -6.59891009e-01 -1.14147916e-01 -1.05588222e+00
-1.13504922e+00 3.03405821e-01 2.78695345e-01 6.96604252e-01
1.19049966e+00 -1.46957910e+00 -1.63357079e-01 -3.92535895e-01
2.33975202e-01 -8.73386741e-01 4.61913764e-01 1.18908322e+00
-5.17603993e-01 8.24425042e-01 2.96682268e-01 -8.53168845e-01
-1.62587345e+00 8.49303901e-01 1.66503862e-01 3.59748244e-01
-6.62553370e-01 6.45541549e-01 9.65492204e-02 3.22374105e-01
5.92584670e-01 -1.67178378e-01 -4.57767218e-01 2.39603266e-01
8.42009604e-01 3.46567720e-01 -2.87058741e-01 -1.26831210e+00
-6.65941060e-01 7.05364823e-01 1.76465705e-01 -4.90981519e-01
7.59150863e-01 -9.45384026e-01 -3.02607179e-01 9.20570850e-01
1.22909009e+00 -1.30005777e-01 -7.29443550e-01 -2.77698815e-01
3.79100710e-01 -4.67063129e-01 -8.95463154e-02 -2.96670675e-01
-6.80018842e-01 5.89088798e-01 4.70925659e-01 3.71317685e-01
1.33452737e+00 5.21378636e-01 7.00737298e-01 4.14790034e-01
3.80874366e-01 -1.03977144e+00 4.51798022e-01 5.41724086e-01
9.39843476e-01 -9.90937650e-01 1.32241724e-02 -4.25018936e-01
-8.63960624e-01 1.45637143e+00 2.98766553e-01 2.89586872e-01
6.47952795e-01 1.23543888e-02 1.47512123e-01 -4.60309118e-01
-6.67031825e-01 -2.67240703e-01 5.44899523e-01 5.47659159e-01
5.66319466e-01 -4.97783750e-01 -1.81099102e-01 1.94633141e-01
2.84830872e-02 1.99286878e-01 3.92860025e-01 1.43359470e+00
-3.83502275e-01 -8.28386486e-01 -5.73544621e-01 1.89254712e-02
-5.08738875e-01 -4.26830091e-02 -5.01850367e-01 8.74214172e-01
-2.48468131e-01 1.18017769e+00 1.67458981e-01 -6.04861677e-01
2.45157465e-01 1.48341402e-01 6.52349412e-01 -3.81987214e-01
-1.14386484e-01 4.74935561e-01 1.24026231e-01 -1.10805321e+00
-1.22647274e+00 -7.87910223e-01 -1.09355295e+00 -2.97436029e-01
-2.91621611e-02 2.36031175e-01 2.54999757e-01 1.12094367e+00
2.19043657e-01 7.26662402e-04 2.45469034e-01 -9.81466174e-01
3.50934654e-01 -4.09438103e-01 -4.41599578e-01 5.65960817e-02
5.20376265e-01 -4.43919271e-01 -1.29421473e-01 8.81626308e-01] | [10.118961334228516, 0.8265277147293091] |
ab26ab44-3cdc-48cd-9068-f9ed1f6f98bd | momentum-calibration-for-text-generation | 2212.04257 | null | https://arxiv.org/abs/2212.04257v1 | https://arxiv.org/pdf/2212.04257v1.pdf | Momentum Calibration for Text Generation | The input and output of most text generation tasks can be transformed to two sequences of tokens and they can be modeled using sequence-to-sequence learning modeling tools such as Transformers. These models are usually trained by maximizing the likelihood the output text sequence and assumes the input sequence and all gold preceding tokens are given during training, while during inference the model suffers from the exposure bias problem (i.e., it only has access to its previously predicted tokens rather gold tokens during beam search). In this paper, we propose MoCa ({\bf Mo}mentum {\bf Ca}libration) for text generation. MoCa is an online method that dynamically generates slowly evolving (but consistent) samples using a momentum moving average generator with beam search and MoCa learns to align its model scores of these samples with their actual qualities. Experiments on four text generation datasets (i.e., CNN/DailyMail, XSum, SAMSum and Gigaword) show MoCa consistently improves strong pre-trained transformers using vanilla fine-tuning and we achieve the state-of-the-art results on CNN/DailyMail and SAMSum datasets. | ['Furu Wei', 'Wayne Xiong', 'Si-Qing Chen', 'Yang Yu', 'Pengcheng He', 'Xun Wang', 'Yiran Liu', 'Xingxing Zhang'] | 2022-12-08 | null | null | null | null | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 4.01494712e-01 2.68492222e-01 -3.69227111e-01 -1.57035798e-01
-8.18910182e-01 -5.72214901e-01 1.03962708e+00 -1.35715172e-01
-2.31967986e-01 1.12914646e+00 3.79887372e-01 -2.91540951e-01
2.81349421e-01 -9.05757129e-01 -9.23288226e-01 -6.55292749e-01
3.38496506e-01 1.00127292e+00 -2.53740624e-02 -3.12597752e-01
3.50049138e-01 -2.55625457e-01 -1.39225411e+00 5.74824095e-01
1.18203521e+00 7.61307061e-01 4.91874367e-01 1.27928388e+00
-4.70761776e-01 1.11448729e+00 -9.17126238e-01 -7.30462193e-01
1.25876158e-01 -9.78539348e-01 -1.03792310e+00 -2.69850791e-01
4.38792944e-01 -2.27186233e-01 -1.35409653e-01 8.50891113e-01
6.19647145e-01 2.37513125e-01 9.31408048e-01 -1.37369728e+00
-1.00779390e+00 1.39879739e+00 -2.84539729e-01 1.62661135e-01
8.91530216e-02 4.23281014e-01 1.30564868e+00 -9.76843894e-01
8.46077800e-01 1.19227850e+00 4.24637169e-01 1.03277004e+00
-1.31219769e+00 -6.65928841e-01 2.53923893e-01 9.14974883e-02
-9.74529743e-01 -2.95447648e-01 3.01183850e-01 -4.12923306e-01
1.34758961e+00 3.15625638e-01 7.12441981e-01 1.47206795e+00
4.84455198e-01 1.32154298e+00 5.22403538e-01 -5.78395069e-01
3.51733536e-01 4.30770628e-02 -2.21679270e-01 6.10907972e-01
-1.33018360e-01 -3.43159698e-02 -1.04762280e+00 -2.10692108e-01
5.08713543e-01 -2.43179083e-01 -1.05452366e-01 1.68249011e-01
-1.24846590e+00 7.94616222e-01 1.11083999e-01 3.03266943e-01
-3.70424777e-01 6.55588388e-01 1.44761011e-01 2.07713813e-01
5.16541660e-01 6.20686829e-01 -5.38276255e-01 -6.34845495e-01
-1.18651557e+00 7.26771772e-01 6.13742352e-01 1.39754975e+00
5.87334871e-01 4.03644800e-01 -8.74838173e-01 8.86085570e-01
2.11862504e-01 4.16960716e-01 1.09539461e+00 -4.14888233e-01
9.29322839e-01 1.05397083e-01 9.95675474e-02 -1.17270537e-01
1.10845841e-01 -4.13274914e-01 -9.59978402e-01 2.66918857e-02
3.83142263e-01 -4.53379393e-01 -1.38585317e+00 1.82080269e+00
-4.48596440e-02 1.54907033e-01 1.01483755e-01 4.36597198e-01
4.53264982e-01 1.11388338e+00 1.07665181e-01 -1.49455234e-01
9.56535459e-01 -1.28445601e+00 -7.67277241e-01 -4.56747979e-01
6.05449975e-01 -6.50737286e-01 1.30664098e+00 6.12612844e-01
-1.53529477e+00 -5.67625284e-01 -1.02393711e+00 6.56101946e-03
-2.46992260e-01 8.14784318e-02 4.70626861e-01 6.13092721e-01
-1.00605404e+00 1.08183324e+00 -7.89339364e-01 8.25171322e-02
5.65795004e-01 -8.21229741e-02 2.34656364e-01 2.71623105e-01
-1.24570560e+00 6.89624727e-01 6.13930643e-01 -9.55575630e-02
-1.24704850e+00 -9.40862715e-01 -5.34089804e-01 2.59099424e-01
-5.94338775e-02 -1.04368269e+00 1.77745569e+00 -9.14752126e-01
-1.83374858e+00 3.91186446e-01 -4.03358161e-01 -7.51095355e-01
9.07302916e-01 -4.36217010e-01 -1.63227990e-01 -4.99516606e-01
8.23994204e-02 9.25937772e-01 1.05495119e+00 -9.28892493e-01
-9.31849122e-01 1.64568484e-01 -4.11416978e-01 1.27075210e-01
-3.09351861e-01 -3.25326353e-01 -5.08745387e-02 -9.94307697e-01
-1.51477188e-01 -8.57858121e-01 -1.37524873e-01 -4.86062288e-01
-7.60742605e-01 -5.65200090e-01 2.33959615e-01 -4.78435069e-01
1.43659008e+00 -1.45640182e+00 3.83392930e-01 -8.82522017e-03
-7.12952465e-02 1.22240014e-01 -1.91903099e-01 6.45762444e-01
1.58552527e-02 5.19802213e-01 -1.57175019e-01 -5.94490290e-01
3.33139747e-01 -4.71046753e-02 -7.33038723e-01 -2.16803074e-01
2.97935188e-01 1.23141170e+00 -1.27938116e+00 -5.24313033e-01
-2.08143413e-01 7.23260269e-03 -6.64499938e-01 3.54687452e-01
-9.68673527e-01 8.95303488e-02 -1.63004652e-01 3.52720588e-01
3.60218972e-01 -5.29939532e-01 1.79990277e-01 3.20913017e-01
1.16112813e-01 6.07426465e-01 -8.18433464e-01 1.80977058e+00
-5.67825973e-01 7.48946428e-01 -7.86464155e-01 -4.47841287e-01
8.62647176e-01 5.22164464e-01 -1.06523618e-01 -4.87639636e-01
-4.71533276e-02 1.10669710e-01 -5.99151812e-02 -2.39335701e-01
1.00120854e+00 -1.63043618e-01 1.69737220e-01 9.84897196e-01
3.76128227e-01 -3.68277580e-01 4.82195079e-01 4.80858386e-01
8.93152714e-01 3.20292592e-01 6.25495166e-02 1.05317287e-01
8.87646824e-02 -1.13486670e-01 3.84350389e-01 1.18656862e+00
4.95238215e-01 8.38272989e-01 5.51641107e-01 -2.08012506e-01
-1.34213638e+00 -1.18489301e+00 1.65627003e-01 1.34828091e+00
-4.67543811e-01 -7.20098495e-01 -8.79581690e-01 -7.84459233e-01
-2.36459970e-01 1.34324491e+00 -7.94320941e-01 -3.41270894e-01
-5.31147957e-01 -1.01225114e+00 6.69471920e-01 7.17786849e-01
1.57514334e-01 -1.61746037e+00 -4.07187611e-01 6.00314140e-01
-4.05216187e-01 -3.93905371e-01 -7.63447881e-01 2.09567443e-01
-9.58131433e-01 -4.41554010e-01 -8.63794088e-01 -6.84774518e-01
6.03971183e-01 -5.17149031e-01 1.51267338e+00 3.18116210e-02
-1.26706183e-01 -4.24568921e-01 -4.81940180e-01 -7.30215132e-01
-8.43120039e-01 5.42081416e-01 -2.20523357e-01 -1.92182779e-01
-1.95225533e-02 -4.64943349e-01 -3.69989961e-01 -4.98422515e-03
-8.37070942e-01 5.41143060e-01 6.46468461e-01 1.22392440e+00
6.39607549e-01 -2.41799682e-01 7.58869767e-01 -1.07327104e+00
1.03482950e+00 -5.09116054e-01 -2.72753328e-01 4.96078223e-01
-9.58781838e-01 5.09100795e-01 8.62119794e-01 -7.67002225e-01
-1.13605285e+00 -2.91587412e-01 -3.33918817e-02 -3.73324752e-01
1.91230670e-01 3.62267643e-01 4.77059484e-02 9.18002665e-01
9.16978598e-01 7.19559789e-01 -3.84056896e-01 -4.00091976e-01
5.98763108e-01 5.42569935e-01 5.00604630e-01 -7.00386524e-01
6.95007026e-01 -7.31058270e-02 -4.16256636e-01 -2.59348810e-01
-8.98486972e-01 2.85906434e-01 -2.54148453e-01 -1.86634541e-01
7.66320348e-01 -6.74194753e-01 -3.94575208e-01 8.10312331e-01
-1.25821471e+00 -8.96852672e-01 -5.36203444e-01 1.75811112e-01
-7.47473001e-01 -1.30844256e-02 -6.46374345e-01 -9.29992557e-01
-1.00475812e+00 -8.26288223e-01 8.38366568e-01 3.56595159e-01
-6.10591829e-01 -1.01699972e+00 2.40796775e-01 -4.55317162e-02
4.55165714e-01 -8.79297033e-02 1.14344704e+00 -5.27880788e-01
-7.17750251e-01 2.66346540e-02 4.62582320e-01 3.05027753e-01
7.24154115e-02 2.09050059e-01 -9.45158422e-01 -2.00754479e-01
-4.88896608e-01 -5.54522693e-01 9.80838239e-01 3.50295484e-01
1.33716834e+00 -7.07069218e-01 -3.45689803e-01 4.76065129e-01
1.00768602e+00 1.34379491e-01 7.82218933e-01 1.64102271e-01
5.04147410e-01 1.79733455e-01 5.99944174e-01 6.65833294e-01
1.82445005e-01 3.64498287e-01 3.76864105e-01 3.61483872e-01
-8.86035785e-02 -9.38292563e-01 5.01841247e-01 7.01704860e-01
9.86096412e-02 -1.03073192e+00 -5.48261285e-01 6.94951415e-01
-2.06588864e+00 -1.41499841e+00 -6.96856435e-03 2.02899790e+00
1.42974567e+00 4.48512584e-01 -5.21528348e-02 6.04113638e-02
6.33283675e-01 8.82179365e-02 -7.02728331e-01 -4.46011722e-01
-1.45841151e-01 6.49568200e-01 2.18561590e-01 4.45865750e-01
-5.93064964e-01 1.17252874e+00 6.50719833e+00 1.22048974e+00
-9.93383408e-01 3.73133868e-02 8.63869309e-01 -6.21315300e-01
-5.42739868e-01 -1.63587600e-01 -1.04982626e+00 8.45211089e-01
9.84218717e-01 -5.69145679e-01 6.28899693e-01 8.14079344e-01
-2.25576144e-02 7.12435693e-02 -1.42287159e+00 7.17670143e-01
-3.26997861e-02 -1.74556839e+00 3.66348594e-01 -1.48661911e-01
1.26747072e+00 -1.18514985e-01 3.91432822e-01 4.76163298e-01
1.06553423e+00 -1.31820107e+00 1.32186425e+00 6.89880252e-01
8.45161915e-01 -7.50467420e-01 4.64527041e-01 7.05769122e-01
-8.14487398e-01 -2.33193375e-02 -4.66833055e-01 2.04462275e-01
3.06257784e-01 6.36519253e-01 -1.54552829e+00 3.92429680e-01
5.58299541e-01 6.04533970e-01 -3.38788360e-01 7.10747838e-01
-5.17476678e-01 9.78531480e-01 2.76733772e-03 -6.75964177e-01
2.53262848e-01 3.36806290e-02 3.81290406e-01 1.26115894e+00
6.48187339e-01 -2.43615448e-01 -2.43355334e-01 1.36353087e+00
-4.37306285e-01 -9.59367454e-02 -3.38900983e-02 -2.45796233e-01
6.99459255e-01 1.21581745e+00 -3.97614300e-01 -7.66439736e-01
3.89663160e-01 9.85094249e-01 3.08995903e-01 2.97842056e-01
-7.63206005e-01 -5.63335121e-01 5.40795088e-01 -3.03881112e-02
4.92478341e-01 2.68987089e-01 -4.76499826e-01 -1.09349382e+00
-1.57626003e-01 -9.62523520e-01 2.88126260e-01 -9.45730507e-01
-1.31510866e+00 6.62231863e-01 -2.33102798e-01 -1.11321735e+00
-1.02772117e+00 -2.15552077e-01 -1.10081756e+00 1.20685649e+00
-1.12946343e+00 -9.13149059e-01 -7.85729513e-02 3.38748038e-01
1.07425702e+00 -3.54481429e-01 6.65487111e-01 -2.70802319e-01
-4.03362423e-01 9.41971064e-01 2.20390007e-01 8.38070735e-02
6.11312747e-01 -1.67219269e+00 1.19224715e+00 6.57444477e-01
2.85502583e-01 4.29226816e-01 7.87801564e-01 -8.42811644e-01
-9.59482968e-01 -1.21944237e+00 1.33137703e+00 -6.68347478e-01
4.51134175e-01 -7.00298667e-01 -7.76197135e-01 8.63180161e-01
4.81624186e-01 -5.54085493e-01 3.86128038e-01 -1.17156848e-01
8.54471326e-03 3.34521592e-01 -9.47822869e-01 8.10408652e-01
1.17743373e+00 -1.49484560e-01 -4.41900879e-01 5.05397677e-01
7.37866402e-01 -6.99962676e-01 -5.46525002e-01 -2.16483995e-01
5.58471859e-01 -7.21857190e-01 6.29344344e-01 -1.03337824e+00
9.58411396e-01 -8.13254528e-03 2.90931046e-01 -1.82433069e+00
-3.58542114e-01 -9.95230734e-01 -5.68961501e-01 1.38205636e+00
9.37956572e-01 -3.03492337e-01 1.00335848e+00 1.71444535e-01
-1.76374167e-01 -9.06054199e-01 -6.96739376e-01 -7.05530107e-01
4.75190043e-01 -4.50580031e-01 1.12720811e+00 6.63723171e-01
2.28072435e-01 5.06829739e-01 -5.62120497e-01 -4.46097881e-01
4.29707527e-01 2.90916357e-02 7.88915157e-01 -9.33458745e-01
-5.34201205e-01 -6.20380342e-01 3.47755164e-01 -1.40848744e+00
9.13971364e-02 -1.39875340e+00 4.66814667e-01 -1.56140745e+00
1.53133661e-01 -4.69701231e-01 -3.93978506e-02 5.20872355e-01
-5.83751321e-01 -1.96328521e-01 2.07451969e-01 5.21177538e-02
-3.78290355e-01 8.54612052e-01 1.24476874e+00 -2.65631646e-01
-2.29403347e-01 1.82726204e-01 -6.47485852e-01 1.24263093e-01
8.00121903e-01 -7.08072960e-01 -4.98712063e-01 -5.97941458e-01
6.40541375e-01 5.40148839e-02 3.09250820e-02 -8.32411945e-01
2.23059118e-01 -2.85213321e-01 7.11097240e-01 -7.07257986e-01
1.19802833e-01 1.61029100e-01 2.81246334e-01 4.84977782e-01
-1.02864575e+00 2.09218934e-01 -1.02073587e-01 4.64787930e-01
2.17771590e-01 -6.80648625e-01 5.27117014e-01 -3.26238781e-01
-7.45241940e-02 3.60396653e-01 -6.66148007e-01 4.11778957e-01
5.05870163e-01 7.54170045e-02 -3.75765324e-01 -6.02514207e-01
-4.75397170e-01 5.50570011e-01 1.54218540e-01 7.02060521e-01
5.53615212e-01 -1.49269998e+00 -1.10828972e+00 1.18497178e-01
-7.55236894e-02 2.52471179e-01 6.51443824e-02 2.38644585e-01
-2.69827038e-01 4.04245049e-01 6.90713152e-02 -4.43498105e-01
-9.43072081e-01 2.02967450e-01 3.34684193e-01 -8.53787601e-01
-3.39477777e-01 1.29060054e+00 -1.05602950e-01 -4.67731714e-01
1.49185553e-01 -4.40712303e-01 1.46533996e-01 -9.58134905e-02
6.27534747e-01 4.26927239e-01 6.29457682e-02 1.27024189e-01
1.68296441e-01 -9.03646350e-02 -5.40916979e-01 -6.26045465e-01
1.12595868e+00 3.60486865e-01 1.92285180e-01 4.45720017e-01
8.00725222e-01 -2.84217000e-01 -1.38711834e+00 -1.54429048e-01
-5.99658974e-02 -4.50968981e-01 -3.33796322e-01 -1.04272914e+00
-9.21995223e-01 9.64506984e-01 2.20210209e-01 1.18482560e-01
5.57945728e-01 -5.70377111e-02 9.88370299e-01 3.94022614e-01
4.51532602e-02 -1.30032504e+00 6.26441002e-01 7.10163176e-01
8.77526462e-01 -6.89180493e-01 -3.78008902e-01 2.08754644e-01
-8.84176314e-01 1.04907465e+00 8.13436508e-01 1.13187090e-01
1.94963500e-01 1.87459618e-01 -1.88593969e-01 2.82911807e-01
-1.63304150e+00 2.20998123e-01 2.46710405e-02 5.97601593e-01
6.99802339e-01 2.18405146e-02 -9.68540907e-02 6.28878236e-01
-9.65787649e-01 1.15545355e-01 5.84217429e-01 8.89286935e-01
-5.61106145e-01 -1.27158129e+00 -1.45442173e-01 8.32478106e-01
-4.02183205e-01 -6.28183663e-01 -3.68406892e-01 2.84083903e-01
3.24164033e-02 8.25850010e-01 3.22922826e-01 -3.49636316e-01
1.61261056e-02 4.58808452e-01 5.18610358e-01 -6.64974630e-01
-9.77867842e-01 -1.68085828e-01 -7.57277943e-03 3.41236107e-02
4.26553220e-01 -8.48217309e-01 -1.32141149e+00 -3.59350741e-01
-4.88564849e-01 3.25044394e-01 5.82315028e-01 7.80442059e-01
2.38869920e-01 8.04444790e-01 5.72209537e-01 -6.77654803e-01
-8.84926081e-01 -1.45665121e+00 -3.18107992e-01 3.04434687e-01
2.09920436e-01 -1.37443626e-02 -1.63754359e-01 3.68946552e-01] | [11.820849418640137, 9.0680570602417] |
ed35f8ce-08d0-4cc3-8070-8d80fc3681bb | an-in-network-data-cleaning-approach-for | null | null | http://www.tandfonline.com/loi/tasj20 | http://www.tandfonline.com/loi/tasj20 | An in-network data cleaning approach for wireless sensor networks | Wireless Sensor Networks (WSNs) are widely used for monitoring physical happenings of the
environment. However, the data gathered by the WSNs may be inaccurate and unreliable due to power
exhaustion, noise and other reasons. Unnecessary data such as erroneous data and redundant data
transmission causes a lot of extra energy consumption. To improve the data reliability and reduce the
energy consumption, we proposed an in-network processing architecture for data cleaning, which
divides the task into four stages implemented in different nodes respectively. This strategy guaranteed
the cleaning algorithms were computationally lightweight in local nodes and energy-efficient due to
almost no communication overhead. In addition, we presented the detection algorithms for data faults
and event outliers, which were conducted by utilizing the related attributes from the local sensor node
and the cooperation with its relaying neighbor. Experiment results show that our proposed approach
is accurate and energy-efficient. | ['Jun Huanga and Haeyoung Baeb', 'Haiyang Bia', 'Ying Xiaa', 'Jianjun Leia'] | 2016-03-17 | null | null | null | journal-2016-3 | ['clustering-algorithms-evaluation'] | ['methodology'] | [ 1.36291340e-01 3.08249831e-01 2.31361762e-01 -5.61744452e-01
-1.05992584e-02 -1.65667161e-01 -1.38985500e-01 9.32358027e-01
-5.29318988e-01 9.62570906e-01 -1.55717835e-01 1.66489020e-01
-5.85701942e-01 -1.21510828e+00 -3.23303729e-01 -1.16971672e+00
-5.18767774e-01 3.19859385e-02 5.24809361e-01 3.81701253e-02
1.74466550e-01 4.23516750e-01 -1.65106976e+00 -4.82211500e-01
8.90434980e-01 1.04802883e+00 3.21778685e-01 8.69469494e-02
-1.31683918e-02 3.15450579e-01 -8.59463274e-01 2.91066557e-01
2.74456870e-02 -7.75677264e-02 -4.43774521e-01 -7.40839839e-02
-9.68754411e-01 -3.63214284e-01 7.78295007e-03 1.05439413e+00
5.13347507e-01 -1.45686731e-01 -5.79691045e-02 -1.82520032e+00
-5.74232154e-02 8.28191698e-01 -8.50239933e-01 -7.71060437e-02
1.52874947e-01 -4.20548469e-01 3.57130095e-02 -3.49924713e-01
3.06657493e-01 1.09612167e+00 8.68472338e-01 3.39303136e-01
-5.74424684e-01 -1.10656846e+00 3.97048928e-02 2.95086622e-01
-1.18535936e+00 -6.42723560e-01 8.50084364e-01 4.19590801e-01
4.02258486e-01 4.93627161e-01 6.60104871e-01 6.05238080e-01
4.41200078e-01 2.97015399e-01 7.24308193e-01 -2.72562385e-01
6.28044128e-01 -1.89079374e-01 1.93404928e-01 1.96421385e-01
1.29490960e+00 -2.84584425e-02 -2.41621196e-01 -2.98018664e-01
1.17282502e-01 6.09316111e-01 -4.99622263e-02 -8.52952972e-02
-1.03000796e+00 4.54958141e-01 5.72151661e-01 4.93386030e-01
-9.02852595e-01 1.67190582e-01 6.06522739e-01 1.82652533e-01
2.73317456e-01 -3.09955627e-01 -6.29062653e-01 2.01926649e-01
-5.24106205e-01 -4.76537049e-01 7.03313410e-01 1.09940231e+00
7.01006353e-01 -4.15520594e-02 3.90527189e-01 3.17587107e-01
6.91334486e-01 6.83638930e-01 2.48224989e-01 -8.20193648e-01
2.22421333e-01 8.94649565e-01 1.42193675e-01 -1.23844957e+00
-9.46806610e-01 -8.64713863e-02 -1.66137183e+00 2.08440319e-01
-2.78406024e-01 -4.98609453e-01 -8.07788670e-01 1.57837868e+00
7.38006175e-01 1.19710237e-01 3.79428476e-01 3.98446202e-01
7.15807617e-01 7.24442124e-01 4.82265472e-01 -1.02002299e+00
1.22613788e+00 -5.85370064e-02 -1.47655225e+00 1.01492457e-01
2.97500402e-01 -5.90520382e-01 1.76208511e-01 3.53421330e-01
-1.00526607e+00 -3.24523091e-01 -1.28014886e+00 4.52380776e-01
-5.41293383e-01 -2.78811425e-01 7.24203110e-01 4.66030151e-01
-8.37743104e-01 3.94884497e-01 -1.21347809e+00 -6.95746243e-01
4.60728317e-01 6.08249724e-01 -3.47595304e-01 2.10475009e-02
-8.53140652e-01 3.60833913e-01 6.59345329e-01 4.16587919e-01
-6.44329131e-01 -3.62658381e-01 -8.44428480e-01 -4.44739871e-02
3.35582346e-01 -3.08145434e-01 8.52278352e-01 -2.88392246e-01
-7.12025642e-01 -2.30839606e-02 -3.13111871e-01 -2.69299388e-01
4.67592999e-02 2.82822222e-01 -7.90673733e-01 -6.04502112e-02
1.52070642e-01 3.74375209e-02 8.04339051e-02 -1.37696266e+00
-8.57175410e-01 -7.21009314e-01 -6.45173550e-01 -1.25101328e-01
-4.77378666e-01 -1.06552072e-01 8.49229190e-03 -3.14476579e-01
7.74843991e-01 -3.77928346e-01 -5.26652038e-01 -3.85155529e-02
-4.77959305e-01 -3.11299354e-01 1.43623686e+00 -5.50394356e-01
1.28562629e+00 -2.17543173e+00 -5.02915740e-01 7.20355451e-01
8.94534513e-02 1.23678416e-01 2.53700316e-01 7.08535671e-01
2.79856026e-01 4.06616658e-01 -3.33297819e-01 -1.94557101e-01
-5.43232203e-01 7.29054868e-01 2.25327775e-01 5.53120196e-01
-2.05204144e-01 3.59936282e-02 -9.76727486e-01 -7.17182755e-01
3.23288769e-01 5.73331773e-01 2.49642476e-01 3.04691762e-01
3.74117732e-01 3.52231354e-01 -8.67045581e-01 9.73478556e-01
1.10527265e+00 1.84509680e-01 2.66720921e-01 -3.85662347e-01
-2.67454892e-01 -7.64917359e-02 -1.78359413e+00 1.43349123e+00
-2.09091172e-01 -5.54778613e-03 9.60114419e-01 -1.12635887e+00
9.35171366e-01 5.53727329e-01 1.10938728e+00 -7.37334251e-01
3.80630046e-01 2.16313958e-01 -3.70174676e-01 -9.39269602e-01
2.69389272e-01 4.27603394e-01 -8.04387182e-02 3.95336598e-01
-5.35230935e-01 5.05804896e-01 2.97129214e-01 3.86532396e-02
1.52147281e+00 -1.87096700e-01 4.04165745e-01 -2.10278228e-01
4.56670582e-01 1.40102625e-01 1.16194725e+00 5.28107464e-01
-1.33816049e-01 -2.83952057e-01 -2.32615665e-01 -3.23060364e-01
-5.60048044e-01 -6.60381138e-01 -9.49731842e-02 6.11717582e-01
7.81662166e-01 -2.28961527e-01 -4.91053462e-01 -2.08751008e-01
-6.36029392e-02 3.92472506e-01 -4.69348550e-01 -2.75190026e-01
-3.49558264e-01 -1.04043150e+00 4.04690683e-01 2.66741157e-01
8.52104187e-01 -1.18676198e+00 -8.96726012e-01 6.87965631e-01
3.56923491e-02 -8.08319807e-01 2.69123197e-01 4.82023805e-01
-1.09986031e+00 -1.50253618e+00 3.41129482e-01 -6.54708207e-01
1.15072346e+00 6.40699863e-01 5.36475241e-01 6.62801802e-01
-2.71559983e-01 -3.05881780e-02 -7.08731830e-01 -1.08270109e+00
-2.96661943e-01 -2.00800896e-01 2.26963907e-01 -3.85705322e-01
3.10574770e-01 -9.27513838e-01 -8.27947557e-01 3.07491690e-01
-1.33380497e+00 -2.96544820e-01 9.28362727e-01 2.11259782e-01
5.65432966e-01 9.50751901e-01 1.02480602e+00 -8.34501386e-01
6.45083487e-01 -1.07332861e+00 -5.09406805e-01 2.48700872e-01
-7.09408820e-01 -1.10680401e-01 6.47082090e-01 3.15333251e-03
-9.93387103e-01 2.55202562e-01 -5.47483563e-02 4.67302620e-01
-2.62414664e-01 1.09819837e-01 -5.42803586e-01 -1.16124973e-01
6.95538800e-03 -1.79236040e-01 4.08412069e-02 -6.82830036e-01
-4.09258246e-01 8.86440873e-01 4.52591568e-01 -7.21096843e-02
9.25172925e-01 6.04968548e-01 5.22792876e-01 -8.01757753e-01
-1.65279597e-01 -2.69232512e-01 -1.53721347e-01 -5.93668371e-02
5.17266452e-01 -8.77911031e-01 -1.13811958e+00 5.00507891e-01
-1.23081946e+00 4.75747317e-01 -5.81318177e-02 4.38274473e-01
5.98285437e-01 4.36509371e-01 -9.02784169e-02 -1.34909475e+00
-8.04540336e-01 -7.76622415e-01 6.45074964e-01 6.20531678e-01
1.81618109e-01 -6.22015834e-01 -3.24398726e-01 -3.62403691e-01
6.40858293e-01 9.78784442e-01 4.21372831e-01 -6.38492823e-01
-3.40202510e-01 -2.89330751e-01 -2.25138381e-01 -3.98557670e-02
7.39107192e-01 3.07579070e-01 -5.60319126e-01 -3.98981720e-01
1.62091255e-01 9.90350097e-02 4.36891735e-01 1.92987978e-01
1.46395659e+00 -5.77274263e-01 -7.68945634e-01 4.67846543e-01
1.69176233e+00 5.99178255e-01 5.06583750e-01 2.13794112e-01
3.44959438e-01 5.75543582e-01 9.15307581e-01 8.29618216e-01
4.71592158e-01 -1.38711184e-01 1.12509811e+00 -3.56771737e-01
3.13486010e-01 1.28684938e-01 1.21608980e-01 9.03932273e-01
7.01071694e-02 -5.77393174e-01 -3.89552146e-01 7.77531624e-01
-1.88437724e+00 -7.28828788e-01 -6.65042639e-01 2.06741500e+00
6.77085280e-01 1.46032527e-01 -2.18367040e-01 1.05296242e+00
7.93780148e-01 -1.26128003e-01 -6.49192393e-01 -1.65614709e-01
8.78012702e-02 8.76609087e-02 9.53610241e-01 5.31566180e-02
-6.14744425e-01 -1.43427715e-01 5.28494978e+00 2.91507691e-01
-5.74705601e-01 2.19116256e-01 8.81140903e-02 2.54174799e-01
-2.13924050e-01 1.20209912e-02 -4.44133490e-01 8.10369432e-01
8.09020758e-01 3.44290249e-02 7.71738440e-02 6.21402979e-01
8.02372158e-01 -6.76976502e-01 -5.45013189e-01 5.67129552e-01
-5.09415627e-01 -8.62763882e-01 -4.92679626e-01 -1.66439768e-02
4.37496752e-01 -9.95897502e-02 -9.20316279e-01 -3.74538630e-01
5.96071482e-01 -4.52555865e-01 2.33520150e-01 4.83406246e-01
3.29102665e-01 -9.04079497e-01 1.17554832e+00 5.26624203e-01
-1.38675845e+00 -3.55991304e-01 -4.13174301e-01 -1.24786742e-01
2.11104572e-01 1.22934008e+00 -4.24524575e-01 1.04304349e+00
1.32727957e+00 3.80764216e-01 -2.85496891e-01 1.09346008e+00
3.06950342e-02 4.30035114e-01 -1.00093508e+00 -1.36751994e-01
-2.93862283e-01 -3.24116312e-02 4.28457528e-01 7.18379259e-01
4.97320086e-01 2.64651865e-01 1.89380243e-01 1.46809369e-01
-2.29120851e-01 -2.02491537e-01 -4.38941777e-01 7.32520819e-01
1.46501374e+00 1.44223177e+00 -5.27936339e-01 -3.54119949e-02
-1.11917175e-01 4.16739881e-01 -3.19098741e-01 7.06447437e-02
-4.57502335e-01 -9.91285086e-01 3.15357924e-01 1.08012311e-01
-1.79479316e-01 -2.00736403e-01 -2.83019274e-01 -2.83578843e-01
4.07066435e-01 -7.26698861e-02 4.93055224e-01 -5.08197069e-01
-9.55986261e-01 3.29872370e-01 9.04239863e-02 -8.80832613e-01
3.27212125e-01 9.71748754e-02 -9.00555015e-01 2.32193083e-01
-1.73669362e+00 -8.41225386e-01 -1.24825716e+00 8.38276327e-01
1.45367086e-01 3.73238057e-01 8.79278302e-01 4.68872607e-01
-8.58747065e-01 1.59616247e-01 2.21971020e-01 -5.85157089e-02
3.10306072e-01 -7.72759438e-01 -3.55663210e-01 1.18373215e+00
-7.78960049e-01 5.13449132e-01 9.61910963e-01 -8.13684404e-01
-1.58584559e+00 -1.09351814e+00 5.76942801e-01 5.22414446e-01
2.22657189e-01 -1.22895665e-01 -7.07293510e-01 3.26511383e-01
1.97209626e-01 2.16085121e-01 5.49013615e-01 -4.09716100e-01
5.25378227e-01 -9.27055418e-01 -2.01659346e+00 -8.79647806e-02
9.22503531e-01 6.66468322e-01 -2.96308964e-01 1.52774394e-01
6.21156037e-01 -1.10185454e-02 -1.13226271e+00 7.30274200e-01
1.57738551e-01 -7.81014562e-01 5.87195218e-01 6.35855198e-02
-5.35974264e-01 -6.85130656e-01 5.68407178e-02 -1.06384993e+00
-1.34597644e-01 -4.76087540e-01 -4.60240021e-02 1.93353415e+00
-8.13318565e-02 -6.53161168e-01 4.58945781e-01 5.14917970e-01
-1.38511598e-01 -3.65483105e-01 -1.09498751e+00 -3.57763141e-01
-1.00397098e+00 -2.46347368e-01 1.05695081e+00 7.33891606e-01
-2.71700531e-01 -3.95759754e-03 -2.08601639e-01 6.75811410e-01
1.18323195e+00 -4.74959612e-01 6.11739874e-01 -1.95338035e+00
8.13480556e-01 3.62679690e-01 -2.37556949e-01 -2.78807610e-01
-4.03699070e-01 -1.55287966e-01 1.62809059e-01 -2.06997347e+00
-1.74646541e-01 -9.87388551e-01 -4.48869169e-01 9.10965383e-01
2.22743917e-02 -3.39307897e-02 -4.83759314e-01 2.72912115e-01
-7.27594495e-01 2.33160958e-01 6.56265676e-01 2.22099841e-01
-2.99496889e-01 1.53754249e-01 -7.14479208e-01 7.22978115e-01
1.23780704e+00 -1.00190997e+00 -5.54960251e-01 -6.20591700e-01
2.77322024e-01 1.72437161e-01 1.47198990e-01 -1.16800952e+00
7.06711948e-01 -3.02913666e-01 5.09698570e-01 -1.01637268e+00
1.54790767e-02 -1.97388935e+00 5.60988963e-01 9.50492680e-01
2.68412709e-01 1.81632355e-01 2.38347463e-02 8.76269758e-01
-4.50705998e-02 6.73110932e-02 5.07430792e-01 7.64928162e-02
-4.55895662e-01 1.56754121e-01 -2.46771544e-01 -6.83657110e-01
1.73241580e+00 -2.16929018e-01 -3.08837414e-01 3.41913700e-02
-3.65690321e-01 8.87569487e-01 3.19258392e-01 1.69343457e-01
6.57534480e-01 -1.09167922e+00 -6.07674241e-01 1.89662710e-01
7.50549957e-02 6.07473850e-01 1.48592085e-01 7.58861780e-01
-2.99496710e-01 -1.01528265e-01 -2.84063905e-01 -2.87233829e-01
-1.36703205e+00 4.16199982e-01 -6.84950426e-02 -6.91264421e-02
-2.58560389e-01 3.12939525e-01 -1.02422559e+00 -1.45866200e-01
5.94955444e-01 -5.32467127e-01 -3.13035995e-01 -7.40524977e-02
4.76526082e-01 9.24751878e-01 2.06224263e-01 -3.20540100e-01
-8.05419207e-01 5.01461744e-01 4.17480737e-01 4.00572658e-01
1.69434130e+00 -8.75702858e-01 -7.03553379e-01 -2.10044980e-02
8.70486140e-01 -2.55191047e-02 -6.29720569e-01 -1.53899208e-01
2.46048123e-01 -1.51563570e-01 8.96996781e-02 -5.23365438e-01
-1.36780632e+00 3.59584466e-02 6.41841769e-01 8.16887259e-01
1.75041020e+00 -3.90289783e-01 9.12139893e-01 3.54417592e-01
7.12216794e-01 -1.14897239e+00 -6.94175899e-01 -3.18402052e-01
1.96824580e-01 -1.20743477e+00 4.65977252e-01 -4.55618918e-01
1.16300337e-01 1.07931268e+00 5.45836210e-01 2.27439106e-02
9.43230391e-01 6.57580972e-01 -2.98957527e-01 -2.52948552e-01
-5.89079320e-01 2.08162934e-01 -8.21709275e-01 1.02269495e+00
-1.53990239e-01 -1.16210684e-01 -7.69406438e-01 7.67575622e-01
1.24418415e-01 8.41919258e-02 6.16580606e-01 1.57287848e+00
-9.90454733e-01 -1.26514685e+00 -6.97162688e-01 5.71333706e-01
-6.72771513e-01 5.86111009e-01 1.63495556e-01 7.59764910e-01
5.65788865e-01 1.87813938e+00 2.47188136e-02 -4.08146709e-01
4.32195365e-01 -4.96953994e-01 -2.38103583e-01 -1.79846704e-01
-3.47669423e-01 -3.21910053e-01 2.44825389e-02 -6.95173502e-01
-1.02789164e+00 -3.98909181e-01 -1.88498807e+00 -5.04210353e-01
-3.13616723e-01 6.10361576e-01 1.38014841e+00 6.92423224e-01
4.50153351e-01 7.72910118e-01 1.13416326e+00 -4.04006690e-01
-2.05201700e-01 -1.01600659e+00 -5.27580917e-01 1.47309020e-01
5.76058686e-01 -4.13380742e-01 -4.05390859e-01 -1.41724914e-01] | [5.930065155029297, 1.7409032583236694] |
9386e334-2508-4deb-b4fc-98c53a457477 | towards-sustainable-deep-learning-for | 2201.09071 | null | https://arxiv.org/abs/2201.09071v1 | https://arxiv.org/pdf/2201.09071v1.pdf | Towards Sustainable Deep Learning for Wireless Fingerprinting Localization | Location based services, already popular with end users, are now inevitably becoming part of new wireless infrastructures and emerging business processes. The increasingly popular Deep Learning (DL) artificial intelligence methods perform very well in wireless fingerprinting localization based on extensive indoor radio measurement data. However, with the increasing complexity these methods become computationally very intensive and energy hungry, both for their training and subsequent operation. Considering only mobile users, estimated to exceed 7.4billion by the end of 2025, and assuming that the networks serving these users will need to perform only one localization per user per hour on average, the machine learning models used for the calculation would need to perform 65*10^12 predictions per year. Add to this equation tens of billions of other connected devices and applications that rely heavily on more frequent location updates, and it becomes apparent that localization will contribute significantly to carbon emissions unless more energy-efficient models are developed and used. This motivated our work on a new DL-based architecture for indoor localization that is more energy efficient compared to related state-of-the-art approaches while showing only marginal performance degradation. A detailed performance evaluation shows that the proposed model producesonly 58 % of the carbon footprint while maintaining 98.7 % of the overall performance compared to state of the art model external to our group. Additionally, we elaborate on a methodology to calculate the complexity of the DL model and thus the CO2 footprint during its training and operation. | ['Carolina Fortuna', 'Marko Meža', 'Mihael Mohorčič', 'Gregor Cerar', 'Blaž Bertalanič', 'Anže Pirnat'] | 2022-01-22 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [-9.65299681e-02 2.09304571e-01 -4.08497542e-01 -2.17406675e-01
-4.45203751e-01 -4.37713265e-01 5.25188923e-01 1.31898791e-01
-6.06415749e-01 1.14596772e+00 -3.01991671e-01 -7.96605051e-01
-3.81549776e-01 -1.16886210e+00 -6.39627457e-01 -6.44940317e-01
-2.82598108e-01 3.26787680e-01 -1.17342971e-01 4.66574758e-01
-1.76916435e-01 7.03090906e-01 -1.24196708e+00 -5.47408998e-01
1.05548298e+00 1.52488494e+00 3.41525257e-01 5.54584086e-01
-8.47347230e-02 4.76053357e-01 -4.56390649e-01 -1.47892669e-01
9.56448838e-02 -2.17949018e-01 -4.84434456e-01 -7.84331143e-01
9.21219662e-02 -3.80579829e-01 -3.72873247e-01 7.07586944e-01
8.87860179e-01 7.31975138e-02 4.71439958e-01 -1.16550684e+00
3.41361053e-02 6.02829337e-01 -1.95728824e-01 -3.08596678e-02
1.94346949e-01 -4.75934327e-01 5.53142250e-01 -5.03954828e-01
1.55839041e-01 3.78309906e-01 1.11699247e+00 2.38442704e-01
-9.41001654e-01 -9.46055710e-01 -1.33891344e-01 4.18563068e-01
-1.44251382e+00 -7.47651517e-01 6.57588243e-01 1.46397039e-01
8.68392706e-01 4.66552317e-01 7.46517539e-01 1.01638746e+00
2.40339905e-01 2.05206186e-01 7.77635276e-01 -4.96485353e-01
6.31575584e-01 3.27053696e-01 -4.51941252e-01 7.52426088e-01
8.42341542e-01 -6.47288486e-02 -1.12173006e-01 -1.35935888e-01
2.34476045e-01 1.94179028e-01 1.12578206e-01 -3.80536586e-01
-1.04979420e+00 4.78335142e-01 6.37075901e-01 6.54706359e-01
-4.67851162e-01 9.05525148e-01 -1.23022936e-01 -2.61241376e-01
2.58095533e-01 2.98673630e-01 -6.32494330e-01 -4.28714722e-01
-1.30282676e+00 -4.35415059e-02 1.00966477e+00 9.27379727e-01
7.68391430e-01 -9.87801403e-02 2.71749020e-01 5.78011692e-01
4.80036885e-01 8.49142611e-01 2.44251683e-01 -1.22857857e+00
1.93891421e-01 3.90936613e-01 4.30130720e-01 -9.60738480e-01
-9.39621449e-01 -1.02731860e+00 -1.10442674e+00 -5.21064460e-01
3.29507142e-01 -7.11239398e-01 -5.33516705e-01 1.61982620e+00
3.45655344e-02 3.47211242e-01 -4.43182349e-01 4.68830951e-03
1.87498033e-01 6.46411717e-01 6.14439920e-02 -2.42062330e-01
9.12974775e-01 -6.82418406e-01 -6.34548724e-01 -1.89677671e-01
5.38850844e-01 -2.86223680e-01 4.62455302e-01 3.12307864e-01
-8.44654918e-01 -4.80590761e-01 -1.30964696e+00 3.79123569e-01
-7.26199865e-01 2.98922043e-02 9.70670819e-01 1.57269228e+00
-1.01904690e+00 6.21818066e-01 -8.47889185e-01 -7.18019783e-01
4.65914220e-01 7.21442401e-01 2.64108807e-01 -9.95988250e-02
-1.11194706e+00 9.48104501e-01 3.55699137e-02 -5.12720570e-02
-6.88377440e-01 -7.10124135e-01 -4.07313645e-01 1.66541368e-01
-3.68301966e-03 -5.92014194e-01 1.31127191e+00 -2.48454452e-01
-1.39818001e+00 9.20719504e-02 -2.90879279e-01 -6.80327833e-01
5.80674827e-01 -3.55667025e-02 -8.90869141e-01 -3.19773525e-01
-8.28798190e-02 4.12081271e-01 3.14944327e-01 -1.20605993e+00
-9.96536016e-01 -7.26351961e-02 8.97243768e-02 -2.34910026e-01
-5.74731052e-01 -5.50799012e-01 -4.17185396e-01 -1.52827263e-01
2.29432628e-01 -1.16467595e+00 -4.42391992e-01 1.53592015e-02
-7.31458664e-02 8.77409354e-02 9.94324803e-01 -6.22003853e-01
1.36473811e+00 -1.64867544e+00 -5.92800796e-01 5.90088964e-01
-4.27200384e-02 2.64982805e-02 3.51365328e-01 2.58504719e-01
4.42701310e-01 3.48897159e-01 -1.32432818e-01 -4.44477826e-01
2.98266917e-01 6.11576326e-02 2.89332390e-01 5.93123019e-01
-7.18870461e-01 6.99129105e-01 -1.06638253e+00 -1.04294196e-01
6.13343358e-01 5.92184484e-01 -3.36353511e-01 -6.94463998e-02
9.41712037e-02 4.65411901e-01 -3.40512395e-01 7.43729115e-01
6.43978953e-01 -1.85400769e-01 5.80395222e-01 -2.01146051e-01
-2.90616393e-01 2.78016448e-01 -9.82380569e-01 1.68150818e+00
-1.26769280e+00 7.06006646e-01 -6.98433891e-02 -1.00477624e+00
4.82796907e-01 2.68565625e-01 9.52059925e-01 -1.20242798e+00
1.25325724e-01 7.10503459e-01 -1.07631817e-01 -3.07579488e-01
1.35921374e-01 3.34357917e-01 -2.49931201e-01 4.21745062e-01
-7.52640516e-02 2.13640824e-01 -3.73266302e-02 -1.99616835e-01
1.37185109e+00 8.41086730e-02 1.84257716e-01 -2.83625066e-01
5.80879152e-01 -5.18798590e-01 4.63398576e-01 9.12216663e-01
-3.59479725e-01 -1.86005577e-01 -1.07158318e-01 -4.51906085e-01
-8.24137270e-01 -9.10532534e-01 -3.40177119e-01 9.02646184e-01
2.31576979e-01 -1.29323110e-01 -9.09247220e-01 -5.90670586e-01
-5.19378781e-02 9.77213621e-01 -5.45236245e-02 -3.46828587e-02
-3.47427130e-01 -1.05345047e+00 8.10155809e-01 2.78247833e-01
1.25779688e+00 -3.37600708e-01 -7.59519875e-01 4.48211104e-01
-2.75672406e-01 -1.17449486e+00 2.36827791e-01 4.57518786e-01
-5.25608420e-01 -7.08068967e-01 -7.33890593e-01 -4.28732336e-01
4.33444977e-01 1.40666932e-01 1.00165188e+00 -8.74290317e-02
-2.87438482e-02 3.30419332e-01 -1.14194863e-01 -6.62908316e-01
1.52307488e-02 5.78155220e-01 4.53094006e-01 -2.16303274e-01
2.31841877e-01 -8.80176842e-01 -1.04869008e+00 2.66056418e-01
-4.27545197e-02 -3.16314623e-02 6.33897901e-01 1.35962725e-01
3.13570946e-01 6.45901561e-01 7.99273908e-01 -5.73635995e-01
1.64621383e-01 -8.62519801e-01 -6.09132349e-01 1.37543321e-01
-1.18064690e+00 -5.80671914e-02 6.51072979e-01 1.37226209e-02
-8.94108713e-01 2.57976890e-01 -2.59051561e-01 4.54978704e-01
-1.30943665e-02 1.95894063e-01 -4.47001934e-01 -6.54598057e-01
3.66437376e-01 3.07799205e-02 -6.15880072e-01 -4.79100168e-01
2.96135664e-01 8.85403216e-01 3.27041090e-01 -3.28591883e-01
9.64374781e-01 4.16971803e-01 3.53376001e-01 -9.78369653e-01
-4.78005648e-01 -2.04597741e-01 -3.06175590e-01 -4.44509298e-01
5.38657725e-01 -8.11727226e-01 -9.15988386e-01 2.37170026e-01
-1.05785334e+00 -2.92251706e-01 2.56492291e-02 3.81633908e-01
-2.29197890e-01 1.28308078e-02 2.47992709e-01 -1.16422248e+00
-3.35451275e-01 -7.92758822e-01 5.06543994e-01 4.01553571e-01
-1.08207196e-01 -1.04793799e+00 -2.10166033e-02 2.72254437e-01
1.13730240e+00 3.73724341e-01 8.56468618e-01 -2.50685304e-01
-6.06652439e-01 -5.62002301e-01 -3.09546053e-01 2.14312419e-01
2.75115371e-01 -5.79832911e-01 -1.24489152e+00 -1.82682246e-01
-2.09863648e-01 5.50052285e-01 3.93384695e-01 4.90973204e-01
1.63718104e+00 -2.78629363e-01 -9.07867312e-01 6.77463174e-01
1.66194892e+00 3.97710353e-01 4.79474038e-01 1.99758396e-01
5.10210097e-01 1.28284946e-01 2.55629033e-01 3.73798579e-01
6.49211407e-01 7.07881987e-01 7.75216520e-01 -1.44637913e-01
-7.19097331e-02 -2.96758503e-01 1.03974588e-01 5.13915062e-01
-3.02186042e-01 -5.99353433e-01 -1.04178643e+00 4.02652532e-01
-1.62207687e+00 -8.82743895e-01 1.87211018e-02 2.50656772e+00
1.71215311e-01 2.96268493e-01 -2.07780510e-01 5.45556962e-01
4.42942709e-01 1.92737848e-01 -6.88720584e-01 -1.98443547e-01
3.75900805e-01 4.56534684e-01 1.40017688e+00 2.57835180e-01
-1.06723499e+00 2.65426457e-01 6.25039244e+00 5.88511288e-01
-1.08258796e+00 5.68943620e-01 5.45613825e-01 -4.72367741e-02
-2.56507009e-01 -3.19735318e-01 -5.02393067e-01 9.27702904e-01
1.80020297e+00 -6.79491758e-02 6.70028031e-01 9.50771689e-01
5.56182981e-01 -4.54625815e-01 -9.67834890e-01 1.21502125e+00
-4.09750611e-01 -1.35192430e+00 -5.43530226e-01 4.90188032e-01
7.87933230e-01 3.52240533e-01 -1.19577155e-01 3.01282227e-01
-2.94707641e-02 -9.42269742e-01 6.77659333e-01 6.59589291e-01
8.36729169e-01 -1.06062436e+00 1.00235534e+00 4.68625218e-01
-1.38261759e+00 -4.52544630e-01 -2.02904299e-01 -1.93347767e-01
1.26600415e-01 1.12232590e+00 -7.60940015e-01 5.96686900e-01
8.42753708e-01 2.27182105e-01 -5.28245986e-01 1.34095573e+00
-9.41834599e-02 6.07056797e-01 -8.33841503e-01 -3.92019182e-01
1.00754619e-01 9.62712839e-02 -1.35796532e-01 1.06349945e+00
1.04130924e+00 -4.55700010e-01 -3.34104151e-01 4.18117464e-01
-4.82610494e-01 -3.74167800e-01 -4.87157494e-01 1.17646158e-01
1.10914874e+00 1.22049201e+00 -5.43975592e-01 4.59697796e-03
-4.41458374e-01 9.77138579e-01 -3.67485464e-01 3.26471388e-01
-1.33914483e+00 -6.45855367e-01 5.58740199e-01 3.35973799e-01
-1.94104221e-02 -5.19146562e-01 -3.58051151e-01 -4.35203999e-01
-1.02387756e-01 -7.95435682e-02 -3.45374584e-01 -3.78876776e-01
-6.25868976e-01 -5.08905239e-02 -3.50972086e-01 -9.31727350e-01
-1.70279250e-01 -4.85946923e-01 -4.12580401e-01 5.89923799e-01
-1.70750952e+00 -9.42947209e-01 -7.05004573e-01 1.79680809e-01
1.31720468e-01 6.53330684e-02 1.25736034e+00 1.08364820e+00
-6.34815514e-01 6.08170569e-01 8.34443688e-01 -3.56598645e-01
2.54533589e-01 -1.16169977e+00 5.14555812e-01 7.58081973e-01
-5.38563766e-02 3.83100122e-01 4.85732079e-01 -1.64841384e-01
-1.60323012e+00 -1.31678319e+00 1.25936258e+00 -2.88908571e-01
3.34483832e-01 -5.39027870e-01 2.37845019e-01 2.93612778e-01
-8.81211646e-03 1.92337662e-01 7.78228104e-01 3.34142298e-02
3.53736550e-01 -7.14170456e-01 -1.57647312e+00 3.55988711e-01
1.31972575e+00 -4.37901616e-01 2.60150611e-01 4.16183859e-01
3.09062481e-01 5.87237440e-02 -7.46215284e-01 4.99104142e-01
1.14223671e+00 -8.85447681e-01 8.63464653e-01 3.63895535e-01
-6.16257608e-01 -3.23641956e-01 -3.83902639e-01 -9.29080069e-01
-2.89447486e-01 -6.95483625e-01 -7.82842159e-01 1.27257323e+00
5.44906199e-01 -6.66278303e-01 1.00608420e+00 5.48373938e-01
1.61237344e-01 -6.71033382e-01 -1.29871631e+00 -9.39727902e-01
-3.59949499e-01 -8.16208899e-01 8.09483349e-01 7.79825747e-01
-4.14645016e-01 -9.28131789e-02 -2.11504713e-01 3.80112648e-01
7.92178750e-01 -5.60652077e-01 4.80836004e-01 -1.57555282e+00
2.84786616e-02 -1.51138797e-01 -3.46671283e-01 -8.70866299e-01
-2.10393503e-01 -6.14438415e-01 -1.20295666e-01 -1.96219051e+00
-2.12794870e-01 -1.01999497e+00 -7.76917756e-01 3.99003804e-01
7.59872139e-01 3.97160202e-01 -3.25676203e-01 -1.65057540e-01
-6.23466730e-01 2.35017225e-01 2.36375690e-01 -3.77712518e-01
-1.34776747e-02 4.56234843e-01 -5.90580285e-01 8.62244368e-01
9.40924466e-01 -4.58046347e-01 -4.78346199e-01 -3.98475230e-01
5.16068757e-01 -5.01514971e-01 2.73799241e-01 -1.89067161e+00
4.04835224e-01 1.76573191e-02 8.21083605e-01 -4.21068937e-01
2.77681708e-01 -1.61443520e+00 7.41300821e-01 9.07330573e-01
3.62958878e-01 -3.94881517e-01 -3.19098718e-02 6.17161036e-01
5.68132222e-01 -4.61437441e-02 6.31275952e-01 1.69633955e-01
-6.81244254e-01 2.58050501e-01 -4.74837899e-01 -8.13469231e-01
1.14124465e+00 -2.72154033e-01 -1.78713381e-01 -5.12287557e-01
-4.55981284e-01 9.33340471e-03 3.79247874e-01 1.60888195e-01
-8.51117671e-02 -1.38042259e+00 2.09815264e-01 -2.25807473e-01
-3.02577376e-01 -3.44717860e-01 6.82746023e-02 7.10427582e-01
-7.20114172e-01 9.42416191e-01 4.93361168e-02 -1.76690623e-01
-5.09050071e-01 3.24731432e-02 8.45421553e-01 -2.59023309e-01
-2.80053671e-02 7.21938908e-01 -7.07989693e-01 -3.79090399e-01
6.42527461e-01 -4.09933388e-01 2.81724244e-01 -1.96897388e-02
1.24910913e-01 1.12147081e+00 4.24544066e-01 -1.52607933e-01
-8.37044239e-01 4.87975925e-01 6.27673924e-01 1.25343710e-01
1.21999061e+00 -3.55059803e-01 1.01730581e-02 1.57891363e-01
1.07280254e+00 2.89508492e-01 -9.38319445e-01 1.74912527e-01
2.35620826e-01 8.26271698e-02 5.45986295e-01 -1.09501708e+00
-9.39864039e-01 5.84652007e-01 1.41869736e+00 4.60867614e-01
1.08294547e+00 -2.40403667e-01 9.04827416e-01 7.48324811e-01
1.17411983e+00 -1.21857345e+00 -6.04629099e-01 7.34157041e-02
1.49127305e-01 -1.02771056e+00 3.62316698e-01 -8.60533044e-02
7.86538780e-01 7.89715111e-01 8.43612924e-02 2.69393146e-01
1.02567315e+00 2.28343576e-01 -5.90588152e-01 2.66576290e-01
8.58182460e-02 1.33027704e-02 -1.77898422e-01 8.40246201e-01
4.12617743e-01 2.94075549e-01 -2.60490328e-01 7.06322372e-01
-1.29098415e-01 3.40562426e-02 -3.54275405e-02 9.38648045e-01
-6.61744416e-01 -1.30449331e+00 -4.99339364e-02 6.58572376e-01
-4.25424069e-01 1.56339213e-01 8.32722709e-02 5.49798369e-01
8.70888650e-01 1.18304491e+00 -1.52824223e-01 -4.82207417e-01
3.16308498e-01 5.16578294e-02 4.10594344e-01 -1.58265859e-01
-6.69850502e-03 -5.82209527e-01 3.40329021e-01 -7.57650852e-01
-4.73890364e-01 -4.23177987e-01 -1.33115840e+00 -6.68734848e-01
-3.65885854e-01 9.52196643e-02 1.59794092e+00 8.93295884e-01
5.04425526e-01 7.42922306e-01 8.28252017e-01 -1.16231930e+00
4.19320613e-02 -7.17004001e-01 -4.68231738e-01 -7.05776870e-01
2.90190965e-01 -6.98557079e-01 -2.87847370e-01 -4.81184721e-01] | [6.393052101135254, 0.9661210179328918] |
68321dfe-46ef-4f18-b913-b1114fa4f819 | the-role-of-context-types-and-dimensionality | 1601.00893 | null | http://arxiv.org/abs/1601.00893v2 | http://arxiv.org/pdf/1601.00893v2.pdf | The Role of Context Types and Dimensionality in Learning Word Embeddings | We provide the first extensive evaluation of how using different types of
context to learn skip-gram word embeddings affects performance on a wide range
of intrinsic and extrinsic NLP tasks. Our results suggest that while intrinsic
tasks tend to exhibit a clear preference to particular types of contexts and
higher dimensionality, more careful tuning is required for finding the optimal
settings for most of the extrinsic tasks that we considered. Furthermore, for
these extrinsic tasks, we find that once the benefit from increasing the
embedding dimensionality is mostly exhausted, simple concatenation of word
embeddings, learned with different context types, can yield further performance
gains. As an additional contribution, we propose a new variant of the skip-gram
model that learns word embeddings from weighted contexts of substitute words. | ['David McClosky', 'Oren Melamud', 'Mohit Bansal', 'Siddharth Patwardhan'] | 2016-01-05 | the-role-of-context-types-and-dimensionality-1 | https://aclanthology.org/N16-1118 | https://aclanthology.org/N16-1118.pdf | naacl-2016-6 | ['learning-word-embeddings'] | ['methodology'] | [-1.65480152e-01 -2.25275517e-01 -5.00236511e-01 -5.33203959e-01
-7.93050349e-01 -6.73320472e-01 7.85463095e-01 2.65294075e-01
-9.40720737e-01 5.95915914e-01 6.80920720e-01 -4.02243376e-01
3.12968977e-02 -6.25139654e-01 -2.89386213e-01 -4.85025436e-01
-4.10916805e-02 1.68406829e-01 8.00523758e-02 -2.73464561e-01
3.32614005e-01 3.03836375e-01 -1.33492982e+00 1.92570295e-02
6.95346951e-01 5.98324835e-01 1.76104710e-01 5.70340037e-01
-3.41590405e-01 -2.24142540e-02 -4.02244031e-01 -5.11283576e-01
3.21409136e-01 -1.52326347e-02 -6.31407022e-01 -4.49736804e-01
4.59051877e-01 -4.49776471e-01 -2.02791080e-01 8.13355803e-01
6.46370173e-01 4.64222372e-01 8.35330069e-01 -8.82170796e-01
-9.29338872e-01 5.39536178e-01 -2.63522178e-01 3.73418093e-01
1.94546089e-01 4.48892087e-01 1.70966434e+00 -1.32240629e+00
3.87962759e-01 1.20019627e+00 6.68907404e-01 4.80065972e-01
-1.47627652e+00 -5.62002957e-01 3.42863053e-01 -1.15999483e-01
-1.21715760e+00 -3.99945289e-01 4.36362237e-01 -2.69471437e-01
1.45715618e+00 -1.30207822e-01 1.65488601e-01 1.32932794e+00
9.83821154e-02 5.37046969e-01 9.75973904e-01 -6.04421437e-01
5.74609861e-02 2.19941095e-01 5.11783004e-01 2.80151784e-01
7.84738541e-01 1.32066995e-01 -4.82036322e-01 -4.89787787e-01
1.73355490e-01 -1.07503667e-01 -1.37355521e-01 -2.54869252e-01
-1.16314006e+00 1.08687615e+00 1.32661656e-01 5.29893458e-01
-4.07096110e-02 4.48909014e-01 4.01587158e-01 2.34268263e-01
5.75206399e-01 1.15013373e+00 -8.45200002e-01 -5.47712266e-01
-5.33173561e-01 2.96205729e-01 7.86868155e-01 7.48043418e-01
9.60793614e-01 -1.73742473e-01 -4.01358992e-01 1.03884053e+00
1.85309663e-01 2.78141916e-01 7.75365174e-01 -6.83748364e-01
5.21850169e-01 4.23185498e-01 2.91636109e-01 -7.64274001e-01
-3.55719239e-01 -3.48158151e-01 6.61642849e-02 -1.17522277e-01
6.93799376e-01 -4.15387928e-01 -9.04603541e-01 2.05958462e+00
5.45483939e-02 5.02675585e-02 6.08381890e-02 8.16305578e-01
4.33670133e-01 4.83184814e-01 4.07933265e-01 2.20273241e-01
1.69009507e+00 -1.00482512e+00 -6.40353918e-01 -7.26631224e-01
1.16336608e+00 -8.54933739e-01 1.78057694e+00 7.97206536e-02
-7.54042447e-01 -3.04530978e-01 -1.01609778e+00 -4.88222927e-01
-7.55766690e-01 1.87559515e-01 8.00719500e-01 8.57892275e-01
-7.92957902e-01 5.01070380e-01 -6.94112778e-01 -5.27527571e-01
5.21843955e-02 2.00433031e-01 -3.64173204e-01 -2.55019456e-01
-1.40935564e+00 1.11799073e+00 3.61316741e-01 -4.48778689e-01
-6.54843211e-01 -8.70162010e-01 -7.81182468e-01 2.22491950e-01
2.64608502e-01 -4.54598516e-01 1.17110932e+00 -5.74065924e-01
-9.75303650e-01 5.88070393e-01 -2.99090117e-01 -1.33613899e-01
-8.01042393e-02 -7.01742947e-01 -3.01076680e-01 -1.62764266e-01
-5.10832258e-02 6.90659821e-01 7.41161108e-01 -9.49756145e-01
-4.80969638e-01 -1.50834158e-01 4.43547606e-01 1.64471939e-01
-1.20496190e+00 5.01311384e-02 -4.53454107e-01 -9.27518010e-01
-4.81422007e-01 -1.02646697e+00 -2.94044882e-01 -2.14275941e-01
8.22393075e-02 -5.64455450e-01 5.60628295e-01 -2.66412497e-01
1.65592396e+00 -2.41795325e+00 5.78709841e-02 -4.74562543e-03
7.49015361e-02 3.47952753e-01 -6.54226065e-01 6.97743118e-01
-6.33946359e-02 4.74269301e-01 -1.02825023e-01 -4.35154259e-01
1.38050869e-01 3.97358984e-01 -2.64700860e-01 3.79820075e-03
3.38870764e-01 9.26568747e-01 -1.11715651e+00 -2.27211207e-01
-2.84308475e-02 4.03598398e-01 -7.21372128e-01 4.29715961e-02
-1.62414625e-01 -2.16810897e-01 -4.88813221e-01 1.71302631e-01
3.65025520e-01 -3.10522676e-01 2.08218947e-01 -1.90380700e-02
1.99224532e-01 7.79247642e-01 -9.81699467e-01 1.54332268e+00
-8.88806880e-01 4.99730259e-01 -1.97735637e-01 -6.88916504e-01
5.46541393e-01 3.07345808e-01 2.09595099e-01 -4.52567428e-01
-5.46470694e-02 3.35273534e-01 2.47388080e-01 -5.85869193e-01
8.80648077e-01 -2.61969745e-01 -2.23864809e-01 8.69898558e-01
2.91843265e-01 -6.61398545e-02 2.70011723e-01 3.30240279e-01
1.15940070e+00 -8.47709402e-02 4.44004744e-01 -3.52733195e-01
1.66322649e-01 -3.81597616e-02 4.53339785e-01 6.60374463e-01
-2.70950586e-01 3.97339016e-01 5.61653674e-01 -2.34238878e-01
-1.11671448e+00 -9.15059865e-01 -3.02955210e-01 1.68254304e+00
-1.48835555e-01 -7.61869848e-01 -3.69655818e-01 -9.58655238e-01
2.73743242e-01 9.52914178e-01 -6.25858665e-01 -2.53163129e-01
-5.56308568e-01 -9.55771148e-01 6.62584841e-01 7.56732523e-01
-2.67205596e-01 -8.10660601e-01 -3.04406941e-01 7.05411509e-02
1.84971586e-01 -1.16795838e+00 -8.39635670e-01 6.37966216e-01
-1.03303170e+00 -8.58841479e-01 -6.51108027e-01 -7.28360534e-01
6.77525461e-01 5.61565042e-01 1.18217051e+00 -7.66216079e-03
-1.13667190e-01 6.53586864e-01 -7.07830548e-01 -1.12188205e-01
-3.12192217e-02 5.84681928e-01 1.89683959e-01 -3.29432935e-01
9.29731369e-01 -5.92173100e-01 -6.61223531e-01 2.31289640e-02
-9.50335562e-01 -5.16242146e-01 6.62425756e-01 1.01598907e+00
8.85354802e-02 -3.47136706e-01 6.27700806e-01 -1.06139660e+00
1.33614933e+00 -6.13508165e-01 -1.24669194e-01 9.20000225e-02
-8.52990985e-01 5.52376807e-01 7.93098092e-01 -6.61182582e-01
-9.14103806e-01 -2.46194288e-01 -8.52004811e-02 -1.79406330e-01
-9.75154899e-03 5.53451478e-01 3.40489745e-02 1.18869863e-01
9.06337559e-01 -2.22146496e-01 -1.92449957e-01 -5.92666388e-01
8.15083385e-01 5.30098379e-01 -8.89272392e-02 -9.75421429e-01
7.66106248e-01 1.28573582e-01 -3.95009696e-01 -7.25332081e-01
-8.68244350e-01 -6.10120296e-01 -4.38459575e-01 4.25219744e-01
9.69890654e-01 -8.74447107e-01 -4.34230380e-02 -9.66790393e-02
-1.03056967e+00 -4.22712773e-01 -2.19079211e-01 7.05710828e-01
-1.72836095e-01 5.18792331e-01 -4.85008776e-01 -6.08202159e-01
-6.16162755e-02 -1.32392216e+00 1.02899337e+00 -1.28980741e-01
-6.82495952e-01 -1.30135477e+00 2.96331406e-01 -5.60525656e-02
5.48460007e-01 -3.18514884e-01 1.25634217e+00 -1.08354318e+00
-2.54556566e-01 -2.05998376e-01 -2.55084872e-01 6.55297697e-01
5.02787352e-01 -1.94871321e-01 -9.70259547e-01 -2.84693927e-01
-2.97195435e-01 -5.13897479e-01 1.10720491e+00 6.19852319e-02
1.03975940e+00 -2.15391949e-01 -3.02727014e-01 4.86813366e-01
1.48969042e+00 -9.67065766e-02 3.10004324e-01 2.64123291e-01
5.75985014e-01 5.17580092e-01 6.86004758e-01 1.66360199e-01
2.58347124e-01 6.55821741e-01 9.19032395e-02 2.51117647e-01
-6.65855557e-02 -1.89246044e-01 4.79407430e-01 8.06555986e-01
1.70170575e-01 -2.20500529e-01 -8.51981819e-01 9.41728354e-01
-1.46204913e+00 -5.37625134e-01 1.19668230e-01 2.33637643e+00
1.01015174e+00 1.76831946e-01 8.09444785e-02 -9.62215886e-02
3.46890271e-01 5.30898571e-01 -3.84752929e-01 -7.50296891e-01
1.17348686e-01 6.97791636e-01 6.85510218e-01 7.02852547e-01
-8.05762827e-01 1.12946475e+00 7.72982073e+00 8.80999744e-01
-8.98853362e-01 2.09314376e-01 2.53205299e-01 -3.46076131e-01
-7.81948268e-01 1.99846849e-01 -8.20650518e-01 4.45513994e-01
8.57766271e-01 -8.94997716e-02 4.32559282e-01 8.74033868e-01
7.56979734e-02 3.71641740e-02 -1.36097264e+00 6.34294748e-01
-1.17461152e-01 -8.67660403e-01 4.11911197e-02 1.29536688e-01
7.04574466e-01 2.42382184e-01 1.60327673e-01 4.59247857e-01
4.44187641e-01 -1.04779053e+00 1.83042720e-01 -5.01656942e-02
7.14563370e-01 -6.26869440e-01 5.69416642e-01 5.11942059e-02
-8.85899663e-01 -1.77089006e-01 -4.73158419e-01 -2.12986663e-01
7.28609040e-02 7.65622377e-01 -7.71265209e-01 2.09729504e-02
2.50413120e-01 3.62941861e-01 -6.65929317e-01 5.27683318e-01
-4.14840341e-01 8.80438387e-01 -2.47268513e-01 -3.74032944e-01
4.68712896e-01 2.89674792e-02 4.37192202e-01 1.63297284e+00
3.77724349e-01 -5.90713583e-02 -1.57624349e-01 4.45951700e-01
-1.93363652e-01 2.36681223e-01 -8.34916413e-01 -5.46187520e-01
7.77462006e-01 1.33670652e+00 -2.40770131e-01 -2.57839322e-01
-7.29285717e-01 5.60065806e-01 6.85279608e-01 6.00125551e-01
-6.51530981e-01 -6.57204330e-01 1.37304163e+00 -3.46992947e-02
5.32645643e-01 -6.92186356e-01 -4.65502650e-01 -1.26770568e+00
6.34938898e-03 -7.08192706e-01 4.74261492e-01 -4.15964305e-01
-1.52378035e+00 1.86908886e-01 1.54540669e-02 -9.13021564e-01
-2.64587015e-01 -1.04388261e+00 -6.77623212e-01 9.30026531e-01
-1.69972241e+00 -6.90404892e-01 6.56670481e-02 1.88254505e-01
4.30765599e-01 -1.44390613e-01 1.00653243e+00 2.31305704e-01
-3.45226914e-01 9.99385893e-01 1.19858220e-01 6.77059442e-02
1.18164957e+00 -1.41529739e+00 5.56246281e-01 7.17125297e-01
3.69891524e-01 1.13519597e+00 8.23495567e-01 -3.95402104e-01
-1.47070491e+00 -8.74479651e-01 1.20068860e+00 -5.45579851e-01
8.16900492e-01 -4.83720332e-01 -7.56627679e-01 6.96326077e-01
2.17819586e-01 8.36642161e-02 1.06688845e+00 6.98459685e-01
-7.05095291e-01 5.46666607e-02 -8.71942103e-01 9.10713196e-01
1.11994100e+00 -8.01863253e-01 -8.10747623e-01 2.12573946e-01
1.16416943e+00 9.36283618e-02 -9.74955440e-01 1.46733403e-01
5.03683507e-01 -4.38335091e-01 1.05515587e+00 -8.56246352e-01
6.50751293e-01 -9.82526597e-03 -4.35358882e-01 -1.62375009e+00
-4.50902849e-01 -3.28686357e-01 -2.27152422e-01 1.16229630e+00
7.69493401e-01 -8.38104129e-01 5.84228933e-01 7.80775309e-01
4.44233790e-02 -8.94906223e-01 -7.32614100e-01 -8.83515477e-01
5.62918127e-01 -6.87956512e-01 6.08708858e-01 1.03111470e+00
1.32984847e-01 5.64846575e-01 -2.86796778e-01 -1.40112609e-01
1.87391549e-01 -2.64013141e-01 7.48744905e-01 -7.86110282e-01
-4.05696481e-01 -4.09135759e-01 -1.30006671e-01 -1.25916755e+00
1.82837024e-01 -1.07030129e+00 -1.08244292e-01 -1.36518300e+00
-9.16292444e-02 -6.15103662e-01 -6.38208210e-01 4.65028733e-01
-7.14768410e-01 1.90753803e-01 1.40442550e-01 -9.22593288e-03
-3.23901713e-01 4.66579288e-01 1.05391717e+00 1.74964622e-01
-1.81748830e-02 -4.55349505e-01 -1.15402400e+00 4.58260000e-01
7.73828089e-01 -5.55155933e-01 -5.40671051e-01 -1.05037069e+00
5.47399223e-01 -5.54719090e-01 -7.74348676e-02 -4.90823001e-01
-1.01816528e-01 -3.90558541e-01 -6.41504861e-03 -1.01024263e-01
4.38109994e-01 -6.49699450e-01 -7.56546736e-01 2.05064982e-01
-5.42151332e-01 2.99825460e-01 1.98064640e-01 6.81616664e-01
-9.92895439e-02 -5.85106134e-01 4.40753311e-01 -8.62694830e-02
-6.98631406e-01 1.29441634e-01 -3.08239490e-01 4.52886879e-01
8.69553447e-01 -1.62967101e-01 -1.48108184e-01 -1.41474500e-01
-5.23119509e-01 1.88309804e-01 5.27712584e-01 7.61292994e-01
2.77238876e-01 -1.48829520e+00 -4.78531927e-01 1.31210014e-02
5.13693035e-01 -4.07324344e-01 -3.49609107e-01 4.92610604e-01
-7.68771544e-02 5.79708338e-01 1.93713471e-01 -1.47929281e-01
-8.88676703e-01 6.58666730e-01 -9.45849717e-02 -4.25812572e-01
-3.83462578e-01 7.63835311e-01 2.42580622e-01 -5.40614963e-01
9.42853093e-02 -5.45378923e-01 -8.67414325e-02 3.19038004e-01
4.16140258e-01 2.35946104e-01 1.14112228e-01 -2.90838957e-01
-3.42889488e-01 3.56293410e-01 -2.62028217e-01 -3.00978184e-01
1.36960483e+00 -5.94913810e-02 2.36849666e-01 4.08198208e-01
1.42672074e+00 2.87670672e-01 -1.12162411e+00 -2.61293530e-01
2.88108706e-01 -7.54840612e-01 2.76177060e-02 -5.01013458e-01
-7.75589883e-01 1.00821471e+00 2.83066124e-01 7.27236867e-02
8.42438698e-01 -1.42839789e-01 9.90483940e-01 5.96204758e-01
3.05458814e-01 -1.24622285e+00 1.92098677e-01 7.87878931e-01
4.26322162e-01 -1.19767046e+00 1.50641590e-01 3.18604633e-02
-6.31891191e-01 1.00063455e+00 6.75162256e-01 -3.60688508e-01
7.08173335e-01 3.02354634e-01 -4.07006070e-02 1.57788687e-03
-1.03246593e+00 -2.47312099e-01 2.13105366e-01 5.45611799e-01
8.98720860e-01 4.14709523e-02 -8.58997643e-01 4.90062594e-01
-1.30218267e-01 -3.40421140e-01 3.20049286e-01 9.00916100e-01
-4.61154193e-01 -1.67099380e+00 -1.06099099e-01 7.21910417e-01
-5.85248172e-01 -4.67922568e-01 -3.37920189e-01 8.06334853e-01
-1.28071383e-02 7.21487164e-01 4.93855998e-02 -2.08270833e-01
1.28892317e-01 5.43825984e-01 4.15663898e-01 -1.05212343e+00
-6.44316912e-01 -4.11327153e-01 4.31897521e-01 -4.02667731e-01
-1.82331204e-01 -5.23282528e-01 -9.26223516e-01 -7.62767568e-02
-4.16104615e-01 1.30695999e-01 5.25246918e-01 9.68437076e-01
4.33934450e-01 2.36106068e-01 3.35472673e-01 -3.93743038e-01
-9.61337864e-01 -1.09853733e+00 -2.80831665e-01 6.19645000e-01
3.09230387e-01 -8.47831309e-01 -6.34670973e-01 -3.70949298e-01] | [10.5645751953125, 8.59926700592041] |
58f18441-f350-4e5c-805c-258797b30aed | rc-qed-evaluating-natural-language | 1910.04601 | null | https://arxiv.org/abs/1910.04601v2 | https://arxiv.org/pdf/1910.04601v2.pdf | R4C: A Benchmark for Evaluating RC Systems to Get the Right Answer for the Right Reason | Recent studies have revealed that reading comprehension (RC) systems learn to exploit annotation artifacts and other biases in current datasets. This prevents the community from reliably measuring the progress of RC systems. To address this issue, we introduce R4C, a new task for evaluating RC systems' internal reasoning. R4C requires giving not only answers but also derivations: explanations that justify predicted answers. We present a reliable, crowdsourced framework for scalably annotating RC datasets with derivations. We create and publicly release the R4C dataset, the first, quality-assured dataset consisting of 4.6k questions, each of which is annotated with 3 reference derivations (i.e. 13.8k derivations). Experiments show that our automatic evaluation metrics using multiple reference derivations are reliable, and that R4C assesses different skills from an existing benchmark. | ['Pontus Stenetorp', 'Naoya Inoue', 'Kentaro Inui'] | 2019-10-10 | r4c-a-benchmark-for-evaluating-rc-systems-to | https://aclanthology.org/2020.acl-main.602 | https://aclanthology.org/2020.acl-main.602.pdf | acl-2020-6 | ['multi-hop-reading-comprehension'] | ['natural-language-processing'] | [ 1.40920699e-01 7.61558473e-01 1.28780693e-01 -5.28051734e-01
-1.24940515e+00 -1.17137444e+00 4.70023632e-01 2.86682606e-01
-2.82400697e-01 7.31950879e-01 6.37652993e-01 -6.87700033e-01
-6.86565191e-02 -4.88918424e-01 -9.82661664e-01 2.81056389e-03
5.56853831e-01 5.52466989e-01 2.95569479e-01 -3.01028669e-01
5.46827078e-01 -1.48692682e-01 -1.44881487e+00 8.02293122e-01
1.56030345e+00 5.81647277e-01 -1.84721842e-01 1.22625804e+00
-1.30492672e-01 1.36510897e+00 -1.02007306e+00 -1.24121499e+00
-1.35361269e-01 -4.27850008e-01 -1.72894394e+00 -5.36293805e-01
6.85225487e-01 -2.92243063e-01 2.72289187e-01 8.47254813e-01
3.35352123e-01 -1.02522194e-01 5.98366320e-01 -1.30053556e+00
-1.45071626e+00 1.13696468e+00 1.57153890e-01 2.88970303e-02
1.02864683e+00 2.36775532e-01 1.26154101e+00 -9.55399811e-01
7.68813550e-01 1.12101424e+00 8.61478150e-01 9.54625607e-01
-1.06620133e+00 -1.89757586e-01 -1.14623621e-01 4.45150286e-01
-9.07259583e-01 -4.07844275e-01 2.59522557e-01 -5.70660651e-01
1.11135137e+00 6.81686878e-01 2.07995728e-01 1.43516767e+00
-1.82412520e-01 7.47518957e-01 1.43655694e+00 -6.04115427e-01
2.46746242e-01 3.51316668e-02 7.48378992e-01 5.19057691e-01
4.23557907e-01 -4.82067764e-01 -7.03056037e-01 -5.74160516e-02
1.37796709e-02 -8.00685823e-01 -6.50866508e-01 9.13916975e-02
-1.15255022e+00 4.82601970e-01 1.78160951e-01 -1.78047031e-01
9.44110081e-02 2.35640213e-01 3.20032030e-01 4.27632719e-01
7.78575912e-02 1.21244586e+00 -7.76320875e-01 -6.48843169e-01
-2.80674726e-01 5.91093659e-01 1.22872221e+00 1.56534755e+00
4.48351204e-01 -4.71414357e-01 -5.46413004e-01 7.25801051e-01
4.84392382e-02 7.46127844e-01 5.99397182e-01 -1.85741377e+00
8.64805698e-01 8.18452179e-01 5.82506239e-01 -6.07262075e-01
-2.87382782e-01 -3.14000815e-01 -1.87901184e-01 -1.88155487e-01
8.56467605e-01 -1.44809455e-01 -2.30744049e-01 1.71316409e+00
8.37010518e-03 -5.35293341e-01 4.24339265e-01 8.31713736e-01
1.33911884e+00 8.42273980e-02 -1.14434334e-02 3.04386556e-01
1.21684229e+00 -1.28849578e+00 -7.23010123e-01 -2.03364775e-01
9.47689235e-01 -7.38240778e-01 1.79609597e+00 5.53364396e-01
-1.32309771e+00 -4.38870281e-01 -9.56339478e-01 -7.33973742e-01
-3.86812299e-01 6.35547936e-02 2.42431909e-01 4.63742375e-01
-1.00780749e+00 3.21891427e-01 -3.05349171e-01 -9.01821852e-02
1.26427710e-01 -3.33844751e-01 -3.18663836e-01 -1.92208230e-01
-1.21349895e+00 1.18493116e+00 1.75949112e-01 -1.12111047e-01
-7.82076299e-01 -7.62819529e-01 -8.17448199e-01 1.29875205e-02
4.15451765e-01 -3.05170149e-01 2.12733841e+00 -8.56688917e-01
-1.32669497e+00 1.00566733e+00 -2.03334868e-01 -3.33717734e-01
8.05983543e-01 -7.60335147e-01 -2.69369960e-01 -3.92794870e-02
3.77721578e-01 4.83290821e-01 2.29697689e-01 -1.11090207e+00
-3.15405965e-01 -5.71562424e-02 6.23868704e-01 3.46652642e-02
-1.10324034e-02 -8.96591842e-02 -2.90337265e-01 -2.97689408e-01
-1.39935508e-01 -7.75222778e-01 2.83872396e-01 -3.72941047e-01
-5.29891431e-01 -7.12128222e-01 -1.12736195e-01 -9.28403735e-01
1.22461653e+00 -1.60518861e+00 2.79019326e-01 -3.37693542e-01
5.62759042e-01 2.23320007e-01 -3.01595718e-01 4.81502265e-01
1.86770126e-01 5.56006432e-01 -1.51550427e-01 -3.22453469e-01
4.38005984e-01 2.10604474e-01 -3.84788334e-01 -1.44180581e-01
3.57874751e-01 1.28008878e+00 -1.20662808e+00 -4.40810531e-01
-4.61020589e-01 -1.58065096e-01 -5.30633271e-01 5.41109502e-01
-6.41931236e-01 1.48879871e-01 -2.86188841e-01 5.40086448e-01
4.67252642e-01 -4.96401042e-01 -4.85723391e-02 3.86744559e-01
6.22031977e-04 7.46186554e-01 -6.04071140e-01 1.82661867e+00
-2.79207706e-01 9.18315291e-01 -4.97855067e-01 -1.41655251e-01
9.08681095e-01 2.73531228e-01 -5.60381711e-01 -7.19900727e-01
-7.31920749e-02 5.48004568e-01 -4.60544310e-04 -1.03810585e+00
6.84061468e-01 4.84191447e-01 -3.27999145e-01 6.49512887e-01
-1.28269829e-02 -2.61865586e-01 3.20455909e-01 4.86861199e-01
1.47226989e+00 1.68334395e-01 1.67365581e-01 -2.75382072e-01
6.97135746e-01 4.52081501e-01 4.67065513e-01 1.02026880e+00
-3.53027970e-01 7.56701112e-01 8.29616785e-01 -2.21514300e-01
-1.02439439e+00 -1.01530719e+00 6.92316443e-02 1.17649329e+00
5.37839942e-02 -7.37693846e-01 -1.26710558e+00 -9.40005124e-01
-2.43067164e-02 1.19927871e+00 -8.58492136e-01 1.30342990e-02
-4.27498668e-01 2.79680677e-02 1.13727021e+00 9.89799500e-01
4.11994278e-01 -9.09974992e-01 -8.41581941e-01 6.51612654e-02
-7.54179478e-01 -1.20406044e+00 -1.36323318e-01 -4.69140895e-03
-4.58475024e-01 -1.50215960e+00 -3.24218988e-01 -5.08473575e-01
6.13656998e-01 1.57412052e-01 1.98311698e+00 6.26749635e-01
2.67509669e-01 6.69845402e-01 -8.85933995e-01 -6.19490266e-01
-6.93186939e-01 3.43550533e-01 -4.22550738e-01 -9.05751884e-01
8.63446414e-01 -8.67296476e-03 -3.79969716e-01 3.57576936e-01
-4.58523512e-01 4.20544922e-01 2.93365568e-01 5.95897615e-01
1.47404566e-01 -7.86355495e-01 7.67631352e-01 -1.36965764e+00
1.07503760e+00 -2.88723201e-01 -4.16315377e-01 8.87710810e-01
-7.84235477e-01 2.50587642e-01 7.00411141e-01 3.71153392e-02
-1.18143141e+00 -3.97603273e-01 3.62400413e-02 3.00991893e-01
-1.82900697e-01 3.71824056e-01 -3.70405875e-02 3.60018432e-01
1.22439945e+00 -1.14779182e-01 -2.45257780e-01 -4.76585388e-01
7.64872909e-01 8.43521118e-01 1.10719597e+00 -1.09518230e+00
6.67864501e-01 -2.46917322e-01 -5.57021379e-01 -1.80392135e-02
-1.49168825e+00 -2.50423074e-01 -5.53565621e-01 -3.25089127e-01
8.74639571e-01 -8.90559077e-01 -8.57262731e-01 2.57090688e-01
-1.60462177e+00 -7.40829229e-01 -1.14232466e-01 7.64273629e-02
-5.17892897e-01 1.76907882e-01 -8.17458034e-01 -5.31779706e-01
-5.05196154e-01 -9.87940490e-01 7.97900558e-01 2.77396351e-01
-8.99997473e-01 -8.44440877e-01 2.78262258e-01 1.07546604e+00
4.62324023e-01 1.18569054e-01 9.63216364e-01 -8.20160270e-01
-3.91780317e-01 -1.17569054e-02 -2.68297315e-01 4.53322381e-01
-4.25239205e-01 2.67160118e-01 -1.14582682e+00 2.48478517e-01
-2.15910673e-01 -9.96144235e-01 3.66966069e-01 -5.36377430e-01
1.47849166e+00 -5.30563414e-01 2.68046051e-01 1.76462814e-01
1.08762586e+00 -3.56107801e-01 7.33024776e-01 5.62096119e-01
5.61471999e-01 9.98426139e-01 6.03764474e-01 -1.55159356e-02
1.05752122e+00 1.85296193e-01 3.20783824e-01 4.65370804e-01
-2.98404217e-01 -5.52882612e-01 3.58309984e-01 1.15454865e+00
-1.83757633e-01 -3.84345204e-01 -1.34943104e+00 6.04600310e-01
-1.82172620e+00 -6.04488611e-01 -9.23890114e-01 1.75503278e+00
1.44088125e+00 3.47945206e-02 -4.04924035e-01 8.18084106e-02
2.93288916e-01 -3.94113243e-01 -6.35251343e-01 -7.73446441e-01
-3.57540756e-01 2.83035219e-01 9.50860232e-02 7.25049376e-01
-4.89551067e-01 7.51547873e-01 6.94145012e+00 3.73775750e-01
-2.44786948e-01 1.01937182e-01 2.83518732e-01 2.27812394e-01
-1.03787386e+00 1.36916608e-01 -5.41114807e-01 3.04576904e-01
1.12740064e+00 -2.17050508e-01 4.59494740e-01 9.05956209e-01
-1.95189580e-01 -8.69062915e-02 -1.47787416e+00 4.82036501e-01
2.31827632e-01 -1.19612193e+00 -1.24854140e-01 -4.45432782e-01
8.78438890e-01 -1.57699868e-01 -1.08467080e-01 6.37507677e-01
7.58269727e-01 -1.30443132e+00 1.14922559e+00 6.62829459e-01
7.94892967e-01 -2.95248002e-01 1.00213361e+00 4.88756061e-01
-4.27682877e-01 1.35934772e-02 -2.67223418e-01 -5.10524988e-01
-2.97986329e-01 3.49748045e-01 -7.37287104e-01 3.04948688e-01
8.61320496e-01 4.37018454e-01 -1.44899523e+00 5.47322869e-01
-1.17990065e+00 8.14572334e-01 3.97439867e-01 -4.94834006e-01
-1.11570656e-01 1.90889806e-01 3.22110541e-02 1.26261592e+00
-4.07668384e-04 3.18540633e-01 -4.26280409e-01 1.11735761e+00
-4.54493731e-01 -3.92942168e-02 -2.76967794e-01 1.23853482e-01
9.59191144e-01 1.18039441e+00 -8.77606571e-02 -5.60718954e-01
-4.69698429e-01 8.44935238e-01 1.03372419e+00 2.25809544e-01
-8.57919395e-01 -5.48063934e-01 3.19945693e-01 -1.55165657e-01
-2.66008854e-01 -4.58875336e-02 -7.49990225e-01 -1.26983082e+00
3.42781872e-01 -1.36841011e+00 2.50262856e-01 -1.47757256e+00
-1.48365796e+00 4.80787843e-01 -2.20981687e-01 -8.10235143e-01
-2.06307650e-01 -7.98762858e-01 -4.86021131e-01 9.00992215e-01
-1.63267553e+00 -6.81015790e-01 -1.11688733e+00 1.72901049e-01
4.88352835e-01 1.67211235e-01 9.79893506e-01 -1.60968900e-01
-5.85444987e-01 8.14154446e-01 -1.88062251e-01 2.80311227e-01
1.09555972e+00 -2.11735082e+00 8.49102378e-01 8.40856135e-01
-7.49192536e-02 9.49902713e-01 8.67633998e-01 -4.84859824e-01
-1.20597410e+00 -8.83816719e-01 1.42871523e+00 -1.60106611e+00
6.73964143e-01 -3.43805075e-01 -1.24529672e+00 9.16356981e-01
5.62792599e-01 -3.38338315e-01 9.16042387e-01 2.24801704e-01
-9.64326382e-01 2.16317192e-01 -8.75712812e-01 6.44843042e-01
1.26923394e+00 -8.87478650e-01 -1.33611035e+00 3.75568032e-01
1.12461722e+00 -8.29014421e-01 -1.01720178e+00 1.47760525e-01
5.36169469e-01 -1.08000267e+00 4.86857206e-01 -8.97266805e-01
1.18139851e+00 -3.09787184e-01 -1.66577086e-01 -1.26016486e+00
-1.39369562e-01 -4.06176537e-01 -1.95577294e-01 1.36730182e+00
9.68885481e-01 -5.49823821e-01 4.33127284e-01 1.40982091e+00
-4.06239361e-01 -6.32906318e-01 -4.98231769e-01 -7.16360092e-01
5.96795082e-01 -6.06488585e-01 1.00051725e+00 8.22557151e-01
3.35786551e-01 3.30309808e-01 2.07753316e-01 9.84386653e-02
4.24499631e-01 1.39359236e-01 1.02594960e+00 -1.16300774e+00
-2.51408279e-01 -2.72376537e-01 2.55331963e-01 -9.66164529e-01
3.41634542e-01 -9.21758771e-01 3.51604044e-01 -1.67978120e+00
3.11095715e-01 -2.82029361e-01 1.86204463e-01 5.23358583e-01
-7.18282998e-01 -1.30408302e-01 -5.02722897e-03 1.69244468e-01
-1.11625242e+00 3.32196742e-01 1.11149704e+00 1.05252311e-01
4.88856107e-01 -5.17946839e-01 -1.05624950e+00 7.46116817e-01
9.35559392e-01 -2.85500228e-01 -4.77336496e-01 -1.14344215e+00
8.94125044e-01 -2.42698714e-01 4.83189881e-01 -8.52487803e-01
3.56879175e-01 -1.03144377e-01 3.01429719e-01 -3.89819652e-01
-2.31706798e-01 -2.07499787e-01 -2.11562261e-01 1.74271956e-01
-1.11961424e+00 4.69861686e-01 -2.04418853e-01 3.33859026e-01
-2.58740429e-02 -7.37289369e-01 1.83815017e-01 -2.82742620e-01
-5.93537509e-01 -6.33068323e-01 -2.18426019e-01 9.30413783e-01
6.42669261e-01 1.22964434e-01 -1.35110867e+00 -4.06969517e-01
-1.88336179e-01 7.21024275e-01 6.44936025e-01 5.02987444e-01
4.15006369e-01 -1.14302909e+00 -9.54911888e-01 -4.63473022e-01
7.36150146e-01 1.70277730e-01 -7.69579113e-02 4.47994232e-01
-8.87507439e-01 5.05069613e-01 -1.95566807e-02 -4.04216766e-01
-1.09902143e+00 3.81879061e-01 2.57535815e-01 -7.04491064e-02
-4.38294977e-01 1.00128388e+00 -5.03274083e-01 -9.07641292e-01
2.51301050e-01 -7.09867477e-01 -3.77376378e-01 -2.77315170e-01
6.77513182e-01 7.84765542e-01 2.39016861e-01 -4.78267446e-02
-1.15988322e-01 3.17777753e-01 3.17424387e-02 -2.31238827e-01
9.12934124e-01 -1.11925505e-01 -3.27517331e-01 5.10188878e-01
7.28724062e-01 2.34196380e-01 -9.83506203e-01 -1.84714094e-01
6.24698043e-01 -3.04038525e-01 -6.92650378e-01 -1.49520612e+00
-1.92701697e-01 7.42853582e-01 -1.89907029e-01 3.03523004e-01
7.81293094e-01 5.00933751e-02 7.46953368e-01 7.89355993e-01
4.63049054e-01 -1.25170100e+00 2.59282291e-01 7.58340538e-01
1.26466560e+00 -1.26554596e+00 -2.42508292e-01 -4.22219664e-01
-8.43315661e-01 1.14030576e+00 1.21083927e+00 2.59169340e-01
-2.94739157e-01 -8.13104361e-02 3.86616409e-01 -2.42471650e-01
-1.26906669e+00 6.84311688e-02 3.71255219e-01 6.84917033e-01
8.39345694e-01 4.04527009e-01 -4.22524482e-01 1.22516859e+00
-9.39739287e-01 -1.62385732e-01 1.12559628e+00 7.55960524e-01
-2.95551449e-01 -9.82964635e-01 -2.64267862e-01 1.70953915e-01
-2.98267692e-01 -4.19419482e-02 -1.17293084e+00 6.91870213e-01
-8.63325298e-02 1.46239841e+00 -4.02745545e-01 -3.18967134e-01
6.85288012e-01 2.52870381e-01 3.99496973e-01 -6.61920786e-01
-8.81606460e-01 -9.82103467e-01 6.11728847e-01 -7.26761758e-01
-3.22979957e-01 -4.42324966e-01 -1.38223159e+00 -4.27708566e-01
-1.46800444e-01 5.05347729e-01 4.18707728e-01 9.33094800e-01
4.10679400e-01 3.64541292e-01 7.24597126e-02 1.73438907e-01
-9.43196297e-01 -1.14548695e+00 1.72422305e-01 8.49587798e-01
1.67121395e-01 -2.74933934e-01 -4.82503891e-01 1.49661750e-01] | [11.05859375, 7.985897541046143] |
850c7a4c-3b20-470e-88b9-61f01c9fd40a | collecting-language-resources-for-the-latvian | null | null | https://aclanthology.org/L16-1202 | https://aclanthology.org/L16-1202.pdf | Collecting Language Resources for the Latvian e-Government Machine Translation Platform | This paper describes corpora collection activity for building large machine translation systems for Latvian e-Government platform. We describe requirements for corpora, selection and assessment of data sources, collection of the public corpora and creation of new corpora from miscellaneous sources. Methodology, tools and assessment methods are also presented along with the results achieved, challenges faced and conclusions made. Several approaches to address the data scarceness are discussed. We summarize the volume of obtained corpora and provide quality metrics of MT systems trained on this data. Resulting MT systems for English-Latvian, Latvian English and Latvian Russian are integrated in the Latvian e-service portal and are freely available on website HUGO.LV. This paper can serve as a guidance for similar activities initiated in other countries, particularly in the context of European Language Resource Coordination action. | ['Raivis Skadi{\\c{n}}{\\v{s}}', 'Andrejs Vasi{\\c{l}}jevs', 'Roberts Rozis'] | 2016-05-01 | collecting-language-resources-for-the-latvian-1 | https://aclanthology.org/L16-1202 | https://aclanthology.org/L16-1202.pdf | lrec-2016-5 | ['miscellaneous'] | ['miscellaneous'] | [-3.06017488e-01 2.23295212e-01 -4.55470890e-01 -3.07738990e-01
-1.41243792e+00 -1.16949022e+00 1.15547645e+00 7.51556233e-02
-6.51118696e-01 1.19177186e+00 7.48662174e-01 -7.62118161e-01
1.77086800e-01 -4.63767558e-01 -1.65237740e-01 -4.86122578e-01
6.80279076e-01 1.18398297e+00 -2.04987898e-01 -5.99279404e-01
3.94283146e-01 5.73263347e-01 -7.10939109e-01 5.33120215e-01
1.05022085e+00 -5.23455963e-02 5.07179916e-01 4.22433227e-01
-2.74432659e-01 3.24089736e-01 -6.02289915e-01 -7.02823758e-01
4.97587740e-01 -4.50675786e-01 -1.42666888e+00 -2.97847074e-02
3.13995272e-01 5.22271022e-02 5.21595664e-02 1.11138010e+00
7.79648721e-01 -1.19863093e-01 5.33669591e-01 -5.74479342e-01
-5.88819087e-01 9.64665413e-01 -1.15767606e-01 6.81827247e-01
4.17363465e-01 -2.14026626e-02 7.90564358e-01 -1.16886926e+00
1.08706331e+00 1.02220118e+00 2.81784654e-01 5.77076852e-01
-9.64010954e-01 -4.71221000e-01 -5.47939897e-01 3.53299417e-02
-1.09745896e+00 -9.07548904e-01 4.21690434e-01 -4.48549062e-01
1.50053000e+00 3.53625417e-01 1.91470534e-01 8.80775332e-01
1.20658681e-01 5.10690331e-01 1.52861965e+00 -1.05551839e+00
-2.06192791e-01 8.23880732e-01 -4.86594215e-02 3.31027806e-01
3.74473423e-01 -3.07619482e-01 -3.52271974e-01 -2.92981446e-01
7.34582305e-01 -8.29956412e-01 2.43563447e-02 7.98289776e-02
-1.39927065e+00 6.75165594e-01 -4.53872174e-01 1.26493394e+00
-5.63518107e-01 -4.96163160e-01 8.52500796e-01 9.96153414e-01
7.77389705e-01 2.86750615e-01 -9.73545492e-01 -5.41094303e-01
-6.35761023e-01 1.17104620e-01 9.10525680e-01 1.32201409e+00
5.74908257e-01 1.29331827e-01 1.67070061e-01 1.22652972e+00
2.57095307e-01 7.80571699e-01 6.32223845e-01 -6.35272086e-01
1.22968829e+00 5.22130966e-01 3.43198538e-01 -3.59706134e-01
-1.73137918e-01 -2.10533187e-01 -2.95864403e-01 -2.98419118e-01
4.00381833e-01 -3.93374205e-01 -4.48447257e-01 1.09953177e+00
3.04314017e-01 -8.85551155e-01 6.10027015e-01 6.17852747e-01
9.93669152e-01 5.06413102e-01 4.79267128e-02 -7.25121200e-01
1.48589218e+00 -8.84956658e-01 -1.01355243e+00 3.18615794e-01
9.30077076e-01 -1.55991447e+00 9.30125952e-01 1.50075093e-01
-1.64144385e+00 -2.92038709e-01 -1.91737458e-01 -1.90980896e-01
-5.31563818e-01 4.42291826e-01 4.78443563e-01 9.09856021e-01
-1.44215631e+00 1.70167401e-01 -7.98187792e-01 -1.17311668e+00
-7.60060400e-02 5.71923673e-01 -5.44922888e-01 3.62099051e-01
-1.03752375e+00 1.45662391e+00 4.21838522e-01 -3.14529613e-02
-4.23205376e-01 -5.29293045e-02 -6.61161423e-01 -5.13425469e-01
-2.39588857e-01 -1.94521174e-01 1.52707529e+00 -9.48755026e-01
-1.36554313e+00 1.44500792e+00 -1.96321964e-01 -2.57054448e-01
7.10936546e-01 4.41743806e-02 -8.68474364e-01 -8.42678547e-02
4.81660783e-01 2.53894955e-01 -2.03752726e-01 -9.22483146e-01
-1.00508356e+00 -4.39164400e-01 -2.17021614e-01 5.13124287e-01
-1.46718130e-01 1.04828596e+00 -2.28693843e-01 -6.50301993e-01
9.89963338e-02 -5.97012877e-01 -2.27029875e-01 -1.05090475e+00
2.32417926e-01 -4.34190989e-01 3.57494712e-01 -1.45306599e+00
1.26886928e+00 -1.31062210e+00 -1.37859583e-01 -5.87689281e-02
-4.75467741e-01 3.81546021e-01 -1.96396753e-01 1.31067336e+00
5.26465550e-02 3.72692019e-01 2.52752244e-01 1.08892038e-01
-7.68046975e-02 2.65921295e-01 -7.04392865e-02 5.97785294e-01
3.62028889e-02 7.56253779e-01 -1.00883138e+00 -7.38852322e-01
2.94852167e-01 4.24108952e-02 1.79219186e-01 -3.30971628e-01
2.13889226e-01 7.81975091e-01 -7.22713590e-01 9.10601079e-01
5.59629142e-01 6.51867807e-01 4.56672311e-01 9.62537453e-02
-8.37450027e-01 1.13151443e+00 -7.25649714e-01 1.65134883e+00
-8.15967441e-01 5.05392551e-01 2.28386208e-01 -6.93191051e-01
1.10458839e+00 1.19316339e+00 4.02267277e-01 -6.84359908e-01
2.51269698e-01 1.03086877e+00 8.80099833e-02 -7.70848036e-01
8.69849801e-01 -1.24886550e-01 -6.73096403e-02 5.83256006e-01
4.32618558e-01 -1.92791820e-01 6.55431092e-01 -2.89534062e-01
5.23945451e-01 3.63520831e-01 5.54899395e-01 -7.74241149e-01
8.96220267e-01 6.79037690e-01 5.05384445e-01 1.27876461e-01
-3.71323407e-01 -2.27908883e-02 1.22490781e-03 -6.47816896e-01
-1.62547529e+00 -4.81133759e-01 -5.39890409e-01 8.82477999e-01
-6.58760428e-01 -3.17926288e-01 -7.76926458e-01 -7.96451688e-01
-8.84053528e-01 6.66610122e-01 -1.62584022e-01 8.46256673e-01
-9.71841276e-01 -5.49316883e-01 6.25155807e-01 -1.18295833e-01
4.60892290e-01 -1.17834342e+00 -1.82007656e-01 5.35536408e-01
-8.84879947e-01 -1.04199588e+00 -2.35051528e-01 -8.91296566e-02
-1.27908993e+00 -7.68221021e-01 -7.17980742e-01 -1.23808277e+00
2.25786552e-01 1.64583296e-01 1.24283743e+00 -2.75242060e-01
4.37736064e-01 1.22489057e-01 -5.72862923e-01 -5.08123100e-01
-1.10656321e+00 5.25658548e-01 1.37311250e-01 -8.10183227e-01
1.21641529e+00 -5.72481453e-01 -1.75016858e-02 5.39523482e-01
-4.18704808e-01 -1.08825028e-01 4.83755350e-01 6.66005194e-01
3.75343144e-01 -4.34871435e-01 7.75917768e-01 -9.44178224e-01
7.33673394e-01 -3.70793730e-01 -5.99054515e-01 4.65580672e-01
-5.07021070e-01 -4.56752181e-01 2.37396285e-01 4.63063717e-02
-1.30968583e+00 -1.64404377e-01 -3.30946594e-01 2.79328078e-01
-3.50601047e-01 4.57279116e-01 -3.69285345e-01 4.54705693e-02
8.99091423e-01 1.83117539e-01 -4.07437623e-01 -8.90534997e-01
1.28732905e-01 1.39842296e+00 2.97988802e-01 -7.90519416e-01
5.31519473e-01 2.08712034e-02 -7.20040381e-01 -7.35337734e-01
3.78318354e-02 -1.01614964e+00 -1.32446992e+00 -2.33676583e-01
5.15343785e-01 -1.03570938e+00 1.83862075e-01 2.44746376e-02
-1.26580751e+00 -2.17582792e-01 -1.85287520e-01 7.75958061e-01
-6.46497607e-01 9.28706527e-02 -6.34023666e-01 -8.70136499e-01
-7.99381733e-01 -9.99343872e-01 8.96207690e-01 -1.40648618e-01
-4.06717539e-01 -1.45620656e+00 5.43069303e-01 9.50282812e-01
2.18599394e-01 -1.40898466e-01 6.90107167e-01 -9.29718137e-01
8.66587907e-02 -1.92152392e-02 2.35705197e-01 1.08823463e-01
3.34373891e-01 -1.87806450e-02 -6.32055759e-01 -2.82114089e-01
2.38262638e-01 -1.97032511e-01 1.61154345e-01 -4.10542153e-02
-1.08851396e-01 -8.45081568e-01 -8.77765343e-02 -2.46908352e-01
1.68183219e+00 4.12670821e-01 6.20329320e-01 7.87358820e-01
1.07319690e-01 1.03092182e+00 8.91568482e-01 1.46247879e-01
3.82313341e-01 7.87513912e-01 -4.69524056e-01 5.34364358e-02
-3.05909336e-01 5.24756266e-03 6.80117130e-01 1.80410802e+00
-7.34916508e-01 2.94400435e-02 -1.35001516e+00 1.10743892e+00
-1.79633176e+00 -1.08571732e+00 -5.89755177e-01 2.13713169e+00
9.73796427e-01 -4.43983018e-01 4.85173047e-01 -1.86613590e-01
8.65572453e-01 -2.43839011e-01 2.65093207e-01 -9.42804754e-01
-4.28474605e-01 3.52717996e-01 5.99464953e-01 8.22989404e-01
-6.99614406e-01 1.61620045e+00 6.73889685e+00 7.79651403e-01
-8.79747033e-01 8.96385372e-01 3.58441353e-01 5.08154631e-02
-4.25194979e-01 1.16508596e-01 -7.88464665e-01 5.63753769e-02
1.38530099e+00 -7.18057454e-01 2.62787730e-01 4.99770761e-01
9.81000602e-01 1.90125585e-01 -6.30775750e-01 5.77915549e-01
-1.34142503e-01 -1.29924834e+00 5.65620773e-02 5.54592073e-01
1.02336395e+00 8.88128102e-01 -4.46583390e-01 2.65775114e-01
6.29953921e-01 -4.45341259e-01 6.60266638e-01 -2.14669928e-01
9.01574910e-01 -6.69123292e-01 1.06297076e+00 2.75135398e-01
-8.69625390e-01 3.81650358e-01 -8.74052048e-01 -1.36928484e-01
2.09900916e-01 2.04020441e-01 -1.10764837e+00 1.09656608e+00
4.70392436e-01 4.21363562e-01 -3.21007013e-01 7.29081511e-01
-2.15219468e-01 9.23935831e-01 6.74984455e-02 -2.00618692e-02
4.82181370e-01 -8.41948628e-01 7.23998487e-01 1.58437049e+00
3.50790799e-01 4.45979321e-03 7.10886270e-02 2.21248716e-01
-5.70227802e-02 1.23724747e+00 -8.34537745e-01 -1.35861605e-01
4.86263186e-01 1.35327458e+00 -5.78478575e-01 -3.87246281e-01
-6.16565406e-01 7.83942521e-01 2.78544456e-01 2.10427016e-01
-1.85885400e-01 -3.44624110e-02 2.84592539e-01 -6.40490130e-02
-3.77464354e-01 -2.16680855e-01 -4.82275456e-01 -1.28890049e+00
1.67201445e-01 -1.44213045e+00 5.44022799e-01 -3.45107466e-01
-7.60999858e-01 9.26811218e-01 4.49244352e-03 -1.24909127e+00
-5.47282875e-01 -3.46508771e-01 -4.54757184e-01 1.28504705e+00
-1.01898396e+00 -1.49521518e+00 5.26858032e-01 3.81843716e-01
1.19167066e+00 -8.94030809e-01 1.18930757e+00 5.41831136e-01
-1.83293447e-01 1.71878621e-01 4.79811251e-01 9.82907042e-02
6.82283580e-01 -1.01036811e+00 6.04427099e-01 7.13656247e-01
3.41190100e-01 5.31539738e-01 7.44380951e-01 -7.27732658e-01
-1.06687021e+00 -8.45932722e-01 2.20453167e+00 -5.89425325e-01
1.09945977e+00 -2.34604940e-01 -3.26714516e-01 7.11121500e-01
9.91710186e-01 -1.02683711e+00 8.55798542e-01 2.63553441e-01
1.59389287e-01 3.19440484e-01 -1.41318572e+00 5.05018950e-01
9.01140809e-01 -5.06463885e-01 -1.03728044e+00 9.57765877e-01
1.52256787e-01 -1.33775920e-01 -1.01243484e+00 -1.60838246e-01
3.11478943e-01 -4.27423686e-01 2.34247372e-01 -8.22495222e-01
3.46071243e-01 -9.05922987e-03 -2.16411039e-01 -1.22977126e+00
-2.68594950e-01 -1.09368432e+00 7.72775590e-01 1.60066760e+00
9.25183058e-01 -8.83563101e-01 4.58875507e-01 2.97564685e-01
-2.95153916e-01 -1.58306703e-01 -1.19020045e+00 -8.51552486e-01
7.08359957e-01 -2.14947894e-01 4.93523240e-01 1.34850490e+00
5.72133601e-01 5.03044009e-01 -2.07470313e-01 -4.91148867e-02
1.93477109e-01 -1.95086882e-01 8.85410011e-01 -7.85980225e-01
-1.87563326e-03 -5.14213562e-01 -2.26861522e-01 -3.76322746e-01
7.76677430e-02 -1.08654153e+00 -4.38440710e-01 -1.95902371e+00
2.64099181e-01 -4.19978857e-01 9.88681540e-02 4.55564111e-01
2.57741243e-01 3.75072747e-01 -5.82560226e-02 7.78380990e-01
1.52708218e-01 -1.73306316e-02 1.25852907e+00 3.15725766e-02
-2.85945088e-01 1.07555501e-01 -6.79776371e-01 5.09910285e-01
1.28933692e+00 -5.04571021e-01 -8.58437866e-02 -8.98063779e-01
7.03161731e-02 3.82718481e-02 -4.93696868e-01 -5.29839694e-01
-3.54564935e-02 -4.78584290e-01 -2.10271195e-01 -8.74222338e-01
-1.12678401e-01 -7.42882252e-01 4.24184412e-01 3.11312586e-01
-1.31729797e-01 8.09513152e-01 2.12174937e-01 -4.16306466e-01
-5.17296672e-01 -6.63094938e-01 4.91778910e-01 -5.59538484e-01
-3.74599814e-01 -2.10779682e-02 -6.25046790e-01 -1.95085064e-01
7.18152165e-01 -3.59034568e-01 -1.49430469e-01 -2.48441994e-01
-7.17033744e-01 1.43941954e-01 6.47334933e-01 3.76521558e-01
-2.64580011e-01 -1.35161853e+00 -1.40524590e+00 -1.69723198e-01
1.04716934e-01 -6.98849499e-01 -1.62118271e-01 9.93783772e-01
-8.22416663e-01 9.72009659e-01 -6.25616550e-01 -1.15481056e-01
-1.58061683e+00 1.98243573e-01 1.93633094e-01 -5.72652996e-01
-3.34609836e-01 6.71145692e-02 -1.80755153e-01 -9.72269475e-01
-2.91011155e-01 1.25973731e-01 -6.38962805e-01 1.69066153e-02
3.26392502e-01 4.37965959e-01 4.18053716e-01 -1.35556269e+00
-2.03624085e-01 3.37081879e-01 -1.84200965e-02 -8.41357470e-01
1.34826803e+00 -6.72644317e-01 -4.55099672e-01 6.09774649e-01
8.34975004e-01 5.90445995e-01 -4.62761745e-02 -2.06637695e-01
3.48844498e-01 -3.32042575e-01 -1.55008629e-01 -1.10018611e+00
-7.55379200e-01 6.44154310e-01 4.41226035e-01 -9.48598906e-02
9.11572397e-01 2.51089875e-02 3.61024648e-01 4.57533121e-01
8.47341061e-01 -1.49578083e+00 -8.55597258e-01 8.13714802e-01
1.06467187e+00 -1.03619552e+00 -2.72466838e-01 -4.69818711e-01
-6.03436291e-01 1.16105628e+00 7.98627064e-02 3.04367810e-01
2.43917897e-01 8.30902308e-02 8.26232553e-01 -1.02266781e-01
-7.66271412e-01 -3.01602274e-01 -4.01778296e-02 8.89516175e-01
1.26725090e+00 1.36771530e-01 -1.85043705e+00 9.08833817e-02
-5.24682701e-01 5.40571138e-02 7.15296924e-01 9.05064404e-01
-2.73000360e-01 -2.01652837e+00 -4.79093075e-01 1.03748962e-01
-1.18806672e+00 -4.99547839e-01 -8.98369312e-01 9.97122884e-01
9.66954082e-02 1.22906756e+00 -2.98256665e-01 6.96852570e-03
3.56477737e-01 2.85236269e-01 5.53300679e-01 -7.61233985e-01
-1.31706011e+00 7.25370109e-01 1.14074969e+00 1.74492270e-01
-7.15907097e-01 -1.04983139e+00 -6.36294305e-01 -4.93725181e-01
-4.57458943e-01 1.03138924e+00 1.15882301e+00 6.19399726e-01
6.85552433e-02 -4.03923810e-01 5.96222401e-01 -4.62277055e-01
-4.28560287e-01 -1.58043814e+00 -2.33412921e-01 1.52539641e-01
-5.04286408e-01 1.36842176e-01 8.43068436e-02 3.86762410e-01] | [11.197053909301758, 10.412286758422852] |
058bb359-510f-40c3-81b6-ea4647e79b11 | learning-to-represent-programs-with-1 | 2012.04188 | null | https://arxiv.org/abs/2012.04188v3 | https://arxiv.org/pdf/2012.04188v3.pdf | Learning to Represent Programs with Heterogeneous Graphs | Program source code contains complex structure information, which can be represented in structured data forms like trees or graphs. To acquire the structural information in source code, most existing researches use abstract syntax trees (AST). A group of works add additional edges to ASTs to convert source code into graphs and use graph neural networks to learn representations for program graphs. Although these works provide additional control or data flow information to ASTs for downstream tasks, they neglect an important aspect of structure information in AST itself: the different types of nodes and edges. In ASTs, different nodes contain different kinds of information like variables or control flow, and the relation between a node and all its children can also be different. To address the information of node and edge types, we bring the idea of heterogeneous graphs to learning on source code and present a new formula of building heterogeneous program graphs from ASTs with additional type information for nodes and edges. We use the ASDL grammar of programming language to define the node and edge types of program graphs. Then we use heterogeneous graph neural networks to learn on these graphs. We evaluate our approach on two tasks: code comment generation and method naming. Both tasks require reasoning on the semantics of complete code snippets. Experiment results show that our approach outperforms baseline models, including homogeneous graph-based models, showing that leveraging the type information of nodes and edges in program graphs can help in learning program semantics. | ['Huangzhao Zhang', 'Kechi Zhang', 'Zhi Jin', 'Ge Li', 'Wenhan Wang'] | 2020-12-08 | learning-to-represent-programs-with | null | null | null | ['code-comment-generation', 'comment-generation'] | ['computer-code', 'natural-language-processing'] | [-1.32803634e-01 3.04058462e-01 -7.13910758e-01 -5.38149893e-01
1.64235234e-02 -8.08500826e-01 1.00823119e-01 7.43301272e-01
2.89519131e-01 9.97382253e-02 2.32325971e-01 -8.52296591e-01
3.64794075e-01 -1.35120857e+00 -9.21225548e-01 -1.53672195e-03
-2.84491092e-01 -5.17388321e-02 5.90776563e-01 -3.59926701e-01
4.45171237e-01 8.51775408e-02 -1.51257539e+00 6.48729503e-01
1.08399808e+00 6.49312437e-01 2.32994273e-01 5.07883012e-01
-1.17121053e+00 1.31769252e+00 -6.13908172e-01 -4.50976312e-01
-5.96836582e-02 -3.22609425e-01 -9.14787412e-01 -3.84971559e-01
3.43380928e-01 -1.60968587e-01 -9.76329222e-02 1.34529305e+00
8.34281370e-02 -2.18471900e-01 4.34836656e-01 -1.53309321e+00
-1.02301455e+00 1.40213501e+00 -6.44747019e-01 8.61019716e-02
5.71167827e-01 7.41062537e-02 1.39405811e+00 -4.40729320e-01
5.40615678e-01 1.31783605e+00 8.76469314e-01 4.29793000e-01
-9.81508732e-01 -4.40627903e-01 3.99578899e-01 2.57767320e-01
-8.50018024e-01 6.44171387e-02 1.00584912e+00 -8.76702309e-01
1.08289158e+00 1.90996915e-01 4.71504956e-01 7.07364023e-01
3.41519088e-01 7.72765875e-01 5.40861607e-01 -2.48688608e-01
1.77814111e-01 -8.04844275e-02 1.07385182e+00 1.20927918e+00
3.46797705e-01 -1.42785445e-01 3.63657847e-02 -4.00089651e-01
3.97703588e-01 1.49619877e-01 -3.00667793e-01 -5.73760271e-01
-9.75391865e-01 1.03418398e+00 9.57323730e-01 2.97359765e-01
2.40816176e-01 3.90775532e-01 8.15708995e-01 4.77238297e-01
-7.79056130e-03 3.49525392e-01 -5.40717721e-01 4.39219624e-02
-3.62648159e-01 -5.85586503e-02 1.06969595e+00 1.44711733e+00
1.37915063e+00 3.44925880e-01 -2.45786935e-01 7.48709142e-01
5.04687965e-01 3.03719372e-01 6.73282444e-01 -4.42305297e-01
1.00503004e+00 1.55532408e+00 -4.56590682e-01 -1.33216453e+00
-4.02089030e-01 -5.83983660e-02 -6.17141068e-01 6.51087686e-02
2.57635266e-01 -3.10270414e-02 -8.92783463e-01 1.49038327e+00
7.61674345e-02 -1.12812631e-01 -1.25321180e-01 3.61438066e-01
1.44058096e+00 6.19140863e-01 -4.32649739e-02 3.52330178e-01
1.05638039e+00 -1.13635170e+00 -5.44247806e-01 -3.78664374e-01
1.40054882e+00 -1.73666954e-01 1.18391848e+00 -1.72346577e-01
-7.14463890e-01 -6.25583708e-01 -9.36458528e-01 -2.53738254e-01
-6.94055259e-01 2.24476540e-03 9.65687692e-01 7.28733659e-01
-1.18918896e+00 4.62161273e-01 -7.11412132e-01 -5.29112257e-02
2.68551499e-01 3.56836021e-01 -2.46527150e-01 6.76951185e-02
-9.37267244e-01 1.69015273e-01 8.34447742e-01 -5.25123887e-02
-7.76050270e-01 -8.78861904e-01 -1.69095993e+00 3.49190772e-01
6.38129532e-01 -6.05775952e-01 1.24940681e+00 -1.23380661e+00
-1.03284240e+00 6.51982188e-01 -7.89712220e-02 -2.18800366e-01
-1.36735260e-01 2.31408328e-01 -2.91104436e-01 -3.49221021e-01
3.79747413e-02 9.06329043e-03 6.29391968e-01 -1.17713392e+00
-5.88500917e-01 -3.41683924e-01 6.98515892e-01 -3.42013061e-01
-3.59704822e-01 3.27474684e-01 -2.76028275e-01 -4.71441776e-01
-6.13993183e-02 -8.19621980e-01 -1.59318656e-01 -1.91949606e-01
-7.54792511e-01 -5.75737119e-01 9.58281934e-01 -6.91228330e-01
1.69603348e+00 -2.17111039e+00 2.19665870e-01 2.15843543e-01
8.61052096e-01 1.08977556e-01 -1.75399691e-01 4.37110215e-01
-1.74761549e-01 6.23481631e-01 -3.98458779e-01 2.65849739e-01
2.61360377e-01 2.82055885e-01 -2.13060617e-01 -2.21534371e-02
2.45020390e-02 1.18710494e+00 -1.05417597e+00 -4.42352206e-01
-1.13327898e-01 -4.90178959e-03 -9.10601020e-01 4.81942624e-01
-6.87468112e-01 -9.53994840e-02 -8.01699281e-01 4.92649227e-01
5.80437541e-01 -4.41559434e-01 2.26413086e-01 -1.03139244e-01
5.69042340e-02 6.19245052e-01 -1.11126590e+00 1.66279554e+00
-8.25520456e-01 4.32382464e-01 -8.71057510e-02 -1.05885684e+00
1.09437430e+00 6.07373798e-03 4.06028517e-02 -3.02853763e-01
-1.40988022e-01 6.09157942e-02 1.85788870e-01 -8.34978580e-01
3.92225504e-01 4.26698804e-01 -4.44543809e-01 6.04602218e-01
-8.47572386e-02 -1.95632860e-01 3.93636614e-01 5.49730957e-01
1.37214553e+00 2.78961986e-01 2.80656636e-01 -2.75184304e-01
7.35677600e-01 -2.36520059e-02 5.76469004e-01 7.64329851e-01
2.65216440e-01 1.86810851e-01 1.27527702e+00 -6.81215644e-01
-7.07229137e-01 -8.13718081e-01 3.09605509e-01 1.40730536e+00
-1.67619213e-02 -1.17848718e+00 -6.83028698e-01 -1.23240387e+00
2.88621970e-02 5.62235415e-01 -8.81520092e-01 -3.93016607e-01
-9.51401055e-01 -5.19469261e-01 5.28601587e-01 8.50611389e-01
3.74287188e-01 -1.17860198e+00 -3.55786562e-01 -3.22285704e-02
4.63796854e-02 -7.45342612e-01 -6.57182992e-01 1.85519248e-01
-9.00664926e-01 -1.43258214e+00 -1.05899811e-01 -1.05022848e+00
8.69563222e-01 -6.71435297e-02 1.50720561e+00 7.90240347e-01
1.51947930e-01 1.72297716e-01 -4.79282171e-01 -2.70568490e-01
-9.01918411e-01 3.75481635e-01 -9.15408075e-01 -2.51594692e-01
2.05734849e-01 -5.73341250e-01 8.69616270e-02 7.47846514e-02
-1.06446564e+00 9.92167220e-02 2.06976995e-01 6.53037906e-01
1.02356546e-01 8.99797529e-02 -9.01469681e-03 -1.60899985e+00
5.38085878e-01 -7.86532342e-01 -8.06925118e-01 2.94018209e-01
-4.51622695e-01 7.46800482e-01 1.20039487e+00 -1.64888486e-01
-1.03924179e+00 -2.08516538e-01 -6.23729266e-02 -1.80465549e-01
-8.35103393e-02 1.21617723e+00 -4.16461021e-01 -1.44486306e-02
7.42804885e-01 6.98228255e-02 -2.70882994e-01 -4.28739756e-01
3.03662091e-01 4.00736511e-01 1.65157989e-01 -1.15825486e+00
7.58863986e-01 -6.96981773e-02 1.56743266e-02 -3.18800658e-01
-4.96248096e-01 -2.09796041e-01 -7.46910930e-01 4.14435655e-01
6.26689374e-01 -4.74040598e-01 -3.59215736e-01 5.77674329e-01
-1.37789536e+00 -6.33394241e-01 1.33716222e-03 -1.66882426e-01
-2.12567374e-01 6.22787178e-01 -7.80765474e-01 -1.88612461e-01
8.12897552e-03 -1.58620441e+00 1.08407998e+00 3.32908891e-02
3.11811380e-02 -1.48407948e+00 3.85775440e-03 -2.39223644e-01
4.20545310e-01 2.72122920e-01 1.78912759e+00 -6.60542548e-01
-8.05747569e-01 5.65907592e-03 -3.21632475e-01 9.82317179e-02
3.97933900e-01 2.87734777e-01 -4.02697057e-01 -8.55406895e-02
-2.34000325e-01 -1.26017421e-01 7.17722178e-01 4.84817363e-02
1.52233291e+00 -6.55072033e-01 -5.61195135e-01 1.05115438e+00
1.64725149e+00 1.60027757e-01 6.80419385e-01 2.04884589e-01
1.45663989e+00 6.12221062e-01 -2.23823741e-01 2.00693518e-01
7.99960196e-01 5.57580590e-01 7.60346234e-01 1.24254659e-01
-1.79683119e-01 -5.09316802e-01 6.37115598e-01 9.23215866e-01
7.76151270e-02 -1.95872992e-01 -1.29700553e+00 4.06332821e-01
-1.65720963e+00 -7.08244681e-01 -9.02299881e-01 2.05531955e+00
1.00322974e+00 -1.44508123e-01 1.26560405e-01 -1.32877603e-01
8.82724583e-01 3.26533347e-01 -3.23484272e-01 -6.23830140e-01
2.38387525e-01 1.68884039e-01 3.55456561e-01 4.54494148e-01
-6.49647236e-01 9.00598288e-01 5.61589909e+00 4.09002274e-01
-1.14531267e+00 -8.65858644e-02 1.00408234e-01 7.70225763e-01
-9.30907428e-01 4.56919014e-01 -8.23010862e-01 5.52913129e-01
1.00526214e+00 -5.06560981e-01 4.46745098e-01 1.16092694e+00
-3.69943172e-01 2.78994113e-01 -1.51965785e+00 6.75081253e-01
7.85637796e-02 -1.32594788e+00 2.20813721e-01 -3.38275641e-01
6.41539872e-01 1.53681338e-01 -3.80864382e-01 7.31825471e-01
8.83498847e-01 -9.47872341e-01 6.10363185e-01 1.38816491e-01
5.75962305e-01 -3.13498050e-01 6.71612859e-01 1.50118083e-01
-1.73950863e+00 -2.92126775e-01 -3.96282941e-01 -9.04711038e-02
-4.51916397e-01 4.63355601e-01 -6.49050951e-01 7.75183797e-01
5.27903616e-01 1.19637918e+00 -1.15346038e+00 1.00344718e+00
-7.16135979e-01 5.37763298e-01 1.57512411e-01 -2.71441996e-01
6.53216392e-02 -1.63938403e-01 1.42439216e-01 1.33203340e+00
2.60817617e-01 -3.81192207e-01 4.13102776e-01 1.51245749e+00
-2.87919402e-01 1.64956957e-01 -9.12066936e-01 -1.19622134e-01
2.80829102e-01 1.15422595e+00 -6.53083503e-01 -4.89145726e-01
-9.18688059e-01 3.30707937e-01 5.81584454e-01 5.19741356e-01
-8.06716502e-01 -7.99730659e-01 1.77895859e-01 1.37189806e-01
2.74871230e-01 -2.29924560e-01 -2.46798366e-01 -1.40666986e+00
3.20762485e-01 -9.59011018e-01 7.26566911e-01 -8.05985272e-01
-9.56693530e-01 6.14490211e-01 1.95861489e-01 -8.50316644e-01
-2.23441750e-01 -6.78635478e-01 -1.01621318e+00 7.86413729e-01
-1.33368647e+00 -1.10808468e+00 -5.44764221e-01 5.71092784e-01
3.41955662e-01 -1.29760444e-01 6.38370037e-01 6.36583641e-02
-6.16289258e-01 7.43647099e-01 -3.56745005e-01 8.56912732e-01
1.38945073e-01 -1.60078859e+00 8.44003260e-01 9.03151155e-01
1.76936224e-01 1.08168304e+00 2.73901999e-01 -8.00328255e-01
-1.83055997e+00 -1.18914211e+00 5.82354069e-01 -5.96920848e-01
9.90914226e-01 -5.25800049e-01 -1.28359151e+00 1.31333840e+00
1.77654028e-01 3.37624609e-01 4.47469562e-01 3.93685281e-01
-1.02493668e+00 1.81680143e-01 -6.32188618e-01 5.02544820e-01
1.20361996e+00 -7.36689925e-01 -7.96746612e-01 1.59540325e-01
1.05707169e+00 -5.78052223e-01 -7.93076217e-01 3.74777794e-01
1.00525938e-01 -1.10851121e+00 6.64193988e-01 -9.08310890e-01
8.22548866e-01 -4.86303359e-01 3.90678830e-02 -1.54481947e+00
-1.85266569e-01 -3.97852123e-01 -2.82837659e-01 1.27321827e+00
6.17778480e-01 -7.67530680e-01 7.08348155e-01 5.14847279e-01
-7.61813939e-01 -4.14763331e-01 -1.04629695e-01 -5.97067952e-01
2.66822129e-01 -4.17911857e-01 1.15634739e+00 1.16363049e+00
1.68032229e-01 4.35139447e-01 2.80884594e-01 1.48800462e-01
1.99944183e-01 6.25615180e-01 9.56274986e-01 -1.51284385e+00
-4.97994870e-01 -6.69848561e-01 -5.61901212e-01 -9.19950902e-01
7.33358026e-01 -1.60284257e+00 -2.14439705e-01 -1.74386835e+00
1.33047268e-01 -7.00810730e-01 1.07391933e-02 9.81869757e-01
-3.97687346e-01 -6.51954353e-01 -8.35088417e-02 -1.52548514e-02
-3.77331197e-01 2.53399104e-01 1.01276469e+00 -7.55879223e-01
-1.39337406e-01 -1.28623456e-01 -6.19832635e-01 7.43951499e-01
6.48639202e-01 -6.81437612e-01 -6.99270427e-01 -8.50080907e-01
7.78298795e-01 3.04531574e-01 2.83298880e-01 -7.50013947e-01
1.94868296e-01 -4.13955390e-01 -3.59897077e-01 -1.83557928e-01
-4.55588371e-01 -8.67803812e-01 1.45328656e-01 4.55434978e-01
-5.17620265e-01 2.89088070e-01 2.67562479e-01 4.12071913e-01
-3.81904721e-01 -7.76180327e-01 3.39572072e-01 -5.18537164e-01
-1.05123734e+00 5.10310233e-01 -7.63889402e-02 3.74424636e-01
8.13886881e-01 -6.72776401e-02 -7.73215950e-01 1.57649945e-02
-4.39914554e-01 1.00879394e-01 8.26315224e-01 9.44169402e-01
3.74878049e-01 -1.19761217e+00 -4.08209801e-01 5.76277673e-01
5.28627455e-01 1.33209705e-01 -8.72033313e-02 5.38958907e-01
-6.78937852e-01 -4.09071259e-02 -2.47267932e-01 -6.11303329e-01
-1.15202820e+00 9.98435676e-01 4.87826496e-01 -2.88548768e-01
-5.42727530e-01 5.65833747e-01 5.80025554e-01 -8.49618614e-01
-1.22359451e-02 -1.00571263e+00 -3.24002445e-01 -2.89611757e-01
4.16050762e-01 -4.12383825e-02 5.99114336e-02 -3.64439964e-01
-1.07121676e-01 6.26111090e-01 -1.09423511e-01 6.98315501e-01
1.16957331e+00 4.35005218e-01 -8.88828695e-01 6.31114125e-01
1.40810490e+00 3.62043262e-01 -7.46821582e-01 -1.82729423e-01
4.22918677e-01 -4.33897704e-01 -2.83203542e-01 -4.74230438e-01
-1.40589702e+00 1.13285029e+00 -1.49636403e-01 6.09605372e-01
7.64072716e-01 1.71882823e-01 5.82500815e-01 5.23725808e-01
6.90921962e-01 -5.02200186e-01 1.41348839e-01 8.74346316e-01
6.90553606e-01 -1.12520993e+00 -3.72067273e-01 -7.40113437e-01
-8.50228071e-02 1.53260493e+00 9.50758517e-01 1.61884073e-02
5.62552035e-01 6.56268537e-01 -1.74068362e-01 -4.68389273e-01
-5.62783241e-01 6.71440288e-02 4.06508356e-01 8.39400291e-01
8.46177518e-01 -1.15633212e-01 -1.69931471e-01 6.34615660e-01
-1.05959743e-01 -1.32280484e-01 9.37082946e-01 1.01414537e+00
-4.08815354e-01 -1.58412683e+00 -1.41477048e-01 8.18415344e-01
-2.69938469e-01 -4.04122084e-01 -3.56568992e-01 8.05723369e-01
3.65057541e-03 6.67546868e-01 -1.15241617e-01 -4.08999741e-01
2.99041986e-01 1.71279479e-02 2.76876897e-01 -1.44135511e+00
-9.37233269e-01 -7.93835282e-01 -3.69221345e-02 -5.07651150e-01
-1.81243494e-01 -2.39826098e-01 -1.66793370e+00 -1.90420344e-01
-1.99530825e-01 3.66609782e-01 3.86082292e-01 6.66865706e-01
4.36412334e-01 8.77265036e-01 4.50082123e-01 -2.76596934e-01
-3.83203626e-01 -6.28252566e-01 -4.09235388e-01 7.44612813e-01
3.85621428e-01 -3.66664678e-01 -4.56623554e-01 2.98240334e-01] | [7.5148725509643555, 7.873141288757324] |
d2fac47b-96db-4597-a548-d229223096cc | optimal-transport-for-unsupervised-1 | 2212.09631 | null | https://arxiv.org/abs/2212.09631v2 | https://arxiv.org/pdf/2212.09631v2.pdf | Optimal Transport for Unsupervised Hallucination Detection in Neural Machine Translation | Neural machine translation (NMT) has become the de-facto standard in real-world machine translation applications. However, NMT models can unpredictably produce severely pathological translations, known as hallucinations, that seriously undermine user trust. It becomes thus crucial to implement effective preventive strategies to guarantee their proper functioning. In this paper, we address the problem of hallucination detection in NMT by following a simple intuition: as hallucinations are detached from the source content, they exhibit encoder-decoder attention patterns that are statistically different from those of good quality translations. We frame this problem with an optimal transport formulation and propose a fully unsupervised, plug-in detector that can be used with any attention-based NMT model. Experimental results show that our detector not only outperforms all previous model-based detectors, but is also competitive with detectors that employ large models trained on millions of samples. | ['André F. T. Martins', 'Pablo Piantanida', 'Pierre Colombo', 'Nuno M. Guerreiro'] | 2022-12-19 | null | null | null | null | ['nmt'] | ['computer-code'] | [ 1.60823837e-01 3.80610108e-01 -2.40171120e-01 5.28759547e-02
-1.01686561e+00 -4.38003719e-01 7.36939728e-01 2.95569357e-02
-7.59780779e-02 7.97653019e-01 3.84910703e-01 -4.88049775e-01
6.60528779e-01 -2.18716323e-01 -9.40932214e-01 -2.93958515e-01
3.14276427e-01 8.81054401e-01 -2.37215802e-01 -3.81800771e-01
1.62704214e-01 7.56605528e-03 -9.56364274e-01 4.91469741e-01
1.21213114e+00 5.61293721e-01 1.43584162e-01 2.80624092e-01
2.83089072e-01 8.78431380e-01 -9.35176849e-01 -9.31690574e-01
4.43009250e-02 -8.61730337e-01 -7.13522613e-01 2.63706982e-01
1.52852803e-01 -3.40764225e-01 -2.44353786e-01 1.47165167e+00
6.58738792e-01 -3.91298681e-01 6.07532024e-01 -1.15315425e+00
-1.47762477e+00 6.50923312e-01 -3.74685615e-01 1.44236177e-01
5.81271827e-01 3.39863956e-01 9.53961790e-01 -1.45288050e+00
7.17115939e-01 1.34502840e+00 6.15910053e-01 7.98508465e-01
-1.21372950e+00 -5.49395442e-01 -2.55870521e-01 3.92686903e-01
-1.24304128e+00 -8.10508668e-01 4.07498419e-01 -2.17992991e-01
1.29867649e+00 2.48448551e-01 6.79467678e-01 2.10706306e+00
6.08413160e-01 1.14645624e+00 1.23583221e+00 -4.12684143e-01
2.83420444e-01 6.01554930e-01 -4.91449207e-01 6.13637209e-01
3.27532530e-01 -1.27978042e-01 -8.20041776e-01 -2.65783012e-01
4.72897023e-01 -3.75927299e-01 -4.46226269e-01 -9.67481360e-02
-1.32810569e+00 8.73553574e-01 1.38321653e-01 2.18844324e-01
-5.25449634e-01 -1.09307244e-01 2.82258838e-01 7.19138563e-01
6.62461281e-01 7.68602908e-01 -1.94546133e-01 -3.34046781e-01
-9.60764050e-01 7.83916488e-02 6.36046946e-01 1.17787373e+00
2.90763289e-01 -4.89960099e-03 -2.40265489e-01 8.89653027e-01
2.35823557e-01 5.65736890e-01 9.89044547e-01 -5.68662345e-01
4.93496776e-01 4.12820488e-01 3.14228714e-01 -8.06772768e-01
1.28562510e-01 -7.68692374e-01 -7.02451169e-01 -1.75283194e-01
-2.32862327e-02 2.13127155e-02 -7.44853616e-01 1.70304370e+00
-1.39565840e-01 -3.34777944e-02 7.37477913e-02 9.59607005e-01
2.00417757e-01 5.66032350e-01 -3.08274508e-01 -4.41060156e-01
1.15520656e+00 -1.14962125e+00 -1.05153275e+00 -6.46204948e-01
6.55662179e-01 -1.01499784e+00 1.20173216e+00 4.01328832e-01
-1.33976102e+00 -7.41146505e-02 -1.28335810e+00 -4.13563624e-02
-1.62678152e-01 -1.64399639e-01 2.15200633e-01 6.20779753e-01
-1.07856965e+00 6.66617811e-01 -7.27995992e-01 -7.49643683e-01
5.49145401e-01 2.18858883e-01 -3.27002138e-01 -1.92214120e-02
-1.40583193e+00 1.68513143e+00 -7.54847303e-02 1.75635904e-01
-1.22624934e+00 2.79767774e-02 -5.73791087e-01 -3.68697681e-02
5.35981692e-02 -7.70497561e-01 1.60479426e+00 -1.25161278e+00
-1.27617466e+00 9.37058926e-01 -4.81720895e-01 -5.74627340e-01
7.31936812e-01 -3.60373557e-01 -5.93933940e-01 -1.42314464e-01
3.05704117e-01 3.24885845e-01 1.23662210e+00 -1.12161112e+00
9.52274799e-02 -2.52438366e-01 -4.51400757e-01 3.20255220e-01
-6.39844239e-01 2.59737283e-01 -1.78517565e-01 -4.87400770e-01
-4.89664339e-02 -8.51780236e-01 -1.65874839e-01 -1.01166755e-01
-9.38130319e-01 4.48448174e-02 6.20999575e-01 -7.58784533e-01
1.21181440e+00 -1.89909530e+00 2.95800805e-01 -2.16273174e-01
3.35719347e-01 4.96732414e-01 -1.56026170e-01 7.37516582e-01
1.86987087e-01 2.11705163e-01 -2.12423354e-01 -6.89214110e-01
1.51004577e-02 3.42450500e-01 -5.01560509e-01 6.32793069e-01
4.68832850e-01 1.24986327e+00 -1.08964586e+00 -3.23810101e-01
-2.08427414e-01 2.05510736e-01 -3.95342141e-01 2.79086560e-01
-3.24768960e-01 4.10632968e-01 -3.30294400e-01 7.43536174e-01
4.30455595e-01 -5.72763085e-01 2.21006781e-01 4.74855036e-01
2.43777946e-01 6.85687959e-01 -1.24040313e-01 1.58657563e+00
-1.69517800e-01 8.75648379e-01 -1.46944150e-01 -8.47591817e-01
7.45779097e-01 7.79053509e-01 -1.02749266e-01 -7.54308522e-01
3.11581433e-01 7.10153699e-01 9.14556161e-02 -6.65624678e-01
6.15154624e-01 -4.56362814e-01 -3.58056463e-02 5.59706092e-01
9.20712277e-02 2.14238644e-01 -2.12221086e-01 1.81018665e-01
1.32123005e+00 -2.54853994e-01 3.91501158e-01 -2.49467138e-02
1.62813459e-02 1.18851997e-01 5.19342244e-01 8.75054479e-01
-3.83654535e-01 7.14821696e-01 5.00905752e-01 -7.35412613e-02
-1.44439030e+00 -8.93732846e-01 2.78961733e-02 4.76197749e-01
1.12962108e-02 -3.67639035e-01 -9.62481618e-01 -5.66929281e-01
-1.06484272e-01 9.35551643e-01 -5.11810601e-01 -6.89717531e-01
-2.70715415e-01 -8.01176786e-01 7.17910707e-01 2.57275164e-01
2.60490298e-01 -1.09422505e+00 -3.43919903e-01 3.65122616e-01
-6.47631228e-01 -1.09041953e+00 -5.49600184e-01 1.53064191e-01
-7.77900040e-01 -4.57019061e-01 -8.35081875e-01 -6.29539430e-01
6.73917651e-01 3.02487046e-01 1.15041757e+00 2.40316078e-01
1.17027901e-01 -1.88531339e-01 -2.98499554e-01 -4.10125047e-01
-1.17423558e+00 4.00204249e-02 4.66247141e-01 1.10610463e-01
6.79978013e-01 -6.99630976e-01 -2.47232914e-01 1.82855234e-01
-6.79005861e-01 2.37949133e-01 9.43982601e-01 1.02786219e+00
5.55256419e-02 -3.59421551e-01 7.96421409e-01 -7.04595804e-01
1.39957356e+00 -7.64593363e-01 -2.45508775e-02 3.03369999e-01
-8.90428007e-01 5.90783395e-02 8.20092499e-01 -5.96483886e-01
-6.43695593e-01 -4.99315560e-01 -6.67663887e-02 -6.82032764e-01
1.60902753e-01 5.52922428e-01 -2.25161403e-01 2.89412826e-01
8.89884591e-01 7.44711757e-01 6.32639453e-02 -5.34934163e-01
2.56643832e-01 1.14468348e+00 3.95797133e-01 -1.71669751e-01
8.28853428e-01 1.56844646e-01 -6.94598913e-01 -5.68027318e-01
-8.28395069e-01 -1.86004326e-01 -3.27128947e-01 -1.20833255e-01
6.07042849e-01 -1.02772212e+00 -1.54071571e-02 2.95357674e-01
-1.61778724e+00 6.01752661e-03 1.34276584e-01 2.61036098e-01
-7.52281845e-01 4.63864386e-01 -8.85218859e-01 -9.37098145e-01
-3.38453829e-01 -1.21193182e+00 9.82282937e-01 -2.32531846e-01
-6.96336865e-01 -7.67785370e-01 2.34207541e-01 5.06146252e-01
4.75765377e-01 -2.16767758e-01 9.59028542e-01 -8.07393670e-01
-5.76183498e-01 -2.34351650e-01 -3.48891243e-02 5.58111906e-01
-1.17422327e-01 -4.43126023e-01 -1.06527829e+00 -3.08326781e-01
4.15724844e-01 -6.01590216e-01 6.91658556e-01 -2.07249627e-01
4.62703854e-01 -7.65561581e-01 -2.50744134e-01 1.97670072e-01
1.08211732e+00 -9.59756300e-02 7.26053774e-01 2.08277479e-01
3.72058719e-01 3.36424917e-01 3.57165545e-01 3.61392200e-01
9.88957062e-02 6.73922658e-01 4.20683503e-01 1.27868960e-02
-8.24515522e-02 -8.10912251e-01 1.04365349e+00 1.14647925e+00
8.92404988e-02 -6.01696432e-01 -7.27554321e-01 7.37410963e-01
-2.12378001e+00 -1.03539848e+00 -2.15168849e-01 2.14958906e+00
9.27438676e-01 2.55920768e-01 -2.21983239e-01 -1.28323570e-01
6.42919540e-01 -1.02308117e-01 -6.42813325e-01 -8.22204947e-01
-4.07061696e-01 2.26050895e-02 3.25337917e-01 2.43129596e-01
-5.93204618e-01 1.09092319e+00 7.06931973e+00 5.06636024e-01
-9.63472843e-01 7.33683228e-01 3.82596374e-01 -1.06007077e-01
-4.46251303e-01 -1.78307310e-01 -3.01706374e-01 6.14653528e-01
1.35448492e+00 -3.51487547e-01 5.64432144e-01 7.64788330e-01
4.90616351e-01 3.16306949e-01 -1.26548386e+00 8.89297307e-01
3.89232963e-01 -1.02221286e+00 4.35806252e-02 3.26811641e-01
8.33186805e-01 3.55141312e-01 5.09617150e-01 3.35662514e-01
2.09641948e-01 -1.16899037e+00 1.10082543e+00 1.06451139e-01
6.00215673e-01 -4.57124770e-01 6.67823613e-01 7.63084471e-01
-1.99273333e-01 2.26748902e-02 -6.90900683e-01 -1.63593680e-01
2.37939775e-01 5.96864164e-01 -1.23968256e+00 1.52404189e-01
2.26661369e-01 8.02968800e-01 -4.50284660e-01 9.53201234e-01
-5.87896407e-01 6.58439875e-01 1.09509133e-01 -1.60390466e-01
2.19622970e-01 1.21967517e-01 8.58540952e-01 9.91365314e-01
6.14043355e-01 -6.17396772e-01 -2.37399355e-01 1.13450670e+00
-5.27248859e-01 8.19220394e-02 -9.14718449e-01 -5.58084667e-01
4.30872262e-01 6.90932333e-01 -1.44285262e-01 -5.76301455e-01
-4.82971221e-01 1.83762109e+00 6.00260317e-01 2.76639670e-01
-8.93184364e-01 3.76475245e-01 8.87197137e-01 5.78418933e-02
7.79542476e-02 -1.27912194e-01 -2.98819512e-01 -1.72871232e+00
4.19799209e-01 -1.31856060e+00 -3.76904964e-01 -9.82222915e-01
-1.45260537e+00 8.53880465e-01 -6.24081552e-01 -1.44733143e+00
-3.90924722e-01 -4.25323486e-01 -4.89952207e-01 8.90766680e-01
-1.28576779e+00 -1.02510726e+00 3.62914175e-01 3.67355943e-01
7.48920619e-01 -1.05955839e-01 1.24584961e+00 2.96437234e-01
-7.62473822e-01 6.92191839e-01 2.74126887e-01 7.11694639e-03
7.79568195e-01 -9.60895061e-01 7.52667665e-01 9.87417042e-01
2.81196505e-01 8.19428682e-01 1.11742222e+00 -7.94523954e-01
-1.50789499e+00 -1.09849894e+00 1.68369138e+00 -9.20490444e-01
7.27670193e-01 -6.25178218e-01 -9.95635986e-01 7.98833609e-01
5.83073974e-01 -4.32048857e-01 5.78767121e-01 1.06171571e-01
-5.63413322e-01 5.89858174e-01 -9.31008279e-01 8.80268991e-01
1.03292644e+00 -8.38701725e-01 -8.58383954e-01 8.51157606e-01
5.47660887e-01 -1.27650976e-01 -4.49598342e-01 -1.05559833e-01
4.44129139e-01 -1.06582868e+00 5.65273583e-01 -8.86376679e-01
7.49205351e-01 1.25976885e-02 7.78887942e-02 -1.64436519e+00
-3.08421314e-01 -1.11855793e+00 -4.17748004e-01 6.44227087e-01
8.38042021e-01 -5.92323542e-01 5.17844915e-01 3.55532944e-01
-2.79596031e-01 -6.92722917e-01 -1.10447264e+00 -1.06111145e+00
2.18262270e-01 -3.28965008e-01 3.79453093e-01 1.15047216e+00
7.10689843e-01 8.82095098e-01 -1.06946766e+00 9.72338393e-02
6.96774900e-01 -2.36800864e-01 3.84434402e-01 -8.43928337e-01
-5.32443225e-01 -5.29204071e-01 -2.78045565e-01 -1.21594894e+00
2.43628472e-02 -1.02921283e+00 2.04594195e-01 -1.42642105e+00
5.99949956e-01 8.30627829e-02 -2.23983929e-01 4.53874171e-01
-1.20155759e-01 2.82581776e-01 -7.72660300e-02 6.53423607e-01
-5.19264877e-01 7.51828134e-01 1.31785631e+00 -2.03266636e-01
3.47907841e-01 -1.56311527e-01 -9.74983156e-01 3.86438221e-01
8.19769502e-01 -7.63132811e-01 -2.84068704e-01 -6.20767534e-01
4.28179383e-01 9.52312946e-02 3.02709401e-01 -7.16590941e-01
1.23068273e-01 6.88217059e-02 2.15242550e-01 -2.48404607e-01
5.27652621e-01 -4.52687919e-01 -3.47618349e-02 4.84294444e-01
-5.11061966e-01 3.11974227e-01 -2.79871851e-01 5.51683068e-01
-6.71219528e-02 -1.96209997e-01 7.88265646e-01 -4.15997624e-01
-3.67086357e-03 7.22484663e-02 -9.50233698e-01 -1.22528546e-01
7.23351240e-01 1.16726719e-01 -3.36611062e-01 -7.93405056e-01
-4.86415923e-01 2.85255611e-02 7.77902961e-01 7.11945534e-01
9.65284467e-01 -1.41711068e+00 -9.60708678e-01 1.59311906e-01
4.61246967e-01 -7.65407622e-01 -2.94958264e-01 1.25849104e+00
-4.18743379e-02 5.60476303e-01 -1.56708866e-01 -2.87570298e-01
-9.88111079e-01 8.15327883e-01 3.18560064e-01 -1.11407958e-01
-5.60563087e-01 7.83451259e-01 -1.34395406e-01 -1.99758470e-01
-9.88886971e-03 -1.24935158e-01 3.30957711e-01 -3.74409735e-01
6.82256401e-01 -1.22740850e-01 2.45034501e-01 -8.34726930e-01
-2.34144896e-01 -4.57541645e-01 -3.86940122e-01 -3.50104064e-01
9.06045556e-01 -1.35085166e-01 -1.85441256e-01 4.63016152e-01
1.04071951e+00 -7.97877386e-02 -6.25447392e-01 -2.37270594e-01
1.39821917e-01 -5.86462796e-01 -1.86901823e-01 -9.86659527e-01
-3.61267865e-01 1.00404286e+00 1.60313979e-01 2.65940011e-01
7.08032787e-01 1.08368516e-01 1.19191194e+00 3.29473943e-01
5.57873368e-01 -1.03053713e+00 2.60846853e-01 3.64015043e-01
1.02001655e+00 -1.26163435e+00 -5.26039958e-01 -4.07434478e-02
-6.70447946e-01 7.05211937e-01 4.21453714e-01 9.51241404e-02
-3.76981348e-02 2.71959063e-02 2.37470537e-01 -1.51546463e-01
-1.42566144e+00 1.48121968e-01 7.60327950e-02 5.24070263e-01
4.79748011e-01 2.17513695e-01 -3.19947094e-01 4.58539307e-01
-7.74202272e-02 -1.65349573e-01 7.40184486e-01 7.08679676e-01
-6.92899823e-01 -1.39629769e+00 -4.73865390e-01 3.23848546e-01
-4.62787747e-01 -5.29145718e-01 -9.37020838e-01 2.64021635e-01
-1.04049578e-01 1.12256455e+00 -3.28858733e-01 -7.75295913e-01
1.61518782e-01 4.92733479e-01 4.80892062e-01 -8.01230848e-01
-5.82451165e-01 5.69492579e-02 3.83606553e-01 -4.14523065e-01
6.54819887e-03 -7.36324728e-01 -7.48894453e-01 -5.05116999e-01
-3.79350752e-01 1.43741533e-01 5.05740047e-01 1.10346937e+00
5.67748427e-01 5.88747524e-02 6.28862202e-01 -3.75171840e-01
-1.18668771e+00 -1.46056104e+00 -2.35143825e-01 4.49801147e-01
4.41141665e-01 -4.29634660e-01 -3.38828832e-01 -2.22023353e-01] | [11.741525650024414, 9.968893051147461] |
e1eb3103-e0f0-4933-bc70-cac73cc3f073 | video-based-pedestrian-attribute-recognition | 1901.05742 | null | https://arxiv.org/abs/1901.05742v2 | https://arxiv.org/pdf/1901.05742v2.pdf | A Temporal Attentive Approach for Video-Based Pedestrian Attribute Recognition | In this paper, we first tackle the problem of pedestrian attribute recognition by video-based approach. The challenge mainly lies in spatial and temporal modeling and how to integrating them for effective and dynamic pedestrian representation. To solve this problem, a novel multi-task model based on the conventional neural network and temporal attention strategy is proposed. Since publicly available dataset is rare, two new large-scale video datasets with expanded attribute definition are presented, on which the effectiveness of both video-based pedestrian attribute recognition methods and the proposed new network architecture is well demonstrated. The two datasets are published on http://irip.buaa.edu.cn/mars_duke_attributes/index.html. | ['Yunhong Wang', 'Zhiyuan Chen', 'Annan Li'] | 2019-01-17 | null | null | null | null | ['pedestrian-attribute-recognition'] | ['computer-vision'] | [ 4.72153053e-02 -8.27490151e-01 -3.75278652e-01 -5.60930848e-01
-7.47028351e-01 -6.89421296e-02 6.26561284e-01 2.14546826e-02
-5.75239897e-01 1.17332017e+00 1.76012993e-01 1.71821490e-01
-1.20487839e-01 -7.02627718e-01 -4.45680231e-01 -8.31136346e-01
-1.70623973e-01 3.97412270e-01 1.14228003e-01 -2.54188865e-01
4.80897240e-02 2.15924025e-01 -1.82675004e+00 2.48211339e-01
8.58730972e-01 1.28669930e+00 2.06141338e-01 7.39589036e-01
1.03026398e-01 8.15373778e-01 -3.64819676e-01 -5.62989593e-01
6.68193251e-02 2.34636832e-02 -7.46251702e-01 -2.54611466e-02
6.26109540e-01 -3.43678981e-01 -8.73302698e-01 8.13616812e-01
7.81175315e-01 7.30139911e-01 6.83427513e-01 -1.44726622e+00
-9.47066486e-01 9.79032516e-02 -2.88842440e-01 4.68265653e-01
6.94495738e-01 1.84129775e-01 8.07962060e-01 -9.17648315e-01
3.87430400e-01 1.14057732e+00 5.42231679e-01 6.06730461e-01
-7.99747348e-01 -5.76601982e-01 4.82988626e-01 1.12072206e+00
-1.82020807e+00 -4.18166637e-01 8.54675770e-01 -5.94072342e-01
8.39998960e-01 4.10406232e-01 1.09779942e+00 1.64937496e+00
-1.11729868e-01 1.16157186e+00 8.81790042e-01 -1.27127185e-01
-2.16104031e-01 -7.95831345e-03 3.83836269e-01 3.46019357e-01
1.27785027e-01 2.01098621e-01 -5.35056114e-01 8.45999345e-02
5.61347663e-01 1.76636666e-01 -3.84300426e-02 -1.67514175e-01
-1.17055702e+00 6.33471370e-01 2.28129417e-01 9.75799412e-02
-5.03950894e-01 1.34160101e-01 8.25848699e-01 8.01448897e-02
5.21488070e-01 -2.04614282e-01 -2.83033639e-01 -3.14067900e-01
-6.31356180e-01 2.47107700e-01 3.00752848e-01 1.22449517e+00
4.46681976e-01 3.17145377e-01 -3.43552142e-01 9.50657010e-01
3.07578892e-01 5.16593635e-01 2.98206210e-01 -7.43002653e-01
2.94501930e-01 2.75255263e-01 -8.85302294e-03 -1.08508229e+00
-1.82744637e-01 -3.38666849e-02 -9.69079256e-01 -3.19300562e-01
3.41808081e-01 -1.56470630e-02 -8.50390315e-01 1.56732130e+00
3.82170260e-01 6.75519466e-01 -2.56372690e-02 9.85076904e-01
1.29639375e+00 7.95214891e-01 8.18641305e-01 -3.14575178e-03
1.48073184e+00 -1.08970630e+00 -9.42200899e-01 -1.65570021e-01
3.28563422e-01 -5.11337936e-01 7.56623268e-01 6.10862784e-02
-7.65562177e-01 -9.77126539e-01 -1.13018215e+00 -5.23390211e-02
-6.74838126e-01 3.71293217e-01 9.57249463e-01 4.97834682e-01
-9.17464316e-01 1.41779572e-01 -4.64935303e-01 -7.25158274e-01
7.02107906e-01 3.84646893e-01 -5.51352501e-01 -1.61065564e-01
-1.63761830e+00 8.12858343e-01 5.48345685e-01 3.26223493e-01
-8.77118945e-01 -2.59362578e-01 -9.37906623e-01 -2.87852317e-01
3.06060404e-01 -8.91809702e-01 1.04679930e+00 -9.65670764e-01
-1.35494220e+00 7.43380249e-01 -3.75856251e-01 -3.85057330e-01
6.00571990e-01 -4.45369273e-01 -7.48314619e-01 1.35373890e-01
1.22289278e-01 6.89853728e-01 6.59583092e-01 -9.09558594e-01
-7.63926148e-01 -4.51658070e-01 -4.31692414e-03 3.38047802e-01
-4.80677783e-01 1.07075170e-01 -6.04211628e-01 -6.89833403e-01
-5.35531044e-01 -8.26186359e-01 -1.92165673e-01 -1.12222396e-01
-2.07301542e-01 -3.49777550e-01 9.76175606e-01 -9.05565202e-01
1.18041074e+00 -1.92449951e+00 2.12283824e-02 -1.97590068e-01
9.82163548e-02 5.46360612e-01 -1.64677352e-01 2.48859629e-01
-6.56018183e-02 -1.44319311e-01 4.03442644e-02 -2.37950355e-01
-1.77651867e-01 1.39243200e-01 2.77417034e-01 6.04411960e-01
3.60540375e-02 9.27872241e-01 -7.64567971e-01 -8.46356392e-01
4.24031079e-01 8.27730358e-01 -2.38296743e-02 1.92128599e-01
1.33134976e-01 6.50115311e-01 -7.91825294e-01 1.11259580e+00
8.31791878e-01 1.44818380e-01 -2.33355328e-01 -4.22245979e-01
-1.35083899e-01 -5.02606034e-01 -1.22082293e+00 1.60530770e+00
-4.41206135e-02 7.94284105e-01 -1.76442206e-01 -1.30958021e+00
9.61181223e-01 3.92096519e-01 7.31772184e-01 -6.90945387e-01
2.04286754e-01 -1.39994875e-01 -3.20253640e-01 -7.05859244e-01
8.75494301e-01 2.93452740e-01 -9.99067351e-02 -2.81877309e-01
1.04469925e-01 6.80686176e-01 4.65605646e-01 1.12830624e-01
4.93770152e-01 6.01049244e-01 4.91930485e-01 -2.05366850e-01
1.10713136e+00 5.44428825e-02 8.78150940e-01 6.99492455e-01
-9.00083184e-01 2.99543530e-01 6.53273836e-02 -8.88924658e-01
-1.23067451e+00 -7.12990761e-01 -3.08278590e-01 1.34177101e+00
2.43305638e-01 -2.81022996e-01 -4.52310890e-01 -4.51373607e-01
3.20061930e-02 3.88204545e-01 -8.13586712e-01 -9.18311179e-02
-6.33599877e-01 -8.35919619e-01 5.31542659e-01 8.36557031e-01
1.08434737e+00 -7.78051496e-01 -1.89738959e-01 2.23505482e-01
-6.09824181e-01 -1.34982908e+00 -5.96969545e-01 -4.75259066e-01
-3.84745955e-01 -1.01017737e+00 -9.53209460e-01 -6.79126084e-01
3.02940875e-01 3.12972784e-01 9.98005211e-01 1.10772669e-01
-3.34813684e-01 6.32252276e-01 -4.87275481e-01 -3.81026894e-01
2.03220084e-01 5.32835126e-02 3.37852269e-01 4.18278903e-01
6.78686619e-01 -3.12527746e-01 -7.27371097e-01 4.72478718e-01
-3.23111236e-01 -6.93494920e-04 6.14940226e-01 1.09010625e+00
6.22329235e-01 -4.83293742e-01 7.52142251e-01 -4.44368243e-01
3.65395606e-01 -5.23420930e-01 -4.06708688e-01 3.50779176e-01
-1.58699825e-01 -4.63327676e-01 3.45510632e-01 -7.04072237e-01
-1.20340657e+00 2.95316160e-01 -4.20918465e-01 -4.00637537e-01
-6.43413305e-01 3.83049697e-01 -4.99900013e-01 -1.23959526e-01
1.43466175e-01 6.30898833e-01 -5.44075072e-02 -3.23869735e-01
2.00887114e-01 7.13034689e-01 6.64760888e-01 -8.06455314e-01
3.30523670e-01 3.26477617e-01 7.58357570e-02 -1.15909278e+00
-6.84918880e-01 -6.12887681e-01 -1.18109965e+00 -7.78770506e-01
1.22042954e+00 -1.11245310e+00 -1.01676631e+00 9.22685444e-01
-1.03918719e+00 -4.96307574e-02 8.05392712e-02 5.55048943e-01
-7.68933415e-01 6.43343329e-01 -6.13355339e-01 -8.49157274e-01
-2.32524812e-01 -1.07475626e+00 6.73919916e-01 2.69987404e-01
1.60260499e-01 -8.74407709e-01 -8.12734738e-02 6.06769621e-01
5.70247829e-01 2.97127903e-01 3.35176319e-01 -9.81944919e-01
-6.43564761e-01 -3.11285228e-01 -3.91120374e-01 8.60523880e-02
-1.99510813e-01 1.32898957e-01 -1.00286150e+00 -2.66206235e-01
-4.20907706e-01 -2.47696221e-01 1.08400023e+00 6.68074429e-01
1.20358574e+00 -1.59257501e-01 -5.18772423e-01 6.79308712e-01
1.22735465e+00 3.96742880e-01 7.11284518e-01 7.68904686e-01
8.48188698e-01 4.18737859e-01 9.48965788e-01 5.70085168e-01
6.61369622e-01 1.17845511e+00 8.71805772e-02 2.31961802e-01
-8.48507509e-02 2.99308803e-02 2.16584355e-01 7.86161125e-01
-6.40093863e-01 -3.24285030e-01 -9.75805640e-01 4.33314770e-01
-2.27695084e+00 -1.60156953e+00 -2.01172277e-01 2.06310558e+00
2.71488905e-01 -2.21125558e-01 4.49622631e-01 -2.16632172e-01
9.96668458e-01 3.23246479e-01 -3.61956090e-01 -8.43370110e-02
-5.17608464e-01 -5.09703815e-01 2.16720060e-01 2.77265698e-01
-2.01015139e+00 9.08241928e-01 5.93055248e+00 1.05595136e+00
-6.32678211e-01 1.85071543e-01 8.10625911e-01 -3.80887277e-02
4.04857904e-01 -2.98075378e-01 -1.09134257e+00 7.23771095e-01
1.12540960e+00 -3.84776801e-01 -1.16882920e-01 9.01180506e-01
1.46065220e-01 -7.16660395e-02 -7.95892477e-01 1.13667870e+00
2.39765227e-01 -1.15762985e+00 3.59913021e-01 -1.54570550e-01
9.76901203e-02 -4.26842421e-01 1.62033021e-01 6.57549083e-01
-2.98806336e-02 -1.01311672e+00 5.63876808e-01 8.97827089e-01
9.41147268e-01 -8.37250829e-01 8.98993373e-01 -7.77644143e-02
-2.04663730e+00 -2.22715303e-01 -4.41280156e-01 -3.08306217e-02
4.08281952e-01 1.00196719e-01 -1.96927518e-01 8.89570713e-01
7.92973578e-01 1.21296656e+00 -7.79128909e-01 1.37627470e+00
6.67131841e-02 3.03823858e-01 -9.91901290e-03 -1.66702464e-01
2.47542679e-01 -3.46279562e-01 6.93823040e-01 1.53317010e+00
4.50625956e-01 4.11213815e-01 4.62738037e-01 3.10531408e-01
3.08017969e-01 1.80658713e-01 -8.29961061e-01 2.33344316e-01
5.93871415e-01 1.32513106e+00 -2.28517339e-01 -4.46436375e-01
-7.85712779e-01 8.21350396e-01 2.80069619e-01 5.57180226e-01
-1.08409333e+00 -1.47249371e-01 1.01639593e+00 -1.04139335e-01
2.83971250e-01 -4.72583055e-01 1.28604233e-01 -1.17008567e+00
-1.47260681e-01 -5.42537928e-01 9.13530827e-01 -6.51732624e-01
-1.33827686e+00 6.90357506e-01 2.49747202e-01 -1.35766029e+00
-2.16063231e-01 -6.16331637e-01 -4.87017065e-01 6.90787375e-01
-1.47645259e+00 -1.73149884e+00 -6.74656034e-01 9.59626555e-01
8.23895633e-01 -5.50679028e-01 7.34019995e-01 9.44715858e-01
-1.03451931e+00 8.97859871e-01 6.57506436e-02 3.50556403e-01
7.22044230e-01 -9.50046539e-01 2.17905417e-01 7.78194308e-01
-3.23135644e-01 6.52148873e-02 5.80757737e-01 -7.76181281e-01
-1.32667351e+00 -1.26236248e+00 7.05112994e-01 -4.72787142e-01
5.78284979e-01 -2.16454998e-01 -9.56321597e-01 7.91640282e-01
4.47761118e-01 6.06765747e-02 8.17426264e-01 -8.45235586e-02
1.77996308e-02 -1.17245726e-01 -9.67412353e-01 3.94437671e-01
1.19322658e+00 -2.24682987e-01 -2.61469215e-01 3.19914281e-01
4.48730618e-01 -1.74057677e-01 -1.01895308e+00 5.06209791e-01
5.33756793e-01 -5.10308206e-01 1.50960958e+00 -7.88004339e-01
1.01537898e-01 -1.93965226e-01 -5.62544823e-01 -1.16719532e+00
-6.48242593e-01 -2.12129071e-01 -2.57937312e-01 1.42825353e+00
1.15052737e-01 -4.26454246e-01 6.97296977e-01 5.97641766e-01
-1.89507708e-01 -7.19660759e-01 -1.20121634e+00 -8.82379830e-01
-2.14244679e-01 -2.83769131e-01 4.46849972e-01 8.24205875e-01
-4.82895106e-01 3.19830805e-01 -8.90580535e-01 -7.89064169e-02
9.33068395e-01 -2.05487430e-01 6.29361928e-01 -1.32744253e+00
1.97579190e-01 -3.23022842e-01 -9.31434393e-01 -9.62808013e-01
3.36452544e-01 -6.59026206e-01 -4.18692112e-01 -1.35817301e+00
5.97733736e-01 -2.05523580e-01 -3.48300219e-01 2.36349985e-01
-3.68592948e-01 6.57274500e-02 1.98482111e-01 2.04766572e-01
-1.01480186e+00 1.03075123e+00 9.69792187e-01 -4.59290326e-01
2.30589345e-01 8.53909329e-02 -1.70644298e-01 5.03884077e-01
8.72073710e-01 3.90778854e-03 -1.27538979e-01 -2.54273862e-01
-4.19603080e-01 2.53367517e-02 6.34329975e-01 -1.25268030e+00
5.29640734e-01 -1.90541536e-01 8.25013101e-01 -8.02742779e-01
9.36718762e-01 -5.81193745e-01 1.69378951e-01 2.20551908e-01
-1.35463610e-01 5.51682055e-01 3.62808913e-01 8.54618251e-01
-4.73149210e-01 1.98326439e-01 8.19254220e-01 -1.43337816e-01
-1.68034303e+00 8.27499807e-01 -5.72064340e-01 -1.39622822e-01
1.52895677e+00 -5.41271091e-01 -3.68073732e-01 -4.24432814e-01
-1.08431435e+00 4.12289739e-01 3.80353220e-02 4.72097516e-01
1.09875119e+00 -2.00313711e+00 -1.02931607e+00 5.87235987e-02
2.98414022e-01 -7.93728292e-01 9.11036193e-01 9.21718180e-01
-2.48104975e-01 5.74588358e-01 -7.59264052e-01 -5.85765302e-01
-1.49365938e+00 7.45698869e-01 2.28811041e-01 1.63947508e-01
-4.77073997e-01 7.26414382e-01 -2.10123196e-01 -2.02357382e-01
3.14917147e-01 7.67875612e-01 -8.14011872e-01 3.86739671e-01
4.87020373e-01 5.58379710e-01 -3.64609092e-01 -1.28456879e+00
-5.43202877e-01 5.66534042e-01 -1.70110703e-01 2.39152879e-01
1.29822803e+00 -3.89890164e-01 1.74166724e-01 3.10850143e-01
1.08482301e+00 -6.73851967e-01 -1.20192480e+00 -2.69956529e-01
-1.49364337e-01 -8.93200278e-01 -2.03846350e-01 -4.00372654e-01
-9.95627582e-01 7.64714599e-01 9.18157935e-01 -2.37014174e-01
1.08222914e+00 -2.80765414e-01 8.30839217e-01 4.56157357e-01
3.05019766e-01 -1.33833230e+00 1.23612680e-01 6.03662968e-01
6.46110177e-01 -1.51475120e+00 8.83616582e-02 -3.96113634e-01
-9.68443513e-01 9.01484251e-01 1.17791355e+00 2.60037184e-01
5.52212894e-01 -7.49276429e-02 -9.15233716e-02 1.64596021e-01
-7.38887250e-01 -4.19269443e-01 3.54345262e-01 9.25075233e-01
3.78234535e-01 8.22705701e-02 -2.89015293e-01 7.11599648e-01
2.91829258e-01 3.15817422e-03 3.95988494e-01 7.28839576e-01
-3.49309385e-01 -1.04140091e+00 -3.18122536e-01 5.53815484e-01
-2.88753897e-01 -3.30769867e-02 3.46363150e-02 7.40553498e-01
1.34401936e-02 7.31625497e-01 3.44782434e-02 -7.20687807e-01
2.76563168e-01 1.69333294e-01 2.84075260e-01 1.56982243e-01
1.89755876e-02 -3.53631526e-01 5.49916625e-01 -5.64856529e-01
-6.69219851e-01 -1.05229557e+00 -5.70078492e-01 -6.81543827e-01
-9.47285723e-03 5.00584655e-02 2.56733984e-01 5.93879044e-01
4.05733049e-01 4.13706362e-01 3.30728412e-01 -1.07320714e+00
-2.05815390e-01 -9.59269404e-01 -3.87538910e-01 5.52905083e-01
1.19447872e-01 -1.16913164e+00 5.82519546e-02 2.17273254e-02] | [14.466917991638184, 0.9694693684577942] |
a7f63a10-4854-469d-b068-32cd879979df | imitation-and-supervised-learning-of | 2111.10488 | null | https://arxiv.org/abs/2111.10488v1 | https://arxiv.org/pdf/2111.10488v1.pdf | Imitation and Supervised Learning of Compliance for Robotic Assembly | We present the design of a learning-based compliance controller for assembly operations for industrial robots. We propose a solution within the general setting of learning from demonstration (LfD), where a nominal trajectory is provided through demonstration by an expert teacher. This can be used to learn a suitable representation of the skill that can be generalized to novel positions of one of the parts involved in the assembly, for example the hole in a peg-in-hole (PiH) insertion task. Under the expectation that this novel position might not be entirely accurately estimated by a vision or other sensing system, the robot will need to further modify the generated trajectory in response to force readings measured by means of a force-torque (F/T) sensor mounted at the wrist of the robot or another suitable location. Under the assumption of constant velocity of traversing the reference trajectory during assembly, we propose a novel accommodation force controller that allows the robot to safely explore different contact configurations. The data collected using this controller is used to train a Gaussian process model to predict the misalignment in the position of the peg with respect to the target hole. We show that the proposed learning-based approach can correct various contact configurations caused by misalignment between the assembled parts in a PiH task, achieving high success rate during insertion. We show results using an industrial manipulator arm, and demonstrate that the proposed method can perform adaptive insertion using force feedback from the trained machine learning models. | ['Daniel Nikovski', 'William Yerazunis', 'Diego Romeres', 'Devesh K. Jha'] | 2021-11-20 | null | null | null | null | ['industrial-robots'] | ['robots'] | [ 3.91593665e-01 6.03604615e-01 1.91633016e-01 1.21729217e-01
-3.29757810e-01 -6.46579325e-01 1.04580685e-01 3.34342048e-02
-3.65013123e-01 6.13500178e-01 -6.89076960e-01 -1.18304558e-01
-6.02486789e-01 -3.83248866e-01 -1.31587839e+00 -8.26328933e-01
1.83753759e-01 9.34758425e-01 1.05712168e-01 -2.40540192e-01
4.35104370e-01 7.71735787e-01 -1.32095993e+00 -3.28926921e-01
7.35538006e-01 5.51138580e-01 1.13268614e+00 7.10695922e-01
5.59572875e-01 8.83785635e-02 -5.59731483e-01 2.44541243e-01
5.49039543e-01 1.93203911e-01 -4.36117500e-01 5.62227130e-01
-1.17850028e-01 -2.93957174e-01 -1.17132559e-01 1.09671104e+00
2.44031742e-01 4.92993236e-01 7.78936386e-01 -1.05954909e+00
-1.53564721e-01 4.76762921e-01 -1.82973921e-01 -4.62587237e-01
5.61371148e-01 3.70071858e-01 2.61405021e-01 -6.80367470e-01
7.98387706e-01 1.21198046e+00 5.29438972e-01 4.65994298e-01
-1.15819907e+00 -1.10186338e-01 2.25785766e-02 -4.53510173e-02
-1.05788946e+00 2.74602205e-01 6.55573606e-01 -7.65341759e-01
5.23753881e-01 -1.29464656e-01 4.89441544e-01 1.06197321e+00
9.19679880e-01 4.29122299e-01 8.23487043e-01 -4.55437988e-01
4.26764756e-01 3.12537372e-01 -3.95904601e-01 5.56282759e-01
1.85786217e-01 2.39446849e-01 2.28324115e-01 4.86022532e-02
1.17730558e+00 3.12860489e-01 -5.73476076e-01 -1.02398026e+00
-1.14013803e+00 6.28707290e-01 5.06592870e-01 2.54410263e-02
-8.33924055e-01 -9.80106741e-02 1.43987596e-01 5.70828319e-01
-2.18313664e-01 1.04548502e+00 -8.59009981e-01 -3.44313622e-01
-1.55077130e-01 4.36250269e-01 1.00455046e+00 1.30388594e+00
3.31701338e-01 1.12011224e-01 1.97117552e-01 3.77315402e-01
7.99339563e-02 3.82122636e-01 2.87005872e-01 -1.23420620e+00
5.74899673e-01 3.95686746e-01 6.88699126e-01 -7.38090098e-01
-5.02293169e-01 -2.03395844e-01 -3.20631146e-01 7.98407555e-01
3.30645561e-01 -5.43448746e-01 -6.61365449e-01 1.40245903e+00
6.47086084e-01 -1.55837357e-01 1.14421688e-01 1.00807905e+00
-3.44462156e-01 6.71973944e-01 -5.61055243e-01 -4.15485054e-01
7.18275666e-01 -7.31273651e-01 -7.30248332e-01 -5.91614582e-02
4.09636170e-01 -4.42168742e-01 1.03821075e+00 9.42441523e-01
-9.49486613e-01 -9.93687391e-01 -1.18190944e+00 5.95310628e-01
-3.16743739e-02 3.50812972e-01 -1.20238274e-01 -2.26266697e-01
-4.12409186e-01 1.15900242e+00 -1.06803071e+00 -2.26059511e-01
-4.76554245e-01 5.16615808e-01 -3.94972265e-01 1.51656419e-01
-6.77871048e-01 1.34933555e+00 4.89181459e-01 5.40316820e-01
-1.18519449e+00 -5.32171905e-01 -6.72944963e-01 -1.16886824e-01
8.22945714e-01 -6.24944925e-01 1.37866199e+00 -4.52058256e-01
-2.02003765e+00 6.07299283e-02 4.23928261e-01 -2.51300514e-01
8.26819658e-01 -6.75529718e-01 2.82812655e-01 1.52520970e-01
-6.44569658e-03 3.56353462e-01 1.31601584e+00 -1.43917799e+00
-4.35677260e-01 -2.38837779e-01 3.23717780e-02 3.05294454e-01
1.56664312e-01 -6.20077789e-01 1.93478033e-01 -2.91397542e-01
3.17308098e-01 -1.55672371e+00 -3.13695967e-01 5.85874729e-02
-4.66302246e-01 -2.61243343e-01 1.10634494e+00 -4.18388009e-01
2.94600368e-01 -2.06989694e+00 7.79974520e-01 2.52398819e-01
-4.52140182e-01 1.63644895e-01 6.30273521e-02 8.16613793e-01
3.18385959e-02 -5.88847578e-01 -3.48507538e-02 2.82400753e-02
-1.09037936e-01 2.98859000e-01 -3.16771090e-01 3.64439696e-01
4.30723242e-02 4.75264668e-01 -1.12491465e+00 3.03374410e-01
3.28316510e-01 2.87313610e-01 -3.20941448e-01 6.22865021e-01
-4.21228260e-01 8.63038301e-01 -6.19587243e-01 1.26957715e-01
2.26061925e-01 3.53442043e-01 1.86568663e-01 -2.72048593e-01
-2.29118556e-01 -1.26483187e-01 -1.41210234e+00 1.50293684e+00
-7.87929952e-01 1.81699544e-01 4.67966080e-01 -1.08839738e+00
1.47023261e+00 5.15490830e-01 2.24459618e-01 2.86988080e-01
3.68833780e-01 2.50110209e-01 3.27530056e-01 -8.62143517e-01
3.56708616e-02 4.31544930e-02 1.15351424e-01 1.08097494e-01
1.05593920e-01 -1.11992753e+00 -2.03727841e-01 -5.07526875e-01
9.49191332e-01 5.40503502e-01 6.11070469e-02 2.19340529e-02
4.75256920e-01 6.00975677e-02 3.72515023e-01 5.77766716e-01
1.36304319e-01 2.92182595e-01 1.25728697e-01 -2.44110107e-01
-1.20204568e+00 -9.98268366e-01 2.87619591e-01 4.98190284e-01
2.54865259e-01 3.87964725e-01 -6.50180399e-01 -4.83686030e-01
3.23053360e-01 8.28776598e-01 -2.73525208e-01 -4.62141186e-01
-8.00478816e-01 1.52206019e-01 -3.50953102e-01 5.67433774e-01
3.91853116e-02 -1.07375586e+00 -1.09770179e+00 5.05229950e-01
3.24149370e-01 -1.01933920e+00 -2.02562734e-01 6.03861392e-01
-1.21618938e+00 -1.38958716e+00 -3.52684826e-01 -1.19241130e+00
9.95376825e-01 -1.10454857e-01 5.66692874e-02 -5.49476802e-01
-2.42063448e-01 8.01629186e-01 -2.51233846e-01 -4.70327646e-01
-7.99122751e-01 -1.92700654e-01 4.41996694e-01 -3.15812647e-01
-4.19829875e-01 -3.76140505e-01 -2.43669435e-01 4.54519868e-01
-4.56522673e-01 -4.23834473e-02 6.60916567e-01 8.94668579e-01
4.23962593e-01 6.58873141e-01 4.95586514e-01 -3.20543766e-01
9.19755280e-01 -8.33666846e-02 -7.73508787e-01 -6.68560639e-02
-3.81933004e-01 -7.55210295e-02 8.47089708e-01 -1.17295372e+00
-1.07210469e+00 5.00640810e-01 8.90555009e-02 -9.17256117e-01
-2.73319066e-01 5.28944254e-01 -1.64190754e-01 3.10958433e-03
3.99016172e-01 1.43223420e-01 6.13708496e-01 -5.58076084e-01
2.10159078e-01 5.17186701e-01 7.04311609e-01 -7.56097674e-01
9.48275030e-01 -5.45589849e-02 4.03279871e-01 -6.18385375e-01
-2.93756843e-01 -2.74074376e-01 -1.19791532e+00 -2.86586225e-01
4.91790026e-01 -4.87579405e-01 -1.20932662e+00 4.68008399e-01
-1.18690181e+00 -5.31847000e-01 -3.78750741e-01 6.67087674e-01
-1.10502791e+00 1.54785454e-01 -7.96948850e-01 -1.03146827e+00
-1.73141032e-01 -1.47429609e+00 1.00614190e+00 1.33146390e-01
-2.30706215e-01 -8.37167382e-01 -2.82224178e-01 5.68257160e-02
1.02966033e-01 5.08886993e-01 1.03324437e+00 -5.91036081e-01
-3.29411298e-01 -7.59893715e-01 7.94750094e-01 5.81319392e-01
4.58864957e-01 -1.14059687e-01 -2.23509014e-01 -8.16168666e-01
7.28434443e-01 -2.69242048e-01 -6.39073476e-02 3.24508518e-01
8.11485350e-01 -3.13860625e-01 -6.14853740e-01 -6.14746213e-02
1.11164010e+00 7.08568275e-01 1.02270193e-01 4.75001812e-01
6.52327955e-01 7.67156601e-01 1.36552131e+00 2.40958005e-01
-1.78525135e-01 8.50800276e-01 7.66551256e-01 6.16601706e-01
4.70788151e-01 -2.57129461e-01 4.25566554e-01 5.68359673e-01
-7.08941817e-02 1.89103842e-01 -4.74018723e-01 4.00243491e-01
-1.90110481e+00 -3.90792638e-01 -1.57846007e-02 2.38733387e+00
6.64346099e-01 2.82415181e-01 -2.19673634e-01 3.01259398e-01
8.71257901e-01 -7.42704153e-01 -8.73343706e-01 -5.45540333e-01
7.60559499e-01 -1.95505276e-01 4.47851062e-01 6.77339435e-01
-6.64214432e-01 4.31325674e-01 5.04276133e+00 3.24746549e-01
-1.18145645e+00 -3.95519674e-01 -3.03364128e-01 3.32785279e-01
3.92678052e-01 -8.90594125e-02 -6.41247988e-01 4.12174344e-01
5.06106913e-01 5.45478910e-02 4.59806234e-01 1.15540552e+00
5.23345113e-01 -1.72777519e-01 -1.51230407e+00 4.20221537e-01
-4.45558503e-02 -6.14694715e-01 -3.54786187e-01 -2.67803609e-01
6.59800947e-01 -7.49028385e-01 9.73595381e-02 3.22984755e-01
-8.55894294e-03 -2.83430427e-01 6.67983353e-01 6.39640152e-01
2.62359142e-01 -5.02806544e-01 6.73511147e-01 1.11900628e+00
-7.31161237e-01 -7.73385525e-01 -3.74884427e-01 -2.73637503e-01
3.58411878e-01 8.79615024e-02 -1.64627349e+00 4.53366280e-01
1.26280367e-01 2.27548122e-01 2.62095779e-01 1.02787864e+00
-5.70752382e-01 8.87203440e-02 -3.40103477e-01 -3.06039929e-01
1.06458806e-01 -4.70291823e-01 1.12419891e+00 4.29746211e-01
5.77424049e-01 -3.20869565e-01 4.84153539e-01 9.66506541e-01
5.48032761e-01 -4.02380884e-01 -7.64717460e-01 4.01045203e-01
4.28359151e-01 1.04571795e+00 -2.32845172e-01 -1.25974178e-01
3.37320775e-01 1.09312665e+00 1.04724735e-01 3.51638764e-01
-6.30715370e-01 -5.42146444e-01 3.29990745e-01 9.79248583e-02
5.45464218e-01 -5.89604497e-01 2.85321116e-01 -4.52833176e-01
2.17719212e-01 -6.68452919e-01 -3.67850363e-01 -1.22706878e+00
-8.80108118e-01 6.54299557e-02 2.68736154e-01 -1.49624240e+00
-7.05760241e-01 -8.20098400e-01 -5.33393502e-01 7.47946441e-01
-8.27794790e-01 -5.59359312e-01 -1.68357775e-01 1.31069660e-01
1.04469728e+00 4.95331641e-03 6.87631190e-01 -4.67409760e-01
-1.78667307e-01 -8.08307827e-02 1.00941353e-01 -3.26132804e-01
6.45002842e-01 -1.35968018e+00 -3.28640151e-03 2.21242428e-01
-3.72137696e-01 7.31627703e-01 1.17866898e+00 -8.48102152e-01
-1.79202437e+00 -1.19283533e+00 2.23125458e-01 -1.94255695e-01
6.97985590e-01 -2.68034965e-01 -8.65316093e-01 9.57388997e-01
-3.43905032e-01 -3.15046370e-01 -3.33925575e-01 -5.89974463e-01
4.57584053e-01 -1.54841125e-01 -1.23631883e+00 6.37152076e-01
4.76648331e-01 -1.13235034e-01 -1.15101135e+00 5.32270730e-01
8.08864594e-01 -9.70150292e-01 -1.00063467e+00 7.03348696e-01
4.94191498e-01 -8.44848603e-02 7.02285945e-01 -5.33075631e-01
1.22848220e-01 -4.03205395e-01 3.97349626e-01 -2.00758410e+00
-2.06467420e-01 -9.24296498e-01 -1.34123385e-01 5.27197599e-01
4.95827161e-02 -5.58981955e-01 6.51216626e-01 3.05752248e-01
-4.49430466e-01 -9.61115956e-01 -9.33839560e-01 -8.97208512e-01
2.16629580e-02 3.17212231e-02 1.16518289e-01 3.69037092e-01
4.30057883e-01 2.29251385e-01 -2.13391244e-01 8.36897194e-01
3.89417142e-01 9.53789800e-02 1.09908247e+00 -1.13095558e+00
-4.91147518e-01 2.62112290e-01 -3.87272894e-01 -1.34456778e+00
3.41807693e-01 -3.18582743e-01 7.63560712e-01 -1.59030473e+00
-3.61293972e-01 -3.22671801e-01 3.25114399e-01 7.62209743e-02
-2.65049879e-02 -6.96974516e-01 3.90933722e-01 1.16500959e-01
3.45981270e-02 4.61910188e-01 1.67835355e+00 -6.82582706e-02
-6.49660766e-01 8.93250883e-01 2.51379400e-01 8.12630355e-01
7.82729864e-01 -2.42331475e-01 -6.50171041e-01 -2.27633581e-01
-1.11157112e-01 7.36457527e-01 1.47659078e-01 -1.20072663e+00
4.18627411e-01 -4.59145531e-02 2.74782062e-01 -3.80032927e-01
4.77873266e-01 -1.52912629e+00 4.45380747e-01 1.04249513e+00
-4.04392570e-01 1.37684703e-01 5.13078831e-02 7.04566538e-01
1.00751318e-01 -5.65495610e-01 3.44399065e-01 -4.37093414e-02
-1.81693614e-01 -1.90598115e-01 -6.02786362e-01 -8.76522779e-01
1.29563510e+00 -2.85832316e-01 -3.33019085e-02 -3.06626111e-01
-1.13842535e+00 3.14233124e-01 2.63382763e-01 6.30517483e-01
6.66717112e-01 -7.63075829e-01 -1.64353743e-01 3.33743155e-01
-2.71627784e-01 2.87303567e-01 -1.84947327e-02 8.03464711e-01
-2.59975195e-01 3.58663231e-01 -2.22553700e-01 -8.77205729e-01
-9.81143713e-01 8.98224294e-01 4.36147392e-01 1.29459038e-01
-8.05329263e-01 3.22454333e-01 -3.92172560e-02 -7.05761552e-01
3.45583230e-01 -8.02462578e-01 4.38836450e-03 -5.70040166e-01
-2.00059637e-02 3.88893962e-01 1.43808350e-01 -2.80020311e-02
4.50268805e-01 6.92950726e-01 1.42266333e-01 8.12495500e-02
1.24813604e+00 -1.22011118e-01 4.05455828e-01 7.89840281e-01
7.18219519e-01 -3.83329332e-01 -1.88309777e+00 1.94256559e-01
-3.51419463e-03 -3.72316867e-01 -5.79692483e-01 -7.85225987e-01
-5.69983959e-01 5.87579846e-01 5.61252356e-01 -2.60048714e-02
5.28976023e-01 -2.21381024e-01 3.75081986e-01 7.93892443e-01
8.47296953e-01 -1.21916127e+00 3.94134790e-01 5.24529755e-01
1.36499894e+00 -8.53350222e-01 -2.99427062e-01 -6.16278648e-01
-5.26535451e-01 1.54285264e+00 8.15889359e-01 -5.46515167e-01
4.80672687e-01 4.14413780e-01 1.53951282e-02 1.04233511e-02
-4.56129909e-01 4.84284282e-01 2.34541342e-01 8.15978944e-01
-2.97453046e-01 9.00586843e-02 -1.16596170e-01 5.14152907e-02
-3.27047147e-02 -9.35028493e-02 8.11943889e-01 1.31248391e+00
-9.15684462e-01 -8.12380791e-01 -7.01843977e-01 5.33343293e-02
1.75424606e-01 7.37454772e-01 2.10275590e-01 1.06175280e+00
4.80098166e-02 7.43452609e-01 1.37743101e-01 -5.87563701e-02
9.49733615e-01 4.91168015e-02 9.10526812e-01 -9.71380889e-01
-2.47184694e-01 2.46255845e-01 -2.75835723e-01 -5.23998797e-01
4.49225903e-02 -5.42267978e-01 -1.45089841e+00 6.88611269e-01
-4.56355780e-01 1.88893005e-01 9.07444000e-01 9.18083906e-01
6.98776990e-02 6.01850629e-01 7.00613737e-01 -1.51467109e+00
-1.50520730e+00 -1.31498146e+00 -6.59348309e-01 2.72900164e-01
4.43196654e-01 -1.18388784e+00 -5.40412188e-01 9.33201462e-02] | [4.815544605255127, 1.3684022426605225] |
c57af8b4-5c19-4cf0-9e25-92fb97f5fcfc | towards-speech-to-text-translation-without | 1702.03856 | null | http://arxiv.org/abs/1702.03856v1 | http://arxiv.org/pdf/1702.03856v1.pdf | Towards speech-to-text translation without speech recognition | We explore the problem of translating speech to text in low-resource
scenarios where neither automatic speech recognition (ASR) nor machine
translation (MT) are available, but we have training data in the form of audio
paired with text translations. We present the first system for this problem
applied to a realistic multi-speaker dataset, the CALLHOME Spanish-English
speech translation corpus. Our approach uses unsupervised term discovery (UTD)
to cluster repeated patterns in the audio, creating a pseudotext, which we pair
with translations to create a parallel text and train a simple bag-of-words MT
model. We identify the challenges faced by the system, finding that the
difficulty of cross-speaker UTD results in low recall, but that our system is
still able to correctly translate some content words in test data. | ['Sameer Bansal', 'Sharon Goldwater', 'Adam Lopez', 'Herman Kamper'] | 2017-02-13 | towards-speech-to-text-translation-without-1 | https://aclanthology.org/E17-2076 | https://aclanthology.org/E17-2076.pdf | eacl-2017-4 | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 6.78361356e-01 1.22103244e-01 1.16377868e-01 -4.36422288e-01
-1.94478691e+00 -8.00239265e-01 8.00065041e-01 -2.40167201e-01
-2.89416641e-01 6.23336017e-01 5.13698459e-01 -7.25956321e-01
3.57344449e-01 -8.38983357e-02 -6.74649000e-01 -3.62964362e-01
2.86698997e-01 1.00303209e+00 6.57658055e-02 -3.82596403e-01
-2.81047583e-01 1.03306651e-01 -1.33842254e+00 8.58858168e-01
5.72332144e-01 5.87397099e-01 3.40868831e-01 8.69109571e-01
-3.17564964e-01 5.71974337e-01 -7.97894299e-01 -3.59275460e-01
2.97085404e-01 -6.86014831e-01 -1.09358692e+00 2.41123527e-01
2.65478075e-01 -1.05830230e-01 -6.72936738e-02 7.43308008e-01
8.82119477e-01 1.00544482e-01 5.20366311e-01 -8.98692369e-01
-2.51565665e-01 9.94205832e-01 -6.59321845e-02 2.74484098e-01
9.16761637e-01 -2.24535927e-01 8.57140899e-01 -1.22231770e+00
7.77832985e-01 1.37185073e+00 5.16457617e-01 5.34213901e-01
-1.48795700e+00 -5.01873374e-01 -3.37478191e-01 1.52516276e-01
-1.53180456e+00 -1.46738875e+00 3.09802294e-01 -1.62417442e-01
1.46924019e+00 6.64890826e-01 2.32438877e-01 1.55800116e+00
-1.16256617e-01 5.79558551e-01 1.10620344e+00 -9.57711518e-01
2.19537094e-01 3.41488928e-01 -5.32204866e-01 1.67861283e-01
-7.49523520e-01 -7.19356164e-03 -9.79537427e-01 -2.77020663e-01
2.08445936e-01 -5.83597124e-01 -2.08868667e-01 3.56313944e-01
-1.48230088e+00 4.40836489e-01 -2.45533183e-01 6.20417416e-01
-4.62923765e-01 -3.36414486e-01 3.77958417e-01 8.99721146e-01
6.91347301e-01 2.32016310e-01 -5.62641025e-01 -4.89082903e-01
-1.25332260e+00 -7.53418654e-02 1.01686859e+00 1.20195031e+00
6.16297603e-01 1.56387344e-01 -1.22671813e-01 1.27563155e+00
1.48409143e-01 9.24827874e-01 1.01004195e+00 -6.61785901e-01
9.45132613e-01 -2.33944267e-01 -1.16191864e-01 -4.43069309e-01
3.14500146e-02 -2.04561561e-01 -2.64974385e-01 -4.62324739e-01
1.01747751e-01 -2.90817291e-01 -9.69670415e-01 1.60409617e+00
3.17019403e-01 -8.04356579e-03 3.84084314e-01 5.55751741e-01
4.25372243e-01 9.46105599e-01 -4.56149906e-01 -6.90406382e-01
1.16322291e+00 -9.09820020e-01 -8.07640553e-01 -6.00680292e-01
7.76604950e-01 -1.50990856e+00 1.03830850e+00 1.42262653e-01
-1.12575865e+00 -3.64011258e-01 -8.45086575e-01 1.64502129e-01
-1.72069192e-01 8.72082859e-02 -1.41970634e-01 6.75276399e-01
-1.34341776e+00 2.21047536e-01 -7.49487340e-01 -8.09734404e-01
-3.44006032e-01 2.46671721e-01 -4.90799785e-01 -1.89597696e-01
-1.12316382e+00 1.07097149e+00 2.06583776e-02 -2.23629475e-01
-9.96461928e-01 -3.08183074e-01 -7.14738548e-01 -1.85691282e-01
2.11706206e-01 -3.04673463e-01 1.74589837e+00 -1.23394024e+00
-1.75644076e+00 8.13780308e-01 -6.96050763e-01 -3.52410316e-01
3.96387875e-01 2.03518674e-01 -8.34034860e-01 2.92416424e-01
2.89374411e-01 5.22803485e-01 1.03014612e+00 -1.09312475e+00
-6.34720743e-01 -1.82744667e-01 -6.31779253e-01 3.92083585e-01
-3.27352911e-01 7.61718452e-01 -2.89970011e-01 -7.79118299e-01
3.03794950e-01 -1.06580627e+00 1.58522785e-01 -8.46625447e-01
-4.47442472e-01 1.15353569e-01 6.26818836e-01 -1.01756227e+00
1.18650913e+00 -2.23853850e+00 5.38312122e-02 2.89017767e-01
-5.12626529e-01 -5.75454719e-02 -3.67908984e-01 8.15042734e-01
-1.92787647e-01 5.37062027e-02 -1.96019128e-01 -7.33477175e-01
-3.17511968e-02 3.92059296e-01 -6.13581836e-01 1.30993560e-01
1.62369937e-01 7.10973084e-01 -6.83741987e-01 -5.78955114e-01
-2.51946300e-01 2.76922852e-01 -1.32417962e-01 1.90078989e-01
-8.65182877e-02 3.76695573e-01 -9.13633034e-02 5.08877099e-01
1.48677349e-01 2.97397524e-01 3.62005204e-01 3.61100197e-01
-1.57349721e-01 1.32475722e+00 -9.41724777e-01 1.64659059e+00
-6.67219639e-01 8.08612049e-01 2.10009769e-01 -8.98256004e-01
9.13781881e-01 9.77712929e-01 3.55095088e-01 -8.05089712e-01
-5.96938320e-02 7.32620239e-01 -1.57676525e-02 -5.97606301e-01
3.67783338e-01 -4.05876160e-01 -9.83243734e-02 7.92229235e-01
2.84278691e-01 -4.45972383e-01 -9.47154015e-02 2.96137594e-02
1.28505170e+00 -2.48209611e-01 -6.50906116e-02 -1.60642993e-02
2.14792624e-01 2.90294647e-01 5.70306778e-02 5.29279649e-01
1.96570724e-01 9.69294310e-01 -6.00237064e-02 1.71617597e-01
-1.15073133e+00 -9.92663085e-01 8.17809850e-02 1.32175648e+00
-6.60945773e-01 -4.57691908e-01 -8.50808620e-01 -3.74231905e-01
-6.41373456e-01 9.73301649e-01 -6.94187284e-02 4.22468968e-02
-6.72453403e-01 -3.28622758e-01 8.74169886e-01 1.10889681e-01
-1.14614226e-01 -9.40881252e-01 1.42069638e-01 5.39192021e-01
-9.83872414e-01 -1.47222698e+00 -9.08499539e-01 5.21059215e-01
-6.83416963e-01 -2.94390321e-01 -6.88712537e-01 -1.15369797e+00
2.57171810e-01 3.38691324e-01 1.12801743e+00 -3.28999400e-01
1.86334713e-03 3.65509719e-01 -6.74333632e-01 -3.71662050e-01
-1.31007409e+00 8.29555467e-02 3.02121222e-01 1.25256419e-01
3.96181732e-01 -6.76711679e-01 1.12859316e-01 5.00137150e-01
-7.17335403e-01 -1.59315810e-01 6.59890890e-01 8.09860587e-01
4.34376329e-01 -2.55930334e-01 7.41464674e-01 -3.78673643e-01
7.72455215e-01 -3.34478110e-01 -1.66585092e-02 3.27313423e-01
-2.86946803e-01 -1.87303066e-01 4.75936413e-01 -7.89445043e-01
-7.38801956e-01 2.35196739e-01 -3.55021626e-01 -4.20099318e-01
-2.16617867e-01 7.20080554e-01 -9.94867161e-02 2.11116225e-01
9.10804868e-01 6.75098300e-01 8.27280059e-02 -5.09114981e-01
2.43190646e-01 1.57680261e+00 5.52200377e-01 -3.90922546e-01
6.04950070e-01 -5.27811833e-02 -6.04142547e-01 -1.05614424e+00
-3.58562529e-01 -6.78595185e-01 -6.23483658e-01 -1.20649211e-01
5.82557142e-01 -1.15330899e+00 2.33071759e-01 -8.78034532e-02
-1.38921046e+00 -3.23317498e-01 -2.50856161e-01 7.54789829e-01
-8.84837389e-01 3.39101762e-01 -5.71114123e-01 -9.05317068e-01
-2.49635726e-01 -1.16485572e+00 1.33711791e+00 -6.93627536e-01
-6.56677127e-01 -6.46316469e-01 2.40459800e-01 6.81148648e-01
4.38343674e-01 -5.32315552e-01 7.64232278e-01 -1.19033170e+00
-2.11016789e-01 -3.08869362e-01 3.57986927e-01 2.18784809e-01
3.12678784e-01 -3.00964117e-01 -1.25765514e+00 -2.23596260e-01
2.04512939e-01 -4.30002928e-01 3.48996431e-01 4.88406271e-02
2.64934212e-01 -6.36275768e-01 -4.50891219e-02 4.90073450e-02
6.85907662e-01 2.57968396e-01 4.94340390e-01 5.81740998e-02
3.37289095e-01 7.57785678e-01 4.61391687e-01 3.97214666e-02
4.25250858e-01 1.09030938e+00 -3.67174417e-01 -3.08395959e-02
-2.62567818e-01 -2.84240782e-01 1.03199136e+00 1.66185260e+00
5.07911980e-01 -2.97838569e-01 -1.14039218e+00 7.77023673e-01
-1.60606134e+00 -8.86899292e-01 -1.64829120e-02 2.26745033e+00
1.27365482e+00 -1.14379330e-02 3.89325738e-01 2.32165724e-01
7.12154448e-01 -3.63607466e-01 1.24935552e-01 -5.53956330e-01
-1.53162982e-02 3.38162661e-01 3.04052979e-01 7.60691106e-01
-7.96611845e-01 1.00383651e+00 7.28817797e+00 1.07378566e+00
-1.28048742e+00 5.99376857e-01 3.61491442e-01 -1.92450881e-01
-3.76077205e-01 -8.81871805e-02 -6.28877878e-01 3.63934219e-01
1.79066205e+00 -1.40418977e-01 8.35289121e-01 2.93398350e-01
4.46608603e-01 2.53476888e-01 -1.32477713e+00 9.60383952e-01
4.40451771e-01 -8.93248379e-01 9.48438272e-02 -1.06718004e-01
4.37091410e-01 3.85392606e-01 -1.23254798e-01 2.86080033e-01
9.64673162e-02 -9.63556528e-01 1.01367748e+00 -7.73576349e-02
1.16050565e+00 -3.58903259e-01 4.99474287e-01 6.14804208e-01
-9.74588633e-01 1.97865620e-01 -1.81287110e-01 2.75912493e-01
2.45237142e-01 2.04659760e-01 -1.67421770e+00 6.62165046e-01
4.39703196e-01 1.49550885e-01 -2.62678355e-01 7.59943008e-01
1.24233022e-01 9.15217936e-01 -6.20296419e-01 -8.59602392e-02
1.37985647e-01 1.19038485e-01 7.79482067e-01 1.47217429e+00
8.74287009e-01 -2.20384225e-01 3.81266892e-01 3.49808991e-01
-1.74262580e-02 5.32134175e-01 -6.51023865e-01 -1.91718131e-01
6.17516816e-01 7.63298035e-01 -6.00901306e-01 -5.14701068e-01
-1.66856512e-01 1.23135710e+00 -8.55710879e-02 3.76104414e-01
-2.54186600e-01 -1.90575853e-01 3.38494062e-01 8.46360549e-02
2.00137407e-01 -3.08146864e-01 -1.77353799e-01 -1.21436560e+00
3.45811576e-01 -1.49548364e+00 7.07748607e-02 -9.52062249e-01
-1.16040587e+00 1.08872783e+00 -1.60821155e-01 -1.47243559e+00
-9.11282659e-01 -1.36437565e-01 -3.74143869e-01 1.11469877e+00
-1.06269407e+00 -1.21749163e+00 3.83730054e-01 6.88204527e-01
1.06151295e+00 -4.56305265e-01 1.25631213e+00 6.59668326e-01
-5.93480244e-02 7.83413947e-01 2.13164106e-01 -2.55864486e-02
1.12933648e+00 -1.01366091e+00 6.75361335e-01 7.68416524e-01
8.21843565e-01 5.10678053e-01 9.85912383e-01 -6.23369396e-01
-1.37844336e+00 -9.42429900e-01 1.72941613e+00 -7.08154678e-01
8.79792690e-01 -8.88125002e-01 -7.41741896e-01 7.36808658e-01
4.42381173e-01 -4.90358233e-01 8.68747652e-01 -8.35302696e-02
-4.00898814e-01 -3.57429832e-02 -7.89203286e-01 6.04405046e-01
8.75544608e-01 -1.16871250e+00 -8.24727952e-01 6.73826694e-01
1.09649265e+00 -3.41108292e-01 -5.82372129e-01 -4.39122356e-02
3.69074881e-01 -2.19879568e-01 6.77773833e-01 -5.93781233e-01
1.08227767e-01 -2.26649716e-01 -6.45861328e-01 -1.69330192e+00
1.78606331e-01 -1.00101423e+00 4.84501392e-01 1.45035112e+00
1.13926387e+00 -5.50656319e-01 2.71779060e-01 2.41048157e-01
-4.76820767e-01 -1.57174170e-01 -1.51054811e+00 -9.30219293e-01
-1.82793796e-01 -6.24758482e-01 4.80107337e-01 9.61200476e-01
3.50842565e-01 7.66576111e-01 -3.02014560e-01 1.97156876e-01
9.87804905e-02 -1.86821580e-01 6.88355505e-01 -6.63629830e-01
-4.70792979e-01 -7.75408894e-02 -2.33537883e-01 -1.00118566e+00
1.56198293e-01 -1.07965970e+00 5.75598657e-01 -1.16042209e+00
-7.48081356e-02 -2.12017626e-01 2.81589776e-01 5.28396904e-01
1.52327478e-01 2.16429114e-01 2.08829753e-02 3.35685343e-01
-3.89303267e-01 3.89181256e-01 5.78176439e-01 -3.88901263e-01
-2.91443825e-01 1.24608383e-01 -3.50935102e-01 2.12913789e-02
5.22227705e-01 -8.87765586e-01 -2.99690962e-01 -6.16147339e-01
-6.93094879e-02 4.26597923e-01 -1.10695794e-01 -7.73513734e-01
3.65323514e-01 9.68728513e-02 5.80866374e-02 -4.46387589e-01
3.67800146e-01 -8.06317151e-01 3.12491298e-01 -2.15726160e-02
-5.41748047e-01 3.19387496e-01 2.68260509e-01 2.03225404e-01
-4.80095893e-01 -3.04406315e-01 4.41906303e-01 2.41618846e-02
-2.06038207e-02 -1.70873106e-01 -1.00577843e+00 -5.91179878e-02
5.40462911e-01 2.68073492e-02 1.79529209e-02 -7.82082140e-01
-8.61614585e-01 -2.19708011e-02 3.00635904e-01 6.17110431e-01
4.42731827e-01 -1.37538135e+00 -9.99231875e-01 4.13009405e-01
2.35898435e-01 -5.05120099e-01 -2.50121117e-01 8.39479148e-01
-1.55668125e-01 4.44289327e-01 2.67080843e-01 -7.41803885e-01
-1.42197764e+00 2.76160926e-01 2.07164854e-01 1.22597165e-01
-3.91496480e-01 7.09482431e-01 -3.71783882e-01 -7.31899977e-01
3.40505719e-01 -5.18268198e-02 2.08083332e-01 4.66494784e-02
6.14285529e-01 -8.87698159e-02 7.84630299e-01 -9.58778858e-01
-3.72228593e-01 2.10916445e-01 9.46696177e-02 -1.21744454e+00
1.18594790e+00 -5.68562329e-01 -1.26271602e-02 8.30715418e-01
1.22498691e+00 2.69789368e-01 -3.19239438e-01 -3.87687504e-01
2.09474325e-01 -2.88520187e-01 -1.71641096e-01 -8.06368530e-01
-1.37024179e-01 7.59354889e-01 5.45648932e-01 3.64280641e-01
9.21921909e-01 1.25733957e-01 8.80045295e-01 7.51695991e-01
4.93834972e-01 -1.28872585e+00 -1.02724567e-01 8.02665174e-01
1.00279653e+00 -1.07217968e+00 -5.78802705e-01 -1.65612996e-01
-6.58420205e-01 1.00904250e+00 -5.91024570e-02 5.95256448e-01
4.47164148e-01 3.99443835e-01 7.34705091e-01 2.32137486e-01
-1.09683526e+00 -2.82772303e-01 2.05884337e-01 6.67236686e-01
4.62894857e-01 1.53889745e-01 6.15284480e-02 3.40669662e-01
-8.39404643e-01 -3.68891060e-01 4.44622189e-01 9.49634790e-01
-3.67466241e-01 -1.31033528e+00 -6.36986613e-01 2.08957747e-01
-6.99970782e-01 -4.08467740e-01 -8.87800872e-01 6.31501228e-02
-3.22657764e-01 1.55812263e+00 2.02865843e-02 -7.51796901e-01
3.95477623e-01 8.34764838e-01 2.06163093e-01 -1.02664912e+00
-6.97992742e-01 7.86275506e-01 5.91030180e-01 -3.04619163e-01
-2.48431236e-01 -9.01497424e-01 -7.80810297e-01 -4.16801050e-02
-4.29272115e-01 4.54731047e-01 1.01071227e+00 1.14547992e+00
3.82895857e-01 3.28237891e-01 1.01497364e+00 -7.59024143e-01
-6.71299040e-01 -1.39945781e+00 -2.54798442e-01 2.25052983e-01
5.89259028e-01 -2.06726585e-02 -4.37904030e-01 5.14381588e-01] | [14.464219093322754, 7.155285835266113] |
06175ad2-3249-4fc7-9a2e-9a4c8e9944bb | explaining-predictions-from-tree-based | 1907.02582 | null | https://arxiv.org/abs/1907.02582v1 | https://arxiv.org/pdf/1907.02582v1.pdf | Explaining Predictions from Tree-based Boosting Ensembles | Understanding how "black-box" models arrive at their predictions has sparked significant interest from both within and outside the AI community. Our work focuses on doing this by generating local explanations about individual predictions for tree-based ensembles, specifically Gradient Boosting Decision Trees (GBDTs). Given a correctly predicted instance in the training set, we wish to generate a counterfactual explanation for this instance, that is, the minimal perturbation of this instance such that the prediction flips to the opposite class. Most existing methods for counterfactual explanations are (1) model-agnostic, so they do not take into account the structure of the original model, and/or (2) involve building a surrogate model on top of the original model, which is not guaranteed to represent the original model accurately. There exists a method specifically for random forests; we wish to extend this method for GBDTs. This involves accounting for (1) the sequential dependency between trees and (2) training on the negative gradients instead of the original labels. | ['Maarten de Rijke', 'Hinda Haned', 'Ana Lucic'] | 2019-07-04 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 6.12324595e-01 7.47221053e-01 -3.39434683e-01 -4.55609918e-01
-4.91728336e-01 -5.01725256e-01 8.74634206e-01 3.50223295e-02
9.27707404e-02 1.19877660e+00 4.06132519e-01 -9.97021139e-01
9.04283114e-03 -1.01680553e+00 -8.75991940e-01 -7.96018958e-01
6.14946149e-02 5.91593206e-01 1.91450819e-01 -1.56836659e-01
4.18866962e-01 3.53949517e-01 -1.41895044e+00 5.64130425e-01
8.66007745e-01 6.04903400e-01 -1.12016201e-01 8.00033987e-01
-1.24084968e-02 9.46616232e-01 -4.02350098e-01 -5.65957785e-01
2.03616843e-01 -8.12071145e-01 -8.71425033e-01 -4.86411601e-02
1.87581986e-01 -1.20485112e-01 1.55052811e-01 7.92529106e-01
6.33761138e-02 -7.50543922e-02 1.01663446e+00 -1.48866653e+00
-3.80806983e-01 8.83690178e-01 -4.84631062e-01 3.48670669e-02
6.34763539e-02 2.16535807e-01 9.43523109e-01 -3.86317372e-01
5.23406446e-01 1.24566293e+00 7.97246933e-01 8.06546748e-01
-1.67991960e+00 -7.50129819e-01 5.35463095e-01 2.60287881e-01
-8.60895991e-01 -4.05375838e-01 6.97692513e-01 -6.10326171e-01
6.95973933e-01 5.53268850e-01 7.01546311e-01 1.22102499e+00
3.81915092e-01 5.28880537e-01 1.61931109e+00 -7.28294671e-01
6.27596915e-01 2.84617543e-01 1.10190302e-01 3.01625192e-01
3.93302798e-01 6.43030107e-01 -4.74906713e-01 -5.25357902e-01
2.72704959e-01 -1.09606519e-01 -2.81554498e-02 -6.13009989e-01
-1.02363706e+00 1.05037010e+00 5.23668766e-01 7.77238980e-02
-5.88504195e-01 2.74225920e-01 5.42940535e-02 2.26769280e-02
7.47694433e-01 5.00335991e-01 -7.77225554e-01 2.11938560e-01
-1.03751886e+00 7.57426202e-01 8.80450487e-01 3.73821318e-01
8.78381491e-01 -1.39189139e-01 -1.92785889e-01 3.25631678e-01
3.08331013e-01 1.29919991e-01 4.92453635e-01 -8.54531825e-01
4.52516913e-01 4.84232157e-01 5.18106997e-01 -6.18721902e-01
-1.68624550e-01 -4.30126190e-01 -6.28774047e-01 4.71042782e-01
6.10629737e-01 -3.73387814e-01 -1.09138632e+00 2.07561970e+00
5.78996301e-01 1.43283501e-01 6.97773993e-02 7.49299288e-01
9.70774814e-02 5.14782012e-01 3.82465631e-01 -2.29916021e-01
7.08270729e-01 -7.11329699e-01 -1.23451456e-01 -6.49645805e-01
9.73655999e-01 -2.98567742e-01 7.70671725e-01 2.04493284e-01
-6.67777717e-01 -2.10371971e-01 -1.10579813e+00 3.86786789e-01
-3.67849141e-01 -2.44031772e-01 7.82209277e-01 6.67435586e-01
-1.05017114e+00 9.66891825e-01 -6.82803035e-01 -1.25028610e-01
2.27202237e-01 2.79327065e-01 -2.14505475e-02 -1.00526065e-02
-1.19200325e+00 1.21320677e+00 2.78500140e-01 6.66857660e-02
-7.56563842e-01 -5.26403964e-01 -5.56574464e-01 8.96890908e-02
1.88186914e-01 -8.91452372e-01 1.27515435e+00 -1.51571095e+00
-1.05580091e+00 6.82605624e-01 -5.64785242e-01 -7.97991753e-01
8.73106718e-01 1.26011550e-01 -1.89845413e-01 -6.35328889e-01
3.89840931e-01 6.81037009e-01 7.62591481e-01 -1.55843592e+00
-6.51874423e-01 -8.58318627e-01 3.91228460e-02 2.56362706e-01
3.36487710e-01 -2.59214252e-01 7.06151664e-01 -5.32685459e-01
2.41877154e-01 -1.16201019e+00 -6.87000990e-01 -2.89447039e-01
-8.94039392e-01 -2.49445904e-02 6.84517324e-01 -5.48602641e-01
1.19199014e+00 -1.61144650e+00 -3.22923660e-01 2.37806424e-01
1.47911325e-01 -8.77710711e-03 1.75718337e-01 3.42975467e-01
-4.18061763e-01 4.56083447e-01 -4.18076515e-01 -1.12324715e-01
-8.63721073e-02 2.45580420e-01 -9.68533933e-01 1.49054706e-01
-1.24696838e-02 6.23332024e-01 -8.02265763e-01 -2.90113270e-01
5.95506988e-02 -7.62826726e-02 -6.03481352e-01 -4.26549390e-02
-4.80561078e-01 5.07587194e-01 -5.25873721e-01 -2.66363993e-02
5.98514557e-01 -2.83252466e-02 5.96958339e-01 4.49981987e-01
-2.09816962e-01 8.30159068e-01 -1.11370850e+00 6.61628723e-01
-4.33454871e-01 3.54882061e-01 -2.94552147e-01 -1.09177923e+00
7.19598353e-01 3.56764585e-01 3.50107029e-02 -2.14902550e-01
-3.65354687e-01 4.11881328e-01 2.34748259e-01 -6.14834800e-02
1.71479627e-01 -9.21294570e-01 9.54175144e-02 6.59156919e-01
-5.53064942e-01 3.13868560e-02 -1.42071545e-01 -4.04868945e-02
1.02361763e+00 4.30466980e-01 7.03497291e-01 -1.96494877e-01
2.19463557e-01 2.72694677e-01 7.62399018e-01 1.13259876e+00
-3.07033258e-03 5.02893031e-01 4.88765270e-01 -7.39820898e-01
-9.99097288e-01 -1.02222455e+00 -1.05241882e-02 8.86272192e-01
-2.76244581e-01 -7.97266215e-02 -6.04804754e-01 -1.05053258e+00
1.59300834e-01 1.49776101e+00 -1.00505590e+00 -2.05400780e-01
-2.84344018e-01 -8.94532263e-01 7.06545040e-02 4.97144967e-01
3.43640596e-01 -9.89385664e-01 -6.07814908e-01 3.71051461e-01
-3.35974544e-01 -3.97566140e-01 -7.87031557e-03 5.11437356e-01
-1.43863702e+00 -1.13896346e+00 -2.80232161e-01 3.19424272e-02
5.59078515e-01 2.14132592e-02 1.19475567e+00 -2.41542496e-02
3.78167361e-01 -2.68672645e-01 -1.11757413e-01 -8.12405646e-01
-7.89968252e-01 -2.12954786e-02 3.58829610e-02 -1.04004489e-02
4.14369375e-01 -8.50064576e-01 -4.15626347e-01 2.63895810e-01
-4.21116829e-01 5.11569321e-01 4.96940404e-01 8.35370064e-01
3.13518226e-01 -1.76280156e-01 7.92947173e-01 -1.43870914e+00
4.69455868e-01 -6.23737037e-01 -2.95240492e-01 3.48770887e-01
-1.00914216e+00 3.24255913e-01 7.31669188e-01 -2.82082975e-01
-1.15503204e+00 1.16859227e-01 7.90451691e-02 2.63180919e-02
-4.32808369e-01 3.39673936e-01 -1.18050292e-01 4.81899381e-01
9.56923246e-01 2.48074442e-01 -3.46993089e-01 -4.66686994e-01
3.05849582e-01 6.72536016e-01 2.15214893e-01 -5.79212546e-01
6.16429389e-01 3.34497869e-01 5.66792861e-02 -1.93641692e-01
-1.04366183e+00 1.74690425e-01 -5.47529519e-01 -1.37548834e-01
5.38460493e-01 -4.69287425e-01 -3.03882658e-01 3.17595787e-02
-1.07291842e+00 -5.69174767e-01 -4.94841605e-01 5.39629579e-01
-6.70214355e-01 -3.96223664e-02 -6.37364686e-02 -1.32306790e+00
-6.29790127e-02 -8.95662189e-01 7.81318247e-01 4.34870273e-02
-7.52272964e-01 -9.44929242e-01 1.83643639e-01 3.70639741e-01
3.02775979e-01 2.98149705e-01 1.41959238e+00 -9.29661214e-01
-4.99317974e-01 -3.12264860e-01 1.08325705e-01 9.63999480e-02
-1.13932952e-01 -1.66982245e-02 -1.17800224e+00 1.08451970e-01
7.68919736e-02 -1.01854756e-01 8.72113585e-01 4.90657300e-01
1.08421636e+00 -7.84977496e-01 -6.93212807e-01 -2.32363213e-03
1.09303498e+00 2.67210305e-01 5.53001761e-01 4.43005025e-01
3.40950668e-01 8.81838381e-01 3.84386212e-01 3.15465629e-01
5.14167547e-01 7.25323677e-01 5.06018639e-01 -1.06935455e-02
1.40281782e-01 -8.08250010e-01 2.92426854e-01 -8.20840597e-02
-6.30190894e-02 -4.92864400e-02 -9.10056233e-01 7.55394399e-01
-2.13744497e+00 -1.14406884e+00 -3.61486316e-01 2.65163136e+00
7.71710157e-01 2.62725562e-01 2.61424575e-02 2.76733696e-01
7.99444079e-01 1.97995588e-01 -8.75127912e-01 -8.57842684e-01
2.00352684e-01 -5.74801341e-02 3.63268197e-01 6.01432741e-01
-8.46251726e-01 7.04631805e-01 6.46511269e+00 5.03006637e-01
-1.05118990e+00 3.42075247e-03 1.03800309e+00 8.20075944e-02
-6.75404489e-01 8.61210704e-01 -7.60392427e-01 5.32394946e-01
9.80945706e-01 -2.94527799e-01 2.54446387e-01 1.08864665e+00
3.34758908e-01 -2.82119393e-01 -1.31938028e+00 -5.05626164e-02
-4.95089859e-01 -1.31362081e+00 2.49931186e-01 3.63519013e-01
8.58453095e-01 -2.36880064e-01 -2.20235318e-01 3.10390353e-01
9.57940936e-01 -1.10766053e+00 8.96595359e-01 4.15921003e-01
4.29327816e-01 -5.38560569e-01 6.99117720e-01 9.61871505e-01
-5.59980094e-01 -1.73239067e-01 -1.26801819e-01 -6.66664243e-01
-1.29719034e-01 8.48178983e-01 -1.26255047e+00 2.92576581e-01
5.37370265e-01 3.20466250e-01 -3.64622891e-01 7.13482797e-01
-5.43275118e-01 1.08547521e+00 -3.05310220e-01 -1.18565019e-02
8.92034769e-02 -5.47082946e-02 5.54639697e-01 8.35208356e-01
2.81537682e-01 -7.64361471e-02 -3.90194841e-02 9.80366409e-01
2.84729183e-01 -1.64508015e-01 -8.89667153e-01 3.17173511e-01
3.55136871e-01 6.83168828e-01 -4.97303963e-01 -5.00480115e-01
-7.97217526e-03 5.39489508e-01 3.22946340e-01 2.49176279e-01
-5.96830964e-01 2.36911684e-01 5.66038847e-01 3.89829844e-01
2.10177869e-01 3.06337535e-01 -9.23642516e-01 -1.14793646e+00
-6.94517568e-02 -7.58097053e-01 4.23773706e-01 -9.83521879e-01
-1.20552039e+00 4.06473815e-01 2.99869180e-02 -9.25025702e-01
-8.52246344e-01 -2.86349654e-01 -1.02940071e+00 1.19322741e+00
-1.16903806e+00 -1.05973184e+00 2.22863272e-01 2.41657600e-01
1.73785582e-01 4.03128773e-01 1.09290695e+00 -6.58875346e-01
-2.93693513e-01 2.46815696e-01 4.26937565e-02 -3.32051873e-01
4.07572210e-01 -1.38571525e+00 4.69534725e-01 9.72714663e-01
2.08701685e-01 5.75798690e-01 1.24948835e+00 -7.45386541e-01
-5.40063143e-01 -1.08258116e+00 1.53525186e+00 -6.91503227e-01
3.33218068e-01 -1.78329393e-01 -7.72997200e-01 1.05508733e+00
-2.65591562e-01 -3.84588420e-01 6.81839347e-01 6.12341225e-01
-4.55142915e-01 -5.01679778e-02 -1.30307126e+00 9.68547940e-01
9.54553187e-01 -1.37549862e-01 -7.27880597e-01 2.96092302e-01
3.04656148e-01 -1.48558319e-02 -1.42327562e-01 4.22735870e-01
7.33651698e-01 -1.35651803e+00 6.58849001e-01 -1.11661327e+00
8.51470828e-01 -2.80221730e-01 -2.33252689e-01 -1.61101699e+00
-3.72731477e-01 -2.15344116e-01 -1.00807816e-01 9.76333499e-01
8.25141788e-01 -9.07372177e-01 1.11640358e+00 1.08662415e+00
1.42951757e-01 -1.03343976e+00 -9.67237055e-01 -4.41924930e-01
3.33117962e-01 -5.81258655e-01 8.62775505e-01 8.17505181e-01
-3.26967705e-03 4.65587974e-01 -3.62515897e-01 8.45440328e-02
5.93600750e-01 6.61029518e-01 9.25201952e-01 -1.31538975e+00
-3.77701700e-01 -1.90395772e-01 -1.25075206e-01 -7.67446876e-01
3.18080932e-02 -6.91291094e-01 1.23144574e-01 -1.48498416e+00
4.00209099e-01 -6.21455252e-01 -8.16781446e-03 6.57012522e-01
-4.51318443e-01 -2.94526890e-02 2.28711694e-01 3.64504874e-01
-2.47093402e-02 4.21904564e-01 9.14643645e-01 2.49586791e-01
-7.86292925e-02 5.76315999e-01 -1.25424051e+00 9.24720347e-01
8.31856966e-01 -8.99926245e-01 -3.37832093e-01 5.55703714e-02
1.55833974e-01 4.08647686e-01 7.14031518e-01 -6.51251376e-01
-8.35150480e-02 -5.58002889e-01 5.92761993e-01 -1.66470587e-01
1.10605009e-01 -6.36358559e-01 4.83173966e-01 6.18734121e-01
-8.49907517e-01 -1.76514953e-01 -1.33497968e-01 7.66409874e-01
1.01740390e-01 -3.30724388e-01 6.27973795e-01 -2.49842539e-01
-3.27902168e-01 -9.10172984e-02 -5.69839537e-01 -1.43392801e-01
7.92129576e-01 -3.92887145e-01 -1.99536622e-01 -6.67175293e-01
-8.79683137e-01 3.77855264e-02 6.22574151e-01 1.26430959e-01
9.70000029e-02 -1.03573143e+00 -6.38251245e-01 -3.01205050e-02
7.99891055e-02 -1.22233614e-01 8.15990642e-02 5.38197637e-01
1.62006602e-01 5.49016714e-01 1.08381778e-01 -9.82045233e-02
-1.01930404e+00 4.71668303e-01 5.62751532e-01 -8.16617370e-01
-2.03197315e-01 5.78244269e-01 5.48295140e-01 -8.55632484e-01
-4.87463892e-01 -1.80115514e-02 -7.36286193e-02 -2.46720463e-01
2.41163328e-01 2.52229214e-01 -9.62201655e-02 -3.33923787e-01
-3.65883887e-01 -1.27345873e-02 -1.18959501e-01 -3.59518051e-01
1.33987057e+00 5.49369305e-02 2.37005744e-02 6.41068459e-01
6.02860391e-01 -1.21889329e-02 -1.35023487e+00 -1.16675464e-03
2.92331547e-01 -4.60136145e-01 -9.63831469e-02 -1.31740630e+00
-5.11407018e-01 9.51209188e-01 1.48034498e-01 3.65482271e-01
8.92296255e-01 -1.76406592e-01 2.16859818e-01 1.44105926e-01
5.24394870e-01 -7.79086471e-01 -6.78449214e-01 1.44777074e-01
9.29231226e-01 -1.11260331e+00 1.12728633e-01 -1.90289497e-01
-7.44283676e-01 9.72896874e-01 5.62771380e-01 -1.56837866e-01
4.24084604e-01 -1.24038771e-01 -1.03207622e-02 3.94863151e-02
-1.22248161e+00 1.02092378e-01 4.59008254e-02 6.76413536e-01
5.72647154e-01 5.50393224e-01 -3.77862185e-01 7.32655168e-01
-5.00513911e-01 3.60666722e-01 4.58508044e-01 7.20408678e-01
-3.42935622e-01 -1.12326145e+00 -4.95899796e-01 9.49564993e-01
-3.97656947e-01 -2.37416178e-01 -7.52927780e-01 7.01696992e-01
1.76673785e-01 1.06329656e+00 -2.37755880e-01 -4.93994772e-01
2.10053921e-01 6.46881402e-01 1.85755387e-01 -7.47671008e-01
-4.65274513e-01 -3.89523357e-01 3.64104599e-01 -3.75064820e-01
-1.83237731e-01 -8.75855088e-01 -8.34439337e-01 -4.17901963e-01
-3.98236245e-01 4.29056615e-01 6.81836426e-01 1.24192560e+00
4.35723454e-01 6.40995130e-02 7.05758929e-01 -9.01190996e-01
-9.24269080e-01 -1.08732247e+00 -4.81100678e-01 2.38553494e-01
1.92174107e-01 -6.33212745e-01 -6.54719591e-01 -1.88701019e-01] | [8.672211647033691, 5.617580413818359] |
1e82bcd4-adea-45e8-933d-2c1ff2959dfe | parser-free-virtual-try-on-via-distilling | 2103.04559 | null | https://arxiv.org/abs/2103.04559v2 | https://arxiv.org/pdf/2103.04559v2.pdf | Parser-Free Virtual Try-on via Distilling Appearance Flows | Image virtual try-on aims to fit a garment image (target clothes) to a person image. Prior methods are heavily based on human parsing. However, slightly-wrong segmentation results would lead to unrealistic try-on images with large artifacts. Inaccurate parsing misleads parser-based methods to produce visually unrealistic results where artifacts usually occur. A recent pioneering work employed knowledge distillation to reduce the dependency of human parsing, where the try-on images produced by a parser-based method are used as supervisions to train a "student" network without relying on segmentation, making the student mimic the try-on ability of the parser-based model. However, the image quality of the student is bounded by the parser-based model. To address this problem, we propose a novel approach, "teacher-tutor-student" knowledge distillation, which is able to produce highly photo-realistic images without human parsing, possessing several appealing advantages compared to prior arts. (1) Unlike existing work, our approach treats the fake images produced by the parser-based method as "tutor knowledge", where the artifacts can be corrected by real "teacher knowledge", which is extracted from the real person images in a self-supervised way. (2) Other than using real images as supervisions, we formulate knowledge distillation in the try-on problem as distilling the appearance flows between the person image and the garment image, enabling us to find accurate dense correspondences between them to produce high-quality results. (3) Extensive evaluations show large superiority of our method (see Fig. 1). | ['Ping Luo', 'Wei Liu', 'Chongjian Ge', 'Ruimao Zhang', 'Yibing Song', 'Yuying Ge'] | 2021-03-08 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Ge_Parser-Free_Virtual_Try-On_via_Distilling_Appearance_Flows_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Ge_Parser-Free_Virtual_Try-On_via_Distilling_Appearance_Flows_CVPR_2021_paper.pdf | cvpr-2021-1 | ['human-parsing'] | ['computer-vision'] | [ 2.98609614e-01 7.09654212e-01 2.26806968e-01 -2.33405992e-01
-5.34327745e-01 -4.37913626e-01 1.95763290e-01 -5.56768596e-01
-2.11123034e-01 6.64140940e-01 -5.60844481e-01 -2.98313260e-01
5.03980935e-01 -9.40124035e-01 -1.26998711e+00 -5.91637313e-01
7.97582209e-01 3.31320286e-01 3.67829055e-01 -1.01038568e-01
-2.77093425e-02 2.72339195e-01 -1.34713101e+00 1.83064386e-01
1.21040237e+00 5.02171814e-01 2.35831112e-01 5.55299580e-01
-1.71592623e-01 6.27262652e-01 -8.03991854e-01 -1.10559785e+00
5.09563565e-01 -8.72716904e-01 -8.45012903e-01 5.54396570e-01
8.46093416e-01 -4.92888808e-01 -2.80870914e-01 1.66596973e+00
1.94273107e-02 -4.13984545e-02 3.92984807e-01 -1.26479995e+00
-9.44402754e-01 4.35469776e-01 -7.49867201e-01 -4.12305266e-01
4.97825921e-01 4.28377628e-01 4.63891953e-01 -4.50080991e-01
6.40643299e-01 1.42820358e+00 6.08733237e-01 9.17278528e-01
-1.24570560e+00 -7.71538615e-01 2.59358793e-01 6.92423806e-02
-1.25211561e+00 -1.39830917e-01 1.03913581e+00 -2.69453049e-01
2.69798636e-01 3.00687671e-01 1.01905704e+00 1.31429112e+00
-1.77566521e-02 1.01823735e+00 1.36590171e+00 -6.05665743e-01
8.59619211e-03 6.36157274e-01 -2.22862408e-01 1.16167462e+00
2.07986936e-01 2.21885547e-01 -2.31497765e-01 2.71313906e-01
1.32848334e+00 -8.23919997e-02 -5.30656040e-01 -3.48442346e-01
-9.23805535e-01 5.44054091e-01 6.81181550e-01 1.46793947e-01
-2.77461797e-01 9.61610600e-02 -1.15478851e-01 1.50844440e-01
3.94291162e-01 2.78778255e-01 -2.85056829e-01 1.33379892e-01
-1.17617488e+00 7.22810924e-02 6.81458235e-01 1.06699061e+00
7.94268012e-01 2.18978316e-01 1.19389340e-01 5.44363797e-01
3.94639134e-01 6.62715554e-01 2.83015877e-01 -1.04070818e+00
3.49190474e-01 7.40310967e-01 2.86923379e-01 -1.02522886e+00
7.79002681e-02 -3.33918571e-01 -7.78462350e-01 5.53779840e-01
9.76626873e-01 -7.02326968e-02 -1.37689221e+00 1.52402663e+00
5.23012578e-01 3.01646382e-01 1.28895063e-02 1.23330998e+00
8.62137139e-01 4.25821096e-01 -7.22972155e-02 -1.27263576e-01
1.38758528e+00 -1.39847648e+00 -7.10261941e-01 -2.48960420e-01
3.04418087e-01 -8.24299395e-01 1.57568157e+00 6.51869893e-01
-1.35055959e+00 -7.90554762e-01 -9.59719241e-01 4.90212403e-02
-3.33098233e-01 4.15188354e-03 4.35991704e-01 9.63075578e-01
-9.41428304e-01 8.82979095e-01 -6.61221862e-01 -2.76381880e-01
5.98417163e-01 1.27368018e-01 -2.37811118e-01 -1.67318806e-01
-9.20460582e-01 9.23234403e-01 3.15445542e-01 2.76225835e-01
-8.37772489e-01 -6.21709824e-01 -7.79463172e-01 -1.62246108e-01
6.68619752e-01 -8.87191176e-01 1.12582648e+00 -1.88550436e+00
-1.84017193e+00 1.11735475e+00 1.62268758e-01 -7.40494803e-02
1.16578996e+00 -1.91618413e-01 -1.35915428e-01 3.22520614e-01
3.17737982e-02 6.81103110e-01 1.13877380e+00 -1.96903479e+00
-2.55620092e-01 -1.65647835e-01 1.50775984e-01 1.61087319e-01
1.58857897e-01 -2.36290842e-01 -7.82746732e-01 -6.65412605e-01
2.32821867e-01 -8.49140048e-01 -2.41625443e-01 5.49691245e-02
-8.20972681e-01 3.11303705e-01 9.77822602e-01 -9.18284953e-01
6.24584138e-01 -2.00977087e+00 -3.12764756e-02 1.93570718e-01
2.47071058e-01 6.07926965e-01 -1.73097089e-01 -9.14535746e-02
-7.51646459e-02 6.40054271e-02 -4.12149876e-01 -4.55427349e-01
-1.55951828e-01 5.64324081e-01 -1.93105578e-01 3.93936306e-01
7.91913047e-02 1.13235354e+00 -1.13655269e+00 -7.99735606e-01
3.83809507e-01 5.34603059e-01 -2.67450601e-01 5.11531293e-01
-2.16063187e-01 7.39630163e-01 -3.72021347e-01 7.71518409e-01
8.70632887e-01 -2.05343172e-01 1.53123781e-01 -3.73170525e-01
3.44203085e-01 -3.10585618e-01 -9.61996198e-01 1.42528307e+00
-3.47739130e-01 2.59785801e-01 1.77005336e-01 -8.14235568e-01
8.38483334e-01 2.14944452e-01 2.87443884e-02 -7.01086104e-01
3.13301772e-01 7.83833414e-02 -1.81262419e-01 -6.85896873e-01
2.74581224e-01 -2.91192293e-01 2.27970958e-01 8.46961215e-02
-1.08531684e-01 -6.15361392e-01 -8.08250159e-02 1.58840671e-01
6.80863023e-01 7.52169669e-01 -1.23825587e-01 1.27028272e-01
5.00599563e-01 1.43676162e-01 6.28101289e-01 7.75534511e-01
-3.96404058e-01 8.50168765e-01 4.50446814e-01 -4.63343292e-01
-1.16098642e+00 -1.05458164e+00 2.10578457e-01 5.03929138e-01
6.19282782e-01 1.13395257e-02 -1.43726444e+00 -1.23877561e+00
-2.42699862e-01 8.45156133e-01 -5.81273377e-01 -1.66282884e-03
-5.25896609e-01 -3.08864176e-01 6.17311835e-01 3.57271612e-01
1.00816953e+00 -1.05682695e+00 -5.68136513e-01 3.22586372e-02
-3.23143065e-01 -1.26811910e+00 -4.76130962e-01 -2.50265658e-01
-6.95843875e-01 -1.32843113e+00 -1.02814126e+00 -7.61333406e-01
1.21324229e+00 1.78709581e-01 1.11000311e+00 4.38994944e-01
-2.79866178e-02 4.56019223e-01 -2.52008229e-01 -2.21925512e-01
-8.04375708e-01 -5.57844460e-01 -1.39340997e-01 1.40780300e-01
3.76797132e-02 -4.55272347e-01 -6.47774458e-01 4.19392854e-01
-9.07523513e-01 6.33714259e-01 7.87981391e-01 8.18823934e-01
6.43718183e-01 9.50686932e-02 1.71970874e-01 -1.22318041e+00
2.27730379e-01 2.71387219e-01 -6.74826980e-01 4.76798654e-01
-5.32170653e-01 -1.45597532e-01 9.13341641e-01 -7.43275404e-01
-1.28289425e+00 3.40230435e-01 -2.95583773e-02 -9.43477333e-01
-4.42406148e-01 -2.64616132e-01 -4.28178459e-01 -4.03463930e-01
5.10772407e-01 3.26644003e-01 8.02521873e-03 -3.45788568e-01
4.91905689e-01 3.53455096e-01 8.78470421e-01 -6.64201438e-01
1.11247528e+00 4.87167776e-01 -2.08247364e-01 -5.63613951e-01
-1.10914969e+00 7.48876259e-02 -7.26390004e-01 -4.91032600e-01
1.03951478e+00 -6.27501786e-01 -7.29483306e-01 5.82067907e-01
-1.38357818e+00 -5.10683477e-01 -5.30449450e-01 2.78549522e-01
-5.87465525e-01 6.26695693e-01 -6.39079452e-01 -8.61838341e-01
-7.01642260e-02 -1.26389992e+00 1.00525022e+00 6.15418494e-01
1.35990947e-01 -9.75360274e-01 -3.53180200e-01 1.08443344e+00
-8.03177059e-03 4.04702276e-01 6.61294758e-01 -3.27368587e-01
-7.82423377e-01 -7.54619464e-02 -4.88707006e-01 6.85123801e-01
-1.40089050e-01 7.68513232e-02 -1.12661684e+00 -1.66087095e-02
6.80065826e-02 -2.08661020e-01 5.49678028e-01 1.51825458e-01
1.11786449e+00 -4.20523256e-01 -1.40498847e-01 4.76628751e-01
1.38286102e+00 -2.61974195e-03 8.84385705e-01 1.07463636e-01
1.18147695e+00 6.12670124e-01 6.23548269e-01 -3.19250375e-01
3.74080241e-01 5.23334563e-01 5.49463809e-01 -7.22810686e-01
-4.92406398e-01 -7.44081080e-01 4.42194015e-01 6.21181965e-01
-3.05405676e-01 -6.02292120e-02 -6.08014286e-01 4.83046234e-01
-1.94810319e+00 -6.84762836e-01 -6.40679419e-01 2.22123289e+00
9.39709961e-01 1.17295526e-01 2.59805322e-02 -1.36453599e-01
8.19808424e-01 -2.22899511e-01 -4.49557573e-01 -3.98560047e-01
9.32136998e-02 1.21104263e-01 5.07179499e-01 4.80045795e-01
-9.10182774e-01 1.29807198e+00 5.75335121e+00 8.05582225e-01
-9.18648422e-01 1.80984184e-01 7.90081263e-01 3.32196951e-01
-2.48144031e-01 1.46761850e-01 -3.69307667e-01 5.79435825e-01
3.95692050e-01 3.23266953e-01 2.91283697e-01 8.57663751e-01
3.11486363e-01 -4.45849508e-01 -1.06340456e+00 9.71422970e-01
2.14311749e-01 -9.42595601e-01 1.17759213e-01 -4.55276808e-03
9.36397612e-01 -7.38418281e-01 -1.76146664e-02 2.18360871e-01
4.05934751e-01 -9.90848839e-01 8.25883865e-01 6.94163442e-01
6.58391953e-01 -6.52702749e-01 7.88480699e-01 5.71865499e-01
-7.68252194e-01 4.13894504e-01 -2.76757419e-01 9.20561776e-02
2.15053782e-01 6.32514775e-01 -8.33655000e-01 6.42432034e-01
7.29417503e-01 1.61407724e-01 -5.26092529e-01 6.71368897e-01
-7.85366535e-01 6.12691522e-01 -1.18793547e-01 3.16948861e-01
4.83371876e-02 -5.49415052e-01 3.64448100e-01 9.55610454e-01
1.24296881e-02 2.44621351e-01 3.57292116e-01 1.42254412e+00
-1.79838404e-01 -6.82505639e-03 -5.53741872e-01 6.20125346e-02
-8.53051022e-02 1.15077507e+00 -9.96648669e-01 -6.84250653e-01
-2.57266313e-01 1.60523295e+00 2.70985328e-02 5.21628261e-01
-1.02892363e+00 -3.15003574e-01 5.28294034e-02 4.28733885e-01
2.35863462e-01 2.27817222e-02 -3.11316669e-01 -1.27658498e+00
5.48752472e-02 -8.62092853e-01 -9.93559510e-02 -1.24293518e+00
-1.22338736e+00 4.61766332e-01 -1.49001583e-01 -1.05553567e+00
9.30928364e-02 -4.44540262e-01 -6.68394148e-01 8.14527988e-01
-1.53426850e+00 -1.57505870e+00 -4.61654395e-01 6.53998673e-01
5.95074713e-01 3.04155201e-01 6.23026907e-01 2.27449700e-01
-6.34121597e-01 7.34303117e-01 -3.67430478e-01 3.50142777e-01
5.47307372e-01 -1.29101896e+00 1.66596532e-01 1.09563410e+00
1.75382271e-01 2.67089337e-01 8.23973417e-01 -8.35770428e-01
-1.08846724e+00 -9.54320908e-01 5.48143268e-01 -6.39980972e-01
3.75214428e-01 -1.55682504e-01 -1.15208828e+00 6.31989241e-01
2.26353630e-01 3.01436335e-02 1.93221971e-01 -4.14046973e-01
-2.58881658e-01 1.71975717e-01 -1.49209344e+00 5.89397371e-01
8.25277150e-01 -2.58945882e-01 -6.27256453e-01 3.11990440e-01
6.37482166e-01 -7.30340660e-01 -6.22562587e-01 9.13367569e-02
4.67946649e-01 -1.20651698e+00 8.04629624e-01 -2.98166692e-01
6.80884600e-01 -4.77642089e-01 3.89304847e-01 -1.22529423e+00
2.28983328e-01 -6.51189148e-01 -2.70147324e-02 1.19973576e+00
1.79280490e-01 -4.66686010e-01 1.08270967e+00 8.81438673e-01
1.37960225e-01 -5.36676049e-01 -6.70427442e-01 -8.03421974e-01
-4.50613312e-02 -2.61518389e-01 4.25712377e-01 1.14129496e+00
-4.04956818e-01 8.74268040e-02 -5.80301225e-01 3.70754629e-01
9.46481407e-01 1.05559811e-01 1.07742798e+00 -9.60159361e-01
-5.74917793e-01 -2.05221012e-01 -2.27532431e-01 -1.17143762e+00
6.34145141e-02 -4.91836369e-01 2.68304557e-01 -1.33949196e+00
1.89251408e-01 -2.57260650e-01 2.74083257e-01 6.91226602e-01
-3.89254332e-01 5.40211380e-01 4.30503309e-01 1.21879160e-01
-4.04773116e-01 4.17053640e-01 2.07199430e+00 -1.27079651e-01
-1.76522240e-01 1.01149336e-01 -4.62318450e-01 1.18112326e+00
4.78526831e-01 -4.86450762e-01 -3.43220949e-01 -2.51863986e-01
-6.60981834e-02 2.59523541e-01 8.07086468e-01 -7.25216448e-01
7.36379102e-02 -2.68572569e-01 5.24194658e-01 -2.03440860e-01
2.80650228e-01 -9.93288457e-01 1.53785512e-01 4.07282829e-01
7.16302842e-02 -3.65465075e-01 5.42471046e-03 5.08978307e-01
-1.73223000e-02 -4.78851169e-01 1.01087916e+00 -5.06393492e-01
-5.02465129e-01 -5.21325879e-02 -1.32536991e-02 -9.03500840e-02
1.03474617e+00 -5.47054887e-01 -6.05631210e-02 -4.72756654e-01
-8.43770802e-01 2.30916412e-04 8.11101019e-01 1.34073198e-01
7.23092854e-01 -1.08595049e+00 -4.22592252e-01 3.62711251e-01
-2.63298392e-01 5.03380835e-01 2.56912053e-01 7.09985018e-01
-7.41551459e-01 -2.63195336e-01 -8.81738290e-02 -5.48551619e-01
-1.32579255e+00 6.04295015e-01 4.88079190e-01 -1.46504983e-01
-8.43075275e-01 8.24686527e-01 3.88257831e-01 -5.34177542e-01
1.84514999e-01 -2.35306352e-01 3.02680969e-01 -5.18129468e-01
2.06662565e-01 1.35030374e-01 -2.10125640e-01 -5.43341696e-01
-2.64755003e-02 6.04990542e-01 -9.15506929e-02 -1.76541097e-02
9.23992395e-01 -1.36449873e-01 2.05320448e-01 1.19007073e-01
7.30554163e-01 3.75438407e-02 -1.58517444e+00 -7.76367188e-02
-5.05118906e-01 -8.45851779e-01 -1.66910533e-02 -1.04649055e+00
-1.44412124e+00 8.60778630e-01 5.26002645e-01 1.87697083e-01
1.14132845e+00 -6.75383434e-02 1.09635210e+00 1.38219912e-02
3.72567147e-01 -1.21605062e+00 3.08273256e-01 -5.12182824e-02
6.07026160e-01 -1.41638458e+00 -3.02283298e-02 -7.52979815e-01
-9.06537712e-01 1.04927886e+00 9.78159428e-01 -1.19844899e-01
8.45688134e-02 1.22122310e-01 4.87172902e-01 -2.29827374e-01
3.48020941e-02 -2.26790309e-01 2.76117861e-01 7.78642118e-01
-1.45759121e-01 1.33707643e-01 5.94237782e-02 5.77799916e-01
-2.76674360e-01 3.26666012e-02 6.65176630e-01 6.78980172e-01
-1.77086994e-01 -1.20225525e+00 -7.59944856e-01 -3.69775593e-02
-1.19634457e-01 1.37706742e-01 -5.97658098e-01 1.00602472e+00
5.38831472e-01 7.85242677e-01 -1.81502476e-01 -2.82930732e-01
5.20871460e-01 4.11835359e-03 7.79227555e-01 -4.40593302e-01
-7.39679992e-01 1.20586008e-01 -1.67581096e-01 -6.61401331e-01
-4.55312639e-01 -1.16252214e-01 -1.18762827e+00 -4.04354751e-01
-5.03662288e-01 -1.60491951e-02 6.40133440e-01 1.00811851e+00
-8.24763626e-02 6.02388799e-01 4.54025954e-01 -8.78109992e-01
-4.38726217e-01 -5.10661244e-01 -5.52883744e-01 8.18163574e-01
3.69733036e-03 -3.63286078e-01 -3.66990387e-01 3.97686750e-01] | [11.168296813964844, -0.4297358989715576] |
8f9936cb-4277-436a-856b-d82faf4a2509 | snu-ids-at-semeval-2019-task-3-addressing | null | null | https://aclanthology.org/S19-2054 | https://aclanthology.org/S19-2054.pdf | SNU IDS at SemEval-2019 Task 3: Addressing Training-Test Class Distribution Mismatch in Conversational Classification | We present several techniques to tackle the mismatch in class distributions between training and test data in the Contextual Emotion Detection task of SemEval 2019, by extending the existing methods for class imbalance problem. Reducing the distance between the distribution of prediction and ground truth, they consistently show positive effects on the performance. Also we propose a novel neural architecture which utilizes representation of overall context as well as of each utterance. The combination of the methods and the models achieved micro F1 score of about 0.766 on the final evaluation. | ['Sang-goo Lee', 'Sanghwan Bae', 'Jihun Choi'] | 2019-06-01 | null | null | null | semeval-2019-6 | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [ 4.94690947e-02 2.27688393e-03 -1.77371517e-01 -1.08436334e+00
-9.29023325e-01 -1.99092180e-01 1.69114381e-01 2.26110488e-01
-4.63157088e-01 9.78392780e-01 1.27812639e-01 2.10047707e-01
3.16449225e-01 -3.92772883e-01 -5.25802135e-01 -4.89698648e-01
1.17324449e-01 2.67128438e-01 -2.66233057e-01 -2.90453345e-01
3.71083736e-01 -5.00141457e-02 -1.98500121e+00 1.12224424e+00
8.19273770e-01 1.68504417e+00 -3.24050605e-01 7.36615479e-01
5.96924722e-02 9.32343960e-01 -1.17995501e+00 -5.78126788e-01
-6.78739473e-02 -4.86955851e-01 -6.45087361e-01 -3.56905401e-01
5.85709035e-01 -2.06413958e-02 8.32704082e-02 1.04259086e+00
9.54100847e-01 3.16805571e-01 7.64491975e-01 -1.54333472e+00
-4.76166695e-01 4.41158146e-01 -6.71177685e-01 3.92414391e-01
3.36201698e-01 -6.10280037e-01 7.40981102e-01 -8.87261689e-01
3.88558090e-01 1.02853894e+00 8.22715163e-01 7.46148825e-01
-7.49340832e-01 -8.78827095e-01 3.04171294e-01 6.87587082e-01
-1.20133018e+00 -4.17862952e-01 9.21056509e-01 -2.63991445e-01
1.32538450e+00 3.79538745e-01 4.90386754e-01 1.31074095e+00
4.17935133e-01 8.40362072e-01 1.19772077e+00 -6.82286680e-01
2.91745573e-01 4.89464164e-01 4.07746941e-01 1.59747869e-01
-2.30550662e-01 -2.81996932e-02 -7.54593015e-01 -1.78252056e-01
-3.43898565e-01 -3.66895556e-01 -6.49132431e-02 1.34122565e-01
-5.53432465e-01 7.06928492e-01 1.02269448e-01 2.73015767e-01
-4.21617687e-01 -2.00196877e-01 8.71376038e-01 4.94662881e-01
9.26831365e-01 4.24070925e-01 -7.79962480e-01 -1.32381946e-01
-8.83935094e-01 4.12667036e-01 5.59326291e-01 6.06443524e-01
2.01283440e-01 2.59144723e-01 -3.16091478e-01 9.30944145e-01
1.13994956e-01 2.72018433e-01 1.01104236e+00 -6.56625688e-01
4.98961538e-01 5.20634651e-01 1.96044818e-01 -1.31591260e+00
-4.12288368e-01 -5.54917455e-01 -8.47966492e-01 6.13749176e-02
-9.87730399e-02 -4.46476877e-01 -7.71944463e-01 1.97315693e+00
3.91314179e-01 1.31938219e-01 4.91940916e-01 7.82731473e-01
1.05302048e+00 8.74100864e-01 3.29791814e-01 -4.82788593e-01
7.77180791e-01 -1.06069851e+00 -1.12453413e+00 -3.50969732e-01
6.06152833e-01 -8.04569304e-01 9.17749107e-01 7.11149991e-01
-1.07059252e+00 -7.48185337e-01 -1.44711316e+00 3.15771490e-01
-5.12972414e-01 5.04160881e-01 3.61361653e-01 7.71887302e-01
-8.92137289e-01 5.97753942e-01 -2.83031404e-01 1.41799867e-01
3.41153473e-01 4.34594780e-01 -4.03611243e-01 2.98480868e-01
-1.51466703e+00 1.11618221e+00 2.71710247e-01 4.42027479e-01
-6.04975343e-01 -4.63679552e-01 -7.04427481e-01 -4.19191923e-03
-3.54917556e-01 7.50549734e-02 1.39654028e+00 -1.77796400e+00
-1.50065958e+00 9.47252393e-01 -1.27585515e-01 -4.37221020e-01
2.73909122e-01 -5.16497254e-01 -8.18516433e-01 -4.74314988e-01
-2.91353464e-01 7.02854633e-01 4.57896829e-01 -8.49591672e-01
-7.50003576e-01 -2.90956736e-01 -4.67945665e-01 4.60534632e-01
-1.03508249e-01 8.08958262e-02 3.37995678e-01 -3.90379637e-01
-6.43968210e-02 -6.55524194e-01 -1.84436277e-01 -8.55350852e-01
-2.09017754e-01 -5.04361331e-01 7.96648264e-01 -8.71863246e-01
1.29420304e+00 -2.11581659e+00 -8.99527892e-02 4.30828147e-02
-3.62109125e-01 3.85398835e-01 -1.46893859e-02 9.03456062e-02
-8.14262688e-01 1.16959646e-01 2.59977039e-02 -2.53900021e-01
1.90270513e-01 -8.41715485e-02 -5.20938993e-01 3.62115353e-01
2.35293999e-01 5.80824137e-01 -4.13380653e-01 -1.53864264e-01
-4.40768749e-02 3.70070249e-01 -5.48223794e-01 5.83189607e-01
-1.49385463e-02 3.39470208e-02 1.08349599e-01 4.52063262e-01
7.71164536e-01 4.82330173e-01 1.87694982e-01 -1.95411742e-01
4.88322020e-01 4.33242172e-01 -1.08498454e+00 1.41357338e+00
-1.95635095e-01 8.46619964e-01 -2.45403543e-01 -1.37064779e+00
1.28512800e+00 7.23897278e-01 2.16075808e-01 -9.93751585e-01
4.32447731e-01 -2.59267166e-02 3.72215472e-02 -6.18305981e-01
4.64720488e-01 -6.54166520e-01 -3.63527924e-01 2.16189072e-01
2.78833538e-01 4.17825393e-03 -1.01112761e-01 -3.35337728e-01
5.72885752e-01 2.20306277e-01 2.76724339e-01 -1.39088720e-01
4.53712702e-01 -1.61693379e-01 1.03765917e+00 5.62372446e-01
-8.01142037e-01 4.61762547e-01 6.38016045e-01 -6.57811701e-01
-7.42292881e-01 -6.66938484e-01 -3.49495187e-02 1.28492963e+00
-7.87530392e-02 -9.17130709e-02 -9.72553670e-01 -9.57927763e-01
-4.03085291e-01 1.04686105e+00 -9.73150313e-01 -5.77418625e-01
-2.18335271e-01 -1.11079907e+00 6.02624297e-01 6.96016908e-01
3.25019121e-01 -1.31634724e+00 -5.54884017e-01 -2.59164609e-02
-6.31152093e-01 -9.14529264e-01 7.55141079e-02 6.70888722e-01
-6.44922197e-01 -7.74076045e-01 -1.61681080e-03 -8.78313482e-01
1.84498817e-01 -5.77980518e-01 1.75487876e+00 -1.65398657e-01
-1.46716669e-01 -1.82785720e-01 -3.75935435e-01 -9.09034014e-01
-2.57308930e-01 -1.57169163e-01 -3.65982950e-02 -8.56509730e-02
9.19661880e-01 -3.50467682e-01 -3.19717407e-01 2.28346512e-01
-5.14667213e-01 -1.44483253e-01 1.96668178e-01 8.94419014e-01
5.93254924e-01 1.58174455e-01 1.19844353e+00 -8.67847681e-01
7.99392283e-01 -5.97089171e-01 -1.04470327e-01 2.21575648e-01
-6.99249327e-01 -3.15667540e-01 4.05466765e-01 -3.92547309e-01
-1.22449923e+00 -1.48676276e-01 -4.37354237e-01 -1.65017545e-01
-4.95630294e-01 3.37710053e-01 -3.59855652e-01 4.83071089e-01
7.77517200e-01 -3.27490330e-01 -4.67274398e-01 -1.68949768e-01
-5.47685251e-02 1.01385403e+00 4.89238530e-01 -3.93081963e-01
-3.90825778e-01 -2.78077245e-01 -3.95332128e-01 -1.84985071e-01
-1.14536881e+00 -1.08535074e-01 -4.06039149e-01 -2.34426871e-01
7.47147620e-01 -9.82845962e-01 -4.95907754e-01 5.52237689e-01
-1.18603837e+00 -1.46089002e-01 -2.10868672e-01 5.11805892e-01
-5.29140890e-01 -2.20092565e-01 -4.29146886e-01 -1.03617227e+00
-5.47411501e-01 -9.09276307e-01 8.28874111e-01 2.58368939e-01
-5.19593358e-01 -7.12478936e-01 4.23054099e-01 2.94938177e-01
2.81601399e-01 6.37991548e-01 1.05359352e+00 -1.19408298e+00
6.61832869e-01 -7.45223284e-01 1.55500770e-01 9.26607251e-01
-1.31019741e-01 4.27099615e-02 -1.54356647e+00 -6.36940747e-02
4.00124460e-01 -8.54375601e-01 6.15189075e-01 3.90793741e-01
1.33348656e+00 -3.02505028e-02 4.48005982e-02 6.80691600e-02
1.41248751e+00 6.48780942e-01 8.17810416e-01 7.13694915e-02
1.28619477e-01 9.45506096e-01 9.73065436e-01 3.90779525e-01
1.05746657e-01 5.66161513e-01 4.37151849e-01 -9.59833153e-03
7.40190744e-02 1.98134612e-02 3.81033570e-01 1.00012469e+00
2.57759988e-01 -4.43102270e-01 -7.62834311e-01 5.53145587e-01
-1.80714917e+00 -1.02104974e+00 -6.42499840e-03 2.09064341e+00
8.63919258e-01 3.04354317e-02 -1.39850602e-01 5.34035265e-01
9.56135571e-01 8.88278708e-02 -4.80206758e-01 -1.29682577e+00
-2.86261857e-01 1.78865865e-01 -1.22975275e-01 5.68458259e-01
-1.27392781e+00 6.74544454e-01 7.86793756e+00 1.03084314e+00
-1.14460671e+00 2.18614906e-01 1.38138938e+00 -4.69833881e-01
5.61030954e-02 -6.73792541e-01 -6.40376270e-01 5.62644482e-01
1.44883120e+00 4.63632755e-02 2.29523201e-02 1.05009604e+00
-1.92800522e-01 -2.09423244e-01 -1.08518338e+00 1.12745965e+00
4.90621507e-01 -6.76669419e-01 -3.20082098e-01 -6.31475925e-01
1.06269395e+00 -1.45598337e-01 1.70557380e-01 9.29909587e-01
-9.61312056e-02 -1.37852705e+00 7.04780996e-01 3.84047985e-01
4.54895824e-01 -1.27015984e+00 1.42226076e+00 2.92793930e-01
-4.62016076e-01 -5.51046617e-02 -4.74190623e-01 -4.66156036e-01
-4.22131568e-01 6.20625317e-01 -7.61044264e-01 3.10153991e-01
9.09380615e-01 4.07062173e-01 -4.11306053e-01 6.15811408e-01
-1.48463696e-01 8.20655048e-01 -7.16269463e-02 -1.72636464e-01
-6.27589971e-02 3.78836662e-01 7.45926127e-02 1.48322487e+00
8.36432502e-02 -1.85619250e-01 4.31240387e-02 2.50628054e-01
-9.15283561e-02 4.93459314e-01 -5.01276970e-01 2.86887765e-01
1.95080176e-01 1.10415912e+00 -2.80229032e-01 -5.16336560e-01
-7.68787786e-02 7.82845914e-01 6.32155538e-01 3.51308197e-01
-1.13053644e+00 -5.74649036e-01 4.91218776e-01 -5.32027602e-01
1.02827214e-01 7.53243208e-01 -6.26497567e-01 -8.96789074e-01
3.35064948e-01 -1.20210361e+00 4.31658894e-01 -8.67480993e-01
-1.35077941e+00 1.03634059e+00 -1.75760955e-01 -1.14045823e+00
-5.39970636e-01 -6.02598011e-01 -6.43353462e-01 7.86519468e-01
-1.21480465e+00 -5.78835607e-01 -1.82411835e-01 2.54004210e-01
4.80620950e-01 -3.29597831e-01 1.24008393e+00 5.48228264e-01
-6.39473557e-01 9.65146124e-01 2.78294217e-02 -1.78659372e-02
1.04674518e+00 -1.35824287e+00 -7.91191310e-02 5.96221089e-01
-5.57786822e-02 -6.34529665e-02 9.79086697e-01 -3.28923345e-01
-5.01611233e-01 -9.18046772e-01 1.15848625e+00 -3.82598162e-01
-9.41776335e-02 -2.46907786e-01 -8.46558034e-01 4.83396202e-01
6.54417217e-01 -9.90897268e-02 1.21033061e+00 4.32405293e-01
-4.14128155e-01 -4.11886692e-01 -1.57682598e+00 8.79140347e-02
3.90530854e-01 -4.39112991e-01 -6.51289165e-01 1.31397858e-01
5.52895069e-01 -6.09793782e-01 -6.75517917e-01 8.16110313e-01
7.04417765e-01 -1.36489034e+00 5.02809405e-01 -1.08741975e+00
9.13223505e-01 8.96738768e-02 -8.81005466e-01 -1.54341483e+00
-5.87482052e-03 2.58308332e-02 -7.23896548e-02 1.42808700e+00
6.72772467e-01 -1.54699937e-01 9.45383668e-01 6.62548721e-01
-1.91106256e-02 -1.22891748e+00 -1.03961539e+00 -1.13359772e-01
2.32060939e-01 -6.24557555e-01 5.99054635e-01 9.81073499e-01
1.49645492e-01 5.17883599e-01 -6.37935281e-01 -1.35976627e-01
1.68155245e-02 8.91084746e-02 3.48395020e-01 -9.45778728e-01
-2.11190537e-01 -2.34903589e-01 -7.70906329e-01 -2.27195412e-01
5.31473279e-01 -6.77878916e-01 4.12335366e-01 -9.52986240e-01
1.98976442e-01 -3.60621139e-02 -9.53472018e-01 2.96977520e-01
-3.03566188e-01 5.36092997e-01 -1.33087978e-01 -4.76154268e-01
-8.19306254e-01 6.37932599e-01 5.96917272e-01 -1.19348943e-01
-5.08595444e-02 9.98341385e-03 -7.95656145e-01 9.07415509e-01
9.95522499e-01 -6.72945142e-01 -4.98743296e-01 -2.78814197e-01
4.26283687e-01 -2.43463833e-02 -3.83160934e-02 -1.14582074e+00
-2.41067171e-01 -8.66644159e-02 8.81362796e-01 -8.15627038e-01
3.26157510e-01 -8.22577596e-01 -3.64366882e-02 7.01011717e-02
-9.52197075e-01 3.62090647e-01 5.59879005e-01 3.94738853e-01
-8.03413987e-01 -4.53942060e-01 9.14364636e-01 1.47610456e-01
-5.04966617e-01 -2.16463238e-01 -4.10620630e-01 1.89173356e-01
1.01804030e+00 1.37102589e-01 -1.99995622e-01 -4.94630665e-01
-9.75824893e-01 -1.57626662e-02 -1.63527757e-01 5.97411156e-01
4.23579335e-01 -1.59811664e+00 -7.22605288e-01 1.94295913e-01
-4.71828580e-02 -3.85621130e-01 4.48185325e-01 6.33199513e-01
-1.23848610e-01 2.44367808e-01 -4.70278114e-01 -4.38298941e-01
-1.58976078e+00 3.83040130e-01 7.52152562e-01 -4.33294296e-01
3.73242497e-01 1.21270061e+00 1.20906055e-01 -8.43598247e-01
5.78364372e-01 -2.14478433e-01 -4.64038312e-01 2.73534060e-01
7.28300512e-01 3.49814743e-01 5.55482805e-01 -6.87258303e-01
-4.73455012e-01 2.38552019e-01 -1.22513287e-01 -1.00056209e-01
1.41396749e+00 -7.73868710e-02 -1.74958974e-01 9.75377381e-01
1.30069315e+00 -2.23710209e-01 -7.62027264e-01 1.66815832e-01
-4.80296351e-02 -3.14446688e-01 1.60488680e-01 -1.59079397e+00
-1.01710725e+00 1.14854097e+00 1.32940328e+00 1.90376997e-01
1.49832010e+00 -5.34500778e-01 4.76465076e-01 3.87997508e-01
-1.82414040e-01 -1.55119801e+00 -6.86947703e-02 4.43543702e-01
8.78441989e-01 -1.48907757e+00 -1.35073334e-01 -9.97066349e-02
-9.42779899e-01 8.76567900e-01 1.00108433e+00 -4.38467056e-01
5.63682199e-01 2.99909204e-01 5.02769470e-01 9.59597602e-02
-1.42010105e+00 5.13359785e-01 4.28742170e-01 7.77268231e-01
8.90655518e-01 1.97084606e-01 -4.30716425e-01 1.43338227e+00
-3.07695359e-01 -8.89051929e-02 2.35812068e-01 6.98305964e-01
-3.95438999e-01 -8.23908448e-01 -2.08745360e-01 2.93431938e-01
-1.06011724e+00 -4.25220989e-02 -5.60834467e-01 3.92227739e-01
5.79459310e-01 1.16726387e+00 2.05615535e-01 -7.34285593e-01
5.38364768e-01 6.23022258e-01 1.77324131e-01 -1.92051291e-01
-8.78443658e-01 -1.59640372e-01 3.40943128e-01 -6.19520962e-01
-5.93975782e-01 -4.06157881e-01 -9.22924042e-01 6.13175742e-02
-4.49405134e-01 4.47977632e-01 8.89174819e-01 1.00113082e+00
5.06211758e-01 6.90810144e-01 9.28040564e-01 -7.79920340e-01
-3.38039160e-01 -1.41257465e+00 -6.11350238e-01 5.23414850e-01
2.83074081e-01 -5.81483424e-01 -4.75302428e-01 -9.26761702e-03] | [13.51669979095459, 5.712762832641602] |
05c54d65-141f-40de-82ba-c058973ef11d | deep-neural-network-based-on-f-neurons-and | null | null | https://www.researchsquare.com/article/rs-2032768/v1 | https://assets.researchsquare.com/files/rs-2032768/v1_covered.pdf?c=1667512127 | Deep neural network based on F-neurons and its learning | Artificial neural networks are widely used in data processing and Data Mining. In contrast to traditional artificial neural networks, deep neural networks employ many artificial neuron blocks and more than three layers. An increase in the number of neuron blocks and layers improves the approximation capabilities but leads to the effects of exploding and vanishing gradients. Deep neural networks often employ piece-wise activation functions like ReLU to overcome the effects of exploding and vanishing gradients. We propose F-neuron as an adaptive alternative to piece-wise activation functions. Like ReLU, F-neuron does not suffer from the effects of exploding and vanishing gradient. F-neuron changes its form during the training, can approximate any currently used function, and synthesize new task-specific functions. We show that F-neuron can synthesize new task-specific functions and achieve higher approximation quality in existing neural network architectures. We evaluate the performance of the F-neuron on two image classification datasets (Fashion-MNIST and CIFAR-10) in two different architectures (LeNet-5 and KerasNet). | ['Serhii Kostiuk', 'Yevgeniy Bodyanskiy'] | 2022-11-03 | null | null | null | research-square-pre-print-2022-11 | ['activation-function-synthesis', 'architecture-search'] | ['methodology', 'methodology'] | [-2.88876861e-01 -2.16156945e-01 5.05674668e-02 -6.36380494e-01
1.80620641e-01 -1.18694127e-01 1.73868388e-01 -2.48944256e-02
-9.34471130e-01 5.40380776e-01 -2.09954426e-01 -3.05807859e-01
1.05510220e-01 -9.19058800e-01 -9.57697809e-01 -3.76395524e-01
-1.12071007e-01 1.31258681e-01 4.05399263e-01 -4.05691952e-01
-4.78305444e-02 6.97290897e-01 -1.36071455e+00 8.02635431e-01
7.61229873e-01 1.33453846e+00 1.44495100e-01 5.54302454e-01
-2.80894935e-01 1.05608964e+00 -8.47919166e-01 -3.54316145e-01
4.72778738e-01 -1.12958729e-01 -6.32420301e-01 -1.65419996e-01
3.17367613e-01 -3.63213956e-01 -6.30491018e-01 9.31303084e-01
3.77604067e-01 2.40602702e-01 5.30555248e-01 -1.22110260e+00
-9.67102885e-01 8.04440200e-01 -2.34654024e-01 4.12712902e-01
-4.65152234e-01 2.42127612e-01 4.94118094e-01 -9.58705187e-01
1.52719960e-01 1.49667275e+00 1.25230837e+00 7.38495529e-01
-1.03814888e+00 -5.35318136e-01 3.37202311e-01 8.59098062e-02
-1.42989194e+00 -2.27605149e-01 4.64086324e-01 5.92027092e-03
1.28874373e+00 1.05720602e-01 6.76570654e-01 8.64825845e-01
2.99971193e-01 1.02860594e+00 5.78324199e-01 -2.72313029e-01
3.05341691e-01 1.57649055e-01 4.90810007e-01 6.64608240e-01
2.51975447e-01 -1.46773174e-01 -8.35447833e-02 2.20166191e-01
1.18876636e+00 4.08869892e-01 -2.91787475e-01 1.34840772e-01
-8.88598919e-01 6.71074271e-01 1.24511099e+00 5.69507122e-01
-7.29318440e-01 7.02037930e-01 5.56116402e-01 6.17294610e-01
2.02170491e-01 8.03663135e-01 -5.77172995e-01 1.42851740e-01
-7.41497695e-01 2.74221480e-01 5.51825047e-01 7.94077575e-01
7.69288421e-01 7.94353843e-01 -2.39181831e-01 1.11951041e+00
-1.78543255e-01 1.84973069e-02 1.08520520e+00 -9.06683266e-01
1.43542752e-01 7.44450688e-01 -2.62239307e-01 -8.13666046e-01
-5.80636084e-01 -7.82737732e-01 -1.16947389e+00 3.14064533e-01
5.61690569e-01 -1.85605541e-01 -1.21486223e+00 1.56643939e+00
-1.45962566e-01 9.48527381e-02 9.00481939e-02 8.97153735e-01
9.72071946e-01 8.05304646e-01 -2.31858063e-02 2.77579010e-01
1.21491539e+00 -1.36590827e+00 -5.95376551e-01 -3.91055733e-01
5.95143080e-01 -2.73334712e-01 1.29706442e+00 5.26670575e-01
-1.41079593e+00 -1.04086077e+00 -1.08474767e+00 -3.48489046e-01
-6.95145071e-01 1.51844352e-01 7.68209040e-01 4.44687545e-01
-1.20961046e+00 8.63396823e-01 -7.03138590e-01 3.04922312e-02
7.02247143e-01 5.82756639e-01 -2.27097839e-01 -3.57577251e-03
-1.15759230e+00 8.28940690e-01 7.11565793e-01 2.02460840e-01
-5.39271712e-01 -9.01144922e-01 -6.58440650e-01 6.21347129e-01
-6.69706464e-02 -5.34084201e-01 1.38593364e+00 -1.56079912e+00
-1.40289187e+00 2.15788320e-01 1.66704193e-01 -1.11413693e+00
3.06861460e-01 -1.98612303e-01 -4.45470095e-01 -1.32256597e-01
-5.98105550e-01 1.12437904e+00 8.21791470e-01 -8.47034514e-01
-5.81524372e-01 -2.02477351e-01 2.09544927e-01 -1.32252067e-01
-9.04443979e-01 -3.85608554e-01 -1.81783497e-01 -9.58506763e-01
2.01693121e-02 -6.37480974e-01 -3.47903162e-01 2.68762439e-01
2.70774495e-02 -3.19997847e-01 8.06025982e-01 -2.86472201e-01
1.32669890e+00 -2.24074101e+00 -3.17330092e-01 8.06428865e-02
2.64737427e-01 6.20785773e-01 -4.23650622e-01 -3.04748297e-01
-1.19547233e-01 -4.50545512e-02 -1.40346020e-01 -2.81444140e-04
-2.99173236e-01 5.83179176e-01 2.35597398e-02 3.08511071e-02
3.61332238e-01 1.15519178e+00 -6.66997731e-01 1.75414886e-02
-1.37313763e-02 6.58286631e-01 -8.19014907e-01 -2.54424155e-01
-2.76183873e-01 -3.57217818e-01 -1.29639298e-01 4.24600303e-01
5.64897478e-01 -1.61533520e-01 -2.80276835e-02 -2.78927743e-01
2.11796016e-02 1.34749472e-01 -8.84052932e-01 1.38296342e+00
-6.09928370e-01 7.52604604e-01 -8.51925388e-02 -1.35392702e+00
1.26068056e+00 1.75537258e-01 1.16558753e-01 -9.71868336e-01
3.58729988e-01 4.11949664e-01 3.21775049e-01 -1.59677595e-01
5.43693483e-01 1.31856859e-01 9.75162387e-02 1.06311187e-01
2.81101972e-01 4.84442145e-01 2.83700913e-01 -2.16289729e-01
1.02948046e+00 -3.01114708e-01 -8.86251777e-02 -4.69327897e-01
4.68988657e-01 -1.34073034e-01 5.71047306e-01 7.41744280e-01
2.10080370e-02 2.55655050e-01 4.07050937e-01 -1.33178139e+00
-1.23506916e+00 -7.61618197e-01 -1.30332112e-01 1.23869646e+00
-3.26246738e-01 -1.74833834e-01 -8.66670668e-01 -5.48887551e-01
2.97244817e-01 6.96969390e-01 -6.15087450e-01 -4.85171109e-01
-8.61131787e-01 -8.02656054e-01 8.44331920e-01 1.05111361e+00
9.28576648e-01 -1.28382683e+00 -7.14322507e-01 4.81180727e-01
2.75817454e-01 -9.69030917e-01 -4.34065670e-01 5.14769435e-01
-1.19694924e+00 -5.73795438e-01 -9.30112302e-01 -1.05744433e+00
9.40751433e-01 1.13762245e-01 1.15935624e+00 1.93777487e-01
-1.18704826e-01 -2.94504583e-01 -9.51821171e-03 -5.36387801e-01
-4.59979892e-01 2.05231860e-01 2.18562558e-01 -2.57811565e-02
4.86397445e-01 -5.64433634e-01 -6.27279043e-01 3.31275046e-01
-9.34442937e-01 -6.18784279e-02 6.97461247e-01 1.13070071e+00
5.79882324e-01 2.96821088e-01 7.78777719e-01 -8.93735051e-01
9.23777461e-01 -3.01625043e-01 -5.17477155e-01 8.02161917e-03
-6.45850420e-01 3.55743617e-01 1.29391599e+00 -9.97456551e-01
-7.66083896e-01 -1.03565179e-01 -2.47597814e-01 -6.88281000e-01
1.72436625e-01 5.24867237e-01 2.99486399e-01 -1.22412361e-01
1.12111855e+00 1.37477532e-01 -3.99832316e-02 -7.38889277e-01
9.27364230e-02 5.66932797e-01 7.52733409e-01 -2.67446905e-01
1.29161268e-01 8.92480165e-02 -2.75812030e-01 -7.12841690e-01
-4.76288170e-01 -1.30929183e-02 -2.40670726e-01 -2.26664040e-02
4.97655869e-01 -8.18043053e-01 -8.54870558e-01 8.17552865e-01
-1.17188764e+00 -5.85904360e-01 -6.43982649e-01 4.73055691e-01
-1.34955170e-02 -1.76641017e-01 -1.11755967e+00 -4.31920320e-01
-6.34535909e-01 -1.13303983e+00 1.94326311e-01 2.84157336e-01
-1.03959791e-01 -9.11400855e-01 -5.77733517e-01 -3.18750888e-01
8.42156410e-01 1.31325505e-03 9.53009963e-01 -4.80993569e-01
-1.16482422e-01 -3.35750967e-01 -3.17418069e-01 9.49051023e-01
-1.14719376e-01 -2.31306076e-01 -9.36626673e-01 -1.89822987e-01
1.14958575e-02 -3.21090400e-01 1.24340904e+00 6.02509916e-01
1.70299089e+00 -7.43426681e-01 9.72154737e-02 8.69201779e-01
1.33404195e+00 4.49258119e-01 8.94253016e-01 5.24192870e-01
5.47203124e-01 1.81100667e-01 -1.69979453e-01 1.17265530e-01
2.36753181e-01 3.14273596e-01 3.14755291e-01 -3.01637322e-01
-1.68360651e-01 4.85637709e-02 3.16180080e-01 7.13419735e-01
-2.52083302e-01 -1.38836101e-01 -8.45937788e-01 4.51735407e-01
-1.78022134e+00 -7.38175213e-01 -1.11042932e-01 1.84563959e+00
7.98187315e-01 5.42314231e-01 7.90140033e-02 3.99060309e-01
3.66053164e-01 -2.90251911e-01 -7.79646397e-01 -9.30953979e-01
-2.17897356e-01 3.39649141e-01 5.37172377e-01 1.91522390e-01
-7.80718386e-01 9.15670574e-01 6.49054766e+00 7.26061702e-01
-1.57042146e+00 1.13253100e-02 1.02087939e+00 -3.13156545e-01
6.55466225e-03 -7.21827626e-01 -7.61421561e-01 4.30230469e-01
1.06006992e+00 4.20103967e-01 5.76650977e-01 1.23192596e+00
-2.50453979e-01 3.03797871e-01 -1.11036038e+00 9.67704892e-01
-2.43704870e-01 -1.59802365e+00 3.59046221e-01 -2.52657622e-01
6.25062346e-01 1.68292746e-01 1.58692166e-01 8.69380891e-01
1.62133098e-01 -1.14537311e+00 8.34393144e-01 3.87299389e-01
6.18014336e-01 -9.71389771e-01 8.20510268e-01 1.99333400e-01
-9.34163928e-01 -6.13011479e-01 -8.66720498e-01 -3.80356044e-01
-3.38558614e-01 4.34444100e-01 -7.05125749e-01 -1.62669167e-01
1.09633303e+00 3.71860862e-01 -4.56315011e-01 9.06687915e-01
8.27208459e-02 5.13782084e-01 -4.65361655e-01 -3.54405105e-01
7.36026645e-01 1.05974056e-01 8.28865692e-02 1.23908985e+00
2.42002636e-01 -9.01556015e-02 -3.78441401e-02 9.16011572e-01
-5.06727040e-01 -1.75979972e-01 -2.36145690e-01 -9.30682477e-03
3.60789239e-01 1.05832887e+00 -5.76370716e-01 -5.98402202e-01
-4.28993851e-01 1.03075564e+00 5.27539968e-01 5.97505808e-01
-7.38968551e-01 -5.77502370e-01 6.83834314e-01 5.30276895e-01
4.97079968e-01 -8.44968408e-02 -5.49568415e-01 -8.14253688e-01
-8.35948065e-02 -9.12915051e-01 2.44947150e-01 -7.35189199e-01
-8.99428010e-01 1.14010525e+00 -4.19964552e-01 -8.50383401e-01
-1.26116171e-01 -1.05520570e+00 -6.33157134e-01 7.54363418e-01
-1.40210164e+00 -7.97505021e-01 -2.69995391e-01 6.33963227e-01
7.22553194e-01 -5.77688694e-01 7.60115981e-01 5.91498733e-01
-7.10783184e-01 9.71471488e-01 1.60925925e-01 4.31647420e-01
2.54902065e-01 -1.01617622e+00 7.68789589e-01 5.16919494e-01
-4.65630852e-02 6.25166714e-01 3.68570745e-01 1.52911781e-03
-1.14429021e+00 -1.30699694e+00 5.99768221e-01 2.25377634e-01
3.00864935e-01 -4.80427176e-01 -1.22629499e+00 7.12465525e-01
-8.61846656e-03 5.11126518e-01 2.70292401e-01 -7.86428079e-02
-4.87559944e-01 -6.40614629e-01 -1.32563102e+00 5.92898190e-01
8.09027910e-01 1.81548838e-02 -2.54578203e-01 -3.24151665e-02
7.48796999e-01 -1.50163069e-01 -8.51085246e-01 4.49671537e-01
6.96706831e-01 -9.16313469e-01 1.07932830e+00 -7.86762953e-01
4.92397010e-01 -2.67365780e-02 -1.50982542e-02 -1.42230904e+00
-8.22549939e-01 -3.42204511e-01 -4.25296783e-01 6.61889195e-01
5.55462241e-01 -8.39372337e-01 8.34958732e-01 4.69164073e-01
-4.02790934e-01 -1.17905068e+00 -6.20385647e-01 -9.34487998e-01
2.67917097e-01 -1.71463087e-01 8.35778296e-01 7.83726394e-01
-3.68559569e-01 4.06237870e-01 -1.29000857e-01 -3.26554924e-01
1.48889914e-01 -3.93128365e-01 4.47732180e-01 -1.23511803e+00
-3.36389959e-01 -7.73330212e-01 -4.11331922e-01 -1.22355449e+00
-1.28128216e-01 -7.75641739e-01 -1.76816240e-01 -1.31466186e+00
-5.04107714e-01 -7.70838201e-01 -7.91405439e-01 9.33851063e-01
7.85656348e-02 4.99013275e-01 2.14987144e-01 -2.89623309e-02
-1.86215162e-01 3.34888250e-01 1.34167552e+00 -2.98792958e-01
-4.13531244e-01 -1.18482718e-02 -6.59580231e-01 9.27049398e-01
8.86423647e-01 -2.78540552e-01 -4.07298476e-01 -8.96917582e-01
1.99727654e-01 -3.55076820e-01 2.78972119e-01 -1.23149657e+00
2.79706478e-01 2.05842108e-01 9.26313579e-01 -3.11220139e-01
2.23340914e-01 -8.21864069e-01 -9.39689502e-02 8.99914443e-01
-2.89815515e-01 4.07852232e-01 6.20085359e-01 3.18136252e-02
-3.22066784e-01 -4.57912624e-01 1.03382730e+00 -4.42318022e-01
-8.43909621e-01 2.48364553e-01 -5.33793986e-01 -3.40477616e-01
7.39172935e-01 -4.56259012e-01 -2.56604403e-01 -1.03249386e-01
-4.45647985e-01 2.29878843e-01 -6.53133094e-02 5.07820368e-01
8.37110341e-01 -1.49224341e+00 -7.59775639e-01 5.17912805e-01
-2.91627824e-01 2.22281098e-01 1.32978871e-01 5.14794528e-01
-9.48502302e-01 3.39330554e-01 -6.58459127e-01 -3.23975623e-01
-7.80136406e-01 7.21422613e-01 8.02589834e-01 -3.26214552e-01
-5.17897248e-01 1.17652547e+00 3.48916471e-01 -4.24480796e-01
5.23718178e-01 -8.87263775e-01 -9.16180462e-02 -2.91699976e-01
7.48595476e-01 4.93599594e-01 1.24056347e-01 -3.22726890e-02
-1.32638827e-01 -5.33787981e-02 -4.43582773e-01 4.41636860e-01
1.46579349e+00 4.55884188e-01 1.95282772e-02 3.06045949e-01
1.13442123e+00 -7.92293787e-01 -1.10841048e+00 -3.62411141e-01
-1.66828528e-01 1.66112229e-01 4.76439714e-01 -8.67037237e-01
-1.58398974e+00 1.02344298e+00 5.55308521e-01 1.35615006e-01
1.32497561e+00 -6.88579321e-01 1.17840660e+00 8.49909782e-01
-2.10989378e-02 -1.12140036e+00 2.99947262e-01 8.37255716e-01
8.63113046e-01 -8.22795212e-01 -2.91171700e-01 5.66864274e-02
-4.37878400e-01 1.30190122e+00 1.02141666e+00 -5.58806598e-01
6.49387777e-01 6.09795749e-01 -2.13889834e-02 4.17557210e-02
-7.65961349e-01 1.64607212e-01 3.06801349e-01 3.93145800e-01
4.17895436e-01 -1.08680189e-01 -2.24770457e-01 8.23000133e-01
-1.76773384e-01 1.74196690e-01 2.43713051e-01 8.01074207e-01
-4.72784281e-01 -8.04243684e-01 -2.56914467e-01 6.89165771e-01
-5.37525952e-01 -3.17783654e-01 9.28245205e-03 7.81747937e-01
2.33979777e-01 2.99193203e-01 6.59603179e-01 -4.01480615e-01
4.32792306e-01 6.39664829e-02 3.91250849e-01 -1.57907590e-01
-1.16157937e+00 -2.27101386e-01 -1.34614199e-01 -2.00656310e-01
7.82737806e-02 1.71722069e-01 -1.56078637e+00 -6.34997964e-01
-3.02002311e-01 -3.09213642e-02 7.71446943e-01 4.69964474e-01
3.14805269e-01 1.12880635e+00 2.93362230e-01 -6.39363706e-01
-7.59482682e-01 -1.17036140e+00 -4.31429088e-01 3.81254822e-01
4.29921210e-01 -3.27185005e-01 9.15567651e-02 -1.70409810e-02] | [8.830869674682617, 2.7464919090270996] |
2b174bc9-c837-4fda-bb99-7e444a684ebc | exploit-fully-automatic-low-level-segmented | 1903.02871 | null | http://arxiv.org/abs/1903.02871v1 | http://arxiv.org/pdf/1903.02871v1.pdf | Exploit fully automatic low-level segmented PET data for training high-level deep learning algorithms for the corresponding CT data | We present an approach for fully automatic urinary bladder segmentation in CT
images with artificial neural networks in this study. Automatic medical image
analysis has become an invaluable tool in the different treatment stages of
diseases. Especially medical image segmentation plays a vital role, since
segmentation is often the initial step in an image analysis pipeline. Since
deep neural networks have made a large impact on the field of image processing
in the past years, we use two different deep learning architectures to segment
the urinary bladder. Both of these architectures are based on pre-trained
classification networks that are adapted to perform semantic segmentation.
Since deep neural networks require a large amount of training data,
specifically images and corresponding ground truth labels, we furthermore
propose a method to generate such a suitable training data set from Positron
Emission Tomography/Computed Tomography image data. This is done by applying
thresholding to the Positron Emission Tomography data for obtaining a ground
truth and by utilizing data augmentation to enlarge the dataset. In this study,
we discuss the influence of data augmentation on the segmentation results, and
compare and evaluate the proposed architectures in terms of qualitative and
quantitative segmentation performance. The results presented in this study
allow concluding that deep neural networks can be considered a promising
approach to segment the urinary bladder in CT images. | ['Jürgen Wallner', 'Peter M. Roth', 'Jan Egger', 'Christina Gsaxner'] | 2019-03-07 | null | null | null | null | ['bladder-segmentation'] | ['medical'] | [ 5.95471442e-01 4.66728300e-01 -2.64452189e-01 -6.35237336e-01
-3.40728283e-01 -2.11874381e-01 3.52221370e-01 4.47459430e-01
-1.13030231e+00 6.77860141e-01 -2.37393931e-01 -3.74955118e-01
4.62292805e-02 -9.91326332e-01 -6.05642259e-01 -5.20769835e-01
-7.18223527e-02 1.04467833e+00 5.30088581e-02 -1.27910171e-02
2.65147448e-01 8.61268699e-01 -1.17405462e+00 2.30281904e-01
9.01522934e-01 8.27703238e-01 6.50453389e-01 5.68640649e-01
-5.71327150e-01 7.16658115e-01 -5.35991549e-01 -2.29143538e-02
1.74434751e-01 -7.36439645e-01 -1.18258131e+00 4.01693285e-01
-1.20086342e-01 -2.70475298e-01 1.67612910e-01 1.10999346e+00
3.99277925e-01 -5.26433215e-02 6.60804868e-01 -7.40673959e-01
-9.64130759e-02 6.57434344e-01 -3.14200521e-01 4.67890888e-01
-1.42978102e-01 -1.12950608e-01 2.86887378e-01 -5.99303424e-01
6.99437082e-01 7.68653750e-01 6.26576245e-01 6.48178756e-01
-1.16781962e+00 -4.35613275e-01 -3.71817023e-01 -6.38018502e-03
-8.58047307e-01 -4.68599126e-02 5.90238512e-01 -6.52644932e-01
3.97556067e-01 2.77299970e-01 1.02507007e+00 6.87979043e-01
1.57641724e-01 8.00298154e-01 1.15700936e+00 -7.81324387e-01
2.96911001e-01 2.78756827e-01 2.58457452e-01 7.42198467e-01
5.01067519e-01 -8.62070471e-02 3.91964495e-01 1.81034669e-01
1.14946890e+00 9.11038443e-02 -6.50397614e-02 -2.68583685e-01
-9.46688950e-01 1.00093758e+00 9.13848758e-01 8.90223682e-01
-7.39697397e-01 2.05048829e-01 6.67141914e-01 -2.21120805e-01
2.70827264e-01 6.04535282e-01 -2.36317459e-02 3.87113869e-01
-1.11936867e+00 -1.03749670e-01 5.06824791e-01 4.21033710e-01
6.52166486e-01 -2.59105712e-01 -2.22543404e-01 9.87833381e-01
3.37907940e-01 -2.90757604e-02 1.02867901e+00 -6.90944910e-01
7.33055100e-02 7.75525391e-01 -1.35769680e-01 -5.21163583e-01
-8.85006547e-01 -3.73157114e-01 -1.01635981e+00 2.78161556e-01
5.88661432e-01 -1.41662613e-01 -1.65618181e+00 1.17732131e+00
2.64629334e-01 -3.30626070e-01 -4.91253920e-02 9.69326615e-01
8.95155668e-01 3.46457243e-01 4.90468413e-01 -5.89505807e-02
1.39652503e+00 -8.99314165e-01 -7.84180045e-01 -2.42081314e-01
8.87779653e-01 -5.58886707e-01 8.05032909e-01 2.93167681e-01
-1.06072772e+00 -5.19569993e-01 -1.13973081e+00 -1.23390093e-01
-3.86538893e-01 3.12079608e-01 6.27522051e-01 9.16693389e-01
-9.31137323e-01 7.83145189e-01 -1.29698467e+00 -4.20751095e-01
7.13846147e-01 6.83090270e-01 -3.31184417e-01 1.22290649e-01
-1.06370449e+00 1.02696276e+00 9.82591450e-01 3.20975721e-01
-4.97920543e-01 -9.51610655e-02 -7.41102636e-01 -1.76029168e-02
1.34048685e-01 -6.90999329e-01 1.33500683e+00 -1.12428999e+00
-1.31986392e+00 1.14156342e+00 2.07488731e-01 -7.40819097e-01
7.53077149e-01 2.64723867e-01 1.53250158e-01 3.46841395e-01
1.64628386e-01 9.38023686e-01 3.77582520e-01 -1.39852345e+00
-6.31736040e-01 -5.59519470e-01 -3.66940379e-01 -4.28212620e-02
-1.79621756e-01 -1.06463589e-01 -7.61391759e-01 -3.76768202e-01
4.75939572e-01 -1.10342479e+00 -7.71821320e-01 -3.04372609e-01
-5.17973661e-01 1.61686968e-02 5.24616182e-01 -7.95582891e-01
7.64103413e-01 -1.72597528e+00 -7.90146589e-02 4.93313491e-01
1.92238018e-02 4.07104611e-01 3.55930030e-01 -1.96786121e-01
-5.32324240e-02 2.79889256e-01 -6.75682783e-01 -4.52916175e-01
-3.92327875e-01 5.51031530e-01 3.68527591e-01 3.81745160e-01
7.17994049e-02 7.68232346e-01 -6.44088864e-01 -9.04276669e-01
5.22011697e-01 3.95020545e-01 -4.27934229e-01 1.49477303e-01
-2.02644318e-01 9.45812285e-01 -5.53562224e-01 3.78184289e-01
6.09531701e-01 -3.53463739e-01 3.11396897e-01 -6.66622259e-03
-6.90587536e-02 -1.53729618e-01 -7.66540706e-01 1.86685908e+00
-4.85645354e-01 3.73018622e-01 -5.85053749e-02 -1.37995434e+00
8.39480877e-01 4.04296368e-01 7.35534549e-01 -9.19942021e-01
6.23824000e-01 5.75044870e-01 2.75264978e-01 -6.59857035e-01
5.34280360e-01 -6.02124274e-01 8.81721079e-02 2.21990779e-01
6.73919842e-02 -3.70024711e-01 7.27629721e-01 -2.79169649e-01
6.38430953e-01 9.04891919e-03 1.30070463e-01 -2.98812121e-01
7.18504250e-01 5.43587089e-01 3.83387178e-01 5.91801465e-01
-1.63712114e-01 8.01987767e-01 4.33542699e-01 -6.31430864e-01
-1.27224541e+00 -5.63862205e-01 -5.75330198e-01 4.46038932e-01
1.05173074e-01 4.82147455e-01 -1.12971246e+00 -5.78721285e-01
-3.40371639e-01 4.09538239e-01 -8.77193034e-01 1.62240177e-01
-7.60632455e-01 -1.21911108e+00 4.40013647e-01 7.47350097e-01
7.06069767e-01 -1.60509062e+00 -9.12549615e-01 3.85408342e-01
-1.65590346e-01 -1.11336732e+00 1.92318469e-01 7.08806992e-01
-1.30202937e+00 -1.09143066e+00 -1.06435251e+00 -9.47262287e-01
1.03070128e+00 -3.50516140e-01 9.12710130e-01 3.85315776e-01
-5.99457264e-01 -1.40877098e-01 -8.91239867e-02 -5.27473629e-01
-9.30915713e-01 3.09381574e-01 -4.97081727e-01 -3.35640430e-01
-3.88880335e-02 -1.87317044e-01 -6.87866569e-01 2.97668632e-02
-1.17683017e+00 2.70333737e-01 9.57291603e-01 8.67104769e-01
9.30254400e-01 -1.43949211e-01 3.33747983e-01 -1.36623263e+00
6.57269955e-01 -3.12222093e-01 -6.60016716e-01 -6.50473684e-02
-4.97595161e-01 1.13711670e-01 5.25550246e-01 -4.20085667e-03
-1.03556979e+00 4.42789942e-01 -7.38586068e-01 -1.28425047e-01
-3.71066958e-01 8.56084585e-01 3.23852658e-01 -6.55337721e-02
7.94405758e-01 -1.04904562e-01 2.17675135e-01 -3.99561703e-01
-1.25623913e-02 4.95041221e-01 6.28644407e-01 -2.13755086e-01
-3.69610824e-03 6.59789085e-01 2.21226960e-01 -8.15470517e-01
-5.59041739e-01 -6.45937443e-01 -9.40050721e-01 -1.56849310e-01
1.37474763e+00 -3.19263875e-01 -3.89827907e-01 2.34687865e-01
-1.07825172e+00 -4.78919894e-01 -3.14659268e-01 6.94227695e-01
-4.48189497e-01 3.15688759e-01 -7.63827741e-01 -4.10170615e-01
-4.89702135e-01 -1.53432691e+00 6.76796317e-01 4.46469069e-01
-1.40023649e-01 -1.20509291e+00 2.31706470e-01 4.07466710e-01
2.03077421e-01 5.31103015e-01 9.18066204e-01 -8.06966662e-01
-2.16201380e-01 -3.80851299e-01 -4.00117874e-01 5.10799706e-01
1.39977515e-01 -3.13686579e-01 -8.42755437e-01 1.35024432e-02
2.29716264e-02 -2.87317276e-01 1.08978820e+00 7.76994526e-01
1.48146927e+00 2.93665946e-01 -5.68776250e-01 5.56843936e-01
1.44654918e+00 5.26416123e-01 6.02185249e-01 7.25248873e-01
6.62245214e-01 5.42733967e-01 4.51619565e-01 1.22974113e-01
-6.50672093e-02 3.55869114e-01 6.55508339e-01 -7.02833772e-01
-1.96229503e-01 1.63825884e-01 -4.87957120e-01 4.56568509e-01
-3.30257982e-01 5.01554832e-02 -1.14645231e+00 6.66233063e-01
-1.53081715e+00 -4.70734328e-01 -3.97701532e-01 2.03723025e+00
7.64238000e-01 2.24002972e-01 1.62505269e-01 2.92553961e-01
7.36661673e-01 -4.66298133e-01 -3.68672132e-01 -4.94315565e-01
3.47032458e-01 5.89900732e-01 6.76189601e-01 3.91672164e-01
-1.25763500e+00 5.78308582e-01 5.75724936e+00 3.35900366e-01
-1.56465948e+00 1.36681035e-01 8.86805296e-01 5.38398027e-01
2.00739905e-01 -3.30176592e-01 -4.19448137e-01 4.07737583e-01
7.72204220e-01 4.43616420e-01 -2.31433511e-01 7.86608100e-01
2.84462988e-01 -4.78731632e-01 -8.74420643e-01 5.61856508e-01
-2.88086921e-01 -1.38948166e+00 -2.95325108e-02 2.13338003e-01
6.07195675e-01 1.23691268e-01 -2.07617685e-01 1.53667212e-01
-3.86545546e-02 -1.20828354e+00 3.92993122e-01 3.71596694e-01
4.41886425e-01 -7.28243887e-01 1.21542299e+00 4.86490160e-01
-6.59016132e-01 1.29377231e-01 -2.70979166e-01 2.36290202e-01
1.95981309e-01 5.08752167e-01 -1.33092821e+00 5.12935460e-01
5.83322287e-01 1.54673308e-01 -5.17880619e-01 1.38751006e+00
-9.65982974e-02 4.16948646e-01 -4.15230304e-01 -3.90699655e-02
5.93233943e-01 -3.58065277e-01 -1.35470673e-01 1.28030431e+00
3.02735567e-01 -1.12345479e-01 2.55427033e-01 1.15449452e+00
-1.83348477e-01 3.23510796e-01 -2.54129171e-01 -1.89008474e-01
-9.27289352e-02 1.39305055e+00 -1.72187841e+00 -5.53052008e-01
-1.10769339e-01 9.82086658e-01 5.89393005e-02 -6.23652041e-02
-6.93500698e-01 -1.94890931e-01 -2.17480421e-01 8.09576362e-02
-4.95446846e-02 1.57537073e-01 -6.93709075e-01 -5.83727658e-01
-3.11362207e-01 -2.65704483e-01 3.96680743e-01 -2.82951087e-01
-6.01967573e-01 7.79661477e-01 -6.41787127e-02 -8.90190303e-01
-3.24277341e-01 -6.67671323e-01 -4.59452361e-01 9.01689827e-01
-1.50453615e+00 -9.69531834e-01 -4.97668445e-01 3.04726422e-01
3.46251875e-01 3.50798160e-01 8.61991048e-01 5.26938200e-01
-6.02198124e-01 -1.06681949e-02 -1.16525516e-02 4.08091217e-01
1.94247440e-01 -1.53527200e+00 1.41996741e-01 6.26715124e-01
5.53364977e-02 4.62190717e-01 5.96611977e-01 -6.48760378e-01
-6.52342319e-01 -1.08700264e+00 5.75377464e-01 1.25587121e-01
9.88625810e-02 1.90395236e-01 -8.88495505e-01 6.88527048e-01
1.18825518e-01 2.11512804e-01 5.87228656e-01 -2.69577533e-01
8.14665020e-01 4.46637690e-01 -1.24240792e+00 3.14921975e-01
2.53956854e-01 -2.31207050e-02 -4.25842524e-01 3.77631813e-01
9.43004712e-02 -7.59055495e-01 -8.39172363e-01 6.73089802e-01
3.18521947e-01 -1.10334802e+00 8.50997865e-01 -2.60899156e-01
5.92931330e-01 -2.48545464e-02 6.38467848e-01 -1.05622208e+00
4.23000343e-02 1.90188974e-01 6.85225248e-01 6.09335124e-01
5.11456490e-01 -3.71192724e-01 1.45673037e+00 5.58032095e-01
-5.08479893e-01 -7.39019632e-01 -7.02449560e-01 -3.92920285e-01
2.00182989e-01 -2.95268893e-01 7.92013556e-02 1.02121294e+00
-2.56267190e-01 -8.42283189e-04 9.05060470e-02 -1.52956039e-01
4.50477868e-01 1.03905439e-01 5.33556700e-01 -1.23740637e+00
1.22791938e-02 -5.88697493e-01 -3.44791830e-01 -5.87454081e-01
-9.56804380e-02 -1.12766933e+00 1.30584210e-01 -2.09100676e+00
2.64917374e-01 -6.18735850e-01 -3.54050338e-01 3.62481594e-01
-4.49747778e-02 7.18120992e-01 2.80931983e-02 2.26172298e-01
3.13103274e-02 1.16206877e-01 1.67796266e+00 -3.73331271e-02
-4.52799618e-01 3.93306434e-01 -1.97163388e-01 9.50416028e-01
7.61434376e-01 -4.97753799e-01 -2.44628936e-01 -1.56892017e-01
-2.15323582e-01 1.36420980e-01 2.21590966e-01 -9.99160409e-01
9.53828469e-02 2.86952913e-01 7.42542386e-01 -5.58669925e-01
2.08701104e-01 -9.53844368e-01 -4.22293041e-03 1.20297134e+00
-3.16894978e-01 -4.40358430e-01 4.06604886e-01 9.12684649e-02
-4.43949491e-01 -8.30397069e-01 1.12288415e+00 -7.86221981e-01
-7.64640152e-01 4.23147082e-02 -6.25297666e-01 -4.50492740e-01
9.67412651e-01 -5.31542122e-01 3.51517528e-01 1.46069303e-01
-1.37593794e+00 1.49823070e-01 2.19355464e-01 -6.01002239e-02
4.27626014e-01 -1.00563192e+00 -3.07850987e-01 1.20237119e-01
-2.65295915e-02 3.94884855e-01 2.33881086e-01 1.05392814e+00
-1.26557660e+00 5.02015173e-01 -6.68778002e-01 -9.11473095e-01
-1.11315441e+00 3.60630006e-01 5.98323762e-01 -6.67448401e-01
-6.11939549e-01 7.59504020e-01 7.33025596e-02 -5.43659329e-01
5.00753522e-02 -9.56687570e-01 -7.05042720e-01 1.47076538e-02
1.67531502e-02 2.64421999e-02 3.98597121e-01 -4.97540295e-01
-9.33738127e-02 1.86588034e-01 -1.64889500e-01 -1.03133164e-01
1.29469287e+00 1.09211698e-01 -1.57936245e-01 1.28383696e-01
9.93115246e-01 -4.66939658e-01 -8.55650246e-01 1.06592342e-01
4.38989371e-01 -1.10799052e-01 3.96815956e-01 -8.55581760e-01
-1.18647873e+00 9.61353064e-01 9.40023124e-01 3.80219936e-01
9.89414871e-01 -1.57520652e-01 7.15689719e-01 1.73181742e-01
7.95576647e-02 -1.17588079e+00 -4.29637134e-02 1.38458788e-01
6.21164501e-01 -1.58029449e+00 -3.53840888e-02 -5.57346761e-01
-4.92972732e-01 1.29689217e+00 5.46776831e-01 -2.56378442e-01
2.99457133e-01 1.98176891e-01 2.13070244e-01 -3.82657230e-01
3.01487833e-01 -3.30163330e-01 2.36351699e-01 1.13563143e-01
8.19099784e-01 -4.40299064e-02 -7.71232784e-01 3.91666353e-01
-9.84755158e-03 3.01798522e-01 5.38958669e-01 1.13287890e+00
-4.77280855e-01 -1.39978778e+00 -6.08307004e-01 4.93758976e-01
-7.43959963e-01 2.84710705e-01 -3.36923569e-01 1.16644835e+00
2.12770388e-01 4.30873603e-01 -1.23229409e-02 1.87216952e-01
2.71222770e-01 5.27731590e-02 7.11305201e-01 -8.10047448e-01
-7.43745148e-01 2.23932788e-01 1.64635256e-02 -1.02203153e-01
-5.77854574e-01 -3.26813608e-01 -1.91028190e+00 2.79116809e-01
-3.56963634e-01 1.56263605e-01 1.22647703e+00 1.23311222e+00
-2.74725646e-01 1.05914962e+00 2.00073332e-01 -1.14780402e+00
-1.17970012e-01 -1.21449816e+00 -5.17407417e-01 5.08219957e-01
8.37041587e-02 -5.19470870e-01 1.90315902e-01 1.72914937e-01] | [14.474397659301758, -2.476074457168579] |
fa9dcd3c-42d0-43a5-a259-87a66d475a27 | head-and-body-unified-detector-and-graph | 2111.13888 | null | https://arxiv.org/abs/2111.13888v1 | https://arxiv.org/pdf/2111.13888v1.pdf | Head and Body: Unified Detector and Graph Network for Person Search in Media | Person search in media has seen increasing potential in Internet applications, such as video clipping and character collection. This task is common but overlooked by previous person search works which focus on surveillance scenes. The media scenarios have some different challenges from surveillance scenes. For example, a person may change his clothes frequently. To alleviate this issue, this paper proposes a Unified Detector and Graph Network (UDGNet) for person search in media. UDGNet is the first person search framework to detect and re-identify the human body and head simultaneously. Specifically, it first builds two branches based on a unified network to detect the human body and head, then the detected body and head are used for re-identification. This dual-task approach can significantly enhance discriminative learning. To tackle the cloth-changing issue, UDGNet builds two graphs to explore reliable links among cloth-changing samples and utilizes a graph network to learn better embeddings. This design effectively enhances the robustness of person search to cloth-changing challenges. Besides, we demonstrate that UDGNet can be implemented with both anchor-based and anchor-free person search frameworks and further achieve performance improvement. This paper also contributes a large-scale dataset for Person Search in Media (PSM), which provides both body and head annotations. It is by far the largest dataset for person search in media. Experiments show that UDGNet improves the anchor-free model AlignPS by 12.1% in mAP. Meanwhile, it shows good generalization across surveillance and longterm scenarios. The dataset and code will be available at: https://github.com/shuxjweb/PSM.git. | ['Bo Ren', 'Wei Wen', 'Bo Ke', 'Ruizhi Qiao', 'Yusheng Tao', 'Xiujun Shu'] | 2021-11-27 | null | null | null | null | ['person-search'] | ['computer-vision'] | [-2.48863265e-01 -4.24000442e-01 -2.04534605e-01 -1.57542869e-01
-4.48260427e-01 -4.22223657e-01 4.05675471e-01 -3.54856178e-02
-4.17044461e-01 3.14625949e-01 4.14532959e-01 3.49541396e-01
8.98213312e-02 -7.76516676e-01 -4.03749943e-01 -4.76384789e-01
-3.67813148e-02 3.17241341e-01 6.25364184e-01 -2.55395293e-01
-3.19710702e-01 1.22583404e-01 -1.38335276e+00 -2.15664089e-01
6.95614636e-01 8.61683726e-01 2.03225806e-01 5.75300992e-01
4.10511464e-01 1.55283913e-01 -5.69753766e-01 -8.87778819e-01
3.37508023e-01 -7.72131281e-03 -5.61474741e-01 1.05565898e-01
8.39303374e-01 -6.52648807e-01 -9.55827534e-01 1.02615452e+00
1.13466561e+00 1.92075878e-01 3.13465565e-01 -1.36561930e+00
-8.08223546e-01 2.71968514e-01 -8.23006570e-01 3.55685681e-01
8.89023483e-01 3.37701350e-01 6.98860288e-01 -7.11183310e-01
3.24088126e-01 1.45302761e+00 9.57040370e-01 9.63567197e-01
-6.45517766e-01 -8.19186449e-01 3.34367186e-01 2.43095398e-01
-1.55246973e+00 -3.09073716e-01 6.22905016e-01 -3.17079753e-01
4.35816497e-01 4.91053522e-01 1.05976295e+00 1.48154485e+00
-2.17892647e-01 1.21234477e+00 5.61717689e-01 -3.81490923e-02
-3.47981930e-01 9.61298719e-02 3.46263051e-01 9.49105620e-01
4.13703591e-01 5.27876988e-03 -5.46756923e-01 -1.78835317e-01
7.38390028e-01 3.77554834e-01 -5.99257827e-01 -2.02114493e-01
-1.18187416e+00 5.72461843e-01 5.54587483e-01 3.06295529e-02
-2.99552307e-02 1.00291431e-01 6.73510849e-01 3.60052735e-02
4.47616816e-01 8.66912603e-02 -2.75762845e-02 2.12343801e-02
-8.20990562e-01 3.47152650e-01 7.55942047e-01 1.17280936e+00
2.08663806e-01 -1.84307888e-01 -4.98991549e-01 1.13112581e+00
3.36945593e-01 9.70972478e-01 2.94554979e-01 -5.08675396e-01
5.89145482e-01 6.87649190e-01 3.95018980e-02 -1.53265214e+00
-4.30720270e-01 -4.78549629e-01 -9.11300540e-01 -6.25794947e-01
3.26634675e-01 -1.58553913e-01 -8.27557504e-01 1.73276103e+00
6.47590876e-01 3.75683814e-01 -3.16638410e-01 1.33206511e+00
1.43562126e+00 5.23417056e-01 8.76619816e-02 1.71085894e-01
1.72957480e+00 -1.34938526e+00 -4.83841479e-01 -3.91505897e-01
1.95809707e-01 -4.78821814e-01 9.75071132e-01 2.39240024e-02
-1.00416136e+00 -7.70274103e-01 -8.47523034e-01 3.27636935e-02
-3.64270687e-01 2.32892379e-01 3.79278243e-01 7.20636308e-01
-9.92645800e-01 1.19747736e-01 -5.95450461e-01 -9.07172501e-01
2.84107745e-01 3.93410414e-01 -3.72094721e-01 -2.74615616e-01
-1.41915393e+00 5.37113369e-01 1.68895647e-01 2.53262460e-01
-6.24279380e-01 -4.02530223e-01 -1.01557398e+00 -2.56891668e-01
6.10286534e-01 -9.78668034e-01 9.57816780e-01 -3.88046354e-01
-9.50299084e-01 9.98647571e-01 -1.06569178e-01 -2.25087836e-01
7.61862457e-01 -2.33462527e-01 -7.88782716e-01 2.43506283e-01
4.09454077e-01 5.56174338e-01 8.65682006e-01 -9.00643885e-01
-5.52154720e-01 -4.86212641e-01 4.21986766e-02 2.22062156e-01
-7.52146423e-01 1.73746347e-01 -1.57508898e+00 -9.23033774e-01
-8.98903012e-02 -1.11079276e+00 1.14380114e-01 2.34228298e-01
-4.39052194e-01 -4.08586711e-01 8.12056661e-01 -9.30605650e-01
1.51149297e+00 -2.00587153e+00 1.94164827e-01 1.55928805e-01
5.20544529e-01 4.26276922e-01 -1.12728029e-01 2.21735612e-01
4.10109371e-01 2.60949098e-02 2.12923706e-01 -7.14687884e-01
7.41071403e-02 -3.06664221e-02 2.89937496e-01 6.32556200e-01
-2.69756794e-01 1.06357467e+00 -9.82186854e-01 -8.13866258e-01
1.30152002e-01 4.36611295e-01 -2.37031370e-01 1.25932679e-01
3.27759713e-01 2.88641602e-01 -4.76454943e-01 1.04740393e+00
7.52890646e-01 -3.79795611e-01 -7.78873935e-02 -4.11369830e-01
2.06147566e-01 -4.47992980e-01 -1.14907551e+00 1.54139793e+00
4.31920812e-02 5.41688442e-01 2.12599114e-01 -5.24859488e-01
7.97115684e-01 1.83031693e-01 4.63410437e-01 -6.79457188e-01
1.20871358e-01 -4.43396159e-02 -4.83060837e-01 -6.93580031e-01
6.02812827e-01 6.16167724e-01 -3.30687791e-01 1.07873194e-01
-1.05801828e-01 4.89015490e-01 3.82973909e-01 2.59753048e-01
9.54419136e-01 -1.58581674e-01 2.67990697e-02 -1.46593124e-01
6.93418384e-01 -4.07886893e-01 6.61135435e-01 8.07959437e-01
-6.49959862e-01 6.16692483e-01 -2.11736917e-01 -5.60258329e-01
-8.16253662e-01 -1.00099552e+00 8.89841393e-02 1.23052073e+00
7.83240199e-01 -6.05833590e-01 -6.97033882e-01 -6.83891177e-01
2.02582851e-01 -1.54392064e-01 -6.20931208e-01 -2.68618912e-01
-6.59859836e-01 -7.11565852e-01 8.01193357e-01 4.69114602e-01
1.19245112e+00 -8.36613894e-01 -2.36197233e-01 -8.36822018e-02
-6.08556986e-01 -1.14384329e+00 -1.23204052e+00 -7.89309740e-01
-2.11065143e-01 -1.30723667e+00 -1.38808370e+00 -1.04348958e+00
5.90251267e-01 5.87082446e-01 9.60369349e-01 3.52940232e-01
-3.81214619e-01 9.52729166e-01 -4.18794364e-01 -2.25388810e-01
1.34237036e-01 2.86738515e-01 4.23321396e-01 2.36097753e-01
5.37997246e-01 -3.07551287e-02 -1.09218502e+00 7.88341045e-01
-5.47052801e-01 -1.03501692e-01 1.49240896e-01 5.57514608e-01
2.88676083e-01 3.77829671e-02 2.21568063e-01 -2.67804444e-01
7.80631602e-01 -5.85279703e-01 -2.37077266e-01 4.91261780e-01
-4.70732093e-01 -6.07898593e-01 1.72333881e-01 -7.73323119e-01
-6.27006829e-01 -2.26787955e-01 -1.46972358e-01 -4.36823130e-01
1.88215375e-02 7.19982535e-02 -2.05355838e-01 -2.93632329e-01
4.24353510e-01 3.83151114e-01 -1.32545471e-01 -5.24617434e-01
-1.55250117e-01 6.99837863e-01 7.09183753e-01 -3.82123679e-01
9.09320474e-01 3.67069453e-01 -2.80125439e-01 -8.61275315e-01
-5.94901443e-01 -1.06748188e+00 -3.21294397e-01 -6.83067381e-01
1.01838398e+00 -1.17387331e+00 -8.69930506e-01 8.76627505e-01
-9.96047318e-01 -8.10637996e-02 2.26709038e-01 3.22718173e-01
1.27961278e-01 7.36242056e-01 -7.39269614e-01 -8.39928448e-01
-6.86659276e-01 -9.08496261e-01 1.25852346e+00 6.63792431e-01
-1.39254034e-01 -9.63650107e-01 -1.46737576e-01 5.81163228e-01
2.12250277e-01 3.73818547e-01 -2.05102619e-02 -4.87069905e-01
-3.67160738e-01 -5.39852560e-01 -3.19212496e-01 -9.39243659e-03
1.74079970e-01 -4.28362906e-01 -6.96605086e-01 -8.32525373e-01
-4.01458919e-01 -6.60468712e-02 1.02014494e+00 2.34659255e-01
8.88980269e-01 -3.80333841e-01 -8.25194478e-01 7.42726564e-01
1.02697992e+00 -1.64640546e-01 3.71704876e-01 4.32135403e-01
1.05742335e+00 4.68185574e-01 5.27119339e-01 4.44767922e-01
8.50545108e-01 1.01020420e+00 1.93382874e-01 -1.56709120e-01
-4.33967829e-01 -4.18640077e-01 4.57828671e-01 5.99747419e-01
-2.01754451e-01 -5.25984645e-01 -9.19816136e-01 6.01460874e-01
-2.05698752e+00 -1.05456054e+00 -4.28483710e-02 1.97136319e+00
6.00686610e-01 -1.27966151e-01 7.22420335e-01 -2.18658410e-02
1.15962911e+00 2.54819781e-01 -6.20690346e-01 4.01088566e-01
-1.50498167e-01 -3.97285610e-01 5.57772696e-01 2.45110884e-01
-1.43154383e+00 8.07836592e-01 5.20964479e+00 8.15512121e-01
-8.41140211e-01 3.24986279e-01 2.43973881e-01 -1.53654262e-01
2.67350674e-01 -6.23596191e-01 -1.30631506e+00 9.64072764e-01
2.89606243e-01 -3.63410041e-02 2.17226222e-01 8.24140370e-01
1.16715461e-01 1.73399717e-01 -9.16943014e-01 1.69587719e+00
5.51712155e-01 -9.42408741e-01 -7.51857683e-02 -1.34454340e-01
4.44058210e-01 -1.53341711e-01 1.06131844e-01 3.97405118e-01
-1.05020650e-01 -6.20564163e-01 5.68671763e-01 5.66800594e-01
8.65309775e-01 -4.99408484e-01 8.22106659e-01 2.23288715e-01
-1.82408333e+00 -2.02749655e-01 -2.60403901e-01 2.88589716e-01
4.24679607e-01 2.18385369e-01 -5.84097028e-01 4.86949444e-01
1.22522557e+00 9.11554873e-01 -1.01368356e+00 1.63095152e+00
8.68077427e-02 3.55237663e-01 -4.43707764e-01 -1.70073465e-01
-1.85590416e-01 3.92769240e-02 9.21453059e-01 1.54883528e+00
2.84898013e-01 -5.05340584e-02 7.90812254e-01 4.56287235e-01
-4.09905799e-02 -3.17242742e-02 -4.39446330e-01 3.81288469e-01
4.72565830e-01 1.10406673e+00 -5.14746547e-01 -3.26736629e-01
-3.66212636e-01 1.41829312e+00 1.24070741e-01 4.47972357e-01
-1.17458487e+00 -1.66529953e-01 5.69145441e-01 4.47396457e-01
1.00697115e-01 -1.58916488e-01 5.26096284e-01 -1.39443409e+00
3.74091357e-01 -8.65370870e-01 7.03070164e-01 -5.93770564e-01
-1.53494680e+00 5.44317961e-01 1.44991785e-01 -1.17255270e+00
8.79920200e-02 -2.81342328e-01 -4.76999342e-01 6.27241969e-01
-1.12217093e+00 -1.66566968e+00 -9.48326588e-01 9.37606514e-01
7.04522848e-01 -3.38483065e-01 4.83456880e-01 8.14942122e-01
-9.83964384e-01 1.17139125e+00 -4.04470652e-01 7.02285230e-01
8.53829026e-01 -9.11421359e-01 4.82173175e-01 9.71589386e-01
-2.73806304e-01 7.56151557e-01 4.75790501e-01 -1.04941559e+00
-1.42793059e+00 -1.31916010e+00 5.62095344e-01 -5.59281766e-01
3.92139107e-01 -5.45432448e-01 -6.27990544e-01 4.28710341e-01
-2.91750338e-02 2.38469976e-04 5.29273510e-01 2.78242026e-02
-3.70559961e-01 -2.05850527e-01 -1.06571400e+00 7.02438712e-01
1.53994179e+00 -4.15298730e-01 -2.92792529e-01 4.83675838e-01
6.73810422e-01 -6.05255187e-01 -8.43114078e-01 4.00396377e-01
6.77617133e-01 -6.91769898e-01 1.34079552e+00 -1.24878339e-01
-1.64854318e-01 -4.08035725e-01 1.54580716e-02 -9.10246134e-01
-5.83424926e-01 -7.44243264e-01 -5.30066490e-01 1.44640505e+00
1.68353152e-02 -7.09287763e-01 8.17326546e-01 7.81847835e-01
1.92433983e-01 -7.56774068e-01 -7.06069052e-01 -9.59027886e-01
-5.46417892e-01 -1.40834585e-01 7.55281031e-01 6.87873840e-01
-3.14308852e-01 2.54763335e-01 -9.46506679e-01 3.13335478e-01
9.21294987e-01 -1.69758976e-01 1.03393030e+00 -1.28273237e+00
-2.78836131e-01 -3.79845560e-01 -5.91362894e-01 -1.52179539e+00
-1.59445018e-01 -8.04866195e-01 -2.18803614e-01 -1.68067431e+00
5.76481044e-01 -4.55368817e-01 -7.07381815e-02 4.08344567e-01
-4.07935649e-01 3.67214650e-01 4.44050491e-01 4.43339050e-01
-1.06595302e+00 4.27396029e-01 1.11513674e+00 -4.99204993e-01
-1.94565639e-01 2.64424652e-01 -5.75431585e-01 6.10421002e-01
7.91628182e-01 -2.98699848e-02 -2.49028414e-01 -5.58037162e-01
1.61345042e-02 -2.42512360e-01 9.35708344e-01 -1.22977960e+00
6.27163291e-01 3.01738858e-01 6.03527248e-01 -6.65426254e-01
5.31852484e-01 -7.59087861e-01 1.51675984e-01 5.30099571e-01
8.56620967e-02 2.75165886e-01 1.82352513e-01 8.54710996e-01
-3.94088477e-02 1.25479952e-01 4.60088998e-01 -2.22221583e-01
-1.10136855e+00 9.06872749e-01 2.27657750e-01 3.15948240e-02
1.03570414e+00 -4.34812814e-01 -4.05734718e-01 -5.05744815e-01
-6.64725780e-01 7.96685517e-01 4.05508071e-01 8.18448722e-01
8.53473425e-01 -1.51851070e+00 -7.82351851e-01 2.91596772e-03
1.82981655e-01 -1.75100341e-01 3.78774166e-01 8.52241933e-01
-3.12493473e-01 1.59724042e-01 1.12033799e-01 -6.64811373e-01
-1.73294818e+00 5.48390329e-01 4.55178946e-01 -8.86059850e-02
-6.68918610e-01 1.05362153e+00 9.73583385e-02 -3.20454568e-01
5.50835788e-01 1.20504141e-01 -3.47776443e-01 1.09700412e-01
7.63552904e-01 6.23382568e-01 -5.46067178e-01 -9.42332745e-01
-6.26114249e-01 9.55596089e-01 -8.49281251e-02 3.78089160e-01
8.84729624e-01 -4.88012612e-01 9.01366696e-02 4.77620214e-03
1.22879755e+00 6.90461472e-02 -1.05834341e+00 -3.59331548e-01
-3.93638015e-01 -6.22947276e-01 -3.50607038e-01 -4.72642601e-01
-1.27985513e+00 3.29349458e-01 8.72685790e-01 1.71845481e-01
1.03577006e+00 1.95343971e-01 1.33947933e+00 1.38391912e-01
5.03433466e-01 -1.14724636e+00 3.46141964e-01 2.51410693e-01
9.39012706e-01 -1.59101653e+00 1.13681965e-01 -5.73900342e-01
-4.07981277e-01 7.14153409e-01 9.65664446e-01 1.33321986e-01
5.12330174e-01 -1.95132911e-01 -2.18283489e-01 -3.50609779e-01
-3.93570811e-02 -3.53934884e-01 6.44716263e-01 8.29303324e-01
-8.94480497e-02 7.94896856e-02 -7.92877749e-02 6.78730845e-01
-2.26606518e-01 -1.98474228e-01 -1.99582130e-01 6.52500510e-01
-1.50384843e-01 -8.54403198e-01 -5.50151467e-01 4.61842060e-01
-2.81395167e-01 -1.26288924e-02 -4.37921315e-01 7.39638567e-01
4.10965204e-01 1.18610418e+00 -2.39351392e-01 -7.42484570e-01
5.43852568e-01 -3.23694170e-01 2.67182082e-01 -2.84320891e-01
-4.99149442e-01 -1.09101802e-01 1.43083036e-01 -5.76001465e-01
-4.88118529e-01 -5.76795161e-01 -7.28757322e-01 -6.73892617e-01
-3.30353409e-01 -1.90096851e-02 6.70178905e-02 4.96960938e-01
3.59845966e-01 2.57985085e-01 1.93508759e-01 -7.66546190e-01
-2.66035378e-01 -8.00855517e-01 -3.05098087e-01 6.48645759e-01
3.82106155e-01 -8.44678521e-01 -9.16645676e-02 -6.56400546e-02] | [14.829538345336914, 0.7823842763900757] |
d5fc1c97-9f21-4649-80b3-8fe81f22fd04 | geometrics-exploiting-geometric-structure-for | 1901.11461 | null | http://arxiv.org/abs/1901.11461v1 | http://arxiv.org/pdf/1901.11461v1.pdf | GEOMetrics: Exploiting Geometric Structure for Graph-Encoded Objects | Mesh models are a promising approach for encoding the structure of 3D
objects. Current mesh reconstruction systems predict uniformly distributed
vertex locations of a predetermined graph through a series of graph
convolutions, leading to compromises with respect to performance or resolution.
In this paper, we argue that the graph representation of geometric objects
allows for additional structure, which should be leveraged for enhanced
reconstruction. Thus, we propose a system which properly benefits from the
advantages of the geometric structure of graph encoded objects by introducing
(1) a graph convolutional update preserving vertex information; (2) an adaptive
splitting heuristic allowing detail to emerge; and (3) a training objective
operating both on the local surfaces defined by vertices as well as the global
structure defined by the mesh. Our proposed method is evaluated on the task of
3D object reconstruction from images with the ShapeNet dataset, where we
demonstrate state of the art performance, both visually and numerically, while
having far smaller space requirements by generating adaptive meshes | ['Scott Fujimoto', 'David Meger', 'Edward J. Smith', 'Adriana Romero'] | 2019-01-31 | null | null | null | null | ['3d-object-reconstruction'] | ['computer-vision'] | [ 2.31605902e-01 5.23623347e-01 2.68649548e-01 -1.31941959e-01
-4.06142712e-01 -4.58669960e-01 7.05507755e-01 5.31708121e-01
1.95599943e-01 2.56378978e-01 1.56016842e-01 -2.18974218e-01
-3.58378552e-02 -1.18436515e+00 -1.05616271e+00 -4.82274354e-01
-3.39535445e-01 7.23319232e-01 3.10540110e-01 -1.57044321e-01
3.27358931e-01 1.26645076e+00 -1.56864583e+00 1.90554440e-01
6.44724786e-01 1.36038220e+00 2.77836084e-01 5.58088243e-01
-4.23986793e-01 6.45317972e-01 -3.51209976e-02 -3.95063043e-01
6.14226162e-01 -1.27272502e-01 -6.61516666e-01 3.82200897e-01
9.08716381e-01 -3.73238921e-01 -3.73898834e-01 8.81452501e-01
2.96195269e-01 4.74723205e-02 7.19428003e-01 -6.89506173e-01
-6.25704110e-01 3.75246942e-01 -4.47366595e-01 -2.81026542e-01
1.53921545e-01 3.49601358e-01 9.83612239e-01 -9.29938436e-01
9.21433687e-01 1.25429428e+00 8.87859106e-01 3.53888959e-01
-1.47123837e+00 -2.74818391e-01 1.10550873e-01 -3.31068128e-01
-1.48769355e+00 -3.99008840e-01 1.16744578e+00 -5.08155048e-01
1.19928765e+00 2.51700431e-01 8.40860963e-01 6.41546667e-01
2.98093498e-01 3.27971965e-01 5.61709821e-01 -1.47624165e-01
3.19963753e-01 -6.63428903e-02 -2.86694825e-01 1.10721457e+00
2.12399900e-01 1.01653695e-01 -2.76525080e-01 -2.96095878e-01
1.48572457e+00 -2.07380444e-01 -3.61260384e-01 -8.91276300e-01
-8.14685583e-01 7.25977302e-01 7.87433028e-01 -1.15655102e-01
-5.66149712e-01 4.93043840e-01 1.87399864e-01 1.28358766e-01
7.86716759e-01 4.50988412e-01 -1.96165740e-01 3.17372054e-01
-9.12388623e-01 4.39389348e-01 7.08101630e-01 1.12262094e+00
1.03759420e+00 2.56433100e-01 -1.05049051e-01 4.88156170e-01
5.09209216e-01 3.04911703e-01 -3.14533532e-01 -1.03435373e+00
2.59425581e-01 9.65528548e-01 2.09965184e-02 -1.29327500e+00
-4.69961137e-01 -4.99621451e-01 -7.26810455e-01 5.46009004e-01
7.95863941e-02 3.67379099e-01 -1.20879030e+00 1.59087849e+00
5.28296232e-01 3.80279392e-01 -2.89636612e-01 7.63757050e-01
1.06776023e+00 4.08847839e-01 -4.24714871e-02 3.04856092e-01
1.04171944e+00 -6.14816368e-01 -3.07316780e-01 -1.37360359e-03
3.73892754e-01 -5.29721260e-01 7.00736105e-01 -7.99863935e-02
-1.49205625e+00 -5.35687327e-01 -1.03434789e+00 -2.79280484e-01
-1.17386945e-01 -2.50437021e-01 7.34511912e-01 3.93288255e-01
-1.37485480e+00 8.09224904e-01 -8.65983725e-01 -1.46787882e-01
7.32594967e-01 5.01581430e-01 -3.25963706e-01 7.18181487e-03
-4.69474107e-01 6.43864274e-01 1.45373240e-01 -1.87421873e-01
-7.37669170e-01 -1.06412423e+00 -1.20526552e+00 4.07129705e-01
1.08049735e-02 -1.06946611e+00 7.16277897e-01 -5.83588004e-01
-1.27162099e+00 1.02492476e+00 2.96615928e-01 -5.33285975e-01
6.24411464e-01 2.10184813e-01 1.48001432e-01 4.28576380e-01
-1.55406818e-01 8.42504919e-01 8.54891300e-01 -1.49234653e+00
-2.99037427e-01 -4.67581719e-01 2.47237206e-01 2.55999744e-01
1.96826711e-01 -6.70940340e-01 -3.40057522e-01 -7.29979932e-01
4.01476264e-01 -7.40464270e-01 -3.86036128e-01 4.00775880e-01
-2.44384691e-01 -4.62678745e-02 7.43154943e-01 -6.80501699e-01
8.87290061e-01 -2.26748848e+00 1.46413803e-01 4.62886363e-01
6.20535970e-01 -3.21929716e-02 -1.70928895e-01 3.89663666e-01
1.06010199e-01 1.76851496e-01 -2.91116178e-01 -6.56648874e-01
-1.01649940e-01 1.74224660e-01 -1.86147854e-01 5.95259368e-01
4.67756093e-01 1.07061589e+00 -8.02367747e-01 -2.67038405e-01
5.19533277e-01 7.55258679e-01 -9.50880349e-01 1.73579305e-01
-5.95263720e-01 4.36754614e-01 -5.76651990e-01 6.09591365e-01
9.62004006e-01 -6.38398349e-01 1.59088448e-02 -5.00896990e-01
1.12066120e-02 3.67548764e-01 -1.26008403e+00 1.72882032e+00
-2.25808382e-01 2.90152192e-01 4.33159351e-01 -9.70393538e-01
9.61528301e-01 2.33290881e-01 7.94744909e-01 -4.39247012e-01
1.20446615e-01 5.98661900e-02 -2.75342673e-01 6.75675869e-02
5.31401336e-01 -3.73254865e-02 3.40026349e-01 2.52988994e-01
1.02048874e-01 -5.10171056e-01 -3.15246820e-01 2.52563447e-01
1.28774548e+00 1.80293739e-01 1.60759164e-03 -5.22031963e-01
9.90022868e-02 6.29009679e-02 1.00638866e-01 6.34109735e-01
3.71543139e-01 8.84343207e-01 1.14594691e-01 -6.37905777e-01
-1.28634405e+00 -1.29361629e+00 -2.18306899e-01 2.70478100e-01
5.28560460e-01 -4.37331349e-01 -5.16240120e-01 -5.10108650e-01
3.34775925e-01 5.84372699e-01 -7.10047662e-01 -6.51868880e-02
-8.17107916e-01 -4.05081473e-02 2.60287195e-01 4.46234077e-01
3.26891929e-01 -8.82967293e-01 -7.14963913e-01 2.73517162e-01
2.84547567e-01 -1.11840224e+00 -4.15026098e-01 2.75707424e-01
-1.20619750e+00 -9.71844316e-01 -3.35911542e-01 -7.48480380e-01
9.56344426e-01 4.83935803e-01 1.38059068e+00 4.47268575e-01
-1.50812194e-01 6.68404996e-01 -1.50493324e-01 -1.36140361e-01
-4.51662064e-01 -9.66521949e-02 -3.94749761e-01 2.12236810e-02
-4.08383787e-01 -9.96071756e-01 -6.65925026e-01 -3.79829155e-03
-9.29287553e-01 5.90349317e-01 4.97116596e-01 5.47567308e-01
8.84697735e-01 -2.88537201e-02 2.14400619e-01 -8.81920516e-01
1.13278612e-01 -4.44901526e-01 -6.95161641e-01 -2.38565332e-03
-5.55956185e-01 2.73669362e-01 5.34791291e-01 -3.01386654e-01
-8.27464819e-01 1.57022253e-01 -1.46013975e-01 -7.63456345e-01
-1.47206681e-02 2.60479003e-01 1.39545025e-02 -5.90465069e-01
5.82959235e-01 1.01139195e-01 1.88731179e-01 -5.47467470e-01
4.34760094e-01 2.89365221e-02 2.77168959e-01 -7.16833532e-01
6.76435232e-01 7.26611018e-01 4.99797761e-01 -9.21825886e-01
-3.60589147e-01 -2.09573165e-01 -4.59459066e-01 -2.75306314e-01
7.50448406e-01 -8.60102832e-01 -3.85833263e-01 1.97195321e-01
-1.23168302e+00 -4.65245217e-01 -6.06455445e-01 1.18401513e-01
-7.91652858e-01 2.31661975e-01 -6.79786325e-01 -5.69988787e-01
-3.56657237e-01 -1.11164701e+00 1.49832547e+00 -6.47406876e-02
-6.45651817e-02 -1.04173362e+00 -3.07732612e-01 3.44845057e-02
5.01327395e-01 7.18880296e-01 1.28628433e+00 -1.86220556e-01
-1.24204314e+00 -2.92926896e-02 -4.40511048e-01 1.14296138e-01
1.37993053e-01 -1.22877307e-01 -7.48307168e-01 -3.28852475e-01
-1.55474365e-01 1.29046617e-02 5.16906679e-01 5.35265088e-01
1.18154037e+00 -3.16730112e-01 -3.39027643e-01 9.35882509e-01
1.61533308e+00 -3.37429613e-01 6.04104102e-01 -7.77587965e-02
1.06376278e+00 4.13542837e-01 -8.76493827e-02 5.42646587e-01
3.32232922e-01 5.98008215e-01 9.14471984e-01 -1.36485338e-01
-7.25485027e-01 -4.72869307e-01 -2.30153784e-01 5.92361093e-01
-1.37761116e-01 -2.48337165e-01 -7.82751143e-01 5.68538249e-01
-1.70661759e+00 -5.58530211e-01 -2.32218847e-01 2.18142247e+00
4.08356667e-01 1.79213405e-01 -1.81021720e-01 -2.65317887e-01
5.25770605e-01 3.18513453e-01 -5.48359811e-01 -2.58483648e-01
-3.09447311e-02 5.87285876e-01 5.78091204e-01 5.52242756e-01
-7.23567426e-01 7.60222614e-01 6.51073980e+00 6.20948374e-01
-1.35002649e+00 -2.49799803e-01 5.18363535e-01 3.21188793e-02
-6.67999923e-01 -2.98188590e-02 -4.36700255e-01 2.26897836e-01
4.72358495e-01 -1.00097179e-01 6.28609121e-01 7.09713697e-01
-2.10119054e-01 3.18908840e-01 -1.20824313e+00 9.99648333e-01
2.08255649e-02 -1.86564994e+00 7.05918729e-01 3.90589714e-01
7.09199429e-01 1.65199041e-01 -7.08208606e-02 -7.34042674e-02
1.78439558e-01 -1.17282891e+00 1.08461273e+00 4.23582554e-01
8.41033280e-01 -5.89505017e-01 2.51422614e-01 2.64558613e-01
-1.34807050e+00 4.43176270e-01 -3.18169951e-01 5.52775078e-02
5.72177619e-02 4.79100138e-01 -9.19227242e-01 8.32100213e-01
4.89209533e-01 5.61935902e-01 -4.31015700e-01 1.02426589e+00
1.14857621e-01 4.35358882e-01 -5.16653717e-01 3.37263793e-01
3.62590730e-01 -1.70730889e-01 7.56177068e-01 9.98403907e-01
2.99508750e-01 1.73520386e-01 2.92900175e-01 1.27559233e+00
-2.87577718e-01 6.98625073e-02 -8.98029804e-01 -1.07932463e-03
4.16765124e-01 1.09772325e+00 -1.06730247e+00 -1.38504893e-01
-3.93532902e-01 7.30762362e-01 6.80532038e-01 2.33535483e-01
-5.61648488e-01 9.84450355e-02 6.47820175e-01 7.56032348e-01
7.56609976e-01 -5.52004874e-01 -8.04041564e-01 -7.51657307e-01
-3.69780399e-02 -4.20873821e-01 4.09474783e-02 -7.95404613e-01
-1.17382348e+00 5.44813216e-01 -1.54241711e-01 -1.07333183e+00
1.15265355e-01 -6.30023479e-01 -3.08621198e-01 8.98419142e-01
-1.39517260e+00 -1.27932346e+00 -4.62485582e-01 2.46902928e-01
2.71735191e-01 2.83837587e-01 7.54596353e-01 2.18030155e-01
2.42634162e-01 2.96562821e-01 -4.31205958e-01 -5.40572777e-02
-6.59925342e-02 -1.23974371e+00 7.73922682e-01 4.06690359e-01
2.90976018e-01 4.53176618e-01 5.62297940e-01 -8.85572493e-01
-1.76619017e+00 -1.20170426e+00 4.70476657e-01 -3.45864028e-01
1.23603798e-01 -3.95714641e-01 -1.02300239e+00 7.31194675e-01
-1.98190600e-01 2.65099257e-01 2.16606215e-01 -1.55110687e-01
-2.52022535e-01 3.46965402e-01 -1.51671600e+00 6.47949398e-01
1.28988028e+00 -3.73934180e-01 -2.16337368e-01 1.39440641e-01
7.59065926e-01 -8.08224678e-01 -9.69809115e-01 7.41927862e-01
4.96155173e-01 -9.65512037e-01 1.10715961e+00 -4.50320631e-01
5.18087983e-01 -2.96115696e-01 -3.87105286e-01 -1.12343204e+00
-7.04087555e-01 -4.25129145e-01 -4.27628636e-01 8.00628603e-01
1.65084153e-01 -3.91224831e-01 9.72181022e-01 6.05674207e-01
-4.72267956e-01 -1.17381406e+00 -9.22666490e-01 -4.43794489e-01
-1.76909417e-01 -3.43574971e-01 8.12057555e-01 8.52378547e-01
-6.46116495e-01 2.51775146e-01 3.32121784e-03 4.84540164e-01
7.35090971e-01 3.05788070e-01 7.94553459e-01 -1.32728755e+00
-1.04448117e-01 -6.29210591e-01 -7.37939239e-01 -1.51418436e+00
-1.28938839e-01 -1.09825599e+00 -8.14354643e-02 -1.74592948e+00
-1.77938774e-01 -7.05439568e-01 -1.73335895e-02 1.14887066e-01
9.42473561e-02 3.19864184e-01 -4.16554091e-03 -3.63920033e-02
-3.71623039e-01 6.94268644e-01 1.52852118e+00 -5.30389249e-02
-8.70148689e-02 -4.38771933e-01 -6.11462295e-01 6.46657765e-01
4.14816737e-01 -1.19153619e-01 -4.14610475e-01 -8.59250307e-01
1.80757672e-01 2.19689161e-01 4.85357344e-01 -8.77027571e-01
7.26760626e-02 8.86001363e-02 4.47686046e-01 -7.06030309e-01
6.37808859e-01 -9.76264417e-01 6.16824746e-01 1.50766894e-01
-4.54456955e-02 -2.80342344e-02 3.35943878e-01 6.86275959e-01
2.30088621e-01 5.76799922e-02 9.81239915e-01 -4.23274547e-01
-4.34476107e-01 6.81926608e-01 2.07054511e-01 -2.18321681e-01
7.91131914e-01 -5.33544838e-01 -8.61370005e-03 -2.24303067e-01
-5.73316813e-01 2.23546289e-03 1.16158819e+00 2.19243988e-01
7.92922914e-01 -1.46903133e+00 -6.31090879e-01 5.26889980e-01
-5.95597327e-02 4.49033260e-01 1.97803199e-01 3.57504100e-01
-9.62122560e-01 5.83230145e-02 -1.38974547e-01 -8.30353916e-01
-9.52483475e-01 3.65359068e-01 3.79422486e-01 -1.79607987e-01
-1.27039385e+00 8.43734205e-01 4.68161821e-01 -4.40265566e-01
1.91006988e-01 -5.96476972e-01 1.99156925e-01 -4.11463082e-01
1.17674530e-01 1.91481635e-01 3.66306931e-01 -6.76344573e-01
-1.09969102e-01 5.71838319e-01 1.29420102e-01 4.04892445e-01
1.50458932e+00 -3.35134789e-02 -3.95039544e-02 1.24840610e-01
9.93678451e-01 1.77459523e-01 -1.46224117e+00 -2.48060152e-01
-2.26342559e-01 -6.41544044e-01 3.57139975e-01 -5.80816031e-01
-1.12023866e+00 5.19147038e-01 3.51690888e-01 4.50045407e-01
7.34556615e-01 2.40022257e-01 9.89752591e-01 -6.22782931e-02
8.15774560e-01 -4.29105729e-01 -3.09998952e-02 4.05802429e-01
9.61527705e-01 -9.42664921e-01 8.61628428e-02 -8.28978539e-01
1.03902690e-01 9.96273160e-01 3.49548817e-01 -6.89147651e-01
7.76754439e-01 3.59458566e-01 -3.52316827e-01 -8.86985004e-01
-5.44344008e-01 -1.41606361e-01 4.32856649e-01 4.53549236e-01
1.31891325e-01 1.04278684e-01 1.22616865e-01 2.31846288e-01
-1.65335909e-01 -4.74998266e-01 8.12694356e-02 8.29559326e-01
-3.89644802e-01 -7.48224139e-01 -1.20594472e-01 7.60501444e-01
-2.19791606e-01 -5.35071567e-02 -4.05166864e-01 6.99158669e-01
8.15765038e-02 4.65133280e-01 2.84410596e-01 -2.34701484e-01
4.34055418e-01 -3.46638352e-01 9.59460735e-01 -8.63696158e-01
-3.89501929e-01 -1.80520982e-01 4.25792448e-02 -7.01378226e-01
-1.44248277e-01 -3.45462859e-01 -1.34891951e+00 -3.08552414e-01
-8.78508538e-02 -1.17989533e-01 6.94137812e-01 6.97968006e-01
8.29824030e-01 6.17033005e-01 5.11645913e-01 -1.36038852e+00
-2.43268698e-01 -6.45086229e-01 -7.37309337e-01 5.75108111e-01
2.94794023e-01 -8.36355746e-01 -3.63736153e-01 -1.09688833e-01] | [8.527009963989258, -3.631091594696045] |
2b872471-761a-4a24-9f63-576a6744ec3e | exploiting-data-characteristics-for-document | null | null | https://openreview.net/forum?id=tVF2KuTau5g | https://openreview.net/pdf?id=tVF2KuTau5g | Exploiting Data Characteristics for Document-level Event Extraction | Document-level event extraction (DEE) extracts structured information of events from a document. Previous studies focus on improving the model architecture. We propose to exploit data characteristics: 1) we utilize more coreference information to obtain better document-level entity representations; 2) we manually identify core roles of each event type and propose the hybrid extraction to shallow the memory and alleviate error propagation. Experiments on a large dataset demonstrate that our methods significantly improve model performance on both the role-level and record-level metrics. Our code is available at https://github.com/coszeros/CAB.
| ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['document-level-event-extraction'] | ['natural-language-processing'] | [-3.65761598e-03 1.64286956e-01 -6.92402124e-01 -5.25118649e-01
-9.47778285e-01 -7.20025241e-01 6.93004489e-01 7.88708091e-01
-5.62408030e-01 7.01752841e-01 7.92484641e-01 -2.19524298e-02
-2.14349292e-02 -8.59050751e-01 -6.22181296e-01 -1.22047644e-02
-1.48577243e-01 2.80808479e-01 5.12136340e-01 1.37093171e-01
2.75690496e-01 4.32891935e-01 -1.40477645e+00 6.84953153e-01
6.17404401e-01 6.52310252e-01 -8.14184397e-02 3.77454072e-01
-1.82062432e-01 1.32739770e+00 -5.72368681e-01 -2.52137691e-01
-1.14761122e-01 -3.02486986e-01 -1.01463294e+00 -4.09028083e-01
-7.42819682e-02 -1.22640043e-01 -5.62577844e-01 7.92636991e-01
4.19768363e-01 1.36585772e-01 4.06827807e-01 -1.13156307e+00
-1.69654101e-01 1.19470608e+00 -5.42408764e-01 6.45153761e-01
4.42357510e-01 -4.64496374e-01 1.20364892e+00 -1.07104266e+00
1.04208386e+00 1.05055976e+00 5.26203871e-01 3.16855371e-01
-1.13977909e+00 -7.45092332e-01 2.29120314e-01 4.57953960e-01
-1.62579274e+00 -8.83532047e-01 6.87540233e-01 1.69515759e-02
1.46630192e+00 2.99413741e-01 2.72491068e-01 1.17136657e+00
1.32502183e-01 8.90558779e-01 5.24298608e-01 -5.46698511e-01
1.52545318e-01 -1.19491421e-01 7.53535867e-01 5.90040922e-01
5.44069052e-01 1.57422305e-03 -7.69968688e-01 -3.88011158e-01
3.58450115e-01 -2.21093725e-02 -2.29436561e-01 3.45654711e-02
-1.10223818e+00 6.26123130e-01 1.52899534e-01 5.90756893e-01
-6.87777102e-01 2.14835539e-01 5.75686812e-01 1.11628592e-01
1.83735102e-01 5.93113780e-01 -4.70884442e-01 -4.71739054e-01
-8.39985967e-01 3.89778018e-01 9.62843359e-01 1.08785152e+00
4.01517749e-01 -3.24789196e-01 -2.85485804e-01 8.99717987e-01
2.07562670e-01 -1.65940505e-02 4.78701949e-01 -1.03422379e+00
6.75565124e-01 8.92022550e-01 1.80635795e-01 -9.06793773e-01
-6.54165566e-01 -3.49117100e-01 -3.44094127e-01 -5.91103435e-01
-2.70141549e-02 -4.13438641e-02 -6.53722346e-01 1.71143901e+00
1.48289129e-01 1.97018310e-01 6.01858050e-02 5.40750563e-01
1.10560977e+00 6.82983160e-01 4.32714552e-01 -2.46677116e-01
1.57193220e+00 -7.28322148e-01 -1.18856204e+00 -4.75905269e-01
1.00357115e+00 -6.12922311e-01 4.55653787e-01 5.80059402e-02
-1.39569867e+00 -1.59026921e-01 -1.04929566e+00 -1.78558528e-01
-3.66644025e-01 2.56101578e-01 7.06119180e-01 2.53032923e-01
-5.23967028e-01 2.23600239e-01 -1.27681065e+00 -2.38817573e-01
2.25101843e-01 1.52012244e-01 -3.72192532e-01 1.86237141e-01
-1.50258410e+00 8.63590419e-01 7.59677529e-01 -1.88561857e-01
-4.88110781e-01 -5.77428520e-01 -1.00015485e+00 3.82908255e-01
6.41552746e-01 -4.57706273e-01 1.50780976e+00 -2.51187831e-01
-7.39679039e-01 6.15912616e-01 -7.21168876e-01 -5.37164032e-01
-1.69898316e-01 -4.10867333e-01 -6.51830971e-01 2.68984765e-01
1.24739662e-01 4.05590981e-01 -7.01143518e-02 -1.10927308e+00
-8.83802772e-01 -2.36523449e-01 3.00611109e-02 1.15878031e-01
-5.16376138e-01 4.89134669e-01 -8.10813427e-01 -9.28529799e-01
1.59683719e-01 -7.46308863e-01 -6.58213198e-02 -5.48699439e-01
-2.14051977e-01 -3.22945744e-01 5.50089121e-01 -6.19702876e-01
2.04032278e+00 -2.06152010e+00 -5.90830892e-02 1.10535048e-01
3.03471982e-01 -6.78100111e-03 5.37633374e-02 7.62781680e-01
-2.21940607e-01 2.81637698e-01 -1.11526169e-01 -4.19839740e-01
-2.76085138e-01 1.79994956e-01 -3.65982682e-01 5.72705716e-02
6.50366694e-02 8.10080409e-01 -8.37264538e-01 -8.27854753e-01
-2.79784530e-01 1.85103506e-01 -6.94822729e-01 1.00291021e-01
-1.25135615e-01 -1.63551271e-01 -6.88628793e-01 5.58535159e-01
3.79754990e-01 -4.37845290e-01 6.45964205e-01 -3.96709770e-01
-1.12320393e-01 1.13734019e+00 -1.48966277e+00 1.56543267e+00
-2.61086851e-01 4.25697029e-01 -1.43496469e-01 -7.46892273e-01
6.57356143e-01 5.15622437e-01 5.45409322e-01 -5.85743427e-01
-5.00431582e-02 -1.59881338e-02 -7.88378865e-02 -3.52851540e-01
8.90239120e-01 2.83415645e-01 -3.80837977e-01 4.83198047e-01
9.20373574e-02 5.28990448e-01 5.15708447e-01 6.88735247e-01
1.41971219e+00 -1.69370040e-01 7.36339808e-01 -1.16517685e-01
2.95793176e-01 1.53089494e-01 1.13355303e+00 7.69735873e-01
1.65085405e-01 3.08718234e-01 5.95406413e-01 -6.43825457e-02
-4.82020140e-01 -8.44740570e-01 -1.53964907e-01 1.12028527e+00
9.11135077e-02 -1.29206252e+00 -4.34946865e-01 -7.70537913e-01
-6.74735233e-02 1.19102752e+00 -4.97457385e-01 -2.59985149e-01
-1.02378666e+00 -8.01865458e-01 8.99873912e-01 1.09885609e+00
2.38404334e-01 -8.89945984e-01 -6.47451580e-01 3.84796053e-01
-6.84115112e-01 -1.09586203e+00 -4.05000418e-01 3.37302357e-01
-9.15391743e-01 -1.06311285e+00 1.60095379e-01 -5.08253038e-01
4.55126524e-01 1.70427244e-02 1.20597339e+00 2.45657662e-04
-1.32344425e-01 1.28319070e-01 -5.88942647e-01 -3.85399669e-01
-3.52811217e-01 3.15276921e-01 -1.31811574e-01 -3.50914180e-01
8.64542246e-01 -5.70556045e-01 -3.84310901e-01 3.29781413e-01
-8.53935063e-01 -6.08996637e-02 4.68019873e-01 5.04302442e-01
6.22855723e-01 1.17594814e-02 6.57023609e-01 -1.12312782e+00
7.75877535e-01 -7.66048074e-01 -4.64997143e-01 3.35697144e-01
-9.39867318e-01 2.10902482e-01 2.25956082e-01 -2.57591635e-01
-1.35955608e+00 -1.77356437e-01 5.16612530e-02 -1.70569628e-01
-2.23693103e-01 8.77964556e-01 -4.72363792e-02 7.78361142e-01
6.84835851e-01 6.35969490e-02 -6.92167103e-01 -6.15007937e-01
1.39255911e-01 7.07764387e-01 6.22636139e-01 -8.88319552e-01
2.93793112e-01 3.48815829e-01 -2.99645096e-01 -4.21973020e-01
-7.71415651e-01 -6.36735141e-01 -3.63986909e-01 1.52783483e-01
5.59500098e-01 -1.00039625e+00 -5.12014687e-01 -1.06571153e-01
-1.24210632e+00 -3.51034664e-02 -1.47422463e-01 6.95333421e-01
-1.20897405e-02 1.13233425e-01 -8.44278634e-01 -5.78364193e-01
-3.69555742e-01 -7.08638489e-01 9.65735555e-01 2.66181797e-01
-6.56843126e-01 -6.89136386e-01 3.53905916e-01 6.35425150e-02
1.37500744e-02 3.68939340e-02 8.53200078e-01 -9.72698927e-01
-3.88177514e-01 -1.73396394e-01 -1.51687562e-01 -3.63489270e-01
6.43351823e-02 -1.18910447e-01 -6.19870842e-01 -1.31578386e-01
-1.66989475e-01 -9.99304131e-02 9.69418347e-01 1.71075299e-01
1.21888685e+00 -3.71996671e-01 -9.09405589e-01 4.55793619e-01
1.03734696e+00 5.35534799e-01 5.25786519e-01 4.32999462e-01
4.32351381e-01 3.48854989e-01 7.22695708e-01 7.22133219e-01
6.78878307e-01 8.29975009e-01 -8.52337703e-02 2.63727725e-01
-2.06504822e-01 -4.60344374e-01 2.32894778e-01 7.56917894e-01
-7.30532184e-02 -4.94773835e-01 -1.18858504e+00 9.64501977e-01
-2.13998055e+00 -1.17115927e+00 -1.35143369e-01 1.82255328e+00
1.08291996e+00 3.91187042e-01 -2.61256266e-02 1.66567102e-01
6.51386797e-01 2.70266086e-01 -2.72136867e-01 -2.16589183e-01
-2.61880308e-02 1.03740118e-01 5.89694977e-01 3.12256157e-01
-1.09795821e+00 9.47102845e-01 6.28424788e+00 4.34150159e-01
-6.09008610e-01 1.65327325e-01 6.24064496e-03 -5.45682132e-01
-3.07127506e-01 1.93351492e-01 -1.35903609e+00 4.62220252e-01
1.36129296e+00 -5.20017147e-01 3.84849869e-02 6.22222364e-01
8.94892439e-02 7.43748294e-03 -1.36360276e+00 7.38628268e-01
1.38063394e-02 -1.68016768e+00 7.27995932e-02 3.60039435e-02
2.65628636e-01 9.82908756e-02 -5.42292118e-01 4.84364361e-01
4.52069610e-01 -5.47738135e-01 7.47727931e-01 5.81509352e-01
4.02536005e-01 -7.34350264e-01 5.12796819e-01 2.49994054e-01
-1.53504288e+00 -1.30327508e-01 -2.50959639e-02 1.64482579e-01
4.43486422e-01 4.98748600e-01 -9.57003117e-01 7.54251659e-01
6.81752741e-01 8.72576118e-01 -7.46469498e-01 9.29902732e-01
-3.90359759e-01 9.60604548e-01 -4.64765966e-01 1.44300073e-01
-7.86843747e-02 4.19047445e-01 7.47045100e-01 1.76277649e+00
-5.25311520e-03 4.69130725e-01 6.38688207e-02 8.44845831e-01
-3.68708074e-01 -1.77180812e-01 -3.70062739e-01 -2.51830131e-01
1.25372410e+00 1.05396533e+00 -6.43521130e-01 -3.64256948e-01
-5.14922619e-01 6.19019210e-01 4.05022711e-01 2.51314789e-01
-8.86852264e-01 -7.01268435e-01 5.81329644e-01 1.23200938e-01
3.78095567e-01 -1.16708890e-01 -1.60530120e-01 -1.32837653e+00
7.11128563e-02 -8.50347340e-01 1.17066705e+00 -6.28526747e-01
-6.44139707e-01 4.29709315e-01 4.62464601e-01 -8.76771748e-01
-6.88584626e-01 -9.23653468e-02 -4.89203334e-01 4.15114820e-01
-1.10866666e+00 -7.93470502e-01 -1.51271686e-01 5.11477292e-01
4.88129437e-01 2.01388359e-01 9.50562716e-01 6.76160991e-01
-8.22766483e-01 7.27464139e-01 -3.11364114e-01 5.74381530e-01
7.90479124e-01 -1.08106720e+00 4.40424085e-01 1.07234418e+00
3.67298871e-01 1.17919910e+00 4.84530568e-01 -9.12711859e-01
-1.25897479e+00 -9.56027508e-01 1.46422970e+00 -5.28851092e-01
5.15668809e-01 -3.11974287e-01 -9.35672641e-01 1.34329736e+00
1.05136357e-01 -2.29593128e-01 8.86081100e-01 5.46933055e-01
-5.24727345e-01 3.13014165e-02 -9.32175696e-01 4.80655849e-01
1.21807659e+00 -5.57909548e-01 -1.20888913e+00 1.27316508e-02
6.05609417e-01 -3.88510168e-01 -1.16866791e+00 4.81411576e-01
4.51654255e-01 -3.04899216e-01 7.57016838e-01 -7.47450233e-01
3.77728105e-01 -4.02637422e-01 -2.49616295e-01 -9.53970492e-01
-3.41180831e-01 -4.32109445e-01 -9.09283876e-01 1.49450409e+00
7.28502095e-01 -5.31055450e-01 4.20355022e-01 7.40426779e-01
-1.03778623e-01 -5.25967002e-01 -6.47513568e-01 -6.85285926e-01
-4.82090116e-01 -4.05767649e-01 7.70179451e-01 1.10574090e+00
4.62757260e-01 5.88194966e-01 -1.14095606e-01 4.96546537e-01
2.81167984e-01 2.96355009e-01 3.24695528e-01 -1.23522854e+00
-2.79874682e-01 -3.15871149e-01 3.98150124e-02 -8.53855133e-01
2.11189002e-01 -9.80241239e-01 -1.39252990e-01 -1.70662689e+00
4.40629065e-01 -2.66774297e-01 -6.11484766e-01 9.20397460e-01
-1.95535719e-01 -2.75084615e-01 -3.68710011e-02 3.10933381e-01
-8.70268583e-01 2.30685890e-01 3.77152711e-01 1.14442013e-01
-2.68948168e-01 -3.00213784e-01 -9.19376135e-01 7.35674560e-01
8.67305756e-01 -1.01458848e+00 -2.59271801e-01 -5.94534159e-01
3.14050734e-01 3.01903874e-01 5.65070398e-02 -8.19224179e-01
5.63832402e-01 -5.70990890e-02 3.17632377e-01 -7.60250390e-01
2.52882719e-01 -5.69432139e-01 2.20392659e-01 3.66272539e-01
-7.14348197e-01 1.74466282e-01 4.52953786e-01 4.49236721e-01
-3.82980853e-01 -4.81255740e-01 4.04511124e-01 3.93156745e-02
-7.35713065e-01 -1.93423912e-01 -4.28066313e-01 1.23654135e-01
7.78007150e-01 2.99961090e-01 -6.01232588e-01 -3.27450305e-01
-6.57003343e-01 3.12199354e-01 2.70294607e-01 6.27092063e-01
4.85435724e-01 -1.28139150e+00 -5.69342732e-01 -7.75937885e-02
4.28656518e-01 -2.71035016e-01 -1.63850337e-01 5.93554318e-01
-7.43941441e-02 6.92914784e-01 9.54768062e-02 -1.51904196e-01
-1.44594705e+00 3.35048109e-01 2.21985355e-01 -6.04920149e-01
-5.66575706e-01 7.00541615e-01 -6.29090443e-02 -1.45532653e-01
2.54102975e-01 -3.25073808e-01 -4.11211640e-01 2.27274671e-01
8.38028193e-01 5.15197515e-01 2.73599118e-01 -4.47884113e-01
-8.62313867e-01 -7.08561689e-02 -5.57886779e-01 -2.43899122e-01
1.51533163e+00 6.88314289e-02 -5.20632416e-02 2.48888060e-01
9.85023499e-01 2.39292473e-01 -6.94027781e-01 -2.79326051e-01
5.73138356e-01 -2.97389746e-01 4.01912689e-01 -9.29413736e-01
-7.37396061e-01 1.63821191e-01 1.95098951e-01 2.30984744e-02
1.12367105e+00 2.36782238e-01 6.99976742e-01 5.71246266e-01
2.51983523e-01 -1.02385843e+00 -3.93839955e-01 4.35803115e-01
7.28445113e-01 -7.02635467e-01 2.06801340e-01 -3.86945277e-01
-5.29143155e-01 7.32832134e-01 6.41647696e-01 1.87076777e-01
6.28194094e-01 7.02816784e-01 -3.02809387e-01 -4.22536641e-01
-1.26859736e+00 -1.25468776e-01 1.96383163e-01 -1.64484587e-02
8.61392856e-01 -1.17128655e-01 -7.94537723e-01 1.19449162e+00
8.17433372e-02 2.09429234e-01 4.51635003e-01 1.37948501e+00
-2.73252666e-01 -1.41001928e+00 -1.70431912e-01 4.99412566e-01
-7.89631784e-01 -1.69216663e-01 -5.56058466e-01 7.73539364e-01
-1.59838513e-01 1.06677020e+00 1.19565040e-01 -2.03275308e-01
6.99414372e-01 3.11859369e-01 3.36425394e-01 -7.74873674e-01
-7.37162113e-01 8.23666677e-02 7.74541557e-01 -8.45008969e-01
-4.81488317e-01 -8.04675460e-01 -1.69218719e+00 -1.23890266e-01
-1.86558694e-01 4.71640021e-01 3.45356643e-01 6.08731091e-01
9.30417955e-01 7.41090477e-01 2.24894375e-01 -2.46223226e-01
-1.73200637e-01 -1.07168722e+00 -2.46956602e-01 4.12930518e-01
1.80918351e-02 -7.57190049e-01 -1.03731766e-01 1.91012174e-01] | [9.167539596557617, 9.20262622833252] |
a834a685-d2f7-427f-8138-c178f0299184 | a-gold-standard-to-measure-relative | null | null | https://aclanthology.org/W18-4601 | https://aclanthology.org/W18-4601.pdf | A Gold Standard to Measure Relative Linguistic Complexity with a Grounded Language Learning Model | This paper focuses on linguistic complexity from a relative perspective. It presents a grounded language learning system that can be used to study linguistic complexity from a developmental point of view and introduces a tool for generating a gold standard in order to evaluate the performance of the learning system. In general, researchers agree that it is more feasible to approach complexity from an objective or theory-oriented viewpoint than from a subjective or user-related point of view. Studies that have adopted a relative complexity approach have showed some preferences for L2 learners. In this paper, we try to show that computational models of the process of language acquisition may be an important tool to consider children and the process of first language acquisition as suitable candidates for evaluating the complexity of languages. | ["M. Dolores Jim{\\'e}nez-L{\\'o}pez", 'Leonor Becerra-Bonache', 'Henning Christiansen'] | 2018-08-01 | null | null | null | ws-2018-8 | ['grounded-language-learning'] | ['natural-language-processing'] | [-3.07367086e-01 2.67800897e-01 -1.42447904e-01 -4.66931850e-01
-3.88033628e-01 -5.21205783e-01 3.86812985e-01 1.05100572e+00
-6.62808657e-01 1.17074572e-01 2.29814291e-01 -6.48877263e-01
-4.40331042e-01 -8.87331605e-01 -3.87094617e-01 -1.13944344e-01
1.36895075e-01 4.78900015e-01 2.01898977e-01 -4.95069116e-01
5.91301322e-01 6.46392941e-01 -1.93958044e+00 2.33067781e-01
1.20799029e+00 -4.33802754e-02 5.55857897e-01 1.97303697e-01
-6.05103612e-01 9.01715815e-01 -5.94297409e-01 -4.54801023e-01
-2.12562189e-01 -4.27946955e-01 -1.07814002e+00 -8.48409254e-03
1.76357388e-01 -2.22018883e-01 5.29211700e-01 1.16639602e+00
4.99291837e-01 -1.33620411e-01 4.17809516e-01 -7.02945054e-01
-4.98725742e-01 1.10394704e+00 4.37555641e-01 1.68296739e-01
9.91037428e-01 1.13727115e-01 6.27647281e-01 -6.17949843e-01
5.03769279e-01 1.46412492e+00 3.36582601e-01 5.18929183e-01
-1.09863567e+00 -3.11898112e-01 3.55086952e-01 2.26249173e-01
-1.22270548e+00 -2.59951651e-01 6.34517491e-01 -9.34281766e-01
8.31791997e-01 -1.53220370e-02 1.21084702e+00 3.85992676e-01
-4.73196208e-02 3.73079479e-01 1.63337767e+00 -1.18167603e+00
3.35176080e-01 7.76408195e-01 3.71792197e-01 5.39041162e-01
8.00210595e-01 1.20676406e-01 -6.90999627e-01 4.32106286e-01
4.02179867e-01 -7.17291832e-01 1.08740792e-01 -1.87086672e-01
-6.20038271e-01 8.52999926e-01 -5.60545400e-02 1.28216743e+00
-6.18897676e-02 -3.15029263e-01 3.94729942e-01 5.74346900e-01
2.27559134e-01 8.53370547e-01 -2.23738790e-01 -4.78022844e-01
-6.84379697e-01 1.35460332e-01 8.92327070e-01 5.38258135e-01
2.61140168e-01 1.36002615e-01 1.10837571e-01 5.46697557e-01
9.26906347e-01 2.38079339e-01 4.39416230e-01 -7.15479791e-01
5.24965189e-02 1.07000470e+00 -4.32856709e-01 -6.37802839e-01
-1.80644900e-01 -1.06511921e-01 3.03882033e-01 7.80074179e-01
6.38175011e-01 1.20981976e-01 -4.31298018e-01 1.72926188e+00
1.89673543e-01 -7.37390578e-01 1.27635568e-01 4.14057195e-01
9.75296140e-01 5.01460850e-01 4.55082089e-01 -4.52183664e-01
1.30657434e+00 -3.66133958e-01 -6.68142080e-01 2.46507511e-01
1.19748139e+00 -7.37713873e-01 1.64477789e+00 7.15646684e-01
-1.60855663e+00 -4.42330569e-01 -1.11073840e+00 -1.51209980e-01
-6.34143531e-01 -2.45502349e-02 5.39085805e-01 1.30795181e+00
-1.28046882e+00 4.70406204e-01 -8.35932493e-01 -6.14850163e-01
-3.96911770e-01 2.32234672e-01 -1.56393945e-02 6.64396137e-02
-9.19353545e-01 1.29139078e+00 7.37806141e-01 -3.17069367e-02
-5.75661778e-01 -2.91262895e-01 -8.16954792e-01 3.91596556e-02
5.80364577e-02 -2.27496937e-01 1.07503378e+00 -1.10003293e+00
-1.60057735e+00 1.27829123e+00 1.07792914e-01 1.69902958e-03
3.07566494e-01 -3.92360203e-02 -2.76232839e-01 1.26363814e-01
1.75162554e-01 1.40859131e-02 2.19246466e-02 -1.15574419e+00
-5.56814849e-01 -2.02296793e-01 6.94364250e-01 3.99901003e-01
-3.74470443e-01 6.75413907e-01 2.26393864e-02 -4.73268807e-01
1.50397912e-01 -5.88542700e-01 1.80377048e-02 -1.41768754e-01
4.65118259e-01 -5.85544944e-01 -1.08306393e-01 -4.42417830e-01
1.74130833e+00 -1.87556624e+00 7.96586499e-02 1.55695960e-01
3.01648021e-01 5.75525284e-01 1.94354311e-01 7.16380537e-01
6.51836069e-03 5.94506979e-01 2.10843250e-01 8.94165859e-02
9.26705003e-02 -1.90354928e-01 5.28465509e-01 3.32014970e-02
-9.73380581e-02 2.40942791e-01 -1.30485368e+00 -6.00062370e-01
3.80345851e-01 4.73185718e-01 -5.96817434e-01 1.96396694e-01
-4.85136472e-02 3.06434214e-01 -2.36608371e-01 3.97007883e-01
2.38407761e-01 3.73981595e-01 5.48888028e-01 2.94094950e-01
-8.67451608e-01 7.37013280e-01 -1.11572373e+00 1.20096457e+00
-8.64084840e-01 8.12899292e-01 -1.91440657e-01 -8.31431031e-01
8.16695094e-01 8.25448394e-01 -8.75853375e-02 -7.30236590e-01
2.81663537e-01 6.38247550e-01 8.89919519e-01 -6.36065781e-01
4.55447957e-02 -3.41166317e-01 3.86268198e-02 4.88822818e-01
1.68191850e-01 -6.34408653e-01 5.20402372e-01 -3.04880679e-01
4.52567488e-01 1.97629631e-01 4.98125583e-01 -8.65497112e-01
1.13799298e+00 -2.48679981e-01 2.66059458e-01 2.14588508e-01
3.70938107e-02 -3.69333774e-01 3.63929003e-01 -4.43313271e-01
-7.09846735e-01 -9.89400685e-01 -2.14296028e-01 9.29119885e-01
-3.72448534e-01 -4.55091804e-01 -1.27547336e+00 -1.12032093e-01
-8.19794893e-01 1.12730455e+00 -1.35839522e-01 5.24961501e-02
-5.56271255e-01 -8.47991109e-02 1.30040511e-01 -4.16634381e-02
2.41249889e-01 -1.20215833e+00 -1.00831831e+00 3.41894954e-01
2.66410261e-01 -7.06143856e-01 3.92674237e-01 -8.59326795e-02
-9.74777639e-01 -7.33492970e-01 -3.25240821e-01 -1.15016353e+00
9.07335281e-01 9.55331838e-04 1.05788076e+00 8.17459524e-01
7.80391842e-02 6.54902101e-01 -7.21449137e-01 -8.73510778e-01
-9.83905315e-01 -5.54895923e-02 6.20290339e-02 -6.80551946e-01
7.10652471e-01 -4.96547550e-01 -3.57213281e-02 -2.65653044e-01
-8.91815066e-01 -8.90313387e-02 1.90822646e-01 -1.17515502e-02
1.97173227e-02 3.78173053e-01 2.72083879e-01 -6.39943242e-01
1.04456449e+00 -7.04777315e-02 -7.07506180e-01 5.23533642e-01
-8.56479585e-01 -2.58503063e-03 4.20399338e-01 -2.80902088e-01
-1.09666765e+00 -4.48761731e-01 -4.18465406e-01 4.04475272e-01
-4.78250891e-01 8.35799515e-01 -3.53746355e-01 -3.59760642e-01
7.30770707e-01 7.08238035e-02 -6.74210116e-02 -3.72105807e-01
-9.68290865e-02 4.19736683e-01 -1.48068801e-01 -1.10018468e+00
8.99011672e-01 -2.79168159e-01 -9.03136060e-02 -1.26581264e+00
-5.23416281e-01 -2.40799055e-01 -9.25124884e-01 -6.71479523e-01
8.74599814e-01 -8.59078169e-01 -6.59056544e-01 5.07458508e-01
-9.89397228e-01 -5.31177640e-01 -3.70788604e-01 8.67926300e-01
-5.21595240e-01 1.74848378e-01 -2.96250284e-01 -1.04416192e+00
-9.99695212e-02 -1.23058534e+00 4.19652760e-01 3.52054656e-01
-4.21855152e-01 -1.58416164e+00 -2.13685147e-02 2.85450518e-01
2.86757737e-01 2.43765995e-01 1.39809537e+00 -6.94518149e-01
-5.07829428e-01 1.75284117e-01 5.57097554e-01 3.35421205e-01
1.66497707e-01 3.78935099e-01 -4.74194199e-01 -8.15121531e-02
4.03047055e-01 -3.32467645e-01 -1.52389184e-01 -5.94443083e-02
3.92493099e-01 -2.94746250e-01 2.21336812e-01 1.41230658e-01
1.94396305e+00 6.13777995e-01 5.01371622e-01 6.12500846e-01
2.64556885e-01 1.15742409e+00 7.63530791e-01 -4.17390205e-02
5.73816597e-01 5.77545404e-01 -2.56939977e-01 8.52115676e-02
-1.12238765e-01 -2.10995212e-01 4.79600787e-01 1.57948816e+00
-1.79866105e-01 -1.34111347e-03 -1.68511260e+00 5.35211921e-01
-1.05220425e+00 -9.35369432e-01 -1.35804057e-01 2.46108150e+00
8.83929014e-01 5.23721516e-01 4.39349651e-01 6.81509435e-01
6.04118645e-01 -3.29012036e-01 5.46711624e-01 -1.15311110e+00
2.98083156e-01 4.50758010e-01 -3.61599505e-01 1.05537105e+00
-1.33522615e-01 9.81328964e-01 6.62271500e+00 4.27895546e-01
-1.42973292e+00 5.12296483e-02 2.25783467e-01 3.83654028e-01
-7.53751457e-01 1.19488858e-01 -7.06467211e-01 4.22880352e-01
9.85600531e-01 -7.28344619e-01 2.72986323e-01 4.76787210e-01
4.25675988e-01 -3.32426846e-01 -1.14706147e+00 3.56982410e-01
-1.19096234e-01 -7.46916234e-01 2.60788500e-01 9.01852995e-02
2.16511130e-01 -4.28535134e-01 -3.55881788e-02 1.57217294e-01
-8.88624191e-02 -1.06264794e+00 1.21974313e+00 5.02893925e-01
5.32529414e-01 -5.50878108e-01 5.91479003e-01 5.73977292e-01
-1.09363925e+00 2.34012872e-01 1.18566267e-01 -8.40573668e-01
-9.68178883e-02 -4.33405414e-02 -8.98405492e-01 -4.47912887e-03
4.70483184e-01 -3.06824654e-01 -7.38907158e-01 1.01590896e+00
-3.74374449e-01 6.19704664e-01 -3.03293765e-01 -6.70310318e-01
5.63588068e-02 -3.12830031e-01 3.76144052e-01 1.01183450e+00
4.00217593e-01 2.29170009e-01 -4.07518148e-02 7.91922987e-01
9.03678179e-01 9.19126511e-01 -7.81104207e-01 -2.26858482e-01
6.98315144e-01 7.67987490e-01 -1.21816969e+00 -1.59862861e-01
-6.41498446e-01 1.40008405e-01 2.18380943e-01 -2.49140471e-01
-2.19899446e-01 -2.54692793e-01 1.11682735e-01 5.04209459e-01
-1.64427191e-01 -4.12154466e-01 -3.00728887e-01 -9.75366473e-01
-7.10514486e-02 -1.12959063e+00 -1.59585059e-01 -3.16029400e-01
-5.45496404e-01 5.57787001e-01 5.52950382e-01 -9.03760076e-01
-3.03953886e-01 -7.93495834e-01 -7.50148296e-01 1.04608166e+00
-8.05313826e-01 -9.07560587e-01 -1.15887970e-01 8.05218816e-02
4.96947706e-01 -1.06555477e-01 9.32688594e-01 1.29513498e-02
-2.27598950e-01 3.24754298e-01 -4.14228797e-01 -2.01514155e-01
-1.90778002e-01 -1.38214421e+00 2.29080953e-02 1.00497913e+00
-3.90311666e-02 8.77345443e-01 9.27130938e-01 -5.43335319e-01
-9.24710810e-01 -5.89048974e-02 1.60316014e+00 -2.53661245e-01
3.97739559e-01 -1.95053458e-01 -8.95341456e-01 2.02907637e-01
3.25419515e-01 -7.85000801e-01 1.12053084e+00 7.87097886e-02
2.80789703e-01 7.49872476e-02 -1.25259197e+00 7.50725508e-01
9.72238958e-01 -6.61621988e-01 -1.06845653e+00 8.76671672e-02
5.29127002e-01 -7.91789815e-02 -1.27656031e+00 2.63018876e-01
4.19111997e-01 -1.33215582e+00 5.75211406e-01 1.58095255e-01
1.71554580e-01 -1.07756585e-01 6.46753833e-02 -9.72043157e-01
-2.65204966e-01 -5.32579720e-01 6.41175747e-01 1.60321021e+00
4.90454108e-01 -6.13363206e-01 4.61008579e-01 5.91084599e-01
3.80488411e-02 -7.30943561e-01 -8.70999157e-01 -5.56467950e-01
6.46859825e-01 -5.49216688e-01 5.13293326e-01 9.02092695e-01
3.46783817e-01 3.20508838e-01 5.97875178e-01 -5.67065505e-03
2.66504109e-01 -6.11610651e-01 4.51319873e-01 -1.66259229e+00
1.06790895e-02 -8.65526259e-01 -6.27593338e-01 -1.44896181e-02
4.64976244e-02 -7.34518111e-01 -2.21557513e-01 -1.70196593e+00
-4.72220063e-01 -3.06170464e-01 -1.96737293e-02 -2.13247061e-01
3.40805436e-03 -4.48921144e-01 6.36667430e-01 -2.27327690e-01
-7.47536309e-03 -1.34245321e-01 9.65605021e-01 4.59713459e-01
-4.58491594e-01 -1.00471161e-01 -8.50057960e-01 1.16940725e+00
9.25963283e-01 -4.08345193e-01 -8.24222922e-01 -6.16690636e-01
7.30662644e-01 -1.31826535e-01 -3.08503389e-01 -1.30059874e+00
6.69411346e-02 -4.67108577e-01 -2.68427938e-01 -8.18261877e-02
-2.70381212e-01 -1.04906619e+00 1.05567545e-01 9.00043666e-01
-3.53837878e-01 2.47196719e-01 2.18118772e-01 -5.78211844e-01
-2.14025974e-01 -1.17315722e+00 5.99314809e-01 -4.10853237e-01
-6.05262995e-01 -2.36756265e-01 -8.03762197e-01 1.13391198e-01
1.14252257e+00 -6.29102111e-01 1.93175197e-01 -2.13504955e-01
-6.55975103e-01 9.92597640e-02 8.54960561e-01 1.45559385e-01
4.89848197e-01 -1.01943755e+00 -4.63143438e-01 -1.32101506e-01
1.78090319e-01 -3.53190392e-01 -1.50645837e-01 3.37728322e-01
-1.39642560e+00 5.30406296e-01 -4.56721008e-01 -2.34567776e-01
-1.30397737e+00 7.10120320e-01 4.59637940e-01 -1.54839888e-01
-1.74818650e-01 6.83391809e-01 -5.06814457e-02 -3.61990899e-01
2.75114596e-01 -5.11043429e-01 -8.49571824e-01 4.25045937e-01
4.71165597e-01 5.19873321e-01 -7.53348321e-02 -9.88770366e-01
-1.78578719e-01 1.04811645e+00 3.48197252e-01 -5.65699518e-01
9.85793769e-01 -1.92393094e-01 -3.46275121e-01 1.07268393e+00
8.27335417e-01 5.86147785e-01 -1.79644421e-01 6.56329608e-03
6.10855758e-01 -3.86575997e-01 -5.62513713e-03 -7.07963943e-01
-3.90033454e-01 9.89765346e-01 7.10327983e-01 4.86277431e-01
1.19162881e+00 -6.38906136e-02 -2.88377285e-01 4.51971024e-01
5.17922938e-01 -1.42535210e+00 -2.28210002e-01 3.73450667e-01
9.94769812e-01 -7.80422866e-01 1.07503638e-01 -7.62039542e-01
-1.56270057e-01 1.46391451e+00 6.71621442e-01 1.17893949e-01
8.60033035e-01 1.42467633e-01 2.50600070e-01 -1.65475845e-01
-8.44669580e-01 -6.39834702e-01 3.26213360e-01 8.98755014e-01
1.18315434e+00 1.29509434e-01 -1.44827747e+00 2.92745650e-01
-6.48106277e-01 1.34568453e-01 7.05561519e-01 1.05875576e+00
-8.29252124e-01 -1.63503134e+00 -4.92925882e-01 -1.00825066e-02
-7.29339838e-01 -5.73829226e-02 -4.09834236e-01 1.02841687e+00
6.90377653e-01 9.51040208e-01 -1.83378264e-01 -4.98064533e-02
5.32001853e-01 1.81095704e-01 8.33944321e-01 -1.19851935e+00
-1.01411164e+00 5.94504774e-02 3.63500267e-01 -8.77939388e-02
-6.99574411e-01 -7.49242306e-01 -1.01075184e+00 -2.87292153e-01
-1.32438168e-01 3.87358606e-01 7.70143926e-01 1.07115817e+00
-5.09115994e-01 3.40732664e-01 3.41737092e-01 -2.29790673e-01
-1.13475852e-01 -7.94396162e-01 -3.26142192e-01 -5.53079508e-02
-2.82792281e-03 -3.77695858e-01 -3.11189264e-01 -1.21636741e-01] | [10.570271492004395, 9.936635971069336] |
074173dc-23b9-4333-8d4d-999155022882 | multi-granularity-prediction-for-scene-text | 2209.03592 | null | https://arxiv.org/abs/2209.03592v2 | https://arxiv.org/pdf/2209.03592v2.pdf | Multi-Granularity Prediction for Scene Text Recognition | Scene text recognition (STR) has been an active research topic in computer vision for years. To tackle this challenging problem, numerous innovative methods have been successively proposed and incorporating linguistic knowledge into STR models has recently become a prominent trend. In this work, we first draw inspiration from the recent progress in Vision Transformer (ViT) to construct a conceptually simple yet powerful vision STR model, which is built upon ViT and outperforms previous state-of-the-art models for scene text recognition, including both pure vision models and language-augmented methods. To integrate linguistic knowledge, we further propose a Multi-Granularity Prediction strategy to inject information from the language modality into the model in an implicit way, i.e. , subword representations (BPE and WordPiece) widely-used in NLP are introduced into the output space, in addition to the conventional character level representation, while no independent language model (LM) is adopted. The resultant algorithm (termed MGP-STR) is able to push the performance envelop of STR to an even higher level. Specifically, it achieves an average recognition accuracy of 93.35% on standard benchmarks. Code is available at https://github.com/AlibabaResearch/AdvancedLiterateMachinery/tree/main/OCR/MGP-STR. | ['Cong Yao', 'Cheng Da', 'Peng Wang'] | 2022-09-08 | null | null | null | null | ['scene-text-recognition'] | ['computer-vision'] | [ 4.57210779e-01 -4.51249987e-01 -1.69329330e-01 -1.29512146e-01
-7.57093787e-01 -3.66617769e-01 9.38572288e-01 1.05679035e-01
-5.36825359e-01 3.25081974e-01 6.27049953e-02 -3.01716924e-01
3.32724273e-01 -5.98945379e-01 -5.35277665e-01 -7.69936085e-01
9.12629902e-01 1.98945895e-01 3.80498528e-01 -7.66896009e-02
5.39988697e-01 3.14922452e-01 -1.46132755e+00 4.41063613e-01
1.03291202e+00 8.01748097e-01 5.53799093e-01 6.30805016e-01
-4.93217349e-01 1.07708907e+00 -3.00066262e-01 -5.61818957e-01
-9.40424800e-02 -3.76471043e-01 -4.82623279e-01 2.78999567e-01
4.23021883e-01 5.08023314e-02 -5.35721779e-01 1.01844966e+00
3.37994099e-01 -1.27026796e-01 7.18599617e-01 -8.31200123e-01
-8.72700989e-01 5.66878259e-01 -8.80372345e-01 2.55539063e-02
2.62741446e-01 2.59013146e-01 1.00888216e+00 -1.35986400e+00
2.84907311e-01 1.33015263e+00 5.36486864e-01 1.69854566e-01
-1.06510329e+00 -2.47693062e-01 3.20640504e-01 6.46586418e-01
-1.46588182e+00 -4.53180462e-01 9.42422211e-01 -4.09354925e-01
9.60122764e-01 2.00703278e-01 4.16760832e-01 1.29104984e+00
8.30178633e-02 1.29620492e+00 1.20982718e+00 -9.46456790e-01
-7.29574114e-02 2.19138488e-01 4.49593455e-01 8.43584836e-01
2.13957757e-01 -3.00093502e-01 -6.31829023e-01 2.83638060e-01
6.53946936e-01 -1.13571389e-02 -3.33213508e-01 -1.47885039e-01
-1.27338314e+00 7.07608402e-01 2.61121213e-01 5.53626716e-01
-2.49066353e-01 1.15195006e-01 4.82736707e-01 -8.74692798e-02
3.14173698e-01 -2.91249398e-02 -2.34697297e-01 -6.14790395e-02
-9.79332447e-01 -1.31512612e-01 6.78257585e-01 7.27098525e-01
6.76652491e-01 3.85413200e-01 -3.51226956e-01 1.16895115e+00
5.47887087e-01 7.09006131e-01 7.34326124e-01 -4.59013730e-01
6.24550641e-01 7.18242049e-01 -2.38040805e-01 -1.03673065e+00
-8.53210464e-02 -4.95438188e-01 -9.51718330e-01 -1.97672665e-01
2.67727822e-01 1.86978340e-01 -1.13165748e+00 1.10197389e+00
1.68445766e-01 2.24238634e-01 1.23282552e-01 7.00778723e-01
9.08381402e-01 9.05731857e-01 1.00264199e-01 -1.31756604e-01
1.57446098e+00 -1.24594116e+00 -6.34852231e-01 -4.47405428e-01
5.16774178e-01 -9.53069031e-01 1.10068572e+00 5.39668500e-01
-7.76872933e-01 -7.86441922e-01 -1.00068915e+00 -1.87581509e-01
-4.53653216e-01 6.14773810e-01 2.98321038e-01 6.54080451e-01
-1.09478199e+00 5.30319810e-02 -6.88329637e-01 -5.72145343e-01
2.87959486e-01 1.60520878e-02 -1.76496819e-01 -2.71645099e-01
-9.53193843e-01 9.44925904e-01 6.81706488e-01 2.90941864e-01
-6.18138671e-01 -1.53991386e-01 -9.01250064e-01 -1.71523854e-01
6.25974596e-01 -6.79690182e-01 1.14276469e+00 -9.74925756e-01
-1.63649642e+00 8.19687366e-01 -3.92082453e-01 -5.07651985e-01
4.80227083e-01 -3.16257238e-01 -3.54779124e-01 2.63883471e-01
-1.88761994e-01 4.99887288e-01 1.22773910e+00 -1.26185644e+00
-5.47947764e-01 -3.35204959e-01 -2.66605198e-01 4.25252914e-01
-5.02149999e-01 2.10170999e-01 -9.26205814e-01 -8.19457769e-01
-3.89935076e-02 -6.49756253e-01 -7.72695467e-02 -8.85727480e-02
-5.67706108e-01 -3.08551759e-01 7.75490344e-01 -6.87826455e-01
1.23046410e+00 -2.06422043e+00 1.83465123e-01 -1.57851920e-01
7.19515011e-02 7.37182081e-01 -7.83587769e-02 6.11232162e-01
2.25109577e-01 -1.86989784e-01 -4.54832941e-01 -5.60993910e-01
3.60089056e-02 1.18212454e-01 -3.82631779e-01 3.72244984e-01
1.22385532e-01 1.11773777e+00 -5.78175187e-01 -6.66886628e-01
6.76867783e-01 5.60548604e-01 -2.36201912e-01 -1.22813269e-01
-3.06430340e-01 1.31325588e-01 -6.42952204e-01 6.97138906e-01
7.22090662e-01 -3.07241738e-01 2.12480444e-02 -1.57964885e-01
-4.69734490e-01 -8.64846185e-02 -1.02636600e+00 1.46691179e+00
-2.35949516e-01 5.48243463e-01 -1.94205865e-01 -1.32404351e+00
1.18205404e+00 1.05299018e-01 1.03797577e-01 -6.86377466e-01
1.94524124e-01 1.29495978e-01 -2.91442752e-01 -5.97285092e-01
6.01077318e-01 6.62795175e-03 5.79042174e-02 2.56373677e-02
-7.37498254e-02 -3.20794731e-02 1.78108037e-01 -1.08427376e-01
7.68948853e-01 2.08706230e-01 4.44209069e-01 1.27233446e-01
1.17354262e+00 1.10605769e-01 2.89741129e-01 8.88373494e-01
-2.40217164e-01 5.48720062e-01 1.28290161e-01 -1.64661840e-01
-8.69412899e-01 -8.56423795e-01 -8.10435787e-02 9.75678623e-01
3.35572138e-02 -3.59709144e-01 -6.81722701e-01 -4.63622034e-01
-8.06378350e-02 7.90434837e-01 -2.89472997e-01 7.46437833e-02
-5.90431690e-01 -6.80361092e-01 7.97386944e-01 4.85222608e-01
9.24436510e-01 -1.24214339e+00 -3.48880589e-01 1.46583572e-01
-1.72751427e-01 -1.47249210e+00 -3.64438087e-01 3.31575871e-02
-7.68317699e-01 -6.83692634e-01 -9.29236233e-01 -8.84915590e-01
3.96057785e-01 4.25581038e-01 5.77343941e-01 -1.04110062e-01
-3.86976957e-01 6.25659704e-01 -6.46873534e-01 -3.75865251e-01
-3.19790930e-01 -4.22816128e-02 -1.96856096e-01 5.46803057e-01
5.50565302e-01 -1.82117552e-01 -3.83677602e-01 1.07549287e-01
-9.00916100e-01 4.55798984e-01 8.65923226e-01 8.45845282e-01
5.94602883e-01 -2.57741570e-01 3.44770670e-01 -6.96147799e-01
5.37182689e-01 -9.20482948e-02 -6.83753669e-01 4.44005400e-01
-3.43237013e-01 -9.94183496e-02 8.01307261e-01 -3.36963028e-01
-1.25234795e+00 1.59915462e-01 -3.59389395e-01 -5.13885021e-01
-4.73187089e-01 8.57353449e-01 -9.97591391e-02 -7.67615959e-02
2.55275637e-01 1.20173073e+00 -1.19593129e-01 -5.88994980e-01
4.81117636e-01 1.00512326e+00 5.21290779e-01 -4.55073327e-01
7.92564034e-01 4.11206990e-01 -3.20181429e-01 -1.43139124e+00
-8.46388161e-01 -6.78782463e-01 -7.51882493e-01 -1.46036267e-01
7.98136652e-01 -9.23362970e-01 -3.91289830e-01 1.01576161e+00
-1.19702840e+00 -2.57466942e-01 3.39148119e-02 4.80738282e-01
-4.82869118e-01 1.01206779e+00 -5.85637391e-01 -1.06161571e+00
-4.20589954e-01 -1.11087358e+00 1.09820390e+00 2.96592981e-01
3.49697411e-01 -9.65936840e-01 -5.21437684e-03 6.68462455e-01
2.10694134e-01 -1.98274031e-01 8.20626438e-01 -6.89957023e-01
-6.06287897e-01 -9.21562091e-02 -5.68280041e-01 4.91057992e-01
2.84894891e-02 1.30718008e-01 -1.09098721e+00 -1.49059162e-01
1.03603765e-01 -1.99163288e-01 1.20596814e+00 3.58856142e-01
1.09230924e+00 -8.66068006e-02 -1.90566003e-01 4.73622710e-01
1.58221626e+00 1.63073853e-01 7.80113697e-01 3.93265218e-01
9.94294822e-01 3.67013633e-01 4.63136315e-01 4.70239878e-01
6.41470909e-01 7.88317680e-01 2.29180098e-01 1.36712799e-02
-2.30951309e-01 -2.55259633e-01 6.62617207e-01 1.19442654e+00
4.11491022e-02 -4.67589468e-01 -1.11632347e+00 2.93043733e-01
-2.07240510e+00 -8.05687547e-01 -3.82319301e-01 2.05969453e+00
5.99259913e-01 2.01207370e-01 -1.13876976e-01 2.72548050e-01
8.31940651e-01 4.05388206e-01 -4.12002057e-01 -4.23494071e-01
-2.73564428e-01 3.39572951e-02 2.86993146e-01 4.68545467e-01
-1.19928432e+00 1.41763818e+00 5.26188231e+00 1.14281166e+00
-1.16152036e+00 -1.36911631e-01 3.81637692e-01 5.39327264e-01
-1.01833250e-02 -1.12092786e-01 -1.16559482e+00 3.85549664e-01
6.92560256e-01 -1.23617724e-01 1.57708362e-01 6.51627839e-01
3.31609666e-01 -1.39799044e-02 -7.32182562e-01 1.24278092e+00
5.25020957e-01 -1.22007930e+00 5.22988379e-01 -8.44674781e-02
3.48971784e-01 1.25690147e-01 1.07453004e-01 4.10641193e-01
-1.99997067e-01 -8.21657419e-01 7.09096849e-01 6.56973064e-01
7.58822262e-01 -3.95426661e-01 6.67373776e-01 6.70851529e-01
-1.40443420e+00 -9.00479630e-02 -4.41931278e-01 2.94461716e-02
7.43599376e-03 5.14320016e-01 -6.66509926e-01 8.93226624e-01
3.82030487e-01 1.00908899e+00 -8.48930001e-01 1.13985145e+00
-3.43164921e-01 8.04508090e-01 -2.18385816e-01 -1.77287176e-01
4.69201356e-01 -2.66108423e-01 5.60178041e-01 1.56830227e+00
1.95333749e-01 -2.14533105e-01 3.84808719e-01 6.54858649e-01
-3.10289534e-03 4.39223766e-01 -5.83647311e-01 -2.58965641e-02
2.04759479e-01 1.27904987e+00 -6.67239189e-01 -4.15887833e-01
-7.47174799e-01 1.08953738e+00 2.35607311e-01 3.69338065e-01
-8.31702232e-01 -2.22115517e-01 1.69676289e-01 -3.90732288e-01
7.11945355e-01 -4.11068439e-01 -2.40730450e-01 -1.39355516e+00
2.10614741e-01 -8.75742674e-01 3.70389074e-01 -8.62032831e-01
-1.26970971e+00 4.55030233e-01 -2.47679591e-01 -1.18722343e+00
6.27881363e-02 -9.43544388e-01 -4.68712717e-01 8.43738198e-01
-1.97330964e+00 -1.43887162e+00 -3.01027060e-01 7.54070044e-01
1.16840529e+00 -2.58338630e-01 6.66249812e-01 1.51216865e-01
-8.54116023e-01 7.24889278e-01 3.36358517e-01 2.93607771e-01
5.17256439e-01 -9.48472500e-01 3.67311001e-01 1.09937060e+00
5.75122833e-01 3.70457560e-01 4.24516916e-01 -5.55835366e-01
-1.84000397e+00 -1.09566259e+00 9.40861225e-01 -3.97509038e-01
8.70288491e-01 -3.26752871e-01 -1.04533863e+00 4.01014715e-01
2.37300187e-01 -2.43298322e-01 3.24424118e-01 -3.40938985e-01
-6.17128074e-01 -9.50020701e-02 -7.04889894e-01 9.38166797e-01
7.56579638e-01 -5.39032280e-01 -8.77427638e-01 9.98530015e-02
5.78825772e-01 -1.30591452e-01 -3.89228970e-01 2.73922294e-01
3.50056469e-01 -8.02237332e-01 9.29621458e-01 -3.99606228e-02
3.21295559e-01 -4.87550795e-01 -3.41217697e-01 -7.34360516e-01
-2.74947643e-01 -3.64197314e-01 -2.74385035e-01 1.28628075e+00
2.09931746e-01 -8.07809532e-01 5.46425462e-01 -1.83544369e-04
-2.98711210e-01 -6.73116982e-01 -9.19794023e-01 -7.21717000e-01
-4.02261643e-03 -6.73182428e-01 -8.55807215e-02 8.32376301e-01
-1.07246518e-01 5.26846707e-01 -4.32469904e-01 1.15831949e-01
7.18966961e-01 1.56300291e-01 6.21420562e-01 -1.04119062e+00
-3.16472441e-01 -6.83978021e-01 -3.40370506e-01 -1.51078284e+00
1.26298472e-01 -9.94897068e-01 1.88328587e-02 -1.80565071e+00
2.50507593e-01 -3.31938900e-02 -3.46201241e-01 5.99997103e-01
-2.46939078e-01 3.13232392e-01 5.29051900e-01 3.00929874e-01
-6.21729255e-01 8.18688035e-01 1.22678435e+00 -3.46946657e-01
1.64553188e-02 -6.24743290e-02 -6.12935364e-01 9.30725336e-01
7.83416331e-01 -2.22173631e-02 -1.71723887e-01 -5.55082083e-01
-1.44753665e-01 -4.19278704e-02 4.48293865e-01 -9.58633780e-01
5.76255500e-01 -1.06259376e-01 2.72193611e-01 -8.45345378e-01
5.20526826e-01 -7.16037452e-01 -2.55716085e-01 4.75440860e-01
-1.68404818e-01 -1.83200955e-01 1.57027349e-01 6.31880403e-01
-3.60375226e-01 -3.48008752e-01 7.35821486e-01 -9.96846706e-02
-1.06514001e+00 1.35563672e-01 -6.09594524e-01 -2.03425318e-01
7.78311372e-01 -4.96644557e-01 -5.10368884e-01 1.17955515e-02
-3.00588876e-01 1.17265575e-01 2.72063881e-01 4.94483322e-01
9.05019224e-01 -9.65143383e-01 -9.44049239e-01 1.23072537e-02
3.79110336e-01 -4.08227563e-01 3.19046199e-01 9.82754350e-01
-4.79725927e-01 7.14712322e-01 1.71503961e-01 -8.16076815e-01
-1.28292000e+00 6.47534251e-01 7.40659013e-02 -3.21559459e-01
-9.23627555e-01 5.94741881e-01 2.57382721e-01 -2.24740833e-01
2.75712430e-01 -1.11982644e-01 -4.92336392e-01 3.59419473e-02
5.07454038e-01 1.56121746e-01 -2.65865438e-02 -8.32155645e-01
-3.86590719e-01 9.25855219e-01 -3.55808139e-01 3.74435000e-02
1.09791708e+00 -2.22438261e-01 -4.02328670e-02 6.76693916e-01
7.79227376e-01 -1.20135777e-01 -9.46714640e-01 -5.90633631e-01
2.12942138e-01 -3.18093747e-01 1.07971393e-01 -7.68827677e-01
-6.67068839e-01 1.09118283e+00 3.75997692e-01 1.76332042e-01
1.04244089e+00 -6.71758950e-02 6.09559596e-01 6.20019376e-01
3.42345834e-01 -1.02268946e+00 2.06429109e-01 8.88508737e-01
8.52823794e-01 -1.21914744e+00 -5.69745712e-02 -4.53027487e-01
-8.32273781e-01 1.21933973e+00 4.18958902e-01 -3.67172472e-02
3.97868186e-01 1.37200877e-01 2.49101222e-02 6.29094467e-02
-6.23930514e-01 -4.50127304e-01 2.93119848e-01 4.37787235e-01
4.81936038e-01 -3.20826512e-05 -3.28404754e-01 3.94923031e-01
2.29699075e-01 2.21958440e-02 4.28783476e-01 7.86148250e-01
-7.06695199e-01 -1.15840399e+00 -5.34487844e-01 3.49194050e-01
-3.36550385e-01 -4.59917486e-01 -4.12656486e-01 5.41590691e-01
-1.87188581e-01 9.30041909e-01 -4.20531183e-01 -2.24125654e-01
2.60274947e-01 2.10575849e-01 4.27390248e-01 -6.13379180e-01
-4.51749235e-01 3.67552966e-01 1.70634333e-02 -1.78637192e-01
-4.69773203e-01 -6.37541234e-01 -9.79697824e-01 1.75058618e-02
-3.26020271e-01 -2.02922359e-01 6.43579602e-01 1.05956972e+00
1.52687699e-01 3.85460764e-01 5.49067318e-01 -8.76945853e-01
-4.59009945e-01 -8.37878048e-01 -3.33017826e-01 -1.29845977e-01
1.75055265e-01 -3.59051645e-01 -2.44500369e-01 3.40350419e-01] | [11.791440963745117, 2.115125894546509] |
80f74173-05c6-4369-a0cf-18f24a6fc870 | shadocnet-learning-spatial-aware-tokens-in | 2211.16675 | null | https://arxiv.org/abs/2211.16675v2 | https://arxiv.org/pdf/2211.16675v2.pdf | ShaDocNet: Learning Spatial-Aware Tokens in Transformer for Document Shadow Removal | Shadow removal improves the visual quality and legibility of digital copies of documents. However, document shadow removal remains an unresolved subject. Traditional techniques rely on heuristics that vary from situation to situation. Given the quality and quantity of current public datasets, the majority of neural network models are ill-equipped for this task. In this paper, we propose a Transformer-based model for document shadow removal that utilizes shadow context encoding and decoding in both shadow and shadow-free regions. Additionally, shadow detection and pixel-level enhancement are included in the whole coarse-to-fine process. On the basis of comprehensive benchmark evaluations, it is competitive with state-of-the-art methods. | ['Shuqiang Wang', 'Chi-Man Pun', 'Xiaodong Cun', 'Xuhang Chen'] | 2022-11-30 | null | null | null | null | ['shadow-removal', 'shadow-detection', 'image-shadow-removal'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 6.50470793e-01 -5.29325247e-01 9.34544280e-02 -2.78560668e-01
-3.62996250e-01 -5.40685177e-01 5.74609280e-01 -9.14054215e-02
-1.39583364e-01 7.31746197e-01 1.52538598e-01 -6.49304867e-01
2.19476819e-01 -5.68269074e-01 -4.75596160e-01 -8.41655791e-01
3.26732963e-01 -1.04709320e-01 5.05328655e-01 -1.43758491e-01
4.22028571e-01 5.94124317e-01 -1.24778903e+00 6.27130032e-01
1.08050168e+00 9.78948712e-01 5.70378542e-01 8.42062354e-01
-2.96438545e-01 6.48425698e-01 -8.44494104e-01 -2.95288205e-01
9.68785286e-02 -1.89086601e-01 -1.29127368e-01 -6.57901168e-04
5.10722220e-01 -5.95035434e-01 -6.08688354e-01 8.02082896e-01
6.14679337e-01 8.23708698e-02 4.87289995e-01 -1.00227642e+00
-1.21498966e+00 1.04227073e-01 -6.31740868e-01 2.30523869e-01
2.74156213e-01 1.09433115e-01 8.28138769e-01 -1.03345096e+00
5.12518227e-01 1.26084447e+00 6.50070965e-01 2.43932471e-01
-8.71362269e-01 -4.72489178e-01 4.87166733e-01 5.03968894e-01
-1.12215483e+00 -3.46136361e-01 7.18489051e-01 -4.47954834e-02
1.01066840e+00 5.12202442e-01 6.92622066e-01 1.04813027e+00
7.02090919e-01 1.06000400e+00 1.41396785e+00 -5.43151557e-01
8.87878537e-02 2.89323814e-02 1.19507246e-01 6.94123864e-01
3.78779292e-01 2.43704463e-03 -7.43547082e-01 1.30442560e-01
4.83425707e-01 1.43725932e-01 -5.97375095e-01 -2.26289198e-01
-7.87257850e-01 3.85606438e-01 4.22875285e-01 5.91189303e-02
-1.11261137e-01 1.86520502e-01 1.46279573e-01 -2.47626945e-01
5.59698999e-01 1.15384512e-01 -2.80893356e-01 1.84058189e-01
-1.28731477e+00 2.78014421e-01 5.84636927e-01 8.38500082e-01
3.28396022e-01 8.68389308e-02 -8.77721012e-01 6.71489716e-01
2.81630039e-01 8.88860047e-01 1.00988686e-01 -5.93325138e-01
4.35881138e-01 5.30071735e-01 1.21915452e-01 -1.11195076e+00
-1.53336138e-01 -5.02918601e-01 -8.69464457e-01 3.84700537e-01
1.88234925e-01 2.16519222e-01 -1.29589605e+00 1.08219135e+00
9.91899371e-02 2.93659151e-01 -2.12662026e-01 8.50672305e-01
9.12695765e-01 7.55385220e-01 -2.24741161e-01 -1.18805945e-01
1.24749410e+00 -1.25174177e+00 -1.12069499e+00 -4.37918037e-01
2.97170039e-02 -9.82808828e-01 1.25877631e+00 6.94721222e-01
-9.89265203e-01 -3.67446512e-01 -1.32686043e+00 -3.36638421e-01
-6.66836798e-01 2.72065759e-01 6.26999795e-01 9.66612339e-01
-9.48641598e-01 4.07028377e-01 -4.05814677e-01 -3.21272649e-02
6.53529108e-01 1.05796747e-01 1.98157623e-01 -5.56728482e-01
-1.08831429e+00 9.00225580e-01 -5.52461483e-02 4.44583535e-01
-7.33478427e-01 -4.76708680e-01 -3.24015528e-01 2.04642370e-01
4.10430998e-01 -3.25971723e-01 1.13884068e+00 -6.64942622e-01
-1.45989418e+00 2.62296379e-01 -4.05459344e-01 -2.92955160e-01
7.03150570e-01 -4.74663407e-01 -4.34189796e-01 -4.29700091e-02
-2.49912828e-01 1.85821936e-01 1.12164688e+00 -1.87970531e+00
-7.25445628e-01 -3.69930387e-01 1.55616820e-01 4.86523002e-01
-4.59182501e-01 4.56138439e-02 -1.21376121e+00 -8.93207431e-01
-6.92769811e-02 -7.88372934e-01 1.11500949e-01 3.67377013e-01
-6.09948993e-01 1.91672355e-01 1.45689428e+00 -1.12662601e+00
1.47143269e+00 -1.97454250e+00 -2.31957853e-01 8.03347901e-02
1.95623398e-01 5.25825143e-01 -1.15206204e-01 1.57612786e-01
3.81346166e-01 -2.16506273e-01 -2.05465972e-01 -7.46880114e-01
2.01093316e-01 -7.81579316e-03 -5.57777107e-01 6.21799469e-01
-2.48629317e-01 8.53131294e-01 -5.32072842e-01 -5.72196066e-01
3.94319981e-01 8.45979154e-01 -1.46706939e-01 1.51565447e-01
-2.47286454e-01 -5.55323139e-02 -2.38816708e-01 1.15457642e+00
1.11757708e+00 -1.69950128e-01 3.98871005e-02 -1.70213878e-01
-3.67501825e-02 -4.34177928e-03 -1.03282762e+00 1.49866664e+00
-5.36637127e-01 1.29794216e+00 3.61815318e-02 -2.24893242e-02
5.35945117e-01 -4.24668938e-02 7.56587982e-02 -1.07041168e+00
9.86828357e-02 -6.27952442e-02 -3.11246872e-01 -1.59910068e-01
1.13795149e+00 4.90300298e-01 2.19421044e-01 3.44466090e-01
-7.60896146e-01 2.29898375e-02 1.04097188e-01 1.38686165e-01
1.13793290e+00 3.94329995e-01 4.55692671e-02 1.09497542e-02
3.40221167e-01 -2.61678547e-01 4.11915720e-01 9.07404423e-01
-2.45345727e-01 8.36358666e-01 2.72907108e-01 -1.85440138e-01
-8.68957043e-01 -1.11114049e+00 -1.65636152e-01 1.36444497e+00
4.98759031e-01 -1.67004794e-01 -1.04552388e+00 -5.50931513e-01
8.75065327e-02 8.71782303e-01 -7.50620425e-01 8.19727108e-02
-5.45963943e-01 -8.39432299e-01 4.98904854e-01 5.63988984e-01
5.48858225e-01 -1.12816739e+00 -8.88681173e-01 -3.04982308e-02
1.43922735e-02 -1.35152709e+00 -4.92610157e-01 2.94488281e-01
-4.64947969e-01 -1.00889540e+00 -8.95130873e-01 -4.29411918e-01
6.34263098e-01 8.61541867e-01 1.03232646e+00 4.39671129e-01
-4.88217235e-01 -3.86374295e-02 -4.26799744e-01 -7.54825234e-01
-8.24512467e-02 -2.07353756e-01 -4.37084734e-01 -1.33125305e-01
-2.74136439e-02 -9.06451198e-04 -7.75888205e-01 1.52229190e-01
-9.07679141e-01 1.40781045e-01 8.19027007e-01 7.16339827e-01
3.90285492e-01 3.88806850e-01 -8.68105441e-02 -8.88750911e-01
9.63019669e-01 1.11541962e-02 -5.96140265e-01 5.87676108e-01
-8.70386839e-01 -3.45976174e-01 6.35861337e-01 -2.72977144e-01
-1.50927341e+00 -5.49819618e-02 4.81920958e-01 -2.48600617e-01
-4.32910360e-02 -2.04011872e-02 -4.81612027e-01 -1.37190804e-01
3.57812405e-01 4.14939404e-01 -5.94283819e-01 -5.01698077e-01
3.42628986e-01 7.25282431e-01 5.88045895e-01 -1.09383248e-01
8.40459406e-01 7.69610345e-01 -1.83811486e-01 -8.12032640e-01
-8.11300516e-01 -2.50085950e-01 -5.38585901e-01 -2.49729276e-01
6.00196838e-01 -5.89894235e-01 -5.32935917e-01 8.20642948e-01
-1.35278940e+00 -5.62719584e-01 1.96024701e-01 -1.91232324e-01
4.91839536e-02 5.73705792e-01 -4.36282277e-01 -1.01999676e+00
-5.34980118e-01 -1.16019320e+00 1.16517079e+00 3.25293422e-01
3.11207026e-01 -7.40149260e-01 -1.87892720e-01 3.89709234e-01
4.87562507e-01 -1.43334242e-02 7.96910882e-01 1.95436642e-01
-9.79022086e-01 -2.64157295e-01 -9.28738654e-01 3.80454153e-01
2.71145254e-01 9.92304832e-02 -1.29567361e+00 -3.66794407e-01
-2.82301813e-01 2.44527340e-01 1.29399526e+00 6.77157223e-01
1.58144748e+00 -2.28984490e-01 -5.99535763e-01 6.70213103e-01
1.45091486e+00 2.08118126e-01 1.06213450e+00 5.32309115e-01
9.50230122e-01 3.20753217e-01 7.27255046e-01 6.55502558e-01
2.73409128e-01 6.59816861e-01 4.68714654e-01 -4.92741644e-01
-6.69823170e-01 1.91740632e-01 2.14520842e-01 2.73579866e-01
1.26114950e-01 -1.09217167e+00 -8.52843761e-01 3.43863219e-01
-1.67614031e+00 -8.21847916e-01 -2.17980579e-01 1.81554890e+00
6.20002985e-01 3.72014105e-01 -5.51296651e-01 3.04919124e-01
6.48026645e-01 5.91712832e-01 -5.17346799e-01 -4.18507129e-01
-4.60299730e-01 -1.64497033e-01 8.74253750e-01 4.16254938e-01
-9.94336605e-01 1.05695975e+00 6.56198835e+00 9.21955705e-01
-8.83375943e-01 -2.75167581e-02 4.87531990e-01 -1.04778685e-01
-4.36771572e-01 -1.23122381e-02 -9.14418995e-01 7.21620560e-01
4.50267315e-01 4.32890505e-01 6.42337799e-01 3.71148407e-01
2.54836351e-01 -5.09371161e-01 -7.94157386e-01 8.46399903e-01
5.23495078e-01 -1.24679101e+00 -1.05389073e-01 -1.40155237e-02
8.30713153e-01 -2.20112517e-01 4.10378635e-01 7.07225502e-02
6.80119991e-02 -1.23437381e+00 8.72917533e-01 7.20878303e-01
8.37072611e-01 -6.54853284e-01 7.17340887e-01 -9.54066440e-02
-1.23341286e+00 -2.25604042e-01 -1.74984425e-01 2.59449869e-01
1.92185298e-01 8.92809808e-01 -8.03479075e-01 3.84071559e-01
9.44432557e-01 1.93371013e-01 -8.35140526e-01 1.15046406e+00
-3.68827820e-01 4.34830457e-01 1.86058953e-02 -6.57264516e-02
1.69896916e-01 -1.27540175e-02 2.74716973e-01 1.75409019e+00
1.79712012e-01 -8.31225440e-02 -4.54931892e-02 6.59140408e-01
-1.72810107e-01 -1.10021770e-01 -2.73591608e-01 1.91257238e-01
5.54268539e-01 1.12560081e+00 -9.23030317e-01 -2.37563416e-01
-4.46134537e-01 1.35196579e+00 6.82534534e-04 6.33274436e-01
-9.78814960e-01 -2.75100529e-01 3.29402506e-01 2.76724007e-02
4.95279312e-01 -2.41464779e-01 -9.24530983e-01 -5.83471775e-01
2.85053730e-01 -1.00780714e+00 -1.24735549e-01 -1.09545743e+00
-7.94515133e-01 3.74539793e-01 -2.39747971e-01 -9.01917040e-01
5.62568903e-01 -7.83257782e-01 -6.79361224e-01 7.43173420e-01
-1.97418976e+00 -1.31714594e+00 -9.75930929e-01 3.80580783e-01
7.53790021e-01 -1.70489371e-01 6.20652676e-01 2.95731723e-01
-6.81120396e-01 5.19021571e-01 7.05356121e-01 -1.42069533e-01
8.33772004e-01 -1.22408199e+00 6.88714743e-01 1.15246308e+00
-7.43424222e-02 3.62848669e-01 8.14168036e-01 -7.53753424e-01
-1.57477963e+00 -1.01928997e+00 5.74273765e-01 -3.56292009e-01
2.51653403e-01 -5.47747910e-01 -8.47832263e-01 1.97957903e-01
4.63354826e-01 1.20585002e-02 3.66264582e-01 1.91537812e-02
-3.25071424e-01 -2.20797420e-01 -9.95462954e-01 8.94631565e-01
7.73102939e-01 -5.09966075e-01 -1.34363011e-01 4.67499763e-01
6.18692875e-01 -6.88856781e-01 -1.46340236e-01 1.33559152e-01
7.74636209e-01 -1.13768971e+00 1.14858365e+00 6.88009188e-02
4.15188432e-01 -3.83027464e-01 -3.94290417e-01 -1.09708893e+00
-2.92614162e-01 -5.47864795e-01 -5.27596354e-01 1.11277103e+00
1.55260399e-01 -2.03511134e-01 8.33264768e-01 3.27207625e-01
-1.52671024e-01 -9.15995002e-01 -5.61964214e-01 -6.68472409e-01
-3.58083397e-01 -4.57204401e-01 7.69583106e-01 3.59043092e-01
-7.45121658e-01 6.19419254e-02 -6.11409307e-01 2.78666466e-01
7.53168762e-01 3.90990496e-01 6.83940649e-01 -1.02941179e+00
-2.72454649e-01 -4.64033365e-01 1.65151939e-01 -1.02428937e+00
-9.44149855e-04 -4.06946301e-01 4.14586455e-01 -2.14057350e+00
4.69400555e-01 -3.45247090e-01 -4.36738461e-01 5.02154469e-01
-4.50388014e-01 3.97544801e-01 3.56498450e-01 6.82799146e-02
-6.63001239e-01 5.94693780e-01 1.37319851e+00 -3.93998861e-01
-2.26908356e-01 -1.71728924e-01 -5.83011746e-01 7.45577991e-01
8.49426866e-01 -3.40039492e-01 -3.91239941e-01 -7.55657852e-01
-7.40270615e-02 -3.46109152e-01 5.43364108e-01 -8.47372711e-01
1.67665929e-01 -3.84767652e-01 8.94217789e-01 -1.13342440e+00
7.50025451e-01 -7.82863736e-01 -3.80098790e-01 4.91317928e-01
-2.27783188e-01 -3.39160822e-02 2.04278156e-01 9.15033877e-01
2.97856927e-01 -6.91202283e-03 7.34151602e-01 1.20952480e-01
-7.67259419e-01 8.82148743e-02 -2.55242020e-01 -2.36734554e-01
8.38452756e-01 -5.17952442e-01 -7.32348204e-01 -2.02320322e-01
1.17861591e-01 -1.39675125e-01 4.91220146e-01 5.02379358e-01
9.36432958e-01 -9.45181787e-01 -6.51681781e-01 2.23361850e-02
-9.56248268e-02 -1.67646199e-01 3.84723574e-01 5.60518563e-01
-6.09394193e-01 6.31688714e-01 -1.31753102e-01 -2.92785496e-01
-1.61921513e+00 4.06292468e-01 -1.57369021e-02 -2.62725681e-01
-6.43481731e-01 8.66202354e-01 4.03050691e-01 6.84676394e-02
6.59582615e-01 -4.19518590e-01 1.50852399e-02 -2.80522585e-01
6.13927603e-01 5.90508044e-01 3.26644957e-01 -4.62479919e-01
-3.77010584e-01 2.12697342e-01 -1.34826884e-01 1.13518141e-01
1.07100511e+00 -2.82830060e-01 -1.63227320e-04 9.53640118e-02
9.13093686e-01 1.02442734e-01 -1.60608006e+00 -1.49692535e-01
-4.07651931e-01 -8.55234921e-01 5.54729998e-01 -1.26468945e+00
-1.07649970e+00 8.81760776e-01 8.56103361e-01 -2.59736292e-02
1.33634079e+00 -6.72380745e-01 1.03498793e+00 4.91201222e-01
-7.53489137e-02 -1.71035624e+00 2.01434746e-01 5.57680607e-01
1.02298319e+00 -1.10784352e+00 6.50851309e-01 -5.02006173e-01
-5.40242374e-01 8.52666616e-01 4.10900503e-01 2.10864916e-01
3.13952178e-01 7.26366520e-01 2.05135215e-02 9.99408662e-02
-6.80850744e-01 7.73128048e-02 5.86536407e-01 7.84576952e-01
3.55719447e-01 8.57402086e-02 -4.89372537e-02 2.31602132e-01
1.02383941e-01 -3.81948091e-02 4.20347452e-01 1.16542542e+00
-6.66973591e-01 -8.87539387e-01 -7.70185828e-01 4.70148295e-01
-2.02865586e-01 -5.34961641e-01 -5.48417270e-01 5.00998497e-01
1.48630202e-01 1.07327747e+00 -3.47975582e-01 -9.84941274e-02
2.07657740e-01 -4.81478453e-01 5.13424754e-01 -2.00632483e-01
-4.15936232e-01 1.81046978e-01 -2.31517348e-02 -5.10596633e-01
-1.46448435e-02 -5.33102155e-01 -1.19512868e+00 -5.20866454e-01
-5.94014585e-01 -5.94057560e-01 9.61401761e-01 9.88287330e-01
1.98224276e-01 1.05319083e+00 4.53654289e-01 -9.80846167e-01
-3.43950808e-01 -7.20112085e-01 -6.81184411e-01 1.25705143e-02
3.57267201e-01 -6.60489917e-01 -9.84511226e-02 6.90062493e-02] | [10.86196231842041, -4.042856693267822] |
04222e42-955e-4f87-b9d9-039940c45d71 | vast-a-vision-audio-subtitle-text-omni | 2305.18500 | null | https://arxiv.org/abs/2305.18500v1 | https://arxiv.org/pdf/2305.18500v1.pdf | VAST: A Vision-Audio-Subtitle-Text Omni-Modality Foundation Model and Dataset | Vision and text have been fully explored in contemporary video-text foundational models, while other modalities such as audio and subtitles in videos have not received sufficient attention. In this paper, we resort to establish connections between multi-modality video tracks, including Vision, Audio, and Subtitle, and Text by exploring an automatically generated large-scale omni-modality video caption dataset called VAST-27M. Specifically, we first collect 27 million open-domain video clips and separately train a vision and an audio captioner to generate vision and audio captions. Then, we employ an off-the-shelf Large Language Model (LLM) to integrate the generated captions, together with subtitles and instructional prompts into omni-modality captions. Based on the proposed VAST-27M dataset, we train an omni-modality video-text foundational model named VAST, which can perceive and process vision, audio, and subtitle modalities from video, and better support various tasks including vision-text, audio-text, and multi-modal video-text tasks (retrieval, captioning and QA). Extensive experiments have been conducted to demonstrate the effectiveness of our proposed VAST-27M corpus and VAST foundation model. VAST achieves 22 new state-of-the-art results on various cross-modality benchmarks. Code, model and dataset will be released at https://github.com/TXH-mercury/VAST. | ['Jing Liu', 'Xinxin Zhu', 'Mingzhen Sun', 'Zijia Zhao', 'Qunbo Wang', 'Handong Li', 'Sihan Chen'] | 2023-05-29 | null | null | null | null | ['audio-captioning', 'video-captioning', 'video-question-answering', 'video-retrieval', 'image-captioning', 'cross-modal-retrieval', 'zero-shot-cross-modal-retrieval'] | ['audio', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'miscellaneous', 'miscellaneous'] | [ 2.76358128e-01 -1.83802262e-01 -1.91446349e-01 -2.10116550e-01
-1.35663092e+00 -5.32520652e-01 7.25875020e-01 -3.39348704e-01
-2.63443649e-01 4.73123997e-01 7.20904052e-01 -1.49010584e-01
3.06838274e-01 -1.11029215e-01 -1.21301162e+00 -3.17465186e-01
3.21335852e-01 2.85057813e-01 1.09978877e-01 7.73489997e-02
-4.23283018e-02 -3.67268682e-01 -1.87327945e+00 7.97519505e-01
5.36933064e-01 1.17901897e+00 4.82700497e-01 1.14692628e+00
-2.81996608e-01 1.03765178e+00 -3.28534335e-01 -3.54271591e-01
-1.82996675e-01 -2.81046152e-01 -6.26045883e-01 1.57814890e-01
9.40857053e-01 -7.26846695e-01 -8.16019952e-01 7.91828156e-01
7.56634951e-01 6.99760094e-02 6.12391472e-01 -1.60825086e+00
-9.68998134e-01 5.69652498e-01 -5.81010222e-01 3.37928608e-02
1.00020623e+00 3.83996695e-01 7.72264719e-01 -1.14667976e+00
6.13728464e-01 1.31199586e+00 3.04824680e-01 7.50927329e-01
-5.04795015e-01 -6.48610413e-01 1.15918897e-01 3.55170816e-01
-1.28170633e+00 -7.36037910e-01 5.06742656e-01 -5.21624923e-01
6.86805964e-01 2.87891954e-01 4.79015648e-01 1.68073726e+00
-4.99395058e-02 1.31381679e+00 6.07838929e-01 -3.59650522e-01
-2.66970992e-01 9.13612470e-02 -1.00288518e-01 6.07098758e-01
-3.09253275e-01 -1.11395478e-01 -1.00714421e+00 2.05023289e-01
5.83134890e-01 -9.54091251e-02 -6.32584751e-01 7.91206583e-02
-1.57422221e+00 4.77537930e-01 -6.86687827e-02 2.62288470e-02
-3.19417536e-01 2.74379373e-01 6.42610848e-01 2.30280936e-01
1.53269336e-01 -1.53804943e-01 -1.34068325e-01 -5.25449753e-01
-8.36877108e-01 -8.79205316e-02 3.17309916e-01 1.31557715e+00
3.80575448e-01 5.21758245e-03 -7.17587709e-01 9.47254837e-01
6.74099684e-01 1.07185423e+00 7.45016813e-01 -1.07216358e+00
8.60289395e-01 1.05546199e-01 4.40120250e-02 -4.96779025e-01
-5.56944646e-02 4.46484759e-02 -6.17585778e-01 -5.28536558e-01
-5.05590588e-02 -2.73411274e-01 -1.15567529e+00 1.63394630e+00
3.05833042e-01 5.80325305e-01 2.37695470e-01 1.12985802e+00
1.71380782e+00 1.23244679e+00 2.12462842e-01 -4.49397489e-02
1.43722343e+00 -1.38905239e+00 -7.72657990e-01 -1.15337394e-01
3.98833424e-01 -1.09324884e+00 1.13357019e+00 2.92697042e-01
-1.23419559e+00 -7.90369630e-01 -5.30449927e-01 -2.37580076e-01
-1.40151635e-01 3.34852993e-01 2.45123371e-01 2.02528030e-01
-1.16504025e+00 -3.84042650e-01 -4.66871530e-01 -4.37245697e-01
2.80183405e-01 -1.70470729e-01 -4.99435484e-01 -5.71977258e-01
-1.21926939e+00 3.81817400e-01 4.41712707e-01 1.37630373e-01
-1.53263676e+00 -6.53654873e-01 -1.15905285e+00 -1.19358927e-01
3.88744414e-01 -9.11374032e-01 1.47270858e+00 -9.48481023e-01
-1.37810659e+00 8.33926201e-01 -2.73085058e-01 -2.53829360e-01
2.62964845e-01 -2.89049625e-01 -6.97812200e-01 7.63811111e-01
8.82177204e-02 1.24617386e+00 1.03380001e+00 -1.09599507e+00
-7.04321027e-01 1.52202826e-02 3.31698507e-01 5.52606046e-01
-6.06529057e-01 2.04450399e-01 -1.19040501e+00 -3.31934214e-01
-4.28847462e-01 -8.13845038e-01 4.45816040e-01 -4.55135815e-02
-5.67674637e-01 -1.38427809e-01 8.78345132e-01 -7.26240516e-01
1.02876306e+00 -2.44177389e+00 1.26632676e-01 -4.73860174e-01
7.68171549e-02 2.76978761e-01 -8.19214165e-01 4.69537973e-01
1.97002709e-01 -2.47238487e-01 3.14344287e-01 -5.55880964e-01
7.54126608e-02 5.10515347e-02 -3.99078459e-01 4.85778563e-02
-9.30805132e-02 9.71378982e-01 -8.72941971e-01 -8.76383960e-01
5.23880899e-01 7.54928768e-01 -3.85058582e-01 4.06979203e-01
-4.36757803e-01 3.41856241e-01 -5.56552649e-01 9.78378594e-01
3.87353271e-01 -3.97413880e-01 -5.12038291e-01 -3.81246030e-01
-4.20386307e-02 -1.04428187e-01 -7.34628499e-01 2.26510978e+00
-6.31118238e-01 1.00386071e+00 2.16197923e-01 -6.70597017e-01
4.81529534e-01 9.68729675e-01 5.14573574e-01 -6.92446172e-01
2.33614653e-01 6.52481839e-02 -7.56056845e-01 -1.13550425e+00
6.63606286e-01 2.73143560e-01 -7.23263398e-02 1.10991441e-01
5.86641610e-01 -2.82087699e-02 4.00478005e-01 5.31247199e-01
9.40328777e-01 2.87137598e-01 -4.09516215e-01 6.54590666e-01
6.38468146e-01 -5.05020237e-03 -1.44468412e-01 6.93256199e-01
-2.23717034e-01 1.03019929e+00 5.99766485e-02 1.96175963e-01
-7.42315650e-01 -1.11653996e+00 2.37809837e-01 1.43594670e+00
1.03022151e-01 -5.62827170e-01 -5.31028152e-01 -3.23681206e-01
-3.34472805e-01 5.04702926e-01 -3.54777217e-01 -6.39604330e-02
-2.82914862e-02 -1.40845716e-01 7.53986299e-01 3.26534539e-01
4.88585830e-01 -1.11586857e+00 -2.22716033e-01 2.12499481e-02
-1.00857794e+00 -1.71263635e+00 -9.26619291e-01 -5.61283886e-01
-3.51307690e-01 -8.81644905e-01 -1.16735113e+00 -1.03525436e+00
2.96406925e-01 7.39001036e-01 1.02599299e+00 -4.25335020e-01
-1.79576382e-01 1.29764616e+00 -5.85615337e-01 -3.38704139e-01
-2.83468902e-01 -2.67526537e-01 -1.29587501e-01 3.31529140e-01
2.50427872e-01 -2.72420943e-01 -4.56529945e-01 2.02041969e-01
-1.19009566e+00 5.25108814e-01 7.02219069e-01 6.58672810e-01
6.59840822e-01 -4.95966792e-01 6.31063819e-01 -2.11508363e-01
3.56678426e-01 -7.90548325e-01 -2.50314802e-01 3.96417171e-01
2.74713963e-01 -4.89827365e-01 4.79069024e-01 -6.80194855e-01
-1.12548101e+00 1.39278233e-01 -1.99364468e-01 -1.23390973e+00
-3.89074355e-01 7.64502764e-01 -2.09226102e-01 1.77609712e-01
2.98300505e-01 6.68137133e-01 -1.52917370e-01 -2.78803587e-01
5.68911314e-01 1.25579095e+00 1.08010232e+00 -5.81241071e-01
6.41239941e-01 4.26978767e-01 -3.78722191e-01 -1.03316259e+00
-1.05026412e+00 -7.78628349e-01 4.07523513e-02 -7.33489394e-01
1.17915106e+00 -1.62962830e+00 -6.12244546e-01 4.83707577e-01
-1.34721684e+00 -2.59421527e-01 3.91562395e-02 7.88020015e-01
-7.43893921e-01 3.43972325e-01 -6.64063036e-01 -6.08493030e-01
-3.90322238e-01 -1.21310389e+00 1.47889137e+00 4.53779101e-01
2.74011135e-01 -7.66038716e-01 -5.86149050e-03 1.14952064e+00
2.07013458e-01 -1.38926664e-02 1.57307908e-01 -3.85055035e-01
-6.79247916e-01 -2.52706677e-01 -3.84056449e-01 2.94762969e-01
-3.01710039e-01 7.24356622e-02 -1.05382812e+00 -1.81841910e-01
-5.13171077e-01 -1.05790222e+00 8.45701218e-01 5.25378406e-01
1.18283463e+00 -3.99699770e-02 -2.49074046e-02 6.26032710e-01
1.16894984e+00 7.33587667e-02 5.67291856e-01 1.30451038e-01
8.43575060e-01 3.78537595e-01 9.37043130e-01 5.66348672e-01
7.69343138e-01 4.38100427e-01 5.71675956e-01 1.22198842e-01
-4.35022026e-01 -5.98733902e-01 8.12477529e-01 1.23555481e+00
1.62423149e-01 -5.71238935e-01 -6.99333370e-01 8.09113979e-01
-1.76890624e+00 -1.01172602e+00 -7.93877169e-02 1.66647100e+00
7.94240415e-01 -2.33568192e-01 -3.90942842e-02 -3.29904824e-01
7.60803521e-01 1.48381740e-01 -4.02619898e-01 -1.71719398e-02
-1.78995520e-01 -4.25441742e-01 7.74348155e-02 2.92021751e-01
-1.02335453e+00 7.37425923e-01 5.32208300e+00 1.07732081e+00
-1.11228180e+00 3.62303704e-01 2.71829158e-01 -4.30411488e-01
-4.86976773e-01 -4.31994379e-01 -7.64540255e-01 5.10414541e-01
1.29726815e+00 -1.36921838e-01 4.39088076e-01 5.22811770e-01
4.66663569e-01 3.17698494e-02 -1.11095238e+00 1.45968425e+00
5.40507078e-01 -1.32007813e+00 3.36209297e-01 -4.07080948e-01
6.89388096e-01 3.47832203e-01 1.22164831e-01 7.31269479e-01
-2.18937933e-01 -8.31094503e-01 9.35980320e-01 4.71938163e-01
1.33228886e+00 -2.41864622e-01 4.87420768e-01 1.65422067e-01
-1.45275557e+00 -1.85664818e-02 -9.79029834e-02 3.60423118e-01
4.46331382e-01 2.13393271e-01 -4.57188010e-01 6.61356449e-01
8.23723555e-01 9.44796145e-01 -4.45472121e-01 1.21293676e+00
-3.05436384e-02 6.74906611e-01 -3.02692592e-01 1.29864112e-01
3.97996128e-01 2.48502553e-01 6.47633910e-01 1.33580995e+00
4.90339875e-01 -6.67506531e-02 1.77728266e-01 3.64012510e-01
-2.45351106e-01 8.52438435e-03 -6.04939282e-01 -3.41827601e-01
5.42175293e-01 1.24927199e+00 -3.05129532e-02 -5.10794878e-01
-9.77730095e-01 8.57972085e-01 -3.34614158e-01 7.22133458e-01
-1.23327529e+00 -2.42290035e-01 3.69737446e-01 -2.21822456e-01
1.46338195e-01 5.02257459e-02 4.37849641e-01 -1.52701068e+00
4.26977873e-02 -1.02494740e+00 5.10032117e-01 -1.82967269e+00
-1.09636331e+00 6.15752995e-01 1.04384832e-01 -1.34260750e+00
-3.06952059e-01 -5.04926264e-01 -3.86141568e-01 5.41205049e-01
-1.64194620e+00 -1.49270678e+00 -6.03969634e-01 1.29807878e+00
1.11626697e+00 -2.09757924e-01 4.72672671e-01 6.85306728e-01
-4.68372136e-01 4.94098216e-01 1.73228860e-01 1.91828668e-01
1.10180199e+00 -6.12602770e-01 -1.89529717e-01 6.41429067e-01
1.90439090e-01 2.87415069e-02 7.30410159e-01 -3.68523300e-01
-1.84928787e+00 -1.36591792e+00 6.27926111e-01 -4.51244920e-01
8.11013162e-01 -2.28330538e-01 -4.49060708e-01 8.47308934e-01
3.97701919e-01 -5.31715453e-02 5.94823658e-01 -4.46696013e-01
-3.42615724e-01 3.11035831e-02 -6.51434720e-01 5.66879690e-01
8.66746008e-01 -8.74076545e-01 -5.88274121e-01 5.65890610e-01
1.28261197e+00 -5.41299105e-01 -7.68281758e-01 1.31745964e-01
6.47754669e-01 -6.10877633e-01 1.18666649e+00 -2.92416185e-01
8.96286905e-01 -1.83258086e-01 -3.99011642e-01 -1.00415909e+00
2.81377733e-01 -6.36539400e-01 -3.26409847e-01 1.48762691e+00
3.22710633e-01 -2.05177605e-01 3.40386748e-01 1.40260994e-01
-6.96721375e-01 -3.45710605e-01 -9.54742134e-01 -5.92864215e-01
-4.31084454e-01 -8.09468806e-01 2.21103266e-01 7.05605268e-01
-5.12888320e-02 5.45824766e-01 -7.85685062e-01 1.90892011e-01
6.32761359e-01 4.46777344e-02 8.45355213e-01 -7.75327981e-01
-2.87322044e-01 -1.04718260e-01 -2.33166397e-01 -1.40846229e+00
1.79248422e-01 -7.22769976e-01 1.95273876e-01 -1.73953485e+00
3.79191458e-01 2.90423393e-01 -2.50158548e-01 5.40351331e-01
7.58120716e-02 5.34637034e-01 4.63801026e-01 2.27627993e-01
-1.35205185e+00 8.44881773e-01 1.46693385e+00 -4.67387289e-01
-2.05406826e-02 -3.71784449e-01 -5.25596917e-01 4.33846116e-01
3.74993980e-01 -8.38107169e-02 -5.67163706e-01 -8.34357023e-01
4.50099632e-02 8.50266278e-01 4.98207897e-01 -1.07562709e+00
2.86866903e-01 -2.00059071e-01 1.73776671e-01 -8.01242292e-01
8.40116680e-01 -7.11981893e-01 -4.22957093e-02 -8.59794766e-02
-5.37091613e-01 -1.29741535e-01 4.11562532e-01 6.46220505e-01
-6.96638465e-01 -1.53419435e-01 2.15146437e-01 1.14713490e-01
-9.93934751e-01 5.33949196e-01 -5.08153319e-01 2.96405017e-01
8.34696829e-01 8.15439504e-03 -6.89460456e-01 -7.52597988e-01
-7.40859151e-01 6.75660968e-01 8.86609554e-02 8.83319080e-01
9.94356275e-01 -1.44620931e+00 -9.37950015e-01 -1.06643870e-01
4.51503545e-01 -1.96720943e-01 8.40876460e-01 8.83207500e-01
-2.57975101e-01 7.78174818e-01 -2.12730125e-01 -9.43820000e-01
-1.37856436e+00 5.47759712e-01 -1.27056167e-01 2.65842527e-01
-4.26932573e-01 7.71843612e-01 4.03201222e-01 -1.57859147e-01
5.96666217e-01 -1.57016128e-01 -3.02516103e-01 -9.24708843e-02
7.43534029e-01 9.83616859e-02 -4.52845454e-01 -9.04569507e-01
-1.79229259e-01 6.33423388e-01 2.97766894e-01 -3.57664227e-01
9.93505061e-01 -6.74808681e-01 3.35939080e-01 4.82930005e-01
1.25867820e+00 -3.30880880e-01 -1.16862917e+00 -2.60545820e-01
-7.89291739e-01 -2.99751878e-01 1.28509700e-01 -7.82996237e-01
-9.99964893e-01 1.10997474e+00 4.68868852e-01 -1.73298895e-01
1.42149007e+00 1.89889401e-01 1.15352523e+00 2.45868579e-01
1.83450401e-01 -8.88672411e-01 4.34639990e-01 5.17561972e-01
8.88582349e-01 -1.45313537e+00 -5.26458561e-01 -1.69512510e-01
-8.31517637e-01 9.42878842e-01 8.18449080e-01 5.46663940e-01
2.64742494e-01 -5.84808998e-02 2.20211998e-01 1.07894570e-01
-1.21303356e+00 -2.93728173e-01 4.18133050e-01 7.23204017e-01
3.14742833e-01 -1.20240048e-01 3.02077383e-01 5.24489462e-01
1.17681786e-01 4.36792195e-01 8.09582174e-01 7.50703096e-01
-4.16313946e-01 -4.92720544e-01 -7.19543755e-01 1.52333289e-01
-3.86370391e-01 -2.29738846e-01 1.08871553e-02 4.68059748e-01
-1.18628740e-01 1.35806656e+00 -6.89612702e-03 -4.25081849e-01
6.86707422e-02 1.86265171e-01 3.73191178e-01 -4.22859550e-01
-1.48586661e-01 2.61372656e-01 1.39265805e-01 -5.80195487e-01
-6.12319052e-01 -3.64910930e-01 -1.08175778e+00 -1.09629542e-01
-2.10348610e-02 1.01355940e-01 8.35550725e-01 8.35940123e-01
4.14012313e-01 6.23401999e-01 3.07075590e-01 -9.73426700e-01
-2.10751101e-01 -1.11638606e+00 -1.47118837e-01 3.23506773e-01
5.06448269e-01 -4.35803473e-01 -3.38477790e-01 4.91277426e-01] | [10.549964904785156, 1.0510869026184082] |
190b2102-12a5-4a5e-a0fc-929129fc7e00 | competing-models | 1907.03809 | null | https://arxiv.org/abs/1907.03809v5 | https://arxiv.org/pdf/1907.03809v5.pdf | Competing Models | Different agents need to make a prediction. They observe identical data, but have different models: they predict using different explanatory variables. We study which agent believes they have the best predictive ability -- as measured by the smallest subjective posterior mean squared prediction error -- and show how it depends on the sample size. With small samples, we present results suggesting it is an agent using a low-dimensional model. With large samples, it is generally an agent with a high-dimensional model, possibly including irrelevant variables, but never excluding relevant ones. We apply our results to characterize the winning model in an auction of productive assets, to argue that entrepreneurs and investors with simple models will be over-represented in new sectors, and to understand the proliferation of "factors" that explain the cross-sectional variation of expected stock returns in the asset-pricing literature. | ['Mallesh M. Pai', 'Andrea Prat', 'Jose Luis Montiel Olea', 'Pietro Ortoleva'] | 2019-07-08 | null | null | null | null | ['small-data'] | ['computer-vision'] | [-2.06704125e-01 5.79807222e-01 -6.35926664e-01 -2.71141618e-01
-2.58130074e-01 -4.95192945e-01 5.31453133e-01 -2.78331131e-01
-4.62419271e-01 8.43849003e-01 2.75133163e-01 -6.16780102e-01
-5.94026864e-01 -7.04176009e-01 -3.79776955e-01 -5.58975875e-01
-2.64405981e-02 8.61170173e-01 -2.25618929e-01 -7.49787614e-02
6.51233196e-01 -2.18318589e-02 -9.75742102e-01 4.40284573e-02
5.84396482e-01 9.24593925e-01 2.20568866e-01 3.58903140e-01
3.61860245e-01 1.18125892e+00 -4.20876175e-01 -8.60117912e-01
6.67842388e-01 -3.33102286e-01 -4.59667534e-01 -9.28317010e-02
-1.37344778e-01 -3.48878294e-01 6.44442067e-02 9.61275637e-01
8.51359367e-02 -1.68818042e-01 1.11808145e+00 -1.04889190e+00
-8.35312426e-01 8.99840772e-01 -2.05183208e-01 3.11154276e-01
6.55662790e-02 2.73087025e-01 1.47891104e+00 -3.94369870e-01
6.08527601e-01 1.24180317e+00 8.04533124e-01 4.73267764e-01
-1.71005535e+00 -5.93026161e-01 9.43021551e-02 -1.72721520e-01
-7.85386205e-01 -3.43375832e-01 4.26784247e-01 -9.87651408e-01
7.68544734e-01 2.21147418e-01 7.34901369e-01 1.23903883e+00
6.19470358e-01 1.46653384e-01 1.20840561e+00 -3.67822319e-01
1.73519745e-01 6.28786087e-01 9.95431319e-02 1.51379839e-01
9.24865723e-01 6.50901675e-01 -5.25395632e-01 -4.40976858e-01
9.85871613e-01 1.76176950e-01 -1.81202561e-01 -4.84330088e-01
-1.17016912e+00 1.34798050e+00 -2.27018446e-01 8.89895931e-02
-8.26991618e-01 2.80128010e-02 -2.11330559e-02 8.60699177e-01
2.32528552e-01 1.13045049e+00 -7.42667317e-01 -2.71206141e-01
-3.83585393e-01 5.14637291e-01 1.34958518e+00 6.07226551e-01
4.41362441e-01 1.03956707e-01 1.63642302e-01 4.45765823e-01
3.62700939e-01 4.70105529e-01 7.72256851e-01 -1.51715302e+00
5.97068548e-01 2.86380351e-01 7.82228529e-01 -8.13813448e-01
-5.42372525e-01 -6.19252324e-01 -4.04035717e-01 6.35078430e-01
4.94143665e-01 -7.08758354e-01 -3.89617011e-02 1.89169168e+00
-4.89416242e-01 -3.71595502e-01 3.86538684e-01 7.82783210e-01
-3.01379263e-01 4.53282267e-01 2.00414270e-01 -6.69307709e-01
1.01334763e+00 -7.11140990e-01 -1.37250826e-01 -7.04760849e-01
5.37987232e-01 -2.17664748e-01 6.10859156e-01 4.86163557e-01
-1.27827036e+00 -3.14932197e-01 -8.35224807e-01 5.54360032e-01
3.18871051e-01 -2.89171457e-01 8.45662713e-01 6.47949576e-01
-7.01938748e-01 9.20890927e-01 -5.71851850e-01 4.42342833e-02
3.39096487e-01 4.08772916e-01 1.45735079e-02 6.87476754e-01
-9.81610119e-01 1.39809179e+00 1.60318762e-02 -4.99933571e-01
-7.64515579e-01 -8.23717654e-01 -4.14507166e-02 5.88021696e-01
4.07164216e-01 -9.76807117e-01 1.56975615e+00 -1.45638037e+00
-1.22821748e+00 5.47603145e-02 -4.92764600e-02 -7.37482071e-01
7.61277854e-01 2.09133998e-02 -1.72363773e-01 -4.34234858e-01
3.47443312e-01 -2.39062548e-01 6.43353343e-01 -1.33305550e+00
-8.43775511e-01 -5.43906987e-01 -2.55762935e-02 2.30786294e-01
1.36593189e-02 -9.95419472e-02 6.17909253e-01 -5.90787828e-01
2.29365766e-01 -9.73080456e-01 -5.44286013e-01 -7.86848903e-01
-1.18159257e-01 9.35025886e-03 -1.95874110e-01 -3.88384461e-01
9.53746200e-01 -1.89600992e+00 -1.72458351e-01 5.52402437e-01
3.41336906e-01 -6.29237950e-01 1.45451888e-01 4.32538956e-01
-1.98479682e-01 5.93435466e-01 2.31643751e-01 -1.98155880e-01
6.26563489e-01 -2.35768128e-02 -5.20288825e-01 2.92246312e-01
-2.10449807e-02 6.13356709e-01 -3.37425649e-01 6.32267352e-03
-3.08280379e-01 -2.70370573e-01 -3.24122250e-01 -1.53639674e-01
2.53862161e-02 -1.30468523e-02 -8.61958742e-01 2.56804347e-01
3.53366494e-01 -4.21502650e-01 5.12844563e-01 5.85190296e-01
-2.03373164e-01 4.38745618e-01 -1.07048738e+00 4.32260990e-01
-1.06115408e-01 7.59768605e-01 -1.71361282e-01 -1.25642371e+00
6.96815193e-01 3.86116475e-01 6.14636354e-02 -6.11827075e-01
9.57876742e-02 4.25423920e-01 4.90391195e-01 -4.97516900e-01
3.91585320e-01 -7.89693236e-01 7.76783973e-02 7.33296096e-01
-4.14872795e-01 2.52924562e-01 -2.23229248e-02 -1.03464574e-01
9.77086484e-01 -4.30225879e-01 3.96484554e-01 -4.75208879e-01
-1.55786544e-01 4.38666865e-02 9.83914256e-01 1.34883010e+00
8.11941177e-02 2.86496073e-01 1.33918154e+00 -4.66302991e-01
-1.31360888e+00 -7.88387358e-01 -1.98321089e-01 9.20791805e-01
-1.28281983e-02 1.32891282e-01 -1.32593215e-01 -3.24879706e-01
6.75648689e-01 1.16271544e+00 -8.91285419e-01 1.74861521e-01
5.11647435e-03 -9.63678002e-01 -4.75414395e-02 4.65156466e-01
-1.21844910e-01 -9.60158765e-01 -6.75139010e-01 2.53688395e-01
3.30585539e-01 -5.02694249e-01 1.44720078e-01 6.07536793e-01
-8.39993119e-01 -9.30775523e-01 -4.73932594e-01 -3.09439003e-01
4.95650262e-01 -1.60416409e-01 1.27283144e+00 -4.70143296e-02
4.33623224e-01 1.19193316e-01 -1.05312459e-01 -9.98168409e-01
-7.44724870e-01 -5.14839543e-03 3.08330595e-01 -3.23450953e-01
6.94867969e-01 -3.68063807e-01 -5.26076794e-01 4.20533478e-01
-1.52945682e-01 3.98415178e-02 7.19490170e-01 7.99268663e-01
5.87155223e-02 2.86966741e-01 9.41801071e-01 -9.92043436e-01
9.35244858e-01 -8.10700536e-01 -7.49395549e-01 3.71902704e-01
-1.08340335e+00 1.04586713e-01 5.45189381e-01 -7.64892042e-01
-1.12808228e+00 -4.27816093e-01 6.61352515e-01 1.50749832e-01
1.54773191e-01 5.76322019e-01 2.61959583e-01 2.06522211e-01
7.30031967e-01 5.82396127e-02 2.72627473e-01 -5.96484900e-01
-2.88818240e-01 4.74800617e-01 -1.24135189e-01 -5.27296066e-01
9.55024064e-01 2.57060200e-01 1.34033084e-01 -2.53984839e-01
-6.89058304e-01 2.76861846e-01 -1.06363274e-01 2.96827376e-01
5.52413225e-01 -9.71047819e-01 -9.34842467e-01 7.04505993e-03
-8.96720588e-01 -4.69299674e-01 -4.31185633e-01 1.34202433e+00
-9.36407983e-01 -1.50780126e-01 -4.97188479e-01 -1.15315461e+00
2.12456286e-01 -1.04232216e+00 2.05837473e-01 3.80175740e-01
-4.85539883e-01 -8.57514977e-01 2.29651198e-01 3.29024285e-01
4.49807733e-01 -3.51694852e-01 1.09702969e+00 -1.46598232e+00
-7.36105680e-01 -2.34702989e-01 3.00782379e-02 8.17614123e-02
-1.07804559e-01 -2.00503826e-01 -4.89701927e-01 -1.09463543e-01
4.87372369e-01 -3.78703773e-02 6.98843479e-01 8.83524716e-01
7.43866801e-01 -6.10518396e-01 -3.34427297e-01 3.44914943e-01
1.29570878e+00 6.72291934e-01 3.23769897e-01 8.50484252e-01
-1.96430743e-01 9.34347153e-01 4.45126534e-01 8.82895350e-01
2.54817635e-01 4.17348444e-01 2.57297516e-01 4.60047543e-01
9.86024916e-01 -2.62230694e-01 3.28220993e-01 2.04640910e-01
-5.33064246e-01 -8.06428716e-02 -6.82846606e-01 3.15811992e-01
-2.12546611e+00 -1.60853434e+00 1.82806239e-01 2.14979482e+00
4.85162705e-01 4.10577476e-01 4.49378163e-01 -3.25794816e-01
4.02069360e-01 -1.10452548e-01 -6.25335991e-01 -3.60329539e-01
-2.28165284e-01 -9.07865912e-02 7.98228920e-01 7.01004624e-01
-6.81292951e-01 3.76660258e-01 7.37141418e+00 4.57226038e-01
-5.77658772e-01 -1.54784948e-01 9.35923576e-01 -2.61172622e-01
-5.84049344e-01 3.50985914e-01 -8.81815851e-01 7.29800344e-01
9.28623974e-01 -7.14574218e-01 2.56647438e-01 1.15731025e+00
3.19310635e-01 -1.45587862e-01 -1.38259709e+00 3.15649182e-01
-1.03154354e-01 -1.39184511e+00 -3.18845451e-01 6.76260531e-01
8.49212289e-01 -4.53756005e-02 2.43296906e-01 4.18025643e-01
7.77358890e-01 -1.35471797e+00 1.03470671e+00 9.74861264e-01
-1.83738023e-01 -5.47586441e-01 9.78286922e-01 6.81914926e-01
-3.28229070e-01 -7.60505378e-01 -6.98400319e-01 -6.74715638e-01
-1.38423964e-01 2.04379454e-01 -7.80868888e-01 -1.84541121e-02
2.25594059e-01 1.87048331e-01 -2.47869134e-01 7.25890517e-01
3.37572485e-01 4.69854027e-01 -3.93147022e-02 -3.73915136e-01
-5.34716509e-02 -6.62236273e-01 1.14690930e-01 5.98629653e-01
7.30612636e-01 4.56440568e-01 -2.28512257e-01 1.09468329e+00
2.30204999e-01 -1.04323521e-01 -4.83606309e-01 -1.58219993e-01
4.50287849e-01 5.84917486e-01 -2.78940797e-01 -3.06195170e-01
-7.70229220e-01 1.87026113e-01 -1.38194546e-01 4.09236312e-01
-4.54523325e-01 -9.82805565e-02 5.88235676e-01 1.92125157e-01
2.25113124e-01 -6.23989776e-02 -6.11938655e-01 -1.08360314e+00
4.15721051e-02 -1.00277781e+00 9.44191217e-02 -7.13832200e-01
-1.66435373e+00 6.47080168e-02 -2.12170407e-01 -8.19243908e-01
-6.15072787e-01 -7.54557014e-01 -5.70665777e-01 1.10898507e+00
-1.19620311e+00 -3.92558634e-01 3.34757179e-01 8.42489153e-02
5.30003786e-01 -8.03552330e-01 5.79510391e-01 -5.55921733e-01
-3.64514232e-01 6.36434332e-02 6.38181329e-01 5.34939319e-02
2.38531426e-01 -1.33771670e+00 1.86951712e-01 2.14338914e-01
4.83883619e-02 7.22027659e-01 8.18144143e-01 -7.20239758e-01
-1.13625705e+00 -3.54239464e-01 8.57929528e-01 -7.14688897e-01
1.05597627e+00 3.58201146e-01 -5.75318456e-01 1.21096492e+00
1.42122000e-01 -6.13485694e-01 7.81002283e-01 3.83914948e-01
-2.97942925e-02 -1.77970126e-01 -9.40277576e-01 6.63966119e-01
5.72392166e-01 -3.04204792e-01 -8.78882170e-01 1.08410321e-01
4.74816263e-01 4.72238064e-01 -9.22706425e-01 1.06599502e-01
8.47337246e-01 -1.45715344e+00 9.12201583e-01 -9.03947175e-01
4.79559690e-01 3.09271157e-01 -2.95857161e-01 -1.41305268e+00
-1.08112419e+00 -6.12835467e-01 3.70521605e-01 8.22767377e-01
9.06377435e-01 -1.13029099e+00 1.01267195e+00 1.05581641e+00
5.98318458e-01 -5.09854674e-01 -7.83895314e-01 -8.00757647e-01
5.10234237e-01 -2.17216671e-01 7.55517960e-01 9.83008683e-01
3.11709732e-01 3.38629365e-01 -2.40341321e-01 1.17504178e-03
8.74700189e-01 4.16753083e-01 7.31224060e-01 -1.98905599e+00
-5.90673447e-01 -7.11047113e-01 -2.46690795e-01 -9.30386364e-01
2.50158727e-01 -4.31757390e-01 -3.78232837e-01 -8.96308899e-01
7.54784524e-01 -6.08201325e-01 -2.06968009e-01 -1.30292196e-02
1.32387578e-01 -4.26796943e-01 4.61085945e-01 5.74298084e-01
-2.37016976e-01 5.84107786e-02 8.77409458e-01 1.83323890e-01
-1.87783763e-01 3.92596930e-01 -1.32081246e+00 1.09124553e+00
9.32305217e-01 -4.80087370e-01 -3.29474181e-01 -7.15807304e-02
7.00414240e-01 6.64207637e-01 5.85361242e-01 -4.51369375e-01
-5.24981096e-02 -1.09117031e+00 7.36429513e-01 -1.57672446e-02
2.46677265e-01 -8.96074712e-01 6.92441583e-01 7.34489858e-01
-7.71441996e-01 2.07929268e-01 -5.46242595e-01 4.42832053e-01
3.34282726e-01 -8.17854226e-01 4.65773672e-01 -5.21324635e-01
-5.74541204e-02 -7.92321190e-02 -6.02183282e-01 -2.12423801e-02
1.00034630e+00 -4.01611486e-03 -6.22683108e-01 -7.44434774e-01
-8.26712012e-01 -1.14968993e-01 7.33817697e-01 -3.42415385e-02
1.49469018e-01 -1.08585989e+00 -8.39985013e-01 5.39666228e-02
-2.11596191e-01 -6.44604921e-01 -1.57494664e-01 6.07414722e-01
-1.55610099e-01 5.18213987e-01 -1.92151949e-01 2.23518774e-01
-8.09461892e-01 3.34823936e-01 5.68496048e-01 -3.54646981e-01
-2.01277167e-01 6.47455394e-01 6.87712610e-01 2.62814909e-01
-1.72671422e-01 -1.51722029e-01 -1.51749447e-01 3.18496525e-01
2.29351670e-01 3.66992533e-01 -6.58825874e-01 -3.92580718e-01
6.31972924e-02 4.03731138e-01 -4.72178012e-02 -3.52749854e-01
1.54064524e+00 -2.55257428e-01 1.25581712e-01 6.82045937e-01
6.27883911e-01 1.58521608e-01 -1.32104421e+00 -1.67696387e-01
3.06225717e-01 -7.86837220e-01 -2.32901141e-01 -6.65358365e-01
-7.66581118e-01 5.35156906e-01 8.48134086e-02 1.07719660e+00
2.06991270e-01 -3.94787416e-02 -1.57312691e-01 5.96720755e-01
3.79908085e-01 -1.31328547e+00 5.45324665e-03 1.32234290e-01
8.52071762e-01 -1.36570978e+00 3.59180033e-01 2.27566794e-01
-1.35890090e+00 9.57474649e-01 3.27642977e-01 -2.06399664e-01
6.98750377e-01 -8.30798000e-02 -1.20101526e-01 -1.22924291e-01
-1.34493577e+00 1.56749234e-01 -1.86922222e-01 5.71812749e-01
1.18468262e-01 4.37451541e-01 -3.98994386e-01 1.32566881e+00
-7.10395455e-01 -1.22487865e-01 1.02583516e+00 4.07794714e-01
-6.94813728e-01 -9.41017687e-01 -3.28447640e-01 1.04621148e+00
-8.79358292e-01 -1.90591463e-03 -2.91747808e-01 8.70577455e-01
-1.59041345e-01 7.77870655e-01 4.55214739e-01 -1.08687416e-01
2.21451610e-01 2.09893510e-01 2.43480876e-02 -5.40599525e-01
-3.07754844e-01 7.77961016e-02 2.52976328e-01 -6.91846982e-02
-2.03014061e-01 -1.13990259e+00 -6.15048289e-01 -6.44838870e-01
-4.12781179e-01 2.09542423e-01 4.26021159e-01 6.87878609e-01
1.04288571e-01 -3.44204344e-03 1.02568614e+00 -6.14336550e-01
-1.48730087e+00 -9.64813948e-01 -1.19381249e+00 2.05601528e-01
3.56162488e-01 -6.49407387e-01 -1.08247626e+00 -2.76876479e-01] | [4.409146785736084, 3.085740566253662] |
61d64ed0-8565-42cb-92f0-5523bbe39d75 | halting-in-random-walk-kernels | null | null | http://papers.nips.cc/paper/5688-halting-in-random-walk-kernels | http://papers.nips.cc/paper/5688-halting-in-random-walk-kernels.pdf | Halting in Random Walk Kernels | Random walk kernels measure graph similarity by counting matching walks in two graphs. In their most popular form of geometric random walk kernels, longer walks of length $k$ are downweighted by a factor of $\lambda^k$ ($\lambda < 1$) to ensure convergence of the corresponding geometric series. We know from the field of link prediction that this downweighting often leads to a phenomenon referred to as halting: Longer walks are downweighted so much that the similarity score is completely dominated by the comparison of walks of length 1. This is a naive kernel between edges and vertices. We theoretically show that halting may occur in geometric random walk kernels. We also empirically quantify its impact in simulated datasets and popular graph classification benchmark datasets. Our findings promise to be instrumental in future graph kernel development and applications of random walk kernels. | ['Karsten Borgwardt', 'Mahito Sugiyama'] | 2015-12-01 | null | null | null | neurips-2015-12 | ['graph-similarity'] | ['graphs'] | [ 7.34373406e-02 3.22608322e-01 -1.50605112e-01 -2.26121694e-01
-3.87745142e-01 -7.68624008e-01 4.06592637e-01 7.29378402e-01
-2.87694484e-01 4.26299959e-01 -2.31263831e-01 -7.57409990e-01
-2.56673843e-01 -1.38181674e+00 -5.79559207e-01 -5.15062630e-01
-8.98927212e-01 4.51339006e-01 6.73502564e-01 -4.94539440e-02
2.56605476e-01 2.33892590e-01 -8.41226041e-01 -1.39404252e-01
7.27635145e-01 5.18287778e-01 -3.53518069e-01 9.11062777e-01
-1.66130364e-01 5.83661973e-01 -1.60221159e-01 -9.28601503e-01
4.69348878e-01 -5.03427804e-01 -9.24284697e-01 -2.48318434e-01
3.84355724e-01 2.37478446e-02 -4.06943053e-01 1.36401570e+00
3.95611264e-02 2.48519376e-01 8.49399388e-01 -1.25705445e+00
-5.17501712e-01 8.48427832e-01 -9.21161175e-01 2.17960879e-01
5.02124429e-01 -1.10476226e-01 1.42303967e+00 -5.03837287e-01
6.19854331e-01 7.84550607e-01 1.17020428e+00 -1.29445344e-02
-1.68373370e+00 -4.79887664e-01 -1.95597738e-01 -7.34835193e-02
-1.45086432e+00 4.74826023e-02 6.39178753e-01 -3.45045000e-01
9.51354325e-01 4.24859673e-01 7.56870508e-01 3.06885034e-01
3.15314829e-01 2.12323904e-01 7.48640954e-01 -3.87717843e-01
1.28256753e-01 -2.23575413e-01 5.29245555e-01 1.08502305e+00
7.04802990e-01 6.34229416e-03 -6.04214706e-02 -4.86311436e-01
5.49351752e-01 -1.00714182e-02 -1.90589651e-01 -6.60127819e-01
-8.73723507e-01 1.14837289e+00 5.93832076e-01 1.17398210e-01
-1.91244230e-01 6.57447755e-01 4.10079598e-01 6.12984538e-01
4.97711718e-01 3.96586537e-01 -5.43726683e-02 -5.53123886e-03
-7.28057504e-01 3.44508648e-01 1.00112605e+00 1.02917445e+00
1.18739450e+00 -4.75951433e-01 2.20426306e-01 5.92674851e-01
-1.74626801e-03 3.08756053e-01 3.46150659e-02 -6.26366258e-01
2.89218962e-01 7.89076686e-01 -2.50008088e-02 -1.30601490e+00
-4.12462413e-01 -1.29036739e-01 -8.02256823e-01 1.12836577e-01
1.00232661e+00 1.65885031e-01 -6.70991302e-01 1.51840246e+00
6.45503541e-03 2.33065143e-01 -5.31351864e-01 5.59198856e-01
1.70988381e-01 3.28460544e-01 2.57304937e-01 -6.02418743e-02
1.11749721e+00 -5.80806911e-01 -8.38284716e-02 2.03405824e-02
9.72844422e-01 -1.04673874e+00 1.24145901e+00 -2.87790187e-02
-1.08552015e+00 6.86041685e-03 -9.93669868e-01 3.25254083e-01
-4.91654664e-01 -5.62621117e-01 1.07758856e+00 8.45892012e-01
-1.02521658e+00 1.04625762e+00 -7.91569114e-01 -5.14879644e-01
2.91128427e-01 2.13641241e-01 -2.34164268e-01 -1.09346390e-01
-1.06056893e+00 7.41172731e-01 -5.70241585e-02 -3.72223079e-01
-6.54482320e-02 -7.52899826e-01 -6.78533852e-01 4.84754182e-02
2.25561291e-01 -4.66573358e-01 8.55854630e-01 -7.10690498e-01
-6.80676162e-01 1.02432644e+00 3.14438939e-02 -8.90720367e-01
3.86612743e-01 1.75431103e-01 -3.73247236e-01 -2.20251232e-02
1.00414366e-01 -1.94331378e-01 8.29308748e-01 -8.01002979e-01
-2.43025601e-01 -2.08118007e-01 1.10920653e-01 -2.10290849e-01
-7.89590552e-02 -3.47975254e-01 -1.94400579e-01 -5.38677156e-01
1.71025530e-01 -1.11134851e+00 -3.33000243e-01 -2.46254012e-01
-5.26836634e-01 -1.03152990e-01 4.02037412e-01 -2.73282737e-01
1.60974777e+00 -1.71507108e+00 -2.55724102e-01 9.17000949e-01
8.86529982e-01 1.25659987e-01 -1.34101748e-01 7.91794658e-01
-3.12584817e-01 2.96879798e-01 -2.98969030e-01 3.84287179e-01
-1.13618344e-01 -1.73409998e-01 -2.61271626e-01 7.48703837e-01
-1.48656145e-01 1.06869483e+00 -1.20370126e+00 -2.42288932e-01
-6.74838573e-02 2.40158439e-01 -4.54644144e-01 -3.00013453e-01
1.29127383e-01 -6.80664241e-01 -3.01198632e-01 3.04313600e-01
7.61973381e-01 -6.38289928e-01 2.10171536e-01 -2.64662236e-01
1.45294890e-01 3.85086149e-01 -1.05227053e+00 1.14911377e+00
-4.68339734e-02 4.32316989e-01 -5.33140838e-01 -9.72351432e-01
5.79186499e-01 -2.58619726e-01 4.12762165e-01 -2.72360891e-01
-7.48696998e-02 6.73036184e-03 1.54606521e-01 2.04296783e-03
3.85316193e-01 -3.02934676e-01 -3.44969518e-02 7.56151438e-01
-3.27027351e-01 -2.94332772e-01 3.87408465e-01 6.85426474e-01
1.99326789e+00 -4.38184917e-01 3.66542667e-01 -5.55689633e-01
2.66002148e-01 1.46709323e-01 4.60277982e-02 8.45842898e-01
-2.79994160e-01 2.83029288e-01 9.88309801e-01 -5.43424547e-01
-1.20504761e+00 -1.64165199e+00 1.47359356e-01 1.15452659e+00
3.31770569e-01 -7.60655880e-01 -8.50156903e-01 -7.79932141e-01
3.55900139e-01 5.42644858e-01 -7.27581024e-01 -5.25231123e-01
-3.06895554e-01 -8.14002752e-01 6.44176841e-01 3.68677497e-01
3.15007865e-01 -6.51296258e-01 -2.57991612e-01 2.10254833e-01
2.10553825e-01 -7.20402777e-01 -1.11797976e+00 2.73199230e-01
-9.01447892e-01 -1.59321439e+00 -5.16417980e-01 -6.87854409e-01
9.14536178e-01 5.55012107e-01 1.27067518e+00 2.99148113e-01
-5.89759588e-01 2.65813470e-01 -4.20758039e-01 1.62640452e-01
-3.25999707e-01 2.10028976e-01 -1.39349148e-01 -2.40602434e-01
6.55245006e-01 -6.08337045e-01 -8.72482598e-01 3.98736358e-01
-7.91571438e-01 -4.63436693e-01 3.23935449e-01 7.61300147e-01
3.29390347e-01 2.37414956e-01 1.32415863e-02 -1.11749232e+00
1.07121408e+00 -4.81623083e-01 -7.63003945e-01 3.48084301e-01
-9.03572381e-01 4.72046167e-01 5.34461558e-01 -7.24243283e-01
-1.68685704e-01 -1.94048911e-01 2.99954474e-01 -1.15326591e-01
4.79935944e-01 4.79381084e-01 4.99646246e-01 -5.93568683e-01
8.72684658e-01 2.29338501e-02 8.04515556e-02 5.19035570e-02
5.07766545e-01 -5.02251089e-03 -1.93778742e-02 -3.89726281e-01
1.01802790e+00 4.41804618e-01 4.34677303e-01 -7.97802687e-01
-3.06573868e-01 -4.96102333e-01 -2.28405342e-01 -1.06716745e-01
6.18927896e-01 -3.41784537e-01 -9.39143419e-01 2.82660633e-01
-8.03719938e-01 -3.73970211e-01 -3.30260634e-01 3.83604825e-01
-4.45323616e-01 5.20563245e-01 -7.33468533e-01 -7.14176774e-01
-1.61492705e-01 -7.28408992e-01 5.15846848e-01 1.12920344e-01
-5.61943412e-01 -1.16024351e+00 3.62281770e-01 -2.83569217e-01
3.25389981e-01 9.65510681e-02 1.37702179e+00 -7.19636023e-01
-2.92852968e-01 -7.25301981e-01 -5.29948473e-01 -1.66031495e-02
1.65186137e-01 4.79985736e-02 -2.67570406e-01 -2.15683892e-01
-6.24602735e-01 4.25621159e-02 8.65105629e-01 3.82867754e-01
6.11748338e-01 -5.56808673e-02 -6.18096352e-01 4.35417026e-01
1.53730893e+00 -1.52816370e-01 6.88336432e-01 3.44329188e-03
6.14309192e-01 3.34853113e-01 1.00391299e-01 -2.42706221e-02
1.36232957e-01 4.15936530e-01 1.16699837e-01 8.03678203e-03
1.57911986e-01 -4.89967674e-01 -3.17082293e-02 6.79741025e-01
-2.71735877e-01 1.17950486e-02 -1.14505577e+00 4.31088388e-01
-1.69349551e+00 -1.16830206e+00 -5.88172495e-01 2.73035789e+00
4.61276948e-01 6.03926897e-01 3.87972236e-01 1.62293941e-01
9.95117724e-01 1.83637992e-01 -2.55070299e-01 -6.45588934e-01
2.05430835e-01 5.72782874e-01 1.35674286e+00 8.27045679e-01
-7.77596295e-01 9.92667496e-01 6.83399582e+00 9.22187686e-01
-6.33017004e-01 -1.03095442e-01 5.49004972e-01 1.15173861e-01
-4.31124210e-01 4.31957096e-01 -3.81504983e-01 5.54970384e-01
7.35571146e-01 -5.11301517e-01 7.03588545e-01 7.54916668e-01
-1.87497646e-01 -3.07514399e-01 -7.58056521e-01 7.12717652e-01
-4.96869057e-01 -1.38749862e+00 -8.37065578e-02 2.51453966e-01
4.97037619e-01 5.00255562e-02 -1.37544423e-01 1.63855985e-01
9.46063161e-01 -1.04575682e+00 2.27411687e-02 3.16236764e-01
7.00472414e-01 -1.01418722e+00 4.56628114e-01 -8.20833147e-02
-1.80840254e+00 3.46374184e-01 -4.86158967e-01 -8.26572925e-02
1.57441303e-01 1.05619025e+00 -1.02429509e+00 4.23659384e-01
2.56229192e-01 2.54678130e-01 -4.88222659e-01 1.01267672e+00
7.38699064e-02 7.54212499e-01 -4.96623307e-01 -1.85947269e-01
1.62927032e-01 -6.03155434e-01 4.63106513e-01 1.03390038e+00
-2.03936477e-03 -5.87323755e-02 -1.25848085e-01 6.89562261e-01
3.48648056e-02 1.87867671e-01 -9.46593046e-01 -3.93116117e-01
4.61119264e-01 1.06584787e+00 -1.22039294e+00 -2.22212970e-01
-6.77853644e-01 1.07481253e+00 5.97511888e-01 3.80127639e-01
-9.00104165e-01 -8.45772624e-01 8.38277400e-01 7.67826617e-01
2.31107071e-01 -4.25691903e-01 -2.06889674e-01 -8.83018196e-01
-7.09730834e-02 -2.10436106e-01 4.16696459e-01 -3.86238724e-01
-1.56192398e+00 4.72525597e-01 -1.19451985e-01 -9.25879478e-01
3.83699127e-02 -4.72186208e-01 -8.83852065e-01 8.39515090e-01
-8.38787735e-01 -8.23391974e-01 -2.39811558e-03 5.50868034e-01
-1.32373378e-01 3.41799706e-01 8.88561904e-01 1.76850140e-01
-8.36976096e-02 8.11509728e-01 -8.77522007e-02 2.15458304e-01
6.57491267e-01 -1.32098186e+00 9.41708088e-01 5.50260544e-01
3.71493608e-01 8.34462345e-01 9.59656417e-01 -9.70936716e-01
-1.27262318e+00 -9.40080285e-01 8.94904613e-01 -2.08719060e-01
1.22006190e+00 -2.14812517e-01 -9.22063351e-01 5.51562965e-01
-3.48854393e-01 2.39727229e-01 7.65442252e-01 3.09721500e-01
-7.94530094e-01 5.96072190e-02 -1.17822528e+00 7.20182776e-01
1.34282219e+00 -7.03096092e-01 -1.22711964e-01 1.57605663e-01
4.65691149e-01 1.47663280e-01 -9.64009404e-01 4.21032429e-01
8.35905373e-01 -1.15240276e+00 1.08154380e+00 -6.93601966e-01
1.39886141e-01 3.11869960e-02 -5.23011349e-02 -1.28662169e+00
-4.33186740e-01 -8.97742271e-01 -1.59852412e-02 6.08324289e-01
6.58101976e-01 -7.99336612e-01 1.05731106e+00 3.94737840e-01
6.07396483e-01 -6.74881518e-01 -6.75306976e-01 -1.02955198e+00
2.00422257e-01 -4.41033363e-01 4.62752223e-01 9.42071319e-01
5.13783097e-01 1.81291595e-01 1.38457809e-02 -5.07440791e-02
9.70950723e-01 -1.47800222e-01 5.74641466e-01 -1.18974030e+00
-2.82023281e-01 -8.98431540e-01 -9.97627378e-01 -8.80707800e-01
-1.66552752e-01 -1.05552197e+00 -3.40563029e-01 -1.08130717e+00
2.47380003e-01 -5.52467227e-01 -7.92163387e-02 -1.48649812e-01
-2.36330479e-01 4.64607090e-01 -2.01291665e-01 -1.07485980e-01
-5.15958369e-01 -1.54800057e-01 9.15918589e-01 1.53601557e-01
-2.64871955e-01 9.59524140e-02 -4.24543023e-01 7.66725004e-01
8.66849124e-01 -5.06309450e-01 -5.05638897e-01 2.33272299e-01
7.44438827e-01 -5.14550246e-02 3.53664815e-01 -6.90970004e-01
9.92763937e-02 -1.54007254e-02 -4.26151976e-02 -3.81973803e-01
-1.80686519e-01 -5.53259552e-01 2.57778227e-01 6.46247685e-01
-4.03907269e-01 3.80713165e-01 -1.17936917e-01 8.70999515e-01
3.09658706e-01 -7.53517076e-02 8.93637598e-01 -4.46616597e-02
-2.65667588e-01 3.75259489e-01 -3.79708201e-01 3.69427294e-01
1.39848971e+00 -2.67665356e-01 -2.06116438e-01 -5.08474946e-01
-6.75043702e-01 -5.33625521e-02 7.17546165e-01 -3.62826586e-02
4.62020785e-01 -1.46685302e+00 -2.29101300e-01 2.85398632e-01
3.87140095e-01 -4.88441020e-01 -2.09079474e-01 8.92318249e-01
-8.25194418e-01 1.61186710e-01 1.50731549e-01 -2.52113521e-01
-1.25114238e+00 5.30740917e-01 2.75880128e-01 -5.19130230e-01
-6.32672310e-01 1.03532243e+00 -2.22510472e-02 -2.70526141e-01
4.18419912e-02 -2.81256676e-01 6.22753859e-01 -1.47679776e-01
4.41781491e-01 7.22802877e-01 -8.04320797e-02 -2.76544213e-01
-4.07519102e-01 6.86713636e-01 -1.84474289e-01 -8.95908400e-02
8.93208623e-01 5.23998290e-02 -2.08240911e-01 3.68212044e-01
1.38313413e+00 3.04363221e-01 -8.25142741e-01 -5.89296818e-02
3.18167537e-01 -4.65283245e-01 -4.94909465e-01 -1.97237596e-01
-6.93902850e-01 3.19064915e-01 2.71977127e-01 1.00699651e+00
7.04457223e-01 2.49803215e-01 9.69006896e-01 2.26527736e-01
5.50680041e-01 -8.94875944e-01 -1.15504012e-01 3.66387039e-01
2.31027976e-01 -1.00327671e+00 1.71569631e-01 -6.74128592e-01
-3.14167410e-01 9.63500321e-01 1.79319993e-01 -8.17704797e-01
1.13547409e+00 3.90034407e-01 -4.06584293e-01 -5.02322197e-01
-3.93669188e-01 -4.13897872e-01 2.09398687e-01 5.64723849e-01
5.77919781e-01 3.18021387e-01 -6.92887723e-01 1.39479816e-01
-2.36160621e-01 -2.84662634e-01 3.42077404e-01 7.41408169e-01
-7.29175806e-01 -1.28334403e+00 6.14865869e-03 9.76391792e-01
-2.18466938e-01 -5.37242353e-01 -3.90212655e-01 7.75781751e-01
-3.38268876e-01 6.22054398e-01 9.64951143e-02 -6.25001132e-01
1.56953424e-01 3.28688025e-01 7.20854342e-01 -3.33994925e-01
-3.47258568e-01 -4.73253906e-01 1.19274974e-01 -6.46178722e-01
1.13938011e-01 -2.96697438e-01 -1.36576343e+00 -1.05923223e+00
-4.45620924e-01 2.97807246e-01 4.02744710e-01 5.35990477e-01
7.98728839e-02 1.37162387e-01 3.84103626e-01 -3.47440541e-01
-5.68255186e-01 -6.81510091e-01 -1.00539529e+00 6.44675672e-01
8.84567201e-02 -5.18916011e-01 -7.64189780e-01 -1.28558278e-01] | [6.975478649139404, 5.4732666015625] |
2688f7a5-1576-4ccb-9ffc-74d8f8521a1a | collaboratively-boosting-data-driven-deep | 2010.02451 | null | https://arxiv.org/abs/2010.02451v1 | https://arxiv.org/pdf/2010.02451v1.pdf | Collaboratively boosting data-driven deep learning and knowledge-guided ontological reasoning for semantic segmentation of remote sensing imagery | As one kind of architecture from the deep learning family, deep semantic segmentation network (DSSN) achieves a certain degree of success on the semantic segmentation task and obviously outperforms the traditional methods based on hand-crafted features. As a classic data-driven technique, DSSN can be trained by an end-to-end mechanism and competent for employing the low-level and mid-level cues (i.e., the discriminative image structure) to understand images, but lacks the high-level inference ability. By contrast, human beings have an excellent inference capacity and can be able to reliably interpret the RS imagery only when human beings master the basic RS domain knowledge. In literature, ontological modeling and reasoning is an ideal way to imitate and employ the domain knowledge of human beings, but is still rarely explored and adopted in the RS domain. To remedy the aforementioned critical limitation of DSSN, this paper proposes a collaboratively boosting framework (CBF) to combine data-driven deep learning module and knowledge-guided ontological reasoning module in an iterative way. | ['Yongjun Zhang', 'Song Ouyang', 'Yansheng Li'] | 2020-10-06 | null | null | null | null | ['segmentation-of-remote-sensing-imagery'] | ['miscellaneous'] | [ 2.25585774e-01 3.49747986e-01 -2.18778178e-01 -6.70500994e-01
-1.10706620e-01 -1.45464659e-01 5.18296063e-01 4.55789566e-02
-4.78199661e-01 3.65302205e-01 -1.91548586e-01 -2.56357431e-01
-3.17358255e-01 -1.04649282e+00 -4.93013024e-01 -4.72421616e-01
3.18629354e-01 4.74902958e-01 3.14245403e-01 -5.96909106e-01
3.19348186e-01 2.00926259e-01 -1.72850323e+00 1.32443145e-01
1.48967683e+00 1.20420682e+00 6.12041593e-01 1.34333640e-01
-4.59294081e-01 1.05289948e+00 -3.17670137e-01 -3.76103669e-01
-1.90608460e-03 -5.64665079e-01 -1.09581959e+00 6.05991744e-02
-9.03443024e-02 -1.79251075e-01 -1.78168118e-01 1.12289774e+00
3.04374009e-01 2.24539757e-01 5.60956359e-01 -1.13031983e+00
-7.84467578e-01 3.53231668e-01 -3.48344266e-01 1.84944764e-01
2.85018563e-01 2.74973866e-02 8.51235151e-01 -5.72021246e-01
2.64982134e-01 1.14820540e+00 4.33837175e-01 5.16472220e-01
-7.11172223e-01 -3.65082145e-01 3.47011864e-01 4.31941807e-01
-1.19689226e+00 -1.35930166e-01 9.38425660e-01 -3.82549167e-01
5.27323544e-01 -4.24370058e-02 6.15467489e-01 8.89666080e-01
-3.12504023e-01 1.04851341e+00 1.22171485e+00 -3.37260902e-01
4.03592020e-01 1.42526522e-01 1.66998193e-01 7.85947978e-01
4.24039327e-02 4.07973416e-02 -5.71492195e-01 4.28126633e-01
8.77977967e-01 1.63093939e-01 -3.61605291e-03 -3.24350387e-01
-1.13695478e+00 7.38040149e-01 1.12425709e+00 4.97219980e-01
-6.38475955e-01 -7.49354064e-02 2.44495898e-01 6.66677952e-03
2.76027232e-01 3.86649281e-01 -5.00965357e-01 7.14149103e-02
-8.60808015e-01 3.49201681e-03 5.86018682e-01 6.86038852e-01
8.66613686e-01 9.87015143e-02 -6.68270737e-02 7.21519768e-01
4.18288648e-01 2.53646046e-01 6.18052185e-01 -9.70739663e-01
1.81304425e-01 1.04163289e+00 -7.53347650e-02 -1.01910591e+00
-3.40105116e-01 -6.48801565e-01 -8.74815762e-01 7.98316449e-02
1.56899944e-01 1.33043258e-02 -1.10271013e+00 1.76329064e+00
4.33450162e-01 1.69614866e-01 2.85734296e-01 1.24045205e+00
1.11606574e+00 3.60602409e-01 4.93439138e-01 2.39949346e-01
1.56597769e+00 -1.07841885e+00 -4.35171753e-01 -7.70520926e-01
2.54629463e-01 -1.19856745e-01 9.79623318e-01 4.48028922e-01
-7.77431428e-01 -8.66857052e-01 -1.15048194e+00 -2.12418586e-01
-5.84608138e-01 1.57672167e-01 1.03752398e+00 5.43659687e-01
-9.59796906e-01 3.65254998e-01 -5.91583192e-01 -4.43699807e-01
8.91158342e-01 1.42907172e-01 -1.30784661e-01 -2.06973597e-01
-1.49843740e+00 1.02513885e+00 9.09694612e-01 4.82316434e-01
-1.02645504e+00 -2.55320847e-01 -6.85671687e-01 -6.33468330e-02
6.72162175e-01 -9.23123896e-01 1.11892772e+00 -1.38814962e+00
-1.60558057e+00 1.04292929e+00 1.31532764e-02 -4.97231454e-01
4.73724574e-01 -1.86843947e-01 -2.08818540e-01 4.83403474e-01
1.58153191e-01 7.95193553e-01 9.59332526e-01 -1.39931333e+00
-6.97740436e-01 -5.97093761e-01 4.95741785e-01 4.60110009e-01
-2.60297477e-01 -4.12503958e-01 -3.32449198e-01 -4.39670444e-01
5.06432712e-01 -5.54835021e-01 -3.64758790e-01 -2.49896899e-01
-3.06383431e-01 -4.36632842e-01 4.78097588e-01 -7.35355735e-01
8.36951494e-01 -2.03036666e+00 1.81352258e-01 1.62498072e-01
2.04315156e-01 5.55940092e-01 1.58356681e-01 5.13214432e-02
2.11977959e-01 2.28079762e-02 -3.87935609e-01 1.76233575e-01
9.06270966e-02 3.59695613e-01 -4.50176708e-02 1.80666316e-02
1.77947640e-01 9.60373342e-01 -1.28180885e+00 -6.17891729e-01
3.86503905e-01 1.70250729e-01 -3.18244249e-01 3.23901474e-01
-4.53633010e-01 7.72471428e-01 -8.47280204e-01 6.15309238e-01
5.13526797e-01 -4.26980138e-01 2.89242625e-01 -2.70097107e-01
2.85607755e-01 9.16517898e-02 -8.44518960e-01 1.94186735e+00
-7.51146019e-01 2.81469405e-01 -5.39459810e-02 -1.60458386e+00
1.13208640e+00 2.68440574e-01 2.86570162e-01 -1.12288892e+00
2.39287794e-01 2.70620286e-01 2.30117608e-02 -7.80328333e-01
1.13674678e-01 -2.82091200e-01 -1.05562717e-01 1.23576365e-01
1.69077128e-01 -9.44272652e-02 -7.43334442e-02 8.02888423e-02
7.63147593e-01 4.34558690e-01 2.78702796e-01 -1.04829736e-01
8.65874648e-01 2.90018767e-01 6.21400595e-01 6.65128171e-01
-2.00295717e-01 2.59837776e-01 2.86045790e-01 -5.23987889e-01
-7.39398837e-01 -8.72002661e-01 -8.42096955e-02 1.38281786e+00
6.27385259e-01 3.32807899e-02 -1.11758256e+00 -6.23270810e-01
-1.01500958e-01 8.22549641e-01 -5.33257544e-01 -1.84693247e-01
-2.23581553e-01 -6.09438479e-01 2.72888303e-01 7.34985948e-01
1.39998984e+00 -1.42719269e+00 -6.71554625e-01 1.79854244e-01
-3.07418317e-01 -1.32190216e+00 4.19797927e-01 4.01526578e-02
-8.13597798e-01 -1.05372989e+00 -4.75297391e-01 -9.10584807e-01
7.74140775e-01 3.23594153e-01 1.15615892e+00 4.54312116e-01
-6.75811917e-02 9.51302573e-02 -5.77665567e-01 -4.51512307e-01
-8.21539685e-02 1.97681218e-01 -2.88496703e-01 7.21403584e-02
6.22217298e-01 -5.80585182e-01 -8.00711513e-01 4.05185699e-01
-9.78396595e-01 3.69220257e-01 8.39491546e-01 7.74369895e-01
3.39060336e-01 5.57130933e-01 9.03861940e-01 -7.32646585e-01
4.66718912e-01 -4.61084038e-01 -2.88518935e-01 2.99439967e-01
-5.45831203e-01 -4.54081967e-02 5.97730935e-01 3.71772386e-02
-1.52658451e+00 -1.34703852e-02 -3.29916716e-01 -1.45808756e-01
-4.63570267e-01 7.37303734e-01 -3.21405172e-01 -3.77818868e-02
5.83048344e-01 2.80102968e-01 -5.72963469e-02 -5.36229134e-01
5.35299778e-01 8.22345078e-01 7.86478281e-01 -8.81436884e-01
5.49947858e-01 5.23708463e-01 -2.42799476e-01 -5.05464017e-01
-1.45399988e+00 -3.69010359e-01 -8.83805752e-01 -1.63605928e-01
1.25648022e+00 -9.51206565e-01 -7.54151642e-01 6.03682458e-01
-9.78281677e-01 -3.21102202e-01 -1.28826082e-01 2.34371513e-01
-5.54620802e-01 2.50859290e-01 -3.25778455e-01 -6.89845264e-01
-1.85743585e-01 -9.10564780e-01 9.49029744e-01 6.26639605e-01
1.25758588e-01 -9.20470536e-01 -4.91724759e-01 1.01187837e+00
3.44197035e-01 1.56390786e-01 9.89847839e-01 -8.30182374e-01
-5.88181376e-01 -1.17616609e-01 -5.47140062e-01 5.74170828e-01
-5.15059978e-02 -4.57098365e-01 -1.10129941e+00 8.06653127e-02
1.91198781e-01 -5.63337982e-01 8.63843381e-01 1.78347424e-01
1.34788573e+00 1.29512444e-01 -1.70335487e-01 3.96531343e-01
1.41302717e+00 2.05712393e-01 6.05574131e-01 6.26684964e-01
7.77438462e-01 8.68591905e-01 7.50175118e-01 1.39445379e-01
8.97221982e-01 3.76159251e-01 6.14098549e-01 -2.57429510e-01
1.62234947e-01 -4.75218147e-01 -1.03649966e-01 4.72442448e-01
-3.52935076e-01 -6.15561679e-02 -1.09438622e+00 4.44839090e-01
-2.05198097e+00 -7.19824076e-01 4.05475572e-02 1.81291246e+00
7.30104268e-01 3.09574366e-01 -1.45946607e-01 8.30354244e-02
7.77655065e-01 8.69718194e-02 -8.25044394e-01 -3.01298976e-01
1.13860331e-01 8.35489258e-02 1.93592072e-01 3.84563692e-02
-1.04532599e+00 1.30510485e+00 5.64649534e+00 1.06117988e+00
-9.44025576e-01 1.28182098e-01 6.28569603e-01 6.85654700e-01
-2.85082072e-01 4.00309488e-02 -4.64404702e-01 5.00304401e-01
7.42859185e-01 2.46196598e-01 4.48803872e-01 1.04492629e+00
4.08440717e-02 -2.66837537e-01 -8.72390389e-01 8.97106051e-01
-4.64536734e-02 -1.04158080e+00 -7.02625960e-02 -7.48065189e-02
6.38868213e-01 -2.25317970e-01 -1.32481847e-02 6.04086041e-01
3.48164141e-01 -1.29128659e+00 6.64192140e-01 6.53558254e-01
3.01911920e-01 -6.45613313e-01 1.02206981e+00 8.37320447e-01
-9.74216104e-01 -3.31936687e-01 -3.34878504e-01 -1.46486446e-01
-4.16973867e-02 5.49742520e-01 -5.56309581e-01 9.46933806e-01
6.97868288e-01 6.50983691e-01 -6.47576511e-01 6.72097862e-01
-8.35666716e-01 5.04529178e-01 -1.18119985e-01 9.99655798e-02
5.27718544e-01 -4.49922740e-01 -3.24420668e-02 7.74901628e-01
1.07443653e-01 3.72854412e-01 3.53475899e-01 8.45923781e-01
1.28519267e-01 -1.47295579e-01 -2.39919335e-01 -1.01548091e-01
3.00009519e-01 1.20140219e+00 -9.28793371e-01 -4.74496007e-01
-3.09553623e-01 9.71017778e-01 2.88461357e-01 4.55319554e-01
-6.96559131e-01 -2.24260092e-01 2.03795120e-01 -1.32414833e-01
3.57470363e-01 -3.37703638e-02 -5.09176373e-01 -1.00710273e+00
3.23157087e-02 -8.98324907e-01 2.48702362e-01 -9.70823705e-01
-1.32000339e+00 6.06347978e-01 -1.52825370e-01 -8.52202058e-01
-1.89313546e-01 -6.48546994e-01 -5.32907963e-01 8.17335725e-01
-1.81792581e+00 -1.39777100e+00 -7.79598892e-01 6.03074789e-01
4.11814570e-01 -1.71155229e-01 7.30022013e-01 1.24876402e-01
-6.02740943e-01 2.38236859e-02 -1.88002408e-01 4.43666130e-01
1.04260825e-01 -1.24615335e+00 1.35075733e-01 6.69679701e-01
-3.31957154e-02 5.94667196e-01 5.67329764e-01 -4.21938300e-01
-1.06075966e+00 -9.02547181e-01 4.83925581e-01 -1.40351102e-01
3.60378206e-01 -1.35142475e-01 -1.01495218e+00 4.94991004e-01
-6.71787094e-03 -3.32029052e-02 7.05252588e-01 1.95542350e-01
-2.03422695e-01 -2.70885527e-01 -1.22411191e+00 4.91742760e-01
1.24045026e+00 -7.06360638e-01 -1.08038008e+00 3.33709776e-01
5.77360988e-01 -2.32062146e-01 -7.98020244e-01 4.97391254e-01
3.11240435e-01 -1.18978679e+00 1.18364739e+00 -4.80994612e-01
5.02618849e-01 -4.39781934e-01 -6.65599480e-02 -1.09788561e+00
-1.03409067e-01 -9.84237343e-02 1.77552924e-01 1.22787356e+00
1.02094941e-01 -6.42164528e-01 8.35391283e-01 5.58015943e-01
-2.05140606e-01 -7.40511417e-01 -5.88383377e-01 -6.51309609e-01
-1.74758270e-01 -6.11167490e-01 8.20365906e-01 9.08605635e-01
-2.00562805e-01 4.15300488e-01 -3.21474597e-02 2.14400649e-01
5.36059678e-01 4.48704094e-01 5.17599225e-01 -1.47711968e+00
3.51034328e-02 -4.52923059e-01 -5.34778655e-01 -1.24897635e+00
3.72430921e-01 -9.09515262e-01 2.84989417e-01 -1.96891308e+00
6.04105853e-02 -6.55704975e-01 -5.41358531e-01 4.59363252e-01
-3.84996235e-01 1.07392512e-01 4.75779660e-02 2.45756894e-01
-8.45570385e-01 5.33025444e-01 1.50842547e+00 2.38219532e-03
2.01225597e-02 -1.70351695e-02 -1.08670998e+00 1.12723947e+00
7.64929950e-01 -1.78999484e-01 -7.58680165e-01 -3.85824680e-01
3.16113919e-01 1.14716135e-01 8.00580502e-01 -9.90449607e-01
4.26747918e-01 -2.43735328e-01 4.08322304e-01 -5.40494442e-01
1.02529906e-01 -8.79568040e-01 -1.85784027e-01 2.48815984e-01
-2.45244622e-01 -5.98158121e-01 -2.50669628e-01 5.81577003e-01
-5.34014940e-01 -3.65522772e-01 6.69083893e-01 -5.74527144e-01
-1.44849098e+00 2.70404637e-01 -1.73187256e-01 1.68186963e-01
9.41908717e-01 -4.42783356e-01 -2.10485607e-01 -1.71899617e-01
-7.21646607e-01 4.57795173e-01 4.03961927e-01 3.55652899e-01
5.72704017e-01 -1.12811601e+00 -3.79660189e-01 1.25364557e-01
3.81910726e-02 5.13678551e-01 5.37229955e-01 7.33694375e-01
-6.13953233e-01 3.23692173e-01 -4.50957268e-01 -6.67682350e-01
-3.71022701e-01 4.84205037e-01 4.28035915e-01 -9.64665487e-02
-3.80515009e-01 1.02486050e+00 2.23254085e-01 -6.69820487e-01
5.32471761e-02 -6.42291456e-02 -4.87581223e-01 1.27542853e-01
2.86526829e-01 1.07713491e-01 2.00242586e-02 -6.00093067e-01
-4.83606398e-01 4.59116369e-01 1.88959554e-01 1.05097108e-01
1.30422711e+00 -1.42299101e-01 -1.37200475e-01 1.31223649e-01
8.07073534e-01 -7.18703985e-01 -1.41714776e+00 -4.54822898e-01
2.52551109e-01 -3.70776206e-01 3.20458025e-01 -9.86528635e-01
-1.18051815e+00 1.22374368e+00 4.46994632e-01 3.22866946e-01
1.32085145e+00 -5.35230823e-02 9.20038760e-01 3.80923003e-01
6.12452626e-01 -1.35899782e+00 3.68299097e-01 2.66499251e-01
4.83973891e-01 -1.56783044e+00 -4.35156450e-02 -4.19006288e-01
-9.24946904e-01 9.71859932e-01 8.03462327e-01 -1.08443744e-01
4.96655673e-01 -2.96478897e-01 -2.20563728e-02 -4.52570975e-01
-2.47458711e-01 -7.30719507e-01 2.90067852e-01 8.28387856e-01
1.39103785e-01 2.08854582e-02 -7.90627822e-02 7.24357486e-01
-1.73619941e-01 2.45831326e-01 -2.24181414e-01 8.57523501e-01
-7.77961969e-01 -8.32295716e-01 -1.73614666e-01 2.42131144e-01
-2.43802011e-01 -5.95906004e-02 -1.12073541e-01 5.82183719e-01
5.57857037e-01 1.08345926e+00 -4.92864028e-02 -3.31823498e-01
2.05486983e-01 -4.00506817e-02 3.08842212e-01 -5.37763357e-01
-3.37973565e-01 -4.49738711e-01 -1.11789098e-02 -5.11871219e-01
-8.38000476e-01 -2.14443818e-01 -1.60371375e+00 -2.19573036e-01
-1.53329410e-02 4.84587960e-02 7.44882643e-01 1.54747009e+00
6.92545027e-02 7.43141174e-01 5.92622578e-01 -7.52081513e-01
-3.01434934e-01 -7.54903972e-01 -4.91191804e-01 5.80128491e-01
8.02298635e-02 -7.78193772e-01 -2.18884736e-01 6.43288046e-02] | [9.679390907287598, 0.6810242533683777] |
c2fe9dda-0f0d-463c-a164-5358a1892676 | where-will-players-move-next-dynamic-graphs | 2211.12217 | null | https://arxiv.org/abs/2211.12217v2 | https://arxiv.org/pdf/2211.12217v2.pdf | Where Will Players Move Next? Dynamic Graphs and Hierarchical Fusion for Movement Forecasting in Badminton | Sports analytics has captured increasing attention since analysis of the various data enables insights for training strategies, player evaluation, etc. In this paper, we focus on predicting what types of returning strokes will be made, and where players will move to based on previous strokes. As this problem has not been addressed to date, movement forecasting can be tackled through sequence-based and graph-based models by formulating as a sequence prediction task. However, existing sequence-based models neglect the effects of interactions between players, and graph-based models still suffer from multifaceted perspectives on the next movement. Moreover, there is no existing work on representing strategic relations among players' shot types and movements. To address these challenges, we first introduce the procedure of the Player Movements (PM) graph to exploit the structural movements of players with strategic relations. Based on the PM graph, we propose a novel Dynamic Graphs and Hierarchical Fusion for Movement Forecasting model (DyMF) with interaction style extractors to capture the mutual interactions of players themselves and between both players within a rally, and dynamic players' tactics across time. In addition, hierarchical fusion modules are designed to incorporate the style influence of both players and rally interactions. Extensive experiments show that our model empirically outperforms both sequence- and graph-based methods and demonstrate the practical usage of movement forecasting. | ['Wen-Chih Peng', 'Wei-Yao Wang', 'Kai-Shiang Chang'] | 2022-11-22 | null | null | null | null | ['sports-analytics'] | ['computer-vision'] | [ 5.91177009e-02 -2.73501754e-01 -4.68205631e-01 2.20719695e-01
-4.94234897e-02 -5.13658524e-01 6.02626145e-01 -9.19845998e-02
-8.47260505e-02 4.01377797e-01 6.92401826e-01 -2.19525337e-01
-3.98338199e-01 -1.11876988e+00 -3.23660493e-01 -3.38357866e-01
-9.70769674e-02 4.75588590e-01 6.45685136e-01 -8.16478968e-01
5.93439460e-01 4.30801779e-01 -1.80559862e+00 5.08970022e-01
8.90323699e-01 4.71767873e-01 1.46289229e-01 8.91053021e-01
-4.12195563e-01 1.43872428e+00 -7.45639563e-01 -6.11153960e-01
2.95677364e-01 -9.07490075e-01 -6.36069477e-01 -2.75237300e-03
-9.34360325e-02 -1.63938344e-01 -8.37910414e-01 5.10600090e-01
4.11464483e-01 4.79948342e-01 4.93689984e-01 -1.35767519e+00
-2.41617709e-01 1.19197130e+00 -7.31904030e-01 5.57738245e-01
6.08470619e-01 2.98966259e-01 9.09296632e-01 -2.85571396e-01
9.00504768e-01 1.02435637e+00 7.40926862e-01 3.89785320e-01
-6.02037549e-01 -6.29439950e-01 4.24149096e-01 8.20090830e-01
-8.24929595e-01 9.90696698e-02 1.09924006e+00 -6.98957860e-01
5.67810833e-01 4.97207761e-01 1.36573827e+00 9.28676605e-01
1.14244640e-01 1.25581884e+00 7.49481380e-01 -3.20708364e-01
-1.95655093e-01 -6.32694244e-01 2.96076089e-01 5.40158629e-01
-1.40417263e-01 5.12007214e-02 -8.79775465e-01 1.51406810e-01
7.85221517e-01 2.02154115e-01 -2.75790393e-02 3.65881883e-02
-9.73926306e-01 6.49271846e-01 2.18150690e-01 4.20159757e-01
-5.09344339e-01 2.11045239e-02 3.78364682e-01 9.34123546e-02
4.26514208e-01 2.80295074e-01 9.87923592e-02 -8.88351977e-01
-1.06351805e+00 6.63943410e-01 8.55436385e-01 7.98785746e-01
4.96854514e-01 7.58948699e-02 -4.79074240e-01 6.41169429e-01
-1.56891376e-01 -9.66774672e-02 4.92861629e-01 -6.57011092e-01
4.97953892e-01 1.31206119e+00 -3.32316965e-01 -1.71519732e+00
-6.45237505e-01 -5.57263553e-01 -7.01233506e-01 -1.82745874e-01
3.89284343e-01 2.93877665e-02 -5.85502803e-01 1.32756650e+00
2.51789868e-01 7.28412867e-01 -4.03440118e-01 7.67390847e-01
1.02796626e+00 4.79316890e-01 -1.19024999e-01 -1.21113226e-01
1.08863103e+00 -1.26679242e+00 -7.02435374e-01 -2.25990355e-01
7.78916776e-01 -5.08106530e-01 8.94432604e-01 3.76437187e-01
-1.16532993e+00 -7.29391336e-01 -8.10469389e-01 2.09248081e-01
-4.49385196e-02 6.18351102e-02 7.41760492e-01 3.76530588e-01
-5.66301525e-01 7.04129815e-01 -7.22790718e-01 -8.33216906e-02
2.99583256e-01 1.60566330e-01 1.85386062e-01 2.12409183e-01
-1.24804080e+00 8.26247394e-01 2.92130351e-01 3.27859670e-01
-5.28316975e-01 -7.26597309e-01 -4.77942050e-01 5.57738310e-03
9.63105500e-01 -5.07383883e-01 9.94118690e-01 -7.45701790e-01
-1.47426629e+00 3.71317267e-01 8.91752914e-03 -4.72132921e-01
7.63811350e-01 -1.57129154e-01 -3.58475417e-01 -1.83042496e-01
-4.71783355e-02 -1.17395222e-01 1.78363234e-01 -1.04857969e+00
-1.06740212e+00 -4.90292668e-01 4.30404305e-01 6.26556516e-01
-3.60335201e-01 1.26745645e-02 -9.44019258e-01 -7.10777223e-01
6.16861181e-03 -1.02907801e+00 -3.12656045e-01 -7.32554376e-01
-5.32265544e-01 -3.06860685e-01 5.43850780e-01 -7.98677564e-01
2.24634337e+00 -1.82540095e+00 6.04028463e-01 2.92586982e-01
5.16275227e-01 4.03232783e-01 2.70083863e-02 1.06353319e+00
4.31280255e-01 4.26654890e-03 1.85479417e-01 2.22258434e-01
6.89276680e-02 2.71459013e-01 -2.75570244e-01 -6.09870851e-02
-5.99490166e-01 1.17062545e+00 -9.39847469e-01 -4.93399143e-01
2.58494735e-01 6.21851236e-02 -5.61015546e-01 2.18001515e-01
9.03336108e-02 3.72681051e-01 -6.76176965e-01 4.82474476e-01
2.51853436e-01 1.83377117e-02 4.96927351e-01 1.76685471e-02
-2.75534600e-01 7.56588951e-02 -1.20130515e+00 1.39675093e+00
-1.10160366e-01 5.03101766e-01 -2.88661629e-01 -1.01266110e+00
9.44657385e-01 -2.41530567e-01 7.54413307e-01 -6.32033944e-01
2.80809645e-02 -1.60748616e-01 3.92180741e-01 -7.09439039e-01
8.27722669e-01 1.49493963e-01 -3.31724167e-01 5.25975287e-01
-2.84355164e-01 3.30807477e-01 7.78129697e-01 5.75971901e-01
1.37396061e+00 4.52558756e-01 2.31382012e-01 3.48219573e-01
4.20179307e-01 4.33730513e-01 8.12669158e-01 9.21068132e-01
-5.06006889e-02 1.59835503e-01 7.29311168e-01 -5.88159442e-01
-6.13072991e-01 -5.85996151e-01 7.74628282e-01 1.47674155e+00
5.52424490e-01 -9.78130162e-01 -7.28584647e-01 -4.77495402e-01
-6.67571723e-02 3.70951086e-01 -8.35865915e-01 -2.28392795e-01
-1.05137038e+00 -6.35539174e-01 7.32746840e-01 6.68943048e-01
4.55833346e-01 -1.21416950e+00 -4.91159320e-01 4.87498820e-01
-3.54051799e-01 -7.88925052e-01 -5.94293296e-01 -5.55760026e-01
-6.65121377e-01 -1.49047136e+00 -2.76015848e-01 -4.93897617e-01
-4.63979542e-02 3.52676690e-01 1.08340180e+00 4.23271298e-01
-3.52691025e-01 2.79990047e-01 -7.02017605e-01 -3.41051579e-01
-2.74497688e-01 2.77983189e-01 -1.88927382e-01 3.79442200e-02
3.36408854e-01 -7.41947293e-01 -6.93136454e-01 6.13746881e-01
-6.30542994e-01 6.40357852e-01 5.00518680e-01 6.12952650e-01
2.82612413e-01 2.06812933e-01 2.83458412e-01 -1.15229130e+00
1.07546175e+00 -5.22700608e-01 -1.57415774e-03 4.59735841e-01
-4.27426070e-01 -3.32772166e-01 4.30507988e-01 -6.99327707e-01
-9.68217194e-01 -1.44933805e-01 7.95735568e-02 -3.25191647e-01
1.44264653e-01 7.02894509e-01 -7.89103508e-02 1.91012651e-01
5.88599741e-01 5.53671956e-01 -9.10442621e-02 -3.21931720e-01
5.55457115e-01 3.82426053e-01 5.99124670e-01 -5.09981453e-01
6.87700868e-01 3.32343310e-01 -3.26541141e-02 -5.60328484e-01
-7.19281316e-01 -6.35624588e-01 -8.46409380e-01 -1.10596001e+00
6.62621319e-01 -5.86366296e-01 -1.24137580e+00 5.72340846e-01
-7.59550869e-01 -2.50096798e-01 -2.64419168e-01 2.68922806e-01
-5.51672220e-01 5.01617193e-01 -7.18757510e-01 -9.88247395e-01
-4.12410358e-03 -1.01643431e+00 4.86157119e-01 5.34351766e-01
-3.37539554e-01 -1.01450729e+00 1.95695683e-01 7.03529537e-01
1.35731429e-01 4.35070515e-01 6.83957338e-01 -6.41127408e-01
-6.20885253e-01 -3.39657784e-01 2.32916385e-01 -3.79585207e-01
-5.74641824e-02 5.87506890e-02 -9.77484882e-02 2.64572859e-01
-4.55100685e-01 2.02658862e-01 7.41901994e-01 5.07410407e-01
9.29183722e-01 -1.77420601e-01 -6.03883505e-01 5.13666391e-01
8.24461520e-01 4.58785057e-01 7.85493493e-01 5.53051054e-01
1.15381753e+00 9.08940256e-01 7.63339400e-01 6.42180443e-01
7.41560817e-01 1.00213349e+00 2.64294952e-01 2.47664168e-01
-2.48009697e-01 -7.00780988e-01 2.16657251e-01 1.06241155e+00
-9.16490853e-01 -4.51696813e-01 -8.76768231e-01 2.89402217e-01
-2.41459441e+00 -1.61842465e+00 -6.65861189e-01 1.67058659e+00
3.25626224e-01 2.01057658e-01 8.27063680e-01 3.49955916e-01
7.57991493e-01 3.74490976e-01 -5.50082326e-01 -3.14518601e-01
-6.57531768e-02 -8.53693858e-02 3.91565949e-01 2.39899427e-01
-6.69864357e-01 1.21761513e+00 6.18595457e+00 1.36324990e+00
-6.25598192e-01 -1.28009945e-01 2.32324496e-01 -2.76678890e-01
-3.71257991e-01 1.70784831e-01 -8.14613760e-01 6.05921209e-01
4.11447555e-01 -4.61194873e-01 4.24998134e-01 4.64132220e-01
3.69173050e-01 -3.02383763e-04 -5.14423192e-01 9.59454954e-01
1.58183023e-01 -1.65638745e+00 1.88839763e-01 3.07020903e-01
4.59935963e-01 -6.26651764e-01 -2.92400211e-01 7.40726233e-01
6.91401064e-01 -8.76289189e-01 7.15528667e-01 1.03856063e+00
2.86156498e-02 -8.75882745e-01 4.20328975e-01 8.54350507e-01
-1.64114010e+00 -4.45954114e-01 4.79921661e-02 -6.56265378e-01
4.79791969e-01 2.20999330e-01 -5.74295044e-01 1.08846569e+00
3.94815058e-01 1.02378941e+00 -4.33609128e-01 8.65060568e-01
-2.90222347e-01 9.21143413e-01 2.08317004e-02 -2.69090891e-01
2.78100401e-01 -5.74759901e-01 6.73158228e-01 9.91516113e-01
4.17687565e-01 6.04456246e-01 5.27608991e-01 4.72467750e-01
3.28590691e-01 3.07653248e-01 -4.61956263e-01 -2.89442986e-01
4.98489618e-01 1.19219303e+00 -8.93800259e-01 -3.22004437e-01
-2.11989120e-01 7.25756288e-01 3.44789118e-01 1.16843201e-01
-8.46583188e-01 8.64406675e-03 5.30684710e-01 6.96304440e-01
-8.83877184e-03 -3.11921418e-01 -3.75678420e-01 -1.09242558e+00
-1.59266010e-01 -9.95070398e-01 7.35633790e-01 -4.16415006e-01
-9.41738367e-01 5.70616305e-01 1.25046168e-02 -1.38362563e+00
-4.26734924e-01 -2.14253858e-01 -1.07781565e+00 5.54735124e-01
-7.38654256e-01 -1.28430033e+00 -3.39607120e-01 4.46202606e-01
7.96409309e-01 -3.95575285e-01 1.05248481e-01 4.36599087e-03
-8.15448165e-01 3.81735206e-01 -3.17822397e-01 2.38967866e-01
5.27622998e-02 -1.06454611e+00 2.39699483e-01 9.21651185e-01
3.53018343e-01 3.05996150e-01 6.87780797e-01 -1.01232147e+00
-1.44147503e+00 -5.65388083e-01 6.37747347e-01 -3.60181302e-01
7.96613216e-01 -4.80017178e-02 -7.62595177e-01 5.45131862e-01
-1.17638163e-01 -5.99526763e-01 1.00152981e+00 5.73036790e-01
1.95882350e-01 1.78284958e-01 -2.18786433e-01 9.35634732e-01
1.90125132e+00 -9.25183967e-02 -6.13525331e-01 1.32703871e-01
1.57974809e-01 -6.75121784e-01 -6.79512799e-01 3.50214154e-01
7.86263347e-01 -1.10267830e+00 1.09251380e+00 -9.10364568e-01
6.93284690e-01 -1.31394073e-01 2.94874161e-01 -1.38916695e+00
-5.35390317e-01 -6.76129460e-01 -3.06022763e-01 1.20043504e+00
5.22364629e-03 -9.50650126e-02 1.09591603e+00 3.54156673e-01
-2.80312210e-01 -1.05347323e+00 -5.13269782e-01 -6.07626379e-01
-3.38004768e-01 -7.00236082e-01 7.99823105e-01 8.02375436e-01
4.78048593e-01 2.63363183e-01 -1.01424110e+00 -2.68200010e-01
3.68403256e-01 5.97076297e-01 1.33575904e+00 -1.29849696e+00
-6.49553597e-01 -9.94988918e-01 -7.00510442e-01 -1.18769169e+00
-4.66383174e-02 -8.81696641e-01 -3.06696087e-01 -1.83935452e+00
3.13033670e-01 -1.89293861e-01 6.55177683e-02 3.41639698e-01
-4.44655299e-01 1.61176518e-01 5.60210288e-01 5.27188778e-01
-7.21875966e-01 3.40409368e-01 1.59390616e+00 1.46515369e-02
-6.38896942e-01 3.49722564e-01 -6.80482984e-01 9.21480000e-01
5.25569618e-01 -1.88940898e-01 -5.64135492e-01 -1.73321262e-01
5.20579040e-01 5.08096218e-01 4.49994728e-02 -1.00739598e+00
6.25306487e-01 -6.17980599e-01 -7.12994486e-02 -7.61373639e-01
1.95444435e-01 -4.65868525e-02 5.96739113e-01 4.75391656e-01
-4.24649298e-01 1.49945647e-01 -1.03738680e-01 8.17220688e-01
-2.81677276e-01 7.83206299e-02 -6.31680563e-02 -1.14590645e-01
-9.66792524e-01 5.35343766e-01 -5.89255452e-01 3.46156545e-02
1.22964287e+00 -8.12490702e-01 -1.52424067e-01 -6.30973935e-01
-1.10506797e+00 5.08965492e-01 1.37850910e-01 5.56771755e-01
5.02210200e-01 -1.34046745e+00 -7.87982345e-01 -2.35071018e-01
-5.57589196e-02 -2.94352204e-01 8.67725670e-01 9.12559569e-01
-6.23003542e-01 -1.77237123e-01 -4.05028880e-01 -1.83175907e-01
-1.57306290e+00 4.36316371e-01 -1.25219056e-03 -8.96292150e-01
-7.67110467e-01 8.20939958e-01 8.51712450e-02 -2.54387975e-01
-8.04804340e-02 1.02730185e-01 -9.22714710e-01 3.80939275e-01
4.63383466e-01 9.36533153e-01 -4.22635138e-01 -9.30858552e-01
-9.31406543e-02 5.39752483e-01 3.63368564e-03 4.14304137e-02
1.41742015e+00 -4.39034589e-02 1.64675266e-01 6.62832260e-01
2.87392616e-01 1.84586585e-01 -1.21216893e+00 -2.10118011e-01
8.99021998e-02 -6.17781520e-01 -3.18129927e-01 -6.53951287e-01
-1.23575842e+00 8.44571888e-01 -1.53111398e-01 4.67925161e-01
1.14789391e+00 -2.08173469e-01 1.06242239e+00 -2.25689963e-01
3.78280014e-01 -1.23152077e+00 1.52509391e-01 4.97335434e-01
6.63495004e-01 -4.90745038e-01 3.77959246e-03 -5.41814446e-01
-1.08091724e+00 1.10109651e+00 7.93253601e-01 -1.91107810e-01
4.88832116e-01 3.17670017e-01 -1.63069904e-01 -3.88799667e-01
-7.52588630e-01 -5.95411003e-01 5.75054586e-01 3.50624621e-01
1.42811224e-01 2.37173870e-01 -8.68522942e-01 1.09298337e+00
-5.58469474e-01 1.72736198e-01 4.77742642e-01 9.23912883e-01
-5.53219438e-01 -1.34516418e+00 -1.57407343e-01 6.81629956e-01
-1.21456288e-01 1.69632137e-01 -6.83066487e-01 7.07246125e-01
4.28723782e-01 8.90715837e-01 -9.64749306e-02 -1.15506005e+00
7.02438831e-01 -1.68053567e-01 4.45455849e-01 -4.77822840e-01
-1.03661633e+00 -7.45234787e-02 1.65706128e-01 -5.97147584e-01
-4.44411814e-01 -7.30166078e-01 -1.12956071e+00 -8.36690247e-01
-2.11046204e-01 1.18734688e-01 1.64280161e-01 1.06068456e+00
3.05177212e-01 9.14386213e-01 4.55119252e-01 -9.03060555e-01
-5.06439209e-02 -9.70770001e-01 -7.67309666e-01 5.73650599e-01
-6.50276959e-01 -9.00756538e-01 -2.82134861e-03 -1.58265024e-01] | [6.720944881439209, 0.3468337655067444] |
a3d36659-c36f-45cd-a102-13d9944225bd | multimodal-cnn-networks-for-brain-tumor | 2212.09310 | null | https://arxiv.org/abs/2212.09310v1 | https://arxiv.org/pdf/2212.09310v1.pdf | Multimodal CNN Networks for Brain Tumor Segmentation in MRI: A BraTS 2022 Challenge Solution | Automatic segmentation is essential for the brain tumor diagnosis, disease prognosis, and follow-up therapy of patients with gliomas. Still, accurate detection of gliomas and their sub-regions in multimodal MRI is very challenging due to the variety of scanners and imaging protocols. Over the last years, the BraTS Challenge has provided a large number of multi-institutional MRI scans as a benchmark for glioma segmentation algorithms. This paper describes our contribution to the BraTS 2022 Continuous Evaluation challenge. We propose a new ensemble of multiple deep learning frameworks namely, DeepSeg, nnU-Net, and DeepSCAN for automatic glioma boundaries detection in pre-operative MRI. It is worth noting that our ensemble models took first place in the final evaluation on the BraTS testing dataset with Dice scores of 0.9294, 0.8788, and 0.8803, and Hausdorf distance of 5.23, 13.54, and 12.05, for the whole tumor, tumor core, and enhancing tumor, respectively. Furthermore, the proposed ensemble method ranked first in the final ranking on another unseen test dataset, namely Sub-Saharan Africa dataset, achieving mean Dice scores of 0.9737, 0.9593, and 0.9022, and HD95 of 2.66, 1.72, 3.32 for the whole tumor, tumor core, and enhancing tumor, respectively. The docker image for the winning submission is publicly available at (https://hub.docker.com/r/razeineldin/camed22). | ['Franziska Mathis-Ullrich', 'Oliver Burgert', 'Mohamed E. Karar', 'Ramy A. Zeineldin'] | 2022-12-19 | null | null | null | null | ['tumor-segmentation', 'brain-tumor-segmentation'] | ['computer-vision', 'medical'] | [-1.18410833e-01 -1.29706133e-03 1.95185483e-01 -2.26839721e-01
-1.14104950e+00 -6.21047854e-01 5.14820337e-01 4.15946931e-01
-6.84646010e-01 9.47422445e-01 1.78397343e-01 -4.40785617e-01
-4.05232787e-01 -4.70382303e-01 -2.20299080e-01 -1.12714958e+00
-3.57079715e-01 7.32580304e-01 2.65129358e-01 1.42431051e-01
1.18700825e-01 5.15305817e-01 -6.78720593e-01 8.49774033e-02
1.24392712e+00 1.12733400e+00 4.18810666e-01 5.74712574e-01
2.20376611e-01 3.60059828e-01 -3.52575868e-01 -3.19051474e-01
-5.11465743e-02 -1.34166539e-01 -8.18107963e-01 -3.27109396e-01
1.79560795e-01 -2.34222963e-01 -4.32995260e-01 1.16060305e+00
6.66999280e-01 -7.61114359e-02 9.21050847e-01 -8.26059937e-01
-1.17354974e-01 4.21831578e-01 -7.79269814e-01 3.53646100e-01
-2.64392614e-01 1.90476805e-01 5.60309649e-01 -6.97827280e-01
6.95909739e-01 2.43524462e-01 6.11768246e-01 4.69481379e-01
-6.74398780e-01 -7.34814227e-01 -1.47122279e-01 1.51448607e-01
-1.24122834e+00 -2.90445596e-01 -1.57374799e-01 -8.03344369e-01
7.46410191e-01 5.10954082e-01 7.46505439e-01 7.98325896e-01
7.63811767e-01 5.41776776e-01 1.45988512e+00 1.87812671e-01
2.82960624e-01 -5.37525833e-01 1.30207390e-01 7.78990149e-01
3.13947111e-01 -1.54355049e-01 9.76626128e-02 -1.00372016e-01
6.00574613e-01 1.56255197e-02 -5.54236829e-01 -1.26820177e-01
-1.29339564e+00 7.04871058e-01 7.17067599e-01 3.91907156e-01
-5.52587926e-01 -1.70463249e-01 5.44811010e-01 -3.24803144e-01
7.03916967e-01 2.17762552e-02 -2.38702446e-01 -3.29505764e-02
-9.30980742e-01 1.13795556e-01 2.85443127e-01 4.48283136e-01
-1.43636480e-01 -5.09301186e-01 -1.11953445e-01 8.51298571e-01
1.00893252e-01 1.60859525e-01 9.85177159e-01 -4.90231514e-01
2.48260453e-01 2.11139306e-01 -8.04408789e-02 -4.51921940e-01
-1.04320526e+00 -1.03095949e+00 -1.23016620e+00 -2.55630612e-02
5.22284985e-01 -1.75356433e-01 -1.23609471e+00 1.59332705e+00
1.65782988e-01 1.08602397e-01 2.54673306e-02 9.36273158e-01
1.03902459e+00 -1.04647405e-01 4.09193113e-02 -6.73897713e-02
1.40552270e+00 -1.06553984e+00 -4.09487575e-01 1.59427337e-02
1.01044846e+00 -5.89133441e-01 4.80308741e-01 4.19169933e-01
-1.07891095e+00 2.92837173e-01 -6.57356977e-01 3.20654273e-01
-3.41927171e-01 2.23445799e-02 5.93245685e-01 3.73272985e-01
-1.34454823e+00 2.45706901e-01 -1.17135191e+00 -4.11456823e-01
8.47036123e-01 4.59649026e-01 -6.99624956e-01 -2.78650999e-01
-8.46769810e-01 1.07226026e+00 3.67234111e-01 -2.04982147e-01
-1.11804080e+00 -7.24917531e-01 -3.04525912e-01 -3.10894847e-01
-5.58877289e-02 -6.80997729e-01 1.12695062e+00 -4.18699920e-01
-1.05347991e+00 1.05740356e+00 5.28118573e-02 -7.40276158e-01
9.13004816e-01 1.59214228e-01 -2.30646640e-01 2.07285002e-01
2.63038516e-01 7.24047482e-01 -7.90096000e-02 -7.36601710e-01
-6.14968538e-01 -1.04350460e+00 -4.19265896e-01 3.70641410e-01
2.39516735e-01 6.02912419e-02 -1.45145833e-01 -4.93289232e-01
2.27574691e-01 -8.99410427e-01 -5.01533151e-01 -5.07006347e-01
-4.91501808e-01 1.50785809e-02 4.18681264e-01 -1.21622574e+00
6.78116202e-01 -1.91547418e+00 1.32141620e-01 3.10085893e-01
7.19123244e-01 -6.37498125e-02 1.37724876e-01 -3.21825594e-01
-2.74984539e-01 6.61247373e-02 -5.54569125e-01 -4.84208539e-02
-3.41686964e-01 -3.15120220e-01 5.04051924e-01 8.06593180e-01
-3.75708699e-01 9.31290865e-01 -9.01428044e-01 -1.27382353e-01
-1.42060325e-03 4.75845665e-01 -9.03151929e-02 -6.30708039e-02
4.04580623e-01 7.08112180e-01 -2.01659635e-01 5.42398632e-01
6.20859921e-01 -1.27723217e-01 -1.27526939e-01 4.47197855e-02
-1.74795002e-01 -1.45636946e-01 -3.11371505e-01 1.79172885e+00
-8.65621865e-02 6.63654089e-01 2.90307160e-02 -9.21552479e-01
6.56053007e-01 3.45413297e-01 9.29204285e-01 -7.41722107e-01
2.47816131e-01 6.63294017e-01 6.63590908e-01 -2.90695816e-01
-6.34466931e-02 -2.19594344e-01 2.01704517e-01 2.97855705e-01
-5.23007251e-02 -2.50741035e-01 2.88158268e-01 2.03238994e-01
1.49686015e+00 -4.25166100e-01 2.57861614e-01 -4.45377827e-01
3.63578588e-01 -6.09870963e-02 2.97179848e-01 3.68152857e-01
-6.55990064e-01 8.73150587e-01 5.02771437e-01 -2.04669565e-01
-8.21773827e-01 -1.06681597e+00 -7.24310517e-01 3.95226568e-01
-1.56808376e-01 5.31745749e-03 -1.02962744e+00 -8.01056087e-01
-2.39909306e-01 5.93429089e-01 -6.20605409e-01 -2.74386518e-02
-1.48814335e-01 -1.17576337e+00 7.14193106e-01 3.90246838e-01
6.43701017e-01 -5.19348264e-01 -7.09394872e-01 7.45882764e-02
-2.88421720e-01 -1.03283095e+00 -3.74357492e-01 2.52377898e-01
-8.62229943e-01 -1.16543400e+00 -1.27624762e+00 -6.89279795e-01
7.09628224e-01 3.21322158e-02 6.88090622e-01 -4.62162010e-02
-4.14652199e-01 -7.04233795e-02 -6.49471954e-02 -1.98534653e-01
-6.20010272e-02 8.55949149e-02 -1.99726094e-02 -3.83919537e-01
9.27278176e-02 -4.06566173e-01 -8.64963472e-01 4.14472580e-01
-7.07331538e-01 3.34422290e-01 5.39596200e-01 9.35959578e-01
8.25873196e-01 4.05425914e-02 4.52663600e-01 -4.44516659e-01
6.11864984e-01 -6.56974077e-01 -5.43044925e-01 1.96541518e-01
-4.35851246e-01 -5.22268236e-01 2.40512818e-01 9.24834237e-02
-6.71178102e-01 -1.01202555e-01 -1.38043091e-01 6.04547225e-02
-4.21647221e-01 7.37652123e-01 1.62054911e-01 -2.11743161e-01
3.26648355e-01 7.70020708e-02 2.24337444e-01 -6.10610582e-02
-1.20868385e-01 6.62490308e-01 5.98240316e-01 -1.76837906e-01
1.63807973e-01 5.68290472e-01 1.85927711e-02 -5.08509099e-01
-6.05277121e-01 -4.62837398e-01 -5.89852750e-01 -1.78629592e-01
1.09160960e+00 -7.45618165e-01 -2.42171615e-01 8.38571787e-01
-8.44548643e-01 -4.66489017e-01 1.18642442e-01 7.20619798e-01
-4.81391370e-01 2.01303303e-01 -6.76829994e-01 -2.00883031e-01
-8.72708797e-01 -1.85029471e+00 6.99638844e-01 1.89830005e-01
-3.49801153e-01 -1.04432821e+00 6.49059713e-02 6.37299359e-01
6.66815758e-01 6.61141574e-01 8.03962708e-01 -1.20646429e+00
-1.28548071e-01 -4.07597452e-01 -5.18114686e-01 1.24265112e-01
2.02349067e-01 -1.16091713e-01 -7.76718259e-01 -3.58732373e-01
-2.65240043e-01 -1.35634467e-01 8.84967506e-01 9.28348064e-01
1.42626524e+00 2.35757619e-01 -4.13119793e-01 7.74135172e-01
1.18514657e+00 6.80462182e-01 4.92009848e-01 4.67924386e-01
5.27638853e-01 4.69373465e-01 1.98374048e-01 2.97634512e-01
3.92575592e-01 5.16667068e-01 7.63811529e-01 4.94882949e-02
-1.33407876e-01 4.19671416e-01 -1.27773449e-01 6.86572969e-01
-3.02004099e-01 -1.08450420e-01 -1.60628176e+00 6.38546288e-01
-1.49177611e+00 -5.97191215e-01 -4.11489129e-01 2.30341792e+00
5.18637061e-01 1.26503319e-01 6.50058612e-02 -1.63246363e-01
7.52804101e-01 -2.89118379e-01 -5.82497358e-01 6.80825785e-02
-2.95917988e-01 2.12466925e-01 7.42355466e-01 4.48892325e-01
-1.24212527e+00 5.70720792e-01 4.97517014e+00 8.15757275e-01
-1.18473768e+00 5.36172450e-01 1.14641190e+00 -1.60473779e-01
2.03944221e-01 -3.20892155e-01 -2.41986260e-01 5.37964344e-01
1.08164430e+00 -2.60870576e-01 1.50929376e-01 1.81172103e-01
3.28727663e-01 -3.81597191e-01 -5.58351934e-01 8.93302977e-01
-1.38318285e-01 -1.24476337e+00 -4.70909297e-01 6.12317204e-01
8.07679832e-01 8.72455060e-01 2.77772039e-01 -1.32147953e-01
3.00244182e-01 -1.46182811e+00 4.06100124e-01 5.58942258e-01
9.04494047e-01 -8.29091907e-01 1.36762249e+00 1.75134450e-01
-5.86259246e-01 2.97098130e-01 3.70166893e-03 4.26684737e-01
2.71583884e-03 8.84633958e-01 -9.72902596e-01 7.10886776e-01
6.53007567e-01 5.14929831e-01 -5.78196585e-01 1.58448315e+00
5.76324239e-02 4.63708758e-01 -2.25108013e-01 2.88874745e-01
5.49145043e-01 -2.42868885e-01 4.22023296e-01 1.02830112e+00
5.82244813e-01 2.18802705e-01 -1.78106651e-01 4.18079913e-01
-1.16756812e-01 2.12455601e-01 -2.47341141e-01 1.98127598e-01
2.02657193e-01 1.50766897e+00 -1.08449423e+00 -2.79473096e-01
-1.07287318e-01 5.45343161e-01 1.99873865e-01 2.19523028e-01
-9.30131197e-01 -3.11009556e-01 4.92715091e-01 -2.17256304e-02
-2.13650510e-01 -2.81680245e-02 -5.60432851e-01 -9.42105591e-01
-3.20620775e-01 -8.19415510e-01 4.95915800e-01 -7.59049177e-01
-9.90828454e-01 8.93291593e-01 -1.79798305e-01 -6.77626371e-01
1.85429305e-01 -6.80476248e-01 -8.02290022e-01 1.05003655e+00
-9.87056732e-01 -6.76677942e-01 -6.55640304e-01 3.67919832e-01
2.81536072e-01 -3.71413201e-01 8.23867202e-01 1.04129352e-01
-8.19201708e-01 5.11176705e-01 4.24568444e-01 1.61187083e-01
4.21438903e-01 -1.22531152e+00 3.80693823e-02 5.66147685e-01
-5.10316730e-01 2.83335537e-01 3.44847590e-01 -5.07795870e-01
-7.10027218e-01 -1.20577395e+00 4.59552407e-01 -2.18932778e-01
9.57996428e-01 1.26322314e-01 -9.10497963e-01 6.12405837e-01
2.88022518e-01 6.38938844e-02 8.47059608e-01 -3.15840572e-01
1.35209084e-01 1.58587039e-01 -1.43616474e+00 3.98787767e-01
6.97629631e-01 -9.66594517e-02 -1.48191109e-01 8.42424333e-01
2.79562265e-01 -8.54001582e-01 -1.34103572e+00 5.84914148e-01
5.91053247e-01 -9.29959595e-01 6.99521422e-01 -4.08149451e-01
4.04917657e-01 2.95636412e-02 -1.42940909e-01 -1.71599782e+00
-3.31260979e-01 -5.51325940e-02 3.58920395e-01 5.69114387e-01
5.44527471e-01 -9.69952166e-01 6.46455765e-01 3.84867519e-01
-7.54245281e-01 -1.40670156e+00 -1.15683925e+00 -5.08870423e-01
6.51777804e-01 -2.32956558e-01 6.38711572e-01 8.11451852e-01
-1.38458908e-01 -2.40680948e-01 4.93545234e-01 6.14104792e-02
8.09062302e-01 -2.87096798e-01 7.75633082e-02 -1.11636055e+00
1.97829619e-01 -1.09941566e+00 -6.11688375e-01 -2.53043592e-01
8.61843452e-02 -1.25708890e+00 -2.90806800e-01 -1.99061322e+00
6.39699161e-01 -3.80104750e-01 -6.35407209e-01 3.51284444e-01
3.55356447e-02 1.69594392e-01 -2.48506844e-01 2.99924284e-01
-3.28093946e-01 2.53173232e-01 1.29213035e+00 -2.50173450e-01
3.35952818e-01 8.60241652e-02 -5.49141645e-01 8.60883057e-01
1.42182696e+00 -2.02185705e-01 -1.32998601e-01 -5.74351847e-01
-5.62048495e-01 9.79160815e-02 4.04847234e-01 -9.91216779e-01
1.25535056e-01 6.11719936e-02 5.51546276e-01 -6.14548802e-01
5.03638573e-02 -5.41636050e-01 2.79153198e-01 8.90903413e-01
-1.77805811e-01 4.12183553e-01 1.03103228e-01 3.46241742e-02
-1.28839463e-01 -9.22836736e-02 9.65102077e-01 -1.90174952e-01
-3.96262169e-01 7.61255562e-01 -2.83583283e-01 1.04407847e-01
1.26604927e+00 -1.53168619e-01 -6.12523019e-01 -1.18640192e-01
-1.00115526e+00 2.23893598e-01 3.66880864e-01 4.67475615e-02
6.48564160e-01 -1.10153103e+00 -9.59354579e-01 -1.58023566e-01
6.52229413e-02 3.35555375e-01 4.19111967e-01 1.83269954e+00
-7.22467601e-01 6.03332341e-01 -2.65477866e-01 -6.12723112e-01
-1.25886643e+00 -1.87390730e-01 8.32931757e-01 -5.83166361e-01
-5.02483547e-01 9.84805524e-01 4.33701575e-01 -3.63057047e-01
1.19987369e-01 -3.66048664e-01 -9.26795378e-02 -4.82747331e-02
4.85565990e-01 5.18352628e-01 6.72377706e-01 -8.16958427e-01
-5.89261413e-01 3.60535622e-01 -3.99832189e-01 -8.03746805e-02
1.39475524e+00 2.05173865e-01 -5.05981266e-01 -8.05882066e-02
1.11453569e+00 -3.35291833e-01 -8.01113188e-01 4.51788977e-02
1.13592058e-01 -4.97014746e-02 5.21029890e-01 -1.27270091e+00
-1.38700581e+00 6.96768045e-01 1.02998400e+00 5.77145927e-02
1.11670494e+00 2.95282677e-02 7.24159062e-01 -3.31034720e-01
4.20043945e-01 -5.47581911e-01 -3.77515405e-01 5.94206989e-01
8.96020293e-01 -1.16407454e+00 -2.07222730e-01 -4.69316952e-02
-5.41895926e-01 9.57685590e-01 5.96108437e-01 1.61077365e-01
3.81310850e-01 2.85273015e-01 -2.84206271e-02 -3.75712931e-01
-6.90275073e-01 1.01400919e-01 2.92925954e-01 5.25363147e-01
7.13633597e-01 6.88698232e-01 -4.49732572e-01 6.50535643e-01
-4.81316835e-01 -2.02274293e-01 4.49635178e-01 5.68799257e-01
-5.25413692e-01 -6.01701975e-01 -2.30861634e-01 8.86458158e-01
-7.21424639e-01 -1.59470469e-01 -3.84659469e-01 7.35728085e-01
-2.96976492e-02 5.34921885e-01 6.38916790e-02 -1.74028754e-01
1.34973288e-01 2.45716702e-02 4.33384806e-01 -2.50132948e-01
-4.06050056e-01 2.13478401e-01 -7.72283226e-02 -2.67753154e-01
-1.30426958e-01 -7.28571236e-01 -1.50687277e+00 -1.05310373e-01
-2.41468474e-01 4.20973450e-01 9.11777198e-01 9.19232488e-01
2.06643760e-01 6.84395611e-01 2.12449908e-01 -5.36273122e-01
-2.12849662e-01 -1.02702522e+00 -6.72584832e-01 -1.92909576e-02
1.31809026e-01 -7.34250903e-01 -2.79731452e-01 -5.73864818e-01] | [14.619728088378906, -2.460951328277588] |
30d37faa-5667-47f4-a41f-63ff5cb1632e | semi-parametric-video-grounded-text | 2301.11507 | null | https://arxiv.org/abs/2301.11507v1 | https://arxiv.org/pdf/2301.11507v1.pdf | Semi-Parametric Video-Grounded Text Generation | Efficient video-language modeling should consider the computational cost because of a large, sometimes intractable, number of video frames. Parametric approaches such as the attention mechanism may not be ideal since its computational cost quadratically increases as the video length increases. Rather, previous studies have relied on offline feature extraction or frame sampling to represent the video efficiently, focusing on cross-modal modeling in short video clips. In this paper, we propose a semi-parametric video-grounded text generation model, SeViT, a novel perspective on scalable video-language modeling toward long untrimmed videos. Treating a video as an external data store, SeViT includes a non-parametric frame retriever to select a few query-relevant frames from the data store for a given query and a parametric generator to effectively aggregate the frames with the query via late fusion methods. Experimental results demonstrate our method has a significant advantage in longer videos and causal video understanding. Moreover, our model achieves the new state of the art on four video-language datasets, iVQA (+4.8), Next-QA (+6.9), and Activitynet-QA (+4.8) in accuracy, and MSRVTT-Caption (+3.6) in CIDEr. | ['Minjoon Seo', 'Jiyoung Lee', 'Jin-Hwa Kim', 'Sungdong Kim'] | 2023-01-27 | null | null | null | null | ['video-question-answering', 'video-understanding'] | ['computer-vision', 'computer-vision'] | [-4.32770364e-02 -2.45545387e-01 -4.58847106e-01 -2.67461181e-01
-1.34094071e+00 -3.67188901e-01 5.21305621e-01 -3.84146482e-01
-3.34917605e-01 5.13750732e-01 4.80480731e-01 -7.09031075e-02
3.02446812e-01 -3.26939732e-01 -1.14467752e+00 -5.37712872e-01
-2.26474658e-01 5.34237325e-01 2.36152232e-01 1.74219549e-01
-3.15487199e-02 6.30209595e-02 -1.49133456e+00 7.74887860e-01
5.84020197e-01 1.09681678e+00 5.61322033e-01 1.03417706e+00
-1.67120531e-01 1.42133296e+00 -4.12445545e-01 -4.54142988e-01
1.21395022e-01 -4.41458493e-01 -5.64551473e-01 2.14843944e-01
5.88360786e-01 -9.13774610e-01 -7.63334930e-01 7.26419628e-01
4.33304042e-01 1.94499001e-01 4.41818297e-01 -1.59179950e+00
-5.99439740e-01 8.22121799e-01 -5.32487750e-01 1.79024890e-01
7.02221870e-01 2.72044122e-01 8.52794170e-01 -1.09800375e+00
8.65982950e-01 1.70255923e+00 3.29347044e-01 4.84270871e-01
-8.53904486e-01 -7.63938904e-01 3.33427995e-01 6.06471002e-01
-1.66090059e+00 -9.04006720e-01 6.72760904e-01 -5.43068886e-01
8.93376946e-01 2.46043608e-01 7.08524466e-01 1.36449564e+00
-7.43410289e-02 1.21679604e+00 4.24624622e-01 -2.50870556e-01
8.44436362e-02 -2.94200540e-01 -2.95752019e-01 6.22297883e-01
-2.44294763e-01 -5.35615124e-02 -1.08309746e+00 -1.74604848e-01
7.04133034e-01 -1.48001358e-01 -3.39800209e-01 -1.90863594e-01
-1.43794394e+00 8.32845747e-01 -1.44875139e-01 -1.43609881e-01
-6.84544683e-01 6.54445112e-01 5.11803091e-01 6.03442974e-02
4.46181327e-01 2.62449216e-02 -3.87687802e-01 -6.58897400e-01
-1.44660747e+00 3.76784801e-01 6.37500226e-01 1.39659512e+00
5.92800915e-01 2.71436423e-01 -4.85432535e-01 6.13282561e-01
4.10770565e-01 6.26636267e-01 3.27073306e-01 -1.53596020e+00
6.94796383e-01 -1.74359337e-03 2.41656244e-01 -9.15098786e-01
1.47768348e-01 1.66703254e-01 -4.49868262e-01 -4.79076266e-01
1.11043885e-01 -9.09140185e-02 -8.07541370e-01 1.73342240e+00
6.85965493e-02 6.14014864e-01 1.99995376e-02 9.23828125e-01
7.48982847e-01 1.27152574e+00 3.33928555e-01 -6.12100482e-01
1.37003398e+00 -1.21842062e+00 -7.93785989e-01 -1.71493441e-01
4.96806592e-01 -7.38225400e-01 1.03992391e+00 4.72685933e-01
-1.45052087e+00 -6.12742364e-01 -7.51864672e-01 -3.10312539e-01
1.57482907e-01 1.66562334e-01 3.29463661e-01 3.04914355e-01
-1.16790211e+00 -9.07581151e-02 -9.83985066e-01 -2.24291518e-01
3.98093492e-01 1.05207883e-01 -2.42130101e-01 -4.65259701e-01
-1.14411449e+00 3.32081199e-01 6.94138050e-01 -1.33608773e-01
-1.36012244e+00 -8.14912677e-01 -9.94432211e-01 1.02595314e-01
8.51227164e-01 -8.75541449e-01 1.35265028e+00 -1.16791534e+00
-1.36600220e+00 4.00020987e-01 -5.76851726e-01 -7.97104478e-01
5.27604938e-01 -5.48517585e-01 -3.45492035e-01 6.94522560e-01
1.16751336e-01 1.25220382e+00 1.14393568e+00 -1.22935879e+00
-6.76518500e-01 -4.11959784e-03 2.69763649e-01 4.45127010e-01
-2.20729664e-01 3.61886740e-01 -1.26127326e+00 -9.29705858e-01
-3.37778002e-01 -8.58002067e-01 5.31628840e-02 5.94227873e-02
-2.29463160e-01 -1.63610667e-01 9.97474253e-01 -9.84639585e-01
1.50808060e+00 -2.14750171e+00 3.17392439e-01 -2.79902428e-01
1.68892458e-01 8.81573707e-02 -2.05446959e-01 5.14134765e-01
1.74186707e-01 1.41349778e-01 2.71022648e-01 -4.42069530e-01
-7.79646486e-02 1.26308694e-01 -5.21740973e-01 3.07310551e-01
1.57751381e-01 1.04856420e+00 -9.92327631e-01 -9.90347028e-01
1.96624994e-01 5.87703645e-01 -8.41686547e-01 2.82459557e-01
-5.83957732e-01 1.26856565e-01 -5.26298404e-01 7.56554723e-01
2.91645020e-01 -4.02023196e-01 3.88903394e-02 -6.76994443e-01
1.06918484e-01 -2.19103247e-01 -1.04746401e+00 2.10220218e+00
-2.95942992e-01 8.67304921e-01 -7.72176450e-03 -7.00840950e-01
3.24587077e-01 5.38536787e-01 8.04023325e-01 -4.94917244e-01
-7.31125697e-02 -2.08032548e-01 -4.50738490e-01 -8.08554232e-01
7.69615352e-01 2.76455969e-01 1.18126096e-02 2.41557017e-01
2.28434965e-01 2.17676595e-01 6.19480729e-01 6.21799946e-01
8.57753932e-01 4.35791224e-01 6.35649860e-02 1.74958825e-01
3.92572284e-01 -6.98098093e-02 4.54724580e-01 7.25650251e-01
-9.99703929e-02 6.54404998e-01 4.66152668e-01 -2.50016153e-01
-1.02512729e+00 -9.04528797e-01 4.07074064e-01 1.35618126e+00
1.63990453e-01 -1.04822993e+00 -8.50970268e-01 -3.36505890e-01
-3.39469790e-01 7.87506580e-01 -1.74894944e-01 -9.50954556e-02
-7.21327186e-01 -2.47510046e-01 6.32803321e-01 5.63171804e-01
2.92276829e-01 -8.50276649e-01 -5.54142058e-01 3.22303981e-01
-9.65021908e-01 -1.65257204e+00 -9.39799368e-01 -6.09635413e-01
-5.48091352e-01 -9.48070705e-01 -5.85321128e-01 -4.77841675e-01
3.24166059e-01 4.85158741e-01 1.26351583e+00 -1.23450302e-01
-3.68960313e-02 7.48476326e-01 -5.48279285e-01 -6.35235757e-02
-5.04007161e-01 -3.02468985e-01 -1.64456596e-03 2.02185646e-01
2.54388422e-01 -1.03962898e-01 -5.52058160e-01 1.55919626e-01
-9.70573843e-01 4.98446941e-01 4.01949555e-01 6.58432245e-01
7.73076594e-01 -1.59481749e-01 3.99546593e-01 -3.95893842e-01
2.35116124e-01 -6.32469833e-01 -4.36233997e-01 3.46614897e-01
-7.69646317e-02 -1.33898601e-01 5.02735496e-01 -6.64082229e-01
-1.10294950e+00 9.09632966e-02 2.44972691e-01 -1.26484001e+00
1.34225339e-01 6.64461255e-01 -2.10048884e-01 4.29640949e-01
3.62475783e-01 5.39168179e-01 -3.64358090e-02 -1.93748102e-01
6.00047708e-01 5.36367595e-01 7.59534299e-01 -6.93090081e-01
6.12319291e-01 4.85074013e-01 -3.72377098e-01 -9.82388258e-01
-5.56899786e-01 -3.59428763e-01 -2.07329363e-01 -4.72329825e-01
1.05831110e+00 -1.72240591e+00 -6.02574229e-01 2.88450688e-01
-1.27742732e+00 -2.59435743e-01 -1.02903567e-01 5.80546558e-01
-9.98559833e-01 5.67770362e-01 -7.03818977e-01 -7.90611446e-01
-3.69139016e-01 -1.43179750e+00 1.43109405e+00 -1.93453684e-01
-1.07895732e-01 -5.43989658e-01 -3.51373017e-01 6.23977780e-01
5.75154908e-02 3.04890685e-02 6.04539037e-01 -3.83938402e-01
-1.12543404e+00 5.56016224e-04 -2.21182972e-01 2.18204007e-01
-2.30828151e-01 2.51080245e-01 -7.29688883e-01 -4.15869892e-01
-2.07113609e-01 -4.93334442e-01 7.10563481e-01 5.25468469e-01
1.51018488e+00 -5.56237102e-01 -2.53591865e-01 6.59756243e-01
1.16934347e+00 4.30187255e-01 5.72240949e-01 -6.44348711e-02
8.96561742e-01 2.68746078e-01 9.48958993e-01 7.13152468e-01
7.24645317e-01 8.36116493e-01 3.32290441e-01 3.72517616e-01
-3.78569484e-01 -5.83040297e-01 8.90204430e-01 9.79943693e-01
6.29790267e-03 -7.99548507e-01 -7.57937729e-01 5.19466221e-01
-2.05851674e+00 -1.31063628e+00 1.43080786e-01 1.92352259e+00
6.59969509e-01 -1.96316987e-01 1.54527947e-01 -3.07073623e-01
6.76068008e-01 2.75310814e-01 -5.16728461e-01 2.69121855e-01
-2.55400717e-01 -2.67213255e-01 2.25552678e-01 3.61142635e-01
-1.07500577e+00 1.12409961e+00 5.99297667e+00 1.11379194e+00
-9.07382309e-01 2.07062095e-01 7.27339923e-01 -5.95672250e-01
-1.00868843e-01 -1.14193909e-01 -7.84544945e-01 8.08859110e-01
1.42610562e+00 -3.69721085e-01 6.13545537e-01 6.20720983e-01
6.78895473e-01 -1.45796150e-01 -1.29201591e+00 1.57232201e+00
4.59780306e-01 -1.65939260e+00 6.40108645e-01 -9.59576294e-02
5.42049587e-01 -5.54498248e-02 -6.29431680e-02 3.83547336e-01
-7.46087283e-02 -8.02674651e-01 1.23982656e+00 7.40153849e-01
9.93871391e-01 -6.35311425e-01 2.73032486e-01 3.54890972e-01
-1.38513720e+00 -1.31812871e-01 -1.82820395e-01 3.87176156e-01
7.93314517e-01 2.05492288e-01 -5.86548030e-01 4.98077571e-01
8.10366929e-01 8.52876425e-01 -4.75748330e-01 7.28678703e-01
3.36678445e-01 7.66000688e-01 -3.35942537e-01 3.94864887e-01
2.64164805e-01 1.67356543e-02 5.47920585e-01 1.25273323e+00
7.10719109e-01 2.17356116e-01 5.37487566e-01 4.71237302e-01
-1.95156276e-01 1.42337039e-01 -4.67205197e-01 -3.51688594e-01
5.12250185e-01 8.65978360e-01 -4.43194360e-01 -7.31451929e-01
-6.98844612e-01 9.41555321e-01 7.03415135e-03 6.11798704e-01
-1.35050297e+00 2.49774367e-01 5.93124747e-01 1.34666950e-01
4.37150449e-01 -2.68090755e-01 5.62856793e-01 -1.47093177e+00
8.82598013e-02 -1.16229975e+00 4.77537334e-01 -1.35317099e+00
-8.08706105e-01 4.65542942e-01 4.52930689e-01 -1.33083773e+00
-8.40455532e-01 -2.22976685e-01 5.04701324e-02 3.71084780e-01
-1.15859938e+00 -1.39758742e+00 -2.59002447e-01 9.23784435e-01
1.24123859e+00 -1.96729138e-01 4.68487889e-01 6.87862992e-01
-3.56325179e-01 4.20174122e-01 -1.77328900e-01 1.41029702e-02
8.22697401e-01 -6.61764622e-01 1.32561430e-01 9.89481509e-01
3.85296851e-01 2.93810308e-01 7.06094861e-01 -6.59773886e-01
-2.02657247e+00 -1.25975084e+00 6.23359501e-01 -3.59876484e-01
6.70827806e-01 -3.72417629e-01 -6.86116993e-01 1.00438488e+00
3.04705173e-01 -4.06155661e-02 5.02933204e-01 -5.81624448e-01
-2.57712424e-01 -8.11032057e-02 -6.45746410e-01 7.06424654e-01
1.12053239e+00 -6.63227499e-01 -1.90396965e-01 5.03101885e-01
1.19276381e+00 -6.41189456e-01 -9.10634220e-01 1.39099672e-01
5.66583455e-01 -7.40997732e-01 1.16737807e+00 -5.53799808e-01
4.88361835e-01 -3.62407923e-01 -4.71736044e-01 -6.75999284e-01
-9.91231799e-02 -9.17076230e-01 -9.03169811e-01 1.21648502e+00
-1.71996038e-02 2.00820059e-01 7.26429641e-01 5.37304461e-01
-8.22481140e-02 -5.57133734e-01 -8.76133740e-01 -6.26651406e-01
-5.31621456e-01 -9.89494264e-01 4.37227100e-01 6.31009281e-01
-3.77723157e-01 2.19121575e-01 -8.88853848e-01 2.21401483e-01
4.58518684e-01 -1.10589378e-02 9.11385775e-01 -6.25225663e-01
-3.94987077e-01 -1.28827319e-01 -3.24476480e-01 -1.56718516e+00
4.24519747e-01 -4.83417213e-01 2.55222376e-02 -1.21693170e+00
3.85761738e-01 1.49542138e-01 -1.64210610e-02 2.22403839e-01
-6.72415048e-02 5.08201979e-02 5.90315342e-01 3.69992822e-01
-1.17518425e+00 6.95100307e-01 1.04296148e+00 -2.87266314e-01
-1.65730014e-01 -3.45563442e-01 -4.64715570e-01 5.95875740e-01
3.15935344e-01 -1.98379666e-01 -8.26008797e-01 -6.93504453e-01
1.77903369e-01 6.93660617e-01 4.46294874e-01 -9.19441462e-01
3.55689734e-01 -1.47877008e-01 2.02990368e-01 -8.77586484e-01
9.00571644e-01 -7.81329334e-01 6.04647815e-01 1.13805272e-01
-3.11565697e-01 3.24969649e-01 7.42197707e-02 7.76229918e-01
-4.47058439e-01 5.20508327e-02 3.47009003e-01 -3.54522988e-02
-1.13524044e+00 7.27026165e-01 -4.24632460e-01 1.92052901e-01
1.08224630e+00 -3.00470710e-01 -2.69710720e-01 -9.37990725e-01
-7.58178890e-01 4.40220803e-01 3.52781057e-01 8.05013597e-01
8.72938633e-01 -1.40365708e+00 -7.98302174e-01 -1.18914209e-01
1.67852685e-01 -5.73040508e-02 4.93732363e-01 7.56747425e-01
-5.26632726e-01 4.46166277e-01 2.18030840e-01 -9.10309434e-01
-1.25331211e+00 7.61817634e-01 -5.88200279e-02 -1.22484609e-01
-4.00212437e-01 7.05253363e-01 5.18447995e-01 4.13863838e-01
2.29017422e-01 -3.24877918e-01 7.89855197e-02 1.98227674e-01
7.64314413e-01 4.25910354e-01 -4.37866330e-01 -1.13196492e+00
-5.79118542e-02 3.11585695e-01 -1.08408533e-01 -2.19446212e-01
1.05886841e+00 -4.62450564e-01 3.37809622e-02 4.46272045e-01
1.23861396e+00 -3.46049845e-01 -1.62131131e+00 -2.35816926e-01
-2.83680022e-01 -5.18248439e-01 2.77853180e-02 -4.44465518e-01
-1.08574271e+00 4.75600839e-01 2.95376390e-01 -2.12543100e-01
1.27678204e+00 1.33324265e-01 9.10486519e-01 3.42079520e-01
4.75516170e-01 -8.43255043e-01 2.09213212e-01 3.42862695e-01
9.83865798e-01 -1.07904971e+00 8.59549269e-03 -3.66020471e-01
-7.84629107e-01 1.05495143e+00 6.31005406e-01 3.10258836e-01
4.82136309e-01 1.14976183e-01 -2.16379374e-01 3.92822400e-02
-1.23920095e+00 2.02118680e-01 2.43446678e-01 4.27027255e-01
2.15118319e-01 -5.35409488e-02 7.82224387e-02 5.60868382e-01
4.31649201e-02 3.06250483e-01 5.72721004e-01 7.63329029e-01
-2.49223799e-01 -8.13004494e-01 -3.75408113e-01 3.53125930e-01
-5.82692027e-01 -2.65145361e-01 2.96694130e-01 7.86264956e-01
-5.72309718e-02 1.02157092e+00 2.21515089e-01 -2.05808774e-01
9.58353933e-03 2.05881879e-01 4.68311489e-01 -2.65799046e-01
-9.40827578e-02 6.13487661e-01 1.75523907e-01 -1.09520710e+00
-8.33215475e-01 -8.01096380e-01 -1.17869604e+00 -3.10499728e-01
-6.14484400e-02 2.06756517e-02 4.33568180e-01 7.60416090e-01
6.60671949e-01 2.95947313e-01 2.43209839e-01 -1.03717256e+00
-2.10382327e-01 -7.67899573e-01 -2.32023612e-01 4.24231321e-01
3.04681867e-01 -4.36737418e-01 -1.84280965e-02 8.28257561e-01] | [10.178637504577637, 0.7176439762115479] |
bc02c5d5-85d4-4ebf-adf5-548dca5fea7a | heart-darts-classification-of-heartbeats | 2105.00693 | null | https://arxiv.org/abs/2105.00693v1 | https://arxiv.org/pdf/2105.00693v1.pdf | Heart-Darts: Classification of Heartbeats Using Differentiable Architecture Search | Arrhythmia is a cardiovascular disease that manifests irregular heartbeats. In arrhythmia detection, the electrocardiogram (ECG) signal is an important diagnostic technique. However, manually evaluating ECG signals is a complicated and time-consuming task. With the application of convolutional neural networks (CNNs), the evaluation process has been accelerated and the performance is improved. It is noteworthy that the performance of CNNs heavily depends on their architecture design, which is a complex process grounded on expert experience and trial-and-error. In this paper, we propose a novel approach, Heart-Darts, to efficiently classify the ECG signals by automatically designing the CNN model with the differentiable architecture search (i.e., Darts, a cell-based neural architecture search method). Specifically, we initially search a cell architecture by Darts and then customize a novel CNN model for ECG classification based on the obtained cells. To investigate the efficiency of the proposed method, we evaluate the constructed model on the MIT-BIH arrhythmia database. Additionally, the extensibility of the proposed CNN model is validated on two other new databases. Extensive experimental results demonstrate that the proposed method outperforms several state-of-the-art CNN models in ECG classification in terms of both performance and generalization capability. | ['Jiancheng Lv', 'Juan Zhao', 'Yanan sun', 'Qing Ye', 'Jindi Lv'] | 2021-05-03 | null | null | null | null | ['arrhythmia-detection', 'ecg-classification'] | ['medical', 'medical'] | [ 1.09626845e-01 -4.62754995e-01 3.02665353e-01 -1.83379441e-01
-2.86336690e-01 -4.49747294e-01 -2.84304440e-01 1.93664789e-01
-3.68207723e-01 7.70110369e-01 -4.26781654e-01 -3.68809104e-01
-2.48027220e-01 -6.59780741e-01 -1.29768908e-01 -7.01526046e-01
-2.76853323e-01 3.58207703e-01 -2.42547944e-01 -4.16267514e-02
3.18569213e-01 6.38878882e-01 -1.23643291e+00 1.27300367e-01
9.27018583e-01 1.38569582e+00 -2.34536633e-01 7.31272578e-01
2.02043012e-01 4.66898710e-01 -8.57479811e-01 -8.30742568e-02
1.26207292e-01 -8.05126607e-01 -6.84483349e-01 -1.91611066e-01
-1.94596574e-01 1.29723221e-01 1.89062372e-01 8.37088764e-01
1.06869435e+00 -3.47646743e-01 4.37581867e-01 -8.02846491e-01
-1.34365767e-01 4.25662011e-01 -1.54694960e-01 5.77805519e-01
-1.53207004e-01 8.27216357e-02 7.89274335e-01 -8.63875568e-01
3.68661910e-01 6.58846974e-01 1.14890945e+00 3.30213785e-01
-9.08683300e-01 -7.40271389e-01 -2.38393441e-01 1.24965332e-01
-1.85326982e+00 -2.28861466e-01 1.15039277e+00 -4.86838996e-01
6.65969014e-01 3.15235257e-01 1.15179110e+00 6.00483954e-01
3.62427294e-01 5.97256303e-01 8.05968285e-01 -3.73262197e-01
3.39560270e-01 -3.24931554e-02 1.45943925e-01 7.62847424e-01
3.64823818e-01 -3.13808173e-02 -1.71744317e-01 -2.45073870e-01
8.73677731e-01 -4.16901372e-02 -3.91390145e-01 -1.66664540e-03
-1.21525180e+00 5.07908046e-01 4.60344940e-01 6.23203099e-01
-6.50864005e-01 1.77955508e-01 6.94425702e-01 2.74008572e-01
1.66633114e-01 8.07980001e-01 -4.57262427e-01 -1.30921617e-01
-1.01621819e+00 1.79768026e-01 7.11533189e-01 5.18889368e-01
3.47437114e-01 3.67601484e-01 -3.03576082e-01 7.43540823e-01
5.32025471e-02 3.72931473e-02 6.53889239e-01 -3.21237147e-01
8.27984214e-02 9.19967234e-01 -1.78085551e-01 -1.16057527e+00
-7.44659305e-01 -1.36331344e+00 -1.59012902e+00 -1.48343712e-01
2.56652415e-01 -3.62346381e-01 -7.26927400e-01 1.37636924e+00
1.79573476e-01 3.34395230e-01 7.75851011e-02 1.01060402e+00
1.03869152e+00 2.96064645e-01 1.25779092e-01 -4.26351339e-01
1.40170336e+00 -5.84129035e-01 -7.89100707e-01 3.07893276e-01
6.12609148e-01 -4.13636804e-01 7.23525941e-01 7.01011479e-01
-8.71783018e-01 -8.76329362e-01 -1.32952654e+00 3.27422053e-01
-5.54868840e-02 7.94917822e-01 5.20561337e-01 6.66551530e-01
-8.42156470e-01 6.69446766e-01 -6.37482226e-01 -3.65117401e-01
6.31659567e-01 5.48801720e-01 2.82914229e-02 5.24088323e-01
-1.40853596e+00 5.16981125e-01 5.67091525e-01 6.95736945e-01
-7.08210528e-01 -3.99100542e-01 -3.76423270e-01 2.28704587e-01
-1.33875936e-01 -8.47516954e-01 8.52381051e-01 -1.05982769e+00
-1.48403549e+00 6.71194792e-01 2.27973700e-01 -6.22436285e-01
4.35032994e-01 2.52084821e-01 -4.93821323e-01 1.76160574e-01
-2.67475784e-01 2.40403935e-01 6.23089433e-01 -9.57897782e-01
-5.59701145e-01 -3.23524624e-01 -2.63959914e-01 5.48169687e-02
-5.17182469e-01 -2.06034437e-01 -2.39992723e-01 -8.92029762e-01
2.85588175e-01 -9.06084001e-01 -2.48775810e-01 -1.86693832e-01
-3.91277283e-01 -1.53033704e-01 3.88433069e-01 -5.31902015e-01
1.76174951e+00 -2.20582247e+00 4.09404188e-02 6.88018084e-01
4.25652981e-01 5.26453853e-01 2.43040666e-01 4.16387655e-02
-1.54707655e-01 3.04277360e-01 -3.50756258e-01 1.36108190e-01
-4.37805504e-01 -3.25595252e-02 2.67433941e-01 3.48388910e-01
2.78162479e-01 9.11750972e-01 -5.49549758e-01 -6.98958158e-01
-4.55030203e-02 3.96364510e-01 -4.03623223e-01 2.03501090e-01
2.08287328e-01 7.31502056e-01 -6.00001335e-01 8.85730147e-01
4.47961390e-01 -4.78631556e-01 4.02120709e-01 -5.91643095e-01
-9.04586539e-02 -1.63101703e-01 -1.08921206e+00 1.62386870e+00
-1.78939641e-01 3.43162268e-01 -4.45405751e-01 -1.28026628e+00
1.36958790e+00 7.29368806e-01 5.64730525e-01 -3.99772584e-01
5.82992911e-01 5.56877673e-01 4.96845961e-01 -6.48556232e-01
-8.30571279e-02 -2.20204722e-02 1.42468616e-01 2.90860802e-01
-6.11938424e-02 2.45698854e-01 5.77574745e-02 -5.17754376e-01
9.72737491e-01 -1.64370269e-01 5.00691533e-01 -5.79685390e-01
8.39281201e-01 -2.71936301e-02 7.87034810e-01 6.53712213e-01
-1.91805378e-01 6.53767645e-01 4.12600279e-01 -1.18706954e+00
-8.47428024e-01 -5.09633124e-01 -3.48305285e-01 1.85214341e-01
-7.95518011e-02 -2.19789639e-01 -9.20253396e-01 -5.38102686e-01
-3.65980178e-01 4.11241353e-02 -7.14957654e-01 -1.83192983e-01
-8.35492969e-01 -1.20617342e+00 1.11444187e+00 6.11111760e-01
8.57022285e-01 -1.40689433e+00 -1.18989646e+00 5.50329208e-01
-2.44668484e-01 -8.36942315e-01 -1.00018129e-01 1.81091279e-01
-1.11430180e+00 -1.01428151e+00 -8.05543423e-01 -1.09183705e+00
5.71742058e-01 -5.12559295e-01 1.13363075e+00 8.52825999e-01
-6.02052748e-01 -2.46172324e-01 -3.53104889e-01 -6.10996544e-01
-2.39925504e-01 5.59082627e-01 -1.61942258e-01 4.93256003e-01
1.24689311e-01 -6.73548520e-01 -1.04900587e+00 7.67640769e-02
-5.92568517e-01 1.88471302e-02 7.15142548e-01 1.12122381e+00
7.98181832e-01 3.15542370e-01 1.05079198e+00 -8.33742499e-01
9.65162575e-01 -1.41323149e-01 -3.19769353e-01 2.90547729e-01
-7.91821897e-01 -2.40423545e-01 6.61640584e-01 -3.32756758e-01
-4.84458655e-01 3.64944011e-01 -2.40407929e-01 -5.54876208e-01
-1.22190237e-01 9.35196280e-01 6.89676479e-02 -2.28327066e-01
7.64168561e-01 3.55828702e-01 7.74175972e-02 -2.87850648e-01
-3.74276459e-01 6.01150036e-01 4.71617818e-01 -4.31273401e-01
3.58503133e-01 9.67511386e-02 1.70583829e-01 -7.68315434e-01
-4.03022170e-01 -7.75671601e-02 -6.09588802e-01 -3.79563689e-01
9.65412498e-01 -6.79694593e-01 -9.17166591e-01 5.81526220e-01
-1.21683800e+00 6.47499636e-02 -2.07458213e-02 4.36200649e-01
-1.63802192e-01 2.16952547e-01 -5.80913842e-01 -7.37959921e-01
-1.02103734e+00 -1.20025277e+00 6.68058634e-01 1.47792518e-01
-4.34572548e-01 -8.42867196e-01 -7.09487051e-02 -1.39456496e-01
5.06829083e-01 7.16772020e-01 1.24371946e+00 -7.59812295e-01
-4.03308004e-01 -4.48713988e-01 -7.45658278e-02 3.09001356e-01
1.33352637e-01 -2.00307779e-02 -9.67008412e-01 -1.59050182e-01
1.16400905e-01 4.42741252e-02 4.90310818e-01 4.59129870e-01
1.56874192e+00 -1.26125365e-01 -2.05378279e-01 8.51979673e-01
1.46839488e+00 7.20597327e-01 8.56661439e-01 2.88938433e-01
6.36073291e-01 1.26616031e-01 2.33825892e-01 5.28571963e-01
7.98943266e-02 3.79532903e-01 2.81068891e-01 -6.64041579e-01
1.53893277e-01 2.22359806e-01 -4.70131934e-01 1.11572504e+00
-6.46622837e-01 -9.36511010e-02 -1.09640062e+00 4.57642823e-01
-1.79097605e+00 -6.32814407e-01 -1.52386725e-01 1.88722944e+00
8.72817397e-01 1.30203739e-01 -1.03393141e-02 8.47750783e-01
6.29298210e-01 -4.71307546e-01 -4.41532940e-01 -3.49644631e-01
3.99433076e-03 7.15888798e-01 -1.52640287e-02 -5.56450523e-02
-1.12564015e+00 4.51969385e-01 5.70363045e+00 6.05039001e-01
-1.61311507e+00 -1.22935474e-01 9.77706313e-01 4.18059289e-01
2.02724308e-01 -4.34379935e-01 -3.62798303e-01 4.04017776e-01
5.67195594e-01 1.03976503e-01 2.23689273e-01 5.83491087e-01
1.32059172e-01 4.10440654e-01 -1.07286394e+00 1.33836472e+00
-1.00194477e-01 -1.53173804e+00 -2.52516028e-02 -2.41143763e-01
3.72575402e-01 -5.36554396e-01 -1.24398053e-01 1.16831519e-01
-7.36851394e-01 -1.12036133e+00 6.43938005e-01 5.91550529e-01
8.94239366e-01 -7.99913287e-01 1.32125521e+00 1.31451041e-01
-1.40437257e+00 -2.74522543e-01 1.13184182e-02 5.72759807e-02
-3.00343722e-01 6.55648887e-01 -8.04211617e-01 5.70115685e-01
8.26476455e-01 7.18198180e-01 -7.19453931e-01 1.42567861e+00
1.12862892e-01 8.31988692e-01 4.97552007e-02 -2.71570355e-01
-2.59588636e-03 -1.93795525e-02 3.55542660e-01 1.27322483e+00
6.03964865e-01 2.05530241e-01 1.23452872e-01 9.65986669e-01
3.02520283e-02 4.70722526e-01 -3.06129217e-01 3.38089690e-02
4.13670510e-01 1.38298845e+00 -1.10147929e+00 -3.83515954e-01
2.23085746e-01 5.04707813e-01 -7.75437504e-02 2.90794224e-01
-8.68257523e-01 -7.25199997e-01 1.08658941e-02 1.29286507e-02
1.29706278e-01 2.52779573e-01 -7.55934119e-01 -8.59198272e-01
6.37502894e-02 -1.16456568e+00 3.41238320e-01 -4.53862369e-01
-1.06492126e+00 1.26288903e+00 -4.32648808e-01 -1.65957999e+00
-5.65653518e-02 -2.31295794e-01 -6.34671688e-01 9.11359906e-01
-1.20423353e+00 -7.85552800e-01 -7.12575078e-01 5.65344632e-01
3.16111743e-01 -3.47586066e-01 1.11067688e+00 7.01810479e-01
-6.74084723e-01 7.89544106e-01 -5.57437301e-01 4.02089655e-01
2.25546107e-01 -9.78885233e-01 1.39358550e-01 5.48317492e-01
-9.88921821e-02 6.28091395e-01 3.11523318e-01 -3.48011613e-01
-1.05239058e+00 -1.11493397e+00 8.82181346e-01 1.20103098e-01
7.25993663e-02 -1.63320839e-01 -9.10736144e-01 1.27048329e-01
7.39146098e-02 1.73869282e-01 7.18251407e-01 2.58171316e-02
2.50457555e-01 -5.46236753e-01 -1.00633872e+00 3.89935970e-01
7.27786839e-01 -1.09343782e-01 -2.46253908e-01 2.55747400e-02
1.48962334e-01 -4.67897832e-01 -9.66719329e-01 9.34345901e-01
8.57967913e-01 -7.36268997e-01 7.59257674e-01 -3.86291087e-01
1.11223996e-01 -5.50591767e-01 2.96295404e-01 -1.19455945e+00
-4.55955237e-01 -5.52626491e-01 -1.62925925e-02 7.63772547e-01
5.18697441e-01 -7.01366484e-01 7.07263768e-01 -2.60030944e-02
-2.74332136e-01 -1.38161707e+00 -8.28560531e-01 -3.81475002e-01
-3.33050400e-01 -8.73194262e-02 8.78712714e-01 1.02164745e+00
-2.66722679e-01 4.17057604e-01 -1.85504869e-01 7.65222609e-02
3.37756962e-01 1.36003762e-01 4.49598163e-01 -1.61441374e+00
-1.77231267e-01 -4.33899343e-01 -6.06680095e-01 -4.85844254e-01
-3.97114187e-01 -8.26772451e-01 -2.52141454e-03 -1.23176277e+00
-1.03805661e-01 -7.12207198e-01 -8.73839319e-01 4.28673744e-01
-1.57502145e-01 4.98493940e-01 -3.59056443e-02 3.20021480e-01
-2.94503212e-01 2.34848410e-01 1.32868254e+00 -1.24426953e-01
-4.72408921e-01 1.84779257e-01 -6.12409592e-01 6.17723048e-01
1.11032665e+00 -4.44279164e-01 -4.68163759e-01 -1.64548054e-01
3.32726002e-01 3.47011656e-01 3.25346708e-01 -1.45393014e+00
1.69530019e-01 4.84333307e-01 8.00880611e-01 -5.92794120e-01
-5.42107895e-02 -6.50946200e-01 4.35987115e-01 8.86526406e-01
-3.89474154e-01 5.02298117e-01 1.52448878e-01 2.30003893e-01
-5.24177134e-01 -1.44213170e-01 6.97198808e-01 -2.19024599e-01
-1.83640704e-01 4.63044971e-01 -4.26220149e-01 -6.39316067e-02
7.94559896e-01 -4.55430388e-01 2.52215654e-01 6.49013743e-02
-8.84308517e-01 -2.65881550e-02 -2.27621660e-01 1.60747059e-02
8.63071620e-01 -1.46199667e+00 -7.26258159e-01 4.38540131e-01
1.08657248e-01 1.13422826e-01 2.51271665e-01 1.13804173e+00
-1.04601407e+00 2.47080579e-01 -3.88970613e-01 -9.38226342e-01
-1.18584538e+00 1.64548591e-01 8.04908335e-01 -3.69935662e-01
-6.23108625e-01 6.36945307e-01 -1.07788388e-03 -1.02204539e-01
3.28510314e-01 -6.44174278e-01 -6.61651850e-01 -3.58873568e-02
1.74097881e-01 1.65708557e-01 4.59930003e-01 -1.34700835e-01
-4.72150743e-01 7.57230520e-01 2.74827778e-01 2.92048365e-01
1.09617388e+00 3.13005209e-01 -3.45696986e-01 2.12018609e-01
8.19623828e-01 -5.27070284e-01 -5.29680610e-01 2.05417186e-01
-7.03052729e-02 -1.90613065e-02 -5.55298179e-02 -9.12944078e-01
-1.49249697e+00 1.02101552e+00 1.07434857e+00 2.34657571e-01
1.53156006e+00 -7.68853724e-01 8.20101619e-01 4.22655255e-01
3.45311195e-01 -9.90248263e-01 2.63847187e-02 2.28898510e-01
8.55899453e-01 -7.86252558e-01 -2.16610014e-01 -2.32335776e-01
-4.47755158e-01 1.46549928e+00 6.63270593e-01 -1.92386478e-01
1.03167284e+00 5.69193177e-02 2.50531495e-01 -4.42117810e-01
-4.81872022e-01 2.10834831e-01 3.48741353e-01 2.66030312e-01
5.71951389e-01 5.04994839e-02 -6.48603022e-01 1.03377223e+00
-1.39062241e-01 5.99815786e-01 2.08954215e-01 7.66361892e-01
-1.24632210e-01 -7.78356552e-01 -1.32713363e-01 4.69341636e-01
-6.69464350e-01 -2.21838802e-01 -1.81332558e-01 6.15231276e-01
5.71410596e-01 7.01839566e-01 -1.46738663e-01 -5.23732841e-01
3.13891709e-01 9.99403074e-02 3.08999687e-01 -4.10815924e-01
-1.18885827e+00 2.27871999e-01 -9.70123932e-02 -4.77886796e-02
-4.49878424e-01 -2.45009720e-01 -1.23838782e+00 2.87876815e-01
-3.92170399e-01 3.09056193e-01 5.56894720e-01 8.04220140e-01
5.14864981e-01 1.01639056e+00 6.13109171e-01 -3.47073823e-01
-3.75289381e-01 -9.92673397e-01 -4.42270488e-01 2.65099406e-01
2.24614754e-01 -4.74078387e-01 -1.88661311e-02 1.74455121e-01] | [14.273576736450195, 3.2698471546173096] |
8bd152ee-1bb2-4969-bd7a-52750aef25ba | self-supervised-mean-teacher-for-semi | 2103.03629 | null | https://arxiv.org/abs/2103.03629v3 | https://arxiv.org/pdf/2103.03629v3.pdf | Self-supervised Mean Teacher for Semi-supervised Chest X-ray Classification | The training of deep learning models generally requires a large amount of annotated data for effective convergence and generalisation. However, obtaining high-quality annotations is a laboursome and expensive process due to the need of expert radiologists for the labelling task. The study of semi-supervised learning in medical image analysis is then of crucial importance given that it is much less expensive to obtain unlabelled images than to acquire images labelled by expert radiologists. Essentially, semi-supervised methods leverage large sets of unlabelled data to enable better training convergence and generalisation than using only the small set of labelled images. In this paper, we propose Self-supervised Mean Teacher for Semi-supervised (S$^2$MTS$^2$) learning that combines self-supervised mean-teacher pre-training with semi-supervised fine-tuning. The main innovation of S$^2$MTS$^2$ is the self-supervised mean-teacher pre-training based on the joint contrastive learning, which uses an infinite number of pairs of positive query and key features to improve the mean-teacher representation. The model is then fine-tuned using the exponential moving average teacher framework trained with semi-supervised learning. We validate S$^2$MTS$^2$ on the multi-label classification problems from Chest X-ray14 and CheXpert, and the multi-class classification from ISIC2018, where we show that it outperforms the previous SOTA semi-supervised learning methods by a large margin. | ['Gustavo Carneiro', 'Ian Reid', 'Vasileios Belagiannis', 'Filipe R. Cordeiro', 'Yu Tian', 'Fengbei Liu'] | 2021-03-05 | null | null | null | null | ['semi-supervised-medical-image-classification'] | ['medical'] | [ 5.28847098e-01 5.57860255e-01 -2.36023337e-01 -8.04422975e-01
-1.41799688e+00 -2.88902998e-01 4.61369574e-01 1.30082548e-01
-8.14966261e-01 7.56412625e-01 -3.72300208e-01 -3.23348850e-01
-5.44848382e-01 -4.80057031e-01 -7.76003540e-01 -9.99212801e-01
-1.18759170e-01 7.84591138e-01 2.90749013e-01 4.17523496e-02
-2.32084960e-01 1.75048828e-01 -1.48403835e+00 2.77608484e-01
8.20888460e-01 1.29693162e+00 3.05080652e-01 6.70128882e-01
1.56512544e-01 7.38784969e-01 -2.94677824e-01 -3.59462231e-01
1.48974329e-01 -6.37097001e-01 -1.14851153e+00 3.37055802e-01
4.13524598e-01 -9.44321509e-03 2.90483445e-01 1.04793704e+00
7.61685491e-01 2.14543417e-01 9.55864489e-01 -7.52536595e-01
-2.70189464e-01 4.62219685e-01 -6.80470943e-01 4.13236052e-01
-3.18500310e-01 5.98743856e-02 7.94240177e-01 -7.13960946e-01
3.05456012e-01 8.39150488e-01 7.49711037e-01 6.43010259e-01
-1.04098248e+00 -7.30088472e-01 -2.41021439e-01 1.06304653e-01
-1.35642195e+00 -2.77238518e-01 5.80696046e-01 -5.25179267e-01
7.08296716e-01 1.22779191e-01 4.50334013e-01 7.84779012e-01
6.75358549e-02 7.68350363e-01 1.43103957e+00 -6.89191163e-01
2.87744045e-01 4.71814126e-01 1.40188128e-01 8.84272695e-01
-1.61830500e-01 3.17680448e-01 -1.90035582e-01 1.04604084e-02
6.13256514e-01 -2.29431242e-02 -7.72975683e-02 -4.10372764e-01
-1.02253580e+00 9.63530362e-01 5.60360909e-01 4.23113972e-01
-4.36498940e-01 5.45820445e-02 5.60045660e-01 3.97138327e-01
7.43996024e-01 5.27175069e-01 -6.81896448e-01 2.76470054e-02
-1.34110272e+00 -3.15909147e-01 3.97602260e-01 7.48953640e-01
9.10240173e-01 -2.32427508e-01 4.83157393e-03 1.01061642e+00
3.18083853e-01 4.55089062e-01 8.78351688e-01 -8.60539615e-01
-3.62592787e-02 5.45778871e-01 -2.21495673e-01 -3.58493805e-01
-5.00766158e-01 -9.35416698e-01 -1.10831332e+00 2.21342370e-01
3.31815094e-01 -8.98194015e-02 -1.12693632e+00 1.46270621e+00
4.30031061e-01 -2.37192982e-03 -9.06874165e-02 5.93168259e-01
7.96033144e-01 1.32235289e-01 1.94800064e-01 -6.24536574e-01
1.36805308e+00 -1.21821153e+00 -5.03717482e-01 -6.93712011e-02
1.06466603e+00 -5.59371412e-01 1.06867993e+00 4.96402830e-01
-1.05614305e+00 -8.03347290e-01 -7.62221158e-01 3.19809020e-01
-2.00148121e-01 1.38172075e-01 4.64644611e-01 6.97315097e-01
-1.09623289e+00 5.06623447e-01 -8.62926781e-01 -9.41358656e-02
8.91238809e-01 7.22500503e-01 -3.98968905e-01 -4.91136424e-02
-1.18155813e+00 8.37434590e-01 5.44622302e-01 3.78698334e-02
-1.01161265e+00 -8.00007641e-01 -8.45728278e-01 -1.23210207e-01
5.84640324e-01 -5.59719503e-01 1.41521966e+00 -1.33511710e+00
-1.37363601e+00 1.42197025e+00 2.15760902e-01 -6.46781921e-01
7.71731138e-01 6.70788810e-02 -8.43319371e-02 4.70585406e-01
1.65771678e-01 8.66884351e-01 1.18426979e+00 -1.00117064e+00
-7.70849705e-01 -3.22465420e-01 -2.94760972e-01 2.99971282e-01
-1.30806431e-01 -2.16081187e-01 -1.19394146e-01 -4.99594003e-01
-3.73424590e-02 -9.91591156e-01 -5.32528400e-01 -3.59409660e-01
-2.06953391e-01 -4.85751748e-01 5.42036712e-01 -2.22999096e-01
1.04106832e+00 -2.18843198e+00 -2.41080344e-01 3.03887814e-01
4.44255233e-01 4.56734478e-01 2.50268937e-03 -1.67073756e-01
-4.85017180e-01 -7.91643187e-02 -4.10680383e-01 -1.98687792e-01
-2.51359910e-01 2.68923700e-01 1.33105800e-01 5.32197356e-01
2.95882076e-01 9.69421566e-01 -1.13073123e+00 -1.02264166e+00
1.56011522e-01 3.23962152e-01 -4.00969833e-01 1.44868001e-01
-2.46863980e-02 6.09451056e-01 -3.32455575e-01 3.90045553e-01
3.13700438e-01 -7.46983349e-01 -1.16326576e-02 -2.15786565e-02
3.04998308e-01 -4.43369448e-02 -1.02066898e+00 1.64704609e+00
-5.32498002e-01 2.61957765e-01 -9.77852941e-02 -1.63740611e+00
7.82201707e-01 5.54391563e-01 7.79106975e-01 -6.73563957e-01
2.37631530e-01 5.21532953e-01 -3.88060883e-02 -5.61094344e-01
-1.84509948e-01 -6.44071937e-01 5.99781759e-02 6.33781433e-01
4.34440851e-01 -5.54277122e-01 1.37369379e-01 1.20984606e-01
1.06364942e+00 -1.24980584e-01 4.41847026e-01 -4.28494275e-01
6.78330123e-01 -1.12491801e-01 2.41681904e-01 8.03229153e-01
-3.31267983e-01 4.33595181e-01 1.43599197e-01 -4.75387007e-01
-7.53836989e-01 -8.73794317e-01 -5.72503507e-01 1.23808563e+00
-3.14737022e-01 1.51201829e-01 -8.26201975e-01 -1.45776474e+00
-3.69156539e-01 4.24466014e-01 -9.92055058e-01 -3.26609373e-01
-4.06139642e-01 -9.28682983e-01 4.25535649e-01 5.69483340e-01
4.39529151e-01 -1.26029289e+00 -7.85846472e-01 3.26083228e-02
8.26542731e-03 -7.21047819e-01 -4.51936483e-01 9.40319479e-01
-9.60708380e-01 -1.14228714e+00 -1.08495939e+00 -9.12991345e-01
1.05934870e+00 -1.52908787e-01 1.26523924e+00 8.22246596e-02
-4.59551543e-01 5.48498929e-01 -3.25003386e-01 -5.36294639e-01
-5.65516293e-01 2.69466311e-01 -3.45487893e-02 1.35640148e-02
3.23830813e-01 -4.60783124e-01 -5.94473898e-01 3.93151373e-01
-1.09029078e+00 -3.39245722e-02 1.11086178e+00 1.48080301e+00
1.02631712e+00 2.59525478e-01 8.58134151e-01 -1.37211955e+00
1.56231195e-01 -1.94137082e-01 -4.80821282e-01 2.41562754e-01
-9.56595898e-01 2.32152849e-01 4.56952155e-01 -6.66911304e-01
-9.06226158e-01 3.53088707e-01 -2.27364555e-01 -4.09097195e-01
-2.43170634e-01 4.48411763e-01 2.70622939e-01 -7.36197680e-02
1.05792201e+00 6.10866547e-02 6.24427050e-02 -1.74337983e-01
3.74903023e-01 7.03775108e-01 4.95502263e-01 -2.52257109e-01
5.76801717e-01 5.13467193e-01 1.70811675e-02 -6.06903613e-01
-1.31468308e+00 -7.02997148e-01 -9.40046370e-01 -1.52246624e-01
8.39488804e-01 -7.34653831e-01 -3.72524619e-01 3.41621488e-01
-4.59786564e-01 -7.93546736e-01 -8.48391950e-01 7.92195082e-01
-9.17721033e-01 2.98612058e-01 -3.64743620e-01 -5.62943459e-01
-5.20158231e-01 -1.22857988e+00 1.10673094e+00 1.31868005e-01
-4.55099605e-02 -1.24798417e+00 1.12858616e-01 5.13280034e-01
3.11296374e-01 -3.15955169e-02 7.33948588e-01 -9.58699286e-01
1.69298444e-02 -2.40288943e-01 -2.64555395e-01 7.34910727e-01
2.68196225e-01 -5.33841550e-01 -1.00518966e+00 -4.40496743e-01
2.57702887e-01 -9.23756838e-01 9.10227239e-01 7.24550188e-01
1.14298701e+00 -5.69551364e-02 -3.60716671e-01 5.08353174e-01
1.12892830e+00 -1.74185599e-03 2.92537600e-01 3.54548633e-01
4.38581705e-01 6.73792303e-01 8.08358014e-01 1.31130993e-01
1.73520565e-01 2.85878867e-01 3.97666842e-01 -4.48036373e-01
-1.58959195e-01 2.28193521e-01 -2.12162599e-01 8.42225850e-01
-7.20914155e-02 3.38553846e-01 -1.03867352e+00 6.69040322e-01
-1.43998921e+00 -6.37410402e-01 -7.11647049e-03 2.12311411e+00
1.43514788e+00 1.81704059e-01 1.33555561e-01 3.28352273e-01
5.12029409e-01 -2.71448165e-01 -4.83226746e-01 6.35798182e-03
2.17410266e-01 6.79030836e-01 5.34174740e-01 4.12803590e-01
-1.38949740e+00 6.61556184e-01 5.85195541e+00 1.29507053e+00
-1.19291067e+00 4.15285498e-01 1.00435209e+00 1.39747350e-03
2.04728872e-01 -3.62455875e-01 -6.66516662e-01 3.42498362e-01
1.20460212e+00 3.01784635e-01 -4.22459655e-02 9.35599983e-01
-7.69818723e-02 -4.21316087e-01 -1.19154775e+00 1.10026896e+00
6.21971563e-02 -9.88036275e-01 -3.49334985e-01 -1.63254794e-02
1.06862640e+00 1.22884594e-01 3.13555330e-01 4.55832243e-01
3.90331894e-01 -1.12649369e+00 3.53548735e-01 3.05813640e-01
1.17128110e+00 -5.60892463e-01 9.38469470e-01 6.77107811e-01
-7.66059577e-01 -3.16172354e-02 -2.02354699e-01 4.44331974e-01
-6.86718225e-02 8.19650531e-01 -9.41177726e-01 5.57173371e-01
9.02307630e-01 5.38719535e-01 -6.81978047e-01 7.78814375e-01
-2.44776472e-01 8.34507823e-01 -3.17280978e-01 1.64230406e-01
4.69867945e-01 2.22447515e-01 -2.36133765e-02 1.23205054e+00
4.47947420e-02 1.55254304e-01 1.37870386e-01 3.57011825e-01
8.09765160e-02 1.64269581e-01 -2.09893063e-01 1.52003571e-01
-7.94504583e-02 1.31947911e+00 -9.58639264e-01 -5.85756183e-01
-2.05135971e-01 8.58152092e-01 2.94889122e-01 8.59108344e-02
-6.40812993e-01 -3.33591193e-01 -5.60916245e-01 9.86299589e-02
4.03333634e-01 3.62281799e-01 -1.53882489e-01 -8.15116704e-01
-3.73028129e-01 -8.99983227e-01 7.18748510e-01 -5.56312919e-01
-1.37142956e+00 8.51132691e-01 1.39038742e-01 -1.21129131e+00
-6.37832284e-01 -4.91460502e-01 -1.02819957e-01 6.42117679e-01
-1.75782633e+00 -1.19495034e+00 -1.33683413e-01 6.61007166e-01
5.13047636e-01 -1.16431519e-01 9.04399395e-01 3.89825761e-01
-1.78667337e-01 8.26259673e-01 1.38733119e-01 1.70896620e-01
9.25373077e-01 -1.55978560e+00 -2.74530500e-01 2.52860129e-01
1.83381692e-01 2.29323283e-01 5.34630597e-01 -2.55379230e-01
-5.04617870e-01 -1.13821423e+00 7.17368186e-01 -4.86275136e-01
4.69525516e-01 3.98556590e-02 -9.22786057e-01 6.00521684e-01
1.34673238e-01 7.05266714e-01 9.16876793e-01 -1.26940280e-01
1.26312757e-02 -2.12413788e-01 -1.21443653e+00 7.63560785e-03
6.39054120e-01 -5.32507479e-01 -5.90999842e-01 6.01485968e-01
3.19842279e-01 -3.14706266e-01 -1.02229881e+00 5.90583563e-01
3.31552267e-01 -7.81925797e-01 8.25631499e-01 -4.18496877e-01
2.72105128e-01 8.71710032e-02 2.02824190e-01 -1.31216526e+00
-1.07486583e-01 -3.92033219e-01 1.66828141e-01 7.45477378e-01
5.79191267e-01 -4.61164504e-01 8.55921566e-01 2.11978123e-01
-1.47385404e-01 -1.24746811e+00 -9.08646166e-01 -7.48367608e-01
2.14202583e-01 -3.73496145e-01 -4.56415489e-02 1.05868375e+00
-2.03643844e-01 4.55585867e-01 -2.18536016e-02 -1.76899940e-01
6.99678719e-01 -1.53985739e-01 4.87528950e-01 -1.26599228e+00
-6.10180020e-01 -2.76776224e-01 -2.93888152e-01 -8.80776942e-01
1.38635352e-01 -9.33804810e-01 3.25497508e-01 -1.17632377e+00
4.41762775e-01 -7.75658190e-01 -5.50626814e-01 6.54383242e-01
-4.12810266e-01 5.90972722e-01 -4.28510547e-01 3.20197672e-01
-1.02219534e+00 3.44040871e-01 1.48178232e+00 -2.06418082e-01
-2.11376175e-01 4.80865926e-01 -4.73343849e-01 8.19292068e-01
4.58685428e-01 -7.61024594e-01 -5.82945764e-01 -6.63094595e-02
1.96481735e-01 -2.09741946e-02 2.14188129e-01 -7.14054167e-01
8.46718177e-02 6.83205798e-02 2.52067238e-01 -5.81867576e-01
1.09358348e-01 -9.18097019e-01 -3.07627112e-01 5.92376351e-01
-6.49954379e-01 -3.93643439e-01 -1.86299644e-02 5.81251085e-01
-3.35800052e-01 -5.73876441e-01 1.19423079e+00 -4.88749266e-01
-3.85313839e-01 2.64248103e-01 -9.27273631e-02 2.49832124e-01
1.13193727e+00 -6.90146759e-02 1.54879496e-01 -2.80248463e-01
-1.18552017e+00 2.56752104e-01 -3.66032384e-02 -1.86960220e-01
4.19362307e-01 -1.08256018e+00 -6.16087794e-01 2.26679623e-01
5.51988482e-02 5.68241894e-01 4.07147974e-01 1.33938432e+00
-1.89019516e-01 3.42213511e-01 -9.38809756e-03 -1.07109296e+00
-1.13060045e+00 7.01881051e-01 5.40690422e-01 -8.79968286e-01
-3.61524701e-01 1.26544333e+00 3.75054032e-01 -6.77802682e-01
3.09281826e-01 -1.61533415e-01 -2.08691716e-01 6.04421310e-02
4.55306232e-01 1.68965146e-01 3.26226473e-01 -5.04312813e-01
-1.66592449e-01 5.63633084e-01 -3.91457111e-01 -1.77065834e-01
1.36217606e+00 -6.25058785e-02 -4.28775661e-02 5.36966562e-01
1.35221303e+00 -5.43217123e-01 -1.23132944e+00 -5.78562081e-01
1.64124429e-01 6.07508980e-02 1.67652875e-01 -8.05558860e-01
-1.05517066e+00 1.03372490e+00 8.92566144e-01 8.15053582e-02
1.27568889e+00 3.89524341e-01 3.49400163e-01 4.21262681e-01
2.21534759e-01 -1.14967465e+00 4.55827028e-01 1.72910318e-01
4.29311812e-01 -1.72851479e+00 6.84389248e-02 -2.00865626e-01
-7.21023500e-01 8.97840321e-01 4.96818781e-01 -2.20274869e-02
9.86023068e-01 8.84768739e-02 4.17131335e-01 -4.32188421e-01
-6.43269300e-01 -2.44931653e-01 6.34367287e-01 4.49819475e-01
4.20389920e-01 -3.46956924e-02 -2.55608350e-01 6.62711143e-01
-9.88252908e-02 1.96240827e-01 -7.26164505e-02 9.47593629e-01
-3.01782340e-01 -9.94352102e-01 -1.57960981e-01 8.18355083e-01
-7.59534001e-01 -1.46520197e-01 8.75543207e-02 7.66093850e-01
3.00521255e-01 7.10065544e-01 -1.07681803e-01 -1.97286122e-02
-1.05427820e-02 1.54313460e-01 5.61626554e-01 -9.65517998e-01
-6.55707717e-01 2.80948967e-01 -3.53297889e-01 -2.60901630e-01
-9.14920211e-01 -5.21207511e-01 -1.35908175e+00 4.39569205e-01
-8.90742898e-01 4.46752667e-01 4.57305133e-01 1.20858860e+00
-1.21200889e-01 7.30992019e-01 8.20298374e-01 -6.41971409e-01
-9.72146451e-01 -1.21469414e+00 -7.22132325e-01 4.60270613e-01
3.84770989e-01 -6.40750945e-01 -4.78458792e-01 4.58939105e-01] | [14.693458557128906, -2.2140378952026367] |
ad841b67-4167-4928-8526-7cb648db8e6c | why-so-deep-towards-boosting-previously | 2201.03212 | null | https://arxiv.org/abs/2201.03212v1 | https://arxiv.org/pdf/2201.03212v1.pdf | Why-So-Deep: Towards Boosting Previously Trained Models for Visual Place Recognition | Deep learning-based image retrieval techniques for the loop closure detection demonstrate satisfactory performance. However, it is still challenging to achieve high-level performance based on previously trained models in different geographical regions. This paper addresses the problem of their deployment with simultaneous localization and mapping (SLAM) systems in the new environment. The general baseline approach uses additional information, such as GPS, sequential keyframes tracking, and re-training the whole environment to enhance the recall rate. We propose a novel approach for improving image retrieval based on previously trained models. We present an intelligent method, MAQBOOL, to amplify the power of pre-trained models for better image recall and its application to real-time multiagent SLAM systems. We achieve comparable image retrieval results at a low descriptor dimension (512-D), compared to the high descriptor dimension (4096-D) of state-of-the-art methods. We use spatial information to improve the recall rate in image retrieval on pre-trained models. | ['Ming Liu', 'Darwin Lau', 'Yuxiang Sun', 'M. Usman Maqbool Bhutta'] | 2022-01-10 | null | null | null | null | ['loop-closure-detection', 'visual-place-recognition'] | ['computer-vision', 'computer-vision'] | [-2.78275251e-01 -5.34108281e-01 -1.08126119e-01 -2.52461433e-01
-1.26938355e+00 -5.30115426e-01 8.86783361e-01 5.18315673e-01
-9.64735746e-01 3.91812205e-01 -2.63195425e-01 -8.69264677e-02
-3.86992425e-01 -7.19065964e-01 -7.25617826e-01 -4.57760900e-01
-3.54136378e-01 7.42311537e-01 3.81324738e-01 -4.40782368e-01
5.47347963e-01 8.52395058e-01 -1.69331491e+00 -8.11896771e-02
5.93358636e-01 8.86420965e-01 6.86013043e-01 8.02683294e-01
-7.18150986e-03 7.48173952e-01 -8.49019408e-01 1.22064069e-01
5.20563006e-01 9.33567658e-02 -7.43441463e-01 -3.22079957e-01
7.56671727e-01 -6.32354558e-01 -6.08012557e-01 8.72001708e-01
6.76192343e-01 3.24819505e-01 3.30438077e-01 -1.22167647e+00
-5.86744010e-01 -1.86365262e-01 -4.25973803e-01 9.32724252e-02
6.26767457e-01 -2.73039341e-01 5.31960905e-01 -8.41460645e-01
7.78219342e-01 1.06987715e+00 8.18460286e-01 -1.14050403e-01
-7.33426988e-01 -5.71494281e-01 -6.09107018e-02 3.06028157e-01
-1.95198512e+00 -3.46224129e-01 3.04705769e-01 -1.62697330e-01
1.23129094e+00 1.54461060e-02 6.49029970e-01 4.20085996e-01
5.07980943e-01 6.77700698e-01 8.10503006e-01 -6.29503131e-01
1.16332412e-01 1.12771310e-01 -2.30927572e-01 9.90393281e-01
1.49363592e-01 2.92502820e-01 -8.69659245e-01 -1.19936295e-01
7.84810722e-01 1.16805829e-01 -7.92064425e-03 -7.03097045e-01
-1.39994347e+00 9.18881238e-01 9.36188340e-01 3.94067556e-01
-5.09914219e-01 4.82916683e-01 2.72941411e-01 4.85523611e-01
3.85470390e-01 6.40426874e-01 -2.37562761e-01 2.63987929e-02
-1.36191189e+00 4.74644870e-01 4.80488479e-01 1.26943159e+00
1.33990860e+00 -2.96439946e-01 2.00166762e-01 5.73310554e-01
4.94536072e-01 1.15386105e+00 3.90557885e-01 -9.62890863e-01
1.61656961e-01 3.79965305e-01 3.35973829e-01 -1.44723046e+00
-5.79683185e-01 -3.67844045e-01 -4.70955551e-01 2.95677539e-02
-6.50613606e-02 2.56586403e-01 -1.07949829e+00 1.38218343e+00
2.95665830e-01 6.32037371e-02 3.32361251e-01 9.92437065e-01
7.78202295e-01 7.70391941e-01 -1.57067850e-01 2.60838956e-01
9.46201444e-01 -1.23545790e+00 -6.39199197e-01 -5.60639918e-01
1.03986859e+00 -1.12699759e+00 4.77142602e-01 1.13371007e-01
-5.90728760e-01 -7.43365288e-01 -1.26126051e+00 -1.02715142e-01
-8.13010037e-01 3.34408641e-01 7.62645125e-01 3.64637733e-01
-1.67635524e+00 3.31920147e-01 -7.23695040e-01 -9.63011265e-01
-2.19950229e-01 7.06627131e-01 -7.74641275e-01 -2.45552287e-01
-1.25414026e+00 1.29689980e+00 3.96987528e-01 -4.88814749e-02
-9.66603339e-01 -2.82319576e-01 -1.04036498e+00 -3.10179830e-01
1.47567883e-01 -4.86959964e-01 9.83358264e-01 -3.03784519e-01
-1.26867163e+00 9.72427845e-01 -5.62930964e-02 -6.18496060e-01
4.04800832e-01 -4.31058645e-01 -1.29917547e-01 3.53188843e-01
4.27096516e-01 1.09922886e+00 6.66540444e-01 -1.27705371e+00
-8.41931343e-01 -3.12529624e-01 2.35883698e-01 4.69442010e-01
7.39470720e-02 -2.45624840e-01 -8.27224493e-01 -1.96289003e-01
4.21533793e-01 -1.39560950e+00 -3.94982338e-01 8.42533112e-02
2.61619747e-01 8.78434107e-02 8.27493489e-01 -4.37900126e-01
7.58844376e-01 -2.07318211e+00 -1.19937442e-01 3.33509058e-01
-1.56723663e-01 3.73898566e-01 -5.58917463e-01 7.63421237e-01
4.48113292e-01 -4.22975570e-01 2.34163195e-01 -6.34391427e-01
1.91281606e-02 4.98127103e-01 -1.71104372e-01 8.74840975e-01
-1.13504305e-01 7.79125690e-01 -1.02910697e+00 -5.53552747e-01
7.18361795e-01 4.00088996e-01 -1.37466297e-01 -2.27068160e-02
2.06539094e-01 4.11934078e-01 -3.81901622e-01 6.28064096e-01
9.74569798e-01 -1.31046459e-01 2.68822405e-02 5.39759733e-03
-2.64929146e-01 -2.49362327e-02 -1.27356124e+00 2.59292459e+00
-6.36596859e-01 8.72819483e-01 -1.77079756e-02 -6.94477975e-01
1.21882582e+00 -1.61282808e-01 5.37704527e-01 -1.29675901e+00
-1.78008318e-01 5.07507026e-01 -6.14755750e-01 -3.06480229e-01
1.33530509e+00 4.45389837e-01 -4.50525463e-01 2.70689726e-01
1.82338789e-01 -4.47372884e-01 5.51596433e-02 3.43268901e-01
1.06944275e+00 1.63789421e-01 1.70323759e-01 -4.10056412e-01
5.17898083e-01 4.79262620e-01 2.68919021e-02 1.36698675e+00
-3.92349631e-01 4.20434654e-01 -3.30242246e-01 -7.94292033e-01
-1.08365393e+00 -5.97366512e-01 -2.73805391e-03 1.11592913e+00
9.41581070e-01 -6.49132133e-01 -4.15948689e-01 -3.71296406e-01
1.56858772e-01 3.43354978e-02 -4.24638510e-01 -1.21830128e-01
-4.54257786e-01 -5.86961627e-01 7.06188798e-01 1.47269428e-01
9.28611755e-01 -7.14766681e-01 -9.09700751e-01 2.01797903e-01
-2.33645827e-01 -1.05138266e+00 -4.45179977e-02 1.80799276e-01
-5.05345464e-01 -8.66155326e-01 -8.57763886e-01 -8.58891189e-01
4.96955007e-01 7.32855976e-01 9.66863036e-01 4.00606066e-01
-1.16237171e-01 6.72475219e-01 -5.85989118e-01 -3.71046841e-01
-3.08662027e-01 3.34434360e-01 3.03727776e-01 -5.61405718e-01
2.84139305e-01 1.05712436e-01 -5.67759752e-01 3.09259027e-01
-8.84906828e-01 -2.31160492e-01 8.65233243e-01 7.80653298e-01
6.13892555e-01 -2.99597442e-01 1.24997310e-01 -2.83974349e-01
3.46924722e-01 -7.55940676e-02 -9.21393991e-01 4.73990262e-01
-7.40358770e-01 2.55325526e-01 -1.41127884e-01 -5.08183300e-01
-4.91231531e-01 2.84599930e-01 1.45378634e-02 -3.41353476e-01
8.07886198e-02 6.72485888e-01 5.19635022e-01 -1.02882290e+00
7.57612646e-01 3.41802716e-01 1.14903614e-01 -1.79130033e-01
4.45486754e-01 6.69365883e-01 4.96812224e-01 -3.25706244e-01
7.24815428e-01 6.10235095e-01 1.65550038e-01 -7.51714110e-01
-4.82368916e-01 -9.56446946e-01 -7.52927661e-01 -2.13151723e-01
3.90639275e-01 -1.37461627e+00 -4.89739269e-01 5.75292468e-01
-1.30759847e+00 -3.47829491e-01 2.27889240e-01 8.09798837e-01
-5.76998889e-01 4.56040323e-01 -4.57557529e-01 -6.83223665e-01
-3.76869887e-01 -1.25905120e+00 1.77315223e+00 1.38958916e-01
3.76041263e-01 -7.68501818e-01 5.83061516e-01 6.32605795e-03
7.18112469e-01 -2.91777011e-02 1.30189210e-01 -4.82414693e-01
-1.07552028e+00 -5.37652194e-01 -3.56449008e-01 -3.18653822e-01
-2.05952972e-01 -5.77610254e-01 -8.01061630e-01 -8.36713254e-01
-3.33196819e-01 -2.58210897e-01 8.60307574e-01 1.24184288e-01
2.52226889e-01 1.24511756e-01 -6.48916960e-01 5.73634386e-01
1.62082791e+00 2.05839097e-01 5.90170681e-01 1.09137702e+00
3.21322918e-01 3.21379393e-01 1.27748835e+00 3.46971273e-01
6.18717968e-01 9.02137220e-01 5.64935386e-01 -4.02979583e-01
3.33871990e-02 -1.49376631e-01 1.54640391e-01 4.89345580e-01
3.20136696e-01 -2.17559516e-01 -1.09132278e+00 6.93818331e-01
-2.24149823e+00 -4.65710402e-01 2.57277191e-01 2.15912294e+00
4.19704989e-02 -1.74003765e-01 -4.11079675e-01 -3.30738932e-01
6.21266067e-01 4.12084609e-01 -2.14958906e-01 -8.63177255e-02
-1.87694266e-01 -1.31433353e-01 1.07506776e+00 7.60472357e-01
-1.39498770e+00 1.41035020e+00 6.88804626e+00 8.41221273e-01
-1.28803122e+00 2.97740906e-01 -2.06438154e-01 1.87451139e-01
1.60435900e-01 1.25840530e-01 -7.53912568e-01 -6.70459005e-04
9.34289157e-01 2.73026466e-01 3.78373086e-01 8.79690051e-01
-2.88384646e-01 -7.28284240e-01 -7.65576959e-01 1.42297196e+00
4.70702738e-01 -1.35820222e+00 -2.83396542e-02 1.83167934e-01
7.53270566e-01 7.32702076e-01 1.10855447e-02 5.58404922e-01
4.39317077e-02 -5.49966574e-01 7.96434522e-01 5.66469133e-01
6.68057203e-01 -6.80568755e-01 1.18141854e+00 3.48211110e-01
-1.22461236e+00 1.35315865e-01 -6.73587322e-01 3.51181217e-02
-7.03440830e-02 1.57014832e-01 -1.10209835e+00 9.00379419e-01
8.16728890e-01 5.83204508e-01 -9.99534428e-01 1.22021234e+00
9.35478956e-02 -5.27083695e-01 -8.51575196e-01 -2.00961649e-01
7.45977819e-01 1.49585843e-01 3.99343729e-01 1.16487336e+00
5.62565088e-01 -3.61112148e-01 5.29622018e-01 1.28451303e-01
3.48536968e-01 6.82190061e-02 -1.09394348e+00 4.18677449e-01
5.20616829e-01 9.82977986e-01 -6.32779121e-01 -4.29700822e-01
-2.54230231e-01 1.14736593e+00 3.32962304e-01 2.94594407e-01
-7.85639942e-01 -3.54749054e-01 3.22101474e-01 -1.67764038e-01
5.86338574e-03 -6.75426543e-01 4.90155101e-01 -1.00580621e+00
-1.32598653e-01 -5.86491406e-01 2.75436133e-01 -1.05170548e+00
-6.20821536e-01 7.76242614e-01 3.48817953e-03 -1.22936010e+00
-4.98915464e-01 -2.80620039e-01 2.48049706e-01 6.82782173e-01
-1.86913204e+00 -1.64621520e+00 -5.62410116e-01 5.57551146e-01
2.25308523e-01 -2.80194938e-01 1.13096750e+00 5.57377100e-01
2.21625641e-01 3.38976353e-01 5.94741881e-01 -1.01139180e-01
1.22227943e+00 -8.99403334e-01 2.84514070e-01 8.73519123e-01
5.32791376e-01 6.10543549e-01 6.24433279e-01 -5.71623683e-01
-1.70545805e+00 -9.42476332e-01 9.56482708e-01 -4.73869562e-01
3.40336174e-01 -3.42784464e-01 -4.75599915e-01 6.99152112e-01
4.22211923e-02 -4.06165011e-02 2.13574067e-01 9.14416760e-02
-1.24393634e-01 -2.19695687e-01 -1.08255267e+00 2.30648845e-01
6.43405139e-01 -9.59455907e-01 -3.31782907e-01 6.58911824e-01
5.01175880e-01 -8.03592026e-01 -6.49301350e-01 5.86429238e-01
5.70909739e-01 -8.25173855e-01 1.07118857e+00 2.87921932e-02
-4.97286618e-01 -6.97057366e-01 -4.91270095e-01 -1.13554835e+00
-2.31270000e-01 -3.83054137e-01 2.85928607e-01 6.00181937e-01
1.83725655e-02 -5.23379683e-01 8.16943049e-01 2.19008103e-01
7.72763118e-02 -2.82161027e-01 -9.69920218e-01 -8.43137383e-01
-3.16850781e-01 -1.18088782e-01 6.37918472e-01 8.39503050e-01
-3.36587727e-01 -7.78269069e-03 -5.54152966e-01 7.05256522e-01
4.95820463e-01 2.78681278e-01 1.33333385e+00 -1.07188654e+00
2.82860816e-01 7.91774169e-02 -1.01870990e+00 -1.26498425e+00
1.66833028e-01 -6.54487133e-01 3.18134427e-01 -1.71176720e+00
4.43936437e-02 -7.86212146e-01 -3.81781518e-01 3.26866001e-01
2.83222675e-01 5.25289357e-01 4.43407953e-01 7.12751865e-01
-1.43349218e+00 4.61033434e-01 8.17917943e-01 -3.48512769e-01
-1.64590433e-01 -5.81882894e-01 1.89337172e-02 3.19635212e-01
5.57548463e-01 -5.63625157e-01 -1.23190850e-01 -9.72279310e-01
2.35155389e-01 1.56141952e-01 4.25910473e-01 -1.45717347e+00
9.35178399e-01 2.10079879e-01 3.69140446e-01 -1.09631813e+00
6.85205579e-01 -8.33958387e-01 3.38231832e-01 7.79613972e-01
-4.38561551e-02 5.12524724e-01 4.80594486e-01 6.00169539e-01
-5.98113954e-01 -2.42086560e-01 3.78017157e-01 -2.82564700e-01
-1.50821161e+00 1.66699216e-01 -4.13660407e-01 -8.02567422e-01
8.64040494e-01 9.18561667e-02 -4.97207910e-01 -6.35799229e-01
-4.98755425e-01 3.77326369e-01 8.64061654e-01 5.70567667e-01
6.67221844e-01 -1.29610026e+00 -3.92580479e-01 3.48656029e-01
5.94180584e-01 -1.54902935e-01 2.83638448e-01 6.90099895e-01
-1.31180561e+00 7.61854410e-01 -3.84378642e-01 -9.93646502e-01
-1.09538817e+00 6.41407728e-01 3.69922251e-01 -3.73336822e-01
-3.24205875e-01 6.13558769e-01 -2.14665726e-01 -7.48733580e-01
2.75252074e-01 8.91910195e-02 7.80811384e-02 -2.65215971e-02
5.50793886e-01 8.31574351e-02 4.61892784e-01 -8.27575088e-01
-7.64568210e-01 1.02173793e+00 -5.98198622e-02 -2.36802906e-01
1.12203479e+00 -4.52122182e-01 -1.76269159e-01 9.27740335e-02
1.29413509e+00 -1.49604708e-01 -7.98514009e-01 -3.97006869e-01
7.19587132e-02 -5.97737610e-01 3.32434773e-01 -5.93706429e-01
-6.40112519e-01 5.69168270e-01 1.24154973e+00 -1.37381732e-01
7.51554072e-01 -6.29310533e-02 6.85121000e-01 1.12654161e+00
1.18854368e+00 -1.12438226e+00 1.39385208e-01 7.89616942e-01
6.55274630e-01 -1.69632363e+00 2.35064283e-01 9.45292711e-02
-2.33943239e-01 8.76742840e-01 2.46515140e-01 -3.65653843e-01
4.71190393e-01 1.31709829e-01 4.78690237e-01 -3.37515980e-01
-2.29783103e-01 -5.32190681e-01 1.33272231e-01 5.80473483e-01
-1.97792456e-01 -1.65067837e-01 5.75444140e-02 -3.98370862e-01
8.80634636e-02 -6.53260797e-02 1.40629500e-01 1.41781557e+00
-6.00895286e-01 -1.07404792e+00 -5.15923023e-01 -1.99687123e-01
-1.82891697e-01 -6.20053262e-02 -2.69096434e-01 1.08570349e+00
-3.69498506e-02 8.64778697e-01 3.03197466e-02 -4.72771227e-01
1.97593614e-01 -3.75080258e-01 6.33359730e-01 -2.36791462e-01
-4.12120938e-01 2.70432290e-02 -1.10935129e-01 -9.13288295e-01
-7.71691978e-01 -3.02918702e-01 -9.32531834e-01 -4.49646235e-01
-5.29051781e-01 3.77455235e-01 1.24930072e+00 6.85313761e-01
7.36419678e-01 2.24759132e-02 3.15119594e-01 -1.26843214e+00
-2.20648184e-01 -9.10573661e-01 -5.26016057e-01 1.81626156e-01
7.16452718e-01 -8.77158761e-01 -3.72821502e-02 -4.10152614e-01] | [7.491269111633301, -2.1169350147247314] |
c29ce70f-93ab-4527-9152-625b566262c4 | discourse-representation-structure-parsing-2 | 2306.09725 | null | https://arxiv.org/abs/2306.09725v1 | https://arxiv.org/pdf/2306.09725v1.pdf | Discourse Representation Structure Parsing for Chinese | Previous work has predominantly focused on monolingual English semantic parsing. We, instead, explore the feasibility of Chinese semantic parsing in the absence of labeled data for Chinese meaning representations. We describe the pipeline of automatically collecting the linearized Chinese meaning representation data for sequential-to sequential neural networks. We further propose a test suite designed explicitly for Chinese semantic parsing, which provides fine-grained evaluation for parsing performance, where we aim to study Chinese parsing difficulties. Our experimental results show that the difficulty of Chinese semantic parsing is mainly caused by adverbs. Realizing Chinese parsing through machine translation and an English parser yields slightly lower performance than training a model directly on Chinese data. | ['Johan Bos', 'Xiao Zhang', 'Chunliu Wang'] | 2023-06-16 | null | null | null | null | ['machine-translation', 'semantic-parsing'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.18334490e-01 5.95559657e-01 2.59851688e-04 -7.77908206e-01
-1.18555486e+00 -7.39468753e-01 2.17996791e-01 -1.28700221e-02
-5.97446084e-01 6.86987936e-01 5.95432997e-01 -8.20407510e-01
6.64048314e-01 -8.18863988e-01 -7.19513655e-01 -1.06939234e-01
2.46394321e-01 5.40835977e-01 3.83177423e-03 -2.97564548e-02
4.86658253e-02 -1.77765533e-01 -8.89922023e-01 7.77492523e-01
7.16215849e-01 5.63267946e-01 6.55906141e-01 6.95225954e-01
-8.02381456e-01 8.57492685e-01 -8.28922451e-01 -2.49164313e-01
-1.16521761e-01 -8.10905278e-01 -1.48407447e+00 -2.57542938e-01
5.57548515e-02 -1.95270851e-01 4.68321592e-01 1.08391309e+00
7.68675050e-03 -2.93163151e-01 -6.66787773e-02 -6.18337750e-01
-9.69938934e-01 1.22301388e+00 2.86257327e-01 -3.55580566e-03
4.45289493e-01 -4.18343276e-01 1.20642507e+00 -4.86562103e-01
9.22264159e-01 1.37973213e+00 6.20005906e-01 1.09602094e+00
-9.15368438e-01 -4.17339861e-01 6.98848903e-01 -2.25897089e-01
-8.49709809e-01 -1.85295179e-01 5.24213672e-01 3.04184735e-01
1.64645064e+00 -6.83264658e-02 2.92230874e-01 1.17882609e+00
-2.07764521e-01 9.70851302e-01 1.09742928e+00 -8.97359252e-01
3.04010898e-01 -8.65811482e-02 5.43657482e-01 4.77902293e-01
1.58438273e-02 -2.06482649e-01 -1.62733763e-01 2.26730868e-01
5.91570437e-01 -5.66805184e-01 2.14203700e-01 4.35897619e-01
-9.24235344e-01 9.32802200e-01 3.20960969e-01 4.85209405e-01
-7.29776360e-03 5.37609518e-01 5.00621378e-01 5.26750505e-01
6.48070037e-01 5.90724349e-01 -1.23052704e+00 -3.75914186e-01
-3.05584937e-01 1.11138457e-02 1.06584084e+00 1.35726690e+00
6.55789733e-01 -1.64688379e-02 4.33490247e-01 7.98104763e-01
3.70089673e-02 4.23552990e-01 6.49108171e-01 -1.14928174e+00
7.90639818e-01 3.59005541e-01 -8.91748667e-02 -2.29045466e-01
-6.42926514e-01 2.93359179e-02 2.11087942e-01 -4.77697879e-01
7.03828990e-01 -6.62323713e-01 -7.15370893e-01 2.03135109e+00
-1.00060888e-01 -2.05492139e-01 8.52451682e-01 6.93788707e-01
7.99443901e-01 5.16882300e-01 8.56629491e-01 1.35767549e-01
1.50541747e+00 -1.04327142e+00 -5.86233377e-01 -6.40938580e-01
1.47937715e+00 -6.86415970e-01 1.49348879e+00 -4.65046382e-03
-1.16984475e+00 -3.88036370e-01 -6.12064719e-01 -3.76266718e-01
-4.50625211e-01 2.38200992e-01 9.48093712e-01 8.30046177e-01
-1.39839530e+00 5.33840239e-01 -1.13715267e+00 -9.58430707e-01
1.20924681e-01 1.64094777e-03 -2.66846210e-01 -1.68738589e-01
-1.22592986e+00 9.34038699e-01 8.00713420e-01 -9.64499936e-02
-2.94731706e-01 -2.22531468e-01 -1.21856391e+00 3.94459628e-02
2.41349101e-01 -3.71727645e-01 1.63968682e+00 -1.41684163e+00
-1.39804053e+00 1.15649474e+00 -5.00555158e-01 -3.77275139e-01
-6.36629313e-02 -6.01497710e-01 -1.68722972e-01 2.69357324e-01
4.13926452e-01 1.11692202e+00 1.00956582e-01 -9.98878539e-01
-6.94145143e-01 -3.98520827e-01 1.93957716e-01 2.79720217e-01
-5.41962534e-02 6.39744461e-01 -6.34810805e-01 -7.09112644e-01
3.71479422e-01 -9.11931753e-01 -3.58630389e-01 -6.93359017e-01
-2.77964205e-01 -3.56999576e-01 7.65014172e-01 -9.13185179e-01
8.17711174e-01 -1.84815085e+00 -3.08972239e-01 -3.22697759e-01
-5.61455250e-01 -1.24020219e-01 -5.58732450e-01 3.16299200e-01
-1.92134529e-01 6.22348011e-01 -2.88136899e-01 -4.83383834e-01
-1.68127045e-01 6.93650365e-01 -2.60944039e-01 -2.60044634e-01
9.35923696e-01 1.39971662e+00 -8.73501837e-01 -2.54396677e-01
-2.36872062e-01 -1.78135671e-02 -9.31784034e-01 2.18371853e-01
-5.51318765e-01 4.86226797e-01 -8.39159667e-01 8.11881185e-01
5.28687775e-01 -2.18644619e-01 9.33392823e-01 4.65542853e-01
-7.64494166e-02 1.01452649e+00 -3.99417907e-01 2.17767286e+00
-7.87558496e-01 3.29598278e-01 1.52194768e-01 -9.76451337e-01
5.98322451e-01 1.95503607e-01 -2.83716798e-01 -1.02662694e+00
5.48199378e-02 2.85064250e-01 -2.16519594e-01 -1.50250763e-01
5.96698105e-01 -3.83133411e-01 -7.81148911e-01 6.19178355e-01
1.18644752e-01 -1.65634766e-01 -2.10366905e-01 1.59649141e-02
1.18703616e+00 7.78180301e-01 5.14426976e-02 -5.89005351e-01
2.16054723e-01 6.77073598e-01 4.84758377e-01 7.29055047e-01
-2.38621142e-02 7.22945094e-01 8.09550464e-01 -3.57912928e-01
-9.55289721e-01 -9.66891646e-01 4.51766029e-02 1.42301977e+00
9.87880677e-02 -6.04935586e-01 -1.43928266e+00 -1.16301739e+00
-6.19777203e-01 1.25278521e+00 -3.38763267e-01 3.07896972e-01
-1.18161404e+00 -7.72197366e-01 8.34916890e-01 1.07272279e+00
3.78042370e-01 -1.52699685e+00 -5.95232487e-01 4.49118704e-01
-3.95053715e-01 -1.76975381e+00 9.64746922e-02 5.14944851e-01
-1.20353842e+00 -1.08704710e+00 -1.67062268e-01 -1.30226088e+00
5.84409773e-01 1.29712457e-02 1.36758304e+00 4.63723272e-01
1.90560326e-01 4.40453619e-01 -7.99601376e-01 -4.15396661e-01
-7.86400139e-01 2.87256956e-01 -5.96444011e-01 -1.04034734e+00
9.12932873e-01 -3.18826921e-02 -1.43552914e-01 7.99491256e-02
-6.47031486e-01 2.71655411e-01 3.19213688e-01 8.64920676e-01
3.65247369e-01 -3.89818609e-01 6.03593469e-01 -1.45902371e+00
5.15780330e-01 -4.02145684e-01 -5.23170412e-01 2.06716046e-01
-2.05348000e-01 3.97691637e-01 8.21730971e-01 4.57807779e-02
-1.75200188e+00 1.12015642e-01 -7.65352845e-01 3.91312510e-01
-6.84456408e-01 4.95859802e-01 -2.79352665e-01 5.87961137e-01
3.47187340e-01 -1.54785141e-01 -2.49380276e-01 -9.04168189e-01
6.07006490e-01 4.48584169e-01 5.72250068e-01 -1.39384127e+00
8.37814510e-02 3.32456082e-01 -6.12920046e-01 -7.90755212e-01
-1.08527386e+00 -1.64973810e-01 -7.41825998e-01 6.56746566e-01
1.53798926e+00 -1.13087809e+00 -1.74065009e-01 2.88990647e-01
-1.51331842e+00 -8.37166965e-01 -1.04420386e-01 3.31221431e-01
-5.47638357e-01 3.98649931e-01 -1.29535127e+00 -2.61116624e-01
-2.96166986e-01 -9.40407753e-01 1.30472291e+00 -1.42711058e-01
-4.07283813e-01 -1.31143165e+00 -2.32573882e-01 3.53934199e-01
3.39784235e-01 -2.39534929e-01 1.29264724e+00 -9.70671654e-01
-4.30336267e-01 9.81752947e-02 -6.17817879e-01 3.77261758e-01
-3.87073755e-02 -6.33414447e-01 -1.18685555e+00 9.57487449e-02
4.52176891e-02 -7.31296837e-01 6.45182848e-01 2.75633782e-01
1.03599930e+00 8.88177473e-03 -5.79421297e-02 5.66618621e-01
1.51249194e+00 1.27852604e-01 4.91033435e-01 4.92264390e-01
7.22722232e-01 9.60650861e-01 6.25051141e-01 -1.86539650e-01
7.89948881e-01 7.24439546e-02 6.03432879e-02 -1.29941896e-01
-1.55366212e-01 -5.62889516e-01 3.94794047e-01 8.39356542e-01
2.35011473e-01 -1.85571060e-01 -9.32869613e-01 3.88540268e-01
-1.62025487e+00 -1.27524689e-01 -5.46288081e-02 1.62231469e+00
7.99752474e-01 2.66626954e-01 -5.28432906e-01 -7.63081133e-01
5.77532470e-01 -1.76582918e-01 -4.26838361e-02 -9.30021584e-01
-2.28732675e-02 9.49089527e-01 5.98401129e-01 6.53173566e-01
-1.15571952e+00 2.18658257e+00 7.36921263e+00 4.52728242e-01
-9.48407352e-01 4.38543320e-01 7.32389987e-01 5.15139997e-01
-5.55798471e-01 5.40640533e-01 -9.43212926e-01 2.38223877e-02
1.32213485e+00 4.87721175e-01 3.63632113e-01 1.05614257e+00
6.97851926e-02 -2.43925601e-01 -1.03301370e+00 2.95596451e-01
-1.86976537e-01 -1.19685996e+00 2.27772631e-02 -4.52930719e-01
4.15766895e-01 4.56027001e-01 -7.26634324e-01 5.07049322e-01
6.85027540e-01 -8.48290503e-01 8.44496429e-01 -2.50449777e-01
7.84821630e-01 -4.59302276e-01 7.52331614e-01 2.30496183e-01
-1.13858449e+00 1.56307593e-01 -5.66649437e-01 -5.16202986e-01
3.31520706e-01 -1.30522996e-01 -5.84566176e-01 3.28600049e-01
7.47370839e-01 7.45997250e-01 -6.86115623e-01 -1.85933292e-01
-1.01131415e+00 9.71876740e-01 -2.81191766e-01 -2.32274473e-01
5.88576376e-01 -1.18085660e-01 -3.73875573e-02 1.55335271e+00
2.23117024e-01 3.24750125e-01 4.20706540e-01 1.01552534e+00
-6.37753531e-02 2.69383430e-01 -4.65896279e-01 -2.99401760e-01
3.25367361e-01 9.36079860e-01 -1.25095272e+00 -5.57631373e-01
-8.20995510e-01 1.13767529e+00 7.89544821e-01 3.25827211e-01
-4.27294105e-01 -1.15177341e-01 5.66571057e-01 -4.47974920e-01
3.00635427e-01 -5.32471657e-01 -6.14262998e-01 -1.47593939e+00
-1.15327574e-02 -5.99353015e-01 4.86107022e-01 -9.65887368e-01
-1.11294580e+00 6.85299575e-01 2.59923213e-03 -3.54761600e-01
-5.28879285e-01 -1.08390021e+00 -8.52144122e-01 1.09089291e+00
-1.68127251e+00 -1.36737788e+00 1.07187010e-01 2.05854803e-01
9.64514256e-01 1.43602505e-01 1.38911092e+00 -1.24646381e-01
-5.51406860e-01 4.38126355e-01 -5.11175752e-01 3.59235406e-01
4.35837895e-01 -1.33665848e+00 1.29001582e+00 9.43951070e-01
-3.52387317e-02 8.47355962e-01 3.75244200e-01 -7.73365974e-01
-1.19164276e+00 -9.45007801e-01 1.34070969e+00 -6.84316278e-01
7.35159397e-01 -6.06176496e-01 -9.25160706e-01 1.13348663e+00
2.91724801e-01 -3.16488683e-01 5.65852046e-01 6.01513028e-01
-1.91942811e-01 6.76394284e-01 -8.30700696e-01 2.89349854e-01
1.31370735e+00 -7.44469941e-01 -1.00761127e+00 2.12412015e-01
1.30713677e+00 -3.09251785e-01 -6.02928102e-01 5.74327596e-02
2.13194057e-01 -6.32642686e-01 5.62308073e-01 -1.05424893e+00
6.25113547e-01 1.93451606e-02 -2.84801006e-01 -1.05562639e+00
1.64363518e-01 -6.22203201e-02 8.53434920e-01 1.07123566e+00
8.82447839e-01 -5.37818372e-01 1.13270974e+00 8.59884679e-01
-4.92348522e-01 -1.94546521e-01 -5.54729104e-01 -5.88367164e-01
7.33220458e-01 -9.74030197e-01 4.54094768e-01 1.02966464e+00
2.36884713e-01 5.52287579e-01 3.38580817e-01 6.25369996e-02
2.62870938e-01 2.00708270e-01 2.61293888e-01 -7.17793226e-01
-3.92458439e-01 -1.66137844e-01 1.20183297e-01 -1.12768185e+00
7.79386044e-01 -1.00936794e+00 4.14539874e-01 -1.49770188e+00
1.83903202e-01 -6.50768459e-01 -2.29927003e-02 7.93140054e-01
-3.90349656e-01 1.10787265e-01 2.00733393e-01 -1.67253450e-01
-7.34360456e-01 2.24237040e-01 9.21096444e-01 4.49259132e-01
-6.67749792e-02 -4.14785475e-01 -9.91524279e-01 8.65272701e-01
9.64101970e-01 -5.35982430e-01 -3.24478507e-01 -1.14570093e+00
1.79790422e-01 4.81374897e-02 -5.65541349e-02 -5.60511231e-01
-2.38840103e-01 5.93313947e-02 2.17445195e-01 -2.65409797e-01
-1.68006778e-01 -4.14218783e-01 -5.42968273e-01 3.30800742e-01
-4.33001220e-01 4.08626765e-01 4.41470921e-01 8.97362679e-02
-3.91171694e-01 -5.32348692e-01 4.00521666e-01 -8.40113163e-01
-1.20731807e+00 -4.69216853e-01 -5.95381260e-01 5.70840061e-01
4.00988668e-01 -6.21143021e-02 -3.25309604e-01 -1.50972053e-01
-1.06130469e+00 2.84967870e-01 6.53364480e-01 6.52397454e-01
2.59743333e-01 -7.38875806e-01 -3.55260074e-01 2.29737014e-01
1.28219649e-01 -2.89304112e-03 -1.76563919e-01 9.16041136e-02
-9.72881973e-01 6.33533776e-01 -5.29794246e-02 -3.63674641e-01
-9.26902354e-01 4.86743569e-01 1.24699295e-01 -2.33139500e-01
-7.94757724e-01 8.40282559e-01 4.13290411e-01 -1.00674951e+00
-8.47467631e-02 -8.16705823e-01 -3.38125229e-03 -5.84708810e-01
3.41806024e-01 -2.63374746e-01 9.70796645e-02 -2.63325065e-01
-3.60837817e-01 3.31384212e-01 1.09397575e-01 -2.50993311e-01
1.16505742e+00 -2.37585455e-01 -2.97632486e-01 -3.43018398e-03
1.39322829e+00 -7.56493658e-02 -9.93509054e-01 4.89712693e-02
7.71052122e-01 8.80427137e-02 -3.65555048e-01 -9.58043396e-01
-5.86455762e-01 8.38893831e-01 1.26054808e-01 -9.17886496e-02
9.85515356e-01 5.20546973e-01 9.66335833e-01 8.47674549e-01
6.17854059e-01 -1.47247875e+00 -2.95475274e-01 1.14463651e+00
3.14653873e-01 -1.22378290e+00 -5.76679647e-01 -6.87710464e-01
-8.53739500e-01 1.19089794e+00 7.52462983e-01 -1.27056554e-01
5.37269592e-01 4.69853640e-01 6.05624318e-01 -1.62193537e-01
-7.23510861e-01 -4.49663460e-01 -4.02884662e-01 5.18550217e-01
9.67220724e-01 2.54655838e-01 -6.11065984e-01 1.05958402e+00
-4.78055894e-01 -1.29368976e-01 3.57077032e-01 1.37132132e+00
-3.49944711e-01 -1.69793952e+00 -8.91891271e-02 -8.27720910e-02
-8.81756961e-01 -6.81132853e-01 -6.10875845e-01 9.80943978e-01
-2.09061101e-01 9.95440543e-01 4.48117375e-01 8.51035491e-02
2.42877617e-01 3.89368385e-01 4.66037601e-01 -1.21714294e+00
-5.16515434e-01 2.86503047e-01 9.05829191e-01 -8.76426637e-01
-4.04775143e-01 -6.38069212e-01 -1.98002243e+00 2.98805416e-01
-2.46860124e-02 3.27063471e-01 8.78689349e-01 1.23621082e+00
2.07242638e-01 2.77863443e-01 9.94243845e-02 -5.58510482e-01
-1.91388920e-01 -9.01651025e-01 -2.08230376e-01 4.69040334e-01
-3.41501564e-01 2.65539270e-02 2.57722306e-04 7.11487457e-02] | [10.479972839355469, 9.410812377929688] |
916d5821-f466-4246-975b-ef5c6c79779d | direct-quantification-for-coronary-artery | 1907.10032 | null | https://arxiv.org/abs/1907.10032v3 | https://arxiv.org/pdf/1907.10032v3.pdf | Direct Quantification for Coronary Artery Stenosis Using Multiview Learning | The quantification of the coronary artery stenosis is of significant clinical importance in coronary artery disease diagnosis and intervention treatment. It aims to quantify the morphological indices of the coronary artery lesions such as minimum lumen diameter, reference vessel diameter, lesion length, and these indices are the reference of the interventional stent placement. In this study, we propose a direct multiview quantitative coronary angiography (DMQCA) model as an automatic clinical tool to quantify the coronary artery stenosis from X-ray coronary angiography images. The proposed DMQCA model consists of a multiview module with two attention mechanisms, a key-frame module, and a regression module, to achieve direct accurate multiple-index estimation. The multi-view module comprehensively learns the Spatio-temporal features of coronary arteries through a three-dimensional convolution. The attention mechanisms of each view focus on the subtle feature of the lesion region and capture the important context information. The key-frame module learns the subtle features of the stenosis through successive dilated residual blocks. The regression module finally generates the indices estimation from multiple features. | ['Heye Zhang', 'Shu Zhao', 'Dong Zhang', 'Yanping Zhang', 'Shuo Li', 'Guang Yang'] | 2019-07-20 | null | null | null | null | ['multiview-learning'] | ['computer-vision'] | [-2.39009291e-01 -1.19308777e-01 -3.74809444e-01 -4.54361290e-01
-1.26665008e+00 -3.41274142e-01 2.64971144e-02 5.10720070e-03
-6.11376995e-03 3.49005848e-01 3.89039457e-01 -4.77668405e-01
-1.25156373e-01 -6.95295155e-01 -1.70850590e-01 -7.05469310e-01
-1.10397287e-01 2.82997876e-01 1.87988758e-01 1.71795875e-01
1.70810208e-01 6.43269420e-01 -9.17964339e-01 3.95939201e-01
4.13227797e-01 1.00890243e+00 4.33040738e-01 1.09557927e+00
-6.14478029e-02 1.14374423e+00 -2.23900273e-01 9.11479518e-02
-3.91075574e-02 -5.56519210e-01 -5.53525388e-01 3.09314340e-01
2.76792347e-01 -6.73930228e-01 -3.97054106e-01 9.28187370e-01
9.49692786e-01 -3.30493659e-01 4.18808281e-01 -6.52817190e-01
-5.01602173e-01 3.44431341e-01 -6.98301196e-01 1.00987744e+00
-2.56203830e-01 3.43845397e-01 8.94837439e-01 -8.33542287e-01
6.93600178e-01 1.29353881e+00 4.60978508e-01 1.75644070e-01
-8.83803606e-01 -1.08096316e-01 8.84215385e-02 4.15053487e-01
-1.00387895e+00 -1.63089350e-01 9.34286237e-01 -6.95831895e-01
2.00844288e-01 2.63294429e-01 7.61262417e-01 3.62262666e-01
4.91977215e-01 5.35643995e-01 9.37763929e-01 -1.02908358e-01
-1.12185717e-01 -1.91417083e-01 1.74011633e-01 8.98053586e-01
-1.90737218e-01 3.51133406e-01 1.88357338e-01 -8.07731450e-02
1.33985102e+00 6.06327116e-01 -2.31820628e-01 -3.59623849e-01
-1.40843213e+00 8.91335309e-01 4.76455420e-01 3.23434263e-01
-5.81758380e-01 3.56461883e-01 6.17887080e-01 -3.36920880e-02
2.25536406e-01 8.18049684e-02 -4.21683311e-01 3.24320309e-02
-6.34282529e-01 -2.32886016e-01 1.13426879e-01 3.88724297e-01
5.34413934e-01 -1.21680707e-01 -2.81600684e-01 5.80608130e-01
4.67277497e-01 4.49435234e-01 5.96337557e-01 -1.42550516e+00
4.37140256e-01 7.00678349e-01 -1.08267248e-01 -1.05884504e+00
-3.25317055e-01 -7.59220660e-01 -1.03051090e+00 5.90463042e-01
3.86325508e-01 -1.04558364e-01 -4.04333055e-01 1.33572447e+00
5.77509642e-01 2.64117241e-01 -3.20897311e-01 1.19900787e+00
1.04928267e+00 2.66740352e-01 2.40632862e-01 -3.38257164e-01
1.66187966e+00 -9.99432981e-01 -6.42715752e-01 4.41124514e-02
8.19043577e-01 -8.44950438e-01 9.86128151e-01 -1.44255295e-01
-1.24554658e+00 -7.63181925e-01 -9.30786014e-01 -2.48859331e-01
1.77732229e-01 6.32825375e-01 3.72833282e-01 2.90577948e-01
-5.22602499e-01 6.57738626e-01 -9.45883453e-01 4.29600805e-01
6.88741267e-01 -2.10701935e-02 -3.33954722e-01 -3.73765943e-03
-9.61098373e-01 7.68219411e-01 -2.06314847e-01 1.97083741e-01
-7.62153625e-01 -1.23855126e+00 -7.99947381e-01 9.72515270e-02
4.00824361e-02 -1.12000310e+00 1.07656407e+00 -8.44193816e-01
-1.36206782e+00 1.12775612e+00 -2.98437774e-01 -2.26840705e-01
6.27656758e-01 -3.87889333e-02 -2.36187413e-01 7.91334093e-01
4.87696826e-01 1.16495393e-01 6.79159880e-01 -8.63617361e-01
-6.92990184e-01 -6.16831660e-01 9.72474553e-03 1.22203836e-02
3.85121077e-01 -5.36814425e-03 -3.09251726e-01 -8.16129148e-01
1.87694728e-01 -5.65379918e-01 -7.65727520e-01 5.21556437e-01
-1.65492281e-01 1.07764587e-01 6.91962838e-01 -1.04026103e+00
1.39420414e+00 -2.39380026e+00 1.62736505e-01 3.87839489e-02
8.87625992e-01 2.15040028e-01 2.07226291e-01 -3.20586205e-01
-4.08847183e-01 2.93035984e-01 -3.42940837e-02 2.84353077e-01
-6.80596948e-01 -1.10972270e-01 -1.31305829e-01 5.53026259e-01
8.28980878e-02 1.13484073e+00 -8.65636647e-01 -9.65035260e-01
6.12590373e-01 4.71494466e-01 -3.85487080e-01 2.01669142e-01
1.74094275e-01 8.29914212e-01 -1.19191289e+00 3.41369033e-01
4.64159489e-01 -7.98735380e-01 1.75824955e-01 -5.66579282e-01
-3.08635324e-01 -1.61309838e-01 -1.03550136e+00 1.64874184e+00
-6.54376924e-01 3.89732718e-01 5.00319786e-02 -8.91627967e-01
7.86465287e-01 5.96152842e-01 8.97977412e-01 -7.06388652e-01
3.07020277e-01 2.35981569e-01 1.75016999e-01 -8.66249740e-01
-5.47942936e-01 -5.80161400e-02 2.19276279e-01 4.84464884e-01
-3.26143295e-01 2.05267712e-01 1.10854367e-02 7.62829557e-02
8.37010860e-01 -1.72734354e-02 7.72066116e-01 -1.42949924e-01
1.05064380e+00 -1.51612416e-01 6.59065485e-01 3.16961437e-01
-7.38055289e-01 9.06926751e-01 7.85920918e-01 -1.14512718e+00
-1.31570923e+00 -1.05832267e+00 -1.58310816e-01 2.86862224e-01
3.48184288e-01 -3.09735358e-01 -3.13670784e-01 -7.77759373e-01
-2.14074522e-01 -2.85494048e-02 -7.00939476e-01 -1.09128796e-01
-1.21644676e+00 -4.70855713e-01 -1.10183023e-01 8.72048199e-01
5.16982019e-01 -8.05021167e-01 -8.89825642e-01 3.58086228e-01
-4.07844484e-01 -8.56405854e-01 -7.27731764e-01 -4.71213281e-01
-1.34856975e+00 -1.46328843e+00 -8.14065397e-01 -8.21393788e-01
4.61594582e-01 2.54780829e-01 1.22199261e+00 4.37749892e-01
-9.59777236e-01 -2.08201278e-02 6.42894208e-02 -2.40817629e-02
-4.84516948e-01 -3.31982046e-01 -8.06652606e-01 2.76601076e-01
5.51635353e-03 -8.36519659e-01 -1.37497222e+00 3.56322944e-01
-1.98027149e-01 1.17120832e-01 6.67263210e-01 6.68736339e-01
1.19614255e+00 -4.52766776e-01 6.06290936e-01 -9.65246916e-01
1.01083100e-01 -4.09144819e-01 -4.30733889e-01 2.54155964e-01
-1.50407821e-01 -1.12653665e-01 5.15127301e-01 -1.42368495e-01
-7.37196445e-01 3.01933456e-02 -3.01695466e-01 -7.39720523e-01
-1.62197083e-01 3.24077636e-01 -1.87011942e-01 2.72824407e-01
6.79498553e-01 1.47246182e-01 1.79633766e-01 -3.13943177e-01
5.24475515e-01 5.09564698e-01 4.62910473e-01 -2.37149298e-01
1.58453628e-01 7.63644755e-01 3.20568025e-01 -3.44989449e-01
-1.04830384e+00 -7.69347250e-01 -8.34001660e-01 -4.64682221e-01
1.00899851e+00 -8.28688264e-01 -7.34258473e-01 -2.09755488e-02
-1.16514587e+00 -9.10145696e-03 -5.03346562e-01 7.67848194e-01
-8.48130286e-01 6.04381025e-01 -7.38793671e-01 -1.60983264e-01
-6.57664001e-01 -1.58038127e+00 1.27133167e+00 -1.37015386e-02
2.91438382e-02 -1.18893564e+00 -2.87370756e-02 1.80082589e-01
2.67656952e-01 6.71463549e-01 1.35647559e+00 1.71769056e-02
-8.61438274e-01 -5.51572442e-02 -4.18262154e-01 3.50478232e-01
2.02312022e-01 4.83778305e-03 -5.34208119e-01 1.39850065e-01
1.81935012e-01 2.53072262e-01 6.25128925e-01 1.27265882e+00
1.42346537e+00 2.88750887e-01 -2.73457348e-01 1.00603330e+00
1.50071633e+00 2.10243627e-01 5.57372391e-01 1.86518207e-01
6.93411827e-01 1.86452851e-01 4.44447339e-01 3.59746724e-01
2.06383362e-01 5.58545351e-01 5.30776739e-01 -3.95631582e-01
-7.83458769e-01 1.60754070e-01 -1.43849745e-01 8.82765114e-01
-8.17343369e-02 6.82163060e-01 -7.28594542e-01 4.32052672e-01
-1.74324191e+00 -1.17552054e+00 -3.78006428e-01 1.94124484e+00
5.78959823e-01 -2.04392314e-01 1.07663803e-01 -7.46702123e-03
1.08521533e+00 3.14120770e-01 -7.05397010e-01 -7.44777024e-02
1.16638310e-01 1.31627899e-02 1.75204173e-01 5.62484324e-01
-1.19204295e+00 2.61086673e-01 6.20637894e+00 5.78014791e-01
-1.40229023e+00 1.60843313e-01 1.15972877e+00 3.67489696e-01
-1.81884363e-01 5.54348342e-02 -5.27386427e-01 3.97239834e-01
4.78727698e-01 -1.48472667e-01 -8.68736058e-02 9.42090273e-01
4.93789524e-01 2.93435454e-01 -8.32913995e-01 1.02035809e+00
-2.05105007e-01 -1.86177206e+00 -2.13394575e-02 2.44072247e-02
3.04699421e-01 -9.61516798e-02 1.72739476e-01 -1.82792991e-01
-1.02257513e-01 -6.41027391e-01 3.33435833e-01 9.37745869e-01
1.14365363e+00 -5.22290230e-01 5.38446546e-01 1.74549282e-01
-1.44757545e+00 -1.96884155e-01 -5.24341166e-02 3.89869034e-01
5.35628557e-01 7.17267334e-01 -4.71723139e-01 3.89952302e-01
4.69471961e-01 1.23539519e+00 -5.92222512e-01 1.01379967e+00
-2.03235783e-02 3.60294521e-01 3.64600122e-01 5.85836172e-01
9.41673443e-02 -4.31152791e-01 6.97868347e-01 6.96719944e-01
2.42973343e-01 1.84683919e-01 3.76678020e-01 8.11954558e-01
2.48915270e-01 3.43300045e-01 -3.36617231e-01 5.35402119e-01
1.36888921e-01 1.51296043e+00 -5.81778944e-01 -6.75148606e-01
-6.35966718e-01 4.40874815e-01 1.66914076e-01 3.82255435e-01
-7.47303843e-01 -1.30774081e-01 4.22493041e-01 4.49687421e-01
3.02885175e-01 -4.55206446e-02 -6.43173933e-01 -1.24000037e+00
-8.63758996e-02 -7.95405626e-01 5.67265749e-01 -8.34372520e-01
-1.04351950e+00 5.65394640e-01 -4.76544887e-01 -1.60703659e+00
-2.24013343e-01 -5.03128171e-01 -1.00485623e+00 1.25871003e+00
-1.74993706e+00 -1.07404935e+00 -4.12857085e-01 5.74549556e-01
6.51733637e-01 -1.88034043e-01 8.94975185e-01 5.23400784e-01
-4.98360574e-01 5.70159927e-02 -1.21071868e-01 3.71917725e-01
5.41201532e-01 -1.33870614e+00 1.71642825e-01 5.05199194e-01
-4.01165962e-01 4.21560854e-01 2.51753241e-01 -3.95799190e-01
-8.58486950e-01 -1.31738532e+00 7.46643186e-01 -3.51171225e-01
3.86080891e-01 5.88135302e-01 -5.61550260e-01 5.66251874e-01
-2.83838809e-01 8.22558761e-01 7.63164997e-01 -3.59523505e-01
-1.42224714e-01 -1.60650775e-01 -7.68526614e-01 4.91881818e-01
5.40805757e-01 -4.97709244e-01 -3.52919310e-01 3.57340872e-01
5.68951845e-01 -4.09426153e-01 -1.31613874e+00 3.44864696e-01
6.28344595e-01 -9.47488129e-01 1.48385417e+00 -8.07098508e-01
8.57429266e-01 -3.89654040e-01 4.57861088e-03 -8.18886817e-01
-6.47751391e-01 -2.90569991e-01 -1.71480745e-01 6.90826356e-01
1.33525327e-01 -2.75860280e-01 6.21049583e-01 1.09950423e-01
-2.82720059e-01 -1.22870338e+00 -8.59470069e-01 3.57095823e-02
1.73325315e-01 -1.99607626e-01 3.34528267e-01 5.26047766e-01
-5.10018826e-01 3.34675997e-01 -7.16989264e-02 2.70620644e-01
6.63438857e-01 7.22824097e-01 5.45979440e-01 -1.12568521e+00
-4.32976872e-01 -5.99705696e-01 -5.81907570e-01 -8.17684054e-01
-1.27086163e-01 -1.02066588e+00 -7.78851926e-01 -1.53359842e+00
3.18285704e-01 -2.86190301e-01 -2.87921637e-01 -3.66442353e-01
-4.73220497e-01 2.04765238e-02 1.29130797e-03 3.70617747e-01
-3.34094733e-01 8.75333399e-02 2.17758369e+00 -3.22952829e-02
-1.45236209e-01 3.81140381e-01 -6.21846199e-01 9.64274108e-01
8.04397166e-01 -3.78403068e-01 -3.30697715e-01 -2.25662813e-01
-2.92391293e-02 8.31224978e-01 7.91357040e-01 -6.20595157e-01
-2.70870049e-02 -2.32914202e-02 5.11974573e-01 -6.34011507e-01
7.83403497e-03 -6.34494066e-01 -1.81098908e-01 8.35513473e-01
-5.09638608e-01 2.49863550e-01 -3.80472302e-01 6.32839084e-01
-4.76714253e-01 -1.74320731e-02 1.12335002e+00 -7.36238599e-01
-4.68739033e-01 7.78591871e-01 -3.55114162e-01 4.07872677e-01
1.06277406e+00 3.17786746e-02 -1.66602850e-01 -1.03590451e-01
-1.39316809e+00 2.30534784e-02 -1.66228011e-01 2.18250468e-01
7.85060644e-01 -1.38271689e+00 -1.06370747e+00 2.46927276e-01
-8.83941539e-03 -1.41826021e-02 8.72971475e-01 1.37158597e+00
-9.25844908e-01 2.59626418e-01 -2.23899543e-01 -8.23476732e-01
-1.22861183e+00 7.54308045e-01 1.02935731e+00 -6.46156907e-01
-1.23234558e+00 2.42182657e-01 6.01835907e-01 -8.36158928e-04
-4.73901778e-02 -3.29482526e-01 -6.08090699e-01 -2.11691648e-01
9.43361402e-01 4.52227652e-01 -1.60426661e-01 -8.20474327e-01
-5.46886809e-02 1.26171696e+00 6.44022673e-02 1.72732756e-01
1.13214850e+00 -5.19163787e-01 -1.53915226e-01 3.55269074e-01
1.44284511e+00 -2.21868455e-01 -1.03268301e+00 -4.86091167e-01
-4.16685998e-01 -4.46806103e-01 4.66484725e-01 -5.96735477e-01
-1.46889341e+00 1.55297911e+00 9.74265099e-01 -2.99723774e-01
1.06899691e+00 -5.84785230e-02 9.54462886e-01 -3.25400144e-01
-1.29408449e-01 -4.93910521e-01 1.71200022e-01 -1.22943185e-01
9.61108923e-01 -1.20390904e+00 2.10643008e-01 -6.26252949e-01
-6.24553144e-01 1.66353261e+00 2.84760833e-01 -3.28282893e-01
1.25919795e+00 2.45291051e-02 4.88577336e-01 -4.24456567e-01
-6.30554914e-01 -1.36161566e-01 3.06936651e-01 3.28013480e-01
3.55601311e-01 1.41866297e-01 -3.07124555e-01 6.35006368e-01
4.23321754e-01 1.63840353e-01 1.48683041e-01 4.75665599e-01
-5.52659631e-01 -1.00685930e+00 -5.87531328e-02 3.53073984e-01
-8.92836332e-01 -1.71815485e-01 3.95436466e-01 6.07307494e-01
1.39806464e-01 4.60605592e-01 -1.04766831e-01 2.72996183e-02
4.45255965e-01 -2.03490481e-02 1.82216093e-01 -4.64085251e-01
-6.39477313e-01 3.69587928e-01 -3.16399425e-01 -7.01385856e-01
-3.94789487e-01 -7.74027705e-01 -1.30916560e+00 5.88172898e-02
-3.72843966e-02 -1.37994662e-01 3.91018242e-01 9.07206416e-01
4.31120813e-01 9.86256897e-01 1.13219929e+00 -4.25213635e-01
-4.49525237e-01 -5.84009469e-01 -5.96295178e-01 5.24465144e-01
6.25878155e-01 -4.51183796e-01 -2.29680598e-01 5.11952281e-01] | [14.45499038696289, -2.4046082496643066] |
f32d75ff-53bf-42c4-b008-a7c71be110bd | counterfactual-explanations-can-be | 2106.02666 | null | https://arxiv.org/abs/2106.02666v2 | https://arxiv.org/pdf/2106.02666v2.pdf | Counterfactual Explanations Can Be Manipulated | Counterfactual explanations are emerging as an attractive option for providing recourse to individuals adversely impacted by algorithmic decisions. As they are deployed in critical applications (e.g. law enforcement, financial lending), it becomes important to ensure that we clearly understand the vulnerabilities of these methods and find ways to address them. However, there is little understanding of the vulnerabilities and shortcomings of counterfactual explanations. In this work, we introduce the first framework that describes the vulnerabilities of counterfactual explanations and shows how they can be manipulated. More specifically, we show counterfactual explanations may converge to drastically different counterfactuals under a small perturbation indicating they are not robust. Leveraging this insight, we introduce a novel objective to train seemingly fair models where counterfactual explanations find much lower cost recourse under a slight perturbation. We describe how these models can unfairly provide low-cost recourse for specific subgroups in the data while appearing fair to auditors. We perform experiments on loan and violent crime prediction data sets where certain subgroups achieve up to 20x lower cost recourse under the perturbation. These results raise concerns regarding the dependability of current counterfactual explanation techniques, which we hope will inspire investigations in robust counterfactual explanations. | ['Sameer Singh', 'Himabindu Lakkaraju', 'Sophie Hilgard', 'Dylan Slack'] | 2021-06-04 | null | http://proceedings.neurips.cc/paper/2021/hash/009c434cab57de48a31f6b669e7ba266-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/009c434cab57de48a31f6b669e7ba266-Paper.pdf | neurips-2021-12 | ['counterfactual-explanation', 'crime-prediction'] | ['miscellaneous', 'miscellaneous'] | [ 1.58867076e-01 5.22582054e-01 -5.42924583e-01 -5.40211916e-01
-4.30422336e-01 -4.54953551e-01 7.49406636e-01 1.24083772e-01
-3.75821888e-01 1.18167114e+00 6.86372042e-01 -1.04660869e+00
-4.47754055e-01 -6.65540993e-01 -5.92365801e-01 -4.01626348e-01
-8.11152384e-02 2.30843320e-01 -4.09789890e-01 -2.62234598e-01
8.18865120e-01 4.59182054e-01 -1.49598610e+00 3.49468529e-01
1.01993263e+00 3.83832455e-01 -5.07939279e-01 4.88475084e-01
1.45952091e-01 6.35485649e-01 -4.84808981e-01 -9.22477841e-01
5.18967330e-01 -2.18398958e-01 -8.37502241e-01 -1.36461675e-01
4.37574089e-01 -6.90676033e-01 -6.42948896e-02 9.20944691e-01
5.21493554e-01 2.78821617e-01 7.70729303e-01 -1.58567286e+00
-8.60244870e-01 9.99413490e-01 -6.35974348e-01 6.05294287e-01
4.34663832e-01 5.16943097e-01 1.18147218e+00 -2.73418993e-01
1.68069646e-01 1.56145394e+00 5.21123290e-01 1.05033171e+00
-1.44349301e+00 -1.01120770e+00 4.57606375e-01 1.23017177e-01
-6.46203399e-01 -6.51979506e-01 5.76706052e-01 -2.61958033e-01
9.86061215e-01 7.84735799e-01 3.49499494e-01 9.89726782e-01
1.21862091e-01 4.99135256e-01 1.17668891e+00 -3.49594295e-01
3.88004631e-01 1.90223545e-01 3.61318290e-01 6.22236952e-02
8.77505422e-01 7.42735982e-01 -4.52945441e-01 -7.38673985e-01
2.59000808e-01 5.29153198e-02 -4.55484003e-01 -3.22012067e-01
-6.88351691e-01 1.27902484e+00 3.10397983e-01 -4.34781425e-02
-2.62782574e-01 1.65975958e-01 4.12125498e-01 3.31759274e-01
5.06632507e-01 8.68503451e-01 -5.12793481e-01 -1.53760791e-01
-8.18817854e-01 7.88742602e-01 7.31182873e-01 2.92773873e-01
2.93199807e-01 7.05806911e-02 -2.27587596e-01 4.65435416e-01
2.46330783e-01 3.25619191e-01 4.49410021e-01 -1.09022379e+00
7.89409220e-01 2.65346974e-01 4.85818893e-01 -9.28312004e-01
-3.57308298e-01 -9.72033367e-02 -5.20437419e-01 4.69342828e-01
6.62549973e-01 -2.26899207e-01 -3.49457055e-01 1.78538859e+00
1.98995665e-01 2.41057888e-01 2.23940741e-02 8.21589947e-01
-7.86729455e-02 1.01344727e-01 4.77017283e-01 -5.72338045e-01
1.04225469e+00 -1.94846287e-01 -6.78364754e-01 -3.67158860e-01
8.03058326e-01 -5.59527159e-01 1.16700053e+00 1.15049481e-01
-1.18905497e+00 1.02796219e-01 -8.59353006e-01 4.26859528e-01
-7.79220164e-02 -7.43346214e-01 1.13279152e+00 1.31985462e+00
-6.56850815e-01 1.00750864e+00 -5.74216127e-01 -1.39004469e-01
7.80833125e-01 4.68356818e-01 -1.39667332e-01 1.58668876e-01
-1.14594412e+00 9.94796634e-01 1.59921423e-01 -1.86899036e-01
-4.22141731e-01 -1.12861335e+00 -8.33879411e-01 5.19684315e-01
5.07099569e-01 -5.88170052e-01 1.19951880e+00 -7.79940724e-01
-9.56005037e-01 5.37708223e-01 7.12157786e-02 -9.45127845e-01
7.86588788e-01 -1.53002232e-01 -3.28041673e-01 -2.63724893e-01
4.81543124e-01 2.46342838e-01 5.44443846e-01 -1.01368093e+00
-5.77734113e-01 -5.30666411e-01 3.10590267e-01 1.88752294e-01
-1.81999430e-01 3.93487990e-01 9.34724450e-01 -8.33554447e-01
-2.99544066e-01 -8.68957937e-01 -4.77405459e-01 -2.89767236e-01
-5.66784382e-01 -2.34983843e-02 6.06913030e-01 -2.82735914e-01
1.24876678e+00 -1.82421803e+00 -6.65729403e-01 5.21890111e-02
6.40872493e-02 2.64685363e-01 1.08283870e-01 -7.81224249e-03
-7.32315302e-01 7.67628312e-01 -3.54218274e-01 -4.13118862e-02
3.97738099e-01 -1.67127643e-02 -9.82112050e-01 5.66980422e-01
1.35763824e-01 6.45800889e-01 -8.17251563e-01 2.55713910e-02
1.83095545e-01 -1.25619158e-01 -1.00147116e+00 2.45332763e-01
2.90461749e-01 -8.12852308e-02 -2.66103476e-01 5.04284441e-01
8.61225605e-01 2.43635818e-01 8.31139311e-02 3.03770334e-01
-1.81431890e-01 7.60700285e-01 -1.10855114e+00 6.23679101e-01
-1.51259646e-01 4.83388722e-01 -2.90690064e-01 -1.39663851e+00
3.54842454e-01 3.03797901e-01 -4.00940813e-02 -4.43649977e-01
1.36081250e-02 2.82452017e-01 2.96733290e-01 -3.20849925e-01
5.44030130e-01 -1.13613808e+00 -1.37459844e-01 1.17901635e+00
-6.74602389e-01 -7.75766447e-02 -1.55086577e-01 2.30578437e-01
7.75539637e-01 -4.69000280e-01 7.29384243e-01 -3.57082665e-01
7.41643831e-02 -1.60629496e-01 6.80191159e-01 1.08416975e+00
-5.79469800e-01 5.36840916e-01 6.71534538e-01 -7.29975879e-01
-8.62639308e-01 -9.44032490e-01 -1.62866399e-01 8.44553649e-01
-3.61084580e-01 2.12946907e-01 -4.46959764e-01 -1.15652812e+00
4.39892650e-01 1.58312643e+00 -8.30931127e-01 -6.45275533e-01
-4.38236982e-01 -1.34736896e+00 4.80892450e-01 6.42001092e-01
2.69628048e-01 -8.23693871e-01 -1.01522648e+00 -7.73064420e-02
-6.04123622e-02 -3.91526908e-01 -4.90490913e-01 -2.35240355e-01
-9.41085994e-01 -1.33674192e+00 -4.77386087e-01 4.47435319e-01
4.09738421e-01 4.35067207e-01 1.11058521e+00 3.85697097e-01
-6.86636940e-03 3.62082094e-01 -6.38305321e-02 -7.90590703e-01
-4.84916747e-01 -5.37980497e-01 6.30774021e-01 -2.44083866e-01
6.66267455e-01 -6.20762348e-01 -6.14390075e-01 1.41385868e-01
-7.22931027e-01 -2.50757128e-01 8.06465894e-02 9.12151575e-01
-1.47867158e-01 -1.19521461e-01 9.46108282e-01 -1.19352436e+00
1.06414163e+00 -8.74529660e-01 -5.11732161e-01 2.80232906e-01
-1.03365159e+00 1.88342363e-01 5.80902159e-01 -5.07217765e-01
-1.44470274e+00 -5.11830032e-01 2.34045103e-01 6.28449917e-02
-2.55319923e-01 1.49351224e-01 -1.56837761e-01 2.56662935e-01
8.56174409e-01 -3.69579583e-01 -2.56416291e-01 -3.50859374e-01
4.57055539e-01 7.00647771e-01 2.83341169e-01 -6.54711843e-01
7.40819812e-01 5.95084667e-01 -1.41816542e-01 -2.12397009e-01
-9.43480015e-01 -2.74061412e-02 -4.47904542e-02 6.01258576e-02
5.38200617e-01 -3.29137832e-01 -9.29893076e-01 -1.18681639e-01
-9.47138667e-01 -2.31706977e-01 -4.05477881e-01 6.04676604e-01
-7.34017074e-01 3.98513258e-01 -5.66745475e-02 -1.23171365e+00
-1.00322813e-01 -8.43275130e-01 4.11458105e-01 2.95371711e-01
-7.06890225e-01 -9.07360137e-01 -2.62823030e-02 6.40488982e-01
3.13935250e-01 1.55488327e-01 8.73659670e-01 -1.05151796e+00
-1.84412599e-01 -6.98757023e-02 -1.71373382e-01 -9.04388446e-03
2.84241527e-01 -1.84891634e-02 -1.05327249e+00 -2.56565601e-01
7.81802759e-02 -1.63581029e-01 8.24647844e-01 6.03759408e-01
1.12923753e+00 -1.01605117e+00 -2.61367321e-01 2.96904862e-01
9.77671266e-01 1.98130012e-01 5.29700398e-01 5.19823015e-01
1.83682948e-01 1.02159810e+00 6.96232557e-01 6.81796253e-01
3.15970629e-01 5.75171709e-01 5.12899160e-01 2.17620999e-01
3.39054763e-01 -1.52854010e-01 2.55732983e-01 -3.20156842e-01
-2.35657126e-01 3.51074971e-02 -8.39851618e-01 7.24021971e-01
-1.84072685e+00 -1.76237786e+00 3.97109520e-03 2.61340165e+00
6.29810691e-01 2.35030800e-01 2.11812273e-01 3.47089618e-01
7.53113151e-01 1.75442383e-01 -7.18689203e-01 -1.16904461e+00
1.12098649e-01 -4.15458671e-05 5.59392095e-01 6.02551043e-01
-9.04102445e-01 4.85116631e-01 6.90682888e+00 2.81140983e-01
-7.80007839e-01 -1.57655030e-01 1.27204931e+00 -5.48457026e-01
-7.15526462e-01 2.99523920e-01 -4.43266422e-01 6.64196312e-01
1.11005580e+00 -8.84088993e-01 2.94695973e-01 7.96850681e-01
7.23855913e-01 -6.13201410e-02 -1.31575882e+00 4.11762357e-01
-2.36553371e-01 -1.50730050e+00 2.08795577e-01 2.18378186e-01
7.06296802e-01 -4.96839941e-01 3.25081736e-01 3.34582180e-01
7.25897789e-01 -1.24531841e+00 7.79827237e-01 2.56606013e-01
4.25691962e-01 -9.67444837e-01 7.84068704e-01 4.37598884e-01
-3.66997719e-01 -5.65420270e-01 -5.65960109e-01 -8.92860472e-01
1.54988647e-01 4.02163982e-01 -9.17429805e-01 1.96601510e-01
5.25378287e-01 2.01196536e-01 -1.00694485e-01 8.78248036e-01
-2.82410890e-01 7.66802669e-01 -1.04435068e-02 1.71828240e-01
1.19693771e-01 2.41870522e-01 4.55471873e-01 8.82436037e-01
3.09895337e-01 6.14273489e-01 -5.43669820e-01 1.09414840e+00
7.15684518e-03 -9.15822908e-02 -9.97078359e-01 1.95778340e-01
5.53758442e-01 4.79146630e-01 -5.50732076e-01 -2.35131219e-01
-3.86824578e-01 4.54830110e-01 2.77600229e-01 3.65635246e-01
-6.65156782e-01 6.38525560e-02 1.11722982e+00 2.80310571e-01
-2.71709770e-01 5.22480845e-01 -8.08031678e-01 -1.39899051e+00
-2.00275928e-01 -1.15041947e+00 8.72258365e-01 -5.67974210e-01
-1.34951389e+00 -1.20906346e-01 2.75730163e-01 -9.19342995e-01
-5.15528977e-01 -3.35771739e-01 -1.21757293e+00 9.67445076e-01
-1.47314215e+00 -5.76230824e-01 4.95101631e-01 1.11384198e-01
4.78129268e-01 -1.11584052e-01 6.29044056e-01 -1.76158369e-01
-5.95943511e-01 8.03060710e-01 -2.36068726e-01 -2.73015022e-01
5.85757792e-01 -1.30196726e+00 5.65357149e-01 1.06954002e+00
-1.59369633e-02 1.03157103e+00 1.21084523e+00 -6.46991313e-01
-5.76132715e-01 -7.58815944e-01 1.06602776e+00 -6.08643413e-01
5.96570849e-01 1.70993313e-01 -1.05187452e+00 8.29426169e-01
4.94425371e-03 -1.92753479e-01 8.99907112e-01 4.60722595e-01
-4.03279901e-01 2.77005881e-01 -1.59439373e+00 9.22369778e-01
8.49741101e-01 -2.63869464e-01 -1.23084521e+00 1.00569203e-01
7.13521123e-01 1.70630544e-01 -1.26137152e-01 1.74282640e-01
6.73052073e-01 -1.43144405e+00 1.05010986e+00 -1.33756471e+00
4.86429930e-01 3.10734391e-01 -1.50204495e-01 -1.50773466e+00
-1.32964358e-01 -7.91284621e-01 7.67305419e-02 1.30870771e+00
4.81157243e-01 -9.69761014e-01 8.41644824e-01 1.86546397e+00
2.76391834e-01 -5.74479520e-01 -1.19786763e+00 -7.90366471e-01
5.60746074e-01 -7.99183965e-01 1.22944605e+00 1.39613640e+00
5.64691842e-01 -2.18144372e-01 -5.17183006e-01 1.82573795e-01
8.05264473e-01 2.31438145e-01 5.90884387e-01 -9.42525208e-01
-2.19370574e-01 -6.14812553e-01 -2.20677331e-01 -3.85109127e-01
4.49502558e-01 -6.46616042e-01 -5.93678117e-01 -8.23744833e-01
6.46117508e-01 -3.92578244e-01 -2.38985837e-01 4.83763814e-01
-7.49380827e-01 -5.69952503e-02 3.73635858e-01 -2.03400776e-01
8.86869058e-02 5.93542874e-01 5.52779794e-01 -1.17788166e-01
-1.13773800e-01 4.59423214e-01 -1.67303228e+00 1.03749144e+00
1.18115532e+00 -6.43415272e-01 -4.06164497e-01 -1.24537468e-01
1.28390476e-01 2.18600184e-02 5.88528633e-01 -2.85901546e-01
-6.02712780e-02 -9.60868180e-01 1.05597759e-02 -7.52527639e-02
-9.71170813e-02 -6.04489148e-01 -1.15107603e-01 7.33293116e-01
-6.90612078e-01 -8.08628127e-02 2.65650451e-01 5.71513295e-01
2.39106238e-01 -3.79238576e-01 6.49104655e-01 -2.15912566e-01
-7.91555792e-02 -7.14996979e-02 -3.35668981e-01 1.22363806e-01
1.01819456e+00 -2.63491452e-01 -5.46942055e-01 -6.73592806e-01
-4.93861765e-01 2.43713558e-01 4.51893002e-01 2.89793938e-01
6.24323964e-01 -1.26435184e+00 -8.34669173e-01 -7.01611415e-02
-1.46157816e-01 -7.45361567e-01 3.15430880e-01 4.62152600e-01
8.12070742e-02 6.83936357e-01 -1.59940466e-01 3.63053858e-01
-1.34079838e+00 7.83206880e-01 4.80812669e-01 -2.32509345e-01
-2.70324826e-01 8.08661222e-01 5.81695855e-01 -3.55084121e-01
-1.81129023e-01 -1.60373673e-01 -1.65377989e-01 -1.03632413e-01
8.79700899e-01 8.18496883e-01 -3.59973431e-01 -3.98620546e-01
-3.69260013e-01 -3.31820697e-02 -2.70690560e-01 -9.13761035e-02
1.47780144e+00 -1.80752262e-01 1.02490589e-01 8.62939656e-02
6.75702035e-01 7.02962726e-02 -1.31599581e+00 4.18505073e-01
3.10352474e-01 -1.25430286e+00 -1.87685177e-01 -7.43330538e-01
-8.22345078e-01 8.98649096e-01 3.12253356e-01 4.84420657e-01
1.03369272e+00 -2.36039907e-01 2.72298485e-01 1.77628160e-01
2.90662348e-01 -9.36275959e-01 -4.84753519e-01 -3.20565850e-01
1.07630539e+00 -1.40660870e+00 5.06908894e-01 -3.17640007e-01
-6.19108260e-01 7.31543958e-01 4.22072619e-01 6.38885349e-02
4.18195754e-01 -7.87595510e-02 -4.08197604e-02 -1.51161864e-01
-1.09149659e+00 2.09986016e-01 1.14237815e-01 7.07702816e-01
5.51171184e-01 4.88916129e-01 -7.53811777e-01 8.86562049e-01
-5.54435492e-01 -2.51647353e-01 1.16132569e+00 4.13565367e-01
-1.90429419e-01 -9.46329176e-01 -6.81840837e-01 9.04853880e-01
-9.52100456e-01 -2.53766328e-01 -3.39965582e-01 8.77388418e-01
-1.65309325e-01 1.21660960e+00 -3.80285494e-02 5.45095839e-02
4.17952150e-01 1.93960797e-02 -1.53160572e-01 -6.09146833e-01
-5.70742548e-01 -6.08464658e-01 3.48332822e-01 -6.99597657e-01
-3.94836128e-01 -1.00370514e+00 -1.00658822e+00 -7.78869867e-01
-4.33208138e-01 3.25564206e-01 -4.03953828e-02 1.06852078e+00
1.88163579e-01 6.31486881e-04 6.86620176e-01 -6.40358746e-01
-1.24520659e+00 -6.33248925e-01 -5.71862757e-01 5.41570187e-01
4.50093806e-01 -7.43772209e-01 -9.02316451e-01 -3.70214164e-01] | [8.67993450164795, 5.591487407684326] |
c0441a04-9b6c-4600-90ae-5afcd2a8f0ef | end-to-end-emotion-cause-pair-extraction-with | null | null | https://aclanthology.org/2020.coling-main.17 | https://aclanthology.org/2020.coling-main.17.pdf | End-to-End Emotion-Cause Pair Extraction with Graph Convolutional Network | Emotion-cause pair extraction (ECPE), which aims at simultaneously extracting emotion-cause pairs that express emotions and their corresponding causes in a document, plays a vital role in understanding natural languages. Considering that most emotions usually have few causes mentioned in their contexts, we present a novel end-to-end Pair Graph Convolutional Network (PairGCN) to model pair-level contexts so that to capture the dependency information among local neighborhood candidate pairs. Moreover, in the graphical network, contexts are grouped into three types and each type of contexts is propagated by its own way. Experiments on a benchmark Chinese emotion-cause pair extraction corpus demonstrate the effectiveness of the proposed model. | ['Xiaoqiang Zhang', 'Caicong Wu', 'Shoushan Li', 'Wenjun Hou', 'Ying Chen'] | 2020-12-01 | null | null | null | coling-2020-8 | ['emotion-cause-pair-extraction'] | ['natural-language-processing'] | [-1.01200394e-01 -1.65824831e-01 2.42061969e-02 -7.57467866e-01
-5.06453514e-01 -5.89476883e-01 6.58679247e-01 5.59375346e-01
-9.77678075e-02 6.75750732e-01 6.93079531e-01 -1.99018605e-02
-9.36624929e-02 -7.05515087e-01 -3.94927561e-01 -3.77023309e-01
-1.36212915e-01 -1.16007052e-01 -1.81911811e-01 -5.01969695e-01
1.32595539e-01 2.50999540e-01 -1.25008023e+00 5.34464657e-01
9.73086536e-01 8.16900432e-01 -1.02947220e-01 4.33542997e-01
-5.78601718e-01 1.42893672e+00 -7.73310125e-01 -5.45407534e-01
-5.07378697e-01 -7.46409655e-01 -6.71991825e-01 -2.48984128e-01
6.81511760e-02 5.75974062e-02 -1.92503836e-02 9.94432151e-01
3.09757501e-01 1.21617027e-01 7.23266721e-01 -1.48046196e+00
-5.65868795e-01 9.97559488e-01 -5.63177705e-01 7.46349171e-02
5.54678261e-01 -4.33430105e-01 1.41998518e+00 -8.15718651e-01
6.00826383e-01 1.43607664e+00 3.53736997e-01 4.93419051e-01
-5.36648452e-01 -8.42487752e-01 5.77289402e-01 4.27298188e-01
-1.00735974e+00 2.80797452e-01 1.25623524e+00 1.52682826e-01
1.29313576e+00 2.33790174e-01 6.73195422e-01 9.47789609e-01
3.62908632e-01 1.05987096e+00 7.46053636e-01 -2.97127634e-01
-4.22531329e-02 -2.27984637e-01 6.15777671e-01 4.95058984e-01
-2.35228270e-01 -5.69529533e-01 -5.35367966e-01 -2.10108176e-01
3.08430821e-01 -3.16511065e-01 -2.42103294e-01 2.26467907e-01
-1.09718859e+00 5.60591519e-01 6.51140571e-01 7.21088231e-01
-2.79346883e-01 3.21998894e-01 7.53902674e-01 1.42654553e-01
5.23322999e-01 2.47519851e-01 -6.28869832e-01 -7.03802258e-02
-2.33802527e-01 2.69366503e-01 7.71740615e-01 9.82338250e-01
4.91828859e-01 -3.60978305e-01 -2.84750432e-01 8.90573323e-01
1.72153309e-01 8.04653317e-02 2.53810018e-01 -4.01374623e-02
5.18242300e-01 1.15684593e+00 -9.53157619e-02 -1.62088144e+00
-7.22215652e-01 -4.95827794e-01 -1.11266708e+00 -6.90713823e-01
-1.42170265e-01 -6.06853545e-01 -4.43853527e-01 2.10942936e+00
6.33122444e-01 5.78844011e-01 2.80469269e-01 9.56569850e-01
1.47866356e+00 1.12890434e+00 2.95122683e-01 -3.20665330e-01
1.54796743e+00 -8.39056730e-01 -1.19479573e+00 -3.04043800e-01
6.95712745e-01 -7.35570848e-01 9.25347567e-01 4.04103369e-01
-5.27344286e-01 -3.12350720e-01 -8.36581647e-01 -4.16572630e-01
-5.97977221e-01 2.63058752e-01 7.92119145e-01 -7.79753271e-03
-5.75085223e-01 1.58431441e-01 -2.09499925e-01 -6.96422681e-02
1.00702509e-01 2.43772432e-01 -2.16987550e-01 1.26444817e-01
-1.96192992e+00 7.33665168e-01 3.98492903e-01 5.60345888e-01
-2.21458599e-01 -6.60035670e-01 -1.04800260e+00 2.25955486e-01
3.39919478e-01 -3.63659710e-01 8.24906647e-01 -1.00557232e+00
-9.43199039e-01 4.00246859e-01 -5.26241958e-01 2.04460435e-02
-1.30273014e-01 -4.91257042e-01 -1.00307465e+00 -4.52018753e-02
-1.06176250e-01 3.68142247e-01 4.99221355e-01 -1.29967141e+00
-6.02305532e-01 -1.61565453e-01 -1.75062761e-01 3.40286821e-01
-4.67403442e-01 3.98531735e-01 -7.40563154e-01 -6.20892763e-01
-2.54421711e-01 -4.39150602e-01 -3.10347497e-01 -7.45630443e-01
-6.49561107e-01 -9.14135873e-01 1.07833099e+00 -5.77259123e-01
1.51348245e+00 -2.11263013e+00 1.19245671e-01 2.63384163e-01
3.89982551e-01 -5.66940866e-02 -2.96978801e-01 5.39662540e-01
-3.91130060e-01 1.86823189e-01 -1.49769098e-01 -1.17660873e-01
2.03094453e-01 3.22613358e-01 -3.76935959e-01 -5.73213212e-02
7.02174664e-01 9.30058181e-01 -1.25142145e+00 -6.51779413e-01
1.11967586e-01 7.32484937e-01 -2.37941325e-01 2.89385587e-01
-2.98615694e-01 8.19299594e-02 -5.94042301e-01 3.78199935e-01
7.81302869e-01 -5.62703684e-02 5.68173766e-01 -4.28485066e-01
2.10306272e-01 3.72027844e-01 -1.04702258e+00 1.28940451e+00
-7.38808870e-01 5.24137318e-01 -2.00513601e-01 -5.39751291e-01
1.40653253e+00 3.12899917e-01 3.58100832e-01 -4.75250512e-01
3.64019275e-01 -4.35606530e-03 6.96551614e-03 -5.89637697e-01
7.02512741e-01 -1.02414694e-02 -6.15956128e-01 1.71058849e-01
3.47508825e-02 -9.07336697e-02 3.20953995e-01 6.28329337e-01
1.02296770e+00 -7.10140839e-02 3.69684368e-01 -8.27880576e-02
7.07755864e-01 -1.78868070e-01 1.05604446e+00 1.11595280e-01
-1.91474006e-01 2.11801127e-01 1.25019073e+00 -5.30691147e-01
-3.72183323e-01 -5.73066115e-01 1.67742997e-01 7.89958954e-01
4.42643017e-01 -8.07041764e-01 -4.21034217e-01 -1.12577939e+00
-3.54954928e-01 8.44481826e-01 -8.10549915e-01 -1.65350482e-01
-6.48699641e-01 -6.65700853e-01 3.91887695e-01 6.03459895e-01
5.86142719e-01 -1.27114534e+00 -2.88959026e-01 3.70258480e-01
-5.27893960e-01 -1.26684022e+00 -3.20708036e-01 3.01808387e-01
-3.10489327e-01 -1.18238842e+00 -3.04394692e-01 -1.11470151e+00
6.80152237e-01 -1.90797761e-01 1.50351906e+00 1.11139543e-01
-1.73060432e-01 -7.96183944e-02 -6.54834449e-01 -4.06310499e-01
-1.14356957e-01 1.73623282e-02 -2.67054915e-01 1.09273978e-01
9.61181283e-01 -5.51514745e-01 -2.71219820e-01 6.34799525e-02
-8.41225505e-01 -3.76307666e-02 4.07028764e-01 7.63649523e-01
4.77972299e-01 3.11909318e-01 7.08943367e-01 -9.90525961e-01
1.20702755e+00 -7.21162140e-01 -2.28854269e-01 5.36955655e-01
-1.40399739e-01 -1.81968957e-01 8.98903728e-01 -3.44650537e-01
-1.33995509e+00 1.08943380e-01 -5.51667102e-02 -2.48192236e-01
-4.21757787e-01 8.13687861e-01 -6.52916193e-01 6.82575941e-01
1.97404817e-01 5.53348213e-02 -6.45511985e-01 2.37523858e-03
7.45833993e-01 5.19473970e-01 6.58104002e-01 -5.91972113e-01
3.29706937e-01 -1.34834155e-01 1.00999989e-01 -5.44721723e-01
-8.38084042e-01 -4.83469009e-01 -5.21401465e-01 -3.93445224e-01
9.45515275e-01 -7.88333654e-01 -7.47626722e-01 2.67950267e-01
-1.67420912e+00 1.97522759e-01 2.97210187e-01 2.36709356e-01
2.22020820e-01 1.48399234e-01 -6.85350299e-01 -7.86135137e-01
-4.73864704e-01 -7.88885057e-01 1.05900538e+00 5.84346712e-01
-4.48186159e-01 -1.09169149e+00 -4.96858731e-02 -1.93227366e-01
-1.05238833e-01 6.92770064e-01 1.27374709e+00 -7.13972211e-01
1.18237086e-01 -2.00939924e-01 -3.95673394e-01 1.19001865e-01
4.96837020e-01 3.72349769e-01 -5.84800184e-01 4.39552367e-01
-2.93363661e-01 -2.98810989e-01 7.27983534e-01 4.36845161e-02
1.03584409e+00 -4.55507457e-01 -4.53020930e-01 9.78409424e-02
1.40373981e+00 2.21511006e-01 4.83491570e-01 -2.60155559e-01
9.86208081e-01 7.42801249e-01 7.27624655e-01 5.32384813e-01
4.58953053e-01 3.28298420e-01 5.95420122e-01 -4.68503028e-01
2.09270939e-01 -3.03327888e-01 1.69962302e-01 1.10816145e+00
2.74951190e-01 -5.34819305e-01 -6.78036392e-01 7.09493339e-01
-1.96088862e+00 -7.79020965e-01 -8.52435350e-01 1.31471395e+00
8.47948849e-01 -7.07325786e-02 -7.33652785e-02 -6.11711442e-02
9.95963752e-01 3.54668677e-01 -3.69087934e-01 -7.59469032e-01
-3.72230232e-01 -1.80152189e-02 -2.55864203e-01 3.79382223e-01
-1.01969135e+00 1.08329713e+00 4.94426346e+00 8.21421325e-01
-8.53336632e-01 -3.92162055e-01 9.19766724e-01 1.69807628e-01
-6.93796575e-01 -6.56801462e-02 -5.91336370e-01 2.60379821e-01
2.04374298e-01 3.82212177e-02 3.98164568e-03 8.50541174e-01
1.86245576e-01 6.86497837e-02 -8.00395012e-01 8.66561532e-01
4.96690609e-02 -1.02095222e+00 1.59289375e-01 -4.62392390e-01
5.91211498e-01 -4.92741793e-01 -2.96431690e-01 3.44277501e-01
3.16560388e-01 -6.72786713e-01 3.14761966e-01 4.37557489e-01
3.18966240e-01 -1.64929748e+00 1.06406856e+00 -9.27975550e-02
-1.66961563e+00 1.33107275e-01 -4.83516335e-01 -1.74916744e-01
2.67153442e-01 1.15664566e+00 -7.02510715e-01 9.07179177e-01
6.09293044e-01 1.21904683e+00 -4.74663854e-01 5.55219710e-01
-9.41516101e-01 7.34742820e-01 -1.54167578e-01 -7.77868569e-01
4.30892676e-01 -1.75780967e-01 4.84226525e-01 1.71784914e+00
3.98967750e-02 5.46525896e-01 -1.92520365e-01 1.13271523e+00
-3.31747711e-01 4.72532481e-01 -4.75447595e-01 8.13645497e-02
5.88459373e-01 1.82236803e+00 -8.63389552e-01 -2.08656922e-01
-2.40152761e-01 1.04336369e+00 6.17285132e-01 1.52093709e-01
-1.09095001e+00 -1.00328434e+00 6.77989125e-01 -8.26566577e-01
1.37885138e-01 3.46965529e-02 -1.56527519e-01 -1.15989876e+00
1.73102885e-01 -6.97507381e-01 4.91604000e-01 -9.52406347e-01
-1.68401015e+00 9.16076183e-01 -1.41717687e-01 -1.07719457e+00
-2.14558244e-01 -4.69956100e-01 -1.28443146e+00 8.21555018e-01
-1.42420876e+00 -1.18369162e+00 -3.42155725e-01 7.45549202e-01
3.21108460e-01 3.52308095e-01 7.58201063e-01 1.09082423e-01
-8.74678135e-01 6.08867526e-01 -5.77964067e-01 4.34310645e-01
5.54282844e-01 -1.30766261e+00 2.23116696e-01 1.14039648e+00
2.78072834e-01 9.36641634e-01 5.88056326e-01 -8.56760800e-01
-1.16952133e+00 -1.09228086e+00 1.62284565e+00 -1.58153757e-01
6.78852022e-01 -6.36070907e-01 -8.98561537e-01 2.57972151e-01
7.75222659e-01 9.80588421e-02 6.83314502e-01 5.12879193e-01
-5.25494635e-01 -1.46980479e-01 -9.09523606e-01 7.54870176e-01
8.67653370e-01 -3.55151653e-01 -7.31752992e-01 2.08121091e-01
1.07591224e+00 -3.14486295e-01 -4.48520958e-01 3.99627298e-01
2.09594756e-01 -8.90179098e-01 4.60348010e-01 -6.36330545e-01
1.08297622e+00 -3.41252655e-01 1.85296714e-01 -1.54010415e+00
-2.23360077e-01 -7.40251541e-01 -5.88579848e-02 2.16765094e+00
6.38956606e-01 -1.55262366e-01 2.32133418e-01 7.62056112e-01
-1.17036931e-01 -9.84466553e-01 -5.48998296e-01 -4.19833899e-01
-2.13715881e-01 -6.47610724e-01 1.17282176e+00 1.40179086e+00
4.53503281e-01 8.27520132e-01 -4.28609312e-01 2.55888432e-01
1.48686871e-01 3.88226718e-01 3.72929335e-01 -7.92523384e-01
-5.88105544e-02 -5.26735008e-01 -1.19494177e-01 -7.45072424e-01
6.15400553e-01 -6.65791929e-01 2.71741062e-01 -1.75885212e+00
1.20344587e-01 -3.94769669e-01 -6.51721537e-01 6.73343658e-01
-8.48425210e-01 -1.96515679e-01 -8.20226893e-02 -5.26408792e-01
-6.77338421e-01 7.50058532e-01 1.11943638e+00 -1.37436226e-01
-3.13159913e-01 -5.33087373e-01 -6.55188322e-01 6.07121289e-01
7.43166208e-01 -3.89819741e-01 -5.12683988e-01 -4.12343919e-01
9.12151754e-01 -6.39141276e-02 1.90334573e-01 -3.70167106e-01
1.20766126e-01 -3.41527790e-01 3.46026003e-01 -7.48300374e-01
-1.31443277e-01 -1.02597475e+00 -3.60809714e-01 6.15746342e-03
-5.72447121e-01 2.39044875e-01 3.55905831e-01 4.07124370e-01
-6.38700247e-01 1.08834960e-01 1.52029872e-01 2.46495560e-01
-9.31094348e-01 1.26876077e-03 -2.69194037e-01 9.63894352e-02
8.32940221e-01 4.52669263e-01 -3.27150792e-01 -5.09815574e-01
-3.01173866e-01 4.76634264e-01 -3.12843442e-01 6.84297562e-01
8.52471888e-01 -1.56854606e+00 -7.53131926e-01 -3.42580408e-01
3.68766338e-01 1.45166352e-01 3.75195920e-01 5.92028975e-01
2.69182231e-02 4.32030074e-02 1.96951762e-01 -4.11792248e-02
-1.54367673e+00 5.02199292e-01 2.66441733e-01 -5.89057744e-01
-3.07472527e-01 1.19581079e+00 2.05814987e-01 -4.68994468e-01
1.93786565e-02 -5.82405448e-01 -8.36022854e-01 8.80611315e-02
5.42999029e-01 -7.70190917e-03 -1.21627301e-01 -7.30581760e-01
-7.25167871e-01 4.34479147e-01 -7.90897831e-02 1.58064723e-01
1.25412035e+00 2.98926532e-02 -8.51346612e-01 3.40582997e-01
1.30134439e+00 1.36938065e-01 -6.70140624e-01 -4.86497022e-03
-1.37420353e-02 -2.77606584e-02 1.10616051e-01 -9.99217391e-01
-1.23615313e+00 7.36047864e-01 5.95606677e-03 3.83594222e-02
1.53972685e+00 2.52400469e-02 1.02764690e+00 2.83697516e-01
-2.43804529e-02 -9.61822748e-01 -1.43882066e-01 7.48574853e-01
1.05850160e+00 -8.81845534e-01 -2.50578433e-01 -8.68100762e-01
-7.96807528e-01 1.33308744e+00 1.01708531e+00 -1.11726128e-01
5.14142156e-01 4.01700616e-01 1.92674801e-01 -4.47684944e-01
-9.42766726e-01 -1.80112571e-01 5.79297423e-01 3.99743974e-01
7.44742095e-01 9.41371620e-02 -6.72051072e-01 1.28950155e+00
-5.75792082e-02 -2.09293768e-01 2.78467923e-01 7.33547270e-01
6.56314613e-03 -1.16230488e+00 -5.31940944e-02 2.41528034e-01
-3.59139651e-01 -4.48528439e-01 -1.23619807e+00 7.35663950e-01
4.77439195e-01 1.43257034e+00 7.01577738e-02 -6.77477121e-01
2.85447687e-01 -1.03833541e-01 6.63797790e-03 -2.40931615e-01
-1.15298140e+00 2.42460966e-01 3.53517771e-01 -5.17894745e-01
-5.35964489e-01 -3.01399052e-01 -1.96641326e+00 1.32745057e-01
-2.86604524e-01 3.56182665e-01 3.29185188e-01 1.16303504e+00
4.80221778e-01 9.56950903e-01 9.99458015e-01 -4.93157767e-02
4.87754583e-01 -9.18415368e-01 -6.31829560e-01 6.51571035e-01
7.96250626e-02 -2.80688971e-01 -1.55503616e-01 -2.57181585e-01] | [12.623649597167969, 6.216344356536865] |
a790bec4-3fc2-491e-9fbd-2322d2abb45a | ernie-search-bridging-cross-encoder-with-dual | 2205.09153 | null | https://arxiv.org/abs/2205.09153v1 | https://arxiv.org/pdf/2205.09153v1.pdf | ERNIE-Search: Bridging Cross-Encoder with Dual-Encoder via Self On-the-fly Distillation for Dense Passage Retrieval | Neural retrievers based on pre-trained language models (PLMs), such as dual-encoders, have achieved promising performance on the task of open-domain question answering (QA). Their effectiveness can further reach new state-of-the-arts by incorporating cross-architecture knowledge distillation. However, most of the existing studies just directly apply conventional distillation methods. They fail to consider the particular situation where the teacher and student have different structures. In this paper, we propose a novel distillation method that significantly advances cross-architecture distillation for dual-encoders. Our method 1) introduces a self on-the-fly distillation method that can effectively distill late interaction (i.e., ColBERT) to vanilla dual-encoder, and 2) incorporates a cascade distillation process to further improve the performance with a cross-encoder teacher. Extensive experiments are conducted to validate that our proposed solution outperforms strong baselines and establish a new state-of-the-art on open-domain QA benchmarks. | ['Haifeng Wang', 'Dawei Yin', 'Shuaiqiang Wang', 'Hua Wu', 'Hao Tian', 'Shikun Feng Yu Sun', 'Zhengjie Huang', 'Yunsheng Shi', 'Jiaxiang Liu', 'Yiding Liu', 'Yuxiang Lu'] | 2022-05-18 | null | null | null | null | ['passage-retrieval'] | ['natural-language-processing'] | [-7.96800256e-02 2.76018173e-01 4.34962213e-02 -4.56557870e-01
-1.38786721e+00 -6.52905703e-01 6.44797087e-01 1.59033202e-02
-5.91272712e-01 6.64031148e-01 3.15747082e-01 -7.75577664e-01
2.17370629e-01 -9.05582130e-01 -1.16306400e+00 -2.65944809e-01
4.79935229e-01 1.06582236e+00 2.25207910e-01 -7.16677904e-01
-3.92972618e-01 -2.06124410e-01 -1.12182629e+00 5.82821250e-01
1.37272954e+00 8.00675571e-01 1.86358035e-01 6.04330122e-01
-3.23665053e-01 1.23258674e+00 -6.42557621e-01 -8.98749590e-01
-3.06941476e-02 -4.53508377e-01 -1.28431952e+00 -7.34690011e-01
8.52858961e-01 -8.76080990e-01 -7.01857269e-01 6.37027323e-01
7.81695962e-01 3.03470016e-01 6.81143165e-01 -5.70199013e-01
-1.65584671e+00 8.08502316e-01 -2.76551098e-01 3.89809385e-02
2.93567210e-01 1.45362213e-01 1.33488536e+00 -8.25837851e-01
2.61041909e-01 1.52484190e+00 2.87677556e-01 8.30587685e-01
-1.28448915e+00 -7.56454110e-01 1.15726553e-01 6.63949728e-01
-1.27770770e+00 -2.71333754e-01 7.48420298e-01 -1.11998782e-01
1.11599398e+00 -1.51565537e-01 9.62767974e-02 1.27678120e+00
-1.98820472e-01 1.35486424e+00 9.49523091e-01 -5.47446966e-01
-1.34599134e-01 -8.46530795e-02 2.63133287e-01 8.14987421e-01
9.27690230e-03 -1.71750844e-01 -8.42811763e-02 -1.07068762e-01
3.27954173e-01 -2.11329147e-01 -4.33241069e-01 -2.10529462e-01
-9.26473975e-01 1.08187044e+00 8.09502482e-01 1.65096536e-01
-1.97252735e-01 1.70233235e-01 3.91997367e-01 6.98629618e-01
4.32776332e-01 6.85437799e-01 -6.62172735e-01 1.04308957e-02
-9.83717740e-01 4.56857681e-01 8.44328582e-01 7.23961771e-01
7.65374243e-01 -7.13805854e-02 -7.39068866e-01 1.15477860e+00
3.45380634e-01 4.92397875e-01 7.51555979e-01 -9.19187307e-01
8.13636124e-01 6.60056710e-01 -1.32992625e-01 -2.95173049e-01
1.98943123e-01 -5.11847377e-01 -6.15100324e-01 -3.10733408e-01
2.60071874e-01 -3.23601663e-01 -1.12583292e+00 1.70491004e+00
2.24064574e-01 4.89309311e-01 5.76977313e-01 8.46374631e-01
1.47363877e+00 9.09988105e-01 1.43772334e-01 4.00351405e-01
1.60614073e+00 -1.74461412e+00 -8.93925726e-01 -3.38442475e-01
7.99644887e-01 -5.88159144e-01 1.33451331e+00 2.62465656e-01
-1.28723300e+00 -7.50340223e-01 -1.12243450e+00 -9.79930401e-01
-4.48235959e-01 1.71307772e-01 3.37371618e-01 1.91376626e-01
-1.02888966e+00 4.02056873e-01 -6.79817259e-01 1.30277425e-01
2.69600451e-01 1.02550261e-01 -6.08093031e-02 -5.49047828e-01
-1.69738901e+00 1.22210801e+00 4.16441172e-01 -9.39841866e-02
-1.37778032e+00 -7.88577616e-01 -9.96110439e-01 3.31936389e-01
3.72160375e-01 -1.29173350e+00 1.84449399e+00 -5.65706670e-01
-1.86860263e+00 8.68149281e-01 -1.43411651e-01 -6.73416555e-01
1.69042975e-01 -8.60539615e-01 -1.71490207e-01 1.68449789e-01
-6.26450107e-02 8.69387388e-01 5.35302699e-01 -1.12741148e+00
-3.18397373e-01 -2.45552093e-01 6.63381934e-01 4.11251336e-01
-3.26411366e-01 -3.05681854e-01 -5.03949285e-01 -5.23383021e-01
-3.93806875e-01 -5.39816976e-01 -1.57474965e-01 -3.25015575e-01
-1.01326838e-01 -8.86996150e-01 6.95114255e-01 -1.01343131e+00
1.33696687e+00 -1.82639515e+00 3.20511788e-01 -6.22302294e-01
2.98242778e-01 8.48983526e-01 -6.08822763e-01 4.54666406e-01
2.17522159e-01 -2.65654683e-01 -3.10394913e-01 -6.91258371e-01
1.21718757e-01 5.30726075e-01 -7.16319263e-01 -7.87968487e-02
4.25972909e-01 1.43871760e+00 -1.05674660e+00 -4.75982487e-01
-3.44204605e-02 5.61834574e-01 -8.32197845e-01 8.44866991e-01
-7.08513618e-01 2.04722509e-01 -3.52297902e-01 2.91940719e-01
5.19333124e-01 -4.35732484e-01 -1.82149813e-01 -1.39147863e-01
3.65622401e-01 1.14034986e+00 -5.48129261e-01 2.49262834e+00
-8.85219991e-01 4.83047903e-01 4.15994152e-02 -1.02624500e+00
8.97987604e-01 4.91141558e-01 -3.67203727e-02 -9.81516361e-01
1.13198563e-01 3.64859134e-01 -1.68149900e-02 -4.43557471e-01
7.69095600e-01 3.96676026e-02 6.56579211e-02 4.87616181e-01
6.62505269e-01 -2.64887571e-01 5.15306666e-02 4.55627441e-01
1.13484514e+00 4.33962643e-01 -6.09584339e-02 1.02589905e-01
6.72156572e-01 -1.15722597e-01 6.22473322e-02 7.82614470e-01
-1.47624305e-02 4.01627034e-01 5.08636087e-02 2.40917206e-02
-5.99980116e-01 -1.29553556e+00 -1.37298396e-02 1.28166962e+00
1.44122988e-01 -3.34622383e-01 -8.98048699e-01 -9.52006578e-01
4.08018157e-02 9.52497244e-01 -4.53740805e-01 -3.80782366e-01
-7.19203651e-01 -2.27506995e-01 1.17592371e+00 4.46961373e-01
8.42721701e-01 -7.10101843e-01 -2.96070874e-01 3.49209398e-01
-6.72169507e-01 -1.26278794e+00 -2.22325519e-01 7.30336234e-02
-9.61854041e-01 -7.18815267e-01 -1.02780652e+00 -9.87084210e-01
3.15810084e-01 3.50641012e-01 1.65228045e+00 -1.05743743e-01
1.73086762e-01 4.22873467e-01 -4.91220415e-01 -1.58365548e-01
-3.12308103e-01 3.85191798e-01 -3.75046879e-01 -3.07181329e-01
6.48814321e-01 -5.16142964e-01 -7.58024335e-01 7.05438526e-03
-8.51158977e-01 5.81362396e-02 7.21449554e-01 1.17377150e+00
5.61916232e-01 -4.88114893e-01 7.92894423e-01 -8.99450660e-01
1.08144426e+00 -6.37824595e-01 -3.54236573e-01 7.91161597e-01
-5.79591632e-01 7.08216906e-01 5.56302309e-01 -3.93700182e-01
-1.37113333e+00 -5.93420446e-01 -4.77771431e-01 -6.28196299e-01
-1.07560702e-01 6.13505065e-01 -2.93339919e-02 1.58040076e-01
6.92920923e-01 9.93203446e-02 -3.05752188e-01 -1.00637043e+00
1.15677369e+00 8.94135892e-01 7.67623603e-01 -9.54545915e-01
6.02484822e-01 6.32398650e-02 -6.47520423e-01 -1.94983095e-01
-1.25668561e+00 -5.19837976e-01 -2.57136852e-01 3.84433985e-01
8.10038805e-01 -1.54762793e+00 -6.33060575e-01 2.11329564e-01
-1.59948528e+00 -5.01188040e-01 -2.71080106e-01 2.84614950e-01
-2.48558223e-01 3.17146301e-01 -1.03832948e+00 -4.25797313e-01
-7.54580736e-01 -1.19457471e+00 1.02980757e+00 3.22432548e-01
3.36957276e-02 -1.13151979e+00 4.50733781e-01 1.16123486e+00
6.46848321e-01 -4.95516747e-01 9.03912306e-01 -8.50236833e-01
-6.01429522e-01 1.90159976e-01 -3.55332345e-01 6.60487115e-01
-1.61525130e-01 -6.15554690e-01 -1.27714264e+00 -2.22868785e-01
-3.18466276e-02 -9.86966431e-01 1.10036147e+00 -1.17815621e-01
1.15843070e+00 -4.73627269e-01 -1.08106844e-01 5.29111028e-01
1.21210110e+00 -8.58783200e-02 7.02539027e-01 2.46733159e-01
7.21774042e-01 4.34814960e-01 3.01046580e-01 -2.01020032e-01
1.08894396e+00 6.99092150e-01 4.84851867e-01 -1.23228505e-01
-5.61313748e-01 -6.84783876e-01 5.32195151e-01 1.28363800e+00
1.72979414e-01 -3.11498970e-01 -7.95194209e-01 9.15221632e-01
-1.68541825e+00 -5.28146207e-01 1.00539550e-02 1.76843131e+00
1.44113326e+00 -2.57195413e-01 -4.31107014e-01 -4.46525782e-01
2.19192326e-01 2.47569785e-01 -4.95244920e-01 -5.18413544e-01
9.64368731e-02 9.95892227e-01 1.44200444e-01 7.48595476e-01
-8.11448574e-01 1.19354606e+00 5.41684055e+00 9.54237223e-01
-7.43213356e-01 5.81127584e-01 3.89448434e-01 1.08229533e-01
-6.21016026e-01 -6.67334022e-03 -1.01951969e+00 1.45182922e-01
1.27583289e+00 9.71488953e-02 3.79826933e-01 8.67011607e-01
-5.58312893e-01 2.36811146e-01 -1.22702551e+00 8.26224864e-01
3.11461478e-01 -1.15389502e+00 4.66139376e-01 -3.29149514e-01
9.08465207e-01 3.66588026e-01 1.73792630e-01 1.25927591e+00
8.73769879e-01 -1.10051966e+00 1.73736393e-01 3.19858611e-01
7.26196170e-01 -4.76274908e-01 7.62281775e-01 4.87900466e-01
-7.72600055e-01 -2.42582685e-03 -2.70340145e-01 -1.61349714e-01
3.81609887e-01 2.94568598e-01 -9.19624388e-01 9.78081584e-01
5.47634602e-01 4.82651591e-01 -5.98855317e-01 6.72065020e-01
-8.82159889e-01 9.34236526e-01 -1.32258996e-01 2.78137829e-02
5.37046611e-01 5.76172723e-04 7.28509799e-02 1.03615332e+00
9.09609795e-02 1.83191195e-01 -1.44210711e-01 9.21115875e-01
-6.17766976e-01 6.16377918e-03 -4.22120333e-01 -1.95263460e-01
3.77231807e-01 8.37643802e-01 4.93262172e-01 -6.43377542e-01
-8.46654236e-01 1.29652119e+00 9.04584527e-01 5.33708811e-01
-7.24322975e-01 -6.56184137e-01 7.74687350e-01 -5.72547689e-02
4.43297267e-01 -3.42529505e-01 5.74982353e-02 -1.61122596e+00
-2.71947496e-02 -1.16587269e+00 6.91427171e-01 -8.49595129e-01
-1.51109338e+00 6.79912567e-01 -6.29802719e-02 -7.19284534e-01
-5.87421775e-01 -5.12714446e-01 -3.94730777e-01 1.05019331e+00
-2.21558952e+00 -1.30321050e+00 -6.40205294e-02 7.64980793e-01
6.33713484e-01 -1.02565713e-01 1.17687452e+00 7.21240878e-01
-2.51549125e-01 1.02425838e+00 4.14579719e-01 3.71209472e-01
1.00844741e+00 -1.19115615e+00 4.73069251e-01 6.77147448e-01
5.04916489e-01 7.46851325e-01 3.16567391e-01 -3.01982522e-01
-1.27855551e+00 -8.24734092e-01 1.15866005e+00 -8.04823697e-01
7.42110789e-01 -3.11066628e-01 -1.23273098e+00 9.37964678e-01
7.99151242e-01 -6.44143447e-02 6.92563653e-01 3.59703243e-01
-7.17004538e-01 -1.49927065e-01 -7.35802710e-01 5.61167300e-01
5.92326880e-01 -8.60582471e-01 -1.42125046e+00 2.16797516e-01
1.32286978e+00 -7.36425519e-01 -7.74954319e-01 6.81470871e-01
2.73317784e-01 -5.45403421e-01 1.25634301e+00 -9.02315021e-01
7.52092719e-01 -4.22158241e-02 -5.63215930e-03 -1.45560741e+00
-4.37637344e-02 -4.56843972e-01 -9.11530793e-01 1.20497835e+00
3.73270363e-01 -5.09622037e-01 3.92327130e-01 4.06627148e-01
-3.96823198e-01 -1.01019955e+00 -9.86974239e-01 -6.32632017e-01
6.08971894e-01 -7.00347647e-02 6.43582046e-01 8.79350424e-01
-1.90860569e-01 1.07634163e+00 -2.27603599e-01 4.09347340e-02
2.87449509e-01 9.71207991e-02 8.74957204e-01 -9.82963145e-01
-7.66054928e-01 -2.65859246e-01 1.52626351e-01 -2.15988755e+00
3.22551608e-01 -1.03941643e+00 -8.02610908e-03 -1.97932029e+00
-7.29064867e-02 -3.53551328e-01 -4.00908411e-01 5.18478453e-01
-6.20734155e-01 -1.00716799e-01 -7.73003995e-02 -7.12193847e-02
-8.79000545e-01 1.15117502e+00 1.53434229e+00 -4.66734529e-01
9.55424309e-02 -3.60250980e-01 -8.06411624e-01 2.52795190e-01
4.55026776e-01 -4.07810718e-01 -7.13366985e-01 -1.43952405e+00
5.17230928e-02 1.35331869e-01 2.82779396e-01 -7.66447902e-01
2.55959451e-01 3.90554488e-01 -2.46589378e-01 -3.25061679e-01
6.35102808e-01 -6.43224776e-01 -7.21659780e-01 1.68543741e-01
-5.48397183e-01 4.53748181e-02 4.12172884e-01 4.57984149e-01
-6.52079284e-01 -4.88227129e-01 4.26154822e-01 -2.17121512e-01
-4.72866446e-01 1.24451607e-01 1.35799438e-01 6.48448467e-01
5.27280688e-01 4.97752517e-01 -7.42588162e-01 -5.68214059e-01
-3.66355002e-01 7.51789927e-01 -1.36797234e-01 5.13878286e-01
3.99953097e-01 -1.28500426e+00 -1.02692413e+00 -4.59621027e-02
1.18861213e-01 5.36814034e-01 3.34742725e-01 3.53119671e-01
-4.15511578e-01 8.24104786e-01 1.37787178e-01 -3.42526019e-01
-1.05291104e+00 3.27359438e-01 3.87782931e-01 -1.02437222e+00
-3.42311174e-01 1.42664969e+00 3.37110966e-01 -8.18938673e-01
3.75129640e-01 -3.63710195e-01 -2.51361102e-01 -5.94999455e-02
5.46612084e-01 1.49993792e-01 2.76011884e-01 -2.55802959e-01
7.46717975e-02 2.43539706e-01 -5.93592405e-01 -1.63888201e-01
1.13306022e+00 -1.35820881e-02 3.47937010e-02 2.12692261e-01
1.22050619e+00 -4.18371350e-01 -1.12850869e+00 -8.06178153e-01
-3.21031958e-01 -8.35607499e-02 1.88109994e-01 -1.16986954e+00
-8.32674086e-01 1.62150872e+00 3.97091120e-01 -3.40016603e-01
1.03793168e+00 1.58433720e-01 1.40908897e+00 8.77260566e-01
8.69326666e-02 -7.60307252e-01 4.46690738e-01 9.90833282e-01
7.90342987e-01 -1.29254997e+00 -2.54180759e-01 1.45240888e-01
-5.39978981e-01 7.75909543e-01 1.04787540e+00 -8.07668194e-02
3.45462024e-01 -2.24450082e-01 2.42301092e-01 -1.95058227e-01
-1.11266863e+00 -4.24831241e-01 5.21786809e-01 3.25372189e-01
7.72872984e-01 7.17631653e-02 -3.56209904e-01 9.34643984e-01
-2.58357495e-01 8.64470974e-02 1.39026687e-01 8.14123452e-01
-2.84369498e-01 -1.28766024e+00 -6.47973921e-03 1.06252633e-01
-5.32881379e-01 -5.83771050e-01 -2.48846978e-01 4.77572262e-01
2.37519518e-02 1.05922234e+00 -1.32391537e-02 -2.26978198e-01
4.52030808e-01 4.16674078e-01 6.96771264e-01 -7.56868720e-01
-8.62690985e-01 -5.69115102e-01 2.02527866e-01 -2.57251114e-01
-1.72952101e-01 2.35243645e-02 -1.09121311e+00 3.85402888e-02
-5.36993265e-01 6.34030163e-01 5.57922304e-01 1.14082134e+00
6.81346953e-01 7.91025519e-01 -1.50867524e-02 -9.66157988e-02
-9.70859289e-01 -1.35193181e+00 1.35139152e-01 2.85185784e-01
5.10717034e-01 -6.72033250e-01 -1.10695355e-01 -3.54848057e-02] | [11.241779327392578, 8.015488624572754] |
dd7c7a63-1df3-4c9e-b90d-5e417fdc9c6d | cross-domain-medical-image-translation-by | 2007.07230 | null | https://arxiv.org/abs/2007.07230v1 | https://arxiv.org/pdf/2007.07230v1.pdf | Cross-Domain Medical Image Translation by Shared Latent Gaussian Mixture Model | Current deep learning based segmentation models often generalize poorly between domains due to insufficient training data. In real-world clinical applications, cross-domain image analysis tools are in high demand since medical images from different domains are often needed to achieve a precise diagnosis. An important example in radiology is generalizing from non-contrast CT to contrast enhanced CTs. Contrast enhanced CT scans at different phases are used to enhance certain pathologies or organs. Many existing cross-domain image-to-image translation models have been shown to improve cross-domain segmentation of large organs. However, such models lack the ability to preserve fine structures during the translation process, which is significant for many clinical applications, such as segmenting small calcified plaques in the aorta and pelvic arteries. In order to preserve fine structures during medical image translation, we propose a patch-based model using shared latent variables from a Gaussian mixture model. We compare our image translation framework to several state-of-the-art methods on cross-domain image translation and show our model does a better job preserving fine structures. The superior performance of our model is verified by performing two tasks with the translated images - detection and segmentation of aortic plaques and pancreas segmentation. We expect the utility of our framework will extend to other problems beyond segmentation due to the improved quality of the generated images and enhanced ability to preserve small structures. | ['Yu-Xing Tang', 'Sung-Won Lee', 'You-Bao Tang', 'Yingying Zhu', 'Perry J. Pickhardt', 'Ronald M. Summers', 'Daniel C. Elton'] | 2020-07-14 | null | null | null | null | ['pancreas-segmentation'] | ['medical'] | [ 3.37685972e-01 3.03533673e-01 -8.11112151e-02 -4.12952751e-01
-1.03327239e+00 -6.07387960e-01 3.72025788e-01 1.33218810e-01
-2.23546699e-01 6.20129883e-01 5.86080551e-02 -2.15641543e-01
-1.06090471e-01 -8.87638092e-01 -8.99746239e-01 -7.68257320e-01
-5.09840138e-02 9.83991981e-01 4.58167315e-01 -1.56469420e-01
-2.68754333e-01 5.66439927e-01 -7.33409524e-01 5.69912732e-01
9.93518233e-01 5.76402426e-01 3.45950693e-01 5.89932680e-01
-2.71657526e-01 3.93548459e-01 -4.07019794e-01 -3.24351370e-01
6.45346940e-01 -9.17123139e-01 -9.69311595e-01 3.00312877e-01
4.19037700e-01 -2.25680411e-01 7.33164251e-02 1.14507854e+00
4.83896941e-01 -1.63839445e-01 7.86208212e-01 -9.26033854e-01
-6.69277310e-01 6.14101529e-01 -6.77044630e-01 -1.66121982e-02
-1.38148695e-01 1.42939137e-02 2.82165557e-01 -3.34802389e-01
8.75684321e-01 1.07552743e+00 7.92801499e-01 5.92019558e-01
-1.49862313e+00 -5.07924676e-01 -1.55484766e-01 -2.04734743e-01
-8.71795356e-01 9.44552571e-02 7.14261174e-01 -5.35183907e-01
2.66553044e-01 2.74881721e-01 5.51607728e-01 8.73279691e-01
5.71739733e-01 6.82280958e-01 1.42936325e+00 -3.50095719e-01
1.81794576e-02 2.31624082e-01 -3.35518718e-01 8.49772274e-01
1.37683660e-01 5.05299270e-02 4.84752818e-04 3.36678550e-02
1.13338184e+00 6.27665743e-02 -3.49105626e-01 -7.55848348e-01
-1.49392819e+00 9.04792547e-01 6.30097210e-01 7.36016512e-01
-4.46770608e-01 -3.61158662e-02 5.01431942e-01 2.71278411e-01
5.20180464e-01 4.52006012e-01 -4.03991491e-01 2.18061358e-01
-1.31689298e+00 1.34693608e-01 7.09930241e-01 8.25273871e-01
2.72997677e-01 -1.61721006e-01 -2.24713400e-01 8.25997651e-01
6.18809685e-02 2.29469866e-01 7.64631391e-01 -1.01130855e+00
1.95955321e-01 4.41136628e-01 -2.41831675e-01 -6.10821784e-01
-5.60516357e-01 -5.86244941e-01 -1.12645638e+00 4.65124160e-01
6.74005568e-01 8.57614800e-02 -1.24241221e+00 1.60651481e+00
4.07595575e-01 1.71526060e-01 4.22166772e-02 1.07720375e+00
8.45468044e-01 3.59955311e-01 1.69537887e-01 -1.70412734e-01
1.50093257e+00 -1.12294638e+00 -5.45195162e-01 -9.29064974e-02
6.58888578e-01 -9.95220780e-01 9.51105058e-01 2.74936438e-01
-1.46450484e+00 -5.04760206e-01 -7.94259787e-01 -1.25810757e-01
-1.13624208e-01 -1.23646423e-01 2.90817082e-01 6.71749592e-01
-9.32007611e-01 6.83853209e-01 -1.19154334e+00 -4.22814906e-01
7.42981851e-01 3.46974701e-01 -5.39841771e-01 -2.41450533e-01
-9.52437639e-01 1.10472775e+00 2.74440050e-01 -3.98297071e-01
-6.56954229e-01 -1.11509609e+00 -7.74300158e-01 -2.95461323e-02
2.41382867e-01 -1.13700473e+00 1.12453294e+00 -1.14969993e+00
-1.33110642e+00 1.25556540e+00 2.40009978e-01 -6.68383896e-01
1.00624943e+00 2.53846020e-01 -1.61602989e-01 4.36743855e-01
2.50937372e-01 9.91222739e-01 8.21630895e-01 -1.33058262e+00
-3.77742857e-01 -3.60603809e-01 -2.76132643e-01 7.80371428e-02
1.89266652e-01 -1.14742547e-01 -4.36581075e-01 -9.45609391e-01
2.66768456e-01 -1.23308277e+00 -5.72534263e-01 4.45230901e-01
-1.67670831e-01 5.73490977e-01 8.50215018e-01 -1.03709829e+00
6.54474199e-01 -1.94524324e+00 3.24864566e-01 5.24362661e-02
3.18824917e-01 1.33807406e-01 -1.27455503e-01 -2.80516148e-01
-3.10483456e-01 -5.20761572e-02 -7.45167017e-01 -2.94259667e-01
-1.68981701e-01 4.71609861e-01 5.61514944e-02 4.90737677e-01
-1.36324521e-02 1.15212238e+00 -7.15774357e-01 -9.06757772e-01
3.08366120e-01 5.35356820e-01 -4.78385508e-01 -7.37680346e-02
-1.57707736e-01 1.15515339e+00 -2.91599035e-01 3.15390795e-01
8.65186870e-01 -4.29676116e-01 1.15145952e-01 -2.86906421e-01
2.58953899e-01 -6.55599460e-02 -8.23711097e-01 2.02014208e+00
-4.64186609e-01 2.47974470e-01 2.24922717e-01 -1.21717215e+00
5.35829246e-01 4.60322738e-01 1.08594728e+00 -9.20187950e-01
1.24775998e-01 3.68528068e-01 2.72150755e-01 -4.04754460e-01
8.68431702e-02 -7.35186040e-01 9.07168351e-03 4.80734795e-01
-2.31692027e-02 -5.96356988e-01 1.52265102e-01 1.04990661e-01
6.28413081e-01 2.06371948e-01 2.40430221e-01 -4.15157706e-01
4.55120146e-01 4.18142259e-01 4.49757487e-01 4.37183946e-01
-3.31900448e-01 1.14622033e+00 3.44632328e-01 -4.57318962e-01
-1.34411979e+00 -1.13712370e+00 -3.59597892e-01 6.83195829e-01
6.75835758e-02 2.16922164e-01 -9.96776283e-01 -9.00792837e-01
-2.74474382e-01 5.25573850e-01 -6.24677718e-01 -1.37409255e-01
-8.47697198e-01 -8.52443337e-01 6.00092947e-01 5.35954237e-01
4.81247962e-01 -9.75805819e-01 -7.75382102e-01 4.94196832e-01
-4.68294293e-01 -1.20263088e+00 -6.46354496e-01 1.68529555e-01
-1.36653602e+00 -9.90444005e-01 -1.52289474e+00 -1.14352059e+00
7.27818191e-01 3.04128435e-02 1.35890329e+00 -6.36692792e-02
-5.20172417e-01 3.64361405e-01 -1.32778242e-01 -3.20331693e-01
-1.14832497e+00 -2.22636876e-03 -2.75447965e-01 -3.41190994e-01
-2.90675730e-01 -4.47067916e-01 -6.81517780e-01 4.17391062e-01
-1.35573769e+00 3.49428028e-01 6.49399221e-01 9.93591309e-01
8.45266759e-01 -8.04917440e-02 1.72507480e-01 -1.08393621e+00
5.10697961e-01 -2.52510875e-01 -5.01460314e-01 3.50844890e-01
-3.92835706e-01 2.32843921e-01 4.16631877e-01 -6.29470050e-01
-1.05068600e+00 1.49952605e-01 -1.48900509e-01 -4.41838235e-01
-1.86706200e-01 3.49264205e-01 2.33384505e-01 -2.28249818e-01
7.78068721e-01 2.65416086e-01 4.81457442e-01 -3.87789756e-01
3.76690209e-01 1.87527314e-02 7.44586825e-01 -5.89492738e-01
5.91589689e-01 6.80253267e-01 3.77217561e-01 -4.06297415e-01
-5.09819627e-01 -2.96318412e-01 -7.20826626e-01 1.43071666e-01
1.12982500e+00 -6.14747345e-01 -1.50758743e-01 2.41708770e-01
-1.05668664e+00 -4.26563263e-01 -6.09010816e-01 5.54395735e-01
-8.64717305e-01 5.07556677e-01 -9.71180975e-01 1.24057032e-01
-4.66113240e-01 -1.72244489e+00 1.20386279e+00 9.20687690e-02
5.67288436e-02 -1.43659437e+00 1.18607365e-01 2.46754751e-01
6.00101054e-01 5.11556625e-01 1.19050229e+00 -4.39038992e-01
-5.43941796e-01 2.80896991e-01 -7.81077072e-02 3.69510293e-01
2.69511372e-01 -5.22903979e-01 -3.91770095e-01 -2.61696965e-01
2.24882826e-01 -8.46876763e-03 6.72287285e-01 9.67957318e-01
1.06750989e+00 1.12798102e-01 -4.80276793e-01 7.65451074e-01
1.28659225e+00 1.49016708e-01 6.82312369e-01 3.10466617e-01
5.30434072e-01 7.22806215e-01 5.09039462e-01 -5.81501313e-02
-2.74827518e-02 8.34919989e-01 3.69741589e-01 -8.73630285e-01
-6.31724834e-01 1.52737767e-01 -3.25783379e-02 6.34740114e-01
1.02471620e-01 1.24247529e-01 -9.37488675e-01 7.22121775e-01
-1.58122540e+00 -6.75445855e-01 -4.02615905e-01 1.92985082e+00
9.99171019e-01 -1.73679784e-01 9.96112451e-02 -2.62530833e-01
5.60219109e-01 -3.44002903e-01 -4.50727254e-01 -2.14094430e-01
-5.68751991e-02 5.35425901e-01 6.14418566e-01 4.34572130e-01
-1.18110943e+00 7.29864836e-01 6.19057655e+00 7.92143404e-01
-1.46016014e+00 6.16901338e-01 8.23203981e-01 3.15734744e-01
-8.08481947e-02 -2.31407627e-01 -1.23833582e-01 3.41403902e-01
5.35963655e-01 6.59982935e-02 3.64995040e-02 6.28165901e-01
5.71390614e-02 -8.75155553e-02 -1.04068792e+00 8.17430615e-01
-7.95529783e-02 -1.49888849e+00 -1.65841028e-01 1.14127621e-01
1.01042366e+00 -3.15624066e-02 2.25472942e-01 4.04300764e-02
2.51606584e-01 -9.87375617e-01 4.83707249e-01 3.76973718e-01
8.90451789e-01 -4.31433409e-01 6.08596981e-01 2.57168084e-01
-8.51790071e-01 4.83410984e-01 -2.67084181e-01 7.11658120e-01
3.84264439e-01 6.46167040e-01 -1.20589185e+00 6.89282596e-01
5.46365023e-01 3.35396618e-01 -4.68693137e-01 1.14045095e+00
8.80321488e-02 2.84314752e-01 -2.39609167e-01 7.83144355e-01
3.66536647e-01 -4.53629464e-01 5.95776975e-01 1.17488885e+00
5.32126606e-01 -1.04730159e-01 1.83242008e-01 1.08673799e+00
6.20402880e-02 2.00820118e-01 -5.00489950e-01 4.19486851e-01
-3.29134703e-01 1.15175414e+00 -1.23252010e+00 -6.79013193e-01
-3.73285681e-01 1.33505917e+00 -3.19790483e-01 1.28075972e-01
-1.12649751e+00 2.22639889e-01 2.72040844e-01 3.41477782e-01
2.51605988e-01 -1.09463692e-01 -5.75079322e-01 -1.09827435e+00
-2.50178307e-01 -9.94344711e-01 5.11847496e-01 -6.50281489e-01
-1.26898694e+00 7.37345755e-01 8.67178440e-02 -1.52608132e+00
-2.34343946e-01 -4.18519706e-01 -2.08842471e-01 8.90532851e-01
-1.47497439e+00 -1.62717545e+00 -2.78253317e-01 6.79603279e-01
5.24178624e-01 -1.40916901e-02 7.85970688e-01 6.31218970e-01
3.10033858e-01 3.82817149e-01 2.84262627e-01 1.98395804e-01
1.12805879e+00 -1.29386783e+00 2.38057837e-01 7.09472597e-01
8.19268376e-02 3.73943508e-01 6.46895349e-01 -7.07714975e-01
-7.63215482e-01 -1.21667731e+00 4.51280773e-01 -2.80098647e-01
1.60186931e-01 1.26228839e-01 -1.06152594e+00 7.86841393e-01
2.95958489e-01 3.58116448e-01 6.15426302e-01 -5.53653777e-01
-2.61412337e-02 6.97827637e-02 -1.52919948e+00 4.57959950e-01
5.88802278e-01 2.05468219e-02 -5.96986115e-01 4.57787663e-01
5.67499161e-01 -8.90932739e-01 -1.11116576e+00 3.31854671e-01
5.10227442e-01 -8.07056904e-01 1.29125881e+00 -5.80038786e-01
6.64509356e-01 -2.34806418e-01 1.25525147e-01 -1.46562457e+00
-2.73882329e-01 -2.14482442e-01 3.73320103e-01 8.43587875e-01
2.68356264e-01 -5.62615514e-01 9.05743897e-01 5.71834862e-01
-2.45430082e-01 -3.02010953e-01 -9.25362825e-01 -6.76217914e-01
5.86994529e-01 -2.55639106e-01 4.05950934e-01 1.25407457e+00
-4.78473365e-01 -6.42129853e-02 -2.35862255e-01 8.02155808e-02
6.74834788e-01 4.79614317e-01 5.87917984e-01 -1.08966887e+00
-5.58743060e-01 -5.55089831e-01 -2.54140884e-01 -7.88426280e-01
-1.15279794e-01 -1.17827284e+00 -8.90713036e-02 -1.66499591e+00
3.87270957e-01 -7.34983206e-01 -1.35178551e-01 3.76701742e-01
-3.44037339e-02 6.40422821e-01 4.98934954e-01 2.20734492e-01
-3.13123405e-01 1.78532794e-01 2.07543373e+00 -3.21213901e-01
-1.56208530e-01 1.95755035e-01 -5.65110505e-01 6.84135020e-01
6.41565442e-01 -8.24429691e-01 -2.04072595e-01 -3.67036968e-01
-3.95261765e-01 3.99819702e-01 4.17727739e-01 -8.58342648e-01
3.95235457e-02 8.72959271e-02 4.45575714e-01 -2.50142157e-01
1.87861055e-01 -1.06020725e+00 5.08221388e-01 9.94143546e-01
-1.80187836e-01 -7.00896606e-02 2.50075430e-01 1.73335880e-01
-4.15155083e-01 -1.43170580e-01 1.34883749e+00 -6.29703820e-01
-3.14026117e-01 3.21458817e-01 -2.75582612e-01 -1.88054387e-02
1.15545833e+00 -2.01284900e-01 9.71617699e-02 -3.75507563e-01
-1.23816693e+00 -9.37701017e-02 5.29302537e-01 3.42064202e-01
5.00896931e-01 -1.17189956e+00 -8.75470340e-01 1.55797228e-01
-9.92969796e-02 2.73241699e-01 4.15383905e-01 1.45341134e+00
-9.37982380e-01 4.00033236e-01 -5.91075480e-01 -1.03414392e+00
-1.48789883e+00 6.53667927e-01 6.45866215e-01 -8.61920238e-01
-9.12845433e-01 6.02980494e-01 7.32758224e-01 -6.14637733e-01
-2.33585566e-01 -8.67796957e-01 2.32900247e-01 -2.94629365e-01
2.02216372e-01 9.14738700e-02 3.05736274e-01 -6.20481908e-01
-2.04472825e-01 7.91668892e-01 -8.34313855e-02 -1.20308772e-01
1.33691871e+00 -6.94273338e-02 -2.91792862e-02 -1.18163191e-01
1.15580058e+00 -1.20349854e-01 -1.22363985e+00 -3.51154029e-01
-2.05757707e-01 -3.32337797e-01 7.86070824e-02 -9.74848092e-01
-1.38525391e+00 9.71861184e-01 1.07597101e+00 -1.42733827e-01
1.35785437e+00 1.54810041e-01 9.93782818e-01 -4.94216859e-01
4.45303082e-01 -7.25984275e-01 3.48982066e-02 1.15267202e-01
8.50467861e-01 -1.39428067e+00 1.00670718e-01 -5.88911474e-01
-9.66427624e-01 1.17187357e+00 1.33781374e-01 -4.07191887e-02
4.75707382e-01 6.86169565e-01 3.75405252e-01 -7.71507993e-03
-1.34342715e-01 2.91353427e-02 5.07409751e-01 6.97067559e-01
4.52497065e-01 1.85684144e-01 -5.71081221e-01 3.16158175e-01
4.08949815e-02 1.17804259e-01 4.22712684e-01 9.19209778e-01
-1.94688607e-02 -1.57056224e+00 -8.23956490e-01 2.57440448e-01
-8.31297934e-01 5.90439439e-02 9.36892107e-02 9.02744412e-01
2.58670717e-01 2.81378567e-01 -1.83781356e-01 4.55565721e-01
2.28919834e-01 -3.76927182e-02 8.53872657e-01 -5.31693697e-01
-8.79362285e-01 5.50649285e-01 -4.25389051e-01 -5.24781287e-01
-6.38823152e-01 -6.37094378e-01 -1.42584932e+00 -3.52575001e-03
6.68028966e-02 -1.73586145e-01 8.74429286e-01 8.32944393e-01
2.05484509e-01 9.74573970e-01 -3.15134339e-02 -7.63248026e-01
-3.59303743e-01 -7.44490743e-01 -6.08039856e-01 8.67711186e-01
2.53570586e-01 -4.84986722e-01 2.86670715e-01 6.32433951e-01] | [14.357365608215332, -2.2397756576538086] |
49e90bae-1975-45ca-9577-fb65edc2131a | physics-coupled-spatio-temporal-active | 2108.05385 | null | https://arxiv.org/abs/2108.05385v1 | https://arxiv.org/pdf/2108.05385v1.pdf | Physics-Coupled Spatio-Temporal Active Learning for Dynamical Systems | Spatio-temporal forecasting is of great importance in a wide range of dynamical systems applications from atmospheric science, to recent COVID-19 spread modeling. These applications rely on accurate predictions of spatio-temporal structured data reflecting real-world phenomena. A stunning characteristic is that the dynamical system is not only driven by some physics laws but also impacted by the localized factor in spatial and temporal regions. One of the major challenges is to infer the underlying causes, which generate the perceived data stream and propagate the involved causal dynamics through the distributed observing units. Another challenge is that the success of machine learning based predictive models requires massive annotated data for model training. However, the acquisition of high-quality annotated data is objectively manual and tedious as it needs a considerable amount of human intervention, making it infeasible in fields that require high levels of expertise. To tackle these challenges, we advocate a spatio-temporal physics-coupled neural networks (ST-PCNN) model to learn the underlying physics of the dynamical system and further couple the learned physics to assist the learning of the recurring dynamics. To deal with data-acquisition constraints, an active learning mechanism with Kriging for actively acquiring the most informative data is proposed for ST-PCNN training in a partially observable environment. Our experiments on both synthetic and real-world datasets exhibit that the proposed ST-PCNN with active learning converges to near optimal accuracy with substantially fewer instances. | ['Laurent Cherubin', 'Hanqi Zhuang', 'Ali Muhamed Ali', 'Min Shi', 'Xingquan Zhu', 'Yufei Tang', 'Yu Huang'] | 2021-08-11 | null | null | null | null | ['spatio-temporal-forecasting'] | ['time-series'] | [ 1.45239606e-01 -5.45423441e-02 -9.96238738e-02 -1.74062252e-01
-1.77659646e-01 -4.78521109e-01 9.14517045e-01 3.57213855e-01
5.43735549e-02 9.42575574e-01 1.21028282e-01 -4.43220168e-01
-8.03677619e-01 -9.27297473e-01 -7.72472918e-01 -1.09869659e+00
-7.07283020e-01 4.70460385e-01 4.35282707e-01 -3.68028045e-01
1.16712935e-01 6.10162020e-01 -1.50572765e+00 -7.60877803e-02
1.16551340e+00 9.93273437e-01 4.91137236e-01 7.74263620e-01
-1.37869656e-01 1.08680820e+00 -2.52009064e-01 2.31466219e-01
1.73242345e-01 -3.32360864e-01 -4.06620562e-01 -4.97652739e-02
-1.98469803e-01 -1.38656674e-02 -2.32677490e-01 8.10396135e-01
2.58247484e-03 6.16949618e-01 5.36144555e-01 -1.17302001e+00
-3.63066554e-01 2.74122268e-01 -3.38871032e-01 5.90934992e-01
-1.29233420e-01 3.26169580e-01 7.44841993e-01 -6.42800748e-01
2.96480596e-01 1.06543183e+00 5.74045420e-01 6.47667125e-02
-1.09711516e+00 -5.25074482e-01 3.47430348e-01 1.77217275e-01
-1.13694310e+00 -1.75585568e-01 9.29571688e-01 -8.07621360e-01
7.10591376e-01 2.03688398e-01 8.05167854e-01 1.02716148e+00
4.56851035e-01 1.93776786e-01 1.03699696e+00 -8.19466412e-02
6.08807743e-01 1.43658563e-01 2.56522801e-02 4.55216736e-01
-2.82366178e-03 5.34840107e-01 -7.22221315e-01 -3.73152077e-01
7.67394185e-01 3.08726519e-01 -3.14244926e-01 -5.23703136e-02
-1.02909601e+00 6.98262453e-01 6.50579453e-01 2.36010164e-01
-8.38030398e-01 2.30481382e-02 -2.54264306e-02 -1.43168373e-02
9.98369217e-01 5.55818379e-01 -7.96283305e-01 -3.54638509e-02
-9.77599859e-01 3.53044242e-01 7.44432926e-01 5.22051215e-01
7.45782018e-01 1.17655791e-01 3.62058818e-01 4.28517073e-01
4.94925767e-01 8.15483212e-01 7.40150688e-03 -7.64508903e-01
4.35066611e-01 8.22373033e-01 4.65938419e-01 -1.39394438e+00
-2.92572170e-01 -6.64659500e-01 -1.36876309e+00 2.21443683e-01
3.61641616e-01 -3.80570263e-01 -7.66369998e-01 1.44991386e+00
6.54048026e-01 8.43618155e-01 -2.15918750e-01 9.88039672e-01
3.75136942e-01 1.53047764e+00 -4.68864851e-02 -6.44674122e-01
9.04175460e-01 -5.94873428e-01 -7.35833704e-01 -9.26368758e-02
9.71322432e-02 -3.11093599e-01 7.22258806e-01 2.64391661e-01
-6.74265027e-01 -6.37602925e-01 -6.92470789e-01 4.29894298e-01
-5.21958113e-01 -3.05686980e-01 7.00218618e-01 -1.88211232e-01
-7.27307200e-01 7.04129636e-01 -1.11786759e+00 -2.19118118e-01
6.56292364e-02 1.32545143e-01 -6.03622720e-02 3.59424412e-01
-1.48089337e+00 5.19341767e-01 6.14979088e-01 6.34965897e-01
-1.05250919e+00 -1.12718296e+00 -4.52890486e-01 8.25201795e-02
4.47230399e-01 -3.93187553e-01 9.34412479e-01 -7.06173897e-01
-1.34691978e+00 -5.03637232e-02 -1.69182763e-01 -3.28422874e-01
3.27885360e-01 -1.39624089e-01 -6.21652186e-01 -3.61753181e-02
-1.70169726e-01 5.59034161e-02 9.10889506e-01 -1.27981055e+00
-6.79666340e-01 -9.38530937e-02 -2.22929437e-02 1.69027612e-01
-2.98099905e-01 -2.63626724e-01 -7.32485997e-03 -4.94527400e-01
6.35163859e-02 -9.28420484e-01 -4.51550663e-01 -1.25365168e-01
-1.85669139e-01 -3.24683398e-01 1.09311283e+00 -5.45049489e-01
1.15534401e+00 -1.87289369e+00 2.44978219e-01 1.50973111e-01
4.59693104e-01 3.33572000e-01 2.20669523e-01 8.05198610e-01
1.11712776e-01 -1.24966994e-01 -3.23689163e-01 -1.78424016e-01
-3.80846053e-01 5.16564667e-01 -1.18704498e+00 2.82313913e-01
4.71962571e-01 5.72804213e-01 -1.17660236e+00 -2.12688714e-01
2.88985342e-01 4.37332004e-01 -3.11873436e-01 6.51025891e-01
-5.30211747e-01 1.25316119e+00 -6.53205812e-01 7.25730807e-02
4.85254139e-01 -6.90043390e-01 -1.71507701e-01 3.40644956e-01
-5.38841546e-01 1.70831382e-01 -1.27058959e+00 1.19915223e+00
-4.32460189e-01 5.18431664e-01 6.28429046e-03 -1.25658095e+00
9.19143975e-01 5.57836652e-01 6.06083930e-01 -5.06208599e-01
-2.22232312e-01 1.93430617e-01 7.68024176e-02 -8.95423770e-01
5.12406170e-01 -2.14925990e-01 1.64909691e-01 4.61602211e-01
-1.81033626e-01 8.28692596e-03 -1.38718948e-01 2.32098863e-01
1.06941116e+00 1.19556397e-01 2.29359850e-01 -4.35740918e-01
5.64059556e-01 3.79914552e-01 9.25491571e-01 5.73417187e-01
3.90777998e-02 7.53100738e-02 1.92062169e-01 -8.41699243e-01
-1.01709759e+00 -8.36723745e-01 -1.35850266e-01 7.99888194e-01
7.52931014e-02 2.03112122e-02 -9.42502394e-02 -1.92199588e-01
-1.06153622e-01 6.28534436e-01 -6.87430918e-01 -7.53997564e-02
-5.71959019e-01 -1.10775030e+00 7.39180967e-02 1.73253402e-01
3.02712083e-01 -1.17962301e+00 -6.64135396e-01 4.63880777e-01
-8.13565478e-02 -9.73977506e-01 2.46327356e-01 1.13161922e-01
-9.81249928e-01 -8.20238471e-01 -3.28644782e-01 -1.73189908e-01
6.89485133e-01 1.26752391e-01 9.03702855e-01 -1.43206388e-01
-1.71986267e-01 -4.50612307e-02 -2.07397759e-01 -5.37290573e-01
-4.10186619e-01 3.00780553e-02 2.27988586e-01 3.36704373e-01
-8.62527266e-02 -9.84447122e-01 -5.97173929e-01 2.14617312e-01
-8.82999122e-01 2.94344038e-01 3.48768204e-01 6.80058777e-01
4.58250880e-01 6.55391693e-01 5.27087688e-01 -6.63125694e-01
4.02194440e-01 -1.03741050e+00 -9.65688705e-01 5.56724295e-02
-4.93876487e-01 1.83894783e-02 9.54217076e-01 -6.72280550e-01
-1.34817445e+00 4.76018861e-02 3.26927602e-01 -5.24165094e-01
-3.11725229e-01 8.75676572e-01 4.18723494e-01 4.18665290e-01
6.17455721e-01 4.65036958e-01 -3.21657687e-01 -3.32388848e-01
1.06521912e-01 3.31765443e-01 3.59900862e-01 -5.00262439e-01
1.14797044e+00 6.12970293e-01 2.62629628e-01 -9.85639870e-01
-1.01207364e+00 -4.64048564e-01 -7.29227662e-01 -5.28886795e-01
7.60233343e-01 -8.23669672e-01 -6.25161588e-01 4.24185961e-01
-1.28641450e+00 -5.82985222e-01 -2.67745495e-01 7.98991024e-01
-1.38197631e-01 -2.62130946e-01 -2.81065077e-01 -1.31120932e+00
-1.08466096e-01 -7.03979611e-01 8.24436903e-01 3.18174005e-01
-1.16007820e-01 -1.42690766e+00 7.12788641e-01 8.39343818e-04
5.58573782e-01 3.91269207e-01 8.94457519e-01 -3.26774448e-01
-1.08519471e+00 -6.30702451e-02 -1.24500304e-01 -1.45359218e-01
3.23311806e-01 3.32193166e-01 -1.01366305e+00 -6.96793571e-03
2.54658699e-01 -1.05306312e-01 5.45666993e-01 5.05320013e-01
9.34637666e-01 -5.50569654e-01 -3.36434901e-01 4.74213392e-01
1.29093266e+00 4.58338886e-01 1.18022755e-01 -3.08106124e-01
7.85561800e-01 9.10162389e-01 4.51954573e-01 5.09797096e-01
3.45684469e-01 3.30453217e-01 5.01931548e-01 -2.88420022e-01
3.61789554e-01 -2.56899148e-01 1.19465329e-01 1.27975655e+00
-4.79007453e-01 -3.92822683e-01 -1.32712162e+00 6.39335692e-01
-2.17995310e+00 -1.23796856e+00 -4.84428406e-01 2.05897880e+00
8.00607681e-01 -2.99647097e-02 -2.41632491e-01 9.90278050e-02
5.48997998e-01 2.94709742e-01 -7.87477851e-01 7.16094673e-02
1.08704500e-01 -5.27273975e-02 2.38639563e-01 5.88050544e-01
-1.06346369e+00 5.76307297e-01 5.45473146e+00 5.00977874e-01
-1.50746715e+00 2.09619641e-01 4.56230968e-01 1.13944292e-01
-8.23669508e-02 1.56059891e-01 -7.01876998e-01 5.74859500e-01
1.18195677e+00 -1.07579477e-01 4.90310550e-01 5.42667985e-01
1.02004945e+00 -1.38197139e-01 -7.28437304e-01 5.12427747e-01
-6.14280343e-01 -1.51799703e+00 -2.23537564e-01 2.69875765e-01
9.88017976e-01 2.19913408e-01 -1.04024775e-01 1.10257454e-01
5.29923260e-01 -8.03700924e-01 4.59790617e-01 1.13301539e+00
2.30557859e-01 -3.79948616e-01 4.24741060e-01 9.67936337e-01
-1.33982408e+00 -6.95006028e-02 -3.80940437e-01 -5.70052266e-01
4.32257473e-01 1.10111570e+00 -7.00471997e-01 5.15105009e-01
6.24158323e-01 9.92813408e-01 -8.87178332e-02 9.31007683e-01
-2.09811807e-01 1.16819656e+00 -4.93581086e-01 -1.04176655e-01
1.83507740e-01 -3.79875034e-01 8.38040769e-01 8.06808770e-01
4.92171288e-01 4.85568404e-01 3.09739769e-01 1.09788001e+00
2.91432619e-01 -2.07176208e-01 -8.20244789e-01 -2.78941363e-01
3.79799396e-01 1.28746831e+00 -7.10469723e-01 -2.87690014e-01
-7.48152137e-02 4.40194517e-01 2.82016814e-01 6.14514768e-01
-8.69955599e-01 1.70460105e-01 5.46405971e-01 2.63474941e-01
4.60080020e-02 -6.02398396e-01 -1.59595281e-01 -1.07403886e+00
-1.36313722e-01 -3.12466890e-01 1.03449136e-01 -7.49898493e-01
-1.52930844e+00 6.02081299e-01 2.07782209e-01 -1.39357483e+00
-3.51758033e-01 -2.93441236e-01 -9.58427012e-01 1.01616502e+00
-1.50345278e+00 -8.85362566e-01 -5.19395411e-01 6.11172378e-01
3.28389674e-01 -6.96785003e-02 8.17107320e-01 1.16581202e-01
-7.24330604e-01 -5.19626915e-01 2.07690448e-01 -1.43506616e-01
3.35109234e-01 -1.17214906e+00 2.91404486e-01 1.01300538e+00
2.26084958e-03 6.76740408e-01 1.08995402e+00 -8.67830276e-01
-1.46274662e+00 -1.39233232e+00 9.65823293e-01 -3.55735749e-01
1.19081664e+00 -4.48255122e-01 -1.28513825e+00 2.74775535e-01
1.22778371e-01 4.73071635e-01 3.19397539e-01 8.36741254e-02
-9.21947416e-03 -3.53307962e-01 -6.65889800e-01 2.09887713e-01
9.36814666e-01 -5.25771976e-01 -6.19889379e-01 5.55715561e-01
6.98660195e-01 -2.13908464e-01 -7.07555652e-01 4.18387622e-01
1.37284502e-01 -6.76394641e-01 6.75484538e-01 -7.56978989e-01
7.23794043e-01 -4.38856751e-01 1.71232566e-01 -1.60582602e+00
-4.54018563e-01 -5.77855408e-01 -4.48415041e-01 1.05262411e+00
4.45555508e-01 -8.35548341e-01 4.07906115e-01 5.76605260e-01
-1.78635530e-02 -7.04839349e-01 -9.16808248e-01 -5.49651504e-01
-5.49995713e-02 -6.11219823e-01 5.29437542e-01 1.35405171e+00
-2.62957573e-01 3.56759816e-01 -5.26602924e-01 9.01382744e-01
7.39405334e-01 2.61578947e-01 6.74151838e-01 -1.62270176e+00
-4.49076474e-01 -8.66642520e-02 -1.01229265e-01 -7.10881114e-01
-5.04929870e-02 -3.44906479e-01 1.84209898e-01 -1.32760942e+00
-1.33176163e-01 -7.91577578e-01 -3.81163716e-01 2.01301351e-01
-3.28375816e-01 -2.19937995e-01 -1.57621622e-01 5.63511610e-01
-1.80447742e-01 8.40699077e-01 1.17781401e+00 7.23164752e-02
-4.87206638e-01 2.29172483e-01 4.65267487e-02 6.30310178e-01
6.16807103e-01 -5.81486344e-01 -9.42413092e-01 -3.20442945e-01
6.91842079e-01 3.70326877e-01 7.32069254e-01 -9.43377614e-01
6.01814508e-01 -8.84584904e-01 5.54088056e-02 -8.01288843e-01
3.13840121e-01 -1.00360370e+00 4.75293756e-01 3.22375536e-01
-2.64288306e-01 8.62392709e-02 1.38511389e-01 7.91981339e-01
-3.94333541e-01 2.45657638e-01 4.37313229e-01 -1.60406277e-01
-6.66289926e-01 6.74530268e-01 -5.21535695e-01 5.16140945e-02
9.61990118e-01 1.70283303e-01 -3.24305266e-01 -4.25478876e-01
-5.08816600e-01 3.10821146e-01 -5.45407385e-02 4.02095169e-01
2.89458305e-01 -1.03962409e+00 -6.21918559e-01 2.89555281e-01
-1.40471442e-03 3.78140330e-01 5.01946092e-01 8.50212991e-01
-2.68219262e-01 5.72165072e-01 1.73819721e-01 -8.93159449e-01
-5.45665979e-01 4.19705510e-01 3.73999447e-01 -3.60216409e-01
-8.37154269e-01 6.22500360e-01 3.90118241e-01 -5.89257479e-01
-1.12642586e-01 -5.17491639e-01 -1.84395805e-01 1.04776993e-01
5.38981497e-01 4.12395775e-01 -9.34931263e-02 -4.21353579e-01
-7.52599835e-02 3.51665437e-01 4.54573631e-01 -1.20287180e-01
1.71101880e+00 -1.27085179e-01 -2.18766764e-01 1.07510698e+00
7.95654416e-01 -2.37881348e-01 -1.60254121e+00 -5.66142678e-01
-5.43480068e-02 -1.51951090e-01 3.82993221e-01 -6.12113178e-01
-9.63943243e-01 1.06116211e+00 3.63637775e-01 8.16331029e-01
1.09352958e+00 -1.53645203e-01 6.87658310e-01 4.71613616e-01
4.73889470e-01 -9.52574730e-01 -1.69874996e-01 6.74335837e-01
9.73060489e-01 -1.18566144e+00 9.30000376e-03 -2.44021893e-01
-3.46858710e-01 1.00440109e+00 3.59213531e-01 -2.77562916e-01
1.26908934e+00 3.14898103e-01 -1.82561018e-02 -4.52325046e-01
-1.14472580e+00 1.19614266e-01 5.21886587e-01 3.67935225e-02
-2.50079334e-02 9.97488424e-02 2.57949769e-01 1.42760724e-01
-1.82140246e-02 7.69935399e-02 1.28933296e-01 6.65935814e-01
-2.90672362e-01 -6.23176575e-01 -4.05537397e-01 2.36140609e-01
1.06582478e-01 1.71356872e-01 -1.36225119e-01 5.46305358e-01
4.13844168e-01 8.70271921e-01 2.93232381e-01 -2.05040336e-01
-3.08268704e-02 -3.48642841e-02 -2.47473165e-01 -5.78170359e-01
-3.26377183e-01 4.15465310e-02 -3.42199117e-01 -4.23649490e-01
-7.15699077e-01 -6.42283380e-01 -1.21093094e+00 -4.24025685e-01
-1.35462269e-01 2.44849250e-01 4.75681782e-01 1.42072117e+00
5.34565449e-01 7.23133802e-01 9.36413109e-01 -9.48436677e-01
-2.68144518e-01 -1.07000041e+00 -5.43198049e-01 3.03688914e-01
6.82645321e-01 -8.37653458e-01 -5.65684974e-01 2.24769026e-01] | [6.7163848876953125, 3.025892734527588] |
70b1b1ca-1804-4514-a7a6-1e2c36311ed4 | parrottts-text-to-speech-synthesis-by | 2303.01261 | null | https://arxiv.org/abs/2303.01261v1 | https://arxiv.org/pdf/2303.01261v1.pdf | ParrotTTS: Text-to-Speech synthesis by exploiting self-supervised representations | Text-to-speech (TTS) systems are modelled as mel-synthesizers followed by speech-vocoders since the era of statistical TTS that is carried forward into neural designs. We propose an alternative approach to TTS modelling referred to as ParrotTTS borrowing from self-supervised learning (SSL) methods. ParrotTTS takes a two-step approach by initially training a speech-to-speech model on unlabelled data that is abundantly available, followed by a text-to-embedding model that leverages speech with aligned transcriptions to extend it to TTS. ParrotTTS achieves competitive mean opinion scores on naturalness compared to traditional TTS models but significantly improves over the latter's data efficiency of transcribed pairs and speaker adaptation without transcriptions. This further paves the path to training TTS models on generically trained SSL speech models. | ['Vineet Gandhi', 'Neha Sherin', 'Vishal Tambrahalli', 'Neil Kumar Shah', 'Saiteja Kosgi'] | 2023-03-01 | null | null | null | null | ['text-to-speech-synthesis', 'speech-synthesis'] | ['speech', 'speech'] | [ 4.06913728e-01 7.41362929e-01 -1.95176318e-01 -6.84610009e-01
-1.41251040e+00 -5.39561629e-01 7.79745817e-01 -1.63785830e-01
-2.03162059e-01 4.13124263e-01 8.30877185e-01 -7.79314160e-01
6.04915619e-01 -8.51440132e-02 -5.99725783e-01 -4.52820510e-01
3.78580570e-01 4.76786971e-01 -4.14751060e-02 -3.51044118e-01
-3.11487138e-01 1.58688784e-01 -1.16301155e+00 3.10258597e-01
4.80186552e-01 9.50089276e-01 2.56965309e-01 1.15432680e+00
-2.78905153e-01 8.79642844e-01 -6.06340170e-01 -4.07799095e-01
2.15451509e-01 -5.61871827e-01 -5.28649449e-01 -4.36666347e-02
3.66800278e-01 -6.12144209e-02 -4.19042289e-01 3.94997537e-01
7.04265654e-01 1.54978693e-01 8.55763733e-01 -9.56946135e-01
-8.20262969e-01 6.64679229e-01 2.37956531e-02 -3.18375416e-02
2.35235348e-01 1.78812727e-01 1.34364080e+00 -1.10287011e+00
4.01551813e-01 1.45647311e+00 6.82697535e-01 6.90761626e-01
-1.50168693e+00 -5.74477077e-01 -1.82010740e-01 -9.91818309e-02
-8.81775916e-01 -1.30263066e+00 5.97703576e-01 -3.09124023e-01
1.52109647e+00 3.56767982e-01 4.54153776e-01 1.58456588e+00
-1.92778826e-01 1.18296325e+00 9.81795251e-01 -1.01715398e+00
3.02369684e-01 3.33984703e-01 -3.49939972e-01 2.22056478e-01
-5.49563766e-01 3.63204688e-01 -8.97466004e-01 2.11741388e-01
3.76679152e-01 -6.48292243e-01 2.85910815e-02 9.35791880e-02
-1.20503509e+00 8.16906393e-01 -5.78801483e-02 3.04199189e-01
-2.39998028e-01 -1.86514407e-02 6.16535306e-01 7.29909956e-01
7.85660148e-01 4.55238938e-01 -7.93340206e-01 -4.00187939e-01
-1.44838274e+00 -1.77427337e-01 8.91982675e-01 1.16540384e+00
2.51552463e-01 9.06763375e-01 -1.92921087e-01 1.27858174e+00
5.05954206e-01 8.08013499e-01 9.03622031e-01 -7.86192417e-01
7.01847732e-01 -1.99961532e-02 -2.58422911e-01 -6.50136247e-02
1.55232728e-01 -6.30716681e-01 -5.06947756e-01 -1.02381729e-01
-1.05501413e-02 -2.66812950e-01 -1.10282588e+00 1.53689623e+00
4.89849932e-02 -6.63912967e-02 3.78545135e-01 1.24509856e-01
5.89731753e-01 1.16725683e+00 -2.28559107e-01 -3.44345748e-01
8.38542759e-01 -1.50099862e+00 -9.36223745e-01 -3.79830778e-01
8.00320446e-01 -1.02516150e+00 1.21369350e+00 2.07190126e-01
-1.24936080e+00 -8.28831971e-01 -1.04472864e+00 -1.17863834e-01
-4.00318921e-01 1.24463148e-01 -2.02381089e-01 1.00665033e+00
-1.50842178e+00 4.48206425e-01 -7.26486146e-01 -2.93172956e-01
1.33200914e-01 2.98883468e-01 -2.41503701e-01 2.76649982e-01
-1.17712367e+00 1.23248565e+00 3.19692791e-01 -9.68845189e-02
-9.60360229e-01 -5.31807601e-01 -1.05452538e+00 -1.45015016e-01
1.65197566e-01 -4.61499095e-01 1.89942455e+00 -9.07357633e-01
-2.20760393e+00 5.10131955e-01 -6.07561052e-01 -7.48937130e-01
3.02688420e-01 -1.86940938e-01 -7.78834105e-01 -1.29020661e-01
-6.11437894e-02 7.22405016e-01 1.29700637e+00 -1.06356645e+00
-5.47417104e-01 1.11511096e-01 -6.46020293e-01 3.24374169e-01
-5.64525604e-01 2.60021597e-01 5.96117042e-03 -8.05980027e-01
-4.89795730e-02 -9.41501021e-01 -7.13825896e-02 -4.80978042e-01
-7.54487634e-01 -3.32478613e-01 7.82362819e-01 -8.96182418e-01
1.22543323e+00 -2.13275862e+00 1.73675239e-01 -9.53407064e-02
-3.01840127e-01 7.18257487e-01 -4.70688045e-01 1.08885384e+00
-2.45886236e-01 2.85702180e-02 5.85557427e-03 -1.24195683e+00
4.78609264e-01 3.82941574e-01 -6.55657470e-01 -1.05155014e-01
5.01511514e-01 1.10370469e+00 -7.06313729e-01 -5.18585622e-01
3.65533948e-01 4.45507586e-01 -4.33387935e-01 6.65233552e-01
-1.93092450e-01 6.50066257e-01 2.34256148e-01 5.41836023e-01
-3.91200185e-02 2.88512766e-01 3.54075641e-03 -3.49182412e-02
-3.92796993e-01 1.03498006e+00 -5.55867791e-01 1.40680933e+00
-8.45545948e-01 9.07874227e-01 -2.77591079e-01 -7.87275374e-01
1.10868883e+00 9.31972861e-01 4.32050899e-02 -6.58864021e-01
-2.14419374e-03 7.21037149e-01 -8.14224631e-02 -3.80752295e-01
5.42211831e-01 -5.43494105e-01 2.23811358e-01 1.71395645e-01
6.06316686e-01 -6.49218857e-01 -2.18751013e-01 3.33868787e-02
1.01256108e+00 3.17973465e-01 2.30451286e-01 7.17868581e-02
2.78499931e-01 -1.42538875e-01 8.08192715e-02 5.30358255e-01
1.11638121e-01 7.06147850e-01 -6.96186200e-02 -8.12194794e-02
-1.52213812e+00 -1.25767851e+00 -5.47641329e-03 1.21055603e+00
-8.84637535e-01 -7.01277375e-01 -8.97282839e-01 -5.03611803e-01
-2.25962296e-01 1.21219718e+00 -3.47318411e-01 -2.29904708e-02
-3.95817637e-01 -3.47220041e-02 9.51792598e-01 6.08577192e-01
-9.55984294e-02 -1.13667178e+00 2.48507082e-01 5.25742531e-01
-4.46849912e-02 -1.15201080e+00 -7.89386332e-01 7.96692550e-01
-7.87777722e-01 -5.59261329e-02 -1.01518357e+00 -8.80538940e-01
1.75241351e-01 -2.17059627e-01 7.96408951e-01 -6.74633265e-01
3.68963778e-01 1.08975708e-01 -4.91311312e-01 -6.80834174e-01
-1.23054099e+00 4.66372132e-01 4.85946447e-01 1.54608667e-01
2.21432820e-01 -6.34847522e-01 -9.26541314e-02 1.22377016e-01
-6.48192286e-01 8.63948539e-02 8.92083406e-01 9.22593117e-01
2.70075530e-01 -3.62725228e-01 8.37042093e-01 -6.17152810e-01
4.84086156e-01 -2.11308330e-01 -2.32815489e-01 1.99959353e-01
-6.54828787e-01 2.53038201e-02 7.80944467e-01 -4.63825136e-01
-1.03252625e+00 8.80477801e-02 -5.96767843e-01 -5.11019707e-01
-2.89473116e-01 4.28249359e-01 -1.80129841e-01 2.15419829e-01
6.21214211e-01 4.35012728e-01 4.55244482e-02 -6.58386409e-01
5.64716220e-01 1.49562156e+00 4.65555400e-01 -2.06421152e-01
1.11619937e+00 -2.23743901e-01 -5.38762569e-01 -1.41132247e+00
-8.37670922e-01 -4.58075702e-01 -6.77788138e-01 -7.08873421e-02
6.83657646e-01 -9.28271174e-01 -1.73163518e-01 3.64742279e-01
-1.19431126e+00 -6.53887749e-01 -7.93110490e-01 6.39825404e-01
-7.40306079e-01 1.09325789e-01 -6.71390891e-01 -1.34116757e+00
-1.41312435e-01 -8.54141891e-01 1.20221221e+00 -2.50972152e-01
-5.98380625e-01 -1.06995237e+00 3.91672522e-01 5.53181350e-01
6.99229360e-01 -4.38830793e-01 7.81922221e-01 -1.05951929e+00
-1.85968265e-01 -2.56873995e-01 3.01487237e-01 1.19630706e+00
1.92172617e-01 -1.27405494e-01 -1.54881394e+00 -2.89546788e-01
1.37573346e-01 -4.66797471e-01 3.94642681e-01 3.99326801e-01
4.07813936e-01 -5.49702227e-01 -7.91802257e-03 4.86496747e-01
8.15562189e-01 5.69914222e-01 6.01213574e-01 8.59587565e-02
7.07288980e-01 8.00918996e-01 2.08760247e-01 -1.78995430e-01
3.16870332e-01 7.86226332e-01 -1.97624102e-01 -4.15595323e-01
-4.95466143e-01 -7.62040198e-01 8.16289186e-01 2.08714390e+00
4.60511774e-01 -4.04055238e-01 -7.12677002e-01 6.42406702e-01
-1.53862810e+00 -6.58098519e-01 -2.08712951e-03 2.08858705e+00
1.02233016e+00 3.52217197e-01 1.72890976e-01 3.60101730e-01
5.92799902e-01 3.29877853e-01 -4.90279138e-01 -9.99376237e-01
-2.23281652e-01 3.95861179e-01 3.60302567e-01 6.48208261e-01
-7.71269917e-01 1.20021105e+00 6.72271585e+00 1.17083228e+00
-9.59161043e-01 2.65474439e-01 5.32964468e-01 -3.65270786e-02
-3.85217190e-01 9.58811566e-02 -9.52696204e-01 3.55162889e-01
1.98785806e+00 1.15535468e-01 4.89969373e-01 5.16329050e-01
4.51365590e-01 4.56053048e-01 -1.33945107e+00 7.95887232e-01
2.15997882e-02 -1.08169663e+00 2.12316532e-02 -3.22222486e-02
6.91322505e-01 3.39242518e-01 2.65207082e-01 6.08440161e-01
3.70524675e-01 -1.08057284e+00 9.79639709e-01 5.71154281e-02
1.08876967e+00 -4.49156016e-01 4.77311999e-01 3.28279227e-01
-1.10153663e+00 1.58030868e-01 1.09632602e-02 2.04941601e-01
4.41457242e-01 1.53318390e-01 -1.62724471e+00 5.01229763e-01
1.33994013e-01 5.62961757e-01 -2.52777427e-01 5.54968238e-01
-2.19316021e-01 1.31007195e+00 -2.77406663e-01 -6.43928573e-02
5.19383252e-01 1.39129087e-01 5.50595820e-01 1.55053163e+00
2.61727452e-01 -4.61618721e-01 -9.72094014e-02 3.69800389e-01
3.88607495e-02 3.72521877e-01 -6.98749065e-01 -5.58314443e-01
7.03097105e-01 7.31229544e-01 -2.17480764e-01 -4.76976931e-01
-4.70115155e-01 1.04576659e+00 4.63143140e-02 4.54900771e-01
-3.62793982e-01 -4.84492242e-01 2.18054160e-01 9.54607651e-02
4.70042348e-01 -3.82844299e-01 7.38712400e-02 -9.61446285e-01
-9.22557190e-02 -1.06310403e+00 -3.45859408e-01 -1.02016318e+00
-1.31219113e+00 1.19492948e+00 -2.48809278e-01 -1.29690599e+00
-7.42579043e-01 -4.29132342e-01 -3.80568266e-01 1.13250971e+00
-1.62971056e+00 -1.44027495e+00 5.87110877e-01 1.98426440e-01
1.08074832e+00 -6.26716852e-01 1.13266540e+00 5.51315360e-02
-3.34491879e-01 7.12277412e-01 5.21149278e-01 3.35936695e-02
7.75689602e-01 -1.38509703e+00 1.31226563e+00 6.91942871e-01
6.16607606e-01 4.47714567e-01 7.11758852e-01 -3.59575659e-01
-1.05415750e+00 -1.06150854e+00 1.48659706e+00 -5.61139107e-01
8.23325753e-01 -8.24725747e-01 -7.60622561e-01 8.89807880e-01
7.24404395e-01 -2.65586466e-01 8.18730652e-01 1.16989769e-01
-3.94163847e-01 -1.23383313e-01 -5.90625763e-01 8.20596159e-01
8.76371264e-01 -1.12698364e+00 -1.02193534e+00 3.17340225e-01
1.23690653e+00 -2.34743327e-01 -7.21659958e-01 6.04560524e-02
5.25545657e-01 -6.39064252e-01 8.03014994e-01 -7.03012347e-01
1.99983001e-01 8.89477320e-03 -7.27582097e-01 -1.78357446e+00
4.79509868e-02 -1.43290138e+00 8.18627477e-02 1.46872997e+00
8.89982462e-01 -6.03395522e-01 5.28880239e-01 1.96490780e-01
-7.71456063e-01 -5.54742455e-01 -1.08730841e+00 -1.29610896e+00
5.60569502e-02 -6.48189783e-01 4.51158702e-01 9.08102095e-01
4.04035412e-02 7.05735505e-01 -4.53925818e-01 -2.60301158e-02
4.99575377e-01 -5.93831182e-01 7.65721023e-01 -9.44248617e-01
-5.48859239e-01 -1.45805165e-01 -2.06152096e-01 -1.29050231e+00
2.81598419e-01 -9.81600344e-01 4.44738269e-01 -1.27832842e+00
-6.12679243e-01 -2.23316148e-01 -1.52943671e-01 4.58529055e-01
2.37618938e-01 3.19427490e-01 1.66616276e-01 1.22626700e-01
-1.71606496e-01 9.23265755e-01 8.82924199e-01 6.13069534e-03
-3.52853358e-01 3.08107398e-03 -2.97067434e-01 5.26728511e-01
8.08150649e-01 -5.30163646e-01 -5.53532660e-01 -3.34931254e-01
-2.18834490e-01 2.62002915e-01 -6.06967248e-02 -1.08022416e+00
3.81993264e-01 1.57672971e-01 1.89420879e-02 -5.75778604e-01
8.34456742e-01 -5.26085079e-01 1.29000079e-02 1.70964062e-01
-7.50605345e-01 -7.20466226e-02 1.68857157e-01 3.85384440e-01
-4.27187741e-01 -2.60203034e-01 6.66345060e-01 1.06584691e-02
-9.71661434e-02 -1.02617286e-01 -7.74856567e-01 2.24727631e-01
3.42052490e-01 -4.09300923e-01 -5.85269649e-03 -7.54423678e-01
-7.79229879e-01 -2.50573218e-01 8.36545900e-02 5.67230582e-01
4.89789873e-01 -1.35760343e+00 -7.70293355e-01 5.82154691e-01
2.20673263e-01 -2.94456363e-01 -1.76408902e-01 6.42781079e-01
1.01640180e-01 8.78486693e-01 3.47110569e-01 -5.53473592e-01
-1.10462618e+00 1.85916558e-01 6.89294562e-02 1.40366212e-01
-4.39591169e-01 1.07285476e+00 2.89725568e-02 -8.46053064e-01
4.55285132e-01 -4.55763400e-01 2.39172280e-01 -3.17587666e-02
1.92038149e-01 8.39589164e-02 3.69039237e-01 -9.75200057e-01
-1.74391165e-01 3.26774180e-01 -1.58567414e-01 -1.01571929e+00
1.33007681e+00 -3.26994717e-01 5.92362523e-01 1.17019749e+00
1.44332111e+00 3.28227580e-01 -1.22419548e+00 -4.35890913e-01
2.14483023e-01 -8.78872871e-02 2.15823814e-01 -9.00366306e-01
-5.36417663e-01 1.15342069e+00 1.87314436e-01 4.25060779e-01
5.89218676e-01 -3.69937159e-02 1.22760344e+00 5.20541608e-01
-5.25542647e-02 -1.34172809e+00 1.18947610e-01 6.42851949e-01
8.79643559e-01 -1.08619130e+00 -7.77549565e-01 -1.56951752e-02
-9.49089885e-01 1.08762288e+00 9.45116393e-03 1.31788090e-01
5.86475134e-01 5.35812020e-01 5.06808400e-01 2.87888914e-01
-1.05217147e+00 -1.27423942e-01 4.48235393e-01 7.88541555e-01
5.21561086e-01 4.40995917e-02 5.54342747e-01 2.43350729e-01
-7.25583196e-01 -1.88198194e-01 1.95928171e-01 6.34172678e-01
-4.11675036e-01 -1.50850022e+00 -2.15520769e-01 3.78923684e-01
-1.09006688e-01 -7.12366641e-01 -4.62802708e-01 3.96236688e-01
-2.60767579e-01 1.38522959e+00 -9.95954201e-02 -6.23274207e-01
5.62709987e-01 7.10550010e-01 2.03869402e-01 -9.84604418e-01
-7.77173340e-01 5.66342652e-01 4.00869399e-01 -1.17594160e-01
-1.86239153e-01 -7.76994824e-01 -7.50687301e-01 6.85364008e-02
-5.55769205e-01 4.40415263e-01 1.07189584e+00 1.11289024e+00
1.68750852e-01 6.66956186e-01 9.85349417e-01 -9.61399138e-01
-7.57822692e-01 -1.19447279e+00 -3.72915864e-01 -1.20487720e-01
8.54871988e-01 -5.66847995e-02 -5.19282699e-01 2.22390890e-01] | [14.563797950744629, 7.0904221534729] |
7f08e5fc-7bc1-4a12-92ed-d957357e1c49 | tpa-net-generate-a-dataset-for-text-to | 2211.13887 | null | https://arxiv.org/abs/2211.13887v1 | https://arxiv.org/pdf/2211.13887v1.pdf | TPA-Net: Generate A Dataset for Text to Physics-based Animation | Recent breakthroughs in Vision-Language (V&L) joint research have achieved remarkable results in various text-driven tasks. High-quality Text-to-video (T2V), a task that has been long considered mission-impossible, was proven feasible with reasonably good results in latest works. However, the resulting videos often have undesired artifacts largely because the system is purely data-driven and agnostic to the physical laws. To tackle this issue and further push T2V towards high-level physical realism, we present an autonomous data generation technique and a dataset, which intend to narrow the gap with a large number of multi-modal, 3D Text-to-Video/Simulation (T2V/S) data. In the dataset, we provide high-resolution 3D physical simulations for both solids and fluids, along with textual descriptions of the physical phenomena. We take advantage of state-of-the-art physical simulation methods (i) Incremental Potential Contact (IPC) and (ii) Material Point Method (MPM) to simulate diverse scenarios, including elastic deformations, material fractures, collisions, turbulence, etc. Additionally, high-quality, multi-view rendering videos are supplied for the benefit of T2V, Neural Radiance Fields (NeRF), and other communities. This work is the first step towards fully automated Text-to-Video/Simulation (T2V/S). Live examples and subsequent work are at https://sites.google.com/view/tpa-net. | ['Chenfanfu Jiang', 'Yin Yang', 'Govind Thattai', 'Minchen Li', 'Feng Gao', 'Yuxing Qiu'] | 2022-11-25 | null | null | null | null | ['physical-simulations'] | ['miscellaneous'] | [ 1.51391765e-02 -4.41919655e-01 5.37667274e-01 -3.08409631e-02
-7.14002073e-01 -3.96484673e-01 1.02069604e+00 -9.32875946e-02
-1.17226407e-01 7.12725639e-01 1.43707562e-02 -2.86020875e-01
-5.79328649e-02 -9.25587475e-01 -7.95903802e-01 -7.31792748e-01
-1.07459791e-01 5.69203615e-01 5.71211398e-01 -6.33669674e-01
3.04206640e-01 7.08163559e-01 -2.01436162e+00 3.62533718e-01
7.02657759e-01 9.84746516e-01 4.43331540e-01 9.38142419e-01
-1.30905256e-01 6.97271407e-01 -4.29009506e-03 -3.07840705e-01
2.20919192e-01 -2.63265282e-01 -5.08172035e-01 -1.12215923e-02
4.88015890e-01 -3.12457800e-01 -4.35540706e-01 7.59174466e-01
5.65387368e-01 2.43445545e-01 7.07804263e-01 -1.38920963e+00
-3.12694162e-01 -1.54218882e-01 -4.81796265e-01 -5.33600710e-02
5.72609305e-01 6.30469859e-01 4.18417573e-01 -1.12944996e+00
1.15093648e+00 1.35124695e+00 6.22811258e-01 4.88832474e-01
-8.70927691e-01 -3.07260185e-01 -3.00816745e-01 2.79036373e-01
-1.23487496e+00 -3.37667376e-01 9.19624746e-01 -7.61933625e-01
9.23684537e-01 4.85244811e-01 9.71977055e-01 1.45776713e+00
4.12877142e-01 5.13820231e-01 1.28437996e+00 -2.24490926e-01
3.38805258e-01 -9.22460705e-02 -3.90405983e-01 6.87745035e-01
-3.68522070e-02 4.90518510e-01 -8.12663078e-01 -1.55472636e-01
9.54951644e-01 -2.89397955e-01 -3.68489921e-01 -3.94157737e-01
-1.36902344e+00 4.94946361e-01 2.45161474e-01 -7.08124563e-02
-4.95469034e-01 4.66248244e-01 4.50613618e-01 1.81776345e-01
4.81946319e-01 1.64645456e-03 -4.02707279e-01 -3.57338846e-01
-8.79512310e-01 8.62356126e-01 6.14100754e-01 8.38836014e-01
6.07890964e-01 3.08363408e-01 1.02646291e-01 5.60044885e-01
5.74624896e-01 9.55636263e-01 1.00358911e-01 -1.39124858e+00
3.04136395e-01 4.94976640e-02 2.45271638e-01 -8.58961821e-01
-1.96284965e-01 -1.06614092e-02 -9.10253108e-01 7.75208712e-01
8.03701356e-02 2.05260273e-02 -7.53280818e-01 1.06824374e+00
7.63370991e-01 1.58635020e-01 1.35965735e-01 1.11970818e+00
1.28256679e+00 9.48848248e-01 -1.58393681e-01 -2.20272169e-01
1.17636418e+00 -8.55053723e-01 -5.53570449e-01 1.85988739e-01
4.86522675e-01 -8.13322186e-01 9.68402565e-01 5.03688276e-01
-1.33602262e+00 -5.63171268e-01 -7.07952619e-01 1.64912944e-03
-2.83736885e-01 -4.52361465e-01 4.53958720e-01 2.54865915e-01
-9.75610793e-01 7.88731337e-01 -8.55681002e-01 -3.69782478e-01
3.52639139e-01 -1.53012127e-01 -4.85126197e-01 -1.83350056e-01
-1.41508043e+00 7.14169264e-01 -2.40239948e-01 -5.24392468e-04
-1.19600809e+00 -9.31975126e-01 -8.87463093e-01 -4.57043767e-01
3.43769729e-01 -9.72473741e-01 1.00273502e+00 -4.76715565e-01
-1.44328308e+00 7.88123667e-01 1.01491890e-03 -1.91086426e-01
1.00541866e+00 -1.27377361e-01 -1.97889149e-01 4.24429536e-01
-1.74048811e-01 6.61036074e-01 6.56368911e-01 -1.78530419e+00
-2.33326688e-01 7.87761360e-02 -4.89106541e-03 4.06179219e-01
3.78649294e-01 -1.49692530e-02 -4.94788051e-01 -4.99341965e-01
-2.31724098e-01 -8.13693821e-01 -2.07380071e-01 5.41688263e-01
-2.56423473e-01 1.24499954e-01 9.32112098e-01 -5.98996878e-01
5.76019228e-01 -1.88956141e+00 2.77899474e-01 -3.35044235e-01
1.04377329e-01 3.30810785e-01 4.53554653e-02 9.30710018e-01
2.20234230e-01 5.11402972e-02 -3.52559805e-01 -5.36092460e-01
-5.75046800e-02 1.57547027e-01 -3.28401715e-01 4.06903803e-01
-9.45559591e-02 9.48696017e-01 -7.99654722e-01 -7.03770339e-01
8.72595251e-01 8.13952982e-01 -3.92214894e-01 1.35755256e-01
-5.03247976e-01 7.81902432e-01 -4.63971168e-01 7.35998213e-01
9.31210816e-01 1.71667803e-02 -3.11249852e-01 -3.66867602e-01
-5.77111721e-01 -2.06488773e-01 -1.27632654e+00 1.63112915e+00
-4.11127150e-01 6.56178236e-01 4.71389860e-01 -6.11899912e-01
6.39212132e-01 3.97438139e-01 6.40075028e-01 -7.20563233e-01
1.03353404e-01 2.55613476e-01 -3.15406054e-01 -1.01314378e+00
6.00739360e-01 -2.06585035e-01 2.51276314e-01 -5.34778312e-02
-3.24513495e-01 -8.90654624e-01 7.52421245e-02 4.24666971e-01
8.55942249e-01 8.29708993e-01 -9.26317349e-02 -2.00311005e-01
4.62308139e-01 3.96236271e-01 1.94252491e-01 6.71922147e-01
-1.49231195e-01 1.08280396e+00 1.30424798e-01 -4.78744894e-01
-1.44139922e+00 -1.07913995e+00 -1.89082220e-01 3.35187107e-01
3.22443336e-01 -4.38144892e-01 -5.31310081e-01 6.82168975e-02
3.94702703e-03 4.81276035e-01 -5.12299299e-01 2.15883210e-01
-5.24409831e-01 -5.68771720e-01 4.20440376e-01 2.26369962e-01
6.09972537e-01 -1.27179694e+00 -8.68514359e-01 2.29933366e-01
-2.37319946e-01 -1.38036501e+00 8.17083567e-02 -2.39842713e-01
-8.12386394e-01 -1.00139594e+00 -8.11841011e-01 -1.90014124e-01
2.15244144e-01 3.95069897e-01 1.09003842e+00 3.13377440e-01
-4.75282967e-01 5.75817525e-01 -4.52405483e-01 -2.39436656e-01
-6.61468267e-01 -7.05241501e-01 7.09233060e-02 -1.60837516e-01
-5.51635623e-01 -5.77090025e-01 -6.83658302e-01 4.17040616e-01
-1.11347377e+00 7.19756603e-01 1.67615563e-02 4.88737583e-01
7.35021889e-01 1.54093746e-02 -4.13869694e-02 -4.99193043e-01
1.46822631e-01 -3.59650671e-01 -5.91508567e-01 2.53057536e-02
-2.74569303e-01 -3.97401184e-01 7.70235121e-01 -3.11344087e-01
-1.17344034e+00 -1.67823911e-01 -4.28764135e-01 -7.93303490e-01
-3.19392234e-01 4.66976851e-01 -2.17288397e-02 -1.50322378e-01
6.18075550e-01 4.71885383e-01 -1.32537454e-01 -2.49945566e-01
2.24628747e-01 5.17751217e-01 5.25179923e-01 -6.42041445e-01
1.00143039e+00 8.45056772e-01 3.76985610e-01 -1.33270073e+00
-3.49971235e-01 -1.88597649e-01 -5.37905335e-01 -6.82012558e-01
8.55877578e-01 -8.74669611e-01 -6.67696655e-01 8.83537114e-01
-1.17238057e+00 -9.40281868e-01 -1.43262714e-01 5.62952638e-01
-6.74870014e-01 6.19644761e-01 -5.08886456e-01 -9.75834012e-01
-2.45263591e-01 -1.36150217e+00 1.37506986e+00 3.88067998e-02
2.27432787e-01 -1.04581082e+00 6.90207258e-02 6.00834072e-01
6.20381296e-01 6.91228926e-01 5.94756067e-01 3.20586532e-01
-7.40005374e-01 2.26630941e-01 -1.74872026e-01 5.27438074e-02
-2.68141150e-01 7.02799499e-01 -9.07656431e-01 -1.86565802e-01
-1.39980102e-02 -4.21958834e-01 6.95783794e-01 6.22574508e-01
1.03749979e+00 1.55986890e-01 -2.05535576e-01 9.22809601e-01
1.57922339e+00 -2.91124918e-02 8.51596534e-01 1.53075933e-01
1.05219650e+00 7.16899693e-01 6.76272094e-01 6.03971839e-01
5.05554199e-01 1.03449881e+00 8.81196320e-01 -1.80618554e-01
-5.42314649e-01 1.40410038e-02 3.03372681e-01 8.57783020e-01
-6.73312247e-01 -5.42110622e-01 -9.63231087e-01 3.44821453e-01
-1.77640629e+00 -1.12988806e+00 -1.00165462e+00 2.04276872e+00
4.53291923e-01 1.28817111e-01 -1.98606819e-01 9.37892795e-02
3.51959080e-01 3.12546283e-01 -4.29986328e-01 -3.11799884e-01
-2.51682222e-01 -4.84600291e-02 3.62224162e-01 7.06174731e-01
-7.15045154e-01 9.27927852e-01 5.35826445e+00 1.08946586e+00
-1.20642984e+00 2.07505450e-01 3.02162230e-01 -4.39130850e-02
-5.63480318e-01 7.62404054e-02 -4.54984039e-01 4.36488211e-01
9.48695660e-01 7.86115676e-02 4.68069822e-01 5.33680141e-01
8.61106277e-01 -4.64929044e-01 -6.14203334e-01 1.09178126e+00
-8.38397220e-02 -1.76538146e+00 1.47952765e-01 4.00837511e-02
6.14724278e-01 5.13309836e-01 -2.84120679e-01 3.54342423e-02
2.95410920e-02 -8.13214540e-01 1.04566848e+00 7.53178298e-01
1.18404317e+00 -3.89308095e-01 4.84148651e-01 3.76930624e-01
-1.42795622e+00 5.28950512e-01 -2.49727145e-01 2.51515042e-02
7.58459330e-01 6.70999110e-01 -6.29442260e-02 1.07774663e+00
1.03688240e+00 8.02964151e-01 -3.11839998e-01 8.77437770e-01
1.48456872e-01 3.42202067e-01 -3.61916512e-01 7.58748278e-02
2.51900733e-01 -3.26819748e-01 8.61787677e-01 1.08772409e+00
3.22965235e-01 3.61756563e-01 3.51479985e-02 6.36147201e-01
2.65207767e-01 -4.17066179e-02 -9.14623082e-01 2.87017524e-01
6.11796491e-02 1.39310157e+00 -7.61013150e-01 -3.78786623e-01
-2.87002057e-01 9.15597379e-01 -6.33196160e-02 2.59825587e-01
-1.06158483e+00 5.60391024e-02 5.51848531e-01 3.98350745e-01
1.36150308e-02 -6.73608124e-01 6.04678653e-02 -1.24539578e+00
-4.83992845e-02 -4.77396905e-01 -1.91848785e-01 -1.62268245e+00
-1.15622890e+00 4.89098489e-01 4.04888242e-01 -1.37183833e+00
6.16762824e-02 -6.78225517e-01 -5.07632613e-01 7.40103126e-01
-1.62965786e+00 -1.22780311e+00 -8.55254054e-01 6.37088478e-01
7.29390085e-01 1.55115470e-01 6.25117600e-01 4.12661612e-01
-5.74549511e-02 -2.36532316e-01 3.40410173e-01 -3.32709134e-01
4.68224853e-01 -9.34381068e-01 6.22905314e-01 5.85180104e-01
-1.82620272e-01 2.26554554e-02 9.72619474e-01 -8.62093985e-01
-1.96488833e+00 -9.85673964e-01 3.55621636e-01 -4.62437660e-01
6.16545677e-01 -3.82265091e-01 -1.00020361e+00 5.51458858e-02
2.30154451e-02 4.29165572e-01 2.61335801e-02 -8.03451955e-01
1.08244553e-01 2.74171472e-01 -1.16719019e+00 7.21312225e-01
1.23340154e+00 -4.00561303e-01 -1.75731793e-01 5.69461703e-01
6.88042104e-01 -8.61657083e-01 -7.19663560e-01 4.79491293e-01
6.34286761e-01 -1.42537522e+00 1.15609467e+00 -3.70154642e-02
9.18163180e-01 -4.73647624e-01 -4.41568136e-01 -1.05864143e+00
1.36146814e-01 -7.52736092e-01 -1.84085920e-01 9.31802213e-01
-1.05167866e-01 -3.59651893e-01 4.48604196e-01 4.68733460e-01
-5.06440878e-01 -8.77248347e-01 -1.03235149e+00 -8.18518460e-01
4.33186926e-02 -9.79848206e-01 4.04809475e-01 8.67501676e-01
-5.93065202e-01 -1.21362820e-01 -5.94644427e-01 1.35542110e-01
7.87132442e-01 3.89669538e-02 9.02135193e-01 -1.11741841e+00
-1.26352832e-01 -1.82761580e-01 -2.89407134e-01 -8.11705947e-01
-2.16656581e-01 -6.54696941e-01 -1.03515647e-01 -1.82823825e+00
-9.11776349e-02 -4.36220169e-01 4.31926131e-01 -1.55935986e-02
3.05228233e-01 3.93625796e-01 1.43108606e-01 4.18411314e-01
-2.42098853e-01 8.28946710e-01 1.73911166e+00 1.98645920e-01
3.95957604e-02 -3.71291101e-01 1.21501431e-01 7.13007510e-01
6.14001632e-01 -3.69276822e-01 -2.02742890e-01 -4.23998415e-01
2.46301264e-01 5.33830404e-01 9.20284569e-01 -1.08334196e+00
2.79030744e-02 -6.11836493e-01 1.37958750e-01 -9.39882934e-01
8.00368309e-01 -7.83112526e-01 6.70280159e-01 5.77059150e-01
1.68101825e-02 -1.45328669e-02 1.41587242e-01 4.24442291e-01
6.10905439e-02 -1.68846741e-01 9.00506854e-01 -4.55041200e-01
-7.19475567e-01 5.60357749e-01 -3.87957782e-01 3.82547006e-02
1.01546562e+00 -3.39424610e-01 -4.03965771e-01 -3.96996230e-01
-5.67791700e-01 1.26417950e-01 9.54201996e-01 1.68161690e-01
8.10849547e-01 -1.16209924e+00 -9.65314209e-01 8.76434818e-02
1.13830395e-01 3.99489045e-01 7.93244004e-01 8.09397578e-01
-1.21000266e+00 7.66335055e-02 -1.56040832e-01 -7.86289811e-01
-1.19229472e+00 3.52273285e-01 3.70459020e-01 4.79937941e-02
-1.18984830e+00 5.68432033e-01 2.22902745e-01 -4.80072916e-01
-2.94371098e-01 -1.02212541e-01 -7.30966590e-03 -2.31754914e-01
3.96018893e-01 4.89032418e-01 1.72160879e-01 -9.13935542e-01
-2.96751559e-01 1.03739250e+00 4.95672017e-01 -3.94181281e-01
1.47671199e+00 -1.77387223e-01 2.97247440e-01 5.05666494e-01
8.98452580e-01 -2.80152503e-02 -1.64616752e+00 9.65222716e-02
-7.22460330e-01 -5.41744828e-01 2.41692960e-01 -5.50855219e-01
-1.05930722e+00 1.03405023e+00 4.01506424e-01 1.58281654e-01
7.51419723e-01 -9.06869695e-02 9.69229877e-01 8.04181248e-02
6.60677373e-01 -8.69521320e-01 -7.08371550e-02 5.22674739e-01
1.12706959e+00 -1.29843986e+00 2.99313188e-01 -5.72648346e-01
-7.50420570e-01 1.19986641e+00 5.50879776e-01 -7.15471506e-02
6.61017060e-01 4.56643075e-01 1.32153317e-01 -5.20112514e-01
-9.54946280e-01 4.40526046e-02 -4.41801094e-04 7.32138157e-01
1.85619518e-01 -1.45305067e-01 -4.07440104e-02 -1.09806962e-01
-3.62385921e-02 4.44759056e-02 8.39392245e-01 9.87650156e-01
-2.10473448e-01 -8.65795791e-01 -5.21981537e-01 2.58718550e-01
-9.30319354e-02 -1.47914171e-01 8.79568309e-02 8.12487841e-01
4.63657565e-02 8.45877826e-01 -1.35464057e-01 -1.36919856e-01
2.94076592e-01 -3.38221997e-01 5.92955589e-01 -2.32988507e-01
-5.46299219e-01 1.87004328e-01 9.29957703e-02 -7.70035744e-01
-6.46244526e-01 -6.73805237e-01 -1.51452088e+00 -6.60521805e-01
-7.48827904e-02 -5.73113076e-02 9.44762230e-01 8.49280119e-01
2.50072300e-01 4.88696277e-01 4.19912934e-01 -1.79517484e+00
5.82519509e-02 -7.03199923e-01 -6.09805465e-01 4.95349199e-01
1.97816581e-01 -9.41626728e-01 -5.60613394e-01 2.96464890e-01] | [9.157235145568848, -2.9653632640838623] |
db7b95d3-9bf4-4609-a7c5-b43973122534 | context-aware-alignment-and-mutual-masking | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Jin_Context-Aware_Alignment_and_Mutual_Masking_for_3D-Language_Pre-Training_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Jin_Context-Aware_Alignment_and_Mutual_Masking_for_3D-Language_Pre-Training_CVPR_2023_paper.pdf | Context-Aware Alignment and Mutual Masking for 3D-Language Pre-Training | 3D visual language reasoning plays an important role in effective human-computer interaction. The current approaches for 3D visual reasoning are task-specific, and lack pre-training methods to learn generic representations that can transfer across various tasks. Despite the encouraging progress in vision-language pre-training for image-text data, 3D-language pre-training is still an open issue due to limited 3D-language paired data, highly sparse and irregular structure of point clouds and ambiguities in spatial relations of 3D objects with viewpoint changes. In this paper, we present a generic 3D-language pre-training approach, that tackles multiple facets of 3D-language reasoning by learning universal representations. Our learning objective constitutes two main parts. 1) Context aware spatial-semantic alignment to establish fine-grained correspondence between point clouds and texts. It reduces relational ambiguities by aligning 3D spatial relationships with textual semantic context. 2) Mutual 3D-Language Masked modeling to enable cross-modality information exchange. Instead of reconstructing sparse 3D points for which language can hardly provide cues, we propose masked proposal reasoning to learn semantic class and mask-invariant representations. Our proposed 3D-language pre-training method achieves promising results once adapted to various downstream tasks, including 3D visual grounding, 3D dense captioning and 3D question answering. Our codes are available at https://github.com/leolyj/3D-VLP | ['Yinjie Lei', 'Yulan Guo', 'Yuwei Yang', 'Munawar Hayat', 'Zhao Jin'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['dense-captioning', 'visual-grounding', 'visual-reasoning', '3d-dense-captioning', 'visual-reasoning'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'reasoning'] | [-0.06354461 0.22568032 -0.12398522 -0.4897935 -0.85541993 -0.7337215
0.8293168 0.3214245 0.05469055 0.1282397 0.41935754 -0.6050696
0.02241159 -0.6068243 -1.0451119 -0.16615215 0.3984772 1.0092075
0.24810699 -0.53833306 0.19167371 0.9792589 -1.5303826 0.9088278
0.43989313 0.72846496 0.4800703 0.606921 -0.98020023 0.69521606
-0.36483246 -0.24673444 0.23669405 -0.12735337 -1.0092343 0.3481927
0.7671076 0.02551078 -0.3618355 0.92411524 0.23505498 0.07291031
0.7886424 -1.3884714 -1.2313169 -0.00943021 -0.63088727 -0.03985173
0.8115762 -0.04222046 1.0198946 -1.2144799 0.6818326 1.7022827
0.60272235 0.57933307 -1.2080345 -0.31546894 0.43549064 0.3047357
-1.5188164 -0.14934237 0.9926384 -0.6908258 1.4533191 0.3655125
0.5057347 1.1779726 -0.2067553 0.7702107 1.0205795 -0.4572554
-0.06880517 0.29036504 -0.06802255 0.7053465 -0.07566687 -0.147242
-0.7606918 0.02419038 1.1577414 0.10110985 0.05813153 -0.8395817
-1.4580966 0.63244087 0.83698654 0.43145317 -0.18730073 0.15542835
0.17550577 0.25331554 0.5751578 0.12845625 -0.585614 0.38787407
-0.6713142 0.39833352 0.5148774 1.331724 0.910209 -0.25129232
-0.18923113 0.72643614 0.59702873 0.8283156 -0.12198557 -0.91613895
0.8945073 0.9170727 -0.09332763 -1.2932771 -0.43981108 -0.12656069
-0.86378485 0.27526405 0.15444918 0.7210477 -0.9060922 1.49451
0.48356113 -0.06773122 0.23063914 1.0856577 1.3813007 0.5315579
0.07151202 0.36849684 1.6296736 -0.69736147 -0.5727936 -0.49609753
0.41907713 -1.0308537 1.3646739 -0.21126814 -1.1281629 -0.8591837
-0.69944 -0.93199205 -0.7318224 -0.18645388 0.41934463 -0.04358171
-1.2214828 -0.10316014 -0.5515428 -0.6952815 0.5865003 0.17229733
-0.6623402 -0.3031158 -1.0027237 0.9875458 0.37495837 -0.0080099
-0.54132915 -0.65758914 -1.1227686 -0.40453744 0.4328533 -1.3405014
1.1352763 -0.5915952 -0.9198449 1.8894997 -0.4981344 -0.1453549
0.53242546 -0.18228783 0.07696408 0.20231216 0.5012342 1.0308548
0.843561 -1.9064022 -0.2930177 -0.60357064 0.43781114 0.64241624
0.4375248 -0.22601278 -0.7614313 -0.7023018 0.76655984 -0.56097233
-0.04382211 0.56058526 -0.3645473 -0.42042226 0.8659478 -0.66613317
0.2382435 -1.941097 0.437648 -0.08458979 0.17756097 -0.00761528
-0.2427087 0.47869954 -0.17844582 -0.09177159 -0.26413953 -0.6161334
0.2601785 0.61695665 -0.7150402 0.26467717 0.57777137 1.2987735
-0.9145813 -0.6502975 0.74431354 0.6677006 -0.6287029 0.6104965
-0.67553586 0.8522406 -0.47614637 0.88888806 0.81089455 -0.58367026
-0.19188553 -0.6156338 -0.01942541 0.29116985 -0.9270434 2.5011802
-0.5358147 0.54042834 -0.02050216 -1.2965977 1.0345606 0.2277522
0.36095998 -0.91813713 -0.14190304 -0.09176762 -0.8360236 -0.75938493
0.45035622 -0.17392991 -0.268317 0.22169553 0.13323605 -1.0898185
-0.53052944 0.38687027 0.5373028 0.5756593 0.35394394 0.06119608
0.59533983 0.3681753 -0.27871642 0.639126 0.10491411 0.9600913
0.24026783 -0.46879226 -0.91766644 -1.3846512 0.05950734 0.9676643
0.57624805 -0.28894752 -0.13369934 -0.5433652 0.32105207 0.9728057
-0.5150129 0.13770355 -0.47315615 0.0195648 0.34747225 0.42392647
0.33771843 -0.9294091 -0.26363117 -0.2616714 -0.31260806 -1.6032654
-0.20200655 0.17229244 -0.82461345 -1.0612146 -0.5912036 -0.9330218
0.9077029 0.8187075 1.4548707 0.10111899 -0.14743873 1.072163
-0.3129889 -0.458387 -0.532664 -0.27346632 -0.2282924 -0.41919875
0.49513775 -0.6033954 -0.25708672 0.25623044 -0.6791231 0.4789496
0.43754363 0.42316583 1.0480826 -0.5125586 -0.08695463 -0.5577029
0.22092353 -0.4536999 -0.36207387 0.39739254 0.04540278 0.20624022
0.12611485 -0.23715906 -0.93837065 0.08775886 -0.26127896 -1.0188676
-0.567437 0.13382104 -0.2543257 0.10490864 0.73587084 0.09765872
0.07034854 -0.58335096 0.96662205 0.34976348 0.6525908 -0.8029013
1.0944035 0.85324436 0.19919395 -0.92965585 -1.2091879 -0.6903813
-1.1702063 0.0949351 1.2839209 -1.376176 -0.6662301 -0.13052402
-1.8091004 -0.19812606 -0.41496512 0.03625567 -0.86514956 0.54561305
-0.1038645 -0.62313503 -0.06875679 -0.90520394 1.789703 -0.26485458
-0.32169917 -1.1646512 -0.22821438 0.8799269 -0.03021966 0.30742878
1.1147969 -0.3462209 -0.92160636 0.19963135 -0.6988385 0.17630391
0.21199 -0.5607113 -0.9835991 0.04627355 0.0162852 -0.37195885
0.45701736 0.16544048 1.1351211 -0.09602698 -0.28357542 0.66421014
1.1661832 -0.3101429 0.23338532 0.18612628 1.3025584 0.88320464
0.5748708 0.09105612 0.9941742 0.8239524 0.8535296 -0.35769117
-0.50167745 -0.61626637 -0.06021019 0.6835917 -0.00992869 -0.09912204
-1.2762043 0.40631697 -2.083071 -0.82404715 -0.22297113 1.700957
0.5319317 -0.04214985 -0.19020905 -0.14658165 0.41207257 0.21342748
-0.5809205 -0.20725247 -0.40460706 0.07904631 0.14431417 0.6803686
-0.9633083 1.2600889 4.8923993 0.46803218 -0.77865547 0.32265952
0.1765373 0.06430109 -0.76608735 -0.02202884 -0.7160845 -0.37092084
0.11227823 0.25246322 0.2273591 0.50150985 0.15830803 0.2939151
-1.3609579 1.696878 0.43506023 -1.7174935 0.6095672 -0.1908023
0.47042847 0.08021788 -0.11583047 0.15389259 0.15908878 -1.0673633
1.1311276 0.7630175 0.8890589 -0.3140503 0.28307867 0.5016968
-1.1338762 0.3153117 -0.22572592 0.04870581 0.22113872 0.36324584
-0.9094806 0.9029382 0.9436319 0.9631627 -0.45223528 0.30629155
-0.30095372 -0.24419977 -0.22133067 0.21091641 0.32413444 -0.07243881
0.71529716 1.0461019 0.19537489 0.09512567 0.16330078 1.3296888
0.11406876 -0.02732831 -1.0561117 0.08171517 0.14220539 0.73908675
-0.5936752 -0.12202964 -0.6475271 1.2574368 0.5572702 0.41588438
-0.64168483 0.3868702 0.90450406 0.31898165 0.17201175 -0.73572224
-0.5620316 -1.2816721 0.15958819 -0.6722782 0.39468908 -1.5016347
-1.6973727 0.50066996 0.4099708 -1.3800329 -0.1585912 -0.921385
0.00670905 1.0885473 -1.6264893 -1.8033578 -0.45341218 0.94092584
0.8702341 -0.18342036 0.81771064 0.05259046 0.36752638 0.0809742
-0.5361493 -0.12551937 0.5523221 -1.1008613 0.56784534 0.37710023
0.6769838 0.58676976 0.59847736 -0.5244295 -1.6252997 -1.0218098
1.0870285 -1.1804236 0.5952441 -0.80957144 -1.0462211 0.78636026
0.03008865 0.21184027 0.4637712 0.20285125 -0.78121114 0.21438175
-0.9711371 0.8304276 1.6808054 -1.254929 -0.9845863 0.85975856
1.1116536 -0.9798545 -0.6022996 0.31354588 0.01772296 -0.90442324
1.6833729 -0.65792346 0.1952608 -0.5095584 -0.79949665 -0.70661074
-0.07017221 -0.28736988 -0.01954344 0.95921975 0.12390237 -0.17946719
0.7465809 0.3317725 -0.20689039 -0.35836843 -1.0409184 -0.31795758
0.07059842 -0.93883336 0.515194 1.1127363 -0.46006322 0.68262196
-0.0801153 0.61320746 0.5728788 0.50002533 1.0070668 -1.1423317
-0.03719201 -0.34160417 -0.4711387 -1.608315 0.63686633 -1.37017
-0.1501907 -2.0337453 -0.18019384 -0.39491448 0.18018036 0.5562388
0.23278773 0.11758771 0.36037153 0.39936617 -0.62414277 0.56245863
1.5266662 -0.42782527 -0.08041844 -0.18665722 -0.61041737 0.68984663
0.6174073 -0.2290503 -0.74612254 -1.0130203 0.42804363 0.07708568
1.1867763 -0.62694484 0.1808874 -0.33218247 0.37518606 -1.2072407
0.7933808 -1.2685772 0.11145101 -0.11587417 -0.2729532 0.1801803
0.5068045 0.5825939 -0.20844471 0.13510358 0.25376856 -0.52784806
-0.9877006 0.34037507 0.13690944 0.10868006 0.7123277 -0.3981937
-0.15261507 -0.28831756 -0.874659 0.20452513 0.4306809 0.93446946
0.93313754 -1.3921328 -0.58631635 0.22983056 0.6330161 0.6840299
0.49508566 0.40949747 -0.5884581 0.7260649 -0.10581972 -1.2576853
-1.3083405 0.6411348 0.35191676 0.20147572 -0.94427335 0.9440431
0.5916992 -1.0272164 0.31439698 -0.5680849 -0.08943891 -0.21271114
0.13259608 -0.4030404 0.10559124 -1.2068533 -0.57697594 1.1601923
0.35023794 -0.04864991 1.0654091 -0.43295717 -0.26657805 0.8537246
1.284453 -0.3637438 -1.0433259 -0.66495734 -0.02533239 -0.47370493
-0.09808458 -0.5082247 -0.5656516 1.2021278 0.45778766 0.04078776
0.7397454 0.8140482 0.36808312 0.5356097 0.32871294 -0.47904372
0.43038604 0.7414999 1.6273253 -1.6007767 0.07897025 -0.6486084
-0.8263365 0.958309 0.633719 0.20432654 0.73335105 -0.21676941
0.39096782 -0.69034445 -0.47041672 -0.49650458 0.8230419 1.0488466
0.20815147 -0.0473051 0.53744274 0.2072716 -0.22684486 -0.34319448
-0.28662026 0.79245687 -0.06801387 -0.9706044 -0.58604616 -0.22330633
0.37963715 -0.05543839 -0.5468496 0.84537274 0.29810092 0.7586705
0.43884003 -0.01147283 0.8077652 -0.01915205 0.7397606 -0.8056711
-0.20010632 -0.13331996 -0.07475757 -0.62278205 -0.9522715 -0.53706837
-1.5035262 -0.11790757 0.46023998 -0.21470413 0.7374912 1.0190787
0.59140086 0.3881076 0.13349181 -1.1881579 -0.1133996 -0.42411387
-0.09119914 0.7355066 0.50807595 -0.8474228 -0.23853552 0.17349719] | [8.167732238769531, -3.3418478965759277] |
15a937a3-748e-45aa-b545-32e4d02604d4 | semantic-aware-knowledge-preservation-for | 1904.03208 | null | https://arxiv.org/abs/1904.03208v3 | https://arxiv.org/pdf/1904.03208v3.pdf | Semantic-Aware Knowledge Preservation for Zero-Shot Sketch-Based Image Retrieval | Sketch-based image retrieval (SBIR) is widely recognized as an important vision problem which implies a wide range of real-world applications. Recently, research interests arise in solving this problem under the more realistic and challenging setting of zero-shot learning. In this paper, we investigate this problem from the viewpoint of domain adaptation which we show is critical in improving feature embedding in the zero-shot scenario. Based on a framework which starts with a pre-trained model on ImageNet and fine-tunes it on the training set of SBIR benchmark, we advocate the importance of preserving previously acquired knowledge, e.g., the rich discriminative features learned from ImageNet, to improve the model's transfer ability. For this purpose, we design an approach named Semantic-Aware Knowledge prEservation (SAKE), which fine-tunes the pre-trained model in an economical way and leverages semantic information, e.g., inter-class relationship, to achieve the goal of knowledge preservation. Zero-shot experiments on two extended SBIR datasets, TU-Berlin and Sketchy, verify the superior performance of our approach. Extensive diagnostic experiments validate that knowledge preserved benefits SBIR in zero-shot settings, as a large fraction of the performance gain is from the more properly structured feature embedding for photo images. Code is available at: https://github.com/qliu24/SAKE. | ['Lingxi Xie', 'Huiyu Wang', 'Qing Liu', 'Alan Yuille'] | 2019-04-05 | semantic-aware-knowledge-preservation-for-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Liu_Semantic-Aware_Knowledge_Preservation_for_Zero-Shot_Sketch-Based_Image_Retrieval_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Liu_Semantic-Aware_Knowledge_Preservation_for_Zero-Shot_Sketch-Based_Image_Retrieval_ICCV_2019_paper.pdf | iccv-2019-10 | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 1.82414472e-01 -1.52953953e-01 -4.00912672e-01 -1.72466576e-01
-6.95014477e-01 -3.47692966e-01 8.12258780e-01 -2.21707493e-01
-4.14795309e-01 5.01484275e-01 4.38118696e-01 1.15302637e-01
-5.71350813e-01 -8.47284675e-01 -7.00661361e-01 -7.14670241e-01
3.80590618e-01 -4.59074639e-02 2.80816764e-01 -4.10035670e-01
3.08321267e-01 4.42635417e-01 -1.57125258e+00 8.02943762e-03
6.63656950e-01 9.35788870e-01 3.33532751e-01 1.79898605e-01
-1.27383590e-01 6.81057453e-01 -2.68813055e-02 -5.83765566e-01
5.01162946e-01 -3.45498592e-01 -8.00689697e-01 1.52103111e-01
4.64353114e-01 -4.90383625e-01 -6.05379879e-01 1.02839243e+00
4.12125319e-01 4.24757302e-01 5.46022117e-01 -1.30308557e+00
-8.95521879e-01 2.38256589e-01 -3.13697428e-01 2.38603372e-02
5.90269500e-03 7.43742809e-02 1.40256882e+00 -9.16099310e-01
7.33131945e-01 1.02091157e+00 4.69338655e-01 6.16414189e-01
-1.18801188e+00 -5.42264521e-01 1.58754289e-02 5.65409184e-01
-1.58987212e+00 -4.68366593e-01 1.15740192e+00 -3.52691203e-01
3.68664563e-01 1.22819498e-01 4.22535360e-01 9.96050537e-01
-1.26181349e-01 8.26769471e-01 9.08247173e-01 -5.08882880e-01
2.78830886e-01 3.08381110e-01 1.10036910e-01 5.97944498e-01
2.44541138e-01 1.67578205e-01 -7.34223604e-01 -8.77855625e-03
9.03748631e-01 3.76657367e-01 -4.44414556e-01 -1.06849563e+00
-1.13481784e+00 9.52329397e-01 6.80781305e-01 4.92051303e-01
-3.60224605e-01 1.21680029e-01 2.90066451e-01 3.57004404e-01
3.73696476e-01 4.03197974e-01 -1.71265885e-01 2.28249114e-02
-8.90737057e-01 -6.43044189e-02 5.86813152e-01 8.75798643e-01
9.51642454e-01 -1.40871450e-01 -3.42475414e-01 1.16003990e+00
5.48592545e-02 3.81478220e-01 3.85842979e-01 -8.81973565e-01
3.70885879e-02 5.60747266e-01 -4.99364808e-02 -8.62706900e-01
2.14369237e-01 -3.61805290e-01 -7.36202121e-01 2.20337749e-01
1.94710076e-01 4.23260778e-01 -8.35786462e-01 1.85200047e+00
3.44789565e-01 3.26118261e-01 4.89346460e-02 1.02571166e+00
6.53332233e-01 4.35023069e-01 2.93104015e-02 2.47581854e-01
1.30768359e+00 -9.51641679e-01 -4.23401952e-01 7.56479509e-04
2.96781331e-01 -5.22914708e-01 1.32445586e+00 9.52718928e-02
-6.80603504e-01 -4.54966694e-01 -1.07201302e+00 -1.60685256e-01
-4.92596477e-01 -9.09935161e-02 3.20844561e-01 3.88849050e-01
-9.61904585e-01 7.12070346e-01 -3.60116541e-01 -5.99612296e-01
7.21283138e-01 1.39695346e-01 -5.25516450e-01 -5.31854868e-01
-1.17550361e+00 7.53101349e-01 3.52824360e-01 -2.79576212e-01
-7.51343369e-01 -8.72585773e-01 -7.29710758e-01 3.02058637e-01
6.90962076e-01 -7.60355115e-01 9.73850608e-01 -9.83481705e-01
-1.46915865e+00 6.79355085e-01 8.26310143e-02 -2.14288861e-01
5.25091350e-01 -7.52701983e-02 -8.78867209e-02 4.46186125e-01
2.44372571e-03 6.40688062e-01 1.19551980e+00 -1.33709693e+00
-3.68731648e-01 -3.25544506e-01 3.59277308e-01 1.46689788e-01
-9.18558061e-01 -3.35112840e-01 -6.35353804e-01 -6.66780829e-01
-2.65563309e-01 -9.35166895e-01 2.99113169e-02 5.05889297e-01
4.47952375e-02 -1.86666355e-01 6.81403220e-01 -7.03438580e-01
1.13631856e+00 -2.32823372e+00 1.30937994e-01 2.94879764e-01
1.04972705e-01 5.53025305e-01 -3.92689437e-01 6.42611325e-01
9.21254531e-02 -1.04462385e-01 -3.41109604e-01 -2.35094532e-01
6.24686219e-02 4.28487808e-01 -4.32912052e-01 3.75566602e-01
3.07161361e-01 1.21736133e+00 -9.92264450e-01 -4.09156382e-01
5.28890789e-01 7.83090055e-01 -5.35172701e-01 5.67518286e-02
-2.82507483e-02 4.47750241e-01 -5.59827089e-01 5.56616426e-01
5.04642189e-01 -5.38645446e-01 1.16212793e-01 -2.98913270e-01
1.03756338e-01 -2.44435221e-01 -9.47727263e-01 2.02141166e+00
-6.13559067e-01 4.62118864e-01 -2.58574516e-01 -1.14709949e+00
7.90712178e-01 1.13343626e-01 5.04254103e-01 -1.09023380e+00
5.65300249e-02 -1.42550725e-03 -2.95793951e-01 -3.36419642e-01
4.32297796e-01 -2.58677900e-01 1.53737590e-01 2.52722234e-01
3.12624425e-01 1.45856604e-01 -3.65538150e-03 4.65669334e-01
9.65466797e-01 1.37303099e-01 5.55078626e-01 -2.27026284e-01
6.06773078e-01 -1.15777701e-01 2.86573738e-01 6.76578701e-01
-2.67423928e-01 5.16799986e-01 8.09814632e-02 -2.47315526e-01
-1.24063575e+00 -1.00490463e+00 -1.40941411e-01 1.13637054e+00
5.25684953e-01 -3.73850733e-01 -4.01184410e-01 -7.11483479e-01
1.67819083e-01 7.39369333e-01 -7.53701627e-01 -5.58752716e-01
-2.88385361e-01 -6.98101446e-02 2.00397789e-01 4.46952701e-01
6.47146761e-01 -9.71534193e-01 -4.99609679e-01 -1.51198786e-02
-9.89305899e-02 -9.98231292e-01 -4.86531436e-01 -3.99952620e-01
-6.21313512e-01 -9.91092324e-01 -9.74184394e-01 -5.88917792e-01
5.85925758e-01 8.79584610e-01 8.88554871e-01 2.39957348e-01
-4.84387368e-01 9.95306313e-01 -6.95954144e-01 -8.98446292e-02
-1.03540964e-01 -6.08173423e-02 -3.18476349e-01 3.22848409e-01
3.58301848e-01 -6.76254153e-01 -7.22641766e-01 2.39160880e-01
-1.23045433e+00 -1.98817588e-02 8.69690359e-01 1.12745655e+00
4.29503947e-01 -9.58502740e-02 6.55728102e-01 -5.95333695e-01
3.84895205e-01 -3.79083872e-01 -3.76663089e-01 5.77586174e-01
-7.18291283e-01 2.96956986e-01 4.30324256e-01 -4.65470284e-01
-1.06785798e+00 -7.67639056e-02 5.27949445e-02 -9.33953404e-01
1.18405111e-01 3.52132976e-01 -2.91422248e-01 -4.08164054e-01
4.73017633e-01 5.47057867e-01 1.77810907e-01 -5.07939994e-01
7.32237577e-01 5.42062223e-01 3.59181732e-01 -5.92918277e-01
1.04106593e+00 8.18910301e-01 -5.16741686e-02 -9.70340550e-01
-8.85226905e-01 -9.24926221e-01 -5.33922195e-01 -1.82352573e-01
5.61607540e-01 -9.00551677e-01 -4.10676807e-01 1.53242946e-01
-8.59287143e-01 -1.65602416e-01 -5.24438024e-01 2.81962752e-01
-6.17735088e-01 4.66457427e-01 -1.39100954e-01 -6.05823874e-01
-5.17056227e-01 -7.45133817e-01 1.04072666e+00 1.48535475e-01
1.27547190e-01 -1.08484685e+00 1.95390031e-01 2.72591650e-01
4.62028652e-01 -2.41416954e-02 9.10556853e-01 -6.11857474e-01
-7.25747526e-01 2.93974914e-02 -5.62407136e-01 6.82028532e-01
1.99783787e-01 -4.29549724e-01 -1.09267044e+00 -3.67008418e-01
-2.90762484e-01 -3.35018486e-01 1.18530869e+00 -2.94036232e-02
1.18170822e+00 -2.74596512e-01 -1.36282086e-01 4.47208554e-01
1.75448644e+00 -1.44062370e-01 7.06524193e-01 2.43324071e-01
6.58784568e-01 5.53673863e-01 6.83757961e-01 4.75783437e-01
3.52787137e-01 8.18560958e-01 4.00665104e-01 1.67972550e-01
-5.34049153e-01 -5.36639273e-01 9.51959565e-02 6.36318803e-01
-6.24561161e-02 1.36442855e-01 -6.77858353e-01 7.80943394e-01
-2.00826216e+00 -1.10976171e+00 6.03352845e-01 2.27967048e+00
8.78264070e-01 -2.24030957e-01 -7.86850080e-02 9.16359201e-02
6.38664663e-01 2.80837864e-01 -5.53619385e-01 6.50174767e-02
9.25397500e-02 3.27466518e-01 3.19315642e-01 2.74452865e-01
-8.95962536e-01 1.03814054e+00 4.31160641e+00 1.18968141e+00
-1.13393629e+00 3.26410085e-01 2.12629080e-01 -1.29239038e-02
-3.72497618e-01 7.97120184e-02 -5.23213327e-01 3.62989992e-01
5.42057216e-01 -5.01341879e-01 6.74588323e-01 8.67359340e-01
-1.78886726e-01 1.62626192e-01 -1.06886959e+00 1.09423769e+00
4.08667773e-01 -1.55488360e+00 3.59299749e-01 1.57704577e-01
7.30572999e-01 -1.83513105e-01 8.18428919e-02 3.85357320e-01
1.02940425e-01 -6.09936357e-01 6.48899138e-01 6.03649497e-01
9.59583342e-01 -6.73813462e-01 4.87208843e-01 2.27821350e-01
-1.07091033e+00 -1.15963466e-01 -4.82563704e-01 1.76173136e-01
-5.07944301e-02 5.06331742e-01 -6.94420159e-01 8.17771614e-01
5.79484701e-01 1.04325032e+00 -5.21982729e-01 1.12671483e+00
-3.58952314e-01 4.18062508e-01 -2.45730206e-02 2.00108305e-01
1.32240206e-01 -1.74284756e-01 5.12450635e-01 8.91866863e-01
2.17787787e-01 1.48651421e-01 -3.95980217e-02 8.10739994e-01
-2.53466606e-01 1.61597922e-01 -8.91594172e-01 -9.96849909e-02
4.52820301e-01 1.37614679e+00 -3.95068794e-01 -3.16295594e-01
-5.63012421e-01 1.05002558e+00 4.37585860e-01 3.61257195e-01
-7.00200140e-01 -4.32872385e-01 7.33020365e-01 -3.18524465e-02
7.40879834e-01 -9.12192911e-02 -5.43925464e-02 -1.34173644e+00
7.37990811e-02 -5.75961292e-01 3.38590860e-01 -6.40654147e-01
-1.49201381e+00 2.82996476e-01 -1.26907364e-01 -1.34734952e+00
-6.60456270e-02 -5.09269238e-01 -5.30481696e-01 5.63841820e-01
-2.04314876e+00 -1.49293756e+00 -6.04378819e-01 9.57463741e-01
5.85437417e-01 2.21694466e-02 6.41866207e-01 4.03549731e-01
-3.47213894e-01 7.81503081e-01 2.94314802e-01 1.08594291e-01
8.26639950e-01 -8.07788014e-01 3.52773666e-02 6.38447106e-01
4.50666845e-01 6.22670352e-01 3.67428482e-01 -4.22565073e-01
-1.59712923e+00 -1.22717547e+00 6.13614798e-01 -3.75958085e-01
8.29444706e-01 -3.15904945e-01 -1.05419362e+00 3.97769600e-01
-9.55069885e-02 1.77239299e-01 6.20009542e-01 -1.84877627e-02
-8.08210015e-01 -2.34056175e-01 -1.10906422e+00 5.26113272e-01
1.25645173e+00 -7.89419770e-01 -7.04607725e-01 8.41142088e-02
7.24995911e-01 2.22395554e-01 -1.01270521e+00 3.59625161e-01
7.91490555e-01 -6.97348952e-01 1.25835812e+00 -5.73074102e-01
5.71805477e-01 -5.16886897e-02 -4.62789446e-01 -1.27934635e+00
-3.88512343e-01 -2.03968108e-01 -1.02162533e-01 1.18602228e+00
-1.55629680e-01 -6.51967227e-01 4.88385051e-01 4.30133164e-01
1.13573208e-01 -6.49986029e-01 -9.18960571e-01 -1.11667478e+00
3.79132330e-02 -9.10807699e-02 4.25960988e-01 1.08155513e+00
-2.20256448e-01 3.64338607e-01 -6.08547926e-01 -6.57001510e-03
7.74786413e-01 2.91323334e-01 8.44733238e-01 -1.29047430e+00
-2.21070856e-01 -3.52239579e-01 -5.58599293e-01 -8.20181966e-01
3.07850301e-01 -9.56669986e-01 -1.12239800e-01 -1.47601855e+00
4.14832741e-01 -3.83472651e-01 -7.64027655e-01 6.45792663e-01
-1.76685572e-01 3.72104704e-01 5.56667507e-01 4.41716224e-01
-8.02797735e-01 9.46431458e-01 1.29518819e+00 -1.80351853e-01
8.71363059e-02 -3.53428274e-01 -8.11767876e-01 3.59988749e-01
7.12135017e-01 -2.11670130e-01 -6.11713469e-01 -3.17414552e-01
-3.29279229e-02 -3.33487421e-01 7.97947347e-01 -7.75296211e-01
3.04389805e-01 -1.79745719e-01 1.06617428e-01 -3.42980288e-02
5.35252213e-01 -9.92404342e-01 -8.27674046e-02 3.29339176e-01
-2.71727562e-01 -6.26275480e-01 -1.60471816e-02 8.11268866e-01
-2.87678838e-01 -1.70031220e-01 8.66825342e-01 -8.17521755e-03
-1.27985585e+00 5.69765747e-01 4.73483175e-01 -6.02537475e-04
1.05728555e+00 -2.16018215e-01 -4.26439643e-01 -3.15191388e-01
-3.92118692e-01 4.22977023e-02 6.25086844e-01 6.24757230e-01
6.79574072e-01 -1.47107470e+00 -5.13838351e-01 2.86244661e-01
7.94505179e-01 -4.45615917e-01 4.59549606e-01 8.20288956e-01
1.12273935e-02 4.26928341e-01 -3.92056435e-01 -4.40702796e-01
-1.10452449e+00 7.43606508e-01 3.78527157e-02 -9.73420441e-02
-9.01539564e-01 6.21423185e-01 4.54030812e-01 -1.57814369e-01
1.12062544e-01 1.15393706e-01 -2.80164722e-02 -1.72276236e-02
6.33646786e-01 4.06094730e-01 3.04374825e-02 -5.34598231e-01
-2.90750265e-01 7.16761172e-01 -1.31639376e-01 2.01262105e-02
1.46608579e+00 -2.43229017e-01 8.19319636e-02 2.32796580e-01
1.32612884e+00 -1.63916081e-01 -1.44527137e+00 -8.30450296e-01
4.13601622e-02 -8.82891893e-01 2.49919489e-01 -7.12178648e-01
-1.09411645e+00 7.75623918e-01 4.52984393e-01 -1.79958448e-01
1.10763574e+00 2.90720493e-01 7.77957618e-01 4.87974256e-01
6.21293485e-01 -9.73603249e-01 4.68898028e-01 2.28487015e-01
1.02366710e+00 -1.39528358e+00 -1.03592664e-01 -2.37809628e-01
-6.51786864e-01 8.06933761e-01 2.56947339e-01 -3.04621220e-01
6.00138307e-01 -4.36519653e-01 -2.69855917e-01 -9.51562524e-02
-5.97265244e-01 -4.65545535e-01 4.55822796e-01 4.16314721e-01
-1.53879777e-01 2.45607719e-02 -2.81095266e-01 5.22327602e-01
1.90556318e-01 2.47431472e-01 1.03159115e-01 9.66212213e-01
-4.98786658e-01 -1.16194916e+00 -6.40467927e-02 3.92518401e-01
4.49704193e-02 -1.52051866e-01 -3.38421196e-01 8.15046608e-01
-3.66739035e-02 6.44920766e-01 -1.90152958e-01 -3.93677354e-01
3.26790810e-01 6.30210415e-02 5.90930223e-01 -4.87829030e-01
-1.12712443e-01 -2.71351069e-01 -2.07643017e-01 -7.63440728e-01
-5.32251656e-01 -3.84965688e-01 -8.36171269e-01 -2.55169868e-01
-2.45393887e-01 -9.62985978e-02 7.02589571e-01 8.46332073e-01
5.43797374e-01 3.16371709e-01 6.94961429e-01 -6.55239820e-01
-9.40038443e-01 -5.81709146e-01 -6.37315333e-01 8.08955431e-01
2.88176000e-01 -9.84175563e-01 -3.47658575e-01 2.23627733e-03] | [11.535587310791016, 0.7832457423210144] |
f8b4ef30-c352-4ab2-96e0-88804f00fce9 | advancing-covid-19-diagnosis-with-privacy | 2111.09461 | null | https://arxiv.org/abs/2111.09461v1 | https://arxiv.org/pdf/2111.09461v1.pdf | Advancing COVID-19 Diagnosis with Privacy-Preserving Collaboration in Artificial Intelligence | Artificial intelligence (AI) provides a promising substitution for streamlining COVID-19 diagnoses. However, concerns surrounding security and trustworthiness impede the collection of large-scale representative medical data, posing a considerable challenge for training a well-generalised model in clinical practices. To address this, we launch the Unified CT-COVID AI Diagnostic Initiative (UCADI), where the AI model can be distributedly trained and independently executed at each host institution under a federated learning framework (FL) without data sharing. Here we show that our FL model outperformed all the local models by a large yield (test sensitivity /specificity in China: 0.973/0.951, in the UK: 0.730/0.942), achieving comparable performance with a panel of professional radiologists. We further evaluated the model on the hold-out (collected from another two hospitals leaving out the FL) and heterogeneous (acquired with contrast materials) data, provided visual explanations for decisions made by the model, and analysed the trade-offs between the model performance and the communication costs in the federated training process. Our study is based on 9,573 chest computed tomography scans (CTs) from 3,336 patients collected from 23 hospitals located in China and the UK. Collectively, our work advanced the prospects of utilising federated learning for privacy-preserving AI in digital health. | ['Tian Xia', 'Carola-Bibiane Schönlieb', 'Zhen Li', 'Jianming Wang', 'Chuangsheng Zheng', 'Joan Lasenby', 'Adrian Weller', 'Daniel Rubin', 'Evis Sala', 'Lorena Escudero Sanchez', 'Lucian Beer', 'Zhongzhao Teng', 'Michael Roberts', 'Weiyang Liu', 'Jianjun Zhang', 'Pattanasak Mongkolwat', 'Jianbo Shao', 'Xiang Wang', 'Xuehua Peng', 'Jia Wu', 'Ihab R. Kamel', 'Neil J. Halin', 'Nagaraj Holalkere', 'Ming-Wei Wang', 'Dehua Yang', 'Yu Zeng', 'Anguo Chen', 'Xi Wang', 'Wenqing Zhang', 'Chuou Xu', 'Dandan Tu', 'Xiaowu Liu', 'Changzheng Zhang', 'Renhao Huang', 'Xinyu Zou', 'Chang Shu', 'Song Bai', 'Jiehua Yang', 'Ke Ma', 'Fan Yang', 'Ziwei Fan', 'Jiefeng Gan', 'Yongchao Xu', 'Liya Ma', 'Hanchen Wang', 'Xiang Bai'] | 2021-11-18 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [ 2.24590644e-01 4.66679543e-01 -3.42088193e-01 -4.10945714e-01
-1.22670019e+00 -6.16281629e-01 2.60688692e-01 3.57887566e-01
-6.48939788e-01 7.19123304e-01 2.42063195e-01 -9.83621240e-01
-6.34012818e-01 -5.50883949e-01 -7.63532579e-01 -9.61770773e-01
-3.69614333e-01 8.06326032e-01 -4.08588141e-01 6.12332344e-01
-3.87034327e-01 6.01167679e-01 -7.09422886e-01 7.42439270e-01
6.11059487e-01 1.19987261e+00 -2.90520996e-01 8.67373109e-01
4.07374203e-01 1.33661187e+00 -4.31477368e-01 -9.63165164e-01
6.14355862e-01 -1.25722989e-01 -1.00972056e+00 -2.01142758e-01
6.25235364e-02 -6.74156785e-01 -4.18373555e-01 6.63391292e-01
7.56193638e-01 -6.62947118e-01 4.27027881e-01 -1.24966455e+00
-3.67376953e-01 8.14293563e-01 -1.98533162e-01 -5.06739952e-02
-2.08781958e-02 5.08983612e-01 9.62397814e-01 -9.78928953e-02
9.84144449e-01 5.35660684e-01 6.05723798e-01 4.90849912e-01
-8.69843721e-01 -9.41860974e-01 -2.74271578e-01 -1.22833475e-02
-1.19049132e+00 -1.79674014e-01 1.38164192e-01 -2.89447814e-01
5.27073264e-01 7.19966471e-01 6.95383430e-01 1.35878134e+00
7.97399938e-01 7.77159989e-01 1.23913634e+00 -2.07600281e-01
4.01169658e-01 3.82378936e-01 -1.77155346e-01 7.22398758e-01
5.73520422e-01 2.51993239e-01 -3.32557112e-01 -1.09249222e+00
5.16303360e-01 2.97754854e-01 -2.37828225e-01 -4.19009566e-01
-1.42215967e+00 9.21800494e-01 6.80145383e-01 -3.02240755e-02
-5.81679046e-01 -2.14132115e-01 7.44010150e-01 3.80168051e-01
2.38010854e-01 3.69000167e-01 -7.23983347e-01 2.31415495e-01
-7.05366910e-01 8.27283189e-02 9.83571172e-01 9.09305215e-01
-6.50745630e-03 -2.76512742e-01 -1.06730551e-01 1.63616747e-01
1.45808220e-01 4.06973809e-01 4.87954021e-01 -9.51957583e-01
3.77119958e-01 3.88109654e-01 -6.05650581e-02 -8.35803270e-01
-5.52234530e-01 -4.04210776e-01 -1.25109828e+00 -9.91133004e-02
3.11633915e-01 -6.32029414e-01 -6.93064511e-01 1.11649549e+00
4.35704470e-01 1.23264119e-02 3.56579244e-01 8.41642141e-01
6.11289501e-01 1.47684932e-01 2.52012849e-01 -1.41686261e-01
1.57666111e+00 -6.37600064e-01 -6.60815239e-01 1.86189502e-01
9.35726523e-01 -3.95243764e-01 3.46371531e-01 7.37887323e-01
-1.05563283e+00 1.81968473e-02 -7.85291016e-01 4.42197055e-01
-1.10028103e-01 -2.14293405e-01 6.58596635e-01 8.54453564e-01
-1.02576399e+00 2.42731914e-01 -1.02236843e+00 -6.46445826e-02
1.17693400e+00 6.20192349e-01 -6.30652547e-01 -4.39633906e-01
-1.06254756e+00 6.56085968e-01 1.03414580e-01 -2.73369521e-01
-1.15802228e+00 -8.75411570e-01 -4.34574723e-01 -1.30577296e-01
4.77531254e-01 -1.18294239e+00 1.21309078e+00 -9.23498154e-01
-8.52399170e-01 9.95597303e-01 6.64785385e-01 -9.69373643e-01
1.08395112e+00 6.15404770e-02 -5.99881768e-01 3.85026693e-01
-5.50829396e-02 4.24178481e-01 5.88908374e-01 -1.14070344e+00
-8.88670325e-01 -5.12297153e-01 -4.72932577e-01 -1.53831452e-01
-2.08829850e-01 3.37216258e-02 -6.50705546e-02 -5.79611480e-01
-4.41457093e-01 -1.09397674e+00 -7.15776980e-01 2.93681443e-01
-3.33314717e-01 2.81228691e-01 7.95062006e-01 -8.34270835e-01
9.86829460e-01 -2.22558260e+00 -4.09616530e-01 6.66073620e-01
7.11681128e-01 2.53066510e-01 2.87615567e-01 8.39884877e-02
-6.02363423e-02 3.52608770e-01 -3.19196075e-01 6.59954920e-02
-3.87159318e-01 4.32950467e-01 1.44425392e-01 7.59230435e-01
-1.15151614e-01 1.04130077e+00 -8.33324015e-01 -8.71602118e-01
-2.06822511e-02 3.08101922e-01 -6.74185276e-01 3.35789531e-01
1.84540600e-01 5.95379591e-01 -7.30339646e-01 8.67471457e-01
5.69321513e-01 -6.80034816e-01 6.35673821e-01 1.99680477e-01
2.51794606e-01 -9.42448154e-02 -7.98810005e-01 1.48520291e+00
-2.10130349e-01 1.65395975e-01 5.24535477e-01 -6.22762382e-01
5.36933064e-01 8.96326303e-01 1.12944174e+00 -3.93823087e-01
3.68959814e-01 3.17214847e-01 3.45042616e-01 -6.26514673e-01
-1.08005561e-01 -1.82789609e-01 -1.66697696e-01 9.08726215e-01
-1.44645020e-01 8.21135789e-02 -9.37036753e-01 3.91679406e-01
1.57251477e+00 -5.75167596e-01 4.54391420e-01 -1.22943431e-01
2.07824871e-01 2.91490376e-01 4.58592594e-01 9.71421123e-01
-6.60636008e-01 5.30379117e-01 3.07367772e-01 -6.79781020e-01
-9.06590760e-01 -8.45732808e-01 -4.97765958e-01 6.52310133e-01
-5.10053515e-01 -3.56619388e-01 -4.91594285e-01 -1.25912535e+00
1.78273901e-01 5.84798336e-01 -5.91515481e-01 -1.66571245e-01
-4.31704670e-01 -4.69385147e-01 8.89164150e-01 4.51113909e-01
3.58606100e-01 -9.86369312e-01 -1.21908319e+00 3.51197153e-01
-1.69629857e-01 -8.70259464e-01 -3.03264409e-01 3.72500300e-01
-6.33082092e-01 -1.12973905e+00 -7.71729589e-01 -2.73514211e-01
5.15652061e-01 -4.94635813e-02 1.05535913e+00 1.03725120e-01
-6.68757379e-01 5.97623765e-01 -1.73190758e-01 -9.74595547e-01
-7.38582075e-01 5.72539866e-04 -1.53491750e-01 4.68405336e-02
2.52999693e-01 8.70504007e-02 -8.32830369e-01 3.10718507e-01
-1.04248345e+00 -1.37934148e-01 7.95488119e-01 1.11066830e+00
5.06547451e-01 -7.74035603e-02 3.96340519e-01 -1.64767468e+00
3.01949412e-01 -9.87392366e-01 -3.92764866e-01 4.07823861e-01
-9.05073524e-01 -3.52802098e-01 5.55360734e-01 -3.34479064e-02
-1.10527217e+00 3.72911602e-01 1.67945907e-01 -4.63306755e-01
-2.16107905e-01 4.71803278e-01 1.82762250e-01 -1.67519778e-01
7.10103571e-01 -2.16773972e-01 4.45685387e-01 -2.29470469e-02
1.19453162e-01 1.23977864e+00 4.81817991e-01 -3.47040713e-01
3.99189413e-01 6.51725233e-01 -1.37963548e-01 -2.98173279e-01
-4.90116298e-01 -3.91477108e-01 -3.63353133e-01 4.53244485e-02
8.36167097e-01 -1.03014970e+00 -8.04397762e-01 1.06634842e-02
-8.11523318e-01 2.11418137e-01 -5.64641714e-01 6.36501431e-01
-4.81251478e-01 -4.28282470e-02 -5.75917542e-01 -5.79331458e-01
-8.73585939e-01 -1.20348525e+00 7.92215705e-01 -3.47409695e-01
-4.71404493e-01 -8.81910563e-01 3.09537537e-02 6.31316900e-01
6.25264645e-01 5.93837976e-01 1.01019681e+00 -1.28413725e+00
-6.69077873e-01 -4.60295975e-01 -1.54129669e-01 1.66510925e-01
2.10052773e-01 -1.11448266e-01 -1.18830621e+00 -5.13410747e-01
3.49999189e-01 -5.39749086e-01 2.64348298e-01 4.89367604e-01
1.49301350e+00 -8.54537785e-01 -5.67808509e-01 7.03114629e-01
1.31668711e+00 3.67793769e-01 3.30989480e-01 3.68750155e-01
5.70237160e-01 5.54765165e-01 5.24534702e-01 9.75632191e-01
2.81069279e-01 -1.49073787e-02 4.58501697e-01 -3.80487114e-01
4.48739052e-01 -5.99036328e-02 -1.50678739e-01 4.49202269e-01
-6.54625613e-03 -5.24209701e-02 -1.38010252e+00 4.33288187e-01
-1.68459654e+00 -4.58089560e-01 1.31703943e-01 2.06852937e+00
7.77265728e-01 -3.11355125e-02 8.62967372e-02 -9.39169973e-02
3.56349200e-01 -1.92270577e-01 -7.75710344e-01 -6.14017963e-01
-4.85869758e-02 2.53413320e-01 1.03622365e+00 -1.80306062e-01
-1.03832769e+00 1.95945948e-01 6.47819471e+00 5.24702191e-01
-1.08174181e+00 2.59566128e-01 1.42438149e+00 -3.10453862e-01
-1.75815612e-01 -4.57570106e-01 -1.01042219e-01 1.78445965e-01
1.62918270e+00 -4.20441270e-01 -9.34508815e-03 1.08140016e+00
-1.75403878e-01 1.88410208e-01 -1.10436249e+00 9.66339648e-01
-1.88477889e-01 -1.82911301e+00 -6.90158084e-02 5.56178093e-01
8.21881711e-01 4.87116843e-01 2.88073093e-01 -1.29021287e-01
8.00358951e-01 -1.25932527e+00 3.65424335e-01 3.46799791e-01
1.18920946e+00 -8.02409112e-01 1.06664395e+00 3.87015224e-01
-5.01448512e-01 -3.87017846e-01 -8.95954743e-02 5.49319685e-01
-3.24200302e-01 1.14765018e-01 -1.54440343e+00 1.11350703e+00
9.67781603e-01 3.85304868e-01 -5.14078617e-01 8.82083714e-01
4.23719347e-01 8.01529586e-01 -2.71474034e-01 2.67616361e-01
4.02632236e-01 2.55396694e-01 1.74458250e-01 1.20465899e+00
-2.62075029e-02 3.45322669e-01 -4.71765958e-02 2.06296638e-01
-1.97771639e-01 2.28320867e-01 -9.17564571e-01 -3.34528796e-02
3.25731933e-01 1.36644948e+00 -3.53564292e-01 -3.73180389e-01
-5.22505522e-01 5.38995683e-01 -7.66352490e-02 -4.92692143e-02
-8.01019430e-01 1.76625922e-01 3.99754554e-01 1.26641974e-01
1.96084511e-02 4.51332688e-01 -5.46836019e-01 -8.69140327e-01
-3.13827604e-01 -1.60425103e+00 1.35033810e+00 -3.11307222e-01
-1.67145741e+00 8.19461882e-01 -1.00092821e-01 -1.37444460e+00
-4.89854634e-01 -4.76442039e-01 -5.37010208e-02 5.87330282e-01
-1.33630180e+00 -1.27727318e+00 4.52788052e-04 1.04093385e+00
-9.81706157e-02 -3.72723043e-01 1.24743938e+00 2.18550280e-01
-3.73828501e-01 9.36222136e-01 1.22598708e-01 2.50096738e-01
8.31014097e-01 -8.16318393e-01 -4.45681885e-02 2.66792446e-01
-1.03041105e-01 5.15376270e-01 -1.35657908e-02 -5.39834082e-01
-1.53820741e+00 -1.25633860e+00 6.45762742e-01 -6.01228058e-01
5.30258000e-01 -1.25443041e-01 -8.42078388e-01 9.71183836e-01
3.04188460e-01 3.89735043e-01 1.30753183e+00 -4.09526646e-01
-1.39890298e-01 -9.78862792e-02 -1.80428338e+00 2.26825103e-01
6.48319125e-01 -4.19909596e-01 -2.17074841e-01 5.68304062e-01
7.83778846e-01 -3.39096218e-01 -1.27051628e+00 2.00973377e-01
5.06764233e-01 -7.01484263e-01 8.71630907e-01 -9.73302841e-01
2.94649839e-01 2.81492323e-01 -1.40138328e-01 -8.85381222e-01
-4.20742780e-01 -8.92481267e-01 1.80584908e-01 6.70146227e-01
4.33079898e-01 -1.01367426e+00 9.54166055e-01 1.32554328e+00
1.54395998e-01 -9.16773915e-01 -1.23817706e+00 -3.38380307e-01
1.54336169e-01 -4.67685163e-01 9.24645305e-01 1.17601800e+00
-2.10085586e-01 -5.22386432e-01 -2.08035201e-01 2.33322144e-01
6.93406761e-01 -1.47619396e-01 5.42289615e-01 -9.37151492e-01
-3.13601434e-01 7.41555616e-02 -2.75307804e-01 -1.28030732e-01
-2.10076869e-01 -1.01111686e+00 -5.72361112e-01 -1.09430230e+00
3.52123529e-01 -8.84347320e-01 -6.94292963e-01 8.07232022e-01
7.04902261e-02 -2.28307039e-01 1.68522969e-01 6.21227562e-01
-3.50274056e-01 -1.19569518e-01 1.26737440e+00 -1.17582142e-01
2.86627501e-01 2.13063210e-01 -9.92579997e-01 5.09735823e-01
6.79667890e-01 -7.53031492e-01 -1.64797500e-01 -3.07302505e-01
-1.85382456e-01 4.92374867e-01 4.97792810e-01 -6.59076631e-01
6.99061751e-01 -1.18568137e-01 5.48788965e-01 -4.98147607e-01
-1.86284661e-01 -1.71494436e+00 6.69429898e-01 1.07730734e+00
-4.64203954e-01 4.63180803e-02 7.84917399e-02 6.45456135e-01
-8.69236737e-02 2.67151505e-01 5.84555030e-01 -3.91241729e-01
-6.71060011e-02 5.62074542e-01 -1.90070555e-01 -1.14166044e-01
1.47172773e+00 -2.07369607e-02 -3.30471247e-01 -2.00362802e-01
-5.26838660e-01 3.79783273e-01 2.51783043e-01 1.52674364e-02
6.39961123e-01 -8.59531760e-01 -1.09297061e+00 6.41156256e-01
2.09765077e-01 2.76956558e-01 3.79724830e-01 8.84682298e-01
-6.16927922e-01 5.65127611e-01 -1.02241397e-01 -7.42357135e-01
-1.16456795e+00 7.37526417e-01 2.72971928e-01 -4.53292459e-01
-8.45735610e-01 7.20575452e-01 4.30326797e-02 -5.32752812e-01
4.00993973e-01 -1.14226807e-02 4.96029198e-01 -1.16341308e-01
4.16470200e-01 3.46678883e-01 4.14447516e-01 -2.91239977e-01
-5.43572009e-01 -2.80281186e-01 -5.40538549e-01 -3.20649403e-03
1.60570300e+00 1.38265118e-01 1.00536766e-02 1.03364401e-01
1.21912229e+00 -2.10050389e-01 -1.01550829e+00 -2.71497577e-01
-5.54768853e-02 -5.36201179e-01 1.57240592e-02 -1.13255203e+00
-1.36700821e+00 3.66552800e-01 6.77687168e-01 1.51132951e-02
1.17449009e+00 2.02377010e-02 7.76682913e-01 1.08753800e-01
6.04581892e-01 -6.59882367e-01 -3.96113336e-01 -2.96421111e-01
5.57147682e-01 -1.29765081e+00 3.37857187e-01 4.84182686e-02
-1.40731263e+00 9.18783784e-01 2.11114705e-01 2.59672225e-01
9.96657312e-01 5.36633968e-01 4.33429480e-01 -3.61728936e-01
-1.10004210e+00 8.19282889e-01 -1.50685348e-02 5.62085152e-01
7.02149943e-02 5.39786696e-01 2.20436037e-01 8.49065781e-01
-1.51048154e-01 3.23285490e-01 3.97073299e-01 1.45489872e+00
3.89541864e-01 -9.48655844e-01 -5.86426020e-01 8.31441224e-01
-8.88038099e-01 6.29699528e-02 -4.98128623e-01 8.02623391e-01
1.61486253e-01 6.77065015e-01 -1.55047759e-01 -2.88099945e-01
3.00996929e-01 -1.97793432e-02 -1.51608527e-01 -3.56895894e-01
-1.35201514e+00 1.87411800e-01 8.35113302e-02 -6.95447326e-01
-1.62146196e-01 -8.00487399e-01 -1.15008140e+00 -3.98568183e-01
-2.20947623e-01 2.60297190e-02 6.09631777e-01 3.23628068e-01
6.38839126e-01 5.00954986e-01 8.91987622e-01 8.05133879e-02
-9.52232540e-01 -2.11458027e-01 -6.15588605e-01 2.88385332e-01
5.35641611e-01 1.11688465e-01 -3.64754051e-01 -4.22636829e-02] | [6.151970863342285, 6.563205718994141] |
bf131733-4cc1-4fb1-882f-2af0ee9bc168 | do-as-i-can-not-as-i-say-grounding-language | 2204.01691 | null | https://arxiv.org/abs/2204.01691v2 | https://arxiv.org/pdf/2204.01691v2.pdf | Do As I Can, Not As I Say: Grounding Language in Robotic Affordances | Large language models can encode a wealth of semantic knowledge about the world. Such knowledge could be extremely useful to robots aiming to act upon high-level, temporally extended instructions expressed in natural language. However, a significant weakness of language models is that they lack real-world experience, which makes it difficult to leverage them for decision making within a given embodiment. For example, asking a language model to describe how to clean a spill might result in a reasonable narrative, but it may not be applicable to a particular agent, such as a robot, that needs to perform this task in a particular environment. We propose to provide real-world grounding by means of pretrained skills, which are used to constrain the model to propose natural language actions that are both feasible and contextually appropriate. The robot can act as the language model's "hands and eyes," while the language model supplies high-level semantic knowledge about the task. We show how low-level skills can be combined with large language models so that the language model provides high-level knowledge about the procedures for performing complex and temporally-extended instructions, while value functions associated with these skills provide the grounding necessary to connect this knowledge to a particular physical environment. We evaluate our method on a number of real-world robotic tasks, where we show the need for real-world grounding and that this approach is capable of completing long-horizon, abstract, natural language instructions on a mobile manipulator. The project's website and the video can be found at https://say-can.github.io/. | ['Andy Zeng', 'Mengyuan Yan', 'Sichun Xu', 'Peng Xu', 'Ted Xiao', 'Fei Xia', 'Vincent Vanhoucke', 'Alexander Toshev', 'Clayton Tan', 'Nicolas Sievers', 'Pierre Sermanet', 'Diego Reyes', 'Jarek Rettinghouse', 'Kanishka Rao', 'Jornell Quiambao', 'Peter Pastor', 'Carolina Parada', 'Linda Luu', 'Yao Lu', 'Sergey Levine', 'Kuang-Huei Lee', 'Yuheng Kuang', 'Dmitry Kalashnikov', 'Ryan Julian', 'Nikhil J Joshi', 'Sally Jesmonth', 'Kyle Jeffrey', 'Rosario Jauregui Ruano', 'Eric Jang', 'Alex Irpan', 'Brian Ichter', 'Julian Ibarz', 'Jasmine Hsu', 'Daniel Ho', 'Alex Herzog', 'Karol Hausman', 'Keerthana Gopalakrishnan', 'Chuyuan Fu', 'Chelsea Finn', 'Byron David', 'Omar Cortes', 'Yevgen Chebotar', 'Noah Brown', 'Anthony Brohan', 'Michael Ahn'] | 2022-04-04 | null | null | null | null | ['robot-task-planning'] | ['robots'] | [ 3.89653817e-02 4.54865456e-01 -1.73864678e-01 -1.79336473e-01
-2.66602457e-01 -7.95022011e-01 6.95955992e-01 1.09621145e-01
-3.43436331e-01 6.35810733e-01 1.16308644e-01 -5.92515171e-01
-2.58596629e-01 -9.18366790e-01 -8.40811253e-01 -2.37924308e-01
-1.80147201e-01 5.54170609e-01 2.41878226e-01 -4.32717025e-01
2.86381066e-01 7.54597962e-01 -1.44630182e+00 1.57354876e-01
5.23272574e-01 5.16627312e-01 9.60723281e-01 5.46288848e-01
-6.10419028e-02 9.93983626e-01 -1.87906414e-01 3.45329940e-01
1.53865725e-01 -2.90807009e-01 -1.24573421e+00 1.45978421e-01
-4.10963833e-01 -3.00784886e-01 -2.34738141e-01 9.47008908e-01
-2.51974404e-01 3.89282972e-01 5.76594532e-01 -1.41057408e+00
-2.39899680e-01 5.46727777e-01 2.36942336e-01 -3.08993995e-01
7.91058660e-01 5.51265657e-01 7.58608937e-01 -3.32107365e-01
8.77846897e-01 1.40820253e+00 2.17102095e-01 6.93320274e-01
-9.90536690e-01 -7.79101104e-02 3.91076654e-01 -1.30870296e-02
-1.13154936e+00 -2.68013328e-01 5.68400800e-01 -4.64171797e-01
1.35020602e+00 -1.31611228e-01 7.62519062e-01 9.65591073e-01
5.81194997e-01 4.90496874e-01 1.05009449e+00 -5.70508242e-01
5.26968002e-01 1.92272104e-02 -4.72177491e-02 8.42209220e-01
3.24084796e-02 5.35108268e-01 -4.98375833e-01 8.99340063e-02
1.13297844e+00 -4.41080034e-02 5.87257277e-03 -7.70807028e-01
-1.49597573e+00 5.45784593e-01 5.07740140e-01 6.03797317e-01
-5.46356797e-01 5.50978005e-01 2.93736368e-01 4.27969217e-01
-2.13025063e-01 1.07158136e+00 -6.01564765e-01 -2.76902735e-01
-1.47874936e-01 2.90879607e-01 1.04149640e+00 1.03533041e+00
9.94722724e-01 -1.78959489e-01 3.33446443e-01 2.51373053e-01
3.52705330e-01 2.80958891e-01 2.33524561e-01 -1.74616778e+00
4.03565794e-01 6.48015499e-01 6.81653678e-01 -7.57246673e-01
-7.07513809e-01 1.23027004e-02 -9.57074612e-02 7.16162324e-01
3.76436412e-01 -2.06455901e-01 -8.80471230e-01 1.94658601e+00
1.48044020e-01 8.58373269e-02 5.19282222e-01 1.11609674e+00
2.11170197e-01 9.81210530e-01 1.61312371e-01 -5.35275452e-02
1.46938229e+00 -8.35391700e-01 -5.51834702e-01 -9.06042874e-01
1.10788321e+00 -2.63515562e-01 1.30497432e+00 3.11505526e-01
-1.22844243e+00 -4.12776768e-01 -1.04161417e+00 -6.68721944e-02
-3.80274534e-01 -1.81541800e-01 1.08576453e+00 -5.70544624e-04
-1.19473934e+00 5.21834731e-01 -1.05870306e+00 -8.70116830e-01
-1.07485816e-01 4.60216820e-01 -5.86076081e-01 -1.23687699e-01
-1.11937737e+00 1.52833593e+00 6.96589112e-01 1.97222859e-01
-1.48631835e+00 -1.48635881e-03 -1.23997092e+00 4.38605510e-02
8.98387730e-01 -9.21923816e-01 1.71096385e+00 -8.23224664e-01
-1.63667905e+00 7.66815662e-01 -2.73900684e-02 -4.27319229e-01
8.99377838e-02 -1.50201395e-01 -7.36184940e-02 2.64618129e-01
3.44947189e-01 9.38901901e-01 2.85815477e-01 -1.33193922e+00
-4.93312657e-01 -2.42797449e-01 1.04738259e+00 5.37589192e-01
1.81054413e-01 3.32611874e-02 -3.83774221e-01 -7.50918835e-02
2.69774497e-01 -1.23185313e+00 -6.54781103e-01 -4.99966107e-02
-2.16643829e-02 -1.37397408e-01 3.25598449e-01 -2.63204694e-01
6.17479503e-01 -1.94626772e+00 5.03253281e-01 1.26017496e-01
-3.34600866e-01 -2.99757898e-01 -2.51480997e-01 7.20377266e-01
1.51681125e-01 1.13705555e-02 -1.48757830e-01 1.00110754e-01
3.94657314e-01 6.21669888e-01 -2.76775271e-01 3.15416723e-01
2.37206548e-01 8.23239207e-01 -1.24974704e+00 -2.59986460e-01
4.16843832e-01 -7.11400509e-02 -6.70474350e-01 3.48225802e-01
-7.62942612e-01 7.05124378e-01 -8.38224590e-01 1.02062680e-01
-1.96613207e-01 -1.03202667e-02 4.26555455e-01 4.02701944e-01
-1.43389016e-01 5.29570758e-01 -1.20114768e+00 2.22273397e+00
-1.09749067e+00 2.52879649e-01 3.42411160e-01 -7.78943956e-01
6.65618539e-01 3.92188370e-01 1.49579570e-01 -3.74912679e-01
1.35876656e-01 2.85367340e-01 9.05192122e-02 -7.17704952e-01
2.99636424e-01 -5.72013497e-01 -6.47921443e-01 6.03627861e-01
-1.96894437e-01 -1.04879832e+00 2.92462111e-01 1.77760884e-01
1.13239682e+00 7.81761467e-01 4.11298811e-01 -2.45313957e-01
4.13890660e-01 4.79923099e-01 3.49021912e-01 5.39410889e-01
1.09359901e-02 -9.77001190e-02 4.43978161e-01 -4.49305296e-01
-9.52566504e-01 -8.32485020e-01 4.49646771e-01 1.07000732e+00
4.77186143e-01 -2.87953049e-01 -6.82667911e-01 -2.53140509e-01
-5.23929536e-01 1.21462405e+00 -3.59166861e-01 -3.40107679e-01
-6.70397341e-01 2.91876376e-01 1.35810494e-01 4.87324417e-01
2.74823725e-01 -1.62093663e+00 -1.52530408e+00 2.65294403e-01
-2.34304592e-01 -1.09802830e+00 -1.98480621e-01 5.31222701e-01
-9.49698448e-01 -9.96012866e-01 -1.33078722e-02 -9.96277809e-01
8.50130856e-01 1.72564127e-02 9.51714993e-01 1.23903573e-01
5.69244511e-02 8.39204967e-01 -2.61213362e-01 -2.76381642e-01
-7.12398469e-01 -1.05555154e-01 1.55695960e-01 -6.88615143e-01
1.24065258e-01 -5.21692932e-01 -8.98662060e-02 2.60352314e-01
-8.41821373e-01 4.42747414e-01 5.01711607e-01 6.35629594e-01
2.34649420e-01 4.19335067e-01 2.82693297e-01 -3.49323362e-01
6.95046902e-01 -3.13153744e-01 -6.19709373e-01 1.17045715e-01
-2.41623089e-01 3.78984839e-01 5.35253286e-01 -3.45773518e-01
-1.26215041e+00 2.36549139e-01 2.11893767e-01 2.89389547e-02
-5.81538856e-01 8.93811107e-01 -3.24930012e-01 3.57738853e-01
5.79704046e-01 5.10603376e-02 -4.47758660e-02 -8.57529193e-02
5.56855619e-01 2.55582392e-01 5.87728202e-01 -1.45666337e+00
5.37230194e-01 2.74581999e-01 2.47676924e-01 -4.06967729e-01
-4.95892048e-01 -2.74714321e-01 -7.17814565e-01 1.49522321e-02
8.73869717e-01 -7.61571229e-01 -9.22946751e-01 1.41963452e-01
-1.21836591e+00 -1.11966074e+00 -3.59508485e-01 5.00797153e-01
-1.46235347e+00 1.67584531e-02 -4.71876532e-01 -8.91923308e-01
4.39419329e-01 -1.39765859e+00 1.01426435e+00 -2.53010001e-02
-6.35132432e-01 -9.48387027e-01 -2.47330904e-01 -4.59623616e-03
1.40281990e-01 1.04410484e-01 1.00732422e+00 -3.82247597e-01
-6.62829459e-01 -2.22467333e-02 2.21614450e-01 8.49091709e-02
3.03312480e-01 -3.60012591e-01 -4.24812645e-01 -1.45009145e-01
1.71973139e-01 -5.36269307e-01 2.53561921e-02 3.10192436e-01
8.54758918e-01 -2.96887010e-01 -5.23706496e-01 -1.21572344e-02
1.27544129e+00 5.14441609e-01 6.85619235e-01 5.61370254e-01
1.53149262e-01 9.57896233e-01 1.09325433e+00 1.50716186e-01
4.59695637e-01 8.75097871e-01 5.22461653e-01 3.08403581e-01
1.51205093e-01 -3.93514216e-01 5.93590856e-01 3.20554912e-01
-2.18072012e-01 1.21146046e-01 -1.16684437e+00 6.97816968e-01
-2.07305121e+00 -9.19008434e-01 2.67614216e-01 1.88777947e+00
8.80375564e-01 2.31505722e-01 -1.65816993e-01 -2.26856902e-01
3.62928599e-01 -1.13539770e-01 -6.66309476e-01 -6.95250630e-01
5.00055194e-01 -1.05456494e-01 1.73490971e-01 8.68304610e-01
-7.43477464e-01 1.24051356e+00 6.00370550e+00 3.62031043e-01
-8.19954097e-01 2.99582668e-02 1.32622302e-01 1.53622851e-01
-2.33482510e-01 5.31932294e-01 -3.57601404e-01 -2.80569401e-02
1.14680386e+00 -3.20950210e-01 8.64972472e-01 9.20397460e-01
5.47926068e-01 -5.17781079e-01 -1.66028273e+00 5.66186786e-01
-4.18345064e-01 -1.19726396e+00 -1.58382505e-01 -4.61387523e-02
2.56423861e-01 -2.20720530e-01 -2.86616445e-01 4.34385955e-01
7.64110863e-01 -1.08077085e+00 1.14787829e+00 3.63590151e-01
5.56950212e-01 -6.55562639e-01 4.85049903e-01 1.04820073e+00
-1.11021304e+00 -3.49941671e-01 -1.31208628e-01 -7.17330694e-01
4.24859434e-01 -2.72019565e-01 -9.14766312e-01 4.71742630e-01
3.71027887e-01 4.07839388e-01 1.48957297e-01 4.65999752e-01
-7.01293707e-01 -1.16400160e-01 -4.18161690e-01 -2.02948481e-01
4.79315281e-01 -2.87109017e-01 4.15013641e-01 6.71460748e-01
4.84533697e-01 4.83960152e-01 5.22817492e-01 9.60920811e-01
4.74689871e-01 -3.59056860e-01 -1.06007361e+00 -1.33620098e-01
3.47913921e-01 7.53174961e-01 -7.64024556e-01 -3.91497314e-01
-2.46095300e-01 7.91973710e-01 2.46405557e-01 5.97223461e-01
-7.06470311e-01 -9.64835435e-02 6.77778900e-01 -3.05670258e-02
-2.54578292e-01 -7.78773546e-01 -1.28240496e-01 -8.76461506e-01
-1.07928112e-01 -9.59401131e-01 3.76959816e-02 -1.53163409e+00
-6.11813843e-01 2.88293451e-01 5.08526444e-01 -1.07317722e+00
-6.96744740e-01 -8.47290397e-01 -3.36867988e-01 7.90778518e-01
-1.26011503e+00 -1.07820189e+00 -1.32356495e-01 4.49341774e-01
6.82871163e-01 1.96426753e-02 1.00021589e+00 -4.63228226e-01
8.16116780e-02 -4.72128570e-01 -8.64130318e-01 -2.54634917e-01
2.41430923e-01 -9.65210617e-01 2.00425535e-01 7.00721383e-01
-7.71376416e-02 1.06374693e+00 1.13559818e+00 -6.42206848e-01
-1.75012326e+00 -8.11535537e-01 7.63122618e-01 -6.80796742e-01
8.52471530e-01 -1.65807948e-01 -6.75923347e-01 1.29281008e+00
2.09043901e-02 -2.21751764e-01 1.33705333e-01 -2.86806338e-02
-4.31984216e-02 2.49618322e-01 -1.15187621e+00 1.05769110e+00
1.38821757e+00 -6.06168747e-01 -1.25157297e+00 4.55878586e-01
8.98197770e-01 -6.22676253e-01 -6.51781738e-01 2.17505366e-01
3.24881226e-01 -6.72987103e-01 9.10851479e-01 -8.14441562e-01
4.17494595e-01 -7.05808342e-01 -2.60684162e-01 -1.41001582e+00
-3.25200230e-01 -5.96208334e-01 1.78603500e-01 6.42205417e-01
4.50567245e-01 -8.56504261e-01 4.23533916e-01 9.99767661e-01
-4.84447420e-01 -4.30357873e-01 -7.55827785e-01 -1.07562780e+00
2.25183636e-01 -8.62190902e-01 6.82767034e-01 6.57650828e-01
7.37723351e-01 3.05925995e-01 3.86230573e-02 5.60912490e-01
2.80708559e-02 2.15894580e-01 7.99655795e-01 -9.11596775e-01
-3.25584173e-01 -3.35144281e-01 -1.09630845e-01 -1.23610759e+00
7.94708908e-01 -9.23943877e-01 5.91033280e-01 -2.02014637e+00
-1.11434147e-01 -7.26356149e-01 7.64229149e-02 9.40287292e-01
5.34821987e-01 -5.42686284e-01 4.91497129e-01 1.78258002e-01
-4.94215310e-01 3.45400155e-01 1.44718623e+00 -7.98227340e-02
-3.44849229e-01 -1.85877264e-01 -5.81238747e-01 9.89721179e-01
8.58272374e-01 -2.41400957e-01 -7.66266048e-01 -5.18362880e-01
5.96947849e-01 4.95170981e-01 4.89109039e-01 -9.23018634e-01
4.82016861e-01 -8.86376560e-01 -2.21710771e-01 1.87599510e-01
5.64425588e-01 -1.31877172e+00 4.11073983e-01 8.35498452e-01
-4.52903807e-01 7.23149925e-02 4.89007533e-01 3.80089611e-01
-1.54285282e-01 -4.59252179e-01 3.71458679e-01 -6.83162034e-01
-1.36485124e+00 -2.00245097e-01 -7.27782905e-01 -4.01995271e-01
1.39953804e+00 -2.94699132e-01 -2.02157706e-01 -4.58741099e-01
-9.44836438e-01 5.36716282e-01 9.82142448e-01 4.06844229e-01
6.41797841e-01 -1.04819465e+00 -1.39493644e-01 -4.59843613e-02
2.40829229e-01 1.13290690e-01 -8.60146582e-02 4.85074520e-01
-5.25494874e-01 5.95711529e-01 -4.96389091e-01 -3.81492138e-01
-7.61994779e-01 8.97035420e-01 3.84030551e-01 -1.38444290e-01
-7.20917106e-01 4.11083221e-01 5.83698452e-01 -7.05347419e-01
-1.13941476e-01 -9.02876496e-01 -8.89182761e-02 -5.52962661e-01
2.88459241e-01 -1.85055137e-01 -3.57896030e-01 -4.90645707e-01
-6.00648403e-01 8.18265557e-01 5.62985122e-01 -5.45816541e-01
1.19987857e+00 -2.63228595e-01 -3.45684141e-01 6.19271100e-01
4.65941727e-01 -2.93362051e-01 -1.39779222e+00 2.94075022e-03
1.71950623e-01 -1.51545241e-01 -2.51830995e-01 -7.92462230e-01
-3.66686523e-01 8.49042118e-01 -1.06974229e-01 6.57838434e-02
9.25360858e-01 3.56256813e-01 3.48870248e-01 9.73201096e-01
1.37620533e+00 -1.11472726e+00 2.83256024e-01 7.85957456e-01
1.20944476e+00 -8.63459110e-01 -1.11673787e-01 -4.01039273e-01
-7.11368680e-01 1.11073971e+00 6.83605850e-01 -5.25143966e-02
2.19157070e-01 4.07462031e-01 1.69008467e-02 -3.81828696e-01
-9.50627327e-01 -2.09074587e-01 -2.18191296e-01 7.59678662e-01
-4.93951328e-03 1.79818958e-01 1.95403025e-03 4.21238244e-01
-2.82691389e-01 2.87004054e-01 8.33706856e-01 1.35966992e+00
-6.93359733e-01 -1.07288671e+00 -4.44232553e-01 -2.70690434e-02
6.33443743e-02 2.64983565e-01 -1.54786348e-01 7.94979453e-01
1.36226818e-01 1.10816002e+00 -5.03461547e-02 4.40107845e-02
4.10348028e-01 1.91710740e-01 7.95434713e-01 -1.35147119e+00
-1.36033803e-01 -2.27199405e-01 5.83016038e-01 -1.01083219e+00
-5.27432799e-01 -6.23162270e-01 -2.19256687e+00 -2.11260282e-02
2.54569143e-01 1.58109382e-01 6.33618832e-01 1.23009551e+00
1.68182343e-01 4.99419838e-01 6.26652688e-02 -1.28595555e+00
-4.16804969e-01 -5.93697548e-01 -3.89160663e-01 4.00057226e-01
2.59297878e-01 -7.64441192e-01 -2.42785573e-01 3.55648100e-01] | [4.449063777923584, 0.9351276755332947] |
991f5bfd-d7cb-4c60-9129-d457492d8303 | editable-graph-neural-network-for-node | 2305.15529 | null | https://arxiv.org/abs/2305.15529v1 | https://arxiv.org/pdf/2305.15529v1.pdf | Editable Graph Neural Network for Node Classifications | Despite Graph Neural Networks (GNNs) have achieved prominent success in many graph-based learning problem, such as credit risk assessment in financial networks and fake news detection in social networks. However, the trained GNNs still make errors and these errors may cause serious negative impact on society. \textit{Model editing}, which corrects the model behavior on wrongly predicted target samples while leaving model predictions unchanged on unrelated samples, has garnered significant interest in the fields of computer vision and natural language processing. However, model editing for graph neural networks (GNNs) is rarely explored, despite GNNs' widespread applicability. To fill the gap, we first observe that existing model editing methods significantly deteriorate prediction accuracy (up to $50\%$ accuracy drop) in GNNs while a slight accuracy drop in multi-layer perception (MLP). The rationale behind this observation is that the node aggregation in GNNs will spread the editing effect throughout the whole graph. This propagation pushes the node representation far from its original one. Motivated by this observation, we propose \underline{E}ditable \underline{G}raph \underline{N}eural \underline{N}etworks (EGNN), a neighbor propagation-free approach to correct the model prediction on misclassified nodes. Specifically, EGNN simply stitches an MLP to the underlying GNNs, where the weights of GNNs are frozen during model editing. In this way, EGNN disables the propagation during editing while still utilizing the neighbor propagation scheme for node prediction to obtain satisfactory results. Experiments demonstrate that EGNN outperforms existing baselines in terms of effectiveness (correcting wrong predictions with lower accuracy drop), generalizability (correcting wrong predictions for other similar nodes), and efficiency (low training time and memory) on various graph datasets. | ['Xia Hu', 'Soo-Hyun Choi', 'Rui Chen', 'Li Li', 'Kaixiong Zhou', 'Shaochen Zhong', 'Zhimeng Jiang', 'Zirui Liu'] | 2023-05-24 | null | null | null | null | ['fake-news-detection', 'model-editing'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.21006322e-01 8.41699362e-01 -3.35010231e-01 -3.15200567e-01
1.74957111e-01 -2.86968231e-01 4.32866782e-01 4.22340482e-01
-3.56129818e-02 8.28426838e-01 -3.36143374e-01 -4.57114160e-01
6.77425563e-02 -1.30040312e+00 -1.05124283e+00 -3.37862402e-01
-9.25078392e-02 3.74404252e-01 3.22435379e-01 -3.10986578e-01
2.27220014e-01 4.48758870e-01 -7.10771501e-01 6.17999248e-02
1.10630035e+00 7.80073762e-01 -2.76826531e-01 4.53751475e-01
-9.29934084e-02 1.10784149e+00 -7.12090790e-01 -1.17198598e+00
1.19997874e-01 -4.25241262e-01 -6.51769221e-01 -2.94358999e-01
3.77205342e-01 -1.89622626e-01 -5.64501762e-01 1.49425268e+00
1.20544575e-01 -2.56667528e-02 6.20617151e-01 -1.58893836e+00
-1.16705596e+00 1.17619669e+00 -7.79720366e-01 -1.47598544e-02
-9.02178138e-02 -1.60911933e-01 1.08154380e+00 -7.53039360e-01
5.91522515e-01 1.10760987e+00 1.25883901e+00 4.83951986e-01
-1.15462863e+00 -8.89698386e-01 3.76002580e-01 2.95878835e-02
-1.43584001e+00 -1.71732798e-01 8.66524279e-01 -1.34418949e-01
9.44049120e-01 3.01555336e-01 6.78701878e-01 1.30067682e+00
5.54911435e-01 5.63049018e-01 5.11449337e-01 -3.05151016e-01
-3.84787135e-02 2.39314526e-01 1.50625795e-01 9.34068382e-01
7.58377016e-01 1.68441623e-01 -4.14245635e-01 -1.56787038e-01
8.04150403e-01 1.47531182e-01 -1.63946643e-01 -4.35488559e-02
-7.33787715e-01 8.97617340e-01 1.01880443e+00 5.38286716e-02
-4.27043647e-01 4.66748655e-01 2.67781258e-01 4.67698127e-01
9.25900698e-01 4.53638136e-01 -3.37885857e-01 2.45650187e-01
-8.88767838e-01 1.33935660e-01 8.83851349e-01 8.91179025e-01
8.06419849e-01 5.68482816e-01 -2.83696763e-02 8.50461066e-01
3.35790813e-01 4.08831716e-01 2.53871560e-01 -5.51787913e-01
4.84672695e-01 1.00631368e+00 -3.51024240e-01 -1.86575723e+00
-5.10013342e-01 -1.04143143e+00 -1.46674812e+00 1.41963601e-01
3.41208100e-01 -3.55132967e-01 -8.80404949e-01 1.65243161e+00
1.36618823e-01 5.21567225e-01 -1.68501139e-01 5.59404314e-01
8.52693617e-01 5.33151627e-01 1.54040918e-01 9.35313404e-02
7.13254035e-01 -1.26721239e+00 -4.26573843e-01 -5.25084198e-01
9.01718855e-01 -3.86414707e-01 7.31158316e-01 3.05265754e-01
-7.00682342e-01 -2.04885677e-01 -9.51736033e-01 2.49149501e-01
-5.64328134e-01 -8.52873251e-02 7.00556278e-01 7.73142099e-01
-1.20819378e+00 1.17993057e+00 -6.25886500e-01 -2.50213265e-01
7.46122777e-01 4.88076121e-01 -2.47938022e-01 -3.46761458e-02
-1.35654449e+00 7.37064958e-01 4.00963724e-01 3.66492331e-01
-5.43252647e-01 -6.43689752e-01 -7.40877211e-01 1.28650591e-01
4.55299139e-01 -4.40423250e-01 7.77334929e-01 -1.31441140e+00
-1.16290009e+00 5.13549447e-01 1.70209259e-01 -7.50196815e-01
7.50310183e-01 1.34646848e-01 -8.80087435e-01 -3.22631598e-01
-7.76493549e-02 4.70541775e-01 9.27626729e-01 -1.13637543e+00
-2.66431034e-01 -2.25846440e-01 -8.15035552e-02 -8.16064104e-02
-5.30543089e-01 -3.79413843e-01 -2.06913814e-01 -1.07665145e+00
2.54281610e-01 -9.58786726e-01 -2.76759952e-01 -8.61412361e-02
-7.88489938e-01 -1.54339103e-02 7.84145355e-01 -7.82461405e-01
1.36935294e+00 -1.76846397e+00 -2.51745284e-01 7.51179099e-01
8.49976242e-01 5.05328119e-01 -1.82223842e-01 4.11789268e-01
-6.50556609e-02 5.06610155e-01 -2.04839826e-01 -2.95186371e-01
-2.07021117e-01 3.37840050e-01 -1.47751123e-01 3.78204882e-01
4.19229306e-02 1.28664064e+00 -8.49476218e-01 -3.19811761e-01
-1.97969645e-01 4.16813165e-01 -6.17862463e-01 -2.74118572e-01
-2.30606645e-01 1.07587732e-01 -2.62837261e-01 7.21843958e-01
7.68753052e-01 -7.32464135e-01 3.21137249e-01 -6.28391951e-02
6.57572150e-01 5.96158952e-02 -9.82885361e-01 8.45084965e-01
6.19857088e-02 4.70118523e-01 -9.39914137e-02 -1.15728724e+00
1.19406223e+00 -1.14402361e-01 2.08277777e-01 -5.76883733e-01
5.36394790e-02 1.46341503e-01 3.97290066e-02 -5.44268712e-02
6.58107817e-01 2.22964898e-01 1.81425169e-01 5.25460124e-01
-1.67016894e-01 2.79070914e-01 -1.00103855e-01 5.24202704e-01
1.23356843e+00 -2.42963389e-01 1.51376426e-01 6.83538541e-02
2.06493005e-01 -3.25489417e-02 5.57057202e-01 1.11008871e+00
-9.85465422e-02 3.67708087e-01 7.13857055e-01 -4.62240785e-01
-8.77159476e-01 -7.28158474e-01 3.99934739e-01 1.17115927e+00
1.10739537e-01 -4.03984308e-01 -8.32539916e-01 -1.14900863e+00
1.19168103e-01 8.53957832e-01 -7.42798805e-01 -6.02508247e-01
-5.04122257e-01 -1.03739142e+00 1.02920306e+00 4.40922379e-01
6.44155502e-01 -1.30960059e+00 3.48765820e-01 3.51023883e-01
8.95511806e-02 -8.97257090e-01 -2.85772622e-01 -1.44199327e-01
-9.80231404e-01 -1.07556498e+00 -3.83315504e-01 -6.59418344e-01
1.08137858e+00 6.04331493e-02 1.23810208e+00 7.36259997e-01
3.55026007e-01 -1.13120407e-01 -3.75114024e-01 -4.44206536e-01
-8.68754983e-01 4.93832141e-01 -3.24600898e-02 -3.05403434e-02
2.98855126e-01 -7.28189707e-01 -1.81254342e-01 2.54534900e-01
-7.46809661e-01 1.42602399e-01 6.70835435e-01 8.48776281e-01
2.88339823e-01 1.06733374e-01 8.24302793e-01 -1.79340541e+00
9.16114867e-01 -6.17050767e-01 -4.70190763e-01 3.91772509e-01
-1.36892021e+00 -1.79946333e-01 8.19425046e-01 -4.34118599e-01
-5.75087249e-01 -4.38826412e-01 -5.37242703e-02 -3.02972168e-01
3.64980280e-01 8.96095991e-01 2.02292562e-01 -5.80278635e-01
9.07985270e-01 1.28293633e-01 4.47695628e-02 -2.83713937e-01
4.94210757e-02 2.97946185e-01 3.36525947e-01 -5.89351691e-02
9.38276887e-01 5.22382706e-02 8.74543637e-02 -5.57784736e-01
-6.90860868e-01 3.52023602e-01 -3.67944002e-01 -3.43289047e-01
2.47767732e-01 -6.65506005e-01 -5.76089144e-01 6.98071957e-01
-1.21126974e+00 -3.33601326e-01 -9.45822895e-02 1.94581136e-01
2.47968987e-01 3.94830376e-01 -8.43610704e-01 -5.56089938e-01
-4.97687310e-01 -6.39295638e-01 2.81141609e-01 2.35740349e-01
-2.79020041e-01 -1.39115739e+00 -3.74914169e-01 1.84409693e-01
5.48543334e-01 3.67199898e-01 1.01444399e+00 -1.04807365e+00
-4.62941229e-01 -5.74556947e-01 -7.01197982e-01 5.08232951e-01
-3.87775782e-03 4.87902649e-02 -7.68310845e-01 -3.02633852e-01
-5.55171192e-01 1.43371031e-01 8.18020344e-01 2.93974280e-01
1.23775351e+00 -7.15990365e-01 -6.03847504e-01 4.38765407e-01
1.14666378e+00 -1.65039018e-01 5.57448864e-01 8.15384686e-02
1.26621997e+00 2.20647156e-01 -4.88706902e-02 1.87419683e-01
3.91996622e-01 1.62493736e-01 8.94498289e-01 -2.46399865e-01
-1.12300597e-01 -5.71098328e-01 3.38059008e-01 7.44439781e-01
-1.58658490e-01 -7.81674147e-01 -9.54811633e-01 3.20304960e-01
-1.83703470e+00 -6.80745482e-01 -4.62440908e-01 2.03004861e+00
5.98960817e-01 5.39667070e-01 -1.80099890e-01 1.99849624e-02
1.01217914e+00 3.21439177e-01 -7.29518890e-01 -4.32412773e-01
-1.62762001e-01 -4.81483787e-02 8.39839995e-01 6.24949694e-01
-8.04680228e-01 1.10429788e+00 5.51050043e+00 7.82384694e-01
-1.04417503e+00 1.00848794e-01 8.89346302e-01 2.57200181e-01
-4.16565835e-01 2.53080558e-02 -7.11974263e-01 4.46163088e-01
8.84114146e-01 -1.89372510e-01 6.33751214e-01 9.33613360e-01
9.80111286e-02 2.60617375e-01 -8.89961302e-01 5.94005644e-01
1.27190724e-01 -1.73552942e+00 4.22082096e-01 5.88880628e-02
8.45907748e-01 2.33283237e-01 -8.36595222e-02 5.00563145e-01
6.90034091e-01 -1.27600503e+00 4.91989821e-01 6.58261716e-01
5.49595237e-01 -6.64904892e-01 9.25618052e-01 4.71544147e-01
-9.07778680e-01 9.96625274e-02 -5.06093144e-01 -2.26614684e-01
-5.54517889e-03 9.97217834e-01 -1.16119361e+00 5.93045473e-01
4.37246025e-01 9.29837942e-01 -8.28851223e-01 7.71692634e-01
-2.78310657e-01 9.22138095e-01 -2.00257942e-01 -2.41099611e-01
1.28462076e-01 -2.01077759e-01 6.59037769e-01 9.20092404e-01
1.92248762e-01 -1.21747255e-01 2.95297932e-02 1.07630265e+00
-8.05919707e-01 -8.98208562e-03 -5.78459322e-01 -1.68991476e-01
5.07014692e-01 9.38700736e-01 -7.34263778e-01 -3.74234200e-01
-1.99753121e-01 1.01821399e+00 6.27511501e-01 5.71339130e-01
-1.01828420e+00 -5.00070214e-01 9.59516093e-02 4.24379975e-01
3.18402387e-02 1.34160876e-01 -5.07114351e-01 -9.70396638e-01
7.74875656e-02 -6.46235049e-01 2.50810444e-01 -5.73885918e-01
-1.46623909e+00 7.15944529e-01 -4.63081658e-01 -9.03776705e-01
-1.59449391e-02 -3.81559968e-01 -6.62689865e-01 7.47053981e-01
-1.38362885e+00 -1.19500899e+00 -3.60045969e-01 3.94441277e-01
6.39647022e-02 -2.94745535e-01 6.05262876e-01 4.44364607e-01
-7.99078643e-01 1.07957566e+00 3.09101958e-02 4.81636524e-01
5.62038422e-01 -1.05661809e+00 7.92069077e-01 9.70043719e-01
2.03376934e-01 4.17640477e-01 5.67848563e-01 -1.24112213e+00
-7.29305327e-01 -1.68070817e+00 1.19765759e+00 -2.13159382e-01
7.09252477e-01 -2.56960511e-01 -1.17505991e+00 9.97961462e-01
-2.19942331e-01 2.98628420e-01 3.14549804e-01 -1.66246016e-02
-3.76822114e-01 -9.27079767e-02 -1.24912679e+00 6.71904206e-01
1.20324957e+00 -2.80651003e-01 2.22544335e-02 4.87700105e-01
8.47478867e-01 -4.22836065e-01 -8.25030744e-01 3.52969378e-01
3.36415350e-01 -9.67655301e-01 8.01870346e-01 -7.70472467e-01
5.25198519e-01 1.25392988e-01 1.79320917e-01 -1.42252195e+00
-3.50714952e-01 -5.23038983e-01 -4.27233696e-01 1.18228316e+00
9.45506871e-01 -1.06987214e+00 1.28839731e+00 7.18458951e-01
-2.63584461e-02 -7.30514586e-01 -6.29210055e-01 -5.16335785e-01
5.40327393e-02 -3.43080938e-01 7.00791955e-01 1.44481421e+00
-3.33637178e-01 1.52080223e-01 -7.15224981e-01 2.93778360e-01
7.64692962e-01 -3.40847999e-01 6.10010326e-01 -1.65444279e+00
-1.87593758e-01 -4.18179095e-01 -3.49812239e-01 -8.50459635e-01
1.94056869e-01 -1.15674138e+00 -5.40429354e-01 -1.44976997e+00
-1.78895801e-01 -7.25346446e-01 -2.64865249e-01 8.55603218e-01
-1.58626571e-01 5.29596984e-01 6.50511906e-02 2.64086396e-01
-5.05040765e-01 3.56876075e-01 1.19038045e+00 -2.73096293e-01
-2.17871338e-01 2.87264615e-01 -8.02277744e-01 8.13425779e-01
7.70753920e-01 -8.52329016e-01 -3.78831476e-01 -5.00978708e-01
7.34905541e-01 -1.64694831e-01 6.24298692e-01 -9.18824732e-01
3.50151837e-01 2.76663806e-02 4.84237313e-01 -2.25889221e-01
-8.93608257e-02 -6.93290114e-01 3.64463955e-01 6.21583402e-01
-2.70502090e-01 1.64133996e-01 -7.08947852e-02 8.94368887e-01
5.83088212e-02 -1.71392590e-01 5.96699357e-01 -8.10540244e-02
-5.38906217e-01 4.73533839e-01 -1.79421946e-01 -2.85301618e-02
7.73611963e-01 -4.13393885e-01 -7.02791750e-01 -5.35306811e-01
-7.97164857e-01 1.97177678e-01 3.16469043e-01 4.81252998e-01
5.42310596e-01 -1.16038322e+00 -6.94528162e-01 4.63287681e-01
-1.09527625e-01 -6.48925826e-02 2.06446722e-01 8.58786702e-01
-6.53480351e-01 -3.92191447e-02 7.07496777e-02 -2.73736566e-01
-9.55323160e-01 1.50441229e-01 5.78765213e-01 -5.32284558e-01
-4.80749130e-01 1.05522561e+00 -2.89009959e-01 -9.66218889e-01
1.26272559e-01 -1.86089873e-01 -2.07009926e-01 -2.57975399e-01
7.53790587e-02 5.66221774e-01 2.24283099e-01 -4.41641182e-01
-2.17647776e-01 -7.08218515e-02 -5.34749568e-01 7.32356369e-01
1.44638824e+00 -7.34929442e-02 -3.83611023e-01 2.80493163e-02
8.35099995e-01 2.19827052e-02 -9.59794462e-01 -3.95326108e-01
4.13555168e-02 -1.92412555e-01 7.31284544e-02 -1.01092303e+00
-1.50854492e+00 4.81563300e-01 1.02169730e-01 6.40057504e-01
7.43671775e-01 -2.93101996e-01 7.83838511e-01 4.00558114e-01
2.44840875e-01 -9.52363610e-01 -6.83494583e-02 5.11054039e-01
7.11249590e-01 -1.17281222e+00 2.49302953e-01 -7.60306776e-01
-5.24912834e-01 9.51473892e-01 7.63661444e-01 -3.02342385e-01
8.86344790e-01 -1.05121106e-01 -2.05223903e-01 -3.71572614e-01
-4.09957081e-01 6.93397641e-01 2.06087977e-02 7.09853709e-01
1.25775412e-01 1.54472128e-01 -4.32905108e-02 7.63068557e-01
-3.27023566e-01 -7.42887706e-02 7.55538523e-01 4.30607438e-01
-2.19488457e-01 -1.00406957e+00 1.89286042e-02 1.01729143e+00
-3.74768496e-01 -4.31383193e-01 -6.99769020e-01 7.86663353e-01
-1.01082623e-01 7.63160348e-01 -2.50450909e-01 -6.82933152e-01
3.34036708e-01 -8.89154598e-02 -1.37164667e-01 -3.96794617e-01
-9.08160508e-01 -4.69014943e-01 2.67756671e-01 -5.02708912e-01
-9.60093439e-02 -1.70797393e-01 -1.08229017e+00 -9.86010194e-01
-5.71466386e-01 8.66331831e-02 2.65918434e-01 8.30815792e-01
5.85883141e-01 7.32175291e-01 3.80243391e-01 -5.48583269e-01
-5.17047107e-01 -1.00519335e+00 -8.81330609e-01 1.91407263e-01
1.19015351e-01 -4.50958401e-01 -5.00822723e-01 -2.93458223e-01] | [6.998199939727783, 6.111227512359619] |
96266538-9610-4481-853f-b46502afb9c0 | dialogue-act-classification-in-team | null | null | https://aclanthology.org/W19-5946 | https://aclanthology.org/W19-5946.pdf | Dialogue Act Classification in Team Communication for Robot Assisted Disaster Response | We present the results we obtained on the classification of dialogue acts in a corpus of human-human team communication in the domain of robot-assisted disaster response. We annotated dialogue acts according to the ISO 24617-2 standard scheme and carried out experiments using the FastText linear classifier as well as several neural architectures, including feed-forward, recurrent and convolutional neural models with different types of embeddings, context and attention mechanism. The best performance was achieved with a {''}Divide {\&} Merge{''} architecture presented in the paper, using trainable GloVe embeddings and a structured dialogue history. This model learns from the current utterance and the preceding context separately and then combines the two generated representations. Average accuracy of 10-fold cross-validation is 79.8{\%}, F-score 71.8{\%}. | ['Ivana Kruijff-Korbayova', 'Tatiana Anikina'] | 2019-09-01 | null | null | null | ws-2019-9 | ['dialogue-act-classification'] | ['natural-language-processing'] | [ 8.02538693e-02 8.97652507e-01 5.36483049e-01 -4.53670353e-01
-4.22272712e-01 -1.00178085e-01 9.35613215e-01 3.88219714e-01
-1.01674759e+00 9.75005209e-01 1.04718208e+00 -2.89713234e-01
-3.35140228e-02 -6.13248885e-01 1.23905733e-01 -5.47376513e-01
-4.90970105e-01 8.62862647e-01 -5.58000132e-02 -8.43236148e-01
4.12012488e-01 1.54230356e-01 -7.52887785e-01 4.55526710e-01
2.28778288e-01 8.14778209e-01 1.02544077e-01 1.29892993e+00
-1.60176486e-01 1.49014103e+00 -1.10319412e+00 -3.12896430e-01
9.50760245e-02 -4.10472751e-01 -1.77631676e+00 -1.77175969e-01
-5.25149941e-01 -5.86688638e-01 -7.08867908e-01 2.71717042e-01
8.43133926e-01 4.89880592e-01 9.41915810e-01 -8.09464276e-01
-6.87739789e-01 6.52564049e-01 1.12489857e-01 2.36503333e-01
8.26715708e-01 1.25529850e-02 7.17692673e-01 -8.46880078e-01
5.39887607e-01 1.25413239e+00 5.19979417e-01 9.01320636e-01
-9.21334147e-01 -2.28782848e-01 -3.41431379e-01 4.65307906e-02
-8.35945904e-01 -2.17692167e-01 3.64422560e-01 -4.57504839e-01
1.96161413e+00 -1.31646156e-01 9.71086770e-02 1.22446787e+00
2.01663047e-01 7.31881797e-01 8.30828011e-01 -7.65238702e-01
-2.11992823e-02 2.13051111e-01 5.96473455e-01 5.69823325e-01
-5.00593483e-01 -1.43854454e-01 -1.11623421e-01 -4.81932670e-01
4.83236313e-01 -1.46058172e-01 -3.64394337e-01 4.84955728e-01
-1.13675070e+00 1.13778019e+00 4.58297431e-01 6.31422341e-01
-5.33774316e-01 -1.61557436e-01 9.76447105e-01 6.30984128e-01
4.61747020e-01 6.51661456e-01 -6.87759876e-01 -5.28687119e-01
8.89213011e-02 1.53688461e-01 1.23596752e+00 6.87630653e-01
4.41546321e-01 1.24497704e-01 -2.30696663e-01 1.14845645e+00
1.00002833e-01 -2.36926883e-01 8.83918166e-01 -8.86785448e-01
7.22101271e-01 6.60219252e-01 4.12315398e-01 -8.38682652e-01
-8.86232257e-01 3.08769882e-01 -9.09459651e-01 1.37135059e-01
3.78383428e-01 -9.07408357e-01 -5.91157258e-01 1.28444910e+00
-4.55266982e-02 -3.37086171e-01 8.92541051e-01 6.14214838e-01
1.20066071e+00 9.65463996e-01 2.96697736e-01 -5.23257405e-02
1.32122350e+00 -1.07482517e+00 -8.85733902e-01 -9.65769403e-03
1.01299310e+00 -6.15732074e-01 7.51595736e-01 3.12178195e-01
-8.79518270e-01 -6.20504200e-01 -9.83849585e-01 -2.21920028e-01
-4.99192268e-01 2.32504651e-01 1.87717035e-01 3.14861715e-01
-1.18475008e+00 5.70226252e-01 -4.74577785e-01 -5.38856387e-01
-2.79927224e-01 6.16902053e-01 -7.87761807e-01 2.09030107e-01
-1.44676781e+00 1.51125062e+00 7.08204985e-01 1.92996338e-01
-7.76426256e-01 1.57466426e-01 -9.87865806e-01 -5.17830253e-02
-1.40728861e-01 -6.00807704e-02 1.50256467e+00 -5.67361712e-01
-1.92016673e+00 7.36267149e-01 2.72739798e-01 -8.75691473e-01
2.59234846e-01 -5.52600205e-01 -1.45788893e-01 1.99901953e-01
-2.15670347e-01 5.81562400e-01 7.00738579e-02 -9.42357063e-01
-5.39844692e-01 -1.86435699e-01 3.92577320e-01 4.77066487e-01
-2.21511930e-01 3.16818267e-01 3.42947185e-01 -3.30793381e-01
-2.83092856e-01 -8.55253816e-01 -3.60542715e-01 -7.13611305e-01
-1.07186750e-01 -7.23178864e-01 7.77060449e-01 -9.70867515e-01
1.12390137e+00 -1.94686925e+00 2.94297844e-01 -3.23876560e-01
9.29892212e-02 4.72299993e-01 -1.17768981e-02 1.18914545e+00
-1.52063310e-01 -1.07619233e-01 -2.61193871e-01 -4.72027361e-01
5.83018921e-02 5.75709760e-01 -5.27825430e-02 2.80288398e-01
2.97851592e-01 4.15319622e-01 -9.41904485e-01 -1.92871004e-01
4.39950228e-01 3.72142822e-01 -1.47121832e-01 1.04091918e+00
1.63137466e-02 3.17251623e-01 -3.04681152e-01 -8.00946876e-02
9.30023007e-03 2.88187623e-01 3.32372636e-01 2.80667305e-01
4.38688993e-02 5.06305695e-01 -6.86527669e-01 1.56156707e+00
-7.81838834e-01 8.94359767e-01 1.84998840e-01 -1.15736806e+00
1.37614465e+00 1.08172274e+00 2.89179176e-01 -3.74206454e-01
5.79715312e-01 -3.16989385e-02 1.59179717e-02 -9.40173030e-01
8.43020022e-01 -6.15961701e-02 -4.88296509e-01 7.12481618e-01
5.70270300e-01 -1.16681363e-02 -5.82064241e-02 1.83637857e-01
1.39180887e+00 -1.01306140e-01 5.18344998e-01 -1.97910927e-02
6.63087189e-01 2.47756988e-02 1.05868299e-02 4.12332952e-01
-3.42239946e-01 5.42144418e-01 7.20649421e-01 -9.12487864e-01
-1.05172515e+00 -4.74399954e-01 2.25786015e-01 1.46438634e+00
-4.43236738e-01 -2.48562947e-01 -7.09254742e-01 -7.63698161e-01
-6.25004411e-01 9.73587275e-01 -9.24216926e-01 -3.70956697e-02
-7.59393394e-01 -6.43008530e-01 9.96831894e-01 6.12690210e-01
7.10850358e-01 -1.75558698e+00 -8.70830297e-01 5.84970534e-01
-3.09341788e-01 -1.10644579e+00 3.29953842e-02 5.49440444e-01
-5.09990871e-01 -1.15073466e+00 -5.46986639e-01 -1.02632642e+00
3.74088794e-01 -2.92050213e-01 9.81161833e-01 9.76274088e-02
-1.65866762e-01 4.93256241e-01 -1.06261206e+00 -3.22293729e-01
-6.60775483e-01 8.02579336e-03 -1.58504285e-02 -2.89304078e-01
4.84134465e-01 -1.94420308e-01 -1.71150133e-01 1.83224797e-01
-6.56301558e-01 -2.55408406e-01 5.05534351e-01 1.35538411e+00
-5.21243513e-01 -2.18115568e-01 6.50143027e-01 -7.37808049e-01
1.39824522e+00 -6.30057514e-01 1.83434844e-01 5.96166626e-02
-1.81958616e-01 1.71045110e-01 7.08189487e-01 -1.95544325e-02
-1.19657433e+00 -6.79735467e-02 -5.34920454e-01 2.56508321e-01
-6.73500299e-01 3.26537579e-01 1.72946781e-01 5.55259645e-01
1.11003280e+00 1.26345590e-01 1.40850440e-01 -4.37419742e-01
2.40098551e-01 1.51605010e+00 6.37836933e-01 -5.20449817e-01
-2.18115970e-02 -1.04096055e-01 -6.42424107e-01 -1.10666525e+00
-2.61081159e-01 -6.17787778e-01 -8.73810172e-01 -1.57556415e-01
1.35524416e+00 -5.54908156e-01 -8.34929347e-01 4.27531064e-01
-1.82469380e+00 -7.91041553e-01 -1.65319256e-02 6.73668981e-01
-6.65892661e-01 5.25271416e-01 -9.38052952e-01 -1.21202099e+00
-8.49738240e-01 -9.26003098e-01 6.86007917e-01 1.81684196e-01
-6.68629766e-01 -1.12930131e+00 2.91167945e-01 2.49495760e-01
3.33057284e-01 4.35589820e-01 8.50108445e-01 -1.59149420e+00
5.43990552e-01 -4.33509290e-01 -1.51120901e-01 7.07310259e-01
7.67873675e-02 -3.92783761e-01 -1.03440607e+00 -2.36687660e-02
2.65915692e-02 -8.81041884e-01 3.94542098e-01 -2.04263628e-01
3.99448782e-01 -6.81685746e-01 1.28999084e-01 -3.77106160e-01
1.00495970e+00 5.85705340e-01 8.04442644e-01 4.55912650e-01
1.77436620e-01 1.02566361e+00 4.94544089e-01 7.76637077e-01
4.70974505e-01 4.36905831e-01 3.38706285e-01 3.12179744e-01
3.78056705e-01 1.17508948e-01 4.39020157e-01 6.37695372e-01
-2.62726635e-01 -4.96336907e-01 -1.35343242e+00 5.32471299e-01
-1.89351428e+00 -9.87921119e-01 -9.71321836e-02 1.86233747e+00
8.97894084e-01 7.92936534e-02 1.38852030e-01 3.33507240e-01
4.51022595e-01 6.84342086e-02 3.98607105e-01 -1.31108618e+00
3.01292032e-01 4.15226042e-01 1.80891514e-01 8.42414618e-01
-1.21406209e+00 8.64123404e-01 6.00105572e+00 2.05773905e-01
-7.47766197e-01 1.77022025e-01 3.72038573e-01 4.84530598e-01
4.98868376e-01 -4.07482535e-01 -4.38873470e-01 1.70608625e-01
1.56723607e+00 3.46856415e-02 1.81974009e-01 9.00772989e-01
-3.26307230e-02 -2.02779740e-01 -9.23070848e-01 5.74793339e-01
2.71157295e-01 -1.21799672e+00 -4.17950720e-01 -2.05506027e-01
2.47368410e-01 1.09039672e-01 -5.71704865e-01 8.55676353e-01
8.31597745e-01 -1.30307853e+00 1.18074007e-01 3.33645433e-01
5.45693099e-01 -5.60012221e-01 1.53650570e+00 6.48628294e-01
-8.14100146e-01 -4.04028654e-01 -3.60253960e-01 -4.09361988e-01
1.15090080e-01 -3.39835912e-01 -1.56555092e+00 8.29510272e-01
6.21082962e-01 3.12099934e-01 -9.43202004e-02 4.43512321e-01
-3.44353944e-01 3.18129659e-01 -1.47183463e-01 -4.25573409e-01
6.92893207e-01 4.80645560e-02 2.14386731e-01 1.79843974e+00
-3.21851164e-01 6.78060710e-01 2.73604810e-01 -1.21324129e-01
1.12560138e-01 1.60807416e-01 -7.92270362e-01 1.05983451e-01
3.28058928e-01 1.03072762e+00 -3.69127929e-01 -4.10164714e-01
-1.89867452e-01 9.48922575e-01 5.17086208e-01 -5.04564820e-03
-4.40115869e-01 -9.92518604e-01 2.07845062e-01 -3.97614807e-01
-7.34600201e-02 -4.56780612e-01 9.40146223e-02 -5.86569607e-01
-3.87516826e-01 -6.75879836e-01 4.76495117e-01 -6.30335987e-01
-1.13040268e+00 1.45463216e+00 3.52725744e-01 -9.47517872e-01
-7.64438331e-01 -9.93388772e-01 -7.78079510e-01 8.81146371e-01
-8.07246745e-01 -8.86324167e-01 -3.41652364e-01 3.81578088e-01
9.98330355e-01 -6.55214787e-01 1.76228428e+00 2.05764100e-02
-3.98436427e-01 3.09972763e-01 -1.05590850e-01 8.80596519e-01
5.11232138e-01 -1.45917547e+00 1.51578501e-01 6.40644804e-02
-3.62627625e-01 4.16408747e-01 7.04618573e-01 -2.17365593e-01
-7.45996833e-01 -7.67527521e-01 1.23274076e+00 -4.02383268e-01
6.11684263e-01 -2.24858493e-01 -9.14701164e-01 7.30988443e-01
1.04241800e+00 -4.58557755e-01 8.94136667e-01 -6.40735552e-02
3.16020437e-02 4.75827247e-01 -1.28563559e+00 3.48751038e-01
3.74796420e-01 -4.90797460e-01 -1.32586837e+00 4.86511588e-01
5.75318813e-01 -5.61405122e-01 -1.11562121e+00 1.70975551e-01
3.03682059e-01 -8.81401837e-01 5.79293549e-01 -8.55034471e-01
4.87089068e-01 3.12034070e-01 -2.51662910e-01 -1.29181659e+00
-5.78321218e-02 -7.35613227e-01 2.46176258e-01 9.45867300e-01
4.52137798e-01 -5.81415355e-01 3.97421598e-01 8.03091049e-01
-5.66108227e-01 -6.91287577e-01 -9.01183367e-01 -1.99613780e-01
2.66660601e-01 -1.20618515e-01 1.48840457e-01 7.63892353e-01
6.30754471e-01 7.61914551e-01 -6.41590297e-01 -2.52423555e-01
-2.69355893e-01 -7.56020367e-01 9.40081298e-01 -9.64750528e-01
-1.35576338e-01 -1.68864250e-01 -5.27124822e-01 -9.51998293e-01
4.09511089e-01 -2.92665303e-01 2.89869040e-01 -1.80743766e+00
-4.72343057e-01 -1.73922718e-01 1.38505539e-02 7.25046873e-01
1.08079918e-01 -1.34570569e-01 -1.34779536e-03 1.18725188e-02
-3.43653917e-01 6.98180914e-01 3.89396280e-01 -1.20989554e-01
-1.33070484e-01 -2.70702858e-02 -2.06373364e-01 5.66896915e-01
1.09809244e+00 -3.04782987e-01 -2.56022900e-01 -5.09929299e-01
-3.56745303e-01 4.65100318e-01 -1.96394362e-02 -8.71005893e-01
1.67941481e-01 1.74891382e-01 8.42506438e-02 -1.94606721e-01
6.71713054e-01 -6.04743719e-01 -3.44170094e-01 5.95584691e-01
-7.89026082e-01 4.07665014e-01 1.45536944e-01 5.37375808e-01
-4.20893103e-01 -6.61040306e-01 5.95777571e-01 -3.61307412e-01
-6.70765460e-01 -3.85727823e-01 -8.15717757e-01 -8.62031356e-02
1.06992650e+00 3.23681086e-02 -2.74561375e-01 -6.58068180e-01
-1.02804077e+00 3.56376469e-01 -2.03693762e-01 4.90597844e-01
6.38813019e-01 -8.44233334e-01 -8.61601532e-01 -1.33239433e-01
-6.36452064e-02 -5.59667908e-02 1.97916720e-02 5.37585199e-01
-9.14417088e-01 4.80108440e-01 -4.85238403e-01 -4.19259667e-02
-1.21666777e+00 1.01449467e-01 5.15793443e-01 -6.21204972e-01
-5.57383597e-01 8.88626873e-01 -4.95033979e-01 -8.30454767e-01
5.68507135e-01 -2.39905156e-02 -9.77638543e-01 1.08437963e-01
5.25914967e-01 5.84583879e-01 -1.76630076e-02 -9.11493361e-01
-1.92050651e-01 1.95214190e-02 -6.21699616e-02 -4.21076387e-01
1.42509890e+00 1.62698850e-01 -1.05718404e-01 6.21372998e-01
1.35258472e+00 -6.38022363e-01 -7.63397753e-01 -2.03292340e-01
2.64563113e-01 -2.78008860e-02 -4.75122958e-01 -7.79370427e-01
-3.67067069e-01 1.04861557e+00 5.01962543e-01 7.16947973e-01
5.82668841e-01 -3.86930048e-01 5.97596109e-01 9.24631655e-01
2.84123480e-01 -1.28823256e+00 3.56279582e-01 1.20450187e+00
1.25316441e+00 -1.24069655e+00 -2.70049702e-02 3.03773075e-01
-1.20864320e+00 1.62138116e+00 6.14694834e-01 -5.06678581e-01
6.55120909e-01 1.73560325e-02 3.08583111e-01 -2.31965393e-01
-9.80684757e-01 1.30810337e-02 -3.16687882e-01 7.50362575e-01
8.53651643e-01 -3.11189946e-02 -5.09406328e-01 5.62103748e-01
-2.21605003e-01 -1.62310600e-01 7.86475182e-01 1.17755759e+00
-5.76669991e-01 -1.10325003e+00 -2.05277085e-01 1.53890714e-01
-5.16301215e-01 2.52543807e-01 -7.89482892e-01 1.02088571e+00
-1.30923659e-01 1.38642776e+00 4.14150283e-02 -7.81717181e-01
6.18808448e-01 4.73569125e-01 -8.91563520e-02 -1.11487913e+00
-1.32817042e+00 -4.21383858e-01 7.82103956e-01 -6.14124686e-02
-3.65678966e-01 -2.54406571e-01 -1.65226507e+00 -2.00656280e-01
-1.10775083e-01 6.59790099e-01 5.15494645e-01 1.03368247e+00
9.67817232e-02 3.62823933e-01 8.18605304e-01 -1.17435920e+00
-7.82908380e-01 -1.98870409e+00 -1.89354479e-01 4.09905046e-01
1.09451838e-01 -4.61200118e-01 -3.53029490e-01 -5.95001169e-02] | [12.812413215637207, 7.781406879425049] |
eccf4154-aa00-470b-a61e-d7f9b5d38781 | toward-creating-subsurface-camera | 1810.12271 | null | http://arxiv.org/abs/1810.12271v1 | http://arxiv.org/pdf/1810.12271v1.pdf | Toward Creating Subsurface Camera | In this article, the framework and architecture of Subsurface Camera (SAMERA)
is envisioned and described for the first time. A SAMERA is a geophysical
sensor network that senses and processes geophysical sensor signals, and
computes a 3D subsurface image in-situ in real-time. The basic mechanism is:
geophysical waves propagating/reflected/refracted through subsurface enter a
network of geophysical sensors, where a 2D or 3D image is computed and
recorded; a control software may be connected to this network to allow view of
the 2D/3D image and adjustment of settings such as resolution, filter,
regularization and other algorithm parameters. System prototypes based on
seismic imaging have been designed. SAMERA technology is envisioned as a game
changer to transform many subsurface survey and monitoring applications,
including oil/gas exploration and production, subsurface infrastructures and
homeland security, wastewater and CO2 sequestration, earthquake and volcano
hazard monitoring. The system prototypes for seismic imaging have been built.
Creating SAMERA requires an interdisciplinary collaboration and transformation
of sensor networks, signal processing, distributed computing, and geophysical
imaging. | [] | 2018-10-29 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [ 4.25008565e-01 -1.20180883e-01 9.92896557e-01 -2.30788499e-01
-6.88018575e-02 -4.66895610e-01 3.67707163e-01 -2.41453841e-01
-3.02327991e-01 5.14140837e-02 3.01366895e-01 -2.21243545e-01
-1.51571348e-01 -1.34061110e+00 -3.17737281e-01 -9.85388339e-01
-6.23784363e-01 1.82714731e-01 7.25299478e-01 -4.33784723e-01
6.05849326e-01 9.02417600e-01 -1.49300909e+00 1.74925894e-01
1.24280475e-01 1.31788945e+00 5.56007981e-01 8.32526445e-01
-3.29128914e-02 2.03039154e-01 -2.19727904e-01 7.82412708e-01
2.77133495e-01 -8.03604200e-02 -4.61089134e-01 2.70452648e-02
-2.73905009e-01 -5.24963021e-01 -4.05534916e-02 1.15760338e+00
6.19199395e-01 -6.30000383e-02 5.19499838e-01 -9.20603752e-01
-2.86102504e-01 2.76928425e-01 -8.77854288e-01 -4.80354279e-02
4.86661464e-01 2.35988736e-01 2.12924749e-01 -1.26987338e+00
5.10974705e-01 1.34219193e+00 7.83318281e-01 -3.43076512e-02
-7.33013570e-01 -3.20869625e-01 -1.62796468e-01 -1.11883923e-01
-1.10644460e+00 -3.87795001e-01 6.99244440e-01 -5.22853255e-01
9.30598319e-01 3.88644695e-01 8.77003670e-01 6.85402989e-01
4.47795331e-01 1.14878705e-02 9.53358829e-01 -5.62324405e-01
7.28773713e-01 -5.74151814e-01 1.00457422e-01 4.47637856e-01
4.93996739e-01 1.38218910e-03 -7.50207007e-01 -3.88971537e-01
1.18468726e+00 1.89352885e-01 -6.95232868e-01 9.64984819e-02
-1.06749630e+00 5.93339503e-01 3.69267655e-03 2.67660469e-01
-6.68368220e-01 2.51143903e-01 3.25506151e-01 5.69549680e-01
2.39894897e-01 3.07052046e-01 -2.82758743e-01 -2.61412915e-02
-6.43626153e-01 1.11340612e-01 7.92625666e-01 3.30241621e-01
8.36399615e-01 1.29868016e-01 6.89995944e-01 6.12951100e-01
9.90339041e-01 1.23002434e+00 5.49788773e-01 -1.10328710e+00
4.35174555e-02 3.95337403e-01 1.42627761e-01 -1.25635195e+00
-5.29358983e-01 6.63894951e-01 -8.79535079e-01 6.80548072e-01
-3.38424951e-01 -4.22079444e-01 -8.96517634e-01 6.42635703e-01
3.15045655e-01 3.24435264e-01 4.61939633e-01 1.20995235e+00
8.05258334e-01 1.15746748e+00 -3.93445402e-01 -1.06575496e-01
1.45099509e+00 -7.14016557e-02 -6.66605413e-01 -2.70899177e-01
2.20686108e-01 -4.90994483e-01 8.14808607e-01 6.77920580e-01
-8.08400631e-01 -9.94000360e-02 -8.95050526e-01 5.71817756e-01
-3.92677873e-01 -3.52985501e-01 5.49941540e-01 4.76026297e-01
-1.11703646e+00 3.67422938e-01 -1.34644508e+00 -5.55933177e-01
-2.24336043e-01 -5.72872795e-02 -4.25626993e-01 2.48682704e-02
-1.15822852e+00 5.01555085e-01 -4.50743251e-02 8.19353640e-01
-1.07813001e+00 -5.22929490e-01 -7.19244659e-01 4.95449342e-02
-1.21248201e-01 -3.76493365e-01 8.83623838e-01 -2.93428212e-01
-1.92291784e+00 6.57117605e-01 2.95581460e-01 -6.89407140e-02
1.62896976e-01 -2.21905634e-01 -5.50534487e-01 3.91318232e-01
1.02123956e-03 -8.38012025e-02 2.57490456e-01 -1.17750847e+00
-4.44760561e-01 -6.60886347e-01 -5.85408628e-01 1.16678059e-01
-1.28361076e-01 2.67575979e-01 8.69832486e-02 -1.31536826e-01
1.01801360e+00 -5.01325369e-01 -5.30369520e-01 8.45826864e-02
-2.47831922e-02 5.42251527e-01 1.37743258e+00 -4.32965219e-01
5.20240426e-01 -2.19091797e+00 -2.08465606e-01 6.79020405e-01
-1.93773612e-01 -5.21933511e-02 1.88580006e-01 9.85800922e-01
2.04063490e-01 -3.24695222e-02 -5.13583958e-01 1.06754094e-01
-3.37225527e-01 6.84340447e-02 -4.33340639e-01 6.79335117e-01
-2.93156564e-01 -5.59667423e-02 -5.82902670e-01 -1.81128108e-03
2.59869516e-01 3.24175984e-01 -2.49476612e-01 3.83514583e-01
8.17488506e-03 5.60678184e-01 -9.31891561e-01 8.23426247e-01
1.35998476e+00 1.03420898e-01 1.36934578e-01 -1.33116871e-01
-9.24916148e-01 -3.67941678e-01 -1.89053988e+00 1.33553410e+00
-3.45112205e-01 5.65770090e-01 1.04551733e+00 -9.72266793e-01
1.61461711e+00 6.12814784e-01 4.73880649e-01 -6.52245581e-01
-5.56341670e-02 4.94485974e-01 -8.17362249e-01 -1.61438215e+00
5.98120213e-01 7.98095092e-02 6.67363331e-02 3.77852112e-01
-6.48652375e-01 -5.93611479e-01 -4.43770617e-01 -7.73911253e-02
1.27813530e+00 5.07965535e-02 -1.91181630e-01 -4.57070261e-01
3.29671383e-01 1.74466848e-01 4.08496261e-01 6.44052267e-01
3.01480412e-01 7.58791268e-01 1.93297610e-01 -5.90523064e-01
-9.31790292e-01 -9.00279164e-01 -4.88721654e-02 7.81529069e-01
5.30030787e-01 4.46559727e-01 -2.93812305e-01 5.39257884e-01
-2.29817387e-02 -2.71456270e-03 -2.07833096e-01 2.24306315e-01
-3.82113516e-01 -8.66911173e-01 5.46597660e-01 1.77689984e-01
7.49984741e-01 -1.11670589e+00 -1.43265200e+00 5.22598565e-01
1.71407402e-01 -8.17165434e-01 2.98730284e-01 3.63142490e-02
-1.12058198e+00 -8.09496582e-01 -3.95377666e-01 -5.81552744e-01
4.66431707e-01 5.78658700e-01 4.91410702e-01 3.21525261e-02
-2.45806456e-01 4.44872141e-01 -5.97085416e-01 -4.65283424e-01
-1.77260086e-01 -8.00214112e-01 -1.40269294e-01 3.26378107e-01
-1.62416175e-01 -8.26581419e-01 -8.04487348e-01 1.44884482e-01
-1.14100361e+00 -5.20143807e-02 9.47443768e-02 3.99301112e-01
1.81737512e-01 -1.50880367e-01 2.47796968e-01 -7.19861865e-01
6.13473833e-01 -7.24536479e-01 -9.68503654e-01 1.13473564e-01
-1.29835621e-01 -4.68308419e-01 1.65917292e-01 1.82895306e-02
-1.41987741e+00 1.26283497e-01 -7.32727945e-02 1.19902544e-01
-3.93498540e-01 1.15498102e+00 -1.14540912e-01 -1.36030659e-01
7.52255678e-01 1.36807889e-01 2.30437949e-01 -4.65526849e-01
-2.45025948e-01 1.26007140e+00 7.50948846e-01 -3.46753091e-01
2.26993859e-01 1.07787383e+00 7.73425922e-02 -1.33878434e+00
2.28199556e-01 -5.88349521e-01 -2.88149655e-01 -7.22492635e-01
7.50562727e-01 -1.02270591e+00 -5.52548170e-01 1.21238053e+00
-1.33570445e+00 -2.41372705e-01 2.78848857e-01 7.48306096e-01
-1.59980766e-02 2.41051137e-01 -5.72583497e-01 -1.05911148e+00
-4.70998764e-01 -1.06165218e+00 1.10988140e+00 4.26169902e-01
2.15988293e-01 -7.91162550e-01 4.40401196e-01 -2.23976269e-01
8.40997756e-01 6.81875348e-01 1.08056381e-01 5.87839931e-02
-6.96390510e-01 -2.46657565e-01 -8.41140375e-02 -6.49849884e-03
2.37519041e-01 4.55850422e-01 -1.08720589e+00 -1.82366863e-01
3.18705469e-01 -1.38006777e-01 8.32665622e-01 6.18088901e-01
7.79005229e-01 4.51170132e-02 -4.71894562e-01 8.32622290e-01
1.73183668e+00 5.92497349e-01 8.40642869e-01 3.70415628e-01
1.69814125e-01 6.03025198e-01 4.04654652e-01 8.32963824e-01
1.03338473e-01 8.90638307e-03 8.15352798e-01 -6.30924925e-02
5.26195168e-01 3.26659799e-01 5.46152532e-01 6.64546072e-01
-4.69231665e-01 -9.98212323e-02 -1.36496806e+00 5.43773234e-01
-1.65373361e+00 -9.38962519e-01 -6.69620693e-01 2.02710271e+00
1.03302307e-01 -1.99342459e-01 -6.69161081e-01 -1.56379435e-02
7.66306698e-01 1.06478140e-01 -2.06010982e-01 -3.18770885e-01
-6.95801750e-02 8.66545066e-02 9.12120104e-01 5.36465049e-01
-5.49500346e-01 5.27598917e-01 6.29224777e+00 -2.14358345e-01
-1.82472265e+00 4.82378118e-02 1.40241962e-02 5.44106185e-01
-7.48186886e-01 1.83100596e-01 -3.42556357e-01 7.03164786e-02
7.85295010e-01 1.75092310e-01 2.47463374e-03 6.59523964e-01
1.07716930e+00 -6.58483803e-01 -3.65061373e-01 7.14308560e-01
-3.52748126e-01 -1.69640255e+00 -1.88995510e-01 -9.24792886e-02
4.31332707e-01 4.21794355e-01 -8.11068714e-01 -5.66538334e-01
5.08884668e-01 -5.30071855e-01 8.43678772e-01 8.61612022e-01
6.84249461e-01 -3.18071127e-01 6.95359588e-01 1.10186420e-01
-1.22559202e+00 -9.76482034e-02 -3.11478049e-01 -3.17596883e-01
6.19578242e-01 8.38124990e-01 -3.84887040e-01 4.34313536e-01
1.36886191e+00 3.37170392e-01 1.36638477e-01 1.02964938e+00
-6.50246767e-03 6.85404897e-01 -6.55203342e-01 -1.75051332e-01
3.07364911e-01 -5.24961650e-01 7.14174986e-01 1.13823247e+00
9.18236792e-01 7.76171565e-01 4.67414483e-02 6.45604312e-01
6.07550502e-01 -2.34306455e-01 -8.89098704e-01 3.94423872e-01
9.82464135e-01 1.25666475e+00 -7.92913496e-01 -2.14724794e-01
-2.64296561e-01 4.63477433e-01 -6.68903708e-01 3.94069135e-01
-4.74078774e-01 -7.74126828e-01 3.93255174e-01 3.08393866e-01
1.64205015e-01 -5.80280542e-01 -2.69661188e-01 -7.72122622e-01
-2.00595215e-01 -3.84873956e-01 1.72695935e-01 -1.39552605e+00
-9.05962884e-01 2.42063984e-01 1.37641862e-01 -1.07535601e+00
2.28939429e-01 -7.90189505e-01 -1.17326832e+00 7.29784369e-01
-1.35857284e+00 -7.95979798e-01 -9.83178020e-01 5.42316794e-01
-3.20530944e-02 -1.77282572e-01 7.83760011e-01 -1.15943067e-02
-3.47755522e-01 -6.09692633e-01 4.23006386e-01 -3.53681296e-03
3.72082531e-01 -6.29483461e-01 4.12590742e-01 1.00795650e+00
-7.11056769e-01 1.43968627e-01 9.04719830e-01 -7.83571959e-01
-2.04439712e+00 -7.84832180e-01 2.90325731e-01 5.89590967e-01
7.81131506e-01 -1.37719989e-01 -1.16352391e+00 5.67620039e-01
1.46172807e-01 3.32123429e-01 1.76526383e-01 -5.94666123e-01
-2.55417842e-02 -3.84796381e-01 -1.43232942e+00 2.68200845e-01
6.26981020e-01 -1.50241509e-01 -4.48649138e-01 1.52802825e-01
6.06171906e-01 -4.40037817e-01 -9.34039772e-01 1.89427495e-01
6.34378910e-01 -1.11497366e+00 5.71426511e-01 1.99777782e-01
2.71213293e-01 -5.14063895e-01 -4.34123456e-01 -1.12483501e+00
-3.24282274e-02 -6.56516790e-01 6.71103895e-01 7.71899581e-01
2.21802831e-01 -1.22658074e+00 6.53186083e-01 3.88955712e-01
-6.77129388e-01 1.02499515e-01 -1.01207125e+00 -4.37656105e-01
-4.00480837e-01 -6.29958212e-01 7.12165654e-01 8.60869467e-01
1.92927226e-01 -2.66326636e-01 5.68076037e-02 1.19719994e+00
7.43236423e-01 -1.19695961e-02 5.23922384e-01 -1.37301672e+00
-4.77083540e-03 3.75497192e-02 -5.21966517e-01 -9.11588907e-01
-7.72560716e-01 -1.09705530e-01 3.09440583e-01 -1.71879947e+00
-3.50321084e-01 -3.66997629e-01 2.59216785e-01 4.43962961e-01
7.29463100e-01 -7.73627311e-02 -2.21009955e-01 4.64100778e-01
2.64142543e-01 2.80757815e-01 1.04049194e+00 1.33919850e-01
-2.72705376e-01 -3.13607216e-01 -1.70327872e-02 8.44019413e-01
4.70257580e-01 -4.35414582e-01 7.17804879e-02 -9.57296312e-01
4.39666390e-01 7.67517149e-01 6.57967448e-01 -1.05042779e+00
8.35963249e-01 -5.94192147e-01 -5.11700995e-02 -6.17581785e-01
8.28311965e-02 -1.05097497e+00 8.00360620e-01 7.80040383e-01
2.18207195e-01 1.93802103e-01 -1.71451420e-01 4.81398225e-01
-5.38836002e-01 -2.93759435e-01 8.33490551e-01 -6.71694696e-01
-1.17730606e+00 2.55622482e-03 -9.94642019e-01 -5.05873919e-01
1.01656687e+00 -6.88580811e-01 -3.73809248e-01 -1.89762309e-01
-6.16537511e-01 2.70600915e-01 2.32280478e-01 -3.01424991e-02
1.15659308e+00 -8.33560884e-01 -7.84968317e-01 7.40654051e-01
-2.10373238e-01 5.40758014e-01 5.36282003e-01 4.86138254e-01
-1.48766041e+00 -2.99018949e-01 -3.80022883e-01 -7.86309421e-01
-9.14225459e-01 -5.42661309e-01 7.32287109e-01 2.79636145e-01
-8.35860252e-01 7.62379766e-01 -3.31429690e-01 -4.49441493e-01
-2.07339257e-01 -4.82003748e-01 -2.57794678e-01 1.99782811e-02
8.95236790e-01 6.57567561e-01 2.85437942e-01 -1.45315871e-01
-4.80728358e-01 9.27240849e-01 6.96858704e-01 -5.59021235e-01
2.07936096e+00 -3.98007423e-01 -6.03847742e-01 4.48838681e-01
8.95944774e-01 -1.08120933e-01 -1.30056226e+00 4.21488918e-02
5.68001643e-02 -3.79669279e-01 4.76201117e-01 -3.30047339e-01
-1.44261920e+00 5.77779353e-01 5.87449849e-01 5.22585571e-01
1.31154096e+00 -1.51202530e-01 5.01859665e-01 5.29847085e-01
5.94409108e-01 -1.12365925e+00 1.51192853e-02 7.92562544e-01
1.15050554e+00 -7.91238308e-01 -1.53509025e-02 -9.97629091e-02
-2.76776820e-01 1.67044961e+00 2.16107979e-01 -2.25179836e-01
1.17667937e+00 1.08324432e+00 3.79805684e-01 -8.16897452e-01
-5.47878146e-01 5.92599988e-01 -9.17220592e-01 6.81707263e-01
2.36499503e-01 5.84517531e-02 1.20196983e-01 2.93092072e-01
3.43409121e-01 2.22574979e-01 1.04027891e+00 1.23748398e+00
-1.17070460e+00 -4.12891626e-01 -1.26821387e+00 5.17985411e-02
-8.69253948e-02 8.20219442e-02 2.48691127e-01 4.48539048e-01
-1.16897441e-01 7.54443884e-01 4.66454476e-01 -6.89163804e-02
6.55524731e-01 -4.38394874e-01 -2.45673910e-01 -4.53492105e-01
-2.65662640e-01 1.52414069e-01 8.71120393e-02 -4.84377176e-01
-5.16390622e-01 -8.09097409e-01 -1.89286864e+00 -9.98665988e-02
-2.28666350e-01 2.67950386e-01 1.13061535e+00 6.43204629e-01
2.95276254e-01 8.40393379e-02 9.32922482e-01 -1.35763264e+00
-3.67872357e-01 -1.05553699e+00 -1.19185591e+00 -1.59341007e-01
3.71982813e-01 -2.96368808e-01 -6.83654785e-01 4.47480917e-01] | [6.918558120727539, 2.4451820850372314] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.